ABSTRACT
In fish, the capacity of thermal acclimation to preserve cardiac mitochondrial function under future warming scenarios is important to understand given the central roles that cardiac energy metabolism and performance play in this taxa's thermal tolerance. We acclimated Atlantic salmon to 12 and 20°C (for >2 months), and investigated the effects of acute and chronic warming on cardiac mitochondrial respiration and reactive oxygen species (ROS) production (release rate) using high-resolution fluorespirometry. Further, we compared the sensitivity of mitochondrial respiration to nitric oxide (i.e. the NO IC50), and assessed the mitochondrial response to anoxia–reoxygenation (AR). Acute exposure to 20°C increased maximal mitochondrial respiration by ∼55%; however, the mitochondria's complex I respiratory control ratio was 17% lower and ROS production was increased by ≥60%. Acclimation to 20°C: (1) preserved mitochondrial coupling and aerobic capacity; (2) decreased the mitochondria's ROS production by ∼30%; (3) increased the mitochondria's NO IC50 by ∼23%; and (4) improved mitochondrial membrane integrity at 20°C. AR did not affect mitochondrial function at 12°C, but acute exposure to 20°C and AR depressed maximal mitochondrial respiration (by ∼9%) and coupling (by ∼16%) without impacting ROS production. Finally, warm acclimation did not improve the capacity of mitochondria to recover from AR, indicating that there was no ‘cross-tolerance’ between these challenges. Our findings provide compelling evidence that thermal plasticity of cardiac mitochondrial function contributes to the Atlantic salmon's capability to survive at ≥20°C for prolonged periods, but call into question whether this plasticity may allow them to withstand high temperatures when combined with other stressors.
INTRODUCTION
With climate warming, temperatures in many aquatic environments may reach levels that have direct negative effects on fish physiological performance (Sandblom et al., 2016 and references therein). As a fish's critical thermal maximum is approached, one of the first organs to fail is the heart (Farrell, 2002; Iftikar and Hickey, 2013; Iftikar et al., 2014; Haverinen and Vornanen, 2020), making cardiac function a primary determinant of fish thermal tolerance (Eliason and Anttila, 2017). Cardiac functionality is highly dependent on ATP production and, therefore, on the capacity of mitochondria to produce ATP (Rodnick and Gesser, 2017). Impairment of cardiac mitochondrial function at elevated temperatures is a threat to myocardial oxidative function, and the importance of mitochondrial plasticity in limiting cardiac dysfunction (i.e. heart failure) and shaping thermal tolerance has been suggested by several studies (Iftikar and Hickey, 2013; Iftikar et al., 2014; Chung et al., 2017; Christen et al., 2018; O'Brien et al., 2018). Acute warming first results in an increase in oxidative phosphorylation (OXPHOS) that enables tissues to meet their higher metabolic demands and to sustain cellular energy (ATP) production. However, mitochondrial capacity becomes compromised (i.e. plateaus or declines) at temperatures close to the fish's upper thermal limit. This is due to alterations in crucial mitochondrial processes during acute warming such as the rate of proton leak, coupling efficiency (respiratory control ratio; RCR), the integrity of mitochondrial membranes and the functionality of protein complexes (notably complex I; Hilton et al., 2010; Lemieux et al., 2010; Iftikar and Hickey, 2013; Iftikar et al., 2014; Rodnick et al., 2014; Christen et al., 2018; O'Brien et al., 2018). The effect of chronic warming on mitochondrial physiology is less understood, but long-term acclimation under laboratory conditions may help to improve our understanding of the plasticity of cardiac mitochondrial function and its role in fish thermal tolerance (Iftikar et al., 2015; Baris et al., 2016; Chung et al., 2017; Leo et al., 2017; Pichaud et al., 2017; Howald et al., 2019; Pichaud et al., 2019). Moreover, the effect of thermal acclimation on the oxidative status (e.g. the level of reactive oxygen species production) experienced by cardiac mitochondria is largely unknown, despite the important role of oxidative processes in defining a fish's responses to environmental change (Birnie-Gauvin et al., 2017). An increase in ROS production during acute warming was reported for permeabilized cardiac fibers from a New Zealand wrasse (Notolabrus celidotus; Iftikar and Hickey, 2013) and isolated cardiac mitochondria of Arctic char (Salvelinus alpinus; Christen et al., 2018). Nevertheless, the effect of chronic exposure to increased temperatures on reactive oxygen species (ROS) production rates by cardiac mitochondria has only been investigated in permeabilized cardiac fibers of N. celidotus, and was comparable between 15°C- and 21°C-accclimated animals (Iftikar et al., 2015). In addition, although the combined effects of warming (acute or chronic) and O2 limitation on cardiac mitochondrial function have been explored, most data on how these combined factors affect fish mitochondrial function have been conducted using the liver as the model tissue (e.g. Onukwufor et al., 2016; Sappal et al., 2015a,b).
Nitric oxide (NO) plays an important role in adjustments of cardiac function and aerobic metabolism (Fago and Jensen, 2015; Imbrogno et al., 2018) by inducing vasodilation (Fago et al., 2012), reversibly inhibiting mitochondrial respiration (Gerber et al., 2019) and controlling breathing (Perry et al., 2016) in fish. However, the role of NO in regulating mitochondrial respiration when ectotherms are warmed, and tissues likely experience hypoxic stress, has not been examined. It is possible that the widespread actions of NO interfere with mitochondrial bioenergetic processes during thermal acclimation. For instance, we recently showed that mitochondrial sensitivity to inhibition by NO was altered by hypoxic acclimation (Gerber et al., 2019), a link between thermosensitivity and responsiveness to NO has been suggested in homeotherms (Simon, 1998), and the involvement of NO in heat defense (i.e. the control of breathing and the related hypoxic ventilatory response) has been demonstrated in rats (Nakano et al., 2001).
