The Movement of Cadmium Through Freshwater Trout Branchial Epithelium and its Interference with Calcium Transport

Ca²⁺ uptake in freshwater fish mainly occurs via the gills (Flik et al. 1985). Ca²⁺ transport across the gills, a tight ion-transporting epithelium, follows a transcellular route (Perry & Flik, 1988). Micromolar concentrations of cadmium (Cd) in the water inhibit branchial Ca²⁺ uptake (Verbost et al. 1987a, Reid & McDonald, 1988) and induce hypocalcaemia (Giles, 1984; Pratap et al. 1989). We have advanced circumstantial evidence that the inhibition of Ca²⁺ uptake by waterborne Cd²⁺ may result from a competitive inhibition by cytosolic Cd²⁺ of the Ca²⁺ pump in the basolateral membrane of the Ca²⁺-transporting cells in the gills (Verbost et al. 1988).

Cd²⁺ uptake from the water in freshwater fish mainly occurs via the gills (Williams & Giesy, 1978). Like Ca²⁺, Cd²⁺ enters the fish predominantly via a transcellular route because of the tight character of the branchial epithelium (Pärt, 1983). Interference of Cd²⁺ with the Ca²⁺ influx route may occur at at least three sites: the apical membrane, where Ca²⁺ enters the cell via Ca²⁺ channels (Perry & Flik, 1988), the intracellular Ca²⁺ buffering systems (Flik et al. 1985) and the basolateral membrane, where Ca²⁺ is translocated to the blood by a high-affinity Ca²⁺ pump (Flik et al. 1985). Levels of Cd²⁺ in fresh water that cause hypocalcaemia (0·1 μmol l⁻¹ Cd in water containing 0·7 mmol l⁻¹ Ca) do not instantly inhibit Ca²⁺ influx (Verbost et al. 1987a), indicating that no significant competition between waterborne Cd²⁺ and Ca²⁺ occurs for the cell entrance step (the concentration ratio Cd/Ca being 1·4×10⁻⁴ under these conditions). The same conclusion was reached by Part et al. (1985), who showed that changes in water Ca²⁺ concentration (0–10 mmol l⁻¹) had no effect on ¹⁰⁹Cd accumulation in the gills (the concentration ratio Cd/Ca ranged from 0·7×10⁻⁴ to 7·0×10⁻⁴). Intracellular Ca²⁺ interacts with a series of more-or-less Ca²⁺-specific ligands, such as calmodulin, Ca²⁺-binding proteins (CaBPs) and Ca²⁺-ATPase. Calmodulin is a Ca²⁺-dependent regulator protein with a high Ca²⁺ affinity and is present in all eukaryotic cells. The affinity of calmodulin for Cd²⁺ is comparable to that for Ca²⁺ (Chao et al. 1984; Flik et al. 1987). The affinities of the CaBPs in fish gills for Cd²⁺ and Ca²⁺ have not been determined so far, but the vitamin-D-dependent CaBPs from rat kidney and pig duodenal mucosa share with calmodulin the property of having the same high affinity for Cd²⁺ and Ca²⁺ (Richardt et al. 1986). To affect the Ca²⁺ buffering capacity of these proteins in the cell, the intracellular concentration of Cd²⁺ must reach that of the cytosolic Ca²⁺ concentration. In comparison with calmodulin or CaBPs, the Ca²⁺ site of the plasma membrane Ca²⁺-ATPase in fish gills has an affinity for Cd²⁺ at least 100 times higher than that for Ca²⁺ (Verbost et al. 1988). It follows that the Ca²⁺ extrusion pump is a very sensitive target of Cd²⁺ in fish gills. Although an inhibition of the basolateral Ca²⁺ pumps by Cd²⁺ could explain the diminished branchial Ca²⁺ uptake following exposure to Cd²⁺, the possibility that cytosolic Cd²⁺ also impedes the movement of Ca²⁺ through apical membrane Ca²⁺ channels cannot be excluded.

