Ca2+ currents activated by depletion of Ca2+ stores in Xenopus oocytes were studied with a two-electrode voltage clamp. Buffering of cytosolic Ca2+ with EGTA and MeBAPTA abolished ICl(Ca) and unmasked a current in oocytes that was activated by InsP3 or ionomycin in minutes and by thapsigargin or the chelators themselves over hours. At −60 mV in 10 mM extracellular CaCl2, the current was typically around −90 or −160 nA in oocytes loaded with EGTA or MeBAPTA, respectively. This current was judged to be a Ca2+-selective current for the following reasons: (a) it was inwardly rectifying and reversed at membrane potentials usually more positive than +40 mV; (b) it was dependent on extracellular [CaCl2] with Km = 11.5 mM; (c) it was highly selective for Ca2+ against monovalent cations Na+ and K+, because replacing Na+ and K+ by N-methyl-d-glucammonium did not reduce the amplitude or voltage dependence of the current significantly; and (d) Ca2+, Sr2+, and Ba2+ currents had similar instantaneous conductances, but Sr2+ and Ba2+ currents appeared to inactivate more strongly than Ca2+. This Ca2+ current was blocked by metal ions with the following potency sequence: Mg2+ << Ni2+ ≈ Co2+ ≈ Mn2+ < Cd2+ << Zn2+ << La3+. It was also inhibited by niflumic acid, which is commonly used to block ICl(Ca). PMA partially inhibited the Ca2+ current, and this effect was mostly abolished by calphostin C, indicating that the Ca2+ current is sensitive to protein kinase C. These results are the first detailed electrophysiological characterization of depletion-activated Ca2+ current in nondialyzed cells. Because exogenous molecules and channels are easy to introduce into oocytes and the distortions in measuring ICl(Ca) can now be bypassed, oocytes are now a superior system in which to analyze the activation mechanisms of capacitative Ca2+ influx.
In nonexcitable cells, intracellular calcium release mediated by activation of phosphoinositide metabolism is followed by a “capacitative calcium influx” (Putney, 1986; Berridge, 1995). Activation of this form of calcium influx is long-lasting, which is probably vital for some physiological functions, such as activation of lymphocytes (Lewis and Cahalan, 1995). Electrical currents corresponding to the calcium influx have been well characterized in mast cells and lymphocytes with whole-cell recording methods. The current has been termed ICRAC (calcium release-activated calcium current)1 because it can be activated by releasing Ca2+ from internal stores through several mechanisms. They include physiological liberation of calcium from internal store by InsP3 and pharmacological depletion of the store calcium by inhibitors of endoplasmic reticulum Ca2+-ATPase, calcium ionophores and high levels of calcium chelators (Hoth and Penner, 1992, 1993; Fasolato, 1994). ICRAC is highly selective for Ca2+ over monovalent cations (Hoth and Penner, 1993). Single channel currents are not resolved using patch–clamp techniques because single-channel conductances appeared to be very small according to noise analysis (Hoth and Penner, 1993; Zweifach and Lewis, 1993; Lepple-Wienhues and Cahalan, 1996).
Similarly, calcium influx can be induced in Xenopus oocytes by agonists that stimulate metabolism of phosphoinositides, by InsP3 and related inositol polyphosphates, and by thapsigargin (Parker et al., 1985; Parker and Miledi, 1987; DeLisle et al., 1995; Petersen and Berridge, 1994). Xenopus oocytes are advantageous in some aspects to study activation mechanisms of the calcium influx. Their giant size facilitates many experimental manipulations. Much data have been accumulated with Xenopus oocytes as model cells to study inositol phosphate-mediated Ca2+ release and Ca2+ homeostasis. The oocytes possess a natural calcium indicator, a calcium-activated chloride current (ICl(Ca)) (Barish, 1983; Miledi and Parker, 1984), with which one can monitor Ca2+ release and Ca2+ influx conveniently. Although ICl(Ca) remains a sensitive approach to detect Ca2+ influx, a quantitative analysis of the Ca2+ influx has been hampered because the relation between ICl(Ca) and calcium influx is complex and incompletely defined (Parker and Yao, 1994). In addition, ICl(Ca) is subject to various modulators, including membrane voltage (Arreola et al., 1996; Hartzell, 1996) and intracellular molecules (Hilgemann, 1995).
To better study Ca2+ influx into oocytes, we explored approaches to record the calcium influx current directly. Injection of Ca2+ chelators was found to be a simple and efficient way to unmask the Ca2+ current by blocking the endogenous ICl(Ca), whereas blockers of anion currents interfered with the Ca2+ current. This Ca2+ current was characterized and shown to be similar in most but not all aspects to ICRAC described in mast cells and lymphocytes (Hoth and Penner, 1992, 1993; Premack et al., 1994; Lewis and Cahalan, 1995). The store-operated Ca2+ current in Xenopus oocytes will be referred to as ISOC to avoid implying that is exactly the same as the previously described ICRAC.
