The Erythrocyte Na⁺ /H⁺ Exchangers of Eel (Anguilla Anguilla) and Rainbow Trout (Oncorhynchus Mykiss): a Comparative Study

The amiloride-sensitive Na⁺ /H⁺ exchanger is a plasma membrane transport protein that is activated by a number of external signals (osmotic pressure, the presence of spermatozoa, phorbol esters, growth factors, hormones and neurotransmitters) and, through such activation, plays a major role in such processes as cell volume regulation, intracellular pH homeostasis, mitogenesis and secretion. The molecular processes involved in activation and regulation of Na⁺ H⁺ antiporters are still not well understood (Tse et al. 1993). Nucleated erythrocytes obtained from non-mammalian species represent a convenient model to study the regulation of such a membrane transport protein because of their easy availability and capacity to respond to a wide range of physiological stimuli, including hormones. Moreover, acquiring mRNA from circulating erythrocytes allows the transport proteins specifically expressed in these cells to be cloned, thus enabling structure–function analyses to be carried out.

The data on Na⁺ /H⁺ exchangers, obtained from the only two types of nucleated erythrocytes extensively studied so far, show that the regulatory properties of these exchangers are very different. In the salamander (Amphiuma sp.) red blood cell, Na⁺ /H⁺ exchange has been found to be strongly activated by cell shrinkage and intracellular acidification but not by catecholamines and other hormones (Cala, 1983; Cala and Maldonado, 1994). Moreover, in response to swelling, a K⁺_int/H⁺_ext exchange is activated, allowing cell volume regulation. It is the only example, to our knowledge, involving activation of K⁺ /H⁺ exchanges during a regulatory volume decrease and it has been suggested that these exchanges are mediated by the Na⁺ /H⁺ antiporter while it is functioning in a different mode (Cala, 1986). In contrast, in the fish (trout) erythrocyte, no evidence of a K⁺ /H⁺ exchange mediated by the Na⁺ /H⁺ antiporter has been obtained (F. Borgese, F. Garcia-Romeu, H. Guizouarn and R. Motais, unpublished results). The exchange has been shown to be insensitive to intracellular acidification, poorly activated by cell shrinkage but strongly activated by catecholamines (Baroin et al. 1984; Nikinmaa and Huestis, 1984; Cossins and Richardson, 1985; Borgese et al. 1986), by cyclic AMP (Mahé et al. 1985) and, to a lesser extent, by phorbol esters (Motais et al. 1992; Guizouarn et al. 1993). In addition, after activation by hormones or cyclic AMP, the antiporter rapidly desensitizes and remains for hours in a refractory state (Garcia-Romeu et al. 1988; Guizouarn et al. 1993), a regulatory process not so far described in other Na⁺ /H⁺ antiporters. The activation of the red cell antiporters in trout is controlled by the level of blood oxygen (Motais et al. 1987; Claireaux et al. 1988; Fiévet and Motais, 1991), the role of the antiporter being to increase the oxygen-transporting capacity of erythrocytes in hypoxic conditions (Nikinmaa, 1982; Nikinmaa and Huestis, 1984; Cossins and Richardson, 1985; Boutilier et al. 1986; Claireaux et al. 1988).

The trout red cell antiporter, termed βNHE because it is activated by β-agonists, has been cloned. This isoform has two cyclic-AMP-dependent protein kinase consensus sequences in the cytoplasmic C terminus (Borgese et al. 1992, 1994). Cloning of the Amphiuma isoform, which is very volume-sensitive, or that of another species with the same characteristics, would provide an insight into the mechanisms involved in activation by cell shrinkage and would also permit an evaluation of the putative ability of this Na⁺ /H⁺ antiporter to perform K⁺ /H⁺ exchange.

With the aim of increasing our knowledge of the divergent regulatory patterns of nucleated red cell Na⁺ /H⁺ antiporters, we made parallel analyses of the capacities of eel and trout red cell antiporters to be activated by various stimuli and of the characteristics of their deactivation and desensitization. These results were then compared with those obtained from the Amphiuma antiporter by Cala.

