Time course of red blood cell intracellular pH recovery following short-circuiting in relation to venous transit times in rainbow trout, Oncorhynchus mykiss

Abstract

Accumulating evidence is highlighting the importance of a system of enhanced hemoglobin-oxygen (Hb-O₂) unloading for cardiovascular O₂ transport in teleosts. Adrenergically stimulated sodium-proton exchangers (β-NHE) create H⁺ gradients across the red blood cell (RBC) membrane that are short-circuited in the presence of plasma-accessible carbonic anhydrase (paCA) at the tissues; the result is a large arterial-venous pH shift that greatly enhances O₂ unloading from pH-sensitive Hb. However, RBC intracellular pH (pH_i) must recover during venous transit (31–90 s) to enable O₂ loading at the gills. The halftimes (t_1/2) and magnitudes of RBC β-adrenergic stimulation, short-circuiting with paCA and recovery of RBC pH_i, were assessed in vitro, on rainbow trout whole blood, and using changes in closed-system partial pressure of O₂ as a sensitive indicator for changes in RBC pH_i. In addition, the recovery rate of RBC pH_i was assessed in a continuous-flow apparatus that more closely mimics RBC transit through the circulation. Results indicate that: 1) the t_1/2 of β-NHE short-circuiting is likely within the residence time of blood in the capillaries, 2) the t_1/2 of RBC pH_i recovery is 17 s and within the time of RBC venous transit, and 3) after short-circuiting, RBCs reestablish the initial H⁺ gradient across the membrane and can potentially undergo repeated cycles of short-circuiting and recovery. Thus, teleosts have evolved a system that greatly enhances O₂ unloading from pH-sensitive Hb at the tissues, while protecting O₂ loading at the gills; the resulting increase in O₂ transport per unit of blood flow may enable the tremendous athletic ability of salmonids.

INTRODUCTION

The hemoglobins (Hb) of most vertebrates exhibit some pH sensitivity, and typically Hb-oxygen (O₂) binding affinity decreases with a reduction in pH, termed the Bohr effect (5). The Bohr effect permits a high Hb-O₂ affinity at the gas-exchange surface, where pH is high, and a low affinity at the tissue capillaries, where metabolically produced CO₂ diffuses into the blood and pH is low. In this way, the Bohr effect can increase the difference between arterial and venous Hb-O₂ saturation and elevate the capacitance of the blood for O₂, thus allowing more O₂ to be transported per unit of blood flow. The magnitude of the Bohr effect is a function of the pH sensitivity of Hb (expressed as the Bohr coefficient) and the arterial-venous pH change that Hb, within the red blood cell (RBC), experiences at the tissue capillaries (ΔpH_a-v). Teleost Hb typically has a large Bohr coefficient (4), and in addition, many species exhibit a Root effect, where a reduction in pH will prevent Hb from becoming fully O₂ saturated even at O₂ tensions above atmospheric (44, 52). To exploit this extreme pH sensitivity, teleosts have evolved dedicated morphological structures (retia mirabilia) that create a large ΔpH_a-v locally and can generate the high partial pressures of O₂ (Po₂) that are necessary to fill the swim bladder at depth and to deliver O₂ to the avascular retina in the eye (59).

However, during a blood acidosis, pH-sensitive Hb-O₂ binding may compromise O₂ loading at the gills and reduce the O₂-carrying capacity of teleost blood (36). Enhanced O₂ unloading by acidifying Hb at the tissues can only increase the capacitance of the blood for O₂ on a systemic level, if O₂ loading at the gas exchange surface is not compromised; consequently, a large ΔpH_a-v must be localized to the tissues. At the swim bladder and the eyes of teleosts, this is enabled by the retes, and thus it was predicted that enhanced Hb-O₂ unloading was restricted to these specialized structures. More recently, however, a novel mechanism has been proposed that localizes a large ΔpH_a-v to sites of plasma-accessible carbonic anhydrase (paCA) in the circulation, thus potentially enhancing Hb-O₂ unloading to all tissues that possess paCA (42, 49).

Many teleost species that have a Root effect can actively regulate RBC intracellular pH (pH_i) and thus protect Hb-O₂ binding at the gills during an acidosis. In stressful situations, such as exercise or hypoxia, that may generate an acidosis, catecholamines are released into the blood that bind to β-adrenergic receptors on the RBC membrane and activate sodium-proton-exchangers (β-NHEs). These β-NHEs actively pump H⁺s from the RBC cytosol into the plasma and thus increase RBC pH_i. In the plasma, these H⁺s will combine with ${\text{HCO}}_3^{-}$ to form CO₂ that can freely diffuse into the RBCs where it is rapidly converted back into H⁺ and ${\text{HCO}}_3^{-}$, a reaction that is catalyzed by the abundant carbonic anhydrase (CA) pool within RBCs (30); this passive transfer of H⁺ across the RBC membrane via CO₂ is termed the Jacobs-Stewart cycle (25). In the absence of extracellular CA activity, the Jacobs-Stewart cycle is limited by the uncatalyzed rate of CO₂ production in the plasma, which allows β-NHE activity to create an H⁺ gradient across the RBC membrane. However, the addition of extracellular CA activity removes this limitation, and H⁺ extruded by β-NHEs will immediately combine with ${\text{HCO}}_3^{-}$ in the plasma to form CO₂ that reacidifies the RBCs. Thus, in the presence of extracellular CA, H⁺ extrusion is futile, and β-NHE activity is effectively short-circuited (33). Teleosts lack soluble CA in the plasma and paCA at the gills (reviewed in Ref. 20), enabling β-NHE activity and active RBC pH_i regulation. At the tissue capillaries of teleosts, where paCA is likely present (23), β-NHE activity will be short-circuited, rapidly transferring H⁺s into the RBCs and creating a large ΔpH_a-v that enhances the unloading of O₂ from Hb. However, short-circuiting of β-NHE activity will only increase the capacitance of blood for O₂ on a systemic level if RBC pH_i is restored during venous transit, enabling O₂ loading at the gills; thus, the recovery of RBC pH_i will effectively localize the ΔpH_a-v experienced by Hb to the tissues.

