Transcellular Intestinal Calcium Transport in Freshwater and Seawater Fish and its Dependence on Sodium/Calcium Exchange

The Na⁺/Ca²⁺ exchanger is a plasma membrane protein crucial to calcium homeostasis in, for example, muscles and neurones (DiPolo, 1989). During the last decade, the opinion prevailed that the protein might be typical of excitable cells (Baker, 1986). Non-excitable cells normally do not develop the high intracellular ionic calcium concentrations characteristic of the former. Therefore, it was thought that the contribution of the Na⁺/Ca²⁺ exchanger, which has a relatively low Ca²⁺ affinity, to cellular calcium homeostasis of such cells was of minor importance. In recent years, however, evidence has accrued that the exchanger can be found in quite a number of non-excitable cells. Cells in ion-transporting epithelia were repeatedly shown to display Na⁺/Ca²⁺ exchange activity involved in transcellular ion transport (Flik et al. 1990; Taylor, 1989). The Na⁺/Ca²⁺ exchanger can serve different purposes in non-excitable cells, as shown by its involvement in volume regulation and Na⁺ extrusion in canine erythrocytes (Parker, 1988; Schoenmakers and Flik, 1992). Basolateral membranes of enterocytes of the freshwater fish Oreochromis mossambicus Peters (tilapia) have been shown to display a weak ATP-dependent Ca²⁺ pump activity and a vigorous Na⁺/Ca²⁺ exchange activity (Flik et al. 1990; Schoenmakers et al. 1992a).

The issue of the importance of the Na⁺/Ca²⁺ exchanger in non-excitable cells is interlaced with the topic of its regulation. In vitro the functioning of the Na⁺/Ca²⁺ exchange protein has been modulated in many ways: membrane composition, intracellular Ca²⁺ concentration and cellular energy status are factors known to influence Na⁺/Ca²⁺ exchange activity (DiPolo, 1989; Philipson and Ward, 1985). These studies were performed on cells or tissues in a steady state. Few studies have concentrated on the question of how altered physiological requirements are reflected by changes in the activity of the Na+/Ca2+ exchanger (Tibbits et al. 1989).

We provide data on the importance and regulation of the Na⁺/Ca²⁺ exchanger in an ion-transporting cell. We used the intestine of the tilapia as a model of an ion-transporting epithelium. The tilapia is a euryhaline fish able to live in fresh water as well as in full-strength sea water. Freshwater fish are hyper-osmoregulators in which the active uptake of ions is needed for metabolism and growth. They drink little and take up most of the monovalent ions they need (such as Na⁺, K⁺ and Cl⁻), as well as most of their calcium, via their gills (Mayer-Gostan et al. 1987; Fenwick, 1989). Seawater fish, however, are hypo-osmoregulators, confronted in particular with osmotic water loss. Therefore, seawater fish drink large amounts of sea water and excrete, via the gills and kidneys, the excess ions taken up in the intestine (Maetz, 1976).

Freshwater fish need extra calcium during growth and ovarian maturation or when the calcium concentration in the water is low, as in soft fresh water (Flik et al. 1986). Under the latter circumstances, fish experience extra calcium loss as a result of an increased epithelial permeability to calcium (Hõbe et al. 1984). When faced with such adverse conditions, they supplement branchial calcium uptake with an increase in intestinal uptake (Berg, 1970).

Experiments on tilapia in our laboratory have shown that the net branchial uptake of calcium, sufficient for growth and homeostasis of fish in fresh water containing around 1 mmol l⁻¹ Ca²⁺, increases when the fish are transferred to sea water: branchial Ca2+ influx is not affected, but branchial Ca²⁺ efflux is drastically reduced (G. Flik, personal observations). Drinking of sea water containing around 10mmol l⁻¹ calcium will lead to an excess influx of calcium in seawater fish unless the intestinal uptake of calcium is strictly controlled. Paracellular influx of calcium in the intestine is counteracted by the generation of an inside-positive transepithelial potential (Maetz, 1976). Here, we concentrate on another means of decreasing net calcium uptake, i.e. the reduction of the active transcellular transport of calcium across the intestine.

