Allosteric regulation by cytosolic Ca2+ of Na+/Ca2+ exchange activity in the Ca2+ efflux mode has received little attention because it has been technically difficult to distinguish between the roles of Ca2+ as allosteric activator and transport substrate. In this study, we used transfected Chinese hamster ovary cells to compare the Ca2+ efflux activities in nontransfected cells and in cells expressing either the wild-type exchanger or a mutant, Δ(241–680), that operates constitutively; i.e., its activity does not require allosteric Ca2+ activation. Expression of the wild-type exchanger did not significantly lower the cytosolic Ca2+ concentration ([Ca2+]i) compared with nontransfected cells. During Ca2+ entry through store-operated Ca2+ channels, Ca2+ efflux by the wild-type exchanger became evident only after [Ca2+]i approached 100–200 nM. A subsequent decline in [Ca2+]i was observed, suggesting that the activation process was time dependent. In contrast, Ca2+ efflux activity was evident under all experimental conditions in cells expressing the constitutive exchanger mutant. After transient exposure to elevated [Ca2+]i, the wild-type exchanger behaved similarly to the constitutive mutant for tens of seconds after [Ca2+]i had returned to resting levels. We conclude that Ca2+ efflux activity by the wild-type exchanger is allosterically activated by Ca2+, perhaps in a time-dependent manner, and that the activated state is briefly retained after the return of [Ca2+]i to resting levels.
- persistent calcium activation
- store-operated channels
- calcium transient
cytosolic calcium is an allosteric regulator of the cardiac Na+/Ca2+ exchanger (NCX1.1) (2, 5). This mode of regulation involves the interaction of Ca2+ with two well-defined binding sites within the exchanger's central hydrophilic domain (15). Most studies indicate that these regulatory sites must be filled with Ca2+ for the exchanger to work in any of its known modes of operation: Na+/Ca2+, Na+/Na+, or Ca2+/Ca2+ exchange. Experiments with excised patches from cardiac myocytes, or frog oocytes expressing NCX1.1, indicate that the half-maximal Ca2+ concentration (Kh) for the regulatory activation of exchange activity is 200–600 nM (3, 7, 14, 15). In contrast, measurements of exchange activity in intact cells, including cardiac myocytes, have often yielded lower values (22–125 nM) (6, 17, 18, 22).
Recently, we presented evidence that Kh for NCX1.1 in transfected Chinese hamster ovary (CHO) cells is ∼300 nM (19). This evidence was produced by analysis of exchange activity operating in the reverse, Ca2+ influx mode. We also demonstrated in this study that after activation of the exchanger with a high concentration of cytosolic Ca2+ ([Ca2+]i), the activated state of the exchanger was retained for several tens of seconds after the return of [Ca2+]i to values well below Kh for activation. We called this phenomenon “persistent Ca2+ activation” and presented evidence that it was due in part to the persistence of elevated Ca2+ concentration beneath the plasma membrane after the return of bulk [Ca2+]i to resting values. In this study, we extend our analysis of allosteric Ca2+ activation to the operation of the exchanger in the Ca2+ efflux mode.
For practical reasons, allosteric Ca2+ activation has been studied most often in the “reverse” mode of exchanger operation, i.e., Ca2+ entry coupled to Na+ efflux. In the “forward” mode, i.e., Ca2+ efflux coupled to Na+ entry, it is experimentally difficult to distinguish the effects of Ca2+ acting as an allosteric activator from the effects of Ca2+ acting as a transport substrate, especially because allosteric Ca2+ activation occurs at concentrations substantially lower than the Km for Ca2+ at its translocation sites (∼5 μM) (16). Given the dual functions of cytosolic Ca2+ during Ca2+ efflux, it is not surprising that there is little information regarding allosteric Ca2+ activation of the exchanger operating in the forward mode. Only two studies have addressed the issue. In the first study, Matsuoka et al. (15) reported that a mutant with a greatly elevated Kh for allosteric Ca2+ activation also displayed an appropriately elevated Km for Ca2+ efflux (48 μM), indicating that in the mutant, the rate of Ca2+ efflux was limited by the concentration dependence for allosteric Ca2+ activation. In the second study, Kappl and Hartung (11) used flash photolysis to produce a rapid jump in Ca2+ concentration at the cytosolic surface of excised patches and reported that steady-state inward exchange currents developed with a time constant of ∼0.6 s, presumably reflecting the rate of allosteric Ca2+ activation under the conditions in their experiments. Other investigators in studies of reverse-mode exchange activity have been unable to resolve the time course for allosteric Ca2+ activation, suggesting that activation of this mode of exchange activity is virtually instantaneous and occurs in <50–200 ms (9, 22).