Materials And Methods
Xenopus laevis were purchased from Xenopus I (Ann Arbor, MI), NASCO (Fort Atkinson, WI) and Xenopus Express (Beverley Hill, FL). Several lobes of ovaries were surgically removed from adult females anesthetized with 0.15% 3-aminobenzoic acid ethyl ester (MS-222; Sigma Chem. Co., St. Louis, MO). Oocytes at stages V and VI (Dumont, 1972) were dissected from the ovaries. They were treated with collagenase (0.5–1 mg/ml) at room temperature for 1 h in Barth's medium, which contained (in mM): 88 NaCl, 1 KCl, 2.4 NaHCO3, 0.33 Ca(NO3)2, 0.41 CaCl2, 0.82 MgSO4, 5 HEPES, pH 7.4, supplemented with 50 μg/ml genta-micin (GIBCO BRL, Gaithersburg, MD). Oocytes were released from all external envelopes, except for the vitelline layer, by rolling them on a poly-l-lysine-coated culture dish with a fire-polished glass bar. The oocytes were maintained at 18°C in Barth's medium.
Defolliculated oocytes were placed in a chamber of 200 μl volume and superfused with Ringer solutions. Normal Ringer solution (NR) had the composition (in mM): 95 NaCl, 1 KCl, 1 CaCl2, 1 MgCl2, 5 HEPES, titrated to pH 7.2 with NaOH. Ca2+-free Ringer contained (in mM): 95 NaCl, 1 KCl, 5 MgCl2, 5 HEPES, titrated to pH 7.2 with NaOH. Iso-osmolar MgCl2 medium (Mg70) was used to reduce leak current, which contained (in mM): 70 MgCl2, 10 HEPES, pH 7.2, titrated to pH 7.2 with NaOH. In experiments studying Ca2+ selectivity over monovalent cations, Ringers contained (in mM): 55 XCl, 30 CaCl2, 10 HEPES, X = Na, K, and N-methyl-d-glucamine (NMDG), titrated to pH 7.2 with NaOH, KOH, and HCl, respectively. Cl−-free Ringer was prepared to examine the contribution of Cl− to currents, which contained (in mM): 30 Ca(CH3SO3)2, 55 KCH3SO3, 10 HEPES-Na, pH 7.2. In experiments comparing conductivity of divalent cations, Ringers contained (in mM): 70 MCl2, 10 HEPES-Na, pH 7.2, M = Ca, Sr, and Ba, respectively. In experiments studying dose-dependence of currents on extracellular concentrations of Ca2+, the Ringers had fixed concentration (in mM): 10 MgCl2 and 10 HEPES-Na, pH 7.2, while CaCl2 and NaCl varied in pairs as 1 and 100, 3 and 100, 10 and 90, 30 and 60, 100, and 0 to achieve approximately matched osmolarities.
Membrane currents were measured using a conventional two-electrode voltage clamp (Axoclamp-2B; Axon Instruments, Foster City, CA). Voltage and current electrodes were pulled and filled with 3 M KCl to have a resistance between 0.5–2 MΩ. Current output was lowpass-filtered with an eight-pole Bessel filter (Frequency Devices, Inc., Haverhill, MA) at 200 Hz in most occasions and 1 kHz in recordings applying voltage-step command. Data acquisition and membrane voltage control were performed with a PC software and a data interface (pCLAMP 6.0.2 and Digidata 1200; Axon instruments). Digital format of current traces was exported to a technical graphics and data analysis software (Origin, Microcal Software, Inc., Northampton, MA) for curve fitting and plotting. Recordings were taken from oocytes with input resistances from 0.3 to 2 MΩ. Oocyte membrane potential was held at −60 mV. Intracellular injections were made using a pneumatic pressure ejection device (PV800; WPI, Inc., Sarasota, FL). All experiments were done at room temperature.
Thapsigargin, ionomycin, InsP3, PMA, and calphostin C were purchased from Calbiochem Novabiochem (La Jolla, CA). EGTA (>99% pure) was obtained from Fluka (Buchs, Switzerland). For comparison with EGTA, we used MeBAPTA, a derivative of BAPTA with one extra methyl group in the 5-position, because it was already available in the laboratory (Tsien, 1981; Adams et al., 1988) in larger quantity and greater confidence of purity than commercial BAPTA. Purity was a concern because Parekh and Penner (1995) reported that 10 mM BAPTA (Sigma) blocked ATP-dependent inactivation of ICRAC, possibly by interfering with protein kinase C. In some experiments, we used BAPTA (Molec-ular Probes, Eugene, OR) and obtained similar results as with MeBAPTA. Niflumic acid (Aldrich, Milwaukee, WI) was dissolved in ethanol to yield a stock solution of 250 mM. Thapsigargin, ionomycin, PMA, and calphostin C were dissolved in DMSO. Calphostin C solution was prepared and injected into albino oocytes under dim light. Final DMSO concentrations did not exceed 0.1%.
Direct Recording of ISOC by Abolishing ICl(Ca) with MeBAPTA and EGTA
Most batches of oocytes did not have significant spontaneous Ca2+ influx. To induce Ca2+ influx, oocytes were incubated with 2 μM thapsigargin in Ca2+-free Ringer for over 3 h (Petersen and Berridge, 1994). Alternatively, the Ca2+ influx was induced by InsP3, ionomycin and Ca2+ chelators (see below).
