Oxygen-sensitive regulatory volume increase and Na transport in red blood cells from the cane toad, Bufo marinus

Red blood cells (RBCs) in nearly all vertebrate classes ranging from agnathans to mammals restore volume upon hyperosmotic shrinkage by stimulation of Na uptake through the Na/H exchanger (NHE) or the Na/K/2Cl co-transporter(NKCC) (Cossins and Gibson,1997; Gibson et al.,2000). The regulatory volume increase (RVI) in RBCs from mammals,most teleosts and amphibians is mediated by increased NHE activity, resulting in a net gain of inorganic osmolytes followed by osmotically obliged water(Cala, 1980; Parker et al., 1991; Orlov et al., 1994; Romero et al., 1996; Gusev et al., 1997; Weaver et al., 1999) whereas bird RBCs activate the NKCC (Kregenow,1981; Muzyamba et al.,1999). In addition to shrinkage, the activity of the NHE and NKCC and other RBC ion transporters is modulated by oxygen, catecholamines or acidification [NHE (Motais et al.,1987; Salama and Nikinmaa,1988; Nielsen,1997; Virkki et al.,1998) and NKCC (Palfrey and Greengard, 1981; Muzyamba et al., 1999; Flatman,2005; Berenbrink et al.,2006)]. Thus, in various vertebrates, the NHE, NKCC and/or RVI response are inhibited in air-equilibrated cells compared with cells at low oxygen partial pressure (P_O2)(Gibson et al., 2000; Brauner et al., 2002; Jensen et al., 2002; Drew et al., 2004).

Plasma osmolality increases naturally in animals during dehydration,physical exercise and upon movement from freshwater to seawater(Shoemaker, 1964; Maxime et al., 1991; Madsen et al., 1996; McKenna et al., 1997; Hyndman et al., 2003; Peterson and Greenshields,2001). Because of their water-permeable skin, amphibians are prone to dehydration, but the dependence on water differs among species. Terrestrial anurans, such as Bufo, tend to be rather tolerant to dehydration(Thorson, 1955), and the cane toad, Bufo marinus, tolerates severe dehydration with 30% loss of body mass (Shoemaker, 1964; Shoemaker, 1965). However,little is known about the RVI response from terrestrial amphibians. Furthermore, the cardiac shunts of the amphibian circulatory system cause desaturation of arterial blood (Johansen and Ditadi, 1966; Wang et al.,1998) and it is of interest, therefore, to examine how RBC volume regulation is affected by oxygen in these animals.

Catecholamines are potent activators of the NKCC and NHE in many vertebrate RBCs (Palfrey et al., 1981; Motais et al., 1987; Nikinmaa, 1992; Weaver et al., 1999; Pedersen and Cala, 2004; Berenbrink et al., 2005) with maximal stimulation at low P_O2 levels(Motais et al., 1987; Salama and Nikinmaa, 1988). The influence of adrenergic stimulation on isotonic volume responses and ion fluxes has been examined in some amphibian RBCs, but responses were only significant in the presence of phosphodiesterase inhibitors; thus, the physiological significance of the responses remains unknown(Rudolph and Greengard, 1980; Tufts et al., 1987a; Tufts et al., 1987b; Kaloyianni et al., 1997).

The present work examines the P_O2 dependency of the Na transport mechanism induced by hyperosmotic shrinkage orβ-adrenergic stimulation in washed B. marinus RBCs using ouabain-insensitive unidirectional ²²Na influx. In parallel experiments, the influence of oxygenation on the ability to restore cell volume after hyperosmotic shrinkage is studied in both whole blood cells and washed RBCs. The Na influx and RVI mechanisms are characterized pharmacologically.

The experiments were performed on 26 cane toads, Bufo marinus L.,of undetermined sex and a body mass of between 149 and 360 g (214±9 g;mean ± s.e.m.). They were purchased from a commercial supplier (Exotic Tropicals, St Michael, Barbados) and kept in containers (100×50×75 cm; L×W×H) with access to water and dry areas at approximately 25°C. The animals were kept in the holding facilities for at least two weeks before use and were fed mealworms and liver pieces several times a week.