We measured unidirectional calcium fluxes in stripped intestinal epithelium of freshwater- and seawater-adapted fish. The isolation of basolateral plasma membrane vesicles allowed us to test which of the two extrusion mechanisms for Ca²⁺ (the ATP-dependent Ca²⁺ pump or Na⁺/Ca²⁺ exchange) is more involved in transcellular calcium uptake under these conditions.

Male freshwater tilapia from laboratory stock, weighing around 250g, were held in 100l tanks. The aquaria were supplied with running tap water (0.7mmol l⁻¹ Ca²⁺, 25°C) under a photoperiod of 12h of light alternating with 12h of darkness. Animals were fed Trouvit fish pellets (Trouw and Co., Putten, The Netherlands) at 1.5% body weight per day.

Artificial sea water was prepared by dissolving natural sea salt (Wimex, Krefeld, Germany) to a density of 1.022 gl⁻¹. Fish were first transferred to half-strength sea water. After 1 day, they were transferred to full-strength sea water. Fish were adapted to sea water for at least 8 weeks before experimentation. Food was supplied at 1.5% body weight per day.

The intestinal epithelium was stripped of its underlying muscle layers as described previously (Flik et al. 1990; Groot et al. 1979). Briefly, fish were killed by spinal transection and pithed. The peritoneal cavity was opened and the intestinal tract removed. The intestine was flushed with saline (117.5 mmol l⁻¹ NaCl, 5.7 mmol l⁻¹ KCl, 25 mmol l⁻¹ NaHCO₃, 1.2mmol l⁻¹ NaH₂PO₄, 1.25 mmol l⁻¹ CaCl₂, 1.2 mmol l⁻¹ MgSO₄, 5 mmol l⁻¹ glucose and 28 mmol l⁻¹ mannitol) and attached strands of fat were carefully removed. The intestine was cut open lengthwise and dipped up and down in saline to remove any remaining food residues. Next, the intestine was put on a Parafilm-covered glass plate mucosa-side-up. The mucosa of the proximal 15cm was stripped free of underlying muscle layers (Groot et al. 1979). The stripped epithelium was mounted on slides (the exposed tissue area was 0.16cm2) and incubated for approximately 20min in gassed saline (95% O₂/5% CO₂; pH7.3±0.1).

The slides were then mounted as the partition between two 4ml half-chambers filled with saline. At time zero, 45CaCl₂ (specific activity: 19 GBqmmol⁻¹) was added as a tracer to either side of the tissue to yield a radioactive concentration of 0.33MBqml⁻¹. After 45min, samples (100 μl) were taken from both half-chambers simultaneously every 15min, for a total period of 180min. The samples were instantly replaced by 100 μl of fresh Ringer’s solution. In pilot experiments it was established that unidirectional fluxes are constant over this period.

Using symmetrical conditions (saline on both sides) should keep passive fluxes of ⁴⁵Ca at a minimum, allowing a good assessment of active calcium transport processes (Barry and Diamond, 1984). The transepithelial potential (TEP) of stripped intestinal preparations from freshwater tilapia was 0.93±0.66mV (N=24). This small TEP indicates that passive calcium flux is low in this Ussing chamber apparatus.

