H⁺ -ATPase Activity in Crude Homogenates of Fish Gill Tissue: Inhibitor Sensitivity and Environmental and Hormonal Regulation

The Na⁺/H⁺ exchanger in fish gill epithelium was postulated to be the major pathway for Na⁺ uptake and acid excretion (Wright and Wood, 1985). The sodium concentration in fresh water, however, is usually lower than 1mmol l^-1, and the intracellular sodium concentration in the gill epithelial cell, although lowered by the Na⁺/K⁺-ATPase in the basolateral membrane, may be much higher than 1mmol l⁻¹ (intracellular sodium ion activity in frog skin epithelium was measured by Harvey and Ehrenfeld, 1986, as 6.2mmol l⁻¹ using double-barrelled ion-sensitive microelectrodes). It seems unlikely that the sodium gradient across the apical membrane could drive the Na⁺/H⁺ exchange. An alternative mechanism which will account for Na⁺ and H⁺ transport in opposite directions is an electrogenic H⁺-translocating ATPase coupled with a sodium conductive channel (Avella and Bornancin, 1989; Lin and Randall, 1991), as demonstrated in freshwater frog skin (Ehrenfeld and Garcia-Romeu, 1977; Ehrenfeld et al. 1985). This so-called proton pump will consume ATP, actively exclude hydrogen ions across the membrane and generate a negative potential inside the apical membrane, which will then drive sodium influx via the sodium channel.

The existence of the H⁺-ATPase is well documented not only in freshwater frog skin, which has the same Na⁺ uptake function as freshwater fish gills, but also in other tight epithelia, such as turtle urinary bladder (Steinmetz and Andersen, 1982) and mammalian renal collecting tubule (Gluck and Al-Awqati, 1984; Ait-Mohamed et al. 1986). N-ethymaleimide, a covalent SH-reactive compound, is an H⁺-ATPase inhibitor commonly used to identify H⁺-ATPases in different organisms (Pedersen and Carafoli, 1987).

Our previous in vivo studies (Lin and Randall, 1991) showed that proton excretion across the gill epithelium of freshwater trout was sensitive to external pH, P_CO2, vanadate (a plasma membrane ATPase inhibitor) and acetazolamide (a carbonic anhydrase inhibitor), but was not sensitive to 0.1mmol l⁻¹ amiloride, which blocked the Na⁺ influx across the gills of rainbow trout completely (Wright and Wood, 1985). All these characteristics are typical of H⁺ transport mediated by H⁺-ATPases in other tight epithelia and thus indicate the presence of a proton pump in the gill epithelium. The first objective of these studies was to measure H⁺-ATPase activity directly in crude homogenates of gill tissue.

Influx of sodium in freshwater fish is affected by different environmental factors, such as hypercapnia (Goss et al. 1992) and water Ca²⁺ levels (Avella et al. 1987). If the proton pump does provide the driving force for sodium uptake in freshwater fish, these factors could act on H⁺-ATPase activity and variations in Na⁺ influx would be the secondary outcome. It has also been suggested that the molecular target of the mineralocorticoid hormone aldosterone on urine acidification is H⁺-ATPase in the kidney collecting tubule of mammals (Mujais, 1987; Garg and Narang, 1988; Khadouri et al. 1989). The second objective of these studies, therefore, was to examine the effects of the environment and hormones on H⁺-ATPase activity in fish gills.

Rainbow trout, Oncorhynchus mykiss (Walbaum), weighing 300–500g were kept in aerated, dechlorinated Vancouver tap water ([Na⁺], 0.89mmol l^-1; [Ca²⁺], 0.03mmol l^-1; [Cl^-], 0.92mmol l^-1) at 10–15°C ambient temperature. Some animals were acclimated to sea water (34–38‰) for 8–10 weeks. All animals were fed commercial trout pellets twice a week; feeding was terminated at least 1 day before usage or treatment.

A crude homogenate of rainbow trout gill tissue was prepared using a method modified from Zaugg (1982). Fish were killed with a blow to the head. The gills were perfused through the heart with heparinized saline in order to clear red blood cells. Gill filaments (approximately 1g wet mass) were trimmed from supporting arches and immersed in 2ml of a cool homogenate medium I containing 300mmol l⁻¹ sucrose, 2mmol l⁻¹ EGTA, 1 mmol l⁻¹ dithiothreitol and 100mmol l⁻¹ Tris–HCl at pH7.3. Tissue was then homogenized with a Kontes micro ultrasonic cell disrupter for 20 strokes, 2ml of distilled water was added to the homogenates and another 20 strokes were employed to ensure that all filaments were disintegrated. The diluted homogenates were centrifuged for 7min in a Janetzli laboratory table centrifuge (model T32c) at about 4000revsmin⁻¹ (2000 RCF). Supernatant solutions were discarded and pellets suspended in 1ml of homogenate medium II (homogenate medium I containing 6% Chaps, a zwitterionic detergent) were homogenized twice for 20 strokes each. The resulting homogenates were centrifuged as before and supernatant solutions were removed for ATPase assay. We found that the ATPase activity in the supernatant was stable for at least 1 month when stored in a freezer at -80°C. The protein concentration of the supernatant was around 6–8mgml^-1.

