HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) All Versions of this Article: 575/2/443    most recent jphysiol.2006.109850v1 Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Almulla, H. A. Articles by Flatman, P. W. Search for Related Content PubMed PubMed Citation Articles by Almulla, H. A. Articles by Flatman, P. W. Related Collections Cellular

¹ Centre for Integrative Physiology, College of Medicine and Veterinary Medicine, The University of Edinburgh, Edinburgh EH8 9XD, UK

	Abstract

Top Abstract Introduction Methods Results Discussion References

 
The objective of our study was to investigate how Mg²⁺ enters mammalian cardiac cells. During this work, we found evidence for a previously undescribed route for Mg²⁺ entry, and now provide a preliminary account of its properties. Changes in Mg²⁺ influx into rat ventricular myocytes were deduced from changes in intracellular ionized Mg²⁺ concentration ([fMg²⁺]_i) measured from the fluorescence of mag-fura-2 loaded into isolated cells. Superfusion of myocytes at 37°C with Ca²⁺-free solutions with both reduced [Na⁺] and raised [Mg²⁺] caused myocytes to load with Mg²⁺. Uptake was seen with solutions containing 5 mM Mg²⁺ and 95 mM Na⁺, and increased linearly with increasing extracellular [Mg²⁺] or decreasing extracellular [Na⁺]. It was very sensitive to temperature (Q₁₀ > 9, 25--37°C), was observed even in myocytes with very low Na⁺ contents, and stopped abruptly when external [Na⁺] was returned to normal. Uptake was greatly reduced by imipramine or KB-R7943 if these were added when [fMg²⁺]_i was close to the physiological level, but was unaffected if they were applied when [fMg²⁺]_i was above 2 mM. Uptake was also reduced by depolarizing the membrane potential by increasing extracellular [K⁺] or voltage clamp to 0 mV. We suggest that initial Mg²⁺ uptake may involve several transporters, including reversed Na⁺–Mg²⁺ antiport and, depending on the exact conditions, reversed Na⁺–Ca²⁺ antiport. The ensuing rise of [fMg²⁺]_i, in conjunction with reduced [Na⁺], may then activate a new Mg²⁺ transporter that is highly sensitive to temperature, is insensitive to imipramine or KB-R7943, but is inactivated by depolarization.

(Received 16 March 2006; accepted after revision 20 June 2006; first published online 22 June 2006)
Corresponding author P. W. Flatman: Centre for Integrative Physiology, The University of Edinburgh, Hugh Robson Building, George Square, Edinburgh EH8 9XD, UK. Email: peter.flatman{at}ed.ac.uk

	Introduction

Top Abstract Introduction Methods Results Discussion References

 
Magnesium is essential for the maintenance and regulation of cellular functions. It affects membrane permeability and excitability (Fry & Proctor, 1993) and it influences the activity of many enzymes, especially those involved in energy generation and utilization (Cowan, 2002). It is necessary for high-fidelity transcription and translation, and it plays important roles in cell growth, division and development (Touyz & Yao, 2003; Rubin, 2005; He et al. 2005). Studies, predominantly with fluorescent indicators, show that only 0.8 mM of the total 17 mM Mg²⁺ in the cytoplasm of rat myocytes is free and ionized (Page & Polimeni, 1972; Murphy et al. 1989; Hongo et al. 1994; Handy et al. 1996; Watanabe & Konishi, 2001), the rest being bound to ligands such as ATP or sequestered in mitochondria (McGuigan et al. 2002). The concentration of ionized Mg²⁺ in rat extracellular fluids is about 0.6 mM (Altura et al. 1995). Thus, in the quiescent myocyte or during diastole, there is a strong inward electrochemical gradient (Mg²⁺ equilibrium potential, E_Mg = –4 mV) favouring Mg²⁺ entry, whereas during the plateau of the action potential, a small outward gradient exists. Recent work suggests that the concentration of cytoplasmic ionized Mg²⁺ ([fMg²⁺]_i) is maintained below electrochemical equilibrium by a Na⁺–Mg²⁺ antiport in the sarcolemma (Fry, 1986; Handy et al. 1996; Tashiro & Konishi, 2000; Tursun et al. 2005; Almulla et al. 2006). However, the identity and many of the properties of the influx pathway(s) remain a mystery. Such pathways must exist to allow the cells to grow and divide, or to regain Mg²⁺ after a loss.

During our studies on the mechanisms that remove Mg²⁺ from heart cells, we developed a protocol to load cells with Mg²⁺ (Handy et al. 1996). Like others (Fry, 1986; Buri et al. 1993; McGuigan et al. 2002), we found that it is very difficult to change [fMg²⁺]_i in cardiac myocytes by simply incubating them in high-[Mg²⁺] solutions. However, cells could be induced to take up Mg²⁺ by removing both Ca²⁺ and Na⁺ from the medium (Handy et al. 1996). Under these conditions, and with external [Mg²⁺] increased to 30 mM, [fMg²⁺]_i could be raised from 0.8 to above 4 mM within 15 min (Almulla et al. 2006). These results could be explained by a number of hypotheses: (1) removal of Na⁺ and Ca²⁺ may reduce competition for movement through Na⁺, Ca²⁺ or non-specific cation channels, allowing significant uptake of Mg²⁺; (2) a Mg²⁺-selective channel may become activated; and (3) Mg²⁺ may be taken up by reverse-mode Na⁺–Mg²⁺ antiport, or by the Na⁺–Ca²⁺ antiporter operating in reverse mode and transporting Mg²⁺ instead of Ca²⁺. In this paper, we explore the mechanisms responsible for Mg²⁺ uptake further by testing these hypotheses. We follow changes in [fMg²⁺]_i, measured with the fluorescent indicator mag-fura-2, in myocytes superfused with solutions of varying ionic composition both in the presence and absence of transport inhibitors. Previous work by ourselves and others has established that the changes in [fMg²⁺]_i seen under the conditions described here are due to changes in Mg²⁺ transport across the plasma membrane, rather than to changes in Mg²⁺ binding to cytoplasmic ligands like ATP or sequestration by organelles (Handy et al. 1996; Tashiro & Konishi, 2000; Tursun et al. 2005). We provide a detailed description of Mg²⁺ entry into cardiac myocytes, including a preliminary account of a new pathway through which large amounts of Mg²⁺ can rapidly cross the cell membrane. Some initial results have been published in abstract form (Steele et al. 2003; Almulla et al. 2003).

