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

This Article Abstract Full Text (PDF) All Versions of this Article: 94/1/441    most recent 01221.2004v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via ISI Web of Science (1) Citing Articles via Google Scholar Google Scholar Articles by Günzel, D. Articles by Schlue, W.-R. Search for Related Content PubMed PubMed Citation Articles by Günzel, D. Articles by Schlue, W.-R.

Mg²⁺-Malate Co-Transport, a Mechanism for Na⁺-Independent Mg²⁺ Transport in Neurons of the Leech Hirudo medicinalis

¹Institut für Neurobiologie, Heinrich-Heine-Universität Düsseldorf, Düsseldorf; and ²Institut für Klinische Physiologie, Charité, Berlin, Germany

Submitted 30 November 2004; accepted in final form 18 March 2005

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Mg²⁺-extrusion from Mg²⁺-loaded neurons of the leech, Hirudo medicinalis, is mediated mainly by Na⁺/Mg²⁺ antiport. However, in a number of leech neurons, Mg²⁺ is extruded in the nominal absence of extracellular Na⁺, indicating the existence of an additional, Na⁺-independent Mg²⁺ transport mechanism. This mechanism was investigated using electrophysiological and microfluorimetrical techniques. The rate of Na⁺-independent Mg²⁺ extrusion from Mg²⁺-loaded leech neurons was found to be independent of extracellular Ca²⁺, K⁺, NO₃^–, HCO₃^–, SO₄^2–, HPO₄^2–, and of intra- and extracellular pH. Na⁺-independent Mg²⁺ extrusion was not inhibited by 4,4'-diisothiocyanatostilbene-2,2'-disulfonic acid (DIDS), furosemide, ouabain, vanadate, iodoacetate, 4-amino-hippurate, or -cyano-4-hydroxycinnamate and was not influenced by changes in the membrane potential in voltage-clamp experiments. Na⁺-independent Mg²⁺ extrusion was, however, inhibited by the application of 2 mM probenecid, a blocker of organic anion transporters, suggesting that Mg²⁺ might be co-transported with organic anions. Extracellularly, of all organic anions tested (malate, citrate, lactate, -ketoglutarate, and 4-amino-hippurate) only high, but physiological, concentrations of malate (30 mM) had a significant inhibitory effect on Na⁺-independent Mg²⁺ extrusion. Intracellularly, iontophoretically injected malate, citrate, or fura-2, but not Cl^–, -ketoglutarate, glutamate, succinate, or urate, were stimulating Na⁺-independent Mg²⁺ extrusion from those neurons that initially did not extrude Mg²⁺ in Na⁺-free solutions. Our data indicate that Mg²⁺ is co-transported with organic anions, preferably with malate, the predominant extracellular anion in the leech. The proposed model implies that, under experimental conditions, malate drives Mg²⁺ extrusion, whereas under physiological conditions, malate is actively taken up, driven by Mg²⁺, so that malate can be metabolized.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
In all preparations studied thus far, the free intracellular Mg²⁺ concentration ([Mg²⁺]_i) is well below its electrochemical equilibrium, as calculated from the Nernst equation (for review, see Beyenbach 1990; Bijvelds et al. 1998 Günzel and Schlue 2000; Maguire 1990; Romani and Scarpa 2000). In most preparations, [Mg²⁺]_i is in the order of 0.1 to 1 mM and thus only slightly below the extracellular Mg²⁺ concentration ([Mg²⁺]_o). However, due to the negative membrane potential (E_m) there is a steep inwardly directed gradient providing the main driving force for Mg²⁺ influx. To keep [Mg²⁺]_i at its low level, Mg²⁺ is actively extruded from the cytoplasm.

In animal cells, there is currently no evidence for the existence of a primary active (i.e., ATP using) Mg²⁺ extrusion mechanism (for review, see Romani and Scarpa 2000). While the main Mg²⁺ extrusion mechanism in virtually all animal preparations investigated to date is a form of a Na⁺/Mg²⁺ antiport (for review, see Beyenbach 1990; Romani and Scarpa 2000), additional Na⁺-independent [Mg²⁺]_i-regulating mechanisms exist in various cells. Such mechanisms work either as a cation antiport system or an anion co-transport system.

Three forms of Na⁺-independent cation transport have been described: a Ca²⁺/Mg²⁺ antiport in liver cells of the rat (Cefaratti et al. 1998), where it appears to be restricted to the apical membrane of the cells; a Mg²⁺/choline exchange in rat erythrocytes (Ebel et al. 2002); and a H⁺-driven Mg²⁺ extrusion in form of a 1 Mg²⁺/2 H⁺ antiport in epithelial cells from the colon and cecum of the rat (Scharrer and Lutz 1990) and possibly in the rumen of sheep (Leonhard-Marek et al. 1998; see, however, Schweigel et al. 2000).

Anion co-transport systems based on HCO₃^– and Cl^– are present in a variety of cells. Like the Mg²⁺/H⁺ antiport, a Mg²⁺-HCO₃^– co-transport uses the energy of the pH-gradient across the cell membrane. In rat ventricular myocytes, it has been suggested that the 4,4'-diisothiocyanatostilbene-2,2'-disulfonic acid (DIDS)-sensitive Mg²⁺-transport may be produced by this kind of Mg²⁺-HCO₃^– co-transport system (Ödblom and Handy 1999). A furosemide-sensitive 1 Mg²⁺-2 HCO₃^– co-transport has been found in ascites tumor cells and thymocytes (Günther and Vormann 1985; Günther et al. 1986). The physiological function of this co-transport, however, is believed to be the uptake of Mg²⁺ into Mg²⁺-depleted cells rather than the extrusion of Mg²⁺ from the cell. A DIDS- and 4-acetamido-4'-isothiocyano-stilbene-2-2'-disulfonic acid (SITS)-sensitive Cl^–-dependent co-transport has been described in chicken and mammalian erythrocytes (Günther and Vormann 1990), in rat hepatocytes (Romani et al. 1993), and in the basolateral membrane of fish enterocytes (Bijvelds et al. 1996, 1998). Whereas in the erythrocytes the Cl^–-dependent Mg²⁺ transport coexists with a Na⁺/Mg²⁺ antiport, it seems to be the sole Mg²⁺ extrusion system in the fish enterocytes.

This study was carried out on various types of neurons from the CNS of the medicinal leech, Hirudo medicinalis. In the multifunctional Retzius neurons and the pressure-sensitive P neurons from the CNS of this animal, an electrogenic 1 Na⁺/1 Mg²⁺ antiport has already been characterized (Günzel and Schlue 1996; Hintz et al. 1999; Müller et al. 1997; for review, see Günzel and Schlue 2000), but preliminary studies suggested that an additional Na⁺-independent Mg²⁺ transport mechanism may be present at least in some leech neurons (Hintz et al. 1996, 1997). It was the aim of this study to identify and characterize this Na⁺-independent transport mechanism.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Preparation

Experiments were carried out on neurons of the leech, H. medicinalis. Segmental ganglia from the leech CNS were dissected as described by Schlue and Deitmer (1980). Isolated ganglia were transferred to an experimental chamber and fixed ventral side up by piercing the connectives with insect pins. During all experiments, the experimental chamber (volume 0.2 ml) was continuously superfused with saline at a rate of 5 ml/min (20 chamber volumes per minute). Experiments were carried out at room temperature (20–25°C) unless stated otherwise.

