[Frontiers in Bioscience 5, d866-879, September 1, 2000]

HOME
Current Issue
VITAL LINKS
INDEX
PDF FILE
ACCESS PUBMED
REPRINTS
CAVEAT LECTOR

Send correspondence to:

Dr Hector Rasgado-Flores,
Dept. Physiology and Biophysics,
FUHS/Chicago Medical School,
3333 Green Bay Rd.,
North Chicago, IL 60064

Tel: 847-578-3425,
Fax: 847-578-3265,
E-mail: Hector.Flores@
finchcms.edu

KEY WORDS

Magnesium, Calcium, Potassium, Chloride, Muscle, Review

SEARCH FBS

Copyright © Frontiers in Bioscience, 1995

PLASMALEMMAL TRANSPORT OF MAGNESIUM IN EXCITABLE CELLS

Hector Rasgado-Flores 1 and Hugo Gonzalez-Serratos 2

1 Department of Physiology and Biophysics, FUHS/Chicago Medical School, N. Chicago, IL 60064, 2 Department of Physiology, University of Maryland at Baltimore, School of Medicine, Baltimore, MD 21201

TABLE OF CONTENTS

1. Abstract
2. Introduction
3. Previous knowledge about plasmalemmal Mg2+ transport in excitable cells
3.1. Intracellular perfusion or dialysis of barnacle muscle cells or squid giant axons
3.2. Evidence for Na+/Mg2+ exchange in barnacle muscle cells and squid giant axons
4. Recent contributions to the understanding of plasmalemmal Mg2+ transport in excitable cells
4.1. Overcoming the limitations of investigating Mg2+ transport
4.1.1. Demonstration that Mg-dependent Na+ fluxes can be used to study the Mg2+ exchanger
4.1.2. Demonstration that barnacle muscle cells can reliably be used to study Mg2+ fluxes
4.1.2.1. Assessment of activity of the Na/Mg exchanger operating in "reverse" mode of exchange (i.e., Nai/Mgo) in barnacle muscle cells
4.1.2.2. Measurement of activity of the Na/Mg exchanger operating in "forward" mode of exchange (i.e., Nao/Mgi) in barnacle muscle cells
4.1.3. Enhancement of the signal/noise ratio of ionic fluxes mediated by the Mg2+ exchanger in barnacle muscle cells
4.2. Demonstration that, in addition to Na+, the Mg2+ transporter also exchanges Mg2+ for K+ and Cl-
4.2.1. Introduction
4.2.2. Involvement of K+ in the regulation of [Mg2+]i
4.2.2.1. There is an absolute requirement of intracellular K+ for the Mgo-dependent Na+ efflux in squid axons
4.2.2.2. Removal of extracellular Mg2+ produces a simultaneous and equimolar reduction in Na+ and K+ efflux in squid axons
4.2.2.3. Removal of extracellular K+ (Ko) produces an increase in the free and total intracellular Mg concentration in intact barnacle muscle cells
4.2.3. Involvement of Cl- in the regulation of [Mg2+]i
4.2.3.1. There is an absolute requirement of intracellular Cl- for the Mgo-dependent Na+ efflux in squid axons
4.2.3.2. Removal of extracellular Mg2+ produces a simultaneous and equimolar reduction in Na+ and Cl- efflux in squid axons
5. Conclusions and Future Experiments
6. Comparison between the electroneutral Na+K+2Cl (or 2Na+1K+3Cl, in the squid) co-transporter and the postulated 2Na+2K+2Cl/1Mg electroneutral exchanger

1. ABSTRACT

In excitable cells, the concentration of intracellular free Mg2+ ([Mg2+]i) is several hundred times lower than expected if Mg2+ ions were at electrochemical equilibrium. Since Mg2+ is a permeant ion across the plasmalemma, it must be constantly extruded. An ATP-dependent Na/Mg exchanger has been proposed as the sole mechanism responsible for Mg2+ extrusion. However, this hypothesis fails to explain numerous observations including the fact that K+ and Cl- appear to be involved in Mg2+ transport. Until now three main limitations have hampered the studies of plasmalemmal Mg2+ transport: i) 28Mg, the only useful radioactive isotope of Mg2+, has a short half-life and is difficult to obtain; ii) squid giant axons, the ideal preparation to carry out transport studies under "zero-trans" conditions, are only available during the summer months; and iii) the ionic fluxes mediated by the Mg2+ transporter are very small and difficult to measure. The purpose of this manuscript is to review how these limitations have been recently overcame and to propose a novel hypothesis for the plasmalemmal Mg2+ transporter in squid axons and barnacle muscle cells. Overcoming the limitations for studying the plasmalemmal Mg2+ transporter has been possible as a result of the following findings: i) the Mg2+ exchanger can operate in "reverse", thus extracellular Mg2+-dependent ionic fluxes (e.g., Na+ efflux) can be utilized to measure its activity; ii) internally perfused, voltage-clamped barnacle muscle cells which are available all year long can be used in addition to squid axons; and iii) phosphoinositides (e.g., PIP2) produce an 8-fold increase in the ionic fluxes mediated by the Mg2+ exchanger. The hypothesis that we postulate is that, in squid giant axons and barnacle muscle cells, a 2Na+2K+2Cl:1Mg exchanger is responsible for transporting Mg2+ across the plasmalemma and for maintaining [Mg2+]i under steady-state conditions.

2. INTRODUCTION

Intracellular magnesium (Mgi) is the second most abundant intracellular cation (1). It plays an essential role in protein biosynthesis and it is a key cofactor for hundreds of enzymes (1), especially enzymes involved with transfer of phosphate groups (e.g., ATPases, phosphatases, kinases, etc.). Mgi modulates membrane receptors (2), ionic channels (3-5) and transporters (6-9). Activation of FAS on B-cell lymphomas causes an increase in [Mg2+]i that appears to be required for apoptosis (10).

Homeostasis of plasma concentration of Mg2+ in humans is achieved via renal conservation mechanisms (reviewed by Quamme and de Rouffignac, this volume. Ref. 11) and hormonal control of magnesium absorption (reviewed by Schweigel and Martens, this volume, Ref. 12). Changes in the Mg2+ plasma concentration occur during alcoholism (13), central nervous system injury (reviewed by Vink and Cernak, this volume, Ref. 14) and diabetes mellitus (15,16). In fact, a primary defect in [Mg2+]I handling may be a critical effector of non-insulin dependent diabetes mellitus (16,17). In addition, hypomagnesemia may produce nervous hyperexcitability (see, 18,19), tetanic syndrome (20), and meningo encephalic syndrome (21).

In excitable cells, the intracellular free Mg2+ concentration ([Mg2+]i) is several hundred times lower than expected if distributed passively (e.g., 600 times lower for squid axons and barnacle muscle cells). Since plasma membranes are permeable to Mg2+ (22), constant extrusion of this ion occurs. Hormones induce massive efflux of Mgi from cells (reviewed by Romani and Scarpa, this volume, Ref. 23) but the underlying mechanisms of plasmalemmal transport of Mg2+ remain largely unknown.

3. PREVIOUS KNOWLEDGE ABOUT PLASMALEMMAL Mg2+ TRANSPORT IN EXCITABLE CELLS

Several biological preparations have been used to study plasmalemmal Mg2+ transport in cells. These preparations include the following: liver and cardiac cells (reviewed by Romani and Scarpa, this volume, Ref. 23); squid giant axons (e.g., 24-28), helix aspersa neurons (29), frog skeletal muscle (30,31), red blood cells (32), epithelial secretory cells (reviewed by Yago, et al., this volume, Ref. 33); kidney cells (reviewed by Quamme and de Rouffignac, and by Beyenbach, this volume, Refs. 11 and 34); epithelial gastrointestinal cells (reviewed by Schweigel and Martens, this volume, Ref. 12); and barnacle muscle cells (22,35-40). Among these preparations, barnacle muscle cells and squid giant axons offer the advantage that, owing to their large size, they can be internally perfused or dialyzed, and voltage-clamped. Thus, all the relevant parameters of plasmalemmal Mg2+ transport (i.e., composition of the intracellular environment, membrane potential) can be measured and controlled. Furthermore, intracellular perfusion/dialysis allows study of the function of plasmalemmal ionic transporters under "zero-trans" conditions (see below).

