Dissociation of the store-operated calcium current I_CRAC and the Mg-nucleotide-regulated metal ion current MagNuM

Abstract

Rat basophilic leukaemia cells (RBL-2H3-M1) were used to study the characteristics of the store-operated Ca²⁺ release-activated Ca²⁺ current (I_CRAC) and the magnesium-nucleotide-regulated metal cation current (MagNuM) (which is conducted by the LTRPC7 channel). Pipette solutions containing 10 mm BAPTA and no added ATP induced both currents in the same cell, but the time to half-maximal activation for MagNuM was about two to three times slower than that of I_CRAC. Differential suppression of I_CRAC was achieved by buffering free [Ca²⁺]_i to 90 nm and selective inhibition of MagNuM was accomplished by intracellular solutions containing 6 mm Mg.ATP, 1.2 mm free [Mg²⁺]_i or 100 μm GTP-γ-S, allowing investigations on these currents in relative isolation. Removal of extracellular Ca²⁺ and Mg²⁺ caused both currents to be carried significantly by monovalent ions. In the absence or presence of free [Mg²⁺]_i, I_CRAC carried by monovalent ions inactivated more rapidly and more completely than MagNuM carried by monovalent ions. Since several studies have used divalent-free solutions on either side of the membrane to study selectivity and single-channel behaviour of I_CRAC, these experimental conditions would have favoured the contribution of MagNuM to monovalent conductance and call for caution in interpreting results where both I_CRAC and MagNuM are activated.

Store-operated or capacitative calcium entry is a widespread mechanism used by electrically non-excitable cells to support sustained elevations in intracellular calcium concentration (Berridge, 1995; Lewis & Cahalan, 1995; Thomas et al. 1996; Parekh & Penner, 1997). It is mediated by store-operated ion channels in the plasma membrane that are activated through an unknown mechanism following InsP₃-mediated release of Ca²⁺ from intracellular stores. Experimentally, any manoeuvre that causes store depletion through either active Ca²⁺ release (e.g. InsP₃, ionomycin) or passive release by impeding refilling (e.g. thapsigargin, intracellular Ca²⁺ buffers) will activate these channels (Hoth & Penner, 1992; Zweifach & Lewis, 1993; Parekh & Penner, 1996; Broad et al. 1999). The prototypical and best-characterized store-operated mechanism is the so-called ‘calcium release-activated calcium’ current (I_CRAC) (Hoth & Penner, 1992, 1993; Zweifach & Lewis, 1993; Lewis & Cahalan, 1995; Parekh & Penner, 1997). This current is characterized by its high selectivity for Ca²⁺ ions, an inwardly rectifying current-voltage relationship, and an extremely small single-channel conductance in the sub-picoSiemens range (Hoth & Penner, 1993; Zweifach & Lewis, 1993). Even whole-cell currents carried by CRAC channels are very small and specific experimental conditions need to be established in order to reliably measure I_CRAC (Hoth & Penner, 1992; Parekh & Penner, 1996). I_CRAC can be carried by monovalent ions when all divalent ions are removed from the extracellular solution (Hoth & Penner, 1993; Kerschbaum & Cahalan, 1998). Recently, several investigations have taken advantage of this feature to arrive at the first single-channel recordings through CRAC channels (Kerschbaum & Cahalan, 1999; Braun et al. 2000; Fomina et al. 2000). It was found that single-channel conductance of CRAC channels carrying monovalent ions is around 30–40 pS and the number of channels per cell was estimated to be around 100–400 in Jurkat lymphocytes (Kerschbaum & Cahalan, 1999), 260–500 in RBL cells (Braun et al. 2000) and as few as 15 in resting human lymphocytes (Fomina et al. 2000). These investigations suggest that CRAC channels can now be studied at the single-channel level.

We have recently cloned and characterized LTRPC7, a novel ion channel of the LTRPC family of ion channels (Nadler et al. 2001). LTRPC7 is regulated by intracellular levels of Mg²⁺-complexed nucleotides and underlies a current that is strongly activated when Mg.ATP falls below 1 mm (designated MagNuM for magnesium-nucleotide-regulated metal current). LTRPC7 is a widely expressed ion channel that conducts Ca²⁺ and Mg²⁺ at negative membrane potentials. However, at potentials above +50 mV, where the driving force for divalent ions to enter the cell is small, the permeation block exerted by divalent ions is reduced and MagNuM is effectively carried by monovalent cations moving in the outward direction, a property that provides a signature for identifying LTRPC7. The permeation block of divalent ions at negative membrane potentials can be removed by omitting divalent ions from the extracellular solution, and under these conditions, MagNuM, like I_CRAC, can significantly be carried by monovalent ions moving inwards. Since I_CRAC and MagNuM are both present in lymphocytes and RBL cells (Nadler et al. 2001), this provides for some ambiguity in whole-cell and single-channel studies of I_CRAC, as those have been performed under experimental conditions that would also favour activation of MagNuM. We have therefore carried out experiments in RBL cells to assess the relative contributions of I_CRAC and MagNuM to monovalent cation currents by selectively dissociating the currents based on their differences in activation mechanism, Mg²⁺ dependence and pharmacology. We conclude that many of the reported properties ascribed to I_CRAC carried by monovalent ions may require re-evaluation, since MagNuM contributes significantly to membrane currents under experimental conditions previously thought to enhance I_CRAC carried by monovalent ions alone.

METHODS

For patch-clamp experiments, RBL-2H3-M1 cells (a clonal mast cell stably transfected with the muscarinic M1 receptor) grown on glass coverslips were transferred to the recording chamber and kept in a standard modified Ringer solution of the following composition (mm): NaCl 145, KCl 2.8, CsCl 10, CaCl₂ 10, MgCl₂ 2, glucose 10, Hepes-NaOH 10, pH 7.2. In some experiments, the above solution was temporarily replaced by an otherwise identical extracellular solution in which Ca²⁺ was omitted (0 Ca²⁺) and in others both Ca²⁺ and Mg²⁺ were omitted and 2 mm EDTA was added (divalent-free, DVF). Solution changes were performed by pressure ejection from a wide-tipped pipette. In experiments reported in Fig. 1, external Ca²⁺ concentration was 1 mm throughout. Intracellular pipette-filling solutions contained (mm): caesium glutamate 145, NaCl 8, caesium BAPTA 10, Hepes-CsOH 10, pH 7.2. This solution was considered divalent free. Mg.ATP and/or MgCl₂ were added as indicated in the figure legends and free [Mg²⁺]_i was calculated by WEBMAXC v. 2.10 (http://www.stanford.edu/~cpatton/webmaxc2.htm). In all experiments that required buffering of free [Ca²⁺]_i to 90 nm, BAPTA was replaced by a mixture of 10 mm EGTA and 3.6 mm CaCl₂.

Figure 1. Differential activation time course of I_CRAC and MagNuM.

