INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIAL AND METHODS RESULTS DISCUSSION REFERENCES

THE FISH GILL CONTAINS EPITHELIAL cells that are in direct contact with the external environment. These cells are the primary site for osmoregulation. Moreover, the role of the gill epithelium in maintaining ionic homeostasis is evidently enhanced in euryhaline fish, as they have the capacity to support drastic changes in water salinity. As is the case for many euryhaline species, the sea bass Dicentrarchus labrax can survive in an estuarine environment where it can be subjected to physiologically severe alterations in osmotic pressure. After these osmotic changes, the different types of gill cells must react rapidly to compensate for ionic imbalances. However, the cell mechanisms underlying such compensations remain largely unknown.

Of the various effectors involved in the osmotic regulation of epithelia, ion channels have undergone extensive investigation. Nevertheless, the situation in relation to osmoregulation in gill tissue has not been elucidated. This is probably due to the anatomic complexity of the tissue, which limits access to the various cell types present in vivo. To overcome this problem, cultured branchial cells represent an interesting alternative. Using primary cultures of respiratory cells from the gill of the sea bass, we previously described the existence of Cl and K⁺ channels on the apical membrane (9, 10). The characteristics of these channels indicate that they could participate in cell volume regulation. Notably, we found a large conductance stretch-activated K⁺ channel regulated by both mechanical pressure and changes in osmotic gradients (9). In many animal cells, cell volume regulation after exposure to a hypotonic shock has been achieved by the activation of conductive pathways for K⁺ and Cl (16, 19, 26, 31), and stretch-activated channels have been implicated in such regulation (32).

The aim of the present study, therefore, was to evaluate the role of potassium conductances on exposure of cultured gill cells to a hypotonic shock. For this purpose we undertook whole cell patch-clamp experiments together with ⁸⁶Rb⁺ efflux measurement on gill cells in primary culture. The simple geometry and the free access to the apical and basolateral membranes provide experimental conditions suited for efflux measurements. We describe two different pathways for ⁸⁶Rb⁺ effluxes activated during cell swelling. One of these pathways is located on the apical membrane and could correspond to the presence of a stretch-activated K⁺ channel. The other pathway is restricted to the basolateral membrane and could be a Ca²⁺-sensitive maxi K⁺ channel. With ¹²⁵I efflux measurements, a basolateral Cl conductance was also found to be activated during the exposure of cells to a hypotonic solution. It is therefore possible that the observed K⁺ and Cl currents could be implicated in the rapid osmoregulation mechanisms that have been identified in fish after abrupt changes in seawater salinity.

	MATERIAL AND METHODS

TOP ABSTRACT INTRODUCTION MATERIAL AND METHODS RESULTS DISCUSSION REFERENCES

Animals

Primary Cell Culture

The explant technique for whole cell patch-clamp experiments has been described in the companion manuscript (9).

The dissociation technique for ⁸⁶Rb⁺ and ¹²⁵I efflux measurements was performed as follows: the gill arches were trypsinized twice at room temperature [2 × 15 min in 0.05% trypsin and 0.02% EDTA solution (Sigma)]. Cell suspensions were filtered, rinsed, and finally seeded in Costar-Transwell 0.45-µm pore inserts (25 mm diameter, Costar) at a high density as previously described (3).

Explants and inserts were maintained in a precision-controlled low-temperature incubator (Jouan) at 17°C in a humidified air atmosphere (i.e., atmospheric PCO₂). The culture medium was changed every second day.

Whole Cell Patch-Clamp Experiments

Voltage-clamp commands, data acquisition, and data analysis were controlled by a computer equipped with a Digi Data 1200 interface (Axon instruments). Commercially available PCLAMP 6.0 software (Axon instruments) was used to generate whole cell current-voltage (I-V) relationships. For the measurements of potassium currents, cells were held at a holding potential (V_h) of 40 mV and 400-ms pulses from 80 to +80 mV were applied in increments of 20 mV every 2 s.