Trout [Oncorhynchus mykiss (Walbaum)=Salmo gairdneri L.] weighing approximately 250 g, obtained from a commercial hatchery and kept for 1 week in the laboratory, were stunned by a sharp blow on the head, and blood was removed from the caudal vein into a heparinized syringe. Eels (Anguilla anguilla L.), also weighing approximately 250 g, were obtained from the Société Piscicole de la Crau, 13280 Raphèle-les-Arles, France. The blood cells were washed in their respective saline solutions and the buffy coat removed by suction. They were then resuspended at a haematocrit of 15 % and incubated overnight at 4 °C in their saline solutions to ensure that they reached a steady state with respect to ion and water contents before experimental treatment.

For trout red cells, the isotonic saline solution (330 mosmol l^-1 ; pH 7.55) contained (mmol l^-1 ): 145 NaCl, 5 CaCl₂, 1 MgSO₄, 4 KCl, 15 Hepes, 5 glucose. For eel red cells, the isotonic saline solution (220 mosmol l^-1 ; pH 7.55) contained (mmol l^-1 ): 85 NaCl, 5 CaCl₂, 1 MgSO₄, 4 KCl, 15 Hepes and 5 glucose.

Na⁺ -free solutions were obtained by substitution of NaCl by choline chloride. The osmolarity of the experimental medium was changed by varying [NaCl] or [choline chloride] or by the addition of sucrose as indicated in the Results section. All media contained 1 mmol l^-1 ouabain.

The techniques used for measuring intracellular ion and water content, rate of unidirectional ²² Na influx and intracellular pH have been described in various publications from our laboratory (Borgese et al. 1986, 1987). Briefly, the cells were separated from the solution by centrifugation in nylon tubes which were then cut to liberate the pellet and the supernatant, both of which were saved for analysis. The mass of the cell pellet (wet and dry) was determined gravimetrically and expressed in g water g^-1 dry cell solids (d.c.s.). Cell water, ion content and ²² Na uptake were calculated after correction for trapped extracellular fluid. The ion content was expressed in μmol g^-1 d.c.s. All experiments were performed at 15 °C.

In order to measure unidirectional Na⁺ influx, cells were suspended in the experimental saline and at regular time intervals three subsamples of this suspension were taken and incubated for 3 min with ²² Na (7.4×10⁵ Bq ml^-1 ). The samples were centrifuged in small nylon tubes and the pellets separated by cutting the tubes. The wet mass, dry mass and ²² Na content of the pellets were measured, together with the specific activity of the supernatant. Sodium influx, J^Na, was calculated and expressed in μmol g^-1 d.c.s min^-1.

In the experiments involving changes in cell volume, several protocols were employed. Cells were usually transferred directly from iso-osmotic to hyperosmotic salines containing Na⁺, thus causing red cell shrinkage, followed by a progressive recovery to their initial volume resulting from the entry of Na⁺ and Cl^- following activation of the Na⁺ ;/H⁺ exchanger. In a second protocol, cells were shrunk in Na⁺ -free hyperosmotic medium. In this experimental condition, the Na⁺ /H⁺ exchanger was activated, but the driving forces for net movement of ions were small and there was no discernible change in cell volume for 1 h. The activation of the Na⁺ /H⁺ exchanger can be studied, however, by adding ²² Na at a tracer dose and measuring its uptake. In some experiment, cells that had been shrunk in a Na⁺ -free hyperosmotic medium were transferred to a Na⁺ -containing hyperosmotic medium of identical osmolarity, and the time courses of and cell volume recovery were followed. Finally, in a third protocol, the time course of deactivation of the exchanger was followed by transferring the cells from a Na⁺ -free hyperosmotic medium to a Na ⁺ -free iso-osmotic solution.

