Influence of elevated temperature on metabolism during aestivation: implications for muscle disuse atrophy

Temperature is a pervasive and heterogeneous environmental factor that impacts upon organisms at all levels of biological organisation. For ectotherms, environmental temperature plays a large role in determining their body temperature and the rate of their physiological processes. Most temperature-sensitive rates or processes operate within a relatively narrow thermal range whereby the performance of these traits declines outside of this thermal window (Bullock, 1955; Huey and Kingsolver, 1993; Angilletta, 2009). However, physiological traits can often be modified in response to changes in temperature, allowing an animal to respond to or compensate for changes in their thermal environment, e.g. acclimatisation (Wilson and Franklin, 1999; Angilletta, 2009).

Arid and semi-arid environments are typified by highly seasonal and often unpredictable rainfall (Schwinning and Osvaldo, 2004), temperature extremes and marked thermal variability. In these environments, many species of amphibian have evolved a suite of behavioural, physiological and biochemical adaptations, broadly termed aestivation, which allow them to avoid desiccation and survive in an environment that places considerable abiotic pressure on physiological systems (Pinder et al., 1992). Most aestivating frogs moderate their exposure to the harsh external environment by burrowing underground (Bentley, 1966; Pinder et al., 1992); however, even at depths of between 10 and 40 cm below the surface, where burrowing frogs can be found (Thompson et al., 2005; Cartledge et al., 2006), temperatures can still exceed 30°C for extended periods (see Fig. 1) (Shoemaker et al., 1972; Claussen, 1974; McLanahan et al., 1976; Loveridge and Craye, 1979).

Metabolic depression is a conserved feature of all dormancy strategies, including aestivation, and is coordinated across whole-animal, tissue and molecular levels in order to protect endogenous fuel supplies from premature exhaustion (Cowan and Storey, 1999; Cowan et al., 2000; Storey, 2002; Storey and Storey, 2004). In addition, metabolic depression is thought to protect the muscles of (immobile) dormant animals from muscle disuse atrophy (MDA) in part via a decrease in the production of reactive oxygen species (ROS) (Hudson and Franklin, 2002) because ROS are produced in proportion to metabolic rate (Adelman et al., 1988; Turrens, 2003). MDA results in morphological changes that may include reduced cross-sectional area of muscles and/or muscle fibres, fibre-type switching and increased connective tissue and which may result in reduced muscle function and performance (Appell, 1990; Boonyarom and Inui, 2006; Clark, 2009). MDA has been linked to an accumulation of ROS, which causes oxidative stress and damage to muscle tissue (Kondo et al., 1991; Powers et al., 2007). Indeed, a comparison of dormant animals has shown that those with lower mass-specific metabolic rates suffer less MDA over a normalised period of disuse (Hudson and Franklin, 2002). Although the muscles of aestivating frogs are remarkably tolerant of MDA compared with those of other animals (Hudson and Franklin, 2002), previous studies have shown that in Cyclorana alboguttata different skeletal muscles undergo different degrees of atrophy, displaying different patterns of morphological change (Mantle et al., 2009) and differentially regulated metabolic enzymes (Mantle et al., 2010). This suggests that, given the links between metabolic rate and ROS production, the rate of oxygen consumption and metabolic depression of muscles during aestivation may also be muscle specific. As frogs are ectotherms, temperature-related changes in metabolic rate have the potential to increase rates of MDA in dormant immobilised frogs, presumably in a muscle-specific fashion, and impact upon post-emergence locomotor performance and survival. In fact, the metabolic rate of the toad Scaphiopus couchii appears to be more sensitive to temperature during aestivation than while active (Seymour, 1973), suggesting that temperature-induced rates of ROS production may even be higher in the tissues of aestivating animals relative to non-aestivating animals.

In the present study we sought to determine the effect of environmental temperature on the metabolic rate and the magnitude of the metabolic depression of aestivating green-striped burrowing frogs, C. alboguttata, at both the whole-animal and tissue (muscle) levels. We hypothesised that high temperatures during aestivation would increase the metabolic rate of frogs and tissues, and the extent of metabolic depression associated with aestivation would be compromised at higher temperatures. Moreover, we predicted that the metabolic responses of muscles to aestivation would be muscle specific in a way that complements the differential susceptibility of specific muscles to MDA in aestivating C. alboguttata found in previous studies. Following on from this we further hypothesised that the metabolic responses of muscles to temperature during aestivation would also be muscle specific.

Animals were collected in accordance with State and Federal laws pertaining to the use of animals for scientific purposes (Scientific Purposes Permit number WISP03572406). All experiments were carried out with the approval of the University of Queensland's Animal Welfare Committee (permit numbers SIB/442/07/UQ and SIB/602/08/ARC).