Solutions

Standard leech saline (SLS) contained (in mM) 85 NaCl, 4 KCl, 1 MgCl₂, 2 CaCl₂, and 10 HEPES, pH 7.4, adjusted with 5 mM NaOH, increasing the total Na⁺ concentration to 90 mM. Saline with a Na⁺ content reduced to 45 mM was obtained by an equimolar substitution of Na⁺ with NMDG⁺ (N-methyl-D-glucamine⁺; Sigma, Deisenhofen, Germany), and pH was adjusted with NMDG-OH. In solutions with a Mg²⁺ content of 30 mM, the Na⁺ concentration was kept at 45 mM while NMDG-Cl was replaced with MgCl₂. Thus it was possible to avoid changes in osmolarity and in the Na⁺ gradient across the cell membrane during alterations of [Mg²⁺]_o.

Nominally Na⁺-free saline was obtained by an equimolar substitution of Na⁺ with NMDG⁺. For nominally Ca²⁺-free saline, CaCl₂ was omitted without substitution. To obtain a Ca²⁺-free solution, 5 mM EGTA was added to the nominally Ca²⁺-free solution and the pH adjusted with NMDG-OH. In solutions with reduced K⁺ content, KCl was omitted from the solution without substitution. The K⁺ content was increased by equimolar replacement of NMDG-Cl with KCl.

To manufacture Na⁺-free solutions with changed anion content, NMDG-OH was titrated with the appropriate acid (HNO₃, H₂SO₄, H₃PO₄, malic acid, lactic acid, citric acid, -keto-glutaric acid) to obtain a 0.5 M stock solution. NMDG-Cl was replaced by the respective stock solution to obtain the desired concentration (10 mM for SO₄^2–, HPO₄^2–, lactate, citrate, and -keto-glutarate, 10 or 30 mM for malate) keeping osmolarity (rather than ionic strength) constant. In Cl^– free saline, all Cl^– was replaced by NO₃^– rather than by gluconate, because gluconate is known to bind divalent cations and therefore alters the free Mg²⁺ concentration of the solution.

In HCO₃^–-containing solutions, 4 mM NMDG-Cl was replaced by 4 mM choline-HCO₃, and the solution was continuously bubbled with 1% CO₂ in O₂. At these low concentrations, choline did not affect the signal of the Mg²⁺-selective microelectrode.

All other substances [4-amino-hippurate, -cyano-4-hydroxycinnamate (Fluka); amiloride, CCCP (carbonyl cyanide 3-chlorophenylhydrazone), DIDS, furosemide, iodoacetate, ouabain, probenecid, sodium orthovanadate (Sigma)] were added to the Na⁺-free solution without substitution. Of these substances, amiloride and CCCP had to be predissolved in DMSO (dimethyl sulfoxide; final DMSO concentration 1%) and DIDS had to be predissolved in ethanol (final ethanol concentration 1%). Dissolving of furosemide and probenecid was facilitated by ultrasonication. Na⁺ contamination of the solution by the addition of 0.5 mM sodium orthovanadate was considered tolerable.

Temperature control

Some experiments were carried out in a temperature-controlled set-up. Under these conditions all solutions were warmed up/cooled down to the appropriate temperature by passing through a heat-exchange system the temperature of which was controlled by a Lauda Ecoline RE 104 cooling thermostat. Temperature in the experimental bath was continuously monitored with a digital thermometer (GTH 1200, Greisinger, Regenstauf, Germany) equipped with a miniature wire thermo element (GTF 300, Greisinger), and the thermostat was adjusted to obtain the desired bath temperature (accuracy ±1°C). pH of all solutions was adjusted to pH 7.4 at room temperature. Cooling to 5°C (lowest temperature used) increased pH to about 7.5, whereas heating to 35°C (highest temperature used) decreased pH to about 7.3. These changes had no effect on the Mg²⁺-selective electrodes or on the Na⁺-independent Mg²⁺ transport (see RESULTS).

Ion-selective microelectrodes

The intracellular free concentration of Mg²⁺ ([Mg²⁺]_i), Na⁺ ([Na⁺]_i) and the intracellular pH (pH_i) were measured using ion-selective microelectrodes based on the neutral carriers ETH 5214 (cocktail IIb, Fluka, Buchs, Switzerland), ETH 227 (Fluka), and ETH 1907 (cocktail IIa, Fluka), respectively.

Double-barrelled microelectrodes were pulled from borosilicate glass-capillaries (Theta-style, TGC200-15, Clark, Reading, UK). For triple-barrelled microelectrodes, theta-style glass capillaries (Theta-style, TGC200-15, Clark) were glued to conventional capillaries (GC150-15, Clark) and pulled out to obtain a single tip (Günzel et al. 1997). As previously determined by scanning electron microscopy, tip diameters were on the order of 0.2 (double-barrelled) to 0.3 µm (triple-barrelled; Günzel et al. 1999). All capillaries were silanized as described by Günzel et al. (1997, 1999). After the ion-sensitive barrels were filled with the appropriate sensors, they were backfilled with 100 mM MgCl₂ (Mg²⁺-selective microelectrodes), 100 mM NaCl (Na⁺-selective microelectrodes), or the pH 7.67 calibration solution (pH-sensitive microelectrode). The reference channel was usually filled with 3 M KCl. In a few experiments 3 M KNO₃ +8 mM KCl, 3 M LiAcetate +8 mM KCl, or 0.8 M K₂SO₄ +8 mM KCl were used.

Calibration

Before and after each experiment, microelectrodes were calibrated in solutions with an ionic background mimicking intracellular conditions. Mg²⁺ calibration solutions contained (in mM) 95 KCl, 10 NaCl, 10 HEPES, and 10, 2.5, 0.5, or 0 MgCl₂, added from a 1 M stock solution (Fluka), pH 7.3 adjusted with KOH. Na⁺ calibration solutions were composed of (in mM) 95 KCl, 0.5 MgCl₂, 10 HEPES, and 50, 10, 2.5, or 0 NaCl, pH 7.3 adjusted with KOH. In addition, Ca²⁺ was buffered to a free concentration of about 10^–7 M by adding 0.73 mM CaCl₂ and 1 mM EGTA (calculation of CaCl₂ and EGTA concentrations based on Pershadsingh and McDonald 1980). pH-sensitive microelectrodes were calibrated in two calibration solutions containing an ionic background of (in mM) 80 KCl, 10 NaCl, and 0.5 MgCl₂. The calibrating solution at pH 7.67 was buffered with 10 mM HEPES and at pH 6.22 with 10 mM 2-[N-morpholino]ethanesulfonic acid (MES). pH values were adjusted with equal amounts of KOH.

The potential difference between the ion-sensitive channel and the reference channel was plotted against pIon (–log[Ion]). For Mg²⁺-selective microelectrodes, the resulting calibration curves were fitted with the Nicolsky-Eisenman equation as previously described by Günzel and Schlue (1996). pH electrodes were calibrated only at two different pH values, because the pH sensor used behaves linearly between pH 2 and pH 9.

All calibrations were carried out at the appropriate temperature. The temperature range from 5 to 35°C was tolerated well by Mg²⁺-selective microelectrodes. However, for unknown reasons, the slope of the electrodes (change in potential per pMg-unit) in the linear range of the electrodes showed greater temperature dependence than expected from the Nernst equation (Fig. 1).

