Calcium transport in gill plasma membranes of the crab Carcinus Maenas: Evidence for carriers driven by ATP and A Na⁺ Gradient

The branchial epithelium of crustacean gills forms a delicate barrier between the external and internal media and serves an important role in respiration, acid–base balance, volume regulation and the transport of ions. Intensive research on Na⁺ and Cl⁻ handling by crustaceans living in marine and brackish waters has been carried out (Mantel and Farmer, 1983; Lucu, 1990, 1993; Towle, 1993), but only a few reports on branchial Ca²⁺ handling are available at the moment (Greenaway, 1985; Cameron, 1989, 1990).

Studies on Ca²⁺ metabolism in crustaceans have focused on the processes of calcification in the carapace. Roer (1980) has shown that, during the late premoult stage, Ca²⁺ is resorbed from the carapace and subsequently transported across the hypodermis into the haemolymph. In aquatic crustaceans, most of the body Ca²⁺ is lost during this period via the gills and with the exuvia (Greenaway, 1974a, 1985), softening the exoskeleton and allowing ecdysis. In the postmoult stage, calcium remaining in the haemolymph is transported to the carapace across the hypodermis (Greenaway, 1974b, 1985). Moreover, crustaceans mobilize large amounts of calcium from the surrounding water during the postmoult period to calcify the enlarged cuticle and the carapace. This implies that the gills must be involved in inwardly directed (from water to haemolymph) and regulated Ca²⁺ transport.

In the hypodermal epithelium of postmoult blue crab Callinectes sapidus, Ca²⁺-ATPase activity increases compared with that measured in the intermoult stages (Cameron, 1989). However, the kinetic properties of this ATPase (an affinity for Ca²⁺ too low to play a role in intracellular calcium homeostasis) suggest that this activity is that of a non-specific Ca²⁺-and ATP-consuming phosphatase (with unknown function), similar to the one described in fish gills, and not that of a high-affinity Ca²⁺-ATPase involved in the extrusion of Ca²⁺ from the cell (Flik et al. 1984). The inward transport of calcium across the hypodermis of Carcinus maenas was postulated to occur through the action of an ATPase or of a Na⁺/Ca²⁺ exchanger for extrusion of Ca²⁺ from the cells (Roer, 1980). However, no direct biochemical evidence for Ca²⁺ transport across the plasma membrane, and thus for the existence of such carriers, has yet been reported. Morris and Greenaway (1992) have found evidence for high-affinity Ca²⁺-ATPase activity (K_0.5 for Ca²⁺ in the nanomolar range) in the gills of the crab Leptograpsus variegatus. The kinetics of this enzymatic activity indeed favour a role of branchial epithelial cells in Ca²⁺ transport and homeostasis. However, Morris and Greenaway (1992) measured ATPase activity and this was assumed to reflect the transport activity. As shown in a study on ‘pumping and non-pumping’ ATPase activity in rat liver plasma membranes (Lin and Russel, 1988), one has to be cautious when making this assumption.

The hypodermis plays a pivotal role in the recycling of Ca²⁺ from the carapace during moulting and in calcification of the carapace; however, Ca²⁺ uptake from the environment is required to augment the body calcium pool after moulting and to recalcify the carapace rapidly (Greenaway, 1985; Neufeld and Cameron, 1993). Uptake of calcium from the environment could take place via the gills and the gut. Ca²⁺ uptake may reach a maximum rate in postmoult blue crabs (Callinectes sapidus) that have not yet resumed eating (Neufeld and Cameron, 1993) and, in this case, must occur via the gills. It seems reasonable then, to postulate an active Ca²⁺ transport mechanism for uptake via the gills. In this study, we concentrated on Ca²⁺ transport by the gills at the membrane level.

