Interdependence of calcium signaling and protein tyrosine phosphorylation in human endothelial cells.

The signal transduction cascade which initiates transmembraneous influx of Ca2+ into endothelial cells in response to the discharge of intracellular Ca2+ stores is thought to involve a step sensitive to tyrosine kinase inhibition. We investigated the interrelationship between Ca2+ signaling and protein tyrosine phosphorylation following cell stimulation with either the receptor-dependent agonist, bradykinin, or the protein-tyrosine phosphatase inhibitor, phenylarsine oxide. In cultured human endothelial cells phenylarsine oxide instigated a concentration-dependent increase in the intracellular concentration of free Ca2+ ([Ca2+]i). This increase in [Ca2+]i was not associated with the tyrosine phosphorylation of phospholipase C gamma, enhanced formation of inositol 1,4,5-trisphosphate, or the rapid depletion of intracellularly stored Ca2+ but was coincident with the enhanced and prolonged tyrosine phosphorylation of a number of cytoskeletal proteins. In bradykinin-stimulated cells the tyrosine phosphorylation of the same cytoskeletal proteins (most notably 85- and 100-kDa proteins) was transient when cells were stimulated in the presence of extracellular Ca2+, was maintained under Ca2+-free conditions, and was reversed following readdition of extracellular Ca2+. These data suggest that the tyrosine phosphorylation of 2 cytoskeletal proteins is determined by the level of Ca2+ present in intracellular stores thus indicating a critical role for tyrosine phosphorylation in the control of capacitative Ca2+ entry in endothelial cells.

The signal transduction cascade which initiates transmembraneous influx of Ca 2؉ into endothelial cells in response to the discharge of intracellular Ca 2؉ stores is thought to involve a step sensitive to tyrosine kinase inhibition. We investigated the interrelationship between Ca 2؉ signaling and protein tyrosine phosphorylation following cell stimulation with either the receptordependent agonist, bradykinin, or the protein-tyrosine phosphatase inhibitor, phenylarsine oxide. 2؉ ] i was not associated with the tyrosine phosphorylation of phospholipase C␥, enhanced formation of inositol 1,4,5-trisphosphate, or the rapid depletion of intracellularly stored Ca 2؉ but was coincident with the enhanced and prolonged tyrosine phosphorylation of a number of cytoskeletal proteins. In bradykinin-stimulated cells the tyrosine phosphorylation of the same cytoskeletal proteins (most notably 85-and 100-kDa proteins) was transient when cells were stimulated in the presence of extracellular Ca 2؉ , was maintained under Ca 2؉ -free conditions, and was reversed following readdition of extracellular Ca 2؉ . These data suggest that the tyrosine phosphorylation of 2 cytoskeletal proteins is determined by the level of Ca 2؉ present in intracellular stores thus indicating a critical role for tyrosine phosphorylation in the control of capacitative Ca 2؉ entry in endothelial cells.

In cultured human endothelial cells phenylarsine oxide instigated a concentration-dependent increase in the intracellular concentration of free Ca 2؉ ([Ca 2؉ ] i ). This increase in [Ca
There is a certain amount of evidence to suggest that the tyrosine phosphorylation of, as yet unidentified, cellular proteins may be involved in the control of store-regulated or "capacitative" Ca 2ϩ entry following the agonist-induced depletion of intracellular stores in non-excitable cells. The first evidence for such a role of tyrosine kinases in intracellular Ca 2ϩ signaling was obtained in platelets, in which the tyrosine phosphorylation of a group of proteins was found to be transiently elevated following stimulation with thrombin (1)(2)(3). This thrombin-induced protein tyrosine phosphorylation could be mimicked by the depletion of intracellular Ca 2ϩ stores following inhibition of the Ca 2ϩ -ATPase and was sensitive to the chelation of intracellular Ca 2ϩ . Repletion of Ca 2ϩ stores was, on the other hand, associated with a return to basal phosphorylation levels (1). These observations inferred that the depletion of intracellular Ca 2ϩ stores favors tyrosine phosphorylation whereas store refilling and the restoration of homeostatic levels of [Ca 2ϩ ] i favors tyrosine dephosphorylation of specific proteins (1). In support of these findings, protein-tyrosine kinase inhibitors such as genistein were reported to inhibit both the thrombin-induced increase in [Ca 2ϩ ] i as well as the subsequent aggregation (2). Since genistein has been found to attenuate Ca 2ϩ influx following cell stimulation with both receptordependent and -independent agonists it would appear that the tyrosine kinase substrate protein is likely to be intimately involved in the regulation of Ca 2ϩ entry processes rather than being linked to specific cell receptors. Recently we have demonstrated that the bradykinin as well as the thapsigargininduced Ca 2ϩ influx in endothelial cells is also mediated by a tyrosine kinase inhibitor-sensitive mechanism (4). Although the chain of events which result in the enhanced membrane permeability to Ca 2ϩ remain largely unexplained, these observations suggest that a tyrosine-phosphorylated protein might be involved in the regulation of Ca 2ϩ influx. Therefore the aim of the present study was to address the hypothesis that protein tyrosine phosphorylation controls capacitative Ca 2ϩ influx in human endothelial cells. To this end we investigated the effects of agonist stimulation on tyrosine phosphorylation as well as the effects on [Ca 2ϩ ] i of altering the balance between tyrosine kinase/phosphatase activity.
