Feedback Regulation of ATP-induced Ca2+ Signaling in HL-60 Cells Is Mediated by Protein Kinase A- and C-mediated Changes in Capacitative Ca2+ Entry*

Extracellular ATP increases intracellular Ca2+ ([Ca2+] i ) in HL-60 cells. When cells are stimulated with supramaximal concentrations of ATP, although the initial [Ca2+] i increase is similar over a range of 30, 100, and 300 μm ATP, the rate of the return to basal [Ca2+] i level is faster in cells treated with higher concentrations of ATP. This probably results from differences in Ca2+ influx rather than Ca2+release, since the influx of the unidirectional Ca2+surrogates Ba2+ and Mn2+ also exhibit similar responses. Furthermore, while 300 μm ATP had an inhibitory effect on the thapsigargin-induced capacitative Ca2+ entry, 30 μm ATP potentiated the response. However, the inhibitory action of 300 μm ATP was blocked by protein kinase C (PKC) inhibitors, such as GF 109203X and chelerythrine, and the potentiating action of 30 μmATP was blocked by protein kinase A (PKA) inhibitors H89 and Rp-cAMPS. The PKC inhibitors also slowed the decay rate of the Ca2+response induced by 300 μm ATP, and the PKA inhibitors increased it when induced by 30 μm ATP. In the measurements of PKA and PKC activity, 30 μm ATP activates only PKA, while 300 μm ATP activates both kinases. Taken together, these data suggest that the changes in the ATP-induced Ca2+ response result from differential modulation of ATP-induced capacitative Ca2+ entry by PKC and PKA in HL-60 cells.

Extracellular ATP evokes many physiological effects such as platelet aggregation, neurotransmission, inflammation, and muscle contraction in numerous cell types (1). These various effects of ATP are mediated by plasma membrane P 2 purinergic receptors (2). Six subtypes of P 2 purinergic receptors, P 2X , P 2Y , P 2D , P 2T , P 2Z , and P 2U , were identified in pharmacological and functional studies and supported by cloning data (3). It has been reported that in HL-60 cells extracellular ATP increases the intracellular free Ca 2ϩ concentration ([Ca 2ϩ ] i ) 1 via plasma membrane P 2U and P 2X1 type receptors (4,5). We have also shown that extracellular ATP elevates cAMP through a novel type of receptor (6). The P 2U receptor is functionally coupled to phospholipase C (PLC) through pertussis toxin-sensitive and pertussis toxin-insensitive G proteins. PLC hydrolyzes phosphatidylinositol 4,5-bisphosphate to generate inositol 1,4,5trisphosphate (IP 3 ) and diacylglycerol. The IP 3 produced increases the [Ca 2ϩ ] i by mobilizing Ca 2ϩ from the intracellular Ca 2ϩ stores. This Ca 2ϩ mobilization activates the plasma membrane Ca 2ϩ influx pathway through Ca 2ϩ release-activated Ca 2ϩ channels (CRAC) and is termed capacitative Ca 2ϩ entry (7,8). The degree of Ca 2ϩ entry is determined by the filling status of the intracellular Ca 2ϩ store. The P 2X1 receptor triggers entry of cations; however, it has been shown that the activity is very weak in undifferentiated HL-60 cells. Thus, ATP increases intracellular Ca 2ϩ in HL-60 cells by mobilizing it from the intracellular stores and by influx from the extracellular space. We observed a different rate of decrease in the Ca 2ϩ response, while the peak level remained the same, when HL-60 cells were stimulated with supramaximal concentrations of ATP. There are several mechanisms responsible for Ca 2ϩ removal from the cytosol after the elevation of the [Ca 2ϩ ] i . These mechanisms include sequestering of Ca 2ϩ into intracellular stores, binding to various Ca 2ϩ -binding proteins, and actions by the Ca 2ϩ pump and Na ϩ /Ca 2ϩ exchanger (9). Among these, the Ca 2ϩ pump, which transports ions across the plasma membrane and into intracellular stores, plays a critical role in reducing the elevated [Ca 2ϩ ] i . The plasma membrane Na ϩ / Ca 2ϩ exchanger also plays an important role in the control of the intracellular free Ca 2ϩ concentration, exchanging three Na ϩ for one Ca 2ϩ . It appears to have a lower affinity for Ca 2ϩ than the plasma membrane Ca 2ϩ pump and a high capacity for removing increased Ca 2ϩ . Thus it operates efficiently when [Ca 2ϩ ] i is increased beyond 10 Ϫ8 M. A number of Ca 2ϩ -binding proteins are also involved in buffering the cytosolic Ca 2ϩ concentration. We studied the mechanism by which the different patterns of decrease in Ca 2ϩ occur upon stimulation with supramaximal concentrations of ATP in HL-60 cells. Our results suggest that this difference is not due to the cytosolic Ca 2ϩ removal system, but that it is instead mainly due to changes in capacitative Ca 2ϩ entry by actions of PKA and PKC, which are differentially activated by ATP itself.
