A product of its environment: the epaulette shark (Hemiscyllium ocellatum) exhibits physiological tolerance to elevated environmental CO₂

Ocean acidification is predicted to affect the performance of marine species, but little is known about the effects on sharks. We found that long-term exposure to elevated CO₂ did not affect the epaulette shark, possibly because it experiences fluctuating environmental conditions in its shallow coral reef habitat.

Abstract

Ocean acidification, resulting from increasing anthropogenic CO₂ emissions, is predicted to affect the physiological performance of many marine species. Recent studies have shown substantial reductions in aerobic performance in some teleost fish species, but no change or even enhanced performance in others. Notably lacking, however, are studies on the effects of near-future CO₂ conditions on larger meso and apex predators, such as elasmobranchs. The epaulette shark (Hemiscyllium ocellatum) lives on shallow coral reef flats and in lagoons, where it may frequently encounter short-term periods of environmental hypoxia and elevated CO₂, especially during nocturnal low tides. Indeed, H. ocellatum is remarkably tolerant to short periods (hours) of hypoxia, and possibly hypercapnia, but nothing is known about its response to prolonged exposure. We exposed H. ocellatum individuals to control (390 µatm) or one of two near-future CO₂ treatments (600 or 880 µatm) for a minimum of 60 days and then measured key aspects of their respiratory physiology, namely the resting oxygen consumption rate, which is used to estimate resting metabolic rate, and critical oxygen tension, a proxy for hypoxia sensitivity. Neither of these respiratory attributes was affected by the long-term exposure to elevated CO₂. Furthermore, there was no change in citrate synthase activity, a cellular indicator of aerobic energy production. Plasma bicarbonate concentrations were significantly elevated in sharks exposed to 600 and 880 µatm CO₂ treatments, indicating that acidosis was probably prevented by regulatory changes in acid–base relevant ions. Epaulette sharks may therefore possess adaptations that confer tolerance to CO₂ levels projected to occur in the ocean by the end of this century. It remains uncertain whether other elasmobranchs, especially pelagic species that do not experience such diurnal fluctuations in their environment, will be equally tolerant.

Introduction

Anthropogenic CO₂ emissions have caused an increase in atmospheric CO₂ by almost 40% over the past 250 years (IPCC, 2013). The resulting rise from pre-industrialization levels (∼280 ppm) to 400 ppm in 2014 has occurred at a rate unprecedented for the past 800 000–1 000 000 years (Raven et al., 2005; Doney and Schimel, 2007; Lüthi et al., 2008). The oceans have absorbed more than 30% of the additional CO₂ released by human activities, thus tempering the atmospheric rise in CO₂ (Sabine et al., 2004; Sabine and Feely, 2007). However, the resulting rise in seawater CO₂ partial pressure (PCO₂) and the associated reduction in pH, called ocean acidification, is a significant threat to marine organisms and ecosystems (Hoegh-Guldberg et al., 2007; Fabry et al., 2008).

The reduced carbonate saturation state that accompanies lower seawater pH affects the ability of calcifying marine organisms to form carbonate shells and skeletons (Orr et al., 2005; Doney et al., 2009), but rising oceanic CO₂ may also impact the respiratory physiology of many water-breathing organisms. Acid–base disturbances related to elevated environmental CO₂ can reduce oxygen uptake and delivery, which could directly impact metabolic performance. Reductions in an organism's scope for aerobic metabolic performance can result in less energy being available for crucial life-history processes, such as growth and reproduction (Pörtner, 2008; Pörtner and Farrell, 2008). For instance, Humboldt squid (Dosidicus gigas) exhibit a 30% reduction in resting metabolic rate and a 45% decrease in activity upon exposure to projected near-future CO₂ levels, owing to an impaired oxygen transport system, which would be predicted to reduce overall performance and compress their habitable depth range (Rosa and Seibel, 2008). In contrast, teleost fishes are expected to be physiologically well equipped to compensate pH and ion disturbances caused by high CO₂ (Ishimatsu et al., 2008; Brauner and Baker, 2009). Nevertheless, interspecific variation is evident in the physiological responses of teleost fish to elevated CO₂; for example, some fishes exhibit no change in aerobic scope in high-CO₂ environments (Ishimatsu et al., 2008; Melzner et al., 2009a; Couturier et al., 2013), whereas others reduce aerobic scope (Munday et al., 2009) and some even increase aerobic scope (Couturier et al., 2013; Rummer et al., 2013a) when exposed to near-future CO₂ levels. Consequently, the effects of ocean acidification on a broad range of species, including vulnerable and tolerant species, should be investigated in order to identify traits that will be important for individual performance and success in near-future oceans and predict changes in community structure (Melzner et al., 2009b).