Although only investigated in a few studies, it appears that paCA activity is absent in the venous circulation of teleosts (41, 42), a requirement for RBC pH_i recovery in this compartment. The transit time of RBCs through the venous circulation has not been measured in fish (37, 51), and best estimates are based on cardiac output and the volume of the primary circulation. Cardiac output in rainbow trout ranges from 19 ± 4 ml/kg at rest to 56 ± 6 ml/kg during maximal exercise (means ± SE based on data from Refs. 7, 28, and 57). Total blood volume in rainbow trout has been assessed by a variety of methods and on average is 41 ± 3 ml/kg (means ± SE from 13 studies using a variety of methods) (reviewed in Ref. 37). Consequently, average blood transit times through the entire circulation will vary between 130 and 43 s, depending on the level of activity. If the venous blood volume in fish, like in mammals, is ~70% of total blood volume (39, 45), then venous transit time at rest is ~91 s and ~30 s during maximal exercise.

Therefore, we hypothesized that the time course of RBC pH_i recovery, after short-circuiting, must concur with the estimated time of venous transit (30–91 s). To test this hypothesis, we used a two-pronged approach. The time courses of β-NHE activation, short-circuiting, and RBC pH_i recovery were assessed in a closed-system preparation that has been validated previously (47), and that elegantly avoids the difficulty of measuring RBC pH_i, by using changes in closed-system Po₂ as a sensitive proxy measurement. To validate that the observed changes in closed-system Po₂ are, in fact, due to β-NHE short-circuiting, experiments were performed on rainbow trout (Oncorhynchus mykiss) and on white sturgeon (Acipenser transmontanus) that do not possess RBC β-NHEs (4). Although this closed-system preparation can quickly generate qualitative data that is comparable to previous work, RBC pH_i recovery was also measured in a continuous-flow apparatus that more closely resembles the in vivo conditions that RBCs experience in the circulation. Results from the combined approach provide strong evidence that a metabolon of RBC β-NHEs and tissue paCA can create and localize a large ΔpH_a-v at the tissues and greatly enhance the capacitance of teleost blood for O₂.

MATERIALS AND METHODS

Animals and housing.

Rainbow trout, Oncorhynchus mykiss Walbaum (average body mass, series 1: 2,266 ± 207 g; series 2: 2,943 ± 188 g), were obtained from Miracle Springs (Mission, BC, Canada), and white sturgeon, Acipenser transmontanus Richardson (761 ± 34 g) were obtained from Vancouver Island University (Nanaimo, BC, Canada). All animals were maintained at the University of British Columbia aquatic facility in 4,000-liter tanks, supplied with flow-through dechlorinated municipal tap water (Vancouver, BC, Canada) at 12°C, and kept under a 12:12 h photoperiod. Fish were fed to satiation every second day using commercial trout pellets (Orient 4–0, Skretting, Vancouver, BC, Canada), and feeding was suspended the day before surgeries. Animal husbandry and all experiments were conducted according to the guidelines of the Canadian Council on Animal Care and approved by the University of British Columbia Animal Care Committee (Protocol no. A15–0266).

Blood sampling and preparation.

Fish were anesthetized in 0.2 g/l tricaine methanesulfonate (MS222, Argent, Redmond, WA) buffered with Na${\text{HCO}}_3^{-}$. After loss of equilibrium, animals were placed on a surgery table where oxygenated and cooled (12°C) water, with a maintenance dose of anesthetic (0.1 g/l buffered MS222), was irrigated over the gills. Fish were chronically cannulated with a catheter (PE50, BD Intramedic, Franklin Lakes, NJ) into the dorsal aorta according to Soivio et al. (55). Thereafter, fish were recovered by irrigating the gills with fresh water and then transferred to individual holding tanks supplied with flow-through water. Fish were allowed to recover for 48 h after surgery and before blood sampling, and cannulas were flushed with heparinized Cortland’s saline [in mM: 124.1 NaCl, 5.1 KCl, 1.6 CaCl₂, 0.9 MgSO₄, 11.9 NaHCO₃, and 3.0 NaH₂PO₄ at pH 7.4 and 50 IU Na-heparin (Sigma H3149)] twice per day to prevent blood clotting.

Blood sampling was conducted in the mornings, before the lights were turned on. Blood was drawn slowly from the cannula into a heparinized syringe, and care was taken not to disturb the animal; sampling was suspended if the fish responded to the procedure. After sampling was completed, animals were euthanized in 0.2 g/l buffered MS222, and blood was stored on ice for processing. Hematocrit (Hct) was measured in triplicate in 15-µl capillary tubes that were centrifuged at 10,000 g for 3 min. Hct in the samples was then adjusted to 25% by adding Cortland’s saline or by gently centrifuging the blood (500 g for 3 min) and removing plasma. Thereafter, 3 ml of blood were loaded into Eschweiler tonometers and were equilibrated (1 h) to a humidified, custom gas mixture generated by a Woesthoff pump (Bochum, Germany). Gas tensions were chosen to produce ~75% Hb-O₂ saturation, and based on preliminary trials and previous work, gases were mixed at 3.8 mmHg CO₂ (0.5 kPa) and 76 mmHg O₂ (10 kPa) in N₂ for rainbow trout (46, 47) and 3.8 mmHg CO₂ and 38 mmHg O₂ in N₂ for white sturgeon (10).

Series 1: closed-system experiment.

To determine the recovery rate of RBC pH_i after short-circuiting, we used a two-pronged approach. First, recovery rates were determined using a closed-system preparation that has been the standard method to characterize RBC β-NHE short-circuiting in fishes (47). After equilibration of rainbow trout or white sturgeon blood in tonometers, a subsample (150 µl) was removed to measure initial Hct (as described above) and blood pH with a thermostated (12°C) microelectrode (16–705 and 16–702; Microelectrodes, Bedford, NH). Thereafter, 2 ml of blood were loaded into a glass vial that was sealed with a septum, carefully excluding any air. This vial was thermostated at 12°C and fitted with a magnetic stir bar that kept the sample well mixed. Po₂ within the vial was measured continuously with a fiberoptic O₂ microsensor (Loligo, Viborg, Denmark; and PreSens MicroTX3 meter, PreSens, Regensburg, Germany) that was pierced through the septum. Because of the pH sensitivity of Hb, changes in RBC pH_i are reflected in qualitative changes in closed-system Po₂, and this overcomes the challenges of measuring RBC pH_i in real-time.