In experiments with the inhibitors DCCD, DES and PCMBS, stock solutions were made by dissolving the drugs in 100% ethanol and then added to the reaction mixture to the required concentration. Stock solution of bafilomycin (purchased from Dr Altendorf, Fachbereich Biologie/Chemie, U. Osnabruck, FRG) were prepared in dimethylsulphoxide (Bowman et al. 1988). Control samples containing the proper amount of solvent were assayed simultaneously. KNO₃, acetazolamide and sodium vanadate are water-soluble and assays were performed in the same way as with NEM.

Four kinds of external media were prepared by dissolving NaCl and/or CaCl₂ into dechlorinated Vancouver tap water: 100mmol l⁻¹ NaCl; 100mmol l⁻¹ NaCl plus 1 mmol l⁻¹ CaCl₂; 1mmol l⁻¹ CaCl₂; and 10mmol l⁻¹ CaCl₂. Fish were placed into 100 l opaque fibreglass tanks (density less than 25 gl⁻¹) filled with different external media for an acclimation period of 10–14 days. The tanks were well-aerated and temperature was maintained at ambient levels with cooling cords. The external medium was changed daily to prevent ammonia accumulation.

Plasma cortisol levels were elevated in both freshwater-and seawater-adapted trout by means of implanted Alzet mini osmotic pumps containing cortisol (Reid and Perry, 1991). Mini osmotic pumps were loaded with cortisol (hydrocortisone 21-hemisuccinate, Sigma) or saline (in sham treatments) and surgically implanted into the peritoneal cavity of anaesthetized fish. The nominal calculated plasma cortisol concentration was 200ngml⁻¹, and fish were sampled for gill tissue and blood (by caudal puncture) 7 days after the implantation. Plasma cortisol level was measured using a Gammacoat [¹²⁵I]cortisol radioimmunoassay kit (Incstar Corp.).

Rainbow trout were placed into a 100l opaque fibreglass tank supplied with flowing aerated dechlorinated Vancouver tap water (pH5.8–6.2) and allowed to acclimate for 24h. Control fish were sampled immediately before the 48h hypercapnia treatment (2% CO₂ in air, mixed using a Wösthoff gas-mixing pump, water pH5.0–5.5) and experimental fish were sampled at 6, 24 and 48h of hypercapnia. They were then allowed to recover for 24h and samples were taken at 6 and 24h of recovery.

Rainbow trout were fitted with dorsal aortic catheters under MS-222 anaesthesia (1:10000 in NaHCO₃-buffered fresh water). Fish were then allowed to recover for 24h in a sectioned Plexiglas box. 2mlkg⁻¹ bodymass of saline (control group) or 1moll⁻¹ NH₄Cl in saline was injected daily for 2 days into the dorsal aorta of the fish over a period of approximately 5min. 48h after the first injection, blood samples were taken for pH measurement using a microcapillary pH electrode (Radiometer G279/G2) coupled to a PHM84 pH meter, and fish were killed for gill tissue sampling.

All ATPase assays were performed in triplicate and data are presented as mean ± standard error. Student’s two-tailed t-test was used to test for significant (P<0.05) differences.

A substantial amount of the ATPase activity in the crude homogenates of gill tissue, in the presence of azide, ouabain and EGTA, was sensitive to NEM. Fig. 1A shows the ATPase activity of gill tissue in response to various concentration of NEM. NEM causes a dose-dependent inhibition of ATPase activity. Maximal inhibition is observed in 1 mmol l⁻¹ NEM, which accounted for more than 70% of the total ATPase activity (Table 1). The difference between ATPase activity with and without 1mmol l⁻¹ NEM is referred to as the NEM-sensitive ATPase activity; at 15°C, it was determined to be about 1.5±0.09 μmolmg⁻¹ protein h⁻¹ for freshwater-adapted trout (Fig. 2).

