Kinetics of empty store-activated Ca2+ influx in HeLa cells.

The intracellular Ca2+ indicator Indo-1 was used to monitor changes in cytosolic [Ca2+] ([Ca2+]i) in single HeLa cells upon readmission of external Ca2+ after a short incubation in Ca(2+)-free solution. HeLa cells were responsive to histamine but not to caffeine, and their histamine-sensitive store was totally depleted by a 60-min exposure to 2 microM thapsigargin. The resting [Ca2+]i in thapsigargin-treated cells was higher than in control cells and low amplitude [Ca2+]i oscillations were observed in about 20% of the cells. Readmission of external Ca2+ after a brief withdrawal of extracellular Ca2+ resulted in a transient [Ca2+]i rise, which then decayed to the same elevated [Ca2+]i measured before the Ca2+ withdrawal period. The [Ca2+]i rise was associated with an increased rate of Mn2+ entry, measured as the rate of quenching of intracellular Fura-2. The same procedure did not affect the [Ca2+]i in control cells not pretreated with thapsigargin. The amplitude of this [Ca2+]i transient in thapsigargin pretreated cells depended on the duration of prior incubation in Ca(2+)-free medium. The [Ca2+]i rise induced by elevating the extracellular [Ca2+] from 1.5 to 10 mM was more pronounced if the [Ca2+]i during the initial incubation in 1.5 mM Ca2+ was first lowered by depolarizing the cells. We conclude that an empty store stimulates a Ca2+ entry pathway consisting of two components: a continuously elevated basal leak and a second component that is transient due to the high [Ca2+]i-induced inhibition of the Ca2+ entry pathway. This inhibition and the subsequent recovery from it as the [Ca2+]i is brought to resting levels could cause the oscillatory Ca2+ entry that we recorded in a fraction of the thapsigargin-treated cells.

The intracellular Ca2+ indicator Indo-1 was used to monitor changes in cytosolic [Ca2+l ([Ca2+li) in single HeLa cells upon readmission of external Ca2+ after a short incubation in Ca2+-free solution. HeLa cells were responsive to histamine but not to caffeine, and their histamine-sensitive store was totally depleted by a 60min exposure to 2 p~ thapsigargin. The resting [Ca2+li in thapsigargin-treated cells was higher than in control cells and low amplitude [Ca2+Ii oscillations were observed in about 20% of the cells. Readmission of external Ca2+ after a brief withdrawal of extracellular Ca2+ resulted in a transient [Ca2+li rise, which then decayed to the same elevated [Ca2+li measured before the Ca2+ withdrawal period. The [Ca2+li rise was associated with an increased rate of Mn2+ entry, measured as the rate of quenching of intracellular Fura-2. The same procedure did not affect the [Ca2+Ii in control cells not pretreated with thapsigargin. The amplitude of this [Ca2+li transient in thapsigargin pretreated cells depended on the duration of prior incubation in Ca2+-free medium. The [Ca2+li rise induced by elevating the extracellular [Ca2+l from 1.5 to 10 m~ was more pronounced if the [Ca2+li during the initial incubation in 1.5 m~ Ca2+ was first lowered by depolarizing the cells. We conclude that an empty store stimulates a Ca2+ entry pathway consisting of two components: a continuously elevated basal leak and a second component that is transient due to the high [Ca2+Ii-induced inhibition of the Ca2+ entry pathway. This inhibition and the subsequent recovery from it as the [Ca2+li is brought to resting levels could cause the oscillatory Ca2+ entry that we recorded in a fraction of the thapsigargin-treated cells.
Ca2+-mobilizing agonists use I~s P~~ as second messenger to release Ca2+ from internal stores (Berridge and Irvine, 1989;Berridge, 1993). The decreased state of filling of the internal Ca2+ stores is then signaled to the plasma membrane to increase its permeability for Ca2+ (Putney, 1986(Putney, , 1990. Interestingly, a similar phenomenon occurs after treatment of cells with thapsigargin, a specific inhibitor of intracellular Ca2+ pumps (Thastrup et al., 1989). The nature of the signal transfer from store to plasma membrane, as well as the exact pathway for influx, is still unknown.
* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

16-345991.