Animals were anaesthetized by immersion in a solution (1 g l^-1)of benzocaine (ethyl p-amino benzoate). When all reflexes had disappeared, an incision was made in the hind limb and a blood sample(3–4 ml) was obtained immediately after inserting a heparinised cannula(PE-50) into the femoral artery. The cannula was removed and the incision was closed using sutures, whereupon the animal was placed under running tapwater until it restored spontaneous ventilation. Andersen and Wang reported no change in plasma lactate concentration immediately after similar surgical procedure in B. marinus (Andersen and Wang, 2002) and this sampling procedure should therefore not affect whole blood in vitro studies.

A sub-sample of the blood was centrifuged at 5°C for 3 min at 1700 g (Sigma-3MK, Osterode, Germany) to remove plasma and buffy coat. The cells were then washed three times in 10 volumes of ice-cold isotonic saline and kept oxygenated by contact with air at 5°C overnight for use in the experiments on washed blood cells. The standard isotonic saline contained: 105 mmol l^-1 NaCl, 6 mmol l^-1 KCl, 1 mmol l^-1 MgSO₄, 5 mmol l^-1CaCl₂.2H₂O, 5 mmol l^-1d-glucose and 10 mmol l^-1 Hepes. It was matched to the plasma osmolality of B. marinus of 240 mOsm kg^-1 and adjusted to pH 7.83 at 25°C (Andersen and Wang,2003).

After overnight storage at 5°C, the RBCs were washed twice in the isotonic saline and adjusted to a haematocrit (Hct) of 10–20%. For each experiment, 800 μl of RBC suspension was equilibrated with humidified air in a rotating Eschweiler tonometer (obtained from the Dept of Chemistry,Aarhus University, Denmark) at 25°C for 45 min before ouabain was added to a final concentration of 10^–4 mol l^-1. Hct was determined in duplicate and, after an additional 10 min (time zero in the experiment), aliquots of RBC suspension were diluted 10 times by transfer to test tubes containing air-equilibrated isotonic saline, ²²Na(∼16.7 kBq ml^-1), ouabain (10^–4 mol l^-1) and other transport modifiers (specified below). When studying effects of a change in P_O2, the tonometer and test tubes were gassed with the experimental gas mixture for 10 min before transfer of RBCs. The blood cells were subjected to P_O2 values ranging from 0 (pure N₂)to 20.5 kPa (air) under isotonic conditions or a calculated 25% shrinkage with or without 10^–4 mol l^-1 EIPA[5-(N-ethyl-N-isopropyl)amiloride]. In shrunken RBCs incubated under nitrogen, a dose–response curve for EIPA was constructed using final concentrations between 10^–7 and 10^–5 mol l^-1. The effect of 10^–5mol l^-1 isoproterenol with or without 10^–4 mol l^-1 EIPA (to inhibit the NHE) or 10^–5 mol l^-1 bumetanide (to inhibit the NKCC) was studied in cells incubated in isotonic saline under humidified nitrogen. Finally, the effect of extracellular acidification to pH 7.28 by addition of appropriate volumes of a pH 7.2 saline and of 10^–5 mol l^-1 isoproterenol was studied in cells suspended in isotonic saline under air.