The present study assessed the thermal dependence and plasticity of cardiac mitochondrial function when Atlantic salmon were warmed to 20°C and held at this temperature for approximately 2 months. We hypothesized that: (1) although several mitochondrial processes and/or functions would be compromised in this species by acute warming, they would be largely compensated for by warm acclimation; and (2) that the role of NO in regulating mitochondrial respiration would be altered by acclimation to high temperatures. To test our hypotheses, we measured isolated cardiac mitochondrial function following in vivo acclimation to 20°C (for 62–72 days), and in vitro acute warming (from 12°C to 20°C) alone or in combination with an in vitro anoxia–reoxygenation (AR) event (i.e. the absence of oxygen for 10 min). We measured mitochondrial O2 consumption, the thermal sensitivity of respiration (Q10), the coupling and efficiency of oxidative phosphorylation (RCR and P:O ratio), outer membrane integrity using cytochrome c and ROS production/release rate (estimated as H2O2 flux) when complex I alone (owing to its known thermal sensitivity) and both complexes I+II were stimulated. In addition, we investigated whether/how warm acclimation altered oxidative capacity [by measuring citrate synthase (CS) activity], antioxidant defense capacity [by measuring superoxide dismutase (SOD) activity] and mitochondrial sensitivity to inhibition by NO (by measuring NO IC50).
The Atlantic salmon was chosen as a model to study the thermal sensitivity of cardiac mitochondria because many facets of its cardiac response to environmental challenges are well characterized (Anttila et al., 2013, 2014, 2015; Penney et al., 2014; Hvas et al., 2017; Leeuwis et al., 2019). Considerable cardiac and aerobic plasticity has been observed in Atlantic salmon during thermal stress (Anttila et al., 2014; Hvas et al., 2017). Finally, the 12°C-acclimated salmon used in this study have a critical thermal maximum (CTmax) of ∼25.8°C, the temperature where cardiac arrhythmias (fHmax) appear is 22.9°C (A.K.G., K.A.C., R. M. Sandrelli, L.G. and E. F. C. Peroni, unpublished data) and salmon exposed to an incremental temperature increase of 1°C per week begin dying at 22°C (Gamperl et al., 2020). Thus, the temperature used for warm acclimation (i.e. 20°C) is close to, but just below, this species' CTmax, fHmax and incremental temperature tolerance (ITmax).
MATERIALS AND METHODS
Experimental animals and temperature acclimation
Atlantic salmon were obtained from Cape d'Or Sustainable Seafood Inc. (Advocate Harbour, Nova Scotia, Canada) and initially maintained at the Dr Joe Brown Aquatic Research Building [JBARB, Memorial University of Newfoundland (MUN)] in a 4000 liter tank supplied with aerated seawater (32 ppt) at 10±1°C and a 12 h:12 h light:dark photoperiod for 1 month. Thereafter, the fish were randomly distributed into four 2200 liter circular tanks (∼30 fish per tank) receiving 12±1°C aerated seawater and a 14 h:10 h light:dark photoperiod at the Laboratory for Atlantic Salmon and Climate Change Research (LASCCR, Ocean Sciences Centre, MUN). After 2 weeks of acclimation to the tanks, two control tanks were maintained at 12°C, while the temperature was gradually increased in the two other tanks (by ∼0.25°C every day) until the temperature reached 20°C. The tanks were then maintained at 12 or 20°C for 2 months before sampling (the experiments began). Fish were fed a commercial marine fish diet (Europa; Skretting Inc.) at ∼1.5% body mass day−1 during this period.
The experiments below were completed within 10 days; hence the salmon were acclimated between 62 and 72 days to 12 or 20°C. The experiments were performed on pairs of animals daily, i.e. one 12°C- and one 20°C-acclimated fish.
All experimental procedures followed guidelines established by the Canadian Council on Animal Care and were approved by the Institutional Animal Care Committee of MUN (protocol 16-92-KG).
Mitochondrial isolation
We used the protocol outlined in Gerber et al. (2019) with the exception that the composition of the respiration medium was modified by replacing 2 mmol l−1 EDTA with 1 mmol l−1 EGTA. Briefly, fish were killed by a blow to the head before measuring their fork length and body mass. Then, the ventricle was excised, cut in half, rinsed briefly with ice-cold isolation medium (in mmol l−1: 230 mannitol; 75 sucrose; 20 HEPES; 1 EGTA, pH 7.4), blotted dry and weighed. A small piece of ventricle was removed, weighed and snap-frozen in liquid nitrogen prior to being kept at −80°C for later use in enzymatic assays. The rest of the ventricle was thoroughly minced and split into four aliquots to facilitate homogenization. Each aliquot was re-suspended in three volumes of isolation medium in an ice-cooled glass-homogenizer, and gently homogenized using six passes of a loose-fitting motor-driven Teflon pestle. The crude homogenates were then combined and centrifuged at 800 g for 10 min at 4°C to remove cell debris, and the resulting supernatant was centrifuged at 8000 g for 10 min at 4°C to pellet the mitochondria. The lipid layer and the supernatant were carefully removed, and the resulting mitochondrial pellet was washed twice by gentle re-suspension in ice-cold isolation medium containing 10 mg ml−1 of BSA (fatty acid free; Sigma Chemical Co., CAS 9048-46-8), and then centrifuged at 8000 g for 10 min at 4°C. This final pellet was weighed, and then gently re-suspended in four volumes of ice-cold respiration medium [in mmol l−1: 160 KCl; 30 HEPES; 10 KH2PO4; 1 EGTA; 10 mg ml−1 BSA (fatty acid free), pH 7.4]. The protein content of the mitochondrial suspensions was assayed using the Bradford assay with BSA as a standard (Thermo Fisher Scientific, Waltham, MA, USA), and the contribution of BSA to the measured protein levels was accounted for using respiration medium with BSA as a blank. Mitochondrial suspensions were placed on ice for 30 min to rest before being used for high-resolution respirometry.
Experimental protocols
Mitochondrial physiology
The physiology of isolated cardiac mitochondria from 12°C- and 20°C-acclimated salmon was assessed using high-resolution respirometry (O2k-polarographic O2 sensor, Oroboros Instruments, Innsbruck, Austria), fluorometry (O2k-Fluo LED2-module, Oroboros Instruments), amperometry (O2k-MultiSensor system, Oroboros Instruments) and an ISO-NOP NO sensor (World Precision Instruments, Sarasota, FL, USA). We conducted two separate experiments at two assay temperatures (12°C and 20°C) on the mitochondria from 12°C- and 20°C-acclimated fish, which are described in detail below and shown in Figs S1 and S2. The experimental duration varied between 60 and 120 min. O2k-sensor calibration and correction for background oxygen flux were performed at each assay temperature (12 and 20°C) following standard operating procedures (O2k-SOP; Fasching and Gnaiger, 2018; Gnaiger, 2018), and the oxygen solubility factor (FM) of the KCl experimental medium at 12°C and 20°C was used to ensure accurate oxygen concentration measurements (O2k-SOP; Gnaiger, 2010). All parameters were recorded in real-time using DatLab v 7.3 software (Oroboros Instruments). In each experiment, mitochondrial suspensions were injected into calibrated 2 ml O2k-chambers containing 12°C or 20°C respiration medium that was 100% air-equilibrated.
Measurement of coupled mitochondrial respiration and ROS production/release rates
In experiment 1, we measured cardiac mitochondrial respiration and ROS production/release rates simultaneously. ROS production/release rates were estimated by measuring extra-mitochondrial H2O2 using Amplex® UltraRed (AmR, 10 µmol l−1), horseradish peroxidase (HRP, 3 U ml−1), SOD (25 U ml−1) and the Green Fluorescence-Sensor of the O2k-Fluo LED2-Module (with gain and LED intensity set to 1000 and 500 mV, respectively). The ROS signal was calibrated by the addition of H2O2 (0.1 μmol l−1) following standard operating procedures (O2k-SOP; Krumschnabel et al., 2019; Makrecka-Kuka et al., 2015). In experiment 1 (see Fig. S1A), mitochondria (0.25 mg protein ml−1) were added to 2 ml of 100% air-equilibrated respiration medium at 12°C or 20°C. Then the complex I (CI) substrates glutamate (15 mmol l−1) and malate (2 mmol l−1), and ADP (200 μmol l−1), were added to fuel the electron transport system (ETS) prior to measuring oxidative phosphorylation (OXPHOS, State 3 respiration) via complex I (PGM) with a second addition of ADP (200 μmol l−1). After the depletion of ADP, leak respiration (State 4) via CI (LT,GM) was measured. Thereafter, succinate (5 mmol l−1) was added, followed by ADP (200 μmol l−1) to measure OXPHOS respiration (State 3) via CI+II (PGMS). After the depletion of ADP, leak respiration rate (State 4) via CI+CII (LT,GMS) was measured. Then, the effect of a 10 min anoxic bout on mitochondrial respiration and ROS release rates via CI+II was measured. To decrease the O2 level in the chamber, an excess of ADP (700 μmol l−1) was added, and the mitochondria consumed all the ADP and returned to State 4 before they became anoxic. After 10 min of anoxia, the chamber oxygen level was returned to pre-anoxic levels by lifting the chamber stoppers. Then, ADP (200 μmol l−1) was added to measure OXPHOS respiration and leak respiration post-anoxia. This protocol also allowed for the determination of mitochondrial P50 [i.e. partial oxygen pressure of oxygen (in kPa) at which the respiration rate was half of maximal] during leak respiration.
Measurement of the inhibition of mitochondrial respiration by exogenous NO
In experiment 2 (see Fig. S2A), we measured ventricular mitochondrial respiration and NO concentration in the chambers simultaneously. An NO sensor (ISO-NOP, World Precision Instruments) was connected to one of the amperometric channels of the O2k-MultiSensor system following a standard protocol (O2k-SOP; Fasching and Gnaiger, 2016). The NO sensors were maintained in a polarized state using a pre-polarizer (NSA-3, World Precision Instruments) to allow for quick stabilization of the signal once connected to the O2k instrument. The NO sensors were calibrated as described elsewhere (Gerber et al., 2019) and following the manufacturer’s recommendations (WPI instruction manual, Calibration of NO Sensor by Decomposition of SNAP – Method 2). Briefly, the O2k chamber windows were covered with custom-made 3D-printed caps and the lights inside the chambers were turned off. The NO sensors were calibrated daily at 12°C or 20°C in O2k chambers filled with 2 ml of 0.1 mol l−1 CuCl2 solution by the sequential addition of known SNAP concentrations (stock solution 0.05 mmol l−1) over a range of 0–1000 nmol l−1 of NO using the TIP2k syringe. Following calibration, the chamber was thoroughly rinsed and filled with respiration medium, and the NO sensors were allowed 30 min to stabilize under these new conditions (respiration medium at 12°C or 20°C). Then, the mitochondria's sensitivity to NO (NO IC50) was measured using the protocol described in Gerber et al. (2019). Briefly, in this experiment (Fig. S2A), mitochondria (0.25 mg protein ml−1) were added to 100% air-equilibrated 2 ml-calibrated chambers filled with respiration medium at 12°C or 20°C. Substrates of CI and CII (glutamate, 15 mmol l−1; malate, 2 mmol l−1; and succinate, 5 mmol l−1) were added, followed by ADP (200 μmol l−1), to fuel the ETS and assess mitochondrial coupling. Then, an excess amount of ADP (1 mmol l−1) was added, and the NO donor PAPANONOate (20 and 10 μmol l−1 at 12°C and 20°C, respectively) was added to the chamber at ∼60% air saturation to initiate NO release at a rate of ∼20 nmol min−1. This slow increase in NO levels in the chambers gradually inhibited respiration, and allowed for the determination of mitochondrial NO IC50 (the concentration of NO that inhibited 50% of maximal respiration). Once the maximal respiration rate was inhibited by ∼75%, and the O2 concentration in the chamber had plateaued, excess oxyhemoglobin [OxyHb, 20 µmol l−1, see Gerber et al. (2019) for preparation of OxyHb] was added to reverse NO's inhibition of respiration and to confirm that the effect observed was due to the release of NO and not to the addition of the NO donor molecule (PAPANONOate) per se. Excess OxyHb scavenged all the NO and stopped the PAPANONOate-mediated NO release, allowing for calculation of the NO sensors’ zero drift. Reversal of NO inhibition allowed mitochondrial OXPHOS respiration to be re-established, and once leak respiration was established, the chamber oxygen level was returned to 60% air saturation by lifting the chamber stoppers, and OXPHOS respiration was initiated by the addition of excess ADP (1 mmol l−1). This served as a control for comparison with OXPHOS respiration in the presence and absence of NO (not shown). Cytochrome c (10 μmol l−1) was also added during the plateau at maximal respiration to assess outer mitochondrial membrane integrity. PAPANONOate (20 mmol l−1) was prepared daily, and immediately prior use, in ice-cold respiration medium (without BSA) for each experiment. Recordings of NO concentration were corrected for baseline and zero drift before data analysis.
Calculation of mitochondrial respiration coupling, efficiency, integrity and thermal sensitivity
Respiratory control (RCR; i.e. State 3/State 4) and P:O (i.e. [ADPinjected in µmol]/[O2 consumed in µ-atoms]) ratios with CI and CI+II substrates were calculated to evaluate mitochondrial coupling and the efficiency of ATP synthesis, respectively. The effect of cytochrome c on mitochondrial respiration [i.e. (CI+CII OXPHOSCyt c−CI+CII OXPHOS)/CI+CII OXPHOSCyt c] and the temperature coefficient [Q10; i.e. (R2/R1)^{10/(T_2-T_1)}] were calculated to evaluate the integrity of the outer mitochondrial membrane and the thermal sensitivity of mitochondrial respiration, respectively.
Enzyme assays
Enzymatic activities were measured at room temperature (25°C) using a 96-well microplate reader (SpectraMax M2e, Molecular Devices, Sunnyvale, CA, USA).
CS activity was determined on both ventricle homogenates and mitochondrial suspensions from both 12°C- and 20°C-acclimated fish using the method described by Treberg et al. (2007), and used as an indicator of mitochondrial density and aerobic (oxidative) capacity. A snap-frozen piece of ventricle was thoroughly minced, re-suspended in four volumes of ice-cold homogenization buffer (25 mmol l−1 HEPES, 2 mmol l−1 EDTA and 0.5% Triton X-100, pH 7.0) and homogenized using a Polytron homogenizer (Kinematica Inc., Bohemia, NY, USA). The resulting crude homogenate and thawed mitochondrial suspensions were further diluted in 2 and 19 volumes of homogenization buffer, respectively, and centrifuged at 2000 g for 5 min at 4°C. CS activities for the ventricle homogenates and mitochondrial suspensions were expressed as U g tissue−1 and as U mg protein−1, respectively.
SOD activity was assessed using the SOD Assay Kit-WST (Sigma Aldrich) following the manufacturer's protocol. The thawed mitochondrial suspensions were diluted in 9 volumes of homogenization buffer to lyse the mitochondrial membrane. Then, the thawed mitochondrial suspensions were centrifuged at 13,000 g, and the resulting supernatant was further diluted in 13 volumes of dilution buffer (SOD Kit). Using the WST-1 inhibition curve and SOD standards, the concentration of SOD (in U mg−1 mitochondrial protein) in the mitochondrial suspensions was determined.
Statistical analyses
Statistical analyses were performed using Prism 8 software (GraphPad Inc., La Jolla, CA, USA). Effects of complex (or anoxia, Fig. 3), acclimation and assay temperature on all mitochondrial respiration data were assessed using repeated measures three-way ANOVAs. Groups that differed by only one factor (Sidak's multiple comparison tests) are shown in the graphs, whereas P-values and significant interactions between main effects are shown in Table 1. The NO IC50, mito-P50, Q10 and enzymatic activity data were analyzed using repeated measures two-way ANOVAs, followed by Sidak's multiple comparison tests as shown in Table 2 or described in the Results. The morphometric data were analyzed using unpaired two-tailed t-tests. The level of statistical significance for all analyses was P<0.05, and all data are presented as means±s.e.m.
RESULTS
Morphometric variables
There was no significant difference in body mass (831.3±28.9 versus 898.6±60.2 g; t-test, F=4.311, P=0.3411) or in relative ventricular mass (0.066±0.003 versus 0.071±0.00; t-test, F=1.713, P=0.2126) between 12°C- and 20°C-acclimated fish. There was also no difference in cardiac mitochondrial yield, the ratio of mitochondrial pellet mass to ventricular wet mass (0.049±0.002 versus 0.047±0.002; t-test, F=0.8024, P=0.4024) or the protein concentration of the mitochondrial suspension (50.44±3.6 mg ml−1 versus 51.73±2.6 mg ml−1; Mann–Whitney test, P=0.7308) between 12°C- and 20°C-acclimated fish.
Effect of acute and chronic warming on mitochondrial respiration, coupling, efficiency, thermal sensitivity and integrity
Respiration rates during OXPHOS (State 3; Fig. 1A) and leak states (State 4; Fig. 1B) were higher when measured at 20°C as compared with 12°C, and when using substrates of CI+CII as compared with CI alone (by approximately 55 and 85%, respectively; Fig. 1, Table 1). In addition, although acclimation to 20 versus 12°C did not significantly alter mitochondrial respiration rates at 20°C, there were subtle differences in the thermal sensitivity of respiration between acclimation groups and between complexes (i.e. I versus I and II) when expressed as Q10 values. For example, 12°C-acclimated salmon had higher (P<0.05) Q10 values for leak respiration than 20°C-acclimated fish (∼2.25 versus 2.07), but lower Q10 values for OXPHOS respiration (∼1.76 versus 1.93–2.23; Table 2). In 20°C-acclimated fish, the Q10 value for CI-OXPHOS (2.23±0.45) was significantly greater than for CI+CII-OXPHOS (1.93±0.34; Table 2).
With regards to mitochondrial coupling (i.e. RCR values; Fig. 1C) and efficiency (i.e. P:O ratio; Fig. 1D), the biggest influence was that of the mitochondrial complexes; RCR and P:O values for complex I+II were approximately ∼30–40% (∼6–7 versus ∼9–12, respectively) and ∼15% (∼2.3–2.4 versus ∼2.6–2.8, respectively) lower as compared with complex I. There was also an overall effect of test temperature on mitochondrial coupling, and of acclimation temperature on the P:O ratio (see Table 1). RCR values measured at 20°C were lower than those at 12°C; this effect was most notable in 12°C-acclimated fish for CI-fueled respiration (Fig. 1C). The P:O ratio was higher in 12°C-acclimated fish as compared with 20°C-acclimated fish, an effect that was most notable in 12°C-acclimated fish for CI-fueled respiration at 20°C (Fig. 1D). There was no effect of acclimation or test temperature on the PO2 at which leak respiration rate was half of maximal (i.e. the P50), with mito-P50 values of ∼0.04 kPa for all groups (data not shown).
Effect of acute and chronic warming on ROS release rates
There was an overall effect of acclimation temperature on ROS production/release rate during OXPHOS (Fig. 2A), but not for leak respiration (Fig. 2B). This acclimation effect was most notable during CI+CII OXPHOS at 20°C, both when measured as ROS production/release rate (Fig. 2A) and when expressed as a percentage of O2 flux (%ROS/O2; Fig. 2C). With regards to the latter, the mitochondrial ROS production/release rate from 20°C-acclimated fish was ∼30% lower as compared with 12°C-acclimated fish (Fig. 2A,C). In both acclimation groups, ROS release rates during both OXPHOS (Fig. 2A) and leak (Fig. 2B) states were higher (by ∼55–78% and ∼52–62%) when measured at 20°C than at 12°C. Although adding the CII substrate succinate increased ROS production/release rate during leak respiration (Fig. 2B; by ∼90–111%) in both groups and at both test temperatures, it decreased %ROS/O2 flux during OXPHOS in all groups except 20°C-acclimated fish at 20°C (Fig. 2C).
Effect of AR on mitochondrial respiration, ROS production/release rate and coupling
Acclimation did not alter the mitochondrial response to AR (Fig. 3). However, temperature and anoxic exposure had interactive effects on mitochondrial respiration. For example, although anoxic exposure decreased OXPHOS (Fig. 3A; by ∼9–16%) and increased leak respiration (Fig. 3B; by 12%) at 20°C in both groups, it had no effect when the mitochondria were tested at 12°C. Only a limited effect of AR on mitochondrial ROS production was observed when measured as ROS release rate (Fig. 3C,D). However, when expressed as a percentage of O2 flux (i.e. %ROS/O2 flux; Fig. 3E,F), mitochondrial ROS production was increased consistently by anoxic exposure in the OXPHOS state (this effect most notable in 20°C-acclimated fish tested at 20°C). In contrast, ROS production/release rate was generally decreased post-anoxia in the leak state (this effect was most notable for 12°C-acclimated fish at both test temperatures). Finally, RCR values were significantly lower (by ∼10–20%) post-anoxia as compared with pre-anoxia in all groups, with the exception of 20°C-acclimated fish tested at 12°C (Fig. 3G).
Effect of thermal acclimation on mitochondrial sensitivity to nitric oxide
The release of NO (at ∼20 nmol l−1 min−1) using the NO donor PAPANONOate progressively inhibited CI+CII stimulated cardiac mitochondrial OXPHOS respiration in both acclimation groups and at both test temperatures (12°C and 20°C; Fig. 4A and Fig. S2). The concentration of NO that inhibited 50% of maximal mitochondrial respiration (i.e. the mitochondrial sensitivity to NO or NO IC50) was not affected by acclimation when measured at 12°C. However, it was higher in 20°C-acclimated fish as compared with 12°C-acclimated fish when tested at 20°C (71.53 versus 54.73 nmol l−1, respectively; post hoc t=2.497, P=0.0392).
Enzymatic activity
Thermal acclimation did not alter CS activities in the myocardium (ventricle) or in the mitochondria themselves, and mitochondrial SOD activities in the mitochondria were not different (Table 3).
Mitochondrial integrity
The integrity of the mitochondrial outer membrane was altered by test temperature (two-way ANOVA, F=9.137, P=0.0116), but not by acclimation temperature (two-way ANOVA, F=0.4310, P=0.5250). For example, although a comparable increase in OXPHOS was reported after the addition of cytochrome c in 12°C- and 20°C-acclimated fish at 12°C (∼7.2%), at 20°C, mitochondria from 20°C-acclimated salmon displayed the lowest cytochrome c-induced increase in oxygen consumption (∼3.6%; post hoc t=3.158, P=0.0181; Fig. 5).
DISCUSSION
There is growing evidence that mitochondrial capacity may limit cardiac performance and, hence, play a major role in determining fish thermal tolerance (Iftikar and Hickey, 2013; Iftikar et al., 2014; Penney et al., 2014; Christen et al., 2018; O'Brien et al., 2018; Little et al., 2020). Thus, we investigated the response(s) of Atlantic salmon cardiac mitochondria to acute and chronic warming (to 20°C). Our results provide evidence that Atlantic salmon cardiac mitochondria can withstand acute and chronic warming to 20°C, and that this likely contributes to this species' capability for cardiac acclimation to warm temperatures (Anttila et al., 2014) and their high aerobic capacity across a wide range of temperatures (Hvas et al., 2017). Further, it is noteworthy that while acute warming increased ROS production/release rate and decreased RCR, these heat-induced effects were largely mitigated by warm acclimation to 20°C. Overall, our findings support the idea that thermal plasticity of cardiac mitochondrial function contributes to shaping fish cardiac performance and thermal tolerance.
Effects of acute warming to 20°C on cardiac mitochondrial function
Acute warming leads to an increase in physiological and biological rates, and the mitochondrial response to warming plays a key role in meeting the temperature-dependent increase in cardiac energy metabolism and function (Iftikar and Hickey, 2013; Iftikar et al., 2014, 2015; Ekström et al., 2016, 2017; Little et al., 2020). As expected, cardiac mitochondrial respiration was higher at 20°C as compared with 12°C (Fig. 1), and the magnitude of the increase was comparable to that reported in other fish species using permeabilized cardiac fibers (Chung et al., 2017; Hilton et al., 2010; Howald et al., 2019; Iftikar and Hickey, 2013; Iftikar et al., 2014) or isolated mitochondria (Christen et al., 2018; Ekström et al., 2017; Lemieux et al., 2010; O'Brien et al., 2018; Rodnick et al., 2014). Furthermore, the Q10 of 1.75 (for CI+CII OXPHOS between 12°C and 20°C; see Table 2 for details) during acute warming is similar to previously reported values for cardiac performance (i.e. heart rate, Q10 of 1.87 from 12 to 20°C; Anttila et al., 2014) and whole-organism O2 consumption (ṀO2, Q10 of 1.79 from 12 to 20°C; Leeuwis et al., 2019) in this species. This strongly suggests that cardiac mitochondrial function is finely tuned to meet the increased energy demands of the heart and whole organism when faced with elevated temperatures. Yet, some of the parameters measured suggest that mitochondrial functional integrity, and hence the mitochondria's capacity to efficiently produce ATP, was impaired. Most notable was the decrease in CI-dependent RCR following acute warming, which indicates a loss of CI coupling (Fig. 1C). This lower RCR in acutely warmed fish was driven primarily by an increase in leak respiration that was not compensated for by the increase in OXPHOS respiration (Fig. 1A,B), and this agrees with previous studies on fish cardiac mitochondria indicating that CI is a thermally sensitive complex and a putative limiting step of the ETS (Iftikar and Hickey, 2013; Iftikar et al., 2014, 2015; Ekström et al., 2017; Christen et al., 2018). The impairment of CI's capacity for electron flow in response to acute warming could result from a decreased mitochondrial capacity for NADH production that constrains mitochondrial ATP production from NADH-linked substrates. Chung and Schulte (2015) showed that suppression of mitochondrial respiration in the liver of killifish (Fundulus heteroclitus) when acclimated to high temperatures (33°C) was primarily caused by lower respiration rates through CI. However, the RCR of CI+II was not lower when 12°C fish were acutely exposed to 20°C (Fig. 1C), and thus, it appears that CII may exert a higher level of control on mitochondrial respiration during acute warming, and result in mitochondrial function and integrity being fairly independent of temperature from 12 to 20°C in this species. This hypothesis is supported by the unaltered P:O ratios when mitochondria from 12°C-acclimated fish were acutely warmed to 20°C. This could underlie a temporary compensatory response in an attempt to conserve metabolic efficiency despite the loss of CI coupling during acute warming.
Another notable mitochondrial response to acute warming was a large (∼65–78%) increase in absolute ROS release rate (Fig. 2A). An increase in ROS production during acute warming has also been reported in permeabilized cardiac fibers of wrasse from New Zealand (Iftikar and Hickey, 2013) and isolated cardiac mitochondria of Arctic char (Christen et al., 2018), although the increase in the latter species was not until the temperature reached 25°C. The elevated level of ROS production could be due to an imbalance between ROS production rates and antioxidant defense capacities, and possibly lead to the opening of the mitochondrial permeability transition pore (Murphy, 2009). The latter would increase inner membrane permeability, a hypothesis supported by the higher fractional increase in CI+CII respiration following the addition of NADH when mitochondria from 12°C-acclimated salmon were exposed to 20°C (Gerber et al., 2020).
Influence of acclimation to 20°C on cardiac physiology
Relative ventricular mass
Relative ventricular mass (RVM) does not change (Klaiman et al., 2011; Hevrøy et al., 2013, 2015; Keen et al., 2016), or decreases slightly (Anttila et al., 2015), when fish are acclimated to temperatures between 15 and 19°C. These data agree with the present study, and the results of Gamperl et al. (2020), who showed that the RVM of Atlantic salmon was not significantly affected by acclimation to 20°C for periods of up to a month. However, these authors also showed that RVM does increase when Atlantic salmon were pushed further (i.e. to 23°C). Overall, these data suggest that changes in RVM in this species are only seen when temperatures approach their maximum rearing temperature, where an increase in heart (ventricular mass) may help to compensate for the negative effects of high temperature on cardiac pumping capacity.
Mitochondrial respiration, efficiency and ROS production
Although the response of cardiac mitochondria to acute heat stress has been investigated in a number of studies, only limited and very recent information is available on how cardiac mitochondrial function might respond to chronic warming (Iftikar et al., 2015; Baris et al., 2016; Chung et al., 2017; Ekström et al., 2017; Leo et al., 2017; Pichaud et al., 2017, 2019; Howald et al., 2019). In the present study, we acclimated Atlantic salmon for 62–72 days to 20°C, and show that chronic exposure to 20°C mitigated some of the negative effects of an acute change from 12 to 20°C on mitochondrial function. For example, as compared with mitochondria from 12°C-acclimated fish acutely warmed to 20°C, 20°C-acclimated fish had a higher CI-RCR value, a higher Q10 for CI between 12 and 20°C, lower CI and CI+CII ROS production, and the lowest fractional increase in CI+CII respiration after the addition of cytochrome c (indicative of better outer membrane integrity) (Figs 1 and 2, Table 2). Furthermore, warm-acclimation did not alter CS activities in the myocardium or the mitochondria themselves (Table 3), suggesting that mitochondrial abundance and oxidative capacities in the heart were not altered (St-Pierre and Boutilier, 2001; Guderley and St-Pierre, 2002), consistent with the maintained mitochondrial respiratory function. Our observation that CS activity was not altered with warm acclimation is consistent with Chung et al. (2017), who showed that the activity of this enzyme was not affected by acclimating killifish to 33°C as compared with 15°C. Further, it is consistent with the results of Pichaud et al. (2017) for rainbow trout. Although these authors reported a temporary increase in heart CS activity when temperature was increased from 10 to 16°C, CS activity returned to control values after a few days of acclimation.
One benefit of warm acclimation (to 20°C) was the apparent preservation of CI functionality (i.e. maintained CI-RCR; Fig. 1 and Table 1), and this was confirmed in our subsequent study, which reports: improved mitochondrial capacity following 20°C acclimation until acute test temperature reached 26°C; and that enhanced CI functionality compensated for a loss of CII capacity induced by exposure to warm temperatures (Gerber et al., 2020). Collectively, these studies (present study; Gerber et al., 2020) support the central role of CI (i.e. CI capacity, thermal sensitivity and plasticity) in defining thermal tolerance, cardiac dysfunction (Iftikar et al., 2015; Christen et al., 2018) and acclimation potential in fish (Strobel et al., 2013). However, this acclimatory response (i.e. the enhancement of CI-dependent respiration) in cardiac mitochondria of warm-acclimated fish as compared with cold-acclimated fish is in contrast with that reported for other fish species. In another salmonid (the rainbow trout, Oncorhynchus mykiss), a loss of cardiac mitochondrial functionality was primarily due to a loss of CI coupling and oxidative capacity when this species was acclimated to 16°C versus 10°C (Pichaud et al., 2017). This disparity is likely explained by the temperature of acclimation used in Pichaud et al. (2017) versus that used here; i.e. 16°C would be considered a preferred/optimum temperature for these species, whereas 20°C is closer to their ITmax (upper thermal tolerance) (Hvas et al., 2017; Gamperl et al., 2020). However, Iftikar et al. (2015) reported that wrasse cardiac mitochondrial function at 20°C was not higher in 20°C- versus 15°C-acclimated fish when measured at 20°C. Both Baris et al. (2016) and Chung et al. (2017) found that killifish acclimated to 28–33°C versus 12–15°C had a lower OXPHOS-I respiration rate when measured at 28–38°C. Similarly, in the European perch (Perca fluvitalis), the catalytic capacity of CI was decreased by warm acclimation (Ekström et al., 2017; Pichaud et al., 2019). One potential reason that salmon show a compensatory increase in, rather than a downregulation of, mitochondrial respiration is that, unlike many other temperate teleost species, the salmon is cold-active and has a high aerobic/metabolic capacity. It is possible that such species show a different acclimatory response when exposed to temperatures approaching their thermal limits for prolonged periods as compared with species such as the wrasse, killifish and European perch. This hypothesis would be consistent with Hvas et al. (2017), who reported that salmon acclimated to 23°C had a slightly higher maximum metabolic rate at this temperature than at 18 and 15°C. However, this hypothesis needs to be examined experimentally.
It is noteworthy that mitochondrial function was not affected by acute cooling as none of the parameters measured at 12°C in the 20°C-acclimated fish differed from those of the 12°C-acclimated fish. This is in agreement with two recent studies showing that cardiac mitochondria from European perch from a chronically heated habitat and warm-acclimated European seabass did not lose their mitochondrial capacity at lower temperatures (Howald et al., 2019; Pichaud et al., 2019).
The thermal sensitivity of ROS production and its role in the response to environmental changes (Blier et al., 2014; Birnie-Gauvin et al., 2017) are still under investigation, and there is limited information on the effects of chronic warming on mitochondrial ROS production. Chung and Schulte (2015) estimated liver mitochondrial ROS production rates (under State 4; no ADP+succinate+rotenone with or without Antimycin A) and free radical leak in killifish after 4 weeks of acclimation to 5, 15 or 33°C using the same Amplex UltraRed assay that we used here (Treberg et al., 2015), but did not detect any acclimation effect on ROS production or free radical leak. Here, we estimated ROS release rates in response to acute and chronic exposure to 20°C under more experimental/respiratory conditions (i.e. with CI substrates alone and both CI+CII substrates in presence and absence of ADP), and report that a clear benefit of warm acclimation in Atlantic salmon was a significant reduction in cardiac mitochondrial ROS production (Fig. 2 and Table 1). Indeed, relative ROS production/release (per respirational O2 flux) was decreased by ∼30% at 20°C in mitochondria from warm-acclimated fish as compared with mitochondria acutely exposed to 20°C (i.e. from cold-acclimated fish) (Fig. 2). The mitigation in ROS production observed in chronically warmed fish is likely related to the preservation versus loss of CI functionality (i.e. CI coupling) reported in the present study and in Gerber et al. (2020). Indeed, CI is an important site of ROS production, and its impairment is known to induce a burst in ROS production (Murphy, 2009). A reduction in ROS release following warm acclimation could also be the result of a long-term antioxidant response (Banh et al., 2016), but no change in the mitochondrial activity of the antioxidant enzyme SOD was observed following warm acclimation (Table 3). Nonetheless, we have to acknowledge that other important antioxidant enzyme activities, such as those of catalase and glutathione peroxidase, were not assessed.
It is also well known that temperature influences mitochondrial membrane composition and fluidity (Hazel, 1995; Kraffe et al., 2007), which can affect: (1) proton leak (see Seebacher et al., 2010 for a review); (2) the oxidative capacity of mitochondria (see Guderley and St-Pierre, 2002 for a review); and (3) ROS production (see Blier et al., 2014 for a review). During thermal acclimation, considerable changes in fatty acid composition occur in the membranes of fish liver mitochondria (Strobel et al., 2013; Chung et al., 2018). Although cardiac mitochondrial membrane remodeling during chronic warming has not been investigated per se, a re-organization of the heart's lipid profile was observed in European perch following warm acclimation (Ekström et al., 2017) and chronic hypoxia has been shown to alter the lipid composition of cardiac mitochondria in the sablefish (Anoplopoma fimbria; Gerber et al., 2019). Nonetheless, although the fractional increase in O2 flux after cytochrome c (an indication of outer mitochondrial integrity) measured at 20°C was the lowest in warm-acclimated salmon in this study (Fig. 5) and that of Gerber et al. (2020) (∼7% versus 3% at 20°C in 12°C- versus 20°C-acclimated fish), the effect of NADH on O2 flux (an indication of inner mitochondrial membrane integrity) increased with test temperature and was not different (and possibly even marginally higher) in 20°C- versus 12°C-acclimated fish tested at temperatures from 20 to 28°C (Gerber et al., 2020). This latter result strongly suggests that changes in inner mitochondrial fatty acid composition (i.e. membrane integrity) were not related to the lower ROS release rates observed for 20°C-acclimated salmon.
NO effects on the sensitivity of mitochondrial respiration
Although an acute temperature increase from 12 to 20°C did not affect the sensitivity of mitochondrial respiration to NO, we found that the NO IC50 increased by 30% with warm acclimation as compared with both 12°C-acclimated fish measured at 20°C and 20°C-acclimated fish measured at 12°C (Fig. 4). Thus, this research shows that NO's modulation of mitochondrial respiration is sensitive to long-term/prolonged changes in temperature. That NOS/NO signaling in the fish heart is temperature sensitive is supported by two previous studies. In the European eel (Anguilla anguilla), NO modulation of the Frank–Starling response was evident at the fish's acclimation temperature, but not when 20°C (summer) hearts were tested at 10°C (winter) and vice versa (Amelio et al., 2013). In addition, Jørgensen et al. (2014) reported an increase in the expression of inducible NO synthase in the myocardium of Atlantic salmon when acclimated to 19°C for 3 and 8 weeks. The mechanism(s) mediating the loss of mitochondrial NO sensitivity in warm acclimated salmon is/are unknown, but could be the result of transitional changes in cytochrome c oxidase (CCO) function or in the affinity of CCO for its substrates – O2, cytochrome c and NO (see Lesnefsky et al., 2001; Kraffe et al., 2007; Hilton et al., 2010). A diminished NO sensitivity may delay NO's inhibition of mitochondrial respiration and allow oxidative phosphorylation to better satisfy the increased metabolic demands associated with heat stress. However, there is not enough information or data in the literature for us to predict whether this effect (i.e. reduction in NO sensitivity) is detrimental or beneficial with regards to prolonged exposure to elevated temperatures close to a fish's ITmax. Arguments both in favor and against this hypothesis can be envisioned given the ubiquitous role of NO in the regulation of mitochondrial signaling (Erusalimsky and Moncada, 2007; Palacios-Callender et al., 2004; Chouchani et al., 2013) and of the cardiovascular system (Moncada et al., 1991; Flögel et al., 2010; Imbrogno et al., 2018; Carenvale et al., 2020).
Warm acclimation does not lessen the relatively minor impacts of AR on salmon mitochondrial function
The effects of O2 limitation at elevated temperatures on fish mitochondrial respiration and ROS production have received very limited attention (Sappal et al., 2015a,b; Onukwufor et al., 2016). Further, no previous study has looked at the effects of both an acute temperature increase and acclimation to the same high temperature on mitochondrial function after AR. In liver mitochondria from 10°C-acclimated trout acutely tested at 5, 13 and 25°C, Onukwufor et al. (2016) reported that AR resulted in a large decrease in State 3 respiration at 13 and 25°C (by ∼30–40% after 15 min of AR), increases in State 4 respiration of ∼1.8- and 3.1-fold, respectively, and that these changes resulted in a much greater decrease in the mitochondria's RCR at 25°C versus 13°C (from ∼7 to 3 versus 2, respectively). In general, our results for cardiac mitochondria exposed to 10 min of AR agree with those of Onukwufor et al. (2016), in that the impact of AR on mitochondrial function was greater at higher temperatures. However, we show that AR at 12°C had no impact on these parameters, and that even at 20°C the effects were relatively minor (changes of ∼10–15% for all parameters) (Fig. 3). Sappal et al. (2015b) exposed liver mitochondria from 11°C- and 20°C-acclimated fish to AR for 10 min at their acclimation temperature. They showed that exposure of liver mitochondria to AR also resulted in large changes in the above three parameters in both 11°C- and 20°C-acclimated fish, and that acclimation to 20°C versus 11°C worsened the deleterious effects of AR on mitochondrial respiration. In contrast, we showed that AR at 12°C had no effects on mitochondrial respiration, and that the overall effects of AR on mitochondrial respiration at 20°C were relatively minor and not affected by acclimation temperature (Fig. 3A,B,G). These results clearly show that the response of salmon cardiac mitochondria to temperature acclimation and AR are much different than those of trout liver mitochondria (these are both salmonid species with similar upper thermal and hypoxia tolerances; Beitinger et al., 2000; Barnes et al., 2011; Remen et al., 2013; Williams et al., 2019). What factor(s) underlie these differences in response are not known, but perhaps the fact that the majority of the salmonid heart is perfused by venous blood of low oxygen content and PO2 (Farrell and Smith, 2017) plays a role. Further, and importantly, there was no indication at all in the present study that acclimation to high temperatures affects the tolerance of mitochondrial respiration to anoxia exposure and/or O2 limitation. This result agrees with the majority of studies that have examined aspects of fish cardiorespiratory physiology, and show that there is no, or very little, ‘cross-tolerance’ between hypoxia and high temperature exposure in this taxa (Anttila et al., 2015; Motyka et al., 2017; Levesques et al., 2019; McBryan et al., 2016).
Post-anoxia, ROS production/release was unchanged or slightly lower in both acclimation groups, with the exception that fish acclimated to and tested at 20°C had a slightly higher ROS production/release rate (Fig. 3E,F), Given that the latter effect was relatively small, and largely related to the fact that acclimation to 20°C decreased ROS production/release at 20°C as compared to fish acclimated to 12°C, our results agree with those of Onukwufor et al. (2016), who showed that AR had no effect on State 4 ROS production/release in rainbow trout.
Summary and perspectives
This study provides several novel insights into how acclimation alters the Atlantic salmon's mitochondrial response to temperatures (20°C) near its maximum rearing temperature (∼23°C; Gamperl et al., 2020), and key information with regards to the mechanistic basis for this species' cardiac and aerobic plasticity in the face of warming temperatures (Anttila et al., 2014; Hvas et al., 2017). For example, within the limits of this study, we found that: the salmon's cardiac mitochondria are resilient to warming in the range of 12–20°C; and that although acclimation to 20°C does not constrain or diminish mitochondrial function at 12°C, it protects mitochondrial coupling and decreases ROS production at 20°C. Yet, the two acclimation groups differed very little with respect to their cardiac mitochondrial responses to warming and O2 limitation (AR). These data strongly suggest that mitochondrial function does not experience ‘cross-tolerance’ to the major environmental challenges of high temperature and O2 limitation, and questions the ability of acclimation to one of these challenges to preserve cardiac mitochondrial function and integrity under climate change scenarios where hypoxia and high temperatures often co-occur.
The question of whether this thermal plasticity preserves or improves mitochondrial function (and thus, potentially salmon upper thermal tolerance) has been addressed in a companion manuscript to this work (Gerber et al., 2020). In that paper, we confirm many of the findings of this study, but also provide compelling evidence that warm acclimation to 20°C does not result in a large trade-off in Atlantic salmon mitochondrial function (i.e. increased oxidative phosphorylation does not result in heightened ROS production), and that it protects mitochondrial functional capacity until temperatures reach at least 24°C versus 20°C in 12°C-acclimated fish. However, questions remain about the acclimatory mechanisms responsible for the mitigation of ROS production at high temperatures, and why there are so disparate responses between temperate teleost species with regards to the mitochondrial acclimatory response to warm temperatures (i.e. those approaching the maximum temperatures they can tolerate; e.g. the present study versus Iftikar et al. 2015; Ekström et al. 2017; Baris et al. 2016 and Chung et al. 2017). Further, the importance of the adjustment in mitochondrial NO sensitivity with prolonged warming, and its implications for cardiovascular protection and functionality, remain to be determined. Clearly, these are critical questions with regard to mitochondrial function and cardiac performance at high temperatures that are germane not only to our understanding of fish physiology and thermal tolerance, but also to the management and conservation of fish species as the impacts of climate change on aquatic species intensify (Helmuth, 2009; Cooke et al., 2013; Speers-Roesch and Norin, 2016; Norin and Metcalfe, 2019; Little et al., 2020).
Acknowledgements
We thank Ellen Peroni and Rebeccah Sandrelli for their assistance with fish care and husbandry.
Footnotes
Author contributions
Conceptualization: L.G., A.K.G.; Methodology: L.G., K.A.C., A.K.G.; Validation: L.G., A.K.G.; Formal analysis: L.G., K.A.C.; Investigation: L.G., K.A.C.; Resources: A.K.G.; Data curation: L.G., K.A.C.; Writing - original draft: L.G., A.K.G.; Writing - review & editing: L.G., K.A.C., A.K.G.; Visualization: L.G., K.A.C., A.K.G.; Supervision: A.K.G.; Project administration: A.K.G.; Funding acquisition: A.K.G.
Funding
Research funding was provided to A.K.G. by the Ocean Frontier Institute, through an award from the Canada First Research Excellence Fund.
References
Competing interests
The authors declare no competing or financial interests.