Temperature-dependent block of capacitative Ca2+ influx in the human leukemic cell line KU-812.

The mechanism by which depletion of intracellular Ca2+ stores activates Ca2+ influx is not understood. We recently showed that primaquine, an inhibitor of vesicular transport, blocks the activation of the calcium release-activated calcium current (ICRAC) in rat megakaryocytes (Somasundaram, B., Norman, J. C., and Mahaut-Smith, M. P. (1995) Biochem. J. 309, 725-729). Since it is well established that vesicular transport is temperature-sensitive, we have investigated the effect of temperature on both the activation and maintenance of store-mediated Ca2+ and Mn2+ influx in the human leukemic cell line KU-812 using a combination of whole cell ICRAC recordings and measurements of Mn2+ photoquench of fura-2. Activation of ICRAC was temperature-sensitive, showing a nonlinear reduction when the temperature was lowered from 27 to 17 degrees C with an abrupt change at 21-22 degrees C and complete inhibition at 17 degrees C. Once activated, ICRAC also displayed an abrupt reduction at 21-22 degrees C but was not completely blocked even when the temperature was reduced to 14 degrees C, suggesting that at least one of the temperature-sensitive components is exclusively involved in ICRAC activation. Activation of store-mediated Mn2+ influx also showed similar nonlinear temperature sensitivity and complete inhibition at 19 degrees C. However, in contrast to ICRAC measurements, lowering the temperature following maximal activation of the influx pathway at 37 degrees C did not result in any detectable residual Mn2+ entry below 19 degrees C. We conclude that the mechanism of store-mediated Ca2+ influx involves temperature-dependent steps in both its maintenance and activation, suggesting dependence on a lipid membrane environment.

In many nonexcitable cells, depletion of intracellular Ca 2ϩ stores by inositol 1,4,5-trisphosphate activates Ca 2ϩ influx across the plasma membrane, a phenomenon termed "capacitative Ca 2ϩ entry" (Putney, 1986). This Ca 2ϩ entry pathway was first characterized electrophysiologically in mast cells and named "calcium release-activated calcium current" (I CRAC ) 1 (Hoth and Penner, 1992). Store depletion-activated currents with different properties from I CRAC have been described in some cell types (Lü ckhoff and Clapham, 1994;; however, I CRAC appears to be the predominant Ca 2ϩ influx pathway in nonexcitable cells Zweifach and Lewis, 1993;Somasundaram and Mahaut-Smith 1994; for review, see Fasalato et al. 1994). I CRAC is distinguishable from other store-dependent influx currents by being highly selective for Ca 2ϩ and having a single channel conductance below the resolution of patch clamp recordings Zweifach and Lewis, 1993). The mechanism whereby depletion of Ca 2ϩ stores is coupled to the activation and control of I CRAC is still unclear. Several signaling mechanisms have been proposed, including direct coupling of the store and plasma membrane via protein-protein interaction (Irvine, 1992) or release of a small, nonproteinaceous phosphate-containing second messenger from the Ca 2ϩ stores that diffuses to the plasma membrane (Randriamampita and Tsien, 1993;Kim et al., 1995). A number of other biochemical modulators have been implicated in the control of I CRAC , including cytochrome P-450, phosphatases, tyrosine kinases, cGMP (for review, see Sargeant and Sage, 1994), and protein kinase C (Parekh and Penner, 1995).
It has also been postulated that store depletion may cause insertion of I CRAC channels into the plasma membrane from specialized vesicles (Fasalato et al., 1994) following evidence of a role of small GTP-binding proteins in I CRAC activation (Fasalato et al., 1993;Bird and Putney, 1993) and the well recognized role of these molecules in vesicular transport (Pryer et al., 1992). We recently provided evidence to further support this notion by showing that two established inhibitors of vesicle-mediated protein transport, the antimalarial amine primaquine and GTP␥S, block the appearance of I CRAC in response to store depletion in rat megakaryocytes (Somasundaram et al., 1995). Primaquine is far less effective if added after I CRAC has developed. Thus the channels responsible for I CRAC or a molecule activating the channel may be held in a membrane compartment and transported to the plasma membrane following depletion of stores (Somasundaram et al., 1995). It is also well established that vesicular transport is temperature-sensitive and is blocked below temperatures ranging from 16 to 18°C (Matlin and Simons 1983;Tartakoff, 1986;Saraste et al., 1986). Therefore, in this study, we have investigated the effect of temperature on the activation and maintenance of storemediated Ca 2ϩ influx in a human leukemic cell line, KU-812 (Nakazawa et al., 1989), using a combination of whole cell patch clamp recordings and fluorescent photoquench of fura-2. The KU-812 cell line was selected because it displayed a high I CRAC channel density and an ability to load and retain the Ca 2ϩ -sensitive dye fura-2 and because of the avail-* This work was supported by the Biotechnology and Biological Sciences Research Council, the Medical Research Council, and the British Heart Foundation. 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.