	Methods

Top Abstract Introduction Methods Results Discussion References

 
The concentration of cytoplasmic ionized Mg²⁺ was measured ratiometrically at 37°C in isolated rat ventricular myocytes containing the fluorescent Mg²⁺ indicator mag-fura-2, and superfused with appropriate test solutions. In some experiments, the cell's membrane potential was controlled by whole-cell voltage clamp. The methods have been described in detail (Almulla et al. 2006), so only a brief outline will be given here. Solutions were prepared in glass-distilled water with analytical grade reagents (BDH AnalaR, VWR International, Lutterworth, UK and Sigma-Aldrich, Gillingham, UK). KB-R7943 mesylate was from Tocris (Bristol, UK). Normal Tyrode solution contained (mM): NaCl, 134; KCl, 6; Hepes, 10; glucose, 10; CaCl₂, 1; and MgCl₂, 1; and pH was adjusted to 7.4 at 37°C with about 6 mM NaOH. Ca²⁺-free Tyrode solution had the same composition except that no Ca²⁺ (or Ca²⁺ chelator) was added to the solution. Where necessary, Na⁺ was replaced in superfusates by equimolar amounts of N-methyl-D-glucamine (NMDG), and increases in [MgCl₂] were balanced by reductions in [NMDG] to keep the osmolality constant. For instance, the composition of Na⁺-free, 30 mM Mg²⁺ loading solution was (mM): MgCl₂, 30; NMDG-Cl, 95; Hepes, 10; and glucose, 10; with pH adjusted to 7.4 at 37°C with approximately 6 mM KOH. Magnesium clamp solution was similar but with 1 mM MgCl₂ and 140 mM NMDG-Cl. Where necessary the concentrations of ionized Mg²⁺ or Ca²⁺ ([fCa²⁺]) in solutions containing known total concentrations of metals and ligands were calculated using the program Chelator (Schoenmakers et al. 1992).

Hearts were taken from male albino Sprague–Dawley rats after terminal anaesthesia by intraperitoneal injection of sodium pentobarbitone (120 mg kg^–1; Rhone Merieux, Harlow, UK) and heparinization (200 i.u.) according to advice from the UK Home Office. Isolated ventricular myocytes were obtained from these hearts by the collagenase method as previously described (Almulla et al. 2006). They were loaded with mag-fura-2 by incubating them for 30 min at room temperature in normal Tyrode solution containing 5 µM mag-fura-2 AM ester (Teflabs, Austin, TX, USA) and 0.03% Pluronic-F127 (Molecular Probes, Eugene, OR, USA). Cells were washed and stored in the dark at room temperature until required.

Myocytes were alternately illuminated with light at 340 and 380 nm, and the intensities of light emitted at wavelengths greater than 510 nm (F₃₄₀, F₃₈₀) were measured using a microspectrophotometer (Cairn Research Ltd, Sittingbourne, UK). Single, well-shaped ventricular cells were selected and continuously observed on a monitor, and the experiment was discarded if cell shortening or rounding was observed before the end of the protocol. Signals were averaged over 4 s and stored together with any corresponding voltage clamp data. Background fluorescence at both wavelengths was recorded after each experiment, and used to correct each data point prior to calculation of the fluorescence ratio (R; F₃₄₀/F₃₈₀). Fluorescence measurements were calibrated in vitro using Mg²⁺ standards in small droplets under oil at 37°C (Almulla et al. 2006). For experiments at 25°C, mag-fura-2 calibration and the pH adjustment of all superfusates were also made at 25°C. In some experiments, intracellular [Na⁺] was measured ratiometrically (excitation at 340 and 380 nm) with the dye SBFI (Molecular Probes) loaded as the acetyl methyl ester form. The protocols were the same as for mag-fura-2.

In experiments where myocytes were voltage clamped, we used patch electrodes containing an intracellular-like filling solution (mM: K⁺-aspartate, 120; NaCl, 10; MgATP, 5; CaCl₂, 0.2; EGTA, 10; and Hepes, 10; pH adjusted to 7.2 at 37°C with KOH; concentrations of ionized Ca²⁺ and Mg²⁺ estimated as 5 nM and 0.6 mM, respectively) connected to a unity gain amplifier head stage (HS-2A, Axon Instruments Inc., Sunnyvale, CA, USA) and thence to an Axoclamp-2A (Axon Instruments Inc.). Recording of membrane potential was made in bridge mode. Signals were digitized (CED 1401, Cambridge, UK) and captured using Spike 2.2 software (CED) at a rate of 0.25 Hz.

Where possible, data are given with their standard errors. The significance of differences between control and experimental means was assessed with Student's two-tailed, unpaired (paired if stated) t test using GraphPad Prism version 4.03 (GraphPad Software, San Diego, CA, USA). Differences were considered significant if P < 0.05.

	Results

Top Abstract Introduction Methods Results Discussion References

 
We have previously shown that rat myocytes take up substantial amounts of Mg²⁺ when superfused with a Ca²⁺- and Na⁺-free medium containing additional Mg²⁺ (Handy et al. 1996; Almulla et al. 2006). Here we explore the effects of changing medium [Na⁺] ([Na⁺]_o) and [Mg²⁺] ([Mg²⁺]_o) on the process. Superfusate tonicity was maintained at the physiological level with NMDG to prevent changes in cell volume. The effects of [Mg²⁺] above 30 mM were not explored to avoid problems with maintaining ionic strength. Myocytes were initially superfused with normal Tyrode solution until the measurements of [fMg²⁺]_i stabilized. They were then superfused with Ca²⁺-free Tyrode solution for 2–3 min to prevent increases in [Ca²⁺]_i when [Na⁺]_o was subsequently reduced.

Sodium dependence of [fMg²⁺]_i increase

Myocytes were superfused with Ca²⁺-free solutions containing 30 mM Mg²⁺ and between 0 and 95 mM Na⁺. After a few minutes in the low-[Na⁺] solution, [fMg²⁺]_i usually started to rise and then increased almost linearly with time for several minutes (Fig. 1). The rates reported here were measured during this phase of [fMg²⁺]_i rise. On returning to normal Tyrode solution, [fMg²⁺]_i returned to the baseline level of approximately 0.75 mM. We found that the rate of [fMg²⁺]_i rise at any particular [Na⁺]_o was variable between cells, and with a few cells (< 10%) apparently failing to load at all over the time span of an experiment. Initially, this variability prevented kinetic analysis of the effects of [Na⁺]_o on [fMg²⁺]_i increase. To circumvent this problem, we devised a normalization procedure. We found that cells could be subjected to load–unload protocols two or even three times in an experiment. If [fMg²⁺]_i rose significantly within a 5–15 min period (at > 0.1 mM min^–1), then it would rise at the same rate if the cell was allowed to recover in normal Tyrode solution and was then subjected to the same load protocol again (data not shown). Because of this reproducibility with individual cells, we compared the rates of [Mg²⁺]_i rise at two or three different values of [Na⁺]_o in the same cell and then normalized the rates to a specific [Na⁺]_o (with 30 mM Mg²⁺ in the medium, this was 95 mM; the maximum [Na⁺]_o achievable without increasing tonicity). In this fashion, we were able to examine the effects of [Na⁺]_o on [fMg²⁺]_i increase while minimizing the effects of cell variability. The protocol adopted was thus to expose each cell to at least two load protocols, one (or two) with a test [Na⁺]_o and one with 95 mM Na⁺, the order of exposure being randomized.