⁸⁶Rb+ Efflux From Confluent Monolayers

Calculation. From back addition of the radioactivity in the efflux samples to the radioactivity remaining in the cells, the apical and basolateral efflux rate constants were calculated as the fraction of the total radioactivity lost per unit of time, according to Le Maout et al. (24) after the equations

where (K_a)_t and (K_b)_t are the apical and basolateral efflux rate constants at time t and (C_a)_t and (C_b)_t are the radioactivity lost, respectively, by the apical and basolateral sides of the epithelium at time t and during the period of experimentation T. C_ep is the radioactivity remaining in the epithelium at the conclusion of measurements, and tf corresponds to the final time of the experiment.

¹²⁵I Effluxes From Confluent Monolayers

Statistical Comparisons

Solutions and Chemicals

Primary cultures. All solutions were prepared using tissue-culture quality chemicals.

Electrophysiological studies. For whole cell patch-clamp experiments, the pipette solution contained in (mM) 140 KCl, 5 HEPES, 5 ATP, and 5 EGTA (pH 7.4 adjusted with KOH 1M). The standard bath solution contained (in mM) 140 Na glutamate, 5 HEPES, and 0.1 CaCl₂. pH was adjusted to 7.7 with NaOH, and osmolarity to 350 mosmol/kgH₂O by the addition of 60 mM mannitol. The hypotonic solution was obtained by removing the mannitol from the preceding solution (270-280 mosmol/kgH₂O). Lyophilized charybdotoxin (Latoxan) was used as an inhibitor of Ca²⁺ dependent K⁺ channels at a final concentration of 10⁶ M. Quinidine (Sigma) was obtained in its sulfate salt form and was added at a final concentration of 500 µM to the solution. Potassium selectivity was investigated using a solution consisting of (in mM) 140 K-glutamate, 5 HEPES, and 0.1 CaCl₂, adjusted to pH 7.7 with NaOH, and osmotic pressure to 350 mosmol/kgH₂O by addition of 60 mM mannitol. Ionomycin (Sigma) was prepared at 1 mM in DMSO and was added at a final concentration of 1 µM to the perfusion solution.

⁸⁶Rb+ and ¹²⁵I efflux studies. ⁸⁶Rb⁺ in its Cl salt form and ¹²⁵I as an Na⁺ salt were purchased from Amersham. Hypotonic solutions were obtained by 30% dilution of L15 (supplemented with 20 mM NaCl). The quinidine solution was prepared as described above for patch-clamp experiments and added to the final solution. Leirus quinquestriatus hebraeus (LQH) venom (Latoxan, France) was used at 20 µg/ml. The gadolinium solution was prepared from a 1 M GdCl₃ stock solution (Sigma) and directly added to the bath solution. A stock solution of 5-nitro-2-(3-phenylpropylamino) benzoic acid (NPPB) from Calbiochem was prepared at 0.1 M in DMSO and used at 0.1 mM in final solutions. Bumetanide and thapsigargin (Sigma) were used at 0.1 µM.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIAL AND METHODS RESULTS DISCUSSION REFERENCES

Whole Cell Patch-Clamp Experiments

K+ currents induced by a hypotonic shock. To study the effects of changes in osmotic pressure on the development of the K⁺ conductance in primary cultures of seawater fish gill cells, currents were induced by osmotic shock. In these experiments, the pipette solution contained the KCl solution and osmotic pressure was maintained at 290 mosmol/kgH₂O. To eliminate the participation of anions in the outward current, experiments were carried out after replacing Cl in the bath solution by glutamate. After successful gigaseal formation, the whole cell configuration of the patch-clamp technique was obtained in 30% of cases. Figure 1A illustrates a set of current recordings made in a hypertonic bath medium (Na glutamate solution, 350 mosmol/kgH₂0) with test potentials that ranged from 80 mV to +80 mV in increments of 20 mV. The voltage step protocol elicited small time-independent currents that changed linearly with the membrane voltage and had a reversal potential (E_rev) of +13 ± 3 mV (n = 9). Because of their small amplitude, the nature of these currents was not analyzed further.