In this study, we used transfected CHO cells to compare the Ca2+ efflux activities of the wild-type exchanger and a constitutive mutant, Δ(241–680), that is missing much of the exchanger's central regulatory domain and does not require allosteric Ca2+ activation (16). Our results indicate that the wild-type exchanger, in contrast to the constitutive mutant, showed little Ca2+ efflux activity at resting values of [Ca2+]i, consistent with our recent report (19) that the Kh for allosteric Ca2+ activation was ∼300 nM in these cells. We found that Ca2+ efflux by wild-type NCX1.1 became activated as [Ca2+]i approached 100–200 nM, consistent with partial activation of the exchanger under these conditions. Finally, allosteric activation of Ca2+ efflux activity was shown to exhibit persistent Ca2+ activation (19); i.e., after transient exposure to [Ca2+]i levels sufficient to fully activate exchange activity, Ca2+ efflux by wild-type NCX1.1 remained activated for a brief interval after [Ca2+]i returned to resting levels.
CHO T cells [CHO K1 cells expressing the human insulin receptor (12); kindly provided by M. Czech, Univ. of Massachusetts Medical School, Worcester, MA] were transfected with the mammalian expression vector pcDNA3 containing the coding sequence for bovine NCX1.1 (GenBank accession no. LO6438). Cells expressing exchange activity were selected using the ionomycin treatment of Iwamoto et al. (10). Cells expressing the Δ(241–680) deletion mutant were prepared similarly. In this mutant, a large part of the exchanger's central hydrophilic domain is deleted; the Δ(241–680) mutant shows high activity in the absence of allosteric Ca2+ activation (6, 16). Cells were grown in a 5% CO2 atmosphere in F-12 medium supplemented with 10% fetal bovine serum, 2 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, and 20 μg/ml gentamicin. The cell culture medium, including the serum, was obtained from Life Technologies.
Na-physiological salt solution (PSS) contained 140 mM NaCl plus 5 mM KCl, K-PSS contained 140 mM KCl, and Na/K-PSS contained 20 mM NaCl plus 120 mM KCl (20/120 Na/K-PSS). All solutions also contained (in mM) 1 MgCl2, 10 glucose, and 20 MOPS and were buffered to pH 7.4 with Tris at either 37°C or room temperature as appropriate for individual experiments. Biochemicals were purchased from Sigma-Aldrich.
Fura 2 imaging.
Mixtures (∼50:50) of parental CHO cells and transfected CHO cells expressing either the wild-type NCX1.1 or the Δ(241–680) mutant were grown on circular 25-mm coverslips. This approach was chosen to ensure that both types of cells were treated and calibrated identically. The cells were loaded with fura 2 by incubating them for 30 or 40 min at room temperature in Na-PSS containing 1 mM CaCl2, 1% fatty acid free bovine serum albumin, 0.25 mM sulfinpyrazone (to retard fura 2 transport from the cells), and 3 μM fura 2-AM (Molecular Probes). The coverslips were then washed in Na-PSS plus 1 mM CaCl2, placed in a stainless steel holder (∼0.8-ml bath volume; Molecular Probes), and viewed under a Zeiss Axiovert 100 microscope coupled to an Attofluor digital imaging system. Alternating excitation at 334 and 380 nm was obtained using appropriate filters, and fluorescence was monitored with a long-pass filter set for 510 nm. In most cases, 50–70 cells were individually monitored on each coverslip. Data points were taken at ∼1.7-s intervals. At the end of each experiment, the two types of cells were identified by evoking the reverse mode of exchange activity: the cells were treated with gramicidin in the presence of 20/120 Na/K-PSS plus 0.3 mM EGTA, and 1 mM CaCl2 in 20/120 Na/K-PSS was subsequently applied. Under these conditions, cells expressing either the wild-type or the mutant exchanger displayed a large, rapid increase in [Ca2+]i, whereas parental cells did not (4). For the data shown in Figs. 1 and 2 and Table 1, the cells were treated for 90 s with 0.1 mM N,N,N′,N′-tetrakis(2-pyridylmethyl)ethylene diamine, a cell-permeable chelator of heavy metal ions (1), in Na-PSS plus 0.3 mM EGTA before the cells were restored to Na-PSS plus 1 mM CaCl2 and the recordings were begun. This was done to minimize possible interference by endogenous heavy metal ions in the calibration procedures.
Mn2+ influx experiments.
Mixtures of cells were loaded with fura 2 as described above. The cells were treated with 100 μM ATP, a purinergic agonist, plus 2 μM thapsigargin, an inhibitor of the endoplasmic reticulum Ca2+-ATPase (13), to release Ca2+ from internal stores 10 min before beginning the recordings. MnCl2 (5 mM in Na-PSS) was added to the cells, and the rate of fura 2 fluorescence quenching at 350-nm excitation was monitored. When the average fluorescence had decreased by ∼80%, the Mn2+ was removed with Na-PSS plus 1 mM EGTA and the cells were treated with gramicidin in 20/120 Na/K-PSS to identify the cells expressing the exchanger as described above. Finally, 10 mM MnCl2 in Na-PSS was again applied to completely quench fura 2 fluorescence to determine background fluorescence. For each cell, individual background intensities were subtracted from the experimental fluorescence values, and the decay of fura 2 fluorescence for each cell was then computed. Results are presented as the average fluorescence for all the cells of each type on the coverslip.
