Species-specific effects of near-future CO₂ on the respiratory performance of two tropical prey fish and their predator

Abstract

Ocean surface CO₂ levels are increasing in line with rising atmospheric CO₂ and could exceed 900 μatm by year 2100, with extremes above 2000 μatm in some coastal habitats. The imminent increase in ocean pCO₂ is predicted to have negative consequences for marine fishes, including reduced aerobic performance, but variability among species could be expected. Understanding interspecific responses to ocean acidification is important for predicting the consequences of ocean acidification on communities and ecosystems. In the present study, the effects of exposure to near-future seawater CO₂ (860 μatm) on resting (Ṁ O_2rest) and maximum (Ṁ O_2max) oxygen consumption rates were determined for three tropical coral reef fish species interlinked through predator-prey relationships: juvenile Pomacentrus moluccensis and P. amboinensis, and one of their predators: adult Pseudochromis fuscus. Contrary to predictions, one of the prey species, P. amboinensis, displayed a 28 – 39 % increase in Ṁ O_2max after both an acute and four-day exposure to near-future CO₂ seawater, while maintaining Ṁ O_2rest. By contrast, the same treatment had no significant effects on Ṁ O_2rest or Ṁ O_2max of the other two species. However, acute exposure of P. amboinensis to 1400 and 2400 μatm CO₂ resulted in Ṁ O_2max returning to control values. Overall, the findings suggest that: (1) the metabolic costs of living in a near-future CO₂ seawater environment were insignificant for the species examined at rest; (2) the ṀO_2max response of tropical reef species to near-future CO₂ seawater can be dependent on the severity of external hypercapnia; and (3) near-future ocean pCO₂ may not be detrimental to aerobic scope of all fish species and it may even augment aerobic scope of some species. The present results also highlight that close phylogenetic relatedness and living in the same environment, does not necessarily imply similar physiological responses to near-future CO₂.

1 Introduction

The average concentration of carbon dioxide (CO₂) in the atmosphere has increased from approximately 280 ppm in pre-industrial times (Barnola et al., 1987) to >400 ppm in 2013 (Dlugokencky and Tans, 2013) and is projected to exceed 900 ppm by the year 2100 if the current emissions trajectory is maintained (Meinshausen et al., 2011). Because atmospheric and ocean surface pCO₂ are in equilibrium, CO₂ in the ocean is also increasing at approximately the same rate as in the atmosphere (Doney, 2010). Moreover, due to the hydrolysis of CO₂ in seawater, ocean surface pH is 0.1 unit lower today than preindustrial values and is predicted to be a further 0.3–0.4 units lower by 2100, which translates to a 100–150% increase in [H⁺] (Solomon et al., 2007). Some coastal regions could experience changes to [H⁺] that are at least 2–3 times the global average due to CO₂ enhancement from eutrophication (Melzner et al., 2012) and amplification of natural CO₂ and pH variation (Shaw et al., 2013). Such changes in ocean chemistry are predicted to affect physiological functions of many marine organisms, with potentially far-reaching effects on marine diversity and ecosystem processes (Fabry et al., 2007; Gattuso and Hansson, 2011; Pörtner, 2008).

Ocean acidification has been hypothesized to have negative consequences for the performance of marine fishes, primarily through an effect of the capacity for oxygen supply and delivery (Pörtner et al., 2004). Aerobic scope, which represents the oxygen available for any activities beyond that required for basic maintenance (Fry and Hart, 1948; Fry, 1947, 1971), is expected to decline with increasing pCO₂ (Pörtner and Farrell, 2008). Reduced aerobic scope could affect individual fitness, since less energy can be devoted to digestion, growth and reproduction (Munday et al., 2009b; 2012). Reduced aerobic scope could also affect the outcome of key ecological interactions and ultimately the structure of ecological communities (Nilsson et al., 2009; Pörtner, 2008).

In accordance with these predictions, the aerobic capacity of two cardinalfish species from the Great Barrier Reef (GBR) was significantly reduced by exposure to CO₂-acidified water (Munday et al., 2009a). However, it may be expected that not all fish species of coral reef ecosystems will be similarly affected by CO₂-induced ocean acidification. Fishes are the most diverse group of vertebrates and the ontogenetic and lifestyle traits of some species could provide pre-adaptation to high ambient CO₂ (Ishimatsu et al., 2008; Melzner et al., 2009b). Indeed, the aerobic scope of some fish species appears to be unaffected by hypercapnia (Baker and Brauner, 2012; Ishimatsu et al., 2008; McKenzie et al., 2003). Interspecific differences in the response of aerobic scope to near-future seawater CO₂ could have important ecological ramifications, especially for species that interact through competitive or predator-prey relationships (Munday et al., 2012). Many studies implicate species interactions to be an important proximate cause of extinction due to climate change, particularly due to decreases in food availability (Cahill et al., 2012).