View larger version (15K): [in this window] [in a new window]   Figure 1.  Consecutive-loading protocol to assess the effects of external [Na+] on the rate of [fMg2+]i increase A myocyte containing mag-fura-2 was superfused at 37°C. The concentration of cytoplasmic ionized Mg2+ was calculated from the ratio of fluorescence intensities measured at 510 nm following excitation with light at wavelengths of 340 and 380 nm. Bars at the top of the figure represent solution changes and numbers indicate concentrations (mM). Dotted bars below the trace indicate periods when the superfusate was Ca2+ free. Where not specified, solution composition was as in normal Tyrode solution. Changes in [Na+] and [Mg2+] were balanced with NMDG to keep the osmotic pressure constant. The myocyte was first equilibrated in normal Tyrode solution and then Mg2+ loaded twice in succession by superfusion with Ca2+-free solutions containing 30 mM Mg2+. In the first instance, the solution contained 95 mM Na+ and in the second, it was Na+ free. Following Mg2+ loading, superfusion with normal Tyrode solution caused [fMg2+]i to return to the baseline.

View larger version (15K): [in this window] [in a new window]   Figure 2.  Rate of [fMg2+]i rise depends on external [Na+] Rates of [fMg2+]i rise were measured in experiments utilizing the consecutive-loading protocol (e.g. Fig. 1). Each myocyte was loaded once using a solution containing 30 mM Mg2+ and 95 mM Na+ (control), and once with 30 mM Mg2+ and 0, 10 or 45 mM Na+ (test), the order of exposure being randomized. Normalized rates of [fMg2+]i rise (test/control) were calculated for each experiment. Points represent the means ± S.E.M. of n experiments, where n is indicated on the figure. The line drawn by linear regression analysis of all normalized values (r2 = 0.63, n = 28) predicts that there should be no change in [fMg2+]i when [Na+] is 153 mM (extrapolated dotted line). Dashed lines indicate 95% confidence limits of regression. Mean rate of [fMg2+]i rise with 95 mM Na+ and 30 mM Mg2+ was 0.25 ± 0.03 mM min–1 (n = 14).

Using similar loading protocols, the effect of changing [Mg²⁺]_o on the rate of [fMg²⁺]_i rise was studied at a constant [Na⁺]_o of 95 mM. The rate of [fMg²⁺]_i rise was measured at 5, 15 and 30 mM [Mg²⁺]_o (Fig. 3). Consecutive-loading protocols were used and rates were normalized to those measured in the presence of 30 mM Mg²⁺ with the same cell. As [Mg²⁺]_o increased, the rate of [fMg²⁺]_i rise increased. In addition, there was considerably less delay between changing solutions and the first signs of a significant change in [fMg²⁺]_i taking place. The relationship between the normalized rate of [fMg²⁺]_i rise and [Mg²⁺]_o appears linear (Fig. 4), and extrapolation of this line to zero [Mg²⁺]_o predicts that there should be no change in [fMg²⁺]_i, as is the case.

View larger version (13K): [in this window] [in a new window]   Figure 3.  Consecutive-loading protocol to assess the effects of external [Mg2+] on the rate of [fMg2+]i increase A myocyte was twice loaded by removing Ca2+ from the superfusate and reducing [Na+] to 95 mM. During the first load, superfusate [Mg2+] was increased to 30 mM and during the second, to 5 mM. Explanation of bars and composition of solutions as in legend to Fig. 1.

View larger version (14K): [in this window] [in a new window]   Figure 4.  Rate of [Mg2+]i rise depends on external [Mg2+] Rates of [fMg2+]i rise were measured in consecutive-loading experiments (e.g. Fig. 3), each myocyte being loaded using solutions containing 95 mM Na+ and: control, 30 mM Mg2+; test, 5 or 15 mM Mg2+. The order of exposure was randomized. Normalized rates of [fMg2+]i rise (test/control) were calculated for each experiment. Points represent the means ± S.E.M. of n experiments, where n is indicated on the figure. The line drawn by linear regression analysis of all normalized values (r2 = 0.93, n = 18) predicts that there should be no change in [fMg2+]i when external [Mg2+] is –0.1 mM (extrapolated dotted line). Dashed lines indicate 95% confidence limits of regression.

Other workers have either failed to observe an increase in [fMg²⁺]_i (Tashiro & Konishi, 2000) when myocytes are bathed in a solution containing low [Na⁺] and high [Mg²⁺], or have seen only a very slow rise that was in the order of a few micromolar per minute (Tashiro et al. 2002; Tursun et al. 2005). Comparison of methods revealed that the major difference between protocols was temperature, with significant loading occurring in our experiments at 37°C, but little being observed at 25°C by others. We therefore examined the effects of temperature on [fMg²⁺]_i increase. Since temperature affects the Mg²⁺ sensitivity of mag-fura-2 and the pH of solutions, mag-fura-2 was also calibrated at 25°C, and the pH of solutions was adjusted accordingly. The approach adopted was to measure the rate of [fMg²⁺]_i rise at both 25 and 37°C in each myocyte under standard loading conditions (30 mM Mg²⁺, Na⁺ free). Mean rates were 19 ± 9 µM min^–1 at 25°C and 281 ± 49 µM min^–1 at 37°C (n = 11). Rates were much more variable at the lower temperature. In six out of 11 myocytes, a change in [fMg²⁺]_i was barely detectable over the load period at 25°C (< 3 µM min^–1), and in the other five myocytes, the maximum rate observed at this temperature was about 30 µM min^–1. In contrast, substantial increases in [fMg²⁺]_i were observed at 37°C. Thus the changes in [fMg²⁺]_i seen at 25°C are only about 7% of those seen at 37°C, and the Q₁₀ for the uptake process over this temperature range has an unusually high value that is > 9.

Effects of intracellular Na⁺ depletion

An obvious route by which Mg²⁺ might enter myocytes during the loading protocol is by reversal of the Na⁺–Mg²⁺ antiporter as a result of the high [Mg²⁺] and low [Na⁺] in the loading medium. However, since superfusion of cardiac cells with low-[Na⁺] solutions also causes their intracellular [Na⁺] ([Na⁺]_i) to fall rapidly (Ellis, 1977; Ellis & MacLeod, 1985), cells must contain sufficient Na⁺ to exchange for external Mg²⁺ if this mechanism is to be effective. Reduced [Na⁺]_i could limit or prevent uptake by the antiport. We therefore examined whether cells which had been Na⁺ depleted could still load with Mg²⁺ when subsequently superfused with a low-[Na⁺], high-[Mg²⁺] solution.

