Kinetics of ATP- and Na⁺-gradient driven Ca²⁺ transport in basolateral membranes from gills of freshwater- and seawater-adapted tilapia

The major fraction of calcium absorbed by freshwater fish is taken up directly from the water via the gills (Fenwick, 1989). Flux ratio analysis indicates that this uptake is transcellular: the electrochemical potential for a passive Ca²⁺ flux is directed outwards and, at the same time, a significant Ca²⁺ influx occurs (Perry and Flik, 1988). Moreover, the branchial epithelium of freshwater fish is generally considered to be a tight epithelium with a minimal paracellular flux of ions and water (Wendelaar Bonga et al. 1983; Ogasawara and Hirano, 1984). In a current model for transbranchial Ca²⁺ uptake (Flik et al. 1985c), Ca²⁺ enters the ionocytes (also called chloride cells) through Ca²⁺ channels in the apical membrane (the rate-limiting step; Verbost et al. 1993) and is transported across the basolateral membrane to the serosa by a high-affinity Ca²⁺-ATPase. High-affinity Ca²⁺-ATPase activity has been demonstrated in the basolateral membranes of the ionocytes of tilapia, eel and trout (Flik et al. 1985b,c; Perry and Flik, 1988). There is good evidence that stanniocalcin inhibits the transcellular Ca²⁺ influx by limiting the entrance of Ca²⁺ at the apical membrane (Verbost et al. 1993).

In the basolateral membrane of some Ca²⁺-transporting epithelia, ATP-dependent Ca²⁺ transport as well as Na⁺/Ca²⁺ exchanger activity have been demonstrated (Van Os, 1987; Taylor, 1989). The relative importance of exchanger-mediated and ATP-driven Ca²⁺ transport in active Ca²⁺ absorption may differ among tissues. In rat intestinal and renal epithelia, the ATP-driven Ca²⁺ pump is responsible for the major part of the transcellular Ca²⁺ transport (Ghijsen et al. 1983; Van Heeswijk et al. 1984); the exchanger appears to provide an emergency system for removal of excess Ca²⁺ from the cytosol (Van Os, 1987). In the intestine of tilapia, the species studied here, however, the exchanger forms the main driving force for the absorption of Ca²⁺ (Flik et al. 1990).

No data are available on Na⁺/Ca²⁺ exchange activity in branchial Ca²⁺ uptake. We present here the first direct evidence for a basolateral Na⁺/Ca²⁺ exchanger in fish branchial epithelium in studies with vesicles isolated from plasma membranes. In previous studies on the ATP-driven Ca²⁺ pump in fish gills (Flik et al. 1985c) it was shown that the capacity of this pump was sufficient to explain the whole-body uptake. The finding of exchange activity in membranes containing Ca²⁺-ATPase encouraged us to determine its importance relative to that of ATP-driven transport.

Bearing in mind the differences in Na⁺ turnover in freshwater and seawater fish (Maetz, 1974; Evans, 1979) and the presence of a sodium-dependent Ca²⁺ transport mechanism in gills (in addition to the classical ATP-dependent Ca²⁺ pump), we compared these mechanisms in gills of freshwater- and seawater-adapted tilapia. We thus hoped to obtain information about the physiological importance of both transporters in freshwater and seawater gills by examining possible adaptive responses in the enzyme kinetics.

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

Artificial sea water was prepared by dissolving natural sea salt (Wimex, Krefeld, Germany) to a density of 1.022 g l^-1. Fish were first transferred to half-strength sea water and, after 1 day, to full-strength sea water. Fish were kept in sea water for at least 8 weeks before experimentation (8.5 mmol l^-1 Ca²⁺, 25 °C). Food was supplied at 1.5 % body mass per day.

For the analysis of blood plasma calcium, sodium, potassium and choride contents, a blood sample was taken by puncture from the caudal vessels using a heparinized syringe with a 23-gauge needle. Blood cells were removed by centrifugation (1 min, 10 000 g) and the plasma was stored at -20 °C until further analysis. Plasma calcium contents were measured colorimetrically using a calcium kit (Sigma Diagnostics). Plasma sodium and potassium contents were determined with a flame photometer (FLM3, Radiometer) coupled to an autoanalyzer (SA-20, Skalar). Plasma chloride content was determined colorimetrically using the method of Zall et al. (1956).

