Sodium-calcium exchange and store-dependent calcium influx in transfected chinese hamster ovary cells expressing the bovine cardiac sodium-calcium exchanger. Acceleration of exchange activity in thapsigargin-treated cells.

The effects of extracellular Na on store-dependent Ca influx were compared for transfected Chinese hamster ovary cells expressing the bovine cardiac Na-Ca exchanger (CK1.4 cells) and vector-transfected control cells. Store-dependent Ca influx was elicited by depletion of intracellular Ca stores with ionomycin, thapsigargin, or extracellular ATP, a purinergic agonist. In each case, the rise in [Ca] upon the addition of extracellular Ca was reduced in CK1.4 cells compared with control cells at physiological [Na]. When Li or NMDG was substituted for Na, the CK1.4 cells showed a greater rise in [Ca] than control cells over the subsequent 3 min after the addition of Ca. Under Na-free conditions, SK& 96365 (50 μM), a blocker of store-operated Ca channels, nearly abolished the thapsigargin-induced rise in [Ca] in the control cells but only partially inhibited this response in the CK1.4 cells. We conclude that in the CK1.4 cells, Ca entry through store-operated channels was counteracted by Na-dependent Ca efflux at physiological [Na], whereas Ca entry was enhanced through Na-dependent Ca influx in the Na-free medium. We examined the effects of thapsigargin on Ba entry in the CK1.4 cells because Ba is transported by the Na-Ca exchanger, but it enters these cells only poorly through store-operated channels, and it is not sequestered by intracellular organelles. Thapsigargin treatment stimulated Ba influx in a Na-free medium, consistent with an acceleration of Ba entry through the Na-Ca exchanger. We conclude that organellar Ca release induces a regulatory activation of Na-Ca exchange activity.

The effects of extracellular Na ؉ on store-dependent Ca 2؉ influx were compared for transfected Chinese hamster ovary cells expressing the bovine cardiac Na ؉ -Ca 2؉ exchanger (CK1.4 cells) and vector-transfected control cells. Store-dependent Ca 2؉ influx was elicited by depletion of intracellular Ca 2؉ stores with ionomycin, thapsigargin, or extracellular ATP, a purinergic agonist. In each case, the rise in [Ca 2؉ ] i upon the addition of extracellular Ca 2؉ was reduced in CK1. 4

cells compared with control cells at physiological [Na ؉ ] o . When
Li ؉ or NMDG was substituted for Na ؉ , the CK1. 4

cells showed a greater rise in [Ca 2؉ ] i than control cells over the subsequent 3 min after the addition of Ca 2؉
o . Under Na ؉ -free conditions, SK&F 96365 (50 M), a blocker of store-operated Ca 2؉ channels, nearly abolished the thapsigargin-induced rise in [Ca 2؉ ] i in the control cells but only partially inhibited this response in the CK1.4 cells. We conclude that in the CK1.4 cells, Ca 2؉ entry through store-operated channels was counteracted by Na ؉ o -dependent Ca 2؉ efflux at physiological [Na ؉ ] o , whereas Ca 2؉ entry was enhanced through Na ؉ i -dependent Ca 2؉ influx in the Na ؉ -free medium. We examined the effects of thapsigargin on Ba 2؉ entry in the CK1.4 cells because Ba 2؉ is transported by the Na ؉ -Ca 2؉ exchanger, but it enters these cells only poorly through store-operated channels, and it is not sequestered by intracellular organelles. Thapsigargin treatment stimulated Ba 2؉ influx in a Na ؉ -free medium, consistent with an acceleration of Ba 2؉ entry through the Na ؉ -Ca 2؉ exchanger. We conclude that organellar Ca 2؉ release induces a regulatory activation of Na ؉ -Ca 2؉ exchange activity.
