Channel phosphorylation and modulation of L-type Ca²⁺ currents by cytosolic Mg²⁺ concentration

Abstract

Previous studies have shown that inhibition of L-type Ca²⁺ current (I_Ca) by cytosolic free Mg²⁺ concentration ([Mg²⁺]_i) is profoundly affected by activation of cAMP-dependent protein kinase pathways. To investigate the mechanism underlying this counterregulation of I_Ca, rat cardiac myocytes and tsA201 cells expressing L-type Ca²⁺ channels were whole cell voltage-clamped with patch pipettes in which [Mg²⁺] ([Mg²⁺]_p) was buffered by citrate and ATP. In tsA201 cells expressing wild-type Ca²⁺ channels (α_1C/β_2A/α₂δ), increasing [Mg²⁺]_p from 0.2 mM to 1.8 mM decreased peak I_Ca by 76 ± 4.5% (n = 7). Mg²⁺-dependent modulation of I_Ca was also observed in cells loaded with ATP-γ-S. With 0.2 mM [Mg²⁺]_p, manipulating phosphorylation conditions by pipette application of protein kinase A (PKA) or phosphatase 2A (PP_2A) produced large changes in I_Ca amplitude; however, with 1.8 mM [Mg²⁺]_p, these same manipulations had no significant effect on I_Ca. With mutant channels lacking principal PKA phosphorylation sites (α_1C/S1928A/β_{2A/S478A/S479A}/α₂δ), increasing [Mg²⁺]_p had only small effects on I_Ca. However, when channel open probability was increased by α_1C-subunit truncation (α_1CΔ1905/β_{2A/S478A/S479A}/α₂δ), increasing [Mg²⁺]_p greatly reduced peak I_Ca. Correspondingly, in myocytes voltage-clamped with pipette PP_2A to minimize channel phosphorylation, increasing [Mg²⁺]_p produced a much larger reduction in I_Ca when channel opening was promoted with BAY K8644. These data suggest that, around its physiological concentration range, cytosolic Mg²⁺ modulates the extent to which channel phosphorylation regulates I_Ca. This modulation does not necessarily involve changes in channel phosphorylation per se, but more generally appears to depend on the kinetics of gating induced by channel phosphorylation.

Changes in cytosolic free Mg²⁺ concentration ([Mg²⁺]_i) have been found to have marked inhibitory effects on L-type Ca²⁺ current (I_Ca) (1, 18, 31, 41–43). Our recent study (37) also showed large changes in I_Ca around physiologically relevant [Mg²⁺]_i. Mg²⁺ inhibition of I_Ca could occur by alterations of ion permeation rate and/or modulation of channel gating kinetics. Nonetheless, at this point, evidence for significant block of divalent cation permeation through L-type channels by physiologically relevant [Mg²⁺]_i is lacking (25, 42, 44). Thus, at these concentrations, cytosolic Mg²⁺ is thought to modulate I_Ca predominantly by altering channel gating kinetics.

[Mg²⁺]_i effects on I_Ca are very much dependent on channel phosphorylation state (1, 31, 37, 41, 44, 45). Although the mechanism by which channel phosphorylation affects [Mg²⁺]_i actions on I_Ca is not clear, two alternatives have been proposed. [Mg²⁺]_i could affect the level of channel phosphorylation by altering the activity of Mg²⁺-dependent kinases and phosphatases, and it could also modulate the effect of channel phosphorylation on gating kinetics. Previous studies (31) have shown that the former mechanism is possible. In addition, the latter has been discussed as an explanation for the results of Yamaoka and Seyama (44); however, direct experimental support for this mechanism is lacking.

The purpose of this study is to investigate the relationship between channel phosphorylation and [Mg²⁺]_i modulation of L-type Ca²⁺ channel currents. We have studied [Mg²⁺]_i actions on I_Ca under different phosphorylation conditions in cardiac ventricular myocytes and tsA201 cells expressing wild-type L-type Ca²⁺ channels and mutant channels lacking putative protein kinase A (PKA) phosphorylation sites. Our data demonstrate that [Mg²⁺]_i regulation of I_Ca was pronounced under conditions that promote channel phosphorylation but was reduced under conditions antagonizing channel phosphorylation. However, changes in channel phosphorylation level were not necessary for Mg²⁺ actions, and even phosphorylation itself was not a prerequisite for [Mg²⁺]_i-dependent modulation of channel gating. Thus, while our data provide direct support for the hypothesis that cytosolic Mg²⁺ modulates the effect of channel phosphorylation on gating kinetics, the results also suggest that Mg²⁺ actions are, in general, more prominent under conditions that promote channel opening, irrespective of the specific mechanism that alters channel gating kinetics.

METHODS

Isolation of Ventricular Myocytes

Adult rat cardiac ventricular myocytes were isolated enzymatically from male Sprague-Dawley rats (200–225 g), as described previously (28). Animals received an intraperitoneal injection of sodium pentobarbitone (50–100 mg/kg), and after full anesthesia was achieved, a thoracotomy was performed to rapidly remove the heart, in accordance with the procedures approved by the Institutional Animal Care and Use Committee of the University of Medicine and Dentistry of New Jersey. After isolation, myocytes were stored in a refrigerator and used within 1 to 8 h.

