ABSTRACT
The magnesium handling of freshwater teleost fish is discussed, with an emphasis on the role of branchial, intestinal and renal transport. In response to the eminent threat of constant diffusive losses of minerals such as magnesium, freshwater fish have developed efficient mechanisms for magnesium homeostasis. Magnesium losses are overcome by the uptake of magnesium from the food, making the intestine an important route for magnesium uptake. Some evidence suggests that intestinal magnesium uptake in fish is a regulated, cellular process. The ambient water is an additional magnesium source for fish, implicating the gills as a secondary route for magnesium uptake. Certainly, in some species, direct uptake from the water, probably via branchial routes, ameliorates the effects of a low-magnesium diet. The hard tissues, representing over 50 % of the total body magnesium pool, form a reservoir from which magnesium can be recruited to perform its functions in the cellular metabolism of soft tissues such as muscle. In fish, as in terrestrial vertebrates, the balance of a variety of elements becomes disturbed when the magnesium homeostasis of the soft tissues is disrupted. However, fish appear to be less sensitive than terrestrial vertebrates to these perturbations. Magnesium is reabsorbed in the kidneys to minimise losses. For renal cells, part of a cellular pathway has been elucidated that would allow absorptive magnesium transport (a magnesium conductive pathway in renal brush-border membranes). In some euryhaline teleosts, the kidneys appear to switch instantaneously to rapid magnesium secretion upon magnesium loading, a response common to marine fish that are threatened by diffusive magnesium entry. This enigmatic mechanism underlies the capacity of some euryhaline species to acclimate rapidly to sea water. Despite the progress made over the last decade, much of the cellular and molecular basis of magnesium transport in the gills, intestine and kidneys remains obscure. The application of fluorescent, radioactive and molecular probes, some of which have only recently become available, may yield rapid progress in the field of magnesium research.
Introduction
In fish, as in all vertebrates, magnesium is found mineralised in bony tissues and is ionised (Mg2+), complexed or protein-bound in all tissues. The highest magnesium concentration is found in the hard tissues of the dermal and skeletal bones and in the scales: approximately 75 and 60 µmol g−1 fresh mass, respectively (Lutz, 1972; Weiss and Watabe, 1978; Cameron, 1985; Van der Velden et al. 1989; Bijvelds et al. 1996a). Depending on fish mass, 50 % to over 70 % of the total magnesium pool is located in these hard tissues (Van der Velden et al. 1989; Bijvelds et al. 1996a). The magnesium pool of the bones and scales may be used as a reserve to maintain normal magnesium levels in soft tissues when the magnesium intake is low (Cowey et al. 1977; Reigh et al. 1991; Bijvelds et al. 1996a). Of the remaining magnesium pool in the soft tissues, only a small percentage is found in the extracellular fluid. The total magnesium concentration of blood plasma in most cases does not exceed 2 mmol l−1, and the ionic concentration is nromally less than 1 mmol l−1. The largest portion of the soft tissues is made up by muscle tissue. In muscle, the magnesium concentration is 10–15 µmol g−1 fresh mass, accounting for approximately 20 % to the total magnesium pool of the body (Knox et al. 1981; Van der Velden et al. 1989; Bijvelds et al. 1997a). In the cell, magnesium is one of the most abundant cations; it is stored in intracellular compartments or bound to proteins and the phosphate groups of adenosine nucleotides. The ionic magnesium level in the cytoplasm is kept relatively low, i.e. in the submillimolar range, typically representing less than 10 % of the total magnesium content of the cell (Heaton, 1993).
It is thought that the level of free, ionic magnesium (Mg2+) is kept at a relatively stable and low value because it is an important determinant of cellular function through its control over catalytic reactions. Mg2+ activates numerous enzymes and also modifies specific enzyme substrates (Heaton, 1990; Black and Cowan, 1995). In this capacity, Mg2+ controls, among other systems, energy metabolism and protein synthesis. As it is a necessary cofactor in the transfer of phosphate groups and it may also control ATP-dependent ion pumps. It regulates Na+/K+/Cl− and K+/Cl− symport activity (Flatman, 1993), cation channels (Horie et al. 1987; Hartzell and White, 1989; Pusch et al. 1989), as well as membrane permeability and the insertion of proteins into membranes (Ebel and Günther, 1980). Intracellular free Mg2+ is carefully buffered, and changes in the cytoplasmic free Mg2+ level occur only slowly and are moderate (Grubbs and Maguire, 1987; Romani and Scarpa, 1992). It has been suggested that Mg2+ is a second messenger for hormone action, with the slow and minute changes in intracellular free [Mg2+] coordinating and fine-tuning the long-term cellular output of hormones (Rubin, 1976; Maguire, 1990).