One branchial operculum from each fish was removed and incubated for 1 h in a well-aerated 2 µmol l^-1 solution of 2-(dimethylaminostyryl)-1-ethylpyridiniumiodine (DASPEI), which stains the mitochondria-rich chloride cells. After rinsing, the inner opercular epithelium was examined in a Zeiss fluorescence microscope at a magnification of 250×. Cells were counted in 20 different squares of the opercular epithelium with a total surface area of 5 mm² per fish. The density of chloride cells is presented as the number of cells per surface area of the inner operculum.

Plasma membranes were isolated from gill tissue scraped off the gill arches as described in detail previously (Flik et al. 1985c). In brief, after homogenization of gill scrapings, blood cells and cellular debris were separated from the membrane fraction by centrifugation. The membranes (from the supernatant) were collected by ultracentrifugation. The fluffy top layer from the resulting pellet was resuspended in buffer using a Dounce homogenizer. This membrane suspension was centrifuged differentially and the basolateral membranes in the resulting supernatant were collected by pelleting. The procedure yields membrane suspensions in which 47–51 % of the membranes are resealed vesicles. Vesicles were kept cooled on ice and were used on the day of isolation. The protein content of the vesicle preparations was 1.86±0.4 g l^-1 (N=12), as determined with a commercial reagent kit (Bio-Rad) using bovine serum albumin as a reference. Protein recovery was 1.67±0.16 % (N=6) in freshwater preparations; in seawater preparations, protein recovery was slightly, though significantly, lower: 1.23±0.11 % (N=6; P<0.001). Na⁺/K⁺-ATPase activity in fish branchial basolateral plasma membrane vesicles (BLMVs) was determined as described previously for similar preparations of enterocytes (Flik et al. 1990). Na⁺/K⁺-ATPase activity was defined as the Na⁺- and K⁺-dependent ouabain-sensitive phosphatase activity. The orientation of the vesicles was determined on the basis of the acetylcholine esterase activity measured in intact and permeabilized vesicles (reflecting the inside-out oriented vesicles, IOVs) and the specific trypsin sensitivity of the cytosol-oriented part of the Na⁺/K⁺-ATPase (reflecting the right-side-out oriented vesicles, ROVs) as described by Flik et al. (1990). The orientation of the vesicles was similar to previously reported values (Flik et al. 1985c) and not significantly different for freshwater (19.2±4.2 % IOV; 29.8±4.4 % ROV, N=6) and seawater (22.5±4.8 % IOV; 26.8±6.4 % ROV) preparations. In the calculation of Ca²⁺ transport velocities of ATP- and Na⁺-gradient-driven mechanisms, no correction for IOV content was made (see also Discussion). For determination of total Na⁺/K⁺-ATPase and Ca²⁺-ATPase activities, the membrane vesicles were permeabilized with saponin (0.20 mg saponin per mg protein).

⁴⁵Ca²⁺-transport was determined in assay media that contained 0.5 mmol l^-1 EGTA, 0.5 mmol l^-1N-(2-hydroxyethyl)-ethylenediamine-N,N’,N’-triacetic acid (HEEDTA) and 0.5 mmol l^-1 nitrilotriacetic acid (NTA) as a Ca²⁺ buffering system. Free calcium and magnesium levels were calculated using a matrix computer program (Schoenmakers et al. 1992) taking into account the first and second protonations of the respective ligands (ATP, EGTA, HEEDTA, NTA). In this program, the metal–chelator stability constants are corrected for the ionic strength and the temperature of the medium. All incubations were performed at 37 °C, which is the optimum assay temperature.