Ca2+ currents are usually masked by endogenous ICl(Ca) in Xenopus oocytes (Barish, 1983; Miledi and Parker, 1984). Therefore, experiments started with different approaches to blocking ICl(Ca). Initially, niflumic acid, a chloride channel antagonist, was tested as a means to inhibit ICl(Ca) (White and Aylwin, 1990; Parekh et al., 1993). A ramp-voltage was applied periodically to allow rapid collection of I-V relations in the range of −120 to +80 mV (see Fig. 1, inset). Adding 10 mM CaCl2 to the extracellular medium elicited ICl(Ca) (Fig. 1,A), as indicated by a typical outwardly rectifying I-V relation and a reversal potential of −19 mV that was close to the Cl− equilibrium potential (Fig. 1,B, b-a). Currents were substantially reduced by 0.5 mM niflumic acid, and the I-V relation of the residual current in the presence of niflumic acid was still almost linear with a reversal potential of −20 mV (Fig. 1 C, c-d). This residual current in the presence of niflumic acid did not appear to be a Ca2+ current, which ought to be inwardly rectifying and to reverse at much more positive potentials.
A different approach using Ca2+ chelators was then explored. Similar ramp-voltage protocols were run regularly to monitor chronological change of I-V relation before and after injection of the buffers. Four nmol EGTA or MeBAPTA were injected at time indicated in Figs. 2 and 3, respectively, resulting in a final concentration of 4 mM assuming a 1-μl oocyte volume. Most ICl(Ca) was blocked by both chelators within a minute after the injections. The reduced current was still mostly carried by Cl− at that time because the I-V relation and reversal potential were close to that before the injection of the chelators (Figs. 2,B and 3,B). Currents induced by extracellular Ca2+ progressively decreased to less than one-thirtieth of the original peak of ICl(Ca) after two min post-injection. The transient phase of ICl(Ca) was completely abolished after that time. The amplitude of the sustained inward current evoked by 10 mM CaCl2 was typically around 90 nA at −60 mV. I-V relations of the leak-subtracted currents obtained from EGTA- and MeBAPTA-loaded oocytes were plotted respectively (Figs. 2,C and 3,C). Typically, I-V curves of the Ca2+ influx-induced current changed drastically ∼4 min after the injection, from outward to inward rectification and from negative to positive reversal potentials (Figs. 2,C, e-g and 3 C, e-g). The I-V relation remained approximately similar after that time but the reversal potential of the current progressively shifted more positive, and the outward current component became even smaller at 8.5 min after the injection of the chelators (Figs. 2 C, f-g, 3 C, f-g). These results indicated that ICl(Ca) was gradually abolished and a Ca2+ current was finally revealed when Ca2+ buffers diffused throughout in the oocytes. Inhibition of ICl(Ca) was dose dependent on Ca2+ chelators. Injection of 0.4 nmol MeBAPTA per oocyte was sufficient to totally block transient ICl(Ca) evoked by 10 mM CaCl2, yet the current reversed at +18 to −12 mV with prominent outward current component at positive membrane potentials (n = 6), indicating contamination by ICl(Ca). More than 1 nmol MeBAPTA and EGTA per oocyte was required to abolish ICl(Ca) completely (1 mM final assuming 1 μl oocyte volume).
MeBAPTA and EGTA showed some differences. The Ca2+ current increased by 140 ± 20% (n = 5) ∼10 min after the injection of MeBAPTA (Fig. 3), whereas it decreased by 10 ± 3% (n = 3) in EGTA-injected oocytes during this period (Fig. 2). Although larger in amplitude, the I-V relation in MeBAPTA-injected oocytes was similar to that in EGTA-loaded oocytes (Figs. 2,C and 3 C), indicating the Ca2+ current was potentiated by MeBAPTA. In addition, input resistance of the oocyte decreased from 1.04 ± 0.13 to 0.87 ± 0.13 MΩ (n = 5) within 10 min after injection of MeBAPTA. The leak conductance was not identified but was inhibited by extracellular Ca2+. In contrast, the input resistance remained stable in EGTA-loaded oocytes (n = 5), suggesting that the leak conductance was associated with an action of MeBAPTA, rather than thapsigargin treatment.
Step-voltage commands were further used to examine the instantaneous I-V relation. ICRAC in mast cells and lymphocytes had a transient peak on a millisecond time scale during hyperpolarization pulses, which resulted from local Ca2+ feedback inhibition (Hoth and Penner, 1992; Zweifach and Lewis, 1995). This transient component of ISOC could be detected readily with a pulse-voltage protocol in thapsigargin-treated oocytes pre-loaded with EGTA. In experiments shown in Fig. 4, membrane potential was held at −60 mV and stepped to +60 mV for 100 ms before hyperpolarizing to various test potentials in increments of −20 mV (see Fig. 4,A, a). The instantaneous inward currents evoked by large hyperpolarizing pulses decayed and reached steady state within ∼20 ms (Fig. 4,B). The transient current peak and plateau current had similar I-V relations (Fig. 4 C). Also the I-V relations obtained with the ramp-voltage command were close to those at steady state.