Because magnesium plays such an important role in cell metabolism, and intracellular and extracellular Mg2+ levels are maintained within narrow limits, vertebrates must have developed effective mechanisms by which Mg2+ is transported, stored and its concentration regulated. As yet, the basis of these cellular mechanisms has not been fully elucidated. The mechanism of transmembrane Mg2+ transport is poorly understood (Flatman, 1984, 1991), and the transport of Mg2+ across intestinal and kidney epithelia, key processes to the understanding of Mg2+ regulation, is still a very enigmatic topic. Freshwater fish are threatened by diffusive losses of Mg2+ across the body surfaces because the magnesium concentration of fresh water is typically well below 0.5 mmol l−1. These losses have to be compensated for by Mg2+ uptake from the food and from the water. Renal excretion must be limited by reabsorption of filtered Mg2+. This review outlines the progress that has been made in our understanding of Mg2+ regulation in freshwater teleost fish and highlights the specific roles of the intestine, gills and the kidneys in the exchange of magnesium between the fish and the environment.
Role of magnesium in diet and water
Freshwater fish depend primarily on diet for their magnesium. Optimal growth is usually achieved when the magnesium concentration of the food is between 15 and 20 µmol g−1 (Ogino et al. 1978; Gatlin et al. 1982; Shim and Ng, 1988). Prolonged feeding with diets of a lower magnesium content may lead to a decrease in growth rate, magnesium depletion of the tissues, muscle dysfunction, neurological disorders and a high mortality.
The function of water-borne magnesium has been less well established. Early life stages of carp Cyprinus carpio depend on magnesium in the water. Although the eggs, and later the yolk, contain significant amounts of magnesium, a low concentration in the water (<0.01 mmol l−1) leads to a decrease or arrest of the magnesium uptake by eggs, an increased mortality and higher levels of deformation, embolism and tissue necrosis (Van der Velden et al. 1991c). Shearer and Asgard (1992) found that the dietary magnesium requirement of rainbow trout Oncorhynchus mykiss decreased when sufficient magnesium was available in the water. Others have also reported that freshwater fish, when fed a diet with a low magnesium content, take up magnesium from the water (Dabrowska et al. 1991; Bijvelds et al. 1996a). However, for all species investigated, this additional magnesium uptake from the water is insufficient to compensate fully for a low dietary intake.
Several studies showed that magnesium deficiency in fish is accompanied by disturbances in the balance of other important minerals such as calcium, potassium and sodium (Ogino and Chiou, 1976; Cowey et al. 1977; Ogino et al. 1978; Knox et al. 1981; Shim and Ng, 1988; Dabrowska et al. 1991; Van der Velden et al. 1991d, 1992a). It is plausible that this relationship reflects the dependence of cellular ion transport mechanisms, such as the Na+/K+-ATPase (Rude, 1989), the Na+/K+/Cl− symporter (Flatman, 1988) and cation channels (Horie et al. 1987; Hartzell and White, 1989; Pusch et al. 1989; Matsuda, 1991; Dorop and Clausen, 1993), on magnesium. Alternatively, magnesium may influence ion movement across cellular membranes through its action on membrane permeability (Tidball, 1964; Ebel and Günther, 1980).
A low dietary magnesium intake induced high body calcium levels in rainbow trout (Cowey et al. 1977), tilapia (Oreochromis niloticus and O. mossambicus) (Dabrowska et al. 1989; Bijvelds et al. 1997a) and guppy Poecilia reticulata (Shim and Ng, 1988). Mg2+ affects the permeability of the intestinal epithelium to ions (Tidball, 1964; Fordtran et al. 1985), and a low luminal Mg2+ concentration will therefore increase the epithelial permeability to ions, possibly stimulating paracellular Ca2+ absorption (Ebel and Günther, 1980; Karbach and Feldmeier, 1991). It has also been suggested that a low magnesium content of the intestine may enhance Ca2+ absorption because these divalent cations compete for one transport pathway (Karbach and Rummel, 1990), although it is now becoming more accepted that these cations are absorbed via separate routes (Hardwick et al. 1990b; Kayne and Lee, 1993). Magnesium deficiency has also been related to calcification of the heart, blood vessels and kidneys. Minute perturbations to cellular magnesium homeostasis affect the acid–base regulation of the cells involved in bone formation, and this may lead to supersaturation of plasma calcium, resulting in spontaneous calcification processes in the soft tissues (Driessens et al. 1987). In line with this hypothesis, calcification of renal tissue occurred in magnesium-deficient rainbow trout (Cowey et al. 1977).