Different models for explaining the stimulation of the Ca2+ entry were proposed. They can be classified in two groups. A first group of models proposes that the rise in [Caz+li, induced by the emptying of the store, would increase the permeability of the plasma membrane for Ca2+. Luckhoff and Clapham (1992) provided electrophysiological evidence for the existence of a Ca2+-permeable plasma membrane channel that requires high [Ca2+li for activation and that is potentiated by InsP,. The Ca2+-mediated activation (by released Ca2+) of the InsP3 kinase converts InsP3 to InsP, and would open the channel (Luckhoff and Clapham, 1992). Non-selective Ca2+-permeable cation channels that are gated by the rise in [Ca2+li (Swandulla and Partridge, 1990) could similarly represent a pathway for the increased Ca2+ entry. The rise in [Ca2+Ji may in addition activate Ca2+-activated K+ channels and hyperpolarize the plasma membrane, which would increase the driving force for Ca2+ ions (Mertz et al., 1992).
Another series of models explains enhanced Ca2+ entry without relying on a rise in [Ca2+li. A Ca2+ influx through a physical link between the store and the extracellular space (Casteels and Droogmans, 1981;Merritt and Rink, 19871, possibly involving a GTP-dependent pathway (Gill et al., 1986), was proposed to explain store refilling without a concomitant rise in [Ca2+li.
Another possibility is that a plasma-membrane InsP4-binding protein, which might be a Ca2+ channel or a molecule associated with such a channel, makes direct contact with the InsP, receptor that would act as a Ca2+ sensor in the lumen of the store (Irvine, 1991). The recent demonstration of receptors with affinity for both InsP3 and InsP, in the plasma membrane (Khan et al., 1992a(Khan et al., , 1992bFujimoto et al., 19921, could be the molecular basis for patch-clamp recordings of InsP3-sensitive Ca2+ channels in the plasma membrane (Kuno and Gardner, 1987;McDonald et al., 1993). In other cells, evidence was presented favoring the existence of a messenger other than an inositol phosphate to signal the filling state of the internal store to the plasma membrane . This signal may involve a cytochrome P450-dependent mechanism (Alvarez et al., 19911, cyclic GMP (Pandol and Schoeffield-Payne, 1990;Bahnson et al., 19931, a low molecular weight phosphorylated molecule containing hydroxyls on adjacent carbons (Randriamampita and Tsien, 1993) or be coupled to a luminal Ca2+activated tyrosine phosphatase that dephosphorylates a 130-kDa protein. Ca2+ depletion of the store would favor phosphorylation of the 130-kDa protein, thereby gating a Ca2+permeable membrane channel (Vostal et al., 1991). It is known that the activation of a tyrosine kinase in mast cells phosphorylates a 110-kDa protein, which then opens a plasma-membrane Ca2+-permeable channel (Corcia et al., 1988;Hemmerich and Pecht, 1988;Hemmerich et al., 1991). Recently, Hoth and Penner (1992) have reported electrophysiological evidence for a receptor-mediated Ca2+-influx pathway through Ca2+-selective, low conductance channels. This mechanism can only be observed at a very low [Ca2+li, it is inactivated by Ca2+i and is insensitive to InsP4.
In this work, we have further characterized the time course of the [Ca2+li rise induced by reexposing HeLa cells that were 5818 Thapsigargin-stimulated Ca2+ Entry in HeLa Cells shortly incubated in Ca2+-free solution, to extracellular Ca2+.  have observed a transient elevation of the [Ca2+li during the refilling process of empty stores in parotid acinar cells. Their protocol consisted of stimulating intact cells with a saturating dose of agonist in the absence of extracellular Ca2+, then preventing further stimulation by adding a receptor antagonist, and finally readmitting extracellular Ca2+ to stimulate Ca2+ entry. We have used a different protocol, in which store depletion was induced by a 60-min incubation in 2 p~ thapsigargin. This protocol prevented Ca2+ sequestration into internal stores during the phase of Ca2+ readdition. Our results indicate that readdition of external Ca2+ induced a transient [Ca2+li rise. The transient nature of this [Ca2+Ii rise in HeLa cells was due to a time-and Ca2+i-dependent decrease of the rate of Ca2+ entry. This Ca2+i-dependent inhibition of the Ca2+ entry and the subsequent recovery from this inhibition could cause an oscillatory Ca2+ entry.