Vanadate (0.1mmol l⁻¹) suppressed 60% of the ATPase activity in gill tissue (Table 1). When 1mmol l⁻¹ NEM and 0.1mmol l⁻¹ vanadate were applied together, the percentage of ATPase affected was only increased slightly. The combination of 1 mmol l⁻¹ of NEM and 0.1mmol l⁻¹ vanadate suppressed 80% of the ATPase in crude homogenates, which indicates that, of the 60% of the ATPase that was sensitive to vanadate, 50% is from the NEM-sensitive ATPase.

The effect of DCCD was also examined. A maximum of 52% of the total ATPase activity was suppressed by 1mmol l⁻¹ of DCCD (Table 1). A similar dose–response curve to that produced by NEM was observed with DCCD (Fig. 1B). The sensitivity profiles of fish gill ATPase towards both NEM and DCCD are similar to those found for the inhibition of H+-ATPase of rat kidney (Ait-Mohamed et al. 1986).

DES and PCMBS had maximal inhibitory effects at a much lower concentration and accounted for inhibition of 63% and 45% of the total ATPase activity, respectively (Table 1). Bafilomycin, a very specific and potent inhibitor of vacuolar H⁺-ATPase (Bowman et al. 1988), significantly inhibited ATPase activity of fish gills only at concentration above 25 μmoll⁻¹ (Table 1), while vacuolar H⁺-ATPases are completely blocked at concentrations as low as 0.1 μmoll⁻¹ (Bowman et al. 1988). Potassium nitrate, another inhibitor used to distinguish vacuolar H⁺-ATPase from plasma membrane H⁺-ATPase (Bowman, 1983), caused a less than 30% reduction in fish gill ATPase activity (Table 1) at a concentration of 100mmol l⁻¹, a dosage that is sufficient to inhibit 80% of the vacuolar H⁺-ATPase. 0.1mmol l⁻¹ acetazolamide, however, had no effect on the ATPase activity of fish gills.

In rainbow trout acclimated to different external level of Na⁺ and/or Ca²⁺, gill NEM-sensitive ATPase activity decreased as Na⁺ acclimation level increased (Fig. 2). NEM-sensitive ATPase activity was significantly lower in fish acclimated to 100mmol l⁻¹ Na⁺ (with or without Ca²⁺) than in control fish. Seawater-adapted rainbow trout had only one-third of the NEM-sensitive ATPase activity of their freshwater-adapted counterparts. The addition of Ca²⁺ to high-Na⁺ water made no difference to NEM-sensitive ATPase activity. Increasing the Ca²⁺ level in low-Na⁺ media, however, had a marked stimulatory effect and resulted in a twofold increase in NEM-sensitive ATPase activity in fish gills (Fig. 2).

Chronic cortisol infusion into freshwater rainbow trout caused a 170% increase in plasma cortisol level (Fig. 3B) and a 30% increase in NEM-sensitive ATPase activity in gill tissue (Fig. 3A). Seawater-adapted animals, in contrast, showed no increase in NEM-sensitive ATPase activity in response to a similar cortisol treatment, although their plasma cortisol level increased fourfold.

Hypercapnia treatment induced an immediate increase in the NEM-sensitive ATPase activity, which stabilized at a level 70% higher than that during normocapnia (Fig. 4). The elevated NEM-sensitive ATPase activity returned to control levels after 24h of recovery.

A blood acidosis was associated with NH₄Cl injection (Table 2). NEM-sensitive ATPase activity, however, was not altered by NH₄Cl injection. Daily injections of NH₄Cl were given to two fish for 5 days and no change in NEM-sensitive ATPase activity was observed (data not shown).

These studies demonstrate for the first time the existence of an NEM-sensitive ATPase in crude homogenates of fish gill tissue. H⁺-ATPase has been reported to be either bound to membranes or packaged in cytoplasmic vesicles (Arruda et al. 1990). The crude homogenates prepared with the current method contain mainly the membrane fraction of gill cells, and the NEM-sensitive ATPase we detected in this study was released by protein solubilizer from the membrane fraction. Soluble cell material, mitochondria, cytoplasmic vesicles and other organelles would have been discarded in the supernatant of the first centrifugation (2000 g) because much higher relative centrifugal forces are required to spin down this material. Little ATPase activity was found in the discarded supernatant, but protein solubilizer was never applied. If there was H⁺-ATPase packaged in cytoplasmic vesicles, perhaps protein solubilizer was required to release it for subsequent detection.