Agents that promote the production of 1,4,5-inositol trisphosphate (InsP 3 ) 1 give rise to a biphasic increase in cytosolic Ca 2ϩ . The initial, transient phase is primarily due to release of Ca 2ϩ from intracellular stores, whereas the more prolonged plateau phase involves an accelerated influx of extra-cellular Ca 2ϩ . The Ca 2ϩ influx pathway involves low conductance Ca 2ϩ channels and is activated, through a poorly understood mechanism, by the loss of Ca 2ϩ from the InsP 3sensitive stores (1)(2)(3)(4). This process is designated as capacitative Ca 2ϩ entry (3) or store-dependent Ca 2ϩ influx (SDCI) (4). When cells are exposed to a Ca 2ϩ -mobilizing agent in the absence of extracellular Ca 2ϩ , the SDCI pathway remains activated (even after removal of the Ca 2ϩ -mobilizing agent) until Ca 2ϩ is restored and the InsP 3 -sensitive stores refill with Ca 2ϩ . Inhibitors of sarco(endo)plasmic reticulum Ca 2ϩ ATPase such as thapsigargin (Tg) prevent Ca 2ϩ reaccumulation by the InsP 3 -sensitive stores, resulting in prolonged activation of SDCI (5).
The Na ϩ -Ca 2ϩ exchanger is a carrier-mediated transport process that couples the transmembrane movement of 3 Na ϩ ions to the movement of a single Ca 2ϩ in the opposite direction. In cardiac cells, it is widely accepted that the exchanger transports a portion of the Ca 2ϩ released from the sarcoplasmic reticulum out of the cells by Na ϩ o -dependent Ca 2ϩ efflux and in this way regulates the amount of stored Ca 2ϩ available for release during subsequent beats (reviewed in Ref. 6). However, in many other types of cells the downstream response to Ca 2ϩmobilizing agents depends more on the sustained influx of Ca 2ϩ following agonist addition than on the magnitude of the Ca 2ϩ transient itself. For noncardiac cells, there is a wealth of evidence supporting the generalized activity of Na ϩ -Ca 2ϩ exchange in mediating Ca 2ϩ efflux and in modulating the Ca 2ϩ content of intracellular stores (7). However, the precise physiological role of the exchanger and its interactions with other Ca 2ϩ hemeostatic mechanisms is not as clearly defined in noncardiac cells as in myocardial cells.
A major difficulty in investigations of exchanger function in intact cells is the absence of a selective inhibitor for exchange activity. In this report, we utilize transfected CHO cells to bypass this limitation. CHO cells do not normally express Na ϩ -Ca 2ϩ exchange activity, but after transfection with an expression vector coding for the bovine cardiac Na ϩ -Ca 2ϩ exchanger, high levels of activity are observed (8,9). Comparing the effects of Na ϩ on Ca 2ϩ mobilization between transfected and vectortransfected control cells provides a means of identifying functional roles of exchange activity in relation to other cellular mechanisms for intracellular Ca 2ϩ handling. The results of this study indicate that Ca 2ϩ efflux via the Na ϩ -Ca 2ϩ exchanger limits the rise in [Ca 2ϩ ] i during sustained Ca 2ϩ entry, suggesting a potential modulatory function for Na ϩ -Ca 2ϩ exchange in the Ca 2ϩ signaling process. Additional evidence suggests that the exchanger itself undergoes a regulatory activation during Ca 2ϩ release from intracellular stores. EXPERIMENTAL PROCEDURES Cells CK1.4 cells were prepared by transfection of dhfr Ϫ1 Chinese hamster ovary cells with a mammalian expression vector (pcDNA-NEO; Invitrogen) containing a cDNA insert coding for the bovine cardiac Na ϩ -Ca 2ϩ exchanger (8). Control cells were prepared by transfection of the CHO cells with the vector alone (i.e. no insert). The cells were grown in Iscove's modified Dulbecco's medium containing 10% fetal calf serum and 500 g/ml geneticin (G418) as described (8). Unless specified otherwise, biochemicals were obtained from Sigma.