Myocytes were superfused with Na⁺-free Tyrode solution for at least 10 min, which should have removed almost all internal Na⁺ (Ellis & MacLeod, 1985). Then Ca²⁺ was also removed from the superfusate for 2 or 3 min before [Mg²⁺]_o was increased to 30 mM to mimic the normal loading conditions. Figure 5 is a typical example of one such experiment. Following 13 min of Na⁺-free superfusion, increasing [Mg²⁺]_o caused a large and rapid rise in [fMg²⁺]_i. This rise cannot be attributed to cell damage, since [fMg²⁺]_i fully recovered to its initial level when cells were superfused with normal Tyrode solution at the end of the experiment. In addition, cells were continuously monitored for morphological changes throughout experiments. They remained looking healthy, with no sign of cell rounding or membrane blebbing. In this example, [fMg²⁺]_i increased from 0.9 to 11 mM in just 4 min. In parallel studies, changes in cell [Na⁺] were monitored using myocytes loaded with the Na⁺-sensitive dye, SBFI. The fluorescence ratio (F₃₄₀/F₃₈₀) changed as anticipated, indicating a fall in cell [Na⁺] when myocytes were incubated in Na⁺-free media even when these contained Mg²⁺. It is clear that, rather than slowing the rate of [fMg²⁺]_i rise as would be expected if Na⁺–Mg²⁺ antiport played the dominant role in Mg²⁺ loading, depletion of internal Na⁺ may in fact accelerate the loading process. Therefore, reverse Na⁺–Mg²⁺ exchange cannot be the only route for Mg²⁺ entry across the sarcolemma; there must be other pathways capable of carrying large Mg²⁺ fluxes. These pathways were revealed when Na⁺ had been removed from the Ca²⁺-free medium in which the myocytes were incubated, a condition where myocyte [Na⁺] is also very low. How (and indeed whether) these transporters are activated under physiological conditions, where both Na⁺ and Ca²⁺ are present, has yet to be determined.

View larger version (11K): [in this window] [in a new window]   Figure 5.  [fMg2+]i can rise dramatically in Na+-depleted myocytes A myocyte was superfused with a Na+-free Tyrode solution (but containing 1 mM Ca2+) for 13 min to reduce its internal [Na+] to a very low level. Calcium was then removed from the solution for 3 min before [Mg2+] was increased to 30 mM. After 4 min in this high-[Mg2+] medium, the cell was superfused with normal Tyrode solution. Explanation of bars and composition of solutions as in legend to Fig. 1.

In this section, we examine the effects of two transport inhibitors on [fMg²⁺]_i increase: imipramine, an inhibitor of Na⁺–Mg²⁺ antiport, and KB-R7943, an inhibitor of Na⁺–Ca²⁺ antiport. Imipramine has been shown to inhibit Mg²⁺ efflux by Na⁺–Mg²⁺ antiport (forward mode of exchange) in erythrocytes (Féray & Garay, 1988; Flatman & Smith, 1990). It has also been shown to be equally effective at inhibiting Mg²⁺ uptake (reverse mode) by the transporter (Flatman & Smith, 1991) and at inhibiting the antiport in a wide range of other tissues (Wolf et al. 1994; Handy et al. 1996; Cefaratti et al. 2000; Touyz et al. 2001). KB-R7943 was developed as a potent, selective inhibitor of Na⁺–Ca²⁺ antiport (Iwamoto et al. 1996), though increasing use shows that it affects a number of other transporters and ion channels (Tanaka et al. 2002). It inhibits both forward and reverse modes of the Na⁺–Ca²⁺ antiport (Kimura et al. 1999).

In initial experiments, we explored the effects of 0.2 mM imipramine in a two-stage protocol (Fig. 6) where [fMg²⁺]_i increase was first measured in the absence of inhibitor and then in its presence, without an intervening recovery period. Myocytes were exposed to the standard load solution (30 mM Mg²⁺, Na⁺ and Ca²⁺ free) for long enough to establish the rate of [fMg²⁺]_i rise. They were then exposed to a ‘Mg²⁺-clamp’ solution containing only 1 mM Mg²⁺ (but still Na⁺ and Ca²⁺ free). The concentration of cytoplasmic ionized Mg²⁺ stops rising in this solution and remains clamped at the raised level (Handy et al. 1996). Imipramine was added to the clamp solution, giving it time to interact with binding sites. Myocytes were then superfused with the loading solution again, but this time containing imipramine, and the new rate of [fMg²⁺]_i rise was measured. At the end of each experiment, cell viability was assessed by superfusing the cell with normal Tyrode solution. In all experiments reported here, [fMg²⁺]_i returned to baseline. In these experiments, the rate of [fMg²⁺]_i rise was 0.46 ± 0.2 mM min^–1 in the absence of imipramine and 0.45 ± 0.2 mM min^–1 in its presence (n = 3). These rates are not significantly different (P = 0.22). In these experiments, [fMg²⁺]_i had risen to about 2 mM before the imipramine was added, and we speculate that an imipramine-insensitive Mg²⁺ entry route is active under these conditions.

View larger version (14K): [in this window] [in a new window]   Figure 6.  Imipramine fails to prevent [fMg2+]i rise in a Mg2+-loaded myocyte A myocyte was Mg2+ loaded by superfusing it with Ca2+- and Na+-free Tyrode solution containing 30 mM Mg2+. When [fMg2+]i approached 2 mM, the superfusate was switched to a ‘Mg2+-clamp’ solution that contained 1 mM Mg2+ (Ca2+ and Na+ free) to prevented further rise in [fMg2+]i. After 5 min, superfusate [Mg2+] was increased to 30 mM for 5 min and then the myocyte was superfused with normal Tyrode solution. Imipramine (0.2 mM) was added to superfusates from the start of the Mg2+-clamp period until the end of the second loading phase. Rates of [fMg2+]i rise during the two load periods were measured and compared. Explanation of bars and composition of solutions as in legend to Fig. 1.

View larger version (18K): [in this window] [in a new window]   Figure 7.  Effects of KB-R7943 on [fMg2+]i rise Myocytes were superfused with Ca2+- and Na+-free loading solution containing 30 mM Mg2+. With some cells (A), 20 µM KB-R7943 (KBR) was added to the superfusate when [fMg2+]i was well above 2 mM. With others (B), 20 µM KB-R7943 was added to the superfusate once it was clear that [fMg2+]i was increasing, but well before it reached 2 mM. After sufficient time to establish the rate of [fMg2+]i change, the cells were superfused with normal Tyrode solution. In both cases, [fMg2+]i returned to the initial preload level. Explanation of bars and composition of solutions as in legend to Fig. 1.