Procedures developed for the isolation of membranes from fish branchial epithelium (Flik et al. 1984, 1985a,b; Perry and Flik, 1988) were applied to crab branchial epithelium and yielded a plasma membrane vesicle preparation with a high proportion of resealed vesicles (more than 70%), but too low a number of inside-out membrane vesicles (IOVs) to demonstrate Ca²⁺ transport. An additional hypotonic shock, as originally suggested for the preparation of red blood cell membrane IOVs (Sarkadi et al. 1980), yielded 22% IOVs in the crab membrane preparation and allowed us to demonstrate ATP-and Na⁺-gradient-driven Ca²⁺ transport.

We used intermoult stages of the euryhaline crab Carcinus maenas adapted for at least 2 weeks to 50% sea water as Ca²⁺ transport was predicted to be enhanced under these conditions (Neufeld et al. 1980).

Green shore crabs, Carcinus maenas, were collected early in January by netting in the Waddensea, The Netherlands. The animals were transferred to the laboratory and kept in running sea water at 6°C, the temperature of the Waddensea at the time the crabs were caught. The mass of the animals ranged from 10 to 30 g. Only male specimens in intermoult stage were used. At least 14 days before experimentation, the sea water was diluted to 50% with demineralized water over a 24 h period. The crabs were fed cod strips daily and kept at a photoperiod of 16 h:8 h (L:D). Less than 5% mortality was observed. Haemolymph was collected by puncture of an infrabranchial sinus near the base of a walking leg. The haemolymph was centrifuged for 3 min at 9000 g and the plasma stored at -20°C until assay. After the carapace had been opened, posterior gills 6–9 were dissected out and quickly cooled on ice. The posterior gills contain an abundance of mitochondria-rich cells (Compère et al. 1989), which are the equivalent of the ionocytes found in other ion-transporting epithelia. The Na⁺ contents of water and plasma were determined by flame photometry on 200-fold diluted samples using a Technikon model IV Autoanalyzer; LiNO₃ (5 mmol l⁻¹) was used as an internal standard. Water and plasma total calcium were determined with an endpoint colorimetric assay kit (Sigma, no. 587). A combined calcium–phosphorus standard set (Sigma, no. 360-11) was used as reference.

Gills from 2–4 animals were pooled to obtain 2–5 g of tissue. Tissue was disrupted by 20 strokes in a Douncer homogenization device equipped with a loosely fitting pestle in 10 ml g⁻¹ of a hypotonic buffer containing 12.5 mmol l⁻¹ NaCl, 1 mmol l⁻¹ Hepes/Tris (pH 8), 1 mmol l⁻¹ dithiothreitol, 0.5 mmol l⁻¹ EDTA and the serine protease inhibitor aprotinin (280 i.u l⁻¹). The resulting suspension was filtered over cheesecloth to remove particulate material and branchial septa. This filtrate was designated the original homogenate (H₀); 0.5 ml was set aside on ice for further assay and the remainder was centrifuged at 550 g for 15 min to remove nuclei and cellular debris. The resulting supernatant was centrifuged for 30 min at 50 000 g to collect the membranes. The pelletized material consisted of a brownish part fixed against the wall of the centrifugation tube (this part contains most of the mitochondrial fragments, indicated by around 90% of the total branchial succinic acid dehydrogenase activity) and, above this, a whitish fluffy layer. The fluffy layer was carefully collected by mild swirling in a quantity of buffer (10 ml g⁻¹ starting material) containing 0.5 mmol l⁻¹ Tris/HCl (pH 8.5), 20 μmol l⁻¹ EDTA and 50 μmol l⁻¹ β -mercaptoethanol (Sarkadi et al. 1980). The membranes were resuspended in the buffer by 20 strokes in the Douncer and left first for 30 min on ice and subsequently for 15 min at 37°C. Every 5 min during this 45 min incubation period the suspension was vigorously agitated for 15 s. The suspension was then cooled again to 0°C and centrifuged for 10 min at 10 000 g (to remove remaining mitochondrial fragments); the resulting supernatant was centrifuged for 30 min at 50 000 g to sediment the plasma membranes (see below). The pellet (P₃) was resuspended in 400 μl of buffer containing 150 mmol l⁻¹ KCl or NaCl, 0.8 mmol l⁻¹ MgCl₂ and 20 mmol l⁻¹ Hepes/Tris (pH 7.4). Vesiculation was secured by 15 passes through a 23 gauge needle fitted to a tuberculin syringe. Membranes in potassium buffer were used for ATP-dependent Ca²⁺ transport assays, membranes in sodium buffer for Na⁺-gradient-driven Ca²⁺ transport and marker enzyme determinations. Transport assays were carried out on the day of isolation. The membranes were kept on ice. When 0.5 ml of P₃ suspension was prepared, the protein content ranged from 1 to 1.5 mg ml⁻¹ bovine serum albumin (BSA) equivalents (see below).