In contrast to the growing body of knowledge about the effects of ocean acidification on teleost fishes, little is known about the impacts of rising levels of oceanic CO₂ on elasmobranchs. Elasmobranchs buffer a pH disturbance, such as that associated with exposure to high CO₂, in a manner similar to teleosts. Bicarbonate is accumulated in the blood, but in addition, elasmobranchs may also increase branchial ammonia excretion rates to ameliorate the acidosis further (Evans, 1982; King and Goldstein, 1983; Claiborne and Evans, 1992; Brauner and Baker, 2009; Tresguerres et al., 2010). The haemoglobin of elasmobranchs also has a much higher buffering capacity compared with that of most teleosts, and thus, O₂ transport and aerobic performance may be less sensitive to pH disturbances (Berenbrink et al., 2005). Yet, it is thought that the resilience of elasmobranchs to acid–base disturbances is related largely to their sophisticated acid excretion processes at the gill (Wood et al., 1995). If elasmobranchs are notably tolerant to near-future CO₂ conditions, this could potentially increase predation pressure and alter species compositions of marine environments.

The epaulette shark (Hemiscyllium ocellatum) exhibits exceptionally high tolerance to the severe hypoxia (low oxygen) that it routinely experiences while inhabiting shallow coral reef flats (Routley et al., 2002; Nilsson and Renshaw, 2004), and thus, it may not be surprising if this species is also tolerant to near-future CO₂. However, acute responses may differ dramatically from the responses to long-term exposure; studies on H. ocellatum in response to anoxia or hypoxia have been following only minutes to hours of exposure (Wise et al., 1998; Renshaw et al., 2002; Routley et al., 2002; Chapman and Renshaw, 2009; Dowd et al., 2010; Speers-Roesch et al., 2012a). No study, to date, has examined how H. ocellatum responds to prolonged exposure to elevated CO₂. Given that increased uptake of CO₂ by the ocean will affect both the average CO₂ levels and the magnitude of extreme CO₂ levels (Shaw et al., 2013), it is important to consider longer-term responses to elevated CO₂ beyond those that would be experienced on a diurnal basis (e.g. hours; Ohde and van Woesik, 1999; Compagno, 2002; Last and Stevens, 2009; Shaw et al., 2013). Thus, both the physiological sensitivity of the organism and the variations it may already be experiencing in its habitat are important when considering which species will exhibit positive or negative responses to rising ocean CO₂ levels. However, it is also important to consider the relationship between environmental cues and other traits, such as behaviour—which is especially relevant to species like H. ocellatum—when considering the importance of phenotypic plasticity, because this could influence selection over the longer term (Marais and Chown, 2008).

We exposed H. ocellatum to near-future CO₂ conditions for a minimum of 60 days and measured resting oxygen consumption rates and critical oxygen tensions as proxies for resting metabolic rate and sensitivity to hypoxia, respectively. In addition to whole-organism responses, we also measured or calculated haematological and tissue parameters, including plasma ionic (HCO₃⁻, Cl⁻, Na⁺ and K⁺) and urea concentrations, haemoglobin (Hb), mean cell haemoglobin concentration (MCHC), haematocrit (Hct), spleen–somatic index (SSI) and citrate synthase activity in heart, brain and red muscle. The aim was to provide insight into the physiological parameters that may underpin changes in metabolic performance and sensitivity to hypoxia in this species. We hypothesized that H. ocellatum can physiologically tolerate elevated CO₂ because it routinely experiences daily reductions in environmental O₂ (Routley et al., 2002; Nilsson and Renshaw, 2004) and probably elevations in CO₂. However, if CO₂ tolerance is related to the diurnal patterns this species already experiences in their natural habitat, prolonged exposure (60 days) to elevated CO₂ may negatively affect metabolic rate and hypoxia tolerance.