It has been suggested that, in magnesium-deficient rainbow trout, the increase in muscle Na+ content was due to a decrease in cell water content coupled with an increase in extracellular volume (Cowey et al. 1977; Knox et al. 1981). Such changes in the mineral status and water content of tissues are suggestive of changes in cell membrane permeability that could cause an increased ion turnover. The action of Mg2+ on the fluidity of cellular membranes may underlie this phenomenon (Schoffeniels, 1967; Cowan, 1995). The permeability of osmoregulatory epithelia may also be affected since it has been demonstrated that external magnesium and calcium levels influence branchial epithelial permeability to both water and ions (Potts and Fleming, 1971; Dharmamba and Maetz, 1972; Ogawa, 1974; Isaia and Masoni, 1976; Wendelaar Bonga et al. 1983).
Internally, magnesium may have similar actions on membrane permeability and ion turnover in osmoregulatory organs such as the gills. For instance, in Mozambique tilapia O. mossambicus, reduced access to dietary magnesium caused a proliferation of chloride cells, the principal ion-transporting cells in the branchial epithelium (Bijvelds et al. 1996a). Such changes are indicative of an increased turnover of these cells in the gill epithelium. Renewal of the branchial epithelium may be a response to disturbances in ion transport across the gills, since both the epithelial permeability to ions and water (Ogawa, 1974; Wendelaar Bonga et al. 1983) and the activity of cellular ion transporters (Flatman, 1991) are controlled by Mg2+. A low dietary magnesium intake decreased Na+ influx across the gills of Mozambique tilapia (Van der Velden et al. 1992b). Furthermore, in carp Cyprinus carpio, it has been demonstrated that magnesium deficiency is associated with changes in branchial ion regulation (an increase in opercular chloride cell density and a decrease in branchial Na+/K+-ATPase activity) that coincide with an increased bone sodium content (Van der Velden et al. 1992a).
Interestingly, enhanced prolactin activity preceded the occurrence of magnesium deficiency symptoms in Mozambique tilapia, suggesting that the low dietary magnesium intake triggered a prolactin-mediated response (Van der Velden et al. 1992b). In the stickleback Gasterosteus aculeatus and in the Mozambique tilapia, high ambient magnesium levels were related to a considerable reduction in the activity of prolactin cells (Wendelaar Bonga, 1978; Wendelaar Bonga et al. 1983). Prolactin principally controls the permeability to ions and water of epithelia that are closely involved in osmoregulation (Wendelaar Bonga, 1993). Changes in external or internal magnesium levels apparently evoke a prolactin-mediated response to counteract the effects of magnesium on epithelial permeability. However, not all species maintain such a tight permeability control: in goldfish Carassius auratus and carp, the prolactin cell activity is little affected by high ambient and low dietary magnesium levels, respectively (Olivereau et al. 1987; Van der Velden et al. 1992a).
Intestinal magnesium transport
The mechanism of intestinal Mg2+ absorption has been investigated extensively in terrestrial vertebrates (Ebel, 1990; Hardwick et al. 1990a; Kayne and Lee, 1993). The curvilinear relationship between the rate of Mg2+ absorption and the luminal Mg2+ concentration was interpreted as evidence for carrier-mediated transport (Karbach et al. 1991). Alternatively, it was suggested that these apparent saturation kinetics can also be explained by assuming that high luminal Mg2+ levels decrease the paracellular ion permeability and thus paracellular Mg2+ absorption (Kayne and Lee, 1993). Therefore, because Mg2+ decreases epithelial ion permeability (Tidball, 1964), the question of whether Mg2+ is transported via a transcellular route cannot be answered unequivocally by this type of study. Moreover, the results of several studies on the dependence of Mg2+ absorption on Mg2+ concentration are conflicting and suggest that considerable species differences exist (Hardwick et al. 1990b; Karbach and Feldmeier, 1991). In a classic assay for the demonstration of active transport, which measures the radioactive tracer accumulated in everted gut sacs starting from isotopic equilibrium, active intestinal Mg2+ transport has never been demonstrated (Ross, 1962; Hardwick et al. 1990b).