View larger version (30K): [in this window] [in a new window]   Fig. 1.   K+ currents induced by a hypotonic shock in the presence of external Na glutamate solution in cultured gill cells in whole cell patch-clamp configuration. Membrane voltage was held at 20 mM and stepped to test potential values between 80 to +80 mV in +20-mM increments. Pipette was filled with 300 mosmol/kgH2O isosmotic 140 mM KCl solution. Whole cell currents were successively recorded in 350 mosmol/kgH2O hyperosmotic Na glutamate (control conditions) (A), 1 min in 280 mosmol/kgH2O hypotonic Na glutamate solution (B), 2 min in 280 mosmol/kgH2O hypotonic Na glutamate solution (C), 3 min 280 mosmol/kgH2O hypotonic Na glutamate solution (D), in 350 mosmol/kgH2O hypertonic Na glutamate solution (E), and in 280 mosmol/kgH2O hypotonic K glutamate solution (F). G: average current-voltage relationships measured 380 ms after the onset of pulse. Values are means ± SE: , control in external hypertonic Na glutamate solution (n = 46 cells from 17 monolayers); , 1 min in hypotonic Na glutamate solution (n = 22 cells from 16 monolayers); , 2 min in hypotonic Na glutamate solution (n = 19 cells from 14 monolayers); , 3 min in hypotonic Na glutamate solution (n = 28 cells from 16 monolayers); , in hyperosmotic Na glutamate solution (n = 10 cells from 5 monolayers); , in hypotonic K glutamate solution (n = 3 cells from 3 monolayers).

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Effects of K+ channel inhibitors on K+ currents induced by a hypotonic shock in the presence of external Na glutamate solution. Pipette was filled with 300 mosmol/kgH2O isosmotic 140 mM KCl solution. Whole cell currents were successively recorded in 350 mosmol/kgH2O hyperosmotic Na glutamate solution (control condition) (A), 3 min in 280 mosmol/kgH2O hypotonic Na glutamate solution (B), with extracellular perfusion of 500 µM quinidine (C), and with extracellular perfusion of 1 µM charybdotoxin (ChTx; D). E: average current-voltage relationships measured 380 ms after the onset of pulse. Values are means ± SE: , control (n = 46 cells from 17 monolayers); , 3 min in hypotonic 140 Na glutamate solution (n = 28 cells from 16 monolayers); , in the presence of 500 µM quinidine (n = 12 from 12 monolayers); or  in the presence of 1 µM charybdotoxin (n = 6 from 5 monolayers).

K+ currents induced by ionomycin. The conductance blockade observed in the presence of charybdotoxin strongly suggested that a part of the conductance induced by hypotonic shock could be due to K⁺ ions being conducted through maxi Ca²⁺-activated K⁺ channels ("BK channels"). Therefore, in a second series of experiments, the effects of ionomycin were tested on whole cell currents developed with KCl in the pipette and Na glutamate in the bath solutions. The osmotic pressure of the extracellular solution was adjusted to 350 mosmol/kgH₂O to avoid inducing volume-activated currents. Figure 3A shows control currents, and Fig. 3B shows currents recorded 1 min after the addition of 1 µM ionomycin to the bathing medium. In the presence of the ionophore, the currents increased during depolarizing voltage pulses, with a maximum slope conductance of 5.7 ± 0.9 nS and an E_rev of 65 ± 6 mV (n = 11; Fig. 3E). The addition of 500 µM quinidine or 1 µM charybdotoxin to the bath ionomycin solution inhibited the whole cell currents within 2 min (Fig. 3, C-E). Overall, the ionomycin-induced K⁺ currents were quite similar to those induced by hypotonic shock.

View larger version (26K): [in this window] [in a new window]   Fig. 3.   K+ currents induced by extracellular perfusion of ionomycin (1 µM) in whole cell configuration. Pipette was filled with 300 mosmol/kgH2O isosmotic 140 mM KCl solution and the bath contained 350 mosmol/kgH2O hyperosmotic Na glutamate solution. Whole cell currents were successively recorded without ionomycin (control condition) (A), 3 min after the extracellular perfusion of 1 µM ionomycin (B), with extracellular perfusion of 1 µM ionomycin + 500 µM quinidine (C), and with extracellular perfusion of 1 µM ionomycin + 1 µM charybdotoxin (D). E: average current-voltage relationships measured 380 ms after the onset of pulse. Values are means ± SE: , control (n = 12 cells from 10 monolayers); , 3 min with extracellular perfusion of ionomycin (n = 10 from 8 monolayers); , with extracellular perfusion of ionomycin + quinidine (n = 4 from 4 monolayers); , with extracellular perfusion of ionomycin + charybdotoxin (n = 3 from 3 monolayers).

