Magnesium levels in cardiac myocytes change in cardiovascular diseases. Intracellular free magnesium (Mgi) inhibits L-type Ca2+ currents through CaV1.2 channels in cardiac myocytes, but the mechanism of this effect is unknown. We hypothesized that Mgi acts through the COOH-terminal EF-hand of CaV1.2. EF-hand mutants were engineered to have either decreased (D1546A/N/S/K) or increased (K1543D and K1539D) Mg2+ affinity. In whole-cell patch clamp experiments, increased Mgi reduced both Ba2+ and Ca2+ currents conducted by wild type (WT) CaV1.2 channels expressed in tsA-201 cells with similar affinity. Exposure of WT CaV1.2 to lower Mgi (0.26 mM) increased the amplitudes of Ba2+ currents 2.6 ± 0.4–fold without effects on the voltage dependence of activation and inactivation. In contrast, increasing Mgi to 2.4 or 7.2 mM reduced current amplitude to 0.5 ± 0.1 and 0.26 ± 0.05 of the control level at 0.8 mM Mgi. The effects of Mgi on peak Ba2+ currents were approximately fit by a single binding site model with an apparent Kd of 0.65 mM. The apparent Kd for this effect of Mgi was shifted ∼3.3- to 16.5-fold to higher concentration in D1546A/N/S mutants, with only small effects on the voltage dependence of activation and inactivation. Moreover, mutant D1546K was insensitive to Mgi up to 7.2 mM. In contrast to these results, peak Ba2+ currents through the K1543D mutant were inhibited by lower concentrations of Mgi compared with WT, consistent with approximately fourfold reduction in apparent Kd for Mgi, and inhibition of mutant K1539D by Mgi was also increased comparably. In addition to these effects, voltage-dependent inactivation of K1543D and K1539D was incomplete at positive membrane potentials when Mgi was reduced to 0.26 or 0.1 mM, respectively. These results support a novel mechanism linking the COOH-terminal EF-hand with modulation of CaV1.2 channels by Mgi. Our findings expand the repertoire of modulatory interactions taking place at the COOH terminus of CaV1.2 channels, and reveal a potentially important role of Mgi binding to the COOH-terminal EF-hand in regulating Ca2+ influx in physiological and pathophysiological states.

Magnesium is the second most abundant intracellular cation (Elin, 1994). In the heart, the level of free intracellular magnesium (Mgi) is well controlled but alterations are observed in a variety of cardiovascular diseases (Murphy, 2000). For example, during myocardial ischemia, there is a reciprocal relationship between decreased ATP and increased Mgi levels, and Mgi rises 3.5-fold as ATP levels fall (Murphy et al., 1989). Cardiovascular diseases are associated with alterations in Ca2+ homeostasis (Tomaselli and Marban, 1999), and Mgi can modulate multiple proteins involved in Ca2+ transport (White and Hartzell, 1988, 1989; Hartzell and White, 1989; Xu et al., 1996; Wei et al., 2002), including substantial effects on the L-type Ca2+ current density, inactivation, and voltage dependence (White and Hartzell, 1988; Agus et al., 1989; Kuo and Hess, 1993; Yamaoka and Seyama, 1996; Pelzer et al., 2001; Wang et al., 2004). The molecular mechanisms responsible for these modulatory effects of Mgi on L-type Ca2+ currents are unknown. It has been proposed that Mgi could produce its effects directly by interacting with the Ca2+ channel protein (Kuo and Hess, 1993; Yamaoka and Seyama, 1996) or indirectly by altering enzyme activities that require Mgi as a cofactor or regulator, such as protein kinases or phosphoprotein phosphatases (Pelzer et al., 2001).

In ventricular myocytes, L-type Ca2+ currents are conducted by CaV1.2 channels consisting of a pore-forming α11.2 subunit in association with β and α2δ subunits (Catterall, 2000). The α1 subunits are composed of four homologous domains (I–IV) with six transmembrane segments (S1–S6) and a reentrant pore loop in each. Regulatory sites for Ca2+/calmodulin and cAMP-dependent protein kinase are located in the COOH-terminal domain (De Jongh et al., 1996; Peterson et al., 1999; Zuhlke et al., 1999; Hulme et al., 2003), which is also subject to in vivo proteolytic processing (De Jongh et al., 1991, 1996). The COOH-terminal domain contains an EF-hand motif that is expected to bind divalent cations (de Leon et al., 1995). Studies of Ca2+ channel chimeras suggested a role for the EF-hand in Ca2+-dependent inactivation (de Leon et al., 1995), but mutations in the amino acid residues that are required for divalent cation binding have no effect on inactivation (Zhou et al., 1997; Peterson et al., 2000), indicating that the role of the EF-hand in Ca2+-dependent inactivation is structural. A nearby IQ domain is directly implicated in Ca2+-dependent inactivation mediated by Ca2+/calmodulin (Peterson et al., 1999; Zuhlke et al., 1999), and structural changes in the EF-hand may impair the inactivation process indirectly.

Could the COOH-terminal EF-hand motif of CaV1.2 channels be involved in modulation by Mgi? EF-hands are composed of a helix-loop-helix motif, where the loop of 12 amino acid residues forms the cation-binding site (Kawasaki and Kretsinger, 1994). In addition to Ca2+, Mg2+ at physiological concentrations can bind to some EF-hand motifs with appropriate amino acid sequences (da Silva et al., 1995; Houdusse and Cohen, 1996; Lewit-Bentley and Rety, 2000; Yang et al., 2002). Inspection of the amino acid sequence of the CaV1.2 COOH-terminal EF-hand suggested that it could bind Mgi. Therefore, we constructed EF-hand ion-coordination site mutants and compared their electrophysiological properties with wild-type (WT) CaV1.2 channels. Our results support a novel mechanism linking the COOH-terminal EF-hand motif of CaV1.2 with the modulation of CaV1.2 channels by intracellular Mg2+ (Brunet et al., 2004). These findings expand the repertoire of modulatory interactions that take place at the COOH terminus of CaV1.2 channels and suggest that Mgi bound to the COOH-terminal EF-hand may have an important role in regulating Ca2+ influx in physiological and pathophysiological states.