We also assessed the reversibility of the effects of imipramine and KB-R7943. In these experiments, cells were exposed to the inhibitors (in loading solution) for about 10–15 min. This was followed by a period of inhibitor washout in normal Tyrode solution of at least 15min. The cells were then subjected to the normal Mg²⁺ loading procedure (30 mM Mg²⁺, Na⁺ and Ca²⁺ free). Where cells had been previously exposed to 0.2 mM imipramine, the rate of [fMg²⁺]_i rise in standard loading solution (without imipramine) was 0.090 ± 0.011 mM min^–1 (n = 12). This represents a reduction of 65% (P = 0.005) compared to the rate seen in control cells that had not been exposed to the inhibitor. In cells previously exposed to 20 µM KB-R7943, the rate of [fMg²⁺]_i rise in loading solution without inhibitor was 0.157 ± 0.019 mM min^–1 (n = 10). This is a reduction of 38% compared to control cells, but the difference between means was not quite significant (P = 0.051). All cells used in these tests produced a rise in [fMg²⁺]_i of >0.1 mM min^–1 in the absence of an inhibitor. The effects of imipramine and KB-R7943 are thus only partly reversed by superfusing cells with normal Tyrode solution.

Effects of membrane potential depolarization

The way in which ion movement across a membrane is affected by membrane potential reveals important thermodynamic and kinetic properties of the transport pathways involved. Two methods were used to examine the effects of membrane depolarization on [fMg²⁺]_i changes produced by Mg²⁺ loading solutions. Myocytes were either superfused with high-[K⁺]_o solutions or were voltage clamped with microelectrodes.

Figure 8 shows a typical example of what happened when superfusate [K⁺] was increased from the control value of 6 mM to 70 mM (replacing NMDG) during the loading protocol. In this experiment, superfusate [K⁺] was increased at the same time that [Mg²⁺]_o was increased and [Na⁺]_o removed. The concentration of cytoplasmic ionized Mg²⁺ rose very slowly over the next 12 min, but then rose rapidly when [K⁺]_o was reduced to 6 mM. The rate of [fMg²⁺]_i increase subsequently fell dramatically when [K⁺] was increased again. The order in which cells were exposed to high- or control-[K⁺] solutions was randomized, and was found to have no effect on outcome. In each experiment, the rate of rise in high-[K⁺] solution was compared to that with control [K⁺]. The rate of [fMg²⁺]_i rise in the presence of 70 mM K⁺ was 0.13 ± 0.02 mM min^–1, significantly lower than the 0.53 ± 0.06 mM min^–1 measured in the standard loading solution (P < 0.01, n = 8). An important aspect of these experiments was that the effects of depolarization could be examined in cells loaded to low- or high-Mg²⁺ levels ([fMg²⁺]_i < 2 mM or > 2 mM, respectively). Depolarization by high-[K⁺] solutions slowed Mg²⁺ uptake at all Mg²⁺ levels.

View larger version (12K): [in this window] [in a new window]   Figure 8.  High external [K+] slows the rate of [fMg2+]i rise After equilibration in normal Tyrode solution and superfusion with Ca2+-free Tyrode solution for 3 min, a myocyte was superfused with a high-[K+] loading solution containing (mM): Mg2+, 30; Na+, 70; and K+, 70 for 10 min. Superfusate [K+] was then reduced to 6 mM (replaced by NMDG) for 5 min, increased to 70 mM for 4 min, and then reduced to 6 mM for the rest of the experiment. Superfusate [Mg2+] was reduced to 1 mM at 27 min to stabilize [fMg2+]i before superfusion with normal Tyrode solution started at 32 min. Rates of [fMg2+]i rise were measured at both values of [K+]. Explanation of bars and composition of solutions as in legend to Fig. 1.

View larger version (16K): [in this window] [in a new window]   Figure 9.  Membrane depolarization slows [fMg2+]i rise A myocyte containing mag-fura-2 was patch clamped in whole-cell configuration. Its membrane potential (V) was maintained at –80 mV while it was superfused with normal Tyrode solution and then with Ca2+-free Tyrode solution. On changing to a Na+- and Ca2+-free Tyrode solution containing 30 mM Mg2+, and with the potential still clamped at –80 mV, [fMg2+]i started to rise rapidly. When [fMg2+]i reached about 2 mM, the potential was clamped to 0 mV for 2 min, after which it was returned to –80 mV for the rest of the experiment. Patch pipette resistance was 3 M in normal Tyrode solution when filled with internal solution, and series resistance was < 10 M at the beginning of the experiment. The upper panel shows membrane potential and the middle panel clamp current (Im). Explanation of bars and composition of solutions as in legend to Fig. 1.

	Discussion

Top Abstract Introduction Methods Results Discussion References

Data shown in Fig. 2 emphasize the difficulty in changing myocyte [fMg²⁺]_i by simply increasing [Mg²⁺]_o. Similarly, changing [Na⁺]_o alone has little or no effect on [fMg²⁺]_i (Fig. 4). However, superfusion of myocytes with solutions that have both a low [Na⁺] and high [Mg²⁺] causes substantial increase in [fMg²⁺]_i at 37°C. After an initial slow phase, the rate of [fMg²⁺]_i rise becomes steady for many minutes (> 10 min) or continues to increase with time (Figs 1 and 3). There are several plausible explanations for this behaviour. For instance, it would occur if intracellular [Mg²⁺] were heavily buffered by ligands that were close to saturation. Initially, much of the total Mg²⁺ that enters would be buffered and only a portion would be free and ionized, and therefore detected by mag-fura-2. However, as Mg²⁺ binding sites become increasingly saturated, a larger fraction of the Mg²⁺ influx would remain free, causing the rate of [fMg²⁺]_i rise to increase, despite a constant influx rate. Alternatively, Mg²⁺ transporters may become activated as [fMg²⁺]_i rises, causing a rise in influx rate. Unfortunately, distinguishing between these alternatives, or assessing the contribution each makes to the observed behaviour, requires a better understanding of Mg²⁺ buffering in myocytes than we currently have. However, as discussed later, there is independent evidence that some Mg²⁺ transporters may be activated as [fMg²⁺]_i rises.

