Coastal zones are ecologically and economically important and are among those areas that will be strongly affected by global climate change. Increasing atmospheric CO₂ concentrations lead to increasing mean temperatures and a higher frequency of thermal extremes as well as to ocean acidification (OA). These environmental changes pose risks for marine ectothermal organisms that are not yet fully understood [2–4]. Current ecosystem changes are largely driven by temperature, whereas ecosystem impacts of elevated CO₂ are expected, but still have to be demonstrated for moderate oceanic Pco₂ levels that are predicted to occur over the next 100 years. Previous studies indicate that OA, “global warming’s evil twin” (Richard Feely cited by [5]) will have mainly adverse consequences for many calcifying and non-calcifying organisms, and may result in changes to biodiversity, trophic interactions, and other ecosystem processes [4,6–11]. Marine organisms are affected by OA in several, mainly negative ways as shown by disturbances of acid-base regulation, respiration, energy turnover and metabolism, as well as a reduction in growth rates, reproductive success and calcification, and sensory abilities [12–20]. Depending on future CO₂ emission scenarios, predicted average surface seawater pH varies between 7.7 and 8.0 [21]. Deviation of environmental factors from the evolutionary optimum for a given species may result in deleterious impacts on energy homeostasis. Sub-optimal conditions create metabolic energy demands that may exceed energy supplied from food and/or accrued in somatic energy resources, as well as overwhelm the capacity of systemic functions (ventilation and circulation) and the cellular metabolic machinery to provide enough adenosine triphosphate (ATP) to sustain routine metabolism. An organism may compensate for elevated energy demand during the moderate stress by increasing energy intake and assimilation, and/or by elevated metabolic flux to cover ATP demand. However, during extreme stress exposures such compensation may be incomplete or impossible, and an organism can enter a metabolically depressed state to conserve energy and to extend the survival time until the conditions return to the optimum [22]. In either case, a disturbance of energy homeostasis will result in a limitation of aerobic scope of an organism, which in turn reduces the thermal tolerance limits of aquatic ectotherms [23]. Indeed, temperature-induced hypoxemia, transition to anaerobic energy production and oxidative stress were shown to play a critical role in determining the thermal tolerance limits and survival at elevated temperatures in aquatic ectotherms [24–29].

Recently, the concept of oxygen- and capacity-limited thermal tolerance of aquatic ectotherms [23] was extended to incorporate CO₂-driven effects predicting that the acute CO₂ stress will increase sensitivity of an organism to temperature change [1,4]. Indeed, hypercapnia caused a reduction in hemolymph O₂ partial pressure (P_eo₂) and narrowed the thermal tolerance window by a downward shift of the upper critical temperature in crabs, Cancer pagurus (1 kPa Pco₂, [30]) and Hyas araneus (0.3 kPa Pco₂, [31]). Michaelidis et al. [32] found no differences in P_eO₂ levels between control and hypercapnia-exposed mussels, Mytilus galloprovincialis (0.6 kPa Pco₂) when measured at the acclimation temperature. Following OA exposure extracellular pH (pH_e) dropped by 0.2 units and the standard metabolic rate (SMR) of the mussels decreased by about 60% indicating metabolic depression [32]. Metabolic depression acts as a time-limited adaptation strategy to survive unfavourable conditions such as hypercapnia [22]. Long-term performance and thus fitness is key to survival and success of a species. As summarized by Pörtner et al. [15], key physiological processes that are involved in setting the sensitivity to ocean acidification are the regulation of the organisms’ cellular acid-base and ion status and the respective feed back loops on other processes that are associated with individual performance. Regulation of pH_e is thought to “be the first line of defense against hypercapnia induced disturbances of metabolic and tissue functioning” (p. 210) emphasizing a key role of pH_e in metabolic depression [33].