The objective of the present study was to assess species-specific effects of near-future seawater CO₂ on aerobic performance among tropical reef fish species involved in predator and prey relationships. The model predator species investigated was the brown dottyback (Pseudochromis fuscus Müller & Troschel 1849), a common meso-predator on the GBR. Two closely related damselfishes, the lemon damselfish (Pomacentrus moluccensis Bleeker 1853) and the Ambon damselfish (P. amboinensis Bleeker 1868) were chosen as the model prey species. P. fuscus is known to be a major predator of recently settled juveniles of these two damselfishes on the GBR (Holmes and McCormick, 2010). Respirometry was utilized to measure the effects of exposure to elevated CO₂ on resting (ṀO_2rest) and maximum (Ṁ O_2max) oxygen consumption rates. In fish, physiological alterations in response to elevated ambient CO₂ can occur within minutes (i.e., ventilatory responses; Gilmour and Perry, 2006), hours (i.e., blood acid-base regulation; Brauner and Baker, 2009; Heisler, 1993; Esbaugh et al., 2012) to days (i.e., neurological disruptions; Nilsson et al., 2012). In terms of neurological impairments, four days of high-CO₂ exposure has been shown to disrupt a number of sensory systems and alter the behaviour of reef fishes, including the three species examined in the present study (Cripps et al., 2011; Ferrari et al., 2011a; Ferrari et al., 2011b; Nilsson et al., 2012). Longer exposure to elevated CO₂ does not induce further behavioral effects (Munday et al., 2010). Therefore, in our first experiment, we measured oxygen consumption after exposing the three species to ambient or elevated CO₂ for four days, to enable direct comparisons with previous studies on coral reef fish. In a second experiment, we measured oxygen consumption of the damselfishes, following acute exposure to near-future seawater CO₂ to determine if exposure to elevated CO₂ induces an immediate effect on aerobic performance. In both experiments, the elevated CO₂ treatment (860 μatm) was selected to approximate the level predicted for the atmosphere and ocean surface in 2100 under the IPCC A2 emissions scenario (Meehl et al., 2007). Finally, a third experiment was conducted to understand the effects of more extreme fluctuations in seawater pCO₂ and consequently pH that may occur in some coastal habitats, including shallow coral reef flats (Shaw et al., 2013). Four groups of P. amboinensis were exposed to one of four pH levels spanning from the present-day control pH of 8.1 down to pH 7.5 at 0.2 unit increments. The desired pH level was obtained by adding increasing volumes of 100 mM hydrochloric acid to the seawater. The addition of a strong acid to a closed system, such as a closed respirometer, has similar consequences on pCO₂ and pH as equilibrating water of an open system with CO₂ gas (Gattuso and Lavigne, 2009). Corresponding pCO₂ levels were approximately 450 μatm at pH 8.1, 860 μatm at pH 7.9, 1400 μatm at pH 7.7 and 2400 μatm at pH 7.5.

2 Materials and Methods

2.1 Experimental fish

The experiments were conducted at Lizard Island Research Station (LIRS; 14°40′S, 145°28′E) between December and January (austral summer). Juvenile P. moluccensis (mean ±SD, 40.9 ±5.6 mg) and P. amboinensis (53.9 ±12.4 mg) were caught at night using light traps moored two meters below the surface and approximately 100 m off the reef (Meekan et al., 2001). In this location, the fish are trapped immediately before their arrival to the reef at the end of their planktonic larval stage (Meekan et al., 1993). Every morning, juveniles were collected from the traps and transferred to the laboratory where they were exposed to ambient or elevated CO₂ for four days (see section 2.2). Adult P. fuscus (4.51 ±0.78 g) were collected from shallow reefs (<6 m) in the Lizard Island lagoon using a hand-net after lightly anaesthetizing them with a mixture of clove oil, ethanol and seawater (Munday and Wilson, 1997). Captured fish were transported to the research station where they were maintained for two days prior to exposure to CO₂ treatments. Fish were maintained at ambient ocean temperatures, which ranged from 28.3 to 30.4°C (Table 1; 29.4 ±0.1°C) during the experimental period. Damselfishes were fed freshly hatched Artemia nauplii three times daily, and P. fuscus were fed twice daily to satiation with INVE Aquaculture Nutrition pellets. Feeding was discontinued 18 – 24 h prior to resting oxygen consumption measurements (see section 2.3.1). Animal care and experimental protocols complied with regulations at James Cook University and Lizard Island Research Station, and were approved by the James Cook University Ethic Committee (Approval # A1722). Fish were collected under permit G10/33239.1 from the GBR Marine Park Authority

Table 1.

Mean (±SE) seawater parameters in the experimental system. pCO₂ was estimated in the program CO2SYS from measured pH, salinity and total alkalinity (A_T) of water samples.