Cells were frozen rapidly, then thawed for 5 min and refrozen; the pH measurements were made immediately after a second thawing using a micro pH electrode.

The phosphatase inhibitors were okadaic acid from Novodirect (France) and calyculin A from Tebu (France). ²² Na was from Amersham Corporation. All other chemicals used were reagent grade.

Fig. 1A,B depicts the time courses of cell volume changes and unidirectional Na⁺ influx for trout and eel erythrocytes after the addition of 5×10⁻⁷ mol l^-1 isoprenaline in a nitrogen atmosphere. Addition of 10⁻³ mol l^-1 8-bromoadenosine 3’,5’-monophosphate (8Br-cyclic-AMP) gave identical results (data not shown). As previously shown (Baroin et al. 1984; Mahé et al. 1985), these compounds strongly stimulate trout red cell Na ⁺ /H⁺ exchangers via activation of the cyclic-AMP-dependent protein kinase A (PKA), which leads to considerable swelling of the erythrocytes as a result of a net uptake of NaCl via the Cl-/HCO₃^- exchanger functionally coupled to the Na ⁺ H⁺ exchanger. In contrast, the two compounds were found to be without significant effect on eel erythrocyte Na ⁺ /H⁺ exchangers.

Fig. 2A,B depicts the time courses of cell volume changes and unidirectional Na⁺ influx for trout and eel erythrocytes after the addition of phorbol ester, an agonist of protein kinase C (PKC). As previously shown (Guizouarn et al. 1993), the trout red cell exchangers were activated by phorbol ester, even though the activation was less marked than that induced by β-agonists (note the difference in the scales in Figs 1 and 2). Clearly, eel erythrocyte Na ⁺ H⁺ exchangers, unlike trout red cell antiporters, are not significantly activated by either PKA or PKC agonists.

Fig. 3A depicts the time course of the change in cell volume of eel erythrocytes subjected to osmotic stress: cells were transferred to a solution with an osmolarity of 1.7 times the control osmolarity. This procedure resulted in an immediate osmotic shrinkage followed by a regulatory phase in which the volume returned to the control level in about 1 h. Amiloride completely inhibited volume recovery. This volume regulation is the result of a net uptake of NaCl caused by the functional coupling between Cl^- /HCO₃^- exchange and the amiloride-sensitive Na⁺ /H⁺ exchange (Fig. 3B,C). In the presence of ouabain, after 60 min the net sodium uptake was 101.20±2.17 μmol g^-1 d.c.s, the net chloride uptake 86.9±15.13 μmol g^-1 d.c.s, and the net potassium uptake (not shown) was 5.90±5.69 μmol g^-1 d.c.s (means ± S.E.M., N=7); the calculated volume increase due to NaCl entry was 0.52 g water g^-1 d.c.s. compared with the observed volume increase of 0.48 g water g^-1 d.c.s.

Fig. 4 compares the abilities of eel and trout erythrocytes to recover their initial volume after a hyperosmotically induced shrinkage of similar amplitude. In trout erythrocytes, volume recovery was only partial and the process stopped after about 1 h.

Fig. 5A,B show the relationship between the Na⁺ /H⁺ exchange rate and cell volume in eel erythrocytes. In Fig. 5A, the rate of unidirectional influx is plotted together with the corresponding cell volume as a function of time after cell shrinkage, the shrinkage being induced by transferring the cells from an iso-osmotic normal saline (Na⁺ -containing medium) to a hyperosmotic saline (obtained by the addition of 165 mosmol l^-1 NaCl). Fig. 5B depicts a similar experiment but, in this case, the cells were transferred from an iso-osmotic normal saline to a nominally Na⁺ -free hyperosmotic medium (external sodium concentration 0.1 mmol l^-1 ). Under these experimental conditions, addition of ²² Na at a trace dose allows an analysis of the rate of Na⁺ /H⁺ exchange as a function of time, but as significant Na⁺ uptake cannot take place, cell volume recovery is prevented. Note that in both cases, although cell shrinkage was very rapid, the rate of Na⁺ influx only increased gradually, taking about 10 min to reach a maximum value. Thus, activation of antiporters by cell shrinkage is a relatively slow process.

