Introduction

The earth’s oceanic carbonate system (partial pressure of CO2, pH and alkalinity) is currently changing 100-times faster than at any point in the past 650,000 years (Caldeira and Wickett 2003; Siegenthaler et al. 2005). In fact, current oceanic CO2 levels represent an increase of 100 μatm (30 %) over pre-industrial levels, equating to an approximate 0.1 pH drop. At the current rate of change, atmospheric CO2 levels are expected to reach between 750 and 1,000 μatm by the end of the century, and 1,900 μatm by 2300, yielding pH reductions of 0.3–0.4 and 0.7–0.8 units, respectively (Caldeira and Wickett 2003). In addition, current day, regional CO2 levels in upwelling zones have been reported to reach levels up to 2,300 μatm (Cai 2011; Feely et al. 2008, 2010; Thomsen et al. 2010). These unprecedented changes in the marine environment pose potentially dramatic challenges for marine organisms and ecosystems. Marine calcifying organisms likely comprise the most at-risk animal group due to effects that ocean acidification has on the CaCO3 saturation state. Shell formation and maintenance both become more challenging and more energetically costly as saturation state declines (Fabry 2008; Guinotte and Fabry 2008; Orr et al. 2005). Consequently, the effects of ocean acidification on these organisms have been frequently studied leading to reports of significant deleterious effects on coral and crustose coralline algae bleaching and productivity (Albright et al. 2010; Anthony et al. 2008), coral recruitment (Albright et al. 2010), bivalve growth and shell formation (Miller et al. 2009; Talmage and Gobler 2010). Alternatively, high PCO2 levels may aid calcification in coccolithophores, which have increased in size by 40 % over the past 220 years (Iglesias-Rodriguez et al. 2008).

While the general physiological response of fish to hypercapnia (high environmental CO2) is well established (see reviews Brauner and Baker 2009; Perry and Gilmour 2006), only recently have studies begun to focus on environmentally realistic and near future CO2 levels. Such CO2 levels affect larval survival and growth of Medinia beryllina (Baumann et al. 2012) and cause tissue damage in larval Gadus morhua (Frommel et al. 2012). Further effects for reef fish species include olfactory disturbances that may impede habitat selection (Dixson et al. 2010; Munday et al. 2009b), auditory effects (Simpson et al. 2011), and predator–prey interactions (Ferrari et al. 2011a, b; Munday et al. 2010). This group of effects are related to inhibited neurotransmitter function (Nilsson et al. 2012) and are predicted to have significant impacts on reef fish populations and recruitment (Munday et al. 2010). Furthermore, otolith growth in white sea bass was enhanced at approximately 1,000 and 2,500 μatm (Checkley et al. 2009). As of yet, no work has carefully examined the potential effects that near-future hypercapnia levels may have on respiratory gas exchange and acid–base balance. Generally, studies involving hypercapnia and fish have focused on physiological mechanisms related to acid–base balance and respiratory gas exchange and therefore have typically involved very high CO2 levels of 10,000–50,000 μatm (e.g. Baker et al. 2009; Brauner et al. 2004; Gilmour and Perry 1994; Perry et al. 2010; Tzaneva et al. 2011). These levels result in CO2 diffusion into the animal leading to an acidosis in the plasma and intracellular environment. However, fish have been shown to be tolerant of such acidosis owing to a variety of protective mechanisms (see reviews Claiborne et al. 2002; Perry and Gilmour 2006). Generally, fish protect extracellular fluid pH status by exchanging acid–base equivalents with the environment via branchial mitochondrial-rich cells; tissue intracellular pH is protected in a qualitatively similar pattern (Cameron and Kormanik 1982). With respect to acid equivalents, protons are excreted apically via either a sodium–proton exchange protein (NHE) or a V-type H+ ATPase. Both of these mechanisms occur in freshwater fish while marine animals predominantly use NHEs (Claiborne et al. 2002), likely due to favorable thermodynamic gradients (Potts 1994). Freshwater fish transport HCO3 − across the basolateral membrane by a sodium bicarbonate co-transport protein (nbc) (Perry et al. 2003). Little is known about basolateral HCO3 − transport in the gills of marine fish, but nbc is expressed in the gills of gulf toadfish (Taylor et al. 2010) and shows upregulation in response to hypersalinity transfer, presumably to compensate for the associated plasma acidosis.

The current study used three experimental series to examine the effects of low-level hypercapnia on the blood acid–base chemistry of a marine teleost, the gulf toadfish (Opsanus beta). In the first experimental series, we tested the hypothesis that acute exposure to predicted future levels of atmospheric CO2 (560, 750, 1,000 and 1,900 μatm) will have significant impacts on the blood acid–base chemistry of toadfish. In series two we investigated the mechanisms of branchial acid–base compensation initiated by low-level hypercapnia, by comparing acid–base flux measurements, gene expression of relevant proteins, and the effects of HCO3 − free seawater for control animals and those exposed to 1,900 μatm CO2. For the third experimental series, we tested the hypothesis that respiratory acidoses caused by low-level hypercapnia result from a decreased outward gradient of CO2 (1,000 μatm CO2) and are indirectly the result of apparent diffusion limitations of CO2 during branchial capillary transit.