Mn 2ϩ Quenching of Fura-2 Fluorescence-Cells loaded with fura-2/AM as described above were stimulated with ATP in the presence of 2 mM Mn 2ϩ , and fluorescence quenching was measured at an excitation wavelength of 360 nm, which is an isosbestic wavelength, and at an emission wavelength of 500 nm (13).
Measurement of Inositol 1,4,5-Trisphosphate-IP 3 mobilization was determined by competition assay with [ 3 H]IP 3 in binding to IP 3 -binding protein as described previously (14). To determine IP 3 production, 2 ϫ 10 6 cells per sample were harvested and stimulated with ATP. The reaction was terminated by the addition of ice-cold 15% (w/v) trichloroacetic acid containing 10 mM EGTA. After centrifugation at 2,000 ϫ g for 5 min, the supernatant was obtained. The Finally, 20 l of binding protein solution was added. The IP 3 -binding protein was prepared from bovine adrenal cortex according to the method of Challiss et al. (15). The mixture was incubated for 15 min on ice and then centrifuged at 2,000 ϫ g for 10 min. 100 l of water and 1 ml of scintillation mixture were added to the pellet to measure the radioactivity. The IP 3 concentration of the sample was determined by comparison with a standard curve and expressed as picomoles/mg of protein. The total cellular protein concentration was measured by the Bradford method after sonication of 2 ϫ 10 6 cells. Assay of PKA Activity-PKA activity was determined by measuring the incorporation of 32 P from [␥-32 P]ATP into the PKA-specific peptide, Leu-Arg-Arg-Ala-Ser-Leu-Gly (Kemptide), using a procedure described previously (17,18), with some modifications. Briefly, HL-60 cells (1 ϫ 10 7 cells/tube) were harvested and treated with inhibitor mixture containing 1 M GF 109203X and 1 M KN62, inhibitors of PKC and Ca 2ϩ /calmodulin-dependent protein kinase, respectively, for 5 min. They were then stimulated with 30 or 300 M ATP for 3 min. After the stimulation, the cells were washed twice with Locke's solution within another 2 min, then resuspended in 100 l of buffer I containing 20 mM Tris-HCl, pH 7.5, 0.25 M sucrose, 10 mM EGTA, 2 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 10 g/ml pepstatin A, 10 g/ml aprotinin, 200 M sodium pyrophosphate, 200 M sodium fluoride, 1 mM dithiothreitol. The cells were sonicated and centrifuged at 10,000 ϫ g for 10 min at 4°C. The supernatant was saved as the PKA fraction and used for in vitro PKA activity measurements. All of the following procedures were performed on ice unless stated otherwise. The reaction was initiated by the addition of 10 l of cell extract to the 30 l of a test mixture consisting of 10 l of Mg 2ϩ /ATP mixture containing 75 mM MgCl 2 , 500 M ATP, 50 Ci of [␥-32 P]ATP (3,000 Ci/ mmol), 10 l of 500 M Kemptide, and 10 l of inhibitor mixture containing 0.02 M GF 109203X, 0.9 M KN62. 10 M cAMP were added to the reaction mixture with Kemptide for a positive control, and 10 l of buffer II instead of Kemptide was added to determine the endogenous PKA substrate. All assay components were prepared by using buffer II that contained 20 mM MOPS, pH 7.2, 25 mM ␤-glycerol phosphate, 5 mM EGTA, 1 mM sodium orthovanadate, and 1 mM dithiothreitol. The reaction mixture was gently vortexed and placed in a 30°C water bath for 10 min. Then 25 l of the reaction mixture was transferred to 1 ϫ 3-cm P81 phosphocellulose strips, which were immediately immersed into 0.75% phosphoric acid. The strips were washed three times with 0.75% phosphoric acid and then dehydrated in 95% ethanol, air-dried, and placed into liquid scintillation vials. The radioactivity was quantified in a Beckman LS 8000 liquid scintillation counter.