A, average inward (•) and outward (○) I_CRAC and MagNuM at −80 and +80 mV, respectively (n = 8). I_CRAC and MagNuM were activated passively by omission of ATP and inclusion of 10 mm BAPTA in the pipette solution ([MgCl₂]_i = 1 mm, free [Mg²⁺]_i = 780 μm, extracellular [Ca²⁺] = 1 mm). The dotted line represents the inverted and scaled outward current to illustrate the difference in activation time course. B, current-voltage relationship derived from a high-resolution current record in response to a voltage ramp of 50 ms duration that ranged from −100 to +100 mV, taken at 300 s from a representative cell under experimental conditions described in A. Arrows indicate −80 and +80 mV. Inward currents are predominantly I_CRAC and outward currents are predominantly MagNuM. C, average inward (•) and outward (○) I_CRAC and MagNuM at −80 and +80 mV, respectively (n = 3). Experimental conditions and solutions as in A, except that 20 μm InsP₃ was added to the pipette solution to accelerate activation of I_CRAC. The dotted lines represent the scaled inward and outward currents depicted in A to illustrate the selective acceleration of I_CRAC activation time course. D, average inward (•) and outward (○) I_CRAC and MagNuM at −80 and +80 mV, respectively (n = 5). Experimental conditions and solutions as in A, except that the pipette solution lacked MgCl₂. The dotted line represents the scaled outward current taken from A to illustrate the difference in MagNuM activation time course.

Patch-clamp experiments were performed in the tight-seal whole-cell configuration at 24 ± 2 °C. High-resolution current recordings were acquired by a computer-based patch-clamp amplifier system (EPC-9, HEKA, Lambrecht, Germany). Patch pipettes had resistances between 2 and 4 MΩ after filling with the standard intracellular solution. Immediately following establishment of the whole-cell configuration, voltage ramps of 50 ms duration spanning the voltage range of −100 to +100 mV were delivered from a holding potential of 0 mV at a rate of 0.5 Hz over a period of 200 to 400 s. All voltages were corrected for a liquid junction potential of 10 mV between external and internal solutions. Currents were filtered at 2.3 kHz and digitized at 100 μs intervals. Capacitive currents and series resistance were determined and corrected before each voltage ramp using the automatic capacitance compensation of the EPC-9. For analysis, the very first ramps prior to current activation were digitally filtered at 2 kHz, pooled and used for leak subtraction of all subsequent current records. The low-resolution temporal development of currents at a given potential was extracted from the leak-corrected individual ramp current records by measuring the current amplitudes at voltages of −80 and +80 mV. Effects of DVF solutions on background currents other than I_CRAC and MagNuM were negligible (0.52 ± 0.13 pA pF⁻¹), as determined in experiments in which neither current was allowed to activate (i.e. [Ca²⁺]_i was buffered to 90 nm and 6 mm Mg.ATP was added to the intracellular solution).

Ionomycin and GTP-γ-S were purchased from Sigma. 2-Aminoethoxydiphenyl borate (2-APB) was kindly provided by Dr Mikoshiba (Department of Molecular Neurobiology, Tokyo University, Tokyo, Japan).

Where applicable, data are given as means ± s.e.m. with n determinations and statistical significance was assessed by Student's t test.

RESULTS

Activation kinetics of I_CRAC and MagNuM

In a first series of experiments, we established conditions that would cause activation of both I_CRAC and MagNuM. This was achieved by using a standard intracellular solution supplemented with 10 mm BAPTA and devoid of added ATP. The extracellular solution contained 1 mm Ca²⁺, which is sufficient to carry significant current through CRAC channels without unduly suppressing outward MagNuM currents. Immediately following break-in, voltage ramps from −100 to +100 mV were delivered at a rate of 0.5 Hz to continuously monitor membrane currents.

As shown in Fig. 1A, both inward and outward currents (measured at −80 and +80 mV, respectively) developed over 100–300 s. The characteristic current-voltage relationship of the conductance under these conditions is seen in Fig. 1B, which illustrates a high-resolution current record in response to a voltage ramp recorded at 300 s in a representative experiment. Under these experimental conditions, at least initially, the inward current at −80 mV is predominantly I_CRAC, which activates passively as BAPTA prevents reuptake of Ca²⁺ into the stores. The outward current at +80 mV is primarily MagNuM, which activates with a delay as cytosolic ATP decreases due to diffusional exchange between the cytosol and the ATP-free pipette solution. Given that there is also an inward MagNuM current, it is likely that the inward current progressively reflects a mixture of I_CRAC and MagNuM components. Kinetic analysis yields a time to half-maximal activation (t_1/2) of 91 ± 8 s for I_CRAC and 220 ± 6 s for MagNuM (n = 8), suggesting that the two current components are carried by distinct sets of ion channels.

Since I_CRAC is activated by store depletion and MagNuM is activated by Mg.ATP depletion, we reasoned that under otherwise identical conditions, the inclusion of InsP₃ in the pipette solution, due to its active Ca²⁺ release activity, would selectively accelerate store-dependent I_CRAC with little or no effect on the MagNuM time course. As illustrated in Fig. 1C this is indeed the case. The half-maximal activation times in the presence of 20 μm InsP₃ were 31 ± 7 s for I_CRAC and 200 ± 15 s for MagNuM (n = 3). The acceleration of inward I_CRAC was statistically significant (P < 0.002), whereas time courses of outward currents were not significantly different. This confirms that inward currents carried by I_CRAC are store dependent, whereas outward currents are not.

Since both I_CRAC and MagNuM are regulated by [Mg²⁺]_i, we carried out experiments similar to those of Fig. 1A, but omitted Mg²⁺ from the pipette solutions (Fig. 1D). This resulted in a large increase in outward current amplitude and a smaller increase in inward current, suggesting that MagNuM is more susceptible to [Mg²⁺]_i regulation than I_CRAC. In terms of kinetics, t_1/2 was 84 ± 6 s and was not significantly different from that observed in Fig. 1A, where free [Mg²⁺]_i was calculated to be 780 μm. However, the time course of MagNuM was accelerated under [Mg²⁺]_i-free conditions with t_1/2 = 151 ± 3 s (n = 5), and this shift was statistically significant (P < 1 × 10⁻⁶) with respect to the time course of MagNuM shown in Fig. 1A. We attribute this acceleration to a faster decrease in cytosolic Mg.ATP levels when perfusing cells with [Mg²⁺]_i-free intracellular solutions.

Dissociation of I_CRAC and MagNuM by selective inhibition of activation

Based on their different activation mechanisms, we designed experimental conditions that would enable us to activate either of the two conductances in isolation. Since MagNuM is regulated by Mg.ATP, we performed experiments under similar conditions to Fig. 1, but included 6 mm Mg.ATP to suppress MagNuM (Fig. 2A). Under these conditions, cells developed an inward current that had the current-voltage (I-V) signature of I_CRAC, whereas outward MagNuM currents were absent (Fig. 2B). Thus, the presence of high levels of Mg.ATP selectively suppressed MagNuM and unmasked I_CRAC.