Molecular Biology

Expression plasmids encoding the rabbit α_1C/pcDNA (GenBank accession no. X15539 ) (40), rabbit α_1C/S1928A/pCR3 (11), rat β_{2A/S478A/S479A}/pRBG (3), rat β_2A/pGW (GenBank accession no. M80545 ) (33), and rat α₂δ/pcDNA3 (GenBank accession no. NM-012929 ) (36) were supplied by Drs. M. Hosey and D. T. Yue. A truncated α₁-subunit (α_1CΔ1905), which consisted of 5,715 bases, was constructed using a standard PCR overlap extension technique with α_1C/S1928A/pCR3. In brief, α_1C/S1928A/pCR3 was first digested with restriction enzymes XbaI and EcoRV to produce a linear subfragment of α_1C/pCR3 containing 4,542 bases of α_1C. The complementary sequence between bases 4,543 and 5,715 of α_1C was created by standard PCR overlap extension to include a unique XbaI restriction site. The PCR product was then ligated with the linear subfragment of α_1C/pCR3 to yield α_1CΔ1905. This construct was verified by sequence analysis. The pCR3 vector carried both ampicillin- and kanamycin-resistant genes for bacterial selection.

Cell Culture

tsA201 cells, a subclone of the human embryonic kidney cell line (HEK)-293 expressing the simian virus-40 T antigen (a gift of Dr. Roman Shirokov), were maintained in a monolayer culture in DMEM/F-12 Ham’s medium (Sigma, St. Louis, MO), supplemented with 10% fetal bovine serum (Hyclone, Logan, UT), 100 U/ml penicillin G, and 100 μg/ml streptomycin at 37°C in a humidified atmosphere of air with 5% CO₂.

Expression of Channel Subunits

The tsA201 cells were used for 12–14 passages and were transferred every 4 days. Seventy percent confluent cultures of tsA201 cells were transiently cotransfected with a mixture of plasmids containing 20 μg α_1C (wild type/mutant), 20 μg β_2A (wild type/mutant), 20 μg α₂δ Ca²⁺ channel subunits, and 1 μg enhanced green fluorescent protein cDNA using the Ca²⁺ phosphate method (7). Twenty-four hours after transfection, the cells were replated onto coverslips at low density for an additional 2 days before electrophysiological recording to allow for optimal transient expression of heterologous proteins. Enhanced green fluorescent protein was used as a visual indicator for successful transfection of cells with 485-nm light.

Electrophysiology

Cells were voltage-clamped with an Axopatch 1C amplifier using standard patch-clamp techniques, as previously reported (37). Pipettes were pulled to resistances of 1.5–2.0 MΩ for tsA201 cells and 1.0–1.5 MΩ for myocytes when filled with a buffered salt solution (see below). After gigaohm seal formation and patch breakthrough, whole cell currents were elicited by depolarizing voltage steps of 200-ms duration to various test potentials from the holding potential. Transfected tsA201 cells were held at −80 mV and the cell membrane was directly depolarized to various test potentials. Myocytes were held at −70 mV and the cell membrane was depolarized to −40 mV with a 1-s ramp to suppress Na⁺ current before the test depolarization (30). In some cases, 30 μM TTX was also included in bathing solutions to block Na⁺ current. Outward K⁺ currents were blocked by Cs⁺ and tetraethylammonium (TEA) ions in the pipette solution. Experiments were conducted when whole cell voltage-clamps had time constants <300 μs without series resistance or capacitance compensation. Cell capacitance was estimated by integrating current elicited with 5-mV depolarizations from the holding potential.

In general, cells were depolarized every 30 s to a test potential of +20 mV in tsA201 cells and 0 mV in myocytes. Current-voltage (I-V) relationships were also obtained periodically by varying the test potential between −30 and +90 mV (in 10-mV increments) in tsA201 cells and between −30 and +60 mV in myocytes. In this case, depolarizations were elicited every 5 s. Current records were acquired by sampling at 10 kHz and applying a low-pass filter at 5 kHz. Data were analyzed using pCLAMP software, version 8.0 (Axon Instruments, Union City, CA). L-type channel current was measured as a 200 μM Cd²⁺-sensitive difference current. Membrane currents, dis played without linear leak subtraction, were obtained after 5 min in the whole cell configuration, as previously published (37).

Drugs and Chemicals

Solutions.

The pipette solution was composed of (in mM) 100 cesium gluconate, 10 PIPES (cesium salt), 15 TEA-Cl, 0.5 NaH₂PO₄, 0.1 Tris-GTP, 5 EGTA along with ATP, Mg-ATP, citric acid, magnesium citrate, MgCl₂, and CaCl₂ to produce free [Mg²⁺] ([Mg²⁺]_p) of 0.2 mM and 1.8 mM, pH 7.2, at a free [Ca²⁺] of 100 nM (WinMAXC 2.40, obtained at http://www.stanford.edu/~cpatton/maxc.html). The superfusion solution in experiments with cardiac myocytes was a modified Tyrode solution containing (in mM) 145 NaCl, 4 KCl, 2 (or 0.5) CaCl₂, 10 HEPES, 1 MgCl₂, and 10 glucose, pH 7.4, or a high-TEA bathing solution for tsA201 cells containing (in mM) 150 TEA-Cl, 10 CaCl₂, 10 HEPES, 1 MgCl₂, and 10 glucose, pH 7.4. Experiments were performed at room temperature (22–25°C).

Reagents.

Unless specified, reagents were obtained from Sigma. Adenosine 5′-O-(2-thiodiphosphate) tetralithium salt (ATP-γ-S; Calbiochem, San Diego, CA), catalytic subunit of PKA (Sigma), and protein phosphatase 2A (PP_2A; Upstate Biochemicals, Lake Placid, NY) were prepared as concentrated stock solutions that were directly added to pipette solutions before experiments. Several reagents purchased from Calbiochem (BAY K8644, IBMX, and TTX) and forskolin (Biomol, Plymouth Meeting, PA) were prepared as concentrated stock solutions that were applied to bathing solutions 30 min before experiments.

Data Analysis

Data are expressed as means ± SE for the number of cells indicated. Statistical significance was determined with the use of ANOVA and Student’s t-tests in commercial software (SigmaPlot, Jandel Scientific and JMP IN, Duxbury, Pacific Grove, CA). A P value <0.05 was considered statistically significant. The percent change, confidence intervals (95%), and SE for current density ratios were calculated using Fieller’s theorem (13).