Figs 6B and 7 show that the magnitude of activation is a graded function of cell shrinkage. Fig. 6A depicts the time course of cell volume changes and Fig. 6B that of Na⁺ influx when eel erythrocytes were osmotically shrunk by transfer from an iso-osmotic solution (220 mosmol l^-1 ) to different hyperosmotic Na⁺ -free media (+40 mosmol l^-1, +100 mosmol l^-1, +160 mosmol l^-1 ) in which the cells cannot recover their initial volumes after shrinkage. Fig. 7 is a plot of Na⁺ influx as a function of cell volume: the cells were transferred from an iso-osmotic solution to different hyperosmotic media containing 85 mmol l^-1 Na+, and the Na⁺ influxes measured 5–10 min later were plotted as a function of the cell volume measured immediately after shrinkage. All these data show that the degree of activation of antiporters is finely adjusted to the degree of cell shrinkage: the greater the shrinkage, the greater the activation.

Figs 5 and 6 also show that the antiporters, once activated by shrinkage, can be rapidly deactivated, or not, depending on the ability of the cells to recover their initial volume. When the cell recovered its volume (Fig. 5A) the rate of Na⁺ influx, maximal at 10 min, decreased rapidly in an exponential manner to reach a value close to the basal value after 1 h; thus, deactivation started when only 30 % of cell volume had been regained and full deactivation occurred after complete recovery. Conversely, the experiments performed in Na⁺ -free media (Figs 5B, 6) show that antiporters remain activated as long as the cells remain shrunk irrespective of the level of activation imposed by the degree of cell shrinkage. This explains why, as illustrated in Fig. 5C, when erythrocytes have been preincubated in a Na⁺ -free hyperosmotic medium for 40 min before being transferred to a Na⁺ -containing medium with the same osmolarity, the rate of Na⁺ influx does not increase progressively but is immediately maximal (the dashed line in Fig. 5C reproduces the results from erythrocytes not preincubated in a Na⁺ -free medium shown in Fig. 5A).

Fig. 8 indicates clearly that the degree of cell shrinkage controls the deactivation of antiporters. In this experiment, two batches of erythrocytes were transferred from an iso-osmotic saline to a hyperosmotic one; after shrinkage, cells started to recover their initial volume (Fig. 8A) by increasing the rate of Na⁺ influx (Fig. 8B). After 18 min, successive additions of sucrose were made to one of the two batches to counteract cell volume recovery and to maintain these cells at a constant volume (Fig. 8A). Under these conditions, the rate of Na⁺ influx remained high and constant (Fig. 8B), despite the fact that antiporters were causing a net Na⁺ uptake. In conclusion, as long as cells remain in a shrunken conditions, antiporters remain activated, whether they transport ions or not, and deactivation is a graded and inverse function of the volume of shrunken cells; it starts when 30 % of the cell volume has been regained and is complete when the initial volume is reached.

The data in Fig. 9 show the time required for antiporter deactivation when the volume recovery is practically instantaneous. In this experiment, red cells were shrunk in a Na+-free hyperosmotic medium for 30 min (conditions in which antiporters are maximally activated but cells cannot regulate their volume) and subsequently transferred back to an iso-osmotic Na ⁺ -free saline; this transfer results in a very fast return to the normal cell volume by osmotic entry of water. Note that full deactivation occurred only about 20 min later. Subsequent hypertonic shock results in reactivation.