Treatment	pHNBS	Temperature (°C)	Salinity (ppt)	AT (μmol kg−1SW)	pCO2 (μatm)
Control	8.11–8.17	29.4 ±0.1	34.5	2272 ±13	451 ±15
High-CO2	7.90–7.92	29.4 ±0.1	34.5	2267 ±2	860 ±14

2.2 CO₂ exposure

Fish were exposed to either aerated control water (pCO₂ = 451 μatm) or 860 μatm CO₂ water (termed near-future seawater CO₂ in the present study) for four days (Table 1). Near-future seawater CO₂ concentrations were maintained by CO₂-dosing to a set pH_NBS (National Bureau of Standards) following standard techniques for ocean acidification research (Gattuso et al., 2010). Seawater was pumped from the ocean into two 60 l header tanks, one equilibrated with air (ambient control) and the other with CO₂ to achieve the pH expected to correspond to the ocean CO₂ concentration projected for 2100 (Meehl et al., 2007). The pH level was based upon preliminary observations of total alkalinity, salinity and temperature of seawater at Lizard Island. A pH-controller (Aqua Medic GmbH, Bissendorf, Germany) was attached to the CO₂-treated header tank to maintain pH at the desired level. A solenoid injected a slow stream of CO₂ into a submersible pump at the bottom of the header tank whenever the seawater pH rose above the set point. The pump ensured rapid dissolution of CO₂ into the seawater and also served as a vigorous stirrer. The pump in the control seawater header tank was injected with a slow stream of air. Seawater from each header tank was supplied at a rate of ca 500 ml min⁻¹ to four replicate 35 l aquaria for each species. pCO₂ in the aquaria was checked twice daily with a CO₂-permeable membrane connected to an infrared CO₂ probe (Vaisala GMP343, Vaisala, Helsinki, Finland) in a closed loop (Hari et al., 2008). Water samples were collected at the start, middle and end of the experiment in order to precisely determine pCO₂. Total alkalinity (A_T) of seawater was estimated by Gran titration (Gran, 1950; 1952) using certified reference material from Dr. A. G. Dickson (Scripps Institution of Oceanography), and average seawater pCO₂ was calculated in CO2SYS (http://cdiac.ornl.gov/oceans/co2rprt.html) from measured A_T and pH and using the constants of Mehrbach et al. (1973) refit by Dickson and Millero (1987).

2.3 Experimental set-up and protocol

Ṁ O_2rest and Ṁ O_2max were used as proxies for resting and maximum metabolic rates and were measured by respirometry as previously utilized for assessing the effects of climate change variables on other fish (see Ishimatsu et al., 2008 for review), especially tropical reef fish species (Gardiner et al., 2010; Munday et al., 2009a; Nilsson et al., 2007a; Nilsson et al., 2010; Nilsson et al., 2007b; Nilsson and Östlund-Nilsson, 2004). Respirometry chambers were immersed in temperature-controlled (29°C) aquaria continuously supplied with either air- or near-future CO₂-equilibrated seawater.

2.3.1 Resting oxygen consumption

Cylindrical 26.7-ml static respirometers were used for P. moluccensis and P. amboinensis juveniles. After four days of exposure to current day or near-future CO₂ seawater, one fish was transferred to each respirometry chamber. The chamber was left open and the fish left undisturbed to habituate to the chamber for 1 – 2 h whereupon the chamber was cautiously closed without disturbing the fish. Previous experiments have shown that habituation periods longer than 2 h in the chamber do not further reduce Ṁ O₂ (Nilsson et al., 2010; Nilsson and Östlund-Nilsson, 2004). All fishes included in this study settled down rapidly and remained virtually motionless during the measuring period. Once the chamber had been sealed, water oxygen concentration was recorded continuously with an oxygen probe (CellOx 325, WTW, Germany; calibrated daily) connected to an oxygen meter (OXI 340i, WTW, Germany). The oxygen probe was fitted with a magnetic propeller (BOD stirring accessory, WTW, Germany) set in motion with a magnetic stir plate situated outside the aquarium along the glass wall. The propeller ensured gentle water mixing inside the respirometer and water renewal along the O₂ probe membrane during the habituation and recording periods. The oxygen meters were connected to a data acquisition system (PowerLab 4/20, ADInstruments, Colorado Springs, USA). Ṁ O_2rest was calculated from the steady rate of oxygen consumption observed between 100 and 90% of air saturation. The decrease of water oxygen concentration was recorded until it reached approximately 10% of air saturation in order to calculate the critical oxygen concentration (O_2crit), which is the lowest O₂ concentration where the fish is still able to maintain Ṁ O_2rest. O_2crit was reached 2.5 – 3 h after the chamber was sealed. For each species, two parallel setups allowed for the simultaneous recording of one fish in high-CO₂ water and another fish in control conditions. Similar to a number of other prior experiments conducted to determine the O_2crit of a fish using this classic protocol, the measurement of O_2crit required the respirometer to remain closed until almost all O₂ was depleted from the system. Consequently, CO₂ concomitantly increased in the respirometer due to the respiration of the fish and the fish simultaneously experienced hypoxia and increasing hypercapnia during the O_2crit measurement. Assuming a respiratory quotient of between 0.7 and 1.0, the maximum build-up of CO₂ within the respirometer for each 10% fall in O₂ is estimated to be 400–600 μtam. However, most of the excreted CO₂ would convert rapidly to bicarbonate, resulting in the build-up of pCO₂ to be considerably less. Given that Ṁ O_2rest did not differ significantly between the control and near-future CO₂ exposed fish (see Fig. 1A), pCO₂ would have increased by a similar magnitude and in parallel for both treatment groups. However, the near-future CO₂ fish would have consistently experienced a greater level of hypercapnia (by 410 μatm) than the control fish at any point in time, including at O_2crit, due to the elevated pCO₂ of the water at the start of the experiment. For P. fuscus, Ṁ O_2rest was measured in 1615-ml intermittent-flow respirometers. Fish were first habituated to the chambers for 90 min. Preliminary experiments determined that 90 min was ample time for this species to ensure O₂ consumption rates had reached the lowest possible values under the experimental conditions. Beyond this time, O₂ consumption rates did not significantly vary. Submersible pumps supplied a water flow (150 l h⁻¹) from the aquaria through the chambers and after the habituation period, water flow to each chamber was stopped for 15 min every 30 min over a period of 90 min. The time the water flow was interrupted was short enough to ensure O₂ did not fall below 80% of air saturation. Water oxygen concentration (mg l⁻¹) was continuously recorded at a frequency of 1 Hz using oxygen-sensitive REDFLASH dye on contactless spots (2mm) adhered to the inside of each chamber and connected via fiber-optic cable to a Firesting Optical Oxygen Meter (Pyro Science e. K., Aachen, Germany). Data were analyzed using LabChart 6.1.3 (ADInstruments, Colorado Springs, USA).