Unidirectional calcium fluxes were calculated as follows. Specific activity (SA) of the radioactive saline (cis-side) was determined from three 10 μl samples distributed over the whole experimental period and expressed in disintsmin⁻¹ nmol⁻¹. The amount of isotope at the trans-side at a given time [q(t) in disintsmin⁻¹] was calculated from the 100 μl samples. Data were corrected for background, and cumulative counts for the whole Ussing hemi-chamber were calculated. Total calcium transported (in nmolcm⁻²) as a function of time was obtained by calculating q(t)/(SAXA) for each sample, where A is the exposed tissue area (in cm²). Unidirectional steady-state Ca²⁺ fluxes were determined by linear regression of calcium accumulation over periods of at least 45min. To calculate net fluxes (J_net), values for unidirectional fluxes from mucosa to serosa (J_ms) and from serosa to mucosa (J_sm) across adjacent segments from the same fish were used in data analysis. Ca²⁺ fluxes were expressed in nmol h⁻¹cm⁻².

Membrane vesicles were isolated from the proximal 30cm of the intestine as described in detail previously (Flik et al. 1990). Vesicles were kept cooled on ice and used on the day of isolation. The protein content of the membrane vesicle preparations was 2.0±0.4 gl⁻¹ (N=12), as determined with a commercial reagent kit (Bio-Rad) with bovine serum albumin (BSA) as a reference. Protein recovery was around 1.9% in both freshwater and seawater preparations. Na⁺/K⁺-ATPase activity in fish intestinal basolateral plasma membrane vesicles (BLMV) was determined as described earlier (Flik et al. 1990). Na⁺/K⁺-ATPase activity was defined as the Na⁺- and K⁺-dependent, ouabain-sensitive phosphatase activity.

All assay media contained 0.5 mmol l⁻¹ EGTA, 0.5 mmol l⁻¹N-(2-hydroxyethyl)-ethylenediamine-N,N′,N′-triacetic acid (HEEDTA) and 0.5mmol l⁻¹ nitrilotriacetic acid (NTA). Free calcium and magnesium levels were calculated as described in detail by Schoenmakers et al. (1992a,b). All incubations were performed at 37°C.

The ATP-driven transport of 45Ca was assayed using a rapid filtration technique (Flik et al. 1990). One-minute incubations of membrane vesicles in the presence of 3mmoll⁻¹ ATP yielded 45Cauptakes (representing initial velocities of the pump) and were corrected for uptake in the absence of ATP. Free Ca²⁺ concentrations were varied from 50nmol l⁻¹ to 1 μmoll⁻¹. A 14-fold dilution in ice-cold isotonic medium containing 0.1mmol l⁻ LaCl₃ stopped the uptake, and the suspension was filtered (Schleicher & Schüll ME25, pore size 0.45 μm). Filters were rinsed twice with 2ml of ice-cold medium and transferred to counting vials. 4ml of Aqualuma was added per vial, filters were allowed to dissolve (30min at room temperature) and radioactivity was determined in a Pharmacia Wallac 1410 liquid scintillation counter.

Na⁺/Ca²⁺ exchange activity in plasma membrane vesicles from fish enterocytes was assayed as described previously (Flik et al. 1990). Briefly, 5 μl of membrane vesicles equilibrated with 150mmol l⁻¹ NaCl was added to 120 μl of medium containing ⁴⁵Ca and either 150mmol l⁻¹ NaCl (blank) or 150mmol l⁻¹ KCl. Ionic Ca²⁺ concentrations ranged from 250nmol l⁻¹ to 25 μmol l⁻¹. After 5s at 37°C, the reaction was stopped by addition of 1ml of ice-cold isotonic stopping solution containing 1mmol l⁻¹ LaCl₃ and the solution was filtered. The filter was then rinsed three times with 2ml of stopping solution and treated as described above. The difference in ⁴⁵Ca accumulation was taken to represent Na⁺-gradient-driven Ca²⁺ transport.

The K_m and V_max values of the ATP-dependent Ca²⁺ pump and the Na⁺/Ca²⁺ exchanger were derived by non-linear regression analysis of the mean velocities (N=5–6) measured as a function of substrate concentration. The standard deviations of the values given by the computer algorithm indicate the accuracy of the estimated values, but cannot be used for statistical assessment of differences (Press et al. 1986).