Materials and methods

Experimental animals

Epaulette sharks (Hemiscyllium ocelatum) were collected from the northern Great Barrier Reef by Northern Barrier and Cairns Marine (Cairns, Queensland, Australia) and transported to James Cook University (JCU). Five to six individuals were placed in each of six 700 l tanks in a recirculating seawater system. Individuals were measured [standard length, 33.38 ± 7.29 cm (mean ± SD)] to ensure an equal distribution of sizes among tanks. Unique fin clips along the margins of pectoral, pelvic and dorsal fins were used for individual identification. Shelter was provided in the form of PVC pipe sections placed within each tank. Food was provided once every 24 h and consisted of 4% of shark biomass per tank in raw prawn meat. There was no indication that any individuals or treatment groups were eating less than this amount throughout the duration of the study. Sharks were acclimated to laboratory conditions for at least 4 weeks prior to commencing CO₂ treatments.

Experimental CO₂ conditions

The experimental system comprised three 8000 l recirculating seawater systems, each set to simulate one of the following three CO₂ treatments: control (∼390 µatm); medium (∼600 µatm); and high (∼880 µatm). Carbon dioxide levels were achieved and maintained by CO₂ infusion of seawater in 3000 l sumps attached to each recirculating seawater system. The pH_NBS (National Bureau of Standards scale) levels were set to match target CO₂ concentrations and maintained using a CO₂-infusing system (Aqua Medic GmbH, Bissendorf, Germany). If the pH rose above the set point, an electronic solenoid initiated the system to deliver a steady stream of CO₂ into a diffuser within the corresponding sump. Carbon dioxide-equilibrated seawater from each system was delivered to two replicate 700 l tanks (∼25 l min⁻¹) per treatment. Each tank contained five or six sharks, as described above. This central approach of pH manipulation allowed for stability in seawater pH and PCO₂ within the holding tanks. Tanks were covered with transparent plastic sheeting to reduce CO₂ loss to the atmosphere.

The pH_NBS was measured daily (Hach, model #HQ40d) in each tank to ensure that it remained within ±0.05 of desired levels. Temperatures were also measured daily and maintained at 28.5°C by automated heater/chillers attached to each seawater system. Salinity and alkalinity were measured on a weekly basis. Total alkalinity (TA) was estimated using Gran titrations and certified reference materials (Dr A. G. Dickson, Scripps Institution of Oceanography). Average seawater PCO₂ was calculated using these parameters in CO2SYS (Pierrot et al., 2006) and using constants from Dickson and Millero (1987) (Table 1).

Table 1:

Mean values for PCO₂, pH, total alkalinity, salinity and temperature over the course of the CO₂ exposure period

Treatment	Tank number	PCO2 (μatm)	pH	Total alkalinity (μmol kg−1)	Salinity (ppt)	Temperature (°C)
Control	1	397 ± 6.5	8.16 ± 0.006	2145 ± 4.7	35.6 ± 0.07	28.6 ± 0.05
Control	2	384 ± 6.8	8.18 ± 0.006	2145 ± 4.7	35.6 ± 0.07	28.4 ± 0.04
Medium	1	614 ± 16.6	8.00 ± 0.009	2095 ± 5.1	35.9 ± 0.07	28.7 ± 0.05
Medium	2	608 ± 16.5	8.00 ± 0.009	2095 ± 5.1	35.9 ± 0.07	28.6 ± 0.05
High	1	876 ± 14.6	7.86 ± 0.006	2079 ± 5.3	36.0 ± 0.03	28.7 ± 0.03
High	2	861 ± 14.4	7.87 ± 0.006	2079 ± 5.3	36.0 ± 0.03	28.7 ± 0.04

Total alkalinity was measured weekly for each treatment condition, and temperature was measured daily for each tank within each treatment. Means were calculated for each treatment over the entire experimental period and are given for each holding tank, ±SEM. Abbreviation: PCO₂, seawater carbon dioxide partial pressure.

Sharks were introduced to the CO₂ treatments following 30 days acclimation to laboratory holding conditions and were then maintained in their respective CO₂ treatment conditions for a minimum of 60 days prior to physiological experimentation.

