INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS. DISCUSSION REFERENCES

CYTOPLASMIC FREE CONCENTRATION of Mg²⁺ ([Mg²⁺]_i) in mammalian cardiac myocytes has been estimated with various methods and appears to be in the range of 0.5-1.0 mM, well below electrochemical equilibrium across the cell membrane (7, 19, 27, 31). Since it has been reported that submillimolar [Mg²⁺]_i significantly influences many intracellular processes of cardiac muscles, including adenylate cyclase activity (2), K⁺ channels (see Refs. 1 and 25 for reviews), excitation-contraction coupling (36), Ca²⁺ sensitivity of myofilaments (8), and Ca²⁺ binding to intracellular sites (10, 20), [Mg²⁺]_i must be tightly regulated by active extrusion from the cell to balance any passive leak influx of Mg²⁺. Na⁺/Mg²⁺ exchange may play an essential role as an active Mg²⁺ extrusion pathway in many types of cells (for review see Refs. 12 and 30), but experimental evidence so far obtained in cardiac muscle is very controversial. The concept of the existence of Na⁺/Mg²⁺ exchange has been supported by measurements of the net Mg²⁺ fluxes by atomic absorption spectroscopy (29) and also by measurements of [Mg²⁺]_i either by ion-selective microelectrodes (14) or a fluorescent indicator (17). However, other studies with improved ion-selective microelectrodes (6, 7) and fluorescent indicators (6, 18, 26, 31) failed to provide any evidence for a Na⁺ gradient-dependent Mg²⁺ efflux in cardiac muscle.

This study describes [Mg²⁺]_i measurements with a fluorescent indicator carried out to seek evidence for Na⁺/Mg²⁺ exchange in cardiac myocytes. The results show that Na⁺-dependent changes in [Mg²⁺]_i are unmasked after facilitation of passive Mg²⁺ influx by an ionophore and strongly suggest the existence of a Na⁺ gradient-dependent Mg²⁺ efflux.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS. DISCUSSION REFERENCES

General. All experiments were carried out on single ventricular myocytes enzymatically isolated from male Wistar rats (250-300 g), as previously described (18, 20). After enzymatic digestion with 0.2 mg/ml collagenase (collagenase S-1; Nitta Zerachin, Tokyo, Japan) and 0.04 mg/ml protease (type XIV; Sigma, St. Louis, MO) in the presence of 0.6 mg/ml BSA (Sigma), cells were stored in 0.2 mM CaCl₂-containing Tyrode solution at 6°C until used. Cells were placed in an experimental chamber on the stage of an inverted microscope (Diaphot; Nikon, Tokyo, Japan) and superfused with normal Tyrode solution containing (in mM) 138 NaCl, 5.9 KCl, 2.4 CaCl₂, 1.2 MgCl₂, 11.8 glucose, and 5 HEPES (pH 7.4). Only quiescent rod-shaped cells that gave an all-or-none response to a 5-ms field stimulation were used for the experiments. After the indicator was loaded by incubation of the cells with 4 µM furaptra-AM for 10 min at room temperature and washout of the AM ester for at least 10 min, fluorescence measurements were carried out under continuous flow of the perfusate at 25°C.