Otherwise, statistical significance of differences between mean values was tested using a Mann–Whitney U-test. Statistical significance was accepted at P<0.05.

Calcium fluxes in stripped intestinal epithelium of freshwater and seawater tilapia are shown in Fig. 1. The mucosa-to-serosa flux of Ca²⁺ (J_ms) in seawater epithelium was not significantly different from that through freshwater epithelium. The serosa-to-mucosa (J_sm) calcium flux in freshwater fish, 19.6±7.5 nmolCa²⁺h⁻¹ cm⁻² (mean ± S.D.; N=15), was significantly (P<0.05) lower than the flux, 28.6±10.0nmolCa²⁺h-1 cm⁻² (mean ± S.D; N=7), in seawater fish, indicating an increased permeability to Ca²⁺ of the tight junctions in vitro. The net uptake of Ca²⁺ (J_net=J_ms-J_sm) in seawater intestine (5.0±5.2 nmolCa²⁺ h⁻¹ cm⁻²; mean ± S.D.; N=7) was not significantly different from zero. The net uptake was significantly (P<0.001) reduced with respect to the net flux in freshwater fish (17.1±8.9 nmolCa²⁺h⁻¹ cm⁻²; mean ± S.D.; N=15).

Using Na⁺/K⁺-ATPase as a marker to assess basolateral plasma membrane purification we found no significant (P0.15) difference in purification between freshwater and seawater tilapia membranes: Na+/K+-ATPase activity purified 4.8±1.0 (N=5) times for seawater tilapia membranes and 6.2±1.7 (N=6) times for freshwater tilapia membranes. The activities of the three basolateral plasma membrane proteins we characterized are given in Table 1. Na+/K+-ATPase activity in the homogenate of intestinal scrapings from seawater tilapia amounted to 26.9±2.9 μmol P_i h⁻¹ mg⁻¹ (N=5), which is significantly (P<0.001) higher than the activity observed in intestinal scrapings from freshwater tilapia (17.3±2.7 μmol P_i h-1 mg-1; N=6). In purified basolateral plasma membranes from seawater fish, its activity was 127.6±14.1 μmol P_i h⁻¹ mg⁻¹ (N=5), differing significantly (P<0.001) from that in freshwater fish, which amounted to 102.7±10.5 μmol P_i h⁻¹ mg-1 (N=6).

Seawater fish enterocytes have a greatly decreased Na⁺/Ca²⁺ exchange activity in the purified basolateral membrane fraction. The maximal velocity of the Na⁺/Ca²⁺ exchange activity in basolateral plasma membrane vesicles from freshwater-adapted tilapia [18.21±0.40nmolCa²⁺ min⁻¹ mg⁻¹ (N=6)] was reduced by 57% in vesicles from seawater-adapted tilapia [7.85±0.21nmolCa²⁺ min⁻¹ mg⁻¹ (N=5)]. The K_m of Na⁺/Ca²⁺ exchange in freshwater tilapia (2.26±0.13 μmol l⁻¹; N=6) differs only slightly from that of Na⁺/Ca²⁺ exchange in seawater tilapia (1.76±0.16 μmol l⁻¹; N=5).

The ATP-dependent Ca²⁺ pump in the basolateral plasma membrane vesicles did not show such a distinct change in maximal velocity. In vesicles from freshwater and seawater tilapia we measured a slight change in K_m, comparable in relative magnitude to the change in K_m of the Na⁺/Ca²⁺ exchanger. V_max decreased in a similar manner by 28% from 0.78±0.06nmolCa²⁺ min⁻¹ mg⁻¹ in freshwater fish to 0.56±0.02nmolCa²⁺ min⁻¹ mg⁻¹ in seawater fish.

The net uptake of calcium in the proximal intestine calculated on the basis of unidirectional calcium fluxes approaches zero and is significantly lower in seawater fish than in freshwater fish. Apparently, seawater fish have successfully reduced active calcium transport mechanisms and keep net intestinal transcellular calcium influx low. The decrease in net uptake and the increase in paracellular calcium mobility result in an apparently unchanged unidirectional calcium influx.