View larger version (14K): [in this window] [in a new window]   FIG. 1. Temperature dependence of the slope of Mg2+-selective microelectrodes. Maximum slope () of calibration curves of Mg2+-selective microelectrodes depended linearly on the temperature of calibration solutions (continuous line, linear regression; slope, 0.3 mV/°C). This dependence was stronger than expected from the Nernst equation (dotted line; slope, 0.1 mV/°C).

All potentials were measured against the potential of an extracellular reference electrode (agar bridge containing 3 M KCl and Ag/AgCl cell), using voltmeters with an input resistance of 10¹⁵ (2-channel voltmeter FD223 or Intra 767; WPI, Mauer, Germany, or 4-channel voltmeter; Institute of Electrochemistry, Heinrich-Heine University Düsseldorf, Germany). The actual ionic signals, i.e., the differences between the potentials of the ion-selective channels and the reference channel, were obtained directly by means of the built-in differential amplifier of the voltmeter. The output signals were AD-converted and continuously recorded on a personal computer.

Because the values of the transformed ion concentrations were not normally distributed, means ± SD are always given in connection with pIon as suggested by Fry et al. (1990). Mean ion concentrations ([Ion]) were calculated from the mean pIon values. Similarly, the rate of changes in [Mg²⁺]_i was quantified as pMg/min (Günzel and Schlue 2002). All traces shown in Figs. 2–5 are typical examples of 3 to 10 experiments.

View larger version (27K): [in this window] [in a new window]   FIG. 2. Na+ dependence of Mg2+ extrusion from leech neurons. A and B: leech neurons were Mg2+-loaded by exposure to 30 mM [Mg2+]o. After returning to 1 mM [Mg2+]o, Mg2+ extrusion was evaluated in the presence and nominal absence of extracellular Na+. A: recording of [Mg2+]i from an anterior pagoda (AP) neuron. Decrease in [Mg2+]i after Mg2+ loading of the neuron was completely blocked in a nominally Na+-free solution. B: recording of [Mg2+]i from a Leydig neuron. After Mg2+ loading, [Mg2+]i continued to decrease in a nominally Na+-free solution, indicating that a Na+-independent Mg2+ transport mechanism was present in this cell. C: dependence of the rate of Mg2+ extrusion (relative to the rate in SLS45) on [Na+]o from Leydig neurons was compared with the values previously found in P neurons (Hintz et al. 1999). Whereas Mg2+ extrusion from P neurons was almost completely blocked in the absence of extracellular Na+, the rate of Mg2+ extrusion from Leydig neurons was virtually [Na+]o independent at [Na+]o of 27 mM or lower. D: electromotive force (EMF) for Na+ was calculated from the values given in Table 2, using Eq. 1. (1) There was no difference in the [Na+]o dependence of EMF in P neurons (replotted from Hintz et al. 1999) and Leydig neurons, excluding any secondary effect through changes in [Na+]i or Em on the driving force of the Na+/Mg2+ antiport. E: comparison of the rate of Na+-independent Mg2+ extrusion (relative to the rate of Mg2+ extrusion in the presence of 45 mM Na+) in all leech neurons investigated. Na+-independent Mg2+ extrusion was most prominent in Leydig neurons and almost negligible in AP and Retzius neurons.

View larger version (15K): [in this window] [in a new window]   FIG. 5. Rate of fura-2 extrusion from Retzius neurons, estimated from the fluorescence excited at the isosbestic wavelength of 360 nm (F360). A: Retzius neurons were iontophoretically loaded with fura-2. Fluorescence was excited at 360 nm, the wavelength at which fura-2 fluorescence is independent of the intracellular Ca2+ concentration. F360 decreased exponentially over time [i.e., log(F360) decreased linearly], reflecting both bleaching and loss of fura-2 from the cell. Application of 2 mM probenecid or an increase in [Mg2+]o to 30 mM (duration indicated by horizontal line) reduced the rate of F360 decrease, whereas a reduction of [Mg2+]o to nominally 0 mM caused a slight acceleration. White lines, linear regression of section used for determination of the rate of F360 decrease. B: histogram summarizing changes in rates of decrease in the intracellular fura-2 concentration relative to the rate in Standard leech saline (SLS; 100%, dotted line). All changes were significant (*P < 5%, **P < 1%, Student's t-test).

For iontophoretic injection of organic anions, conventional glass microelectrodes (pulled from GC150-15 borosilicate glass capillaries; Clark) were filled with 100 mM of the appropriate solution and connected to the current injection unit of an electrometer (Intra 767, WPI) via chlorided silver wires. Organic anions were injected by applying continuous currents between –1 and –5 nA. For experiments during which the membrane potential was clamped to a set value, the electrometer was combined with a feedback amplifier (Heinrich-Heine-University Düsseldorf). The reference potential was obtained from the reference channel of the Mg²⁺-selective microelectrode while current was delivered from a conventional glass microelectrode filled with 3 M KCl. Control experiments were carried out on micro-droplets to estimate the amount of malate injected into neurons. To this end, droplets of the 0.5 mM Mg²⁺ calibration solution with a mean diameter of 145 µm were pressure-injected from a glass capillary onto a piece of Parafilm to prevent spreading of the droplet and covered with paraffin oil to prevent evaporation. As previously described, these droplets could be treated like cells (Günzel et al. 1999) so that malate or citrate could be injected iontophoretically while the decrease of the free Mg²⁺ concentration due to Mg²⁺ binding to malate and citrate, respectively, could be recorded with a Mg²⁺-selective microelectrode. Using the recently determined apparent dissociation constants of 15.85 mM for Mg-malate (Günzel et al. 2005) and of 0.332 for Mg-citrate, the respective rates of malate and citrate injection could be determined to be in the order of 80 and 15 fmol/min/nA.

Microfluorimetry

Microfluorimetric measurements with the fluorescent indicator fura-2 (Molecular Probes, Eugene, OR) were carried out to investigate the rate of active fura-2 extrusion from leech neurons. The experimental procedures and the set-up have been previously described in detail (Dierkes et al. 1996; Hochstrate and Schlue 1994; Hochstrate et al. 1995). Fluorescence of the iontophoretically fura-2–loaded cells was excited with a wavelength of 360 nm using a commercial microspectrofluorimeter (Deltascan 4000, Photon Technology International, Wedel, Germany) with an objective of high numerical aperture (Fluor 40 Ph3DL, Nikon). The fluorescence light emitted from the preparation (F₃₆₀) was collected by the objective, filtered through a 510/540-nm barrier filter, and measured by a photon-counting photomultiplier tube. At a wavelength of 360 nm, fura-2 fluorescence is independent of [Ca²⁺]_i. Thus variations in F₃₆₀ reflect changes in the intracellular dye concentration.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Resting values of [Mg²⁺]_i and E_m

Neurons in the segmental ganglia of the medicinal leech can be identified from their position within the ganglion (cf. map by Muller et al. 1981) and from their electrical activity (shape and frequency of action potentials). In this study, Mg²⁺ transport was investigated in the multifunctional Retzius neurons, the sensory P (pressure), T (touch), and N (noxious) neurons, the AE (annulus erector) motoneurons, the neuromodulatory Leydig neurons, and the AP (anterior pagoda) neurons (function unknown).