The pattern of activation of trout Na⁺ /H⁺ antiporters by cell shrinkage is very similar to that described for eel antiporters, i.e. slow and progressive activation (taking about 15 min); also, the antiporters did not deactivate in the absence of cell volume recovery in Na⁺ -free medium (data not shown). However, there are two differences between the trout and eel antiporters: for a similar degree of cell shrinkage, the rate of Na⁺ /H⁺ exchange is significantly lower in trout red cells and, in addition, trout antiporters become fully deactivated well before the cell volume has returned to its initial value, resulting in an incomplete volume recovery (see Fig. 11).

Fig. 10 shows that eel erythrocyte antiporters, activated by an osmotic shrinkage and then deactivated 1 h later after the cell volume had recovered, can again be fully reactivated by a second osmotic stress and initiate a new cell volume recovery. Thus, when cells start to regain their initial volume, the Na⁺ /H⁺ exchange activity decreases progressively and finally stops, but the antiporters can at any time immediately respond to a new osmotic challenge. Such a ‘deactivation’ of the antiporter is unrelated to the antiporter ‘desensitization’ observed when trout red cells are stimulated by catecholamine, cyclic AMP or forskolin. It has already been shown that exposure of trout erythrocytes to isoprenaline or cyclic AMP induces a rapid and considerable activation followed by a progressive decrease in exchange activity (see Fig. 1B). After removal of the agonist by washing the cells for 1 h after hormonal stimulation, the antiporters show a reduced ability to respond to a fresh challenge of hormone or cyclic AMP (Garcia-Romeu et al. 1988). In other words, in this case, the decline of activity does not reflect a simple activation–deactivation transition, but the shift of antiporters from an active to a refractory non-responsive state. This process has been termed desensitization (Garcia-Romeu et al. 1988).

We investigated whether the decrease in the exchange activity observed after activation by shrinkage of trout erythrocytes is a result of deactivation (as in eel erythrocytes) or desensitization (as when these cells are hormonally stimulated). Fig. 11 illustrates the results of an experiment similar to that described in Fig. 10. It is clear that trout erythrocytes respond immediately and fully to the second osmotic challenge, as do eel erythrocytes. In conclusion, the trout antiporter f3NHE, which is desensitized after hormonal activation, is deactivated and not desensitized after osmotic activation.

In Amphiuma erythrocytes, volume regulation following cell swelling is due to activation of a K⁺ /H⁺ exchanger which mediates net K⁺ loss. This type of volume regulatory mechanism is unique, and it has been suggested that these K⁺ /H⁺ exchanges may be mediated by the Na⁺ /H⁺ exchanger operating in a different mode. In view of this, we decided to look for the existence of K⁺ /H⁺ exchanges in eel erythrocytes but we failed to observe such a process after osmotically induced swelling (data not shown).

Although in various cell types, including trout erythrocytes (Cossins et al. 1994), K⁺ loss responsible for a regulatory volume decrease (RVD) appears to be activated by dephosphorylation, in Amphiuma red cells K⁺ /H⁺ exchange activation is brought about by phosphorylation-dependent events. Exposing red cells in an iso-osmotic medium to the phosphatase inhibitor calyculin A resulted in a net K⁺ loss (mediated by K⁺ /H⁺ exchanges) and a Na⁺ gain (mediated by Na⁺ /H⁺ exchanges) at nearly identical rates (Cala et al. 1993). Fig. 12A shows that 0.2 μmol l^-1 calyculin A added to eel erythrocytes suspended in an iso-osmotic saline results in the activation of a Na⁺ influx which increases progressively for the first 10 min after addition of calyculin A and then decreases to zero after 1 h. This Na⁺ influx is amiloride-sensitive and promotes external acidification when the anionic exchanger is blocked by DIDS (4,4′-diisothiocyanato-stilbene-2,2′-disulphonic acid; data not shown), indicating that calyculin A activates Na⁺ /H⁺ exchange. This activation resulted in an increase in cell volume (Fig. 12B) due to the net uptake of Na⁺ and Cl^- (Fig. 12D) via the parallel functioning of Na⁺ /H⁺ and Cl^- /HCO₃^- exchanges. It should be noted that the cell K⁺ content remained constant (Fig. 12C), indicating that calyculin A did not promote the activation of a putative K⁺ /H⁺ exchange, but did prevent the K⁺ loss that is normally induced by cell swelling.

