The gills of most teleost fishes lack plasma-accessible carbonic anhydrase (paCA) that could participate in CO2 excretion. We tested the prevailing hypothesis that paCA would interfere with red blood cell (RBC) intracellular pH regulation by β-adrenergic sodium-proton exchangers (β-NHE) that protect pH-sensitive haemoglobin–oxygen (Hb–O2) binding during an acidosis. In an open system that mimics the gills, β-NHE activity increased Hb–O2 saturation during a respiratory acidosis in the presence or absence of paCA, whereas the effect was abolished by NHE inhibition. However, in a closed system that mimics the tissue capillaries, paCA disrupted the protective effects of β-NHE activity on Hb–O2 binding. The gills are an open system, where CO2 generated by paCA can diffuse out and is not available to acidifying the RBCs. Therefore, branchial paCA in teleosts may not interfere with RBC pH regulation by β-NHEs, and other explanations for the evolutionary loss of the enzyme must be considered.

Red blood cells (RBCs) are the functional unit of cardiovascular gas transport in vertebrates and contain the principal O2-carrying protein, haemoglobin (Hb). In teleost fishes, Hb–O2 binding is uniquely pH sensitive, in that proton (H+) binding decreases not only Hb–O2 affinity by the Bohr effect (Bohr et al., 1904; Jensen et al., 1998), but also Hb–O2 carrying capacity by the Root effect (Brittain, 2005; Root, 1931). Teleosts take advantage of this pH sensitivity to greatly enhance O2 unloading to their eyes and the swim bladder, by acidifying the blood locally in specialized counter-current exchangers (Pelster, 1997; Wittenberg and Wittenberg, 1962).

However, during a systemic acidosis, which in fish may occur during exercise or hypoxia, teleost Hbs may lose over 50% of their O2 carrying capacity (Berenbrink et al., 2011). To prevent severe hypoxemia, the RBCs of many teleosts species can regulate their intracellular pH (pHi) against reductions in extracellular pH (pHe) (Berenbrink et al., 2005). Under stressful conditions, adrenaline and noradrenaline are released into the blood (Randall and Perry, 1992) and activate beta-adrenergic sodium-proton exchangers (β-NHE) on the RBC membrane that extrude H+ to protect pHi and Hb–O2 binding (Nikinmaa, 1992). However, the extruded H+ can re-equilibrate across the RBC membrane via the Jacobs–Stewart cycle (Jacobs and Stewart, 1942). In the plasma, these H+ react with bicarbonate (HCO3) to form CO2 that diffuses back into the RBC, where it is once again hydrated to produce H+, a reaction that is catalyzed by the abundant presence of carbonic anhydrase (CA) within RBCs (Maren, 1967). An adrenergic increase in RBC pHi requires that β-NHE H+ extrusion outpaces the re-acidification through the Jacobs–Stewart cycle, which is rate limited by the uncatalyzed formation of CO2 in the plasma (Motais et al., 1989a). Therefore, the absence of CA activity in the plasma of teleosts is thought to be essential for RBC pHi regulation and, in fact, a potent plasma CA inhibitor in most vertebrate species seems to ensure that this requirement is met (Henry et al., 1997).

However, plasma accessible carbonic anhydrase (paCA) is present at the tissue capillaries of most vertebrates, where it serves important functions facilitating CO2, ammonia and lactate excretion from cellular metabolism into the blood (for review, see Harter and Brauner, 2017). Within the capillaries, a CA4 isoform is anchored to the luminal walls of the endothelium (Damsgaard et al., 2020), which is largely unaffected by endogenous CA inhibitors in the plasma (Gervais and Tufts, 1998; Heming et al., 1993; Roush and Fierke, 1992). Thus, when RBCs reach the capillaries, the formation of CO2 in the plasma is catalyzed and the H+ influx into the RBCs exceeds the rate of H+ extrusion, effectively short-circuiting β-NHE activity (Rummer and Brauner, 2011). The result is a sudden decrease in Hb–O2 affinity that enhances the unloading of O2 to the tissues. Once the RBCs leave the capillaries and the site of paCA, β-NHE activity recovers pHi during venous transit to safeguard the renewed uptake of O2 at the gills (Harter et al., 2018a). This mechanism of β-NHE short-circuiting has been shown to increase the partial pressure of O2 (PO2) in the swimming muscles of rainbow trout (Rummer et al., 2013) and is critical to sustain maximal exercise performance in Atlantic salmon (Harter et al., 2019). Therefore, interactions between RBC pHi regulation and paCA may be a fundamental aspect of the teleost O2 transport system (Harter and Brauner, 2021).

Most vertebrates also have paCA at their gas exchange surfaces (Stabenau and Heming, 2003), where it contributes to CO2 excretion by ∼10% in humans (Henry and Swenson, 2000), ∼50% in an elasmobranch (Squalus acanthias) (Gilmour et al., 2001) and 100% in an Antarctic icefish (Champsocephalus gunnari) that has lost RBCs from the circulation (Harter et al., 2018b). The presence of paCA at the gills of the Pacific hagfish (Epatretus stoutii) (Esbaugh et al., 2009) and several elasmobranch species (McMillan et al., 2019) may indicate that this is the ancestral condition. But despite the potential benefits for CO2 excretion, most teleost fishes seem to lack paCA at their gills (Harter and Brauner, 2017). It has been proposed that the evolution of highly pH-sensitive Hbs and the need to safeguard branchial O2 uptake by RBC β-NHE activity created a selective pressure for the loss of branchial paCA (Randall et al., 2014). However, whether branchial paCA actually interferes with RBC pHi regulation has not been tested experimentally, and fundamental differences between the gill and capillary microenvironments have not been addressed.

At the capillaries, metabolically produced CO2 from the tissues diffuses into the blood. This inwardly directed diffusion gradient would prevent any CO2 produced by paCA from diffusing out, making it available to acidify the RBCs. Therefore, from the perspective of the RBCs, the tissue capillaries are a functionally closed system with regards to CO2 (but not O2 that diffuses from the blood to the mitochondria). In contrast, the fish gills are an open system, where most CO2 produced by paCA could simply diffuse into the water and would not be available to acidify the RBCs. Therefore, we hypothesized that the presence of paCA in an open system, such as the fish gill, is not an impediment for effective pHi regulation by RBC β-NHE activity.

