Na(+)-independent Mg(2+) transport sensitive to 2-aminoethoxydiphenyl borate (2-APB) in vascular smooth muscle cells: involvement of TRPM-like channels.

Magnesium is associated with several important cardiovascular diseases. There is an accumulating body of evidence verifying the important roles of Mg(2+)-permeable channels. In the present study, we estimated the intracellular free Mg(2+) concentration ([Mg(2+)](i)) using (31)P-nuclear magnetic resonance ((31)P-NMR) in porcine carotid arteries. pH(i) and intracellular phosphorus compounds were simultaneously monitored. Removal of extracellular divalent cations (Ca(2+) and Mg(2+)) in the absence of Na(+) caused a gradual decrease in [Mg(2+)](i) to approximately 60% of the control value after 125 min. On the other hand, the simultaneous removal of extracellular Ca(2+) and Na(+) in the presence of Mg(2+) gradually increased [Mg(2+)](i) in an extracellular Mg(2+)-dependent manner. 2-aminoethoxydiphenyl borate (2-APB) attenuated both [Mg(2+)](i) load and depletion caused under Na(+)- and Ca(2+)-free conditions. Neither [ATP](i) nor pH(i) correlated with changes in [Mg(2+)](i). RT-PCR detected transcripts of both TRPM6 and TRPM7, although TRPM7 was predominant. In conclusion, the results suggest the presence of Mg(2+)-permeable channels of TRPM family that contribute to Mg(2+) homeostasis in vascular smooth muscle cells. The low, basal [Mg(2+)](i) level in vascular smooth muscle cells is attributable to the relatively low activity of this Mg(2+) entry pathway.

Introduction

From both clinical and epidemiologic aspects, Mg2+-deficiency is related to cardiovascular diseases [1-3]. More Mg2+ intake is recommended to prevent arteriosclerosis and hypertension. It is important, however, to point out that the serum Mg2+ level does not reflect Mg2+-deficiency of the entire body [4]. Cellular Mg2+-deficiency is caused by a malfunction of Mg2+ transporters across the plasma membrane, as well as a fall in the extracellular Mg2+ concentration. It is therefore of great interest to investigate mechanisms driving Mg2+ transport into the cell, especially into vascular smooth muscle cells.

[Mg2+]i is believed to be regulated by two transmembrane Mg2+ pathways: the Na+–Mg2+ exchange driven by the Na+-gradient, and the Na+-independent ‘passive’ Mg2+ transport via Mg2+-permeable channels [5-8]. Since the molecular identification of the latter Mg2+ pathway, such as melastatin-type transient receptor potential (TRPM) homologue channels, there has been an accumulating body of evidence for the crucial role that this pathway plays in Mg2+ homeostasis [9-11]. Also, TRPM homologue channels are bifunctional proteins, which contain a kinase domain in the C-terminus.

[Mg2+]i is known to change slowly, and thereby act as a chronic regulator. In addition, changes in the intracellular milieu, such as the intracellular pH (pHi) and [ATP]i can affect [Mg2+]i regulation. However, the importance of TRPM homologues in [Mg2+]i regulation during relatively short durations has been assessed using fluorescent Mg2+ indicators. In the present study, we thus utilized 31P-NMR to estimate slow changes in [Mg2+]i over several hours in carotid arteries, which are now frequently used as a model to evaluate arteriosclerotic changes, and assessed the contribution of TRPM-like Mg2+-permeable channels.

Materials and methods

Preparation

Porcine carotid arteries were collected at an abattoir. The arteries were stripped of fat and connective tissue, and cut into segments of approximately 30 mm in length. The lumen was exposed by cutting the artery segments into two strips along the longitudinal direction. The endothelium was removed by scratching with a cotton-tipped stick. The resultant pig carotid artery strips (∼2 g wet weight) were mounted in a sample tube of 10 mm in diameter. This study was approved by the institutional committee of animal experiments.

31P-NMR

The methods employed for the 31P-NMR measurements were essentially the same as those previously described [12]. NMR spectrometers (GSX270W: JEOL, Tokyo, Japan; UNITY-500plus: Varian, Tokyo, Japan) were operated at 109.4 and 202.3 MHz, respectively. The temperature of the sample was maintained at 32°C. Radio frequency pulses corresponding to a flip angle of 30 were applied every 0.6 sec. 31P-NMR spectra were obtained by accumulating 2500 signals (free induction decays) over 25 min. Before Fourier transformation, a broadening factor of 20 Hz was applied to enhance the signal-to-noise ratio. Spectral peak resonances (frequencies) were measured relative to that of phosphocreatine (PCr) in p.p.m.

Control spectra were acquired in the absence of Ca2+. Then, experiments were carried out in the absence of extracellular Na+ to rule out the contribution of Na+-coupled Mg2+ transport, that is, Na+–Mg2+ exchange. Six major peaks were observed (Fig. 1): phosphomonoesters (PME), inorganic phosphate (Pi), PCr and the γ-, α- and β-phosphorus atoms of ATP (γ-, α- and β-ATP).

1

Changes in the 31P-NMR spectrum during exposures to a divalent-cation-free, Na+-free solution. After acquiring the control spectrum in a Ca2+-free solution (a), extracel-lular Mg2+ and Na+ were simultaneously removed (0 Ca2+, 0 Mg2+, 0 Na+: K+ substitution) for 125 min. The spectra (b) and (c) were obtained during 25–50 min and 100–125 min periods, respectively. Each spectrum was obtained with 2500 signals accumulated over 25 min. The whole spectrum is shown in (A), and the β-ATP peaks are shown expanded in (B). The vertical line indicates the initial chemical shift of the β-ATP peak. The explanations are the same for the spectra in (C) and (D), but the divalent cation-free, Na+-free solution (0 Ca2+, 0 Mg2+, 0 Na+: K+ substitution) contained 150 μM 2-APB.

Concentrations of phosphorus compounds were estimated by integrating the peak areas (Scion image; Scion Corp., Fredrick, MA, U.S.A.) and by correcting with their saturation factors (Pi, 1.60; PCr, 1.36; β-ATP, 1.07).

Estimation of [Mg2+]i and pHi

Intracellular pH (pHi) was estimated from the chemical shift observed for the Pi peak (δo(PI)), using a Henderson–Hasselbalch type equation:

where pKa is the negative logarithm of the dissociation constant of Pi (= 6.70), and δp(Pi) and δd(Pi) are the chemical shifts for H2PO4− (= 3.15 p.p.m.) and HPO4− (= 5.72 p.p.m.), respectively.The pHi value was used to correct the [Mg2+]i estimation.

Mg2+ usually binds to ATP as a 1 to 1 complex. [Mg2+]i was thus estimated from the chemical shift observed for the β-ATP peak (γoβ) using the following equation [13,14]:

where δfβ and δbβ are the chemical shifts of metal-free and Mg2+-bound forms of β-ATP, respectively. We have previously shown that KD“MgATP”δfβ and δbβ can be described as functions of pH [15] (See, Supplementary Material S1: Details of estimation procedures). KD“MgATP”(pHi) was derived from quadratic pH-functions for KD“MgATP” at 25 and 37°C [16], using the van't Hoff isochore. The pH–functions of δfβ and δbβ were constructed by fitting the data points of model solutions with sigmoid curves [17]. Thus, Eq(2) can be rewritten as:

In Table 2, [Mg2+]i and pHi values were also estimated, from the chemical shifts of β- and γ-ATP [12,17]. For the chemical shift of γ-ATP, an equation analogous to Eq(3) can be written:

2

[Mg2+]i and pHi values estimated using two methods: (1) from the chemical shifts of β-and γ-ATP or 2) from the chemical shifts of β-ATP and Pi.(See Materials and methods for details). [Mg2+]i and pHi values during [Mg2+]i deple- tion were compared in A and B. In C and D, [Mg2+]i and pHi values were during the elevation of [Mg2+]i. Single (*) and double asterisks (**) indicate P<0.05 and P<0.01 versus control, respectively.

(A) Exposure to a divalent cation- and Na+-free solution (n = 7).
	1) from β- and γ-ATP			2) from β-ATP and Pi
[Mg2+]i (mM)		pHi			[Mg2+]i (mM)				pHi
Control (0 Ca2+)	0.70±0.13	7.20±0.11	0.75±0.09	7.09±0.05
0 Ca2+, 0 Mg2+, 0 Na+ (K+) 25–50 min	0.61±0.07**	7.09±0.09**	0.62±0.04**	7.06±0.04**
0 Ca2+, 0 Mg2+, 0 Na+ (K+) 100–125 min	0.43±0.04**	7.01±0.08**	0.46±0.05**	6.92±0.05**
(B) Exposure to a divalent cation- and Na+-free solution containing 2-APB (n = 7).
	1) from β- and γ-ATP			2) from β-ATP and Pi
[Mg2+]i (mM)		pHi			[Mg2+]i (mM)				pHi
Control (0 Ca2+)	0.68±0.04	7.21±0.05	0.74±0.05	7.09±0.04
0 Ca2+, 0 Mg2+, 0 Na+ (K+) +150μM 2-APB 25–50 min	0.65±0.08*	7.09±0.08**	0.68±0.07*	7.03±0.04**
0 Ca2+, 0 Mg2+, 0 Na+ (K+) +150μM 2-APB 100–125 min	0.60±0.12**	7.00±0.11**	0.62±0.08**	6.94±0.05**
(C) Exposure to a Ca2+- and Na+-free, high Mg2+ (6.0 mM) solutions (n = 7).
	1) from β- and γ-ATP			2) from β-ATP and Pi
[Mg2+]i (mM)		pHi			[Mg2+]i (mM)				pHi
Control (0 Ca2+)	0.72±0.10	7.20±0.09	0.78±0.08	7.08±0.12
0 Ca2+, 6.0 Mg2+, 0 Na+ (K+) 25–50 min	1.11±0.28**	7.12±0.10**	1.18±0.28**	7.03±0.05**
0 Ca2+, 6.0 Mg2+, 0 Na+ (K+) 100–125 min	1.63±0.23**	6.99±0.11**	1.79±0.18**	6.92±0.04**
(D) Exposure to a Ca2+- and Na+-free, high Mg2+ (6.0 mM) solution containing 2-APB (n = 6).
	1) from β- and γ-ATP			2) from β-ATP and Pi
[Mg2+]i (mM)		pHi			[Mg2+]i (mM)				pHi
Control (0 Ca2+)	0.70±0.09	7.20±0.11	0.75±0.04	7.10±0.03
0 Ca2+, 6.0 Mg2+, 0 Na+ (K+) +150μM 2-APB 25–50 min	0.91±0.10**	7.06±0.07**	0.93±0.06**	7.02±0.02**
0 Ca2+, 6.0 Mg2+, 0 Na+ (K+) +150μM 2-APB 100–125 min	1.10±0.08**	6.99±0.04**	1.14±0.08**	6.95±0.03**

δfγ and δbγ are the chemical shifts of metal-free and Mg2+-bound forms of γ-ATP, respectively, and these parameters are also expressed as pH-functions. [Mg2+]i and pHi were estimated by solving the Eq(3) and Eq(4) simultaneously.

Similar to the effect of Mg2+, Ca2+-binding to ATP changes the chemical shifts of the ATP peaks. However, basal [Ca2+]i is maintained at ∼100 nM in smooth muscle cells under physiological conditions. Furthermore, experiments in the present study were carried out mainly under Ca2+-free conditions. Therefore, the effect of Ca2+-binding to ATP is considered negligible.

Solutions and chemicals

The extracellular solution used for the ‘normal'solution had the following composition in mM: NaCl, 137.9; KHCO3 5.9; CaCl2 2.4; MgCl2 1.2; glucose 11.8; HEPES 5 (pH adjusted to 7.4–7.5 at 32°C). The ionic composition was modified iso-osmotically. Also, divalent cation-free solutions contained 1 mM EDTA. The solutions used for 31P-NMR measurements were normally aerated with 95% O2/5%CO2. 2-aminoethoxydiphenyl borate (2-APB) was purchased from Calbiochem (San Diego, CA, USA).

RT-PCR

The procedures for RT-PCR were essentially the same as previously described [18]. Total RNA was extracted from porcine carotid arteries. After treatment with RQ1 DNase (Promega, Madison, WI, USA), the total RNA was subjected to an RT (reverse-transcription) reaction. RT was performed using a random hexamer (12 pmol/reaction) and Moloney murine leukemia virus (MMLV) reverse transcriptase 5 (100 U/reaction) according to the manufacture's instructions (Gibco-BRL, Rockville MD, U.S.A.). The cycling condition was 3 min of initial denaturation at 95°C followed by 35 cycles of 95°C for 30 sec, 54°C for 30 sec and 72°C for 35 sec. The RT sample was then used as a template for the PCR reaction. The amplicons (5 μl were run on a 2.5% agarose gel and stained with ethidium bromide. Because the cDNA sequences for porcine TRP (transient receptor potential) homologue cation channels have not been published, the PCR primers were designed by using the conserved sequences in humans and mice. The primers used are listed in Table 1.

1

PCR primers. PCR primers for porcine sample were designed by using the conserved sequences in humans and mice

Clones	Primer Sequence (+): Sense, (−): Antisense	Primer site (in human clones)	Accession No (human) (mouse)
TRPM2	(+): 5′-TTC CAG GAG ATG TTT GAG AC-3′	1604-1938	NM_003307
(−): 5′-TCA GGC TTG TTG GAG ATG AG-3′			NM_138301
TRPM4	(+): 5′-GTG GGA GGG ACT GGA ATT GA-3′	863-1263	NM_017636
(−): 5′-AGC TCA TCC AGG TAG GCT GA-3′			NM_175130
TRPM6	(+): 5′-TGT TGG TGG AGA TGC AGC C-3′	2862-3174	NM_017662
(−): 5′-CCT GCA TGT TGA TTC ACA GC-3′			NM_153417
TRPM7	(+): 5′-GAT GCC CTC AAA GAA CAT GC-3′	794-1269	NM_017672
(−): 5′-GGC TCT GCT GCA TCA GGA AG-3′			NM_021450

Statistics

Numerical data are expressed as the mean (S.D. Differences between groups with different experimental protocols were evaluated by use of ANOVA for repeated measures. When a significant difference was identified between the groups (P<0.05), individual comparisons at the same time point were performed using an unpaired t-test.

Results

Depletion of [Mg2+]ivia Na+-independent Mg2+ pathways

31P-NMR was used to continuously measure phosphorus compounds in porcine carotid artery smooth muscle (Fig. 1). In this study, we mainly estimated [Mg2+]i from the chemical shift of the β-ATP peak, and correction was made by pHi estimated from the chemical shift of the Pi peak.

During exposure to a divalent cation-free solution (i.e. 0 Ca2+, 0 Mg2+), [Mg2+]i decreased from from 0.74±0.11 to 0.49±0.08 mM (n = 7; Fig. 2A, ▪) after 125 min, while pHi did not change significantly (Fig. 2B, •). Essentially the same decrease in [Mg2+]i (from 0.75±0.09 to 0.46±0.05 mM; n = 7; Fig. 2A, □) was observed in the absence of Na+, suggesting that Mg2+-permeable channels make a major contribution to the changes in [Mg2+]i under divalent cation-free conditions. On the other hand, pHi decreased from 7.09 ±0.05 to 6.92 ±0.05 (n = 7) after 125 min in the absence of Na+ (Fig. 2B, ○), presumably due to the inhibition of Na+-coupled pHi regulatory mechanisms, such as Na+-H+ exchange and Na+-HCO3− co-transport.

2

Time course of changes in [Mg2+]i (A: squares) and pHi (B: circles) during exposure to divalent cation-free solutions. After control measurements in a Ca2+-free solution (control), extracellular Mg2+ was also removed for 125 min. Each point was obtained from the accumulation of NMR signals over 25 min. The open (□, ○) and filled symbols (▪, •) represent the data obtained under exposures to Na+-free (0 Ca2+, 0 Mg2+, 0 Na+: K+ substitution; n = 7) and Na+-containing solutions (0 Ca2+, 0 Mg2+; n = 7), respectively. The spectra shown in Fig. 1A and B correspond to this Na+-free experiment. The dotted lines indicate the mean value of [Mg2+]i and pHi before exposure to the divalent cation-free solutions (n= 14 for all preparations shown in this figure). Vertical bars represent S.D. values. Asterisks indicate statistically significant differences compared to the [Mg2+]i and pHi values before removal of extracellu-lar Mg2+ (*, P<0.05; **, P<0.01). Crosses on filled symbols, in the presence of Na+, indicate statistically significant differences compared to the open symbols, in the absence of Na+, at the same time point (†, P<0.05; ††, P<0.01).

Effects of 2-APB in Na+-independent depletion of [Mg2+]i

2-APB is known to block TRPM7 [11]. To substantiate the involvement of analogous Mg2+-permeable channels in vascular muscle cells, the effect of 2-APB was examined (Fig. 1C and D). As shown in Figure 3, application of 150 μM 2-APB to the divalent cation-and Na+-free solution significantly attenuated the depletion of [Mg2+]i (from 0.74 ± 0.05 to 0.62 ± 0.08 mM after 125 min; Fig. 3A, ▪). This inhibitory effect of 2-APB was concentration-dependent (Fig. 3C). The decrease in [Mg2+]i after 125 min was –0.27 ± 0.11 mM at 15 μM (n = 5), (0.22 ± 0.07 mM at 50 μM (n = 5) and –0.12±0.08 mM at 150 μM (n = 7). On the other hand, this drug had little effect on the changes in pHi (unpaired t-test, P >0.05, n = 7; Fig. 3B, •).

3

The effect of 2-APB application during exposure to divalent cation- and Na+-free solutions. (0 Ca2+, 0 Mg2+, 0 Na+). Changes in [Mg2+]i and pHi are plotted in (A; ▪) and (B; •), respectively. After acquiring the control data in a Ca2+-free solution, extracellular Mg2+ and Na+ were simultaneously removed, and 150 μM 2-APB was added (n = 7). The spectra in Figures 1C and D correspond to this experiment. The data obtained in the absence of 2-APB (open symbols: □, ○; the same data shown in Fig. 2, n= 7)are also plotted to clearly show the inhibitory effect of 2-APB. Asterisks indicate statistically significant differences compared to the [Mg2+]i and pHi values before removal of extracellular Na+ (*, P<0.05;**, P<0.01). Crosses on filled symbols indicate statistically significant differences compared to the open symbols at the same time point (†, P<0.05; ††, P<0.01). Bar graphs in (C) indicate effects of 15, 50 and 150 μM 2-APB during 125–150 min (n= 7 in (–) and 150 μM 2-APB; n= 5 in 15 and 50 μM 2-APB).

Increase in [Mg2+]ivia Na+-independent Mg2+ pathways

Next, to demonstrate the increase in [Mg2+]ivia transmembrane Mg2+-permeable channels, the effect of Na+ removal was examined in the presence of Mg2+ (Supplementary Fig. S1, A). Extracellular Ca2+ was again removed to potentiate Mg2+ transport, i.e. to reduce competition between divalent cations at the channel pore.

Changes in [Mg2+]i and pHi during exposure to Ca2+- and Na+-free solutions are shown in Figures 4A and B, respectively. In the presence of 1.2 mM Mg2+ (the ‘normal’Mg2+ concentration in extracellular medium), [Mg2+]i increased from 0.73±0.07 to 1.01±0.09 mM after 125 min (▪; n= 5; P<0.01). When the concentration of extracellular Mg2+ was increased to 6.0 mM, [Mg2+]i increased from 0.78±0.08 to 1.79 ± 0.18 mM after 125 min (□; n= 7; P<0.01). The [Mg2+]i rise was clearly enhanced by raising the extracellular Mg2+ concentration (unpaired t-test, P<0.01).

4

Changes in [Mg2+]i (A) and pHi (B) during exposures to Ca2+- and Na+-free solutions containing Mg2+. After the carotid artery preparations were superfused with a Ca2+-free solution containing 1.2 mM Mg2+, extra-cellular Na+ was removed for 125 min (filled symbols: ▪, •). In the experiments indicated by open symbols (□, ○), extracellular Mg2+ was increased to 6.0 mM during Na+ removal. Asterisks indicate statistically significant differences compared to the [Mg2+]i and pHi values before removal of extracellular Mg2+ (*, P<0.05; **, P<0.01). Crosses on open symbols indicate statistically significant differences compared to the filled symbols at the same time point (†, P<0.05;††, P<0.01).

Effects of 2-APB on the increase in [Mg2+]i

To assess whether the same Mg2+-permeable channels contributed to the inward and outward transport of Mg2+ under Na+-free conditions, we examined the effect of 2-APB in the presence of Mg2+. 2-APB (150 μM) was applied to a Ca2+- and Na+-free solution containing 6.0 mM Mg2+ (Supplementary Fig. S1, B). After 125 min, [Mg2+]i increased from 0.75±0.04 to 1.14±0.08 mM (Fig. 5A, ▪; n = 7; P<0.01), but this increase in [Mg2+]i was significantly smaller than without 2-APB (unpaired t-test, P<0.01). On the other hand, pHi with and without 2-APB, was comparable throughout experiments (not shown).

5

The inhibitory effect of 2-APB on [Mg2+]i rise in Mg2+-containing solutions. In (A), after acquiring the control data in a Ca2+-free solution, the extracellular Mg2+ was increased to 6.0 mM, Na+ was removed, and 150 μM 2-APB was added in the extracellular solution. The data indicated by open symbols (□) represent experiments without 2-APB (the same data shown in Fig. 4A, □). In (B), extracellular Na+ was substituted with NMDG. Crosses on filled symbols (□, ○) indicate statistically significant differences compared to the open symbols at the same time point (†, P<0.05;††, P<0.01). Bar graphs in (C) indicate effects of 15, 50 and 150 μM 2-APB during 125–150 min in the presence of Ca2+ (n= 4 for each experiment).

A gradual, and slow increase in [Mg2+]i was caused by substituting extracellular Na+ with N-methyl-D-glucamine (NMDG), even in the presence of Ca2+ (Fig. 5B, □; n= 4). Application of 150 μM 2-APB again attenuated the increase in [Mg2+]i significantly (Fig. 5B, ▪; n= 4; P<0.01), suggesting that TRPM7-like Mg2+-permeable channels sensitive to 2-APB play a crucial role in Mg2+ regulation (i.e. uptake) under physiological conditions, in which Ca2+ is present. Lower concentrations (15 and 50 μM) of 2-APB only attenuated the increase in [Mg2+]i slightly (P >0.05; Fig.5C).

In the present experiments, Na+ was frequently substituted with equimolar K+. To rule out that Na+–Mg2+ exchangers coupled Mg2+ transport with K+ under Na+-free conditions, we examined the effect of amiloride, which is known to inhibit a broad range of Na+-coupled transporters [19], including Na+–Mg2+ exchange [12, 20]. Application of amiloride (1 mM) had little effect on the increase in [Mg2+]i caused by exposure to a Ca2+- and Na+-free solution containing 6.0 mM Mg2+ (unpaired t-test, P>0.05; see Supplementary Fig.S2).

Estimation of [Mg2+]i and pHi from the β and γ-ATP peaks

To confirm the changes in [Mg2+]i and pHi estimated above, we used a different procedure; specifically, the chemical shifts of β- and γ-ATP were used to solve simultaneous equations for [Mg2+]i and pHi (see, Material and methods). The changes in pHi, as well as [Mg2+]i, were comparable between the two analyses (Table 2). [Mg2+]i estimated from β- and γ-ATP were slightly smaller because of the higher pHi estimated; i.e. the apparent dissociation constant for MgATP is smaller in a higher pH.

High-energy phosphates

ATP is known to affect the activity of TRPM7 and to act as an important intracellular Mg2+ buffer. Correlation analysis revealed that [ATP]i and [Mg2+]i are not correlated regardless of the presence of extracellular Mg2+ (Fig. 6). Also, [PCr]i did not change significantly throughout (Supplementary Table S1).

6

Correlation between [ATP]i and [Mg2+]i in the absence (0 Mg2+; A) and presence of extracellular Mg2+ (6 mM Mg2+; B). Each open () and filled circle () represents a 31P-NMR data point (spectrum accumulated over 25 min) obtained in the absence and presence of 2-APB, respectively. [ATP]i is expressed relative to that in the control. Data points in (A) are from experiments shown in Fig. 3, while those in (B) are from Fig. 5A.

To confirm the transcription of genes for Mg2+-permeable channels, RT-PCR was performed. Because the cDNA sequences for porcine TRP (transient receptor potential) homologue cation channels have not been published, the PCR primers were designed by using the conserved sequences from humans and mice. Of TRPM2, 4, 6 and 7, the TRPM7 was predominant. TRPM6 was also detectable under the same PCR condition (Fig. 7; see also Supplementary Fig. S3).

Discussion

The present 31P-NMR measurements revealed several features of [Mg2+]i modulation via transmem-brane Mg2+-permeable channels, including sensitivity to 2-APB. Simultaneous removal of extracellular Mg2+ and Ca2+ reduced [Mg2+]i to approximately 60% of the control after 125 min even in the absence of Na+ (Fig. 2), under which conditions Na+-coupled Mg2+ transporters i.e. Na+–Mg2+ exchange, do not operate. In addition, the removal of extracellular Na+ in the presence of extracellular Mg2+ increased [Mg2+]i in an extracellular Mg2+-concentration-dependent manner (Fig. 4), and this effect was enhanced by the simultaneous removal of Ca2+. Altogether, these results suggested an important role of transmembrane Mg2+-permeable channels in regulating [Mg2+]i in vascular smooth muscle cells.

As shown in Fig. 5B, the activity of Mg2+-permeable channels are attenuated by extracellular Ca2+, but Mg2+ entry still occurred. In these experiments, extracellular Na+ was substituted with NMDG, and this procedure maintained the negative membrane potential, unlike Na+ substitution with K+, which was used in other experiments. Furthermore, simultaneous removal of extracellular divalent cations, that is, Ca2+ and Mg2+, did not significantly decrease [Mg2+]i when extracellular Na+ was substituted with NMDG (data not shown). Therefore, it is considered that, although Mg2+ loss can be caused via Mg2+-permeable channels under extreme conditions, this mechanism mainly contributes to Mg2+ uptake under physiological conditions, in which the negative resting membrane potential is preserved (probably several tens of mV).

TRPM6 and TRPM7 are known Mg2+-permeable non-selective cation channels. It has been shown that TRPM6 is abundant in the kidney and small intestine, while TRPM7 is expressed ubiquitously in numerous tissues and organs [9, 21, 22]. It is thought that the former and latter are responsible for Mg2+ regulation at the organ and cellular levels, respectively. In line with this notion, RT-PCR examinations revealed that TRPM7 was a major component of TRPM homologues in porcine carotid arteries, but TRPM6 was also detectable (Fig. 7). Previous patch clamp studies have shown that the removal of extracellular divalent cations facilitates TRPM7-like current [9, 11, 23], a finding that is in good agreement with our [Mg2+]i measurements, that is Figure 5.

7

RT-PCR detection of TRPM channel members in porcine carotid arteries. The PCR amplification was performed by 35 cycles. A 100-bp molecular weight marker was used (right column). The size of each PCR product is as expected from the mouse and human sequences.

2-APB is known to block store-operated Ca2+ release-activated Ca2+ (CRAC) channels as well as Mg2+-permeable channels [24]. However, Mg2+ transport via CRAC channels is considered negligible, because these channels have even greater Ca2+ selectivity than voltage-sensitive Ca2+ channels [25]. In the present study, Mg2+ transport via Mg2+-permeable channels demonstrated as Na+-independent [Mg2+]i changes, was not completely suppressed by 2-APB at <150 μM, which is three times greater than the IC50 (∼50 μM) reported for TRPM7-like cation currents [11, 23]. This discrepancy may be explained as follows: (1) The IC50 of 2-APB in TRPM7 may be altered by the intracellular milieu, namely, [ATP]i is maintained in our 31P-NMR measurements, while no ATP was added in the intracellular solution of the previous patch clamp experiments in which the effect of 2-APB was examined [11, 23]; (2) Native TRP channels may act as multimeric proteins of complex compositions [26]; 3) some other, as yet unknown Mg2+ transport mechanisms operate in parallel. Recently, it has been shown that the application of 2-APB does not block, but rather facilitates TRPM6 [27], which is contained as a minor component in porcine carotid arteries. In future studies, it would thus be interesting to examine the effects of drugs, such as 2-APB in TRPM6/7 knockdown cells. NMR measurements provide us with abundant information on [Mg2+]i regulation, that is pHi and [ATP]i which affect intracellular Mg2+-buffering, but require a considerable amount of vascular smooth muscle tissue to achieve reasonable time-resolution. For tissue level experiments, gene-targeting techniques, for example, RNA interference, remain to be improved.

The TRPM homologue channels consist of channel pore and kinase domains. Thus, these channels are referred to as ‘chanzymes’. It has been reported that Mg2+ and MgATP regulate the activity of TRPM7 via the kinase domain [9–11, 28]. In addition to this mechanism, intracellular ATP itself acts as an important Mg2+ buffer. With respect to the inhibitory effect of 2-APB on [Mg2+]i rise in the absence of extracellu-lar Ca2+ and Na+ (in the presence of 6 mM Mg2+; Fig. 4), one may suspect that 2-APB interacts with the kinase domain to elevate its sensitivity to [Mg2+]i. If this mechanism operates, increase in [Mg2+]i would be slowed at around the [Mg2+]i which significantly reduces the open probability of the Mg2+-permeable channels. The present 31P-NMR measurements, however, revealed that changes in [Mg2+]i were suppressed throughout the application of 2-APB (Fig. 6; Supplementary Material IV, Supplementary Table S1). Therefore, the present results suggest that 2-APB directly blocks Mg2+-permeable channels at the channel pore.

Reported values for [Mg2+]i in smooth muscle cells are lower than found in the other two muscle types, that is, skeletal and cardiac muscles. Furthermore, among smooth muscles distributed in numerous tissues and organs, we estimated [Mg2+]i in vascular smooth muscle cells to be the lowest after correcting for the pHi and temperature [8]:(0.8-0.85 mM in taenia caeci; 0.7–0.9 mM in the uterus; 0.8 mM in the urinary bladder and 0.65–0.75 mM in the carotid artery. Under divalent cation-free conditions, [Mg2+]i decreased to ∼60% of the control value after 125 min in porcine carotid arteries, however, the decrease in [Mg2+]i was much slower than that previously observed in intestinal smooth muscle cells (a decrease to tens of μM after 100 min) [6, 17]. The difference in the rate of [Mg2+]i depletion under Na+-free conditions suggests that the activity of Mg2+-permeable channels is lower in vascular smooth muscle cells. This hypothesis may account for the fact that the lowest basal [Mg2+]i level has been estimated in vascular smooth muscle cells and may contribute to our understanding of Mg2+-dependent cellular mechanisms. For example, the physiological significance of Ca2+-induced Ca2+ release from ryanodine receptors in vascular smooth muscle could be attributed to the low [Mg2+]i[29]. Also, Mg2+-permeable non-selective cation channels, such as TRPM7, have been recently shown to play a crucial role in generating spontaneous rhythmicity in pacemaker interstitial cells of intestinal smooth muscle tissues [30, 31]. Taken together, these facts warrant systematic investigation into cell- and tissue-specific roles of TRPM channels in numerous smooth muscle tissues containing pacemaker-like interstitial cells [32-35].

Functional mutations of ion channels are known to cause numerous diseases. Functional mutations of TRPM7 result in a significant effect on cellular Mg2+ homeostasis. It has been experimentally shown that mutations of the kinase domain in TRPM7 modulate the sensitivity to [Mg2+]i. Furthermore, knockout of TRPM7 causes intracellular Mg2+-deficiency in cultured cells [36]. In clinical fields, it has long been recognized that Mg2+ is associated with several important diseases, e.g. diabetes mellitus, hypertension, and cardiovascular and cerebrovascular diseases [37, 38]. Subgroups of these diseases may involve functional mutations of Mg2+-permeable channels contributing to passive Mg2+ transport via the electrochemical gradient. Similarly to the studies on TRPM6 [39, 40], future genetic analyses to clarify the contribution of TRPM7-like Mg2+-permeable channels to cardiovascular diseases will be of great interest. Indeed, a TRPM7 variant has been recently reported for two Guamanian neurodegenerative disorders [41].

In conclusion, the present 31P-NMR data have suggested the existence of Mg2+-permeable channels of TRPM homologues, contributing to [Mg2+]i regulation in vascular smooth muscle cells. The fact that [Mg2+]i decreases slowly under divalent cation-free conditions, presumably corresponding to the low expression level and/or activity of Mg2+-permeable channels, may account for the low basal [Mg2+]i in vascular smooth muscle cells. From the inhibitory effect of 2-APB and the results of RT-PCR, TRPM7 appears to play a predominant role in the passive Mg2+ transport.