Calibration of furaptra signals. The ratio of furaptra fluorescence intensities measured with excitation at 382 and 350 nm [R = F(382)/F(350)] was used as a Mg²⁺-related signal. Slow drift of the optical instruments (e.g., aging of the lamp) was corrected by occasional measurement of R in a Ca²⁺-Mg²⁺-free buffer solution (in mM: 140 KCl, 10 NaCl, 1 EDTA, 1 EGTA, 0.025 furaptra, and 10 PIPES, pH 7.1) as a standard. All values of the measured R were normalized to the standard R value taken with identical optics and were converted to [Mg²⁺]_i with the standard equation
where R_min and R_max are the R values at zero [Mg²⁺]_i and saturating [Mg²⁺]_i, respectively, and K_D is the dissociation constant. We used parameter values previously estimated in smooth muscle cells of tenia at 25°C: R_min = 0.986, R_max= 0.199 and K_D = 5.43 mM (34). The calibration procedure also included a correction for the small offset of R (+0.035) found in cardiac myocytes in a previous study (34). This offset was observed shortly after application of an ionophore cocktail containing Br-A-23187 and other ionophores, and was attributed to an artifact unrelated to [Mg²⁺]_i. Direct effects of imipramine (200 µM) and ionomycin (10 µM) on the furaptra R were tested at 0-4 mM [Mg²⁺] in vitro and were found to be negligible. It has been reported that equimolar substitution of Na⁺ by N-methyl-D-glucamine (NMDG⁺) up to 40 mM has little influence on the furaptra R in solutions containing 0-4 mM [Mg²⁺] (35).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Long-term observation of cell autofluorescence, cell F(), excited at 350 nm (A) and 382 nm (B). Intensities of autofluorescence relative to values measured just before ionomycin application (the first data points in A and B) were plotted as a function of time during application and washout of ionomycin. Ionomycin (10 µM) was added to the high-Mg2+ solution (10 mM) containing either 130 mM Na+ (5 cells) or 130 mM N-methyl-D-glucamine (NMDG+; 5 cells) for the periods indicated by the horizontal bars. Cells were then perfused with Ca2+-free Tyrode solution containing 130 mM Na+. Values of F(350) were 0.551±0.042 (Na+) and 0.544±0.052 (NMDG+) at time = 30 min, and were 0.645±0.047 (Na+) and 0.598±0.032 (NMDG+) at time = 50 min. Values of F(382) were 0.721±0.057 (Na+) and 0.763±0.073 (NMDG+) at time = 30 min, and were 0.780±0.090 (Na+) and 0.773±0.052 (NMDG+) at time = 50 min. Because the ionomycin effect was not significantly different in the presence and in the absence of extracellular Na+, the data from the 10 cells were treated as a single data set; each symbol represents a mean (±SE) value measured from the total of 10 cells. Solid lines are the least-squares fit of the mean values by an exponential (plus constant) function of the form F() = c + a × exp(t/), where t is time after the first data point in the fitted range, and  is a time constant. Constants c and a give, respectively, an offset and a scaling factor. At each wavelength (), four data points between 0 and 30 min () were fitted for the decay during the ionomycin treatment, and three data points between 30 and 50 min (the last  and two ) were fitted for the recovery after washout.

Solutions and chemicals. Ca²⁺-free Tyrode solution contained 0.1 mM EGTA replacing 2.4 mM CaCl₂ of normal Tyrode solution. High-Mg²⁺ Tyrode solution contained 10 mM MgCl₂, 0 mM CaCl₂, and 0.1 mM EGTA, with NaCl concentration reduced to 128 mM. In one experiment, Mg²⁺ concentration was raised to 30 mM by simple addition of 20 mM MgCl₂ to the high-Mg²⁺ Tyrode solution without any osmotic compensation. High-K⁺ (55.9 mM) solutions with various Na⁺ concentrations ([Na⁺]; 0-90 mM) were made by equimolar substitution of NaCl with potassium methanesulfonate and sodium methanesulfonate to keep the [K⁺]×[Cl] product constant. For Na⁺-free conditions, Na⁺ was substituted by equimolar NMDG⁺. Furaptra (tetrapotassium salt of mag-fura 2) and furaptra-AM (mag-fura 2-AM) were purchased from Molecular Probes (Eugene, OR). EGTA was obtained from Sigma Chemical. Ionomycin (Sigma Chemical) was dissolved from a 10 mM stock solution in DMSO (DOTITE Spectrosol; Dojindo, Kumamoto, Japan). Imipramine · HCl (Nacalai Tesque, Kyoto, Japan) was directly dissolved in the perfusates. All other chemicals were reagent grade.