Animals and husbandry

Rainbow trout [Oncorhynchus mykiss (Walbaum 1972)] of both sexes were obtained from Thomas Fish Company (Anderson, CA, USA) and raised in a recirculating freshwater system at 15°C. Juvenile trout (>20 g body mass) were acclimated to seawater by slowly increasing salinity to 35 ppt over the course of 2 weeks. Seawater-acclimated trout were held in 3.5–10 m3 tanks in the Hubbs Hall aquatics facility (Scripps Institution of Oceanography), supplied with flow-through seawater at >95% dissolved O2. Animals were held for 18 months at a 12 h:12 h light:dark cycle and at a constant water temperature of 12°C, to avoid seasonal effects on RBC β-NHE function (Nikinmaa and Jensen, 1986). Fish were fed twice a week with commercial trout pellets (Skretting Classic Trout, 8 mm sinking) and water quality was monitored daily. Feeding was suspended 48 h before experiments. Fish husbandry and experimental procedures were performed in compliance with the Institutional Animal Care and Use Committee (IACUC) guidelines and were approved by the Animal Care program at the University of California, San Diego (protocol no. S10320).

Blood sampling

Rainbow trout (705.6±82.0 g) were anaesthetized in seawater containing 70 mg l−1 benzocaine (Acros 150785000, Waltham, MA, USA; stock solution made up in ethanol). Once unresponsive, a heparinized syringe was used to draw ∼3 ml of blood from the caudal vein and fish were quickly recovered in their tank for future experiments. Blood was centrifuged (500 g for 3 min) and the plasma was removed and stored at 4°C overnight. The remaining RBC pellet was rinsed three times with filter sterilized Cortland's saline (in mmol l−1: 151 NaCl, 5.1 KCl, 1.6 CaCl, 0.9 MgSO4, 6 NaHCO3, 3 NaH2PO4, 5.6 glucose; at pH 7.8; Wolf, 1963), discarding the buffy coat each time to remove white blood cells and platelets. Finally, the RBCs were resuspended in 25 ml of Cortland's saline and stored with a large headspace on a tilt-shaker overnight at 17°C to reverse the adrenergic effects of sampling stress.

Open-system experiments

The potential effects of extracellular CA on β-NHE activity at the fish gills were studied using an in vitro spectrophotometric approach. RBCs were rinsed in Cortland's saline and suspended in native plasma at 5% haematocrit (Hct), which yields an acceptable optical density for spectrophotometric measurements (Harter et al., 2021). The use of native plasma from rainbow trout was critical, as it: (i) minimizes RBC lysis during tonometry, owing to its higher viscosity compared with saline (Jensen et al., 2009); and (ii) contains a plasma CA inhibitor that effectively neutralises any extracellular CA activity from RBC lysis and is ineffective against CA isoforms from other species, including bovine CA2 (Henry et al., 1997). Blood aliquots of 300 µl were treated with either: (i) a carrier control of 0.25% dimethyl sulfoxide (DMSO, VWR BDH 1115; Radnor); (ii) 10 µmol l−1 isoproterenol (ISO; Sigma-Aldrich I6504, St Louis, MO, USA) a β-adrenergic agonist that induces maximal β-NHE activity (Tetens and Lykkeboe, 1988); (iii) ISO plus 10 µmol l−1 CA (ISO+CA) from bovine erythrocytes (Sigma C3934≥2500 W-A units), shown to short-circuit RBC β-NHE activity in vitro (Rummer and Brauner, 2011); or (iv) ISO plus 1 mmol l−1 amiloride (ISO+Am; Sigma A7410) an NHE inhibitor (Mahe et al., 1985). Treated blood aliquots were equilibrated in tonometers (open gas equilibration systems) to a humidified gas mixture (21 kPa PO2, 0.3 kPa PCO2 at 12°C; gas mixing system GMS, Loligo Systems, Viborg, Denmark) that matches arterial PCO2 in normoxic rainbow trout (Brauner et al., 2000). After 1 h, a subsample of 2 µl was taken from the tonometers and Hb–O2 binding was studied spectrophotometrically with a Blood Oxygen Binding System (BOBS, Loligo Systems). The remainder of the blood was used to measure blood parameters as described below. All trials were run at 12°C and 21 kPa PO2, while PCO2 was increased in six steps from 0.3 to 6.0 kPa with 2 min of equilibration in between. During these trials, the absorbance of the blood was recorded once per second at 430 nm, where changes in absorbance correspond to changes in Hb–O2 saturation, and at the isosbestic reference wavelength of 390 nm, where absorbance is independent of oxygenation. To calculate Hb–O2 saturation from raw absorbance data, a calibration was carried out at the beginning and end of each trial, by equilibrating blood to gas tensions that yielded full Hb oxygenation (99.7 kPa PO2, 0.3 kPa PCO2) and deoxygenation (0 kPa PO2, 0.3 kPa PCO2), which was confirmed by examining the absorption spectra.

Closed-system experiments

The effects of extracellular CA on β-NHE activity in the tissue capillaries were verified in vitro using a previously described closed-system preparation (Rummer and Brauner, 2011). Blood sampling and overnight storage were as described above. RBC pellets were rinsed in Cortland's saline and resuspended in native plasma at 25% Hct. The blood was loaded into a tonometer and equilibrated to 7.5 kPa PO2 and 0.3 kPa PCO2 for 1 h at 12°C. This gas tension yields ∼75% Hb–O2 saturation as calculated using previously generated oxygen equilibrium curves. These conditions were chosen so that changes in Hb–O2 saturation could be observed in both positive and negative directions during trials (Harter et al., 2018a; Rummer and Brauner, 2011; Shu et al., 2017).