The protein content of the membrane preparation was determined according to Bradford’s procedure (BioRad, catalogue no. 500-0002), using bovine serum albumin (BSA) as a reference. Two marker enzymes were used to characterize the membrane preparation: succinic acid dehydrogenase (SDH) for mitochondrial membrane fragments and ouabain-sensitive, sodium-and potassium-dependent ATPase (Na⁺/K⁺-ATPase) for basolateral plasma membranes (Flik et al. 1990). K⁺-dependent, ouabain-sensitive p-nitrophenylphosphatase (K⁺-pNPPase) activity (which reflects the dephosphorylation step of the Na⁺/K⁺-ATPase) was determined as the difference between the amounts of inorganic phosphate (P_i) released from pNPP in a medium containing 20 mmol l⁻¹ Tris/Hepes (pH 7.4), 120 mmol l⁻¹ NaCl, 10 mmol l⁻¹ KCl and 3 mmol l⁻¹ Na₂pNPP and the amount of P_i released in the same medium but without KCl and supplemented with 1 mmol l⁻¹ ouabain. The incubation was carried out at 37°C for 30 min.

Three detergents were tested for their suitability to unmask enzymatic activities that become latent as a result of membrane resealing. As shown in Fig. 1, digitonin and 3-[(3-cholamidopropyl)dimethylammonio]-1-propane-sulphonate (Chaps) had no effect, or a slightly inhibitory effect, on Na⁺/K⁺-ATPase respectively, at concentrations where a clear stimulation was obtained with saponin. The optimum saponin concentration was 0.2 mg per milligram membrane BSA equivalents, and saponin at this concentration was used in all instances when determination of optimal activity was required.

Data on recovery (ratio of total activities in H₀ and P₃) and enrichment (ratio of specific activities in H₀ and P₃) of marker enzymes for the isolation procedure are given in Table 1. The protein recovery was 2.8±0.5% BSA equivalents, the Na⁺/K⁺-ATPase activity recovery was 12.9±3.6% and the K⁺-pNPPase recovery 17.7±2.9%. No purification occurred and low recovery was observed for SDH activity. We conclude from these results that the P₃ fraction obtained by this procedure reflects a plasma membrane preparation.

Membrane orientation was analyzed on the basis of K⁺-pNPPase and Na⁺/K⁺-ATPase activity (Flik et al. 1990). Briefly, saponin was used to unmask enzymatic activities that are latent as a result of membrane resealing. The K⁺-pNPPase activity was used to identify inside-out membrane vesicles (IOVs). The K⁺-pNPPase assay replaced the assay of the exoenzyme acetyl cholinesterase that we have successfully applied to fish (Flik et al. 1990) and human red cell (Verbost et al. 1989) preparations; esterase activity proved to be too low to measure in our crab gill plasma membrane preparation. The trypsin-sensitivity of the cytosolic part of the Na⁺/K⁺-ATPase was the basis for the identification of the right-side-out membrane vesicles (ROVs): after trypsin treatment, only ROVs are left with an intact Na⁺/K⁺-ATPase that may subsequently be unmasked by saponin. Inactivation of Na⁺/K⁺-ATPase activity in IOVs and leaky membrane fragments was optimal when they were treated with 4500 BAEE trypsin (Sigma, T0134) units per milligram membrane protein for 30 min at 25°C. Trypsin activity was stopped by addition of 25 mg ml⁻¹ soybean trypsin inhibitor (Sigma, T9253). In controls, the inhibitor was added before the addition of trypsin.