The experiments described here were designed to test the hypothesis that Cd²⁺ enters the Ca²⁺-transporting cell via Ca²⁺ channels, and that cytosolic Cd²⁺ may affect apical Ca²⁺ channels. Two types of experiments were performed, (i) Shortterm accumulation of ⁴⁵ Ca and ¹⁰⁹Cd from the water into the branchial epithelium was determined to evaluate the movements of Ca²⁺ and Cd²⁺ through the apical membranes into the epithelium, (ii) Branchial influx of Ca²⁺ and Cd²⁺ was measured to evaluate the movement of these ions through both apical and basolateral plasma membranes. In these two types of experiments, apical membrane permeability for Ca²⁺ was manipulated by adding exogenous La³⁺ and by injections of the hormone hypocalcin. Recently, it was concluded that this hormone, produced in the Stannius corpuscles of fish (Wendelaar Bonga & Pang, 1986), controls apical membrane permeability to Ca²⁺ (Lafeber, 1988).

Freshwater rainbow trout (Salmo gairdneri) ranging in mass from 20 to 40 g were kept indoors and acclimated to city of Nijmegen tapwater ([Ca] = 0·70 ± 0·02 mmol l⁻¹, N = 20) under the conditions described previously (Verbost et al. 1987a). In experiments with La³⁺, carbonate-free artificial tapwater was used to prevent precipitation of La₂(CO₃)₃. The composition of the artificial tapwater was (in mmol l⁻¹): CaCl₂, 0·7; MgCl₂, 0·2; NaCl, 3·8; and KC1, 0·06 in demineralized water (pH 7·6). In experiments with Co²⁺, CoCl₂ was added to normal tapwater. The ionic content of tapwater was (in mmol l⁻¹): Ca²⁺, 0·7; Mg²⁺, 0·38; Na⁺, 0·61; K⁺, 0·05; Cl⁻, 0-66; SO₄²⁻, 0·32; and HCO₃⁻, 315 (pH 7·6).

The total calcium content of water was determined with a calcium kit (Sigma Chemical). Water total cadmium content was determined by atomic absorption spectrophotometry. Protein was measured with a reagent kit (BioRad) using bovine serum albumin as reference. Tracer content of water samples and tissue digests was determined by liquid scintillation analysis. Aqueous samples (0·5 ml) were mixed with 4·5 ml of Aqualuma scintillation fluid.

Fish were pre-exposed to Cd²⁺ (nominal 1·0 or 0·1 μmol l⁻¹) up to 16 h before experimentation. Water Cd²⁺ [added as Cd(NO₃)₂] concentrations were carefully monitored, and a maximum 10% deviation from the calculated concentrations was accepted. Pre-exposure was followed by a 1 h flux period in which the Cd²⁺ concentration in the water was kept constant (see radio tracer techniques).

Purified hypocalcin and crude extract of Stannius corpuscles were prepared as described by Lafeber et al. (1988b). The hypocalcin content of extracts of Stannius corpuscles, determined by ELISA (Kaneko et al. 1988) was 120–150 μg mg⁻¹ protein. The dose of hypocalcin injected (intraperitoneally) was 2·2–3·0 nmol hypocalcin 100 g⁻¹ fish. Hormone or extract was injected l h before tracer exposure. Injection of saline, used as vehicle, served as control.

Fish were transferred to 3·01 of aerated recirculating water containing l·0MBql^−1·45CaCl₂, 0·9 MBql^−1·109CdCl₂ or 0·2 MBql^−1·22 NaCl in tapwater. After 30min (²²Na) or l h (⁴⁵Ca, ¹⁰⁹Cd) of tracer exposure the fish were quickly (2 min) anaesthetized in bicarbonate-buffered methane sulphonate salt (MS 222, 0·5 l⁻¹, pH 7·4) and injected intraperitoneally with 5000 i.u. of sodium heparin per 100 g of fish. The gills were rapidly cleared of blood by perfusion with saline (10 ml 0·9% NaCl) via the ventral aorta. The perfusate was collected by suction after opening the atrium to determine its tracer content. Next, the gill arches were excised, rinsed for 5 s in demineralized water and blotted on wet tissue paper. Gill epithelium was carefully scraped off onto a glass plate with a microscope slide (approx. 0·4 g wet mass, weighed to ± 0·001 g) and dissolved overnight at 60°C in tissue dissolver (NCS, Amersham). The digested tissue was neutralized with glacial acetic acid, scintillation fluid (9 vols) was added, and the radioactivity determined. The remainder of the fish was processed for the determination of total body radioactivity (see below).