Fig. 12A,B also shows the effects of calyculin A on osmotically shrunken cells. After transfer of erythrocytes to a hyperosmotic saline, the patterns of Na⁺ influx with or without calyculin A were similar, but the magnitude of the flux was consistently higher in the presence of calyculin A (Fig. 12A). As a consequence, shrunken cells treated with calyculin A recovered their initial volume more rapidly than control cells (20–30 min and 60 min, respectively) and continued to swell (Fig. 12B).

It should be noted that another phosphatase inhibitor, okadaic acid, had no significant effect on Na⁺ influx and cell volume either in iso-osmotic saline or in osmotically shrunken cells (Fig. 12A,B).

The trout and eel red blood cell Na⁺ /H⁺ exchangers have been shown to have widely different regulatory properties. Hyde and Perry (1990) demonstrated the adrenergic insensitivity of American eel (Anguilla rostrata) red blood cells in vivo when the eel is exposed to hypoxia or acidosis, conditions known to enhance the catecholamine-mediated erythrocyte response in trout. We confirmed in vitro, using the European eel (Anguilla anguilla), that catecholamine and cyclic AMP do not produce significant activation of the red cell Na⁺ /H⁺ exchangers (Fig. 1). Similarly, phorbol esters, which activate the trout red cell exchanger f3NHE, do not affect the eel exchanger. Furthermore, eel red blood cells show a remarkable capacity for volume regulation: after exposure to 1.7-fold hyperosmotic saline, the initial cell volume recovers fully in 1 h. This volume-induced process, which is faster even than that of Amphiuma erythrocytes, involves the activation of an amiloride-sensitive Na⁺ /H⁺ exchanger that is functionally coupled to the anion exchanger. In contrast, trout red cells exposed to similar variations of osmotic pressure tend to regulate their volume by activating Na⁺ /H⁺ exchange, but activation is only partial since they recover less than half their volume (Fig. 4).

Thus, the regulatory properties of the eel red cell antiporter are very different from those of the trout red cell exchanger but are similar to those described by Cala (1983) for the Amphiuma red cell exchanger. These different regulatory properties probably indicate the existence in nucleated erythrocytes of different isoforms of the exchanger, a hypothesis essentially based on the observation that the sensitivity of μNHE to catecholamines is strictly related to the presence of PKA consensus sites located on the cytoplasmic domain of the antiporter: the point mutation of serine in these sites inhibits catecholamine-induced activation, and grafting these sites to the human exchanger NHE1, normally insensitive to catecholamine, confers a catecholamine-dependence to NHE1 (Borgese et al. 1994).

Our experiments have demonstrated that eel erythrocytes, after osmotic shrinkage, show a remarkable capacity to return to the volume they had in an isotonic medium. This regulatory volume increase (RVI) is mediated by the activation of Na ⁺ /H⁺ exchangers; RVI stops when the initial volume is attained, indicating that all the antiporters are turned off. The term ‘volume set point’ is frequently used to define the volume at which all volume-sensitive transporters are inactive, any deviation from this set point resulting in the activation of shrinkage-activated or swelling-activated pathways.

In a series of careful studies, Cala showed that the shrinkage-activated Amphiuma exchangers possess regulatory properties very similar to those described in this paper for the shrinkage-activated eel exchanger (Cala, 1986; Maldonado and Cala, 1992). The main characteristics of eel antiporter activation and deactivation following exposure to a hyperosmotic medium are listed below.