According to previous work (47), blood within the closed system was acidified by injecting HCl, β-adrenergically stimulated with isoproterenol (Iso, a synthetic β-agonist), and short-circuited with soluble CA, treatments that have large effects on RBC pH_i and thus closed-system Po₂ (47, 54). We extended this protocol by a final injection of C18, a membrane-impermeable CA inhibitor (53) that inhibits the soluble CA injected previously, but that has no significant effect on RBC intracellular CA within 1 h of injection (49). The inhibition of the extracellular CA pool with C18 allows for a renewed recovery of RBC pH_i by β-NHE activity. The following solutions were made up in Cortland’s saline and were injected into the vial (total injection volume was 3% of vial volume) in 5-min intervals (as determined in preliminary trials): 1) 20 µl of 500 mM HCl, to acidify the blood by ~0.3 pH units according to Ref. 60; 2) 20 µl of 2 mM Iso (Sigma I-5627), for a final concentration of 0.01 mM; 3) 10 µl of 0.2 mM CA (CA2 from bovine erythrocytes, Sigma C3934), for a final concentration of 10⁻³ mM; and 4) 10 µl of 40 mM C18 (in 20% DMSO), for a final concentration of 0.2 mM C18 (0.1% DMSO). At the end of the trial, a subsample of blood was removed from the vial to measure final Hct and pH. In a separate trial, solvent controls were run to assess the effect of 0.1% DMSO on ΔPo₂; however, no effects were observed (ΔPo₂ was 0.0 ± 2.6 mmHg, n = 3).

Series 2: continuous-flow apparatus.

Although the results obtained in the closed-system preparation were qualitatively informative, it became apparent that RBC pH_i recovery occurred at a rate that was likely to be confounded by: 1) the rate of mixing within the closed system that delayed the delivery of drugs (Iso, CA, or C18) and the measurement of Po₂ and 2) the response time of the fiberoptic sensors that was in the order of several seconds for the observed ΔPo₂; these constraints have been identified previously (47). Therefore, in addition to the closed-system experiment, RBC pH_i recovery was measured in a continuous-flow apparatus that is robust against these confounding factors (see Validation of the continuous-flow apparatus) and that replicates more closely the dynamic conditions that RBCs experience in the circulation. A similar approach has been used previously to measure the rate of rapid reactions that exceed the response time of standard measuring techniques, such as Bohr and Root shifts in blood and CO₂ diffusion across RBC membranes (14, 16, 17).

After equilibration in tonometers, 3 ml of blood were loaded into a 10-ml, gas-tight, Hamilton syringe at 12°C, excluding any air. As in the closed system, blood in the syringe was acidified (by ~0.3 pH units) and β-adrenergically stimulated with Iso (0.01 mM). The sample was carefully mixed by inverting the syringe, and this process was aided by a small stir bar within the syringe that moved through the sample with every tilt. Thereafter, the syringe was connected to the system (see Fig. 1A for system diagram) via a blunted 18-gauge needle fitted to gas-impermeable tubing [polytetrafluoroethylene (PTFE) no. 20 AWG, Cole-Parmer, Vernon Hills, IL]. Blood was drawn out of the syringe by a peristaltic pump, which retracted the syringe plunger as the volume of blood decreased (ensuring no air exposure). The flow rate of the peristaltic pump was set to 0.2 ml/min, as determined in preliminary trials. Downstream of the peristaltic pump, blood was perfused through a section of glass-capillary tubing (length = 10 mm, inner diameter = 0.15 mm) that was fitted with 10 polymethyl-pentene fibers (Oxyplus, Membrana GmbH, Wuppertal, Germany). These fibers were coated with CA (CA2 from bovine erythrocytes, Sigma C3934) according to Refs. 3 and 26, creating a site of localized CA activity, such as blood may experience during capillary transit in fish. Thus, during the residence time of blood within this section (~8 s), plasma CO₂-${\text{HCO}}_3^{-}$ reactions were catalyzed, and β-NHE activity of stimulated RBCs was short-circuited. When blood left this site, plasma CO₂-${\text{HCO}}_3^{-}$ reactions became, once again, uncatalyzed, and RBC β-NHE activity restored H⁺ gradients across the membrane. Therefore, whether plasma CO₂-${\text{HCO}}_3^{-}$ reactions were catalyzed depended only on the position of the RBCs within the system, and because no inhibitor was involved, the confounding effects of mixing and inhibition kinetics were avoided.

Fig. 1.

A: schematic of the continuous-flow apparatus. 1) Blood from cannulated rainbow trout was equilibrated in tonometers, acidified, and β-adrenergically stimulated with isoproterenol, a β-agonist. 2) Blood was perfused (0.2 ml/min) through a glass capillary containing carbonic anhydrase (CA)-coated polymethyl-pentene (PMP) fibers (3, 26). In the presence of CA, red blood cell (RBC) sodium-proton-exchangers (β-NHEs) are short-circuited, leading to a decrease in RBC intracellular pH (pH_i). 3) When blood leaves the site of CA, β-NHE activity recovers RBC pH_i, which is reflected in a change in the partial pressure of O₂ (ΔPo₂), because of the pH sensitivity of hemoglobin-O₂ binding in rainbow trout. By placing fiberoptic Po₂ sensors at fixed distances from the site of short-circuiting, the ΔPo₂ response was reconstructed using the readings from all sensors, each corresponding to a time point in the response. 4) A blood sample was collected for measurements of final hematocrit (Hct) and pH. B: setup for validation of the continuous-flow apparatus. 5) Buffer was continuously (1 ml/min) equilibrated to 38 mmHg CO₂ in N₂ in a hollow-fiber gas exchanger, to create a pH disequilibrium (at 12°C). 6) Buffer in disequilibrium was immediately perfused through a glass capillary containing CA-coated or uncoated (Ctrl) fibers. 7) Buffer pH was measured downstream with a flow-through pH microelectrode after stopping the flow. Change in pH (ΔpH) during stop flow is indicative of the residual pH disequilibrium in the buffer; an absence of a ΔpH indicates the presence of CA catalytic activity in the system. 8) Perfusate buffer was collected for the analysis of CA activity that was washed out from the coated fibers using the electrometric ΔpH assay (22).

At a continuous blood-flow rate and a constant diameter of the PTFE tubing (inner diameter = 0.81 mm), the distance that blood has traveled after leaving the site of CA will correspond to the time of RBC pH_i recovery. Po₂ in the blood was measured at 5, 15, 30, 60, and 120 s after short-circuiting by placing fiberoptic sensors at fixed distances from the site of CA (32, 96, 193, 385, and 771 mm). Therefore, each Po₂ sensor performed measurements at a constant Po₂ that corresponded to one time point in the RBC pH_i recovery reaction. The system was run continuously for up to 15 min, which allowed sufficient time for complete equilibration of the Po₂ sensors and avoided the issue of slow sensor response times. Finally, after blood had passed through the system, a blood sample was collected at the outlet for measurements of final Hct and pH (as described above).

Validation of the continuous-flow apparatus.