MATERIALS AND METHODS
HeLa (S3) cells, which are epithelial cells derived from an epidermoid carcinoma of the cervix, were obtained from the American Type Culture Collection (Rockville, MD) and grown in Ham's F-12 medium supplemented with 10% fetal calf serum. A7r5 cells were cultured as described by Missiaen et al. (1990). The culture medium was replaced every 2-3 days. The cells were plated at a density of 2500 celldcm2 in coverglass chambers (Nunc Inc., Naperville, IL) and routinely investigated 6 days after plating. The technique for culturing BCBHl cells and for manipulating their state of differentiation was as described by De Smedt et al. (1991a, 1991b).
Single-cell [Ca2+li measurements using a laser-scanning confocal fluorescence microscope (MRC-BOO, Bio-Rad) coupled to an inverted epifluorescence microscope (Nikon Diaphot-TMD-EF), were performed as described (Missiaen et al., 1993). Briefly, aRer removing the culture medium and washing the cells, they were incubated for 30 min with 5 p~ Indo-1-AM dissolved in a modified Krebs solution of the following composition (m): 135 NaCl, 5.9 KCI, 1.5 CaCI2, 1.2 MgCl2, 11.6 Hepes, and 11.5 glucose (pH 7.3). The cells were then incubated for another 1-1.5 h in the absence of Indo-1. During the experiment, the cells were continuously superfused with either the modified Krebs solution or a Ca2+-free Krebs solution containing 2 m EGTA (flow rate of 2 ml rnin-'). The distance from the point where the fluid entered the coverglass chamber and the location of the cells which were imaged, was about 5 mm, causing some latency (5-10 s) of the various responses. K+-rich solutions were prepared by replacing all NaCl by KC1. All experiments were. performed at room temperature. We always monitored the [Ca2+Ii in 15 cells in one coverglass chamber. The [Ca2+li tracings shown in the various figures were representative for experiments performed on four coverglass chambers, i.e. on 60 cells. Mn2+ can be used as a probe for Ca2+ influx in a variety of cells (Hallam and Rink, 1985). The Mn2+-quench experiment was performed using a photomultiplier-based system consisting of an inverted microscope (IM 10, Zeiss, Oberkochen, Germany), filter wheel, amplifier and controller (Luigs and Neumann, Germany), and photomultiplier unit (Hamamatsu, Japan). Single cells were loaded with 2 p~ Fura-2 AM for 20 min at room temperature. Mn2+ uptake was monitored as the rate of quenching of Fura-2 fluorescence measured at the isosbestic point for the Ca2+.Fura-2 complex (excitation 360 nm). Measurements were corrected for autofluorescence.

RESULTS
HeLa cells contain histamine receptors, which are linked to Ca2+ mobilization through InsPs production (Tilly et al., 1990).  (Bootman et al., 1992a) and single HeLa cells (Diarra and Sauve, 1992). The plateau phase, which slowly declined, was curtailed by removal of extracellular Ca2+. When plasma-membrane Ca2+ influx was These data confirm previous findings in HeLa cells as well as in many other cell types, that the changes in [Ca2+Ii elicited by agonists, are due partly to release of internal Ca2+ and partly to influx of extracellular Ca2+. Thapsigargin releases Ca2+ from internal stores by specifically inhibiting the intracellular Ca2+ pumps (Thastrup et al., 1989). Addition of 2 p~ thapsigargin to HeLa cells incubated in a 1.5 m M Ca2+-containing solution, also increased the [Ca2+li ( Fig. IC). A sudden removal of extracellular Ca2+ produced an immediate drop in [Ca2+Ii in the thapsigargin-treated cells. When applied to cells exposed to a Ca2+-free solution, thapsigargin only produced a transient [Ca2+Ii rise (Fig. ID). These findings indicate that part of the thapsigargin response in Ca2+-containing solution must represent Ca2+ entry, confirming the original observations of Takemura et al. (1989).
All the known types of internal Ca2+ pumps, expressed in COS cells, are thapsigargin-sensitive (Lytton et al., 1991), but studies in intact cells have suggested that, under some conditions, not all stores are affected by thapsigargin, e.g. the caffeine-sensitive store of some cell types does not seem to react to thapsigargin (Law et al., 1990;Robinson and Burgoyne, 1991;Foskett and Wong, 1991). It was therefore necessary to check whether thapsigargin indeed depleted the stores under our

FIG.