After equilibration, approximately 1 ml of well-mixed blood was loaded into a Micro Respiration Chamber (Unisense, Aarhaus, Denmark), continuously mixed with a stir bar at 350 rpm (Micro Respiration Rack, Unisense) and kept in a temperature-controlled water bath at 12°C. An O2 microsensor (OX-200, Unisense) was inserted into the closed vial through a pore and PO2 was recorded once per second. After loading, the PO2 reading was allowed to stabilize before sequential (every 5 min) injections of: (i) 15 µl of 200 mmol l−1 HCl to decrease blood pH by 0.3 units, according to Wood et al. (1982); (ii) 10 µmol l−1 ISO to stimulate β-NHE activity; and (iii) 1 µmol l−1 CA to short-circuit β-NHE activity. A 10-fold lower concentration of CA was used in the closed system after preliminary trials showed significant effects; thus, in the open system, CA concentration exceeded that required for successful β-NHE short-circuiting. To validate that a respiratory acidosis in the closed system has the same effect as the metabolic acidosis from HCl injections, an additional experiment was run by injecting 20 µl Cortand's saline saturated with CO2 (i.e. at 100 kPa or 1,000,000 µatm PCO2).

Blood analysis

Blood parameters were measured after equilibration in tonometers. Hct was measured in triplicate using micro-capillary tubes (Drummond Scientific Company, Broomall, PA, USA) centrifuged at 10,000 g for 3 min in a Model MB micro-capillary centrifuge (International Equipment Company, Needham Heights, MA, USA). [Hb] was measured on 10 µl of well-mixed blood in 1 ml aliquots of Drabkin's reagent (Sigma D5941). Absorbance of the sample was measured in a UV-VIS spectrophotometer (Shimadsu UV-1820, Columbia, MD, USA) at 540 nm using an absorption coefficient of 10.99 (van Kampen and Zijlstra, 1983). Mean corpuscular Hb concentration (MCHC) was calculated by dividing [Hb] in mmol l−1 by Hct (as a fraction) multiplied by 100. Extracellular pH (pHe) was measured with a pH microelectrode at 12°C (Fisher Accumet 13-620-850, Hampton, NY, USA; with Denver Instruments UB-10 meter, Bohemia, NY, USA). The remaining blood was centrifuged to separate RBCs and plasma (500 g, 3 min). Plasma total carbon dioxide content (TCO2) was measured in triplicate using a Corning 965 (Corning Instruments, Corning, NY, USA), and HCO3 was calculated using the Henderson–Hasselbalch equation and published pKa and solubility values for CO2 (Boutilier et al., 1984). The RBC pellet was frozen and thawed three times in liquid nitrogen (Zeidler and Kim, 1977) and pHi was measured in the lysate as described above. For RBC fixations, the above tonometry trial was repeated with cells suspended in saline, as plasma proteins can interfere with the quality of antibody staining. RBC fixation and immunostaining were as previously described (Harter et al., 2021), with a custom-made polyclonal antibody raised against rainbow trout β-NHE and a monoclonal mouse anti-Tetrahymena α-tubulin antibody (DSHB12G10).

Data analysis and statistics

All data were analysed in RStudio v.1.4.1106 using R v.4.0.4 (www.rstudio.org) and the ggplot2 package to generate figures (Wickham, 2009). Raw data from the open-system experiment were generated by dividing the absorbance at 430 nm by the isosbestic absorbance recorded at 390 nm. The last 10 absorbance ratios from each PCO2 equilibration step were averaged and used for statistical analysis. Absolute values of Hb–O2 saturation were calculated based on the calibration values for fully oxygenated and deoxygenated Hb and applying a linear correction for drift (Harter et al., 2021). The decrease in Hb–O2 saturation during the respiratory acidosis was analysed by fitting a nonlinear Hill model to the data (tested against Michaelis–Menten and exponential models and selected based on the lowest Akaike's information criterion; AIC). The two parameters estimated by the Hill model represent the PCO2 at which Hb–O2 saturation decreases by 50% of the maximal change (EC50PCO2) and the maximal change in Hb–O2 saturation (max. ΔHb–O2 sat.). The effects of drug additions on EC50PCO2 and max. ΔHb–O2 sat. were tested with a linear mixed-effects model using individual fish as a random effect (repeated measures). If a significant main effect of treatment (drug additions) was detected, post hoc analysis was performed using pairwise t-tests with a Benjamini–Hochberg correction of the significance level. The same statistical methods were applied to analyse the blood parameter data (Hct, [Hb], MCHC, pHe, pHi and plasma HCO3). The raw data from the closed-system experiment were analysed by calculating the change in PO2 (ΔPO2) after drug treatments, by averaging 10 data points before and after additions of a drug. Whether treatments (HCl, ISO or CA) had an effect on closed-system PO2 was tested with paired t-tests. Parametric assumptions were confirmed with the Shapiro–Wilk test of normality and Levene's test for homogeneity of variance, and analyses were performed on transformed data if these assumptions were violated. All data are reported as means±s.e.m., for N=6 fish (unless stated otherwise).

Open-system experiments

When RBCs were exposed to a respiratory acidosis, Hb–O2 saturation decreased significantly in all treatments, as expected owing to an intracellular acidification and the Bohr–Root effect of Hb (Fig. 1A). However, the PCO2 at which Hb–O2 saturation decreased by 50% of the maximal change (EC50PCO2; Fig. 1B) was significantly affected by the drug treatments (P<0.001). Adrenergically stimulated RBCs in the ISO treatment had significantly higher EC50PCO2 (1.56±0.13 kPa) compared with DMSO controls (1.14±0.04 kPa). In contrast, stimulated RBCs treated with the NHE inhibitor amiloride (ISO+Am) had a significantly lower EC50PCO2 (0.96±0.04 kPa) compared with the ISO treatment, and were not different from DMSO controls. Therefore, the increase in Hb–O2 saturation after adrenergic stimulation of RBCs can be attributed to β-NHE activity, likely through its well-described effects on RBC pHi (Nikinmaa, 1992); Na+ channels that are also affected by amiloride are not involved in the RBC β-NHE response (Reid and Perry, 1994). More importantly, however, the addition of extracellular CA to ISO-treated RBCs (ISO+CA) significantly increased EC50PCO2 (1.55±0.07 kPa) compared with DMSO and ISO+Am controls and was not different from RBCs treated with ISO alone. This critical finding supports our hypothesis that the presence of extracellular CA in an open system, such as the fish gill, does not impair pHi regulation by RBC β-NHEs. In fact, both ISO- and ISO+CA-treated RBCs maintained Hb–O2 saturation ∼10% higher compared with paired DMSO controls from the same animal (Fig. 1C). The negative ΔHb–O2 saturation in ISO+AM-treated cells may support the presence of other non-adrenergic NHE isoforms that have been speculated to contribute to RBC pHi protection under routine conditions (Pedersen et al., 2003; Rummer and Brauner, 2011). In all treatments, the protective effects of β-NHE activity on Hb–O2 binding decreased at higher PCO2 exceeding 3 kPa, where there were no differences in ΔHb–O2 saturation between the treatments. This is in line with the finding that the maximal decrease in Hb–O2 saturation (max. ΔHb–O2 sat.) during the respiratory acidosis was not different between the treatments and was on average −47.8±1.4% (Fig. 1D), which is in line with previous Root effect measurements in rainbow trout (Berenbrink et al., 2011).