In contrast to vertebrates, most invertebrates exhibit a low capacity for acid-base regulation such that the changes in acid-base and ion status may directly interfere with the organism’s performance [3,15,34,35]. Thus, the impact of future CO₂ concentrations is expected to be strong in invertebrates that are weak acid-base regulators and are unable to compensate the OA-induced shift in extracellular pH. Among this group, calcifying organisms may be particularly vulnerable to OA, because in addition to the acid-base disturbances they can experience disturbances in biomineralization needed for production of CaCO₃ exo- and endoskeletons [32,36–39]. As shown by Gazeau et al. [40] and Ries et al. [41], acute exposure to OA impairs the calcification of benthic mollusks already at calcium carbonate saturation values above 1 suggesting a decrease by 25% and 10% in mussel and oyster calcification by the end of the century with predicted Pco₂ levels of ~740 ppmv. In the mussel M. galloprovincialis long-term hypercapnia (seawater pH 7.3, ~0.6 kPa Pco₂) caused a reduction in hemolymph pH which was partly compensated by increased [HCO₃ ⁻]_e [32]. The increase in bicarbonate levels was likely linked to shell dissolution as indicated by increased extracellular [Ca²⁺] [32]. The drop in pH_e from 7.55 to 7.36 correlated with a reduction in metabolic rate and was suggested to be responsible for the observed growth rate reduction [32]. In Mytilus edulis chronically exposed to various CO₂ concentrations (seawater pH ranged from 6.7 to 8.1) growth increments were reduced when seawater pH fell below 7.4 [42].

In contrast to the bivalves, hypercapnia (0.4–0.6 kPa Pco₂, seawater pH of 7.1–7.2) did not depress metabolic rate and growth performance in the cephalopod, Sepia officinalis. Extracellular acidosis was compensated for by the accumulation of bicarbonate to a larger degree than in the bivalves but compensation remained incomplete [43]. These data indicate that CO₂ responses and thus tolerance thresholds may vary between animals depending on taxonomic group, acclimation history and degree of the hypercapnic stress.

Oysters are a major group of calcifiers in estuaries and coastal zones that may be strongly affected by OA and climate change. Estuarine and coastal organisms are regularly exposed to variable temperature and pH values in their environment and thus may be less sensitive to acidification and global warming than their open-ocean counterparts that live in more stable thermal and pH environments [42]. However, estuarine and coastal areas are likely impacted the most in terms of the degrees and rates of projected warming and seawater acidification trends. In combination with lower salinity and thus depressed CaCO₃ saturation state these trends may strongly affect ectothermic calcifiers such as oysters [44]. We hypothesized that chronic OA exposure will affect energy metabolism in oysters and will have negative consequences on their temperature tolerance. To test this hypothesis we determined the impact of long-term OA/hypercapnia on temperature-dependent metabolism by analyzing the acid-base and metabolic status of the Pacific oyster, Crassostrea gigas under conditions simulating a future scenario in which CO₂ levels stabilize at ~0.1 kPa Pco₂ and water pH decreases by ~0.5 units [45]. We show that OA exposure leads to a disturbance of acid-base status and changes in the steady-state levels of metabolic intermediates and induces an increase in basal maintenance costs at elevated temperatures. These disturbances of energy metabolism indicate that performance and survival of estuarine invertebrates can be affected by even moderate OA scenarios supporting the recommendation by Turley et al. [46].