New polymethyl-pentene fibers were coated with CA every day and were fitted into the glass capillary before experiments. The presence of CA activity within the capillary was confirmed before running trials by using a buffer in pH disequilibrium and stop-flow measurements of pH according to Ref. 19. Briefly, a continuous flow of assay buffer (in mM: 225 mannitol, 75 sucrose, 10 Tris base; pH adjusted to 7.4 with 10% phosphoric acid) was rapidly loaded with CO₂ in a hollow-fiber membrane gas exchanger (Fig. 1B) by equilibrating the buffer to a humidified, high-CO₂ gas (38 mmHg CO₂ in N₂ from a Woesthoff gas mixing pump). Thereafter, the buffer was immediately perfused over the CA-coated fibers (or noncoated fibers as a control), and pH was measured downstream using a microelectrode (Microelectrodes, Bedford, NH). After equilibration of the pH readings, the flow was stomeasured downstream using a microelectrode (Microelectrodespped and the magnitude and direction of the ΔpH was measured.

To confirm that CA activity was not washed out from the fibers, perfusate buffer (1 ml) was collected from the previous CA-validation trial and analyzed for CA activity using the electrometric ΔpH assay (22). Reactions were in 6 ml of assay buffer (described above) at 4°C using 100 μl CO₂-saturated water as a substrate and in the presence or absence of 100 μl of perfusate sample. The reaction kinetics were assessed as the time for a 0.15 unit ΔpH with a GK2401C electrode and PHM84 meter (Radiometer, Copenhagen, Denmark), and absolute rates were calculated from the buffer curve of the assay buffer over the tested pH range.

Data analysis and statistics.

All data were analyzed in R studio v1.0.153 (R v3.4.1), and figures were generated with the ggplot2 v.2.2.1 package (58). Normality of the data was tested with the Shapiro-Wilk test and homogeneity of variances with the Levene’s test (P < 0.05). The rate of RBC pH_i recovery was assessed by measuring changes in closed-system Po₂ and taking advantage of the pH sensitivity of Hb. Po₂ data were normalized to the initial Po₂ values within each run and expressed as ΔPo₂ over time. To ensure that the RBC pH_i recovery between the two systems were comparable, the continuous Po₂ traces from the closed system were subsampled at those times corresponding to measurements in the continuous-flow apparatus (i.e., 5, 15, 30, 60, and 120 s after short-circuiting) using LabChart v8.1.5 (ADInstruments, Dunedin, New Zealand). In the closed system, the metabolic O₂ consumption (ṀO₂) of RBCs was determined as the average slope (mmHg/s) of the Po₂ trace (over a 40 s interval), immediately before and after treatments (injections of Iso, CA, or C18). In the continuous-flow apparatus, RBC ṀO₂ was determined as the slope of the Po₂ trace for each Po₂ sensor; all ΔPo₂ data were corrected for the average RBC ṀO₂ over the tested period. However, no significant differences (paired-samples t-test; P > 0.05) were observed between the ṀO₂ corrected and the uncorrected data set; the ṀO₂ corrected data are presented here. ΔPo₂ data for all treatments were analyzed by fitting two nonlinear models to each data set, the Michaelis-Menten enzyme kinetics model,

$$\Delta {PO}_2=\frac{at}{\left(b+t\right)}$$ (1)

and the Hill-model,

$$\Delta {PO}_2=\frac{a{t}^b}{\left(c^b+t^b\right)}$$ (2)

where t is the time after treatment, a represents the magnitude of the response (ΔPo_2max), and b (Eq. 1) and c (Eq. 2) represent the halftimes (t_1/2), respectively. The fit of both models was compared with the Akaike information criterion (1), and the model with the lower Akaike information criterion was used for analysis. Representative models were run on the pooled data sets for each treatment and system, and these are depicted in the figures. However, for statistical analysis, each individual trace was analyzed as described above, yielding parameter estimates t_1/2 and ΔPo_2max that are reported as means ± SE (n = 15, unless otherwise indicated). Because of the very rapid and variable ΔPo₂ after acidification (Fig. 2), HCl data were not analyzed by fitting a model. Therefore, t_1/2 was not determined for this treatment, and ΔPo_2max was calculated as the difference between Po₂ at HCl injection and the maximal Po₂ after acidification. Differences in ΔPo_2max between treatments (HCl, Iso, CA, and C18) and t_1/2 between treatments (Iso, CA, and C18) in the closed system were tested with Kruskal-Wallis one-way analysis of variance (P < 0.05, n = 15, unless otherwise indicated) and the kruskalmc function (R pgirmess package) for post hoc analysis. Differences between initial and final blood pH and Hct (n = 15 in rainbow trout and n = 6 in white sturgeon), and differences between ΔPo_2max and t_1/2 between the closed and continuous-flow systems (n = 15), were tested with the Wilcoxon rank-sum test (P < 0.05). Differences between control (Ctrl) and CA treatments in the validation of the continuous-flow apparatus were tested with independent samples t-tests (P < 0.05, n = 11).

Fig. 2.

Representative trace of changes in the partial pressure of oxygen (Po₂, mmHg) in rainbow trout (top, dark trace) and white sturgeon (bottom, light trace) whole blood hematocrit [(Hct) = 25%] in a closed-system trial. Blood was sampled from cannulated animals at rest and equilibrated in tonometers to conditions that mimic venous blood: 76 mmHg (rainbow trout) or 38 mmHg O₂ (white sturgeon), 3.8 mmHg CO₂ in N₂ at 12°C. Blood was loaded into a 2-ml vial (47), and dashed lines indicate injections (through a sealed septum) of: 1) HCl to acidify the sample by ~0.3 pH units; 2) isoproterenol (Iso), a synthetic β-agonist at 0.01 mM; 3) carbonic anhydrase (CA) at 10⁻³ mM; and 4) a membrane-impermeable CA inhibitor (C18) at 0.2 mM.

RESULTS

Series 1: closed-system experiment.

In the closed-system experiment, the initial Hct of rainbow trout blood was on average 25.1 ± 0.7%, and the experimental protocol increased Hct to a final value of 29.9 ± 0.8% (P < 0.001). Initial blood pH was 7.62 ± 0.02 and decreased to 7.34 ± 0.02 by the end of the experimental protocol (P < 0.001), a reduction of 0.28 ± 0.03 that was in line with the planned degree of acidification with HCl. In white sturgeon blood, however, there was no significant increase in Hct because of the experimental protocol, and average Hct was 23.1 ± 0.4% (P = 0.750). Blood pH in white sturgeon decreased significantly (P = 0.005) from 7.63 ± 0.01 to 7.47 ± 0.02, as expected.