RESULTS

Effect of [Mg²⁺]_i on Wild-Type L-Type I_Ca in tsA201 Cells

In studying mechanisms by which changes in [Mg²⁺]_i around its physiological level of 0.6–1.3 mM (6, 21, 38) might affect I_Ca in cardiac myocytes, most of our experiments have been performed with [Mg²⁺]_p in the range of 0.2 to 1.8 mM (37). These same conditions were used in experiments with tsA201 cells heterologously expressing L-type Ca²⁺ channel subunits (α_1C, β_2A, α₂δ). A time diary of these experiments is shown in Fig. 1A.

Fig. 1.

L-type Ca²⁺ current (I_Ca) recorded in tsA201 cells expressing wild-type channels (α_1C/β_2A/α₂δ) dialyzed with low and high pipette free [Mg²⁺] ([Mg²⁺]_p). A: time diaries of I_Ca in tsA201 cells depolarized to +20 mV from a holding potential of −80 mV during dialysis with pipette solutions containing 0.2 mM (●) and 1.8 mM [Mg²⁺] (▲). Time 0 coincides with patch breakthrough. Data are means ± SE. Eight experiments were performed for each [Mg²⁺]_p. Inset, superimposed currents recorded with 0.2 mM (a) and 1.8 mM (b) [Mg²⁺]_p. Horizontal and vertical calibration bar values are 50 ms and 5 pA/pF, respectively. B: current-voltage (I-V) relationships of I_Ca in tsA201 cells dialyzed with 0.2 mM [Mg²⁺]_p (●) and 1.8 mM [Mg²⁺]_p (▲). Currents were recorded 5 min after whole cell patch-clamp configuration was established. Experimental replicates are 7 and 10 for 0.2 and 1.8 mM [Mg²⁺]_p, respectively.

In the tsA201 cells dialyzed with 0.2 mM [Mg²⁺]_p, I_Ca increased for 1–4 min after patch breakthrough and then slowly decreased for the remainder of the recording period. The pattern of the time diary of I_Ca with 0.2 mM [Mg²⁺]_p in tsA201 cells was quite similar to that in cardiac myocytes (37). In the tsA201 cells dialyzed with 1.8 mM [Mg²⁺]_p, the initial current density at the first recording point (10 s) (5.3 ± 1.8 pA/pF, n = 8) was not different from that with 0.2 mM [Mg²⁺]_p (8.7 ± 2.6 pA/pF, n = 8). However, at all time points beyond 30 s after patch breakthrough, I_Ca was much smaller than that in cells patch-clamped with 0.2 mM [Mg²⁺]_p. Although I_Ca in the tsA201 cells dialyzed with 1.8 mM [Mg²⁺]_p also increased for 1–2 min after patch breakthrough, the initial increase of I_Ca was much smaller than that with 0.2 mM [Mg²⁺]_p. The relative amplitude of I_Ca measured with 1.8 mM vs. 0.2 mM [Mg²⁺]_p also changed rapidly within 3–5 min after patch breakthrough and changed slowly thereafter (ranging from 28% to 24% of I_Ca with 0.2 mM [Mg²⁺]_p at times >5 min). These results are consistent with our previous data in cardiac myocytes (37). As a consequence, results reported in subsequent experiments were recorded 5 min after establishing whole cell voltage clamps to allow adequate solution dialysis between pipette and cytosol. Representative currents recorded 5 min after patch breakthrough are shown in Fig. 1A, inset.

The I-V relationship was determined by a series of voltage pulses from −30 to +90 mV, as described in METHODS. With 0.2 mM [Mg²⁺]_p, the threshold for I_Ca activation occurred at −20 mV and peak current was observed at 30 mV (Fig. 1B). The membrane potential dependence of this current is shifted approximately +20 mV compared with I_Ca observed in cardiac myocytes, consistent with other reports, in which I_Ca has been measured in heterologous expression systems (e.g., Ref. 11). When [Mg²⁺]_p was increased from 0.2 to 1.8 mM, peak I_Ca amplitude was decreased from 22.7 ± 3.5 pA/pF (n = 7) to 5.4 ± 0.9 pA/pF (n = 10), a 76 ± 4.5% (n = 7) reduction (Table 1), and the peak of the I-V relationship was shifted 5–10 mV in the negative direction (Fig. 1B).

Table 1.

Effect of [Mg²⁺]_i on I_Ca in tsA201 cells expressing heterologous channel subunits

Channels	Phosphorylation Conditions	ICa, pA/pF		%Reduction of ICa
		0.2 mM [Mg2+]p	1.8 mM [Mg2+]p
	Control	22.7±3.5 (7)	5.4±0.9*(10)	76±4.5
	PKA	36.4±4.0 (7)	5.9±0.8*(7)	83±3.1
α1C/β2A/α2δ	PP2A	12.9±1.7 (12)	6.9±0.5*(7)	46±7.8†
	PKA+ATP-γ-S	34.5±4.2 (8)	8.7±1.2*(8)	73±4.8
α1C/S1928A/β2A/S478A/S479A/α2δ	NA	15.5±2.0 (10)	11.1±1.1*(9)	21±12.5†
α1CΔ19Q5/β2A/S478A/S479A/α2δ	NA	28.4±2.4 (11)	8.6±1.6*(12)	69±4.7

Values are means ± SE. Numbers in parentheses indicate replicates for each experimental condition. I_Ca, L-type Ca²⁺ current; PP_2A, protein phosphatase 2A; [Mg²⁺]_p, pipette free [Mg²⁺]; [Mg²⁺]_i; cytosolic free [Mg²⁺]; NA, not applicable because PKA did not affect currents carried by these mutant channels. I_Ca density in tsA201 cells voltage-clamped with patch electrodes containing the indicated [Mg²⁺]_p. Currents were recorded at 5 min after breakthrough into the whole cell patch-clamp configuration.