Fig. 1.

Respiratory performance (mean ±SE) of P. moluccensis and P. amboinensis. A. Resting oxygen consumption rate (Ṁ O_2rest; mgO₂ kg⁻¹ h⁻¹), B. Critical oxygen concentration (O_2crit; % of air saturation), C. Maximum oxygen consumption rate (Ṁ O_2max; mgO₂ kg⁻¹ h⁻¹) after a 4 d exposure to control seawater (451 μatm;■) or near-future CO₂ seawater (860 μatm;■) and D. Ṁ O_2max after an acute exposure to near-future CO₂ seawater. Letters that differ indicate statistically significant differences (see text for p values). N numbers are indicated at the bottom of each bar.

2.3.2 Maximum rate of oxygen consumption

Ṁ O_2max was measured in custom made cylindrical swimming chambers previously utilized by a large number of studies for small tropical coral reef fish species (see Nilsson et al., 2007b for diagram and detailed description of the set-up; Gardiner et al., 2010; Nilsson et al., 2009, Munday et al., 2009a; Nilsson et al., 2007a; Nilsson et al., 2010; Nilsson and Östlund-Nilsson, 2004). The volume of the chambers ranged between 215 and 267 ml for P. moluccensis and P. amboinensis juveniles and was 1594 ml for P. fuscus. The decrease in [O₂] was measured with an O₂ electrode (WTW CellOx 325, as above). It has been suggested that the aerobically fuelled muscle mass in some fish is not large enough to force them to reach the maximum rate of oxygen uptake during maximal swimming performance (Goolish, 1991). Therefore, fish were fed ad libitum prior to estimating Ṁ O_2max, since it is likely that the combined oxygen needs of digestion and maximal swimming would be high enough to engage the full capacity of the respiratory system (Bennett and Hicks, 2001; Gardiner et al., 2010). Up to three P. amboinensis (to increase the total fish weight so that reliable recordings could be made), one P. moluccensis or one P. fuscus (same fish as for Ṁ O_2rest measurements) were placed in the swimming chamber. The water speed was regulated with a magnetic stirrer located beneath the chamber. As soon as the water was set in motion, the fishes started swimming against the current, apparently guided by landmarks provided by items such as the oxygen electrode and the edges of the surrounding aquarium. The speed was consistently increased to a point where it was assumed that the fishes swam at their maximum speed. The water speed corresponded to the point at which the fishes could just barely maintain a steady position in the chamber. At a slightly higher speed, the fishes were no longer able to maintain position for more than a few seconds and stopped swimming. Nilsson et al. (2007b) showed that water speeds of approximately 50 and 125 cm s⁻¹ could be achieved near the inner and outer wall of the chamber, respectively. The speeds are more than sufficient for each of the species examined to reach their maximal swimming speed. Pre-settlement larvae of both Pomacentrus species as well as P. fuscus, which are more efficient swimmers than the post-settlement larvae and adults studied here (Nilsson and al., 2007b) exhibit an average maximum sustained swimming speed ranging from less than 30 cm s⁻¹ to a maximum of 36 cm s⁻¹ (Fisher et al., 2005). The decrease in oxygen concentration was recorded at a frequency of 1 Hz in the chamber at the maximum swimming speed for up to 6 min, during which time oxygen concentration remained above 90% of air saturation.