⁸⁶Rb Efflux Experiments

Effect of hypotonic media. The Rb efflux technique was used to further study the movements of potassium across the apical and the basolateral membranes of gill cells in primary culture. Notably, the existence of a K⁺ conductance activated by exposure to a hypotonic solution led us to study the influence of cell swelling on the polarization (i.e., confined to a specific membrane) of potassium effluxes. Gill cells were grown on permeable supports, and apical and basolateral membrane effluxes were measured after the cells had been loaded with ⁸⁶Rb⁺. Cell swelling was induced by exposure of the monolayer to a hypotonic medium.

2

1

2

1

View larger version (18K): [in this window] [in a new window]   Fig. 4.   A: time course of 86Rb effluxes stimulated by hypotonic solution in cultured gill cells on permeable filters. Epithelia were preloaded with 86Rb (1-2 µCi/ml, 4-5 h), washed, and placed first in L15 solution (350 mosmol/kgH2O) with a change of apical and basolateral medium every 2 min. Cells were exposed to L15 solution (control condition): , % 86Rb efflux across the apical membrane (n = 4); , % 86Rb efflux across basolateral membrane (n = 4). Cells were exposed to hypotonic solution (hypotonic shock condition, 245 mosmol/kgH2O): , % 86Rb efflux across apical membrane (n = 15); , % 86Rb efflux across basolateral membrane (n = 15). B: effect of quinidine on 86Rb effluxes stimulated by hypotonic shock. Cells were exposed to hypotonic solution (hypotonic shock condition, 245 mosmol/kgH2O): , % 86Rb efflux across apical membrane (n = 4); , % 86 Rb efflux across basolateral membrane (n = 4). Cells were exposed to hypotonic solution plus 500 µM quinidine: ,% 86Rb efflux across apical membrane (n = 4); , % 86Rb efflux across basolateral membrane (n = 4).

Pharmacological properties of ⁸⁶Rb+ efflux stimulated by exposure to a hypotonic solution. To characterize the stimulated ⁸⁶Rb⁺ efflux under hypotonic conditions, the effects of K⁺ channel inhibitors were studied. The action of 0.5 mM quinidine is shown in Fig. 4B, where the channel blocker inhibited significantly the action of the hypotonic medium on ⁸⁶Rb⁺ effluxes across the basolateral (at time = 16 min: 258 ± 33 vs. 149 ± 10%, n = 4, P < 0.05) and apical (at time = 16 min: 139 ± 8 vs. 81 ± 6%, n = 4, P < 0.05) membranes.

View larger version (18K): [in this window] [in a new window]   Fig. 5.   A: effect of Leirus quinquestriatus hebraeus (LQH) venom on 86Rb effluxes stimulated by exposure to hypotonic solution. Cells were exposed to hypotonic solution (hypotonic shock condition, 245 mosmol/kgH2O): , % 86Rb efflux across apical membrane (n = 4); , % 86Rb efflux across basolateral membrane (n = 4). Cells were exposed to hypotonic solution plus LQH venom (20 µg/ml): , % 86Rb efflux across apical membrane (n = 4); , % 86Rb efflux across basolateral membrane (n = 4). B: effect of gadolinium on 86Rb effluxes stimulated by exposure to hypotonic solutions. Cells were exposed to hypotonic solution (hypotonic shock condition, 245 mosmol/kgH2O): , % 86Rb efflux across apical membrane (n = 4); , % 86Rb efflux across basolateral membrane (n = 4). Cells were exposed to hypotonic solution plus 0.1 mM GdCl3: , % 86Rb efflux across apical membrane (n = 4); , % 86Rb efflux across basolateral membrane (n = 4).