A sub-sample of ∼2 ml freshly drawn blood was kept on ice for 1–2 h until used for whole-blood RVI studies. Approximately 0.5 ml (or 0.7 ml if pH was measured; see below) was incubated in each of four rotating Eschweiler tonometers. The tonometers were immersed in a water bath kept at 25°C and supplied with a humidified gas mixture of 1% or 3% CO₂, balanced with air, through a Wösthoff gas mixing pump (Bochum, Germany) to achieve physiologically relevant P_CO2 and pH values for resting and active animals, respectively(Andersen and Wang, 2003). After a 30 min equilibration period, samples were taken for duplicate pH, Hct and haemoglobin concentration ([Hb]) determinations and, five minutes later(time zero in the experiment), the blood cells were subjected to the following manipulations: (1) osmotic shrinkage alone by addition of appropriate volumes of a 2.3 Osm kg^-1 stock solution (sucrose-containing standard saline) to increase plasma osmolarity from ∼240 to ∼320 mOsm l^-1 (calculated to yield 25% cell shrinkage) or with (2)simultaneous addition of amiloride to a final concentration of 10^–4 mol l^-1 to inhibit the NHE or (3)simultaneous addition of bumetanide to a final concentration of 10^–5 mol l^-1 to inhibit the NKCC. In addition,parallel control experiments with no osmotic disturbance were performed to determine the influence of incubation alone on the RBCs.

Duplicate Hct determinations were made at time zero as well as 15, 30, 60,90 and 120 min after treatment, and duplicate samples for [Hb] determinations were taken at 0, 60 and 120 min. Haematocrit was determined after centrifuging heparinized micro-haematocrit tubes at 16 000 g for 3 min in a haemofuge (Heraeus Sepatech, Usingen, Germany). After conversion of Hb to cyanomethaemoglobin by Drabkin's reagent {11.9 mmol l^-1NaHCO₃, 0.61 mmol l^-1 K₃[Fe(CN)₆],0.77 mmol l^-1 KCN and 0.05% v/v Triton X-100}, [Hb] was determined at a wavelength of 540 nm (Ultrospec 2; LKB Biochrom, Cambridge, UK), applying a millimolar extinction coefficient of 11.0. The [Hb] was constant after time zero, so an average value was calculated for the time 0, 60 and 120 min measurements and used to calculate mean cellular haemoglobin concentration(MCHC) as [Hb]/Hct at all sampling points after treatment. MCHC was used as an indicator of red cell volume changes. pH was measured with a Radiometer pH electrode (BMS2, Copenhagen, Denmark) at 25°C.

Washed RBCs, as above, were washed twice in the isotonic saline and adjusted to an Hct value of 20–25% (22.2±0.9%, N=16). The RBC suspension was then transferred to a rotating Eschweiler tonometer kept in a water bath at 25°C and equilibrated to humidified air. After 20 min, ouabain was added to a final concentration of 10^–4 mol l^-1, and 5 min later samples were taken for duplicate determinations of Hct and [Hb]. After an additional 5 min (time zero in the experiment), the suspension was distributed into open test tubes and subjected to the same manipulations and sampling procedures as described above for whole blood. The test tubes were kept at 25°C throughout the experiment and gently swirled regularly to ensure oxygenation and avoid sedimentation. In addition, identical volume recovery experiments on deoxygenated washed RBCs were performed. To achieve deoxygenated conditions, humidified N₂replaced air in the tonometer 10 min before the suspension was transferred to test tubes (at time zero in the experiment), which were also gassed with humidified N₂.

Hyperosmotic shrinkage of RBCs suspended in physiological saline significantly increased a ouabain-insensitive Na influx at all P_O2 values(Fig. 1). Within 5–15 min, there was a strong P_O2 dependency of the shrinkage-induced Na influx, which increased significantly at a P_O2 of 4.1 kPa compared with at 8.2 and 20.5 kPa and reached a maximal value of ∼50 mmol Na l^-1 RBCs h^-1 in deoxygenated cells (0 kPa P_O2) (Fig. 1A). The kinetics of the shrinkage-induced Na influx also depended on P_O2. The Na influx in deoxygenated cells remained high throughout the initial 30 min(Fig. 1A,B), whereupon it returned to basal levels between 45 and 60 min(Fig. 1D). However, while the Na influx at 4.1 kPa P_O2 also remained high(36–42 mmol Na l^-1 RBCs h^-1) during the first 30 min, the influx declined but remained significantly above basal levels throughout the experiment (Fig. 1). At higher P_O2 values, there appeared to be a 15-min lag before maximal Na influx values of ∼30 mmol Na l^-1 RBCs h^-1 were reached between 15 and 45 min(Fig. 1B,C), whereupon the influx declined but remained significantly elevated at 20–25 mmol Na l^-1 RBCs h^-1 (Fig. 1D).