NEM is an alkylating agent that is relatively selective for sulphydryl groups (SH-). It inhibits vacuolar H⁺-ATPase in an ATP-protectable manner at concentrations under 10 μmol l⁻¹ (Forgac, 1989; Pedersen and Carafoli, 1987). Phosphorylated ATPases (including Na⁺/K⁺-ATPase, Ca²⁺-ATPase and plasma membrane H⁺-ATPase) are sensitive to higher concentrations (100 μmol l⁻¹ to 1mmol l⁻¹) of NEM (Forgac, 1989). Since the assay was carried out in the presence of EGTA, a Ca²⁺ chelator that should abolish Ca²⁺-ATPase activity, azide, a mitochondrial H⁺-ATPase inhibitor, and ouabain, a Na⁺/K⁺-ATPase inhibitor, the contribution of unrelated ATPase activity was minimized. Thus, the ATPase activity in the crude homogenate of gill tissue that was sensitive to 1mmol l⁻¹ NEM probably originated from plasma membrane H⁺-ATPase. 30% of the ATPase activity was NEM-insensitive and is of unknown origin. Unidentified NEM-insensitive ATPase has also been detected in mammalian kidney (Ait-Mohamed et al. 1986; Garg and Narang, 1988). Bornancin et al. (1980) presented evidence of Cl⁻/HCO₃⁻-ATPase in gill plasma membrane of rainbow trout. This might account for the NEM-insensitive ATPase activity in the gill tissue crude homogenates.

The gill ATPase sensitive to NEM was also sensitive to vanadate. Orthovanadate, VO₄^3-, acting as a phosphate transition analogue, blocks the formation of phosphorylated intermediates in all P-type ATPases. 0.1mmol l⁻¹ vanadate has been reported to inhibit branchial proton excretion in freshwater trout (Lin and Randall, 1991). Urinary acidification by turtle bladder (Arruda et al. 1981) and proton transport across freshwater frog skin (Ehrenfeld et al. 1985), both mediated by H⁺-ATPase, are also vanadate-sensitive. However, vanadate fails to inhibit NEM-sensitive ATPase activity and proton transport in mammalian kidney (Gluck and Al-Awqati, 1984; Ait-Mohamed et al. 1986), which were believed to be mediated by vacuolar H⁺-ATPase (Forgac, 1989). DCCD inhibits mitochondrial, vacuolar and plasma membrane H⁺-ATPase by binding to the c subunit of the hydrophobic channel portion (Pedersen and Carafoli, 1987). Mitochondrial H⁺-ATPase has the highest sensitivity to DCCD (0.1–0.5 μmol l⁻¹), followed by vacuolar H⁺-ATPase (1–10 μmol l⁻¹) and then plasma membrane H⁺-ATPase (10–100 μmol l⁻¹). The dose–response curves of the ATPase activity in gill tissue for NEM and DCCD are very similar to those reported for rat kidney H⁺-ATPase (Ait-Mohamed et al. 1986). DES and PCMBS were also reported to inactivate the F_o moiety of H⁺-ATPase (Pedersen and Carafoli, 1987). Proton transport mediated by H⁺-ATPase in bovine kidney medulla was completely blocked by 10 μmol l⁻¹ PCMBS (Gluck and Al-Awqati, 1984), which partially inhibited H⁺-ATPase in fish gills.

Bafilomycins are macrolide antibiotics that have a specific and potent inhibitory effect on vacuolar H⁺-ATPase (Bowman et al. 1988). Mitochondrial H⁺-ATPase is resistant to up to 1mmol l⁻¹ bafilomycin, whereas vacuolar H⁺-ATPase is completely inhibited by less than 0.1 μmol l⁻¹ of the antibiotic. Phosphorylated ATPase exhibits intermediate sensitivities, with I₅₀ values between 10 and 100 μmol l⁻¹. The sensitivity of the ATPase in gill tissue to bafilomycin is within this intermediate range.

Vacuolar H⁺-ATPase also demonstrates a unique sensitivity to KNO₃ with an I₅₀ value of about 50mmol l⁻¹. The resistance of fish gill ATPase to nitrate indicates that the H⁺-ATPase we measured is not of the vacuolar type. Taken together, these pharmacological properties indicate that H⁺-ATPase in fish gills is of the plasma membrane type and not the vacuolar type.

Acetazolamide has been demonstrated to inhibit luminal acidification in turtle bladder by stimulating the endocytosis of apical membrane (Dixon et al. 1988; Graber et al. 1989). The inhibition appeared to be independent of cell pH, which ruled out the possibility of a secondary effect due to inhibition of carbonic anhydrase. The inhibitory effect of acetazolamide on in vivo proton excretion (Lin and Randall, 1991) cannot be reproduced in the in vitro ATPase assay, indicating that acetazolamide has no direct effect on H⁺-ATPase itself.