Measurement of [ 3 H]cAMP-Intracellular
Assay of PKC Activity-PKC activity was measured by determining the incorporation of 32 P from [␥-32 P]ATP into histone IIIS as described previously (17,19), with some modifications. HL-60 cells were harvested and treated with inhibitor mixture containing 10 M H89 and 1 M KN62 and then stimulated with 30 or 300 M ATP and 100 nM PMA. After the stimulation, the cells were washed three times with Locke's solution and then resuspended in 200 l of buffer I, which is described in the PKA assay. The cells were sonicated and centrifuged at 100,000 ϫ g for 1 h at 4°C, and the pellet was saved as the membrane fraction and then was solubilized with the above buffer I containing 1% Nonidet P-40. The reaction was initiated by the addition of 10 l of solubilized membrane fraction to the 40 l of reaction mixture containing 10 l of 500 M histone IIIS, 10 l of inhibitor mixture containing 2 M PKI, PKA inhibitor peptide and 0.9 M KN62, 10 l of 500 nM PMA, and 10 l of the Mg 2ϩ /ATP mixture containing 75 mM MgCl 2 , 500 M ATP, and 100 Ci of [␥-32 P]ATP. All assay components were prepared using buffer II described in the assay of PKA. The reaction mixture was incubated at 30°C for 10 min, and 25 l of the reaction mixture was transferred to the P81 phosphocellulose strips. The strips were immersed into the 0.75% phosphoric acid and washed three times for 10 min. After washing, they were rinsed in 95% ethanol, air-dried, and quantified by measuring the radioactivity in a liquid scintillation counter.
Analysis of Data-Data are summarized as the means Ϯ S.E. EC 50 was calculated with the AllFit program (20). We considered differences significant at p Ͻ 0.05.

Effect of Extracellular ATP on Cytosolic [Ca 2ϩ ] i in HL-60
Cells-In HL-60 cells, ATP increased the [Ca 2ϩ ] i in a concentration-dependent manner with maximal and half-maximal effective concentrations (EC 50 ) seen at approximately 10 M and 85 nM, respectively (Fig. 1A). Fig. 1B illustrates the typical changes in [Ca 2ϩ ] i observed in fura-2-loaded HL-60 cells stimulated with maximal concentrations of ATP. Initially, the [Ca 2ϩ ] i increased rapidly to a peak level and then completely returned to the basal Ca 2ϩ level, even if the stimulant remained present. Notably, the changes in cytosolic Ca 2ϩ exhibited a different desensitization pattern in response to supramaximal concentrations of ATP as compared with the lower concentrations. Although the peak levels were similar, the rate of return to the basal [Ca 2ϩ ] i level was faster in cells treated with the higher concentration of ATP. This phenomenon is clearly seen in Fig. 1C. We measured the time from peak response of the Ca 2ϩ signal to the 70% desensitized [Ca 2ϩ ] i level as indicated in the inset of Fig. 1C. The data show that the times for return to 30% of the peak level became less in stimulations with increasing concentration of ATP. In other words, the higher the ATP concentration, the faster the return rate. We further analyzed whether changes in the desensitization rate are elicited by Ca 2ϩ release from intracellular stores or Ca 2ϩ influx from the extracellular space, since ATP increases [Ca 2ϩ ] i via both pathways.