Cell Culture and Expression

TsA-201 cells, a subclone of the human embryonic kidney cell line HEK-293 that expresses the simian virus 40 T-antigen (a gift of Robert Dubridge, Cell Genesis, Foster City, CA), were grown in DMEM/Ham's F-12 medium (Life Technologies), supplemented with 10% (vol/vol) FBS (Hyclone), and incubated at 37°C in 10% CO2. TsA-201 cells were grown to 80% confluence, suspended with trypsin/EDTA, and plated onto 35-mm culture dishes (Corning) at 40% confluence 24 h before transfection. Immediately before transfection, the medium was replaced with fresh DMEM/F-12 supplemented with serum and antibiotics, and the cells were transiently transfected with cDNA encoding rabbit cardiac CaV1.2, β1b, and α2δ subunits at 1:1:1 molar ratio by using the Ca2+ phosphate method or the Fugene method (Invitrogen) in accordance with the manufacturer's protocol. In addition, a 10-fold lower molar concentration of cDNA encoding CD-8 antigen (EBO-pCD-Leu2; American Type Culture Collection) was added to each transfection. The cells were incubated overnight at 37°C in 3% CO2. For Ca2+ phosphate transfection, the medium was replaced with fresh DMEM/F12 after 6–7 h, and the cells were allowed to recover for 15 h. For Fugene transfection, the cells were incubated with Fugene for 24 h at 3% CO2. After recovering from transfection for 15 h or 24 h, the cells were suspended using EDTA, plated in 35-mm dishes, and incubated at 37°C in 10% CO2 for at least 3–4 h before recording.

Electrophysiological Recordings

Immediately before recording, a 35-mm culture dish with transfected cells was stirred for 1 min with latex beads conjugated to an anti-CD8 antibody (Dynal), which bound those cells that had been successfully transfected with the Ca2+ channels plus CD-8 receptor (Margolskee et al., 1993). The extracellular recording solution contained (in mM) 10 BaCl2, 140 Tris, 2 MgCl2, and 10 d-glucose titrated to pH 7.3 with MeSO4H. For some recordings, 1.8 mM CaCl2 was substituted for 10 mM BaCl2 with appropriate adjustment of osmolarity. The control intracellular Mg2+ (0.8 mM free Mg2+) solution contained (in mM) 130 N-methyl-d-glucamine, 60 Hepes, 5 MgATP, 1 MgCl2, and 10 EGTA titrated to pH 7.3 with MeSO4H. The osmolarity of all solutions was adjusted to 295 mOsM with sucrose. The free intracellular Mg2+ concentration was altered by changing the MgCl2 in the intracellular solution. Free [Mg2+]i was calculated by Maxchelator program (Bers et al., 1994).

Voltage clamp recordings were made in the whole-cell configuration using an Axopatch 200A amplifier (Axon Instruments). Linear leak and capacitance were cancelled using the internal amplifier circuitry and >80% of the series resistance was compensated. Pipettes were pulled from VWR micropipettes. After achieving the whole cell configuration, pipette resistances were from 2 to 5 MΩ. The design of this study required recording currents with a broad range of magnitudes under identical conditions. To avoid poor voltage control for cells with large currents, larger pipettes were used in combination with maximal series resistance compensation. Voltage clamp fidelity was assessed for each cell and cells with >5 mV uncompensated voltage error, steep current–voltage relationships, and/or slowly decaying tail currents were omitted from analysis.

Construction of EF-hand Mutants

To construct mutants D1546A/N/S/K, K1539D, and K1543D, mutagenic primers were designed that contained an internal HindIII restriction site. K1539D, K1543D, and D1546A/N/S/K were all amplified in a two-phase manner, using a SacII-XhoI cDNA fragment of CaV1.2 in pBluescript SK+. Phase I generated the 5′ and the 3′ arms using the WT DNA template. PCR products were gel purified. Phase II combined the 5′ and 3′ arms to serve as both initial primers and template. PCR products were precipitated, washed, dried, and resuspended. PCR products and vector were cut with Sac II and Xho I and isolated by ethanol precipitation. The digested vector was treated with calf alkaline phosphatase (CIP, NEB). Digested DNA was run out on a 1.25% agarose gel. Fragment bands were excised and purified either by Spin-X column (Fisher Scientific) or QIAQuick (QIAGEN). Fragments were ligated (Fast-Link, Epicentre). A negative control of vector and no insert was run at the same time. Ligation mixtures were then transformed into competent DH5 cells and samples were plated onto LB plates overnight. DNA was extracted from mutant colonies by basic SDS/sodium acetate method. Samples were screened by cutting with restriction enzymes whose recognition sites were silently built into the mutation primer. Positive samples were further purified (Quantum miniprep kit, Bio-Rad Laboratories) and sequenced (BigDye, ABI). When the sequences were confirmed to be correct, DNA from the original mini-prep was retransformed, amplified, and extracted (QUANTUM maxiprep kit, Bio-Rad Laboratories). Each mutant DNA in pBluescript SK+ (Stratagene) and the full-length WT CaV1.2 α in pCDNA3 (Stratagene) were digested with Sac II and Xho I, ligated, and subcloned. Samples were sequenced again to confirm the sequence of the final construct.

Data Analysis

Voltage-clamp data were compiled and analyzed using Igor (IGOR Pro version 5.0, Wavemetrics Inc.) and Excel (Excel 97, Microsoft). Peak tail currents were measured during repolarization to −40 mV after a 20-ms depolarization to potentials between −40 and 80 mV. For the measurement of the voltage dependence of inactivation, 4-s depolarizations to potentials from −80 to 20 mV were applied to inactivate a fraction of Cav1.2 channels. A standard test pulse of 30 ms to 30 mV was applied, and peak tail currents were measured during repolarization to −40 mV immediately following the test pulse. To control for differences in expression levels between transfections, individual cells recorded with control Mgi (0.8 mM Mgi) were alternated with cells recorded at test Mgi concentrations during each recording period. Activation and inactivation data were fit to a Boltzmann function (Sigmaplot version 7.0, Systat Software). Differences between control and experimental values were assessed using ANOVA and the Student's t test; P values <0.05 were considered statistically significant.

Online Supplemental Material

The mean values for all of the parameters describing the voltage dependence of activation and inactivation for WT CaV1.2 and EF-hand mutants are presented in Table S1 (available at http://www.jgp.org/cgi/content/full/jgp.200509333/DC1).