^*P < 0.05, statistically significant changes of I_Ca, comparing high (1.8 mM) vs. low (0.2 mM) [Mg²⁺]_p.

^†A reduction of I_Ca by [Mg²⁺]_i that is significantly different (P < 0.05) than that produced under control conditions with α_1C/β_2A/α₂δ expression.

Effects of [Mg²⁺]_i on Wild-Type L-Type Channels in tsA201 Cells Under Low and High Phosphorylation Conditions

Previous studies (1, 31, 37, 41, 44, 45) have shown that [Mg²⁺]_i effects on I_Ca are dependent on channel phosphorylation state in cardiac myocytes. To determine whether these effects could be reproduced in tsA201 cells, phosphorylating conditions were manipulated by adding protein kinases and phosphatases to the patch pipette solution. Inclusion of PKA in the pipette solution increased I_Ca amplitude by ~50% (Fig. 2A, middle) compared with control conditions (Fig. 2A, left) with 0.2 mM [Mg²⁺]_p, an indication that inclusion of PKA promoted channel phosphorylation. Under these conditions, increasing [Mg²⁺]_p from 0.2 to 1.8 mM reduced peak I_Ca amplitude by 83 ± 3.1% (Table 1). Conversely, inclusion of PP_2A, a Ca²⁺- and Mg²⁺-independent phosphatase (19), in the patch pipette solution to promote channel dephosphorylation yielded a 40% decrease in I_Ca amplitude with 0.2 mM [Mg²⁺]_p (Fig. 2A, right vs. left panels). Furthermore, increasing [Mg²⁺]_p to 1.8 mM decreased peak I_Ca amplitude by 46 ± 7.8% (Table 1), significantly smaller than that observed under control conditions. Figure 2B shows the I-V relationships for these experiments. In all cases, increasing [Mg²⁺]_p caused a leftward shift in the peak of the I-V relationship. Taken together, the data showed that manipulating channel phosphorylation conditions in the presence of 0.2 mM [Mg²⁺]_i allowed the amplitude of I_Ca to be changed over a threefold range, whereas the same maneuvers with 1.8 mM [Mg²⁺]_i resulted in no significant change in I_Ca amplitude. This demonstration that [Mg²⁺]_i modulation of I_Ca is dependent on channel phosphorylation conditions is consistent with published effects of [Mg²⁺]_i in cardiac myocytes (37, 41). Thus the tsA201 expression system would appear to be a good model to study [Mg²⁺]_i-dependent effects on I_Ca.

Fig. 2.

Effect of [Mg²⁺]_p on I_Ca in tsA201 cells expressing wild-type channels (α_1C/β_2A/α₂δ) under low and high phosphorylation conditions. A: typical I_Ca records at a test potential of +20 mV in cells dialyzed with 0.2 mM (solid symbols) and 1.8 mM [Mg²⁺]_p (open symbols) in control conditions (left) and with 30 U/ml PKA (middle) and 1 U/ml phosphatase 2A (PP_2A; right) added to the pipette solution. B: I-V relationships for I_Ca in cells dialyzed with 0.2 mM (solid symbols) and 1.8 mM [Mg²⁺]_p (open symbols) with control (circle) and PKA (triangle)- and PP_2A (square)-containing pipette solutions. Currents were recorded at 5 min after breakthrough into the whole cell patch-clamp configuration. The number of experiments ranges from 7 to 10.

Effects of [Mg²⁺]_i on Wild-Type L-Type Channels in tsA201 Cells Under Thiophosphorylation Conditions

Our previous studies in cardiac myocytes (37) and our present data above in tsA201 cells demonstrated that under conditions promoting channel phosphorylation, increasing [Mg²⁺]_p from 0.2 to 1.8 mM has a profoundly antagonistic effect on I_Ca amplitude, whereas conditions promoting channel dephosphorylation reduce [Mg²⁺]_i effects on I_Ca amplitude. This phosphorylation dependence of [Mg²⁺]_i effects on I_Ca could be explained by a change in the channel phosphorylation level and/or a modulation of the effects of channel phosphorylation on gating kinetics. To distinguish between these two possible mechanisms, tsA201 cells expressing wild-type L-type channels were voltage-clamped with patch electrodes containing solutions that included the catalytic subunit of PKA and 4 mM ATP-γ-S. Inclusion of ATP-γ-S with PKA should promote the formation of thiophosphorylated proteins (14) that are insensitive to potential changes in the activities of Mg²⁺-dependent phosphatases (19).

As expected, I_Ca amplitude in cells voltage-clamped with 0.2 mM [Mg²⁺]_p and PKA was significantly increased (Fig. 3A). Under these conditions, increasing [Mg²⁺]_p from 0.2 mM to 1.8 mM decreased peak I_Ca amplitude by 73 ± 4.8% (Table 1) and shifted the peak of the I-V relationship in the negative direction (Fig. 3B). These effects are comparable to those under basal (Fig. 1B) and high phosphorylation (Fig. 2B) conditions. Thus the data suggest that Ca²⁺ channel phosphorylation need not change in order for intracellular Mg²⁺ to strongly modulate gating kinetics.

Fig. 3.

Effect of [Mg²⁺]_p on I_Ca in tsA201 cells expressing wild-type channels (α_1C/β_2A/α₂δ) under thiophosphorylation conditions. A: typical I_Ca records at a test potential of +20 mV in cells dialyzed with 0.2 mM (●, ▲) and 1.8 mM [Mg²⁺]_p (○, △) in absence (left) and presence of 4 mM adenosine 5′-O-(2-thiodiphosphate) tetralithium salt (ATP-γ-S) plus 30 U/ml PKA (right). B: I-V relationships for I_Ca in cells dialyzed with 0.2 mM (solid symbol) and 1.8 mM [Mg²⁺]_p (open symbols) in the absence (●, ○) and presence of ATP-γ-S and PKA (▲, △), respectively. Currents were recorded at 5 min after breakthrough into the whole cell patch-clamp configuration. The number of experiment replicates is 7 or 8.