Dependence of Oocyte ISOC on Extracellular Ca2+ Concentration
We then examined dependence of the Ca2+ current on extracellular Ca2+ concentrations in oocytes treated with thapsigargin and loaded with MeBAPTA. As the CaCl2 concentration in perfusion Ringers was varied between 1 and 100 mM, the size of the Ca2+ current changed correspondingly (Fig. 5,A). The I-V relation of the current was measured at each concentration of CaCl2 to ensure no ICl(Ca) contamination occurred due to possible depletion of local Ca2+ buffers. Intervals between two CaCl2 applications could be minimized as no inactivation of the current was seen at this time scale of recording. The current increased slightly after each previous application of CaCl2 in oocytes loaded with MeBAPTA. Therefore application of CaCl2 started from low to high concentration and then returned in the opposite direction to obtain an average value for further evaluation of dose dependence. A Michaelis-Menten function was used to fit the current amplitude obtained at each extracellular concentration of CaCl2 assuming the oocyte ISOC pathway had no cooperative binding of Ca2+. The best fit yielded an apparent activation constant Km = 11.5 mM for CaCl2 (Fig. 5 B). A similar value of Km = 10.9 mM was obtained in a series of separate experiments, in which CaCl2 was directly added to NR with a similar set of values without compensation of osmolarity.
Ion Selectivity and Conductivity
As the ion composition inside the oocyte was not controlled under our experimental procedures, biionic approaches (Hess et al., 1986) could not be applied to study the ion selectivity of ISOC. In addition, exact measurement of reversal potential of oocyte ISOC was not warranted in our conditions, because the I-V curve approached zero asymptotically, so that apparent reversal potential was strongly affected by choice of leak current for subtraction. One endogenous background current of monovalent cations (Ic) was evident in oocytes in Ca2+ -free Ringer and was inhibited by both extracellular Ca2+ and Mg2+ (Arellano et al., 1995). Thus Ca2+-free Ringer was replaced by Mg70 medium to reduce the leak current induced by removal of Ca2+. Oocytes appeared healthy for at least 20 h in this saline lacking both monovalent cations and Ca2+. Resting potentials and membrane input resistance remained stable during this period. However, the leak current measured in Mg 70 medium appeared still larger than that in the presence of 10 mM or higher extracellular Ca2+, so that the subtracted current showed an artefactual inward current phase at potentials more positive than +30 mV in most oocytes. To reduce interference of the leak conductance, La3+ was used to block ISOC. The current acquired after adding La3+ was then taken as the leak current for subtraction. La3+ is by no means a selective antagonist of ISOC. Like another lanthanide ion, Gd3+ (Arellano et al., 1995), it could additionally inhibit Ic. The La3+-sensitive difference current might thus contain residual Ic that was not blocked by Ca2+. In this case, the current would reverse at a potential less positive than the pure ISOC.
To examine the selectivity of ISOC channels for Ca2+ over Na+ and K+, all extracellular monovalent cations were replaced alternately by pure Na+, K+, or NMDG. Also, the contribution of Cl− was assessed by replacement with CH3SO3−. Recordings were made in thapsigargin-treated and MeBAPTA-loaded oocytes. Currents induced in the above solutions had almost similar amplitudes at a holding potential of −60 mV (Fig. 6,A), indicating that Na+, K+, and Cl− ions did not contribute to the current significantly. Further, I-V relations were obtained in the above solutions (Fig. 6 B). No significant differences in I-V relation were observed when Na+, K+, and Cl− were removed totally from the extracellular medium, indicating that the Ca2+ current was highly selective for Ca2+ over Na+ and K+, and Cl− did not affect the Ca2+ current.
Ca2+ was then replaced by Ba2+ and Sr2+ to examine their permeability through this oocyte ISOC pathway. Membrane currents carried by Ca2+, Sr2+, and Ba2+ were measured at membrane potential of −60 mV (Fig. 7,A). Ca2+ current increased slowly while Ba2+ current decreased with time during the perfusion. In most oocytes (69%, n = 26 for total number of oocytes measured), peak amplitudes of Ca2+, Sr2+, and Ba2+ currents were about equal. In the remaining oocytes (31% of total), Ba2+ current was smaller than Ca2+ and Sr2+ currents. Sr2+ current was smaller than Ca2+ current in 19% of total oocytes measured. Variability in size of Sr2+ or Ba2+ current versus Ca2+ current may result from their poorer buffering by Ca2+ chelators and greater ability to inactivate their own permeability. In the few seconds required for bath turnover, Sr2+ or Ba2+ current might have already been inactivated to various extent. The fast inactivation of Sr2+ and Ba2+ current was studied with the pulse-voltage protocol (Fig. 7, B and C). The I-V relations of instantaneous Sr2+ and Ba2+ currents induced by membrane hyperpolarization steps were monotonic with voltage (Fig. 7, B b, and C b, symbol ×) and similar to those of the Ca2+ current. Yet, the I-V relation obtained at the end of a 200-ms voltage pulse showed a maximum at −80 mV (Fig. 7, B b and C b, symbol ×). Several hypotheses such as direct voltage dependence of blockade might account for the crossover of the current traces, but this phenomenon has not been further explored experimentally.