Effect of Extracellular ATP on Ca 2ϩ Release and IP 3 Gener-ation-To measure Ca 2ϩ release from the intracellular stores, cells were stimulated with ATP in the absence of extracellular Ca 2ϩ . Subsequently, 3 mM Ca 2ϩ was introduced to the medium to assess the activity of the Ca 2ϩ influx. As shown in Fig. 2A, when cells were stimulated with 30, 100, and 300 M ATP in Ca 2ϩ -free medium, there were no significant differences in Ca 2ϩ release, but large differences occurred in Ca 2ϩ influx between the different ATP concentrations. Ca 2ϩ influx stimulated by 300 M ATP was 62.5% less than when stimulated with 30 M ATP. The data indicate that the differences in the falling state of the Ca 2ϩ responses caused by the supramaximal concentration of ATP resulted from changes in Ca 2ϩ influx.
Since the amount of Ca 2ϩ release was small, it is possible that undetectable differences could exist between those stimulations. Thus, we measured the IP 3 production of cells stimulated with ATP. Fig. 2B shows the time course for IP 3 production when cells were stimulated with 30  Thus, from the result of IP 3 production, it seems likely that 300 M ATP stimulation can cause more Ca 2ϩ release than 30 M ATP or, at least, can trigger the release of a similar amount of Ca 2ϩ from the internal stores, while IP 3 produced by 30 M ATP is enough to maximally mobilize Ca 2ϩ . Therefore, we conclude that the 300 M ATP-induced Ca 2ϩ release is not less than the 30 M ATP-induced one and that the difference detected in the Ca 2ϩ decay rate is due to changes in the Ca 2ϩ influx from the extracellular space.
Effect of Extracellular ATP on Mn 2ϩ Quenching and Ba 2ϩ Uptake-To test whether differences in the falling state of the Ca 2ϩ response are due to modulation of the Na ϩ /Ca 2ϩ exchanger and Ca 2ϩ /ATPase activity, we measured Mn 2ϩ and Ba 2ϩ influx after the addition of ATP. Mn 2ϩ and Ba 2ϩ are good Ca 2ϩ surrogates, since they are not pumped out of the cell, so they can be considered as selective tracers for entry (21,22). Mn 2ϩ uptake was estimated by the quenching of the fura-2 fluorescence when excited at the 360-nm wavelength, which is an isosbestic wavelength and insensitive to variations in Ca 2ϩ concentration. Ba 2ϩ uptake was estimated by the increase in the fura-2 fluorescence ratio when excited at the 340-and 380-nm wavelength. Fig. 3A shows the fluorescence quenching by Mn 2ϩ influx when cells were stimulated with 30, 100, and 300 M ATP. As 2 mM Mn 2ϩ was applied to the medium, it The data indicate that the lower concentration of ATP activates the divalent cation influx. This result was also supported by the Ba 2ϩ uptake. To measure Ba 2ϩ influx, cells were stimulated with ATP in the absence of external Ca 2ϩ . When the Ba 2ϩ was added to the medium, it caused an increase in the fluorescence intensity reflecting Ba 2ϩ uptake. The influx of Ba 2ϩ elicited by ATP shows a concentration-dependent pattern with the higher concentrations of ATP triggering less Ba 2ϩ uptake. This is similar to the result of Ca 2ϩ influx as shown in Fig. 2A. These results suggest that ATP regulates the amount of Ca 2ϩ influx, but does not modulate the activity of Na ϩ /Ca 2ϩ exchanger and Ca 2ϩ /ATPase. To investigate the involvement of CRAC, we tested the effect of various metal ions on ATP-induced Ca 2ϩ signaling, because metal ions are known to block CRAC. Cells were treated with 30 M La 3ϩ , Cd 2ϩ , Co 2ϩ , or Ni 2ϩ for 1 min and then stimulated with ATP in Ca 2ϩ -containing medium. The difference between the 30 and 300 M ATP-induced Ca 2ϩ signals disappeared in the presence of 30 M La 3ϩ , whereas the other metal ions had little or negligible effects (data not shown). The data, thus, suggested that changes in CRAC activity could be the main cause for the rapid desensitization of the Ca 2ϩ response induced by higher concentrations of ATP.