Fura-2 Assays
Cells grown in 75-cm 2 plastic culture flasks were washed three times with Na-PSS, which contains (in mM) 140 NaCl, 5 KCl, 1 MgCl 2 , 1 CaCl 2 , 10 glucose, and 20 mM Mops, buffered to pH 7.4 (37°C) with Tris. PSS prepared by substituting 140 mM NaCl with 140 mM LiCl is designated as Li-PSS, and nominally Ca 2ϩ -free PSS refers to PSS in which 1 mM CaCl 2 has been omitted. The cells were released from the flask by the addition of 5 mM EDTA to Ca 2ϩ -free Na-PSS, centrifuged, and resuspended twice in Na-PSS. Aliquots (300 l) of cells were loaded for 30 min in Na-PSS containing 3 M fura-2 AM (Molecular Probes) and 0.25 mM sulfinpyrazone (to retard transport of fura-2 out of the cells; Ref. 8). After loading, the cells were centrifuged and resuspended in 100 l of Ca 2ϩ -free Na-PSS containing 0.3 mM EGTA and, as indicated, either 10 M ionomycin or 200 nM thapsigargin. After 1 min, the cells were centrifuged, resuspended in the desired medium (specified in the individual experiments), and placed in a fluorescence cuvette; 0.25 mM sulfinpyrazone was included in the cuvette solutions. Occasional variations in this protocol are designated in individual experiments. All experiments were conducted at 37°C. Fluorescence was monitored at 510 nm emission with excitation at 340 and 380 nm in a Photon Technology International RF-M 2001 fluorometer. Calibration of the 340/380 ratios was accomplished by adding 10 M digitonin to the cuvettes in the presence of excess EGTA or Ca 2ϩ for determination of R min and R max , using the formula of Grynkiewicz et al. (10) to calculate the corresponding values of [Ca 2ϩ ] i . No differences in calibration between CK1.4 and control cells were observed. All fluorescence values were corrected for autofluorescence determined for each set of experiments using unloaded cells. For experiments involving Ba 2ϩ entry, the excitation wavelengths were 350 and 390 nm (11); calibrations were not conducted for the Ba 2ϩ experiments. The results are presented for multiple experiments as mean values Ϯ S.E. (error bars shown in figures). Significance testing was carried out using Student's t test (two-tailed) for unpaired samples.

Ca Uptake
Cells were grown in 24-well plastic dishes. The medium was replaced with 1 ml of nominally Ca 2ϩ -free Na-PSS, and the cells were preincubated for 30 min at 37°C. Midway through the preincubation period, Tg (50 nM) or vehicle (0.1% dimethyl sulfoxide) was added to the desired wells. The preincubation medium was replaced with the 200 l of assay medium (either Na-PSS or NMDG-PSS, pH 8.0, containing 1 mM 45 CaCl 2 ), and after the desired intervals, the wells were washed 4 times with 1 ml of termination medium (100 mM MgCl 2 ϩ 10 mM LaCl 3 ϩ 5 mM Mops/Tris, pH 7.4). The contents of the wells were extracted with 0.1 N HNO 3 and counted.

Protocols
Refilling of Ca 2ϩ Stores-Ca 2ϩ stores were depleted by treating 0.1 ml of fura-2-loaded cells for 1 min with 10 M ionomycin in Ca 2ϩ -free Na-PSS containing 0.3 mM EGTA (37°C). The cells were centrifuged and resuspended in nominally Ca 2ϩ -free PSS (with substitutions for Na ϩ as indicated) containing 0.3% fatty acid free bovine serum albumin to scavenge residual ionomycin. After 30 s, 1 mM CaCl 2 was added to the medium, and [Ca 2ϩ ] i was monitored. After the desired interval, 3 mM EGTA and, 10 or 30 s later, 0.3 mM ATP were added to bring about Ca 2ϩ release from the InsP 3 -sensitive stores (8). The peak of the Ca 2ϩ transient was taken as a measure of the amount of Ca 2ϩ in the InsP 3sensitive pool.
Tg Treatment-Cell suspensions (0.1 ml) were loaded with fura-2 and treated with 200 nM Tg in Ca 2ϩ -free Na-PSS containing 0.3 mM EGTA for 1 min at 37°C. The cells were centrifuged and suspended in nominally Ca 2ϩ -free PSS containing 0.3 mM EGTA (with or without Na ϩ substitutes as indicated) in the cuvette. CaCl 2 (1 mM) was added as indicated, and [Ca 2ϩ ] i was monitored.