Each cell was calibrated individually using the procedure of Grynkiewicz et al. (8) as described in detail (19). Briefly, we treated cells with 10 μM ionomycin and 50 μM EGTA-AM (Molecular Probes) and measured Rmin as the lower limit of the ratio attained after ∼30 min of incubation, averaged over 10 data points. A 1:4 mixture of 100 mM CaCl2 and K-PSS was then applied (along with an additional 10 μM ionomycin), and, after the fura signal stabilized, the average ratio over 10 data points was taken as Rmax. Finally, 10 mM MnCl2 in 1 ml of Na-PSS was applied to the cells to quench fura 2 fluorescence, and the final fluorescence intensities at 334- and 380-nm excitation, averaged over 10 data points, were taken as background fluorescence. For each cell, individual background fluorescence values were subtracted from the experimental fluorescence values at 334 and 380 nm of excitation. The corrected intensities were used for computing all ratios. Ratios were converted to [Ca2+]i using the expression of Grynkiewicz et al. (8): [Ca2+]i = KD,fura(Sf/Sb)(R − Rmin)/(Rmax − R), where Sf/Sb is the ratio of corrected fluorescence intensities at 380 nm under conditions for determination of Rmin (Sf) and Rmax (Sb). KD,fura was assumed to be 274 nM at room temperature (21) and 224 nM at 37°C (8).
Tests of significance were done using Student's two-tailed t-test for paired or unpaired samples as appropriate.
Effect of exchange activity on resting [Ca2+]i.
Our initial experiments focused on the question of whether the presence of NCX1.1 would have a measurable influence on resting [Ca2+]i levels in transfected cells. We assumed that under physiological conditions, Ca2+ efflux by the exchanger would tend to reduce resting [Ca2+]i. Our experimental design involved mixing nontransfected CHO cells with cells expressing either wild-type NCX1.1 or a deletion mutant, Δ(241–680), in which exchange activity is constitutive, i.e., not allosterically regulated by [Ca2+]i. This approach ensures that both types of cells were treated and calibrated identically and facilitates comparisons between the cell types (see experimental procedures and discussion). In the following experiments, we compared the effects of expressing the two types of exchangers on the resting [Ca2+]i and on Ca2+ efflux activity. Cells expressing the exchanger were identified by inducing reverse-mode exchange activity as described in experimental procedures.
The data in Fig. 1 compare resting values of [Ca2+]i for nontransfected CHO cells and cells expressing either the wild-type NCX1.1 or the constitutively active mutant. The results represent the average of five coverslips and display [Ca2+]i values in the presence and absence of 1 mM external CaCl2. In the presence of extracellular Ca2+, [Ca2+]i was not significantly different when the parental CHO cells (66 nM) and cells expressing NCX1.1 (64 nM) were compared. A more stringent test of significance is to compare CHO and NCX1.1 cells on the individual coverslips. As shown in Table 1, differences between CHO cells and NCX1.1-expressing cells in the presence of extracellular Ca2+ were significant for only one of the five coverslips tested. In contrast, for the cells expressing the Δ(241–680) mutant, the average resting [Ca2+]i for the pooled results from five coverslips was significantly reduced (P < 0.01) compared with nontransfected CHO cells when extracellular Ca2+ was present (Fig. 1B). This difference was highly significant for each of the five individual coverslips examined (Table 2).
In the absence of extracellular Ca2+, [Ca2+]i was slightly but significantly lower (P < 0.05) in the cells expressing NCX1.1 than in the nontransfected cells (Fig. 1A). Comparison of the two cell types on individual coverslips revealed a significant difference in the absence of extracellular Ca2+ on two of the five coverslips (Table 1). The nontransfected CHO cells showed a small decline in [Ca2+]i when extracellular Ca2+ was removed, although the difference was not significant for the five coverslips shown in Fig. 1A; however, the difference was significant for the CHO cells in Fig. 1B (P = 0.002, paired t-test). When the measurements from all 10 coverslips were combined, [Ca2+]i in the nontransfected CHO cells decreased from 61.9 ± 2.9 to 56.3 ± 2.5 nM when external Ca2+ was removed (mean ± SE; P = 0.0008, paired t-test). For the constitutive mutant, removal of extracellular Ca2+ induced a sharp reduction in [Ca2+]i, from 46 nM to 26 nM (Fig. 1B, Table 2).
We conclude that in contrast to the constitutive mutant, wild-type NCX1.1 had little if any effect on resting [Ca2+]i, probably because under resting conditions, [Ca2+]i was far below the Kh for allosteric Ca2+ activation (300 nM) (19), and the exchanger was relatively inactive under these conditions. The reduction in [Ca2+]i when extracellular Ca2+ was removed suggests that the entry of Ca2+ from the external medium had a small but significant effect in maintaining resting [Ca2+]i levels in the parental and NCX1.1-expressing cells. The influence of external Ca2+ was greater in the cells expressing Δ(241–680) mutant, presumably because continual Ca2+ entry was needed to offset the continual Ca2+ efflux by the constitutive exchanger.
Exchange activity and [Ca2+]i during Ca2+ release from internal stores.