In all preparations saponin enhanced the K⁺-pNPPase activity. Expressed as a percentage, the enhancement (mean values for specific activity V_spec were 13.6±5.9 and 16.0±6.8 μmol P_i h⁻¹ mg⁻¹ protein, in the absence and in the presence of saponin, repectively) was 18.6±2.6% (N=5), indicating 19% of IOVs in the preparation. The trypsin-insensitive Na⁺/K⁺-ATPase activity (V_spec 22.7±4.3 μmol P_i h⁻¹ mg⁻¹ protein) was 33±4% (N=5) of the total activity (V_spec 69.6±12.9 μmol P_i h⁻¹ mg⁻¹ protein), indicating 33% of ROVs in the preparation. In untreated membranes, the percentage resealing in P₃, indicated by the increase in Na⁺/K⁺-ATPase (from 32.2±6.1 to 79.3±11.7 μmol P_i h⁻¹ mg⁻¹ protein) upon saponin treatment was 59±12%. This value did not differ (P>0.15) from the value calculated for control-treated membranes. In the latter, the percentage resealing indicated by the increase in Na⁺/K⁺-ATPase (from 31.8±5.5 to 69.6±12.9 μmol P_i h⁻¹ mg⁻¹ protein) upon saponin treatment was 55±10%; from this value we calculated that there were 45±10% leaky membrane fragments in the preparation and 22±8% IOVs (percentage of resealed vesicles minus percentage of ROVs). The value for IOVs derived from Na⁺/K⁺-ATPase activity is in good agreement with the 18.6±2.6% IOVs determined directly by the K⁺-pNPPase activity assay (Table 2). The average configuration of the membrane preparation was therefore 22% IOVs, 33% ROVs and 45% leaky membrane fragments.

Applying the basolateral plasma membrane isolation method for fish gills to crab gills, we found a consistent, high degree of membrane resealing in the plasma membrane fraction, 73.5±4.2%. However, the percentage of ROVs (71.1±4.3%) did not differ significantly (P>0.15) from the percentage of resealed membranes. The calculated percentage of IOVs in these preparations was 3.8±2.2% (N=5). In accordance with this low percentage of IOVs, we failed to show ATP-driven Ca²⁺ transport in these preparations (data not shown).

Ca²⁺ transport was determined as described in detail by Flik et al. (1990). In brief, a rapid filtration procedure was used to determine uptake of ⁴⁵Ca²⁺ into vesicles. The composition of the assay medium was (final concentrations, in mmol l⁻¹): Hepes/Tris, 20 at pH 7.4; NaCl, 150; Tris–ATP, 0 or 3; EGTA, 0.5; N-(2-hydroxyethyl)-ethylenediamine-N,N′,N′-triacetic acid (HEEDTA), 0.5; nitrilo-triacetic acid (NTA), 0.5; free Mg²⁺, 0.8; free Ca²⁺, 5X10⁻⁵ to 2.5X10⁻³; NaN₃, 1.0; and oligomycin B (5 μg ml⁻¹); free Mg²⁺ and Ca²⁺ concentrations were varied by the addition of MgCl₂ and CaCl₂ and the actual free ion concentrations were calculated using the program Chelator (Schoenmakers et al. 1992b). The first and second protonations of the ligands in the calcium buffer (ATP, EGTA, HEEDTA and NTA) were taken into account and the stability constants were adjusted in accordance with the pH, temperature and ionic strength of the medium. The ⁴⁵Ca radioactive concentration was 1.0–1.6 MBq ml⁻¹. Incubation was carried out at the optimum in vitro temperature of 37°C. ATP-driven Ca²⁺ transport was defined as the difference in Ca²⁺ uptake in the presence and in the absence of ATP. The incubation was terminated by quenching the reaction with 1 ml of ice-cold isotonic stop buffer (150 mmol l⁻¹ KCl, 20 mmol l⁻¹ Hepes/Tris, at pH 7.4, and 1.0 mmol l⁻¹ LaCl₃). The filters (Schleicher & Schüll, ME 25) with retained ⁴⁵Ca²⁺ containing vesicles were washed three times with 2 ml of stop buffer. After 30 min in scintillation fluid (Aqualuma, Lumac), the filters had dissolved and the radioactivity collected on them was subsequently determined by liquid scintillation counting.