For rainbow trout (Shearer and Asgard, 1992) and Mozambique tilapia (Bijvelds et al. 1996a), the fractional absorption of Mg2+ is a function of the dietary magnesium intake, suggesting that Mg2+ absorption is a regulated process. More concrete evidence for cellular Mg2+ transport comes from studies showing that the absorptive Mg2+ flux is sensitive to inhibitors of cellular ion transport such as ouabain and bumetanide (Partridge et al. 1987; Van der Velden et al. 1990). In the Mozambique tilapia, transepithelial Mg2+ influx is larger than can be explained by solute-linked transport and is coupled to cellular Na+ transport (Van der Velden et al. 1990).
What drives Mg2+ transport across the basolateral plasma membrane of the enterocyte is not known. The energy for extrusion could be derived from the hydrolysis of high-energy compounds or from energy carried by ion gradients. Mg2+ efflux is dependent on Na+ antiport activity in a number of different cell types and, in some cases, requires ATP, either for activating or for energising transport (Féray and Garay, 1986; DiPolo and Beaugé, 1988; Frenkel et al. 1989; Günther and Höllriegl, 1993a; Xu and Willis, 1994; Handy et al. 1996). An ATP-consuming Mg2+ pump has been described for some bacteria (Maguire, 1992; Smith and Maguire, 1995). We have demonstrated, using a fluorescent Mg2+ probe, that isolated enterocytes of Mozambique tilapia require external Na+ to reverse the Mg2+ loading that occurs in medium containing 5 mmol l−1 Mg2+ (Fig. 1A). Clearly, this hinted at the presence of Na+/Mg2+ antiport activity in these cells. However, we have been unable to confirm this in studies conducted at the molecular level using a radiotracer to follow Mg2+ movement across the plasma membrane (Fig. 1B; Bijvelds et al. 1996b). Neither could the presence of an ATP-driven Mg2+ transporter be demonstrated. Since Mg2+ transport did not appear to be directly coupled to Na+ movement, it was hypothesised that it may, therefore, be coupled to the transport of other ions that are accumulated in the enterocytes via Na+-dependent mechanisms. Because several anion-coupled Mg2+ transport mechanisms have been described (Günther et al. 1986; Günther and Vormann, 1989; Günther and Höllriegl, 1993b), the ability of the transmembrane gradients of several anions to stimulate Mg2+ transport in plasma membrane preparations prepared from enterocytes was examined. The results showed that Mg2+ uptake into plasma membrane vesicles was stimulated by some membrane-permeable anions (Bijvelds et al. 1996b) and that this transport was sensitive to stilbene derivatives. In this respect, the enterocyte Mg2+ transport system resembles the Cl− symport mechanism found in erythrocytes (Günther and Vormann, 1989).
Coupling of Mg2+ extrusion to anion symport may result in an electroneutral mode of transport. Thus, anion symport may be a way of overcoming the large electrical barrier across the plasma membrane that opposes the extrusion of positive charge from enterocytes. The initial observations that Mg2+ absorption is inhibited by bumetanide and ouabain (Partridge et al. 1987; Van der Velden et al. 1990) may be explained by assuming that these substances disturb cellular Cl− transport by inhibiting the Na+/K+/2Cl− cotransporter and the Na+/K+-ATPase (which maintains the transmembrane Na+ gradient that drives Cl− uptake), respectively.