Resting values for [Mg²⁺]_i and E_m were determined in neurons equilibrated in SLS and are summarized in Table 1. Neurons were Mg²⁺-loaded in a solution containing 30 mM Mg²⁺ for 5–20 min. In this solution, the extracellular Na⁺ concentration ([Na⁺]_o) was reduced to 45 mM to maintain osmolarity. To avoid overlapping effects by concomitant changes in the extra- and intracellular Na⁺ concentration, [Na⁺]_o was reduced to 45 mM (SLS_45Na, Na⁺ replaced by NMDG⁺) 10 min before the increase in the extracellular Mg²⁺ concentration ([Mg²⁺]_o). As shown in Table 1, exposure to SLS_45Na caused no significant changes in [Mg²⁺]_i and E_m (Student's t-test) in any of the neurons.

After neurons were Mg²⁺-loaded by exposing them to 30 mM [Mg²⁺]_o, the bath solution was changed back to SLS_45Na (1 mM [Mg²⁺]_o), and the cells started to extrude Mg²⁺ as judged by the decrease in [Mg²⁺]_i. [Na⁺]_o was reduced to 27, 9, or nominally 0 mM or increased to 67.5 or 90 mM. Osmolarity was maintained by an equimolar replacement of Na⁺ with NMDG⁺. As illustrated in Fig. 2A, Mg²⁺ extrusion from AP neurons was almost abolished in a Na⁺-free solution. In contrast, in Leydig neurons (Fig. 2B), there was an extrusion of Mg²⁺ in Na⁺-free solutions and [Mg²⁺]_i slowly returned to the resting value. To quantify this Na⁺-independent Mg²⁺ extrusion, the rate of the decrease in [Mg²⁺]_i was evaluated as pMg/min (see Günzel and Schlue 2002). To allow comparison between experiments, the relative rate of extrusion, v_rel, was calculated from these values as percentage of the rate at 45 mM [Na⁺]_o. The Na⁺ dependence of v_rel in Leydig neurons was compared with the previously published data from P neurons (Hintz et al. 1999). As shown in Fig. 2C, Leydig neurons and P neurons showed a strikingly different dependence of the rate of Mg²⁺ extrusion on [Na⁺]_o. While in P neurons, as in AP neurons, the rate of Mg²⁺ extrusion decreased to almost zero with decreasing [Na⁺]_o, the rate of Mg²⁺ extrusion from Leydig neurons was virtually [Na⁺]_o independent at [Na⁺]_o of 27 mM or lower. As shown in Fig. 2D, the driving force of the Na⁺/Mg²⁺ antiport calculated from the values of [Na⁺]_i and E_m at the different [Na⁺]_o given in Table 2 (Eq. 1; Fig. 2D) was similar in P and Leydig neurons, so that the [Na⁺]_o independence of Mg²⁺ extrusion in Leydig neurons was not due to any secondary effect on the driving force.

Extending the experiments to T, N, AE, and AP neurons showed that the different types of neurons varied greatly with respect to their ability to extrude Mg²⁺ in the nominal absence of extracellular Na⁺, as shown in Fig. 2E. These findings led to the hypothesis that the additional Na⁺-independent Mg²⁺ extruding mechanism exists at least in Leydig, AE, N, and T neurons.

Na⁺-independent Mg²⁺ extrusion is not affected by inorganic ions

In theory, the observed [Mg²⁺]_i decrease might be due to Mg²⁺ extrusion from the cell and/or to Mg²⁺ uptake into intracellular stores. As the rate of [Mg²⁺]_i decrease was not affected by the application of 5 µM CCCP, a mitochondrial uncoupler, major uptake into mitochondria seemed unlikely and the working hypothesis was adopted that the observed [Mg²⁺]_i decrease truly reflected Mg²⁺ extrusion.

In the following effort to identify the nature of the Na⁺-independent Mg²⁺ extrusion mechanism, the effects of changes in extracellular ion concentrations, changes in E_m, and the application of various transport inhibitors on the rate of Mg²⁺ extrusion were tested in the nominal absence of extracellular Na⁺.

Although to date there is no evidence for the existence of a primary active Mg²⁺ extrusion mechanism (Mg²⁺ pump) in animal cells, the possibility cannot be completely ruled out. However, Na⁺-independent Mg²⁺ efflux from leech neurons could not be inhibited by the application of 0.5 mM ouabain, 0.5 mM vanadate, or 10 mM iodoacetate. While a hypothetical Mg²⁺ pump may not be sensitive to ouabain or vanadat, 10 mM iodoacetate has previously been shown to decrease ATP levels effectively in leech neurons (Frey et al. 1998) and should therefore affect any primary ion pump. Because iodoacetate did not affect Mg²⁺ extrusion, the existence of a Mg²⁺ pump in leech neurons is unlikely.

Secondary Mg²⁺ extrusion mechanisms couple the transport of Mg²⁺ against its electrochemical gradient to the transport of one or more other ion species along their electrochemical gradient. As calculated from the ion gradients across the cell membrane of leech neurons (Günzel and Schlue 2000), the only cation-coupled transport mechanism that would be able to extrude Mg²⁺ in the absence of extracellular Na⁺ is a Ca²⁺/Mg²⁺ antiport. It was found, however, that the rate of Na⁺-independent Mg²⁺ extrusion was independent of the presence of extracellular Ca²⁺, even if Ca²⁺ had been removed by the addition of 5 mM EGTA to a nominally Ca²⁺-free saline (Fig. 3A), ruling out the presence of a Ca²⁺/Mg²⁺ antiport.

View larger version (29K): [in this window] [in a new window]   FIG. 3. Original recordings of [Mg2+]i with Mg2+-selective microelectrodes, showing the effects of changes in Em, Ca2+, Cl–, malate, and probenecid on Na+-independent Mg2+ efflux from Mg2+-loaded leech neurons. During all recordings, neurons were loaded with Mg2+ by exposure to 30 mM [Mg2+]o. After returning to 1 mM [Mg2+]o, Mg2+ extrusion was evaluated in the presence and nominal absence of extracellular Na+. A: Na+-independent Mg2+ extrusion from an AE neuron was not inhibited by exposing the neuron to Ca2+-free saline (Ca2+ buffered with 5 mM EGTA). B: 2-electrode voltage clamp was employed to change the membrane potential of a Leydig neuron from –40 to –20 mV while [Mg2+]i was recorded with a Mg2+-selective microelectrode. Recording shows that Na+-independent Mg2+ efflux was voltage-independent. C: exposure of a Leydig neuron to a nominally Cl–-free solution inhibited the Na+-independent Mg2+ efflux. Intracellular reference electrode was filled with 3 M KCl. D: experiment on a Leydig neuron as in C, except that the intracellular reference electrode was filled with 3 M KNO3 (+8 mM KCl). Here, Na+-independent Mg2+ efflux was even slightly accelerated (compare lines) during the application of a nominally Cl–-free solution. E: Na+-independent Mg2+ efflux from a Leydig neuron was inhibited by the application of 30 mM malate. F: Na+-independent Mg2+ efflux from a T neuron was inhibited by 2 mM probenecid, a blocker of organic anion transporters.