Figure 2. Selective activation of I_CRAC and MagNuM.

A, average inward (•) and outward (○) currents at −80 and +80 mV, respectively (n = 4), recorded under experimental conditions that suppress MagNuM and favour I_CRAC activation. I_CRAC was activated passively by inclusion of 10 mm BAPTA in the pipette solution and MagNuM was suppressed by 6 mm Mg.ATP ([MgCl₂]_i = 0, free [Mg²⁺]_i = 550 μm). B, I-V relationship of I_CRAC derived from high-resolution current records in response to voltage ramps of 50 ms duration that ranged from −100 mV to +100 mV. Data are taken from a representative cell under experimental conditions described in A. C, average inward (•) and outward (○) currents at −80 and +80 mV, respectively (n = 7), recorded under experimental conditions that suppress I_CRAC and favour MagNuM activation. MagNuM was activated passively by omission of ATP from the pipette solution and I_CRAC was suppressed by buffering [Ca²⁺]_i to 90 nm using 10 mm EGTA and 3.6 mm CaCl₂ ([MgCl₂]_i = 1 mm, free [Mg²⁺]_i = 760 μm). D, I-V relationship of MagNuM derived from high-resolution current records measured with the same pulse protocol as in B. Data are taken from a representative cell under experimental conditions described in C. E, average inward (•) and outward (○) currents at −80 and +80 mV, respectively (n = 5), recorded under experimental conditions that suppress I_CRAC and favour MagNuM activation. MagNuM was activated passively by omission of ATP from the pipette solution and I_CRAC was suppressed by buffering [Ca²⁺]_i to 90 nm using 10 mm EGTA and 3.6 mm CaCl₂ (free [Mg²⁺]_i = 0). I_CRAC was activated by brief application (2–3 s) of 20 μm ionomycin applied at the time indicated by the arrow. F, I-V relationships derived from high-resolution current records, measured as in B. Data are taken from a representative cell under experimental conditions described in E before ionomycin application (300 s, thin line) and after I_CRAC had fully developed (400 s, thick line).

In order to activate MagNuM in isolation, we omitted ATP from the pipette solution and suppressed I_CRAC by clamping cytosolic free Ca²⁺ to 90 nm using appropriate mixtures of EGTA and Ca²⁺. As can be seen in Fig. 2C, this resulted in selective activation of predominantly outward currents with the I-V signature of MagNuM (Fig. 2D). Note, however, that MagNuM also has an inward component, which develops in parallel with the outward component.

Based on the differential kinetics and the selective suppression of currents, the experiments so far suggest that the major portions of inward and outward currents are carried by different ion channel mechanisms, namely I_CRAC and MagNuM, respectively. However, one might argue that these differences are due to changes in conductance properties of only one mechanism, namely I_CRAC, and that MagNuM reflects an unusual behaviour of outward currents through CRAC channels. We have therefore performed experiments in which we first allowed MagNuM to fully activate under experimental conditions that suppress I_CRAC and then applied the Ca²⁺ ionophore ionomycin to actively induce store depletion. As illustrated in Fig. 2E, this resulted in a selective increase in inward current without a corresponding increase in outward current. The I-V relationships of cells prior to and following ionomycin-induced store depletion, illustrated in Fig. 2F, demonstrate in more detail the selective recruitment of an inwardly rectifying conductance with the signature of I_CRAC. The additivity of I_CRAC and MagNuM establishes, without reasonable doubt, that the two conductances are indeed distinct entities.

Monovalent I_CRAC and MagNuM

I_CRAC is a highly specific Ca²⁺ current, but can support a large monovalent current when divalents are removed from the extracellular solution (Hoth & Penner, 1993; Kerschbaum & Cahalan, 1998). Figure 3A illustrates this phenomenon under conditions where MagNuM has been suppressed by inclusion of 6 mm intracellular Mg.ATP. Under these conditions I_CRAC, but not MagNuM, activates due to passive depletion of Ca²⁺ from intracellular stores. Upon removal of divalents, there is a transient increase in inward I_CRAC as previously described (Hoth & Penner, 1993). Note, however, that outward currents do not increase proportionally so that inward rectification of I_CRAC is preserved (Fig. 3B). Also note the rather strong inactivation of the monovalent inward current, whose underlying mechanism is at present not understood but has been proposed to be related to the presence of [Mg²⁺]_i (Kerschbaum & Cahalan, 1998). However, as will be shown below, our data do not support this interpretation.

Figure 3. Monovalent I_CRAC and MagNuM.

A, average inward (•) and outward (○) currents at −80 and +80 mV, respectively (n = 3), recorded under experimental conditions that suppress MagNuM and favour I_CRAC activation. I_CRAC was activated passively by inclusion of 10 mm BAPTA in the pipette solution and MagNuM was suppressed by 6 mm Mg.ATP ([MgCl₂]_i = 1 mm, free [Mg²⁺]_i = 1.1 mm). Cells were perfused with divalent-free (DVF) extracellular solution for the time indicated by the bars. B, I-V relationship derived from a high-resolution current record in response to a voltage ramp of 50 ms duration that ranged from −100 to +100 mV, taken at the peak of monovalent inward current (206 s) from a representative cell under experimental conditions described in A. Note the strong inward rectification of monovalent I_CRAC. C, average inward (•) and outward (○) currents at −80 and +80 mV, respectively (n = 4), recorded under experimental conditions that suppress I_CRAC and favour MagNuM activation. MagNuM was activated passively by omission of ATP from the pipette solution and I_CRAC was suppressed by buffering [Ca²⁺]_i to 90 nm using 10 mm EGTA and 3.6 mm CaCl₂ ([MgCl₂]_i = 1 mm, free [Mg²⁺]_i = 760 μm). Cells were perfused with DVF extracellular solution for the time indicated by the bars. D, I-V relationship derived from a high-resolution current record in response to a voltage ramp as detailed in B, taken at the peak of monovalent inward current (390 s) from a representative cell under experimental conditions described in C. Note the linear I-V relationship of monovalent MagNuM currents.

We have previously shown that MagNuM can also sustain a large monovalent current upon removal of all divalent ions from the extracellular solution (Nadler et al. 2001). Figure 3C illustrates this under experimental conditions where buffering of [Ca²⁺]_i to 90 nm suppresses I_CRAC activation, while omission of ATP selectively leads to MagNuM activation. During application of divalent-free extracellular solution, there is a large increase in inward MagNuM current, but in contrast to the behaviour of I_CRAC, there is also a significant increase in outward current even at potentials below +50 mV, which we interpret to be due to the removal of permeation block of divalent ions. As outward currents are no longer impeded by divalent ions, the MagNuM I-V relationship becomes almost linear (Fig. 3D). Furthermore, monovalent MagNuM currents also exhibit some inactivation, most prominently of the outward current, albeit to a lesser degree than the inactivation of monovalent I_CRAC.