Methods

Experimental animals

Adult gulf toadfish (N = 73; 30–121 g) were obtained from commercial fisherman in Biscayne Bay, Florida, USA. Once obtained, animals were sorted by size and transferred to 40 l aquaria (>50 g fish, 8–10 per tank; <50 g fish, 20 per tank) supplied with flow through aerated, circulating filtered seawater (22–25 °C) at Rosenstiel School of Marine and Atmospheric Science, University of Miami. After initial transfer to the holding facility fish were treated with Malachite Green (Aquatic Ecosystems) to prevent ectoparasite infection (McDonald et al. 2003). Each tank was supplied with short lengths of polyvinylchloride tubing for use as shelters, and animals were fed twice a week with frozen squid. Food was withheld for at least 48 h prior to experimentation. All experiments were performed using protocols approved by the University of Miami Institutional Animal Care and Use Committee.

Hypercapnia exposure system

Exposure CO2 concentrations were controlled using a commercially available PCO2/pH feedback controller (DAQ-S; Loligo Systems Inc.) connected to a wtw pH 3310 meter and SenTix 41 pH electrode (Loligo Systems Inc) and controlled using CapCTRL software (Loligo Systems Inc). This automated system used measured pH data and an input PCO2–pH standard curve to determine water PCO2. When chamber PCO2 dropped below the set point, as determined by pH measurements, the system would add pure CO2 gas via an airstone until the PCO2 was returned to the desired level. To prevent excessive fluctuations in PCO2 and over exposure, the CO2 gas rates were manually set to the lowest possible levels that still allowed the desired PCO2 level to be obtained. The PCO2–pH standard curve was generated by equilibrating seawater with normal air or custom CO2–O2 mixed gas (3,090 ppm) and measuring the resulting pH. Periodic measurements of total CO2 were also taken to verify PCO2 concentrations using CO2SYS. A header tank design was not used for exposures because it resulted in excessive variance in water CO2, but it should be noted that equilibrated tanks of either size showed very little within-tank variation in pH/PCO2, as detected using a second pH meter/electrode.

Surgical procedures

Gulf toadfish (N = 42) were affixed with caudal artery catheters to assess the effects of varying levels of hypercapnia on blood acid–base status. This procedure involved anesthetizing the animal using MS-222 (0.5 g/l) buffered with NaHCO3 (1 g/l) and implanting a heparin-saline (0.9 % NaCl) filled catheter (PE-50) into the caudal artery, as outlined by Wood et al. (1997). Animals were allowed to recover for 24–48 h prior to experimentation. Toadfish were equipped with a rectal collection sac to isolate the branchial contribution to acid–base balance, according to the method described by Genz et al. (2008). After surgery, animals were placed directly in exposure tanks, as collection of gut contents would begin immediately after the collection sac was equipped.

Series I: the effects of low-level hypercapnia on blood acid–base chemistry

Cannulated animals were placed in individual numbered holding tanks (approximately 750 ml) with numerous 2 cm holes to allow for water flow. These holding tanks were placed within larger exposure tanks. Both 30- and 6-l plastic containers with flow-through sand-filtered seawater and aeration were used as exposure tanks. The tank size was determined on the basis of the number of animals exposed at a given time, with the exception of 560 μatm experiments that were all performed in the 6-l tank due to the finer control necessary to reliably maintain the exposure levels. One experimental series consisted of an initial blood sample through the arterial catheter followed by a 24-h exposure period to a given CO2 after which a final blood sample was taken. The tested CO2 exposure treatments were 380 (control), 560, 750, 1,000 and 1,900 μatm CO2. Each cannulated animal was exposed to control conditions and only one of the above CO2 treatments. A second time trial experimental series was performed at 1,000 and 1,900 μatm CO2 and consisted of 0-, 0.25-, 0.5-, 1-, 2-, 4-, 8-, 24-, and 48-h sampling points. A 200-μl blood sample was taken for all sampling points and analyzed for blood pH and plasma total CO2 (see analytical methods). For time trial experiments involving nine sampling points, the removed red blood cells were isolated, resuspended in saline, and re-injected into the animal via the catheter, to prevent confounding effects related to possible anemia. At the termination of experiments, the animals were euthanized by spinal transection.

Series II: branchial contributions to acid–base compensation during low-level hypercapnia exposure

To isolate the branchial contribution to acid–base compensation in toadfish, animals were first equipped with a rectal collection sac, as described above. The branchial acid–base flux was determined on individual animals in 1.5 l exposure containers, and PCO2 was controlled as described above. The animal was allowed to recover for 4 h during which the exposure containers received seawater flow. After recovery, seawater flow was stopped and a 15-ml initial water sample was taken. After 24 h, a final sample was taken followed by a 4-h chamber flush period, after which another initial sample was taken. Experiments lasted 72 h in total with exact flux times, animal mass and chamber volumes recorded. Both HCO3 − and NH3 concentrations (see analytical techniques) were determined for all initial and final samples and acid–base equivalent flux rates were calculated for each interval. After 72 h animals were euthanized by an overdose of MS-222 followed by a spinal transection; all intestinal and rectal fluids were collected for other research purposes.