The P_O2 dependency of the shrinkage-induced Na influx was maximal during the first 5–15 min(Fig. 1A), decreased between 15 and 30 min (Fig. 1B), and between 30 and 45 min there was no significant P_O2 dependency of the shrinkage-induced Na influx (Fig. 1C). In the 45–60 min interval, the P_O2 dependency even reversed, with maximal shrinkage-induced Na influxes at 8.2 and 20.5 kPa P_O2 (Fig. 1D). In the 5–15 min interval, a P₅₀value of 5.5±1.0 kPa (N=4) was determined for shrinkage-induced Na influx.

The shrinkage-induced Na influx was completely abolished by 10^–4 mol l^-1 EIPA, a selective inhibitor of the NHE, at all P_O2 values(Fig. 1). A dose–response curve for EIPA is shown in Fig. 2, where the effect of EIPA was studied in deoxygenated RBCs exposed to a calculated 25% shrinkage. An IC₅₀ of∼10^–6 mol l^-1 EIPA can be estimated from the dose–response curve.

The shrinkage-induced Na uptake was associated with RVI in washed RBCs kept under conditions identical to those in Fig. 1 and exposed to 0 or 20.5 kPa P_O2(Fig. 3A). The RVI response was faster in deoxygenated compared with oxygenated RBCs, with MCHC recovery corresponding to 50% of the initial change attained after ∼23 and 70 min,respectively, under the two conditions(Fig. 3B). The faster RVI response of deoxygenated RBCs (Fig. 3B) reflects that the Na influx was already maximal in the initial two time intervals under those conditions(Fig. 1A,B), whereas maximal Na influx values were not attained until after 15–45 min in oxygenated cells (Fig. 1B,C), delaying the RVI response (Fig. 3B). Maximal volume recovery of deoxygenated cells was attained after ∼30 min(Fig. 3B), after which the Na influx was reduced and reached basal levels after 45–60 min(Fig. 1C,D). The volume recovery was slower in oxygenated cells, but after 90 min the extent of volume recovery (∼60%) was similar to that attained in deoxygenated cells(Fig. 3B), reflecting the more sustained Na influx in cells kept at 20.5 kPa P_O2 (Fig. 1B–D). The RVI response was substantially reduced by the NHE inhibitor amiloride (10^–4 mol l^-1), whereas the NKCC inhibitor bumetanide (10^–5 mol l^-1) had no effect (Fig. 3C).

The above experiments were carried out on RBCs suspended in a physiological saline to allow comparisons with the Na influx data. However, RVI experiments on whole blood were also carried out under similar conditions of pH and temperature. Fig. 4 depicts the time courses in MCHC of whole blood at 19.9–20.3 kPa P_O2 after different treatments at 1%CO₂ (P_CO2 ∼1 kPa) and 3%CO₂ (P_CO2 ∼3 kPa), corresponding to pH 7.86±0.05 (N=7) and 7.55±0.03 (N=4),respectively, as measured immediately before shrinkage. These are typical P_CO2 and pH values for resting and active or burrowing animals, respectively (Boutilier et al., 1979). MCHC of untreated control cells at both P_CO2 values did not change significantly during the experiment (Fig. 4A,B). At pH 7.86, the RVI response after a calculated 25% shrinkage of the cells was statistically significant from 30 min onwards, and after 120 min volume recovery had attained 89% (Fig. 4A). Amiloride (10^–4 mol l^-1)significantly reduced RVI, whereas bumetanide (10^–5 mol l^-1) did not have any effect on the extent of RVI(Fig. 4A). Whole blood, at pH and P_CO2 values more typical for an active or burrowing animal, showed an overall lowering of the MCHC from ∼4.7 to∼3.9 mmol Hb₄ l^-1 RBCs and, thereby, had a new increased steady-state RBC volume (Fig. 4B). The time course and extent of volume recovery after shrinkage were similar to those at 1% CO₂ but, due to larger differences in MCHC values between individual experiments, there was no difference between MCHC values of control and shrunken cells from 60 min onwards(Fig. 4B).