Branchial magnesium transport
In freshwater environments, the gills are an important route for active uptake of ions such as Na+, Ca2+ and Cl− (Evans, 1980; Flik et al. 1985). Some evidence suggests that Mg2+ may also be absorbed via the gills. In fish fed a low-magnesium diet, the deposition of magnesium in the body can exceed the dietary magnesium intake (Shearer, 1989; Dabrowska et al. 1991; Bijvelds et al. 1996a), indicating that the food cannot be the only source of magnesium and that, consequently, Mg2+ may be obtained directly from the water. Assuming that water ingestion is negligible in freshwater fish (to prevent water loading), these observations implicate the integument as an alternative route for Mg2+ uptake. In Mozambique tilapia, when dietary Mg2+ intake is reduced by providing a low-magnesium diet, the extra-intestinal (presumably branchial) Mg2+ uptake contributes significantly (approximately 30 %) to the total amount of magnesium accumulated (Bijvelds et al. 1996a). Moreover, when access to water-borne magnesium is also reduced, a marked depletion of the magnesium stores of the body occurs (Shearer and Asgard, 1992; Bijvelds et al. 1996a). Thus, it is reasonable to conclude that it must be branchial Mg2+ uptake that enables the Mozambique tilapia to maintain a positive magnesium balance (a net magnesium accumulation) when the dietary magnesium intake is restricted. In fresh water, a slightly negative transepithelial potential (approximately −5 mV) is maintained across the branchial epithelium, which would not support passive Mg2+ uptake (Dharmamba et al. 1975; Perry and Flik, 1988; Young et al. 1988). Therefore, although such a mechanism has not hitherto been demonstrated, we predict that an active transcellular Mg2+ transport mechanism is present in the branchial epithelium.
Only a few studies have measured intake or uptake (=intake minus loss) of magnesium from the water. In white suckers Catostomus commersoni of approximately 500 g, the rate of branchial magnesium uptake was insignificant (2±13 µmol h−1 kg−1; Hõbe et al. 1984). In 50 g carp, the rate of magnesium uptake from the water was approximately 50 nmol h−1 or 1 µmol h−1 kg−1 (Van der Velden et al. 1991b). For juvenile (6.5 g) as well as 50 g Mozambique tilapia, an integumental magnesium intake of approximately 2 µmol h−1 kg−1 was measured (Van der Velden et al. 1991a; Bijvelds et al. 1996a).
Magnesium loss via the gills is usually low. In rainbow trout injected with a large dose of magnesium, the extrarenal net magnesium efflux was below the level of detection of 1 µmol h−1 kg−1 and less than 1 % of the amount excreted renally (Oikari and Rankin, 1985). In Mozambique tilapia fed a low-magnesium diet, the magnesium deposition in the body was equal to the intake (from food and from the water) minus the renal excretion, implying that integumental magnesium loss was negligible (Bijvelds et al. 1996a). This indicates that freshwater fish maintain a very low integumental permeability for magnesium and thus effectively minimise diffusive losses.
Renal magnesium transport
Magnesium excretion proceeds almost exclusively via renal pathways (Hickman, 1968; Oikari and Rankin, 1985). Freshwater fish produce a voluminous and substantially hypotonic urine. In this way, excess water taken in osmotically is excreted, whilst salt losses are minimised by reabsorption of electrolytes in the renal tubules and the bladder. Magnesium concentrations of 0.2–1.3 mmol l−1 have been found in the urine of freshwater teleosts (Oguri, 1968; Hõbe et al. 1983; Nanba et al. 1987; Wales and Gaunt, 1987; Bijvelds et al. 1996a). Reported mean urine flows in freshwater teleosts vary from 1.5 to 4.5 ml h−1 kg−1 (Hickman and Trump, 1969; Hunn, 1982; Hõbe et al. 1983; Wales and Gaunt, 1987). From these values, we calculate a mean urinary magnesium loss of approximately 2 µmol h−1 kg−1. Since the amount of Mg2+ filtered in the glomeruli can exceed the urinary Mg2+ output, Mg2+ must be reabsorbed across the renal epithelium (Hickman and Trump, 1969; Schmidt-Nielsen and Renfro, 1975). Mg2+ reabsorption in the bladder is probably insignificant (Foster, 1975). In Mozambique tilapia, the renal magnesium output decreased (without concomitant changes in plasma magnesium levels) when fish were fed a diet with a low magnesium content, suggesting that renal magnesium reabsorption is a regulated process that enables fish to minimise magnesium losses under these circumstances (Bijvelds et al. 1996a).