Figure 2 compares the changes in [Ca2+]i that were observed in CHO and NCX-expressing cells when Ca2+ was released from internal stores upon application of 10 μM ionomycin, a Ca2+ ionophore, and thapsigargin (TG), an inhibitor of the Ca2+-ATPase of the endoplasmic reticulum (13). These experiments were performed in Ca2+-free Na-PSS. The [Ca2+]i transient was greatly attenuated in both amplitude and duration in cells expressing either the wild-type exchanger or the Δ(241–680) mutant. In each coverslip examined, the difference in [Ca2+]i between nontransfected and transfected cells was highly significant. Control studies (data not shown) indicate that there was little or no difference between the nontransfected and transfected cells when similar experiments were performed in the absence of extracellular Na+ (K-PSS), a condition in which Na+/Ca2+ exchange activity would be blocked. Thus the amount of releasable Ca2+ in intracellular stores was similar in both kinds of cells, and the differences noted in Fig. 2 primarily reflect Ca2+ efflux by NCX1.1.
The results in Fig. 2 suggest that the constitutive mutant may be more efficient than the wild-type NCX1.1 in clearing Ca2+ from the cytosol during ionomycin-induced Ca2+ release. However, the differences between the wild type and the mutant were significant only for the first data point after Ca2+ release was initiated (at 1.8 s in Fig. 2, B and C) and may partially reflect differences in calibration, the rate of solution exchange, or the amount of exchanger in the plasma membrane. We conclude that both types of exchangers carry out rapid Ca2+ efflux under these conditions, and we cannot discern any differences in their behavior.
Figure 2D shows the average values of [Ca2+]i >100 s after Ca2+ release (see brace in Fig. 2A). For the cells expressing NCX1.1, [Ca2+]i was 42.9 ± 3.2 nM, and for the cells expressing the constitutive mutant, [Ca2+]i was 11.6 ± 4.2 nM (n = 4); these values are significantly less than those found in the same cells in the absence of extracellular Ca2+ before Ca2+ release [56.6 ± 4.1 nM (P < 0.05) and 25.7 ± 2.1 nM (P < 0.01), paired t-test; see Fig. 2 legend]. The large standard error for the nontransfected cells in Fig. 2D reflects the fact that the [Ca2+]i transient had not completely subsided for many of these cells at the time the measurements were made. The results suggest that Ca2+ leakage from filled internal stores makes a significant contribution to maintaining [Ca2+]i in these cells.
Exchange activity during Ca2+ entry through store-operated channels.
Because [Ca2+]i increased rapidly when Ca2+ was released from internal stores, and because allosteric Ca2+ activation was also a rapid process, we could not discern any differences in the behavior of the wild-type and constitutively active exchangers. In the experiments described below, we compared the two types of exchangers under conditions in which the rise in [Ca2+]i was slower and less extensive. For this purpose, we measured Ca2+ entry through store-operated Ca2+ channels, in which the rate of Ca2+ entry could be controlled by changing the external Ca2+ concentration.
Initial control experiments were performed to determine whether the expression of Na+/Ca2+ exchange protein altered store-operated channel activity. Surprisingly, we found that channel activity was reduced in cells expressing the exchanger. In the experiments shown in Fig. 3, mixtures of nontransfected CHO cells and cells expressing either the wild-type exchanger (Fig. 3A) or the Δ(241–680) mutant (Fig. 3B) were treated with 100 μM ATP, a purinergic agonist, and 2 μM TG in Ca2+-free medium 10 min before the start of the recording. This treatment rapidly released Ca2+ from the endoplasmic reticulum and activated store-operated channels. MnCl2 (5 mM) in Na-PSS was applied to the cells, and the rate at which Mn2+ entered the cells and quenched fura 2 fluorescence was monitored. This assay has been used widely to measure store-operated channel activity (20) and has the advantage that Mn2+ is not transported by NCX1.1 (Reeves JP, unpublished observations). The experiment shown in Fig. 3A shows that Mn2+ entry into cells expressing NCX1.1 was 26% slower than that of the nontransfected CHO cells (P < 0.02). In cells expressing the Δ(241–680) mutant, the rate of Mn2+ entry was 54% less than that of the nontransfected CHO cells (P ≈ 10−9). The pooled results of five experiments with NCX1.1 and four experiments with the Δ(241–680) mutant are shown in Fig. 3C. Cells expressing NCX1.1 and Δ(241–680) showed rates of Mn2+ entry that averaged 31 ± 2.4 and 54 ± 2.6% less, respectively, than those for the nontransfected CHO cells on the same coverslips. The mechanism underlying the reduced store-operated channel activity in cells expressing exchange activity is unknown.