Following heavy rains, green striped burrowing frogs, C. alboguttata Günther, were collected from agricultural areas in southeast Queensland, Australia. Frogs were transported to The University of Queensland and maintained in individual housing, which was cleaned weekly. Frogs were fed once a week with live food (crickets or cockroaches). Frogs were assigned randomly to treatment temperatures of 20, 24 or 30°C for whole-animal metabolic rate (oxygen consumption rate, ) determination, and to 24 or 30°C for muscle determination. Treatment temperatures were within the range of soil temperatures frogs would be expected to experience in the wild based on data from temperature loggers buried at 10, 20 and 40 cm depth in areas with known C. alboguttata populations (Fig. 1).

Frogs were maintained in controlled-temperature rooms at their treatment temperatures for 2 weeks without food prior to beginning the measurements. Mass-specific was measured using closed-system respirometry (N=10 per treatment group). Frogs were placed into individual respirometers (550–1020 ml volume) with moist paper towels 24 h prior to the first measurements and left in darkened conditions. Aestivation was induced by allowing the paper towelling to slowly dry out (see Flanigan et al., 1991; Withers, 1993; Fuery et al., 1998; Withers and Thompson, 2000; Bayomy et al., 2002; Kayes et al., 2009a). To prevent dehydration, small amounts of water (∼0.5 ml) were added weekly to the paper in each respirometer. Animal behaviour (adoption of water-conserving posture and formation of cocoon) and condition were monitored under a red light. Any frogs showing signs of stress or those not entering aestivation were removed from the experiment.

The metabolic rate of isolated skeletal muscles was measured in 24°C non-aestivating (N=7), 24°C 6 month aestivating (N=7), 30°C non-aestivating (N=7) and 30°C 6 month aestivating (N=9) frogs. In order to facilitate 6 months of aestivation, frogs were placed in 4 l ice cream containers filled with wet clay mud. Frogs were then left in the dark in constant temperature rooms at their respective treatment temperature. Frogs were checked once a week and were deemed to be in aestivation once they had burrowed into the mud block and remained there. The mud was allowed to dry slowly over the course of the experimental period. Note that, unlike for animals aestivating in respirometers, no data were collected at 20°C as animals would not remain burrowed or enter aestivation. Non-aestivating animals were maintained at their treatment temperature and were fed and cleaned weekly. After 6 months, aestivating frogs (removed from their clay burrows) and non-aestivating frogs from each temperature group were immediately killed by double pithing. Five skeletal muscles were dissected from the left hindleg of each frog: gastrocnemius, sartorius, iliofibularis, gluteus magnus and cruralis. Each muscle was weighed immediately and placed in air-saturated Ringer solution (111 mmol l^–1 NaCl, 2.5 mmol l^–1 KCl, 1.8 mmol l^–1 CaCl₂·2H₂O, 1 mmol l^–1 MgCl, 5 mmol l^–1 Hepes, 10 mmol l^–1 glucose, adjusted to pH 7.4 with NaOH). All solutions were maintained, and experiments run, at the treatment temperature of the frog that the tissues were obtained from. Each muscle was cut longitudinally into thin slices (<0.5 mm) with a razor blade.

Rates of tissue oxygen consumption were measured using a 24-channel oxygen meter (sensor dish® reader, SDR2, PreSens GmbH, Regensburg, Germany). Vials (5 ml) were equipped with a planar oxygen-sensing spot (5 mm diameter, type Sp-PSt5-NAU-D5-YOP, PreSens GmbH) that measured oxygen (% air saturation) via optical fluorescence. The vials were randomly distributed in a 24-well rack secured to the SDR2. The SDR2 was placed on an orbital shaker set at 288 r.p.m. in a temperature-controlled chamber (set to the treatment temperature of the source frog) and connected to a computer. Recordings were made every 2 min for vials of oxygenated Ringer solution, in the absence of tissue, to allow the oxygen concentration reading to stabilise. Tissue slices were then added to the vials, suspended above the sensor in a small soft mesh basket, and vials were then topped up with oxygenated Ringer solution and sealed. Vials containing Ringer solution only served as blanks to correct for background oxygen consumption. Oxygen concentration readings were recorded every 2 min until muscle respiration rate stabilised. All measures of tissue oxygen consumption were made at the treatment temperature of the frog from which the tissues were collected. At the termination of the experiment tissue slices were removed from vials, gently blotted on paper towels and weighed. All tissues were handled in a standardised manner throughout all experimental recordings.