Influx was determined on the basis of total body radioactivity. Fish bodies and scraped gill arches were microwave-cooked (2 min) and homogenized in a blender with distilled water (volume: 65% of body mass). Triplicate samples of the homogenate (approx. 0·4 g weighed to ± 0·001 g) were processed for determination of radioactivity, as described above for the gill scrapings. Total body tracer content included the combined activities of homogenate, gill scrapings and perfusate.

To check whether the tracer associated with the gill scrapings reflected accumulation into the epithelium, and not adsorption to the external side of the epithelium, six trout were exposed to ¹⁰⁹Cd and six to ⁴⁵ Ca for 1 h and the gills perfused and excised as described above. Then, the left gill arches were rinsed in 1 mmol l⁻¹ EDTAand those from the right side in demineralized water (control). No difference in radioactivity was found between samples from either side for either isotope, indicating that the tracers were associated with the internal face of the epithelial cells.

Influx of Cd²⁺ or Ca²⁺ was calculated from the total body radioactivity after 1 h of exposure to ¹⁰⁹Cd or ⁴⁵Ca, respectively, and the respective mean tracer specific activities of the water. Fluxes were normalized to fish mass by linear extrapolation and expressed in nanomoles (Cd²⁺) or micromoles (Ca²⁺) per hour per 100 grams of fish (Lafeber et al. 1988b).

Na⁺ influx was calculated from the plasma ²²Na activity (blood samples were taken directly after anaesthetizing the fish), the apparent radiospace at 30min for Na⁺ (approx. 125 ml kg⁻¹) and the specific activity of the water (Payan & Maetz, 1973; Bath & Eddy, 1979).

The amount of Cd²⁺ accumulated in gill soft tissue was calculated from the tissue tracer content (activity per gram wet mass, <q_g′) and the mean specific activity of ¹⁰⁹Cd in the water (SA_W) using the equation: Cd²⁺ accumulated in l h = q_g′/SA_W (in nanomoles Cd²⁺ per gram wet mass). For this calculation we have assumed that during the experiment no significant transflux or backflux from blood to cells or from cells to water occurred. The Ca²⁺ accumulation in the cellular compartment could not be calculated from the tracer accumulation because of the multicompartment behaviour of cells when exposed to the Ca²⁺ tracer (Borle, 1981). Therefore, to compare the accumulation of ¹⁰⁹Cd and ⁴⁵Ca in gill tissue (Fig. 1A), relative values are presented (q_g′/q_w′; where q_w′ is the water activity per millilitre). Values found for control fish were designated as 100%.

Results are presented as means ± S.E. For statistical evaluation the Mann-Whitney U-test was used; significance was accepted for P ⩽ 0·05.

Fig. 1A summarizes the effects of several treatments that affect apical Ca²⁺ permeability on ⁴⁵Ca and ¹⁰⁹Cd accumulation in the gills. La³⁺ (1 μmol l⁻¹) and Co²⁺ (100 μmol l⁻¹) in the water reduced the Ca²⁺ accumulation in gill soft tissue by 71% and 33%, respectively, compared with controls. Because of the limited availability of purified hypocalcin we used both Stannius corpuscle extract and purified hypocalcin. Injection of Stannius extract or hypocalcin reduced ⁴⁵Ca accumulation in gills by 63 % and 37 %, respectively.