Immunoblotting-Confluent primary cultures of human umbilical vein endothelial cells were washed twice in HEPES-Tyrode solution and incubated at 37°C with or without various receptor-dependent and -independent stimuli as described under "Results." Thereafter, cells were washed with ice-cold HEPES buffer containing sodium fluoride (100 mM), Na 4 P 2 O 7 (15 mM), Na 3 VO 4 (2 mM), leupeptin (2 g/ml), pepstatin A (2 g/ml), trypsin inhibitor (10 g/ml), and phenylmethylsulfonyl fluoride (44 g/ml) and harvested by scraping. The cell suspension was centrifuged at 13,000 ϫ g for 60 s, cells contained in the pellet were then lysed in buffer containing 1% (v/v) Triton X-100, left on ice for 5 min, and centrifuged at 10,000 ϫ g for 10 min. Approximately 30 g of protein from the resulting supernatant or from the Triton X-100 insoluble fraction was separated by 10 or 7% SDS-polyacrylamide gel electrophoresis, respectively, as described (4). Tyrosine-phosphorylated proteins were detected with a mouse monoclonal anti-phosphotyrosine antibody (1 g/ml) (Upstate Biotechnology Inc.) and were visualized by enhanced chemiluminescence using a commercially available kit (Amersham). Prestained molecular weight marker proteins (Bio-Rad) were used as standards for the SDS-polyacrylamide gel electrophoresis.
Assay of Inositol 1,4,5-Trisphosphate-Confluent cultures of human umbilical vein endothelial cells were washed twice with HEPES-Tyrode solution and allowed to equilibrate for 20 min. Thereafter, cells were stimulated with either bradykinin (100 nM) or phenylarsine oxide (10 M) for the times indicated under "Results." The incubation was stopped by the aspiration of the HEPES-Tyrode and addition of 400 l of ice-cold 6% (v/v) trichloroacetic acid. Samples were then left on ice for 30 min, cells and supernatant were then harvested by scraping and the precipitates centrifuged at 2000 ϫ g (15 min, 4°C). The supernatants were then extracted 4 times with 5 volumes of water-saturated diethyl ether and were neutralized by titration with NaHCO 3 . Inositol 1,4,5trisphosphate (IP 3 ) contained in the samples was then assessed using a commercially available kit (Biotrak, Amersham).
Protein-tyrosine Phosphatase Assay-Protein-tyrosine phosphatase activity was assayed in whole cells lysates from cells stimulated as described under "Results." Activity was determined by monitoring the dephosphorylation of the phosphopeptide RRLIEDAEpYAARG using a commercially available kit (Upstate Biotechnology Inc.).
Statistical Analysis-Unless otherwise indicated data are expressed as mean Ϯ S.E. Statistical evaluation was performed using Student's t test for unpaired data, one-way analysis of variance (ANOVA) followed by a Bonferroni t test, or ANOVA for repeated measures where appropriate. Values of p Ͻ 0.05 were considered statistically significant.