Whole-body calcium influx and efflux for a 250g freshwater tilapia can be calculated according to Flik et al. (1985). Influx of calcium will be around 4.3 μmol h⁻¹ while calcium efflux is around 0.7 μmol h⁻¹, resulting in a net whole-body calcium influx of 3.6 μmol h⁻¹. The length of intestine that performs the measured net influx of 17.1nmolcm-2 h⁻¹ is not known. If one assumes that this transport occurs over a length of 50cm (on the basis of gut morphology), the calculated net influx (with a width of 1cm) is 0.86 μmol h⁻¹. Intestinal calcium influx would then contribute significantly to the whole-body influx.

The net uptake of calcium in tilapia intestine under freshwater conditions has been demonstrated by Flik et al. (1990). In that study we also showed that this net uptake depended on the presence of a sodium gradient across the basolateral plasma membrane of the enterocyte. This indicated that Na⁺/Ca²⁺ exchange, shown to be present in these membranes, was important for transcellular calcium uptake (Flik et al. 1990). The lower net intestinal calcium uptake in seawater intestine shown here prompted an investigation into the molecular basis of this change.

Seawater adaptation induced significant changes in two of the three plasma membrane ion-transporting mechanisms we studied. The increase in Na⁺/K⁺-ATPase activity we found is a normal phenomenon for intestine and gills of seawater-adapted fish (Utida et al. 1972; Maetz, 1976; Pagliarani et al. 1991). Around 90% of the cells in the proximal intestine are of the absorptive cell type. Therefore, the observed increase in Na⁺/K⁺-ATPase activity per milligram membrane protein represents an increase in activity per cell, not an increase in the numbers of a specific cell type. The increase in the ratio of the activities of Na⁺/K⁺-ATPase to Ca²⁺-ATPase in the purified basolateral plasma membrane vesicles, 3.80X103 in seawater vs 2.19X103 in freshwater fish [dividing the Na+/K+-ATPase activity (μmol P_i h⁻¹ mg⁻¹) by the Ca²⁺-ATPase activity (nmolCa²⁺ h⁻¹ mg⁻¹) and assuming that 1 P_i is formed per Ca²⁺ transported], shows that the increase in Na⁺/K⁺-ATPase activity is not due to a purification effect (Flik et al. 1984). The increase is in line with ultrastructural observations on the intestine of seawater tilapia showing more extensive basolateral membrane invaginations and more mitochondria than in freshwater fish and giving the impression of very active cells (G. Flik, personal observations). The elevated Na⁺/K⁺-ATPase activity is consistent with the water transport function of the intestine of seawater fish (Skadhauge, 1969; Utida et al. 1972), assuming that solvent drag is the motive force for this water transport.

The ATP-dependent Ca²⁺ pump activity in plasma membranes from freshwater tilapia is slightly higher than that observed in membranes from seawater tilapia. If we combine these data with the respective Na⁺/K⁺-ATPase purification data, we calculate equal amounts of Ca²⁺ pump activity per milligram cell protein in the whole-cell homogenate (direct measurements of this activity in such a preparation are complicated) in freshwater (0.13nmolCa²⁺ min⁻¹ mg⁻¹ cellprotein) and in seawater enterocytes (0.12nmolCa²⁺ min⁻¹ mg⁻¹ cellprotein). Apparently, freshwater- and seawater-adapted enterocytes contain equal amounts of ATP-dependent calcium pump proteins.