For the experiments shown in Figs. 4 and 5, we measured net Ca2+ entry through store-operated channels. Cell mixtures were pretreated with ATP plus TG in Ca2+-free Na-PSS 10 min before the recordings were begun. As indicated by the bars in Fig. 4A, 0.3 mM CaCl2 in Na-PSS was applied, followed by a 3-min interval in Ca2+-free Na-PSS, and then 1.0 mM CaCl2 in Na-PSS was applied. For the CHO cells, [Ca2+]i increased slowly during the application of 0.3 mM CaCl2 and much more rapidly when 1.0 mM CaCl2 was applied; in the latter case, peak [Ca2+]i values of ∼1.2 μM were attained (data not shown). For the cells expressing NCX1.1, the increase in [Ca2+]i was initially similar to that seen with the CHO cells, but the rise in [Ca2+]i stopped rather abruptly at 100 nM (0.3 mM CaCl2) or at 180 nM (1.0 mM CaCl2). In both cases, this was followed by a decline in [Ca2+]i until steady values were attained. Table 3 presents the average peak and steady-state values for the pooled results from five coverslips. The data in Fig. 4A are shown on an expanded time scale in Fig. 4, B and C. Inspection of the individual time points reveals that the rise in [Ca2+]i for the cells expressing NCX1.1 lagged slightly behind [Ca2+]i in the nontransfected cells until the peak value was attained. After application of 1 mM CaCl2, the initial rate of increase in [Ca2+]i for the NCX1.1 cells was 41% lower than that for the nontransfected cells (13 ± 4.1 and 22 ± 1.5 nM/s, respectively; P < 10−7), a value similar to the reduction in store-operated channel activity. This suggests that the wild-type exchanger initially showed little or no Ca2+ efflux activity and that Ca2+ efflux increased dramatically after the peak [Ca2+]i values were attained.
Figure 5 presents the results of an identical experiment for a coverslip containing mixtures of nontransfected CHO cells and cells expressing the Δ(241–680) mutant. In this case, [Ca2+]i values for the Δ(241–680) cells were substantially lower than those for the CHO cells at every time point. Note in Fig. 5C that the difference in [Ca2+]i between the cells with or without Δ(241–680) was much greater during the early time points after Ca2+ addition than for the comparable data with NCX1.1 in Fig. 4C. After application of 1 mM CaCl2, the rate of rise in [Ca2+]i for the cells expressing the Δ(241–680) mutant was 83% less than that for the nontransfected cells. Importantly, [Ca2+]i values for the Δ(241–680) cells showed a gradual rise toward a steady value rather than the peak and subsequent decline observed with the cells expressing the wild-type exchanger (cf. Fig. 4).
We interpret these results in the following way: the rise in [Ca2+]i was attenuated in the cells expressing the exchanger because Ca2+ entering the cells through store-operated channels was quickly transported out by the exchanger operating in the Ca2+ efflux mode. Thus the difference in [Ca2+]i between cells expressing the exchanger and the parental CHO cells provides a measure of Ca2+ efflux by the exchanger. The reduction in store-operated Ca2+ channel activity accounts for part of the difference in the initial rates of Ca2+ entry between the cells expressing the exchanger and nontransfected cells, but this difference cannot account for the very large differences in [Ca2+]i that develop during the steady state. Moreover, there is a qualitative difference in the time course of net Ca2+ entry into the NCX1.1 cells, in which [Ca2+]i rises to a peak value and then declines, compared with cells expressing the Δ(241–680) mutant, which show a gradual rise in [Ca2+]i to a steady-state value. This difference cannot be attributed to a reduced rate of Ca2+ entry through store-operated channels in cells expressing the Δ(241–680) mutant. The peak-and-decline behavior of the wild-type cells was evident at both concentrations of CaCl2 in Fig. 4 (see also Table 3), although the rate of Ca2+ entry at 0.3 mM CaCl2 was only 20% of that observed at 1.0 mM CaCl2 (Fig. 4).
We conclude that the wild-type exchanger displayed only a low rate of Ca2+ efflux until [Ca2+]i had attained values at which allosteric activation of the exchanger occurred (100–200 nM; Fig. 4). These values are somewhat below the reported Kh for exchange activity (300 nM). However, local [Ca2+]i beneath the plasma membrane may be higher than global [Ca2+]i under these conditions (see discussion). Moreover, it seems likely that only partial activation of the exchanger would be necessary to offset the relatively low rate of Ca2+ influx under these conditions. Indeed, in previously published experiments, we induced reverse-mode exchange activity in cells expressing NCX1.1 after the application of 0.3 mM CaCl2 in Na-PSS; the results provided strong evidence for partial, but not complete, activation of the exchanger under these conditions (see Fig. 1 in Ref. 19).
Persistent activation of Ca2+ efflux.
We previously reported that after the wild-type exchanger becomes activated by a transient rise in [Ca2+]i, the activated state persists for tens of seconds after the return of [Ca2+]i to resting levels (19). This conclusion was based on assays of the “reverse,” or Ca2+ influx, mode of exchange activity, and it is important to determine whether a similar phenomenon can be observed for the exchanger operating in the Ca2+ efflux mode. To this end, we applied ATP plus TG to NCX1.1 or Δ(241–680) cells to release Ca2+ from internal stores. We then applied 1 mM CaCl2 in Na-PSS immediately after [Ca2+]i had returned to resting levels, i.e., 40 s after the ATP plus TG addition. As shown in Fig. 6A for cells expressing wild-type NCX1.1, the time course of the rise in [Ca2+]i after the addition of 1 mM CaCl2 displayed the characteristics normally seen for the Δ(241–680) mutant (Fig. 6B). That is, [Ca2+]i for the NCX1.1 cells increased gradually to a steady value instead of rising to a peak and subsequently declining as described in Fig. 4. Indeed, when the cells were returned to Ca2+-free media and retested after 10 min with a second addition of 1 mM CaCl2, they showed the typical peak-and-decline behavior of cells expressing the wild-type exchanger. In control experiments, we found that the rate of Mn2+ entry 40 s after ATP plus TG addition was the same as that observed at 10 min (data not shown); thus store-operated Ca2+ channels were activated equally at the two time points.