Fig. 1.

Results from the open-system experiment. (A) Changes in haemoglobin–oxygen saturation (Hb–O2 sat.; %) of rainbow trout whole blood during a respiratory acidosis (0.3–6 kPa PCO2). Blood was treated with either: (i) a carrier control (DMSO; 0.25%), (ii) the β-adrenergic agonist isoproterenol (ISO; 10 µmol l−1), (iii) ISO plus amiloride (ISO+Am; 1 mmol l−1), an inhibitor of sodium-proton exchangers (NHE), or (iv) ISO plus carbonic anhydrase (ISO+CA; 10 µmol l−1). (B) The PCO2 at which Hb–O2 saturation decreased by 50% of the maximal change (EC50PCO2; kPa). (C) Changes in Hb–O2 saturation relative to paired controls from the same animal (ΔHb–O2 sat.; %). (D) The maximal reduction in Hb–O2 saturation owing to acidification (max. ΔHb–O2 sat.; %). Superscript letters that differ indicate significant differences between treatments. Individual data points, means±s.e.m. and 95% confidence intervals in C (N=6–7).

Fig. 1.

Results from the open-system experiment. (A) Changes in haemoglobin–oxygen saturation (Hb–O2 sat.; %) of rainbow trout whole blood during a respiratory acidosis (0.3–6 kPa PCO2). Blood was treated with either: (i) a carrier control (DMSO; 0.25%), (ii) the β-adrenergic agonist isoproterenol (ISO; 10 µmol l−1), (iii) ISO plus amiloride (ISO+Am; 1 mmol l−1), an inhibitor of sodium-proton exchangers (NHE), or (iv) ISO plus carbonic anhydrase (ISO+CA; 10 µmol l−1). (B) The PCO2 at which Hb–O2 saturation decreased by 50% of the maximal change (EC50PCO2; kPa). (C) Changes in Hb–O2 saturation relative to paired controls from the same animal (ΔHb–O2 sat.; %). (D) The maximal reduction in Hb–O2 saturation owing to acidification (max. ΔHb–O2 sat.; %). Superscript letters that differ indicate significant differences between treatments. Individual data points, means±s.e.m. and 95% confidence intervals in C (N=6–7).

Our experiments showed robust effects of β-NHE activity on Hb–O2 saturation during the spectrophotometric measurements (Fig. 1A–C), dissipating initial concerns about a desensitization of the transporter over the experimental period (Garcia-Romeu et al., 1988). This is in line with previous results where adrenergically stimulated rainbow trout blood was successfully short-circuited by CA additions in vitro more than 60 min after it was rinsed to remove ISO (Shu, 2019). Also, RBCs from another teleost, white seabass (Atractoscion nobilis), showed β-NHE activity throughout the same protocol used here (Harter et al., 2021). Despite the low Hct of 5%, we observed a trend (P<0.10) for RBC swelling in the ISO and ISO+CA treatments, which is indicative of β-NHE activity (Nikinmaa, 1992), and was absent in the ISO+Am or DMSO controls (Fig. S1A). Finally, we found evidence for a translocation of β-NHE protein in adrenergically stimulated rainbow trout RBCs, from the cytoplasm into the membrane (Fig. 2), which is consistent with previous findings in white seabass (Harter et al., 2021) and work describing an increase in radiolabelled β-NHE binding sites in hypoxia-exposed RBCs (Reid and Perry, 1994). The translocation of β-NHE protein in adrenergically stimulated trout RBCs was less prominent than that observed in white seabass RBCs (Harter et al., 2021), perhaps indicating interspecific differences in this novel cellular mechanism that are worthy of investigation. The 3D reconstructions of trout RBC images provide novel insights indicating that β-NHE protein is not homogeneously distributed throughout the RBC membrane, but is most abundant along the marginal band of the cells (Fig. 2). An involvement of cytoskeletal structures, such as α-tubulin, in the intracellular translocation of RBC membrane proteins seems likely, yet remains to be substantiated in dedicated experiments.

Fig. 2.

Immunocytochemical localisation of the β-adrenergic sodium proton exchanger (β-NHE) in rainbow trout red blood cells (RBCs). Blood was incubated in tonometers with either: (A–C) a carrier control (0.25% DMSO) or (D–F) the β-adrenergic agonist isoproterenol (ISO; 10 µmol l−1) for 1 h. Fixed cells were immuno-stained with an anti-α-tubulin antibody to visualize the marginal band (green), with DAPI to visualize the cell nuclei (blue), and with an anti-β-NHE antibody (magenta) raised against the rainbow trout protein. (G) Insets show 2D images highlighting the differences in membrane staining for β-NHE protein between control and ISO-treated RBCs. Representative images for N=3 fish.

Fig. 2.

Immunocytochemical localisation of the β-adrenergic sodium proton exchanger (β-NHE) in rainbow trout red blood cells (RBCs). Blood was incubated in tonometers with either: (A–C) a carrier control (0.25% DMSO) or (D–F) the β-adrenergic agonist isoproterenol (ISO; 10 µmol l−1) for 1 h. Fixed cells were immuno-stained with an anti-α-tubulin antibody to visualize the marginal band (green), with DAPI to visualize the cell nuclei (blue), and with an anti-β-NHE antibody (magenta) raised against the rainbow trout protein. (G) Insets show 2D images highlighting the differences in membrane staining for β-NHE protein between control and ISO-treated RBCs. Representative images for N=3 fish.