To investigate branchial gene expression and enzyme activity changes in response to low-level hypercapnia, toadfish (N = 31; 30–40 g) were placed in a 30-l tank with flowing seawater and allowed to acclimate for 24 h. Control animals (N = 8) were sampled for gill and muscle tissue, after which the 1,900-μatm CO2 exposure was initiated. Subsequent samples were taken at 8 h (N = 8), 24 h (N = 8) and 72 h (N = 7) post-exposure. This time window has previously demonstrated gene expression changes in teleost gills and intestinal tissue following environmental manipulation (Guffey et al. 2011; Ivanis et al. 2008; Sattin et al. 2010; Taylor et al. 2010). Animals were euthanized with an overdose of MS-222, after which muscle tissue (approximately 100 mg) was excised from the dorsal tail region, freeze clamped using liquid nitrogen and stored at −80 °C. The gills were perfused by making a ventral incision to expose the heart and aorta. A 23-gauge needle equipped to a 10-ml syringe filled with heparin-saline was passed through the ventricle and into the aorta after which the atrium was severed. The syringe was slowly depressed to clear the body of blood; successful perfusion was determined based on gill color, and if deemed incomplete was repeated. The procedure took approximately 5 min after which the gill arches were sampled for mRNA and ATPase activity, frozen in liquid nitrogen and subsequently stored at −80 °C. Samples for ATPase activity were frozen in 100 μl of SEI buffer (described in analytical techniques).

For a final set of experiments, animals were exposed to 1,900 μatm CO2 while being held in nominally HCO3 −-free water (N = 6). The HCO3 − free water (469 mM NaCl, 28.2 mM MgSO4, 10.3 mM KCl, 24.6 mM MgCl2) was buffered with 3 mM HEPES free acid and the pH was set to 8.1 using NaOH. A separate PCO2/pH standard curve was created with this water, as described above, which was qualitatively similar to regular seawater in the magnitude of pH change. Additionally, CO2 gassed water was allowed to return to normal under standard aeration to ensure that hypercapnia did not permanently alter pH and thereby give overestimates of PCO2. Post-experiment a 50-ml aliquot of water was analyzed for HCO3 − by double endpoint titration. Remaining experimental procedures were as described for series I.

Series III: the effects of carbonic anhydrase infusion on the response to low-level hypercapnia

A third experimental series examined the effects of carbonic anhydrase infusion on blood acid–base chemistry after low-level hypercapnia exposure. Animals were equipped with a caudal artery catheter, as described above, and allowed to recover for 24 h. An initial blood sample of 200 μl was taken and analyzed as in series one, after which 5 mg/kg of bovine carbonic anhydrase was infused through the catheter. Volumes ranged from 113 to 176 μl per animal. The animals were allowed to equilibrate for 1 h after which the PCO2 exposure was initiated. A final blood sample was taken 24 h after exposure and analyzed as described above. An additional 500 μl of blood was removed for analysis of plasma CA activity. Experiments were performed at both 1,000 μatm (outward CO2 diffusion gradient; N = 4) and 1,900 μatm (neutral CO2 gradient; N = 4), and the response was analyzed in relation to the response of series one animals.

Molecular techniques

Total RNA was extracted from toadfish gill tissues using RNA Stat-60 reagent (Tel-test Inc, TX, USA) according to manufacturer guidelines, with homogenization performed using motor driven tissue homogenizer. Total RNA was quantified using an ND-1,000 (Thermo Fisher Scientific, DE, USA) spectrophotometer at a wavelength of 260 nm. Prior to cDNA synthesis, a subsample of RNA was DNase-treated with amplification grade DNase I (Invitrogen, CA, USA; manufacturer specifications) to remove potential DNA contamination. Subsequent cDNA synthesis was performed using RevertAid MULV reverse transcriptase (Fermentas, MD, USA), according to manufacturer’s specifications. Real-time PCR was performed on an Mx3000P real-time PCR system (Stratagene, CA, USA.) using the Brilliant SYBR green master mix kit (Stratagene, CA, USA.; 12.5 μl reactions). The gene-specific primers are listed in Table 1; new primer sets (slc4a2, slc26a3) were designed using the Fastpcr freeware program. Both the thermocycler set-up and reaction composition were performed according to manufacturer guidelines, and disassociation curves were used to assess the primer specificity of each reaction. The PCR efficiency of each primer pair was calculated using a cDNA standard curve. PCR efficiencies ranged from 85 to 103 % with an R 2 ≥ 0.98. Relative mRNA expression was calculated using the delta–delta ct method using elongation factor 1α (EF1α) as an internal control and the fasted treatment as the relative control (Pfaffl 2001). Successful DNase treatment was verified using a no-reverse transcriptase control for each tissue set.

Table 1 List of real-time PCR primers used for branchial gene expression analysis
Full size table

Analytical techniques

Blood pH was analyzed using a custom-built gastight, thermostated sleeve equipped with a pH electrode (Radiometer) attached to a MeterLab portable pH meter (Radiometer). Plasma total CO2 was determined using a Corning total CO2 analyzer; red blood cells and plasma were separated by low-speed centrifugation for 1 min using a bench top centrifuge (VWR). Plasma PCO2 was calculated from total CO2 and pH according to the Henderson-Hasselbach equation using the appropriate toadfish plasma constants (Boutilier et al. 1984). Levels of HCO3 − were assessed using a double-endpoint titration procedure (Hill 1973) as previously described (Genz et al. 2008). Ammonia was determined using colorimetric assays (Verdouw et al. 1978). Na+/K+ ATPase and V-type ATPase enzyme levels were assessed as previously described (Lin and Randall 1993; McCormick and Bern 1989) and normalized to total protein as determined by the Bradford assay. Plasma CA levels were assessed using the delta pH method (Henry 1991). Intracellular muscle pH was determined using the homogenate method (Portner et al. 1990) and the gastight pH apparatus described above.