Effect of [Mg²⁺]_i on Mutant L-Type Channels Lacking Putative PKA Phosphorylation Sites in tsA201 Cells

Is channel phosphorylation an absolute requirement for Mg²⁺ modulation on the L-type Ca²⁺ channels? To answer this question, Ca²⁺ channels lacking putative PKA-sensitive phosphorylation sites were expressed in tsA201 cells to determine whether Mg²⁺ would modulate gating in these channels. Although the L-type Ca²⁺ channel contains many potential Ser/Thr kinase sites, only three serines actually appear to become phosphorylated by PKA in situ, Ser1928 in the α-subunit and Ser478, Ser479 in the β-subunit (3, 11, 18, 29, 32). For this reason, α- and β-subunits with all three PKA-sensitive Ser mutated to Ala (α_1C/S1928A, β_{2A/S478A/S479A}) were expressed with the α₂δ-subunit in tsA201 cells. This triple Ser to Ala mutant channel was designed to mimic low phosphorylation conditions induced by PP_2A in wild-type channels.

With the triple Ser to Ala mutant channel (Fig. 4A, left), current densities (15.5 ± 2.0 pA/pF, n = 10) were lower than those observed with wild-type channels (22.7 ± 3.5 pA/pF, n = 7) when cells were voltage-clamped with patch pipettes containing 0.2 mM Mg²⁺. Inclusion of PKA in the pipette solution had no effect on I_Ca amplitude at 0.2 mM [Mg²⁺]_p (Fig. 4B), confirmation that the Ala mutations rendered the channel insensitive to this kinase. Peak I_Ca amplitude was 11.1 ± 1.1 pA/pF (n = 9) in cells expressing this mutant channel and voltage-clamped with 1.8 mM [Mg²⁺]_p. This current density was calculated to be only 21 ± 12.5% (n = 9) less than that observed in cells voltage-clamped with electrodes containing 0.2 mM [Mg²⁺]_p (Fig. 4B; Table 1), a much smaller decrease in current than observed with wild-type channels expressed in these cells. Thus these data show that the lack of PKA-sensitive phosphorylation sites resulted in a significantly reduced effect of [Mg²⁺]_i on channel function, analogous to the effect of PP_2A on [Mg²⁺]_p modulation of I_Ca.

Fig. 4.

Effect of [Mg²⁺]_p on I_Ca in tsA201 cells expressing mutated Ca²⁺ channels. A: typical I_Ca records at a test potential of +20 mV with 0.2 mM (solid symbols) and 1.8 mM (open symbols) [Mg²⁺]_p in cells expressing α_1C/S1928A/β_{2A/S478A/S479A}/α₂δ (left) and α_1CΔ1905/β_{2A/S478A/S479A}/α₂δ (right) channel subunits. B: I-V relationships in cells expressing α_1C/S1928A/β_{2A/S478A/S479A}/α₂δ (●, ○) and α_1CΔ1905/β_{2A/S478A/S479A}/α₂δ (▲, △) channel subunits with 0.2 mM (solid symbol) and 1.8 mM [Mg²⁺]_p (open symbol). Experiments in which PKA was included in the patch pipette solution are indicated for cells expressing α_1C/S1928A/β_{2A/S478A/S479A}/α₂δ (■) and α_1CΔ1905/β_{2A/S478A/S479A}/α₂δ (♦) channel subunits with 0.2 mM [Mg²⁺]_p. Currents were recorded at 5 min after breakthrough into the whole cell patch-clamp configuration. The number of experiments ranged from 7 to 12.

In a second series of experiments, an α-subunit truncated proximal to Ser1928 (α_1CΔ1905) was expressed with the β-subunit containing the double Ser to Ala substitution (β_{2A/S478A/S479A}) and α₂δ. This truncation of the α-subunit has been reported previously to dramatically increase channel opening (10, 39), analogous to the effects of PKA phosphorylation on channel gating. In tsA201 cells expressing channels with α_1CΔ1905 (Fig. 4A, right), Ca²⁺ current densities with 0.2 mM [Mg²⁺]_p (28.4 ± 2.4 pA/pF, n = 11) were larger than those observed with wild-type channels. Inclusion of PKA in the pipette solution had no effects on I_Ca amplitude at 0.2 mM [Mg²⁺]_p (Fig. 4B), again confirming that removal of phosphorylation sites in α- and β-subunits was effective at abolishing the action of PKA on channel function. Interestingly, increasing [Mg²⁺]_p from 0.2 to 1.8 mM reduced the peak I_Ca amplitude by 69 ± 4.7% (Table 1). The magnitude of this [Mg²⁺]_i-dependent reduction was not significantly different than that observed with wild-type channels. As a consequence, these data directly demonstrate that PKA-dependent channel phosphorylation need not occur in L-type channels for [Mg²⁺]_i to have a pronounced effect on I_Ca.

Relevance of Intracellular Mg²⁺ Effects in Cardiac Myocytes

A pertinent question at this point is whether the results obtained in tsA201 cells are applicable to cardiac myocytes. To determine whether a change in channel phosphorylation is necessary for [Mg²⁺]_i effects, thiophosphorylation experiments were repeated in myocytes preincubated for 30 min with 10 μM forskolin and 300 μM IBMX in modified Tyrode solution containing 0.5 mM Ca²⁺ (37). I_Ca was then measured 5 min after patch breakthrough with a pipette solution containing 4 mM ATP-γ-S. Under these conditions, increasing [Mg²⁺]_p from 0.2 mM to 1.8 mM decreased the peak I_Ca amplitude from 16.4 ± 2.3 pA/pF (n = 6) to 4.4 ± 0.4 pA/pF (n = 6), producing a marked reduction of peak I_Ca (69 ± 2.6%, n = 6) (Fig. 5, A and B), comparable to that under basal and high phosphorylation conditions (see Ref. 37). These data, therefore, suggest that Ca²⁺ channel phosphorylation need not change for intracellular Mg²⁺ to strongly modulate gating kinetics in myocytes.