Figure 2 shows representative traces from two closed-system trials on rainbow trout and white sturgeon blood, respectively. In rainbow trout, the ΔPo₂ because of Iso, CA, and C18 injections was analyzed on 15 individual traces by fitting nonlinear models to the data, and representative models were fitted to the pooled data sets shown in Fig. 3; the parameter estimates for all representative models are summarized in Table 1. The pooled data sets were best described by a sigmoidal Hill curve for the Iso response (Fig. 3A) and a Michaelis-Menten curve for the CA (Fig. 3B) and C18 responses (Fig. 3C).

Fig. 3.

Closed-system change in the partial pressure of oxygen (ΔPo₂, mmHg) measured in rainbow trout whole blood (see Fig. 2 caption for blood preparation) after injections of the following. A: isoproterenol (Iso, 0.01 mM), a synthetic β-agonist. B: carbonic anhydrase (CA, 10⁻³ mM). C: a membrane-impermeable CA inhibitor (C18, 0.2 mM). Individual data points are indicated by light circles and means ± SE (n = 15) by dark circles, and the models of best fit are indicated by continuous lines (Hill-model for Iso and Michealis-Menten models for CA and C18). The response halftime (t_1/2) and magnitudes (ΔPo_2max) are indicated by vertical and horizontal dashed lines as means ± SE, and were determined from the models fitted to the pooled data (these do not exactly resemble the parameter estimates used for statistical analysis that represent averages of individually analyzed traces).

Table 1.

Nonlinear models of best-fit and parameter estimates for changes in closed-system Po₂ of rainbow trout whole blood, in response to stimulation with Iso, short-circuiting with extracellular CA and recovery of red blood cell intracellular pH after inhibition of CA with C18 in a closed system, or absence of CA in a continuous-flow apparatus

	Parameter Estimates
Treatment	a	b	c	Model
Closed system
Iso	−27.64 ± 1.71*	2.50 ± 0.392*	63.65 ± 4.74*	Hill
CA	31.84 ± 3.48*	13.02 ± 3.88†		Michaelis-Menten
C18	−33.12 ± 6.23*	71.84 ± 26.75†		Michaelis-Menten
Continuous-flow apparatus	−36.13 ± 2.89*	16.97 ± 4.53*		Michaelis-Menten

C18, membrane-impermeable CA inhibitor; CA, carbonic anhydrase; Iso, isoproterenol; Po₂, partial pressure of oxygen.

^*P < 0.001,

^†P < 0.01.

Based on the analysis of individual traces, the stimulation of RBC β-NHEs with Iso had a ΔPo_2max of −27.5 ± 3.2 mmHg with a t_1/2 of 63.9 ± 2.3 s. The addition of extracellular CA had a ΔPo_2max of 29.5 ± 2.0 mmHg with a t_1/2 of 12.8 ± 1.3 s. The inhibition of extracellular CA with C18 had a ΔPo_2max of −27.4 ± 2.6 mmHg with a t_1/2 of 53.5 ± 5.2 s. Significant differences were detected between the average t_1/2 for Iso, CA, and C18 responses in rainbow trout (Fig. 4A). The ΔPo_2max after HCl addition was on average 46.0 ± 4.0 mmHg and significantly larger compared with the responses after Iso, CA, and C18 injections, which were not significantly different from one another (Fig. 4B). In white sturgeon, the ΔPo₂ because of acidification was significantly lower compared with rainbow trout (P < 0.001), and on average 10.8 ± 1.1 mmHg; subsequent injections of Iso, CA, or C18 had no effect on ΔPo₂ in white sturgeon (Fig. 5).

Fig. 4.

Halftimes (t_1/2, s) (A) and magnitudes (ΔPo_2max, mmHg) (B) of a closed-system change in the partial pressure of oxygen (ΔPo₂) in rainbow trout whole blood, after the addition of isoproterenol (Iso, 0.01 mM), carbonic anhydrase (CA, 10⁻³ mM), and a membrane-impermeable CA inhibitor (C18, 0.2 mM). Individual data points are indicated by circles with means ± SE. The t_1/2 and ΔPo_2max were determined by fitting nonlinear models to individual traces. Significant main effects of treatment were detected for t_1/2, but not for ΔPo_2max as determined with a nonparametric Kruskal-Wallis test (P < 0.05, n = 15), significant differences between treatments are indicated by letters a and b.

Fig. 5.

Magnitude (ΔPo_2max, mmHg) of a closed-system change in the partial pressure of oxygen (ΔPo₂) in white sturgeon whole blood (see Fig. 2 caption for blood preparation), after the addition of HCl to acidify the sample by ~0.3 pH units, isoproterenol (Iso, 0.01 mM), carbonic anhydrase (CA, 10⁻³ mM), and a membrane-impermeable CA inhibitor (C18, 0.2 mM). Individual data points are indicated by circles with means ± SE. A significant main effect of treatment was detected with a nonparametric Kruskal-Wallis test (P < 0.05, n = 15), significant differences between treatments are indicated by letters a and b.

Series 2: continuous-flow apparatus.

In the continuous-flow apparatus, the experimental protocol increased Hct of rainbow trout blood from 23.6 ± 0.4% to 32.1 ± 0.6% (P < 0.001) and decreased blood pH from 7.65 ± 0.02 to 7.39 ± 0.02 (P < 0.001); no significant difference in pH (P = 0.309) or Hct (P = 0.833) was observed between the closed-system and the continuous-flow experiments. RBC pH_i recovery in the continuous-flow apparatus was analyzed on 15 individual traces by fitting nonlinear models to the data. The pooled data set was best described by a Michaelis-Menten curve, which is depicted in Fig. 6, with the previous closed-system RBC pH_i recovery data (C18 treatment) for comparison. RBC pH_i recovery in the continuous-flow apparatus had a t_1/2 of 17.2 ± 2.5 s and was significantly faster compared with the closed system (Fig. 7A). The ΔPo_2max in the continuous-flow apparatus was −34.4 ± 3.0 mmHg and not significantly different from that in the closed system (Fig. 7B).

Fig. 6.