View larger version (18K): [in this window] [in a new window]   Fig. 6.   A: effect of bumetanide on 86Rb effluxes stimulated by exposure to hypotonic solutions. Cells were exposed to hypotonic solution (hypotonic shock condition, 245 mosmol/kgH2O): , % 86Rb efflux across apical membrane (n = 4); , % 86Rb efflux across basolateral membrane (n = 4). Cells were exposed to hypotonic solution plus 0.1 mM bumetanide: , % 86Rb efflux across apical membrane (n = 4); , 86Rb efflux across basolateral membrane (n = 4). B: effect of 5-nitro-2-(3-phenylpropylamino)benzoic acid (NPPB) on 86Rb effluxes stimulated by exposure to hypotonic solutions. Cells were exposed to hypotonic solution (hypotonic shock condition, 245 mosmol/kgH2O): , % 86Rb efflux across apical membrane (n = 4); , % 86Rb efflux across basolateral membrane (n = 4). Cells were exposed to hypotonic solution plus 0.1 mM of NPPB: , % 86Rb efflux across apical membrane (n = 4); , % 86Rb efflux across basolateral membrane (n = 4).

Effect of calcium. The results obtained with whole cell patch-clamp technique clearly showed that hypotonicity was able to activate Ca²⁺-dependent K⁺ currents. To test whether this activation was due to a release of Ca²⁺ from internal stores, monolayers were preincubated with thapsigargin for 1 h before their exposure to a hypotonic solution. As illustrated in Fig. 7A, this treatment did not modify the classical stimulation response of ⁸⁶Rb efflux induced by hypotonic shock.

View larger version (18K): [in this window] [in a new window]   Fig. 7.   A: effect of thapsigargin on 86Rb effluxes stimulated by exposure to hypotonic solutions. Cells were exposed to L15 solution (control condition without hypotonic shock): , % 86Rb efflux across the apical membrane (n = 4); , % 86Rb efflux across basolateral membrane (n = 4). Cells were exposed to hypotonic solution plus 0.1 µM thapsigargin added 1 h before the experiment: , % 86Rb efflux across apical membrane (n = 4); , % 86Rb efflux across basolateral membrane (n = 4). B: effect of ionomycin on 86Rb effluxes. Cells were exposed to L15 solution (control condition without hypotonic shock): , % 86Rb efflux across the apical membrane (n = 4); , % 86Rb efflux across basolateral membrane (n = 4). Cells were exposed to L15 solution plus 1 µM of ionomycin: , % 86Rb efflux across apical membrane (n = 6); , % 86Rb efflux across basolateral membrane (n = 6).

Effect of an asymmetric hypotonic shock. In its natural environment, the fish gill epithelium forms a barrier between two compartments containing radically different concentrations of ionic species. It was therefore of interest to perform experiments in which just the apical or the basolateral membrane was exposed to the hypotonic shock. The results of these experiments are summarized in Fig. 8. Control experiments (Fig. 8A) were performed in which both sides of the monolayer were exposed to hypotonic medium. As already described, this exposure led to an increase in ⁸⁶Rb⁺ efflux across both basolateral and apical membranes. Surprisingly, applying the hypotonic shock to the apical compartment only did not elicit apical or basolateral ⁸⁶Rb⁺ effluxes (Fig. 8B). In contrast, application of the hypotonic shock to the basolateral side induced the classical increase in both apical and basolateral ⁸⁶Rb⁺ effluxes (Fig. 8C).

View larger version (19K): [in this window] [in a new window]   Fig. 8.   Time course of 86Rb effluxes stimulated by symmetrical or asymmetrical hypotonic shocks. A: cells exposed to symmetrical hypotonic solution (hypotonic shock condition, 245 mosmol/kgH2O): , % 86Rb efflux across apical membrane (n = 4); , % 86Rb efflux across basolateral membrane (n = 4). B: cells exposed to apical hypotonic solution (hypotonic shock condition, 245 mosmol/kgH2O): , % 86Rb efflux across apical membrane (n = 4); , % 86Rb efflux across basolateral membrane (n = 4). C: cells exposed to basolateral hypotonic solution (hypotonic shock condition, 245 mosmol/kgH2O): , 86Rb efflux across basolateral membrane (n = 4); , % 86Rb efflux across basolateral membrane (n = 4).