A three-way ANOVA comparing the whole-blood experiments(Fig. 4A) to parallel RVI experiments on oxygenated washed RBCs (Fig. 4C) showed that the extent and inhibitor sensitivity of the RVI were not different under similar conditions of temperature and pH. Haemolysis was not observed in whole blood up to 120 min of experimentation, whereas haemolysis started to develop after 90 min in washed cell suspensions. Therefore, data at 120 min for washed cells are not included.

In the 15–30 min interval, reduction of P_O2 to 4.1 and 0 kPa induced a 7- and 16-fold increase in the ouabain-insensitive basal flux, respectively, relative to that measured in fully oxygenated cells (Fig. 1B). In cells exposed to a P_O2 of 0 kPa, a 10-fold increase in basal flux relative to that of oxygenated cells was still present in the 30–45 min interval(Fig. 1C). Deoxygenation per se increased cell volume, as illustrated by significantly lowered MCHC values compared with oxygenated RBCs from 60 min onwards(Fig. 3A), corresponding well with the increased Na influx in deoxygenated RBCs in the 15–45 min interval (Fig. 1B,C).

Extracellular acidification (to pH 7.28) of oxygenated RBCs suspended in isotonic saline did not change the Na influx compared with basal values. Thus,basal Na influxes in oxygenated RBCs in the 5–15, 15–30 and 30–45 min intervals were 3.9±2.1, 0.4±1.3 and 2.2±1.2 mmol l^-1 RBCs h^-1 (N=6) whereas the Na influxes in acidified RBCs were 1.5±2.9, 3.5±2.8 and 1.3±1.0 mmol l^-1 RBCs h^-1 (N=6).

The β-adrenergic agonist isoproterenol significantly increased the ouabain-insensitive Na influx in deoxygenated RBCs from ∼6 and ∼11 mmol Na l^-1 RBCs h^-1 to ∼25 and ∼35 mmol Na l^-1 RBCs h^-1, respectively, within the first two time intervals (Fig. 5A). The effect of isoproterenol on Na influx was statistically significant for up to 45 min,whereupon the influx was reduced to basal levels(Fig. 5A). Bumetanide did not significantly reduce the isoproterenol-stimulated Na influx at any time,whereas EIPA completely blocked the effects of isoproterenol on Na influx(Fig. 5A). In air-equilibrated RBCs, on the other hand, isoproterenol had no effect on Na influx within 45 min of stimulation (Fig. 5B).