It is now well established that Na⁺–Mg²⁺ antiport is responsible for keeping myocyte [fMg²⁺]_i at the physiological level by balancing inward leaks of Mg²⁺ (Handy et al. 1996; Tashiro & Konishi, 2000; Almulla et al. 2006), so an obvious explanation for the rise in [fMg²⁺]_i is that the antiport reverses direction to provide a route for Mg²⁺ influx. The antiport has been shown to be reversible and to carry Mg²⁺ into cells (Flatman & Smith, 1991; Günther & Vormann, 1995; Schweigel et al. 2000), though its capacity may be limited (Tashiro et al. 2005). Like the forward mode, the reverse mode is inhibited by imipramine (Flatman & Smith, 1991; Schweigel et al. 2000). Assuming 1:1 stoichiometry, the antiport is close to equilibrium under normal physiological conditions (Handy et al. 1996; Almulla et al. 2006), so any reduction of [Na⁺]_o or increase in [Mg²⁺]_o should cause it to transport Mg²⁺ into the cell. As shown above, [fMg²⁺]_i does increase when [Na⁺]_o is reduced, but only if [Mg²⁺]_o is increased as well, and the relationship between the rate of [fMg²⁺]_i rise and [Na⁺]_o is linear (Fig. 2). Also, [fMg²⁺]_i rises if [Mg²⁺]_o is increased, but only if [Na⁺]_o is also reduced, and again the relationship between the rate of [fMg²⁺]_i rise and [Mg²⁺]_o appears linear (Fig. 4). Since increases in [fMg²⁺]_i from the physiological level are much reduced by imipramine, a compound that inhibits Na⁺–Mg²⁺ antiport, it seems plausible that the transporter involved is reverse Na⁺–Mg²⁺ antiport. However, an important question remains: if the antiport is at equilibrium under physiological conditions, why is it necessary to change both [Mg²⁺]_o and [Na⁺]_o to elicit a rise in [fMg²⁺]_i? Part of the answer may lie in the way [Na⁺] is regulated in cardiac myocytes. When [Na⁺]_o is reduced with the intention of reducing the transmembrane Na⁺ gradient (source of energy for Mg²⁺ transport), the myocyte [Na⁺] falls rapidly to a new level within a few minutes. This level is linearly related to [Na⁺]_o over a wide range (Ellis, 1977; Ellis & MacLeod, 1985; Ödblom & Handy, 2001). Thus the change in Na⁺ gradient is much less than expected, but has been characterized. With 1 mM Mg²⁺ in the medium, the predicted maximum changes in [fMg²⁺]_i on reducing [Na⁺]_o are small, so changes within the time frame of our experiments would be undetectable unless amplified by Na⁺ loading the cells (Tashiro et al. 2005). Failure of increased [Mg²⁺]_o alone to cause a rise in [fMg²⁺]_i may also be due to Na⁺, which competes with Mg²⁺ for occupancy of external sites on the antiport (Flatman & Smith, 1990; Flatman & Smith, 1991; Vormann & Günther, 1993). This competition, rather than changes in the Na⁺ gradient, may explain the linear relationship between the rate of [fMg²⁺]_i rise and [Na⁺]_o (Fig. 2). Thus, it may only be possible to generate measurable increases in [fMg²⁺]_i within the experimental time frame, and without significant changes to tonicity, by simultaneously increasing [Mg²⁺]_o (Fig. 4) and decreasing [Na⁺]_o (Fig. 2).

The stoichiometry of Na⁺–Mg²⁺ antiport has been the cause of much debate, with suggested values ranging from 1 Na⁺:1 Mg²⁺ (Flatman & Smith, 1990; Günzel & Schlue, 1996) to 3 Na⁺:1 Mg²⁺ (Féray & Garay, 1988; Tashiro & Konishi, 1997). Analysis of the experiments reported here can also yield a stoichiometry for the antiporter, if it is the mechanism responsible for the rise in [fMg²⁺]_i. Table 1 shows the direction of expected net fluxes and the effects of membrane depolarization on these fluxes, if Mg²⁺ uptake is through a channel or through antiports with stoichiometries of 1, 2 or 3 Na⁺ per Mg²⁺ ion. All the mechanisms considered could provide a route for net Mg²⁺ uptake when myocytes are superfused with a Na⁺-free medium containing 30 mM Mg²⁺ (Table 1). Shortly after this superfusion begins, however, cell [Na⁺] will fall to low levels (see above), altering the driving force on Mg²⁺ movement if this is on an antiport. This will greatly limit Mg²⁺ uptake by systems with 3:1 stoichiometries unless the membrane potential also depolarizes substantially. Particularly important to the issue of stoichiometry is our finding that large rises in [fMg²⁺]_i occur when myocytes are superfused with solutions containing 95 mM [Na⁺] and 30, 15 or even 5 mM [Mg²⁺] (Figs 3 and 4). Thermodynamic analysis (Table 1) suggests that such increases can only occur if Mg²⁺ enters through a channel or an antiport with 1:1 stoichiometry. It predicts that Mg²⁺ should exit if an antiport transports 2 or 3 Na⁺ per Mg²⁺ ion. Indeed, antiports with these stoichiometries can only give rise to Mg²⁺ influx if the Na⁺ gradient is greatly reduced, and this would only have occurred in our experiments with Na⁺-free media (with the exception that very limited uptake by a 2:1 antiport may be possible when [Na⁺] is 10 and [Mg²⁺] 30 mM). Under physiological conditions (Table 1), activation of these systems would cause net uptake through a channel but net loss through any antiport that has a stoichiometry of 2 or more Na⁺ per Mg²⁺ ion. A 1 Na⁺:1 Mg²⁺ antiport would be at equilibrium, and its activation would not affect [fMg²⁺]_i.

Our experiments strongly suggest that substantial amounts of Mg²⁺ can enter myocytes by a route in addition to reversed Na⁺–Mg²⁺ or Na⁺–Ca²⁺ antiport. For instance, large [fMg²⁺]_i rises occur in myocytes that have been Na⁺ depleted by incubation in Na⁺-free media prior to loading (Fig. 5). Myocyte [Na⁺] falls to such low levels under these conditions (Ellis & MacLeod, 1985; Ödblom & Handy, 2001) that it could not sustain the Mg²⁺ influx necessary to raise [fMg²⁺]_i to the levels observed by reverse antiport whatever the stoichiometry. In addition, if [fMg²⁺]_i is allowed to rise above 2 mM before imipramine or KB-R7943 is applied, then neither affects the rate of Mg²⁺ uptake (Figs 6 and 7A). Taken together, these data suggest that an imipramine- and KB-R7943-insensitive pathway is activated under our loading conditions. Similar behaviour has been seen in red blood cells (Flatman & Smith, 1996), where activation of a Mg²⁺-permeable route was shown to require both a doubling of [fMg²⁺]_i and a fall in cell [Na⁺]. Although the effects of imipramine and KB-R7943 were not tested in the red blood cell work, the route was found to be insensitive to 1 mM amiloride, a compound that, like imipramine, inhibits Na⁺–Mg²⁺ antiport (Flatman & Smith, 1990) and, like KB-R7943, inhibits Na⁺–Ca²⁺ antiport (Kaczorowski et al. 1985).