Na⁺-dependent Ca²⁺ transport across plasma membranes was assayed as the difference in ⁴⁵Ca²⁺ accumulation upon transfer of membrane vesicles equilibrated in 150 mmol l⁻¹ NaCl to a medium containing either 150 mmol l⁻¹ NaCl (blank) or 150 mmol l⁻¹ KCl. In preparing the media, a 25-fold dilution of the vesicle suspension was taken into account to yield the following composition (final concentrations in mmol l⁻¹): NaCl or KCl, 150; Hepes/Tris, 20 at pH 7.4; EGTA, 0.5; HEEDTA, 0.5; NTA, 0.5; free Mg²⁺, 0.8; and free Ca²⁺, 5X10⁻⁴ to 2.5X10⁻². Free Ca²⁺ and Mg²⁺ concentrations were calculated as described above. The ⁴⁵Ca radioactive concentration was 0.5–0.8 MBq ml⁻¹. A 5 μl vesicle suspension was mixed with 120 μl of medium, both prewarmed to 37°C. After 5 s of incubation, the reaction was stopped by addition of 1 ml of isotonic ice-cold stop buffer (150 mmol l⁻¹ NaCl, 20 mmol l⁻¹ Hepes/Tris at pH 7.4, 1.0 mmol l⁻¹ EDTA). Membrane vesicles with retained ⁴⁵Ca²⁺ were collected by filtration as described above.

Kinetic data were analyzed using Enzfitter, a non-linear regression data analysis program (Leatherbarrow, 1987). Significance of differences between mean values was assessed by the Mann–Whitney U-test. Data are presented as mean values ±1 standard deviation (S.D.). Significance was accepted when P<0.05.

Crabs kept for 2–4 weeks in 50% sea water, containing 5.3 mmol l⁻¹ Ca²⁺, had mean haemolymph Ca²⁺ levels of 9.77±1.13 mmol l⁻¹ (N=21). Apparently, these animals hyperregulated their internal Ca²⁺ levels. The low variation in plasma Ca²⁺ levels was taken as circumstantial evidence for the intermoult stage.

Membrane vesicles, characterized as the basolateral plasma membrane fraction, were isolated from posterior gills 6–9 of these intermoult crabs (see Materials and methods). As shown in Fig. 2, uphill (against a [Ca²⁺] gradient) ⁴⁵Ca²⁺ accumulation in the vesicles occurred in the presence of ATP and was abolished by the ionophore A₂₃₁₈₇. Addition of the ionophore resulted in a loss of ⁴⁵Ca²⁺ from the loading vesicles. The apparently linear increase in ⁴⁵Ca²⁺ uptake observed during the first minute indicated that the uptake at 1 min reflected the initial velocity of the transporter.

Kinetic analysis of the ATP-driven Ca²⁺ transport (1 min determinations) obeyed Michaelis–Menten kinetics for a single enzyme (Fig. 3). The K_m for Ca²⁺ was 149 nmol l⁻¹ and the maximum velocity was 1.73 nmol min⁻¹ mg⁻¹ protein.