¹²⁵I Efflux Experiments

View larger version (21K): [in this window] [in a new window]   Fig. 9.   A: time course of 125I effluxes stimulated by hypotonic solution in gill cells cultured on permeable filters. Epithelia were preloaded with 125I (10 µCi/ml, 4-5 h), washed, and placed first in L15 solution (350 mosmol/kgH2O) with a change of apical and basolateral medium every 2 min. Cells were exposed to L15 solution (control condition): , % 125I efflux across the apical membrane (n = 4); , % 125I efflux across basolateral membrane (n = 4). Cells were exposed to hypotonic solution (hypotonic shock condition, 245 mosmol/kgH2O): , % 125I efflux across apical membrane (n = 4) ; , % 125I efflux across basolateral membrane (n = 4). B: effect of NPPB on 125I effluxes stimulated by exposure to hypotonic solutions. Cells were exposed to hypotonic solution (hypotonic shock condition, 245 mosmol/kgH2O): , % 125I efflux across apical membrane (n = 4); , % 125I efflux across basolateral membrane (n = 4). Cells were exposed to hypotonic solution plus 0.1 mM of NPPB: , % 125I efflux across apical membrane (n = 4); , % 125I efflux across basolateral membrane (n = 4). C: effect of quinidine on the 125I effluxes stimulated by exposure to hypotonic solutions. Cells were exposed to hypotonic solution (hypotonic shock condition, 245 mosmol/kgH2O): , % 125I efflux across apical membrane (n = 4); , % 125I efflux across basolateral membrane (n = 4). Cells were exposed to hypotonic solution plus 500 µM of quinidine: , % 125I efflux across apical membrane (n = 4); , % 125I efflux across basolateral membrane (n = 4).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIAL AND METHODS RESULTS DISCUSSION REFERENCES

In the companion paper (9) we demonstrated the existence of stretch-activated K⁺ channels in the apical membrane of gill cells in primary culture. This channel was also activated by hypotonic shock, suggesting that it could be implicated in cell volume regulation. In this study, we further investigated the issue of K⁺ ion movements in gill cells in primary culture during exposure to hypotonic media.

The results obtained in whole cell configuration demonstrate that cell swelling was responsible for the activation of a K⁺ conductance in cultured gill cells. To characterize more precisely this K+ conductance activated by hypotonic shock, specific K⁺ channel inhibitors were used. This conductance was completely blocked by quinidine and strongly reduced by charybdotoxin. This toxin, contained in LQH scorpion venom, is known to inhibit the activation of high-conductance Ca²⁺-activated K⁺ channels (34). This observation led to discrimination between two K⁺ components activated by hypotonic shock: one major, charybdotoxin sensitive and the other minor, charybdotoxin insensitive.

To identify the nature of the charybdotoxin-sensitive component, whole cell experiments were performed without any hypotonic shock and in the presence of a calcium ionophore. The results show that ionomycin mimicked the action of hypotonicity by eliciting the development of K⁺ currents whose characteristics were very similar to swelling-activated K⁺ conductances. Moreover, the total ionomycin-stimulated K⁺ current was fully blocked by charybdotoxin. Taken together, these results suggest that hypotonicity was able to activate Ca²⁺-sensitive K⁺ currents. A swelling-induced increase in K⁺ conductance has already been reported for a number of cell types (5, 15, 19, 24, 31). Nevertheless, the nature of K⁺ channels implicated in this phenomenon is still under discussion and could depend on the type of tissue involved. The K⁺ channels implicated in the regulatory volume decrease (RVD) after a hypotonic shock in a variety of cells including epithelial cells, are Ca²⁺-sensitive K⁺ channels, or BK channels to be more specific. These channels are activated by a rise in cytosolic Ca²⁺ elicited by cell swelling (12, 24, 27, 28, 33). In contrast, the swelling-activated K⁺ conductance in other cell types such as Ehrlich cells (18), hepatocytes (29), lymphocytes (16, 17), or cerebellar granule neurons (30), has been shown to be Ca²⁺ independent. Conflicting data have also been reported for fish epithelia, with studies on the proximal renal tubule resulting in opposite conclusions concerning the role of Ca²⁺ in controlling K⁺ movements induced by hypotonicity (21, 35). Different types of K⁺ channels have, however, been described in other fish epithelia (25). Notably, in teleost intestinal epithelium, a voltage-dependent Ca²⁺-activated K⁺ channel has been described (25), while basolateral K⁺ channels were detected in shark rectal gland tubules (13, 14). To the best of our knowledge, however, the role of these channels during hypotonic shock has not yet been addressed.