Statistical analysis

Paired student’s t tests were used for all comparisons on the effects of varying levels of hypercapnia on blood acid–base properties. Time trial analysis of blood gas properties at 1,000 and 1,900 μatm CO2 were assessed by repeated measures analysis of variance (ANOVA). The effects of CA injection and HCO3 − free water on the hypercapnia response were assessed using two-way repeated measures ANOVA with CO2 status as one factor and the experimental treatment as the other. Effects of hypercapnia on NKA and V-type ATPase levels, muscle pHi and gene expression were assessed using ANOVA analyses. Branchial acid–base flux data were assessed using an unpaired student’s t test. The level of significance for all statistical analyses was P < 0.05. Also note that power tests were performed on all negative results in an effort to assess whether increased sample size would affect the outcome. The number of animals required to meet statistical significance at a power level of 0.8 at a constant mean difference and standard deviation was deemed excessive.

Results

Series I: The effects of low-level hypercapnia on blood acid–base chemistry

The PCO2/pH feedback controller system was successfully able to control 560-, 750-, 1,000- and 1,900-μatm exposure treatments. In general, the measured PCO2 exposures did not exceed the desired set point by more than 5 % for the 560-μatm treatment, 10 % for the 750- and 1,000-μatm treatments and 15 % for the 1,900-μatm treatment. The duration between CO2 bursts was variable between treatments, but was generally less than 90 s. These treatments resulted in seawater pH decreases of approximately 0.15, 0.2, 0.4 and 0.6 for 560, 750, 1,000 and 1,900 μatm, respectively.

Initial experiments tested the effects of 24-h exposure at 560-, 750-, 1,000- and 1,900-μatm CO2 on the blood acid–base chemistry of gulf toadfish (Fig. 1). The 750- (N = 4), 1,000- (N = 6), and 1,900 (N = 6)-μatm exposures all had significantly higher plasma HCO3 − and PCO2 at 24 h of exposure compared with control measurements. Neither the 560-μatm exposure (N = 5) nor the control exposure (380 μatm; N = 3) had significant effects on any blood acid–base parameter (paired t test); pH was unchanged by 24 h of exposure to any treatment. The mean plasma HCO3 − concentration appears to increase in a dose-dependent manner with environmental CO2. It should be noted that control plasma HCO3 − values were significantly lower in the 1,000-μatm treatment than the 560- and 1,900-μatm treatments, likely owing to intrinsic differences between the groups of fish used. No differences were detected between control groups for plasma PCO2 or blood pH. Intracellular pHi measurements were performed on white muscle samples from toadfish exposed to 1,900 μatm for 8 h and 24 h, as well as unexposed animals (Fig. 2). Exposure resulted in significantly increased pHi at the 24-h time point compared with unexposed animals, while no change was observed at the 8-h time point.

Fig. 1
figure 1

The effects of 24-h exposure to low-level hypercapnia on a blood pH, b plasma HCO3 − and c plasma PCO2 in the gulf toadfish, Opsanus beta. Values represent mean ± SEM, N = 3, 5, 4, 6 and 6 for 380, 560, 750, 1,000 and 1,900 μatm CO2, respectively. A significant difference between control (black) and 24-h (gray) blood variables is denoted by an asterisk (paired t test, P ≤ 0.05)

Full size image
Fig. 2
figure 2

The effects of 1,900 μatm CO2 on intracellular white muscle pH (pHi) in the gulf toadfish, Opsanus beta. Values represent mean ± SEM, N = 7. A significant difference from time 0 values is denoted by an asterisk (ANOVA, P ≤ 0.05)

Full size image

A time course experiment was performed to more fully characterize the blood acid–base disturbance at both 1,000 and 1,900 μatm CO2 (Fig. 3; N = 4). Significant decreases in blood pH relative to initial samples were observed as early as 15 min after the initiation of hypercapnia at both 1,000 and 1,900 μatm (Fig. 3a). Blood pH returned to pre-hypercapnia levels by 2 and 4 h in the 1,900- and 1,000-μatm treatments, respectively. The plasma PCO2 increased by 15 min and remained elevated for the duration of the time trial in the 1,900-μatm treatment (Fig. 3b). In the 1,000-μatm treatment a significant difference was not detected until 8 h after exposure was initiated. Plasma HCO3 − began to increase by 1 h post-exposure in the 1,900-μatm treatment and reached statistical significance by 4 h, remaining elevated for the duration of the exposure (Fig. 3c). The 1,000-μatm treatment showed a peak response in HCO3 − concentration at 8 h post-exposure. When displayed as PCO2/pH/HCO3 − diagrams (Fig. 3d, e) the 1,900-μatm data show a response typical of a respiratory acidosis followed by compensation. While the 1,000-μatm data show a similar response over the first 8-h period, the 24- and 48-h data suggest a secondary acidosis in this treatment.