Effects of intracellular Mg²⁺ on I_CRAC and MagNuM

The majority of experiments presented so far were carried out under conditions where intracellular Mg²⁺ concentrations were kept in the physiological range of 0.5–0.8 mm (Romani & Scarpa, 2000). Previous studies of I_CRAC have suggested that [Mg²⁺]_i contributes to inactivation and inward rectification of I_CRAC through a voltage-dependent block of outward currents (Kerschbaum & Cahalan, 1998). We have therefore attempted to directly compare monovalent I_CRAC and MagNuM under similar free [Mg²⁺]_i. The left panel of Fig. 4A illustrates that I_CRAC carried by Ca²⁺, in this case activated by 20 μm InsP₃, activates rapidly and maximally, while MagNuM remains suppressed by 6 mm Mg.ATP. Upon switching to divalent-free extracellular solution, there is significant increase in inward current that rapidly inactivates to a steady-state level that is similar to the current amplitude seen with Ca²⁺ as charge carrier. As illustrated in the right panel of Fig. 4A, the same level of 1.1 mm free [Mg²⁺]_i largely suppressed activation of MagNuM and consequently there is little monovalent current when establishing divalent-free conditions.

Figure 4. Monovalent I_CRAC and MagNuM in reduced intracellular Mg²⁺.

In all panels, average inward (•) and outward (○) currents at −80 and +80 mV were acquired under defined free [Mg²⁺]_i. Left panels represent experimental conditions that suppress MagNuM and favour I_CRAC activation by inclusion of 6 mm Mg.ATP, 20 μm InsP₃ and 10 mm BAPTA in the pipette solution, whereas right panels represent conditions in which I_CRAC was suppressed by buffering [Ca²⁺]_i to 90 nm using 10 mm EGTA and 3.6 mm CaCl₂. MagNuM developed by omission of ATP. In all experiments, appropriate amounts of MgCl₂ were added to arrive at defined [Mg²⁺]_i, as indicated above each pair of panels. Cells were perfused with divalent-free extracellular solution for the times indicated by the bars. A, left panel: average responses of five cells ([MgCl₂]_i = 1 mm, free [Mg²⁺]_i = 1.1 mm). Right panel: average responses of nine cells ([MgCl₂]_i = 1.4 mm, free [Mg²⁺]_i = 1.1 mm). B, left panel: average responses of four cells ([MgCl₂]_i = 0, free [Mg²⁺]_i = 550 μm). Right panel: average responses of eight cells ([MgCl₂]_i = 740 μm, free [Mg²⁺]_i = 550 μm). C, left panel: average responses of five cells in [Mg²⁺]_i-free conditions where ATP and MgCl₂ were omitted from the pipette solution and MagNuM was suppressed by 100 μm GTP-γ-S. I_CRAC was activated by 20 μm InsP₃. Right panel: average responses of nine cells in [Mg²⁺]_i-free conditions. D, average responses of four cells in [Mg²⁺]_i-free conditions where ATP and MgCl₂ were omitted from the pipette solution. Therefore, both InsP₃-mediated I_CRAC and ATP-dependent MagNuM were activated. The dotted trace represents data taken from the left panel of C to illustrate the contribution of I_CRAC. E, I-V relationships derived from high-resolution current records in response to a voltage ramp of 50 ms duration that ranged from −100 to +100 mV, taken at the peak of monovalent inward currents from representative cells under experimental conditions described in C. Note the linear I-V relationship of monovalent MagNuM currents and inward rectification of I_CRAC.

We next reduced the free [Mg²⁺]_i to 550 μm, which represents the minimal concentration achievable when using 6 mm Mg.ATP to suppress MagNuM and study I_CRAC in isolation. Figure 4B illustrates that under these conditions, InsP₃-induced I_CRAC carried by Ca²⁺ is not affected in any significant way compared to 1.1 mm free [Mg²⁺]_i. When switching to divalent-free solution, there is a large increase in inward current, which, as with 1.1 mm [Mg²⁺]_i, inactivates rapidly and strongly. Thus, the reduction in [Mg²⁺]_i increases the monovalent current through I_CRAC, but has little impact on inactivation. The corresponding experiment, in which MagNuM was activated by ATP depletion while [Mg²⁺]_i was fixed to 550 μm and [Ca²⁺]_i was buffered to 90 nm to avoid activation of I_CRAC, is illustrated in the right panel of Fig. 4B. It demonstrates that exposing cells to divalent-free extracellular solution greatly amplifies monovalent MagNuM currents, confirming that [Mg²⁺]_i regulates this conductance as well. The magnitude of monovalent inward MagNuM current is about half of that of I_CRAC, but exhibits less-pronounced inactivation.

A further increase in monovalent MagNuM currents was achieved by omitting Mg²⁺ from the intracellular solution, as illustrated in the right panel of Fig. 4C. It should be noted that in order to suppress I_CRAC, the intracellular solution was buffered to 90 nm free [Ca²⁺]_i, and this presumably limited the maximal amount of monovalent MagNuM current, since [Ca²⁺]_i causes some inhibition of MagNuM (Nadler et al. 2001). Nevertheless, it is clear that omitting [Mg²⁺]_i significantly increases monovalent inward MagNuM currents.

Unfortunately, it is not possible to completely remove all free Mg²⁺ from the intracellular solution and measure I_CRAC in isolation from MagNuM, since MagNuM suppression requires Mg.ATP (Nadler et al. 2001). However, we have found that GTP-γ-S at 100 μm can inhibit MagNuM in a Mg²⁺- and ATP-independent manner, possibly through a G protein-mediated regulation, and this allowed us to study monovalent I_CRAC in isolation from MagNuM under [Mg²⁺]_i-free conditions. In this experiment, illustrated in the left panel of Fig. 4C, intracellular solutions contained 100 μm Na-GTP-γ-S in the absence of Mg²⁺ and I_CRAC was activated by 20 μm InsP₃. Under these conditions, we observed only a very small, transient increase in outward MagNuM current activation, which peaked at 0.6 pA pF⁻¹ around 150 s into the experiment and subsequently decayed to 0.1 pA pF⁻¹ due to GTP-γ-S block. Thus the inhibition of MagNuM by GTP-γ-S was essentially complete by the time DVF solution was applied, since the magnitude of the corresponding outward MagNuM current under Mg²⁺-free conditions is 3.2 pA pF⁻¹ (right panel of Fig. 4C).