Modulation of CaV1.2 Channels by Intracellular Mg2+

To determine whether Mgi modulates cloned CaV1.2 channels expressed in tsA-201 cells, we examined the electrophysiological properties of WT CaV1.2 channels over a range of intracellular Mgi concentrations from 0.1 to 7.2 mM. Changes in Mgi in this range modulated the peak current amplitude in all cases and altered the voltage dependence of activation and inactivation at the extreme concentrations. We were unable to study lower intracellular Mg2+ concentrations because the recordings became unstable. Higher intracellular Mg2+ concentrations depressed the Ba2+ current too much for accurate measurements. As each measurement was necessarily made on a separate cell, we interleaved cells at 0.8 mM Mgi and other test concentrations to be certain that no artifactual drift in electrophysiological parameters occurred during a recording period. The principal experimental results are presented in the figures and text below, whereas mean values for all of the parameters describing the voltage dependence of activation and inactivation are presented in Table S1 (available at http://www.jgp.org/cgi/content/full/jgp.200509333/DC1).

Changes in Mgi modulated the peak Ba2+ current conducted by CaV1.2 channels (Fig. 1 A). Reduction in Mgi from the control level of 0.8 mM increased Ba2+ current, and increases in Mgi decreased Ba2+ current. Parallel changes were observed in the amplitudes of tail currents recorded after repolarization to −40 mV (Fig. 1 A). We used the tail currents as a measure of Ba2+ conductance to analyze inhibition by Mgi at different test pulse potentials as well as to determine the voltage dependence of activation and inactivation. In the presence of the 0.8 mM Mgi, the tail current first was observed at −25 mV, increased progressively with more depolarized test potentials, and reached a plateau at potentials more positive than +30 mV (Fig. 1, A and B). Higher concentrations of Mgi significantly reduced tail current amplitude across the entire voltage range, while lower Mgi significantly increased it (Fig. 1, A and B). Thus, increases in Mgi cause primarily voltage-independent inhibition of peak Ba2+ currents through CaV1.2 channels.

Effects of Mgi on the voltage dependence of activation and inactivation were observed at the highest and lowest Mgi concentrations tested. The highest Mgi tested (7.2 mM) led to a shift of −7.5 mV in the activation curve (Fig. 1 C), while reducing Mgi to 0.1 mM shifted the activation curve by +10.1 mV (Fig. 1 C). In contrast, changes in Mgi between 0.26 and 2.4 mM did not cause changes in the voltage dependence of activation (Fig. 1 C and Table S1). High Mgi concentrations (2.4 and 7.2 mM) also caused negative shifts (−5 and −10 mV, respectively) of the voltage dependence of inactivation (Fig. 1 D), whereas lower Mgi did not alter the voltage dependence of inactivation but made inactivation incomplete (Fig. 1 D).

Concentration-dependent Inhibition of Ba2+ and Ca2+ Currents by Mgi

Although changes in Mgi caused shifts in the voltage dependence of both activation and inactivation, peak Ba2+ currents reached a plateau at voltages more positive than 30 mV for all Mgi concentrations (Fig. 1 B). Therefore, the reductions of peak Ba2+ current with increasing Mgi at test potentials more positive than 30 mV were readily separated from effects on voltage dependence. The requirement to measure Ba2+ currents at different Mgi concentrations in separate cells introduced variability due to the different expression levels of Ca2+ channels among cells. Despite this variability, we found that the mean values for peak Ba2+ currents were significantly different as a function of Mgi, and could be approximated by a single binding isotherm with an apparent Kd value of 0.65 mM (Fig. 2 A). Although the fit to a single binding site model is only approximate, a one-site model was significantly better than a fit to a two binding site model in a global fit of all of our results with WT and mutant channels (see below).

To determine whether the effect of Mgi would be observed under physiological conditions, we performed similar experiments with Ca2+ (1.8 mM) as charge carrier (Fig. 2 A). Similar to the experiments with Ba2+, we found that peak Ca2+ currents were significantly decreased as a function of Mgi, and the inhibition could be approximated by a single-site binding isotherm with an apparent Kd value of 0.41 mM (Fig. 2 A). As for Ba2+ currents, we also observed shifts in the voltage dependence of activation and inactivation of Ca2+ currents. Lower Mgi (0.26 and 0.1 mM) positively shifted the voltage dependence of activation by ∼10 mV (6.6 ± 0.7 mV, n = 10, at 0.26 mM and 7.1 ± 0.8 mV, n = 11, at 0.1 mM vs. −3.1 ± 0.4 mV, n = 35, at 0.8 mM, P < 0.001), while higher Mgi (7.2 mM) negatively shifted the voltage dependence of activation by ∼8 mV (−11.3 ± 0.3 mV, n = 9, vs. −3.1 ± 0.4 mV, n = 35, P < 0.001). As activation was complete at +30 mV, the measurements of peak Ca2+ currents (Fig. 2 A) were unaffected by the changes in voltage dependence. Because Ca2+ binds with far higher affinity to the pore of CaV1.2 channels than Ba2+, the similarity of apparent Kd values for inhibition of Ca2+ and Ba2+ currents suggests that the effects of Mgi are not mediated by competitive binding to the Ca2+ coordination sites in the ion selectivity filter of the pore.

Effects of Neutralization of a Negative Charge at Position 1546 in the COOH-terminal EF-hand

To determine whether interaction of Mgi with the COOH-terminal EF-hand (Fig. 2, B and C) was responsible for mediating the electrophysiological effects on CaV1.2 channels, we tested the mutations of D1546 (−z position) to A, N, or S, which are predicted to reduce affinity for Mgi (da Silva and Reinach, 1991; Kawasaki and Kretsinger, 1994; da Silva et al., 1995). Compared with WT CaV1.2 channels, increasing Mgi had a much smaller effect on the peak tail current amplitude of D1546A/N/S mutant channels, as illustrated in Fig. 3 A for D1546N. In the presence of 0.8 mM Mgi, the tail current activated at −25 mV, progressively increased with more depolarized test potential up to +20 mV and reached a plateau at more positive test potentials as for WT (Fig. 3 B). However, in contrast to WT CaV1.2 channels (Fig. 3 B, WT range), changes of Mgi from 7.2 to 0.26 mM had much smaller effects on peak tail currents of the D1546A/N/S mutants, as illustrated in Fig. 3 B for D1546N. These results are consistent with reduced affinity for binding of Mgi to the altered EF-hands of these mutants.