2. Nature of the intracellular CaZ+ stores in HeLa cells. In A, a HeLa cell was incubated in a 2 mM EGTA-containing Caz+-free medium as indicated and then stimulated with 10 mM caffeine for the rest of the experiment. 20 PM histamine was subsequently added. The cell in B was pretreated with 2 p~ thapsigargin in a 1.5 m M CaZ+containing medium for 60 min. Addition of 20 p~ histamine to this thapsigargin-pretreated cell failed to affect the [Ca2+l,. experimental conditions. We first investigated whether HeLa cells express a caffeine-sensitive Ca2+ store. Fig. 2 A , which is representative for 60 different cells of various batches, illustrates that HeLa cells did not release internal Ca2+ in response to 10 mM caffeine. Because caffeine can interact with Ca2+sensitive dyes (Iino, 19891, it was necessary to consider this possible artifact. The absence of a caffeine-induced [Ca2+li rise was not due to the inability of Indo-1 to monitor increases in [Ca2+Ii in the presence of caffeine, since the subsequent addition of a high dose of histamine (20 PM) produced the expected [Ca2+Ii rise. We then investigated whether thapsigargin would completely deplete the histamine-sensitive Ca2+ store. Fig. 2B illustrates that after a 60-min preincubation in a 1.5 mM Ca2+containing medium supplemented with 2 p~ thapsigargin, 20 PM histamine did no more elicit a response and that the [Ca2+li remained at the same elevated steady-state level (compare with Fig. l . 4 ) . These data therefore indicate that thapsigargin completely emptied the histamine-sensitive store in HeLa cells. The steady-state [Ca2+lj in cells pretreated for 1 h with thapsigargin ( Fig. 2 B ) was higher than in control cells (Fig. 2 A ) . The exposure of thapsigargin-pretreated cells to a Ca2+-free medium lowered the [Ca2+li (Fig. 3 A ) . Readmission of extracellular Ca2+ produced a rapid and pronounced [Ca2+li rise, which then progressively declined to the control level observed before the Ca2+ withdrawal. The presumed Ca2+ entry blocker SKF 96365 (50 PM) (Merritt et al., 1990) blocked this [Ca2+Ii rise, indicating that it reflected entry of extracellular Ca2+ (Fig. 3B ). Washing out of the blocker resulted in the development of the [Ca2+Ii spike. This protocol of removal and readdition of extracellular Ca2+ did not have an effect on the [Ca2+Ii of control cells that were not pretreated with thapsigargin and that therefore had filled stores (Fig. 3C). This finding excludes aspecific effects of Ca2+ removal at the level of the plasma membrane.
We can assume that the measured free [Ca2+Ij in the presence of thapsigargin represented the balance between the rate of Ca2+ entry into the cell and the rate of Ca2+ extrusion across the plasma membrane. We therefore investigated whether the decrease of the [Ca2+Ij following the initial [Ca2+Ii peak would be due to a delayed activation of the Ca2+ extrusion mechanism at the level of the plasma membrane, or to an inactivation of a component of the Ca2+ entry mechanism. Fig. 4  brane. The rate of [Ca2+li decline at point 1 was 15 2 2 nM s-l (n = 15). That at point 2 was 12 2 1 nM s-l (n = 15). It was indeed necessary to compare the rate of [Ca2+li decline at the same level of [Ca2+Ii, since the Ca2+ pump is a Ca2+-activated transporter. The lower [Ca2+Ii at the plateau following the spike was therefore not caused by a delayed and persistent activation of the Ca2+ extrusion. Mn2+ can be used as a probe for Ca2+ influx in a variety of cells (Hallam and Rink, 1985). We have measured the unidirectional uptake of Mn2+ into single Fura-2-loaded HeLa cells to investigate whether the decrease of the [Ca2+li following the initial [Ca2+li peak represented the inactivation of a component of the Ca2+ entry pathway. We have studied the effect of a short lasting withdrawal of extracellular Ca2+ on the subsequent rate of Mn2+ uptake in the presence of 1.5 mM Ca2+ in a thapsigargin-pretreated cell (Fig. 5). The rate of Fura-2 quenching induced by 1.5 mM Mn2+ added after a 1-h exposure to a 1.5 mM Ca2+ containing modified Krebs solution, is shown in Fig. 5a. Fig. 5b shows that the same procedure induced a much more accelerated rate of Mn2+ quenching if the Mn2+ and Ca2+ addition was preceded by a 10-min incubation in a Ca2+-free medium. The very pronounced [Ca2+Ii rise induced by a short preincubation in Ca2+-free medium (Fig. 3 A ) was therefore associated with a n increased rate of Mn2+ entry. This finding indicates that the rate of Ca2+ entry is highest immediately following the incubation in Ca2+-free medium and that this rate progressively decreases leading to the declining phase of the transient.