Curve fitting and statistical analysis. Nonlinear least-squares fitting was carried out with the program Origin (version 5.0J; Microcal Software, Northampton, MA) that uses the Levenberg-Marquardt algorithm. Statistical values were given as means ± SE. The two-tailed Student's t-test was used for statistical comparison with the significance level set at P < 0.05.

	RESULTS.

TOP ABSTRACT INTRODUCTION METHODS RESULTS. DISCUSSION REFERENCES

All [Mg²⁺]_i measurements were carried out in Ca²⁺-free conditions (0.1 mM EGTA) to minimize any complications caused by changes in cytoplasmic [Ca²⁺] ([Ca²⁺]_i; see DISCUSSION) and also to avoid Ca²⁺ overloading of the cells. Values of basal [Mg²⁺]_i thus measured in the Ca²⁺-free Tyrode solution showed a Gaussian distribution (not shown) with a mean value of 0.71 mM (±0.01 mM, n = 128).

Effect of extracellular Na+ on [Mg²⁺]_i in intact myocytes. Figure 2 shows the results of our initial experiments, in which [Mg²⁺]_i was continuously measured from the same myocytes. Removal of extracellular Na⁺, with extracellular [Mg²⁺] ([Mg²⁺]_o) increased to 10 mM, did not cause any significant change in [Mg²⁺]_i during a period of 30 min (Fig. 2A). Following Na⁺-free perfusion, cytoplasmic [Na⁺] ([Na⁺]_i) is expected to fall rapidly to a low level (3), dissipating the Na⁺ gradient across the cell membrane. High [Mg²⁺]_o was employed so that the inward driving force of Mg²⁺ and consequently Mg²⁺ influx should be enhanced. One myocyte treated with even higher [Mg²⁺]_o (30 mM) also showed no clear increase in [Mg²⁺]_i (Fig. 2B), indicating no supporting evidence for the Na⁺ gradient-dependent Mg²⁺ transport. However, absence of any clear change in [Mg²⁺]_i in these experiments can also be explained if Mg²⁺ permeability of the cell membrane is very low and therefore passive Mg²⁺ influx is so slow (even with the increased driving force) that inhibition of Mg²⁺ extrusion fails to cause detectable changes in [Mg²⁺]_i in the time scale of these measurements. We therefore carried out experiments in which passive Mg²⁺ influx was greatly facilitated by the aid of ionomycin, an ionophore for divalent cations.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Cytoplasmic concentration of Mg2+ ([Mg2+]i) measurement with extracellular perfusion of the Na+-free solution containing 10 mM Mg2+ (A) or 30 mM Mg2+ (B). Cells were initially perfused with Ca2+-free Tyrode solution, and extracellular Na+ and Mg2+ concentrations were changed for the periods indicated by the horizontal bars. Values of means ± SE from 3 cells are plotted in A, whereas values from single myocyte are plotted in B.

Effect of extracellular Na+ on ionomycin-induced rise of [Mg²⁺]_i. Under Ca²⁺-free conditions, ionomycin is expected to facilitate Mg²⁺ influx (37), because Mg²⁺ is the only divalent cation present in the extracellular space. Perfusion of the myocytes with 10 µM ionomycin plus high [Mg²⁺]_o (10 mM) caused a gradual and nearly linear increase in [Mg²⁺]_i, probably due to the increased influx of Mg²⁺ (Fig. 3). Changes in [Mg²⁺]_i ([Mg²⁺]_i) by 30 min treatment of ionomycin were clearly smaller in the presence of 130 mM extracellular Na⁺ than in its absence (Fig. 3). The rate of Na⁺-dependent [Mg²⁺]_i calculated as the difference between the values with or without extracellular Na⁺ at 30 min (Fig. 3) was, on average, 0.76 mM/30 min or 0.42 µM/s. The ionomycin-induced [Mg²⁺]_i was significantly reduced by lowering [Mg²⁺]_o, as expected from the extracellular origin of the increased [Mg²⁺]_i (Fig. 4). The rise of [Mg²⁺]_i was substantial only in the absence of extracellular Na⁺ at normal [Mg²⁺]_o (1.2 mM) and was virtually absent at 0.4 mM [Mg²⁺]_o independent of extracellular Na⁺ (Fig. 4). These observations are qualitatively similar to those in smooth muscle cells of guinea pig tenia (see Fig. 3 of Ref. 35) and are consistent with the hypothesis that Mg²⁺ is extruded through a Na⁺ gradient-dependent pathway to cause smaller net influx of Mg²⁺. This hypothesis was further tested by observation of the Na⁺ dependence of net Mg²⁺ efflux in the following series of experiments.