To evaluate the effects of the aforementioned treatments on Cd²⁺ accumulation in the gills, a water Cd concentration had to be established that did not influence the apical permeability to Cd²⁺ during the 1 h ¹⁰⁹Cd accumulation measurement. Since even a 4h exposure to 0·1 μmol l⁻¹ Cd had no effect on the Cd²⁺ (or Ca²⁺) accumulation (see Figs 2, 3), this concentration was used. The results showed that treatments that decreased ⁴⁵Ca accumulation had a similar, and proportional, effect on ¹⁰⁹Cd accumulation in the gills (Fig. 1A).

Fig. IB summarizes the effects of the treatments on Ca²⁺ and Cd²⁺ influx. Ca²⁺ influx decreased by 77% when LaCl₃ was added to the water. Injections of Stannius extracts and of hypocalcin reduced Ca²⁺ influx by 55 % and 43 %, respectively. Comparable effects were also found on the Cd²⁺ influx (Fig. IB). These results show that for both Ca²⁺ and Cd²⁺ a reduction in tracer accumulation in the gills is accompanied by a reduction in whole-body influx of the ions.

Fig. 2A shows the effects of time of exposure to exogenous Cd²⁺ on the ⁴⁵Ca accumulation rate in gill epithelium. A time-related inhibition of ⁴⁵Ca accumulation was found, with a significant 25% inhibition after 6h of exposure to 0·1 μmol l⁻¹ Cd, and with 60% and 70% inhibition after 9 and 17 h, respectively, of exposure to 0·1 μmol l⁻¹ Cd (exposure time to Cd²⁺ included pre-exposure and the l h flux period). When fish were exposed to 1·0 μmol l⁻¹ Cd, however, ⁴⁵Ca accumulation was inhibited by 25 % within 1 h (without pre-exposure), whereas a maximum inhibition of 60% was observed after 3–4 h of exposure. The same maximum inhibition of ⁴⁵Ca accumulation was reached for 0·1 and 1·0 μmol l⁻¹ Cd. Exposure of trout for 16·5 h to 1·0 μmol l⁻¹ Cd did not affect ²²Na accumulation in branchial tissue.

The effects of exogenous Cd²⁺ on whole-body Ca²⁺ and Na⁺ influx are shown in Fig. 2B. Exposure to 0·l μmol l⁻¹ Cd for up to 6h had no significant effect on Ca²⁺ influx, but after 9 and 17 h Ca²⁺ influx was inhibited by 67% and 87%, respectively. With 1 μmol l⁻¹ exogenous Cd, Ca²⁺ influx was significantly (49%) inhibited within 1h of exposure, and after 3–4 h of exposure a maximum 80% inhibition was observed. Exposure of fish to 1 μmol l⁻¹ Cd for 16·5 h had no effect on Na⁺ influx.

The effects of exogenous Cd²⁺ on the ¹⁰⁹Cd accumulation in gill epithelium are shown in Fig. 3A. Exposure to 0·1 μmol l⁻¹ Cd for up to 4 h had no significant effect, but ¹⁰⁹Cd accumulation decreased by 64 % after 6 h and by 75 % after 17 h of exposure to Cd²⁺. With 1·0 μmol l⁻¹ Cd in the water, six times more Cd²⁺ accumulated in gill soft tissue than with 0·1 μmol l⁻¹ Cd in the water. If we designate Cd²⁺ accumulation after l h as 100%, an 80% inhibition occurred during the second hour of incubation, which was almost the maximum inhibition observed after longer exposure (up to 17 h).

In Fig. 3B the effects of exogenous Cd²⁺ on whole-body Cd²⁺ influx are shown. The Cd²⁺ influx at 0·1 μmol l⁻¹ external Cd was around 3·4 nmol h⁻¹100 g⁻¹ fish. Exposure to Ol^moll^-1 Cd for 4h did not significantly affect Cd²⁺ influx. However, after 6h or more Cd²⁺ influx decreased by 55%. In fish exposed to l·0 μmol l⁻¹ Cd for l h, Cd²⁺ influx was around 60 nmol h⁻¹100 g⁻¹ fish, 18 times higher than in those exposed to 0·1 μmol l⁻¹ Cd. Cd²⁺ influx during the second hour of exposure to 1 μmol l⁻¹ Cd and thereafter decreased by 80 %.