The lack of effect of changes in all major extracellular cations made it more likely that the Na⁺-independent Mg²⁺ extrusion was due to a co-transport with an intracellular anion. Mg²⁺-Cl^– co-transport could be ruled out, because the Cl^– gradient in leech neurons is directed into the cells, so that such a mechanism would cause a Mg²⁺ influx. As bath solutions contained Cl^– as the sole inorganic anion, the gradients of all other anions present in the cytoplasm are outwardly directed and might thus drive Mg²⁺ extrusion. Intracellular inorganic anions that might exist in a sufficiently high concentration to be responsible for Na⁺-independent Mg²⁺-extrusion are phosphate (intracellular concentration about 1–2 mM, Hoeger et al. 1989), bicarbonate, and possibly sulfate. However, the rate of Mg²⁺ extrusion was not affected by the extracellular presence of 10 mM SO₄^2–, 10 mM HPO₄^2–, or 4 mM HCO₃^– (solution equilibrated with 1% CO₂ in O₂). Furthermore, the 1 Mg²⁺-2 HCO₃^– co-transport found in ascites tumor cells and thymocytes (Günther and Vormann 1985; Günther et al. 1986) was reported to be furosemide-sensitive, but 1 mM furosemide did not inhibit Na⁺-independent Mg²⁺ extrusion in leech neurons, making a HCO₃^–-driven Mg²⁺ extrusion even more unlikely. Because the application of a bath solution containing CO₂/HCO₃^– is known to cause an intracellular acidification without affecting the resting [Mg²⁺]_i (Günzel et al. 1997), this again implies that the Na⁺-independent Mg²⁺ extrusion is pH_i independent.

Due to the inwardly directed Cl^– gradient in leech neurons, a Mg²⁺-Cl^– co-transport would not be able to extrude Mg²⁺. However, Cl^– might still be involved in a more complex mechanism. Therefore experiments were carried out during which Cl^– in the bath solution was completely replaced by NO₃^–. Surprisingly, the effects depended on the filling solution of the intracellular reference electrode. When reference electrodes filled with 3 M KCl were used, i.e., if the Cl^– gradient was reversed by keeping [Cl^–]_i high through Cl^– leaking from reference electrode, an inhibition of Na⁺-independent Mg²⁺ extrusion was found (rate of Mg²⁺ extrusion reduced to 7 ± 11%, n = 4, Fig. 3C). However, when the intracellular reference electrode contained 3 M KNO₃ + 8 mM KCl, the rate of Na⁺-independent Mg²⁺ extrusion was even slightly increased (148 ± 46%, n = 5, Fig. 3D). A similar acceleration was observed when the reference electrode was filled with 3 M LiAcetate + 8 mM KCl (2 experiments, both 146%), whereas 0.8 M K₂SO₄ + 8 mM KCl as reference electrolyte had no significant effect (96 ± 44%, n = 10).

To further elucidate the role of Cl^– in the mechanism of Na⁺-independent Mg²⁺ extrusion, the effects of the Cl^– transport inhibitor DIDS was tested, because Cl^–-dependent Mg²⁺ transport systems described in chicken and mammalian erythrocytes (Günther and Vormann 1990), in rat hepatocytes (Romani et al. 1993), and in the basolateral membrane of fish enterocytes (Bijvelds et al. 1996, 1998) have been found to be DIDS- and SITS-sensitive. However, Na⁺-independent Mg²⁺ extrusion was not inhibited by the application of 1 mM DIDS.

Na⁺-independent Mg²⁺ extrusion is affected by organic anions

Under physiological conditions, leech blood contains high concentrations of malate (in the order of 30 mM, Hoeger et al. 1989). Initial indications that organic anions might be involved in the Na⁺-independent Mg²⁺ extrusion from leech neurons originated from experiments in which a solution containing 30 mM malate was applied. Under these conditions, an inhibition of the Na⁺-independent Mg²⁺ extrusion was observed (Fig. 3E: the rate of Mg²⁺ extrusion was reduced to 12 ± 13%, n = 4). However, the Cl^– concentration of this solution had to be reduced to 25 mM, so it was not clear whether the observed effect was due to the reduction in Cl^– or to the presence of malate, especially because 10 mM malate, 10 mM lactate, and 10 mM citrate were all ineffective.

To test, whether Na⁺-independent Mg²⁺ extrusion was indeed affected by organic anions rather than by changes in extracellular Cl^–, we used probenecid, an inhibitor of organic anion transporters (Munsch and Deitmer 1995). The application of 2 mM probenecid caused a complete but reversible inhibition of the Na⁺-independent Mg²⁺ extrusion (Fig. 3F). In several cases, [Mg²⁺]_i actually increased in the presence of probenecid, possibly due to Mg²⁺ leaking into the cells, since a direct effect of probenecid on the Mg²⁺-selective microelectrode was ruled out in control experiments. In contrast to probenecid, the application of 1 mM -cyano-4-hydroxycinnamate, 5 mM 4-amino-hippurate, or 10 mM -keto-glutarate did not affect Na⁺-independent Mg²⁺ extrusion.

Activation of Na⁺-independent Mg²⁺ extrusion by injection of organic anions

The use of polyvalent organic anions such as malate and citrate imposed problems, because it is not possible to manufacture solutions containing these anions and at the same time keep the extracellular Cl^– concentration, osmolarity, and ionic strength of the solutions constant. Therefore the strategy was reversed, and Retzius neurons were chosen, which under our standard experimental protocol (observing Mg²⁺ extrusion from Mg²⁺-loaded cells in a nominally Na⁺-free solution at room temperature) had not exhibited any Na⁺-independent Mg²⁺ extrusion and which are large enough to be penetrated simultaneously by two electrodes. In these neurons an attempt was made to activate the Na⁺-independent Mg²⁺ extrusion by the iontophoretic injection of various organic anions. As shown in Fig. 4, this scheme was indeed successful: Na⁺-independent Mg²⁺ extrusion could be activated by the iontophoretic injection of malate (Fig. 4A) and citrate, whereas the injection of Cl^– (Fig. 4B), -keto-glutarate (Fig. 4C), succinate, glutamate, and urate had no effect.

View larger version (33K): [in this window] [in a new window]   FIG. 4. Stimulation of Na+-independent Mg2+ efflux from Retzius neurons by the injection of organic anions. Experiments shown in A–D are original recordings with Mg2+-selective microelectrodes. In all cases, Retzius neurons were loaded by exposure to 30 mM [Mg2+]o. After returning to 1 mM [Mg2+]o the extracellular Na+ was removed (nominally 0 mM Na+) and Na+-independent Mg2+ extrusion evaluated as the rate of [Mg2+]i decrease. A: left: under control conditions, [Mg2+]i remained almost constant in the nominal absence of extracellular Na+, indicating that under these conditions no Na+-independent Mg2+ efflux was detectable. Right: same cell was subsequently injected iontophoretically with malate. Under these conditions, [Mg2+]i continued to decrease in the absence of extracellular Na+. Gap: 40 min. This decrease could not be explained by mere Mg2+ buffering through malate. The maximum decrease in [Mg2+]i expected from buffering was calculated from Eq. 2 and 3 and is depicted as a dotted line (2) (3) [Mg]t, [Malate]t, total intracellular magnesium, and malate concentration, respectively. Kd, apparent dissociation constant, 15.85 mM. B: during a control injection of Cl–, Mg2+ efflux is inhibited in the nominal absence of extracellular Na+, indicating that Na+-independent Mg2+ efflux cannot be stimulated by an iontophoretic injection of Cl–. C: iontophoretic injection of -ketoglutarate was not effective in the stimulation of Na+-independent Mg2+ efflux. D: iontophoretic injection of fura-2 stimulated Na+-independent Mg2+ efflux, i.e., Mg2+ efflux was observed in the nominal absence of extracellular Na+. E: histogram summarizing the effect of iontophoretic injection of various ions on Na+-independent Mg2+ efflux. There was a significant stimulation of Na+-independent Mg2+ efflux during the iontophoretic injection of malate, citrate, and fura-2, but not during the injection of chloride, -ketoglutarate, succinate, glutamate, and urate (bars, mean + SD; white, Na+-independent Mg2+ efflux before the injection; gray, during the injection of the respective substance; *significant increase in Na+-independent Mg2+ efflux, paired t-test, P < 0.05). Experiments were carried out between the months June and March and are shown in chronological order.