The results for the separate additions of 1 mM CaCl2 are compared in Fig. 7, A and B, for cells expressing wild-type NCX1.1 and the Δ(241–680) mutant, respectively. Note in Fig. 7B that for cells expressing the Δ(241–680) mutant, the initial rates of net Ca2+ entry 40 s and 10 min after the application of 1 mM CaCl2 were very similar, consistent with our finding that the store-operated Ca2+ entry mechanism was activated to the same extent at both time points. We conclude that the lower rate of Ca2+ entry at 40 s vs. 10 min for the cells expressing NCX1.1 is caused entirely by an increased rate of Ca2+ efflux, reflecting the activated state of the exchanger. Figure 7C directly compares cells expressing the two types of exchanger; the time dependence of the rise in [Ca2+]i for NCX1.1 was indistinguishable from that for the Δ(241–680) mutant when CaCl2 was added 40 s after Ca2+ release was initiated with ATP plus TG. In other experiments, 1 mM CaCl2 was applied 90 and 120 s after adding ATP plus TG. As shown in Fig. 8, the time course of the rise in [Ca2+]i displayed a progressive reversion to the typical behavior of NCX1.1 cells and had completely reverted when Ca2+ was added 120 s after ATP plus TG. Thus the Ca2+-activated state was maintained for tens of seconds after the return of [Ca2+]i to resting levels.
For the reasons given in the Introduction, information regarding the allosteric regulation of NCX1.1 by Ca2+ has come primarily from studies of the exchanger operating in the reverse, Ca2+ influx mode. Previous work at our laboratory showed that activation of the reverse mode occurs with a Kh of ∼300 nM and that the activated state persists for an interval after the return of [Ca2+]i to resting values (19). The results presented here indicate that allosteric activation of the forward, Ca2+ efflux mode of exchange activity is similar in many respects to activation of the reverse mode. Thus Ca2+ efflux by the wild-type exchanger appears to be very inefficient under resting conditions but becomes very robust when [Ca2+]i increases to higher values. Moreover, the efflux mode of the exchanger also displays persistent Ca2+ activation (19) when [Ca2+]i declines to resting values after transient exposure to high [Ca2+]i.
The pooled results of our measurements of resting [Ca2+]i in CHO cells with or without expression of wild-type NCX1.1 failed to reveal a significant difference between the two cell types in the presence of extracellular Ca2+ (Fig. 1A). The protocol chosen for these experiments, in which CHO cells and exchanger-expressing cells are both present on the same coverslip, allows a rigorous comparison to be made because the experimental treatments and calibration procedures are identical for both cell types. To illustrate the pitfalls in comparing cells calibrated separately, the [Ca2+]i values for CHO cells in the absence of extracellular Ca2+ were significantly different for the two sets of coverslips [i.e., CHO cells plated with NCX1.1 cells (61.6 ± 3.3 nM) or with the Δ(241–680) cells (51.0 ± 1.7 nM, P = 0.02) (Tables 1 and 2)]. Because the parental CHO cells are the same on both coverslips, this difference is clearly fortuitous and merely reflects the imprecision of calibrations conducted separately.
Thus the most stringent comparisons between nontransfected CHO cells and cells expressing NCX1.1 involve mixtures of cells grown on the same coverslip (Table 1). In the presence of extracellular Ca2+, only one of the five coverslips tested showed a significant difference in resting [Ca2+]i, whereas two coverslips showed a significant difference in the absence of extracellular Ca2+. In contrast, every coverslip showed significantly lower [Ca2+]i for the cells expressing the Δ(241–680) mutant than for the nontransfected cells (Table 2). We conclude that expression of the wild-type NCX1.1 does not lower [Ca2+]i in CHO cells under resting conditions. On the basis of this finding in conjunction with the data obtained for the constitutive mutant, we also conclude that NCX1.1 shows little or no Ca2+ efflux activity under resting conditions, undoubtedly because resting [Ca2+]i (∼65 nM) is below the Kh for allosteric Ca2+ activation (300 nM) (19).
It is interesting to note that when extracellular Ca2+ was removed, the Δ(241–680) cells showed a sharp drop, from 46 nM to 26 nM (Fig. 1B, Table 2); the corresponding drop for the CHO cells (from 62 nM to 56 nM) and the cells expressing NCX1.1 (from 64 nM to 56 nM), although significant (see results), was less substantial. This suggests that under resting conditions, Ca2+ influx from the extracellular medium only modestly contributes to maintaining resting [Ca2+]i in the latter cells. Why does the removal of extracellular Ca2+ produce a greater fall in [Ca2+]i for the cells expressing the constitutively active mutant? It seems likely that maintaining [Ca2+]i would require greater Ca2+ influx into these cells to match the constitutive Ca2+ efflux by the mutant exchanger. The mechanism of Ca2+ influx is unclear, however. Store-operated Ca2+ channels appear to be the most likely candidate, but the amount of Ca2+ in intracellular stores is not markedly reduced for the cells expressing the Δ(241–680) mutant, and, as indicated in Fig. 3, store-operated channel activity is actually reduced by >50% in these cells compared with nontransfected CHO cells. Thus the explanation for this effect remains uncertain.