During the first days of CO₂ incubation oysters C. gigas showed delayed behavioral defense responses (e.g., were slow to close or did not close their shells in response to a touch; data not shown). However, the behavioural responses normalized during long-term CO₂ exposure (up to 55 days), and only one oyster out of 23 animals died after 22 days of CO₂ exposure, which equals a 4.3% mortality rate. No mortality was observed in the control group. Chronic hypercapnia resulted in significant changes in hemolymph parameters of oysters (Table 1). CO₂-exposed oysters showed elevated P_eco₂ accompanied by lower hemolymph pH (decrease in pH_e by ~0.5 units) compared to controls. A small but significant increase in [HCO⁻ ₃]_e was also observed in oysters following chronic hypercapnia. Chronic hypercapnia also affected concentrations of [Na⁺]_e, [K⁺]_e and [Ca²⁺]_e in oyster hemolymph (Table 1). Hemolymph oxygenation was slightly affected by long-term hypercapnia with an approximately 20% reduction in hemolymph Po₂ in CO₂-exposed oysters (Table II; unpaired t-test, P = 0.07). Long-term hypercapnia also affected the body condition index of oysters that decreased marginally by ~20%, from 3.5 ± 0.6 (control) to 2.7 ± 0.7 (CO₂-exposed group)(unpaired t-test, N = 5–6, P = 0.052).

The ¹H-NMR spectra of the different tissues showed clear and reproducible specific metabolic profiles. Most of the signals did not change significantly between control and CO₂ exposed oysters in the various tissues, whereas concentrations of some metabolic intermediates were strongly affected by long-term hypercapnia in mantle and gill tissue (Figures 1b, c). In comparison the metabolic profile of muscle tissue was unaffected (Figure 1a). A strong depletion in alanine and ATP levels and a slight (non-significant) decline in glycogen concentrations was found in the mantle of CO₂-exposed oysters (Figure 1b). In gills, succinate levels were significantly elevated in CO₂-exposed oysters compared to their control counterparts (Figure 1c). A significant increase in succinate concentrations was also found in hepatopancreas of CO₂-exposed oysters compared to controls (data not shown).

SMR was similar in control and CO₂-exposed animals when measured at the acclimation temperature of 15 °C (Figure 2). With warming SMR rose significantly in both groups revealing a stronger rise in SMR of CO₂-exposed compared to control animals as shown by higher Q₁₀ values in the former group (Figure 2). As a result, SMR of CO₂-exposed oysters was significantly higher than in the controls at 20 °C and especially at 25 °C.

Contrary to whole animal SMR, we found no temperature-dependent effect of hypercapnia on respiration rates of isolated gill cells (Figure 3). Cellular respiration was significantly higher at 25 °C compared to 15 °C but the increase was similar in cells isolated from control and CO₂-exposed animals. Ouabain-sensitive respiration indicating energy demand for ion regulation via Na⁺/K⁺-ATPase comprised 33–37% of the total cellular oxygen demand and was not changed by temperature or by long-term hypercapnia (P > 0.05, data not shown).

Environmental change can be considered stressful if an organism needs to increase energy expenditure on maintenance, defense or repair. Stress-induced metabolic adjustments are aimed at reinstating the metabolic balance (i.e., energy homeostasis) and thus ensuring survival of the individual and, most importantly, of the population. Energy homeostasis implies that the energy demand is covered by sufficient energy supply. Furthermore, there is a net energy gain to invest into production (somatic growth and reproduction). Therefore, any environmental disturbance that reduces this energy investment into production will have direct consequences for the organism’s fitness.

Earlier studies in calcifying marine invertebrates including mollusks indicate that biomineralization is an energetically costly process. In molluskan shells consisting of inorganic crystals (calcite and/or aragonite) and an organic (mostly proteinaceous) matrix [72], production of the organic matrix was proposed to be the main cost-intensive process in shell growth [73]. Compared to the estimated total cost for inorganic shell material (1–2 J·mg⁻¹ CaCO₃), total costs for protein synthesis for shell formation are much higher (29 J·mg⁻¹) and may explain the observed inverse relationship between rates of calcification and proportion of organic matrix in shells of marine mollusks [73,74]. However, this is a conservative estimate of the cost of the inorganic shell material in mollusks because it does not take into account such energy-dependent aspects of calcification as the production of enzymes involved in calcium carbonate deposition (e.g., carbonic anhydrase), transport of CaCO₃ crystals by hemocytes and acid-base regulation at the site of CaCO₃ deposition [75–77]. Even with this conservative estimate, energetic costs for calcification could account for 75% of the total energy needed for somatic growth and may be up to four times higher than the amount of energy invested in reproduction, as shown in a rocky shore archeogastropod, Tegula funebralis [74]. After the shell is deposited, mollusks spend additional energy on its maintenance in order to counteract dissolution and erosion; in some intertidal limpets of the genus Patella the annual costs of shell erosion accounted for 8–20% of the total energy invested in production (somatic and shell growth and gonadal output)[78]. Such high energy expenditure for shell deposition and maintenance makes it likely that energetic trade-offs can also occur between shell, somatic growth and reproduction.