3.1. Intracellular perfusion or dialisis of barnacle muscle cells or squid giant axons

Operation of the Na+/Mg2+ exchanger requires that Na+ binds to the "cis"-side of the transporter protein while Mg2+ binds to the "trans"-side. Subsequently, the ions are translocated to the opposite side of the membrane. Demonstration that the putative Na+/Mg2+ exchanger is the mechanism responsible for transporting Mg2+ across the plasmalemma can be provided by showing that the flux (influx or efflux) of each ion involved is a function of the electrochemical gradient of the other ion. This can be determined by measuring the unidirectional influx or efflux of Na+, simultaneously with the unidirectional flux of Mg2+ going in the opposite direction. Ideally, experiments should be conducted under conditions in which the ions to be translocated are only present in the side of the membrane where they bind to the transporter (i.e., "zero-trans" condition). This permits avoidance of the following possible errors (41): i) there is no contamination of the fluxes mediated by the Mg2+ exchanger by fluxes mediated by other mechanisms; ii) there is no possibility of exchange fluxes mediated by the Mg2+ exchanger; and iii) any possible regulatory effects of the ions to be transported acting on the trans-side of the membrane are prevented. These ideal conditions can be attained in giant cells using intracellular perfusion or dialysis techniques.

The methods for intracellular perfusion and dialysis in barnacle muscle cells and squid giant axons have been described in detail (42-48). Unidirectional fluxes under "zero-trans" conditions are attained by adding the ion (and its radioactive isotope) whose transport is to be measured, only at either the intracellular or extracellular fluids. Subsequently, effluxes or influxes of the labeled ion are measured accomplished by taken aliquots from the fluid at the opposite side of the plasma membrane where the ion was originally added, and measuring their radioactive content.

3.2. Evidence for Na+/Mg2+ exchange in barnacle muscle cells and squid giant axons

Numerous observations in injected and dialyzed (or perfused) squid axons and barnacle muscle cells have led to the suggestion that a Na/Mg exchanger is responsible for extruding Mg2+ from these cells. The evidence shows that Mg2+ efflux is:

  • Largely dependent on extracellular Na+ (Nao) (25,26,28);
  • Reduced by metabolic poisoning (25,26);
  • Highly sensitive to temperature (26);
  • Unaffected by changes in membrane potential (Vm) (26); and
  • Promoted by the presence of extracellular K+ (Ko) (39,40)

DiPolo and Beaugé (24) have confirmed and extended these observations reporting that:

  • Unidirectional Mg2+ efflux from dialyzed squid axons is completely dependent on ATP and external Na+ (Nao/Mgi exchange).
  • ATP-g -S substitutes for ATP for activation of this exchanger. This suggests that ATP is a substrate for kinases rather than for ATPases;
  • In the presence of ATP and the absence of Nao, extracellular Mg2+ (Mgo) stimulates (tracer) Mg2+ efflux (Mg/Mg exchange); and
  • Mg2+ efflux is insensitive to membrane potential (from -80 to -30 mV)

There is also a (tracer) Na/Na exchange in squid axons that is ouabain insensitive and is not mediated by the Na/Ca exchanger (since it does not require the presence of activating intracellular Ca2+) (49). Since this exchange has an absolute requirement for ATP (49), it is unlikely to be mediated by the Na/H exchanger and is therefore, likely to be mediated by the (ATP-dependent) Na/Mg exchanger. In summary, the published results suggest that Na/Mg exchange:

  • Is the only mechanism responsible for extruding Mg2+ from squid axon and barnacle muscle;
  • Has an absolute requirement for ATP;
  • Like Na/Ca exchange, seems to be able to operate in several modes of exchange: Nao/Mgi; intracellular Na+ (Nai)/extracellular Mg2+ (Mgo); (tracer) Mg/Mg, and (tracer) Na/Na exchange modes; and
  • It is voltage-insensitive;

Several critical observations, however, remain unexplained by the Na/Mg exchanger hypothesis:

a. Extracellular Na+ (Nao) activates Mg2+ efflux (25) and extracellular Mg2+ (Mgo) activates Na+ efflux (50) with Michaelis-Menten kinetics suggesting a stoichiometry of 1Na+:1Mg2+. However, Nao-dependent Mg2+ efflux is insensitive to membrane potential (24,26) suggesting that either: (i) the stoichiometry of the exchanger is 2Na+:1Mg2+; (ii) another cation is co-transported with 1 Na+ in exchange for Mg2+ (1Na+ + cation/Mg2+ exchange); or (iii) the transporter itself has a negative charge that is neutralized by Na+ (1Na+:1Mg2+ exchange) and the return half-cycle carries a different cation than Na+ (e.g., K+).

b. If the electrochemical potential of Na+ is responsible for maintaining [Mg2+]I under steady-state conditions, [Mg2+]I would be governed by the following equations (26):

For an electroneutral exchange (i.e. 2Na+:1Mg2+):

where the suffixes i and o represent the intra and extracellular compartments, respectively.

For an electrogenic exchange (i.e. a stoichiometry other than 2Na+:1Mg2+):

where n is the number of Na+ ions exchanged per Mg2+ ion, Vm, F, R, z and T have their usual meanings.

  • For the electrogenic 3Na+:1Mg2+ stoichiometry, [Mg2+]o = 44 mM (25,26), [Na+]o = 461 mM, and [Na+]I = 27 mM (27), membrane potential (Vm) = -60 mV, at 16oC, [Mg2+]I would be about 2 m M at equilibrium. This value is at least 1000 times smaller than the measured free [Mg2+]I, which is 2 - 3.5 mM (25,51).
  • For the electroneutral 2Na+:1Mg2+ stoichiometry, [Mg2+]i at equilibrium would be about 0.15 mM which is at least 13 times smaller than the measured free [Mg2+]i.
  • For the electrogenic 1Na+:1Mg2+ stoichiometry, the expected [Mg2+]i at equilibrium would be about 9 mM which is at least 2.6 times larger than the measured [Mg2+]i.

These results are important since they indicate that 2 or more extracellular Na+ ions are required to account for the observed steady-state distribution of [Mg2+]i.

Unfortunately, in spite of the great interest in understanding how Mg2+ is regulated in excitable cells, studies of Mg2+ fluxes have been hindered by three major limitations:

1. 28Mg, the only useful radioisotope of Mg2+ has a very short half-life (21 h) and is currently being produced on only 1 day each year in the United States;

2.The use of squid axons is limited by their seasonal availability (3-4 months/year) and by the fact that, due to the fragility of the squid, the experiments have to be carried out at the place where they are captured (e.g., Marine Biological Laboratory, Woods Hole, MA);

3.Na+-dependent Mg2+ fluxes are very small (< 5 pmoles cm-2 sec-1) (24, and see below). Thus, ability to carry out systematic studies of these fluxes is limited by their intrinsic low signal/noise ratio.

As shown below, our laboratory has developed strategies to overcome these limitations.

4. RECENT CONTRIBUTIONS TO THE UNDERSTANDING OF PLASMALEMMAL Mg2+ TRANSPORT IN EXCITABLE CELLS

The following is a summary of two contributions that our laboratory has made to the field of Mg2+ transport in excitable cells. The first one is technical and consists of overcoming the limitations to study ionic fluxes mediated by the Mg2+ transporter, the second is scientific and consists of showing that, besides the electrochemical gradient of Na+, other ions are also involved in Mg2+ transport:

1. Overcoming the limitations of investigating Mg2+ fluxes:

      a. We have demonstrated that Mg-dependent 22Na fluxes can be used to study the Mg2+ exchanger (50)

      b. We have shown that, in addition to squid axons, barnacle muscle cells can reliably be used to study Mg2+ fluxes (36,37,39)

      c. We have increased the signal/noise ratio of ionic fluxes mediated by the Mg2+ exchanger by increasing the magnitude of the measured fluxes by ~ 8 fold (52).

2. Demonstrating that, in addition to Na+, the Mg2+ transporter also exchanges Mg2+ for K+ and Cl- (53)

4.1. Overcoming the limitations of investigating Mg2+ transport

4.1.1. Demonstration that Mg-dependent Na+ fluxes can be used to study the Mg2+ exchanger (50

To overcome the inherent problems of working with 28Mg, experiments were designed to assess whether the Na/Mg exchanger, like other gradient-driven transport systems (e.g., Na/Ca exchange and Na-K-Cl cotransporter) (46,54), is able to operate in "reverse", i.e., can mediate a Mgo-dependent Na+ efflux (i.e., Nai/Mgo exchange). Assessment of this possibility was initially accomplished in internally dialyzed squid giant axons. The experimental strategy consisted of increasing the intracellular concentration of Na+ ([Na+]i), removing extracellular Na+ (Nao) to prevent Na/Na exchange and comparing the efflux of Na+ in the presence and absence of Mgo. Mgo was isosmotically replaced with Tris or better yet, to maintain the extracellular concentration of divalent cations constant, with Ba2+ or Ca2+.

Figure 1 shows a representative experiment in which the effect of various concentrations of extracellular Mg2+ were studied on Na+ efflux. The external solution was free of Na+ and contained ouabain (to inhibit the Na/K pump) and tetrodotoxin (TTX, to block Na+ efflux via Na+ channels). Solving for Michaelis-Menten equation, the kinetic data from this figure indicate that, the apparent KMgo = 23 + 2 mM and JNa(max) = 4.6 + 0.3 pmolocm-2osec-1.

Figure 1. Effect of various concentrations of extracellular Mg2+ (Mgo) on Na+ efflux in an internally dialyzed squid axon. Concentration of Mgo was diminished in 3 steps: from 25 to 12.5 mM (at a), then to 6.25 mM (at b), and finally to 0 (at c). Subsequently, Mgo was restored in two steps: first to 6.25 mM (at d) and then to 25 mM (at e). Internal fluid contained very low [Ca2+] (~10-10 M) to prevent activation of the Na/Ca exchanger. Discontinuous horizontal lines on graph represent average flux values. Reproduced with permission from the American Physiological Society. (Gonzalez-Serratos & Rasgado-Flores, 1988).

Additional characterization of the Mgo-dependent Na+ efflux indicated the following (50):

  • In these experiments Mgo was replaced with Ba2+, however, the reduction in Na+ efflux was not due to the increase in [Ba2+] but was due instead to the removal of Mgo. Furthermore, replacement of Mgo with Ca2+ or Tris produces similar results (see below);
  • Changes in membrane potential were not responsible for the reduction in Na+ efflux (see below);
  • Mgo-dependent Na+ efflux was blocked by amiloride (IC50= 1.8 mM), but was largely insensitive to bumetanide (5 M) which blocks the Na+K+Cl co-transporter (a discussion about similarities and differences between the Mg2+ transporter and the Na+K+Cl co-transporter is presented below).
  • activation of the Mgo-dependent Na+ efflux by Mgo followed Michaelis-Menten kinetics.

In sum, these results indicate that, in squid giant axons, the Na/Mg exchanger can operate in "reverse" mediating a Mgo-dependent Na+ efflux.

4.1.2. Demonstration that barnacle muscle cells can reliably be used to study Mg2+ fluxes (36,37,39)

At the present time, the internally perfused, voltage-clamped barnacle muscle cell constitutes the only reasonable alternative to the squid giant axon to carry out transport studies under conditions in which the relevant parameters for membrane transport can be measured and controlled (i.e., voltage-clamp and intracellular perfusion or dialysis). As has been demonstrated, plasmalemmal transporters in squid and barnacle are very similar (e.g., Na/Ca exchanger) (55). Furthermore, recent experiments (see below) indicate that the Mg2+ transporter is also very similar for both preparations. Of particular importance is the fact that, barnacle muscle cells are available all year long and can survive indefinitely in an appropriate aquarium. Therefore, an ideal situation to achieve the most rapid and efficient progress on plasmalemmal Mg2+ transport would be if the squid could be used during the summer months and the barnacle during the rest of the year. However, the barnacle muscle preparation has the disadvantage that the measured fluxes are not as stable as in the squid. Therefore, to make the barnacle preparation as useful as the squid, it would be necessary to improve the signal/noise ratio of the measured fluxes in the former preparation. Fortunately, during the past few months our laboratory has been able to activate much larger ionic fluxes mediated by the Mg2+ transporter in barnacle muscle cells thereby increasing the signal/noise ratio (see below).

Two strategies have been used to demonstrate that a putative Na/Mg exchanger is present in internally perfused barnacle muscle cells:

I. measurement of the activity of the Na/Mg exchanger operating in "reverse". This was accomplished by following a protocol identical to the one used in squid axons (see Figure 1, above), i.e., by measuring the counter ion-dependency of the efflux of Na+ activated by extracellular Mg2; and

II. measurement of the activity of the exchanger operating in the "forward" mode of operation. This was accomplished by measuring a Nao-dependent Mg2+ efflux. As mentioned above, these experiments were limited by the difficulties in obtaining 28Mg and by the fact that the production of this isotope has to be purchased in its entirety since there are no other users.

4.1.2.1. Assessment of activity of the Na/Mg exchanger operating in "reverse" mode of exchange (i.e., Nai/Mgo) in barnacle muscle cells

Experiments have been performed in which the effect on Na+ efflux of isosmotic replacement of extracellular Mg2+ by either Ba2+, Ca2+ or Tris was studied. The results indicate that, like in squid axons (see Fig. 1), a "reversed" Nai/Mgo exchange can be readily measured in barnacle muscle cells (37,56). The decreases in Na+ efflux in response to Mgo removal were not due to the addition of the cation substituting for Mg2+ because similar results were obtained when this cation was Ca2+, Ba2+ or Tris. The reductions in Na+ efflux were accompanied with hyperpolarizations of 2.3 to 4 mV. These hyperpolarizations, however, were not responsible for reductions in Na+ efflux because similar experiments performed under voltage-clamp conditions yielded virtual identical results.

4.1.2.2. Measurement of activity of the Na/Mg exchanger operating in "forward" mode of exchange (i.e., Nao/Mgi) in barnacle muscle cells

Experiments have been performed in which extracellular Na+-dependent Mg2+ efflux was measured in internally perfused barnacle muscle cells (37,56). The results show that, removal of Nao (replaced by Tris) in the absence of Mgo produced reductions in Mg2+ efflux of 0.2 to 0.3 pmoles/cm2 sec. Likewise, addition of Nao (replacing Tris) in the presence of Mgo produced increases of 0.25 to 0.35 pmoles/cm2 sec. Interestingly, removal of Mgo (replaced by Cao) in the presence of Nao produced an increase in Mg2+ efflux of 0.3 pmoles/cm2 sec while addition of Mgo in the absence of Nao produced an increase in Mg2+ efflux of 0.17 pmoles/cm2 sec. These results indicate the presence of Nao-dependent Mg2+ efflux both in the presence and absence of Mgo. Likewise, they indicate that Mg/Mg exchange is manifested only in the absence of Nao. In the presence of Nao, removal of Mgo produces an increase in Mg2+ efflux, which may be due to activation of Mg2+ efflux by Nao. In this case, Mgo would appear to act as an inhibitor of Nao-dependent Mg2+ efflux as a consequence of its stimulation of Mg/Mg exchange.

Experiments performed under voltage-clamp conditions indicate that the Nao-dependent reductions in Mg2+ efflux were not due to changes in Vm. However, these results do not rule out the possibility that activity of the Mg2+ exchanger may be electrogenic. Analysis of the Vm changes associated with Mgo-dependent Na+ efflux and with the Nao-dependent Mg2+ efflux show that this possibility is very unlikely:

  • the hyperpolarizations that accompanies the reduction in the efflux of Na+ (when Mgo is removed) could be explained by the transfer of a net positive charge carried by Na+; and
  • the hyperpolarization that accompanies the reduction in Mg2+ efflux (when Nao is removed) requires that the net positive charge be carried by Mg2+ instead.

Consequently, these changes in Vm cannot be attributed to the activity of the exchanger but instead may result from some other effect of the ionic substitution. These results are consistent with the hypothesis that the Mg2+ exchanger is electroneutral.

In sum, the results demonstrate that the internally perfused barnacle muscle cell can be used to study ionic fluxes mediated by the putative plasmalemmal Na/Mg exchanger operating in either the "forward" or "reverse" mode of exchange.

4.1.3. Enhancement of the signal/noise ratio of ionic fluxes mediated by the Mg2+ exchanger in barnacle muscle cells

Figure 1 shows that, in squid axons, the signal/noise ratio for the Mgo-dependent Na+ efflux is 2.0/2 = 1.0. Conversely, in barnacle muscle cells, this ratio is from 5/17=0.3 to 5/30=0.16, respectively. To improve the signal/noise ratio in barnacle muscle cells we attempted to enhance the ionic fluxes mediated by the Mg2+ transporter. Based on the arguments listed below, we reasoned that Phosphatidylinositol-4,5-bisphosphate (PIP2) and/or phosphoarginine (P-Arg) could be strong candidates for activating the Mg2+ transporter:

  • The Mg2+ transporter has an absolute requirement for ATP (24);
  • ATP may activate transporters because it may work (reviewed in Ref. 57): i) as a substrate for ATPases; ii) as a substrate for protein kinases; iii) as a substrate for lipid kinases generating second messengers (e.g., Phosphatidylinositol phosphates); iv) by directly binding to the transporter inducing allosteric effects; iv) by inducing changes of actin cytoskeleton; v) by chelating polyvalent cations; and vi) by activating ATP-dependent phospholipases;
  • In squid axons, the non-hydrolyzable ATP analog, ATP-g -S, can substitute for ATP for activating the Mg2+ transporter. Thus, modulation by this nucleotide does not appear to involve an ATPase but may instead involve a kinase; and
  • In squid axons, the Na/Ca exchanger, which is modulated by ATP, can be activated by PIP2 (58) or P-Arg (59) in the absence of ATP.

Figure 2 shows a diagram depicting several of the possible pathways by which ATP could activate the putative Na/Mg exchanger:

Figure 2. Diagram showing several of the possible mechanisms by which ATP could activate the Mg2+ transporter: a) via a direct effect; b) by being a precursor for P-Arg; or c) by being a precursor for PIP2. PIP2 could in turn activate the transporter by: d) a direct effect; e) being a precursor of PIP3; f) being a precursor of diacylglycerol (DAG); or g) being a precursor of IP3. See text for further details.

a. ATP could be working directly either by being a substrate for an ATPase, by directly producing an allosteric modification of the transporter or another cellular component related to the transporter, or by chelating polyvalent cations;

b. ATP could work by being a substrate for arginine kinase yielding phosphoarginine (P-Arg) which would in turn activate the transporter;

c. ATP could work by being a substrate of phophatidylinositol (PI) kinase yielding PIP2;

d. PIP2 could directly activate the transporter;

e. PIP2 could work by being a substrate of Phosphatidylinositol 3'-kinase (PI3 kinase) yielding Phosphatidylinositol-3,4,5-trisphosphate (PIP3) which would in turn activate the transporter

f. PIP2 could work by being a substrate of phospholipase C (PLC) yielding diacylglycerol (DAG) which would in turn activate the transporter

g. PIP2 could work by being a precursor of inositol 1,4,5-trisphosphate (IP3) which would in turn activate the transporter

The experimental strategy followed to test whether PIP2 and/or P-Arg, could substitute for ATP in

activating the Mg2+ transporter consisted of comparing the Mgo-dependent Na+ fluxes in cells perfused with ATP-free solutions containing either 0.1 mM PIP2 or 5 mM P-Arg. To ascertain the absence of ATP in nominal ATP-free perfusates, 10 U/ml of the ATP-degrading enzyme apyrase (42) were added to the internal solutions.

Control experiments consisted of cells perfused with the ATP-free perfusate containing Apyrase; experimental cells were also perfused with a similar perfusate than experimental cells, but in this instance, the fluid also contained PIP2 or P-Arg.

The results using PIP2 show (52,26) that, in control cells, Mgo removal produced no effect on Na+ efflux. The efflux of Na+ under this condition is very steady and low (average 10 pmoles cm-2 sec-1) because it mainly represents a Na + "leak" in the absence of Nao and presence of ouabain, bumetanide and verapamil. In contrast, in the presence of PIP2, Mgo removal produced a large reduction in Na+ efflux (i.e., 30-75 pmol cm-2 sec-1) that became significantly different than control cells (P<0.05) 2 minutes following the removal.

The results using P-Arg show (52,26) that removal of Mgo in cells perfused with P-Arg produced a large reduction in Na+ efflux (i.e., 45-75 pmol cm-2 sec-1) that became significantly different than control cells 2 minutes after Mgo removal. In control solutions (0-ATP and Apyrase) Mgo removal did not affect Na+ efflux. However, in the presence of PIP2 or P-Arg, this manipulation produced a reversible reduction in Na+ efflux of 30-75 (n=7) and of 45-75 (n=6) pmol cm-2 sec-1, respectively.

These results demonstrate that PIP2 and P-Arg promote an 8-fold activation in Nao-dependent Mg2+ efflux in barnacle muscle cells. Clearly, it is expected that similar results would be obtained in squid giant axons.

Our laboratory is currently attempting to dissect the metabolic pathway by which P-Arg and PIP2 activate the Mg2+ exchanger. Preliminary studies have been performed assessing the ability of the metabolic products of PIP2 (i.e., IP3 and DAG) to activate the transporter. The results show that DAG but not IP3, activates the exchanger. This suggests that Protein Kinase C may be involved in activating the Mg2+ exchanger via a likely phosphorylation. In any event, the new signal/noise ratio of the PIP2- or P-Arg-stimulated Mgo-dependent Na+ efflux in barnacle is about 75/90= 0.8. This value is a great improvement over the 0.16-0.3 values in the absence of the stimulator and is very close to the one obtained using squid axons. Thus, barnacle muscle cells can now reliably be used to measure ionic fluxes mediated by the Mg2+ transporter. Furthermore, making the reasonable assumption that barnacle muscle cells and squid giant axons possess similar transporters (as is the case for the Na/Ca exchanger), both preparations can be used to study the plasmalemmal Mg2+ transporter.

4.2 Demonstration that, in addition to Na+, the Mg2+ tranporter also exchanges mg2+ for K+ and Cl-

4.2.1. Introduction

Numerous observations indicate that there is a strong relationship between the transport of Mg2+ and K+ in cells. Some examples of these observations are the following:

  • There is a direct correlation between the concentrations of intracellular free K+ and Mg2+ in skeletal muscle (31);
  • There is a strong correlation between the total intracellular concentrations of Mg and K in red cells, lymphocytes, and skeletal muscle (reviewed in Ref. 60);
  • Extracellular Mg2+ induces K+ efflux in barnacle muscle cells (40);
  • Mg2+ affects K+ transport in red blood cells and muscle cells (6,60-63);

There is a direct correlation between the changes in plasma Mg and K in pathological conditions such as Bartter's syndrome (64) and Alzheimer disease (65).

Likewise, some evidence, though indirect, suggests that there is a relationship between the fluxes of Cl- and Mg2+ across the plasma membrane of cells:

  • Extracellular Mg2+ and Cl- interact to modulate the tone and contractility of vascular muscle (9);
  • Intracellular Mg2+ inhibits Cl- efflux in barnacle muscle cells (66);
  • Net Mg2+ efflux is dependent on net Cl- efflux in Mg2+-loaded human erythrocytes (67);
  • Ingestion of supplementary dietary Cl- produces a reduction in plasma Mg2+ in rats (68).

In sum, given the fact that K+ and Cl- appear to be involved in regulating [Mg2+]I, attention should be given to the possibility that the electrochemical gradients of these ions could be involved in Mg2+ transport.

Using internally dialyzed squid giant axons and intact barnacle muscle cells we have explored the possibility that, in addition to the electrochemical gradient of Na+, the electrochemical gradients of K+ and Cl- may be involved in the regulation of [Mg2+]I (39,53). A summary of the results is the following:

4.2.2. Involvement of K+ in the regulation of [Mg2+]I

We have found that, in squid giant axons and barnacle muscle cells, the electrochemical gradient of K+ is coupled to Mg2+ transport. Three main observations described below support this:

I. There is an absolute requirement of intracellular K+ for the Mgo-dependent Na+ efflux in squid axons;

II. Removal of extracellular Mg2+ produces a simultaneous and equimolar reduction in Na+ and K+ efflux in squid axons; and;

III. Removal of extracellular K+ (Ko) produces an increase in the free and total intracellular Mg concentration in intact barnacle muscle cells.

A more detailed description of these observations is as follows:

4.2.2.1. There is an absolute requirement of intracellular K+ for the Mgo-dependent Na+ efflux in squid axons

We have assessed the role of intracellular and extracellular K+ for operation of the Mgo/Nai exchanger (53). To perform these experiments, the dialyzed squid axons were incubated in the absence of both intra- and extracellular K+, in the absence of extracellular Na+ (replaced by Tris) and presence of phenylpropyltriethylammonium bromide (PPTEA) to block K+ channels (69). Under these conditions, replacement of extracellular Mg2+ by Ba2+ produced no effect on Na+ efflux. This indicates that either intra- and/or extracellular K+ are necessary for activation of the Na/Mg exchanger. Addition of extracellular K+ promoted a ouabain-insensitive Na+ efflux. This activation, however, was not due to activation of the Na/Mg exchanger since removal of external Mg2+ produced no effect on Na+ efflux. On the other hand, the effect of intracellular K+ (Ki) on the Na/Mg exchanger was tested by adding Ki to the dialysis solution. This manipulation also produced an increase in Na+ efflux but in this instance, this effect is attributed to activation of the Na/Mg exchanger because under this condition, removal of extracellular Mg2+ produced a reversible reduction in Na+ efflux of 0.8 pmoles cm-2 sec-1. This indicates that activity of the Na/Mg exchanger operating in "reverse" (Nai/Mgo exchange) requires intracellular K+.

4.2.2.2. Removal of extracellular Mg2+ produces a simultaneous and equimolar reduction in Na+ and K+ efflux in squid axons

Requirement of Ki for operation of the Nai/Mgo exchange could be due to two possibilities: i) Ki is a necessary co-factor for Nai/Mgo exchange; or ii) Ki is co-transported with Nai in exchange for Mgo. To test these possibilities we determined the effect of removal of Mgo on the simultaneous efflux of K+ and Na+.

Figure 3 is an example of the simultaneous reduction of Na+ and K+ efflux on removal of extracellular Mg2+. The external solution was Na+-free and contained ouabain, TTX and bumetanide. The dialysis solution contained ATP-Mg, EGTA, no added Ca2+, and PPTEA.

Figure 3. Effect of Mg0 on simultaneous efflux of Na+ (O) and K+ (l) in an internally dialyzed squid axon. The external solution was Na+-free (replaced by Tris) and contained (among other components) 25 mM Mg2+ and 10 K+, 0.1 mM ouabain, 0.2 mM TTX and 10 m M bumetanide. The dialysis solution contained among other components (in mM): 39 Na+, 99 K+, 4 ATP-Mg, 5 EGTA, 1 Mg2+, and 20 PPTEA. Fluxes were allowed to equilibrate for 1 hr before sample collection. Solid and dashed lines represent the average of K+ and Na+ efflux in the presence and absence , respectively of Mgo. Reproduced with permission from the American Physiological Society (Rasgado-Flores & Gonzalez-Serratos, 1994).

To simultaneously measure the Mgo-dependent efflux of K+ and Na+, 42K and 22Na were used as tracers because these isotopes have very different half-lives (12.4 hours and 2.6 years, respectively). The magnitude of the fluxes of each ion was determined by subtracting the radioactive content of aliquots of the superfusate and counted immediately (furnishing the counts of both isotopes) from the counts obtained as a result of re-counting the same aliquots after the isotope of short half-life has decayed (giving the counts of the isotope with long half-life, i.e. 22Na). Under these conditions, the steady-state efflux of Na+ (open circles) and K+ (closed circles) were similar (1.6 and 1.4 pmoles cm-2 sec-1, respectively). Removal of Mg2+ (from a to b) produced a reversible reduction in the efflux of both Na+ and K+ (0.7 and 0.65 pmoles cm-2 sec-1, respectively).

The average ratio of the Mg2+-dependent Na+ efflux over the Mg2+-dependent K+ efflux from two independent experiments was 0.95 + 0.2. This experiment, therefore, indicates that extracellular Mg2+ activates the co-transport of stoichiometrically equal amounts of Na+ and K+.

4.2.2.3. Removal of extracellular K+ (Ko) produces an increase in the free and total intracellular Mg2+ concentration in intact barnacle muscle cells

We have observed that in intact barnacle muscle cells:

I. There is an inverse relationship between the concentrations of total intracellular Mg content and extracellular K measured using flame photometry and atomic absorption spectroscopy (39); and

II.Extracellular K+ (Ko) stimulates Mg2+ efflux, measured with differential absorption spectroscopy using eriochrome blue as a [Mg2+]i indicator (39).

These results suggest that K+ influx is coupled to Mg2+ efflux via a K+/Mg2+ exchange.

Involvement of K+ for Mg2+ transport may not be limited to invertebrate cells: in frog skeletal muscle, for example, it has been observed that there is a direct correlation between the levels of [Mg2+]i and intracellular free K+ (see Fig. 7 in Alvarez-Leefmans et al., 1986, Ref. 31). This is consistent with K+/Mg2+ exchange similar to the well-documented direct relationship between the intracellular concentrations of Na+ and Ca2+ in excitable cells resulting from activity of the Na/Ca exchanger (70).

One piece of information apparently contradicts the involvement of K+ on the regulation of [Mg2+]i in excitable cells: In injected squid giant axons, removal of extracellular K+ (Ko) does not affect unidirectional Mg2+ efflux (25). This observation does not rule out a role for Ko because Ko removal (in the presence of Ki) may not completely eliminate the possibility that intracellular K+ could leak from the cell and activate the exchanger from the external surface of the membrane. It has been reported that K+ efflux from the axon (through resting K+ channels) to the extracellular restricted-diffusion space (71) raises Ko to about 0.5 mM above the nominal K+ concentration in the seawater (72). Therefore, if Ko stimulates Mg2+ efflux with high affinity, the lack of effect of Ko removal on Mg2+ efflux is not adequate to rule out a role for Ko in the promotion of Mg2+ efflux.

In sum, the electrochemical gradients of Na+ and K+ appear to be coupled to regulate intracellular Mg2+. Consistently with this, an electroneutral 1Na++ 1K+: 1Mg2+ exchanger could be postulated to regulate [Mg2+]i. In this case, the equilibrium [Mg2+]i is predicted by:

Substituting [K+]i=325 and [K+]o=17 (73,74) and for the other terms as in equation 1, equation 3 gives [Mg2+]i = 49 mM. This value is about 19 times larger than the measured [Mg2+]i (25,51). Consequently, we are led to conclude that an electroneutral 1 Na++ 1 K+: 1 Mg2+ exchanger is unlikely to be responsible for maintaining [Mg2+]i under equilibrium conditions. Following this conclusion, the question arises then as to whether an additional ion may be involved in Mg2+ transport.

4.2.3. Involvement of Cl- in the regulation of [Mg2+]i

We have found that, in squid axons, in addition to the electrochemical gradients of Na+ and K+, the electrochemical gradient of Cl- is coupled to Mg2+ transport. Two main observations support this:

a. there is an absolute requirement of intracellular Cl- for the Mgo-dependent Na+ efflux; and

b. removal of extracellular Mg2+ produces a simultaneous equimolar reduction in Na+ and Cl- efflux. The following is a more detailed description of these observations:

4.2.3.1. There is an absolute requirement of intracellular Cl- for the Mgo-dependent Na+ efflux

We have studied the effect of intracellular Cl (Cli) on the Mgo-dependent Na+ efflux (53). To carry out these studies, the dialyzed squid axon was incubated in the absence of both intra- (substituted with Trizma-base and aspartate) and extracellular (substituted with methanesulfonate) Cl-. The results show that, under these conditions, removal of extracellular Mg2+ failed to induce a reduction in Na+ efflux. However, when the internal fluid was substituted for a solution containing 52 mM Cl-, this manipulation activated the efflux of Na+. That this was due to activation of the Na/Mg exchanger was shown by the fact that removal of external Mg2+ produced reversible reductions in Na+ efflux of 0.85-1.2 pmoles cm-2 sec-1.

4.2.3.2. Removal of extracellular Mg2+ produces a simultaneous equimolar reduction in Na+ and Cl- efflux

Figure 4 is an example of the simultaneous reduction of Na+ and Cl- efflux upon removal of Mg2+. To measure the efflux of Cl- and Na+ simultaneously, a similar strategy to the one used for the simultaneous measurement of K+ and Na+ fluxes was followed (see Fig. 3). In this case, the the radioisotopes used as tracers were 36Cl and 24Na. The half-lives of 36Cl and 24Na are 3 x 105 years and 14.9 hours, respectively.

Figure 4.. Effect of removal of Mg0 on simultaneous efflux of Cl- (O) and Na+ (>l ) in an internally dialyzed squid axon. Among other components, the external solution contained (in mM): 25 Mg2+, 567 Cl-, 10 K+, 0.1 mM ouabain, 0.2 mM TTX and 10 m M bumetanide. Among other components, the dialysis solution contained (in mM): 40 Na+, 52 Cl-, 100 K+, 1 Mg2+, 20 PPTEA, 8 EGTA, no added Ca2+, and 4 ATP-Mg. Fluxes were allowed to equilibrate for 1 hr before collection of the samples. Reproduced with permission from the American Physiological Society (from Rasgado-Flores & Gonzalez-Serratos, 1994). See text for further details.