Refilling of Intracellular Ca 2ϩ
Stores-Intracellular Ca 2ϩ stores in fura-2-loaded cells were depleted by treatment with ionomycin in a Ca 2ϩ -free medium. The cells were then reexposed to Ca 2ϩ , and the state of filling of the InsP 3 -sensitive Ca 2ϩ stores was assessed after the desired intervals by the addition of 3 mM EGTA followed by 0.3 mM ATP. ATP elicits the formation of InsP 3 in these cells (8,12) through its interaction with a P 2U purinergic receptor (12). As shown in Fig. 1 Fig. 1, CaCl 2 (1 mM) was added at 30 s (arrows), and then 3 mM EGTA was added after various intervals, followed by 0.3 mM ATP 30 s later. For the control cells (right panel, Fig. 1), [Ca 2ϩ ] i increased markedly upon the addition of 1 mM CaCl 2 , and the stores refilled rapidly with Ca 2ϩ , attaining a maximal level within 2.5 min. For the CK1.4 cells (left panel, Fig. 1), the increase in [Ca 2ϩ ] i upon the addition of 1 mM CaCl 2 was smaller than for the control cells, and the rate of store refilling was slower, requiring 5 min of exposure to Ca 2ϩ to achieve a maximal level.
For the CK1.4 cells, the rate of store refilling was markedly increased in a Na ϩ -free medium (Li substitution). As shown in Fig. 2 (left panels), the increase in [Ca 2ϩ ] i upon the addition of CaCl 2 in Li-PSS was larger than in Na-PSS, and the [Ca 2ϩ ] i transients elicited by ATP were markedly increased; in contrast, there was no significant difference between the sodiumand lithium-based media for the control cells (right panels, control cells in Na-PSS (150 s) was significant (p Ͻ 0.01); other comparisons showed no significant differences. The larger size of the [Ca 2ϩ ] i transient for the CK1.4 cells in the Na ϩ -free medium was primarily a reflection of the increased amount of Ca 2ϩ in the InsP 3 -sensitive stores. As shown below (see Fig. 8), the ATP-evoked [Ca 2ϩ ] i transient in cells containing equivalent amounts of stored Ca 2ϩ was smaller in Na-PSS than in Li-PSS, but the effect of Na ϩ was small compared with that shown in Fig. 2. The results in Figs. 1 and 2 indicate that extracellular Na ϩ reduces the rate of store refilling in CHO cells expressing the Na ϩ -Ca 2ϩ exchanger.
Tg-induced Ca 2ϩ Entry in CK1.4 Cells-Another method of depleting intracellular Ca 2ϩ stores is by treating the cells with Tg, an irreversible inhibitor of sarco(endo)plasmic reticulum Ca 2ϩ ATPases (13, 14). In cells pretreated with 200 nM Tg, ATP does not induce a [Ca 2ϩ ] i transient either before or after the addition of extracellular Ca 2ϩ (data not shown). This indicates that the InsP 3 -sensitive Ca 2ϩ stores were fully depleted by the Tg treatment and that refilling of the stores does not occur, as expected from the blockade of sarco(endo)plasmic reticulum Ca 2ϩ ATPase activity. As in other cell types, Tg treatment increases the influx of Mn 2ϩ , a Ca 2ϩ surrogate, as measured by the decline in fluorescence due to quenching of fura-2 by intracellular Mn 2ϩ . As shown in Fig. 3, Mn 2ϩ entry was stimulated by Tg to the same extent in vector-transfected and CK1.4 cells, indicating that intracellular store depletion is equally effective in activating store-operated Ca 2ϩ channels in the two types of cells. No difference was observed in the rate of Mn 2ϩ entry between Na-and Li-PSS for either type of cell (data not shown). Mn 2ϩ is not transported by the Na ϩ -Ca 2ϩ exchanger in these cells (data not shown).
The right panel of Fig. 4 shows that the addition of 1 mM CaCl 2 to Tg-treated control cells produced a marked increase in [Ca 2ϩ ] i , which was essentially identical in Na-PSS or Li-PSS. In many types of cells, Ca 2ϩ entry via the SDCI pathway is blocked by SK&F 96365 (15,16). As shown in the right panel of Thus, in the CK1.4 cells a substantial fraction of Ca 2ϩ enters the cell via a pathway that is insensitive to inhibition by SK&F 96365; because this pathway is absent in the control cells, it probably reflects Ca 2ϩ influx via reverse Na ϩ -Ca 2ϩ exchange. Incubation of the Tg-treated cells under Na-free conditions for 5 min prior to adding CaCl 2 reduced the increase in [Ca 2ϩ ] i and increased its sensitivity to SK&F 96365, consistent with a reduced [Na ϩ ] i and decreased contribution of Ca 2ϩ influx via the exchanger (data not shown). Note also that the decline in [Ca 2ϩ ] i after addition of EGTA was slowed by the presence of SK&F 96365 (left panel, Fig. 6), consistent with previous reports of an inhibitory effect of this agent on Ca 2ϩ efflux (16,17).