‡ To whom correspondence should be addressed.
ability of large numbers of cells for population studies of Mn 2ϩ quench.

MATERIALS AND METHODS
Cells and Reagents-KU-812 cells were obtained from the European Collection of Animal Cell Cultures (Centre for Applied Microbiology and Research, Wiltshire, United Kingdom) and cultured in a humidified atmosphere at 37°C and 5% CO 2 in RPMI 1640 medium (Life Technologies, Inc.) supplemented with 10% fetal calf serum, 2 mM glutamine, 10 units/ml penicillin, and 10 mg/ml streptomycin. Cells were harvested by centrifugation, washed, and resuspended in a standard external solution (see below). Thapsigargin (TG), valinomycin, and ionomycin were obtained from Calbiochem, and fura-2/AM and pluronic acid F-127 were from Molecular Probes, Inc., (Eugene, OR). All these agents were prepared as stocks in dimethyl sulfoxide. Cs 4 -BAPTA was from Molecular Probes (Eugene OR), and all other reagents were from Sigma.
Electrophysiology-Patch clamp experiments were performed in the conventional whole cell configuration (Hamill et al., 1981) by means of an Axopatch 200A patch clamp amplifier (Axon Instruments, Inc., Foster City, CA). Pipettes were pulled from borosilicate glass tubing (Clark Electromedical Instruments) and had filled resistances of 2-3 megaohms. Series resistances were in the range of 10 -30 megaohms, and 40 -70% series resistance compensation was used. The cells were stored at 23-24°C in a standard external solution containing 140 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 2 mM CaCl 2 , 10 mM glucose, and 10 mM HEPES, pH 7.4 (adjusted with Tris) before transfer to the recording chamber. All recordings were conducted in an external solution in which Na ϩ and K ϩ ions were replaced by equal concentrations of positively charged nmethyl-D-glucamine and Cs ϩ , respectively. The internal saline contained 60 mM cesium gluconate, 20 mM Cs 4 -BAPTA, 5 mM NaCl, 2 mM MgCl 2 , 2 mM CaCl 2 , and 10 mM HEPES, pH 7.3, (adjusted with Tris). This Ca 2ϩ /BAPTA mixture maintained [Ca 2ϩ ] i near resting cytosolic levels and reduced the chance of spontaneous activation of I CRAC due to leaking of Ca 2ϩ from internal stores that often occurred with BAPTA alone. TG (3 M) was applied from a puffer pipette placed 150 m from the cell. A diffusible factor required for activation of I CRAC is washed out during whole cell recordings from rat megakaryocytes (Somasundaram et al., 1995) and rat basophilic leukemic cells (Fasalato et al., 1993). The rate of dialysis of the cytoplasm will depend on the whole cell access resistance. Therefore, to allow comparison of I CRAC in different cells, the time after breakthrough to the whole cell mode was normalized for series resistance, and thapsigargin was applied at a standard normalized time range of 9 s megaohm Ϫ1 . Cells were held at Ϫ20 mV, and current-voltage relationships were obtained at regular intervals using 300-ms voltage ramps from Ϫ140 to ϩ60 mV. The membrane currents during voltage ramps were low pass filtered at 2 kHz and sampled at 100 s using a Digidata 1200 interface and pClamp software (Axon Instruments). A continuous record of whole cell current was also obtained using a Cairn Research computer interface and associated software (Cairn Research Ltd., Kent, UK) following acquisition at a rate of 60 Hz (low pass filtered at approximately 30 Hz). Liquid junction potentials were measured by reference to a 3 M KCl agar bridge, and membrane potentials were adjusted accordingly. Currents were normalized for cell capacitance, and I CRAC was measured as the amount of inward current that developed in response to TG at a given voltage.
Temperature Control-In the electrophysiological experiments the temperature of the recording chamber was adjusted by a combination of a heating block and a heat exchanger controlled by a Peltier element. A gravity-fed perfusion system was used to exchange the saline in the experimental chamber. With this system the temperature of the recording chamber could be varied from 14 to 30°C. Temperature was measured with a thermistor positioned 1 mm from the cell. At each temperature selected, the cell was allowed to equilibrate for at least 2 min before whole cell configuration was established. In experiments in which the effect of temperature on the maintenance of I CRAC was measured, the whole cell configuration was established at 24°C, and the temperature was lowered at a rate of approximately 0.1°C/s. In the fluorescence experiments the temperature was controlled by a water jacket connected to a heating-cooling system, which enabled the temperature to be varied between 10 and 37°C. In the calcium electrode experiments the temperature was changed between 15 and 30°C using a Peltier element.