First, Ṁ O_2max was measured in the three species that had been maintained in control or near-future CO₂ seawater for four days. Then, in order to assess an acute effect of near-future CO₂, Ṁ O_2max was measured in P. moluccensis and P. amboinensis maintained in control conditions for four days and acutely exposed to near-future CO₂ while in the respirometer. Finally, to further understand the effects of increased CO₂ and low pH in more extreme habitats, Ṁ O_2max of four groups of P. amboinensis was measured after acute exposure of the fish to pH levels of 8.1 (control, present day level, corresponding to approximately 450 μatm CO₂), 7.9 (corresponding to approximately 860 μatm CO₂), 7.7 (corresponding to approximately 1400 μatm CO₂) and 7.5 (corresponding to approximately 2400 μatm CO₂). The desired pH level was obtained by diluting increasing volumes of 100 mM hydrochloric acid in the seawater utilized in the experimental set-up. The water was prepared immediately prior to the experiment, and within 2 min, the fish was placed in the respirometry chamber, which was sealed, thereby preventing the water from equilibrating with the atmosphere. Ṁ O_2max was then measured as described above. Average seawater pCO₂ was calculated in CO2SYS (http://cdiac.ornl.gov/oceans/co2rprt.html) from measured pH and assuming the same A_T as in the control conditions. The addition of a strong acid to a closed system like a closed respirometer has rather similar consequences on water chemistry as equilibrating water of an open system with CO₂ gas, thereby allowing reasonable comparison between both techniques (Gattuso and Lavigne, 2009). However, without the addition of CO₃²⁻ or HCO₃⁻, the technique leads to a slightly lower A_T. For example, Gattuso et al. (2010) report that the A_T of seawater with a salinity of 35 ppt will decrease by 6% when pH is decreased from 8.1 to 7.8 via the addition of HCl in a closed system. This would reduce pCO₂ estimates in our system by 6%. Therefore, our estimates of pCO₂ from hydrochloric acid addition may be marginally higher than what was achieved in the respirometer prior to the start of the experiment and addition of respiratory CO₂ from the fish.

Because the measurements of Ṁ O_2max in the small damselfish (P. amboinensis) required the pooling of individuals (see above), Ṁ O_2rest and Ṁ O_2max had to be measured on different sets of fish. The procedure precluded the calculation of individual aerobic scope in damselfish. As for the measurement of Ṁ O_2rest, two parallel setups allowed for the simultaneous measurement of Ṁ O_2max of fish exposed to near-future or higher CO₂, and those exposed to control conditions.

2.4 Data analyses and statistics

Oxygen consumption (Ṁ O₂ in mgO₂ kg⁻¹ h⁻¹) was calculated for each fish or pool of fish using the following formula:

$$\stackrel{.}{M}{\mathrm{O}}_2=\mathrm{\Delta }[{\mathrm{O}}_2]\times \mathrm{\Delta }{t}^{-1}\times {\text{VOL}}_{\text{resp}}\times {\mathrm{M}}^{-1}$$

where Δ[O₂] is the decrease in water oxygen concentration (Δ mgO₂ l⁻₁), Δt the recording time (h), VOL_resp is the volume of the respirometer minus the volume of the fish (l), and M the mass of the fish (kg). Fish Ṁ O₂ was corrected for background (microbial) respiration measured after each recording. For P. moluccensis and P. amboinensis, Ṁ O_2rest was measured in a closed respirometer from Δ[O₂] data above 90% of air saturation. O_2crit was calculated for each damselfish as the concentration of O₂ at the intersection of the regression lines from the data recorded >80% of air saturation (well above O_2crit) and the data recorded <20% air saturation (well below O_2crit as evident from the break in the curve at O_2crit). For P. fuscus, three slopes (Δ[O₂] × Δt⁻¹) were averaged to calculate Ṁ O_2rest. Ṁ O_2max was calculated for all fish from within the first minute of recording. Unpaired t-tests were utilized to assess the effect of seawater pCO₂ on Ṁ O_2rest and Ṁ O_2max for each species, and the effect of seawater CO₂ concentration on O_2crit in P. moluccensis and P. amboinensis. In the experiment examining the effect of extreme pH on Ṁ O_2max in P. amboinensis, no significant differences were found between the three control groups (P = 0.236). Therefore, the control data were pooled. Then, a one-way ANOVA was used to test for the effect of pH on Ṁ O_2max. In all instances, P<0.05 was considered significant.

3 Results

Ṁ O_2rest of the two damselfish prey species was unaltered after exposure to near-future CO₂ for four days (Fig 1A). Similarly, four days of exposure to near-future CO₂ had no significant effect on O_2crit for either of the two damselfish prey species (Fig. 1B).

Ṁ O_2max of P. moluccensis was unaffected by four days exposure as well as acute exposure to near-future CO₂ (Fig. 1C and D). In contrast, Ṁ O_2max of P. amboinensis was significantly higher after four days (+39%; t_[10] = 4.665; p = 0.009) as well as acute (+ 28%; t_[10] = 2.597; p = 0.029) exposure to near-future CO₂ when compared to controls (Fig. 1C and D). Furthermore, the two CO₂ treatments resulted in a similar increase in Ṁ O_2max. Ṁ O_2max of P. amboinensis measured after four days exposure to near-future CO₂ was not significantly different than Ṁ O_2max measured after acute exposure to near-future CO₂ (P = 0.417). Similarly, Ṁ O_2max of the respective control groups did not differ significantly (P = 0.858).

Exposure to near-future CO₂ for four days had no effect on either Ṁ O_2rest or Ṁ O_2max of the predator, P. fuscus (Fig 2). Consequently, there was no significant effect of near-future CO₂ on net aerobic scope or factorial aerobic scope in this species.

Fig. 2.

Respiratory performance (mean ±SE) of P. fuscus. Resting oxygen consumption rate (Ṁ O_2rest; mgO₂ kg⁻¹ h⁻₁) (○) and maximum oxygen consumption rate (Ṁ O_2max; mgO₂ kg⁻¹ h⁻₁) (○)after a 4 d exp&sure to control seawater (451 μatm; n = 8) or near-future CO₂ seawater (860 μatm; n = 7).