Effect of Extracellular ATP on Thapsigargin-induced Capacitative Ca 2ϩ Entry-In the above experiments, we found that different Ca 2ϩ decay rates were caused by changes in Ca 2ϩ influx. ATP induces capacitative Ca 2ϩ entry through CRAC, which is stimulated by a Ca 2ϩ influx factor liberated from the depleted intracellular Ca 2ϩ stores by action of IP 3 . To study the regulation of capacitative Ca 2ϩ entry, we used thapsigargin, which depletes intracellular Ca 2ϩ stores by inhibiting the microsomal Ca 2ϩ /ATPase and induces Ca 2ϩ influx (23). The differences in Ca 2ϩ influx could also be demonstrated when we measured the effect of ATP pretreatment on thapsigargin-induced capacitative Ca 2ϩ entry. As shown in Fig. 4, cells were incubated with 100 nM thapsigargin in a Ca 2ϩ -free medium, which resulted in a transient [Ca 2ϩ ] i elevation. After reaching a peak, the [Ca 2ϩ ] i decreased slowly to the basal level, which reflects the emptying of the intracellular Ca 2ϩ stores. The subsequent addition of 3 mM Ca 2ϩ to the medium induced a marked and sustained Ca 2ϩ rise (dotted trace). Treatment with 300 M ATP for 1 min prior to the extracellular Ca 2ϩ application significantly diminished the thapsigargin-induced capacitative Ca 2ϩ entry by 25.7% as compared with the untreated control (dotted trace). 100 M ATP had also a slightly inhibitory effect. However, 30 M ATP substantially potentiated the thapsigargin-induced capacitative Ca 2ϩ entry to 152.8% over the control cells. These results indicate that ATP has a biphasic effect on the thapsigargin-induced capacitative Ca 2ϩ entry linked to its concentration. Therefore, we speculated that ATP itself might potentiate and inhibit capacitative Ca 2ϩ entry that it evokes, forming both a positive and a negative feedback loop. At 30 M, ATP potentiates Ca 2ϩ influx, which slows down the desensitization of the [Ca 2ϩ ] i . Whereas, at 300 M, ATP inhibits Ca 2ϩ influx, which speeds up the desensitization of the [Ca 2ϩ ] i .

Effect of PKC and PKA Inhibitors on the ATP Activities in [Ca 2ϩ ] i Rise and Thapsigargin-induced Capacitative Ca 2ϩ
Entry-ATP activates PLC and produces IP 3 and diacylglycerol, which subsequently activates PKC. We have also shown that extracellular ATP triggers elevation of cAMP in HL-60 cells (6). To assess the involvement of PKC and PKA in modulation of capacitative Ca 2ϩ entry, we used inhibitors specific for those kinases. GF 109203X and chelerythrine, selective PKC inhibitors, were used to characterize the inhibitory or stimulatory effect of ATP on the capacitative Ca 2ϩ entry. Fig. 5A shows what effect pretreatment with protein kinase inhibitors has on the Ca 2ϩ transient elicited by thapsigargin. 300 M ATP has a substantial inhibitory effect on thapsigargin-induced capacitative Ca 2ϩ entry as seen in Fig. 4. This inhibitory action was antagonized by pretreatment with 1 M GF 109203X, and Ca 2ϩ influx was even potentiated in the presence of GF 109203X. Similar effects were obtained when 1 M chelerythrine was used in place of GF 109203X. The results suggest that PKC, when activated by 300 M ATP, inhibits thapsigargin-induced capacitative Ca 2ϩ entry.
The involvement of PKA in the 30 M ATP-induced potentiation of capacitative Ca 2ϩ entry was investigated by testing the effect of the PKA inhibitors H89 and Rp-cAMPS. Fig. 5B shows the effect that H89 and Rp-cAMPS has on the enhancement of the thapsigargin-induced capacitative Ca 2ϩ entry by 30 M The effects of protein kinase inhibitors on ATP activity in the desensitization pattern of [Ca 2ϩ ] i were also investigated. Inhibition of PKA by pretreatment with 2 M H89 significantly accelerated the decay rate of the 30 M ATP-induced [Ca 2ϩ ] i level with little effect on the peak Ca 2ϩ level (Fig. 6A). In contrast, pretreatment with 1 M GF 109203X slowed the decay rate induced by 300 M ATP, resulting in a Ca 2ϩ response similar to the 30 M ATP-evoked response (Fig. 6B). The inhibitory action of 300 M ATP was slightly enhanced in the presence of PKA inhibitors, while the potentiating effect of 30 M ATP became even more activated in the presence of PKC inhibitors (data not shown). Thus, the slower decay rate of the 30 M ATP-induced Ca 2ϩ signal may be the result of the potentiating action of PKA as it increases the capacitative Ca 2ϩ entry induced by ATP, whereas the rapid decay rate in the 300 M ATP-induced Ca 2ϩ signal might be the result of an inhibitory action by PKC as it blocks the ATP-induced capacitative Ca 2ϩ entry. Taken together, the different desensitization rates of the ATP-induced Ca 2ϩ signals after peak level could be the result of an interplay between inhibition by PKC and activation by PKA of the capacitative Ca 2ϩ entry.