The rate of red blood cell (RBC) intracellular pH (pH_i) recovery, measured as a change in the partial pressure of oxygen (ΔPo₂, mmHg) in rainbow trout whole blood using two different methods (see Fig. 2 for blood preparation). Results from the continuous-flow apparatus, after blood has left the site of carbonic anhydrase (CA) activity (see Fig. 1), are shown as closed points and lines. Previous results from the closed system (see Fig. 3C, C18 treatment) are shown as open circles and dashed lines, allowing for a direct comparison. Individual data points and means ± SE are indicated by closed and open circles for the two methods, respectively. The response halftime (t_1/2) and magnitudes (ΔPo_2max) are indicated by vertical and horizontal lines as means ± SE, and were determined by fitting a nonlinear model (Michaelis-Menten) to the pooled data sets (n = 15).

Fig. 7.

Halftimes (t_1/2, s) (A) and magnitudes (ΔPo_2max, mmHg) (B) of red blood cell (RBC) intracellular pH (pH_i) recovery, measured as a change in the partial pressure of oxygen (ΔPo₂) in rainbow trout whole blood (see Fig. 2 for blood preparation) in a closed-system preparation after addition of a membrane-impermeable carbonic anhydrase (CA) inhibitor (C18, 0.2 mM) and in a continuous-flow apparatus after blood left the site of CA activity. Individual data points are indicated by circles with means ± SE. The t_1/2 and ΔPo_2max were determined by fitting nonlinear models to individual traces (n = 15). Significant main effects of the system were detected for t_1/2, but not for ΔPo_2max, with a nonparametric Kruskal-Wallis test (P < 0.05, n = 15).

Validation of continuous-flow apparatus.

Buffer that was perfused over Ctrl fibers showed a large negative ΔpH during stop-flow, indicative of an ongoing hydration of CO₂ and production of H⁺s, and thus an absence of CA activity (Fig. 8A). However, in CA-coated fibers, ΔpH was positive and significantly different from Ctrl. This is indicative of a complete hydration of CO₂ within the time the buffer was in contact with the coated fibers and a clear sign of CA catalytic activity within the glass capillary. Buffer without CO₂ loading was run as an additional control and also showed a positive ΔpH. No significant difference in CA activity was detected in perfusates from Ctrl or CA-coated fibers (Fig. 8B), indicating that the enzyme was appropriately bound to the fibers and did not wash out; this is an important prerequisite for the recovery of RBC pH_i after short-circuiting. In addition, the ΔpH assay was validated by measuring a sample of the CA solution that was used to coat fibers (a positive Ctrl), which confirmed the sensitivity of the assay.

Fig. 8.

Validation of the continuous-flow apparatus and the coating of polymethyl-pentene (PMP) fibers with carbonic anhydrase (CA). A: validation of CA activity in uncoated [control (Ctrl)] and CA-coated fibers fitted into a glass capillary that was perfused (1 ml/min) with a buffer in pH disequilibrium; changes in pH (ΔpH) were measured during periods of stopped flow with a pH microelectrode (see Fig. 1B for description of the setup). The system was also perfused with buffer alone (Buffer) to account for the effects of flow on pH measurements. Individual data points are indicated by circles with means ± SE. Differences between treatments were tested with an independent samples t-test (P < 0.05, n = 11). B: CA activity (µmol H⁺/min) measured with the electrometric ΔpH assay (22) on perfusate buffers collected from the previous validation experiment. Differences in CA activity between treatments were tested with an independent samples t-test (P < 0.05, n = 11). No significant difference in CA activity was detected between perfusates from CA and Ctrl fibers, thus indicating that no CA washout occurred from coated fibers. The assay was validated by running positive controls (pos Ctrl) on the CA solution used for coating the fibers (1 mg/ml bovine CA2).

DISCUSSION

It is critical that teleost RBCs recover pH_i during venous transit. Only then can β-NHE short-circuiting enhance Hb-O₂ unloading at the tissues without compromising O₂ uptake at the gill during a blood acidosis, a system that increases the difference between arterial and venous Hb-O₂ saturations and, therefore, the capacitance of the blood for O₂. The transit time of blood through the venous circulation in rainbow trout may vary between 30 and 91 s in exercise and resting scenarios, respectively (as described above). The results of the present work place the t_1/2 of RBC pH_i recovery, after short-circuiting in a continuous-flow apparatus, at 17 s, thus providing strong evidence that β-NHE activity is rapid enough to largely recover Hb-O₂ affinity during venous transit, even under conditions of maximal exercise and a severe acidosis.

The present study used two experimental systems to independently measure the rate of RBC pH_i recovery. The closed-system preparation has been the standard method for studying β-NHE short-circuiting in teleost RBCs (47); the t_1/2 and Po_2max values presented here, for the activation (Iso) and the short-circuiting (CA) of β-NHEs, resemble closely those of previous work. However, the slow mixing kinetics and sensor response times in the closed system may be inadequate to accurately assess the t_1/2 of fast changes in Po₂. In fact, the measured t_1/2 of RBC pH_i recovery was more than threefold slower in the closed system compared with a continuous-flow apparatus (Fig. 7A) that is robust to these confounding factors. These results lend support to the idea that the closed system overestimates t_1/2 measurements for all treatments, and therefore these values will not be discussed in detail; the t_1/2 of 17 s, obtained in the continuous-flow apparatus, is considered a more accurate measure of the rate of RBC pH_i recovery after short-circuiting. Closed-system results, however, provide important insight into the magnitude of HCl, Iso, CA, and C18 responses. Therefore, combined data from both systems can reveal the temporal kinetics and response magnitudes that characterize the teleost system of enhanced Hb-O₂ unloading.

Based on the model of best fit (Table 1), β-NHE activity may recover RBC pH_i by 64% within the venous transit times available during exercise and by 84% at rest; however, these values are likely conservative estimates. The activity of β-NHEs increases RBC pH_i and, therefore, shifts the oxygen equilibrium curve toward a higher Hb-O₂ affinity (left shift). Consequently, Hb binds additional O₂, Hb-O₂ saturation increases, and Po₂ in the closed system decreases. Because of the sigmoidal shape of the oxygen equilibrium curve, the relationship between Hb-O₂ saturation and Po₂ is necessarily nonlinear. Thus, at high levels of Hb-O₂ saturation, changes in affinity will have minor effects on closed-system Po₂ that will underestimate the true degree of RBC pH_i recovery. Consequently, the degree of RBC pH_i recovery during venous transit, as calculated here, is likely a lower boundary, and the measured t_1/2 may permit almost complete recovery of pH_i and thus Hb-O₂ affinity by the time RBCs return to the gills.