Ca 2ϩ Efflux from CK1.4 Cells-The results presented above suggest that Na ϩ o -dependent Ca 2ϩ efflux via the Na ϩ -Ca 2ϩ exchanger is responsible for the attenuation of Tg-induced Ca 2ϩ entry in the presence of Na ϩ . To assess the magnitude of Na ϩ o -dependent Ca 2ϩ efflux from CK1.4 cells, we adopted the following protocol. After loading with fura-2, [Ca 2ϩ ] i was elevated by preincubating a concentrated suspension of the cells for 1 min with Li-PSS containing 200 nM Tg and 1 mM CaCl 2 . The cell suspension was then diluted 30-fold into a cuvette containing 0.3 mM EGTA in either Li-or Na-PSS, and the decline in [Ca 2ϩ ] i was monitored. As shown in Fig. 7, the rate of decline of [Ca 2ϩ ] i was more rapid in Na-PSS than in Li-PSS, consistent with a contribution of Na ϩ -Ca 2ϩ exchange to Ca 2ϩ efflux. We cannot be certain that the decline in [Ca 2ϩ ] i under these conditions is entirely due to Ca 2ϩ efflux from the cell because Ca 2ϩ sequestration by Tg-resistant compartments could also play a role. However, when control cells were used instead of the CK1.4 cells in similar experiments, there was no difference between Na-or Li-PSS, and the rate of decline in [Ca 2ϩ ] i was similar to that observed in Li-PSS for the CK1.4 cells (data not shown). Thus, the Na ϩ -dependent component of the decline in [Ca 2ϩ ] i probably represents Ca 2ϩ efflux via the Na ϩ -Ca 2ϩ exchanger.
The data in the inset of Fig. 7 depict the decline in [Ca 2ϩ ] i during the first 40 s of the experiment, expressed as a first order plot. Nonlinear first order plots were observed in both Na-and Li-PSS. At equivalent concentrations of cytosolic Ca 2ϩ , the rate of decline in [Ca 2ϩ ] i was approximately 2.5-fold greater in Na-PSS than in Li-PSS throughout the concentration range examined.

cells (left panel) or control cells (right panel)
were treated with Tg (filled symbols) in Ca 2ϩ -free Na-PSS and assayed for 45 Ca 2ϩ uptake in either Na-PSS (squares) or NMDG-PSS (circles). The unfilled symbols represent data for cells not treated with Tg. The transport assays were conducted at pH 8.0. The results are the means of two separate experiments for each cell type. ference between Na-and Li-PSS was observed in comparable experiments (cf. Fig. 9 for data with ATP). The results confirm the importance of Na ϩ o -dependent Ca 2ϩ efflux in regulating [Ca 2ϩ ] i in the CK1.4 cells.
The effects of ATP in the presence extracellular Ca 2ϩ are compared for vector-transfected control cells and the CK1.4 cells in Fig. 9. The cuvette solutions in these experiments initially contained 0.3 mM EGTA and PSS with 140 mM NaCl, 140 mM LiCl, or 40 mM NaCl ϩ 100 mM KCl as the principal salts. The effects of adding 0.3 mM ATP plus 1 mM CaCl 2 (upper traces, Fig. 9), ATP alone (center traces, Fig. 9), or CaCl 2 alone (lower traces, Fig. 9) were examined for each cell type. For the control cells (right panels, Fig. 9), the addition of ATP alone produced a transient rise in [Ca 2ϩ ] i , whereas the addition of ATP ϩ 1 mM CaCl 2 elicited a sustained increase in [Ca 2ϩ ] i , which was higher than that evoked by the addition of Ca 2ϩ alone. There were no major differences in the behavior of the control cells among the various media, although there was a tendency for the sustained phase of Ca 2ϩ entry to be reduced in the 40/100 sodium/potassium medium (trace b, upper right panel, Fig. 9); the latter effect is probably due to the reduced driving force for Ca 2ϩ entry in cells depolarized by the high potassium concentration.