Effect of Extracellular UTP on Cytosolic [Ca 2ϩ ] i in HL-60 Cells-Since it has been shown that P 2U receptors were present and coupled to PLC in HL-60 cells, we treated the cells with UTP and measured the return rate of the [Ca 2ϩ ] i level. Fig. 7A illustrates the times for return to 30% of the peak level in response to supramaximal concentrations of UTP: the higher the UTP concentration, the faster the return rate. However, although the phenomenon was similar to that of ATP, the rate of the return to the basal [Ca 2ϩ ] i level was not as remarkable compared with that induced by ATP in Fig. 1C. We further analyzed whether changes in the desensitization rate involve PKC and PKA in the modulation of the UTP-induced capacitative Ca 2ϩ entry using kinase inhibitors. Fig. 7B shows that pretreatment with GF 109203X slowed the decay rate induced by 300 M UTP. However, inhibition of PKA by pretreatment with H89 had no effect on the return rate of the 30 M UTPinduced [Ca 2ϩ ] i level (data not shown). The difference in the desensitization pattern between ATP and UTP might result from different activations of effector enzymes, such as PLC and adenylyl cyclase. In HL-60 cells, UTP has no effect on cAMP production (6). Therefore, the results suggest that the effect of PKA increasing the capacitative Ca 2ϩ entry is not involved in UTP-treated cells and that PKC alone acts in the desensitization of the UTP-induced Ca 2ϩ response.

Effects of ATP on cAMP Generation, IP 3 Production, and
Protein Kinase Activity-To assess agonist concentration-dependent differential activation of PKC and PKA, we measured the production of cAMP and IP 3 . Fig. 8 shows the production of cAMP and IP 3 induced by various concentrations of ATP. The maximal increase of cAMP was obtained with 300 M ATP. The EC 50 value was 19.2 M. Particularly, for 30 and 300 M ATP, respectively, the cAMP levels reached 137.2 Ϯ 9.3 and 190.2 Ϯ 7.7 over the basal cAMP level of 25.3 Ϯ 4.2. The amounts of IP 3 caused by 30 and 300 M ATP were 27.7 Ϯ 5.7 and 67.0 Ϯ 5.5 pmol/mg of protein, respectively, while the basal IP 3 level was 18.0 Ϯ 3.1. There was only a slight increase in the IP 3 level in the response to 30 M ATP, whereas the cAMP level was already significantly increased at that concentration. On the other hand, during stimulation with 300 M ATP, cAMP was produced maximally, and IP 3 was also dramatically increased over the basal IP 3 level, suggesting that both PKA and PKC might be highly activated. We directly measured the activities of the protein kinases induced by different concentrations of ATP. Fig. 9A shows that stimulation with 30 M ATP induced strong activation of PKA similar to 300 M ATP. However, stimulation with 30 M ATP produced a relatively weak activation of PKC (Fig. 9B), indicating that PKA was more significantly activated than PKC during the 30 M ATP stimulation. However, PKC seems to have a dominant effect during the highly activated state of PKA and PKC as occurs with the 300 M ATP treatment, because inhibition of the capacitative Ca 2ϩ entry was only exhibited during the stimulation with 300 M ATP.