Magnesium uptake in myocytes is very sensitive to temperature. The rate of [fMg²⁺]_i rise increases about 15-fold when the temperature is raised from 25 to 37°C. Given that the antiports have a Q₁₀ of about 2 (e.g. 2.3 for the Na⁺–Mg²⁺ antiport; Almulla et al. 2006), this suggests that an alternative route exists with a Q₁₀ > 9. This high temperature sensitivity probably explains the small [fMg²⁺]_i increases reported by some laboratories that have attempted to Mg²⁺ load myocytes at room temperature using the methods described here (Tashiro & Konishi, 2000; Tashiro et al. 2002; Tursun et al. 2005).

Membrane depolarization to about –20 mV by superfusion of myocytes with solutions containing 70 mM K⁺ (Chapman, 1973), or to 0 mV by microelectrode voltage clamp, substantially reduced, and in some cases stopped, the rise of [fMg²⁺]_i. The reduction is in line with predictions for Mg²⁺ uptake by 1 Na⁺–1 Mg²⁺ antiport or by a channel, and excludes mechanisms that involve exchange of Mg²⁺ for 2 or more Na⁺ ions. In principle, it should be possible to distinguish between channel and antiport on the basis of the reversal potential of Mg²⁺ movement. Unfortunately, this requires more precise knowledge of the Na⁺ gradient than we currently have. However, the cessation of Mg²⁺ movement when the membrane potential was depolarized to 0 mV (Fig. 9) cannot simply be explained by the reduction in electrical driving force alone. Under the conditions of the experiment shown in Fig. 9, it would have been necessary to depolarize the membrane potential to more than +30 mV to prevent Mg²⁺ entry through a channel, and to an even more positive potential to prevent uptake on the antiport. The most likely explanation is that depolarization closes a Mg²⁺-permeant channel.

In conclusion, we show that in the absence of Ca²⁺, the simultaneous reduction of external [Na⁺] and increase in external [Mg²⁺] causes myocytes to take up large amounts of Mg²⁺. The initial influx may involve several transporters, including reversal of the Na⁺–Mg²⁺ antiport and transport of Mg²⁺ on reversed Na⁺–Ca²⁺ antiport. The consequent rise of [fMg²⁺]_i, in conjunction with reduced [Na⁺], may then activate a Mg²⁺ transport route that is insensitive to imipramine or KB-R7943, but is inactivated by depolarization. Magnesium transport through this route is highly sensitive to temperature and stops abruptly when external [Na⁺] is returned to the physiological level. Exciting new work, often involving heterologous expression of genes, suggests that there are a number of routes by which Mg²⁺ may enter eukaryotic cells. Where selectivity over Ca²⁺ is important, the unique chemical properties of Mg²⁺ must be exploited, and these transporters often belong to families that lack homology with other transporter families, or are very unusual members of their family (Maguire, 2006). These include members of the SLC41, ACDP and MagT families (Wabakken et al. 2003; Goytain & Quamme, 2005a). In contrast, where Mg²⁺ versus Ca²⁺ selectivity is not an issue, transporters with a divalent cation binding site may suffice, as exemplified by the TRPs, TRPM6 and TRPM7 (Schmitz et al. 2003; Voets et al. 2004; Chubanov et al. 2005; He et al. 2005). Expression of SLC41A1, SLC41A2, ACDP2 and MagT1 in Xenopus oocytes all stimulate Mg²⁺ uptake by mechanisms with channel-like properties (Goytain & Quamme, 2005a,b,c,d). They show inward rectification and are all expressed in the heart, but have different Mg²⁺ selectivities, with MagT1 being highly Mg²⁺ selective, whereas SLC41 and ADCP2 transport a variety of other divalent cations too (Ca²⁺ is a poor substrate). However, in no case was Mg²⁺ uptake affected by removal of Na⁺ from the external medium as might be expected for the transporter we describe (though the effect of internal Na⁺ was not tested). Expression of TRPM6 or TRPM7 in HEK-293 cells also causes the appearance of Mg²⁺ (and Ca²⁺)-permeable channels in the plasma membrane. These allow entry of Mg²⁺ at negative potentials but exit of Na⁺ at positive potentials (Nadler et al. 2001; Voets et al. 2004). This competition between Mg²⁺ and Na⁺ for movement through the channels fits the profile of the cardiac channel; however, Mg²⁺ movement through the TRP channels is voltage independent, and internal Mg²⁺ inhibits rather than stimulates uptake as described here. The finding that coexpression of TRPM6 and TRPM7 produces hetero-multimeric channels (Chubanov et al. 2004) further complicates matters because this implies that channel properties may depend on the precise mix of subunits. Clearly, the properties of these transporters have not yet been sufficiently elucidated to allow them to be identified as constituents of the new influx pathway in the heart. Our data, however, indicate key parameters and properties that need to be established in expression studies to facilitate this task, and in turn these studies should suggest new strategies for experimentation at the functional cellular level.

	References

Top Abstract Introduction Methods Results Discussion References

 
Almulla HA, Bush PG, Steele MG, Flatman PW & Ellis D (2006). Sodium-dependent recovery of ionised magnesium concentration following magnesium load in rat heart myocytes. Pflugers Arch 451, 657–667.[CrossRef][Medline]

Almulla HA, Ellis D & Flatman PW (2003). Effects of imipramine and KB-R7943 on Na⁺ dependent Mg²⁺ transport in isolated rat ventricular myocytes. J Physiol 547.P, C19.

Altura BM, Gebrewold A, Altura BT & Gupta RK (1995). Role of brain [Mg²⁺]_i in alcohol-induced hemorrhagic stroke in a rat model: a ³¹P-NMR in vivo study. Alcohol 12, 131–136.[CrossRef][Medline]

Buri A, Chen S, Fry CH, Illner H, Kickenweiz E, McGuigan JAS, Noble D, Powell T & Twist VW (1993). The regulation of intracellular Mg²⁺ in guinea-pig heart, studied with Mg²⁺-selective microelectrodes and fluorochromes. Exp Physiol 78, 221–233.[Abstract]

Cefaratti C, Romani A & Scarpa A (2000). Differential localization and operation of distinct Mg²⁺ transporters in apical and basolateral sides of rat liver plasma membrane. J Biol Chem 275, 3772–3780.[Abstract/Free Full Text]

Chapman RA (1973). The ionic dependence of the strength and spontaneous relaxation of the potassium contracture induced in the heart of the frog Rana pipiens. J Physiol 231, 209–232.[Abstract/Free Full Text]

Chubanov V, Gudermann T & Schlingmann KP (2005). Essential role for TRPM6 in epithelial magnesium transport and body magnesium homeostasis. Pflugers Arch 451, 228–234.[CrossRef][Medline]

Chubanov V, Waldegger S, Schnitzler MM, Vitzthum H, Sassen MC, Seyberth HW, Konrad M & Gudermann T (2004). Disruption of TRPM6/TRPM7 complex formation by a mutation in the TRPM6 gene causes hypomagnesemia with secondary hypocalcemia. Proc Natl Acad Sci U S A 101, 2894–2899.[Abstract/Free Full Text]