Effect Phenylarsine Oxide on [Ca 2ϩ
] i -In fura-2-loaded human endothelial cells the protein-tyrosine phosphatase inhibitor, phenylarsine oxide induced a slightly delayed and concentration-dependent (1-10 M) increase in [Ca 2ϩ ] i (Fig. 1A). Concentrations of phenylarsine oxide lower that 1 M failed to have any effect on [Ca 2ϩ ] i , while the addition of 3 M phenylarsine oxide resulted in a gradual increase in [Ca 2ϩ ] i and consistently produced low frequency [Ca 2ϩ ] i oscillations which were synchronized throughout the entire cell population. Preliminary evidence indicates that this phenomenon is associated with a concomitant oscillation in membrane potential attributed to the activation of Ca 2ϩ -activated K ϩ channels (not shown). The highest concentration of phenylarsine oxide used (10 M) produced a rapid increase in [Ca 2ϩ ] i which in 40% of the cells tested were biphasic in nature, consisting of an initial Ca 2ϩ peak followed by a second more delayed Ca 2ϩ increase. Oscillations in [Ca 2ϩ ] i were rarely observed in the presence of the highest concentration of phenylarsine oxide (10 M).
The phenylarsine oxide-induced increase in [Ca 2ϩ ] i was completely reversed by the dithiol reagent 2,3-dimercaptopropanol (50 M) which binds to vicinal sulfhydryl groups (Fig. 2). In these experiments, the subsequent addition of bradykinin (100 nM) resulted in a normal Ca 2ϩ response suggesting that the effects of phenylarsine oxide were completely reversible and that the inhibitor did not lead to the permanent uncoupling of the agonist-induced signal transduction pathway.
In the absence of extracellular Ca 2ϩ , phenylarsine oxide (10 M) induced a small, slowly developing, increase in [Ca 2ϩ ] i (after 10 min [Ca 2ϩ ] i had increased from 69.7 Ϯ 12 nM to 106.5 Ϯ 23 nM, n ϭ 8, p Ͻ 0.001; Fig. 3). Subsequent addition of the Ca 2ϩ -ATPase inhibitor, thapsigargin (0.3 M), resulted in an immediate further increase in [Ca 2ϩ ] i ([Ca 2ϩ ] i increased from 104 Ϯ 24 to 218 Ϯ 15 nM; n ϭ 4, p Ͻ 0.01) demonstrating that the protein-tyrosine phosphatase inhibitor did not deplete intracellular Ca 2ϩ stores. In cells stimulated with phenylarsine oxide in the absence of extracellular Ca 2ϩ , the readdition of Ca 2ϩ was associated with an immediate increase in [Ca 2ϩ ] i (Fig. 3).
Effects of Phenylarsine Oxide and Bradykinin on PLC␥ Phosphorylation and IP 3 Production-In order to establish that the effects of the tyrosine phosphatase inhibitor on [Ca 2ϩ ] i were not related to the depletion of intracellular Ca 2ϩ stores the effects of phenylarsine oxide on the tyrosine phosphorylation of PLC␥ and production of IP 3 were compared with those of the receptor-dependent agonist bradykinin.
When tyrosine-phosphorylated proteins were immunoprecipitated from endothelial cells treated with phenylarsine oxide (10 M) and blotted with anti-PLC␥ 1 , a clear signal was apparent in control cells but could not be detected in cells incubated with the tyrosine phosphatase inhibitor for up to 5 min (Fig.  4A). Stimulation of endothelial cells with bradykinin (100 nM), on the other hand, resulted in a rapid increase in the tyrosine phosphorylation of PLC␥ 1 with a 3.5-fold increase in tyrosinephosphorylated protein being detected within 30 s (Fig. 4B). The bradykinin-induced tyrosine phosphorylation of PLC␥ 1 was relatively transient and tyrosine phosphorylation of the PLC␥ 1 returned to near basal levels within 2 min.
In accordance with its effect on the tyrosine phosphorylation of PLC␥ 1 , phenylarsine oxide (10 M) failed to precipitate an increase in intracellular levels of IP 3 at any of the time points measured. In bradykinin-treated endothelial cells IP 3 levels increased 7-fold within 10 s and had returned to baseline values within 1 min (Fig. 4C).