When acutely exposed to pH 7.9 (approximately 860 μatm pCO₂), P. amboinensis exhibited a Ṁ O_2max that was significantly higher than that of the control fish, as well that of fish acutely exposed to pH 7.7 (approximately 1400 μatm pCO₂) or 7.5 (approximately 2400 μatm pCO₂; Fig. 3; P <0.0001). The increase in Ṁ O_2max (+ 32%) was quantitatively similar to the increased Ṁ O_2max displayed by P. amboinensis after the four-day and acute exposures to near-future CO₂. However, at pH 7.7 and 7.5, Ṁ O_2max of P. amboinensis was no longer elevated and was not significantly different from the control (Fig. 3).

Fig. 3.

Maximum oxygen consumption rate (Ṁ O_2max; mgO₂ kg⁻¹ h⁻¹; mean ±SE) of P. amboinensis at different levels of seawater pH. N numbers are indicated at the bottom of each bar.

4 Discussion

CO₂-driven ocean acidification has been predicted to have detrimental effects on marine organisms by reducing the scope for aerobic performance (Pörtner and Farrell, 2008; Pörtner and Knust, 2007; Seibel and Walsh, 2001). However, contrary to expectations, none of the three tropical reef fish examined here, P. moluccensis, P. amboinensis or P. fuscus, exhibited an elevated Ṁ O_2rest or reduced Ṁ O_2max when exposed to the average pCO₂ projected to occur in the ocean surface by the year 2100. Rather, Ṁ O_2rest of the two juvenile damselfish prey species (P. moluccensis and P. amboinensis) and the adult predator (P. fuscus) was maintained after a four-day exposure to near-future seawater CO₂. Early life stages of fish and other marine organisms are believed to be more sensitive to pH changes because of their high metabolic demand (Brown and Sadler, 1989; Pörtner et al., 2005). The unchanged Ṁ O_2rest of juvenile P. moluccensis and P. amboinensis after four days exposure to near-future CO₂ thus suggests that the hypothesized metabolic costs of living in a high CO₂ environment, namely altered acid-base balance, ionoregulation and cardiorespiratory function (Pörtner et al., 2004), were insignificant for these species at rest. Nevertheless, the unchanged Ṁ O_2rest of the species examined after four days exposure to near-future CO₂ does not preclude that physiological changes occurred, including compensatory ones. For example, gill ionoregulatory machinery is rapidly altered (within 8 h to 2 d) in response to hypercapnia exposure in the eelpout (Zoarces viviparous), without measurable effect on resting metabolic rate over the same time period (Deigweiher et al., 2008). Similarly, long-term (4–12 months) exposure to hypercapnia leads to upregulated gill Na⁺/K⁺-ATPase activity and protein expression in the Atlantic cod (Gadus morhua), but resting and low activity metabolic rates remain unaltered (Melzner et al., 2009a). Furthermore, the gilthead bream (Sparus auratus) exhibits a shift from aerobic to anaerobic metabolic pathways during hypercapnic exposure (Michaelidis et al., 2007). Whether similar compensatory physiological changes also occur in coral reef fish during exposure to near-future climate change CO₂ levels remains to be investigated.

Like observed for Ṁ O_2rest, Ṁ O_2max of P. moluccensis and P. fuscus was also unaffected by exposure to near-future CO₂. In contrast, Ṁ O_2max of P. amboinensis was higher under near-future CO₂ conditions than under control conditions. The lack of a detrimental effect of increased CO₂ on Ṁ O_2rest and Ṁ O_2max, as reported here for P. moluccensis and P. fuscus, is not unprecedented for fish (Ishimatsu et al., 2008; McKenzie et al., 2003; Melzner et al., 2009a). However, to the best of our knowledge, an augmented aerobic capacity of a juvenile marine teleost in response to elevated CO₂ has not been documented previously. The validity of the present results for P. amboinensis are supported by the findings that Ṁ O_2max was increased with elevated CO₂ in separate groups of fish that were exposed to near-future CO₂ using three different experimental techniques. The specific treatments consisted of a four-day exposure to near-future CO₂ maintained by CO₂-dosing to a set pH_NBS, an acute exposure to near-future CO₂ maintained by CO₂-dosing to a set pH_NBS and an acute exposure to near-future CO₂ obtained by the addition of strong acid into a closed system. Across the three methodologies, all 21 individuals exposed to near-future CO₂ and the resulting acidosis exhibited an increased ṀO_2max. Moreover, the magnitude of the increase in Ṁ O_2max was consistent among the different experimental protocols (+28–39%). Finally, Ṁ O_2max, Ṁ O_2rest and O_2crit of P. amboinensis in control conditions were comparable to data from a previous study using similar size fish (Nilsson et al., 2007b), indicating that the elevated Ṁ O_2max under near-future CO₂ did not arise from a comparison to control fish that were underperforming.