Cowan JA (2002). Structural and catalytic chemistry of magnesium-dependent enzymes. Biometals 15, 225–235.[CrossRef][Medline]

Ellis D (1977). The effects of external cations and ouabain on the intracellular sodium activity of sheep heart Purkinje fibres. J Physiol 273, 211–240.[Abstract/Free Full Text]

Ellis D & MacLeod KT (1985). Sodium-dependent control of intracellular pH in Purkinje fibres of sheep heart. J Physiol 359, 81–105.[Abstract/Free Full Text]

Féray JC & Garay R (1988). Demonstration of a Na⁺: Mg²⁺ exchange in human red cells by its sensitivity to tricyclic antidepressant drugs. Naunyn-Schmiedeberg's Arch Pharmacol 338, 332–337.[Medline]

Flatman PW & Smith LM (1990). Magnesium transport in ferret red cells. J Physiol 431, 11–25.[Abstract/Free Full Text]

Flatman PW & Smith LM (1991). Sodium-dependent magnesium uptake by ferret red cells. J Physiol 443, 217–230.[Abstract/Free Full Text]

Flatman PW & Smith LM (1996). Magnesium transport in magnesium-loaded ferret red blood cells. Pflugers Arch 432, 995–1002.[CrossRef][Medline]

Fry CH (1986). Measurement and control of intracellular magnesium ion concentration in guinea pig and ferret ventricular myocardium. Magnesium 5, 306–316.[Medline]

Fry CH & Proctor AV (1993). The effects of magnesium on excitable tissues. In Magnesium and the Cell, ed. Birch NJ, pp. 217–234. Academic Press, London.

Goytain A & Quamme GA (2005a). Functional characterization of human SLC41A1, a Mg²⁺ transporter with similarity to prokaryotic MgtE Mg²⁺ transporters. Physiol Genomics 21, 337–342.[Abstract/Free Full Text]

Goytain A & Quamme GA (2005b). Functional characterization of ACDP2 (ancient conserved domain protein), a divalent metal transporter. Physiol Genomics 22, 382–389.[Abstract/Free Full Text]

Goytain A & Quamme GA (2005c). Functional characterization of the human solute carrier, SLC41A2. Biochem Biophys Res Commun 330, 701–705.[CrossRef][Medline]

Goytain A & Quamme G (2005d). Identification and characterization of a novel mammalian Mg²⁺ transporter with channel-like properties. BMC Genomics 6, 48.[CrossRef][Medline]

Günther T & Vormann J (1995). Reversibility of Na⁺/Mg²⁺ antiport in rat erythrocytes. Biochim Biophys Acta 1234, 105–110.[Medline]

Günzel D & Schlue WR (1996). Sodium-magnesium antiport in Retzius neurones of the leech Hirudo medicinalis. J Physiol 491, 595–608.[Abstract/Free Full Text]

Handy RD, Gow IF, Ellis D & Flatman PW (1996). Na-dependent regulation of intracellular free magnesium concentration in isolated rat ventricular myocytes. J Mol Cell Cardiol 28, 1641–1651.[CrossRef][Medline]

He Y, Yao G, Savoia C & Touyz RM (2005). Transient receptor potential melastatin 7 ion channels regulate magnesium homeostasis in vascular smooth muscle cells. Circ Res 96, 207–215.[Abstract/Free Full Text]

Hongo K, Konishi M & Kurihara S (1994). Cytoplasmic free Mg²⁺ in rat ventricular myocytes studied with the fluorescent indicator furaptra. Jpn J Physiol 44, 357–378.[CrossRef][Medline]

Iwamoto T, Watano T & Shigekawa M (1996). A novel isothiourea derivative selectively inhibits the reverse mode of Na⁺/Ca²⁺ exchange in cells expressing NCX1. J Biol Chem 271, 22391–22397.[Abstract/Free Full Text]

Kaczorowski GJ, Barros F, Dethmers JK, Trumble MJ & Cragoe EJ (1985). Inhibition of Na⁺/Ca²⁺ exchange in pituitary plasma membrane vesicles by analogues of amiloride. Biochemistry 24, 1394–1403.[CrossRef][Medline]

Kimura J, Watano T, Kawahara M, Sakai E & Yatabe J (1999). Direction-independent block of bi-directional Na⁺/Ca²⁺ exchange current by KB-R7943 in guinea-pig cardiac myocytes. B J Pharmacol 128, 969–974.[CrossRef][Medline]

McGuigan JAS, Elder HY, Günzel D & Schlue WR (2002). Magnesium homeostasis in heart: a critical reappraisal. J Clin Basic Cardiol 5, 5–22.

Maguire ME (2006). Magnesium transporters: properties, regulation and structure. Front Biosci 11, 3149–3163.[Medline]

Murphy E, Steenbergen C, Levy LA, Raju B & London RE (1989). Cytosolic free magnesium levels in ischemic rat heart. J Biol Chem 264, 5622–5627.[Abstract/Free Full Text]

Nadler MJS, Hermosura MC, Inabe K, Perraud A-L, Zhu Q, Stokes AJ, Kurosaki T, Kinet JP, Penner R, Scharenberg AM & Fleig A (2001). LTRPC7 is a Mg·ATP-regulated divalent cation channel required for cell viability. Nature 411, 590–595.

Ödblom MP & Handy RD (2001). Effect of external magnesium on intracellular free sodium: Na⁺ flux via Na⁺/Mg²⁺ antiport is masked by other Na⁺ transport systems in rat cardiac myocytes. Magnes Res 14, 3–9.[Medline]

Page E & Polimeni PI (1972). Magnesium exchange in rat ventricle. J Physiol 224, 121–139.[Abstract/Free Full Text]

Rubin H (2005). Magnesium: the missing element in molecular views of cell proliferation control. Bioessays 27, 311–320.[CrossRef][Medline]

Schmitz C, Perraud A-L, Johnson CO, Inabe K, Smith MK, Penner R, Kurosaki T, Fleig A & Scharenberg AM (2003). Regulation of vertebrate cellular Mg²⁺ homeostasis by TRPM7. Cell 114, 191–200.[CrossRef][Medline]

Schoenmakers TJM, Visser GJ, Flik G & Theuvenet APR (1992). Chelator – an improved method for computing metal-ion concentrations in physiological solutions. Biotechniques 12, 870–879.[Medline]

Schweigel M, Vormann J & Martens H (2000). Mechanisms of Mg²⁺ transport in cultured ruminal epithelial cells. Am J Physiol 278, G400–G408.

Steele MG, Flatman PW & Ellis D (2003). Temperature dependence of Mg²⁺ transport in isolated ventricular myocytes. J Physiol 547.P, C18.