Fig. 3
figure 3

A time course of the effects of 1,000 and 1,900 μatm CO2 on a blood pH, b plasma PCO2 and c plasma HCO3 − in the gulf toadfish, Opsanus beta, and expressed as pH/HCO3 −/PCO2 diagrams (D = 1,000 μatm, E = 1,900 μatm, dashed line estimated blood buffer line). Values represent mean ± SEM, N = 4. a, b and c: A significant difference from control values is denoted by an asterisk (repeated measures ANOVA, P ≤ 0.05)

Full size image

Series II: branchial contributions to acid–base compensation during low-level hypercapnia exposure

The branchial contribution to acid–base flux during 1,900 μatm CO2 exposures was assessed after equipping animals with rectal collection sacs, which were implanted to remove intestinal contributions to measured acid–base compensation. There were no detectable differences between titratable acid efflux (69 ± 30 and 4.9 ± 44.8), ammonia efflux (18 ± 6.5 and 17.9 ± 4.5) or total acid efflux (85.3 ± 36.8 and 22.8 ± 45.3) between control and the 1,900-μatm-exposed fish, respectively. Further study of the mechanistic pathway for acid–base compensation was performed by exposing animals to hypercapnia in HCO3 − free (nominally) seawater (Fig. 4). After exposure to 1,900 μatm CO2 for 24 h, toadfish in HCO3 − free seawater (N = 6) were unable to compensate for blood acidosis in a similar manner to animals in normal seawater (N = 10). Toadfish exposed to hypercapnia in HCO3 − free water had significantly lower blood pH after exposure. Plasma HCO3 − was unaffected by CO2 exposure in HCO3 − free water, while showing a significant increase in normal seawater. Plasma PCO2 was significantly increased in both HCO3 − free and control seawater; however, HCO3 − free exposed animals had a lower increase as compared with control water animals.

Fig. 4
figure 4

The effects of HCO3 − free seawater on the blood acid–base response to 24-h exposure to 1,900 μatm CO2 in the gulf toadfish, Opsanus beta. Significant effects of hypercapnia within a treatment are denoted by an asterisk, while a significant difference between treatments is denoted by a dagger (N = 10 and 6 for control and HCO3 − free seawater, respectively; 2-way repeated measures ANOVA, P ≤ 0.05)

Full size image

Branchial gene expression of known HCO3 − transporters, including slc26a3, slc26a6, slc4a2 and the sodium bicarbonate co-transporter (nbc), as well as cytoplasmic carbonic anhydrase (CA) was examined at 0, 8, 24 and 72 h post-exposure to 1,900 μatm CO2. A significant decrease in CA expression was observed at 8, 24 and 72 h post-exposure, while slc4a2 had decreased expression at 8 and 72 h post-exposure. No effects on gene expression were observed for nbc, slc26a3 or slc26a6. Na+/K+ ATPase enzyme activity was significantly decreased at 24 h relative to the 0 h time point and returned to control levels by 72 h of exposure (Fig. 5b). Although not statistically significant, V-type H+ ATPase enzyme activity showed a similar trend (Fig. 5b).

Fig. 5
figure 5

The effects of exposure to 1,900 μatm CO2 on a branchial gene expression, and b Na+/K+ (NKA) and V-type H+ATPase enzyme activity. Significant differences versus control samples are denoted by an asterisk (ANOVA, N = 7−8)

Full size image

Series III: the effects of carbonic anhydrase infusion on the response to low-level hypercapnia

Experiments examining the effects of 5 mg/kg bovine CA infusion into the blood were performed at both 1,000 and 1,900 μatm (N = 4 each; Fig. 6). The blood acid–base profile of CA infused animals after 24-h exposure to 1,900 μatm CO2 conformed to trends of non-CA infused animals for all three blood parameters (Fig. 6a). In contrast, infusion of CA in combination with 24-h exposure to 1,000 μatm resulted in significantly elevated blood pH relative to both paired pre-exposure measurements and non-infused toadfish post-exposure measurements. Furthermore, plasma PCO2 was not significantly elevated after hypercapnia exposure in CA infused animals; plasma HCO3 − conformed to control animal trends with a significant increase after 24 h hypercapnia exposure. The amount of CA in the plasma of injected toadfish after 24 h was 250 ± 120 and 932 ± 191 enzyme units for the 1,900- and 1,000-μatm treatments, respectively. Non-injected animals contained only 60 ± 23 enzyme units, where a single enzyme unit is defined as a doubling of the uncatalyzed rate.

Fig. 6
figure 6

The effects of carbonic anhydrase (CA) infusion (5 mg/kg) on the blood acid–base response to 24-h exposure to a 1,000 μatm and b 1,900 μatm CO2 in the gulf toadfish, Opsanus beta. Significant effects of hypercapnia within a treatment are denoted by an asterisk, while a significant differences between treatments is denoted by a dagger (N = 10 and 4 for control and CA infusion, respectively; 2 way repeated measures ANOVA, P ≤ 0.05)

Full size image

Discussion

The current study has demonstrated that acute exposure to CO2 levels relevant for present time upwelling zones and near-future climate change result in a respiratory acidosis in toadfish and compensation for this respiratory acidosis depends on seawater HCO3 − uptake. We have also demonstrated that the less favorable partial pressure gradients observed at 1,000 μatm can be compensated for through CA infusion into the blood, which acts to lower the chemical equilibrium constraints governing CO2 excretion.