As shown in Fig. 4, significantly more ⁴⁵Ca²⁺ was accumulated when Na⁺-loaded vesicles were transferred to a medium containing an equimolar concentration of K⁺ instead of Na⁺, suggesting the presence of a Na⁺/Ca²⁺ exchanger in these membranes.

Subsequently, we analyzed the Ca²⁺ kinetics of the exchanger (Fig. 5) to evaluate its activity relative to that of the ATP-driven Ca²⁺ pump. The Na⁺/Ca²⁺ exchange activity obeyed Michaelis–Menten kinetics for a single enzyme; V_max was 9.88 nmol min⁻¹ mg⁻¹ protein and K_m was 1.78 μmol l⁻¹. The relative contributions of these two calcium carriers to Ca²⁺ extrusion in a branchial Ca²⁺-transporting cell are presented in Fig. 6.

We provide here the first direct evidence for active Ca²⁺ transport in basolateral plasma membranes of the gills of a crustacean, i.e. an ATP-driven Ca²⁺ uptake and a Na⁺-gradient-driven Ca²⁺ extrusion mechanism. Both activities proved to be homogeneous (indicated by kinetics for a single enzyme), and their half-maximal activation by Ca²⁺ occurred at physiological cell Ca²⁺ concentrations. Two other important conclusions can be drawn from the data presented. First, the plasma membranes from green shore crab gills have a strong tendency to reseal predominantly as ROVs, which may hamper the direct demonstration of Ca²⁺ extrusion activity, which can only be measured in IOVs (Schoenmakers and Flik, 1992). However, when the membranes were subjected to a strong hypotonic shock (Sarkadi et al. 1980), sufficient IOVs were formed to measure such activities. Second, the activity of the Na⁺/Ca²⁺ exchanger exceeds that of the Ca²⁺-ATPase in the branchial plasma membrane and may, therefore, be predicted to play an important role in Ca²⁺ transport in this epithelium.

Euryhaline crabs (e.g. the green shore crab Carcinus maenas and the blue crab Callinectes sapidus) maintain high plasma osmotic values when confronted with decreasing salinities (Siebers et al. 1982; Mantel and Farmer, 1983; Lucu and Siebers, 1986; Shetlar and Towle, 1989). Neufeld and Cameron (1993) have given a convincing electrochemical evaluation of branchial Ca²⁺ exchange in the blue crab and postulated that active Ca²⁺ transport mechanisms must exist in its branchial epithelium. There is evidence that haemolymph Ca²⁺ levels can be maintained hyperionic to the medium when crabs are acclimated to dilute sea water (Carcinus maenas, Greenaway, 1976; this study), indicating that net uptake of Ca²⁺ from the water must occur. Indeed, in intermoult Carcinus mediterraneus, which is closely related to the green shore crab used in this study, kept under identical water conditions to those described in this paper, whole-body (branchial) Ca²⁺ influx (6.57±1.95 μmol h⁻¹ g⁻¹) significantly exceeded the efflux (3.15±2.63 μmol h⁻¹ g⁻¹; means ± S.D., N=6; C. Lucu, unpublished data). The resulting net Ca²⁺influx via the gills could explain the high Ca²⁺ concentration of the haemolymph. The transcellular movement of Ca²⁺ that must underlie this influx warranted further studies on Ca²⁺ transport mechanisms in the gills.

Na⁺/K⁺-ATPase activity has been localized exclusively in the basolateral plasma membrane of blue as well as green shore crab gill cells (Towle and Kays, 1986; Morris and Greenaway, 1992). Its enrichment in the membrane fraction used in this study, therefore, indicates that this fraction is enriched in basolateral plasma membranes. Moreover, SDH activity was not enriched in this fraction, indicating that mitochondrial membrane fragments had selectively been removed.