Regulation of renal Mg2+ reabsorption may involve stanniocalcin. This hormone is produced by the corpuscles of Stannius (which are embedded in the kidneys of holostean and teleost fish) and is chiefly known for its hypocalcaemic actions. Both the gills and the kidneys are targets for stanniocalcin, where it controls epithelial Ca2+ transport by a mechanism not yet fully elucidated. Besides causing an increase in plasma Ca2+ concentration, the removal of the corpuscles of Stannius from the European eel Anguilla anguilla and the American eel A. rostrata caused an increase in K+ concentration and a decrease in Na+ and Mg2+ concentration (Fontaine, 1964; Butler, 1969; Kenyon et al. 1980; Butler and Cadinouche, 1995), indicating that stanniocalcin exerts some action on the handling of other major electrolytes apart from Ca2+. In stanniectomised American eel, renal magnesium excretion increased, and it was suggested that stanniocalcin, or some other putative substance produced by the corpuscles of Stannius, directly or indirectly reduces Mg2+ reabsorption (Butler, 1993; Butler and Cadinouche, 1995). Ovine prolactin decreased urine magnesium levels in starry flounder Platichthys stellatus, but did not affect renal magnesium excretion (Foster, 1975). This confirms that prolactin primarily decreases water reabsorption and has no direct effect on Mg2+ transport. Chan (1975) reported that urotensin I from the white sucker Catostomus commersoni increased urine magnesium levels and renal magnesium excretion in silver eels (A. rostrata). In the Japanese eel A. japonica, administration of extracts from the urophysis (producing urotensin I and II) of carp increased renal excretion of K+, Ca2+ and Mg2+ by increasing the urine volume, without affecting ion concentrations in the urine. Although the urotensins apparently affect renal Mg2+ excretion in eels, the evidence for direct effects of any product released by the urophysis on tubular Mg2+ transport is very limited. Urotensin I and II have complex interactions with the hypothalamo-pituitary-interrenal axis and may thus influence osmoregulory signals, e.g. cortisol activity (Wendelaar Bonga, 1993). This could explain some of their effects on renal ion and water excretion.
In mammals, the proximal tubule and the ascending limb of the loop of Henle are the principal sections of the nephron involved in Mg2+ transport (Ryan, 1990; Quamme, 1993). Mg2+ reabsorption proceeds predominantly through solute-linked paracellular transport and is driven by the transepithelial potential (Di Stefano et al. 1993; De Rouffignac and Quamme, 1994). A minor transcellular component may also be involved (Ryan, 1990). In freshwater fish proximal tubules, similar transepithelial potentials have been reported (Nishimura and Imai, 1982), and they may constitute a driving force for Mg2+ reabsorption. However, reabsorption of water must be limited in freshwater fish, and this will reduce the extent to which paracellular solute-linked transport of salts occurs. Therefore, active transcellular Mg2+ reabsorption may be more prominent in fish than in terrestrial vertebrates.
In apical plasma membrane preparations of renal tissue of Mozambique tilapia (Bijvelds et al. 1997b) and rainbow trout (Freire et al. 1996), electrodiffusive Ca2+-sensitive Mg2+ transport routes have been identified that may mediate passive Mg2+ transport. Because of the prevailing electrochemical potential across the luminal membrane, this pathway may be involved in absorptive Mg2+ transport. Considering that the renal tubules of euryhaline species fulfil opposite roles in fresh water (reabsorption of salts) and sea water (secretion of salts), the Mg2+ transport route in these euryhaline freshwater-acclimated fish is remarkably similar to the apical Mg2+ transport described for marine winter flounder Pseudopleuronectes americanus (Renfro and Shustock, 1985). Indeed, proximal tubules of freshwater and seawater killifish Fundulus heteroclitus displayed very similar Mg2+ transport characteristics in vitro (Cliff and Beyenbach, 1992), and seawater acclimation did not affect the Mg2+ transport characteristics of renal plasma membrane preparations from the Mozambique tilapia (Bijvelds et al. 1997b).
A high water magnesium concentration (up to 50 mmol l−1) lowers the integumental permeability to water and ions, and this elicits a compensatory hormonal response (involving a decrease in prolactin activity) in stickleback Gasterosteus aculeatus (Wendelaar Bonga, 1978) and in Mozambique tilapia (Wendelaar Bonga et al. 1983). Only relatively small changes in plasma Ca2+ and Mg2+ levels occurred, and a high external magnesium concentration had no apparent deleterious effects in stickleback (Wendelaar Bonga, 1978), Mozambique tilapia (Wendelaar Bonga et al. 1983) or goldfish Carassius auratus (Olivereau et al. 1987). Whether this is because freshwater fish can restrict their integumental magnesium influx and/or, like sea water species, eliminate excess magnesium through renal pathways is not known. Certainly, it has been shown that some euryhaline species maintain the capacity for rapid and effective renal Mg2+ excretion in fresh and brackish waters when challenged by a large magnesium dose administered directly to the body (Natochin and Gusev, 1970; Hirano, 1979; Oikari and Rankin, 1985). Apparently, the secretory mechanism present, albeit normally silent, can be activated instantaneously, probably triggered by an increase in plasma magnesium concentration. As far as we know, no comparable studies have been conducted with stenohaline freshwater fish. Therefore, it is not known whether the renal magnesium secretion mechanism found in some euryhaline species is also present in species restricted to freshwater habitats.