Measurement of Free Ca 2ϩ Concentration in Permeabilized Cells-Ca 2ϩ was determined with a Ca 2ϩ -sensitive minielectrode suspended in 60 l of permeabilized cell suspension solution stirred by a magnetic agitator. The Ca 2ϩ minielectrodes were prepared using a protocol similar to that of Clapper and Lee (1985). The electrode membrane was made with a Ca 2ϩ electrode mixture (Fluka Chemical Corp.) containing 4.15 mg of ETH 1001 Ca 2ϩ ionophore, 37 mg of S-nitrophenyloctyl ether, 0.42 mg of sodium tetraphenyl borate, and 24 mg of polyvinyl chloride dissolved in 240 l of tetrahydrofuran. The electrode was filled with a 10 mM CaCl 2 solution and connected to a Genway 3040 pH and ion analyzer. The minireference (3 M KCl) electrode used (ULTRA-WICK TM ) was purchased from World Precision Instruments, Inc. The electrode was calibrated at 15, 20, and 30°C using various Ca 2ϩ -EGTA buffers and displayed a linear response within the range of 0.1-100 M free Ca 2ϩ (25 mV/pCa 2ϩ unit). Cells were washed in a solution containing 120 mM KCl, 25 mM HEPES, 3 mM MgCl 2 , and 1 mM EGTA, washed again in the same solution without EGTA, and finally resuspended in this non-EGTA-containing solution at a final concentration of 2 ϫ 10 8 cells/ml. The cells were then stored on ice until use. For each experiment, 25 l of the cell suspension was warmed to 37°C, and digitonin was added to give a final concentration of 40 M and incubated for 2 min. The cell suspension was made up to 60 l with non-EGTA-containing solution, phosphocreatine (10 mM), creatine kinase (20 units/ ml), and 3 mM ATP and then transferred to form a droplet into which the calcium electrode tip was placed. The temperature was adjusted as required before experimentation.
Fluorescence Recording-Fura-2 fluorescence measurements were carried out on cell populations in a final volume of 1.5 ml (10 6 cells/ml) using a Cairn spectrophotometer system. Excitation wavelengths of 340, 360, and 380 nm were provided by a filter wheel rotating at 35 Hz in the light path. Emitted light was filtered by a 485 nm long pass filter and samples averaged to give a data point every 500 ms. The background-corrected 340:380 ratio was multiplied by the dissociation constants for fura-2 at different temperatures obtained from Shuttleworth and Thompson (1991) to give an indication of [Ca 2ϩ ] i . The isosbestic excitation wavelength (360 nm) was used to monitor the fura-2 photoquench by Mn 2ϩ ; an established measure of influx through the Ca 2ϩ store-activated influx pathway. Cells were loaded with fura-2 by incubation with 1 M fura-2/AM and 0.025% Pluronic F127 for 30 min at 24°C in standard external solution containing 140 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 10 mM glucose, 2 mg/ml apyrase, 10 mM HEPES, pH 7.4 (adjusted with Tris), and 1 mM CaCl 2 . The apyrase was used to minimize purinergic receptor activation, which may be caused by ATP or ADP released from damaged cells. The cells were spun down and resuspended in the standard external solution with 2 mM CaCl 2 and split into 100-ml aliquots each containing 10 6 cells and stored at room temperature until use. Prior to experimentation the cells were spun and resuspended in 100 ml of nominally Ca 2ϩ -free standard external medium with the K ϩ ionophore valinomycin (1 mM). Valinomycin was then present throughout the experiment allowing the cell membrane potential to be "clamped" close to the reversal potential of K ϩ (approximately Ϫ80 mV). No differences in Mn 2ϩ influx were observed (at 16 and 37°C) when valinomycin was added immediately or 6 min before measurement of Mn 2ϩ photoquench (n ϭ 3). To study the effect of temperature on activation of store-dependent Mn 2ϩ influx, the 100-l cell aliquot was added to 1.4 ml of medium at the required temperature. After a 30-s delay, 100 nM TG was added, followed 6 min later by 100 M Mn 2ϩ . To measure basal Mn 2ϩ influx, 100 M Mn 2ϩ was added in place of TG, after a 30-s incubation at the required temperature. To study the effects of temperature on the maintenance of the store-mediated Mn 2ϩ influx pathway, the stores were emptied first by adding 100 nM TG to the 100-l cell suspension in nominally Ca 2ϩ -free standard external medium and incubated for 6 min at 37°C. The cell suspension was then added to the cuvette containing 1.4 ml of nominally Ca 2ϩ -free standard external medium at the required temperature. After a 30-s delay, 100 M Mn 2ϩ was added, and the quench was measured. In controls, the 100-l cell suspensions were incubated at 37°C without TG for the same length of time but with 1 mM added Ca 2ϩ to prevent any passive store depletion during incubation. To quantify Mn 2ϩ entry, the slopes of the fura-2 photoquench over the first 15 s after Mn 2ϩ addition were determined. The very initial, sharp drop in fluorescence observed in some experiments, due to external free dye, was not included in the calculations.
The degree of TG-induced store depletion was assessed at both 34 and 18°C by comparing the ionomycin-induced (5 M) Ca 2ϩ rise before and after a 6-min incubation with TG. To eliminate the contribution of Ca 2ϩ influx to these measurements, 0.1 mM EGTA was added to the external medium immediately before addition of ionomycin.