Fig. 5.

Effect of [Mg²⁺]_p on I_Ca in rat cardiac ventricular myocytes under thiophosphorylation conditions. A: I-V relationships for I_Ca in myocytes dialyzed with 0.2 mM (●) and 1.8 mM (▲) [Mg²⁺]_p with 4 mM ATP-γ-S in the pipette solution and inclusion of 10 μM forskolin and 300 μM IBMX in the 0.5 mM Ca²⁺-containing superfusion solution. External Ca²⁺ was 0.5 mM to avoid voltage errors due to high current density induced by forskolin and IBMX (and by BAY K8644 in Table 2). Our published data (37) showed that [Mg²⁺]_i effects are the same with 0.5 and 2.0 mM Ca²⁺-containing superfusion solutions. Currents were recorded at 5 min after breakthrough into the whole cell patch-clamp configuration. Six experiments were conducted for each [Mg²⁺]_p. B: typical I_Ca records at a test potential of 0 mV with 0.2 mM [Mg²⁺]_p (●) and 1.8 mM [Mg²⁺]_p (▲).

Data from tsA201 cells suggest that the channel phosphorylation is not necessary for [Mg²⁺]_i-dependent modulation of gating, if channel opening is promoted by other mechanisms. To test this idea, myocytes were exposed to 0.1 μM BAY K8644 to increase channel opening, independent of PKA phosphorylation. As expected, the addition of BAY K8644 to the superfusion solution significantly increased I_Ca compared with control conditions in cells voltage-clamped with 0.2 mM [Mg²⁺]_p (Table 2). In experiments with 0.2 and 1.8 mM [Mg²⁺]_p, peak I_Ca amplitude was 16.7 ± 0.9 pA/pF (n = 4) and 3.9 ± 0.6 pA/pF (n = 4), respectively, i.e., increasing [Mg²⁺]_p produced a 76 ± 2.9% decrease. To ensure that channel phosphorylation state did not affect this result, experiments were repeated in myocytes voltage-clamped with patch electrodes containing PP_2A. For purposes of comparison, Table 2 includes our previously published data (37) showing that, with PP_2A (1 U/ml) alone, increasing [Mg²⁺]_p from 0.2 mM to 1.8 mM produces a relatively small decrease in I_Ca. In the presence of PP_2A, 0.1 μM BAY K8644 still produced a significant increase in I_Ca to 11.6 ± 1.4 pA/pF (n = 5), but with 1.8 mM [Mg²⁺]_p, peak I_Ca density was reduced by 75 ± 2.9% (Table 2). This reduction in current was significantly larger than the effect of [Mg²⁺]_i in the presence of PP_2A alone, but similar to the reduction of I_Ca observed under control conditions in the absence of PP_2A (Table 2), with BAY K8644 alone and in high phosphorylation conditions produced with forskolin/IBMX/okadaic acid (see Ref. 37). Thus these results suggest that [Mg²⁺]_i effects are not uniquely dependent on channel phosphorylation state but gating kinetics that promote channel opening, consistent with the results in tsA201 cells.

Table 2.

Effect of [Mg²⁺]_i on I_Ca in cardiac myocytes with PP_2A and BAY K8644

	Conditions	ICa, pA/pF		%Reduction of ICa
		0.2 mM [Mg2+]p	1.8 mM [Mg2+]p
[Ca2+]o (2.0 mM)	Control	17.0±2.0 (12)	5.6±1.0*(10)	64±2.8
	PP2A	9.8±1.3† (12)	6.7±0.4*(16)	25±3.4†
[Ca2+]o (0.5 mM)	Control	8.0±0.7 (6)	1.9±0.3*(8)	75±2.4
	BAY K8644	16.7±0.9‡ (4)	3.9±0.6*(4)	76±2.9
	BAY K8644 + PP2A	11.6±1.4† (5)	2.5±0.3*(5)	75±2.9

Values are means ± SE. The numbers in parentheses indicate replicates for each experimental condition. I_Ca density in rat ventricular myocytes voltage-clamped with patch electrodes containing the indicated [Mg²⁺]_p. [Ca²⁺]_o (2.0 mM) values were from Ref. 37. Currents were recorded at 5 min after breakthrough into the whole cell patch-clamp configuration. PP_2A was included in patch electrode solutions at a concentration of 5 U/ml. BAY K8644 (0.1 μM) was added to superfusion solutions with the indicated [Ca²⁺]_o.

^*P < 0.05, statistically significant changes of I_Ca between 0.2 and 1.8 mM [Mg²⁺]_p.

^†P < 0.05, a reduction of I_Ca by [Mg²⁺]_i that is significantly different than that produced under control conditions.

^‡P < 0.05, statistically significant difference between control and experimental condition at the same [Mg²⁺]_p.