The observation that fish acclimated to water with low Na⁺ level have higher H⁺-ATPase activity than fish acclimated to water with a high Na+ level indicates that the functional significance of the H⁺-ATPase in fish gills is to generate an electrochemical gradient for Na⁺ uptake from a dilute medium. Ehrenfeld et al. (1985) demonstrated that sodium absorption across freshwater frog skin was mediated by an active proton pump indirectly coupled with a sodium channel, instead of Na⁺/H⁺ exchange. The proton pump in frog skin was inhibited by DCCD and vanadate. Since the gill epithelium in freshwater fish is a tight epithelium similar to that of freshwater frog skin, it is reasonable to suppose that they have the same mechanism to solve the same osmoregulatory problem. When external Na⁺ level is high, H⁺-ATPase is down-regulated, possibly by endocytosis of membrane protein into intracellular vesicles (Schwartz and Al-Awqati, 1986). The residual H⁺-ATPase in fish gills may play a role in acid–base regulation, although it is not required for sodium absorption or sodium excretion in fish that live in high-sodium environments.

The stimulatory effect of Ca²⁺ on H⁺-ATPase could be explained using the cellular model of Wendelaar Bonga et al. (1992), who proposed the existence of an apical Ca²⁺ channel. When the external medium has a high Ca²⁺ but a low Na⁺ level, Ca²⁺ might compete with Na⁺ by entering the cell via the Ca²⁺ channel, reducing the potential gradient generated by the H⁺-ATPase and therefore Na⁺ influx. The resulting high Ca²⁺ concentration in the cell might stimulate the insertion of proton pumps from intracellular vesicles into the apical membrane in order to maintain Na⁺ influx. Exocytosis of vesicles containing proton pumps into the cell membrane is Ca²⁺-dependent in turtle bladder epithelium (Adelsberg and Al-Awqati, 1986).

Variations in Ca²⁺ level in freshwater environments have been reported to affect gill morphology and sodium influx in rainbow trout (Avella et al. 1987). Na⁺ influx increased 2.5-fold in fish acclimated to fresh water+10mmol l⁻¹ CaCl₂ for 15 days and new globular chloride cells appeared and proliferated in the secondary lamellae. Fish acclimated to fresh water+5mmol l⁻¹ CaCl₂ for 5 days showed no change in gill morphology or sodium flux, perhaps because 5 days was too short for morphological modification.

We have also investigated the effects of chronic infusion of cortisol on H⁺-ATPase activity in gill tissue in freshwater and seawater rainbow trout. Aldosterone treatments, either long-term (7 days) or short-term, have been reported to stimulate H⁺ secretion mediated by H⁺-ATPase in the collecting duct of mammalian kidney (Garg and Narang, 1988; Mujais, 1987). The functionally parallel steroid hormone in fish is cortisol. Perry and Laurent (1989) have shown that plasma cortisol level rose transiently in fish exposed to deionized water. Daily intramuscular injections of cortisol for 10 days caused an increase in Na⁺ uptake. The 30% increase in H⁺-ATPase activity observed in freshwater trout following chronic cortisol treatment is probably responsible for this increased Na⁺ uptake. Similar cortisol treatment has no effect on seawater-acclimated rainbow trout, indicating that Na⁺ concentration is the predominant regulator of the H⁺-ATPase in fish gills.

Another possible function of the H⁺-ATPase in fish gill epithelium is acid–base regulation. We induced respiratory acidosis through hypercapnia treatment and observed an increase in H⁺-ATPase activity in fish gills. Similar hypercapnia treatment in freshwater catfish was reported to cause a marked increase in Na⁺ influx, which might be correlated with the increased H⁺-ATPase activity. The elevated H⁺-ATPase activity could be induced by the CO₂-mediated insertion of proton pumps via exocytosis, as demonstrated in turtle bladder epithelium (Cannon et al. 1985; Arruda et al. 1990). High CO₂ levels reduced the intracellular pH of the proton-secreting cells, which increased the intracellular Ca²⁺ concentration and, in turn, stimulated the fusion of cytoplasmic vesicles containing proton pumps into the apical membrane. This would then correct the intracellular acidosis.

A chronic metabolic acidosis was induced in the fish by NH₄Cl injection; it resulted in no significant change in H⁺-ATPase activity. This indicates that high plasma hydrogen ion levels alone do not stimulate H⁺-ATPase activity in gill tissue. We conclude that elevated CO₂ levels increase H⁺-ATPase activity via a depression of epithelial pH. A metabolic acidosis, however, will only activate H⁺-ATPase activity if the acidosis is transferred into the gill intracellular compartment. If we follow this argument, then presumably NH₄Cl infusion has little or no effect on epithelium pH because H⁺-ATPase activity was unchanged. Unfortunately, we were not be able to measure the effects of NH₄Cl injection on gill epithelial pH.