These observations demonstrate that the phenylarsine oxide-induced increase in [Ca 2ϩ ] i is not due to the rapid depletion of intracellular stores and that activation of a transmembraneous influx accounts for most of the tyrosine phosphatase inhibitor-induced increase in [Ca 2ϩ ] i . As a consequence of its effects on fura-2 fluorescence it was not possible to repeat these ex-periments using a second widely used tyrosine phosphatase inhibitor, sodium orthovanadate.
Effect of Tyrosine Phosphatase Inhibitors on Triton-soluble Proteins-Since the overall cellular level of tyrosine phosphorylation is determined by the delicate balance between the activity of tyrosine kinases and tyrosine phosphatases, inhibition of dephosphorylation would be expected to result in a net increase in detectable phosphotyrosine-containing proteins. In cultured human endothelial cells, phenylarsine oxide (10 M) induced a clear time-dependent increase in the tyrosine phosphorylation of a triplet of bands centered at ϳ80 kDa as well as a 42/44-kDa doublet (Fig. 5A). The latter proteins were identified in immunoprecipitation experiments as the 42-and 44-kDa isoforms of the mitogen-activated protein kinase (MAP kinase: not shown). The enhanced tyrosine phosphorylation of MAP kinase was evident 2 min after addition of the inhibitor and was maximal after 5-10 min. These effects of phenylarsine oxide were not observed in endothelial cells pretreated with the tyrosine kinase inhibitor genistein (100 M, 10 min: not shown).
In order to evaluate the role of [Ca 2ϩ ] i in phenylarsine oxidestimulated tyrosine phosphorylation, cells were either incubated in a nominally Ca 2ϩ -free buffer or pretreated with the intracellular Ca 2ϩ chelator BAPTA (10 M, 30 min). Phenylarsine oxide-induced tyrosine phosphorylation of the 42/44-kDa doublet was largely Ca 2ϩ -dependent since a slight increase in tyrosine phosphorylation was observed in cells stimulated in the absence of Ca 2ϩ but no increase was observed in BAPTAtreated cells. The tyrosine phosphorylation of the ϳ80-kDa triplet was, on the other hand, largely Ca 2ϩ -independent (Fig.  5B).
A second protein-tyrosine phosphatase inhibitor, sodium orthovanadate (0.3 mM), also resulted in a genistein-sensitive, time-dependent increase in the tyrosine phosphorylation of the 42-and 44-kDa isoforms of MAP kinase as well as the ϳ80-kDa triplet (Fig. 6).
Effect of Tyrosine Phosphatase Inhibitors on Triton-insoluble Cytoskeletal Proteins-Protein-tyrosine phosphatase inhibition had much more rapid and pronounced effects on levels of phosphotyrosine containing proteins in the Triton X-100-insoluble (cytoskeletal) fraction. Within 30 s phenylarsine oxide induced phosphorylation of 5 distinct bands, corresponding to estimated molecular masses of ϳ70, 75, 77, and a doublet of ϳ220 kDa, while a series of bands of approximately 100, 130, and ϳ265 kDa were apparent after 1 to 2 min (Fig. 7A).
Pretreatment of endothelial cells with the tyrosine kinase inhibitor genistein tended to attenuate the phenylarsine oxideinduced tyrosine phosphorylation although the effects were by no means prevented by the inhibitor (Fig. 7B). Almost complete reversal of the phenylarsine oxide-induced increase in tyrosine phosphorylation was achieved by subsequent addition of 2,3dimercaptopropanol (50 M ; Fig. 7C).
The removal of extracellular Ca 2ϩ or the chelation of intracellular Ca 2ϩ failed to alter the phenylarsine oxide induced increase in tyrosine phosphorylation of cytoskeletal proteins (Fig. 8).
Effect of Bradykinin on Triton-soluble and -insoluble Proteins-In order to investigate the putative link between Ca 2ϩ signaling and the tyrosine phosphorylation of specific proteins, we studied the effect of extracellular Ca 2ϩ removal and readdition on bradykinin-stimulated tyrosine phosphorylation.