DISCUSSION

[Mg²⁺]_i Inhibition of L-type Ca²⁺ Channels

Data in this study demonstrated that increasing [Mg²⁺]_p from 0.2 to 1.8 mM had a marked inhibitory effect on wild-type L-type Ca²⁺ channels expressed in HEK-tsA201 cells, decreasing peak amplitude of I_Ca and shifting the peak of the I-V relationship 5–10 mV in the negative direction. These effects were in general similar to [Mg²⁺]_i modulation of I_Ca observed in our previous study with cardiac myocytes (37). Some previous studies (31, 42–45) have shown that cytosolic Mg²⁺ inhibition of I_Ca was manifested in the micromolar concentration range, whereas under defined experimental conditions that favor high phosphorylation, inhibition of I_Ca was only observed at [Mg²⁺]_i well >1 mM (31, 44, 45). In addition to the present data, some reports also show that changes of [Mg²⁺]_i around physiological levels can regulate L-type Ca²⁺ channels (1, 41); however, the degree of current inhibition is considerably smaller than that observed here and by the study of Wang et al. (37). Taken together, our data make the case that intracellular Mg²⁺ can be a potent modulator of I_Ca at physiological concentrations.

Channel Phosphorylation Affects [Mg²⁺]_i Inhibition of L-Type Ca²⁺ Channels

Under conditions that antagonized channel phosphorylation in tsA201 cells (with PP_2A in the patch pipette solution), an increase in [Mg²⁺]_p from 0.2 to 1.8 mM had a less pronounced effect on I_Ca than in basal conditions or those promoting channel phosphorylation (with PKA in the pipette solution), consistent with our previous studies in cardiac myocytes (37). These findings are also consistent with reports that cytosolic Mg²⁺ inhibits I_Ca with high affinity (in micromolar range) under basal, presumably low, phosphorylation conditions (43) but with low apparent affinity, i.e., >1 mM, under high phosphorylation conditions (31, 44, 45). In light of these findings, we would conclude that 0.2 mM [Mg²⁺]_i produces substantial inhibition of I_Ca under conditions antagonizing channel phosphorylation, so that increasing [Mg²⁺]_i to 1.8 mM would have only modest effects on I_Ca.

The importance of channel phosphorylation in [Mg²⁺]_i effects on I_Ca was further supported by data derived from Ca²⁺ channels, in which phosphorylation sites were mutated to Ala residues. Many potential phosphorylation sites are found in the α₁ (23, 26, 27, 35, 46) and β-subunits of L-type Ca²⁺ channels (12, 16); however, the major sites of PKA-dependent protein phosphorylation in heart muscle appear to be located at Ser1928 of the α_1C-subunit (8, 11, 29, 32) and two serine residues (Ser478 and Ser479) of the β_2A-subunit (3). Consistent with previous reports, our data demonstrated that these three serine residues are functional targets for phosphorylation because mutation of these three amino acids to Ala produced PKA-insensitive currents. In channels with all three Ser residues mutated to Ala (α_1C/S1928A/β_{2A/S478A/S479A}/α₂δ) increasing [Mg²⁺]_p from 0.2 to 1.8 mM produced a much smaller effect on the peak I_Ca amplitude than that observed in wild-type channels. [Mg²⁺]_i effects on these channels were, however, comparable to those observed in the presence of PP_2A. Thus, from these data, we conclude that cytosolic Mg²⁺ in the concentration range tested has only marginal (~20–25%) inhibitory effects in unphosphorylated channels.

Interestingly, with 0.2 mM [Mg²⁺]_p, current density in cells expressing α_1C/S1928A/β_{2A/S478A/S479A}/α₂δ Ca²⁺ channels was less than that observed when wild-type channels were expressed. If these mutant channels mimic the condition of low phosphorylation in wild-type channels produced by PP_2A, which reduces channel opening (15), this lower current density makes sense. However, an alternative explanation is that expression level of this mutant channel protein is lower. We have no evidence pertaining to expression levels, but irrespective of the reasons for the lower current density, these data, and the effects of PP_2A in cells expressing wild-type Ca²⁺ channels, suggest that current density is maintained by some level of channel phosphorylation that probably accounts for [Mg²⁺]_i actions under basal conditions.

Approximately 20–25% inhibition of peak I_Ca by high [Mg²⁺]_p was still observed on increasing [Mg²⁺]_p from 0.2 to 1.8 mM with Ca²⁺ channels containing the triple Ser to Ala mutation. One explanation for this result is that channel phosphorylation by other kinases such as PKC (24) and PKG (22) could also be affected by [Mg²⁺]_i. Another explanation may be that [Mg²⁺]_i blocks ion permeation through the channels. Although previous studies (25, 42, 44) with L-type Ca²⁺ channels suggest that cytosolic Mg²⁺, at the concentrations used in this study, would not affect divalent cation permeation rate, [Mg²⁺]_i in the 1 mM range has been reported to produce extensive block in voltage-dependent Na⁺ and K⁺ channels (4, 17). Thus the possibility that physiologically relevant [Mg²⁺]_i blocks ion permeation cannot be entirely neglected.

Does Channel Phosphorylation Need to Change for [Mg²⁺]_I Effects on I_Ca?

Two mechanisms have been proposed to explain the phosphorylation dependence of [Mg²⁺]_i effects on I_Ca. First, [Mg²⁺]_i might change the level of channel phosphorylation by regulating the activity of the many Mg²⁺-dependent kinases, phosphatases, and phosphodiesterases present in the cell (31, 41). Second, Mg²⁺ might change the effect of phosphorylation on channel gating kinetics (44). To distinguish between these two possible mechanisms, [Mg²⁺]_i effects on I_Ca were investigated under thiophosphorylation conditions induced by inclusion of ATP-γ-S in the pipette solutions used for experiments with tsA201 cells and myocytes. The addition of ATP-γ-S is reported to produce phosphatase-resistant thiophosphorylated channels (14) that would be insensitive to [Mg²⁺]_i-induced changes in enzyme activities. Under these conditions, increasing [Mg²⁺]_p from 0.2 mM to 1.8 mM produced a similar reduction of I_Ca amplitude to that observed in basal conditions and high phosphorylation conditions in both tsA201 cells and cardiac myocytes. Thus these data suggest that the level of Ca²⁺ channel phosphorylation need not change for intracellular Mg²⁺ to regulate Ca²⁺ channels. Thus we must consider the hypothesis that Mg²⁺ might change the effect of phosphorylation on channel-gating kinetics.