Measurement of Cytosolic ATP-Cytosolic ATP concentrations were measured using an ATP assay kit (Calbiochem) based on the firefly luciferase-catalyzed oxidation of D-luciferin in the presence of an ATPmagnesium salt and oxygen. A cell suspension (3 ϫ 10 5 cells/ml) in standard external medium was incubated at the desired temperature for 8 min, spun, and resuspended in HEPES buffer, pH 7.75, containing a permeabilizing reagent (Calbiochem). To assess ATP released from the cytoplasm, luciferase was added, and the peak amount of photons emitted was measured using a photomultiplier tube (Thorn EMI Electron Tubes Ltd., Ruislip, Middlesex, UK) coupled to a spectrophotometer (Cairn). Luminescence following addition of luciferase was measured and subtracted from luminescence in the absence of cells. The cytosolic ATP concentration was calculated using a measured mean cell diameter of 10 m and assuming the cytosolic volume to be 20% of the total cell volume.
Calculation of Thermodynamic Parameters-Temperature coefficient (Q 10 ) values were determined for current densities and Mn 2ϩ influx rates within a given temperature range using Q 10 ϭ (P 1 / P 2 ) 10/(T1 Ϫ T2) , where P 1 and P 2 are current densities at temperature T1 and T2, respectively. Activation energy (E a ) was determined from an equation describing the slope of the Arrhenius plot: E a ϭ R (lnP 1 Ϫ lnP 2 )/(1/T1 Ϫ 1/T2), where R is the gas constant.

RESULTS
Effect of Temperature on I CRAC Activation-To study I CRAC in KU-812 cells, conditions were selected that have been shown to enhance this small current and largely eliminate contributions from other ionic currents (Somasundaram et al., 1995;see "Materials and Methods"). Briefly, the external medium contained 2 mM Ca 2ϩ and no K ϩ or Na ϩ , and any outward K ϩ currents were blocked by replacing the internal K ϩ with Cs ϩ . The [Ca 2ϩ ] 1 was strongly buffered with 20 mM BAPTA. Under whole cell voltage clamp, depletion of internal Ca 2ϩ stores by 3 M TG, an endoplasmic Ca-ATPase inhibitor, evoked an inward current at a holding potential of Ϫ50 mV (Fig. 1A). The currents generated by voltage ramps from Ϫ140 to ϩ 50 mV before TG application and at various times during the development of the inward current are shown in Fig. 1B. This inwardly rectifying current showed little or no reversal within the voltage range studied, developed without detectable single channel events, was selective for Ca 2ϩ (data not shown), and was blocked by 1 mM Zn 2ϩ (data not shown), all of which are characteristic of I CRAC reported in other cells (Fasalato et al., 1994;Somasundaram et al., 1995). To study the effect of temperature on the activation of I CRAC , cells were held at various temperatures ranging from 15 to 27°C, and the current evoked by TG was recorded. The currents activated at a holding potential of Ϫ20 mV at four different temperatures are shown in Fig. 1C. The maximum TG-induced current density was reduced with decreasing temperature such that no measurable current was activated at or below 17°C. The speed of development of the current was also reduced, and the times to halfmaximum current were 82 Ϯ 12, 72 Ϯ 8, 55 Ϯ 9, and 50 Ϯ 7 s at 19, 20, 23, and 26°C, respectively. At maximal activation, the current-voltage relationship generated by voltage ramps from Ϫ140 mV to ϩ50 mV at the different temperatures did not show any obvious shift in the reversal potential along the x axis (Fig. 1D). It should be noted that it was not technically possible to perform whole cell patch clamp recordings from cells held at temperatures greater than 27°C due to rapid deterioration of glass membrane seals.
The TG-evoked I CRAC densities, measured at Ϫ120 mV, in individual cells at different temperatures are shown in Fig. 2A. Despite the considerable variation in current density between cells, there is a clear positive correlation between current density and temperature. This relationship is further illustrated in Fig. 2B, in which the mean current density values at Ϫ120, Ϫ70, and Ϫ20 mV are plotted against temperature. Over the temperature range of 27 to 21°C the current density at all three voltages displayed little change, whereas further lowering of temperature caused a significant decrease in current, which became blocked completely at 17°C. To describe quan- titatively the temperature dependence of the current density, we have presented the same data in the form of Arrhenius plots (Fig. 2C). The current density decreases linearly as temperature is lowered from 27 to 22°C, and from the slope, apparent E a values of 4.3 kcal/mol (Q 10 ϭ 1.11), 5.1 kcal/mol (Q 10 ϭ 1.26), and 5.8 kcal/mol (Q 10 ϭ 1.18) were calculated at Ϫ120, Ϫ70, and Ϫ20 mV, respectively. On lowering the temperature from 21 to 17°C, there was an abrupt increase in the E a by a factor of approximately 33 to give E a values of 161 kcal/mol (Q 10 ϭ 196), 178 kcal/mol (Q 10 ϭ 67), and 167 kcal/mol (Q 10 ϭ 180) at the respective voltages, reflecting a very high energy barrier at the lower temperatures. For both the high and low temperature ranges, the activation energies were similar at different membrane potentials, suggesting that these steps were not voltage-dependent.