Exposure to elevated CO₂ levels and associated acidification of seawater can inhibit biomineralization rates and increase shell dissolution in marine mollusks [38,40,41,79] or as shown in some species of mollusks, crustaceans and echinoderms, calcification rates increased with decreasing seawater pH up to a certain point [41,51,80]. This may result in elevated metabolic costs of shell deposition and maintenance. The present investigation did not reveal an increase in basal metabolism in OA-exposed oysters when measured at acclimation temperature suggesting: (i) no OA-induced rise in costs or (ii) a shift in energy budget between tissues where a cost depression in one tissue is compensated for by the cost increment in another (see below). In the brittlestar Amphiura filiformis [80] the OA-induced increase in calcification rates coincided with an increase in respiration rates and was accompanied by a decrease in muscle mass [80]. The authors interpreted the loss in muscle mass as an energy source for the OA-induced elevated energy demands indicating a trade-off between structure (morphological integrity) and function (arm movement). In contrast, OA-exposed cephalopods S. officinalis gained soft body mass at similar rates as control animals although calcification rates were significantly higher in OA-exposed than in control animals [43]. In a following study the authors noted a reduced incorporation of organic matrix in the calcified structure of OA-exposed cephalopods [51] indicating again that a compensatory increase in calcification rates with increasing acidification comes at a cost.

Similar to previous reports on mussels [81] mild CO₂ accumulation according to OA scenarios had no impact on SMR of oysters at the acclimation temperature (15 °C) indicating neither elevated energy demand nor metabolic rate depression in CO₂-exposed animals. However, SMR of CO₂-exposed animals was significantly above that of the normocapnic controls during acute warming indicating that hypercapnia resulted in elevated energy demand when combined with temperature stress. Moreover, the lowered condition index of CO₂-exposed animals suggested reduced growth efficiency due to a likely shift in energy budget in OA-exposed animals (see discussion above). Furthermore, it has been shown that hypercapnia interferes with protein turnover resulting in lowered protein synthesis in marine organisms [80,82,83], which may explain reduced growth under elevated CO₂ levels despite unchanged respiration rates.