Inhibitory Action on Oocyte ISOC by Metal Ions and Niflumic Acid
Whereas Sr2+ and Ba2+ permeated this oocyte ISOC pathway readily, some transition metal ions, Ni2+, Co2+, Mn2+, Cd2+, and La+ blocked the current. The inhibition was reversible, allowing effects of all metal ions to be compared in single oocytes (Fig. 8). The Ca2+ current was elicited by 10 mM CaCl2 as control. The oocyte ISOC was reduced by 24 ± 4% (n = 6), 26 ± 4% (n = 6), 27 ± 2% (n = 6), 65 ± 3% (n = 6) of the control by 1 mM Ni2+, Co2+, Mn2+, or Cd2+, respectively. Zn2+ and La3+ blocked ISOC completely at 1 mM, and concentrations for half-inhibition (IC50) were ∼40 μM for Zn2+ (n = 4) and 0.3 μM for La3+ (n = 3), respectively. In experiments evaluating the inhibitory effect of Mg2+, ISOC was recorded in Mg2+-free solution (in mM): (10 CaCl2, 90 NMDG-Cl, 10 HEPES-Na, pH 7.2) and Mg2+-containing solution (Mg70 medium plus 10 mM CaCl2). Mg2+ had a very weak inhibitory effect on ISOC, inhibiting by only 24 ± 4% at 70 mM MgCl2 (n = 5).
The action of niflumic acid on ISOC was tested because it appeared to have side effects on oocyte ISOC pathway in initial experiments. At of 0.5 mM, niflumic acid inhibited ISOC with only partial reversibility (Fig. 9,A). The blocking action of niflumic acid was slow, reaching half-inhibition in ∼3 min (n = 6). Only a small portion of ISOC recovered after washing for as long as 5 min. The blockade appeared direct for oocyte ISOC, rather than due to effects on a contaminating ICl(Ca), because the I-V relation remained similar before and after the inhibition (Fig. 9 B).
Activation of Oocyte ISOC by InsP3, Ionomycin and EGTA
InsP3, ionomycin, and EGTA all deplete intracellular Ca2+ stores by mechanisms different from thapsigargin. We tested whether these agents can also induce oocyte ISOC, as expected if the latter represents capacitative influx. Oocyte membrane potential was held at −60 mV. A ramp-voltage command similar to the inset of Fig. 1 was repetitively applied to monitor change of I-V relation before and after injection of EGTA.
The Ca2+ current induced by InsP3 was illustrated in Fig. 10. A bolus of InsP3 was injected at time marked by an arrow to evoke Ca2+ release, which was indicated by a large ICl(Ca) in Ca2+-free Ringer. No significant Ca2+ influx was induced by 10 mM Ca2+ before the injection of InsP3. About 2 min after the injection of InsP3, Ca2+ influx was clearly activated as indicated by a large ICl(Ca) when bath solution was switched from Ca2+-free Ringer to NR with Ca2+ concentration added to 10 mM. EGTA was subsequently injected to block ICl(Ca). About 1 min after the injection, Ca2+ influx-induced ICl(Ca) appeared to be almost completely abolished, because the reversal potential of the current was positive and the current showed inward rectification (Fig. 10,B, a-c). This relatively quick effect of EGTA was probably due to the local action of InsP3 as the two injection pipettes were close to each other. The Ca2+ current increased by 40% in the following 6 min and outward current component was further suppressed (Fig. 10,B, b-c). The time course of Ca2+ current activation by InsP3 in oocytes pre-injected with MeBAPTA is depicted in Fig. 10 C.
Ionomycin induced Ca2+ influx quickly. Ca2+ influx was not significant before bath application of ionomycin as monitored by switching from Ca2+-free Ringer to NR with 10 mM Ca2+. Ionomycin induced Ca2+ release as indicated by ICl(Ca) of several μA in Ca2+-free Ringer (Fig. 11,A). A sustained Ca2+ influx activity was then recorded for longer than 10 min after washing out bath ionomycin. A typically inwardly rectifying I-V relation of the Ca2+ current was obtained after injection of EGTA (Fig. 11,B, a-b). Full activation of the Ca2+ current in oocytes pre-injected with MeBAPTA took only several min (Fig. 11 C). To ascertain that the ionomycin-induced Ca2+ current was not due to Ca2+ transport by residual ionomycin, the oocytes treated with thapsigargin were exposed to ionomycin for 3 min to examine whether ionomycin could induce an additional Ca2+ current. No significant change of Ca2+ current was measured after incubation of 2 μM ionomycin (n = 4), indicating that ionomycin is an electroneutral carrier and does not mediate significant Ca2+ current by itself. The concentration dependence of the Ca2+ current induced by ionomycin on extracellular Ca2+ and the inhibitory potency of the metal ions were identical to those measured in thapsigargin-treated oocytes (n = 3), which further indicated that the ionomycin-induced Ca2+ current was not carried by ionomycin and was activated at a step downstream to depletion of Ca2+ store.