For the CK1.4 cells (left panel, Fig. 9), dramatic differences were observed among the different media when ATP and Ca 2ϩ were added together. In 140 mM Na ϩ (trace c; upper left panel, Fig. 9), the initial [Ca 2ϩ ] i transient was followed by a sustained phase that was slightly but significantly reduced (p Ͻ 0.05) compared with that observed in the control cells. The levels of [Ca 2ϩ ] i attained in the Na-free or 40 mM Na ϩ media (traces a and b; upper left panel, Fig. 9) were much higher than those produced by the addition of Ca 2ϩ alone (lower left panel, Fig. 9) and were markedly elevated compared with the sustained [Ca 2ϩ ] i levels in the vector-transfected control cells, particularly in the case of Li-PSS. Ca 2ϩ entry was similarly enhanced when NMDG was used as the Na ϩ substitute (data not shown). As in the case of Tg-treated cells, the sustained increases in [Ca 2ϩ ] i in the low [Na ϩ ] media were only partially inhibited by SK&F 96365 (data not shown), suggesting that a portion of the Ca 2ϩ entry under these conditions was conducted by the Na ϩ -Ca 2ϩ exchange system. Intracellular Ca 2ϩ stores refilled rapidly after ATP addition (assessed by ionomycin-induced Ca 2ϩ release; data not shown) indicating that Ca 2ϩ sequestration was not blocked under these experimental conditions. The pronounced increase in [Ca 2ϩ ] i in the low [Na ϩ ] media suggests that exchange activity might have been accelerated during the release of Ca 2ϩ from InsP 3 -sensitive stores. This possibility is Barium Influx in Tg-treated CK1.4 Cells-Ba 2ϩ is transported by the cardiac-type Na ϩ -Ca 2ϩ exchanger (18) and reportedly enters several types of cells via the SDCI pathway (11, 20 -22), although in either pathway it is less effectively transported than Ca 2ϩ . Importantly, however, Ba 2ϩ is not sequestered by intracellular organelles such as the endoplasmic reticulum or the mitochondria (20,23). Experiments to be reported elsewhere confirm the exchanger's ability to transport Ba 2ϩ and the absence of organellar Ba 2ϩ sequestration in the CK1.4 cells. 2 Because Ba 2ϩ is not accumulated by intracellular organelles, the fura-2 signal should provide a relatively direct assessment of Ba 2ϩ influx.
The data in Fig. 10 (right panel) show the effects of adding 1 mM BaCl 2 to control cells with or without prior treatment with Tg. Ba 2ϩ entry produces an initial abrupt rise in the 350/390 ratio in these experiments, which is probably due to small amounts of extracellular fura-2. The subsequent gradual rise in the 350/390 ratio reflects Ba 2ϩ entry into the cells and is only slightly enhanced in Tg-treated cells (trace a, Fig. 10) compared with untreated cells (trace b, Fig. 10). No difference between Li-PSS and Na-PSS was observed, and so the results with both media have been combined in each trace. Ba 2ϩ influx therefore occurs only weakly through the SDCI pathway in these cells. Note that after the addition of EGTA at 180 s, only a small decline in the fura-2 signal was observed, indicating that Ba 2ϩ is not readily transported out of the cell under these conditions. In CK1.4 cells (left panel, Fig. 10), Ba 2ϩ entry in Li-PSS is substantially increased by Tg (trace a, Fig. 10) compared with untreated cells (trace b, Fig. 10); the fura-2 ratios attained in trace a (Fig. 10) are significantly higher than observed in the vector-transfected control cells. SK&F 96365 (50 M) had no effect on Ba 2ϩ entry in the CK1.4 cells (data not shown), consistent with the absence of significant barium entry via storeoperated channels. In Na-PSS, Ba 2ϩ influx (trace c, Fig. 10) was only slightly less than in the nontreated cells in Li-PSS Fig. 10); no difference was observed between Tgtreated and untreated cells in Na-PSS, and both conditions have been combined in trace c (Fig. 10). Because a significant acceleration of Ba 2ϩ influx is observed only in Tg-treated CK1.4 cells under Na ϩ -free conditions, we conclude that Tg treatment accelerates Na ϩ -Ca 2ϩ exchange activity.