Hypercapnia as an experimental treatment has been extensively studied, both in freshwater and marine fish. However, the majority of studies in this area have utilized hypercapnia as a means to study mechanisms of acid–base and ion transport, or cardiorespiratory physiology, and have used environmentally unrealistic CO2 levels ranging from 1 to 5 % to highlight physiological effects and mechanisms. Only recently has focus turned to the possible impacts of environmentally relevant CO2 levels related to climate change and ocean acidification or current regionally elevated CO2 levels, with documented impacts of 1,000 μatm CO2 on embryo survival and growth in M. beryllina (Baumann et al. 2012), olfactory responses and aerobic scope in several reef fish species (Dixson et al. 2010; Munday et al. 2009a, b), as well as otolith growth in white sea bass (Checkley et al. 2009) and tissue damage in larval Atlantic cod (Frommel et al. 2012). The results of the current study further support the hypothesis that predicted future levels of atmospheric CO2, levels that are estimated to occur within the next 100 years will pose physiological challenges to marine teleosts.

Exposure to 24-h hypercapnia at levels of 750-, 1,000- and 1,900 μatm CO2 resulted in significantly increased plasma HCO3 − and PCO2. It seems likely that the increase in HCO3 − observed at 750 μatm is the result of blood non-HCO3 − buffering, assuming a non-HCO3 − buffer value of 8–10 as is typical of teleosts (Tufts and Perry 1998). However, the findings at 1,000 and 1,900 μatm are consistent with the long-established response to hypercapnia of respiratory acidosis followed by metabolic compensation owing to raised plasma HCO3 − (Lloyd and White 1967). This pattern is further supported by time trial experiments at both 1,000 and 1,900 μatm CO2, where acidosis occurred in both treatments within 15 min followed by progressively increased HCO3 − concentrations until the acidosis was alleviated. Interestingly, while blood acid–base chemistry parameters remained stable at 1,900 μatm after compensation (after 8 h), there was more substantial variation in the 1,000-μatm treatment. More specifically, plasma HCO3 − peaked at 8 h and subsequently fell at 24 and 48 h of exposure; plasma pH was significantly depressed at 24 h. This suggests an incidence of secondary respiratory acidosis at these time points, likely indicating that after initial compensation toadfish stopped retaining HCO3 − until plasma pH again began to decrease. Of greater interest, however, is that low levels of hypercapnia affect blood acid–base chemistry in an apparent dose–response pattern, whereby higher levels of hypercapnia resulted in larger changes in plasma HCO3 − and PCO2 (Fig. 1). This pattern has been previously documented for higher levels of hypercapnia in a number of species (Baker et al. 2009; Brauner et al. 2004; McKenzie et al. 2002, 2003; Perry et al. 2010), but this is the first report of such low levels of hypercapnia having similar effects.

Exposure to 1,900 μatm CO2 also resulted in elevated white muscle intracellular pH (pHi). Similar results for muscle pHi have been observed for higher levels of hypercapnia in heart, brain and liver of the white sturgeon (Baker et al. 2009), although not in the white muscle. The increased muscle pHi likely stems from metabolic compensation to an initial acidosis similar to that observed in the plasma. In fact, 48 h of hypercapnia (1.5 kPa; ≈ 15,000 μatm) in white sturgeon increased white muscle intracellular HCO3 − threefold (Baker et al. 2009), while that of 24 h resulted in a twofold HCO3 − increase in the white muscle of the armored catfish (Brauner et al. 2004); pHi was unaffected in both species. It is somewhat surprising that muscle pHi was affected at such low levels of hypercapnia; however, because pHi is typically closer to the CO2 equilibrium constant (pKI) than plasma pH, a small increase in HCO3 − will have a greater impact on muscle pHi. Conversely, a smaller intracellular PCO2 increase will also have a more profound impact. The increased muscle pHi may reflect active regulation by the muscle via HCO3 − uptake from the extracellular fluid in response to intracellular acidosis, or it could also reflect passive HCO3 − uptake facilitated by the greater extracellular concentrations resulting from extracellular pH compensation.

A typical response to plasma acidosis is the net efflux of acid equivalents into the environment (Claiborne et al. 2002). However, no change in branchial titratable acid excretion, ammonia excretion or net acid excretion was observed over a 72-h exposure to 1,900 μatm CO2. Although the current study took effort to eliminate the intestinal contributions of acid–base balance (reviewed by Taylor et al. 2011), no effort was made to prevent renal contributions. However, gulf toadfish are aglomerular, so renal contributions to acid–base balance are likely negligible, as previously demonstrated for hypersalinity-induced acidosis (Genz et al. 2011). As such these results are likely representative of branchial flux. It may seem surprising that no net efflux of acid equivalents was detected given the clear compensatory responses described above; however, the variability in these measurements likely made detecting small differences difficult in the non-paired experiments.