Two major conclusions are drawn from this study. Cd²⁺ in the water at a concentration that provokes a specific hypocalcaemia inhibits both Ca²⁺ and Cd²⁺ influx, but not Na⁺ influx. The inhibitory effect of Cd²⁺ is concentration-and time-dependent. Treatments that inhibit branchial Ca²⁺ influx (exposure to La³⁺ in the water or hypocalcin injections) inhibit branchial Cd²⁺ influx from the water.

The main purpose of our study was to examine how Cd²⁺ passes across the branchial epithelium. We tested the hypothesis that Cd²⁺ follows the Ca²⁺ route through the epithelium. This idea was prompted by the similarity in charge and ionic radius of Ca²⁺ and Cd²⁺.

Whole-body Cd²⁺ influx amounted to 3·4 nmol h⁻¹100 g⁻¹ fish at OTgmoll^-1 external Cd and 61·9 nmol h⁻¹100 g⁻¹ fish at 1·0 μmol l⁻¹ Cd. These values for whole-body Cd²⁺ influxes are of the same order as the values found by Part & Svanberg (1981) for Cd²⁺ influx in isolated head preparations of trout. These results extend the observations of Williams & Giesy (1978) that the gills are the predominant site of Cd²⁺ influx. Our finding that inhibition of Cd²⁺ influx is accompanied by a decrease in Cd²⁺ accumulation in gill soft tissue establishes that the Cd²⁺ influx is transcellular.

Our earlier experiments with the perfused isolated head technique indicated that Cd²⁺ may be transferred across the apical membrane via La³⁺-inhibitable Ca²⁺ channels (Verbost et al.1987a). The present results on intact trout confirm and extend these earlier findings: both La³⁺ treatment and hypocalcin treatment decrease Ca²⁺ and Cd²⁺ accumulation in the gills. These observations indicate an effect of hypocalcin on apical membrane Ca²⁺ channels, and thus also extend the data of Lafeber et al. (1988a), who showed that hypocalcin inhibits transepithelial Ca²⁺ influx. Moreover, the observation that both Ca²⁺ and Cd²⁺ accumulation are inhibited by these treatments strongly suggests that these ions enter the cells via the same pathway.

As the decrease in tissue accumulation was proportional to the decrease in transcellular Ca²⁺ and Cd²⁺ flux, both whole-body Ca²⁺ and whole-body Cd²⁺ influx depend on passage of the ions through the apical membranes of the epithelium. Ca²⁺ and Cd²⁺ transfer across the epithelium should not be expected to be the same, because of the differences in tissue status of the ions (at the beginning of an experiment the gill cells contain millimolar concentrations of Ca and no Cd). However, if epithelial cells were unable to discriminate between Ca²⁺ and Cd^{2+, 109}Cd could be used as Ca²⁺ tracer (and ⁴⁵Ca as Cd²⁺ tracer). The tracer uptake studies show, however, that the epithelial cells do discriminate between Ca²⁺ and Cd²⁺ (¹⁰⁹Cd being differentially accumulated in the gills compared with ⁴⁵Ca, using similar amounts of radioactivity), which indicates a difference in cellular buffering and/or a difference in basolateral transfer.

To understand transepithelial Cd²⁺ uptake it is important to know how Cd²⁺ passes through the basolateral membrane. Because we obtained no evidence for an ATP-driven translocation from experiments with basolateral membrane vesicles in our laboratory (results not shown), and because the transepithelial Cd²⁺ influx is too high to be explained by diffusion through a pure lipid bilayer, some mechanism of facilitated diffusion seems likely. A similar conclusion was drawn for Cd²⁺ transfer through basolateral membranes in rat duodenum (Foulkes, 1986).

Short-term regulation of transepithelial Ca²⁺ influx takes place at the level of the apical membrane, which is under the control of fast-acting hypocalcin (Lafeber et al. 1988a). Indeed, the transepithelial influx of Ca²⁺ proved to be proportionally related to Ca²⁺ accumulation in the gills. The rate of transport of Cd²⁺via the gills appears to depend on the permeability of the apical membrane to Cd²⁺, as well as on passage through the cytosol and/or the basolateral membrane of the cell.