Upon exposure to divalent-free extracellular solution, there was again a large increase in inward current, which also inactivated rather quickly. The degree of inactivation was significantly larger than the one observed with MagNuM, amounting to 73 % at the end of the DVF challenge, whereas the corresponding decrease in MagNuM was only 35 % (right panel of Fig. 4C). Monovalent I_CRAC retained considerable inward rectification, whereas the corresponding monovalent MagNuM currents were rather linear (Fig. 4E). Finally, we performed experiments in which both I_CRAC and MagNuM were allowed to activate under [Mg²⁺]_i-free conditions and perfused cells with internal solutions that included InsP₃ but were devoid of Mg²⁺ and ATP (Fig. 4D). Here, the activation of I_CRAC proceeded rapidly and when the extracellular solution was changed to divalent-free there was a sudden massive increase in both inward and outward monovalent current, which was probably due to the combined contributions of I_CRAC and MagNuM, followed by a slower, steady increase during the application period of divalent-free solution. We interpret this slow increase to be primarily due to MagNuM, since its activation proceeds during the divalent-free episode, as witnessed by the increased levels of outward current upon readmission of divalent-containing extracellular solution. This suggests that the contribution of MagNuM masks that of I_CRAC, which is likely to be similar to the response seen in the left panel of Fig. 4C and which is superimposed here for comparison as a dotted line. Since I_CRAC is subject to inactivation and MagNuM continues to increase, the latter is likely to carry the major component of monovalents under these experimental conditions.

Selectivity of I_CRAC and MagNuM

We next took advantage of differences in selectivity as a further means to distinguish I_CRAC and MagNuM. I_CRAC is well known for the high selectivity for Ca²⁺ over any other divalent ion, including Mg²⁺ (Hoth & Penner, 1993). MagNuM, on the other hand, is due to divalent-selective cation channel that conducts both Ca²⁺ and Mg²⁺ at membrane potentials below 0 mV (Nadler et al. 2001). The permeation of divalent ions such as Ca²⁺ and Mg²⁺ obstructs any significant monovalent conductance at negative membrane voltages. Only above 0 mV will MagNuM be carried by monovalent cations, since relief from this divalent permeation block gradually decreases as the reversal potential for Ca²⁺ is approached. Figure 5 illustrates that I_CRAC and MagNuM indeed behave differently when exposed to extracellular solutions that are Ca²⁺ free, but contain the standard Mg²⁺ levels. In the case of I_CRAC, illustrated in Fig. 5A, removal of Ca²⁺ completely blocks any inward current and there is no change in outward current. Conversely, when suppressing I_CRAC and activating MagNuM in isolation (Fig. 5B), the removal of extracellular Ca²⁺ suppresses the inward current only partially, since Mg²⁺ can sustain inward MagNuM currents. At the same time, there is a significant augmentation of monovalent outward current, which is consistent with a reduced permeation block as Ca²⁺ permeation is reduced.

Figure 5. Differential permeation of Ca²⁺ and Mg²⁺ through I_CRAC and MagNuM.

A, average inward (•) and outward (○) currents at −80 and +80 mV, respectively (n = 3), recorded under experimental conditions that suppress MagNuM and favour I_CRAC activation. I_CRAC was activated passively by inclusion of 10 mm BAPTA in the pipette solution and MagNuM was suppressed by 6 mm Mg.ATP ([MgCl₂]_i = 1 mm, free [Mg²⁺]_i = 1.1 mm). Cells were perfused with nominally Ca²⁺-free, but Mg²⁺-containing extracellular solution for the time indicated by the bars. Note the complete block of I_CRAC when Ca²⁺ was removed. B, average inward (•) and outward (○) currents at −80 and +80 mV, respectively (n = 3), recorded under experimental conditions that suppress I_CRAC and favour MagNuM activation. MagNuM was activated passively by omission of ATP and MgCl₂ from the pipette solution and I_CRAC was suppressed by buffering [Ca²⁺]_i to 90 nm using 10 mm EGTA and 3.6 mm CaCl₂ (free [Mg²⁺]_i = 0). Cells were perfused with nominally Ca²⁺-free, but Mg²⁺-containing extracellular solution for the time indicated by the bars. Note that MagNuM inward current is not abolished, but carried by Mg²⁺ and that outward currents are enhanced.

Pharmacology of I_CRAC and MagNuM

Finally, we sought to dissociate I_CRAC and MagNuM pharmacologically. 2-Aminoethoxydiphenyl borate (2-APB) has been reported to be a selective inhibitor of store-operated Ca²⁺ entry (Ma et al. 2000). Its inhibition of InsP₃ receptors has been interpreted in support of the conformational coupling model of I_CRAC activation, where InsP₃ receptors are hypothesized to directly interact with and gate CRAC channels (Irvine, 1990; Berridge, 1995). By the same token, the inhibitory effect of 2-APB on single channels of RBL, recorded under divalent-free conditions, has been invoked as major supporting evidence for identifying these channels as being CRAC channels (Braun et al. 2000). We have re-investigated the effects of 2-APB under conditions where I_CRAC and MagNuM are activated in the same cell. Figure 6A illustrates the InsP₃-mediated activation of I_CRAC carried by 10 mm Ca²⁺. As I_CRAC is almost fully activated, the delayed activation of MagNuM starts, as evidenced by the small increase in outward current. After the inward I_CRAC reached its maximum, 2-APB (100 μm) was applied extracellularly, and this resulted in a strong and irreversible inhibition of I_CRAC. At the same time, the outward MagNuM current, which had just started to activate, was also inhibited. This inhibition, however, was not irreversible, since outward MagNuM currents recovered from 2-APB block. After significant activation of MagNuM, we re-applied 2-APB and again observed a significant, albeit not complete, inhibition of outward currents.

Figure 6. Differential pharmacology of I_CRAC and MagNuM.

A, average inward (•) and outward (○) I_CRAC and MagNuM at −80 and +80 mV, respectively (n = 4). I_CRAC and MagNuM were activated passively by omission of ATP and inclusion of 20 μm InsP₃ and 10 mm BAPTA in the pipette solution ([MgCl₂]_i = 1 mm, free [Mg²⁺]_i = 780 μm). After full activation of I_CRAC, 100 μm 2-APB was applied extracellularly for the time indicated by the bar and a second application was performed after MagNuM had developed. Note the differential reversibility of 2-APB-mediated inhibition of outward and inward currents. B, average inward (•) and outward (○) currents at −80 and +80 mV, respectively (n = 3), recorded under experimental conditions that suppress MagNuM and favour I_CRAC activation. I_CRAC was activated passively by inclusion of 10 mm BAPTA in the pipette solution and MagNuM was suppressed by 6 mm Mg.ATP ([MgCl₂]_i = 1 mm, free [Mg²⁺]_i = 1.1 mm). After full activation of I_CRAC, 10 μm Gd³⁺ was applied extracellularly for the time indicated by the bar. C, average inward (•) and outward (○) I_CRAC and MagNuM at −80 and +80 mV, respectively (n = 3). I_CRAC and MagNuM were activated passively by omission of ATP and MgCl₂ and inclusion of 10 mm BAPTA in the pipette solution. After full activation of both I_CRAC and MagNuM, 10 μm Gd³⁺ was applied extracellularly for the time indicated by the bar. Note the complete inhibition of inward current and the lack of effect on outward current.