Rainbow trout blood was equilibrated to gas tensions that represent venous conditions during aerobic exercise (3.8 mmHg CO₂ and 76 mmHg O₂), resulting in an extracellular pH (pH_e) of 7.63 ± 0.01 (averaged data from both systems) that corresponds closely to in vivo measurements (8, 28). The addition of HCl lowered blood pH_e further, to 7.34 ± 0.01, a value consistent with a severe metabolic acidosis that in vivo may occur because of glycolytic ATP production at the exercising muscle. Final pH_e was measured at the end of the trial and after the activation of β-NHE, whereby H⁺ extrusion from the RBCs into the plasma will have contributed to the low pH_e measured. The initial acidification of blood by injection of HCl occurred in the absence of β-NHE activity; thus, pH_e and pH_i were coupled (21), and the acidosis was transferred into the intracellular compartment by the Jacobs-Stewart cycle. The reduction in pH_i because of HCl addition caused a ΔPo₂ of 46 mmHg, a result of the high pH sensitivity of rainbow trout Hb (Bohr coefficient of −0.91) (48).

The large ΔPo₂ observed after acidification of rainbow trout blood is the result of a large reduction in Hb-O₂ affinity and illustrates the dangers of a pH-sensitive Hb that, in teleosts, may fail to become fully oxygenated at the gas exchange surface. Therefore, during a blood acidosis in vivo, catecholamines are released into the blood that activate RBC β-NHEs and restore Hb-O₂ affinity; in vitro, RBCs were stimulated with Iso at concentrations that induce maximal β-NHE activity in rainbow trout (47). The addition of Iso caused a ΔPo_2max of 28 mmHg (Fig. 4B) that was significantly smaller than the ΔPo_2max observed after acidification with HCl. Clearly, more H⁺ entered the RBC during acidification than were subsequently removed by β-NHE activity, and thus RBC pH_i was not recovered to its original set point. As secondarily active transporters, β-NHEs build up Na⁺ and H⁺ gradients across the RBC membrane that are ultimately driven by the activity of the Na⁺-K⁺-ATPase (NKA). As extracellular H⁺s and intracellular Na⁺ accumulate (18), β-NHE activity slows, despite an increase in NKA activity during the β-adrenergic response (6, 13, 38). The acidification of the plasma increases the H₂CO₃ pool and thus the rate of uncatalyzed ${\text{HCO}}_3^{-}$ dehydration, accelerating the Jacobs-Steward cycle and the passive transfer of H⁺s into the RBCs. Eventually the apparent H⁺ fluxes of the β-NHEs and the Jacobs-Stewart cycle become equal, and net H⁺ extrusion comes to a halt. Thus, the degree of pH_i recovery will depend on transmembrane ion gradients, NKA activity, and the buffer capacities of the intra- and extracellular medium. Teleost Hb, the primary buffer within RBCs, typically has a low buffer capacity, and thus fewer H⁺s need to be extruded for a given change in pH_i (34). Under severely acidotic conditions, such as those induced here, high extracellular H⁺ concentrations may set thermodynamic limits upon β-NHE activity (40), limiting complete recovery of pH_i. The maximal H⁺ gradient that can be produced by β-NHE activity may be rather constant for a given species’ blood characteristics, and this H⁺ gradient will dictate the maximal ΔpH_a-v that can be created upon β-NHE short-circuiting to enhance Hb-O₂ unloading.

A critical finding of the present work was that the stimulation of β-NHEs with Iso, the short-circuiting with CA, and RBC pH_i recovery after addition of C18 produced a similar ΔPo_2max (Fig. 4B). This indicates that the entire H⁺ gradient across the RBC membrane that was initially established by β-NHE activity is available to enhance ΔpH_a-v and thus Hb-O₂ unloading at the tissues of teleosts, where paCA is present. In addition, when RBCs leave the site of paCA activity, β-NHE activity can reestablish the original H⁺ gradient that was present before short-circuiting. Importantly, this indicates that Hb-O₂ loading at the gill is protected but also that the same H⁺ gradient is available for renewed short-circuiting when RBCs reach the tissues once again. Thus, potentially, RBCs in rainbow trout can undergo repeated cycles of short-circuiting and recovery with every pass through the circulation and thus maintain an elevated capacitance of the blood for O₂ during a β-adrenergic response.

The initial stimulation of β-NHEs with Iso resulted in a time lag before changes in Po₂ were observed; thus, these data were best described by a sigmoidal Hill model that can accommodate a delayed response onset (Fig. 3A). The mechanism underlying the delayed Iso response is likely found in the excitation cascade that translates the stimulation of the RBC β-adrenergic receptor by catecholamines into an activation of the transporter, which involves adenylate cyclase and the accumulation of cAMP as a secondary messenger (29, 35). In contrast, the inhibition of CA by C18 was best described by a Michaelis-Menten curve, indicating the absence of a time lag (Fig. 3C). Conceptually, the addition of CA to stimulated RBCs will accelerate the Jacobs-Steward cycle and abolish the H⁺ gradient across the RBC membrane, despite ongoing β-NHE activity. Clearly, the large decrease in closed-system Po₂, caused by C18 addition, is evidence that β-NHE activity is ongoing, but futile, in the presence of CA. Therefore, β-NHEs are already activated when blood leaves the capillaries after short-circuiting, and this may be of physiological significance, as it will allow for a rapid onset of pH_i recovery.

In the continuous-flow apparatus, the recovery of RBC pH_i was approximately threefold faster compared with the closed system, which likely overestimated the t_1/2 of all treatments. If CA short-circuiting was confounded to a similar degree, then the t_1/2 was likely in the order of 4 s (3-fold faster than the measured t_1/2 = 12.8 s). This is perhaps a conservative estimate, as confounding factors may have a relatively larger contribution to the measured t_1/2 of fast responses (CA) compared with slower ones (Iso, C18); evidence from other studies indicates that the short-circuiting of β-NHE activity by CA is likely an extremely rapid process. The addition of extracellular CA to blood catalyzes the CO₂-${\text{HCO}}_3^{-}$ reactions in the plasma, which, because of the extremely high turnover rate of the enzyme (human CA2 at 25°C, k_cat = 10⁶/s), can be considered instantaneous in physiologically relevant time-scales (15, 27), and the same applies to intracellular CO₂-${\text{HCO}}_3^{-}$ reactions that are catalyzed by the RBC CA pool (30). Therefore, the intracellular acidification of RBCs in the presence of paCA is likely rate-limited by the diffusion of CO₂ across the RBC membrane, a rapid process with a t_1/2 of < 1 ms (at 37°C in human RBCs) (14, 56). In fact, when extracellular CA (0.02 g/l) was added to eel RBCs suspended in a solution that was in pH disequilibrium, the resulting Bohr shift and release of O₂ from Hb (ΔPo₂ 19 mmHg) had a t_1/2 of ~0.5 s (17). Thus, without venturing into an exact estimate of the t_1/2 of CA short-circuiting in rainbow trout, the combined data from both systems, and data from previous work, indicate that it may be sufficiently rapid to enhance Hb-O₂ unloading during capillary transit (<4 s) (24, 31), which is in line with the results from in vivo work on rainbow trout (49).