Data are presented as means ± s.e.m. Whole-animal () data were tested between temperature groups using one-way ANOVA for each time point. Data within temperature treatment groups were compared between consecutive time points using matched pair analysis. Magnitude (%) of metabolic depression data were arcsine transformed and the difference between temperature groups tested using one-way ANOVA. Tissue data were log transformed to meet equality of variance assumptions. A two-way ANOVA model was fitted to the transformed data considering temperature (two levels: 24 and 30°C), metabolic state (two levels: controls and aestivators), and the interaction between temperature and metabolic state as main effects. Frog snout–vent length (SVL) was included as a covariate. The model was run for each muscle separately. Where significant effects were found they were localised with Tukey's honestly significantly different (HSD) test. Planned comparisons between 24 and 30°C within the aestivating group were run for each muscle using Student's t-test or Welch ANOVA where appropriate. Muscle-specific within treatments was tested using a general linear model considering muscle type as the main effect with source animal as a random effect to account for the non-independence of data when analysing data for different muscles obtained from the same animal. Where general linear model analyses detected significant differences, they were localised using Tukey HSD. All analyses were conducted using JMPIN v4 or JMP v8 software with significance set at P<0.05.

The body mass of frogs used in this study ranged from 19.6 to 41.1 g (mean 27.2±0.9 g) and SVL of these animals ranged from 51.05 to 63.04 mm (mean 56.8±0.6 mm).

In all temperature treatments (20, 24 and 30°C), C. alboguttata exhibited a significant decrease in over the 10 week measurement period, which corresponded to a cessation of movement and the formation of a cocoon, typical of aestivation in this species (20°C P=0.0024, 24°C P=0.001, 30°C P=0.002; Fig. 2). The highest metabolic rates were recorded from non-aestivating, resting frogs at week 1. By week 5, the of C. alboguttata had decreased significantly, and it continued to decrease in the following weeks, with the lowest mean metabolic rate observed after 10 weeks in all groups (Fig. 2).

The metabolic rate of C. alboguttata was significantly influenced by temperature at rest and during aestivation across the three temperature treatments (ANOVA, P<0.001; Fig. 2 and Fig. 3A). Interestingly, however, temperature did not have a significant effect on the of frogs between 20 and 24°C during the 10 week sampling period (Tukey HSD, P>0.05; Fig. 2). At rest (week 1, Fig. 2), the of C. alboguttata at 30°C was significantly higher than the of frogs at 20 and 24°C and these frogs maintained the highest oxygen consumption rates during aestivation, relative to animals at the lower temperatures (weeks 5, 8 and 10; Fig. 2 and Fig. 3A). The mean minimum of aestivating frogs recorded at 30°C was 22.3±1.4μl O₂ g^–1 h^–1, which was more than double that of aestivating frogs at 20°C (=9.9±1.6 μl O₂ g^–1 h^–1; P<0.01; Fig. 3A).

Exposure to a higher temperature decreased the magnitude of the maximum metabolic depression that was achievable during aestivation by C. alboguttata (Fig. 3B). Despite the frogs at 30°C being inactive and encased in a cocoon, they were only able to depress their metabolic rate by 58±3% relative to resting levels. In contrast, at 20 and 24°C, the frogs were able to depress their metabolic rate by 75±5% and 66±4%, respectively.

The covariate SVL was not significant in any model. The of tissue slices recorded from five hindlimb muscles from aestivating animals was always significantly lower than that recorded from non-aestivating animals, indicating that tissue-level metabolic depression is associated with aestivation in C. alboguttata (Fig. 4). Temperature had a significant effect on all muscles except the iliofibularis, with rates of oxygen consumption significantly higher at 30°C than at 24°C (Fig. 4). There was a significant interaction between temperature and metabolic state for the gastrocnemius (F_1,24=15.0627, P=0.0007) and cruralis (F_1,24=6.6197, P=0.0167) muscles whereby the rates of oxygen consumption in muscles from frogs aestivating at 30°C were substantially higher than those of frogs aestivating at 24°C (all P<0.05, Fig. 4).

Planned comparisons between the of muscles from frogs aestivating at 24 and 30°C demonstrated that oxygen consumption was significantly higher in the cruralis (F_1,13=17.7526, P=0.001), gastrocnemius (Welch F_1,8.6981=33.7527, P=0.0003), and sartorius (F_1,13=5.0472, P=0.0427) muscles from frogs aestivating at 30°C relative to those from frogs aestivating at 24°C. The of the gluteus magnus was also consistent with this pattern but just failed to reach significance (Welch F_1,11.249=4.7816, P=0.0507). The rate of oxygen consumption of the iliofibularis from aestivating frogs was not influenced by temperature at all (Welch F_1,8.1251=1.0821, P=0.3282).