The ATP-driven transport of ⁴⁵Ca²⁺ was determined in 1 min incubations (yielding initial pump velocities; Flik et al. 1985c) of membrane vesicles (45 µg per assay) in the presence of 3 mmol l^-1 ATP, 150 mmol l^-1 KCl, 20 mmol l^-1 Hepes/Tris, pH 7.4, 0.8 mmol l^-1 Mg²⁺ and the Ca²⁺ buffers mentioned above. ⁴⁵Ca²⁺ uptakes were corrected for uptake in the absence of ATP. Free Ca²⁺ concentrations were varied from 50 nmol l^-1 to 1 µmol l^-1. The uptake was quenched by a 14-fold dilution of incubate in ice-cold isotonic buffer containing 20 mmol l^-1 Tris–HCl (pH 7.4), 150 mmol l^-1 KCl or NaCl and 0.1 mmol l^-1 LaCl₃. Vesicles were collected by a rapid filtration technique (Van Heeswijk et al. 1984; using Schleicher & Schüll ME25 filters, pore size 0.45 µm). Filters were rinsed twice with 2 ml of ice-cold medium and transferred to counting vials. 4 ml of Aqualuma was added per vial, filters were allowed to dissolve (30 min at room temperature) and radioactivity was determined in a Pharmacia Wallac 1410 liquid scintillation counter.

Na⁺/Ca²⁺ exchange activity was assayed as described previously (Flik et al. 1990). Briefly, 5 µl of prewarmed membrane vesicles (9–12 µg protein), equilibrated with 150 mmol l^-1 NaCl, was added to 120 µl of medium containing ⁴⁵Ca²⁺ and either 150 mmol l^-1 NaCl (blank) or 150 mmol l^-1 KCl. Ionic Ca²⁺ concentrations ranged from 250 nmol l^-1 to 25 µmol l^-1. After 5 s at 37 °C, the reaction was stopped by addition of 1 ml of ice-cold isotonic stopping solution (as above) containing 1 mmol l^-1 LaCl₃. Vesicles were collected as described above for the ATP-dependent Ca²⁺ transport assay. The difference in ⁴⁵Ca²⁺ accumulation after transfer to Na⁺ or K⁺ medium 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 from best fits of the curves by non-linear regression analysis (with the Levenberg–Marquardt method) of the velocities measured as a function of substrate concentration (computer program by Leatherbarrow, 1987). Values for preparations of individual fish were averaged (N=5–7) and used for statistical assessment of differences.

Statistical significance of differences between mean values was tested by the Mann–Whitney U-test. Statistical significance was accepted at P<0.05.

Plasma osmolarity is 6 % higher in tilapia held in sea water than in those held in fresh water; plasma calcium, sodium and chloride concentrations are slightly, but not significantly, different from the values found in freshwater tilapia (Table 1), illustrating the euryhalinity of this species.

As shown in Table 2, the total and specific Na⁺/K⁺-ATPase activities in the branchial epithelial homogenate of freshwater- and seawater-adapted tilapia are not significantly different. The Na⁺/K⁺-ATPase specific activity in the basolateral plasma membrane (BLM) fraction was 46 % higher in seawater preparations than in freshwater preparations but the total Na⁺/K⁺-ATPase was not significantly different for freshwater gills and seawater gills (indicating better purification of BLMs from seawater gills). Using the purification values for Na⁺/K⁺-ATPase activity in Table 2, we calculate a 15.20/9.36=1.6 times higher purification for this BLM marker of seawater gills than of freshwater gills (P<0.05). This means that the BLMs from seawater gills are better purified than those from freshwater gills when they are isolated using the same procedure.

The numbers of chloride cells in the opercular membrane, which give a good reflection of the numbers in the gills (Wendelaar Bonga et al. 1990), are not significantly different for freshwater and seawater tilapia (Table 2).

To compare the Ca²⁺ transport capacity in freshwater and seawater fish, we have expressed exchange relative to Na⁺/K⁺-ATPase activity (as a more specific reference). Using this comparison, the V_max values (V_max,corr) are not significantly different for freshwater- and seawater-derived BLMs (V_max,corr=V_max,_Ca/V_{spec,Na/K-ATPase}; 19.56± 4.80 nmol Ca²⁺µmol^-1 P_i for FW and 16.92±4.80 nmol Ca²⁺µmol^-1 P_i for seawater gills), which means that the density of the exchanger relative to Na⁺/K⁺-ATPase activity is similar. The affinities for Ca²⁺ (K_m) of the branchial Na⁺/Ca²⁺ exchangers are also similar in BLMVs from freshwater- and seawater-adapted tilapia (Fig. 1).