In vitro studies of β-NHE function do not accurately replicate the conditions of flow, shear and residence time of the RBCs at the gills, factors that warrant some careful consideration. Our findings are generally in line with two in vivo studies that injected extracellular CA into rainbow trout exposed to hypoxia (Lessard et al., 1995) or during maximal exercise (Wood and Munger, 1994) and found no changes in arterial O2 transport that would be expected owing to β-NHE short-circuiting at the gills. However, the injections of CA in vivo also caused significant changes in blood PCO2, pHe and the release of catecholamines, which are known to affect β-NHE activity in rainbow trout (Nikinmaa, 1992). Owing to these confounding factors, it can be difficult to isolate the effects of CA on β-NHE function in vivo, highlighting the importance of mechanistic in vitro experiments. For instance, Motais et al. (1989a,b) found that additions of extracellular CA to adrenergically stimulated rainbow trout blood in tonometers, which are open systems, did not impair the protective effects of β-NHE function on RBC pHi; this is in line with our findings. However, very similar experiments by Nikinmaa et al. (1990) showed that the addition of extracellular CA abolished the H+ gradient across the RBC membrane, which indicates β-NHE short-circuiting. The discrepancy between these results has not been resolved experimentally, but may be related to the variable diffusive characteristics of tonometry systems.

Upon β-NHE activation in tonometers, H+ ions are extruded from the RBC, which increases pHi and decreases pHe. In the plasma, H+ ions react with HCO3 to form CO2, which can take one of two routes: some CO2 may diffuse out of the open system, driving the dehydration of more HCO3, while the consumption of H+ leads to a progressive recovery of pHe; and some CO2 may diffuse into the RBCs, where it is hydrated by intracellular CA to form H+ and HCO3, which are exported back into the plasma by the β-NHE and anion exchanger, respectively. This recycling of H+ through the RBC continuously fuels the β-NHE system and delays the recovery of pHe beyond the time course predicted by the uncatalyzed formation of CO2 (Motais et al., 1989a), which has also been reported in rainbow trout in vivo (Perry and Thomas, 1993). Because the Jacobs–Stewart cycle is rate-limited by the slow formation of CO2 in the plasma, the H+ influx into RBCs is slower than H+ extrusion by the β-NHE, and pHi remains protected. When extracellular CA is added to a closed system, the Jacobs–Stewart cycle accelerates sufficiently to overwhelm β-NHE H+ extrusion, leading to a decrease in RBC pHi. However, in an open system, the diffusive loss of CO2 diverts H+ away from the RBCs, which may prevent changes in RBC pHi. Although tonometers are nominally open systems, they are not particularly efficient gas exchangers, as evidenced by transient changes in blood PO2 and PCO2 upon adrenergic stimulation (Perry and Thomas, 1993). Therefore, the divergent results of Nikinmaa et al. (1990) and Motais et al. (1989a,b) could be explained by different gas transfer efficiencies between the blood and the equilibration gas, where the degree of RBC short-circuiting upon CA addition will depend on how much CO2 is lost from the system and how much is available to acidify the cells.

To avoid the diffusive limitations of tonometers, we used a spectrophotometric approach to study β-NHE function through direct measurements of Hb–O2 binding. Therefore, 2 µl of blood was spread onto a temperature-controlled glass plate and the large surface-area-to-volume ratio of these samples allowed for an efficient equilibration of gases, as indicated by Hb–O2 saturations that responded within seconds after gas tensions were changed. Under these improved experimental conditions, the presence of extracellular CA had no effect on the protective function of β-NHE activity on Hb–O2 binding. The situation at the fish gill is likely similar, where RBCs squeeze through thin gill lamellae with large surface areas and most CO2 diffuses from the blood to the environment within seconds (Hughes, 1984).

Closed-system experiments

Our main finding that RBC β-NHE activity can protect pHi in an open system, even in the presence of extracellular CA, does not contradict previous work on RBC short-circuiting in closed systems (Harter et al., 2018a; Rummer and Brauner, 2011; Shu et al., 2017). In fact, we were able to replicate previous findings with our seawater-acclimated rainbow trout and using the same drug aliquots used in the open-system experiments, and even 10-fold lower (1 versus 10 µmol l−1) CA concentrations (Fig. 3A). CA additions to adrenergically stimulated RBCs in a closed system caused a significant increase in PO2 (ΔPO2 31.7±7.2 mmHg; P<0.003; Fig. 3B), indicating that β-NHE activity was unable to protect Hb–O2 binding. Furthermore, to validate that β-NHE activity during a respiratory acidosis can also be short-circuited by CA, a separate experiment acidified RBCs by injecting a CO2-saturated saline (Fig. 3C). The results were qualitatively similar to the HCl injections, albeit with a smaller ΔPO2. This corresponded with the lower acid load from CO2 (0.9 µmol l−1 H+ from CO2 versus 3.0 µmol l−1 H+ from HCl), which is constrained by the solubility of the gas in water. The combined results support that both metabolic and respiratory acidosis in a closed system can create conditions that enable β-NHE short-circuiting.

Fig. 3.

Results from the closed-system experiment. (A) Representative trace of the partial pressure of oxygen (PO2) in rainbow trout whole blood after sequential injections of: (i) 200 mmol l−1 HCl to acidify the blood by 0.3 pH units, (ii) the β-adrenergic agonist isoproterenol (ISO; 10 µmol l−1) and (iii) carbonic anhydrase (CA; 1 µmol l−1); all had significant effects on PO2 (P<0.05). (B) Changes in PO2 (ΔPO2) owing to drug additions, showing individual data points and means±s.e.m. (N=6). (C) Representative trace of a closed-system measurement of PO2 during a respiratory acidosis induced by injecting 20 μl CO2 saturated saline (i.e. 100 kPa PCO2), followed by ISO and CA injections as described above.

Fig. 3.

Results from the closed-system experiment. (A) Representative trace of the partial pressure of oxygen (PO2) in rainbow trout whole blood after sequential injections of: (i) 200 mmol l−1 HCl to acidify the blood by 0.3 pH units, (ii) the β-adrenergic agonist isoproterenol (ISO; 10 µmol l−1) and (iii) carbonic anhydrase (CA; 1 µmol l−1); all had significant effects on PO2 (P<0.05). (B) Changes in PO2 (ΔPO2) owing to drug additions, showing individual data points and means±s.e.m. (N=6). (C) Representative trace of a closed-system measurement of PO2 during a respiratory acidosis induced by injecting 20 μl CO2 saturated saline (i.e. 100 kPa PCO2), followed by ISO and CA injections as described above.