In the experiments described above, the cells had been pretreated with Tg in Ca 2ϩ -free Na-PSS, and it is conceivable that increased Na ϩ entry might be responsible for the subsequent acceleration of Ba 2ϩ entry by the Na ϩ -Ca 2ϩ exchanger. Therefore, we determined whether Tg could accelerate Ba 2ϩ entry when added to the cells in a Na ϩ -free medium. CK1.4 cells were loaded with fura-2 and placed in a cuvette containing either Ca 2ϩ -free Na-PSS or Li-PSS and 0.3 mM EGTA. After 30 s, either 200 nM Tg or vehicle (dimethyl sulfoxide) was added to the cuvette, and 1 mM BaCl 2 was added 2 min later. As shown in Fig. 11, in the absence of Tg treatment, there was little or no difference between Na-or Li-PSS for barium entry in the CK1.4 cells (A) or the control cells (C). For both types of cells (Fig. 11,  B and D), the addition of Tg resulted in a slowly developing transient rise in the 350/390 fluorescence ratio, which undoubtedly reflects the gradual release of Ca 2ϩ from internal stores. (The choice of 350/390 excitation wavelengths optimizes the Ba 2ϩ signal but still allows detection of increases in [Ca 2ϩ ] i .) For the Tg-treated control cells (Fig. 11D), no difference was observed in Ba 2ϩ entry between the Na-PSS and the Li-PSS; the ratios were not significantly higher than those observed in the absence of Tg, again indicating that Ba 2ϩ does not enter these cells efficiently by the SDCI pathway. However, the Tgtreated CK1.4 cells (Fig. 11B)  wise identical experiments, similar results were obtained, i.e. ionomycin stimulated Ba 2ϩ uptake in Li-PSS but not in Na-PSS (data not shown). DISCUSSION The expression of the cardiac Na ϩ -Ca 2ϩ exchanger in a cell type that does not normally exhibit this activity confers several new attributes to cellular Ca 2ϩ homeostasis. In the CK1.4 cells, physiological concentrations of extracellular Na ϩ retard store refilling and attenuate Tg-induced Ca 2ϩ entry compared with control cells. Upon the addition of Ca 2ϩ o to Tg-treated cells, the initial rise in [Ca 2ϩ ] i is identical in CK1.4 cells and in control cells until [Ca 2ϩ ] i reaches 100 nM, when the increase slackens markedly in the CK1.4 cells (inset, Fig. 4). Na ϩ o -dependent Ca 2ϩ efflux provides the simplest explanation for these results. We suggest that a portion of the Ca 2ϩ entering the cell through store-operated Ca 2ϩ channels is transported back out of the cell by the exchanger, thereby reducing net Ca 2ϩ entry and attenuating the rise in [Ca 2ϩ ] i during SDCI. In this view, the exchanger generates circulatory movements of Ca 2ϩ across the plasma membrane during Ca 2ϩ channel activity; this circulation could play an important role in Ca 2ϩ signaling processes, as suggested by Alkon and Rasmussen (27).
The effects of Na ϩ o in stimulating Ca 2ϩ efflux (Fig. 7) and reducing the [Ca 2ϩ ] i transients elicited by ATP or ionomycin (Fig. 8) Table I). Under these conditions, Ca 2ϩ efflux via the exchanger is blocked and Ca 2ϩ enters the cell both by SK&F-sensitive channels (Fig. 6) and by "reverse mode" Na ϩ -Ca 2ϩ exchange (Na ϩ i -dependent Ca 2ϩ influx), a process that is insensitive to the SK&F compound. An unexpected feature of our results is that exchange activity is accelerated during Ca 2ϩ release from internal stores. In Na-free or low [Na ϩ ] media, the plateau levels of [Ca 2ϩ ] i following the simultaneous addition of ATP and Ca 2ϩ o to CK1.4 cells were greatly enhanced compared with either the addition of Ca 2ϩ alone or the responses of the control cells (Fig. 9). Although these observations are consistent with an increase in exchange activity, alterations in other Ca 2ϩ handling pathways could also contribute to the results.