Oxygenation-dependent RBC membrane transporters have now been described in nearly all vertebrate classes, ranging from lamprey to mammals(Motais et al., 1987; Nielsen et al., 1992; Gibson et al., 1998; Virkki et al., 1998; Muzyamba et al., 1999; Berenbrink et al., 2000; Berenbrink et al., 2006; Flatman, 2005) (reviewed by Gibson et al., 2000). This also applies to the transporters involved in volume regulation(Jensen, 1995; Nielsen, 1997; Gibson et al., 2000; Brauner et al., 2002; Drew et al., 2004). In Bufo marinus, there is a strong P_O2dependency of the shrinkage-induced Na influx, with maximal stimulation at low P_O2 during the initial 30 min(Fig. 1). The shrinkage-induced Na influx was completely abolished by the specific NHE inhibitor EIPA(10^–4 mol l^-1), showing that the Na uptake was entirely due to activation of the NHE (Fig. 1). The P₅₀ value of 5.5 kPa for half-maximal shrinkage-induced Na influx in the first time interval is similar to the in vitro Hb P₅₀ value of 5.3 kPa that can be estimated for B. marinus at a similar pH and temperature [based on data in table 1 of Andersen et al.(Andersen et al., 2001)]. Therefore, with matching oxygen affinities of NHE activation and Hb, it is conceivable that Hb is involved in oxygen sensing and modulation of NHE activation in B. marinus RBCs. This role of Hb has been suggested for modulation of NHE and KCC (K/Cl co-transporter) activity in rainbow trout RBCs(Motais et al., 1987; Nielsen et al., 1992). However, other studies on rainbow trout and crucian carp RBCs indicate that the P₅₀ of KCC activation is significantly different from the P₅₀ of Hb, implying that the effect of oxygen is not regulated by bulk Hb oxygen saturation(Berenbrink et al., 2000; Berenbrink et al., 2006). In rainbow trout, the effect of oxygen on transporter activity seems to be mediated through hydroxyl radicals, which stimulate KCC activity, whilst they inhibit the βNHE at high P_O2(Bogdanova and Nikinmaa, 2001; Nikinmaa et al., 2003). The oxygen dependency of NKCC activity in RBCs of some tetrapods such as turkey and ferret and in crucian carp, however, is consistent with modulation by the Hb oxygenation state (Muzyamba et al.,1999; Flatman,2005; Berenbrink et al.,2006).

Shrinkage of B. marinus RBCs was attended by a pronounced amiloride-sensitive RVI response (Fig. 3C, Fig. 4C) that was much faster under deoxygenated conditions(Fig. 3B), reflecting the time course of activation and the magnitude of the Na influxes under the two levels of P_O2 (Fig. 1). The RVI response was, however, not complete under either of the conditions within the 90–120 min studied here. It appears that the fast RVI response in deoxygenated B. marinus RBCs is due to immediate and large NHE activation, whereas the slower RVI in oxygenated cells reflects the fact that the Na influx is not maximal until 15–45 min after shrinkage activation (Fig. 1, Fig. 3B). In European flounder,the RVI is progressively faster with decreasing levels of Hb oxygen saturation(Jensen et al., 2002). In light of this, studies carried out at high P_O2values, as for example during air equilibration, may mistakenly underestimate the ability of RBCs to perform RVI under conditions prevailing in the venous circulation or, in animals with cardiac shunts, even in arterial blood.

The NHE also mediates RVI in Rana ridibunda RBCs, and a flux of 11.6±1.1 mmol Na l^-1 RBCs h^-1 under oxygenated conditions was measured within the first 30 min after a calculated 31%shrinkage (Gusev and Ivanova,2003). This Na influx is lower than the influx between 0 and 30 min in B. marinus RBCs, which at P_O220.5 kPa can be estimated to be 19.1±2.0 mmol Na l^-1 RBCs h^-1 (N=4) and it suggests that, at least initially, the NHE activity upon shrinkage is larger in the terrestrial toad than in the semi-aquatic frog. These data, however, are too sparse to allow a realistic comparison of the RVI ability of terrestrial and semi-aquatic amphibians because, as shown here, the RVI responses depend on blood P_O2 and time frame.

The mechanism of RVI in B. marinus RBCs seems to be the same at low and high P_O2 as the response was significantly reduced by amiloride but insensitive to bumetanide under both conditions (Fig. 3C and Fig. 4C, respectively). At a concentration of 10^–4 mol l^-1, amiloride did not completely block the RVI response, which was also found in shrunken Rana temporaria RBCs (Jørgensen,1995). Based on the Na influx experiments(Fig. 1), it is likely that 10^–4 mol l^-1 EIPA would abolish the RVI completely, but at present we cannot rule out that other, Na-independent,mechanisms may be involved in the RVI response.