The time course of the ISOC activation by EGTA was slow and the current size quite variable (Fig. 12). Data acquired from oocytes from three different donors were pooled together to characterize the variability in activation of ISOC by EGTA. Oocyte ISOC appeared to be fully activated in 3 h after the injection of 3–5 nmol EGTA although the final size of the Ca2+ current varied severalfold in oocytes from different frogs (Fig. 12). The leak conductance did not vary systematically during this period (Fig. 12). Little or no Ca2+ current was seen in the first several minutes after the injection of EGTA.
Ion replacement experiments similar to those described above were also performed to examine the Ca2+ currents induced by InsP3, ionomycin and EGTA. No marked difference in ion selectivity could be detected among the Ca2+ currents activated by these different means. No obvious ISOC was found in oocytes incubated for 20 h in Ca2+-free Ringer supplemented with 0.1 mM EGTA (n = 3), although the transient Ca2+-influx-dependent ICl(Ca) described by Petersen and Berridge (1994) was seen. When the Ca2+-free Ringer was temporarily replaced with normal oocyte Ringer, resting potentials were around −50 mV and input resistance was 1 MΩ.
Modulation of ISOC by Protein Kinase C
PMA inhibited ICRAC in RBL-2H3 cells (Parekh and Penner, 1995), while PMA exerted biphasic actions on Ca2+ influx-mediated ICl(Ca) in oocytes, characterized by an initial potentiation and a subsequent inhibition of the current (Petersen and Berridge, 1994). To determine the direct action of kinase C on Ca2+ influx in oocytes, Ca2+ current was measured before and after bath application of the phorbol ester PMA. The only effect of PMA on ISOC was inhibitory in oocytes activated by ionomycin (n = 9) or thapsigargin (n = 19). ISOC declined monotonically with time during perfusion of PMA. The inhibitory rate increased with concentration of PMA. Typically, ISOC was reduced by 49.7 ± 3.4% (n = 8) 6 min after perfusion of PMA 1 μM (Fig. 13 A). The leak conductance was simultaneously decreased but by an unknown mechanism. Calphostin C was used to confirm that the inhibitory effect of PMA resulted from activation of protein kinase C (Kobayashi et al., 1989). Calphostin C was injected into oocytes to reach a final concentration of 2 μM and kept under room fluorescent light for more than 0.5 h before recording started, as recommended by Bruns et al. (1991). ISOC was reduced by only 14.4 ± 2.9% (n = 5) 6 min after bath application of PMA 1 μM in the calphostin C–injected oocytes, indicating that inhibitory action of PMA was mostly blocked. ISOC was 93 ± 15 nA (n = 5) and 104 ± 14 nA (n = 4) in calphostin C–injected and control oocytes, respectively, suggesting that ISOC was not significantly modulated by protein kinase C at the resting state.
This study describes ISOC in Xenopus oocytes using a conventional two-electrode voltage-clamp technique. This method preserves the oocyte cytosol during prolonged recording, in contrast to the whole-cell patch– clamp technique used in other studies (Hoth and Penner, 1992, 1993; Premack et al., 1994; Lewis and Cahalan, 1995). ISOC was isolated by blocking native ICl(Ca) with microinjected Ca2+ chelators (Figs. 2 and 3). Compared with ICRAC described formerly in mast cells and Jurkat lymphocytes (Hoth and Penner, 1992, 1993; Premack et al., 1994; Lewis and Cahalan, 1995), oocyte ISOC had a similar inwardly rectifying I-V relation (Figs. 2–4), high Ca2+ selectivity over Na+ and K+ (Fig. 6), and sequence of inhibitory potency by other ions, Mg2+ << Ni2+ ≈ Co2+ ≈ Mn2+ < Cd2+ << Zn2+ << La3+ (<< represents about one order of magnitude difference or more) (Fig. 8). The oocyte ISOC pathway had similar instantaneous conductances for Ca2+, Sr2+, and Ba2+, yet Ba2+ and Sr2+ currents appeared to inactivate more strongly than Ca2+ current (Fig. 7). By contrast the ICRAC pathway conducts Ca2+ about twice as well as Ba2+ and Sr2+ (Lewis and Cahalan, 1995). In addition, the dependence of ISOC on extracellular CaCl2 concentration had an apparent Km =11.5 mM in the oocytes (Fig. 5), higher than that of ICRAC in mast cells (Km = 3.3 mM; Hoth and Penner, 1993) and Jurkat lymphocytes (Km = 2.1 mM; Premack et al., 1994). Activation of ISOC by EGTA and thapsigargin in oocytes was slow, usually requiring several hours to complete (Fig. 12), which might reflect the slowness of passive leak of Ca2+ from internal stores of oocytes. Total ISOC in oocytes was three to four orders of magnitude larger than that in mast cells and Jurkat cells, which was probably due merely to immense size of oocytes. In terms of current density, 200 nA current in an oocyte corresponds to 1pA pF−1 because an oocyte has a membrane capacitance of 200 nF. Thus the current density in oocytes is roughly similar to that in small mammalian cells. The gigantic size of the oocytes is an attractive feature. It allows microinjection to be easily performed. Besides, direct recording of ISOC should be possible on a giant membrane patch with size of around 30 mm (Hilgemann, 1995). This may offer the opportunity to separate the plasma membrane ISOC pathway from the internal Ca2+ store and to test the putative diffusible messengers directly in an excised membrane patch configuration (Randriamampita and Tsien, 1993; Kim et al., 1995).