The Ba 2ϩ experiments provide firm support for activation of the exchanger during Ca 2ϩ release. The fura-2 signal for Ba 2ϩ provides a better index of divalent cation influx than that for Ca 2ϩ because Ba 2ϩ is not sequestered by intracellular organelles, and its cytosolic concentration should therefore be unaffected by blockade of sarco(endo)plasmic reticulum Ca 2ϩ ATPase activity. Moreover, Ba 2ϩ entry via store operated Ca 2ϩ channels appears to be minimal in CHO cells (Figs. 10 and 11), allowing a better assessment of alterations in exchange activity. With CK1.4 cells, Tg differentially stimulates Ba 2ϩ entry in Li-PSS compared with Na-PSS but has no such effect with the control cells (Figs. 10, 11). Similar results were obtained using ionomycin to release Ca 2ϩ from internal stores (data not shown). Recent experiments indicate that ATP also evokes a transient increase in exchange-mediated Ba 2ϩ influx in the CK1.4 cells. 3 It is important to distinguish between regulatory activation of exchange activity and increased activity that is simply due to a rise in the concentration of Na ϩ or Ca 2ϩ as a transport substrate. Although increased Ca 2ϩ efflux via the exchanger following organellar Ca 2ϩ release has been reported on numerous occasions, to our knowledge a linkage between Ca 2ϩ release and regulatory activation of exchange activity has not been shown previously in any cell type. In our experiments, enhanced exchange activity was measured as Na ϩ i -dependent Ba 2ϩ influx and could be observed in a Na ϩ -free medium (Fig.  11). We therefore conclude that exchange activity is activated by a regulatory process during Ca 2ϩ release from internal stores.
The mechanism of activation is uncertain. Two modes of regulation of the cardiac exchanger have been described: ATPdependent regulation and secondary Ca 2ϩ activation (28 -31). A phosphorylation mechanism has been suggested for ATP-dependent activation of exchange activity in squid giant axons (28), but experiments with sarcolemmal membrane patches (29 -31) and our previously published results with CK1.4 cells (9) raise doubts as to the relevance of this mechanism for the cardiac exchanger. Moreover, preliminary experiments indicate that the Tg-induced increase in Ba 2ϩ entry is not inhibited by the nonspecific protein kinase inhibitor staurosporine (1 M; data not shown). Thus, it seems unlikely that exchange activity is accelerated by a protein kinase-dependent mechanism. Other suggested modalities of ATP-dependent regulation, such as aminophospholipid translocase activity (30) or cytoskeletal alterations (9), have not yet been tested.
Secondary activation by Ca 2ϩ would seem to be a plausible mechanism, because acceleration of exchange activity is linked to Ca 2ϩ release from internal stores. This mode of regulation involves the interaction of Ca 2ϩ with regulatory sites distinct from the transport sites for Ca 2ϩ that lie within the central hydrophilic domain of the exchanger. The molecular identity of these sites has recently been described by Philipson and his colleagues (32,33). Secondary Ca 2ϩ activation has been extensively studied in squid giant axons (28,34), barnacle muscle (35), myocardial cells (36), and cardiac sarcolemmal membrane patches (29,31,37), but its physiological importance is not well understood.
The data in Fig. 9 (ATP) and Fig. 11 (Tg) are consistent with secondary Ca 2ϩ activation as the mechanism accelerating exchange activity, because in both cases the increased Ca 2ϩ or Ba 2ϩ influx was associated with an elevation in [Ca 2ϩ ] i . However, under conditions where the cells were pretreated with Tg, there does not appear to be an elevation of [Ca 2ϩ ] i compared with untreated cells prior to adding extracellular Ca 2ϩ (Figs. 4 and 6) or Ba 2ϩ (Fig. 10). Thus, it appears that accelerated exchange activity does not necessarily correlate with increased cytosolic Ca 2ϩ . Experiments currently under way suggest that Ca 2ϩ i -dependent changes in exchange activity involve complex interactions between cytosolic Ca 2ϩ and the Ca 2ϩ content of internal stores. Additional studies will be required to resolve these issues.
Regardless of the precise mechanism(s) involved, our findings indicate that Ca 2ϩ release from intracellular stores is coupled to regulatory activation of Na ϩ -Ca 2ϩ exchange activity. The exchanger is potentially a focal point for a variety of regulatory influences, as suggested by reports that the sensitivity of the exchanger to secondary activation by Ca 2ϩ can itself be modulated by an ATP-dependent mechanism (19).
Activation of exchange activity during Ca 2ϩ release could therefore provide an adjustable negative feedback mechanism for controlling the amount of Ca 2ϩ that is resequestered by the sarco(endo)plasmic reticulum and, as demonstrated in this report, for limiting net Ca 2ϩ entry into the cell during Ca 2ϩ channel activity. Na ϩ -Ca 2ϩ exchange activity thus provides a potentially rich source of regulatory control for cellular Ca 2ϩ traffic.