Although depletion of Ca 2ϩ stores at 17°C failed to induce an inward current, an increase in temperature several min after TG application at 17°C resulted in I CRAC activation, as shown by the experiment in Fig. 3. The current-voltage relationships at 17°C 2 min after TG application (Fig. 3B, trace b) showed no evidence of a store-dependent current, whereas this current slowly developed following a temperature increase to 20°C (Fig. 3B, trace c). A further abrupt increase in temperature to 28°C produced a parallel increase in I CRAC (Fig. 3B, trace d), and furthermore, this temperature-dependent effect was reversible, as shown by the currents obtained at 20°C before and after increasing the temperature to 28°C (Fig. 3B, traces c and e).
Effect of Temperature on I CRAC Maintenance-We next studied the effect of temperature on I CRAC following full activation at 24°C, to determine whether the temperature dependence arose primarily from an effect on the activation mechanism of the current or via a direct effect on the influx pathway itself. In the experiments of Fig. 4, I CRAC was maximally activated by TG-induced store depletion at 24.5°C before the temperature was lowered to 14.5°C at a rate of 0.1°C/s and then returned to 24.5°C. A continuous record of I CRAC at a holding potential of Ϫ60 mV is shown in Fig. 4A, and current-voltage relationships at different temperatures are shown in Fig. 4B. Fig. 4C shows the average current density at Ϫ120 and Ϫ70 mV calculated from 10 cells at different temperatures following protocols similar to that of Fig. 4A and shows the complete reversibility of this temperature block.
When the temperature was dropped from 24.5 to 22°C, the inhibition of current density was of a level (Fig. 5A) very similar to that observed when measuring temperature effects on I CRAC activation (Fig. 2B). However, reduction in temperature from 22 to 14.5°C resulted in incomplete inhibition of current. Even at 14.5°C, the I CRAC current density remained at 44% of maximum, in contrast to the complete inhibition at this temperature observed for I CRAC activation (Fig. 2B). Indeed, even after prolonged incubations of up to 3 min at 14°C, cells FIG. 2. Analysis of the temperature dependence of I CRAC . A, I CRAC density measured at Ϫ120 mV as a function of activation temperature. Each point represents maximal I CRAC determined from one cell. Maximal I CRAC densities were determined using voltage ramps from Ϫ120 to ϩ60 mV applied after maximal activation of the current by TG-induced store depletion and corrected for background current measured prior to TG application. B, average maximal I CRAC densities at Ϫ120, Ϫ70, and Ϫ20 mV as a function of activation temperature. Bars, S.E. C, Arrhenius plots of the average maximal current densities at the three different voltages. Curves in B and C are the result of a four-variable logistic fit. pF, picofarad. still maintained substantial amounts of I CRAC (data not shown). This was in contrast to the complete block of I CRAC activation observed at this temperature ( Fig. 2A). To quantitatively describe the temperature dependence of I CRAC maintenance, the mean current densities were plotted against temperature in the form of an Arrhenius plot (Fig. 5B) from which E a values were calculated for temperatures from 24.5 to 22°C of 5 kcal/mol (Q 10 ϭ 1.4) at Ϫ120 mV and 2.2 kcal/mol (Q 10 ϭ 1.1) at Ϫ70 mV. From 22 to 14.5°C, the E a values increased to 23 kcal/mol (Q 10 ϭ 3.5) at Ϫ120 mV and 35 kcal/mol (Q 10 ϭ 5.6) at Ϫ70 mV.