In the presence of extracellular Ca 2ϩ , bradykinin (100 nM) induced an immediate increase in the tyrosine phosphorylation of 3 Triton-soluble proteins (60, 77, and 86 kDa). Phosphorylation of these proteins was maximal after 30 s and returned to control levels over 5 min (Fig. 9A). In the same cells bradykinin also induced tyrosine phosphorylation of the 42/44-kDa MAP kinase doublet which was detectable 2 min after agonist stimulation and was maximal after 5 min, as described previously (4). Stimulation of endothelial cells in the absence of extracellular Ca 2ϩ failed to alter phosphorylation of the 60-, 77-, or 86-kDa proteins but resulted in only a transient increase in the tyrosine phosphorylation of the MAP kinases which was no longer apparent after 5 min (Fig. 9B). Phosphorylation of the 42/44-kDa doublet reappeared 1 to 2 min after the readdition of extracellular Ca 2ϩ to these cells (Fig. 9C). Pretreatment of endothelial cells with BAPTA abrogated the bradykinin-induced tyrosine phosphorylation of the p42 and p44 (data not shown).
Bradykinin stimulation resulted in an immediate but transient increase in the phosphorylation of 4 cytoskeletal proteins with estimated molecular masses of 85, 100, 110, and 125 kDa which returned to baseline levels after 5-10 min (Fig. 10A). In the absence of extracellular Ca 2ϩ , bradykinin induced a distinct increase in the phosphorylation of the 85-and 100-kDa proteins which was immediately reversed upon readdition of extracellular Ca 2ϩ (Fig. 10B).
Effect of Thapsigargin on Triton-soluble and -Insoluble Proteins-Thapsigargin (100 nM) induced tyrosine phosphorylation of the 42-and 44-kDa isoforms of the MAP kinase. Tyrosine phosphorylation of the 42/44-kDa doublet was not evident in cells stimulated in the absence of extracellular Ca 2ϩ but a distinct tyrosine phosphorylation of both bands was detected 5 min after the readdition of Ca 2ϩ to the incubation medium (not shown). These findings are in line with previously published results (4).
Thapsigargin (100 nM) induced the tyrosine phosphorylation of 4 proteins in the Triton-insoluble fraction corresponding to molecular masses of 85, 100, 110, and 125 kDa. However, the tyrosine phosphorylation of these bands was not transient, as was observed following cell stimulation with bradykinin, but was maintained for up to 10 min. Removal of extracellular Ca 2ϩ did not influence the pattern of tyrosine phosphorylation which emerged following stimulation with thapsigargin, and the readdition of extracellular Ca 2ϩ to depleted cells was not associated with a visible change in the phosphorylation pattern (Fig. 11).
Effects of Phenylarsine Oxide and Bradykinin on Tyrosine Phosphatase Activity-Human endothelial cells were found to express a basal tyrosine phosphatase activity which was attenuated in the presence of both phenylarsine oxide and orthovanadate. Bradykinin induced a 2-fold increase in phosphatase activity which was inhibited in cells pretreated with either phenylarsine oxide or sodium orthovanadate (Table I). DISCUSSION Over the last few years there have been several reports that the transmembraneous influx of Ca 2ϩ is selectively attenuated in a number of cell types following inhibition of tyrosine kinases (1, 8 -10), thus suggesting that cellular levels of tyrosine phosphorylation play a determinant role in regulating Ca 2ϩ entry in non-excitable cells. In the present study the proteintyrosine phosphatase inhibitor, phenylarsine oxide, induced a concentration-dependent increase in [Ca 2ϩ ] i and elicited the tyrosine phosphorylation of a number of endothelial proteins with the most marked effects being apparent in the Triton X-100-insoluble, or cytoskeletal, fraction. Both the increase in [Ca 2ϩ ] i and the enhanced tyrosine phosphorylation were attenuated in cells pretreated with the tyrosine kinase inhibitor, genistein, supporting the hypothesis that a tyrosine-phosphorylated protein may be involved in the regulation of [Ca 2ϩ ] i in endothelial cells.