Thus, aside from a slightly lower efficacy in suppressing MagNuM, the major difference of 2-APB-mediated effects on I_CRAC and MagNuM appears to be the degree of reversibility. Furthermore, although we are not concerned with the activation mechanism of I_CRAC, this result demonstrates that 2-APB can inhibit I_CRAC after it has been fully activated and it does so with a rapidity that is consistent with a direct pharmacological inhibition (Braun et al. 2000). We observed the same type of inhibition in cells where I_CRAC was activated in an InsP₃-independent manner through passive depletion (data not shown). Furthermore, the 2-APB-induced block of I_CRAC only occurred when the compound was applied extracellularly and inclusion of the same concentration of 2-APB in the pipette solution was ineffective, consistent with an interpretation that 2-APB acts as an extracellular pore blocker of CRAC channels (Braun et al. 2000). In any event, the 2-APB-induced block of I_CRAC is not selective and can therefore not be considered an identifying signature for CRAC channels.

We sought to establish a more discriminating pharmacological assay to distinguish between I_CRAC and MagNuM and tested for selective Gd³⁺ sensitivity of the two conductances. As illustrated in Fig. 6B, I_CRAC is potently inhibited by low concentrations of Gd³⁺ (10 μm). However, Gd³⁺ at this concentration did not significantly affect outward MagNuM currents (Fig. 6C), while the inward current, in this case both I_CRAC and MagNuM, was completely suppressed. Additionally, we confirmed that 10 μm Gd³⁺ could partially block inward but not outward currents in HEK-293 cells over-expressing recombinant LTRPC7 (data not shown). Thus, although Gd³⁺ does not inhibit outward MagNuM currents, it does not seem to be an ideal tool to distinguish between I_CRAC and MagNuM, since inward MagNuM currents, as with I_CRAC, are suppressed by Gd³⁺.

DISCUSSION

The present study was designed to distinguish the Ca²⁺ release-activated Ca²⁺ current (I_CRAC) from the recently discovered magnesium-nucleotide-regulated metal ion current (MagNuM) (Nadler et al. 2001). The underlying ion channels are encoded by the LTRPC7 gene, a member of TRP family of ion channel genes (Harteneck et al. 2000). LTRPC7 is widely expressed in numerous tissues and cells (Nadler et al. 2001). MagNuM shares some features with I_CRAC, most notably a large monovalent conductance in the absence of divalent charge carriers such as Ca²⁺ and Mg²⁺. Furthermore, MagNuM is activated under experimental conditions that have traditionally been used to study I_CRAC. Finally, since MagNuM is also present in RBL cells and lymphocytes (Nadler et al. 2001), the model systems used to study I_CRAC, it is important to establish experimental conditions that rule out ambiguities arising from the parallel activation of I_CRAC and MagNuM in these cells. This is particularly important for recently developed protocols designed to measure single-channel currents through CRAC channels (Kerschbaum & Cahalan, 1999; Braun et al. 2000; Fomina et al. 2000), as they involve Mg.ATP and divalent-free solutions, which greatly favour the activation of MagNuM. In the following, we shall discuss the salient features of I_CRAC and MagNuM.

I_CRAC is a store-operated current that is activated by Ca²⁺ release from intracellular stores (Hoth & Penner, 1992; Parekh & Penner, 1997). Physiologically, this is achieved by receptor-mediated InsP₃ formation that causes store depletion. MagNuM, due to its relatively low sensitivity to Mg.ATP (millimolar), is constitutively active at low levels in resting cells, but is strongly activated by cytosolic ATP depletion below 1 mm (Nadler et al. 2001). Given the differential activation mechanisms, it is possible to isolate both conductances by appropriately composed intra-cellular solutions. Thus, I_CRAC can be effectively suppressed by buffering [Ca²⁺]_i to resting levels around 100 nm (Fasolato et al. 1993; Krause et al. 1999) and MagNuM can be inhibited by increasing Mg.ATP and/or [Mg²⁺]_i (Nadler et al. 2001). However, studies on I_CRAC originating from our own, as well as numerous other, laboratories have traditionally used low Mg.ATP concentrations or have completely omitted ATP from intracellular solutions. Therefore, I_CRAC recordings in most of these studies have probably been contaminated to various degrees with MagNuM currents.

Kinetics of activation of I_CRAC and MagNuM

Experimentally, one can trigger I_CRAC activation by a variety of conditions that lead to store depletion (Hoth & Penner, 1992; Parekh & Penner, 1996). MagNuM is activated by depletion of Mg.ATP and appears to be store independent, since active store depletion by InsP₃ does not significantly affect its activation kinetics (see Fig. 1). Several electro-physiological studies have used a passive store depletion protocol, where I_CRAC is activated by perfusing cells with pipette solutions that contain high concentrations of Ca²⁺ chelators such as EGTA and BAPTA. This results in the gradual depletion of intracellular stores due to basal leakage of Ca²⁺ out of the stores into the cytosol and the inability to refill these stores, since released Ca²⁺ is immediately captured by the buffers. As a result, store-operated I_CRAC is activated and reaches half-maximal amplitude at around 80 s in RBL cells (see Fig. 1). In these cells, ATP-free internal solutions cause activation of MagNuM with a considerably longer delay. The average half-maximal activation times of I_CRAC and MagNuM differ by about 140 s (see Fig. 1). There is certainly some degree of variability in the kinetics of activation of both conductances, as would be expected from their activation mechanisms (store depletion and ATP depletion, respectively). It is conceivable that both currents might activate in parallel, depending on how efficiently the two activation mechanisms are engaged under various experimental conditions. Factors that may influence this include (i) the metabolic state (e.g. cells with reduced Mg.ATP levels could activate MagNuM earlier), (ii) the cell type under investigation (e.g. lymphocytes have a significantly smaller cytosol volume compared to RBL cells and may equilibrate more rapidly with the pipette solution), (iii) the series resistance of a particular patch-clamped cell (e.g. higher resistances will slow diffusional exchange of cytosol and pipette solutions). One study has attempted to correlate inward and outward currents activated by store depletion and this study reported a parallel activation of inward and outward currents (Hoth, 1996). Our present data provide no evidence for parallel activation and although, on rare occasions, we have seen similar activation time courses, in no case was there an identical correspondence; MagNuM consistently lagged behind activation of I_CRAC. It is difficult to directly compare our data with those of the above investigation, since that study did not provide a quantitative assessment of average current kinetics and illustrated only anecdotal records from individual cells. Kerschbaum & Cahalan (1998) have also considered kinetic aspects of inward and outward currents through I_CRAC and observed parallel activation. However, those experiments were carried out under conditions that excluded divalent ions from either side of the membrane and, as will be discussed below, are not unambiguous as they do not directly address the issue of whether the monovalent currents were carried by I_CRAC and/or MagNuM.