In order to find out whether the trigger for activation of the entry mechanism following the incubation in Ca2+-free medium was the decreased [Ca2+lo or the decreased [Ca2+li, we have tested the effect of different [Ca2+J0 (Fig. 6). In Fig. 6 A , plasmamembrane Ca2+ entry in a thapsigargin-pretreated HeLa cell was increased by raising [Ca2+lo from 1.5 to 10 mM. Thereupon, the cell was incubated in a Ca2+-free medium for 150 s, and then challenged again with the same 10 mM Ca2+-containing solution. This second application of 10 mM Ca2+ produced a more pronounced [Ca2+li rise than the first application, indicating that the entry mechanism must have been partially inactivated during the first 10 mM Ca2+ application. In Fig. 6B, a rather similar protocol has been followed, except that the first 10 mM Ca2+ challenge was preceded by a 5-min incubation in a 7 7 Thapsigargin-pretreated . 6. Effect of the initial [Ca2+Ii on the amplitude of the subsequent spike in a thapsigargin-pretreated HeLa cell.
[Ca2+], (expressed in mM) was vaned as indicated below the tracing. In A, the first application of 10 mM Ca2+ produced a less pronounced [Ca2+li rise than the second application. The cell in B was preincubated for 4 min in a 140 mM K+-containing modified Krebs solution to reduce the [Ca2+],, and from time zero onward exposed for another 1 min (indicated by the horizontal bar above the tracing) to this depolarizing medium. The first application of 10 mM Ca2+ (in the usual 5.9 mM K+-containing modified the second application. Krebs solution) now produced a [Ca2+], rise of similar amplitude than medium containing 1.5 mM Ca2+ and 140 mM K+ to depolarize the cell and hence to decrease the driving force for Ca2+ ions across the plasma membrane. The free [Caz+Ii during this incubation in depolarizing medium (Fig. 6 B ) was therefore lower than during incubation in normal medium (Fig. 6A). Subsequently raising the [Ca2+], to 10 mM in the 5.9 mM K+-containing solution after incubation in this depolarizing solution, produced a [Ca2+li rise of similar magnitude to that following an incubation in the EGTA-containing Ca2+-free medium (Fig.  6 B ) . The [Ca2+li rise upon raising the [Ca2+l, from 1.5 to 10 mM was therefore higher if the [Ca2+Ii was initially lowered by raising the external K+ concentration. These data suggest that it was cytosolic Ca2+, and not external Ca2+, which inactivated the entry mechanism.
The amplitude of the [Ca2+li rise upon readdition of external Ca2+ to thapsigargin-pretreated HeLa cells bathed in Ca2+-free solution depended on the duration of the prior incubation in Ca2+-free solution (Fig. 7). A brief withdrawal of external Ca2+ was followed by a much less pronounced [Ca2+Ii rise than a long-lasting omission of extracellular Ca2+. Note that under both conditions, the [Ca2+li dropped to the same level. These results could point to a time-dependent activation of the entrymechanism upon lowering [Ca2+li.
Our finding of a time-dependent inactivation of the entry mechanism at high free [Ca2+Ii and a time-dependent reactivation at low [Ca2+Ii, implies that thapsigargin-treated HeLa cells should have the capability to oscillate. Fig. 8A illustrates that such oscillatory behavior was occasionally observed. About 20% of the cells oscillated in the presence of thapsigargin. We always monitored 15 cells in a field. We never observed two cells, including neighboring cells, oscillating in synchrony. For comparison, a typical Ca2+ oscillation induced by 1 p~ histamine in a HeLa cell not pretreated with thapsigargin (Tilly et al., 1990;Bootman et al., 1992b;Missiaen et al., 1993) is shown in Fig.  8B. The thapsigargin-induced Ca2+ oscillation had a lower amplitude, exhibited a slower rising phase and individual spikes in the oscillation lasted longer than the histamine-induced ones.