The values for enzyme enrichment and recovery of plasma membranes calculated on the basis of Na⁺/K⁺-ATPase and K⁺-pNPPase were comparable, which strongly suggested that these activities indeed reflect the activity of the same enzymatic entity. This allowed a direct comparison of these activities, as used in the membrane orientation assays. The value of 22% for IOVs, calculated indirectly on the basis of the trypsin treatment of membranes, was strongly supported by the comparable value of 19% of IOVs determined directly on the basis of the K⁺-pNPPase activity (Table 2). The configuration of the plasma membrane fragments in the preparation was therefore 22% IOVs, 33% ROVs and 45% leaky fragments. These values compare very well with those obtained for plasma membrane vesicles prepared from fish tissues (Flik et al. 1985a, 1990), which have been shown to be suitable for the demonstration of membrane Ca²⁺ transport. It should be noted that others (Towle and Hølleland, 1987; Shetlar and Towle, 1989; Hølleland and Towle, 1990) have succeeded in isolating plasma membrane preparations from blue and green shore crab gills with sufficient IOVs to demonstrate energized Na⁺ transport, applying procedures not very different from those normally applied to fish gill epithelium, but that yielded insufficient IOVs for measurements of Ca²⁺ transport. This suggests that species differences may exist with respect to membrane behaviour during isolation procedures and, therefore, that these procedures need careful evaluation.

Our kinetic studies indicated that both the ATP-driven and the Na⁺-gradient-driven Ca²⁺ transport obeyed Michaelis–Menten kinetics and thus were the expression of a single class of carriers. This was taken as further evidence that the enzymatic activities in this membrane fraction reflect those present in the basolateral plasma membrane of the ion-transporting cells. The maximum velocities of the transporters are of the same order of magnitude as those found in, for instance, fish gills and intestine (Flik et al. 1990; Verbost et al. 1994) and suggest a Ca²⁺ transport capacity sufficient to explain the Ca²⁺ influx seen in these tissues.

Clearly, the Na⁺/Ca²⁺ exchange activity in the crab gills may be an important mechanism for Ca²⁺ transport. Fig. 6 shows the activities of the two plasma membrane Ca²⁺ extrusion mechanisms as a function of cytosolic Ca²⁺ levels. For the exchange activity, it was assumed that the Na⁺ gradient was not limiting its activity. At cytosolic Ca²⁺ concentrations up to 500 nmol l⁻¹, the rate of Ca²⁺-ATPase-mediated transport exceeds that of exchanger-mediated transport. With cytosolic Ca²⁺ concentrations exceeding 500 nmol l⁻¹, the rate of export of Ca²⁺ from the cell depends predominantly on the activity of the exchanger. This situation is strongly reminiscent of that in the tilapia enterocyte, where Na⁺/Ca²⁺ exchange determines the rate of extrusion of Ca²⁺ from the cell (Flik et al. 1990; Schoenmakers et al. 1992a, 1993). The crab gill may represent a second tissue in which Ca²⁺ transport is, to a large extent, mediated by a Na⁺/Ca²⁺ exchanger. The occurrence of these two Ca²⁺ transport mechanisms in a crustacean tissue lends support to the prediction from physiological experiments (Roer, 1980) that Ca²⁺ transport in the green shore crab hypodermis must be mediated by ATP-dependent as well as Na⁺-dependent transport mechanisms. Further studies on the characterization of Ca²⁺ transport mechanisms in crustacean tissues seem warranted. In particular, the regulation of these mechanisms in gills and hypodermis during the moulting cycle may yield interesting new insights into the functioning of Ca²⁺ transporters. Euryhaline crabs may provide a particularly powerful model for the study of the Na⁺-dependence of Ca²⁺ transport (when determined by the exchanger). The hyperregulation of plasma Na⁺ in (these) crustaceans suggests that these animals may depend on a guaranteed inward Na⁺ gradient for Ca²⁺ extrusion, irrespective of the ion composition of the ambient medium.

Wies Vonck is gratefully acknowledged for organizing the mid-winter collection of crabs and their transport to the laboratory in Nijmegen and Tom Spanings for his outstanding organization of the crab husbandry. C. Lucu received a travel fund from the University of Zagreb.