In this group of mutants, the voltage dependence of activation was similar to WT at 0.8 mM Mgi. However, reduction to 0.26 mM Mgi caused a positive shift in the conductance–voltage relationship, while increase to 2.4 mM Mgi caused a negative shift for D1546A/N/S (Fig. 3 C, D1546N). The voltage dependence of inactivation of D1546A/N/S was indistinguishable from WT, and included the shift of approximately −5 mV in the voltage dependence of inactivation at 2.4 mM Mgi (Fig. 3 D, D1546N).

As for WT Cav1.2 channels, the changes in voltage dependence of activation and inactivation caused by the mutations were small, and peak Ba2+ currents reached a plateau at the test pulses more positive than 30 mV. We plotted the mean tail currents for several cells expressing each mutant versus the concentration of Mgi in order to estimate the apparent Kd for reduction of peak Ba2+ currents by Mgi (Fig. 4). These results show that the Kd increased to ∼1.8 mM for D1546N, 2.6 mM for D1546A, and 9.9 mM for D1546S, ∼3- to 16.5-fold higher than WT. However, the limited range of Mgi concentration accessible for investigation prevented us from defining more precise Kd values for the mutants.

Effects of Substitution of a Positive Charge for D1546 in the COOH-terminal EF-hand

To further disrupt the electrostatic interaction between Mgi and the EF-hand, we introduced a positive charge at position 1546 (D1546K, −z position), and the electrophysiological properties of this mutant and effects of Mgi on its function were tested. Substitution of the positively charged K for a negatively charged D that directly interacts with the bound cation in the EF-hand would be expected to completely prevent binding of Mg2+ and other divalent cations because of charge–charge repulsion. As for WT channels, in the presence of 0.8 mM Mgi, the tail current conducted by mutant D1546K was detectable at −25 mV, progressively increased with more depolarized test potential up to +20 mV, and reached a plateau at more positive test potentials (Fig. 5, A and B). However, in contrast to WT CaV1.2 and D1546A/N/S channels, changes of Mgi from 7.2 to 0.26 mM had much smaller effects on peak tail currents of the D1546K mutants (WT range; Fig. 5 B). These results are consistent with reduced affinity for binding of Mgi to the altered EF-hand of this mutant.

The voltage dependence of activation of this mutant was negatively shifted compared with WT at 0.8 mM Mgi. Reduced Mgi (0.1 mM) caused a positive shift in the conductance–voltage relationship similar to WT (Fig. 5 C); however, in contrast to WT CaV1.2 channels, higher Mgi (7.2 mM) did not negatively shift the voltage dependence of activation (Fig. 5 C). The voltage dependence of inactivation was modified by Mgi similarly to WT (Fig. 5 D).

Compared with the WT CaV1.2 channels, Mgi had a much smaller effect on the peak tail current amplitude of D1546K mutant channel across the full range that we were able to study (Fig. 6). The peak tail currents recorded from 0.26 mM Mgi to 7.2 mM Mgi were fit by a straight line with zero slope (Fig. 6), and a single binding site isotherm did not give an improved fit (not presented). These results indicate that the effect of Mgi is completely lost in this charge-reversal EF-hand mutant over the physiologically relevant concentration range.

Effect of Addition of a Negative Charge at Position 1543 in the COOH-terminal EF-hand

To determine whether the affinity of Mgi for the COOH-terminal EF-hand of CaV1.2 could be increased by the addition of a negative charge, D was introduced at two ligand coordination positions, K1543 (−x position, Fig. 2, B and C) and K1539 (z position, see below), and the electrophysiological properties and effects of Mgi were tested. Mutation K1543D increased the effects of Mgi. Increasing Mgi from 0.8 to 2.4 mM led to a substantial reduction of normalized peak tail current amplitude (Fig. 7, A and B). Moreover, reducing Mgi to 0.26 mM caused a much smaller increase in tail currents than for WT, while further reduction to 0.1 mM Mgi gave a more substantial increase in peak tail currents. These results are consistent with increased affinity for inhibition by Mgi (see below).

In contrast to the D1546A/N/S mutants, the voltage dependence of activation of K1543D was shifted −6 mV compared with WT (Fig. 7 C). Increasing Mgi concentration to 2.4 mM did not cause a further shift in the activation relationship, while decreasing Mgi positively shifted activation (Fig. 7 C). The voltage dependence of inactivation was unchanged from WT for K1543D at 0.8 mM Mgi, but 2.4 mM Mgi negatively shifted the voltage dependence of inactivation as for WT CaV1.2 channels (Fig. 7 D). Lower Mgi (≤0.26 mM) caused voltage-dependent inactivation to be strikingly incomplete at positive test pulse potentials (Fig. 7 D).

The smaller inhibition by increasing Mgi compared with WT coupled with the larger increase in IBa with decreasing Mgi indicate increased apparent affinity for inhibition of peak currents by Mg2+. Because WT and all four mutants substituting A, N, S, and K at position 1546 give a similar extrapolated peak IBa when fit to a single binding site isotherm (Figs. 4 and 6), we have plotted the results for K1543D in a similar manner (Fig. 6). The experimental results are well fit by a single binding site model with an apparent Kd of 0.16 mM for Mgi and a similar maximum IBa. Thus, these results are consistent with ∼4.1-fold higher affinity for inhibition of peak CaV1.2 currents by Mgi in the K1543D mutant. Alternative interpretations of these results are considered in DISCUSSION.

A Single Site Model for Mg2+ Inhibition

To test the significance of the single binding site model we have used to fit our results (Figs. 4 and 6), we used a global fit protocol to analyze simultaneously the fit of all of our results for inhibition of WT and mutant CaV1.2 channels by Mgi to a binding isotherm with a variable Hill slope, constraining that value to be the same for all of the experimental results. The resulting fit yielded an estimate of Hill slope of 0.77 with 95% confidence limits of ±0.16. These results are consistent with a single-binding-site model for Mgi inhibition and are significantly different from a two-binding-site model. This analysis validates use of a single binding site model for inhibition of CaV1.2 channels by Mgi and provides support for our estimates of apparent Kd values based on a single binding site model.

Effects of Addition of a Negative Charge at Position 1539 in the COOH-terminal EF-hand

In contrast to K1543D, changes in Mgi from 0.1 to 2.4 mM did not have a concentration-dependent effect on peak tail currents of K1539D, as if the effect of Mgi was saturated at these concentrations (Fig. 8 A). Further reduction in Mgi to 0.05 mM significantly increased peak tail currents, indicating that Mgi inhibition of this mutant was complete at 0.1 mM and above (Fig. 8 A).