To demonstrate that the large changes in Po₂ observed here were, in fact, a direct consequence of the H⁺ gradients established by β-NHE activity, a second set of experiments was run on white sturgeon, a species that lacks an RBC β-NHE (4). The activation of β-NHEs leads to the intracellular accumulation of Na⁺ and Cl⁻ that causes osmotic swelling of RBCs and thus an increase in Hct. In rainbow trout blood, the addition of Iso increased Hct from 24.4 ± 0.4% to 31.0 ± 0.5%, which is in line with results from previous work (11, 47, 54). As expected, white sturgeon RBCs showed no β-adrenergic swelling, and Hct was unaffected by Iso addition. The evolution of β-NHE activity in teleosts correlates with the advent of highly pH-sensitive Hbs (4). White sturgeon have a moderate Root effect, however, that is only expressed at low pH values that are likely never encountered in the circulation in vivo (43); thus, in white sturgeon, β-NHE activity is not vital to protect Hb-O₂ loading at the gills. Acidification of white sturgeon blood, to a similar degree as in rainbow trout, resulted in a ΔPo₂ of only 11 mmHg (compared with 46 mmHg in rainbow trout), a result of the lower pH sensitivity of white sturgeon Hb (Bohr coefficient −0.4) (10, 12). Thus, in white sturgeon, the potential benefits of the Bohr effect for Hb-O₂ unloading are limited by the low pH sensitivity of Hb and by the absence of RBC β-NHEs that, when short-circuited at the tissues, create a large ΔpH_a-v that magnifies the Bohr effect in teleosts.

It has been proposed that other transporters on the RBC membrane may create H⁺ gradients that can be short-circuited in the presence of CA (47, 50). The idea that other taxa, which lack β-NHEs, may benefit from a similar mechanism of enhanced Hb-O₂ unloading is intriguing and may bear consequences for our understanding of O₂ transport in all vertebrates (9). However, in white sturgeon, the absence of a ΔPo₂ after CA addition (Figs. 2 and 5) indicates that H⁺s were distributed passively across the RBC membrane and refutes the idea that other nonadrenergic transporters may create H⁺ gradients that are available to enhance Hb-O₂ unloading. Based on these results, it appears that short-circuiting of H⁺ gradients across the RBC membrane is a mechanism that is exclusive to the teleost clade through a metabolon that requires β-NHE activity and a heterogeneous distribution of paCA in the circulation, conditions that are not met by most vertebrates. An acidosis, such as that induced by the injection of HCl, or a moderate pH sensitivity of Hb, such as observed in white sturgeon, does not appear to be sufficient to enhance Hb-O₂ unloading in the presence of paCA.

In the closed system, the addition of CA to stimulated RBCs caused an increase in Po₂ of 30 mmHg, which is in line with a previous study that found a ΔPo₂ of 25 mmHg under similar conditions (47). These reproducible findings add to an increasing body of literature that illustrates the great potential for β-NHE short-circuiting to enhance Hb-O₂ unloading in teleosts. This mechanism may explain why rainbow trout can sustain higher red muscle Po₂ compared with mammals, even during spontaneous struggling, exercise, or hypoxia (32), or why stressed striped bass (Morone saxatilis) have arterial Po₂ that is higher than those in water (36). When paCA in the circulation was inhibited with C18 in vivo, red muscle Po₂ in hypercapnic rainbow trout decreased significantly, indicating a role of paCA in maintaining elevated tissue Po₂ (49). Similarly, the injection of C18 into the circulation of swimming Atlantic salmon either resulted in collapse, or at lower speeds, in a large compensatory increase in cardiac output, indicating that enhanced Hb-O₂ unloading, mediated by paCA, facilitates O₂ transport during exercise (Harter et al., unpublished data). In the heart of coho salmon, paCA activity may enhance Hb-O₂ unloading to the spongy myocardium that relies exclusively on the O₂-depleted venous return (2). More broadly, recent work indicates that β-NHE short-circuiting may be important for migratory salmonids and perhaps teleosts in general (54), potentially enhancing Hb-O₂ unloading by over 70% relative to what mammals can achieve with their moderate Bohr effect alone (48).

In conclusion, the present study provides strong evidence that in rainbow trout: 1) the rate of RBC pH_i recovery after short-circuiting is sufficiently rapid to largely restore Hb-O₂ affinity during venous transit, 2) the short-circuiting of β-NHE activity by paCA may be sufficiently rapid to enhance Hb-O₂ unloading during capillary transit, and 3) after short-circuiting, RBCs reestablish the initial H⁺ gradient across the membrane, which protects O₂ loading at the gills, and sets the system up for renewed short-circuiting. Thus, potentially, β-NHE short-circuiting may enhance Hb-O₂ unloading with every pass through the circulation and maintain an elevated capacitance of the blood for O₂ during a β-adrenergic response. In addition, based on the results on white sturgeon, we found no indications that other nonadrenergic transporters create H⁺ gradients across the RBC membrane that can be short-circuited to enhance Hb-O₂ unloading. The temporal kinetics of β-NHE short-circuiting and RBC pH_i recovery corroborate that a system of enhanced Hb-O₂ unloading is not only feasible, but likely a major pillar of the teleost O₂-transport system. Thus, approximately half of all vertebrates may enhance cardiovascular O₂ transport by selectively creating and eliminating H⁺ gradients across the RBC membrane.

GRANTS

This study was supported by Natural Sciences and Engineering Research Council of Canada Accelerator Supplement (446005–13) and Discovery Grant (261924–13) to C. J. Brauner, and an NIH: Ambulatory Assist Lung for Children Award (NIH R01HL135482) to W. J. Federspiel.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

T.S.H. and C.J.B. conceived and designed research; T.S.H. performed experiments; T.S.H. analyzed data; T.S.H. and C.J.B. interpreted results of experiments; T.S.H. prepared figures; T.S.H. drafted manuscript; T.S.H. and C.J.B. edited and revised manuscript; A.G.M. and W.J.F. provided PTFE fibers for CA-coating; C.T.S provided C18; all authors approved final version of manuscript.