Experimental protocol

Resting oxygen consumption rates

Resting O₂ consumption rates (ṀO_2Rest) were determined for sharks following 60–68 days of exposure to control (n = 10), medium (n = 12) and high (n = 11) CO₂ conditions and a 48 h fasting period using an intermittent-flow respirometry system with purpose-built respirometry chambers. Animals were transferred individually into the cylindrical 11 or 15 l respirometry chambers (depending on animal body size) submerged in a temperature-controlled aquarium (28.5°C) within each animal's respective experimental CO₂ conditions and habituated to the chamber for 12 h before oxygen consumption measurements commenced. Submersible pumps were fitted to each chamber to supply a continuous water flow (1300 l h⁻¹; WEIPRO WH-2000; Yongcheng Aquarium Co., Ltd, Guangdong, China) from the surrounding water bath through the chambers. During respirometry trials, a digital relay timer (MFRT-1 Multi Function Recycling Timer; Xiamen SUPERPRO Technology Co., Ltd, Xiamen, Fujian, China) was used to stop water flow for 15 or 20 min and then resume flushing for 15 min over a total period of 12 h. The intervals of interrupted water flow were short enough to ensure that oxygen within the chambers did not fall below 80% saturation at any time, while flush periods were long enough to eliminate accumulation of metabolic CO₂ and allow water oxygen levels to return to 100% saturation (Steffensen et al., 1984; Steffensen, 1989). A second pump (1300 l h⁻¹; WEIPRO WH-2000) was connected to each respirometry chamber to recirculate water continuously within the chamber, thus ensuring complete mixing and homogeneous water PO₂ (P_WO₂). Contactless spots (2 mm) with oxygen-sensitive REDFLASH dye were adhered to the inside of glass tubes connected to the recirculating pumps on each respirometer. These spots were then linked to a Firesting Optical Oxygen Meter (Pyro Science e. K., Aachen, Germany) via 5 m fibre-optic cables to record continuously (0.5 Hz) the temperature-compensated O₂ concentration (in milligrams per litre) of the water within each chamber over the 12 h period of time. The 0 and 100% oxygen levels of the Firesting oxygen meter were calibrated using 0 and 100% air-saturated seawater. At the end of each trial, the wet mass was taken for each shark [232.47 ± 117.98 g (mean ± SD)] prior to release back to experimental holding conditions.

Critical oxygen tension

Upon completion of ṀO_2Rest measurements, sharks were permitted to recover in their respective CO₂ treatment conditions for ∼3 weeks. Then, the same respirometers used to determine ṀO_2Rest were used to determine the critical oxygen tension (P_crit) for the same sharks exposed to control (n = 9), medium (n = 12) and high CO₂ (n = 10). By this point, sharks would have been exposed to their respective experimental conditions for 85–92 days. Prior to measurements, sharks were fasted for 48 h before being introduced to the cylindrical 11–15 l respirometry chambers. Then, the ṀO₂ of each animal was monitored for a minimum of 4 h using an intermittent flush cycle (15 min flush–15 min closed) so that stable ṀO_2Rest was achieved prior to commencing the hypoxia experiment. The respirometers were then sealed by turning off flush pumps and closing previously installed ball-valves downstream of the flush pumps. Oxygen levels in the chamber were monitored continuously (0.5 Hz) and allowed to decrease to at least 0.8 mg l⁻¹ to ensure that the critical oxygen tension for each individual was recorded (based on estimates from Routley et al., 2002). The changes in water pH and PCO₂ that occur when using closed respirometry for a short period of time have been shown previously to have no effect on P_crit in fish (Henriksson et al., 2008). After this oxygen concentration was achieved, the aforementioned flush cycle was reinstated such that O₂ levels within each respirometer could quickly return to normoxic conditions (100% air-saturated seawater).

Haematological and tissue analyses

Following P_crit measurements, animals were returned to their treatment tanks to recover for ∼1 week. After this time, blood was sampled from sharks exposed to control (n = 8), medium (n = 10) and high (n = 8) CO₂ conditions by inserting a 23 gauge needle posterior to the cloaca into the caudal vein and collecting the blood (<1% of body volume) into heparinized syringes. Animals were then euthanized by severing the spinal cord using the method described by Speers-Roesch et al. (2012b) so that tissues could be sampled. Whole blood [Hb] was determined using the HemoCue^® (Hb 201 System, Australia Pty Ltd) with 10 µl of whole blood and was reported as grams per 100 ml using a calibration curve according to Clark et al. (2008) corrected for tropical reef species by Rummer et al. (2013b). The Hct was determined by centrifuging 60 µl of whole blood in heparinized microcapillary tubes for 3 min at 17 000g and calculated as the ratio of packed red blood cells to total blood volume (as a percentage). Both [Hb] and Hct were used to calculate the MCHC. The spleen was dissected from each shark and weighed to the nearest 0.001 g. The SSI was calculated as the ratio of the spleen to body mass (as a percentage). Plasma was flash frozen immediately in liquid nitrogen and then stored at −80°C until analysis for [HCO₃⁻] via colorometrically linked enzyme assay and for [Na⁺], [K⁺], [Cl⁻] (1:1 dilution with deionized water) and [urea] (1:19 dilution with deionized water) via ion-specific electrodes (ISE; Beckman Coulter System AU480). Heart, brain and red muscle samples were also collected and frozen in liquid N₂ for citrate synthase enzyme analysis according to McClelland et al. (2005). Briefly, frozen tissues were homogenized in a standard buffer solution containing 5 mm EDTA, 0.1% Triton X-100, 0.2 mm dithiothreitol and 50 mm Hepes (adjusted to pH 7.4) and stored at −80°C. The citrate synthase assay buffer contained (mm): 20 Tris (pH 8.0), 0.1 5,5-dithiobis (2-nitrobenzoic acid) and 0.3 acetyl-CoA. The reaction was initiated by the addition of 0.5 mm oxaloacetate, and absorbance was measured for 5 min at 412 nm. Control samples were assayed without oxaloacetate to control for background hydrolase activity.