In the calculations of Ca²⁺ accumulation and Cd²⁺ influx, we have assumed that no backflux from cytosol to water or from blood to cytosol occurred. This assumption seems justified as the influx of Cd²⁺ is very low (62 nmol h⁻¹ 100 g⁻¹ fish in water containing 1·0 μmol l⁻¹ Cd and 0·7 mmol l⁻¹ Ca) and the Cd²⁺ space is large. Also, backflux of Cd²⁺ through the apical membrane will be negligible because of the electrical potential difference between cytosol and water (the cytosol being negative; Perry & Flik, 1988).

Chloride cells are generally considered as the sites of Na⁺ uptake in the gills (Avella et al. 1987). The Na⁺/K⁺-ATPase activity, located in the basolaferal membrane of these cells, is regarded as the most important driving force for transepithelial Na⁺ uptake in fish gills (deRenzis & Bornacin, 1984). Branchial Na⁺ influx is not affected by Cd²⁺ exposure, in contrast to Ca²⁺ influx, which also depends on ATPase activity (Ca²⁺-ATPase; Flik et al. 1985) colocalized with Na⁺/K⁺-ATPase. This suggests that Na⁺/K⁺-ATPase activity is not inhibited by Cd²⁺ concentrations that inhibit Ca²⁺-ATPase. This is in agreement with in vitro studies, which indicate a much lower sensitivity of the Na⁺/K⁺-ATPase than the Ca²⁺-ATPase for Cd²⁺. The concentration of Cd²⁺ causing 50% inhibition in vitro of Na⁺/K⁺-ATPase from various origins, for instance rabbit proximal tubule (Diezi et al. 1988), cultured vascular smooth muscle cells (Tokushige et al. 1984) or rat brain synaptosomes (Lai et al. 1980) is in the micromolar range. We found for trout gills, that the branchial para-nitrophenyl phosphatase (pNPPase) activity, which reflects the K⁺-dependent dephosphorylation step in the Na⁺/K⁺-ATPase reaction cycle, was inhibited by 50 % by 0-25 /zmol I^-1 Cd²⁺ (G. Flik, unpublished results). These data indicate that the concentration of free cytosolic Cd²⁺ that causes inhibition of Ca²⁺ influx must be below the micromolar range, as Na⁺ influx was not affected. Such a low cytosolic Cd²⁺ concentration also excludes significant binding of Cd²⁺ to calmodulin, as the K_m of calmodulin for Cd²⁺ is in the micromolar range (Flik et al. 1987). The formation of Cd²⁺-calmodulin complexes has been proposed as the primary cause of cellular Cd²⁺ toxicity (e.g. Suzuki et al. 1985). The present results do not support such a calmodulin-mediated toxicity mechanism for Cd²⁺.

Our data show that Ca²⁺ and Cd²⁺ enter the gills via the same route. This implies that Cd²⁺ is concentrated in the ion-transporting cells of the gills, the chloride cells, since these cells account for the branchial Ca²⁺ transport (Fenwick, 1989). As a consequence, inhibition of Ca²⁺ transport will occur as soon as Cd²⁺ has accumulated to a level sufficient to inhibit the Ca²⁺-ATPase transport system (Verbost et al. 1988). However, the present results also show a reduction in the rate of accumulation of Ca²⁺ in gills after prolonged Cd²⁺ exposure. We suggest that this effect is caused by a decrease in the permeability of the apical membrane to Ca²⁺, possibly by an indirect blockage of Ca²⁺ channels by cytosolic Cd²⁺. This hypothesis is supported by several observations. First, the inhibition by Cd²⁺ is not acute, in contrast to the inhibition by La³⁺ (Verbost et al. 1987a). Moreover, short-term exposure to Cd²⁺ (0 · 1 μ mol l^{− 1} Cd) has no effect on the rate of accumulation of Ca²⁺ in the tissue, whereas long-term exposure (17 h) decreases the rate of accumulation of Ca²⁺. Second, long-term exposure to Cd²⁺ decreases the rate of accumulation of Cd²⁺ in the gills, whereas short-term exposure to Cd²⁺ has no effect. These observations indicate that Ca²⁺ channels become blocked when Cd²⁺ accumulates in the epithelial cells, as happens after a prolonged exposure to waterborne Cd²⁺. These findings also support our conclusion that external Cd²⁺ cannot cause the inhibition of Ca²⁺ influx by competition with Ca²⁺ at mucosal sites, because the inhibition of influx is not instantaneous.