Effect of Temperature on the Activation and Maintenance of Mn 2ϩ Influx-Although patch clamp recordings of I CRAC give a direct measure of Ca 2ϩ influx, the current is measured under nonphysiological ionic conditions. In addition, as a result of the whole cell patch configuration, dialysis of the cytosol may lead to removal of intracellular components controlling Ca 2ϩ influx. These potential problems were minimized by investigating the effect of temperature on store-mediated Ca 2ϩ influx in intact cells using Mn 2ϩ as a surrogate permeable ion and measuring the rate of the Mn 2ϩ -induced photoquench of cytosolic fura-2 at different temperatures. The experimental protocol is illustrated in Fig. 6A and shows a greatly increased Mn 2ϩ quench in cells depleted at 37°C than in cells depleted at 16°C. TGinduced store release led to a substantial increase in Mn 2ϩ influx compared with nondepleted cells at 32°C (Fig. 6B, traces c and d). At 10°C, however, the photoquench in store-depleted cells was indistinguishable from that in nondepleted cells (Fig.  6B, traces a and b). The differences in the Mn 2ϩ photoquench between store-depleted and nondepleted cells at 10, 16, 22, and 32°C are shown in Fig. 6C. From this basal-subtracted photoquench, the rate of store-mediated Mn 2ϩ entry at different temperatures was calculated (Fig. 6D). When the temperature was decreased from 37 to 28°C, there was a small reduction in the rate of the Mn 2ϩ quench, but on decreasing the temperature further from 28 to 10°C, there was a drastic reduction in the rate of Mn 2ϩ influx, with a complete block at 18°C. To examine the temperature effects on Mn 2ϩ influx quantitatively, the rate of the store-mediated Mn 2ϩ quench was plotted as an Arrhenius plot (Fig. 6E). The store-mediated Mn 2ϩ influx shows a linear decrease in the rate of influx from 37 to 25°C with a E a of 2 kcal/mol (Q 10 ϭ 1.2). At about 25°C there is an abrupt increase in the sensitivity of Mn 2ϩ influx to temperature. Below this temperature, the E a increases to 86 kcal/mol (Q 10 ϭ 51), reflecting a much higher energy barrier and similar to the increase observed when Ca 2ϩ influx was measured as I CRAC .
To establish whether this temperature sensitivity was a result of its effect on the activation mechanism of influx or on the influx pathway itself, we depleted the stores at 37°C prior FIG. 4. Effect of temperature on the maintenance of I CRAC . A, effect of temperature (top trace) on the whole cell current (bottom trace) after maximal activation at 23°C by 3 M TG at a holding potential of Ϫ40 mV. B, membrane currents generated by voltage ramps from Ϫ140 to ϩ40 mV at different times during the experiments shown in A: a, at 23°C before TG; b, after maximal activation of I CRAC at 23°C; c, after lowering the temperature to 16°C; and d, after reheating to 23°C. C, mean I CRAC densities at Ϫ120 and Ϫ70mV at different temperatures following activation by 3 M TG at 24.5°C. The data were obtained from similar experiments as described in A and B and were corrected for background current measured prior to TG application. Each column is the mean from 10 cells. Bars, S.E.; pF, picofarad. to measuring the rate of Mn 2ϩ influx at various temperatures. The difference in the rate of the Mn 2ϩ photoquench between store-depleted and nondepleted cells measured in this way is shown in Fig. 7A. The maintenance of Mn 2ϩ influx was affected by temperature in a manner similar to activation. The Arrhenius plot (Fig. 7B) of the rate of the store-mediated Mn 2ϩ quench component shows a change of E a from 1.3 kcal/mol (Q 10 ϭ 1) between 37 and 23°C to 119 kcal/mol (Q 10 ϭ 61) between 23 and 10°C. Thus, in contrast to the patch clamp recordings, in which there was found to be a temperature effect on both activation and maintenance, using the Mn 2ϩ photoquench we could only detect a temperature-dependent block on the maintenance of the store-mediated influx pathway.
Effect of Temperature on TG-induced Store Release-The observed temperature-dependent block of store-mediated Ca 2ϩ entry may have been a consequence of temperature sensitivity in the ability of TG to release Ca 2ϩ from internal stores. This appears unlikely, however, since the temperature effects on Mn 2ϩ influx were equivalent whether TG was added at 37°C or at lower temperatures (compare Fig. 6, D and E, with Fig. 7, A  and B). To further assess whether Ca 2ϩ stores were being emptied at low temperatures, we used both fura-2 and a Ca 2ϩ minielectrode to measure TG-induced Ca 2ϩ release at different temperatures. As illustrated in Fig. 6A, TG was added to cells in nominally Ca 2ϩ -free external medium at different temperatures, and the peak [Ca 2ϩ ] i during the subsequent 6 min was assessed using the 340:380-nm fura-2 ratio, corrected (see "Ma-terials and Methods") for temperature effects on the K d of the dye (Fig. 8C). These results suggest that there was no significant difference in the level of Ca 2ϩ released at different temperatures by TG over 6 min. Further confirmation of this is illustrated in Fig. 8D. The initial pool size was assessed by measuring the peak [Ca 2ϩ ] i following the addition of 5 M ionomycin at both 18 and 34°C. The degree of pool depletion (as a percentage of initial pool content) was measured by the ionomycin-induced [Ca 2ϩ ] i rise following a 6-min incubation with TG and was not shown to be significantly different at these two temperatures. Since fura-2 measurements only indicate the net balance between Ca 2ϩ influx and efflux, we resorted to Ca 2ϩ -sensitive minielectrode measurements on digitonin-permeabilized cells to directly assess TG-induced Ca 2ϩ store release at different temperatures. The protocol by which this was carried out is shown in Fig. 8A (also see "Materials and Methods"). The amount of Ca 2ϩ released by TG at different temperatures was measured as a percentage of the Ca 2ϩ released by ionomycin and is shown in Fig. 8B. There was no significant reduction in the percentage of Ca 2ϩ released by TG at lower temperatures. However, this method may also have limitations, since ionomycin may show some temperature sensitivity. Taken together, these two independent methods of testing the temperature sensitivity of TG-induced store release suggest that there is no significant difference in the degree of store release within the temperature range studied, and, hence, the temperature effects on store-mediated Ca 2ϩ and Mn 2ϩ influx are unlikely to be due to incomplete store release.