We finally investigated whether the transient [Ca2+li rise that we observed upon readdition of external Ca2+ to a thapsigargin-pretreated HeLa cell (Fig. SA)  types. In A7r5 smooth-muscle cells, we never observed such a transient (Fig. 9B). BCBHl cells represent a myogenic cell line, that can reversibly be switched from an undifferentiated to a differentiated state (De Smedt et al., 1991a, 1991b. While differentiated cells also failed to respond with a [Ca2+li transient (Fig. 9C), undifferentiated cells exhibited a clear [Ca2+Ii transient (Fig. 9D). Various cell types therefore seem to behave differently during the reexposure of external Ca2+.

DISCUSSION
The main finding in this work is that reexposure of HeLa cells with functionally depleted Ca2+ stores to extracellular Ca2+ induced a large transient increase in [Ca2+li. This transient occurred in the presence of thapsigargin, i.e. under conditions where the histamine-sensitive store was proven to be empty, and could therefore not represent the delayed sequestration of cytoplasmic Ca2+ into the internal stores. There are two possible explanations for this transient peak. A first possibility is that the rapid decline of the [Ca2+Ii transient is due to a delayed activation of the Ca2+ extrusion at the level of the plasma membrane. An enhanced extrusion following an initial [Ca2+li rise has been reported for smooth-muscle cells, but this activation diminished over a period of 30 s (Becker et al., 1989). The experiment represented in Fig. 4  bility. A second possibility is that the Ca2+ entry mechanism became inactivated. The Mn2+ uptake experiment (Fig. 5) favors the concept of a slow Ca2+i-dependent inactivation of the entry of divalent cations.
The way in which the state of filling of the store, or of a part thereof (Shuttleworth and Thompson, 19921, is signaled to the plasma membrane is not resolved yet, and it is therefore not clear how Ca2+i might act. The transient nature of the [Ca2+Ii rise upon readdition of extracellular Ca2+ was not primarily caused by time-dependent changes in membrane potential, since the [Ca2+li transient also occurred in a voltage-clamped cell at -40 mV2 However, secondary changes in membrane potential as a result of the [Ca2+li rise may modulate the amplitude of the [Ca2+li transient. The inactivation of the entry mechanism by high Ca2+i and its reactivation by low Ca2+& could represent a direct effect of Ca2+i on the Ca2+ entry channel. It should be pointed out that the [Ca2+Ii in the vicinity of the plasma membrane, and especially in the mouth of the channel, is much higher that the [Ca2+li measured in the bulk of the cytoplasm (Stern, 1992). The inactivation of a Ca2+-permeable channel by cytosolic Ca2+ would be in line with observations in mast cells by Hoth and Penner (19921, who   Thapsigargin-stimulated Ca2' Entry in HeLa Cells same range as observed in our study. The inactivation of the entry mechanism by high Ca2+j and its reactivation by low Ca2+j could also represent a Ca2+i effect on the production or metabolism of a putative messenger, or represent an effect of Ca2+i on a critical phosphorylation or dephosphorylation step (Parekh et al., 1993;Randriamampita and Tsien, 1993). The time dependence of the activation (Fig. 7) is in the expected range for such phosphorylation or dephosphorylation events. We cannot finally exclude that the inhibitory effects of [Ca2+Ij would be exerted from within the stores, even though they are functionally empty, since the free luminal [Ca2+] under these conditions should be similar as the [Ca2+Ij. However, it is conceivable that the primary control of "capacitative" Ca2+ entry by luminal Ca2+ occurs in a range of much higher luminal [Ca2+] than the range of 204-134 nM Ca2+ (Fig. 3.4).