To investigate the interrelation between PKA and PKC, thapsigargin-induced capacitative Ca 2ϩ entry was measured in cells treated simultaneously for 1 min with 100 nM PMA and 30 M ATP. As shown in Table I, PMA by itself has an inhibitory effect on the thapsigargin-induced capacitative Ca 2ϩ entry. Moreover, the 30 M ATP-induced potentiation of the capacitative Ca 2ϩ entry also disappeared when cells were simultaneously treated with PMA. Thus it seems likely that strong activation of PKC has a dominant effect on the capacitative Ca 2ϩ entry, even during a state of strongly activated PKA. Similarly, the dominant effect of PKC might cause the rapid decline of the 300 M ATP-induced Ca 2ϩ response. DISCUSSION The present study demonstrates that ATP stimulation with maximal concentrations of ATP causes different decay rates for the Ca 2ϩ signal after obtaining similar peak levels if supramaximal concentrations of ATP are used. We suggest that this is a result of homologous feedback regulation of PKC and PKA activation. ATP increases intracellular Ca 2ϩ by the release of Ca 2ϩ from the intracellular stores and by influx from the extracellular space. Most of the Ca 2ϩ increase was caused by capacitative Ca 2ϩ entry activated by the store depletion. It has also been reported that ATP activates nonselective cation channels permeable for Ca 2ϩ and Na ϩ in dibutyryl cAMP-differentiated HL-60 cells (24). Recently, Buell et al. (5) demonstrated the presence of P 2X1 in HL-60 cells. The current through the P 2X1 was barely detected in undifferentiated HL-60 cells. However, the current was markedly increased in differentiated cells. We cannot exclude the involvement of this nonselective cation channel in the homologous desensitization of the ATPinduced Ca 2ϩ response; however, its contribution would be small, since we used undifferentiated cells.
Our data indicate that the differences in the Ca 2ϩ signal were caused not by Ca 2ϩ release but by influx. However, it is possible that the different rates of desensitization could also be the result of a differential activity of the cytosolic Ca 2ϩ removing system. There are two major pathways by which to decrease the [Ca 2ϩ ] i . One is the pumping out of Ca 2ϩ from the cytosol to the cell exterior by Na ϩ /Ca 2ϩ exchanger and/or by plasma membrane Ca 2ϩ /ATPase. The other is the pumping of Ca 2ϩ into the intracellular stores by Ca 2ϩ /ATPase. The Na ϩ /Ca 2ϩ exchanger may not be directly involved in the phenomena of the present study, because stimulation of cells with ATP in Na ϩ -free medium or in the presence of the Na ϩ /Ca 2ϩ ex-changer blocker benzamil did not affect the desensitization pattern of the Ca 2ϩ responses elicited by supramaximal concentrations of ATP (data not shown). It has been reported that PKC stimulates Ca 2ϩ efflux by activation of plasma membrane Ca 2ϩ /ATPase in neutrophils (25). It seems unlikely that the activation of the Ca 2ϩ efflux was involved in the fast return to basal level at higher concentrations of ATP, because unidirectional Ca 2ϩ surrogates, Mn 2ϩ and Ba 2ϩ , showed a similar pattern as Ca 2ϩ .
It has been reported that the capacitative Ca 2ϩ entry is blocked by metal ions with an efficiency order of La 3ϩ Ͼ Zn 2ϩ Ͼ Cd 2ϩ Ͼ Be 2ϩ ϭ Co 2ϩ ϭ Mn 2ϩ Ͼ Ni 2ϩ Ͼ Sr 2ϩ Ͼ Ba 2ϩ (7). We found the same desensitization pattern between stimulations with 30 and 300 M ATP in the presence of La 3ϩ . This is consistent with the notion that the different desensitization rates of the ATP-induced Ca 2ϩ signals resulted from differential feedback regulation of the capacitative Ca 2ϩ entry evoked by ATP itself.
Using ATP as an agonist, we had to consider whether a particular cellular response was caused by the activation of a P 2 purinergic receptor per se. ATP is rapidly hydrolyzed to adenosine by extracellular ATPase and nucleotidase (26). But adenosine itself is a potent signaling substance. To assess this potential problem, we tested the hydrolysis-resistant ATP analog, adenosine ATP␥S, and obtained the same results as in the stimulations with ATP (data not shown). For example, ATP␥S also showed the differences in the Ca 2ϩ decay rate and the biphasic effect on the thapsigargin-induced capacitative Ca 2ϩ entry. These results indicate that the changes in the decay rate of elevated [Ca 2ϩ ] i were due to ATP and not to its metabolite.