The V_max of the Ca²⁺ pump, expressed relative to Na⁺/K⁺-ATPase activity, does not differ in BLMVs from freshwater and seawater gills (Fig. 2), indicating similar pump densities (V_max,corr=V_max,_Ca/V_{spec,Na/K-ATPase}; 7.50±1.08 nmol Ca²⁺µmol^-1 P_i for freshwater and 7.38±0.84 nmol Ca²⁺µmol^-1 P_i for seawater gills). When the values are not corrected for the difference in purification, the V_max in seawater BLMs appears to be higher than in freshwater BLMs (P<0.005). The affinity of the pump for Ca²⁺ was significantly lower in BLMVs from seawater tilapia than in BLMVs from freshwater tilapia (Fig. 2).

High-affinity Ca²⁺-ATPase activity, determined as Ca²⁺-dependent ATP hydrolysis (Flik et al. 1984), amounts to 2.28±0.64 µmol P_i h^-1 mg^-1 and 3.87±0.85 µmol P_i h^-1 mg^-1 (N=6, ± S.E.M.) for freshwater and seawater BLMs, respectively. The ratio of this Ca²⁺-ATPase activity to Na⁺/K⁺-ATPase activity is not significantly different for BLMs obtained from freshwater and seawater gills (2.28/37.8=0.060 and 3.87/55.3=0.070, respectively), which indicates a similar density of the Ca²⁺-ATPase under both conditions.

The kinetic data presented in Figs 1 and 2 are used to plot the potential activities of both Ca²⁺ extrusion systems as a function of intracellular Ca²⁺ concentrations in Fig. 3. These results predict that, at resting levels of intracellular Ca²⁺ of approximately 10⁻⁷ mol l^-1, the Ca²⁺ pump activities exceed the Na⁺/Ca²⁺ exchange activities in freshwater and seawater BLMs by factors of 1.8 and 1.5, respectively. At 10⁻⁶ mol l^-1 Ca²⁺, the activity of the exchanger exceeds that of the pump by a factor of 1.2 in freshwater and in seawater BLMs.

The activities of three ion carriers in the basolateral membrane from gill epithelium of freshwater and seawater tilapia were measured in vitro. The Na⁺/Ca²⁺ exchange activity in this membrane displayed similar kinetics (V_max, K_m) in freshwater and seawater gills. The ATP-driven Ca²⁺ pump showed a somewhat lower affinity for Ca²⁺ in BLMs isolated from seawater gills than in BLMs from freshwater gills, whereas V_max was the same. The significance of this change in K_m for cellular calcium transport is discussed below (Ca²⁺transport across the basolateral membrane). The V_max values were expressed relative to Na⁺/K⁺-ATPase activity to eliminate purification-dependent differences between BLM preparations. This correction allows a comparison of the relative density of the carriers between freshwater and seawater gills. Because the V_max values of the exchanger and the Ca²⁺ pump do not differ in BLMs from freshwater and seawater fish, it follows that the density of these carriers is similar in freshwater and seawater gills. Total exchanger capacities, calculated as the product of V_max and total protein in the homogenate (and adjusted to 100 % IOV and 100 % recovery), amount 2297±566 and 2294±650 µmol Ca²⁺ h^-1 at 37 °C in freshwater and seawater gills, respectively. Total Ca²⁺ pump capacities, calculated in the same way, amount to 877±125 and 998±113 µmol Ca²⁺ h^-1 in freshwater and seawater branchial epithelium (average total protein in homogenates was 92 mg for freshwater gills and 103 mg for seawater gills). The branchial Ca²⁺ influx in vivo for a tilapia of the same size (approximately 7.5 µmol h^-1 at 25 °C; Flik et al. 1985a; G. Flik, personal observations) is a factor 212 [carrier capacity (2297+877) corrected for the temperature difference assuming an activation energy of 57.9 kJ mol^-1, as reported for the eel gill (Flik et al. 1984) divided by 7.5] lower than the total Ca²⁺ transport capacity in the plasma membrane compartment. This indicates a significant overcapacity to transport Ca²⁺.