As shown in Fig. 2A, [Ca2+]i declined further for the cells expressing either the wild-type exchanger (41 nM) or the constitutive mutant exchanger (12 nM) (Fig. 1D) after the [Ca2+]i transient induced by ionomycin had subsided. The data for the parental CHO cells in Fig. 2D could not be used in this comparison because the [Ca2+]i transient for many cells had not decayed completely at the time the measurements were made; this is reflected in the large standard error for the CHO cells. However, the initial values for [Ca2+]i in Fig. 4 were obtained 10 min after Ca2+ release with ATP plus TG and show that both the CHO cells and the NCX1.1 cells had identical values of ∼25 nM for [Ca2+]i; replications of this experiment showed essentially the same results. We conclude that Ca2+ leakage from filled internal stores makes a substantial contribution to maintaining resting [Ca2+]i in each of the cell types examined.
The wild-type exchanger was quickly activated during the ionomycin-induced release of Ca2+ from internal stores, and its activity markedly reduced the amplitude and duration of the [Ca2+]i transient (Fig. 2). The rapidity and extent of the rise in [Ca2+]i under these conditions did not allow us to detect an effect of allosteric Ca2+ activation on Ca2+ efflux, so we designed experiments in which [Ca2+]i would increase at a slower rate. For this purpose, we examined Ca2+ influx through store-operated channels so that the rate and extent of Ca2+ influx could be controlled by adjusting the external [Ca2+]. Store-operated channel activity was reduced in cells expressing the wild-type exchanger (by 30%) or the Δ(241–680) mutant (by 54%; Fig. 3). While the reasons for this behavior remain unknown, the reduced channel activity cannot account for the very large reductions in [Ca2+]i in cells with exchange activity after the application of external Ca2+ or for the qualitative differences in the time course of net Ca2+ influx between the two cell types. Thus, when external Ca2+ was applied to cells expressing the constitutively active exchanger mutant, [Ca2+]i values were markedly below those of the nontransfected cells at every time point and increased gradually to a steady-state value (Fig. 5). The gradual increase in [Ca2+]i undoubtedly reflects an initial mismatch between the rates of Ca2+ influx and efflux at low [Ca2+]i. As [Ca2+]i increases, exchange activity also increased because [Ca2+]i is far below the Km for Ca2+ translocation (∼5 μM; Ref. 16). In the steady state, the rates of Ca2+ efflux and influx are equal so that net Ca2+ entry ceases.
In contrast to the behavior of the constitutive mutant, when external Ca2+ was applied to cells expressing the wild-type exchanger, [Ca2+]i values were only slightly less than those in the nontransfected cells until a threshold value was attained, whereupon the rise in [Ca2+]i abruptly ceased (Fig. 4). The threshold value was ∼110 nM in 0.3 mM CaCl2 and 180 nM in 1.0 mM CaCl2 (Table 3) and undoubtedly corresponds to the point at which sufficient activation of the exchanger occurred so that the rate of Ca2+ efflux would match (or exceed; see below) the rate of Ca2+ entry. The sudden cessation of net Ca2+ influx would be consistent with a high degree of cooperativity of the activation process. An average Hill coefficient of 2.8 for allosteric Ca2+ activation of reverse-mode activity was obtained in our previous study (19), although we also presented reasons for caution in evaluating this result. In any event, the peak value of [Ca2+]i was followed by a decline until a steady-state value of [Ca2+]i was reached 15–20 s later. The threshold values are somewhat less than the Kh for Ca2+ activation reported previously (300 nM; Ref. 19). However, it is likely that only partial activation of the exchanger would be required to counter the rather low rate of Ca2+ influx in these experiments. Another potentially important factor is that the [Ca2+] in the submembrane region during Ca2+ influx might be substantially higher than indicated by measurements of global [Ca2+]i (see below).
The steady-state value of [Ca2+]i (78 ± 6 nM; Table 3) for NCX1.1 cells during the application of 0.3 mM CaCl2 was similar to resting [Ca2+]i (64 ± 5 nM; Table 1) in these cells, yet Ca2+ efflux was obviously occurring under the former condition but not under resting conditions. This comparison suggests that the local [Ca2+] during Ca2+ influx was greater than bulk [Ca2+]i, perhaps because Ca2+ diffusion from the cortical regions to the bulk cytosol is restrained in some manner. Under resting conditions, Ca2+ entry appears to contribute very little to maintaining [Ca2+]i (see above), so the [Ca2+] beneath the plasma membrane would be expected to be nearer the bulk [Ca2+]i. This interpretation is consistent with the apparent absence of wild-type exchange activity under resting conditions.