Figure 4 shows that, in the presence of Mgo, the steady-state efflux of Na+ (closed circles) and Cl- (open circles) were 4.8 and 13 pmoles cm-2 sec-1, respectively. Removal of Mgo (from a to b and at c) produced a reversible reduction in the efflux of both Na+ and Cl- (1.02 + 0.2 and 1.13 + 0.2 pmol cm-2 sec-1, respectively). The ratio of the Mg2+-dependent Na+ efflux over the Mg2+-dependent Cl- efflux was 0.9 + 0.1 (n=2). This suggests that extracellular Mg2+ activates the co-transport of stoichiometrically equal amounts of Na+ and Cl-.

Consistent with these results, an electrogenic 1Na++ 1K++ 1Cl-:1 Mg2+ exchanger could be postulated to regulate [Mg2+]i. Under steady-state conditions, [Mg2+]i would be predicted by the following equation:

By substituting in the above equation the values (in mM) of [Cl-]i=120, and [Cl-]o=470 (75) and the rest of the terms as in equations 2 and 3, equation 4 gives an expected value of [Mg2+]i of 44 mM. This value is about 17 fold larger than the measured [Mg2+]i (25,51). Thus, thermodynamically, this stoichiometry is not feasible.

On the other hand, if an electroneutral 2Na++2K++2Cl-:1Mg2+ exchanger is postulated, the predicted value of [Mg2+]i under steady-state conditions would be determined by the following equation:

Substituting in this equation the appropriate values used before, gives an expected value of [Mg2+]i at equilibrium of 3.6 mM which is precisely in the range of measured [Mg2+]i (2-3.5 mM) (25,51). Thus, this stoichiometry is possible if [Mg2+]i is distributed nearly at equilibrium.

5. CONCLUSIONS AND FUTURE EXPERIMENTS

In an attempt to explain our results and other available information about Mg2+ transport, we suggest that the electrochemical potentials of Na+, K+, Cl- and Mg2+ are coupled through an electroneutral ATP-dependent

Na+K+Cl/Mg exchanger with a stoichiometry of 2Na++2K++2Cl-:1Mg2+ that operates near its thermodynamic equilibrium. This hypothesis is consistent with the observations that:

I. intracellular K+ and Cl- are required for operation of Nai/Mgo exchange (see above);

II.Ki and Cli are co-transported with Nai in stoichiometrically equal amounts in exchange for Mgo (Figs. 3 and 4);

III. the activation curves of extracellular Mg2+ for Mgo-dependent Na+ efflux follow Michaelis-Menten kinetics (50);

IV. Nao-dependent Mg2+ efflux is voltage-insensitive (24,26);

V. Nao-dependent Mg2+ efflux is ATP-dependent (24); and

VI. The steady-state [Mg2+]i is predicted by such exchanger.

This model is shown in Figure 5. The diagram shows that 2 Na+, 2K+ and 2Cl- ions bind to the exchanger at the extracellular side of the membrane while 1 Mg2+ binds to the intracellular side. This model is consistent with all the major observations on Mg2+ transport in excitable cells. Thus, we believe that this proposed exchanger is a reasonable working hypothesis that should be tested. The following is a list of experiments that should be performed to advance our understanding of the plasmalemmal Mg2+ transporter:

Figure 5. Schematic representation of the proposed 2 Na+ + 2K+ + 2Cl-: 1Mg2+ exchanger. See text for details.

1. Demonstration that in barnacle muscle cells, presence of intracellular K+ and Cl- are essential for operation of the Mgo-dependent Na+ efflux, and that intracellular K+ and Cl- are co-transported with Nai in exchange for Mgo. A positive result will indicate that the postulated 2Na++2K++2Cl-:1Mg2+ exchanger is not unique for squid axons.

2. Confirmation that P-Arg and PIP2 stimulate the plasmalemmal Mg2+ transporter in squid axons. A positive result will indicate that the transporter in squid and barnacle share basic characteristics;

3. Demonstration that the fluxes of Na+,K+,Cl- going in one direction across the plasmalemma are coupled with the flux of Mg2+ going in the opposite direction. This can be determined by showing that the flux (influx or efflux) of each of the four ions involved is a function of the electrochemical gradient of the other three ions;

4. Testing if the stoichiometry of the Mg2+ exchanger is: 2Na++2K++2Cl-:1Mg2+. This can be accomplished by measuring the stoichiometric ratio of the Mg2+-dependent unidirectional fluxes of Na+, K+ and Cl- as well as of the Na++K++Cl--dependent unidirectional fluxes of Mg2+ under "zero-trans" conditions;

5. Assessing whether the electrochemical gradients of Na+, K+ and Cl- determine the intracellular free and total Mg2+ concentrations under steady-state conditions. This can be accomplished by measuring the effect of changes in the electrochemical gradients of Na+, K+, or Cl- and of combinations of these three ions on the total and free intracellular concentrations of Mg, Na, K and Cl in either intact or voltage-clamped cells which have been perfused/dialyzed only during the first hour of the experiment in order to establish a desired composition of the intracellular environment.

6. Evaluation of whether the 2Na++2K++2Cl-:1Mg2+ exchanger operates as a result of a simultaneous or consecutive binding of the ionic species involved. This information could be assessed by studying the specific requirements for the occurrence of each of the possible unidirectional exchanges mediated by the Mg2+ exchanger between any combination of the three ionic species transported (i.e. Nao/Mgi; Nai/Mgo; Ko/Mgi; Ki/Mgo; Clo/Mgi; Cli/Mgo; Nao/Ki; Nai/Ko; Nao/Cli; Nai/Clo; Ko/Cli; Ki/Clo and (tracer) Na/Na, Mg/Mg, K/K, and Cl/Cl exchange). These experiments could be performed providing that specific intra and extracellular ionic requirements are met for each exchange mode (76); and

7. Assessment of whether the putative 2Na++2K++2Cl-:1Mg2+ exchanger and the Na+K+Cl co-transporter are either the same protein entity operating in different "modes" of operation or are members of the same gene family. Discussion of this subject merits special attention and is presented below.

6. COMPARISON BETWEEN THE ELECTRO-NEUTRAL Na+K+2Cl (OR 2Na+1K+3Cl, IN THE SQUID) COTRANSPORTER AND THE POSTULATED 2Na+2K+2Cl/1Mg ELECTRONEUTRAL EXCHANGER

Postulation of the putative 2Na++2K++2Cl-:1Mg2+ exchanger may raise the question of whether a protein transporting all these ions could in fact exist in the plasma membrane. Answering this question brings to mind the existence of another, well established and already cloned transporter: the Na+K+Cl cotransporter (NKCC) (reviewed in Ref. 41). This cotransporter operates with a Na+K+2Cl stoichiometry in most cells but with a 2Na+K+3Cl stoichiometry in squid giant axons (77). There are some striking similarities between the putative 2Na++2K++2Cl-:1Mg2+ exchanger and the NKCC including the fact that both transporters:

  • effect the translocation of Na, K, Cl from the extracellular to the intracellular side of the membrane;
  • are electroneutral;
  • can operate in various modes of exchange;
  • involve intracellular Mg2+ (Mgi) for their operation: NKCC is activated by Mgi (independent of ATP) (78) while the 2Na++2K++2Cl-:1Mg2+ exchanger transports this divalent cation; and
  • have an absolute dependence on intracellular ATP

Having so many similarities, the possibility could be raised that both transporters are in fact the same protein moiety with the ability to "switch" modes of operation. Although this possibility cannot be discarded at present, there is an important difference between both transporters: NKCC is highly sensitive to bumetanide (i.e., K0.5= 10-7 M) (54) while the 2Na++2K++2Cl-:1Mg2+ exchanger appears to be much less sensitive to this loop diuretic since it is able to operate in the presence of this concentration of the loop diuretic (53).

For many years it was thought that the Na+Cl cotransporter, the K+Cl cotransporter and the NKCC constituted different "modes" of operation of a single transport moiety. However, it has now been established that all three of these cotransporters are separate proteins encoded by the same gene family (79-81). In fact, it has been proposed that this gene family be termed the cation chloride cotransporter gene family (82). These three cotransporters share numerous characteristics but are somewhat distinguishable by pharmacological means. Bumetanide inhibits with much lower potency the K+Cl cotransporter (i.e., K0.5= 10-4 M) (83) as compared to the NKCC (see above) and does not appear to inhibit the Na+Cl cotransporter (84).

Clearly, it is very tempting to speculate that the putative 2Na++2K++2Cl-:1Mg2+ exchanger and the NKCC are either the same protein entity or are members of the same gene family. Answer to these question will be vigorously pursued.

REFERENCES

1. Ebel, H. and T. Gunther: Magnesium metabolism: A review. J.Clin.Chem.Clin.Biochem. 18:257-270 (1980)

2. Kleckner, N.W. and R. Dingledine: Regulation of hippocampal NMDA receptors by magnesium and glycine during development. Mol.Brain Res. 11:151-159 (1991)

3. Kozlowski, R.Z. and M.L.J. Ashford: ATP-sensitive K+-channel run-down is Mg2+ dependent. Proc.R.Soc.Lond.[Biol.] 240:397-410 (1990)

4. Agus, Z.S. and M. Morad: Modulation of cardiac ion channels by magnesium. Annu.Rev.Physiol. 53:299-307 (1991)

5. Matsuda, H: Magnesium gating of the inwardly rectifying K+ channel. Annu.Rev.Physiol. 53:289-298 (1991)

6. Delpire, E. and P.K. Lauf: Magnesium and ATP dependence of K-Cl co-transport in low K+ sheep red blood cells. J.Physiol.(Lond.) 441:219-231 (1991)

7. Zhang, A., B.T. Altura, and B.M. Altura: Endothelial cells are required for contractile responses induced by reduction in extracellular magnesium and sodium ions in rat aortic smooth muscle. Microcirc.Endoth.Lymphatics 6:427-435 (1990)

8. Altura, B.T., A. Zhang, and B.M. Altura: Sodium-calcium exchange mechanism in vascular smooth muscle tissue as revealed by manipulating external magnesium. Magnesium Trace Elem. 9:163-175 (1990)

9. Zhang, A., A. Carella, B.T. Altura, and B.M. Altura: Interactions of magnesium and chloride ions on tone and contractility of vascular muscle. Eur.J.Pharmacol. 203:223-235 (1991)

10. Chien, M.M., K.E. Zahradka, M.K. Newell, and J.H. Freed: Fas-induced B cell apoptosis requires an increase in free cytosolic magnesium as an early event. J.Biol.Chem. 274:7059-7066 (1999)

11. Quamme, G. and C. De Rouffignac: Epithelial magnesium transport and regulation by the kidney. Frontiers in Bioscience 5:d694-d711 (2000)

12. Schweigel, M. and H. Martens: Magnesium transport in the gastrointestinal tract. Frontiers in Bioscience 5:d666-d677 (2000)

13. Flink, E.B. : Alcoholism, liver disease and magnesium. Magnes.Bull. 3:209-216 (1981)

14. Vink, R. and I. Cernak: Regulation of intracellular free magnesium in central nervous system injury. Frontiers in Bioscience 5:d656-d665 (2000)

15. Paschen, K., M.G. Bachem, and B. Strobel: Magnesium-stoffwechsel beim diabetes mellitus. Magnes.Bull. 3:307-316 (1981)

16. Tosiello, L: Hypomagnesemia and diabetes mellitus. Archives of Internal Medicine 156:1143-1148 (1996)

17. White, J.R. and R.K. Campbel: Magnesium and diabetes: a review. Ann.Pharmacother. 27:775-780 (1993)

18. Altura, B.M. and B.T. Altura: Magnesium ions and contraction of vascular smooth muscles: relationship to some vascular diseases. Federation Proc. 40:2672-2679 (1981).

19. Altura, B.M. and B.T. Altura: Role of magnesium ions in contractility of blood vessels and skeletal muscles. Magnesium 3:102-114 (1981)

20. Fehlinger, R., L. Franke, E. Glatzel, E. Meyer, M. Michalik, S.M. Rapoport, M. Rüstow, Ch. Schulz, and G. Schumann: Klinische studienzur magnesium-behandlung des tetanischen syndroms. Magnes.Bull. 3:298-306 (1981)

21. Chhaparval, B.C., S. Mehta, and S.D. Singh: Meningoencephalitic syndrome (infantile tremore syndrome). Magnesium cum nutritional deficiency syndrome. Magnes.Bull. 5:15-18 (1983)

22. Montes, J.G., R.A. Sjodin, A.L. Yergey, and N.E. Vieira: Simultaneous bidirectional magnesium ion flux measurements in single barnacle muscle cells by mass spectrometry. Biophys.J. 56:437-446 (1989)

23. Romani, A. and A. Scarpa: Regulation of cellular magnesium. Frontiers in Bioscience 5:d720-d734 (2000)

24. DiPolo, R. and L. Beauge: An ATP-dependent Na/Mg countertransport is the only mechanism for Mg extrusion in squid axons. Biochim.Biophys.Acta. 946:424-428 (1988)

25. Baker, P.F. and A.C. Crawford: Mobility and transport of magnesium in squid giant axons. J.Physiol.(Lond.) 227:855-874 (1972)

26. De Weer, P: Axoplasmic free magnesium levels and magnesium extrusion from squid giant axons. J.Gen.Physiol. 68:159-178 (1976)

27. Caldwell-Violich, M. and J. Requena: Magnesium content and net fluxes in squid giant axons. J.Gen.Physiol. 74:739-752 (1979)

28. Mullins, L.J., F.J.Jr. Brinley, S.G. Spangler, and R.F. Abercrombie: Magnesium efflux in dialyzed squid axons. J.Gen.Physiol. 69:389-400 (1977)

29. Leefmans, F.J.A., S.M. Gamiño, and T.J. Rink: Intracellular free magnesium in neurones of helix aspersa measured with ion-selective micro-electrodes. J.Physiol.(Lond.). 354:303-317 (1984)

30. Blatter, L.A: Intracellular free magnesium in frog skeletal muscle studied with a new type of magnesium-selective microelectrode: Interactions between magnesium and sodium in the regulation of [Mg]i. Pflugers Arch. 416:238-246 (1990)

31. Alvarez-Leefmans, F.J., S.M. Gamiño, F. Giraldez, and H. Gonzalez-Serratos: Intracellular free magnesium in frog skeletal muscle fibres measured with ion-selective micro-electrodes. J.Physiol.(Lond.) 378:461-483 (1986)

32. Flatman, P.W. and L.M. Smith: Magnesium transport in magnesium-loaded ferret red blood cells. Pflugers Arch. 432:995-1002 (1996)

33. Yago, M.D., M. Manas, and J. Singh: Intracellular magnesium: transport and regulation in epithelial secretory cells. Frontiers in Bioscience 5:d602-d618 (2000)

34. Beyenbach, K.W: Renal handling of magnesium in fish: from whole animal to brush border membrane vesicles. Frontiers in Bioscience 5:d712-d719 (2000)

35. Ashley, C.C. and J.C. Ellory: The efflux of magnesium from single crustacean muscle fibres. J.Physiol.(Lond.) 226:653-674 (1972)

36. Chandler, R.E., H. Gonzalez-Serratos, and H. Rasgado-Flores: Further evidence for a Na/Mg exchange in excitable cell membranes. Proc.Intl.Union Physiol.Sci. 16:127a (1986)

37. DeSantiago, J., R. Modak, H. Gonzalez-Serratos, and H. Rasgado-Flores: Coupled fluxes of sodium, magnesium and manganese in barnacle muscle cells. Biophys.J. 61:A390 (1992)

38. Hoyle, G: The giant muscle cells of barnacles. In: Barnacle Biology. A.J. Southward, Ed: Balkema A.A, Rotterdam. 213-217 (1987)

39. Montes, J.G., R.A. Sjodin, H. Gonzalez-Serratos, and H. Rasgado-Flores: Evidence for potassium-activated magnesium extrusion in barnacle muscle cells. Biophys.J. 53:344a (1988)

40. Montes, J.G., R.A. Sjodin, Y. Wu, J.-S. Chen, A.L. Yergey, and N.E. Vieira: Regulation of potassium and magnesium effluxes by external magnesium in barnacle muscle fibres. Magnes.Res. 3:239-248 (1990)

41. Russell, J.M: Sodium-Potassium-Chloride Cotransport. Physiol.Rev. 80:211-276 (2000)

42. Nelson, M.T. and M.P. Blaustein: Properties of sodium pumps in internally perfused barnacle muscle fibers. J.Gen.Physiol. 75:183-206 (1980)

43. Brinley, F.J.Jr. and L.J. Mullins: Sodium extrusion by internally dialyzed squid axons. J.Gen.Physiol. 50:2303-2331 (1967)