View larger version (11K): [in this window] [in a new window]   Fig. 3.   Effects of extracellular Na+ on rise of [Mg2+]i induced by ionomycin plus high extracellular [Mg2+]. Changes in [Mg2+]i are expressed as difference from basal level ([Mg2+]i) and plotted as a function of time. Cells were initially perfused with Ca2+-free Tyrode solution, and 10 µM ionomycin was introduced with raised [Mg2+] (10 mM) either in the presence of 130 mM Na+ (+Na, ) or in the absence of extracellular Na+ (Na, ) for the period indicated by horizontal bar. All cells treated with ionomycin in the absence of Na+ and a part of the cells (8/42) treated with ionomycin in the presence of 130 mM Na+ () were then reperfused with Ca2+-free Tyrode solution. Other cells were perfused with solutions of different compositions for the experiments shown in Figs. 5-7. Each symbol represents mean ± SE (n = 42 for , n = 8 for , n = 12 for ).

View larger version (10K): [in this window] [in a new window]   Fig. 4.   Extracellular Mg2+ concentration ([Mg2+]o) dependence of ionomycin-induced [Mg2+]i. With [Mg2+]o set at 0.4, 1.2, and 10 mM either in the absence of Na+ (open bars) or in the presence of 130-140 mM Na+ (solid bars), [Mg2+]i values at 30 min after application of 10 µM ionomycin were obtained from the experiments of the type shown in Fig. 3. Each column shows mean ± SE; from left to right, n = 6, 6, 5, 6, 42, and 12.

Net Mg²+ efflux from loaded myocytes. After the [Mg²⁺]_i was raised by treatment with ionomycin and high [Mg²⁺]_o (see above), washout of ionomycin by the Ca²⁺-free Tyrode solution containing 1.2 mM Mg²⁺ caused a rapid decrease in [Mg²⁺]_i toward the basal level, indicating net Mg²⁺ efflux (Fig. 3). Figure 5 shows the results of experiments designed to specifically study the rate of decline of [Mg²⁺]_i from the Mg²⁺-loaded cells. Myocytes were initially loaded with Mg²⁺ by treatment with ionomycin in the high-Mg²⁺ solution containing 10 mM Mg²⁺ and 130 mM Na⁺, as described above. Because the effects of ionomycin on cellular Mg²⁺ loading is quite variable from cell to cell, we appropriately adjusted the concentration (9-12 µM) and treatment time (25-60 min) of ionomycin to achieve similar Mg²⁺ loading: 0.4-0.7 mM (0.57±0.01 mM, n = 93) above the basal level. Washout of ionomycin and reduction of extracellular Mg²⁺ back to the normal level quickly decreased [Mg²⁺]_i in the presence of 140 mM extracellular Na⁺ (open inverted triangles in Fig. 5A). On the other hand, [Mg²⁺]_i did not decrease, or even slightly increase, in the absence of extracellular Na⁺ (open circles in Fig. 5A); this small increase in [Mg²⁺]_i is probably due to the continued Mg²⁺ influx driven by membrane potential. [Mg²⁺]_i started to decrease after reintroduction of 140 mM Na⁺ (open circles in Fig. 5A). The decay of [Mg²⁺]_i in the presence of extracellular Na⁺ was approximately linear during the first 10 min but slowed thereafter as [Mg²⁺]_i approached the basal level. We therefore analyzed the [Mg²⁺]_i over the initial 10 min in the following analysis.