The simplest explanation of our data is that empty stores induce two changes at the level of the plasma membrane. The first would result in an increased rate of Ca2+ entry, contributing to the steady-state elevation of the [Ca2+li in thapsigarginpretreated cells incubated in the presence of 1.5 mM Ca2+. The second component is a rapidly inactivating entry pathway under control of cytosolic Ca2+. The latter pathway was transiently activated by a short incubation in Ca2+-free medium. Interestingly, the second component was not present in some cell types and even seemed differentiation-dependent, at least in BCSHl cells (Fig. 9).
Our occasional observation of a Ca2+ oscillation in thapsigargin-pretreated cells was in a certain way unexpected, since current models to explain oscillations in non-excitable cells all rely on the presumed periodic release and reuptake of Ca2+ from internal stores (Rooney et al., 1989;Berridge, 1990). The internal stores in HeLa cells were depleted by thapsigargin, because histamine failed to mobilize Ca2+ in the presence of thapsigargin. This work is, however, not the first report of a Ca2+ oscillation in thapsigargin-pretreated cells, since rather similar observations were reported earlier for parotid acinar cells . However, the latter oscillations were assumed to be due to the release and reuptake of Ca2+ from a caffeine-sensitive but thapsigargin-insensitive Ca2+ store . We (Fig. 2 A ) and others (Diarra and Sauve, 1992) found no evidence for a caffeine-induced Ca2+ release in HeLa cells. We therefore propose that the plasma membrane of a non-excitable cell has the intrinsic capability of periodically increasing its permeability for Ca2+. It remains to be determined to what extent the Ca2+,-dependent inactivation of the entry mechanism could be involved in setting up the oscillatory behavior in the presence of thapsigargin. We also want to stress that it is not our intention to imply that the oscillations observed in the presence of thapsigargin are based on the same mechanism as the histamine-induced Ca2+ oscillation in control cells. The fact that their characteristics are different (Fig. 8) suggests that both are unrelated.
It is not clear what net effect Ca2+-mobilizing agents have on the activity of the Ca2+ extrusion mechanism in intact cells. The plasma-membrane Ca2+ pump is a Ca2+-activated enzyme, implying that the agonist-induced [Ca2+Ii rise will stimulate the Ca2+ pump (e.g. transient increases in the rate of Ca2+ efflux have been observed during each [Ca2+lj spike of an oscillation in single pancreatic acinar cells) (Tepikin et al., 1992). It remains to be determined whether the pump activity at a fwced [Ca2+Ii will also be affected by agonists. Experiments on the purified Ca2+ pump protein have shown that the agonistinduced breakdown of phosphatidylinositol 4-phosphate and phosphatidylinositol 4,5-bisphosphate would tend to inhibit the pump (Choquette et al., 1984;Missiaen et al., 19891, while the activation of the protein kinase C branch of the signaling pathway would stimulate it (Smallwood et al., 1988). The rise in cyclic GMP concentration, which often accompanies agonist stimulation (Pandol and Schoeffleld-Payne, 1990), would also stimulate the Ca2+ pump (Vrolix et al., 1988). The net effect of agonists on the rate of Ca2+ extrusion in intact cells is less clear. Experiments on intact cells have suggested that Ca2+-mobilizing agents stimulate Ca2+ extrusion at the level of the plasma membrane (Duddy et al., 1989;Zhang et al., 1992). However, possible compartmentalization of the dye could seriously affect the interpretation of such experiments, a t least in hepatocytes (Glennon et al., 1992). In the present work, we have found that histamine failed to affect the elevated steady-state [Ca2+Ii in the presence of thapsigargin (Fig. 2B). The lack of any [Ca2+Ii rise in response to histamine indicates that the internal stores were depleted and, as a consequence, that the measured [Ca2+Ii represented the balance between the rate of Ca2+ entry and extrusion across the plasma membrane. The lack of effect of histamine on the steady-state [Ca2+Ii under these conditions seems to suggest that this agonist had only a minor effect on the Ca2+ extrusion system in HeLa cells.
In conclusion, depleting the internal stores with thapsigargin increased the rate of Ca2+ entry into the cell, as judged from the increased resting [Ca2+Ij. A short period of lowering [Ca2+li induced a dramatic activation of the entry mechanism. Our findings are compatible with the hypothesis that some component of the entry mechanism is inactivated by high [Ca2+Ii and, therefore, that a low [Ca2+Ii is needed to fully activate the Ca2+ entry pathway.