The data might also suggest that these Mg²⁺-dependent enzymes are not necessarily involved in [Mg²⁺]_i regulation on the Ca²⁺ channels, consistent with our previous result that Ca²⁺-calmodulin-dependent kinase II is not involved in [Mg²⁺]_i regulation of I_Ca (37); however, the data do not exclude the possibility that, under the appropriate conditions, such kinases, phosphatases and phosphodiesterases could participate in [Mg²⁺]_i modulation of L-type Ca²⁺ channels.

Is Channel Phosphorylation Absolutely Necessary for [Mg²⁺]_i Effects on I_Ca?

Previously published studies (1, 31, 37, 41, 44, 45) and our present data clearly demonstrate that channel phosphorylation affects [Mg²⁺]_i regulation of I_Ca. Nonetheless, one motivation for this study was to determine whether channel phosphorylation is necessary for [Mg²⁺]_i regulation of I_Ca. To address this question, we made a truncated α_1C subunit (α_1CΔ1905) that lacked 265 COOH-terminal amino acids including Ser1928. Channels containing this truncated α_1C-subunit have been reported to display increased current amplitude due to increased channel opening probability rather than by elevated channel expression (10, 39). We took advantage of this characteristic of α_1CΔ1905 and coexpressed the subunit along with a β-subunit containing the double Ser to Ala substitution (β_{2A/S478A/S479A}) to obtain a PKA-insensitive channel that displayed high channel opening. With this channel, increasing [Mg²⁺]_p from 0.2 to 1.8 mM produced a marked reduction of peak I_Ca amplitude similar to that observed under high phosphorylation conditions. These data suggested that channel phosphorylation is not absolutely necessary for [Mg²⁺]_i regulation on I_Ca.

Considering that truncation of α_1C-subunit is, in fact, functionally very similar to cAMP-dependent phosphorylation, i.e., it promotes channel openings (47), we supposed that [Mg²⁺]_i regulation on I_Ca might simply be dependent on the functional state of the channels. To test this idea, we used the dihydropyridine agonist BAY K8644 to induce channel opening (20) and PP_2A to inhibit channel phosphorylation at the same time in cardiac myocytes. Under these conditions, increasing [Mg²⁺]_p from 0.2 to 1.8 mM produced a marked inhibitory effect on I_Ca, which was much bigger than that under low phosphorylation conditions, i.e., PP_2A alone, but very similar to that under basal and high phosphorylation conditions produced by forskolin/IBMX/okadaic acid (37). Together with the data discussed above, these results suggest that [Mg²⁺]_i modulation of I_Ca channels is promoted by conditions that favor channel opening, including channel phosphorylation and pharmacological agonists.

[Mg²⁺]_i Effect on I_Ca and Channel Gating Modes

Three patterns of gating kinetics displayed by L-type Ca²⁺ channels have been classified as “gating modes” (20). In mode 0, channels fail to open during depolarization. In mode 1, channels display brief openings with a low frequency, whereas mode 2 gating is characterized by long-lasting and frequent channel openings (20). Transitions among these three gating modes are thought to be modulated by, among other factors, channel phosphorylation and dihydropyridines. The β-adrenergic receptor-PKA signaling pathways enhance transitions to mode 1 or mode 2 gating (2, 47). Long-lived, mode 2 gating is also promoted by pharmacological agents, such as BAY K8644 (9, 20) and structural modification of channels, such as the truncation of α_1C-subunit (10, 39). Conversely, transitions from mode 2 to mode 1 or mode 0 gating are stimulated by channel dephosphorylation induced by maneuvers such as cytosolic application of PP_2A (15).

Our data show that [Mg²⁺]_i-dependent effects were enhanced by seemingly any maneuver, including channel phosphorylation, pharmacological agents (BAY K8644), α_1C-subunit truncation (α_1CΔ1905) that increased channel availability, and/or favored mode 2 gating but were lessened by maneuvers, such as channel dephosphorylation (PP_2A), that decreased mode 2 gating. Thus one explanation for intracellular [Mg²⁺]_i modulation of channel gating is that Mg²⁺ shifts channel gating to low open probability states (mode 1) and decreases channel availability (mode 0). This hypothesis is quite plausible because, in addition to our data, Yamaoka and Seyama (42) reported that channel availability increased on lowering cytosolic Mg²⁺ from millimolar to micomolar concentrations, reflecting the shift of channels from silent (mode 0) to active modal states (mode 1).

The three experimental maneuvers that promote [Mg²⁺]_i inhibition of I_Ca, i.e., channel phosphorylation, BAY K8644 and truncation of α_1C-subunits, also promote channel opening. The degree to which these maneuvers share a common mechanism in altering channel gating kinetics is unknown. Nonetheless, it is parsimonious to assume that cytosolic Mg²⁺ acts via a common mechanism to decrease I_Ca. In this regard, much attention has been focused on intracellular structures of the Ca²⁺ channel that might be involved in the inactivation process (34). Several domains of the protein have been shown to be critical in current inactivation and facilitation, including the COOH-terminal tail of the α-subunit. In the COOH-terminal tail, one of those domains, an EF-hand motif, has recently been suggested as a cytosolic binding site at which Mg²⁺ can regulate channel gating (5). Given the proximity of the EF-hand motif to the other protein domains that are also critical for channel gating (34), such a conclusion does not at all seem at odds with our finding that Mg²⁺ can modulate I_Ca under a wide variety of conditions that promote channel opening. At the same time, considerable caution must be exercised in defining the structural basis for cytosolic Mg²⁺ actions given the apparent complexity of channel gating properties and the reported contributions by many protein domains in the gating process.