The results from all injection experiments are summarized in Fig. 4E and support the hypothesis that Mg²⁺ is extruded together with organic anions by a transport mechanism that accepts malate, citrate, and fura-2 but not -keto-glutarate, succinate, glutamate, or urate. In Fig. 4E, experiments are shown in chronological order (extending from June to March). Interestingly, during the course of these experiments, the basal rate of Na⁺-independent Mg²⁺ extrusion changed (white bars). Whether this indicates a seasonal variation or differences in the feeding status of the animals remains unknown.

Fura-2 extrusion inhibited by high [Mg²⁺]_o and probenecid

The action of a Mg²⁺ organic anion co-transport should not only extrude Mg²⁺ from the neurons in an organic anion dependent manner, but should conversely remove organic anions from the cytoplasm in a Mg²⁺-dependent manner. However, neither malate nor citrate concentration could be monitored during such experiments. In contrast, changes in fura-2 concentration can be easily measured microfluorometrically by recording changes in fluorescence excited at the isosbestic wavelength of 360 nm. This method was adopted to investigate fura-2 extrusion from Retzius neurons. As expected from Munsch and Deitmer (1995), application of 2 mM probenecid inhibited fura-2 extrusion from Retzius neurons (reduction of the extrusion rate to 29 ± 11%, n = 4; P = 0.04, paired t-test). In these experiments, the effect of probenecid was only partially reversible (recovery to 45 ± 9% of the initial value). Furthermore, the rate of fura-2 extrusion was reduced by the application of 30 mM [Mg²⁺]_o (reduction of the extrusion rate to 73 ± 8%, n = 4; P = 0.003, paired t-test). This reduction was fully reversible (recovery to 97 ± 10% of the initial value). A decrease in [Mg²⁺]_o to nominally 0 mM caused a slight increase in fura-2 extrusion (increase of the extrusion rate to 115 ± 3%, n = 3; P = 0.05, paired t-test), indicating that the fluxes of these ions may indeed be coupled (Fig. 5).

Effect of malate on pH_i in the presence and absence of [Mg²⁺]_o

Under physiological conditions, leech blood contains high concentrations of malate (Hoeger et al. 1989), so the organic anion-coupled Mg²⁺ transport should mediate a Mg²⁺ driven malate uptake into leech neurons. Because we had no direct method to show Mg²⁺-dependent malate uptake into leech neurons, we attempted an indirect verification of this hypothesis by measuring intracellular pH using pH-sensitive microelectrodes. Munsch et al. (1995) had observed small intracellular alkalinizations in solutions containing high malate concentrations. In these experiments, it was assumed that the postulated Mg²⁺-malate co-transport should transport malate as the unprotonated form, malate^2–, and that it should only be taken up into the cells if Mg²⁺ was present in the extracellular medium. In the cell, a proportion of the entering malate should be protonated according to its pK value, thus causing a small intracellular alkalinization. [Mg²⁺]_i would not be expected to change significantly, because the Mg²⁺ entering the cell should be buffered in the cytoplasm and removed by Na⁺/Mg²⁺ antiport.

Application of a bath solution containing 30 mM malate and 2.4 mM Mg²⁺ (1 mM free Mg²⁺; Günzel et al. 2005) caused a reversible intracellular alkalinization by 0.15 ± 0.08 pH units (n = 45). This alkalinization was significantly reduced in the nominal absence of extracellular Mg²⁺ to 0.08 ± 0.06 pH units (n = 43, P < 1%, unpaired Student's t-test). Thus the Mg²⁺ and malate-dependent alkalinization amounted to 0.07 pH units.

Although these results appear to support the hypothesis that the movement of organic anions and Mg²⁺ across the cell membrane are coupled, some doubts remain. At physiological extra- and intracellular pH values of about 7.4 and 7.2, respectively, most of the malate, as judged from its pK values (4.71 and 3.24, Martell and Smith 1977) should be in the form of malate^2–, with only 0.2–0.3% being present as H-malate^– and even less as H₂-malate. Malate entering the cell through Mg²⁺-malate co-transport would be expected to be in the form of malate^2– but should redistribute in the cytoplasm, leading to an intracellular alkalinization, the total decrease in H⁺ concentration being equal to the amount of H-malate^– formed during this process. The cytoplasmic pH buffer capacity (total change in H⁺ concentration/observed change in pH_i) in leech neurons has been determined by Schlue and Thomas (1985) to amount to 16.7 mM. The observed Mg²⁺- and malate-dependent alkalinization of 0.07 pH units should therefore correspond to the formation of about 1 mM H-malate^– and thus to the influx of 500 mM malate^2–, an unrealistically high value. The major part of the observed intracellular alkalinization has therefore to be considered a secondary effect of malate uptake (e.g., through an activation of metabolic processes).

Estimate of intracellular Mg²⁺ buffering by organic anions

Before a Mg²⁺-organic anion co-transport mechanism could be suggested for leech neurons, it had to be excluded that the decrease in [Mg²⁺]_i observed during the injection of organic anions was simply due to increasing the intracellular Mg²⁺ buffering capacity. It has recently been shown in titration experiments that both malate and citrate bind Mg²⁺ at physiological pH values (apparent dissociation constants 15.85 and 0.332 mM, respectively; Günzel et al. 2005). Fura-2, according to Grynkiewitz et al. (1985), has an apparent dissociation constant for Mg²⁺ of 9.8 mM. It was therefore attempted to estimate intracellular malate buffering from the rate of malate injection, namely 80 fmol/min/nA as measured in electrolyte droplets (see METHODS). Injection of organic anions was always started before the neurons were exposed to the bath solution containing 30 mM Mg²⁺ for intracellular Mg²⁺ loading (see Fig. 4), so that the injection had always been going on for 15–20 min before the cells were exposed to the Na⁺-free solution. At an injection current of –3 nA, the amount of malate injected into the neuron is on the order of 3.6 pmol. Retzius neuron diameter has previously been determined to be 81.4 µm (Günzel et al. 2001), so malate should have accumulated to concentrations of 12–16 mM. Taking Fig. 4A as an example, malate had been injected for about 20 min and [Mg²⁺]_i had reached a value of 1.3 mM when the neuron was exposed to the Na⁺-free solution, indicating that the total Mg²⁺ concentration in the neuron would be 2.6 mM, if malate were the sole intracellular Mg²⁺ buffer (Eq. 2; Fig. 4). During the 7-min exposure, [Mg²⁺]_i decreased to 0.7 mM. As indicated by the dotted line in Fig. 4A, the increase in malate of 6 mM during this time period should only cause a [Mg²⁺]_i decrease to about 1.15 mM (Eq. 3; Fig. 4). In view of the fact that malate is by no means the sole buffer in Retzius neurons (Günzel et al. 2001), this estimate can be considered a worst case scenario. It thus can be ruled out that the observed decrease in [Mg²⁺]_i is only due to increasing the intracellular Mg²⁺ buffering capacity during the injection of organic anions.