Which mechanism underlies the blockage by Cd²⁺ of the permeability of the apical membrane to Ca²⁺? Once Cd²⁺ has entered the Ca²⁺-transporting cell it could affect the Ca²⁺ channels in several ways. The basolateral Ca²⁺ pump has an extremely high affinity for Cd²⁺ (I₅₀ = 3 nmol l⁻¹; Verbost et al. 1988) and, therefore, nanomolar intracellular Cd²⁺ concentrations inhibit Ca²⁺ extrusion. We suggest that inhibition of the Ca²⁺ pump leads to an increased intracellular Ca²⁺ concentration, [Ca²⁺]_i. Possibly, [Ca²⁺]_i rises to levels that close the apical membrane Ca²⁺ channels. Thus, [Ca²⁺]_i serves as a feedback signal to control Ca²⁺ entry at the apical membrane. An analogous model was proposed for rat small intestine enterocytes (Van Os, 1987). Both an overcapacity of Ca²⁺ pumps and buffering of Cd²⁺ by cytosolic binding proteins could explain the delay in the inhibition of Ca²⁺ uptake by Cd²⁺.

To test the hypothesis that Cd²⁺ closes Ca²⁺ channels by increasing [Ca²⁺]_i would require examination with Ca²⁺ fluorochromes to show the predicted rise in [Ca²⁺]_i. However, this was not feasible because Cd²⁺ quenches Quin2 fluorescence and interferes with Ca²⁺-dependent fluorescence of Fura-2 and Indo-1 (P. M. Verbost & G. Visser, personal observations). The Ca²⁺ binding sites of Fura-2 and Indo-1 are EGTA-type sites (Tsien, 1980) with a much higher affinity for Cd²⁺ than for Ca²⁺ (Sillen & Martell, 1964). There is, however, indirect evidence for a rise in [Ca²⁺]_i upon exposure to Cd²⁺ of various other cell types: erythrocytes show an acceleration in age-related changes (Kunimoto et al. 1985), protein phosphorylation in human platelets is increased (Pezzi & Cheung, 1987), water absorption in rat duodenum is reduced (Toraason & Foulkes, 1984) and in rat skeletal muscle cells the degradative enzymes phosphorylase-b kinase, phospholipases and proteases are activated (Toury et al. 1985). All these phenomena are known to be mediated by a rise in [Ca²⁺]_i. In human lymphocytes Cd²⁺ produces an increase in the accumulation of ⁴⁵ Ca and in the rate of mitogenesis (Parker, 1974). The latter effect is also dependent on a rise in [Ca²⁺]_i (Parker, 1974).

Although apical membrane Ca²⁺ channels may also be affected in fish exposed to Cd²⁺, we conclude that Ca²⁺ transport in the gills becomes inhibited primarily because the basolateral membrane Ca²⁺ pump has such an extremely high sensitivity to Cd²⁺. This conclusion for fish gills may be extended to explain intoxication of mammalian Ca²⁺-transporting epithelia, such as the kidneys and the intestine, by Cd²⁺. Recent data on Cd²⁺ inhibition of Ca²⁺ pumps from rat kidney and rat intestinal tissues (Verbost et al. 1987b) confirm the applicability of the gill model.

This study was supported by a grant from the Foundation for Fundamental Biological Research (BION), which is subsidized by the Dutch Organisation for Scientific Research (NWO). Mr F. A. T. Spanings is thanked for his excellent fish husbandry.