A very significant change was found in the Na⁺/Ca²⁺ exchange activity. The maximum velocity of this carrier decreased by 57% in seawater tilapia. When we calculate activities in the whole-cell homogenate (as for the ATP-dependent Ca²⁺ pump above), there is a large difference between freshwater enterocytes (2.9nmolCa²⁺ min⁻¹ mg⁻¹ cellprotein) and seawater enterocytes (1.6nmol Ca²⁺ min⁻¹ mg⁻¹ cellprotein). Seawater enterocytes clearly contain less Na⁺/Ca²⁺ exchange activity. The observed 57% reduction in Na⁺/Ca²⁺ exchange activity in purified plasma membranes is in line with the 71% decrease in net transepithelial calcium uptake. We take this observation as evidence for the involvement of the Na⁺/Ca²⁺ exchanger in the active transcellular uptake of calcium in both freshwater and seawater fish.

This conclusion is further supported when we plot the activities of the two Ca²⁺-transporting membrane proteins under both conditions (Fig. 2). This graph does not take the inside-negative cell membrane potential of around −60mV (Bakker and Groot, 1988) into account. The Na⁺/Ca²⁺ exchange is an electrogenic process, swapping 3 Na⁺ for 1 Ca²⁺ (Flik et al. 1990), and would be even more active than shown here. In the physiological intracellular Ca²⁺ concentration range (around 85nmol l⁻¹; Schoenmakers et al. 1992a) the Na⁺/Ca²⁺ exchanger clearly dominates basolateral Ca²⁺ extrusion from the enterocytes in freshwater fish. Under seawater conditions, both the absolute calcium transport capacity of the extrusion mechanisms and the contribution of the Na⁺/Ca²⁺ exchanger to this transport have decreased.

The distinct decrease in V_max of the Na⁺/Ca²⁺ exchange activity probably reflects a regulatory principle: presumably, the expression rate of the exchanger’s gene has decreased and there are fewer exchanger molecules per cell. Table 1 shows that the percentage decreases in K_m of the ATP-dependent Ca²⁺ pump and the Na⁺/Ca²⁺ exchanger are similar. This suggests the existence of a mechanism that has a similar effect on both membrane proteins. We suggest that seawater fish develop a more fluid plasma membrane. Increased membrane fluidity has been shown to decrease the K_m of Na⁺/Ca²⁺ exchange (Philipson and Ward, 1985). The lipid composition of gill membranes changes when fish adapt to seawater conditions (Hansen, 1987). The changed endocrine status of seawater fish may well underlie such a phenomenon. Cortisol and prolactin levels change upon adaptation to sea water (Utida et al. 1972). Actions of cortisol on calcium homeostasis appear to involve primarily a stimulation of calcium uptake mechanisms (Utida et al. 1972; Flik and Perry, 1989). Prolactin, however, influences plasma membrane fluidity and enzyme activity in rat tissues in a complex manner (Dave et al. 1981; Dave and Witorsch, 1985). Thus, the plasma membrane fluidity of the enterocytes of seawater fish may have increased as a result of the decreased prolactin levels.

We here show that seawater fish develop more intestinal Na⁺/K⁺-ATPase activity, which is in line with one of the main functions of the intestine in seawater fish, i.e. water uptake by solute-linked water transport, and with the lower prolactin levels in seawater fish, as prolactin is known to exert an inhibitory control over Na⁺/K⁺-ATPase activity (Pickford et al. 1970). Also, this study is the first to give evidence for the regulation of transepithelial calcium flux in the tilapia intestine. Seawater fish drastically reduce active intestinal uptake of calcium. The reduction in Na⁺/Ca²⁺ exchange activity in purified basolateral plasma membranes, which parallels the reduction in net transepithelial calcium uptake, supports the notion that the Na⁺/Ca²⁺ exchanger is the transport protein chiefly responsible for transcellular uptake of calcium in fish intestine.

We wish to thank Mr Klaas Dekker and Dr J. A. Groot (University of Amsterdam) for their expert advice on practical and theoretical aspects of using the Ussing-chamber technique with stripped tilapia intestine. Th. J. M. Schoenmakers is supported by the Foundation for Fundamental Biological Research (BION) and the Netherlands Organization for Scientific Research (NWO).