We therefore conclude that leech neurons possess a Mg²⁺-organic anion co-transport mechanism.

Temperature dependence of Mg²⁺ transport

In Retzius neurons, the contradictory evidence remained that Na⁺-independent Mg²⁺ efflux could not be observed during the removal of extracellular Na⁺, but only when additionally stimulated by the injection of organic anions. We therefore attempted to prove its general presence in Retzius neurons by stimulating the transport at higher temperatures. To this end, we investigated the temperature dependence of [Mg²⁺]_i increases and decreases (expressed as change in pMg/min, cf. Günzel and Schlue 2002) over the range of 5 to 35°C in steps of 5°. The rate of change in [Mg²⁺]_i induced by an increase in [Mg²⁺]_o to 30 mM was found to be independent of temperature between 5 and 25°C (ANOVA, P = 0.79) but approximately doubled at 35°C (Fig. 6A). The rate of Na⁺-dependent Mg²⁺ extrusion was temperature-dependent (Fig. 6A) over the whole temperature range investigated. The Arrhenius plot (Fig. 6B) revealed two different levels of activation energy above and below 25°C for this transport. The rate of Na⁺-independent Mg²⁺ extrusion was so low that it was not significantly different from 0 over most of the temperature range tested. Only at 35°C could a significant (Student's t-test, P < 0.01) Na⁺-independent Mg²⁺ extrusion be detected. From this result, we conclude that Na⁺-independent Mg²⁺ extrusion is indeed present in Retzius neurons but that its activity at room temperature was too low to be reliably detected.

View larger version (18K): [in this window] [in a new window]   FIG. 6. Effect of temperature on Mg2+ transport in Retzius neurons. A: temperature dependence of the rate of [Mg2+]i increase {, expressed as pMg/min = –log([Mg2+]i)/min and therefore negative } during the exposure of Retzius neurons to 30 mM [Mg2+]o, and the subsequent decrease (positive values) in the presence () and nominal absence () of extracellular Na+ (1 mM [Mg2+]o). Each data point is the mean of 5–16 experiments. B: Arrhenius-plot of the absolute values shown in A. Whereas the values for [Mg2+]i increase () and [Mg2+]i decrease in nominally 0 mM [Na+]o could be fitted with one straight line, [Mg2+]i decrease in the presence of extracellular Na+ showed a break at 20°C, indicating that the activation energy needed for Na+/Mg2+ antiport decreases by a factor of 4 at temperatures >20°C. Values in the figure are apparent activation energies derived from the slopes of the linear regression lines. Each data point is the mean of 5–16 experiments. C: [Mg2+]i of Retzius neurons equilibrated in SLS was temperature-independent in the range of 10–35°C (ANOVA, P = 0.33). Each data point is the mean of 14–21 experiments.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Na⁺-independent Mg²⁺ transport system

It was the aim of this study to characterize the Na⁺-independent Mg²⁺ transport system observed in leech neurons. During this attempt, various types of transport systems previously found in different preparations (human, rat, and chicken erythrocytes, Ebel et al. 2002; Günther and Vormann 1990; rat hepatocytes, Cefaratti et al. 1998; rat ventricular myocytes, Ödblom and Handy 1999; rat distal colon and caecum, Scharrer and Lutz 1990; ascites tumor cells, Günther et al. 1986; fish enterocytes, Bijvelds et al. 1996) could be ruled out. These systems comprise Ca²⁺/Mg²⁺ antiport, Mg²⁺/H⁺ antiport, Mg²⁺-HCO₃^– co-transport, DIDS- and SITS-sensitive Cl^–-dependent Mg²⁺ transport, and K⁺-dependent Mg²⁺ extrusion. Detailed investigations as to whether the Na⁺-independent Mg²⁺-extrusion from leech neurons might be explained by the recently suggested Mg²⁺/choline exchanger found in rat eythrocytes (Ebel et al. 2002) could not be carried out, because choline interferes at the Mg²⁺-selective microelectrode (McGuigan et al. 1993). The presence of such an antiport seemed unlikely, because it would depend on the presence of extracellular choline. Even if no choline was added to the bath solutions, it cannot be ruled out that there might be a small amount of endogenous choline from neighboring cholinergic neurons. Indirect evidence against the presence of a Mg²⁺/choline exchanger, however, is based on the experiments carried out in HCO₃^–-containing solutions that, in addition, contained 4 mM choline (see METHODS). This choline concentration was too low to cause interference at the Mg²⁺-selective electrode and no effect on Na⁺-independent Mg²⁺ extrusion was observed.

There was also no evidence for a Mg²⁺ pump or for a contribution of mitochondria, which are considered to be a major intracellular Mg²⁺ store.

Organic anion-coupled Mg²⁺ transport system

The involvement of organic anions in Na⁺-independent Mg²⁺ transport was deemed possible for three reasons. First blood of the medicinal leech is known to have a very high content of organic anions, especially the divalent anion malate (Hoeger et al. 1989). Second, it has been known for some time that leech neurons are able to actively extrude the organic anion fura-2 in a probenecid-sensitive manner (Munsch and Deitmer 1995). Third, Mg²⁺-dependent organic anion transporters are known to exist in bacteria (Mg²⁺-citrate transport CitM; Warner and Lolkema 2002) and plant vacuoles (Mg²⁺-dependent malate uptake; Dietz et al. 1990).

The results presented here support the hypothesis that Mg²⁺ transport is coupled to the transport of organic anions in leech neurons. 1) High extracellular malate-concentrations block Na⁺-independent Mg²⁺ extrusion from Mg²⁺-loaded cells. 2) Injection of organic anions (malate, citrate, fura-2) stimulates Na⁺-independent Mg²⁺ extrusion and Mg²⁺ buffering through the injected anions is found to be too small to account for the observed [Mg²⁺]_i decrease. 3) High [Mg²⁺]_o blocks fura-2 extrusion and low [Mg²⁺]_o stimulates fura-2 extrusion. 4) The organic anion transport blocker probenecid inhibits both Na⁺-independent Mg²⁺ extrusion from Mg²⁺-loaded neurons and fura-2 extrusion from fura-2-loaded neurons.

In contrast to all other leech neurons investigated in this study, Mg²⁺ loaded Retzius and AP neurons did not show any Mg²⁺ extrusion in Na⁺-free solutions at room temperature. However, experiments on Na⁺-independent Mg²⁺ extrusion during the injection of organic anions and at high temperatures showed that Na⁺-independent Mg²⁺ extrusion is also present in Retzius neurons. Thus Na⁺-independent Mg²⁺ extrusion is likely to be generally present in leech neurons. Differences in activity of this transport in different neurons may reflect differences in intracellular organic anion concentrations and thus in cellular metabolism.