The voltage dependence of activation of K1539D was shifted +10 mV from WT at 0.8 mM Mgi. Increased Mgi caused a negative shift (−6 mV), whereas decreased Mgi caused a positive shift (+11 mV) of the voltage dependence of activation (Fig. 8 B). Changes of Mgi did not alter the voltage dependence of inactivation substantially, but reduced Mgi (0.1 mM) caused strikingly incomplete inactivation at positive membrane potentials as for K1543D (Fig. 8 C).

Although K1539D has altered voltage dependence of activation and inactivation, we were able to accurately measure the plateau of peak Ba2+ tail currents following test pulses to potentials more positive than 30 mV. However, unlike the results with other mutants, these results did not fit a single binding site isotherm well. Comparison of these values to the binding isotherm for mutant K1543D (Fig. 8 D) suggests the possibility that K1539D does indeed have increased apparent affinity for inhibition of peak current by Mgi, comparable to K1543D, but K1539D also has a reduced maximum extent of inhibition leading to a plateau at ∼19% of maximum current at Mgi > 0.1 mM.

Mg2+ Concentrations Change in Cardiovascular Diseases

Intracellular Mg2+ is primarily bound, and free Mg2+ levels are altered in response to changes in the levels of macromolecular Mg2+ binding sites and soluble Mg2+ chelators such as nucleotides. The transport mechanisms that allow Mg2+ to cross the plasma membrane and control the levels of intracellular Mg2+ in order to maintain homeostasis are poorly understood. However, given the high concentration of Mg-ATP in cells, it may be expected that conditions that substantially alter ATP levels would also alter free Mgi. Consistent with this expectation, free Mgi in cardiac myocytes increases from 0.6–0.7 mM to 2.1–2.3 mM in parallel with the decrease in ATP levels during myocardial ischemia in rat heart (Murphy et al., 1989; Headrick and Willis, 1991). Free Mgi also increases during transient decreases in pH (Freudenrich et al., 1992), which may contribute to the increase observed in ischemia as intracellular H+ concentration rises due to lactic acid production. Increased extracellular Mg2+ is cardioprotective for patients during episodes of ischemia (McCully and Levitsky, 1997). Many mechanisms may contribute to this cardioprotective effect, but inhibition of Ca2+ entry via CaV1.2 channels would be a likely candidate for a major role in cardioprotection because of the large L-type Ca2+ currents in myocytes and the high sensitivity of myocytes to the cytotoxic effects of excess cytosolic Ca2+.

In contrast to the increase in free Mgi during acute episodes of myocardial ischemia, chronic experimental heart failure in dogs is accompanied by a decrease in intracellular free Mg2+ from 1.06 to 0.49 mM (Haigney et al., 1998). The mechanism of this prolonged effect on free Mgi is unknown, but it is associated with altered ventricular repolarization, which may predispose to dangerous arrhythmias. Together with the results on myocardial ischemia, these results on heart failure support the conclusion that there are substantial variations in myocardial free Mgi during pathological conditions.

Intracellular Mg2+ Inhibits Transfected Cav1.2 Channels

White and Hartzell (1988) showed that exposure of cardiac myocytes to increasing concentrations of intracellular Mg2+ (0.3–3 mM) reduces the L-type Ca2+ current density. The reduction of L-type Ca2+ current in isolated cardiac myocyte preparation has been reproduced by many other laboratories (Agus et al., 1989; Yamaoka and Seyama, 1996; Pelzer et al., 2001; Wang et al., 2004). Here we show that this effect is also observed for cloned Cav1.2 channels expressed in human embryonic kidney tsA-201 cells. These results indicate that the inhibitory effect of Mgi on Cav1.2 channels does not require proteins that are specific to cardiac cells and provide a model system for analysis of the molecular mechanism of the effect of Mgi.

Mgi Acts through the COOH-terminal EF-hand to Inhibit Cav1.2 Channels

The mechanism responsible for the effects of Mgi on Ca2+ currents in cardiac cells has been controversial. Some reports support direct interaction of Mgi with the cardiac Ca2+ channel (Kuo and Hess, 1993; Yamaoka and Seyama, 1996) while others support an indirect mode of interaction for Mgi via modification of the activities of protein kinases or phosphoprotein phosphatases (McGowan and Cohen, 1988; Pelzer et al., 2001; Wang et al., 2004). By analysis of the effects of mutations in a potential Mg2+-binding EF-hand motif, our experiments support direct action of Mg2+ on the Ca2+ channel protein that reduces peak L-type Ba2+ currents through the Ca2+ channel by binding to the COOH-terminal EF-hand of CaV1.2 channels.

The ligand-binding amino acid residues in EF-hands are well known (Kawasaki and Kretsinger, 1994), allowing design of mutations that are known to affect divalent cation binding directly and specifically in other structurally defined, EF-hand–containing proteins. We found that mutations of ion-coordinating amino acids of the COOH-terminal EF-hand lead to alterations in Mgi sensitivity. Mutations at position 1546 (D1546A/N/S) that are predicted to reduce Mg2+ affinity caused a substantial reduction in the sensitivity of Mgi inhibition, consistent with 4.1- to 16.5-fold increase in the apparent Kd for Mg2+ binding. In addition, introducing a positive charge at that position in mutant D1546K completely removed Mgi sensitivity over the concentration range that we tested. Our results are in agreement with previous experimental observations showing that D at the −z position (here D1546) of EF-hand ion-coordination binding sites is a key amino acid residue for the binding of Mgi to EF-hand motifs (Kawasaki and Kretsinger, 1994; da Silva et al., 1995; Yang et al., 2002).

The complementary mutation K1543D that is predicted to increase affinity for Mgi caused a substantial increase in sensitivity to Mgi inhibition, consistent with 4.1-fold reduction in apparent Kd for Mg2+ inhibition. Moreover, our results with the complementary mutation K1539D also suggest an increase in apparent affinity, although we could not quantitate the change because the concentration dependence did not conform to a single binding isotherm. Altogether, these results support a model in which Mgi directly interacts with the COOH-terminal EF-hand of the CaV1.2 channel to reduce current amplitude.

Because there is only a single EF-hand in the CaV1.2 channel, effects of Mg2+ acting through that site would be expected to conform to a single binding-site model. Using a global fitting protocol, we found that our results for WT, D1546A/N/S/K, and K1543D mutants could be successfully fit to a single binding site model with a Hill coefficient of 0.77 and 95% confidence limits of 0.16. This fit was significantly better than a fit to a two-site model. Therefore, our results support the conclusion that binding of a single Mg2+ ion to the COOH-terminal of EF-hand CaV1.2 channels substantially reduces peak Ba2+ or Ca2+ currents.

Our global fitting protocol also allows us to address the possibility that altered expression levels of the mutants compared with WT had a major effect on our results. We extrapolated the inhibition curves to Mgi = 0 to eliminate the different effects of Mgi on WT and mutants. This extrapolation was made without any normalization of the data at different Mgi concentrations. Comparison of these extrapolated values showed that none differed significantly from WT (P > 0.5), providing further support for our estimates of apparent Kd values for inhibition of CaV1.2 channels by Mg2+ binding to the EF-hand.

Comparison to Previous Work on Regulation of CaV1.2 Channels by Mg2+ and Protein Phosphorylation

The findings reported here are in agreement with those of other groups who concluded that the ion-coordinating residues of this EF-hand do not serve as a Ca2+ sensor for Ca2+-dependent inactivation of CaV1.2 channels (Zhou et al., 1997; Bernatchez et al., 1998; Peterson et al., 2000). Based on mutagenesis analysis of structural residues that do not participate in ion coordination in the EF-hand, Peterson et al. (2000) proposed that the F-helix (Fig. 2 B) serves as a transducer in mediating Ca2+-dependent inactivation. If this idea is correct, Mg2+ binding to the EF-hand may indirectly alter Ca2+-dependent inactivation. Further experiments will be required to assess this possibility. However, careful design and interpretation of such experiments will be necessary because we have found that Mg2+ binding to the COOH-terminal EF-hand substantially increases Ca2+-independent, voltage-dependent inactivation, which would complicate measurements of effects on Ca2+-dependent inactivation (Brunet et al., 2005).

The COOH-terminal domain of Cav1.2 channels has previously been established as an important locus for regulation by β-adrenergic receptor-stimulated phosphorylation by cAMP-dependent protein kinase and by Ca2+/calmodulin binding (Catterall, 2000). Our results add Mgi to the growing list of regulators that control Ca2+ channel activity through interaction with the COOH-terminal domain.

Previous work (Kuo and Hess, 1993) has shown that Mgi blocks unitary L-type inward currents through single Ca2+ channels in cultured pheochromocytoma cells, suggesting that direct pore block is an alternative mechanism through which Mgi could reduce peak Ba2+ currents in our experiments. However, the Kd for pore block in those experiments (Kuo and Hess, 1993) varied from ∼40 mM at −70 mV to 8 mM at +20 mV. In contrast, the block of Cav1.2 channels we observe here has a Kd of 0.7 mM and the Kd does not vary significantly between −20 and +80 mV. In addition, we found that block of Ca2+ and Ba2+ currents had approximately the same apparent Kd for Mgi, which is inconsistent with competitive block of the pore because the ion coordination site in the pore has much higher affinity for Ca2+ than Ba2+. Therefore, direct pore block by Mgi is unlikely to contribute significantly to our measurements and also appears unlikely to occur under physiological conditions in cardiac myocytes.

Although our results establish an important role for direct binding of Mgi to the COOH-terminal EF-hand in modulation of L-type Ca2+ currents conducted by Cav1.2 channels, they do not exclude additional important effects of Mgi that may be mediated indirectly through changes in the activities of protein kinases or phosphoprotein phosphatases (White and Hartzell, 1988; Agus et al., 1989; Yamaoka and Seyama, 1996; Pelzer et al., 2001; Wang et al., 2004). The effects of kinases and phosphatases might be more pronounced in cardiac myocytes, where their concentrations are higher and specific targeting to Ca2+ channels by anchoring proteins or scaffolding proteins may enhance their effects. Thus, it will be important to examine the range of functional effects in cardiac myocytes that can be ascribed to the direct actions of Mgi through the EF-hand of Ca2+ channels as described here and to determine whether there are additional regulatory effects that are related to modification of the enzymatic activities of kinases and phosphatases.

Effects of Mg2+ on the Voltage Dependence of Activation and Inactivation of Cav1.2 Channels

Significant shifts in the voltage dependence of activation and inactivation were observed at high (7.2 mM) and low (≤0.26 mM) Mgi concentrations. Most of these effects likely result from screening of negative charges on the intracellular surface of the Ca2+ channel protein or surrounding membrane because increasing Mgi causes negative shifts of voltage dependence in most cases, consistent with an intracellular charge-screening mechanism. In contrast to these charge-screening effects, at low Mgi (0.26 and 0.1 mM Mgi) the extent of voltage-dependent inactivation at positive membrane potentials was substantially reduced, especially in the K1539D and K1543D mutants. These results suggest that the introduction of a negative charge in the ion-coordination site in the COOH-terminal EF-hand alters voltage-dependent inactivation through an allosteric effect on the inactivation gating process. Binding of Mg2+ to this EF-hand in WT CaV1.2 channels may be a necessary structural element for normal voltage-dependent inactivation.

Potential Physiological Role of Inhibition of Cav1.2 Channels by Mg2+ Binding

What is the physiological role of inhibition of cardiac Ca2+ channels by Mg2+ binding? We propose that Mgi modulation of CaV1.2 may be an important part of a cardiac stress response to reduced energy metabolism designed to maintain Ca2+ homeostasis. In ischemic stress, rising Mgi as a result of increased Mg-ATP hydrolysis is sensed by at least three key proteins important for EC-coupling: the Cav1.2 channel, the type-2 ryanodine-sensitive Ca2+ release channel RYR2, and the sarcoplasmic reticulum Ca2+ pump SERCA2a. Elevated Mgi inhibits Cav1.2 (White and Hartzell, 1988, 1989; Hartzell and White, 1989; Agus et al., 1989; Yamaoka and Seyama, 1996; Pelzer et al., 2001; Wang et al., 2004) and RYR2 (Copello et al., 2002) but stimulates SERCA2a (Mahey and Katz, 1990). Thus, the overall effect of increased Mgi would be to maintain intracellular Ca2+ at a low level. The Mgi increase is terminated when ATP production is reestablished and buffers Mgi back to pre-ischemic levels. Therefore, we propose that by sensing intracellular Mgi the Cav1.2 channel is able to monitor the metabolic status of the cardiac myocyte, adjust the amount of trigger Ca2+ that enters the cell, and ultimately fine tune the contractile function and cardiac output of the heart under conditions of stress.

In addition to the effects of changes in Mgi levels during ischemic stress, Lehnart et al. (2004) described several mutations of the RYR2 that have reduced Mgi sensitivity. Interestingly, these mutations are associated with catecholaminergic polymorphic ventricular tachycardia (CPVT) (Priori et al., 2002), in which patients have arrhythmic episodes and sudden death mainly during periods of exercise-induced stress. Evidently, the decreased Mgi sensitivity of RYR2 may generate arrhythmias in response to rises in Cai during exercise-induced stress, in addition to the altered inhibitory role of FKBP12.6 (calstabin2), which is also caused by these mutations (Wehrens et al., 2004). Thus, defects in Mgi homeostasis and/or altered sensitivity of proteins to Mgi could be associated with negative clinical outcome in patients under conditions of increased stress, including ischemia, exercise, cardiac hypertrophy, and heart failure. Our findings indicate that modulation of Cav1.2 channels by Mgi binding to the COOH-terminal EF-hand may contribute to the multiplicity of effects of this crucial intracellular ion in both physiological and pathophysiological states.

We thank Dr. Robert Kretsinger (University of Virginia, Charlottesville, VA) for valuable discussions of the structure and function of EF-hand motifs.

The authors gratefully acknowledge the financial support provided by the National Institutes of Health (P01 HL 44948) to W.A. Catterall and R. Klevit, and a postdoctoral research fellowship from the American Heart Association to S. Brunet.

Olaf S. Andersen served as editor.