At the tissue capillaries, metabolically produced CO2 diffuses into the blood. Thus, any CO2 produced in the plasma cannot diffuse out until blood PCO2 becomes so high that it reverses the diffusion gradient from the tissues. Therefore, the tissue capillaries act as a functionally closed system for CO2, where paCA can effectively short-circuit β-NHE activity. The resulting decrease in RBC pHi during capillary transit greatly enhances the unloading of O2 from pH-sensitive Hb, and may be a fundamental aspect of the O2 transport system in many teleost fishes (Harter and Brauner, 2017). The present work adds to this growing body of literature by showing that paCA alone is not sufficient to disrupt RBC pHi regulation and that the diffusive characteristics of the system must also be considered.

Conclusion and perspectives

The sequence of evolutionary events that led to the complex cardiovascular O2 transport system in teleosts has been described in detail (Berenbrink et al., 2005), but the timeline for a loss of branchial paCA remains largely unresolved. Interestingly, however, bowfin (Amia calva), a basal actinopterygian without a RBC β-NHE, does not have paCA at the gills (Gervais and Tufts, 1998), perhaps indicating that the loss of the enzyme preceded the evolution of a β-NHE. Therefore, in addition to our mechanistic data, this independent line of evidence supports the idea that the selective pressures that led to the loss of branchial paCA in teleosts did not include the need to protect RBC pHi regulation by β-NHEs. Future work should investigate additional species that bracket the transition from basal actinopterygians to early teleosts, and consider alternative selective pressures for the notorious loss of paCA at their gills.

The authors thank Phil Zerofski, Christine Steinke, Ryan Myers and Dan Jio for their help with animal care.

Author contributions

Conceptualization: T.S.H.; Methodology: T.S.H.; Validation: T.S.H.; Formal analysis: T.S.H., E.A.S.; Investigation: T.S.H., E.A.S.; Resources: M.T.; Data curation: T.S.H., M.T.; Writing - original draft: T.S.H.; Writing - review & editing: T.S.H., E.A.S., M.T.; Visualization: T.S.H., E.A.S., M.T.; Supervision: T.S.H., M.T.; Project administration: T.S.H., M.T.; Funding acquisition: M.T.

Funding

This study was funded by National Science Foundation grant no. 1754994 to M.T. T.S.H. was supported by a Natural Sciences and Engineering Research Council of Canada Banting Postdoctoral Fellowship BPF-181632. Open Access funding provided by University of California, San Diego. Deposited in PMC for immediate release.

Data availability

All raw data and R source code are available from figshare at https://doi.org/10.6084/m9.figshare.23646405

ECR Spotlight

This article has an associated ECR Spotlight interview with Till Harter.