The physiological mechanism(s) underlying the increased Ṁ O_2max of P. amboinensis under near-future seawater CO₂ conditions remain to be elucidated. However, the consistent increase of Ṁ O_2max exhibited by P. amboinensis across the three experimental protocols, two of which constituted an acute exposure to near-future CO₂, suggests that the phenomenon did not arise from physiological acclimation to the elevated CO₂, but rather from the consequences of an acute exposure to increased CO₂ on the physiology of the fish. One explanation for the increased Ṁ O_2max of P. amboinensis when exposed to near-future CO₂ is that the maximum swimming speed of the fish was greater. Indeed, it is well established that Ṁ O_2max of fishes is positively correlated with swimming speed (Fry, 1971; Bushnell et al., 1984; Lee et al. 2003; Smit et al., 1965; Torres and Childress, 1983). A greater swimming speed could arise from a number of possibilities.

A greater maximal swimming speed could stem from a change in the motivation of the fish to swim fast. Recent studies have revealed that tropical reef fish exhibit an array of behavioral and sensory disruptions when exposed near-future CO₂ for several days. These range from reversal or loss of olfactory and auditory preferences, loss of behavioral lateralization to increased boldness and activity levels (Dixson et al., 2010; Domenici et al., 2011; Munday et al., 2010; Simpson et al., 2011; Nilsson et al., 2012). The behavioral changes are believed to arise from alterations of the normal flow of Cl⁻ and HCO₃⁻ through GABA-A receptors caused by disruptions of transmembrane Cl⁻ and HCO₃⁻ ion gradients (Nilsson et al. 2012). Nevertheless, we find it unlikely that the increased Ṁ O_2max of P. amboinensis arises from a motivational drive to swim faster due to the effects of CO₂ exposure on neurotransmitter function because the previously documented behavioral alterations only occurred after several days’ exposure to increased CO₂ (Dixson et al., 2010; Domenici et al., 2012; Munday et al., 2010; Simpson et al., 2011; Nilsson et al., 2012). Short-term fluctuations in CO₂ did not induce behavioral effects (Munday et al. 2010). In comparison, in the present study, P. amboinensis displayed an elevated Ṁ O_2max immediately (within minutes) of exposure to near-future CO_2. Moreover, a recent study investigating acid-base balance of the gulf toadfish (Opsanus beta) during exposure to near-future seawater CO₂, reported that full compensation of the respiratory acidosis and elevation of plasma HCO₃⁻ did not occur until after two hours of exposure (Esbaugh et al., 2012). Again, the time course of physiological change is at odds with alacritous increase of Ṁ O_2max displayed by P. amboinensis in the present study.

Alternatively, a greater swimming speed could arise from an increased oxygen delivery to the swimming muscles and oxygen uptake at the gills. Recent in vitro and in vivo studies have revealed that moderate acidosis can serve to increase the delivery of oxygen to the red muscle of teleost fish (Rummer and Brauner, 2011; Rummer et al., 2013). Briefly, in fish exposed to a stressor or mild acidosis, catecholamine release stimulates the activation of Na⁺/ H⁺ exchange across the erythrocyte membrane, thereby increasing red blood cell intracellular pH relative to the plasma and thus facilitating haemoglobin-O₂ binding at the gills (Boutilier et al., 1986; Nikinmaa, 1986). However, if plasma accessible carbonic anhydrase, which catalyzes the reversible conversion fo HCO₃⁻ and H⁺ to CO₂, is present in muscle capillaires, it serves to short-circuit the Na⁺/ H⁺ exchange, reduce red blood cell pH and haemoglobin-O₂ affinity and enhance O₂ unloading. For rainbow trout (Oncorhynchus mykiss) exposed to less than 1% CO₂, the mechanism increased red muscle pO₂ by 65% (Rummer et al., 2013). In the present study, the possibility exists that the combined exposure of P. amboinensis to near-future CO₂ and maximal exercise led to catecholamine release. If carbonic anydrase is present in the red muscle of P. amboinensis, an increased oxygen delivery to the swimming muscles would have also likely ensued. Moreover, external CO₂ can rapidly induce cardiorespiratory responses via gill chemoreceptors (Reid et al., 2005). In particular, environmental CO₂ elicits increased ventilation in most water-breathing fishes. An increased oxygen delivery to the muscles combined with an increased ventilation rate could translate to the observed greater oxygen uptake.