Perspectives
Much of the function of the intestine and kidneys in magnesium handling has been elucidated over the last few decades, and it is evident that these organs are crucial for maintaining the magnesium balance. Furthermore, there is substantial evidence for the involvement of branchial routes in Mg2+ uptake (Fig. 2). Nevertheless, many questions concerning magnesium handling in fish have remained unanswered. Recent progress notwithstanding, our knowledge of the cellular and molecular basis of magnesium transport across epithelia and, in particular, across cell membranes is limited. Furthermore, relatively little is still known about the hormonal control mechanisms that govern magnesium transport in fish and in terrestrial vertebrates. Indeed, when compared with our understanding of the physiology of other major cations such as Na+, K+ and Ca2+, one can only conclude that much of the physiology of magnesium is still uncharted terrain.
The paucity in our understanding of magnesium physiology is surely related to the limited use that has been made of radioactive magnesium isotopes. Radiotracer techniques are highly sensitive and allow an assessment of unidirectional transport, making these techniques virtually indispensable for the study of ion transport at the tissue and (sub)cellular levels. Unfortunately, the only radioisotopes of magnesium applicable in biological research have a relatively short half-life (9.46 min for 27Mg and 20.9 h for 28Mg), which has restricted their availability for the life sciences (Flik et al. 1993). Natural magnesium consists of three stable isotopes, which may also be used as tracers, using mass spectrometry or neutron activation to quantify Mg2+. For instance, this principle has been applied to study the distribution of magnesium in human blood after oral ingestion of 25Mg (Stegmann and Karbach, 1993).
Probably because radioactive probes are scarce, fluorescent Mg2+ indicators have become increasingly popular. Like radioactive probes, they allow an assessment of dynamic processes at the cellular level and, for this reason, are particularly suitable for studying cellular Mg2+ transport and regulation. One intrinsic limitation of the technique, however, is that it does not permit an assessment of unidirectional Mg2+ transport since only relative changes in intracellular Mg2+ activity are measured. Because Mg2+ is strongly buffered in the cell, proportionally large transmembrane Mg2+ fluxes may be necessary to elicit detectable changes in the intracellular Mg2+ concentration. For this reason, fluorescent Mg2+ probes may be a less sensitive method of monitoring Mg2+ transport. Another drawback is that fluorescent Mg2+ probes have a relatively high affinity for Ca2+ (which, since they were developed from similar Ca2+ probes, is not surprising). Therefore, the possible interference between Ca2+ dynamics and Mg2+ measurements should always be taken into account (McGuigan et al. 1993). Despite being considered an accurate method for measuring intracellular Mg2+ activity, microelectrodes have been used less often than fluorescent dyes, partly because Mg2+-selective microelectrodes suffer from the same limitations, such as interference from other ions (Alvarez-Leefmans et al. 1986; McGuigan et al. 1993). The invasive nature of the technique and the laborious calibration procedures may also have limited its use.
ACKNOWLEDGEMENTS
As more sensitive and specific Mg2+ probes become available, we foresee rapid progress in the field of magnesium physiology. In this respect, a molecular approach may be fruitful for future research on cellular Mg2+ transport in vertebrates. The genetic characterisation of Mg2+ transporters in Salmonella typhimurium and, successively, in other bacteria has provided some interesting and promising perspectives that warrant a search for phylogenetically related eukaryotic counterparts (Hmiel et al. 1989; Smith and Maguire, 1995). Some progress has been made with the demonstration of an mRNA transcript homologous to the Escherichia coli CorA Mg2+ transport system in rainbow trout kidney (Freire, 1996). Indeed, considering the formidable Mg2+ transport capacity of the (marine) fish kidney, molecular screening of renal tissue may prove especially rewarding.