44. Mullins, L.J. and F.J.Jr. Brinley: Magnesium influx in dialyzed squid axons. J.Membr.Biol. 43:243-250 (1978)

45. Rasgado-Flores, H., E.M. Santiago, and M.P. Blaustein. Kinetics and stoichiometry of coupled Na efflux and Ca influx (Na/Ca exchange) in barnacle muscle cells. J.Gen.Physiol. 93:1219-1241 (1989)

46. Rasgado-Flores, H. and M.P. Blaustein: Na/Ca exchange in barnacle muscle cells has a stoichiometry of 3 Na/1 Ca. Am.J.Physiol. 252 (Cell Physiol.21):C499-C504 (1987)

47. Espinosa-Tanguma, R., J. DeSantiago, and H. Rasgado-Flores: a -Chymotrypsin deregulation of the sodium-calcium exchanger in barnacle muscle cells. Am.J.Physiol. 265 (Cell Physiol.34):C1128-C1137 (1993)

48. Rasgado-Flores, H., R. Espinosa-Tanguma, J. Tie, and J. DeSantiago: Voltage-dependence of Na/Ca exchange in barnacle muscle cells: I. Na-Na exchange activated by a-chymotropsin. Ann.NY Acad.Sci. 779:236-248 (1996)

49. DiPolo, R. and L. Beauge: Characterization of the reverse Na/Ca exchange in squid axons and its modulation by Cai and ATP. J.Gen.Physiol. 90:505-525 (1987)

50. Gonzalez-Serratos, H. and H. Rasgado-Flores: Extracellular magnesium-dependent sodium efflux in squid giant axons. Am.J.Physiol. 259 (Cell Physiol.28):C541-C548 (1990)

51. Scarpa, A. and F.J.Jr. Brinley: In situ measurements of free cytosolic magnesium ions. Federation Proc. 40:2646-2652 (1981)

52. Shirazi, R., E. Dessner, E. Kim, C. Pena-Rasgado, and H. Rasgado-Flores: Second messenger stimulation of Na/Mg exchange in skeletal muscle. FASEB J. 13:126 (1999)

53. Rasgado-Flores, H., H. Gonzalez-Serratos, and J. DeSantiago: Extracellular Mg-dependent Na, K, and Cl efflux in squid giant axons. Am.J.Physiol. (Cell Physiol. 266):C1112-C1117 (1994)

54. Altamirano, A.A. and J.M. Russell: Coupled Na/K/Cl Efflux "Reverse" unidirectional fluxes in squid giant axons. J.Gen.Physiol 89:669-686 (1987)

55. Blaustein, M.P. and W.J. Lederer: Sodium/Calcium exchange: Its Physiological implications. Physiol.Rev. 79:763-854 (1999)

56. Shirazi, R., R. Modak, J. Tie, P.P. Yee, J. DeSantiago, C. Pena-Rasgado, H. Gonzalez-Serratos, and H. Rasgado-Flores: Modulation of extracellular Mg2+ (Mgo)-dependent Na+ efflux in skeletal muscle. FASEB J. 12:A548 (1998)

57. Hilgemann, D.W: Cytoplasmic ATP-dependent regulation of ion transporters and channels: Mechanisms and messengers. Annu.Rev.Physiol. 59:193-220 (1997)

58. Hilgemann, D.W. and R. Ball: Regulation of cardiac Na+,Ca2+ exchange and KATP potassium channels by PIP2. Science 273:956-959 (1996)

59. DiPolo, R. and L. Beaugé: Phosphoarginine stimulation of Na+-Ca2+ exchange in squid axons- a new pathway for metabolic regulation? J.Physiol.(Lond.). 487.1:57-66 (1995)

60. Flatman, P.W: Regulation of potassium transport by magnesium. In Regulation of potassium transport across biological membranes. L. Reuss and J.M. Russell, editors. University of Texas, Austin. 191-212 (1990)

61. Flatman, P.W: Mechanisms of magnesium transport. Annu.Rev.Physiol. 53:259-271 (1991)

62. Ellory, J.C., P.W. Flatman, and G.W. Stewart: Inhibition of human red cell sodium potassium transport in ferret and potassium by divalent cations. J.Physiol.(Lond.) 340:1-17 (1983)

63. Flatman, P.W: The effects of magnesium on potassium transport in ferret red cells. J.Physiol.(Lond.). 397:471-487 (1988)

64. Cushner, H.M., T.P. Peller, T. Fried, and C.S. Delea: Does magnesium play a role in the hypokalemia of Bartter's syndrome. Am.J.Kidney Dis. 16:495-500 (1990)

65. Borella, P., A. Giardino, M. Neri, and E. Andermarker: Magnesium and potassium status in elderly subjects with and without dementia of the Alzheimer type. Magnes.Res. 3:283-289 (1990)

66. Russell, J.M. and M.S. Brodwick: The interaction of intracellular Mg and pH on Cl fluxes associated with intracellular pH regulation in barnacle muscle fibers. J.Gen.Physiol 91:495-513 (1988)

67. Gunther, T. and J. Vormann: Removal and reuptake of intracellular magnesium. Magnes.Bull. 2:66-70 (1985)

68. Kaup, S.M. and J.L. Greger: Effect of various chloride salts on the utilization of phosphorus, calcium, and magnesium. J.Nutr.Biochem. 1:542-548 (1990)

69. Armstrong, C.M: Interaction of tetraethylammonium ion derivatives with the potassium channels of giant axons. J.Gen.Physiol. 58:413-437 (1971)

70. Blaustein, M.P: Sodium/calcium exchange and the control of contractility in cardiac muscle and vascular smooth muscle. J.Cardiovasc.Pharmacol. 12 Suppl. 5:S56-S68 (1988)

71. Frankenhaeuser, B. and A.L. Hodgkin: The after-effects of impulses in the giant nerve fibres of Loligo. J.Physiol.(Lond.) 116:449-472 (1956)

72. De Weer, P: Effects on intracellular adenosine-5'-diphosphate and orthophosphate on the sensitivity of sodium efflux from squid axon to external sodium and potassium. J.Gen.Physiol. 56:583-620 (1970)

73. Bear, R.S. and O.F. Schmidt: Electrolytes in the axoplasm of the giant nerve fibers of the squid. J.Cell.Comp.Physiol. 14:205-215 (1939)

74. Steinbach, H.B. and S. Spiegelman: The Na and K balance in squid nerve axoplasm. J.Cell.Comp.Physiol. 22:187-196 (1943)

75. Russell, J.M: Chloride in the squid giant axon. Curr.T.Membr.Transport 22:177-193 (1984)

76. Lytle, C., T.J. McManus, and M. Haas: A model of Na-K-2Cl cotransport based on ordered ion binding and glide symmetry. Am.J.Physiol. (Cell Physiol. 274:C299-C309) (1998)

77. Russell, J.M: Cation-coupled chloride influx in squid axon: role of potassium and stoichiometry of the transport process. J.Gen.Physiol. 81:909-925 (1983)

78. Flatman, P.W: The effects of magnesium on potassium transport in ferret red cells. J.Physiol.(London) 397:471-487 (1988)

79. Gamba, G., S.N. Saltzberg, M. Lombardi, A. Miyanoshita, J. Lytton, M.A. Hediger, B.M. Brenner, and S.C. Hebert: Primary structure and functional expression of a cDNA encoding the thiazide-sensitive, electroneutral sodium-chloride cotransporter. Proc.Natl.Acad.Sci.USA 90:2749-2753 (1993)

80. Gillen, C.M., S. Brill, J.A. Payne, and B. Forbush III: Molecular cloning and functional expression of the K-Cl cotransporter from rabbit, rat and human. A new member of the cation-chloride cotransporter family. J.Biol.Chem. 271:16237-16244 (1996)

81. Park, J.H. and M.H.Jr. Saier: Phylogenetic, structural and funcional characteristics of the Na-K-Cl cotransporter family. J.Membr.Biol. 149:161-168 (1996)

82. Haas, M: Tha Na-K-Cl cotransporters. Am.J.Physiol.Cell Physiol. 267 (Cell Physiol. 36):C869-C885 (1994)

83. Lauf, P.K., J. Bauer, N.C. Adragna, H. Fujise, A.M.M. Zade-Oppen, K.H. Ryu, and E. Delpire: Erythrocyte K-Cl cotransport:properties and regulation. Am.J.Physiol.( Cell Physiol. 263):C917-C932 (1992)

84. Stokes, J.B., I. Lee, and M. D'Amico: Sodium chloride absorption by the urinary bladder of the winter flounder. J.Clin.Invest. 74:7-16 (1984)