Activation of the Na⁺ /H⁺ exchanger is induced by cell shrinkage and occurs even in the absence of Na⁺ in the hypertonic medium (Fig. 5A,B), but the magnitude of Na⁺ /H⁺ exchange activation is a graded function of cell shrinkage. The activation is not instantaneous: after shrinkage, both in Na⁺ -containing and in Na⁺ -free medium, the Na⁺ /H⁺ exchange rate gradually increases, reaching a maximum value after about 10 min (Fig. 5A,B). This time lag between osmotic shrinkage and Na⁺ /H⁺ exchange activation is roughly the same irrespective of the degree of osmotic shrinkage (Fig. 6).

Deactivation, like activation, is induced by volume change: as cells progressively swell during RVI, the rate of Na ⁺ /H⁺ exchange decreases simultaneously (Fig. 5A). If cells are kept shrunken either by successive small additions of hypertonic medium to cancel RVI (Fig. 8) or by shrinking cells in a Na ⁺ -free medium in which no net ion uptake is possible (Fig. 5B), their antiporters remain activated and the rate of Na ⁺ /H⁺ exchange is maintained constant. Like activation, deactivation also occurs after some delay, as can be seen in Fig. 9. Here, cells shrunk in a Na ⁺ -free hypertonic medium for 30 min were resuspended in a Na ⁺ -free isotonic medium in which their volume was rapidly restored. The Na ⁺ /H⁺ exchange rate decreased, but deactivation was only complete 20 min later.

Volume change is clearly the signal for both activation and deactivation of antiporters. However, perception of the signal by cells and its transduction are complex phenomena, as illustrated by the following example. Fig. 5 shows that the magnitude of activation of Na⁺ /H⁺ exchange is the same in erythrocytes shrunk by preincubation in a Na⁺ -free hypertonic medium and then transferred to a Na⁺ -containing hypertonic saline of the same osmolarity (no lag time for activation) as in erythrocytes transferred directly to the Na⁺ -containing hypertonic saline (progressive activation). This indicates that the magnitude of activation is strictly controlled by the initial degree of cell shrinkage, but is subsequently unaffected by the partial cell volume recovery that occurs during the progressive activation (about 10 min).

βNHE, the trout antiporter, is much less sensitive to cell shrinkage than the eel antiporter: for a shrinkage of similar magnitude, βNHE carries fewer ions. In addition, it is deactivated prior to restoration of the normal control volume, leaving cell volume regulation notably defective (Fig. 11). It has been suggested that in human lymphocytes a similar failure of volume regulatory capacity could be due to pH-dependent, and not to volume-dependent, deactivation of antiporters (Grinstein et al. 1985). Whether this could explain the situation in βNHE is, as yet, unknown.

A number of findings point to the involvement of phosphorylation–dephosphorylation reactions in the control of volume-activated transport systems. Work on red blood cells has led to the hypothesis that a protein kinase plays a pivotal role. Shrinking the cell stimulates a kinase, leading to activation of transporters involved in regulatory volume increase (RVI), such as the Na⁺ /H⁺ exchanger and Na⁺ /K⁺ /2Cl^- cotransport. Swelling inhibits the kinase, leading to activation of the transporters involved in regulatory volume decrease (RVD), such as K+/Cl^- cotransport (Jennings and Al-Rohil, 1990; Parker et al. 1991; Cossins, 1991; Starke and Jennings, 1993). The target of the kinase may be the transporters. The two types of transporter, however, always appear to be regulated in a coordinated fashion: agents stimulating one type inhibit the other and vice versa (Parker et al. 1990), suggesting that a ‘regulator’ controls the two types of transport system in a reciprocal manner and orchestrates the responses of the cell (Parker, 1993). In this context, the regulator might also be a putative target of the kinase. Phosphatase inhibitors, by shifting the phosphorylation–dephosphorylation equilibrium towards phosphorylation, were shown to have major effects on volume-sensitive transport systems, i.e. stimulation of shrinkage-activated transporters such as the Na⁺ /H⁺ exchanger and inhibition of swelling-activated transporters such as K⁺ /Cl^- cotransport (Jennings and Schulz, 1990; Bianchini et al. 1991; Parker et al. 1991; Cossins et al. 1994; Guizouarn et al. 1995).