As described earlier, marine teleosts are known to compensate for blood acidosis by excreting protons through the combined action of apical NHE, basolateral NBC and intracellular CA found in the mitochondrial rich cells of the branchial epithelium (Claiborne et al. 2002; Perry and Gilmour 2006). Interestingly, exposure to 1,900 μatm CO2 in nominally HCO3 − free seawater (182 μM) prevented metabolic compensation by toadfish, as shown by the persistent acidosis and unchanged plasma HCO3 − in combination with a significantly increased PCO2 (Fig. 4). While the increased PCO2 resulting from 1,900 μatm exposure is not equivalent between the seawater and HCO3 − free treatments, the HCO3 − free blood parameters are qualitatively similar to those described for the pre-compensation time points (15 min) of the time trial experiment (Fig. 3). It seems likely that the higher PCO2 in control seawater results from an inability to excrete the CO2 that is generated during pH compensation after HCO3 − uptake. Overall, these results suggest that toadfish, in contrast to hypotheses, rely on environmental HCO3 − uptake to drive pH compensation.

It is not easy to reconcile HCO3 − uptake from seawater with current models of seawater branchial ion transport (Fig. 7). To further examine this phenomenon the effects of low-level hypercapnia on the mRNA expression of a suite of HCO3 − transporters in the gill were examined. While time points for transcriptional studies can often be difficult to identify, studies of osmoregulatory and acid–base disturbances have routinely reported transcriptional responses at 24 h post-exposure (e.g. Gilmour et al. 2011; Ivanis et al. 2008). The expression of cytoplasmic CA, the cytoplasmic isoform found in the gills of teleosts (Esbaugh et al. 2005; Sattin et al. 2010), was also examined. Both cytoplasmic CA and slc4a2 were slightly but significantly down regulated in response to 1,900 μatm hypercapnia (Fig. 5a). The slc4a2 anion exchanger is an electroneutral transporter, presumably in the apical membrane that would only function in the Cl− uptake/HCO3 − excretion direction in seawater (Romero et al. 2004), so down regulation in response to an acidosis is obviously beneficial. The role of branchial CA is to hydrate CO2 to provide H+ and HCO3 − for transport processes (Gilmour and Perry 2009), which in the case of an acidosis involves excreting H+ apically and reabsorbing HCO3 − across the basolateral membrane. Interestingly, down regulation of CA supports the observation that compensation is due to HCO3 − uptake from the environment, since CA activity would have no benefit to such a pathway and may only interfere with the transcellular movement of HCO3 − from the environment to the plasma.

Fig. 7
figure 7

Mechanistic model of branchial ion transport in gulf toadfish, Opsanus beta. During an acidosis both intracellular carbonic anhydrase (CA) and slc4a2 expression are down regulated, and Na+ K+ ATPase (NKA) activity is reduced. Decreased basolateral NKA activity could act to increase cytoplasmic Na+ to aid in HCO3 − uptake via NBC and reduce H+ uptake through NHE1. Downregulation of slc4a2 would prevent HCO3 − loss, while down regulated CA would prevent CO2 trapping supporting the finding that compensation is the result of environmental HCO3 − uptake. Gray circles are ATPases, black circles are electroneutral transporters and white circles are electrogenic transporters

Full size image

No significant effects of 1,900 μatm CO2 were observed on gene expression of the available branchial electrogenic HCO3 − transporters (Fig. 5a); however, examination of the enzyme activity of NKA revealed a significant decrease after 24 h of exposure. This reduction returned to control levels by 72 h of exposure, and a similar but not significant pattern was observed for V-type ATPase activity. Such decreases in NKA and V-type ATPase were observed for mRNA expression in Atlantic salmon exposed to 1 and 3 days of 20 matm CO2 (Seidelin et al. 2001), and exposure to 10 matm CO2 also resulted in decreased expression of the a1a subunit of NKA in eelpout (Deigweiher et al. 2008). Although these results do not give clear indication of the mechanisms of HCO3 − uptake, they do provide insight into basolateral aspects of transport. It is relatively well established that basolateral HCO3 − transport occurs via nbc (Marshall 2002; Perry et al. 2003; Perry and Gilmour 2006; Sussman et al. 2009), where NHE1 is also localized to move protons into the plasma. It seems likely that lower NKA activity at the basolateral membrane could lessen the Na+ gradient by elevating cytoplasmic Na+ levels. This would reduce proton movement back into the plasma via NHE1 and aid the gradient for HCO3 − transport via nbc. It is important to note that lowered NKA activity could have detrimental effects on Na+ efflux (Marshall 2002; Marshall and Grosell 2006). This trade off may explain why NKA activity returns to control levels by 72 h after the onset of exposure and may also suggest that an alternate pathway—such as the NHE pathway—could become more dominant during chronic exposures. Although it cannot be excluded that the time points chosen failed to capture early and transient expression changes of some of the selected genes, changes in mRNA expression in teleost gill and intestinal tissue following environmental manipulations are often reported to occur at the time points chosen in the present study (e.g. Gilmour et al. 2011; Guffey et al. 2011; Ivanis et al. 2008; Sattin et al. 2010; Taylor et al. 2010).