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Effect of extracellular Na+ and imipramine on [Mg2+]i. A: after cells were loaded with Mg2+ by 9-12 µM ionomycin + 10 mM Mg2+, the perfusate was changed, at zero time, to one of the test solutions containing 1.2 mM Mg2+ with various Na+ concentrations for the period indicated by a horizontal bar. Cells were then washed by Ca2+-free Tyrode solution containing 1.2 mM Mg2+ and 140 mM Na+. Changes in [Mg2+]i (ordinate) are expressed as the difference from the level achieved at the end of loading: generally 1.2-1.3 mM [Mg2+]i (for details, see text). The test solution contained 0 mM Na+ (, n = 8), 140 mM Na+ (, n = 8), 0 mM Na+ + 200 µM imipramine (, n = 3), or 140 mM Na+ + 200 µM imipramine (, n = 6). B: imipramine concentration dependence of [Mg2+]i measured at 10 min after application of the test solutions with the protocol shown in A. Solid line indicates least-squares fit of [Mg2+]i data () as a function of imipramine concentration ([X]) by the Hill-type curve with parameters shown in the panel: [Mg2+]i = min + (max  min) {[X]N/(K1/2N + [X]N)}, where max and min denote, respectively, [Mg2+]i in the presence of a saturating concentration of imipramine and in its absence, N is the Hill coefficient, and K1/2 is imipramine concentration that gives a midpoint value of [Mg2+]i between min and max. In the range of independent variable values ([X]), the program iterated the fitting procedure until a set of 4 adjustable variables (max, min, K1/2, and N) that gave the least-squares sum of the error of [Mg2+]i values was found. The 2 dotted lines show the levels of average [Mg2+]i in the absence of imipramine at 0 mM [Na+]o (top) and 140 mM [Na+]o (bottom). In A and B, each symbol represents mean±SE of 3-8 cells.

View larger version (16K): [in this window] [in a new window]   Fig. 6.   Effect of extracellular Na+ and K+ on [Mg2+]i. Changes in [Mg2+]i from the loaded level (ordinate) were measured at 5.9 mM [K+]o (A) or 55.9 mM [K+]o (B). After cells were loaded with Mg2+, one of the test solutions containing 1.2 mM Mg2+ with various Na+ concentrations (marked in mM near the symbols) was introduced at zero time for 20 min. Each symbol represents mean±SE of 4-9 cells.

View larger version (14K): [in this window] [in a new window]   Fig. 7.   Analysis of [Mg2+]i from the experiments of the type shown in Fig. 6. A: extracellular Na+ dependence of [Mg2+]i measured at 10 min after application of the test solutions containing either 5.9 mM K+ () or 55.9 mM K+ (). Solid line indicates the least-squares fit of the data at 5.9 mM [K+]o by the Hill-type curve (see below). B: Na+-dependent [Mg2+]i calculated by subtraction of the values at zero [Na+]o for each data set and plotted as a function of [Na+]o. Solid line indicates the least-squares fit of the data at 5.9 mM [K+]o () by the Hill-type curve with parameters shown in the panel: [Mg2+]i = max + (min  max) · {[Na]oN/(K1/2N + [Na]oN)}, where min and max denote, respectively, minimum and maximum values of Na+-dependent [Mg2+]i, N is the Hill coefficient, and K1/2 is [Na+]o that gives a midpoint value of the Na+-dependent [Mg2+]i between min and max.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS. DISCUSSION REFERENCES