Many RVI studies are only carried out on cells suspended in defined,artificial media containing Na⁺/K⁺-ATPase inhibitors(Cala, 1977; Siebens and Kregenow, 1985; Romero et al., 1996; Weaver et al., 1999). Our study shows that the RVI response was not significantly different in whole blood and washed RBCs from B. marinus(Fig. 4). RVI did therefore not depend on washing and storage procedures or activity of the Na⁺/K⁺-ATPase. Cala also reported that RVI did not depend on the Na⁺/K⁺-ATPase(Cala, 1977), and Caldwell et al. showed that effects of β-adrenergic stimulation in trout RBCs taken via caudal puncture and then washed and stored for up to 96 h were not significantly different from those of whole blood taken viacannulae 48 h after surgery (Caldwell et al., 2006). Haemolysis was not detected in B. marinuswhole blood during the course of the experiment, making it appear less fragile than the washed RBCs. However, the influence of, for instance, stress hormones can be avoided when replacing the plasma and incubating the cells overnight(Bourne and Cossins,1982).

Volume changes due to the Haldane effect and associated shifts in the Donnan equilibrium are very fast (Hladky and Rink, 1977; Borgese et al.,1991) and can therefore not explain the slow volume increase in B. marinus RBCs caused by deoxygenation alone(Fig. 3A). The same treatment caused a significant increase in the basal Na influx(Fig. 1B,C), which could explain the volume increase. The P₅₀ for half-maximal deoxygenation-induced Na influx in the 15–30 min time interval is in the same range as the values for Hb and shrinkage-induced Na influx (see above)and, thus, support that Hb oxygenation could also modulate basal NHE activity. The mechanism of the increased deoxygenation-induced Na influx was not addressed specifically, but 10^–4 mol l^-1 EIPA lowered the shrinkage-induced Na influx in deoxygenated cells to basal values measured in oxygenated RBCs (Fig. 1B,C), so it likely involved the NHE. Weaver et al. showed that deoxygenation per se activates the NHE in European flounder RBCs(Weaver et al., 1999).

Elevation of P_CO2 caused swelling of the RBCs (Fig. 4B), but since extracellular acidification in washed RBCs did not influence the Na influx(see above), swelling is probably caused by a change in the Donnan equilibrium across the membrane and not a pH regulatory Na transport. In rainbow trout RBCs, extracellular acidification stimulates the NHE in the acidic pH range,whereas intracellular acidification stimulates the NHE in the alkaline pH range (Borgese et al., 1987). Intracellular acidification also stimulates the NHE in Amphiuma tridactylum and R. temporaria RBCs(Cala and Maldonado, 1994; Jørgensen, 1995). This regulation of intracellular pH may help maintain the oxygen affinity of haemoglobin. As at 1 kPa P_CO2, bumetanide did not affect RVI at 3 kPa P_CO2, indicating also that, under conditions typical for an active or burrowing animal, the NKCC did not contribute to volume regulation.

Stimulation by the β-adrenergic agonist isoproterenol at 0 kPa P_O2 elicited a pronounced increase in the EIPA-sensitive and bumetanide-insensitive Na influx(Fig. 5A). At present, we do not know how widespread β-adrenergic NHE activation is within amphibians,but it was not found in deoxygenated Xenopus laevis RBCs(Berenbrink et al., 2005). Adrenergic stimulation increased cAMP and caused swelling of deoxygenated RBCs from R. ridibunda, but the mechanism(s) involved in the response were not characterized (Kaloyianni and Rasidaki, 1996). In preliminary experiments on B. marinusRBCs there was a slower, but significant, increase in the isoproterenol-stimulated Na influx at high P_O2and there also appeared to be a prolonged lag phase in the Na influx in shrunken RBCs (data not shown). The reason for the variable lag time between experiments is currently not understood.

The magnitude of the β-adrenergic response of B. marinus RBCs is comparable to that measured in the initial 5 min after stimulation in some teleosts such as chub (36±6 mmol Na l^-1 RBCs h^-1; N=3) and striped bass (40±8 mmol Na l^-1 RBCs h^-1; N=4), but much higher values of more than 200 mmol Na l^-1 RBCs h^-1 have been measured in other teleost RBCs(Berenbrink et al., 2005). However, in teleosts, maximal activation of the βNHE is observed within 1–2 min followed by a rapid desensitization(Motais et al., 1992) (P.K. and M.B., personal observations), whereas in B. marinus the Na influx remained maximal for up to 30 min and was still elevated significantly above basal level for up to 45 min (Fig. 5A).