Agus, Z.S., E. Kelepouris, I. Dukes, and M. Morad.
1989
. Cytosolic magnesium modulates calcium channel activity in mammalian ventricular cells.
Am. J. Physiol.
256
:
C452
–C455.
Bernatchez, G., D. Talwar, and L. Parent.
1998
. Mutations in the EF-hand motif impair the inactivation of barium currents of the cardiac α1C channel.
Biophys. J.
75
:
1727
–1739.
Bers, D.M., C.W. Patton, and R. Nuccitelli.
1994
. A practical guide to the preparation of Ca2+ buffers.
Methods Cell Biol.
40
:
3
–29.
Brunet, S., T. Scheuer, R. Klevit, and W.A. Catterall.
2004
. Modulation of CaV1.2 channels by Mg2+ binding to an EF-hand motif in the C-terminal domain.
Biophys. J.
86
:
274a
.
Brunet, S., T. Scheuer, and W.A. Catterall. 2005. Intracellular magnesium enhances voltage-dependent inactivation of CaV1.2 by binding to the C-terminal EF-hand. Biophys. Soc. Abst. 976-Plat.
Catterall, W.A.
2000
. Structure and regulation of voltage-gated Ca2+ channels.
Annu. Rev. Cell Dev. Biol.
16
:
521
–555.
Copello, J.A., S. Barg, A. Sonnleitner, M. Porta, P. Diaz-Sylvester, M. Fill, H. Schindler, and S. Fleischer.
2002
. Differential activation by Ca2+, ATP and caffeine of cardiac and skeletal muscle ryanodine receptors after block by Mg2+.
J. Membr. Biol.
187
:
51
–64.
da Silva, A.C., and F.C. Reinach.
1991
. Calcium binding induces conformational changes in muscle regulatory proteins.
Trends Biochem. Sci.
16
:
53
–57.
da Silva, A.C., J. Kendrick-Jones, and F.C. Reinach.
1995
. Determinants of ion specificity on EF-hands sites. Conversion of the Ca2+/Mg2+ site of smooth muscle myosin regulatory light chain into a Ca2+-specific site.
J. Biol. Chem.
270
:
6773
–6778.
De Jongh, K.S., C. Warner, A.A. Colvin, and W.A. Catterall.
1991
. Characterisation of the two size forms of the α1 subunit of skeletal L-type calcium channels.
Proc. Natl. Acad. Sci. USA.
88
:
10778
–10782.
De Jongh, K.S., B.J. Murphy, A.A. Colvin, J.W. Hell, M. Takahashi, and W.A. Catterall.
1996
. Specific phosphorylation of a site in the full-length form of the α1 subunit of the cardiac L-type calcium channel by adenosine 3′, 5′-cyclic monophosphate-dependent protein kinase.
Biochemistry.
35
:
10392
–10402.
de Leon, M., Y. Wang, L. Jones, E. Perez-Reyes, X. Wei, T.W. Soong, T.P. Snutch, and D.T. Yue.
1995
. Essential Ca2+-binding motif for Ca2+-sensitive inactivation of L-type Ca2+ channels.
Science.
270
:
1502
–1506.
Elin, R.J.
1994
. Magnesium: the fifth but forgotten electrolyte.
Am. J. Clin. Pathol.
102
:
616
–622.
Freudenrich, C.C., E. Murphy, L.A. Levy, R.E. London, and M. Lieberman.
1992
. Intracellular pH modulates cytosolic free magnesium in cultured chicken heart cells.
Am. J. Physiol.
262
:
C1024
–C1030.
Haigney, M.C., S. Wei, S. Kaab, E. Griffiths, R. Berger, R. Tunin, D. Kass, W.G. Fisher, B. Silver, and H. Silverman.
1998
. Loss of cardiac magnesium in experimental heart failure prolongs and destabilizes repolarization in dogs.
J. Am. Coll. Cardiol.
31
:
701
–706.
Hartzell, H.C., and R.E. White.
1989
. Effects of magnesium on inactivation of the voltage-gated calcium current in cardiac myocytes.
J. Gen. Physiol.
94
:
745
–767.
Headrick, J.P., and R.J. Willis.
1991
. Cytosolic free magnesium in stimulated, hypoxic, and underperfused rat heart.
J. Mol. Cell. Cardiol.
23
:
991
–999.
Houdusse, A., and C. Cohen.
1996
. Structure of the regulatory domain of scallop myosin at 2 Å resolution: implications for regulation.
Structure.
4
:
21
–32.
Hulme, J.T., T.W. Lin, R.E. Westenbroek, T. Scheuer, and W.A. Catterall.
2003
. β-Adrenergic regulation requires direct anchoring of PKA to cardiac CaV1.2 channels via a leucine zipper interaction with A kinase-anchoring protein 15.
Proc. Natl. Acad. Sci. USA.
100
:
13093
–13098.
Kawasaki, H., and R.H. Kretsinger.
1994
. Calcium-binding proteins. 1: EF-hands.
Protein Profile.
1
:
343
–517.
Kuo, C.C., and P. Hess.
1993
. Block of the L-type Ca2+ channel pore by external and internal Mg2+ in rat phaeochromocytoma cells.
J. Physiol.
466
:
683
–706.
Lehnart, S.E., X.H. Wehrens, P.J. Laitinen, S.R. Reiken, S.X. Deng, Z. Cheng, D.W. Landry, K. Kontula, H. Swan, and A.R. Marks.
2004
. Sudden death in familial polymorphic ventricular tachycardia associated with calcium release channel (ryanodine receptor) leak.
Circulation.
109
:
3208
–3214.
Lewit-Bentley, A., and S. Rety.
2000
. EF-hand calcium-binding proteins.
Curr. Opin. Struct. Biol.
10
:
637
–643.
Mahey, R., and S. Katz.
1990
. A non-specific Ca2+ (or Mg2+)-stimulated ATPase in rat heart sarcoplasmic reticulum.
Mol. Cell. Biochem.
96
:
175
–182.
Margolskee, R.F., B. McHendry-Rinde, and R. Horn.
1993
. Panning transfected cells for electrophysiological studies.
Biotechniques.
15
:
906
–911.
McCully, J.D., and S. Levitsky.
1997
. Mechanisms of in vitro cardioprotective action of magnesium on the aging myocardium.
Magnes. Res.
10
:
157
–168.
McGowan, C.H., and P. Cohen.
1988
. Protein phosphatase-2C from rabbit skeletal muscle and liver: an Mg2+-dependent enzyme.
Methods Enzymol.
159
:
416
–426.
Murphy, E.
2000
. Mysteries of magnesium homeostasis.
Circ. Res.
86
:
245
–248.
Murphy, E., C. Steenbergen, L.A. Levy, B. Raju, and R.E. London.
1989
. Cytosolic free magnesium levels in ischemic rat heart.
J. Biol. Chem.
264
:
5622
–5627.
Pelzer, S., C. La, and D.J. Pelzer.
2001
. Phosphorylation-dependent modulation of cardiac calcium current by intracellular free magnesium.
Am. J. Physiol. Heart Circ. Physiol.
281
:
H1532
–H1544.
Peterson, B.Z., J.S. Lee, J.G. Mulle, Y. Wang, M. de Leon, and D.T. Yue.
1999
. Calmodulin is the Ca2+ sensor for Ca2+-dependent inactivation of L-type calcium channels.
Neuron.
22
:
549
–558.
Peterson, B.Z., C.D. DeMaria, J.P. Adelman, and D.T. Yue.
2000
. Critical determinants of Ca2+ -dependent inactivation within an EF- hand motif of L-type Ca2+ channels.
Biophys. J.
78
:
1906
–1920.
Priori, S.G., N. Napolitano, M. Memmi, B. Colombi, F. Drago, M. Gasparini, L. DeSimone, F. Coltorti, R. Bloise, R. Keegan, et al.
2002
. Clinical and molecular characterization of patients with catecholaminergic polymorphic ventricular tachycardia.
Circulation.
106
:
69
–74.
Tomaselli, G.F., and E. Marban.
1999
. Electrophysiological remodeling in hypertrophy and heart failure.
Cardiovasc. Res.
42
:
270
–283.
Wang, M., M. Tashiro, and J.R. Berlin.
2004
. Regulation of L-type calcium current by intracellular magnesium in rat cardiac myocytes.
J. Physiol.
555
:
383
–396.
Wehrens, X.H., S.E. Lehnart, S.R. Reiken, S.X. Deng, J.A. Vest, D. Cervantes, J. Coromilas, D.W. Landry, and A.R. Marks.
2004
. Protection from cardiac arrhythmia through ryanodine receptor-stabilizing protein calstabin2.
Science.
304
:
292
–296.
Wei, S.K., J.F. Quigley, S.U. Hanlon, B. O'Rourke, and M.C. Haigney.
2002
. Cytosolic free magnesium modulates Na/Ca exchange currents in pig myocytes.
Cardiovasc. Res.
53
:
334
–340.
White, R.E., and H.C. Hartzell.
1988
. Effects of intracellular free magnesium on calcium current in isolated cardiac myocytes.
Science.
239
:
778
–780.
White, R.E., and H.C. Hartzell.
1989
. Magnesium ions in cardiac function. Regulator of ion channels and second messengers.
Biochem. Pharmacol.
38
:
859
–867.
Xu, L., G. Mann, and G. Meissner.
1996
. Regulation of cardiac Ca2+ release channel (ryanodine receptor) by Ca2+, H+, Mg2+, and adenine nucleotides under normal and simulated ischemic conditions.
Circ. Res.
79
:
1100
–1109.
Yamaoka, K., and I. Seyama.
1996
. Modulation of Ca2+ channels by intracellular Mg2+ ions and GTP in frog ventricular myocytes.
Pflugers Arch.
432
:
433
–438.
Yang, W., H.W. Lee, H. Hellinga, and J.J. Yang.
2002
. Structural analysis, identification, and design of calcium-binding sites in proteins.
Proteins.
47
:
344
–356.
Zhou, J., R. Olcese, N.N.F. Qin, F. Noceti, L. Birnbaumer, and E. Stefani.
1997
. Feedback inhibition of Ca2+ channels by Ca2+ depends on short sequence of the C terminus that does not include the Ca2+-binding function of a motif with similarity to Ca2+-binding domains.
Proc. Natl. Acad. Sci. USA.
94
:
2301
–2305.
Zuhlke, R.D., G.S. Pitt, K. Deisseroth, R.W. Tsien, and H. Reuter.
1999
. Calmodulin supports both inactivation and facilitation of L-type calcium channels.
Nature.
399
:
159
–162.

Abbreviations used in this paper: Mgi, intracellular free magnesium; WT, wild-type.

Supplementary data