Whole animal respiration rates gives an estimation of the overall sum of all energy consuming processes, making it likely to oversee small changes in specific processes or if the change in one process is compensated for by another. In line with observations cited above the rise in calcification costs may compensate for metabolic depression in other tissues like muscle or gills such that a net effect of OA on SMR of the whole animal does not occur. In this context, OA-induced alterations in metabolite profiles of oyster gills and mantle indicate that hypercapnia has an impact on metabolic pathways in these tissues. In gills and hepatopancreas, the most notable alteration was an increase in succinate concentration during prolonged exposure to elevated CO₂ levels. Succinate accumulation is considered an indicator of anaerobic metabolism in bivalves including oysters [84,85]. Succinate accumulation in CO₂-exposed oysters in our study may only be due to a transition to partial anaerobiosis in the respective organs. We cannot exclude such oxygen limitation to set in as we found a small reduction in hemolymph oxygenation at constant SMR during long-term hypercapnia at 15 °C. This indicates that tissues may not become fully hypoxemic under these conditions but tissue-specific microperfusion may have been disturbed. However, further study is needed to corroborate our conclusions. In contrast to familiar patterns of anaerobic metabolism, no alanine accumulation was observed in gills or hepatopancreas (data for hepatopancreas not shown). Alanine is an early indicator of acute anaerobiosis in marine bivalves including oysters and its accumulation typically precedes that of succinate [84–86]. Hypercapnia also resulted in a strong (by ~80%) depletion of alanine levels in the mantle tissues of oysters. Given that osmotic and hypoxic stress, two major modulators of amino acid levels in bivalves, can be ruled out in our study, the drop in alanine levels of mantle tissue of OA exposed oysters may reflect increased gluconeogenesis. In fully aerobic cells this process may occur and compensate for the putative accumulation of alanine in anaerobic cells. An OA-induced shift in metabolic pathways favoring gluconeogenesis due to an OA-induced impairment of glycolysis was also suggested in fish [87]. We suggest that alanine is transaminated to pyruvate, which together with ATP will be used to build up phosphoenolpyruvate (PEP), with PEP as substrate entering the gluconeogenetic pathway. The glucose synthesized from alanine in the mantle can then be used to fuel the metabolic demand of mantle and most likely transported to other tissues like the gills. However, a nonsignificant trend (due to the high variability between individuals) for a decrease in the glycogen level was observed in the mantle tissue of hypercapnic oysters, which would be consistent with elevated glycolytic flux. Hypoxemia may prevent enhanced gluconeogenesis to be effective in replenishing glycogen levels. Further investigations are needed to determine activities of enzymes involved in glycolysis and gluconeogenesis and/or metabolite fluxes in different tissues to fully unravel the mechanisms of the observed metabolite shifts and their physiological consequences at the wholeorganism level.

Extracellular pH has been identified as an important determinant of metabolic rate of muscle tissue and (vertebrate) liver, and metabolic rate can drop if pH_e decreases below a certain threshold level [33]. OA exposure caused a marked acidosis in extracellular fluids of C. gigas, which remained largely uncompensated as bicarbonate levels increased only slightly. Along with the relative stability of bicarbonate levels, extracellular [Ca²⁺] also did not change during hypercapnia in C. gigas indicating no dissolution of shell or tissue calcium carbonate. In contrast, hypercapnia-induced acidification of seawater (pH = 7.3) resulted in a decrease of pH_e in M. galloprovincialis and a significant increase in hemolymph Ca²⁺ levels which was interpreted as partial buffering of pH_e by shell dissolution [32]. Our findings are in line with recent studies in blue mussels M. edulis and the Greenland smoothcockle, Serripes groenlandicus, where no significant change in extracellular [HCO₃ ⁻] and [Ca²⁺] was noted with increasing seawater Pco₂ up to 0.4 kPa [88–90]. Possibly, the degree of involvement of the shell and tissue CaCO₃ in buffering of the hemolymph pH depends on the degree of acidification and may become notable when the environmental pH decreases below a certain species-specific threshold. Furthermore, as discussed by Hardewig et al. [91] “metabolism itself may contribute to the regulation of acid-base balance”. Pörtner [92] suggested that PEP can be carboxylated under hypercapnia resulting in an increase of succinate (as observed in this study) and propionate concentrations accompanied with the consumption of bicarbonate or CO₂, an additional benefit under hypercapnic conditions [88].