Berenbrink
,
M.
,
Koldkjaer
,
P.
,
Kepp
,
O.
and
Cossins
,
A. R.
(
2005
).
Evolution of oxygen secretion in fishes and the emergence of a complex physiological system
.
Science
307
,
1752
-
1757
.
Berenbrink
,
M.
,
Koldkjaer
,
P.
,
Wright
,
E. H.
,
Kepp
,
O.
and
Da Silva
,
A. J.
(
2011
).
Magnitude of the Root effect in red blood cells and haemoglobin solutions of fishes: a tribute to August Krogh
.
Acta Physiol.
202
,
583
-
592
.
Bohr
,
C.
,
Hasselbalch
,
K.
and
Krogh
,
A.
(
1904
).
Über einen in biologischer Beziehung wichtigen Einfluss, den die Kohlensäurespannung des blutes auf dessen Sauerstoffbindung übt
.
Skand. Arch. Fuer Physiol.
16
,
402
-
412
.
Boutilier
,
R. G.
,
Heming
,
T. A.
and
Iwama
,
G. K.
(
1984
).
Physicochemical parameters for use in fish respiratory physiology
. In
Fish Physiology
(
ed.
W. S.
Hoar
and
D. J.
Randall
), pp.
403
-
426
.
New York
:
Academic Press
.
Brauner
,
C. J.
,
Thorarensen
,
H.
,
Gallaugher
,
P.
,
Farrell
,
A. P.
and
Randall
,
D. J.
(
2000
).
The interaction between O2 and CO2 exchange in rainbow trout during graded sustained exercise
.
Respir. Physiol.
119
,
83
-
96
.
Brittain
,
T.
(
2005
).
Root effect hemoglobins
.
J. Inorg. Biochem.
99
,
120
-
129
.
Damsgaard
,
C.
,
Lauridsen
,
H.
,
Harter
,
T. S.
,
Kwan
,
G. T.
,
Thomsen
,
J. S.
,
Funder
,
A. M.
,
Supuran
,
C. T.
,
Tresguerres
,
M.
,
Matthews
,
P. G.
and
Brauner
,
C. J.
(
2020
).
A novel acidification mechanism for greatly enhanced oxygen supply to the fish retina
.
eLife
9
,
e58995
.
Esbaugh
,
A. J.
,
Gilmour
,
K. M.
and
Perry
,
S. F.
(
2009
).
Membrane-associated carbonic anhydrase in the respiratory system of the Pacific hagfish (Eptatretus stouti)
.
Respir. Physiol. Neurobiol.
166
,
107
-
116
.
Garcia-Romeu
,
F.
,
Motais
,
R.
and
Borgese
,
F.
(
1988
).
Desensitization by external Na+ of the cyclic AMP-dependent Na+/H+ antiporter in trout red blood cells
.
J. Gen. Physiol.
91
,
529
-
548
.
Gervais
,
M. R.
and
Tufts
,
B. L.
(
1998
).
Evidence for membrane-bound carbonic anhydrase in the air bladder of bowfin (Amia calva), a primitive air-breathing fish
.
J. Exp. Biol.
201
,
2205
-
2212
.
Gilmour
,
K. M.
,
Perry
,
S. F.
,
Bernier
,
N. J.
,
Henry
,
R. P. H.
and
Wood
,
C. M.
(
2001
).
Extracellular carbonic anhydrase in the dogfish, Squalus acanthias: a role in CO2 excretion
.
Physiol. Biochem. Zool.
74
,
477
-
492
.
Harter
,
T. S.
and
Brauner
,
C. J.
(
2017
).
The O2 and CO2 transport system in teleosts and the specialized mechanisms that enhance Hb-O2 unloading to tissues
. In
Fish Physiology
(
ed.
A. K.
Gamperl
,
T. E.
Gillis
,
A. P.
Farrell
and
C. J.
Brauner
), pp.
1
-
107
.
New York
:
Academic Press
.
Harter
,
T. S.
and
Brauner
,
C. J.
(
2021
).
Teleost red blood cells actively enhance the passive diffusion of oxygen that was discovered by August Krogh
.
Comp. Biochem. Physiol. A. Mol. Integr. Physiol.
253
,
110855
.
Harter
,
T. S.
,
May
,
A. G.
,
Federspiel
,
W. J.
,
Supuran
,
C. T.
and
Brauner
,
C. J.
(
2018a
).
The time-course of red blood cell intracellular pH recovery following short-circuiting in relation to venous transit times in rainbow trout, Oncorhynchus mykiss
.
Am. J. Physiol. Regul. Integr. Comp. Physiol.
315
,
R397
-
R407
.
Harter
,
T. S.
,
Sackville
,
M. A.
,
Wilson
,
J. M.
,
Metzger
,
D. C. H.
,
Egginton
,
S.
,
Esbaugh
,
A. J.
,
Farrell
,
A. P.
and
Brauner
,
C. J.
(
2018b
).
A solution to Nature's haemoglobin knockout: a plasma-accessible carbonic anhydrase catalyses CO2 excretion in Antarctic icefish gills
.
J. Exp. Biol.
221
,
jeb190918
.
Harter
,
T. S.
,
Zanuzzo
,
F. S.
,
Supuran
,
C. T.
,
Gamperl
,
A. K.
and
Brauner
,
C. J.
(
2019
).
Functional support for a novel mechanism that enhances tissue oxygen extraction in a teleost fish
.
Proc. R. Soc. B Biol. Sci.
286
,
20190339
.
Harter
,
T. S.
,
Clifford
,
A. M.
and
Tresguerres
,
M.
(
2021
).
Adrenergically induced translocation of red blood cell β-adrenergic sodium-proton exchangers has ecological relevance for hypoxic and hypercapnic white seabass
.
Am. J. Physiol. Regul. Integr. Comp. Physiol.
321
,
R655
-
R671
.
Heming
,
T. A.
,
Vanoye
,
C. G.
,
Stabenau
,
E. K.
,
Roush
,
E. D.
,
Fierke
,
C. A.
and
Bidani
,
A.
(
1993
).
Inhibitor sensitivity of pulmonary vascular carbonic anhydrase
.
J. Appl. Physiol.
75
,
1642
-
1649
.
Henry
,
R. P.
and
Swenson
,
E. R.
(
2000
).
The distribution and physiological significance of carbonic anhydrase in vertebrate gas exchange organs
.
Respir. Physiol.
121
,
1
-
12
.
Henry
,
R. P.
,
Gilmour
,
K. M.
,
Wood
,
C. M.
and
Perry
,
S. F.
(
1997
).
Extracellular carbonic anhydrase activity and carbonic anhydrase inhibitors in the circulatory system of fish
.
Physiol. Zool.
70
,
650
-
659
.
Hughes
,
G. M.
(
1984
).
General anatomy of the gills
. In
Fish Physiology
(
ed.
W. S.
Hoar
and
D. J.
Randall
), pp.
1
-
63
.
New York
:
Academic Press
.
Jacobs
,
M. H.
and
Stewart
,
D. R.
(
1942
).
The role of carbonic anhydrase in certain ionic exchanges involving the erythrocyte
.
J. Gen. Physiol.
25
,
539
-
552
.
Jensen
,
F. B.
,
Fago
,
A.
and
Weber
,
R. E.
(
1998
).
Hemoglobin structure and function
. In
Fish Physiology
(
ed.
S. F.
Perry
II
and
B. L.
Tufts
), pp.
1
-
40
.
New York
:
Academic Press
.
Jensen
,
F. B.
,
Agnisola
,
C.
and
Novak
,
I.
(
2009
).
ATP release and extracellular nucleotidase activity in erythrocytes and coronary circulation of rainbow trout
.
Comp. Biochem. Physiol. A Mol. Integr. Physiol.
152
,
351
-
356
.
Lessard
,
J.
,
Val
,
A.
,
Aota
,
S.
and
Randall
,
D.
(
1995
).
Why is there no carbonic anhydrase activity available to fish plasma?
J. Exp. Biol.
198
,
31
-
38
.
Mahe
,
Y.
,
Garciaromeu
,
F.
and
Motais
,
R.
(