Role of Cl^–

Although the presence of a Mg²⁺-Cl^– co-transport could be excluded on grounds of the inwardly directed electrochemical Cl^– gradient alone, Cl^– clearly played a role in Na⁺-independent Mg²⁺ extrusion. As shown in Fig. 3, C and D, Na⁺-independent Mg²⁺ extrusion was inhibited in a Cl^–-free solution, if the Cl^– gradient was reversed by keeping [Cl^–]_i high through Cl^– leaking from a 3 M KCl filled reference electrode. This is in contrast to what would be expected for a Mg²⁺-Cl^– co-transport, which should be stimulated under these conditions. From this observation it was concluded that Mg²⁺ and Cl^– are transported in opposite directions.

In addition, the lack of inhibition of Na⁺-independent Mg²⁺ extrusion in a Cl^–-free solution when using a KNO₃-filled reference electrode indicates that NO₃^– can substitute for Cl^– at the postulated intracellular binding site. Equation 6 (Fig. 7) attempts to describe such a system. In the equation, in analogy to the Nicolsky-Eisenman equation (Amman 1986), a selectivity coefficient for the binding site, K_Cl/NO3, was defined. To explain the results, a K_Cl/NO3 < 1 has to be assumed, indicating a preference of the binding site for Cl^– over NO₃^–.

View larger version (16K): [in this window] [in a new window]   FIG. 7. Proposed mechanism for the Na+-independent Mg2+ transport system in leech neurons. [Mg2+]i at the thermodynamic equilibrium, [Mg2+]eq, of an electroneutral transport system can be derived from the gradients of the involved ions and gives an estimate of the lowest [Mg2+]i that could be reached if the system worked with 100% efficiency and if no other Mg2+ transport system was present. [Mg2+]eq for a 1 Mg2+-1 malate2– co-transport can be calculated from Eq. 4. Such a system is not in agreement with experimental evidence as it cannot explain the observation that Na+-independent Mg2+ extrusion was blocked by extracellular malate concentrations of 30 mM but not of 10 mM (4) [Mg2+]eq for a 1 Mg2+-2 malate2–/2 Cl– transport (transporting 1 Mg2+ ion and 2 malate2– ions in the same and 2 Cl– ions in the opposite direction) is also voltage independent and can be calculated from Eq. 5. This mechanism explains both the observed malate and Cl– dependence of Na+-independent Mg2+ transport. Under standard experimental conditions (0 mM extracellular malate) such a system would extrude Mg2+ ([Mg2+]eq << [Mg2+]i), whereas under physiological conditions (high extracellular malate concentration, low extracellular Cl– concentration), it would mediate a Mg2+-driven malate uptake ([Mg2+]eq > [Mg2+]i) (5) To explain the effect of NO3–, Eq. 5 has to be expanded, defining a selectivity coefficient KCl/NO3 as a measure for the preference of the binding site for Cl– over NO3– (in analogy to the Nicolsky-Eisenman equation; Amman 1986) (6) Calculations were based on the following values: [Mg2+]o 1 mM; [Mg2+]i 0.3 mM; [Cl–]i /[Cl–]o about 1/10 (Müller et al. 2003), [Malate2-]o 30 mM (Hoeger et al. 1989), [Mg2+]i 2–5 mM (minimum and maximum value reported in yeast during cell cycle; Wittmann et al. 2005).

The present voltage-clamp (cf. Fig. 3B) experiments showed that Na⁺-independent Mg²⁺ transport was electroneutral. This would be in agreement with the co-transport of one Mg²⁺ ion together with one malate^–, because according to the pK values of malate (Martell and Smith 1977), >99% of malate at pH values between 7.2 and 7.4 is in the form of malate^2–. However, the observed Na⁺-independent Mg²⁺ extrusion was inhibited at extracellular malate concentrations of 30 mM but not of 10 mM. For a 1:1 Mg²⁺-malate^2– co-transport, such a dependence on the extracellular malate concentration would only be expected if the intracellular malate concentration was >60 mM (calculated from Eq. 4; Fig. 7), which is unlikely. Furthermore, the mechanism would leave the effect of Cl^– unexplained. We therefore postulate that the Na⁺-independent Mg²⁺ transport translocates malate^2–, Mg²⁺, and Cl^– across the cell membrane at a stoichiometry of 2:1:2 as depicted in Fig. 7. At the assumed intra- and extracellular concentrations for Mg²⁺ of 0.3 and 1 mM and for Cl^– of 10 and 100 mM, respectively (Müller et al. 2003), such a transport mechanism would extrude Mg²⁺ at an extracellular malate concentration of 10 mM as long as the intracellular malate concentration remained above 2 mM, whereas it would be inhibited or even reversed at extracellular malate concentrations of 30 mM (calculated from Eq. 5; Fig. 7).

Differences in the activity of Na⁺-independent Mg²⁺ transport between different types of neurons may reflect differences in expression or differences in electrochemical gradients. The latter, however, seems more likely, as the presence of Na⁺-independent Mg²⁺ transport could clearly be shown in Retzius neurons at higher temperatures and by modifying the gradients during the injection experiments.

Physiological implications

Malate is the predominant extracellular anion in the leech (Hoeger et al. 1989), reaching concentrations of almost 30 mM. This may be due to the feeding behavior of leeches, which are able to take up several times their own body weight and then live on this meal for several months. It may also be an adaptation to cope with limited oxygen availability in their habitat (Hildebrandt 1992). Similarly high concentrations in organic anions have also been found in other invertebrates, such as insects (Scholz and Zerbst-Boroffka 1998). Scholz and Zerbst-Boroffka (1998) found very high (>80 meq/l) normoxic malate levels in the hemolymph of a freshwater midge larvae (Chaoborus crystallinus) and assumed an adaptive significance for the hypoxia tolerance of this species.

Under physiological conditions, Na⁺-independent Mg²⁺ transport is most likely to work as a Mg²⁺-driven malate uptake mechanism given that extracellular malate concentrations are sufficiently high, whereas extracellular citrate concentrations are negligible (<0.1 mM; Hildebrandt and Zerbst-Boroffka 1992). Malate is known to be metabolized under hypoxic conditions. Under these conditions, the extracellular malate concentration decreases dramatically, approaching values of about 20 mM in H. medicinalis and of about 10 mM in C. crystallinus, and thus values close to the estimated equilibrium of the postulated transporter (Eq. 5). Extracellular succinate/propionate (hypercapnic conditions) or lactate concentrations (hypocapnic conditions) increase in parallel with the decrease in malate concentration (Hildebrandt 1992; Scholz and Zerbst-Boroffka 1998), but neither succinate nor lactate interfered with Na⁺-independent Mg²⁺ transport in the present experiments. Coupling malate uptake to Mg²⁺ transport may be advantageous, because Mg²⁺ is well buffered, both intra- and extracellularly, so that initially the Mg²⁺ gradient would be expected to remain almost constant and to disintegrate only slowly even over more extended periods of hypoxia.

To our knowledge this is the first report of a transport system in animal cells that, under physiological conditions, uses the Mg²⁺ gradient for the "tertiary active" transport of other ions/compounds (cf. Fig. 7).

	GRANTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS