Temperature but not ocean acidification affects energy metabolism and enzyme activities in the blue mussel, Mytilus edulis

Abstract In mosaic marine habitats, such as intertidal zones, ocean acidification (OA) is exacerbated by high variability of pH, temperature, and biological CO2 production. The nonlinear interactions among these drivers can be context‐specific and their effect on organisms in these habitats remains largely unknown, warranting further investigation. We were particularly interested in Mytilus edulis (the blue mussel) from intertidal zones of the Gulf of Maine (GOM), USA, for this study. GOM is a hot spot of global climate change (average sea surface temperature (SST) increasing by >0.2°C/year) with >60% decline in mussel population over the past 40 years. Here, we utilize bioenergetic underpinnings to identify limits of stress tolerance in M. edulis from GOM exposed to warming and OA. We have measured whole‐organism oxygen consumption rates and metabolic biomarkers in mussels exposed to control and elevated temperatures (10 vs. 15°C, respectively) and current and moderately elevated P CO2 levels (~400 vs. 800 µatm, respectively). Our study demonstrates that adult M. edulis from GOM are metabolically resilient to the moderate OA scenario but responsive to warming as seen in changes in metabolic rate, energy reserves (total lipids), metabolite profiles (glucose and osmolyte dimethyl amine), and enzyme activities (carbonic anhydrase and calcium ATPase). Our results are in agreement with recent literature that OA scenarios for the next 100–300 years do not affect this species, possibly as a consequence of maintaining its in vivo acid‐base balance.

Responses of marine molluscs to OA are highly variable (Sokolova et al., 2015 and references therein). Inter-and intrapopulation variability in OA sensitivity has been shown depending on habitat, scales of environmental variability, and other concomitant stressors (Parker et al., 2011;Stapp et al., 2017;Vargas et al., 2017;Waldbusser et al., 2015). In mosaic environments, local adaptation, as well as temporally and spatially varying selection, can select for metabolically plastic, stress-tolerant genotypes that can maintain optimal phenotypes (including energetic sustainability) in a broad range of environmental conditions. Importantly, bioenergetic responses can predict tolerance limits under environmentally realistic scenarios of stress exposure  providing a common denominator to integrate responses to multiple stressors.
Quantifying the independent and interactive effects of multiple stressors to identify metabolic tipping points is essential to determine the impact of global climate change on marine organisms and ecosystems (Boyd & Brown, 2015).
This study aims to determine the interactive effects of elevated temperature and P CO2 on energy metabolism and biomineralizationrelated enzymes in an ecologically and economically important bivalve mollusk, the blue mussel Mytilus edulis Linnaeus 1758. It is a critical foundation species in coastal ecosystems (Seed, 1969) that increasingly faces risk of local extinction along the USA east coast (Jones et al., 2009;Sorte et al., 2017). We were particularly interested in mussel populations from intertidal zones of Gulf of Maine (GOM), USA, for this study. GOM is a hot spot of global climate change with >60% (range 29%-100%) decline in mussel population over the past 40 years (Sorte et al., 2017). Within GOM, the previous decade has witnessed an average warming of >0.2°C/year (Sorte et al., 2017). Satellite observations of SST within the GOM show that the region was warming at a faster rate than 99% of the global ocean (Pershing et al. 2015), with the highest average annual values exceeding over 150 years of observations held in NOAA's Merged Land-Ocean Surface Temperature Analysis database (Salisbury & Jönsson, 2018). Combined with poor buffering capacity of the seawater in the coastal GOM regions (Ekstrom et al., 2015), temperature alone produces an annual change of 0.013 in pH and 1.06 in Ω AR (Salisbury & Jönsson, 2018). Modeling data indicate that these rapid changes in the GOM together with physical processes in the area (e.g., strong tides, wind-driven mixing, coastal currents, and adjacent equilibrium with Northwestern Atlantic) alter ocean carbonate parameters but also mitigate the decrease, or even raise pH (Salisbury & Jönsson, 2018). Mussels inhabiting GOM are thus exposed to one of the fastest rates of warming in the world, in addition to ocean acidification.
The metabolic plasticity of M. edulis to the combined effects of elevated P CO2 and temperature from GOM is not yet fully understood.
Here, we utilize the bioenergetic underpinnings of stress physiology to identify limits and mechanisms of stress tolerance in M. edulis. We tested the hypothesis that elevated temperature will exacerbate the effects of ocean acidification, resulting in elevated basal energy metabolism, which may decrease the amount of energy reserves and reduce enzyme activities involved in energy-demanding process of shell formation. To this end, we measured whole-organism oxygen consumption rates and metabolic biomarkers in mussels exposed to control and moderately elevated temperatures (10 vs. 15°C) and P CO2 (~400 vs. 800 µatm) to mimic a realistic scenario of ocean warming and acidification. Standard metabolic rate (SMR) represents the basal energy cost for maintenance and is widely used to assess stress response (Pettersen et al., 2018). To account for possible tissue-specific variation in responses, we conducted a comprehensive analyses of the bioenergetic health index by measuring energy-related biomarkers (cellular and tissue energy reserves), metabolic profiles, and specific enzyme activities (acid-base regulating enzyme and energy-demanding ion transport enzymes) in different tissues depending on their physiological role. In mollusks, including M. edulis, carbonic anhydrase (CA) is the major acid-base regulating enzyme (Li et al., 2016;Ramesh et al., 2020;Wang et al., 2017). It plays a key role in gas exchange, acid-base regulation, calcification, and ion transport. Given the central role of CA, its activity is likely to play an important part in physiological mechanisms of acclimatization and tolerance of marine calcifiers, including M. edulis, to CO 2 -induced shifts in ocean chemistry as well as to other stressors such as temperature, salinity, or pollution. However, the fundamental information about the tissue distribution and thermal sensitivity of CA kinetics in M. edulis is not known. We, therefore, analyzed the tissue distribution and temperature sensitivity (determined by apparent activation energy and Arrhenius breakpoint temperature) of CA activity here in a separate set of experiment. This will provide the muchneeded data about the nuanced role of CA in regulation of acid-base balance and biomineralization in M. edulis.

| Chemicals
Unless otherwise indicated, all chemicals and enzymes were purchased from Sigma-Aldrich, Roche, or Fisher Scientific and were of analytical grade or higher.

| Animal collection, maintenance and experimental design
Blue mussels M. edulis were collected from Biddeford Pool, Gulf of Maine (43°26′50.6N, 70°21′19.0W) in early summer 2011 and shipped on ice by an overnight delivery to the University of North Carolina at Charlotte. Mussels were kept in tanks with recirculating artificial seawater (ASW) (Instant Ocean ® , Kent Marine) at 9.6 ± 0.3°C and 30 ± 1 salinity (practical salinity units, PSU), aerated with ambient air for 10 days. Mussels were then randomly assigned to four treatment groups, and each group was exposed for 4 weeks to one of the four possible combinations of two levels of P CO2 and two temperatures. The two selected P CO2 levels were representative of the present-day conditions (~400 µatm P CO2 ; normocapnia) and atmospheric P CO2 concentrations predicted by a moderate scenario of the Intergovernmental Panel for Climate Change (IPCC) for the year 2100 (~800 µatm P CO2 ; hypercapnia) (IPCC, 2014). The temperatures were chosen to represent the average water temperature of mussels at the time of collection (10°C), and a +5°C increase predicted for the year 2100 by an IPCC scenario 'business-as-usual' (15°C). Both experimental temperatures are within the environmentally relevant range for the studied mussel population in the GOM. For the bivalves exposed to elevated temperature, water temperature in the tanks was slowly raised from 10°C by 1°C per day until 15°C was achieved and the experimental exposures began.
Two replicate tanks were set for each experimental treatment.
Water in normocapnic treatments was bubbled with ambient air whereas for hypercapnic treatments ambient air was mixed with 100% CO 2 (Roberts Oxygen, Charlotte, NC,USA) using precision mass flow controllers (Cole-Parmer, Vernon Hills, IL, USA). The air-CO 2 mixture flow rate was set up to maintain the respective systems at a steady-state pH. Animals were fed ad libitum on alternative days with 2 ml per tank of commercial algal mixture containing Isochrysis spp., Pavlova spp., Thalassoisira weissflogii, and Tetraselmis spp. with 5-20 μm cells (Shellfish Diet 1800, Reed Mariculture, Campbell, CA, USA). Mortality was checked daily and animals that gaped and did not respond to a mechanical stimulus were recorded as dead and immediately removed. Artificial Sea Water (ASW) for all exposures were prepared using the same batch of Instant Ocean ® (Kent Marine, Acworth, GA, USA) salt to avoid potential variations in water chemistry.
Carbonate chemistry of seawater was determined periodically during experimental exposures as described elsewhere (Beniash et al., 2010).
Seawater temperature and chemistry data are shown in Table 1.

| Standard metabolic rate
Standard metabolic rate was measured as resting oxygen consumption (ṀO 2 ) of mussels at their respective acclimation temperature and P CO2 using microfiber optic oxygen probes (Tx-Type, PreSens GmbH, Germany, www.presens.de) as described in Matoo et al., 2013. Two-point calibration was performed at each temperature and P CO2 concentration. Mussels were placed into flowthrough respiration chambers and allowed to recover overnight. To avoid interference with postprandial metabolism and feces excretion, animals were fasted for 24 hr prior to the start of Ṁ O 2 recordings. Water flow (20-25 ml/min) was adjusted so that animals consumed less than 25% of O 2 at all times to avoid potential inhibitory effects of low oxygen levels on respiration rate. 10 biological replicates for each temperature and P CO2 group were randomized across chambers and respirometry runs. After each individual run for ~12 hr, the animals were dissected to determine wet tissue mass. Dry tissue mass was calculated from the wet tissue mass assuming an average water content of 80%.
After acclimation, a subset of mussels was dissected, tissues shock-frozen, and stored in liquid nitrogen for analyses of energy reserves and enzyme activities. Due to limited amount of tissues, we divided samples for different assays depending on the physiological function of a given tissue. The energy reserves (lipids, glycogen, and/or adenylates) were measured in hepatopancreas and adductor muscle that serve as reserve storage sites in bivalves (Cappello et al., 2018). Metabolite profiles were explored by untargeted metabolomics in two metabolically active aerobic tissues, the gills, and the muscle. The effects of warming and OA on biomineralization were assessed by the activities of three key enzymes involved in shell formation (carbonic anhydrase (CA), plasma membrane calcium (Ca 2+ ) ATPase, and proton (H + ) ATPase) in the mantle tissue as the main organ involved in shell formation.

| Energy reserves
Lipid content was determined using the chloroform extraction method as described in Ivanina et al., 2013. Concentration of lipids was expressed as g/g wet tissue mass. Concentrations of glycogen and adenylates (μmol g − 1 wet tissue mass) were measured using standard NADH-or NADPH-linked spectrophotometric tests described in Ivanina et al., 2013. Adenylate energy charge (AEC) was calculated as follows:

| Metabolic profiling based on 1 H-NMR spectroscopy
Samples of the muscle and gill tissues were extracted as described elsewhere for untargeted metabolic profiling using NMR spectroscopy (Dickinson et al., 2012;Lannig et al., 2010).
Frozen tissues were homogenized under the liquid nitrogen and extracted with ice-cold perchloric acid (PCA, 0.6 M). Samples were homogenized by ultrasonic treatment (0°C, 360 W) and centrifuged (0°C, 2 min, 16,000 g) to remove precipitated protein. Supernatants were neutralized to pH 7.0-7.5 using potassium hydroxide and centrifuged to remove precipitated potassium perchloride. Extracts were freeze-dried and shipped on dry ice to the Alfred Wegener Institute (Bremerhaven, Germany) for NMRbased metabolic profiling. Untargeted metabolic profiling using 1 H-NMR spectroscopy was performed using a method modified from Schmidt et al. 2017. Samples were dissolved in D 2 O containing 1% trimethylsilylpropanoic acid (TSP) as internal standard to achieve a constant extract/TSP ratio. Typically, a sample volume of 70 µl was transferred to a HRMAS rotor and placed into a triple tuned high-resolution magic angle spinning (HRMAS) probe of a wide-bore 400 MHz NMR spectrometer (9.4T Bruker Avance III HD, Bruker Biospin). NMR spectra were collected at a temperature of 4°C and a spinning rate of 3,000 Hz. A set of 1D-1H-NMR spectra were collected from every sample including a classical one-pulse with water saturation protocol (Bruker's zgpr), a Carr-Purcell-Meiboom-Gill (cpmg) sequence used for quantification, a NOESY protocol and a J-resolved (JRES) protocol for signal identification with parameters as described in (Schmidt et al. 2017). For each sample, 32 scans with a spectral width of 5,000 Hz were collected in 64K data points. The measuring protocol lasted ~28 min per sample. 1 H-NMR spectra from the cpmg protocol were used to analyze the metabolic profiles. Spectra were processed with an exponential multiplication (lb = 0.3) and automatically phase-and baseline-corrected using Topsin 3.2 (Bruker Biospin). A total of 24 metabolites was identified from processed tissue spectra and quantified using Chenomx NMR suite 8.1 (Chenomx Inc.). Metabolic profiles were analyzed and tested for statistically significant changes between groups using

| Enzyme activities
Activities of biomineralization-related enzymes were measured in the mantle edge tissue (a 2-3 mm wide edge along the ventral shell margin functionally specialized for biomineralization) of M. edulis exposed to different temperature and P CO2 conditions. Mantle edge expresses biomineralization-related genes and is involved in shell deposition in mussels and other bivalves (Bjärnmark et al., 2016;Gazeau et al., 2014). Protein concentrations were determined using Bradford assay (Bradford, 1976) and used to standardize enzyme activities.
In a separate set of experiments, CA activity was measured at different temperatures in an environmentally relevant range (5-35°C) in the gill, mantle, adductor muscle, and hepatopancreas of the control mussels to characterize the tissue-dependent capacity and temperature sensitivity (determined by apparent activation energy (E a ) and Arrhenius breakpoint temperature (ABT)) of CA activity. E a was determined from an Arrhenius plot of ln(V max ) against 1/T (K −1 ), and ABT was determined as a point when the slope of Arrhenius plot significantly changed using an algorithm for multi-segment linear regression (Oosterbaan, 2011). 2.6.2 | Calcium (Ca 2+ ) and proton (H + ) ATPases Mantle edge tissue was homogenized and activities of Ca 2+ -ATPase (EC 3.6.3.8) and H + -ATPase (EC 3.6.3.6) were assayed as described in Ivanina et al., 2020. Inorganic phosphate (P i ) was measured using malachite green assay kit (ab65622, Abcam) and ATPase activities were expressed as µmol of P i µg protein −1 hr −1 .

| Statistical analyses
Effects of temperature, P CO2 , and their interaction were assessed for all studied traits using generalized linear model (GLM) ANOVA.
All factors were treated as fixed and post hoc tests (Fisher's least significant difference) were used to test differences between group means. The number of biological replicates was 5-10 for all experimental groups. Regression analysis for ABT and Arrhenius plots for CA were done using GraphPad Prism ver. 4.03 (GraphPad Software, Inc.) and SegReg software (Oosterbaan, 2011).  (Tusher et al. 2001). A SAM plot compares the observed relative differences by the expected relative differences from the permutation results. Metabolites that showed a particular pattern were identified using PatternHunter analysis within MetaboAnalyst (Pavlidis & Noble, 2001).
Unless otherwise indicated, data are shown as means ± standard errors of means (SEM). Differences were considered significant if probability of type I error was < 0.05.

| RE SULTS
ANOVA analysis showed a significant effect of the acclimation temperature of the SMR, total lipid content of hepatopancreas, and activities of CA and Ca 2+ -ATPase in the mantle tissues of M. edulis (Table 2). No significant effects of the acclimation temperature were found for glycogen levels in the hepatopancreas, ATP, ADP, AMP or AEC in the adductor muscle, or H + -ATPase activity in the mantle. Acclimation P CO2 or temperature × P CO2 interactions had no significant effect on any of the studied bioenergetics-or biomineralization-related traits in M. edulis (Table 2).
OA did not significantly affect SMR in mussels (p = 0.070), although a trend of elevated SMR was observed in OA-exposed mussels.

| Tissue energy status under warming and OA
Warming significantly increased the total lipid content in hepatopancreas under normocapnia (p = 0.05) and hypercapnia (p = 0.001) ( Figure 1b; Table 2). OA did not significantly change the lipid content in hepatopancreas, regardless of the temperature (p = 0.608 and 0.288 at 10 and 15°C, respectively).
No significant changes were observed for glycogen ( Figure 1c) and adenylates (Figure 2) content under warming, OA, or OA combined with warming (OWA) in the muscle of M. edulis (Table 2). These two metabolites were also identified as significantly different between the temperature groups by SAM analysis (Figure 4b). Unlike gills, metabolite profiles of adductor muscle were not affected by temperature or P CO2 (see PLS-DA in Figure S3).

| Effects of warming and OA on enzyme activity
Carbonic anhydrase activity in the mantle of M. edulis showed an elevated trend under warming (p = 0.001 for temperature effects) but did not under OA (p = 0.30) (Figure 5a; Table 2). When data for normocapnia and hypercapnia were pooled, a significant increase in CA activity at 15°C compared to 10°C (p = 0.018) was detected ( Figure S1). Tissue-specific CA activity over a broad temperature range (5-35°C) showed a significant effect of two-factor (temperature x tissue) interactions (p < 0.0001). CA activity was significantly higher in the hepatopancreas compared to other tissues (Figure 5b).
Irrespective of the tissue, CA activity monotonously increased with increasing temperatures with similar E a (32.5-43.8 kJ/mol K −1 ) and no ABT in the studied tissues (data not shown).  Table 2). At 10°C, hypercapnia led to a decrease in the H + -ATPase activity, but this decrease was nonsignificant (p = 0.06).

| D ISCUSS I ON
Our study demonstrates that adult M. edulis from GOM are metabolically resilient to moderate OA (P CO2 ~ 800 µatm) but responsive to warming as seen in changes in whole-body metabolic rate, energy reserves, metabolite profiles, and enzyme activities. The combination of ocean warming and acidification (OWA) did not elicit TA B L E 2 ANOVA results of the effects of exposure temperature, P CO2 , and their interaction on energy-related indices and enzyme activities in Mytilus edulis F I G U R E 1 Oxygen consumption rates of Mytilus edulis exposed to different temperatures and P CO2 . The traits were measured in mussels groups at their respective acclimation temperatures (10 vs. 15°C) and P CO2 (~400 vs. 800 µatm). Within each graph, different letters indicate means are significantly different (p < 0.05). Vertical bars represent SEM. N = 8-10 detrimental metabolic changes in mussels beyond the effect of warming. This indicates that under these conditions, temperature is the dominant factor in determining species' physiology. Intertidal pools, in general, are very dynamic environments with variations of temperature, pH and CO 2 levels, where seawater pH can fluctuate from 8.1 to 6.9-6.5 (and P CO2 can change from ~ 400 μatm to more than 10,000-40,000 μatm) due to biological CO 2 production and/or freshwater inflow, and temperature fluctuations can exceed 15-20°C during diurnal, tidal and seasonal cycles with local geographical factors moderating these variables (Burnett, 1997;Chapman et al., 2011;Ringwood & Keppler, 2002). The physiological responses of mussels in our study might be related to their ecology in GOM. There is a high temporal variability in temperature (annual SST range 15.5°C) in temperate rocky intertidal pools of GOM throughout the year (Salisbury & Jönsson, 2018). Thus, M. edulis in this region are predominantly exposed to large variation in temperature which might explain their plasticity to thermal stress. Earlier studies in M. edulis populations from thermal clines of GOM showed high phenotypic plasticity in physiology despite lack of population genetic structure and local adaptation (Lesser, 2016;Lesser et al., 2010) consistent with the findings of metabolic plasticity to warming found in our study.

| Effect of warming and OA on bioenergetics
We observed a ~2-3-fold increase in SMR of mussels exposed to 5°C warming under normocapnia or hypercapnia, indicating a strong temperature effect on metabolism with Q 10 ~ 4-6. A 5°C increase is well within the range of temperature fluctuations experienced by mussels in GOM (Salisbury & Jönsson, 2018). Blue mussels are eurythermal and well adapted for 5-20°C range, with an upper thermal tolerance limit of ~29°C for adults (Gosling, 1992).
Therefore, temperatures in our study are within the ecological relevant and even optimal range for this species. On the other hand, Pejus temperatures represent thermal transitions where organisms mount compensatory mechanisms that divert away from performance parameters with direct fitness correlates (Zittier et al., 2015). Previous studies have shown that the circulatory system of M. edulis reached its temperature-induced capacity limit above 25°C, indicating the onset of the Pejus range at this temperature (Zittier et al., 2015). The elevated SMR at 15°C seen in the present study, therefore, reflects the rate-enhancing effects of temperature on physiological and biochemical reactions such as activities of metabolic enzymes, ion channels, mitochondria, and other important bioenergetic processes (Hochachka, 1973). Recent meta-analysis also indicates that temperature threshold for long-term survival of M. edulis is ~20°C (Lupo et al., 2021).

F I G U R E 2
Therefore, 20°C appears to be close to the metabolic optimum of M. edulis, so rate-enhancing effects of temperature dominate over the potentially negative impacts on metabolism as long as warming occurs below the 20°C threshold, as in our study.
Unlike temperature effect, modest OA had no effect on SMR regardless of the temperature. Metabolic response to OA in M. edulis F I G U R E 3 PLS-DA analysis of metabolite profiles in the gill tissues of Mytilus edulis exposed to different temperatures and P CO2 . (a) 3D-loading plot of first three components separating the metabolic profiles from Warming and OWA groups (blue crosses and teal diamonds) from control (red triangles). The OA group (green plus) did not separated from control. (b) Important metabolites identified by PLS-DA. The colored squares on the right show group-specific relative changes in metabolite concentration. Groups: Ctr-control (acclimated at 10°C and normocapnia), OA-ocean acidification (acclimated at 10°C and hypercapnia), W-warming (acclimated at 15°C and normocapnia), OWA-ocean warming and acidification (acclimated at 15°C and hypercapnia) can vary and depend on magnitude of P CO2 , food availability, and population (Fitzer et al., 2014;Hüning et al., 2013;Lesser, 2016;Thomsen et al., 2013;Thomsen & Melzner, 2010;Zittier et al., 2015). Fitzer et al. (2014) showed that for M. edulis at 1,000 µatm and beyond, biomineralization continued but with compensated protein metabolism and shell growth indicating that ~1,000 µatm could be an OA metabolic tipping point for M. edulis. In our study, P CO2 levels were below this threshold and could explain the physiological tolerance of mussels seen here. Our hypothesis that P CO2 could exacerbate the increase of SMR caused by elevated temperature was not supported in this study. The outcome of OWA on metabolic rate in bivalves is commonly additive; albeit in other cases, the effects of temperature or P CO2 dominate (see Lefevre, 2016 and references therein). Furthermore, metabolic responses of bivalves to OWA are dependent on the degree of temperature or P CO2 stress. For example, Lesser, 2016 showed that mussels from GOM showed metabolic depression as a protective response when exposed to combined stress of higher warming (22°C) and modestly elevated P CO2 (560 µatm).
Temperature (but not OA) had a marked effect on lipid content in M. edulis. We observed an increase in lipids under warming in hepatopancreas, under both normocapnia and hypercapnia. In bivalves, lipids are primarily stored in hepatopancreas (Giese, 1966) and synthesis, storage, and use of lipids show pronounced seasonal cycles. Specifically, lipids are accumulated during summer (at high temperature and food availability) and used for metabolism and initiation of gametogenesis during winter (at low temperature and food F I G U R E 4 (a) Tissue levels (given in normalized concentrations) of DMA and glucose in gill tissues of Mytilus edulis exposed to different temperatures and P CO2 . Groups: Ctr-control (acclimated at 10°C and normocapnia), OA-ocean acidification (acclimated at 10°C and hypercapnia), W-warming (acclimated at 15°C and normocapnia), OWA-ocean warming and acidification (acclimated at 15°C and hypercapnia). (b) Significance Analysis of Microarray (SAM) plot of gill tissue. Scatter plot showing observed relative differences on the axis of ordinates against the expected relative differences estimated by data permutation on the abscise using a delta of 0.5 (dotted lines). The green dots are highlighting significant differences and correspond to DMA and glucose availability) (Fokina et al., 2015). Laboratory studies indicate that lipid accumulation is a direct response to temperature in mussels (Fokina et al., 2015;Wu et al., 2021) and might reflect a metabolic adjustment for anticipated reproduction (which requires high energy investment as well as lipid deposition into developing gametes) in mussels.
Unlike lipids, the glycogen content did not change in response to warming in mussels from GOM. M. edulis from Baltic Sea also showed no change in glycogen content during warming from 10 to 15°C and from 15 to 20°C (Wu et al., 2021). Modest hypercapnia likewise had no effect on the glycogen content of adductor muscle of mussels in our present study. Earlier studies show that impacts of temperature and OA on glycogen reserves of mussels are threshold dependent. Thus, mussels from GOM showed a marked decrease in the glycogen content of adductor muscle when exposed to higher temperature (22°C) alone and combined with modest acidification (560 µatm P CO2 , pH 7.9) (Lesser, 2016). The concentrations of adenylates and AEC in adductor muscle remained at steady-state levels in all exposures, indicating that cellular energy balance in mussels was maintained under all temperatures and OA scenarios. Overall, our findings indicate that temperature and OA used in our present study are not energetically stressful to GOM mussels (at least under ad libitum feeding conditions).
Warming alone or OWA altered the metabolite profile of M. edulis in a tissue-dependent manner. Warming from 10 to 15°C increased glucose and decreased dimethyl amine (DMA) in the gills, whereas adductor muscle metabolites were not affected. Increased glucose could reflect mobilization of energy reserves to meet increased tissue energy demand. Furthermore, warming could increase glucose levels via gluconeogenesis by channeling of amino acids like serine, alanine, and glycine into pyruvate and increased activity of the enzyme phosphoenolpyruvate carboxykinase (PEPCK) (Ellis et al., 2014;Le Moullac et al., 2007). In our study, we saw a trend for decreased glycine and serine in gills under warming and OWA, suggesting a potential for increased flux through gluconeogenesis that warrants further investigation. The decrease in DMA, a common organic osmolyte found in gills of bivalves (Zhang et al., 2011), might osmotically compensate for elevated glucose in M. edulis gills during warming.
In M. edulis, OA had no effect on the metabolite profile in gills or muscle tissue. Similarly, metabolite profiling studies with P CO2 ≤ ~1,000 µatm (pH ≥ ~7.8) reported no OA-induced alteration in metabolite levels of bivalves (Dickinson et al., 2012;Ellis et al., 2014;Wei et al., 2015) whereas higher P CO2 (≥1,500 µatm, pH ≤ ~7.7) led to a shift in metabolite profiles with alterations in energy metabolism (Ellis et al., 2014;Lannig et al., 2010;Wei et al., 2015). These findings are consistent with the notion that modest OA is not a metabolic F I G U R E 5 Carbonic anhydrase (CA) activity in tissues of Mytilus edulis. (a) CA activity in mantle edge exposed to different temperature and P CO2 . (b) Tissue-specific variation in specific activities of carbonic anhydrase with temperature.

| Effect of warming and OA on enzyme activity
In mollusks, CA plays a key role in the maintenance of acid-base homeostasis of all tissues as well as biomineralization in the mantle (Li et al., 2016;Ramesh et al., 2020;Wang et al., 2017 OA in bivalves including mussels (Ivanina et al., 2020). Similar to our findings, CA activity in the mantle of C. virginica and M. mercenaria remained unchanged after exposure to elevated P CO2 (800 µatm) for 2-15 weeks . In contrast, CA activity in mantle of M. edulis decreased after prolonged (6-months) exposure to elevated P CO2 (750 µatm) exposure (Fitzer et al., 2014). In oysters, CA accumulated along mantle edge in response to P CO2 exposure (2,622 µatm, pH 7.50), suggesting an active role of CA in ion-regulation and acidbase balance (Wang et al., 2017). However, the effect of OA on CA activity is threshold dependent. At P CO2 < 1,000 µatm mussels try to compensate for intracellular acid loads instead of decreasing their metabolism (Fitzer et al., 2014;Hüning et al., 2013;Thomsen & Melzner, 2010). Taken together, these data suggest that M. edulis in the GOM can upregulate acid-base balance contributing to their metabolic plasticity toward warming, but moderate OA has no effect on this trait.
As a consequence of maintaining acid-base balance, OA may change concentrations of H + , HCO 3 − , Ca 2+ , Mg 2+ , and Cl − in calcifiers (Fitzer et al., 2014;Ramesh et al., 2017). Ion transport is an important contributor to energy budget of biomineralization because Ca 2+ transport and removal of excess protons from the site of biomineralization are ATP-dependent (Ivanina et al., 2020). In GOM mussels, activity of Ca + -ATPase and H 2+ -ATPase-in the mantle remained stable under moderate warming and OA (except for a modest but significant decline in Ca 2+ -ATPase under OWA). This indicates that the mussels can maintain ion regulatory fluxes at least in the mantle edge despite variations in temperature and P CO2 relevant to nearfuture climate change. Mussels from other environments acidified like Kiel Fjord (pH < 7.5) also build and maintain their shells despite fluctuating P CO2 and pH, partially owing to enhanced ion transport . Furthermore, when mussel larvae were raised under OA between 500 and 1,500 µatm P CO2 , ∆H + at the calcification site when compared to seawater remained constant, irrespective of P CO2 . In C. virginica and M. mercenaria activities of Ca 2+ ATPase and H + ATPase, as well as the cellular energy costs of Ca 2+ and H + transport in the biomineralizing cells (mantle and hemocytes) were insensitive to ocean acidification (pH 7.8) (Ivanina et al., 2020). This indicates that intertidal species (such as mussels, oysters, and clams) that are adapted to variable temperature and pH in their habitat are generally tolerant against moderate warming and OA, predicted by the climate change models.

| CON CLUS IONS
In this study, we report that adult M. edulis from GOM are sensitive to warming but tolerant to moderate acidification scenario predicted by IPCC for the year 2100. This result also provides an insight in the natural history of GOM mussels given that in the last decade (2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014) GOM was characterized by an extreme warming trend (Salisbury & Jönsson, 2018). Although this study is limited to adults and does not consider larval stage sensitivity of M. edulis, our results support earlier reports that acidification scenarios for the next 100-300 years do not affect this species (Telesca et al., 2019). Taken together, our study provides important data about extant levels of plasticity in physiology of mussels as well as insights into potential sensitivity of mussels to future global change.

ACK N OWLED G M ENTS
The authors thank Dr. Markus Frederich (University of New England) for assistance with mussel collection in Gulf of Maine. This work was supported by the US National Science Foundation (NSF) award IOS-1557870 to I.M.S.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest. funding acquisition (lead); methodology (equal); project administration (lead); resources (lead); writing -review and editing (supporting).