Recently a putative store-operated current in an oocyte injected with InsP3 and BAPTA was revealed (Fig. 14 of Hartzell, 1996) as the difference between currents in 0 and 10 mM Ca. Although the reversal potential could not be determined and no other analysis was presented, the amplitude and inward rectification were roughly similar to those presented here (e.g., Fig. 3, f-g and Fig. 10, b-c), suggesting that we are studying the same pathway that Hartzell (1996) first detected.
Similarities and Differences Among Niflumic Acid, EGTA and MeBAPTA as Blockers of ICl(Ca)
Parekh et al. (1993) reported a niflumic acid–resistant current in cell-attached membrane patches in oocytes stimulated by serotonin. Similar properties of niflumic acid-resistant current were observed in this whole-cell study. This current reversed at about −20 mV and had a prominent outward current component at positive membrane potentials. These characteristics of the current were attributed to increases of both Ca2+ and K+ permeability (Parekh et al., 1993). However, Ca2+ chelator-loaded oocytes in the present study gave quite different results. The current activated by depletion of Ca2+ store showed little outward current, no K+ permeability increase, and high calcium selectivity over monovalent cations. While the exact ionic components of niflumic acid–resistant current remain to be studied more thoroughly, we feared that niflumic acid was not a specific antagonist for ICl(Ca) because this drug also inhibited oocyte ISOC irreversibly (Fig. 9). Niflumic acid has been reported to have multiple actions on different ionic pathways. It was used originally to block anion transporters (Cousin and Motais, 1979), but also found to affect one K current (Busch et al., 1994) expressed in the oocytes, and to block ICRAC in rat basophilic leukemia 2H3 cells (Reinsprecht et al., 1995).
The calcium chelators block ICl(Ca) not by direct channel blockade but by binding and removing Ca2+, so that free Ca2+ ions appearing at calcium channel pores are closely confined without spreading to activate Cl(Ca) channels (Roberts, 1993). Ca2+ chelation has likewise helped unmask voltage-gated Ca2+ currents from exogenous channels in oocytes (Charnet et al., 1994). To totally saturate 4 nmol buffer, the amount typically injected into the oocytes (Btotal), a sustained Ca2+ current (I) of 200 nA should last for time t = Btotal · z · F/I = 4 × 10−9 mol · 2 · 9.65 × 104 coul · mol−1/(2 × 10−7A) ≈ 4,000 s, where z is the valence of Ca2+ and F is Faraday's constant. Therefore sustained recording of ISOC in oocytes should be possible as long as local depletion of Ca2+ buffers does not occur.
The slow buffer, EGTA, was found to have no effect at 1 mM on Ca2+-activated K+ current in saccular hair cells (Roberts, 1993), while a similar dose of EGTA strongly inhibited ICl(Ca) in oocytes. This might arise from the smaller unitary conductance of CRAC or SOC channels and possibly larger distance between Cl(Ca) channels and SOC channels in oocytes. ISOC was found to be consistently larger in oocytes injected with MeBAPTA than that with EGTA. Also, in oocytes buffered with MeBAPTA, ISOC did not instantly level off within each exposure to high extracellular Ca2+ (see Figs. 3, 5–9), whereas in oocytes injected with EGTA, ISOC immediately reached a flat plateau within each pulse (see Figs. 2 and 11). These differences might result from the difference in Ca2+ binding kinetics of the buffers, because local feedback inhibition by Ca2+ on SOC channels is better attenuated by BAPTA due to its faster binding rate (Zweifach and Lewis, 1995). Inhibition of tonically active protein kinase C (Parekh and Penner, 1995) does not readily explain the enhancement of ISOC by MeBAPTA, because calphostin C did not mimic MeBAPTA, though this kinase blocker could inhibit PMA effects (Fig. 13). The slight increase in leak current induced by MeBAPTA was not desirable especially when a prolonged recording of ISOC was needed. EGTA would then be the better choice in this case, because neither ISOC nor the membrane leak was significantly affected by this chelator.