It is important to note that previous studies have, in fact, observed kinetic differences in the activation of I_CRAC carried by Ca²⁺ ions and inward currents carried by monovalent ions in the absence of divalents (Kerschbaum & Cahalan, 1999; Braun et al. 2000). Thus, Cahalan and colleagues illustrate I_CRAC carried by Ca²⁺ that require about 80 s for half-maximal activation, whereas monovalent currents in the same study consistently activated half-maximally at around 200 s (Kerschbaum & Cahalan, 1999). Both of these values are in very close correspondence to the kinetics observed in our present study. Similarly, Putney and colleagues observed characteristic kinetic delays of about 1.5- to 2-fold for monovalent currents compared to Ca²⁺ currents (Braun et al. 2000). In the light of the newly discovered MagNuM conductance, which supports large monovalent currents and matches the delayed activation kinetics of monovalent currents in the above studies, it is difficult to avoid the conclusion that at least part of the monovalent currents in those studies, if not all, reflect MagNuM rather than I_CRAC.

Selectivity of I_CRAC and MagNuM

MagNuM comprises inward current in the voltage range where I_CRAC is carried by Ca²⁺, i.e. in the range of −100 mV to 0 mV. Thus, depending on the expression levels of a given cell type and the specific experimental conditions established by the composition of extra- and intracellular solutions, MagNuM can contribute significantly to inward currents and this contribution develops with a delay that coincides with the peak activation of I_CRAC. This makes it very difficult to assess accurately the contribution of each current. The typical size of total inward current carried by divalent ions at −80 mV in RBL cells under conditions that are normally used to study I_CRAC amounts to about 3 pA pF⁻¹. In the absence of Mg.ATP and under conditions that suppress I_CRAC, MagNuM typically generates around 0.5–1 pA pF⁻¹, and, although these conditions do not match exactly, we can estimate that MagNuM can contribute around 15–30 % of total inward current under standard conditions used to study I_CRAC.

MagNuM currents become relatively prominent above +50 mV, where they are carried by monovalent ions in the outward direction, but many investigations of I_CRAC cut off data presentation obtained at these positive voltages. We are aware of one publication that specifically addresses the outward currents of RBL cells in relation to I_CRAC (Hoth, 1996). This study suggested that outward currents recorded above +50 mV were carried by monovalent ions through CRAC channels. These outward currents were only observed when the intracellular solution contained K⁺, but not Cs⁺, as the primary monovalent cation and it was concluded that CRAC channels were permeable to K⁺ at positive voltages. This investigation was carried out with intracellular solutions containing sub-millimolar ATP levels (0.5 mm), which in our hands will lead to significant activation of MagNuM. Our present study was exclusively conducted with Cs⁺-based intracellular solutions and this cation is clearly permeant through LTRPC7 to the same extent as K⁺ (Nadler et al. 2001). At present, we cannot offer an explanation for the discrepancies between these findings. However, since we can suppress all outward currents by including Mg.ATP in our intracellular solutions without suppressing I_CRAC, we conclude that CRAC channels do not carry significant outward currents per se and that under our experimental conditions the predominant part of the outward currents is MagNuM.

I_CRAC is normally a highly selective Ca²⁺ current (Hoth & Penner, 1992, 1993). However, removal of divalent ions from the extracellular solution leads to a characteristic loss in selectivity, so that Na⁺ ions can permeate (Hoth & Penner, 1993; Kerschbaum & Cahalan, 1998). This typically results in a large increase in inward current. However, this increase is only transient and monovalent currents quickly inactivate by an unknown mechanism (Hoth & Penner, 1993). It has been proposed that this inactivation of I_CRAC is regulated by intracellular Mg²⁺, since removal of [Mg²⁺]_i reduces the amount of inactivation (Kerschbaum & Cahalan, 1998), but, as discussed below, this notion may need revision, as it is based on experiments in which MagNuM may have contributed to monovalent inward currents.

MagNuM is normally also a divalent-selective cation current, at least over the physiological voltage range of −100 to +50 mV, where it is carried by both Ca²⁺ and Mg²⁺. This characteristic difference between I_CRAC and MagNuM is highlighted by our experiments illustrated in Fig. 5, where the selective removal of extracellular Ca²⁺ completely suppresses I_CRAC, while the remaining extracellular Mg²⁺ partially supports inward MagNuM currents. Only at very positive potentials, where the permeation block of these divalents is reduced, MagNuM is carried by monovalent ions outwardly (physiologically this would be K⁺, but experimentally Cs⁺ permeates equally well). MagNuM, like I_CRAC, can lose its selectivity for divalents when both Ca²⁺ and Mg²⁺ are removed from the extracellular solution and this results in a significant augmentation of inward currents. Unlike I_CRAC, however, MagNuM supports large outward currents, and this occurs even when intracellular Mg²⁺ is present, whereas outward I_CRAC is virtually absent (see Fig. 4). Given that several investigations have studied monovalent currents in the absence of [Mg²⁺]_i, we need to consider how intracellular Mg²⁺ affects I_CRAC and MagNuM.

Effects of [Mg²⁺]_i on I_CRAC and MagNuM

Previous studies have suggested that I_CRAC is regulated by intracellular Mg²⁺, since monovalent inward currents are greatly enhanced and their inactivation is reduced when [Mg²⁺]_i is removed (Kerschbaum & Cahalan, 1998, 1999; Braun et al. 2000; Fomina et al. 2000). Data presented in Fig. 4 of the present investigation confirm that monovalent currents carried by I_CRAC are larger than those carried by Ca²⁺ when free [Mg²⁺]_i is lowered from 1.1 mm to 550 μm. However, inward currents carried by Ca²⁺ are not significantly inhibited by [Mg²⁺]_i in this concentration range. We even failed to observe a significant inhibition of I_CRAC carried by Ca²⁺ at concentrations up to 2.8 mm (data not shown). By contrast, free [Mg²⁺]_i levels above 1 mm essentially abolish both inward and outward MagNuM currents, irrespective of the ion species carried. We therefore conclude that MagNuM is significantly more sensitive to inhibition by [Mg²⁺]_i than I_CRAC, at least with respect to currents carried by divalent ions.