In this study, we have compared freshwater tilapia with fish adapted to sea water for 8–12 weeks. The plasma composition (Table 1) shows slight, insignificant, increases in calcium, sodium and chloride concentrations which, collectively, appear to be responsible for the 6 % increase in plasma osmolarity in seawater tilapia. These observations indicate successful ionoregulation under these widely differing water conditions and thus illustrate the euryhalinity of this fish species.

Mixed populations of of right-side-out (ROV), inside-out (IOV) and leaky BLMVs were used in the in vitro determinations. One could argue that the activities of the pump and the exchanger should only be compared directly if the Na⁺/Ca²⁺ exchange activity is measured in IOVs only, as is the case for the ATP-dependent Ca²⁺ transport. This has indeed been shown in several mixed-vesicle preparations, in Ca²⁺ affinity studies (Schoenmakers and Flik, 1992) and in an experiment in which chemical blocking of the cytosolic side of the exchanger completely inhibited Na⁺/Ca²⁺ exchange in a mixed population of IOVs and ROVs (Li et al. 1991).

We can only speculate about the reason for the difference in branchial BLM purification between seawater and freshwater gills (Table 2). It is plausible that the looser character of the branchial epithelium in sea water (Sardet et al. 1979; Foskett et al. 1981) leads to a more effective dispersion of the cells in the isolation procedure, which subsequently results in a higher yield of basolateral membranes. Changes in the lipid fraction of the gills after a shift from fresh water to sea water, as reported for eels (Hansen, 1975), could also influence the effectiveness of our BLM isolation procedure.

Although the V_max of the exchanger is roughly 2.5 times higher than that of the Ca²⁺ pump, the exchanger’s relative contribution to Ca²⁺ excretion at a physiological cellular Ca²⁺ concentration will be lower than that of the pump because of its low affinity for Ca²⁺. Only at concentrations of 10⁻⁶ mol l^-1 or more does the Na⁺/Ca²⁺ exchanger becomes a major transport system (from cytoplasm to blood) (Fig. 3). We conclude, on the basis of kinetic data obtained from isolated plasma membranes, that Ca²⁺ transport may depend for a substantial part on the Na⁺/Ca²⁺ exchanger (giving the model depicted in Fig. 4), because the relative contribution of the exchanger to the transport of Ca²⁺ is intermediate when compared with its minor role in rat intestine and rat kidney (Ghijsen et al. 1983; Van Heeswijk et al. 1984) and its major role in tilapia intestine (Flik et al. 1990). At concentrations higher than 10⁻⁶ mol l^-1, which are unlikely to occur intracellularly, the exchanger becomes the major Ca²⁺ transport system. In this in vitro study, the presence of the Ca²⁺ transporters and their kinetic properties have been determined. Clearly, other studies with intact epithelia are needed to obtain direct evidence for the contribution to Ca²⁺ transport of each transporter in vivo. The effect of ouabain on net calcium transport should be evaluated, as was done for rat small intestine (Ghijsen et al. 1983) and fish intestine (Flik et al. 1990), to determine the Na⁺-dependency (serosal) of the Ca²⁺ uptake.

The lower Ca²⁺ affinity of the Ca²⁺ pump in BLMs from seawater gills compared with that of freshwater gills suggests a more important role for the exchanger in Ca²⁺ influx in sea water (but with no change in affinity). This may mean that, in sea water, the Ca²⁺ influx depends on the Na⁺ status to a greater extent than it does in fresh water. However, this extra contribution of exchange activity in seawater BLMs is only marginal at intracellular Ca²⁺ concentrations around 10⁻⁷ mol l^-1. The difference would only become manifest if the intracellular Ca²⁺ concentration rose (to 10⁻⁶ mol l^-1 or more; Fig. 3). We realize, however, that at present we can only evaluate the contribution of the exchanger to Ca²⁺ extrusion on the basis of its Ca²⁺ affinity. Changes in Na⁺ affinity may occur, raising the possibility of another mode of regulation. It would, therefore, be very interesting to evaluate the Na⁺-dependency of the exchanger to see whether the regulation is at the Na⁺ site.