In the experiments with UTP, the rate of return to the basal Ca 2ϩ level increased as the UTP concentration was raised, but the decay rate was not as prominent as with ATP. Recently, we (6) showed that ATP, but not UTP, elevates the cAMP level in HL-60 cells, maybe through a novel subtype of P 2 receptor. Therefore, the difference in the desensitization pattern between ATP and UTP might be due to the effects of the nucleotides on the activity of the protein kinases. While ATP activates both PKC and PKA, UTP activates only PKC, which would result in a difference in Ca 2ϩ signaling.
Until now, little is known about the signaling between the intracellular Ca 2ϩ store and the plasma membrane CRAC. It has been proposed that the signal is mediated via Ca 2ϩ entry factors, which include calcium influx factor (27,28), heterotrimeric G protein (29), and small G protein (30), or is mediated via direct interaction between the IP 3 receptor and the plasma membrane Ca 2ϩ channel (31). Recently, there were some reports describing the cloning and functional expression of a mammalian homologue to the Drosophila eye-specific trp gene (32)(33)(34). It was identified as a Ca 2ϩ -permeable cation channel that is activated by calcium store depletion. Although it was not clearly shown that TRP is the I CRAC protein, there is a possibility that TRP should be classified as one of the CRAC family (35). The molecular regulatory mechanism of signaling between the Ca 2ϩ store depletion and CRAC is controversial and complicated. There is some evidence that protein phosphorylation is involved in the regulation of capacitative Ca 2ϩ entry. In Xenopus oocytes and lymphocytes, protein phosphatase inhibitor potentiates Ca 2ϩ influx (36). Tyrosine kinase inhibitor blocks the bradykinin-and thapsigargin-induced Ca 2ϩ influx in lymphocytes and in human foreskin fibroblast cells (37,38). Protein kinase C-dependent phosphorylation plays a key role in the modulation of the capacitative Ca 2ϩ entry, too. In the insulin-secreting cell line RINm5F, PKC activates capacitative Ca 2ϩ entry (39). On the contrary, PKC stimulation has been shown to inhibit capacitative Ca 2ϩ entry in thyroid cells   (40). In human neutrophils, formyl-methionyl-leucyl-phenylalanine (41) and PMA (42) inhibited capacitative Ca 2ϩ entry, which was mediated by PKC. Capacitative Ca 2ϩ entry caused by Drosophila photoreceptor activation is inhibited by PKC as well (43). Here, we suggest that 300 M ATP preferentially inhibits capacitative Ca 2ϩ entry by PKC activation. However, little is known about the involvement of PKA in the capacitative Ca 2ϩ entry. It has been reported that activation of PKA had a biphasic effect on Ca 2ϩ entry-evoked currents in thapsigargin-treated Xenopus oocytes. Application of dibutyryl cAMP at 1-10 M inhibited the current, whereas at 1-10 mM potentiated the current (44). We show here that 30 M ATP preferentially activates capacitative Ca 2ϩ entry by relatively strong PKA activation rather than PKC activation. This activation is not the result of the further emptying the intracellular stores, because Ca 2ϩ stores may be fully depleted after thapsigargin treatment for 10 min, and no Ca 2ϩ increase was detectable upon subsequent ATP treatment. We also tested the effect of prostaglandin E 2 on the thapsigargin-induced capacitative Ca 2ϩ entry. Prostaglandin E 2 activates adenylyl cyclase and increases intracellular cAMP concentration, but there were no detectable changes in the Ca 2ϩ signal and IP 3 generation in HL-60 cells. 2 Prostaglandin E 2 also potentiates the thapsigargin-induced capacitative Ca 2ϩ entry as shown in the stimulation with 30 M ATP. This result also suggests that PKA activates the capacitative Ca 2ϩ entry in HL-60 cells. In many cell types, functional effects elicited by extracellular ATP are related to a Ca 2ϩ increase. Therefore, the Ca 2ϩ increase must be tightly regulated to maintain cellular homeostasis and to exert physiological effects. The fine regulation of the ATP-induced Ca 2ϩ signal could be achieved in a feedback mode with PKC and PKA, which are differentially activated according to the extent of stimulation caused by different ATP concentrations.