The results obtained with Amphiuma red cells do not fit well into this general scheme. In these erythrocytes, the shrinkage-activated pathway is a Na⁺ /H⁺ exchanger and the swelling-activated pathway a K⁺ /H⁺ exchanger. A K⁺ /H⁺ exchanger as a volume-sensitive transport system has only been described in Amphiuma red cells (Cala, 1985). Cala (1986) also suggested that, in these cells, the Na⁺ /H⁺ and K⁺ /H⁺ exchangers could represent alternative forms of the same transport moiety, an interesting possibility which, to our knowledge, has never been considered in relation to other isoforms of Na⁺ /H⁺ antiporters. As in other cells, the two pathways are normally regulated in a coordinated fashion, i.e. when one is turned on, the other is turned off. However, addition of calyculin A, a phosphatase inhibitor, to erythrocytes suspended in an isotonic medium activated Na⁺ /H⁺ exchange but also, unexpectedly, swelling-activated K⁺ /H⁺ exchange, resulting in net K⁺ loss and Na⁺ gain at nearly identical rates (Cala et al. 1993).

In view of similarities between the Na⁺ /H⁺ antiporters of eel and Amphiuma erythrocytes, it was of interest to analyze the effects of phosphatase inhibitors on eel erythrocytes, especially in relation to the expression of a putative K⁺ /H⁺ exchange.

Exposing eel red cells in iso-osmotic medium to calyculin A resulted in cell swelling due to the activation of the Na⁺ /H⁺ exchanger, but the red cell K⁺ content remained constant, indicating that the phosphatase inhibitor had not induced a K⁺ loss (Fig. 12). Thus, the calyculin-A-dependent phosphorylation does not reveal a silent K⁺ /H⁺ exchange in the eel erythrocyte, as in Amphiuma erythrocyte. Furthermore, we also failed to activate K⁺ /H⁺ exchange by osmotically induced cell swelling (data not shown).

When eel and trout red cell Na⁺ /H⁺ antiporters have been activated by cell shrinkage, their activity decreases progressively as a function of time, but they can immediately be restimulated by a new osmotic challenge (Figs 10, 11). Thus, after volume activation, the observed decline of activity reflects a progressive deactivation of the antiporters.

In contrast, the decline of activity observed after hormonal stimulation of βNHE, the trout antiporter (Fig. 1B), does not reflect such a simple activation–deactivation transition, but a transition from an active to a refractory state: βNHE can no longer be immediately reactivated by a fresh challenge with catecholamine or cyclic AMP, and several hours without stimulation are necessary for the exchanger to recover its ability to respond to cyclic AMP or catecholamine. This desensitization is controlled by a phosphorylation– dephosphorylation reaction and may be the result of withdrawal of the protein from the cell surface and its sequestration in an altered conformation (Guizouarn et al. 1993, 1995).

The same Na⁺ /H⁺ exchanger βNHE, once activated, can be either desensitized or simply deactivated depending on the nature of the stimulus. The reason for such divergent behaviour is unknown, but could be related to whether the antiporter is, or is not, phosphorylated: there is quite convincing evidence that catecholamine-induced activation is associated with phosphorylation of βNHE, since the point mutation of only one serine of the PKA consensus sites located on the cytoplasmic domain of the exchanger drastically inhibits activation (Borgese et al. 1994). In contrast, shrinkage-induced activation of NHE1, the human antiporter, has been shown not to be associated with phosphorylation of the transport molecule (Grinstein et al. 1992). This finding does not invalidate the model in which the kinase stimulation caused by shrinkage is likely to be responsible for the Na⁺ /H⁺ exchange. The substrate for the shrinkage-induced phosphorylation would not be the exchanger itself but another element, such as the regulator proposed by Parker (1993).

We thank Dr Peter Cala for helpful discussions. This work was supported by grants from the Centre National de la Recherche Scientifique (URA 1855).