Calculations and statistical analyses

Raw text files created for the Firesting recordings were imported offline into LabChart version 6.1.3 (ADInstruments, Colorado Springs, CO, USA), which was used to analyse data. A modified version of equations from Bushnell et al. (1994) and Schurmann and Steffensen (1997) was used to calculate ṀO_2Rest (in milligams per kilogram per hour). To do this, the average of the shallowest 10% of slopes [change in O₂ concentration over a period of 15–20 min (in milligams of O₂ per litre per second) in between flushing cycles] was determined for each individual shark. From this, the appropriate proportion of background O₂ consumption, which was measured 2–3 h before and after each trial for each respirometer and assumed linear, was subtracted. This value was then multiplied by the volume of the respirometer (in litres; minus the volume of the fish), all of which was divided by the mass of the fish (in kilograms). Respirometers were cleaned daily to ensure that background (microbial) respiration did not exceed 10% of the ṀO_2Rest of the sharks. Means and SEM for ṀO_2Rest were calculated for each of the three CO₂ treatments.

A similar data extraction and calculation protocol was followed for determining P_crit. Again, ṀO_2Rest was calculated for each shark from the shallowest 10% of slopes that were recorded prior to sealing the respirometer. Then, the mean slope for every 5 min period of time while the respirometer was sealed was extracted (usually 20–30 slopes), and ṀO₂ values were calculated from those slopes. To determine P_crit, all ṀO₂ values were plotted against the oxygen concentration within the chamber for each shark. A horizontal (regression) line was fitted to the mean ṀO_2Rest prior to sealing the respirometer. Then, a linear regression was applied to all of the points that consecutively fell below ṀO_2Rest once the respirometer had been sealed. The point at which both regression lines intersected was reported as the critical oxygen tension or P_crit (in milligams of O₂ per litre) for that individual (Fig. 1; Ott et al., 1980; Nilsson et al., 2004; Collins et al., 2013). Means and SEM for P_crit were calculated for each CO₂ treatment.

Figure 1:

Representative trace illustrating the changes in oxygen consumption rate of an individual epaulette shark (Hemiscyllium ocellatum) as the oxygen concentration of the water decreased and the time over which this occurred. The parallel line represents the resting oxygen consumption rate. After the respirometry chamber was sealed, the oxygen consumption rate began to decrease below resting levels. The diagonal line is a trend line, with the intersection of both lines demarcating the critical oxygen tension (P_crit).

Nested ANOVA, with holding tanks nested within CO₂ treatments, was first used to test whether there was a significant effect of holding tank on mean ṀO_2Rest (in milligams per kilogram per hour) or mean P_crit (in milligams of O₂ per litre). As there was no significant effect of tanks on either parameter, data from the two tanks within treatments were pooled for further analyses. ANCOVA was used to compare ṀO_2Rest among the three CO₂ treatments, with standard length as a covariate. To compare P_crit among treatment groups, a robust regression analysis was performed with standard length as a covariate. Robust regression analysis was chosen over ANCOVA for P_crit analysis due to potential outliers that could otherwise be solely responsible for significant outcomes. The removal of such outliers was rejected owing to the relatively small sample size. Instead, robust regression weighs values differently based on their chance of being an outlier. Hence, the further away a single data point was from the mean, the less influential it became for the statistical outcome of the analysis. There was no interaction between the main effect (CO₂) and the covariate in either analysis; therefore, to increase statistical power, the analyses were run again without this term included. Standard length was not included in haematological and tissue data analyses because it had no significant effect on the outcomes. ANOVAs were then used together with Holm–Sidak post hoc tests to compare haematological and tissue parameters between animals acclimated to control, medium and high CO₂ conditions. Statistical significance was accepted when P < 0.05. All analyses were carried out using S-Plus (TIBCO Software Inc., Palo Alto, CA, USA).