A major conclusion drawn from these results is that the fully adapted seawater fish possesses amounts of the exchanger, the Ca²⁺ pump and the Na⁺/K⁺-ATPase that are comparable to those of freshwater fish. Apparently, these mechanisms suffice to accomplish Na⁺/Cl^- extrusion in sea water and to maintain a similar Ca²⁺ influx to that in freshwater fish [the rates of branchial Ca²⁺ influx do not differ for freshwater and seawater fish; data from Mayer-Gostan et al. (1983) on killifish and G. Flik (unpublished data) on eel]. The similar levels of Ca²⁺ transport capacity indicate that, in both freshwater and seawater gills, Ca²⁺ may be absorbed transcellularly. This assumption is further supported by two findings. First, stanniocalcin, the hypocalcaemic fish hormone, circulates at higher concentrations and has a higher turnover rate in seawater fish (Hanssen et al. 1992, 1993) and it is known to exert its inhibitory control through a reduction in transcellular Ca²⁺ uptake, also in tilapia (Verbost et al. 1993). Second, the transepithelial potential (TEP) measured in seawater fish is never less than the resting potential for Ca²⁺ in the gill epithelium of fish in sea water, thus excluding a passive influx of Ca²⁺ in seawater fish (resting potential is 21 mV calculated according to the Nernst equation with 8.5 mmol l^-1 Ca²⁺ outside and 1.5 mmol l^-1 Ca²⁺ inside, whereas the TEP in seawater fish is greater than 21 mV, based on TEP values given by Maetz and Bornancin, 1975; Young et al. 1988).

The total branchial Na⁺/K⁺-ATPase activity of freshwater tilapia was similar after adaptation to sea water, in contrast to what has been reported for killifish (Epstein et al. 1967), eels (Epstein et al. 1971; Bornancin and de Renzis, 1972; Utida and Hirano, 1973), coho salmon (Zaugg and McLain, 1970) and some other species (reviewed by Foskett et al. 1983). It is believed that the capacity of fish to tolerate hyperosmotic media is linked with their ability to increase the production of this enzyme, thereby gaining a more efficient system for ridding the body of excess salts (Maetz, 1971). When tilapia are gradually transferred to sea water (as performed in this study), the Na⁺/K⁺-ATPase activity increases only slightly (51 % after stepwise acclimation to sea water in 8 weeks; Dange, 1985) or does not increase at all (this study). We did not measure a difference in Na⁺/K⁺-ATPase activity 8–12 weeks after the transfer, although there is a transient increase in the activity (2 weeks after transfer, 40±8 % increase, P<0.01, N=8; G. Flik, personal observations). The densities of chloride cells in the opercular membrane of the freshwater and seawater tilapia were also alike. These cell numbers match data on enzymatic activity and are in agreement with data by Foskett et al. (1981) showing that, in tilapia opercular membranes, the number of chloride cells increases only transiently after transfer to sea water. Microscopical evaluation indicated a subsequent hypertrophy of the chloride cells, which was suggested to be correlated with the amount of chloride secretion (Foskett et al. 1981), but the corresponding Na⁺/K⁺-ATPase activity was, unfortunately, not measured. One may conclude from the available data that the differences in branchial Na⁺/K⁺-ATPase are not necessarily indicative of osmoregulatory responses during seawater adaptation in tilapia (Dange, 1985). In any case, the quantity of Na⁺/K⁺-ATPase available for Na⁺ uptake in freshwater gills suffices to sustain the Na⁺ and (indirectly) Cl^- excretion in seawater gills, which is much more rapid than the Na⁺ uptake in fresh water (Maetz, 1971). This implies that, in freshwater conditions, an unknown but small proportion of the Na⁺/K⁺-ATPase capacity is used in vivo. This, in turn, suggests that the difference in gill physiology between the freshwater and seawater situations is dictated by properties of the epithelium other than the capacity of these plasma membrane carriers. Such other properties could include transepithelial potential (Marshall and Bern, 1980), the permeability of the apical membrane and the presence of other carriers (we have preliminary evidence that, in sea water, a basolateral bumetanide-inhibitable Na⁺/K⁺/Cl^- co-transporter is activated and accounts for the high rate of Cl^- extrusion, as originally proposed by Silva et al. 1977).

The research of P.M.V. has been made possible by a fellowship of the Royal Netherlands Academy of Arts and Sciences (KNAW).