Relationship Between ISOC and Ca2+ Influx-mediated ICl(Ca) in Oocytes
ICl(Ca) has been widely used to indicate Ca2+ influx activity in the oocytes. It is a sensitive measure of Ca2+ influx activity, as peak ICl(Ca) is more than 10 times larger than the underlying calcium influx current (see Figs. 2 and 3). However, the relation between ICl(Ca) and ISOC is not simple. Several discrepancies are obvious. First, ICl(Ca) evoked by Ca2+ influx has an initially large transient component that rises and decays in hundreds of milliseconds to several seconds, followed by a relatively sustained component. Second, the Ca2+ entry-dependent transient ICl(Ca) is a highly nonlinear function of membrane hyperpolarization and extracellular Ca2+ concentration (Parker et al., 1985; Petersen and Berridge, 1994). Third, most of the Ca2+ influx-induced transient ICl(Ca) inactivates and recovers in about one minute. Although different mechanisms were proposed, approximately similar characteristics of the ICl(Ca) were also observed in oocytes that were either injected with InsP3 (Yao and Parker, 1993) or incubated with calcium ionophores, A23187 (Boton et al., 1989) and ionomycin (Yao, Y., unpublished data). Thus, in oocytes injected with InsP3 or stimulated with agonists, Ca2+ influx promoted by hyperpolarization pulses evoked a large Ca2+-dependent transient ICl(Ca), Tin (Parker et al., 1985). Tin appeared to rise and decay more rapidly than the transient ICl(Ca) in thapsigargin- and the ionophore-treated oocytes. In addition, Tin appeared to have a different onset shape. Simultaneous recording of ICl(Ca) and Ca2+ fluorescence showed that Tin reflects InsP3-dependent Ca2+-induced Ca2+ release (Yao and Parker, 1993). Consistent with this, the hump component of the ICl(Ca) could be elicited by membrane depolarization in oocytes expressing voltage-gated Ca2+ channels together with InsP3 application (Yao and Parker, 1992). This indicated that transient ICl(Ca) does not require Ca2+ influx via the ISOC pathway per se. The InsP3-dependent Ca2+-induced Ca2+ release mechanism, however, fails to explain the transient ICl(Ca) induced by Ca2+ influx in thapsigargin-treated oocytes since Ca2+ store had been depleted, as indicated by the lack of further Ca2+ release in response to InsP3 (Petersen and Berridge, 1994) and ionomycin (Yao, Y., unpublished observation). To explain the supralinear relation between transient ICl(Ca) and extracellular Ca2+ concentration or membrane hyperpolarization, a positive feedback regulation at the level of the Ca2+ influx pathway has been proposed (Petersen and Berridge, 1994). However, no sign of regenerativity was seen with oocyte ISOC in this study. Besides, injection of as little as 120 pmol slow buffer EGTA (final 120 μM assuming 1 μl oocyte volume) was sufficient to totally block the transient ICl(Ca) evoked by 10 mM extracellular CaCl2 in thapsigargin-treated oocytes (Y. Yao, unpublished data). This suggests that any regenerativity lies between Ca2+ influx and transient ICl(Ca), rather than between stores depletion and Ca2+ influx.
Alternatives to regenerativity to interpret the nonlinearity of Ca2+ influx-induced ICl(Ca) should be considered. First, diffusion and buffering processes could contribute to the nonlinearity. Local free [Ca2+] profiles are dependent on intensity of Ca2+ current source and Ca2+ buffers. Mobility, binding and dissociation kinetics and concentration of the endogenous Ca2+ buffers are all critical variables to shape local [Ca2+]. Qualitatively, with a small Ca2+ influx, the endogenous buffers would be sufficient to bind and remove Ca2+ so that no or low free Ca2+ could reach Cl(Ca) channels and hence no or small ICl(Ca). With large influx of Ca2+, the Ca2+ buffers would be depleted locally, so that more free Ca2+ would spread to Cl(Ca) channels to cause a large ICl(Ca), probably in a nonlinear manner. Second, ICl(Ca) appears to be an increment detector of cytosolic Ca2+ because ICl(Ca) was found to correspond to the rate of rise of intracellular free Ca2+ rather than to its steady state levels of oocytes (Parker and Yao, 1994). The initial transient of ICl(Ca) might thus result from a large rate of rise of cytosolic Ca2+ at the beginning of Ca2+ influx.
To conclude, ICl(Ca) is a quantitatively unreliable measure of the underlying Ca2+ current as the relation between them is quite complex. Further detailed studies on ICl(Ca) channels (Hartzell, 1996) and diffusion-buf-fering processes will further increase the complexity. Therefore, direct measurement of ISOC in oocytes should greatly facilitate quantitative and molecular analysis of capacitative Ca2+ entry mechanisms.
We would like to thank Dr. Ian Parker and Dr. Pierre Vincent for their critical discussion and comments on the manuscript. We appreciate Dr. Stephen Adams for providing MeBAPTA.
This study was supported by grant RG520/95 from the Human Frontier Science Program and by the Howard Hughes Medical Institute.
Abbreviations used in this paper: BAPTA, 1,2-bis(2-aminophenoxy)ethane-N,N,N ′,N′-tetraacetic acid; MeBAPTA, 1-(2-amino-5-methylphenoxy)-2-(2-aminophenoxy)ethane-N,N,N′,N ′-tetraacetic acid; NMDG, N -methyl-d-glucammonium; InsP3, inositol 1,4,5-trisphosphate; PMA, phorbol-12-myristate-13-acetate; NR, normal Ringer; ICRAC, mammalian calcium release-activated calcium current; ISOC, Xenopus stores-operated calcium current.
Address correspondence to Dr. Roger Y. Tsien, Department of Pharmacology/HHMI, CMM-West, Room 310, University of California San Diego, La Jolla, CA 92093-0647. Fax: 619-534-5270. E-mail: email@example.com