Results

Resting oxygen consumption rates

There were no significant differences in ṀO_2Rest values between CO₂ treatment groups (F_2,28 = 0.578; P = 0.568). However, ṀO_2Rest depended on the standard length of the individuals, with larger animals having a higher ṀO_2Rest than smaller animals (F_2,28 = 6.70; P = 0.0151; Fig. 2A). Values for ṀO_2Rest ranged from 46.8 to 95.4 mg O₂ kg⁻¹ h⁻¹ with a mean of 65.2 ± 2.13 mg O₂ kg⁻¹ h ⁻¹ across all treatments.

Figure 2:

The relationship between resting oxygen consumption rate (A) and critical oxygen tension (P_crit; B) with the standard length of individual sharks from control, medium and high CO₂ treatment groups. Please refer to Materials and methods for further details regarding calculations.

Critical oxygen tension

The P_crit values did not differ significantly between CO₂ treatment groups (t_4,26 = − 0.170; P = 0.866). However, standard length had a significant effect on the P_crit of individuals (t_4,26 = 2.26; P = 0.0323; Fig. 2B), with larger animals reaching P_crit at a higher seawater O₂ concentration than smaller animals. The P_crit values ranged from 1.32 to 5.07 mg O₂ l⁻¹ with a mean of 2.51 ± 0.122 mg O₂ l⁻¹ across all treatment groups.

Haematology and tissue samples

No significant differences were detected in Hct between CO₂ treatment groups (F_2,22 = 0.214; P = 0.809; Fig. 3A). There was a significant increase in [Hb] between the control and the medium CO₂ treatment groups (F_2,23 = 3.447; P = 0.048; Fig. 3B), an elevation that was maintained with the high CO₂ treatment group for MCHC values (F_2,21 = 5.067; P = 0.0160; Fig. 3C). Although not significant, there was a trend toward decreased SSI with high CO₂ exposure (F_2,22 = 2.050; P = 0.153; Fig. 3D). There was a significant increase in plasma [HCO₃⁻] in both the medium and high CO₂ treatment groups (F_2,21 = 10.893; P < 0.001; Fig. 4A). However, there was no difference in plasma [Na⁺], [K⁺], [Cl⁻] or [urea] between control and CO₂ treatment groups ([Na⁺], F_2,21 = 1.543, P = 0.237, Fig. 4B; [K⁺], F_2,21 = 0.247, P = 0.783, Fig. 4C; [Cl⁻], F_2,21 = 1.697, P = 0.207, Fig. 4D; and [urea], F_2,21 = 2.907, P = 0.077, Fig. 4E). Citrate synthase activity did not change significantly between control and CO₂ treatment groups in red muscle (F_2,16 = 0.371; P = 0.696), heart (F_2,18 = 0.0238; P = 0.976) or brain (F_2,19 = 0.131; P = 0.878; Fig. 5).

Figure 3:

Changes in haematocrit (A), haemoglobin concentration (B), mean cell haemoglobin concentration (MCHC; C), and spleen–somatic index (D) after sharks were exposed to control, medium or high CO₂ for ∼90 days. Different letters within a panel demarcate significant differences between treatment groups, and statistical significance is noted in the top right corner of each panel. Abbreviation: NS, not significant.

Figure 4:

Changes in plasma parameters after sharks were exposed to control, medium or high CO₂ for ∼90 days. Different letters within a panel demarcate significant differences between treatment groups, and statistical significance is noted in the top right corner of each panel. Abbreviation: NS, not significant.

Figure 5:

Changes in red muscle, heart and brain citrate synthase enzyme activity after sharks were exposed to control, medium or high CO₂ for ∼90 days. Statistical significance is noted in the top right corner of each panel. Abbreviation: NS, not significant.