The postpeak decline in [Ca2+]i in cells expressing wild-type NCX1.1 suggests that the allosteric activation of Ca2+ efflux is a time-dependent process. Although it is possible that store-operated channels become partially inactivated during the postpeak interval (23), the cells expressing the constitutively active mutant did not display the peak-and-decline behavior (Fig. 5), so channel inactivation is not a likely explanation for the postpeak decline seen with the NCX1.1 cells. It is possible that the results reflect time-dependent changes in local submembrane Ca2+ gradients, although it seems likely that any such gradients would be established quite rapidly during Ca2+ influx. On the other hand, if allosteric activation of Ca2+ efflux were not instantaneous but occurred with a delay, the rise in [Ca2+]i would overshoot the value that would provide enough activation of Ca2+ efflux to balance the rate of Ca2+ influx. In this case, exchange-mediated Ca2+ efflux would gradually reduce [Ca2+]i from its peak value until the rates of influx and efflux were equal (i.e., steady-state value; Table 3). Kappl and Hartung (11) reported that inward exchange currents in excised patches were activated with a time constant of 0.6 s after the release of ∼100 μM cytosolic Ca2+ by flash photolysis. The rate of activation at lower [Ca2+] was not examined but would be expected to be slower. Considering the implications of this issue for the physiological role of Na+/Ca2+ exchange in the heart, further investigations of the time and concentration dependence of allosteric Ca2+ activation of the forward mode of exchange activity are clearly warranted.
The term “persistent Ca2+ activation” refers to maintenance of the activated state of the wild-type exchanger after the return of [Ca2+]i to values well below the Kh for allosteric Ca2+ activation. It is not clear whether this phenomenon is related to the slow relaxation of the activated state observed in excised patches. In the patches, removal of cytosolic Ca2+ in the continued presence of cytosolic Na+ leads to a decay of outward exchange currents with a time constant of ∼10 s (9, 15), although much more rapid (0.5 s) decay was observed in one study in which a “macropatch” from cardiac myocytes was used (7). The presence of ATP was shown to slow still further the decay of outward exchange currents (9). The experiments with excised patches were performed in the absence of extracellular Na+ and in the presence of high concentrations (100 mM) of cytosolic Na+; little information is available concerning relaxation of Ca2+ activation in patches at lower cytosolic Na+ concentrations.
Persistent Ca2+ activation was described in our recent report on allosteric Ca2+ activation of reverse-mode exchange activity in transfected CHO cells (19), in which we argued that the maintenance of Ca2+ activation at low [Ca2+]i reflected, at least in part, the persistence of local Ca2+ gradients beneath the plasma membrane. In the present study, we have shown that persistent Ca2+ activation applies to the Ca2+ efflux mode of exchanger operation as well. Shortly after ATP-induced Ca2+ release, the cells expressing NCX1.1 behaved essentially the same as cells expressing the constitutive mutant when they were challenged with 1 mM CaCl2 (Figs. 6 and 7). That is, the rise in [Ca2+]i did not exhibit the peak-and-decline behavior typical of NCX1.1 cells but showed a gradual approach to a steady-state value typically seen in cells expressing the Δ(241–680) mutant. Thus NCX1.1 retained its state of allosteric Ca2+ activation despite the fall in [Ca2+]i to low values just before external Ca2+ was applied (∼30 nM in the experiment in Fig. 7C). The activated state was demonstrated to have relaxed during a period of tens of seconds because extending the interval between the initiation of Ca2+ release and the application of external Ca2+ resulted in the gradual restoration of wild-type behavior (Fig. 8). The results are entirely analogous to those presented previously for reverse-mode exchange activity (19).
In summary, we have shown that allosteric Ca2+ activation of NCX1.1 serves to limit its activity under resting conditions so that resting [Ca2+]i remains substantially higher than the value determined by the thermodynamic driving force for exchange activity (∼10 nM). Other factors contributing to maintaining [Ca2+]i at resting levels include Ca2+ influx across the plasma membrane and, more important, leakage of Ca2+ from internal stores, i.e., the endoplasmic reticulum. A rise in [Ca2+]i activates Ca2+ efflux activity by NCX1.1, and our results suggest that this may occur in a time-dependent manner. Finally, regulatory activation of Ca2+ efflux persists for tens of seconds after the restoration of [Ca2+]i to resting levels or below. Local Ca2+ gradients beneath the plasma membrane might be important in each of these aspects of the allosteric activation of Ca2+ efflux. Further studies are required to determine whether allosteric Ca2+ activation of the forward mode of exchange activity is indeed time dependent and to delineate the mechanism of persistent Ca2+ activation. Time-dependent regulation of exchange activity would have important physiological implications for cardiac Ca2+ homeostasis, because in this case the exchanger would monitor and regulate the integral of [Ca2+]i across multiple beats rather than beat-to-beat variations in [Ca2+]i.
This work was supported by National Heart, Lung, and Blood Institute Grant HL-49932 and American Heart Association Grant 0151201T.
We thank Larissa Bonilla for excellent technical assistance and Dr. Roman Shirokov for comments on the manuscript.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 the American Physiological Society