In this regard, the species-specific responses to near-future CO₂ in terms of Ṁ O_2max may indicate different response times for catecholamine release, regulatory capacities or tolerance to changes in blood H⁺ and/or differences among the species with regards to the presence of plasma accessible carbonic anhydrase in the red muscle. The lack of increase of Ṁ O_2max of P. amboinensis when the fish were exposed to pH 7.7 (corresponding to approximately 1400 μatm CO₂) or pH 7.5 (corresponding to approximately 2400 μatm CO₂) could arise from negative physiological effects of acidosis at the more extreme lower pH levels (Brauner and Baker, 2009) nullifying the enhanced physiological capacity to swim faster. This possibility could explain why many previous studies that exposed fish to much higher CO₂ levels (3000 – 60000 μatm) than those employed in the present study did not report any positive effects on Ṁ O_2max or aerobic scope (McKenzie et al., 2003; Melzner et al., 2009a). The juvenile damselfish and adult P. fuscus studied were much too small to enable blood sampling to test the above hypotheses. An alternate explanation for the increased Ṁ O_2max of P. amboinensis under near-future CO₂ conditions is that maximal swimming speed was unchanged at Ṁ O_2max, but an additional demand for oxygen to maintain homeostasis arose that was not apparent at Ṁ O_2rest or during maximal swimming under control conditions. For example, the ‘osmo-respiratory compromise’ almost doubles with exercise (Randall et al. 1972; Nilsson 1986). In this scenario, P. amboinensis would have incurred a greater cost to swim at its maximum speed under near-future CO₂ conditions. Clearly, future studies incorporating the measurement of swimming speed at Ṁ O_2max are needed to differentiate the possible explanations for the increased Ṁ O_2max of P. amboinensis under near-future CO₂. The use of a swimming flume was not feasible in the present study because such a system would have required a much larger volume of water than appropriate to accurately measure oxygen consumption of the extremely small juvenile fish (Steffensen, 1989). Rather, a cylindrical swimming chamber with a small volume, but with the capacity to swim the fish at their maximal swimming speed (Fisher et al. 2005; Gardiner et al., 2010; Nilsson et al., 2009, Munday et al., 2009a; Nilsson et al., 2007a; Nilsson et al., 2010; Nilsson and Östlund-Nilsson, 2004) was utilized to obtain reliable measurements of oxygen consumption.

Regardless of the possible mechanistic determinant(s) of the elevated Ṁ O_2max under near-future CO₂, the differing response to near-future CO₂ between P. amboinensis and its congener P. moluccensis could foreseeably have consequences for ecological interactions and the relative abundance of species within coral reef fish communities. With a higher aerobic metabolic capacity (and potentially maximum swimming speed), P. amboinensis would have the potential for increased individual performance in any energetically demanding behavior, such as swimming against a current, repaying O₂ debt after repeated anaerobic burst-swimming escapes from a threat, foraging or digesting (i.e., specific dynamic action). Concurrently, P. amboinensis should still be able to enter and remain in the hypoxic waters found deep inside coral colonies at night in order to escape predation (Nilsson et al., 2007a). The unchanged O_2crit of P. amboinensis after exposure to near-future CO₂ suggests that no trade-off exists for this species between its higher aerobic capacity and its hypoxia tolerance. In addition, in face of warming ocean surface temperatures, the enhanced Ṁ O_2max of juvenile P. amboinensis under near-future CO₂ may enable it to maintain its thermal tolerance window, and perhaps geographical distribution, as thermal tolerance is thought to be guided by aerobic scope in many species (Pörtner and Knust, 2007). On the contrary, if P. amboinensis incurs a greater metabolic cost while swimming at its maximum speed under near-future CO₂ conditions, energy expenditure for any other energetically demanding process would be reduced. Consequently, individual performance would be decreased in any energetically demanding behavior, leading to potentially negative consequences for the species.

Interestingly, the results of a recent experiment that examined mortality rate of a number of juvenile damselfish species when facing P. fuscus in a mesocosm after a four-day exposure to CO₂-acidified water found that P. amboinensis showed a similar mortality rate to P. moluccensis, as well as to other damselfish species (Ferrari et al., 2011b). The findings suggest no benefit or disadvantage of the increased Ṁ O_2max displayed by P. amboinensis under near-future CO₂ conditions. It may be that larger scale and longer term studies that encompass a variety of other variables such as temperature, current, life stages and the presence of other predators are required to reveal the implications of species-specific Ṁ O_2max responses to near-future CO₂ on fitness and mortality rate. Another recent study reported that in CO₂ acidified water, P. fuscus had a slower response to prey detection than in control water, but higher activity levels (Cripps et al., 2011). The higher activity level had been suggested to compensate for slower prey detection by increasing the chance of prey encounter (Cripps et al., 2011). The present findings indicate that a higher activity level in P. fuscus in CO₂-acidified water is not linked to a higher aerobic performance (aerobic scope). The higher activity levels of P. fuscus might be one of the consequences of increased neural excitation in CO₂-acidified water (Nilsson et al., 2012).

5. Concluding remarks

In summary, the present study reports an increased aerobic capacity of a juvenile marine teleost, P. amboinensis, in response to near-future pCO₂, but no effect on its congener P. moluccensis or on their predator P. fuscus. While the mechanistic basis for the species-specific responses and potential for differences in the swimming performance of the fish under near-future pCO₂ remains to be investigated, the results emphasize that being of the same genus, sharing similar ecology and life history, or living in the same environment, does not necessarily imply similar physiological responses to near-future CO₂. The results highlight that understanding interspecific variability is an important component of predicting the consequences of ocean acidification on marine communities and ecosystems. Additional studies assessing the effects of a number of other environmental factors in conjunction with near-future CO₂ exposure are required to more fully understand the interesting finding of the increased Ṁ O_2max of P. amboinensis in response to near-future seawater CO₂. Most importantly, future experiments should assess how maximum swimming ability is affected and if species-specific responses to elevated CO₂ persist with elevated temperature. Elevated pCO₂ in the future will not occur independently of temperature. Likewise, investigations into the effects of near-future CO₂ and temperature on different life stages of tropical coral reef species, as well as the capacity of both prey and predators to adapt to ocean acidification over the long-term, are needed to understand key ecological interactions and ultimately how the structure of ecological communities will be affected by climate change variables.