Tufts et al. found elevated plasma catecholamine levels after forced activity in B. marinus (Tufts et al., 1987a), and it is tempting to speculate that, as in teleosts(Nikinmaa, 1992), the physiological role of βNHE activation in deoxygenated RBCs is to safeguard intracellular pH and oxygen binding under conditions of general acidosis. Tufts et al., however, found no evidence of in vivoβNHE activation as cell volume remained unaltered after 30 min of vigorous exercise (Tufts et al.,1987a).

In rainbow trout RBCs, the βNHE is partially inhibited at high P_O2 values(Motais et al., 1987; Nikinmaa et al., 2003). We did not find βNHE activation at 20.5 kPa P_O2for up to 45 min in B. marinus RBCs(Fig. 5B). Tufts et al. were similarly unable to show in vitro β-adrenergic changes in intracellular pH and water content of air-equilibrated B. marinusRBCs kept under 5% CO₂ (extracellular pH of ∼7.5)(Tufts et al., 1987a) or of air-equilibrated A. tridactylum RBCs kept under 4 or 8%CO₂ (extracellular pH of ∼7.65 or 7.44, respectively)(Tufts et al., 1987b). Air-equilibrated RBCs from R. ridibunda and R. pipiens,nevertheless, increase cAMP levels and swell when stimulated by isoproterenol,but these responses have only been demonstrated under non-physiological conditions in the presence of the phosphodiesterase inhibitor EDTA(Rudolph and Greengard, 1980; Kaloyianni and Rasidaki,1996). In air-equilibrated RBCs, therefore, it seems thatβNHE stimulation is lacking within amphibians.

In conclusion, this study demonstrates a pronounced P_O2 sensitivity of the shrinkage-induced RVI response and the underlying Na influxes mediated through the NHE in RBCs from the terrestrial anuran Bufo marinus. Deoxygenation revealsβ-adrenergic activation of the NHE mediated Na influx, which has been missed in earlier studies on air-equilibrated amphibian RBCs. The regulation of the RBC βNHE seems to differ from that in teleosts, where rapid desensitization of the transporter may be the rule, whereas the activation was prolonged for up to 45 min in B. marinus. The RVI response in B. marinus RBCs may protect the HbO₂-affinity in vivo by reducing intracellular concentrations of Hb and organic phosphates and it may thus contribute to the high tolerance to elevated plasma osmolarity following dehydration in this species (Shoemaker,1964; Shoemaker,1965). A description of the relationship between RBC volume and HbO₂-affinity in amphibians with different tolerance to dehydration would help understand the role of RVI.

List of abbreviations and symbols

 
• βNHE

  β-adrenergically stimulated Na/H exchanger

 
• cAMP

  cyclic AMP

 
• DIDS

  4,4′-diisothiocyanatostilbene-2,2′-disulphonic acid

 
• EDTA

  ethylenediaminetetraacetic acid

 
• Hb

  haemoglobin

 
• Hct

  haematocrit

 
• Hepes

  4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

 
• KCC

  K/Cl co-transporter

 
• MCHC

  mean cellular haemoglobin concentration

 
• NHE

  Na/H exchanger

 
• NKCC

  Na/K/2Cl co-transporter

 
• P_CO2

  carbon dioxide partial pressure

 
• P_O2

  oxygen partial pressure

 
• RBC

  red blood cell

 
• RVI

  regulatory volume increase

 
• TCA

  trichloroacetic acid

The authors would like to thank Dr Ole Bækgaard Nielsen for kindly allowing us to use his facilities at Aarhus University, Denmark, for radioisotope work.