Unlike previous studies examining hypercapnia the levels used in the current study largely represent outward CO2 diffusion gradients (Table 2). On this basis the current findings are surprising and suggest that CO2 diffusion is limited under these circumstances owing to the decreased partial pressure gradients. It is important to remember that even under normal conditions CO2 transport in fish is constrained by an apparent diffusion limitation, which is imposed by access to red blood cell CA via AE1 anion exchange (Esbaugh and Tufts 2006; Perry 1986; Perry et al. 2009; Tufts and Perry 1998). This is simply because the vast majority of the blood total CO2 load exists as plasma HCO3 −, which must be transported into the red blood cell through AE1 where it is hydrated by CA to form CO2 before subsequently diffusing across the gill (see reviews Perry et al. 2009; Tufts and Perry 1998). A number of studies have previously shown that this apparent diffusion limitation caused by chemical equilibrium constraints during capillary transit can be alleviated by CA infusion into the blood stream (Desforges et al. 2001, 2002; Julio et al. 2000), thereby removing the rate limiting step to chemical equilibration.

Table 2 The partial pressure of carbon dioxide (PCO2) in seawater equilibrated to various environmental carbon dioxide levels and mean pre-exposed Opsanus beta plasma
Full size table

We sought to investigate whether the respiratory acidosis caused by low-level hypercapnia was caused by a true diffusion limitation, or an apparent diffusion limitation. In other words, would decreasing the chemical equilibrium constraints during capillary transit prevent the respiratory acidosis. To investigate this, toadfish were infused with 5 mg/kg of CA via a caudal artery catheter and exposed to either 1,900 μatm CO2 (Fig. 6a) or 1,000 μatm CO2. We predicted that CA infusion would prevent the respiratory acidosis at 1,000 μatm due to the outward diffusion gradient of CO2, while having little to no effect during exposure to 1,900 μatm. Previous experiments have shown that 5–10 mg/kg doses of bovine CA cause respiratory alkalosis within 1 h post infusion (Desforges et al. 2001; Wood and Munger 1994). As predicted, the infusion of CA had no effect relative to controls on the blood acid–base parameters during exposure to 1,900 μatm CO2. In contrast, exposure to 1,000 μatm CO2 (Fig. 6b) in combination with CA infusion resulted in a significantly different response in both plasma pH and PCO2. This response, as predicted, is consistent with CA infusion ameliorating the chemical equilibrium constraints that limit CO2 diffusion. The observed respiratory acidosis is therefore the net result of a lower outward CO2 diffusion gradient combined with chemical equilibrium constraints that limit the mobilization of plasma HCO3 − during capillary transit. This is similar to the results of Julio et al. (2000), which showed that increased arterial PCO2 stemming from reduced gill surface area was corrected by CA infusion. To our knowledge this is the first study to show that CO2 chemical equilibrium constraints can impact excretion when confronted with only marginally less favorable diffusion gradients.

It is noteworthy that plasma HCO3 − also increased with CA infusion, suggesting some degree of metabolic compensation. This is corroborated by the increased pH, since a steady PCO2 would result in constant pH provided HCO3 − was also stable. These results may be indicative of physiological sensing of elevated PCO2 by externally oriented chemoreceptors, which could ultimately initiate HCO3 − uptake pathways. Similar findings have been reported for the cardiovascular effects of hypercapnia on rainbow trout (Perry and McKendry 2001); however, to our knowledge the connection between CO2 sensing and cellular acid–base compensation mechanisms has yet to be made.

Significance

Although the idea of hypercapnia as a physiological stress is not new, the current results are the first to demonstrate significant impacts on acid–base and respiratory physiology of low-level hypercapnia, levels currently found regionally and are predicted globally within 50 years. However, the potential concern to this group of animals is not likely to stem directly from these systems. Instead, the potential concerns lie in the downstream impacts of chronic exposure, most notably the energetic cost and tradeoffs associated with defending acid–base status in the presence of elevated CO2. For example, red blood cell and tissue pHi must be protected to prevent potential effects on oxygen uptake and other biochemical pathways. Acid–base disturbances can also impact cardiorespiratory variables such as ventilation rate, blood pressure and heart rate (Gilmour and Perry 2006). Furthermore, because otolith growth is proportional to HCO3 − concentrations of the medium surrounding the saccular epithelium (Tohse and Mugiya 2001), it seems likely that elevated plasma HCO3 − stemming from pH regulation may account for the otherwise counter-intuitive observations of increased otolith growth in white sea bass at 1,000 and 2,500 μatm (Checkley et al. 2009). Furthermore, chronically elevated plasma HCO3 − is likely to dramatically affect gastrointestinal HCO3 − excretion, as well as Cl−, Na+ and water uptake owing to the kinetics of the rate limiting basolateral nbc (Taylor et al. 2010)—effects that are currently under investigation. Admittedly, over time fish may adapt to a low-level hypercapnia environment and therefore not rely on acid–base compensation processes. Such adaptation could involve morphological changes in the gill to alter CO2 permeability, or as theoretically demonstrated here, fish may alter the CO2 chemical equilibrium constraints in the blood that lead to an acidosis. Such adjustments could include increased haematocrit, increased red blood cell anion exchange capacity or decreased branchial diffusion distance. Clearly, the physiological implications of imminent levels of hypercapnia require further study.