1985
).
Inhibition by amiloride of both adenylate-cyclase activity and the Na+/H+ antiporter in fish erythrocytes
.
Eur. J. Pharmacol.
116
,
199
-
206
.
Maren
,
T. H.
(
1967
).
Carbonic anhydrase: chemistry, physiology, and inhibition
.
Physiol. Rev.
47
,
595
-
781
.
McMillan
,
O. J. L.
,
Dichiera
,
A. M.
,
Harter
,
T. S.
,
Wilson
,
J. M.
,
Esbaugh
,
A. J.
and
Brauner
,
C. J.
(
2019
).
Blood and gill carbonic anhydrase in the context of a chondrichthyan model of CO2 excretion
.
Physiol. Biochem. Zool.
92
,
554
-
566
.
Motais
,
R.
,
Fievet
,
B.
,
Garcia-Romeu
,
F.
and
Thomas
,
S.
(
1989a
).
Na+-H+ exchange and pH regulation in red blood cells: role of uncatalyzed H2CO3 dehydration
.
Am. J. Physiol.
256
,
C728
-
C735
.
Motais
,
R.
,
Fievet
,
B.
and
Garcia-Romeu
,
F.
(
1989b
).
Effect of Na+-H+ antiport activation on pH for erythrocytes suspended in a HCO3 containing saline
.
Stud. Biophys.
134
,
121
-
126
.
Nikinmaa
,
M.
(
1992
).
Membrane transport and control of hemoglobin-oxygen affinity in nucleated erythrocytes
.
Physiol. Rev.
72
,
301
-
321
.
Nikinmaa
,
M.
and
Jensen
,
F. B.
(
1986
).
Blood oxygen transport and acid-base status of stressed trout (Salmo gairdnerii): pre- and postbranchial values in winter fish
.
Comp. Biochem. Physiol. A Physiol.
84
,
391
-
396
.
Nikinmaa
,
M.
,
Tiihonen
,
K.
and
Paajaste
,
M.
(
1990
).
Adrenergic control of red-cell pH in salmonid fish: roles of the sodium proton-exchange, Jacobs–Stewart cycle and membrane potential
.
J. Exp. Biol.
154
,
257
-
271
.
Pedersen
,
S. F.
,
King
,
S. A.
,
Rigor
,
R. R.
,
Zhuang
,
Z.
,
Warren
,
J. M.
and
Cala
,
P. M.
(
2003
).
Molecular cloning of NHE1 from winter flounder RBCs: activation by osmotic shrinkage, cAMP, and calyculin A
.
Am. J. Physiol. Cell Physiol.
284
,
C1561
-
C1576
.
Pelster
,
B.
(
1997
).
Buoyancy at depth
. In
Fish Physiology
(
ed.
D.
Randall
and
A.
Farrell
), pp.
195
-
238
.
New York
:
Academic Press
.
Perry
,
S. F.
and
Thomas
,
S.
(
1993
).
Rapid respiratory changes in trout red blood cells during Na+/H+ exchange activation
.
J. Exp. Biol.
180
,
27
-
37
.
Randall
,
D. J.
and
Perry
,
S. F.
(
1992
).
Catecholamines
. In
Fish Physiology
(
ed.
W. S.
Hoar
,
D. J.
Randall
and
A. P.
Farrell
), pp.
255
-
300
.
New York
:
Academic Press
.
Randall
,
D. J.
,
Rummer
,
J. L.
,
Wilson
,
J. M.
,
Wang
,
S.
and
Brauner
,
C. J.
(
2014
).
A unique mode of tissue oxygenation and the adaptive radiation of teleost fishes
.
J. Exp. Biol.
217
,
1205
-
1214
.
Reid
,
S. D.
and
Perry
,
S. F.
(
1994
).
Quantification of presumptive Na+/H+ antiporters of the erythrocytes of trout and eel
.
Fish Physiol. Biochem.
12
,
455
-
463
.
Root
,
R. W.
(
1931
).
The respiratory function of the blood of marine fishes
.
Biol. Bull.
61
,
427
-
456
.
Roush
,
E. D.
and
Fierke
,
C. A.
(
1992
).
Purification and characterization of a carbonic anhydrase II inhibitor from porcine plasma
.
Biochemistry
31
,
12536
-
12542
.
Rummer
,
J. L.
and
Brauner
,
C. J.
(
2011
).
Plasma-accessible carbonic anhydrase at the tissue of a teleost fish may greatly enhance oxygen delivery: in vitro evidence in rainbow trout, Oncorhynchus mykiss
.
J. Exp. Biol.
214
,
2319
-
2328
.
Rummer
,
J. L.
,
Mckenzie
,
D. J.
,
Innocenti
,
A.
,
Supuran
,
C. T.
and
Brauner
,
C. J.
(
2013
).
Root effect hemoglobin may have evolved to enhance general tissue oxygen delivery
.
Science
340
,
1327
-
1329
.
Shu
,
J. J.
(
2019
).
On the teleost system for enhanced oxygen unloading: time course following catecholamine removal and an investigation into other teleosts
. PhD dissertation, University of British Columbia.
Shu
,
J. J.
,
Harter
,
T. S.
,
Morrison
,
P. R.
and
Brauner
,
C. J.
(
2017
).
Enhanced hemoglobin-oxygen unloading in migratory salmonids
.
J. Comp. Physiol. B
7
,
1
-
11
.
Stabenau
,
E. K.
and
Heming
,
T.
(
2003
).
Pulmonary carbonic anhydrase in vertebrate gas exchange organs
.
Comp. Biochem. Physiol. A. Mol. Integr. Physiol.
136
,
271
-
279
.
Tetens
,
V.
,
Lykkeboe
,
G.
and
Christensen
,
N. J.
(
1988
).
Potency of adrenaline and noradrenaline for b-adrenergic proton extrusion from red cells of rainbow trout, Salmo gairdneri
.
J. Exp. Biol.
134
,
267
-
280
.
Van Kampen
,
E. J.
and
Zijlstra
,
W. G.
(
1983
).
Spectrophotometry of hemoglobin and hemoglobin derivatives
. In
Advances in Clinical Chemistry
(
ed.
A. L.
Latner
and
M. K.
Schwartz
), pp.
199
-
257
.
New York
:
Academic Press
.
Wickham
,
H.
(
2009
).
ggplot2: Elegant Graphics for Data Analysis
.
New York
:
Springer-Verlag
.
Wittenberg
,
J. B.
and
Wittenberg
,
B. A.
(
1962
).
Active secretion of oxygen into the eye of fish
.
Nature
194
,
106
-
107
.
Wolf
,
K.
(
1963
).
Physiological salines for fresh-water teleosts
.
Progress. Fish Cult.
25
,
135
-
140
.
Wood
,
C. M.
and
Munger
,
R. S.
(
1994
).
Carbonic anhydrase injection provides evidence for the role of blood acid-base status in stimulating ventilation after exhaustive exercise in rainbow trout
.
J. Exp. Biol.
194
,
225
-
253
.
Wood
,
C. M.
,
Mcdonald
,
D. G.
and
Mcmahon
,
B. R.
(
1982
).
The influence of experimental anemia on blood acid-base regulation in vivo and in vitro in the starry flounder (Platichthys stellatus) and the rainbow trout (Salmo gairdneri)
.
J. Exp. Biol.
96
,
221
-
237
.
Zeidler
,
R.
and
Kim
,
H. D.
(
1977
).
Preferential hemolysis of postnatal calf red cells induced by internal alkalinization
.
J. Gen. Physiol.
70
,
385
-
401
.

Competing interests

The authors declare no competing or financial interests.

This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution and reproduction in any medium provided that the original work is properly attributed.

Supplementary information