In principle there are two ways by which an agonist can induce capacitative Ca 2ϩ entry. The activation of the transmembraneous Ca 2ϩ influx pathway could be attributed to an indirect, i.e. instigation of capacitative Ca 2ϩ entry following the mobilization of intracellularly stored Ca 2ϩ , or a direct effect, i.e. activation of a Ca 2ϩ influx regulatory protein. In contrast to the receptor-dependent agonist, bradykinin, phenylarsine oxide failed to tyrosine phosphorylate PLC␥ or increase cellular levels of IP 3 . The tyrosine phosphatase inhibitor therefore appeared unable to mobilize intracellular Ca 2ϩ via the classical signaling pathway associated with agonist-induced activation of endothelial cells. Inhibition of the Ca 2ϩ -ATPase is also unable to account for the phenylarsine oxide-induced Ca 2ϩ response. In the absence of extracellular Ca 2ϩ phenylarsine oxide had no immediate effect on [Ca 2ϩ ] i but when incubated with cells for longer periods did induce a slight elevation. Subse- quent addition of thapsigargin to these cells resulted in an immediate further increase in [Ca 2ϩ ] i suggesting that the tyrosine phosphatase inhibitor did not mobilize Ca 2ϩ from intracellular stores. However, since the readdition of Ca 2ϩ to cells stimulated with phenylarsine oxide in the absence of extracellular Ca 2ϩ resulted in a marked and instantaneous increase in [Ca 2ϩ ] i , it would appear that the tyrosine phosphatase inhibitor directly activates a Ca 2ϩ influx pathway which does not appear to be regulated by the filling state of intracellular Ca 2ϩ stores.
Our observation that tyrosine phosphatase inhibitors appear to be able to activate Ca 2ϩ influx pathways without first mobilizing intracellular stores is supported by similar findings in T lymphocytes following the administration of phenylarsine oxide and pervanadate (11,12). These data, together with the results obtained in the present study, suggest that the effects of the protein-tyrosine phosphatase inhibitors on [Ca 2ϩ ] i may instead be a direct consequence of the enhanced tyrosine phosphorylation of a Ca 2ϩ influx regulatory protein. This hypothesis is supported by the report that transfection of T cells with the constitutively active tyrosine kinase v-Src results in elevated basal levels of [Ca 2ϩ ] i as well as in the exaggeration of agonist-stimulated Ca 2ϩ responses (13).
Since the inhibition of protein-tyrosine phosphatases resulted in an increase in [Ca 2ϩ ] i , it would appear likely that phosphatase activity plays a crucial role in damping Ca 2ϩ influx both in unstimulated cells and following agonist stimulation. Indeed, the observation that protein-tyrosine phosphatase inhibitors can themselves induce cellular responses implies that there is a significant basal activity of protein-tyrosine phosphatases in cultured human endothelial cells. This was confirmed by determination of tyrosine phosphatase activity in whole cell lysates. Moreover, the observation that the tyrosine kinase inhibitor genistein attenuated both the increase in [Ca 2ϩ ] i and tyrosine phosphorylation initiated by phenylarsine oxide suggests that a certain basal phosphatase activity is required to counteract phosphorylation by constitutively active protein-tyrosine kinases. Thus the dynamic balance between tyrosine kinase and phosphatase activity may play a central role in the maintenance of homeostatic levels of [Ca 2ϩ ] i in unstimulated cells.
Based on current knowledge, the hypothetically ideal Ca 2ϩ influx-regulatory protein in non-excitable cells should be "activated" immediately after agonist-induced emptying of intracellular Ca 2ϩ stores, even in the absence of extracellular Ca 2ϩ , and to remain so until store filling is accomplished. Ideally this protein should be membrane-associated, either permanently or temporarily, and preferably linked by some manner or means to the cation channel by which Ca 2ϩ enters the cell. To identify proteins which conform with these criteria the tyrosine phosphorylation of proteins from cells stimulated with bradykinin, thapsigargin, and the tyrosine phosphatase inhibitors was monitored in Triton X-100-soluble and -insoluble fractions. In the soluble fraction, bradykinin induced the rapid and transient phosphorylation of 3 proteins which contrasted with the relatively slow tyrosine phosphorylation of the 42-and 44-kDa isoforms of MAP kinase, as described previously (4). Cell stimulation in the absence of extracellular Ca 2ϩ was without effect on the phosphorylation of the ϳ60-, 77-, and 86-kDa proteins whereas agonist-induced phosphorylation of the 42-and 44-kDa isoforms of the MAP kinase was critically dependent on an increase in [Ca 2ϩ ] i . The Ca 2ϩ -ATPase inhibitor, thapsigargin, elicited essentially the same effects on Triton-soluble proteins. Sodium orthovanadate and phenylarsine oxide, however, induced the maintained, Ca 2ϩ -independent phosphorylation of a protein triplet of ϳ80 kDa.