Discussion

Long-term exposure to near-future CO₂ conditions did not significantly affect metabolic performance or hypoxia sensitivity of epaulette sharks. In contrast, changes in [Hb] and MCHC were evident after ∼90 days of exposure to 600 µatm CO₂ levels, and plasma [HCO₃⁻] was elevated in both the moderate and high CO₂ treatment groups, suggesting that physiological adjustments were being made to cope with elevated CO₂ at the level of oxygen transport and ion regulation. However, there was no increase in metabolic capacity at the level of the mitochondria, as indicated by the lack of change in citrate synthase activity. Our findings suggest that, for this reef-inhabiting benthic elasmobranch, neither the energetic costs of basic maintenance nor sensitivity to hypoxia may be compromised in the elevated CO₂ conditions projected for the end of this century.

The compensatory mechanisms used by H. ocellatum to maintain resting metabolic rates in normoxic and hypoxic conditions after prolonged exposure to elevated CO₂ may be linked to maintaining oxygen uptake and delivery and ion regulation. Following ∼90 days of CO₂ exposure, epaulette sharks exhibited a significant increase in [Hb] and MCHC. Short-term changes in haematological parameters have been documented in teleosts and elasmobranchs following capture, cannulation and exercise (Soivio et al., 1973; Wood et al., 1977; Turner et al., 1983; Wells et al., 1986), upon acclimation to elevated temperature (adult horn sharks, Heterodontus francisci; Neale et al., 1977) and in response to anoxia (grey carpet shark, Chiloscyllium punctatum, and epaulette shark; Chapman and Renshaw, 2009). In teleosts, acute changes can be associated with adrenergic red blood cell (RBC) swelling (Caldwell et al., 2006), a mechanism in place to protect RBC pH and oxygen transport during stress, but not known to occur in elasmobranchs (Berenbrink et al., 2005). Both teleosts and elasmobranchs do, however, use their spleen to produce and store RBCs (Turner et al., 1983; Fänge and Nilsson, 1985; Lai et al., 2006) and can contract it to increase the proportion of RBCs in the circulation (Ken-Ichi, 1988; Lai et al., 2006), presumably to aid in oxygen transport (Jensen et al., 1992). We observed a decrease, although non-significant, in the SSI in sharks exposed to both medium and high CO₂, suggesting that splenic contraction(s) may have occurred at some point during the CO₂ exposure period. Periodic splenic contractions could also increase the proportion of immature RBCs in circulation, which could explain the slight increase in MCHC without significant changes in Hct. The temporal scale of splenic RBC release and subsequent increases in erythropoietin, the glycoprotein responsible for regulating RBC numbers, is well understood for teleosts (Lai et al., 2006) and could be similar in elasmobranchs exposed to elevated CO₂ over extended periods of time, which is worth further investigation.

Plasma [HCO₃⁻] was elevated in sharks upon 90 days of exposure to elevated CO₂, which indicates some level of long-term acid–base compensation. This finding is supported by studies by Deigweiher et al. (2008), in which acclimation to elevated CO₂ over 6 weeks in a marine teleost resulted in upregulation of Na⁺/HCO₃⁻ cotransporters (NBC1) and Na⁺–K⁺-ATPase at higher densities. Given the relationship between bicarbonate availability and synthesis of urea (the predominant osmolyte used by most elasmobranchs), acid–base compensatory mechanisms could have affected [urea] and therefore the efficiency of osmoregulatory pathways (Wood et al., 1995). As [urea] did not change with CO₂ exposure, this may not be problematic at the CO₂ levels used here and/or over the 90 day duration. The activity of citrate synthase, the first enzyme of the Krebs cycle located within the mitochondria, can be a good indicator of aerobic capacity. Unchanged citrate synthase activity after prolonged CO₂ exposure further suggests that there is no limitation at the level of aerobic energy production in any of the tested tissues (McClelland et al., 2005). Although there may have been no changes to aerobic capacity, changes may have been occurring in anaerobic pathways (e.g. activity of lactate dehydrogenase, the last enzyme of anaerobic glycolysis) to maintain energy production. This would be worthy of further investigation. As Esbaugh et al. (2012) suggest, species that are already adapted to low levels of CO₂ may no longer rely on traditional short-term acid–base compensation strategies but instead use morphological changes (e.g. gill permeability, diffusion distances) or alter chemical equilibrium constraints in the blood over longer periods to maintain oxygen transport.