Effect of Temperature on Cytosolic ATP Concentration-Lowering the temperature of cells will slow down many metabolic processes, including ATP synthesis. Recent work suggests that ATP depletion by metabolic inhibitors can block store-dependent Ca 2ϩ influx (Gamberucci et al., 1994;Marriott and Mason, 1995). To test whether the temperature-dependent inhibition of Ca 2ϩ influx was due to a depletion of cytosolic ATP, we studied the effect of temperature on the cytosolic ATP concentration. Cytosolic ATP concentrations at different temperatures are shown in Fig. 9. Contrary to expectations, ATP levels increased at lower temperatures, possibly due to a reduction in ATP hydrolysis. It seems unlikely, therefore, that the temperaturedependent block of Ca 2ϩ influx is due to depletion of cytosolic ATP.

DISCUSSION
The present study demonstrates that, in KU-812 cells, storeregulated Ca 2ϩ and Mn 2ϩ entry is strongly dependent on temperature. It is clear from patch clamp recordings that both the activation of I CRAC and its maintenance are temperature-sensitive. The specific temperature effect on activation is apparent from the fact that development of I CRAC is blocked at 17°C, although once activated, lowering the temperature even to 14°C does not totally inhibit the current. The temperature effects on the maintenance of I CRAC are demonstrated by the nonlinear decrease when the temperature is lowered, with an abrupt transition around 22°C. Although Mn 2ϩ quench measurements also showed equivalent temperature-dependent decreases in influx with an abrupt transition temperature, we could only detect temperature effects on maintenance of influx, since these were equivalent regardless of whether the temper-ature was changed before or after store release.
This discrepancy between the two methods may be due to a number of factors. In contrast to Mn 2ϩ influx measurements, patch clamp recordings will lead to dialysis of the cytosol with two possible consequences. The use of internal solutions low in Na ϩ , K ϩ , and Cl Ϫ and high in Cs ϩ may alter temperature-dependent lipid rearrangements, which could explain the incomplete inhibition of I CRAC seen on cooling following activation at 24°C. In addition, any soluble factors that may inhibit I CRAC will be dialyzed out. One such factor, ATP, which will be at much higher levels in intact cells, has recently been shown to inactivate I CRAC possibly via protein kinase C activation (Parekh and Penner, 1995). This may explain the complete temperature-dependent inhibition of Mn 2ϩ entry following store depletion at 37°C.
The inhibitory effects of temperature on Ca 2ϩ influx are unlikely to be due to partial emptying of stores by TG at lower temperatures. Both fura-2 and the Ca 2ϩ -sensitive minielectrode measurements demonstrate that, following a 6-min incubation with TG, the Ca 2ϩ released from stores did not vary significantly over the temperature range studied (10 -37°C). Furthermore, the temperature-dependent inhibition of Mn 2ϩ influx was equivalent whether TG addition was carried out at low temperatures or at 37°C, suggesting equivalent store depletion after 6 min at different temperatures.
Another concern was whether the observed temperature-dependent block of store-dependent Ca 2ϩ influx was due to depletion of cytosolic ATP, since Gamberucci et al. (1994) have shown that a 5% reduction in cytosolic ATP concentration can inhibit store-mediated Ca 2ϩ influx by 50%. However, our results show that cytosolic ATP concentrations at temperatures at which I CRAC and Mn 2ϩ were blocked (4 mM at 17-18°C) were in fact greater than at higher temperatures (3 mM at 37°C), at which the influx was fully activated, implying a temperaturedependent reduction in ATP hydrolysis.