Concerning the origin of the increase of Ca²⁺ leading to BK channel stimulation, two main mechanisms are classically reported in the literature: an influx of external Ca²⁺ through stretch-activated cation channels (8) and/or Ca²⁺ release from internal Ca²⁺ stores (see Ref. 18 for review). In the present study, we found that the activation of swelling-induced or ionomycin-sensitive K⁺ conductances took place in the presence of high concentrations of EGTA in the pipette solution. Moreover, removal of extracellular Ca²⁺ strongly impaired the efficacy of the hypotonic shock (data not shown). However, as this action was irreversible and the gill cell function declined rapidly, we were unable to study further the role of external calcium in these experiments. In any case, these results demonstrate that gill cell K⁺ conductance during swelling is probably not due to the release of Ca²⁺ from internal stores. We could therefore reasonably postulate that a Ca²⁺ influx could participate in activation of K⁺ channels. It remains to be shown how this Ca²⁺ influx can increase the cytosolic concentration of Ca²⁺ in the presence of a high concentration of EGTA. The observations of Evans and Marty (11) do, however, shed light on this problem by indicating that with EGTA as a buffer, a whole region of the cell could escape control by this Ca²⁺ buffer. Because this region could extend to a large part of the plasma membrane (11), a local transient increase in Ca²⁺ could stimulate BK channel activity.

In whole cell configuration, the second minor component was charybdotoxin insensitive. These observations indicate that BK or other Ca²⁺-sensitive channels are probably not the only K⁺ channels implicated in the volume-activated K⁺ conductance observed here. This minor charybdotoxin and Ca²⁺-insensitive component should correspond to the stretch-activated K⁺ channel characterized in the companion paper (9). This channel found on the apical membrane of gill cells in primary culture was not permeable to Ca²⁺ and could not be responsible for the Ca²⁺ influx required during cell swelling.

To obtain further information on K⁺ conductances in both the apical and basolateral membranes, ⁸⁶Rb⁺ efflux experiments were performed on gill-cell monolayers grown on permeable filters. In various epithelial cells (1, 20), ⁸⁶Rb⁺ has been shown to be a good indicator for K⁺ movement. Moreover, the substitution of KCl by RbCl in gill cells did not change the whole cell currents induced by the hypotonic shock (data not given). Hence, ⁸⁶Rb⁺ would seem to be an appropriate substitute for K⁺. The present work shows that stimulated ⁸⁶Rb⁺ effluxes were inhibited by known K⁺ channel blockers, indicating that effluxes measured under these conditions were therefore due to the transport of Rb⁺ through K⁺ channels.

The present study shows that exposure of gill cells in primary culture to a hypotonic shock increased ⁸⁶Rb⁺ efflux through both the apical and basolateral membranes. The pharmacology of these effluxes was investigated by adding to the bathing media known blockers of potassium conductances in epithelial cells. Of these, quinidine was an effective blocker of both apical and basolateral membrane ⁸⁶Rb⁺ fluxes. Its action, reported for a wide variety of tissues, is not specific to a given type of K⁺ channel. However, crude scorpion venom (LQH), strongly inhibited the hypotonicity-stimulated efflux across the basolateral membrane but did not affect the apical efflux. These results clearly demonstrated that the K⁺ efflux takes place through different types of channels in both membranes. This is confirmed by the observation that gadolinium only blocked the apical ⁸⁶Rb⁺ efflux without modifying the basolateral efflux. Because the action of crude venom on K⁺ channels is probably due to charybdotoxin (34), the basolateral K⁺ conductance activated by cell swelling could be mainly due to an increased BK channel activity.