While there were no changes in metabolic performance in the sharks upon long-term CO₂ exposure, there was an unexpected pattern of mass-specific metabolic rates, with larger sharks exhibiting higher mass-specific metabolic rates than smaller sharks. This contradicts the usual pattern exhibited by ectotherms, but may be related to their feeding patterns. For example, we examined sharks ranging in size from ∼20 to 50 cm. However, we used a set 48 h fasting period prior to determining oxygen consumption rates and prior to blood and tissue sampling because of their small size and benthic lifestyle and previous feeding patterns while in captivity. It could have been that 48 h was sufficient fasting time for the smaller animals but not for the larger animals of that size range (Wood et al., 2007). Therefore, the larger animals could have been exhibiting slightly increased metabolic rates due to specific dynamic action, which could also mask any acid–base processes occurring due to CO₂ exposure. The relationship between acid–base disturbances originating from feeding and those due to elevated water CO₂ has ecological relevance and should be investigated in future studies.

Environment and lifestyle play an important role in physiological tolerance to changing environmental conditions (Pörtner and Farrell, 2008), and this study confirms that H. ocellatum is no exception. It is already known that H. ocellatum exhibits the lowest value of P_crit shown for any elasmobranch tested to date, suggesting an exceptional tolerance to short-term hypoxia, which is unique among chondrichthyans (Wise et al., 1998; Routley et al., 2002). Hemiscyllium ocellatum occupies shallow reef platforms that are subject to dramatic diurnal fluctuations in environmental O₂ and CO₂ conditions (Routley et al., 2002; Diaz and Breitburg, 2009; Last and Stevens, 2009). During calm nights, the low O₂ tension encountered on coral reefs can drop below 10% air saturation (Routley et al., 2002), usually as a result of respiration by reef organisms and especially during nocturnal low tides. This can also result in elevations in PCO₂, which have been reported to exceed 1000 µatm on shallow reef flats at night (Ohde and van Woesik, 1999; Shaw et al., 2013). The CO₂ levels may even be higher in caves, reef crevices and restricted-flow habitats, which are used by H. ocellatum for shelter (Compagno, 2002; Last and Stevens, 2009). Indeed, diurnal or acute fluctuations in O₂ and CO₂ may play a role in signalling metabolism in species using such habitats. However, acute responses often differ dramatically from responses to prolonged exposure, and it is important to make this distinction. The increased uptake of CO₂ by the ocean will affect both the average CO₂ level and the magnitude of extreme CO₂ fluctuations (Ohde and van Woesik, 1999; Shaw et al., 2013). This makes our finding that H. ocellatum exhibited no change in metabolic performance, including sensitivity to hypoxia, after prolonged exposure to projected future CO₂ levels even more important.

Adaptation to life on shallow reef platforms and lagoons may be the key to species like H. ocellatum for maintaining performance in projected future CO₂ concentrations (Melzner et al., 2009b). While noteworthy, what was previously known about the physiological tolerance of the epaulette shark to challenging environmental conditions was related to acute exposure of minutes to hours. This is extremely relevant to a shelter-seeking, benthic, reef-dwelling species like the epaulette shark that would experience such conditions burrowing into coral caves to avoid predation or to exploit food sources, activities vital to biological fitness. Pelagic shark species, many of which function as apex predators in their respective environments (Last and Stevens, 2009), however, do not typically exhibit shelter-seeking behaviours in areas that would experience the routine fluctuations in water chemistry experienced by H. ocellatum and therefore may not tolerate prolonged periods of elevated CO₂. Given that increased uptake of CO₂ by the ocean may mean that the high CO₂ levels that the epaulette shark may already routinely experience could be the new average ocean CO₂ levels, some species may be able to tolerate future conditions better. Future studies should investigate the importance of fluctuating environmental conditions in shaping an organism's tolerance. Differential effects on functional groups could impact predator–prey dynamics, affect the population structure of elasmobranchs and other aquatic organisms inhabiting coral reefs and, ultimately, impact ecosystem health. Investigating both sensitive and tolerant species from an array of habitat types would help to tease apart the role of the environment from other factors, including evolutionary history and behaviour, all of which is important when considering conservation measures under future climate change scenarios.

Funding

This work was supported by funding from the School of Marine and Tropical Biology (D.D.U.H., P.L.M.); the School of Earth and Environmental Science (C.A.S., M.R.H.); AIMS@JCU (D.D.U.H., M.R.H.); and the Australian Research Council Centre of Excellence for Coral Reef Studies (J.L.R., P.L.M.).