The main finding of the present study is the nonlinear reduction of I CRAC and Mn 2ϩ influx observed when the temperature was lowered. Most ion channels are known to be sensitive to temperature but exhibit a linear change in activity with change in temperature and have low activation energies (Bamberg and Lauger, 1974;Zeidel et al., 1992). Among Ca 2ϩ channels, the voltage-gated channels from ventricular myocytes (Cavalie et al., 1985) and sensory neurons (Nobile et al., 1990) also show a linear increase in current with increasing temperature due to an increase in open channel probability. However, some ion channels, including the actylcholine channel (Fischbach and Lass, 1978) and Ca 2ϩ channels in neuroblastoma cells (Narahashi et al., 1987), show a nonlinear decrease in conductance when the temperature is lowered, with a transition around 20°C, and it has been suggested that such transition effects are indicative of membrane-dependent regulation of channel gating (Romey et al., 1980). The nonlinearity in the Arrhenius plots of various membrane functions have been correlated with phase transition or lateral phase separation of membrane phospholipids (Raison 1972;Linden et al., 1973;Warren et al., 1975;Chapman, 1975). In the present study, the observed transition of I CRAC and Mn 2ϩ influx at 21-24°C indicates that this influx pathway may also be in close association with a lipid membrane environment. A similar transition of Mn 2ϩ influx at 21°C has also been observed in rat parotid acinar cells (Lockwich et al., 1994). The store-mediated influx pathway shows a much greater increase in E a compared with the acetylcholine channel or the neuroblastoma channels. It should be also noted that such transitional changes also apply to other membrane transport systems, such as transporters and pumps (Rega, 1986), and in view of the fact that the FIG. 7. Effect of temperature on the maintenance of storemediated Mn 2؉ influx. A, mean rate of Mn 2ϩ entry as a function of temperature. B, Arrhenius plot of the data in A. The rate of Mn 2ϩ photoquench was calculated as described in Fig. 6, B and C; however, the Ca 2ϩ stores were first emptied at 37°C using TG before changing to the desired temperature at which the Mn 2ϩ quench was measured (see "Materials and Methods"). Curves in A and B are the result of a fourvariable logistic fit. store-dependent Ca 2ϩ influx pathway has been shown to be of very low conductance with no single channel noise, the possibility of a transporter being involved in capacitative Ca 2ϩ influx should not be ruled out. The possibility also exists that abrupt breaks in Arrhenius plots may reflect intrinsic changes in the protein conformation independent of changes in the membrane phospholipid (Dean and Tanford, 1978;Sondergaard, 1979;Hoffman et al., 1979). This may be occurring in the store-dependent Ca 2ϩ influx pathway, in which a possible conformational coupling between stores and plasma membrane has been postulated (Irvine, 1992;Berridge, 1995).
The other interesting finding is that I CRAC activation is completely inhibited at 17°C but that, once activated, the current can be maintained even at lower temperatures. These observations suggest that, unlike the effect of temperature on Mn 2ϩ entry, there is an additional effect of temperature specifically on I CRAC activation. Although this could be a consequence of intracellular dialysis, it is consistent with the possible involvement of vesicular transport in the activation pathway of I CRAC . It is established that low temperatures, between 15 and 22°C, block the in vivo transport of membrane constituents at different points along the endoplasmic reticulum-golgi apparatusplasma membrane pathway (Matlin and Simons 1983;Tartakoff, 1986;Saraste et al., 1986;Moreau and Cassagne, 1994). This transition temperature is dependent on the various chain lengths of the phospholipid, and typically in animal cells, the golgi apparatus-plasma membrane pathway mediates the transfer of C 20 -C 24 -containing phospholipids, which is blocked at 16°C (Moreau and Cassagne, 1994). Furthermore, this may also explain the fact that the maintenance of I CRAC , although reduced, is not totally inhibited by lowering the temperature. Once the activation of the influx pathway via a vesicular transport process has been achieved (recruitment of channels or regulators of channels to the plasma membrane), then blocking this process should not inhibit Ca 2ϩ influx, since the channels, or channel regulators, are already on the plasma membrane. FIG. 8. Effect of temperature on TG-induced release of stored Ca 2؉ . A, percentage Ca 2ϩ released by TG (3 M) and ionomycin (10 M) in digitonin-permeabilized cell suspensions at 15 and 30°C. The cells were suspended in an internal medium containing an ATP-regenerating system, and Ca 2ϩ was measured by a mini-Ca 2ϩ electrode (see "Materials and Methods"). B, TG-induced Ca 2ϩ store release measured from experiments as in A, given as a mean percentage of ionomycin (Iono)-induced Ca 2ϩ release, plotted as a function of temperature. Bars, S.D.; n ϭ 3. C, effect of temperature on the mean TG-induced peak [Ca 2ϩ ] i measured as 340:380 fluorescence ratio (R 340/380 SF). The ratio was multiplied by the K d of fura-2 at different temperatures (see "Materials and Methods"). Bars, S.D. D, TG-induced Ca 2ϩ release at 18 and 34°C calculated from the difference between peak [Ca 2ϩ ] i rise induced by ionomycin before and 6 min after TG addition (n ϭ 3). This is expressed as a percentage of the ionomycin-induced peak [Ca 2ϩ ] i rise before TG addition.
FIG. 9. Effect of temperature on cytosolic [ATP]. The cytosolic ATP concentration is plotted as a function of incubation temperature; each point is the mean of two experiments. Cell suspensions were incubated for 6 min at different temperatures in a standard external medium containing 2 mM Ca 2ϩ before cytosolic [ATP] was determined using a luciferin/luciferase assay (see "Materials and Methods").
In conclusion, we have provided the first direct evidence demonstrating that both activation and maintenance of the store-mediated Ca 2ϩ influx pathway are exquisitely temperature-sensitive, suggesting a very intimate association with the lipid membrane environment.