In the cytoskeletal fraction bradykinin also evoked a rapid and transient tyrosine phosphorylation of 4 proteins (85, 100, 110, and 125 kDa) which appeared identical to proteins permanently tyrosine phosphorylated following application of phenylarsine oxide. In contrast to the effects seen in the presence of  extracellular Ca 2ϩ , a maintained tyrosine phosphorylation of these proteins was observed in cells stimulated following the removal of extracellular Ca 2ϩ . Readdition of extracellular Ca 2ϩ to these Ca 2ϩ -depleted cells was associated with the transient dephosphorylation and maintained rephosphorylation of the 110-and 125-kDa proteins and the sustained dephosphorylation of the 85-and 100-kDa proteins. In similar experiments using thapsigargin, the reapplication of extracellular Ca 2ϩ to depleted cells did not affect tyrosine phosphorylation of the 125-kDa protein or result in the dephosphorylation of the 85and 100-kDa proteins. This observation was, however, not unexpected since, in the continued presence of the Ca 2ϩ -ATPase inhibitor refilling of intracellular Ca 2ϩ stores is antagonized, thus the store remains empty although the signal for refilling, is sustained. In bradykinin-stimulated cells, however, store refilling could be accomplished following Ca 2ϩ readdition and the tyrosine phosphorylation of the 85-and 100-kDa proteins was transient. It would therefore appear that the tyrosine phosphorylation of the 85-and 100-kDa proteins mirrors the filling state of intracellular Ca 2ϩ stores, thus these two cytoskeletal proteins fit the requirements of the hypothetical Ca 2ϩ influx regulatory protein.
Although similar experiments involving intracellular Ca 2ϩ depletion and repletion in platelets have also demonstrated the existence of tyrosine-phosphorylated proteins apparently sensitive to the filling state of the Ca 2ϩ store (1,14), the reported apparent molecular weights of these proteins differs from that of proteins displaying similar characteristics in the present study.
In addition to the putative tyrosine-phosphorylated Ca 2ϩ influx regulatory protein, a number of other mechanisms have been proposed to regulate capacitative Ca 2ϩ entry. Nonhydrolyzable analogues of GTP, such as GTP␥S, have been shown to interfere with Ca 2ϩ signaling in a number of cell types. The inhibitory effect of these analogues occurs at some point after the release of intracellular Ca 2ϩ and prior to the activation of Ca 2ϩ influx. These effects can be prevented by GTP thus implying that a small G protein is involved in communicating the empty state of intracellular Ca 2ϩ stores to the plasma membrane (15)(16)(17)(18). The role of the "calcium influx factor," a small, phosphate-containing, non-protein factor termed originally isolated from Jurkat T lymphocytes (19), as an exclusive messenger for capacitative Ca 2ϩ entry has recently been questioned since the lymphocyte-derived factor has also been demonstrated to mobilize Ca 2ϩ from intracellular stores (20). Involvement of the cytochrome P-450 monooxygenase in the regulation of Ca 2ϩ influx has also been proposed on the basis of observations that a number of chemically distinct P-450 inhibitors, such as the imidazole anti-fungal agents, potently inhibited Ca 2ϩ influx in endothelial cells and platelets (21)(22)(23). This hypothesis is supported by the findings that the induction of P-450 by ␤-naphtoflavone, potentiated agonist-induced Ca 2ϩ influx and that the P-450 product, 5,6-epoxyeicosatrienoic acid, activated Ca 2ϩ entry into endothelial cells without prior depletion of intracellular Ca 2ϩ (24). Such observations suggest that the regulation of capacitative Ca 2ϩ entry into endothelial cells is a complex process likely to involve the activation of protein tyrosine kinases and phosphatases, small G proteins, serine/ threonine phosphatases, and probably also the cytochrome P-450 monooxygenase.
In summary, in the present study we observed that treatment of endothelial cells with protein-tyrosine phosphatase inhibitors resulted in the prolonged tyrosine phosphorylation of 2 cytoskeletal proteins and an increased Ca 2ϩ influx via a mechanism independent of intracellular Ca 2ϩ store depletion. Our findings strongly suggest that the tyrosine phosphorylation of both cytoskeletal proteins mirrors the filling state of the intracellular Ca 2ϩ store and that they play a central role in the regulation of capacitative Ca 2ϩ entry.