To further emphasize the results of these experiments, efflux measurements made in the presence of ionomycin also demonstrated that ⁸⁶Rb⁺ efflux increased through the basolateral membrane only, confirming the presence of Ca²⁺-sensitive K⁺ channels in this membrane. These observations strengthen the data obtained with the whole cell patch-clamp technique. Although it is reasonable to conclude that the increase in basolateral conductance during osmotic shock was mediated by BK channels, it is necessary to appreciate that this type of channel is not the only Ca²⁺-sensitive K⁺ channel blocked by charybdotoxin. In fact, McCann et al. (27, 28) demonstrated the existence of a low-conductance inwardly rectifying basolateral K⁺ channel in airway epithelium. This channel was inhibited by charybdotoxin and was activated by the cAMP-induced release of intracellular Ca²⁺. However, this type of channel is probably not implicated in the whole cell conductances or ⁸⁶Rb⁺ effluxes examined in the present study because the K⁺ currents were outwardly rectifying and insensitive to cAMP (data not given).

At the apical level, the action of gadolinium suggests that the stimulated K⁺ efflux occurred through stretch-activated channels. In the companion paper (9), we demonstrated the presence of stretch-activated K⁺ channels in the apical membrane of cultured gill cells, which were blocked by gadolinium and sensitive to changes in osmolarity. Such channels could therefore be involved in the increased apical K⁺ efflux that takes place as a result of cell swelling.

The experiments carried out in the presence of thapsigargin demonstrated that ⁸⁶Rb⁺ effluxes through both apical and basolateral membranes were insensitive to Ca²⁺ released from internal stores. Thus, as discussed for whole cell experiments, it now seems clear that the response to hypotonic shock was mediated moreover by an influx of external Ca²⁺.

In a previous study (3), the presence of an Na⁺-K⁺-2Cl cotransporter was found in the basolateral membrane of gill cells in primary culture. To investigate the putative role of this transporter in the K⁺ flux measured here, we tested the effect of bumetanide on ⁸⁶Rb⁺ efflux rates elicited by exposure of cells to hypotonic solutions. The absence of effect observed on the basolateral side shows that it is inconsistent that the Na⁺-K⁺-2Cl cotransporter provides the primary basolateral K⁺ efflux pathway. The slight stimulation effect observed on the apical side is paradoxical and still unclear.

In many cells, swelling induced by exposure to hypotonic solutions is followed by RVD, which is mediated by KCl loss via K⁺ and Cl channels (23, 24, 26, 29, 31). Literature data are less abundant, however, in relation to such properties in fish epithelia. Nevertheless, in proximal renal tubules from Salmo trutta or Carassius auratus (21, 22), RVD after hypotonic swelling appears to be the result of K⁺ efflux through quinidine-sensitive K⁺ channels and to increased Cl conductance.

The observation that the Cl channel blocker NPPB was a potent inhibitor of swelling-activated ⁸⁶Rb⁺ efflux in gill cells in primary culture suggests a relationship between swelling-induced K⁺ and Cl channels in cultured gill cells. The first experiments that we performed to characterize anion movements during osmotic challenge indicated that Cl effluxes also increased during the hypotonic shock. This efflux occurred via a conductive pathway because it was blocked by NPPB. Interestingly, this basolateral ¹²⁵I efflux was not modified by quinidine. To undergo the process of RVD, it is possible that a swelling-sensitive Cl efflux drives the K⁺ efflux by depolarizing the cell membrane.

The effect of exposure of cells to a hypotonic shock is not symmetrical in cultured gill cells. The activation of ⁸⁶Rb⁺ effluxes occurred only when the shock was applied to the basolateral surface of the epithelium. Therefore, it could be hypothesized that the apical membrane is impermeable to water, whereas the basolateral membrane is permeable.

In conclusion, the present study shows that gill cells in primary culture present at least two distinct K⁺ conductances activated by hypotonic shock. One of them, a charybdotoxin-sensitive conductance is located in the basolateral membrane and shares its characteristics with calcium-sensitive BK⁺ channels. A second conductance, present in the apical membrane only, is blocked by gadolinium and could be due to the activation of a large conductance stretch-sensitive K⁺ channel. These two conductances seem to be involved in volume regulation phenomena after hypotonic shock in conjunction with a hypotonicity-sensitive Cl conductance located in the basolateral membrane.

Perspectives

It remains to be shown that gill cells are able to undergo the RVD process in response to hyposmotic shock and the related implication of Cl conductance possibly driving the K⁺ efflux under this condition.