The present experiments used fluorescence signals of a Mg²⁺ indicator furaptra to study Mg²⁺ transport across the cell membrane. This method allows noninvasive measurements of [Mg²⁺]_i from single myocytes over a period of hours. A difficulty of [Mg²⁺]_i measurement by furaptra lies in the calibration of fluorescence signals in terms of [Mg²⁺]_i, because properties of furaptra are likely altered in cytoplasmic environments (21). The present calibration method of furaptra fluorescence signals relied on parameter values previously estimated in smooth muscle cells (34) and gave a mean value of the basal [Mg²⁺]_i of 0.71 mM. Although obtained in Ca²⁺-free Tyrode solution, this value is probably also applicable to the [Mg²⁺]_i in normal Tyrode solution containing 2.4 mM Ca²⁺, because our previous study showed that removal of extracellular Ca²⁺ had little influence on [Mg²⁺]_i (18). The value of 0.71 mM is consistent with estimates of a 0.5-1.0 mM range with various methods (see introduction), but somewhat lower than our previous estimate, on average 1.13 mM calibrated under identical conditions in rat cardiac myocytes (34). The difference may be due to cell-to-cell variation in the small number of myocytes (n = 9) in the previous study and was not further considered here.

Although [Mg²⁺]_i measured in the present study was thought to reflect, in the most part, Mg²⁺ flux across the cell membrane (Fig. 4), alteration of intracellular Mg²⁺ binding or Mg²⁺ sequestration by organelles could also contribute to [Mg²⁺]_i. We therefore minimized the changes in [Ca²⁺]_i and cytoplasmic pH that should affect Mg²⁺ buffering (23); changes in [Ca²⁺]_i are expected to be negligible and changes in pH should also be minimal in Ca²⁺-free conditions (9). The previous study from this laboratory (18) has shown that, after ~10 min perfusion with a 0.1 mM EGTA-containing solution, neither Ca²⁺ release from the sarcoplasmic reticulum (by 25 mM caffeine) nor intracellular acidosis of ~0.4 pH unit (by 5% CO₂) has little influence on [Mg²⁺]_i measured with furaptra. Ca²⁺-free conditions should also minimize Mg²⁺ uptake by mitochondria, as shown by electron probe microanalysis (4), and Mg²⁺ uptake by the sarcoplasmic reticulum as a counter ion of Ca²⁺ release (33). We also avoided experimental conditions that induce Mg²⁺ release from intracellular organelles by stimulation of cAMP production (29, 30) or muscarinic stimulation (38). It is possible, however, that quantification of transmembrane Mg²⁺ flux based on the measured [Mg²⁺]_i suffers, because of uncertainties related to intracellular binding and sequestration of Mg²⁺.

Absence of Na+-dependent changes in [Mg²⁺]_i in intact myocytes. While removal of extracellular Na⁺ (which should inhibit Na⁺ gradient-dependent Mg²⁺ extrusion) in the presence of high [Mg²⁺]_o (which should increase leak influx of Mg²⁺) failed to show any significant changes in [Mg²⁺]_i (Fig. 2), a large increase in cell membrane permeability to Mg²⁺ unmasks the Na⁺ gradient-dependent Mg²⁺ transport (Fig. 3). It is thus conceivable that very controversial findings reported concerning cardiac myocytes (see introduction) may result from variable Mg²⁺ permeability of the cell membrane in different experimental conditions (or cell conditions). Handy et al. (17) reported that Na⁺ withdrawal in the presence of 5 mM Mg²⁺ caused a small but clear increase in [Mg²⁺]_i (~28 µM/min) in rat ventricular myocytes at 37°C. The difference in the results between the present study and Handy et al. (17) could be due simply to the difference in experimental temperature (37°C vs. 25°C), but other experimental conditions that somehow alter the cell membrane permeability to Mg²⁺ could also be involved.

Evidence for Na+/Mg²⁺ exchange. Extracellular Na⁺ suppressed net Mg²⁺ influx into ionomycin-treated cells and facilitated net Mg²⁺ efflux from Mg²⁺-loaded cells. Imipramine markedly inhibited net Mg²⁺ efflux only in the presence of extracellular Na⁺. These results are most likely explained by Na⁺ gradient-dependent Mg²⁺ efflux (or Na⁺/Mg²⁺ exchange). The putative Na⁺-Mg²⁺ exchange may play a role in long-term regulation of [Mg²⁺]_i to prevent Mg²⁺ overloading of cardiac myocytes.

1

2

1

2

1

2

1

2

1

2

1