An important result of the present study is that GTP-γ-S can suppress MagNuM in a Mg²⁺- and ATP-independent fashion. Although GTP-γ-S has also been reported to inhibit the activation of I_CRAC (Fasolato et al. 1993), this effect requires prolonged pre-exposure to GTP-γ-S and delayed store depletion, whereas GTP-γ-S is completely ineffective in suppressing I_CRAC once it has been activated early in the experiment, e.g. when GTP-γ-S is co-perfused with InsP₃. Therefore, it is feasible to measure I_CRAC in isolation under Mg²⁺-free conditions. The results presented in Fig. 4C directly demonstrate that I_CRAC remains subject to inactivation even under [Mg²⁺]_i-free conditions. We can therefore assess the relative contributions of I_CRAC and MagNuM to monovalent membrane currents flowing upon removal of divalent ions when both conductances are activated. The peak monovalent inward I_CRAC and MagNuM currents are roughly equivalent under [Mg²⁺]_i-free conditions (Fig. 4C). Importantly, the rapid inactivation of I_CRAC is evident even under [Mg²⁺]_i-free conditions. It should be noted that the original observation of a reduction in the inactivation of monovalent currents through I_CRAC in [Mg²⁺]_i-free solutions was made under experimental conditions that would inevitably reveal MagNuM as a contaminating, if not dominant, conductance. The experiment depicted in Fig. 4D reproduces this scenario and shows that, indeed, inactivation of monovalent currents is abolished in [Mg²⁺]_i-free conditions. In fact, the monovalent currents continue to increase while exposing cells to divalent-free extracellular solutions. This increase may be attributed solely to MagNuM, since InsP₃-mediated activation of I_CRAC surely was maximal at this time. Even if we assumed that at the time of switching to divalent-free solutions, I_CRAC was the only pathway for monovalent ions, its contribution to inward current would have faded at the end of the divalent-free episode of the experiment due to inactivation. The dotted line in the left panel of Fig. 4C, representing the monovalent inward current through I_CRAC under [Mg²⁺]_i-free conditions, illustrates the rapidly decreasing contribution of I_CRAC to whole-cell monovalent currents, leaving MagNuM as the primary, if not sole, monovalent influx pathway.

Pharmacology of I_CRAC and MagNuM

Investigations on I_CRAC suffer from the lack of specific pharmacological tools. However, 2-APB has been shown to inhibit I_CRAC (Braun et al. 2000; Bakowski et al. 2001; Prakriya & Lewis, 2001). This compound was originally suggested to represent a specific InsP₃ receptor antagonist (Maruyama et al. 1997) and, based on its effects on capacitative Ca²⁺ entry, was subsequently invoked to support the conformational coupling model (Ma et al. 2000), in which InsP₃ receptors are hypothesized to activate and maintain the activity of CRAC channels (Irvine, 1990; Berridge, 1995). However, subsequent studies have recognized that 2-APB may, in fact, be a direct blocker of I_CRAC and its inhibitory effect may be unrelated to its action on InsP₃ receptors (Braun et al. 2000; Bakowski et al. 2001; Prakriya & Lewis, 2001). Nevertheless, since 2-APB is ineffective in suppressing other channels such as L-type Ca²⁺ channels and Trp-3 channels, the compound has been considered as relatively selective for I_CRAC. Since 2-APB was found to inhibit both the whole-cell monovalent currents and single channels, this has been interpreted as additional evidence that they are both carried by CRAC channels (Braun et al. 2000). In the light of the present study, this assessment may need to be revised, since 2-APB was found to also inhibit MagNuM (see Fig. 6) and therefore cannot selectively identify CRAC channels.

Conclusions

In the light of the properties of the newly discovered ion channel LTRPC7, which underlies the Mg-nucleotide-regulated metal ion (MagNuM) current, we may need to re-evaluate some of the findings that were originally attributed to I_CRAC. In particular, this applies to those studies that obtained data under experimental conditions that favour MagNuM activation, notably ATP- and Mg²⁺-free intracellular solutions (Kerschbaum & Cahalan, 1998, 1999; Braun et al. 2000; Fomina et al. 2000). The monovalent whole-cell and single-channel currents presented in these studies may have been due to (i) I_CRAC alone, (ii) MagNuM alone, or (iii) a mixture of I_CRAC and MagNuM, and we will consider each of these alternatives in turn.

The above studies all assumed that monovalent currents were exclusively carried by CRAC channels. Since the intracellular solutions were ATP and [Mg²⁺]_i free, there is little doubt that MagNuM must have been contributing to monovalent currents, at least with respect to whole-cell currents, and therefore these are unlikely to represent I_CRAC alone. Let us next consider that monovalent currents in the absence of any divalents on either side of the membrane are exclusively MagNuM. In this context, it is important to point out that monovalent I_CRAC inactivates rapidly (Hoth & Penner, 1993; Kerschbaum & Cahalan, 1998) and the present study demonstrates that it does so in the absence of intracellular Mg²⁺. Since the inactivation of monovalent I_CRAC proceeds faster than its activation, it seems likely that I_CRAC cannot support significant monovalent current and, since MagNuM is not subject to such strong inactivation, it could provide the predominant, if not exclusive, monovalent influx pathway. In fact, most features of monovalent currents recorded in the investigations above are entirely consistent with MagNuM as the sole pathway of monovalent current under ATP- and [Mg²⁺]_i-free conditions, including kinetics, selectivity, current-voltage relationship and pharmacology.

Finally, we should consider a scenario in which both I_CRAC and MagNuM contribute to monovalent currents. In order for I_CRAC to do so, we would need to postulate that CRAC channels do not inactivate completely in [Mg²⁺]_i-free conditions. If this were the case, then one would expect a sequential activation of monovalent currents, first I_CRAC and then followed by MagNuM. However, in previous studies, the activation kinetics of monovalent currents is consistent with MagNuM and not I_CRAC. In some circumstances, Putney and colleagues have observed an ‘intermediate-sized’ current that was present when cells were pre-activated by thapsigargin (Braun et al. 2000), which may be related to a similar intermediate-sized current Cahalan and colleagues have observed when recording monovalent currents in the presence of intracellular Mg²⁺ (Kerschbaum & Cahalan, 1999) and these currents may have been due to I_CRAC.

The possible contamination of I_CRAC with MagNuM may also have affected the study by Parekh & Penner (1995), where I_CRAC showed little signs of inactivation in ATP-free conditions, whereas moderate to strong inactivation was observed when perfusing cells with 2 mm ATP and ATP-γ-S, respectively. Although the general conclusions of that investigation remain valid, the overall magnitudes of the inactivation observed with these nucleotides may have been slightly overestimated, since both Mg.ATP and Mg.ATP-γ-S would have also suppressed any contaminating component of MagNuM.

In summary, the present study has demonstrated the coexistence of two potential ion channel mechanisms, I_CRAC and MagNuM, which both support large monovalent currents when divalent-free extracellular solutions are used. A reduction or omission of intracellular Mg²⁺ significantly augments the monovalent MagNuM. Since these latter conditions have been used to study selectivity and single-channel properties of I_CRAC, the results of these investigations may have to be reassessed in the light of a possible contamination with MagNuM.