Concomitant with the more acidic pH_e, Michaelidis et al. [32] found reduced respiration rates in OA-exposed mussels. Also in oysters (C. virginica), Willson and Burnett [69] reported a lowered oxygen uptake in animals subjected to high CO₂ partial pressures: high (0.8–1 kPa Pco₂, seawater pH ≤ 7) compared to controls at low (<0.1 kPa Pco₂, seawater pH 8.2). In line with findings at the whole animal level, oxygen uptake of isolated gill tissue from oysters, C. virginica was depressed under the low pH conditions irrespective of the CO₂ levels [69]. Reipschläger and Pörtner [68] and Pörtner et al. [93] provided conclusive evidence that and how metabolic depression in isolated body wall musculature of the peanut worm Sipinculus nudus under hypercapnia was specifically due to lowered extracellular pH rather than intracellular pH, elevated Pco₂ or modified bicarbonate levels. In our experiments, seawater pH fell to 7.7 and no pH_e compensation was observed (pH_e = 7.1). At the temperatures tested CO₂-exposed animals showed no sign of metabolic depression and, similarily, isolated gill cells showed no change in respiration rates when measured in buffer that was adjusted to the respective pH_e [7.6 (control) vs. 7.1 (CO₂-group)]. Different temperature sensitivities suggest, however, that upon cooling, CO₂ may cause metabolic rate to fall below controls. Our findings suggest that within the tested temperature range the pH_e reached during exposure to mild hypercapnia as in this study is found above the threshold triggering overall metabolic rate depression in C. gigas.

Ion regulation capacity via Na⁺K⁺-ATPase showed no difference between gill cells isolated from control and CO₂-exposed animals. The data for Na⁺K⁺-ATPase regulation under hypercapnia are sparse and contradictory suggesting species-specific thresholds in CO₂-induced impact on Na⁺K⁺-ATPase capacity. Thus, in some species elevated CO₂ levels have no effect on activity of Na⁺K⁺-ATPase (2 kPa Pco₂, [94]) whereas in others increased Na⁺K⁺-ATPase activity was observed after acclimation to long-term hypercapnia (1–5 kPa Pco₂, [17,95]). Melzner et al. [3] observed a concentration-dependent effect of CO₂ on Na⁺K⁺-ATPase activity in cod gills with unchanged activity at 0.3 kPa Pco₂, and increased activity at 0.6 kPa Pco₂. In S. nudus, extracellular pH modulates the energy costs for acid-base regulation by a shift from less to more ATP-efficient ion transporters during hypercapnia resulting in a reduction of Na⁺K⁺-ATPase activity due to reduced requirement for sodium regulation [68,93,96]. In CO₂-exposed oysters the observed lowering of extracellular [Na⁺] and elevation of [K⁺] compared to controls is consistent with the hypothesized shift to more ATP-efficient ion transporters. If it holds true in oysters, the associated changes in Na⁺K⁺-ATPase activity must be small and undetectable by the method employed in our present study (determination of ouabainsensitive respiration). To date, many different ion transporters have been identified [97,98] but the complex interplay of these proteins regulation ion- and pH balance is not yet fully understood and requires further investigation.

Our present study demonstrates that CO₂ levels corresponding to expected OA scenarios are likely to interfere with the energy metabolism of oysters. This may reflect vulnerability to OA and temperature extremes. These findings are especially noteworthy because oysters, like other estuarine invertebrates, are normally exposed to broad fluctuations in CO₂ levels, pH and temperature in their habitats and thus should be better adapted to these changes than their deep-water or open-ocean counterparts. Nevertheless, chronic hypercapnia affects energy metabolism even in this eurybiont species especially when combined with elevated temperature. Synergistic effects of elevated temperature and hypercapnia were also identified in other bivalves [89,99] and are consistent with earlier reports that elevated temperature enhanced the sensitivity to a variety of environmental stressors such as pollution, hypercapnia or oxygen deficiency [1,4,100]. However, species-specific physiological differences in biomineralization as well as energy metabolism and acid-base regulation may shape differential sensitivities of various marine invertebrates to OA and elevated temperature thus complicating the picture of ecosystem-level responses of marine organisms to a high CO₂ world. Further studies are critically needed to determine the range of sensitivities of key marine species to OA and global climate change and the mechanisms setting limits to their tolerance to elevated temperatures and low pH in the future oceans. Analyses of energy metabolism as in the present study can provide a useful integrative view of stress effects on physiological performance of a variety of marine species and characterize their tolerance and tolerance limits in the face of global change.