Differences in the rates of oxygen consumption were also apparent between muscle types in non-aestivating frogs at both 24°C (F_4,23=9.0809, P=0.0001) and 30°C (F_4,22=5.5158, P=0.0031) and in frogs aestivating at 24°C (F_4,20=24.0352, P<0.0001). There was, however, no difference between the rates of oxygen consumption of the five muscles from frogs aestivating at 30°C (F_4,30=0.7299, P=0.5787). The rates of oxygen consumption in the smaller muscles (iliofibularis and sartorius) were significantly higher than those of the larger muscles (cruralis and gastrocnemius) (Fig. 4). The gluteus magnus showed a rate of oxygen consumption that was intermediate between the small and large muscles.

For C. alboguttata the impacts of environmental temperature fluctuations are somewhat alleviated by burrowing underground and selecting a thermally buffered microhabitat. Although we have relatively little information regarding natural burrow depths of wild C. alboguttata, studies of closely related cocoon-forming frog species (Cyclorana australis and Cyclorana platycephala) have recorded natural burrow depths of between 2 and 30 cm (Van Beurden, 1984; Tracy et al., 2007). From our laboratory studies, this burrow depth range seems likely to also reflect burrowing tendencies of C. alboguttata. From Fig. 1 it is clear that at shallower depths C. alboguttata would be subject to large daily fluctuations in temperature. At depth, extreme fluctuations are dampened and temperature is more constant; however, even at 40 cm below the surface, soil temperatures can reach 30°C or more and can remain elevated for some time. The thermal regime employed in this study is therefore realistic and may even be somewhat conservative depending on frog burrow depths. Although some non-cocoon-forming species can continue to burrow over the course of their underground aestivation, presumably to follow the soil moisture level, this would seem unlikely to occur in cocoon-forming species like C. alboguttata burrowed in heavy clay soils, which become impenetrable as the soils dry and frogs would remain subjected to the temperature profile of the depth at which they had burrowed.

During aestivation many species of frog substantially reduce their metabolic rate by between 60% and 80% (Seymour, 1973; Flanigan et al., 1991; Withers, 1993) and the magnitude of metabolic depression observed for aestivating C. alboguttata in the present study was consistent with this. These results are also consistent with earlier studies which have shown that as ambient temperature increases, the rate of oxygen consumption by aestivating frogs also increases (Van Beurden, 1980), and the magnitude of metabolic depression achieved decreases (Seymour, 1973; Abe and Garcia, 1991). Although aestivating C. alboguttata significantly depressed metabolic rate at all experimental temperatures, their capacity to do so was clearly impaired at 30°C. As expected, exposure to higher temperatures during aestivation resulted in higher rates of oxygen consumption, which reduced the ability of frogs aestivating at 30°C to suppress metabolic rate to the same extent as animals aestivating at 20°C. Thus, temperature may have a more profound influence on physiological function during aestivation as opposed to while active (not aestivating), consistent with the results of Seymour (Seymour, 1973). The inability of frogs to suppress metabolic rate at 30°C to the same rate of oxygen consumption shown by an animal at 20°C would mean that the energy costs of maintaining physiological function are greater in frogs aestivating at the higher temperature, and during long periods of aestivation these costs could significantly impact their ability to survive.

Interestingly, some species of aestivating frogs, such as Leptodactylus fuscus and Physalaemus fuscomaculatus, show whole-animal oxygen consumption rates that are thermally insensitive between 20 and 25°C (Abe and Garcia, 1991). Thus, the lack of difference in the metabolic rate of non-aestivating and aestivating C. alboguttata kept at 20 and 24°C may also reflect some similar thermal insensitivity of oxygen consumption rate within this temperature range. This may allow for subtle fluctuations in environmental temperature about this range to be ‘absorbed’ to provide some physiological stability. However, at 30°C the effect of environmental temperature on the physiological rate processes of aestivating C. alboguttata was apparently unavoidable and consequently oxygen consumption rates were substantially elevated and remained so, which is probably reflective of the physiological costs of maintaining essential functions at higher temperature.

Although during aestivation at 24°C C. alboguttata skeletal muscles undergo a slow yet selective atrophy (Hudson et al., 2006; Symonds et al., 2007; Mantle et al., 2009), relative to other animals aestivating frogs suffer little MDA and it is postulated that this is related to the protection from ROS damage afforded by metabolic depression (Hudson and Franklin, 2002). Metabolic depression during aestivation is a coordinated strategy that is reflected across levels of biological organisation from whole-animal to tissue and sub-cellular levels (Storey, 1988; Storey and Storey, 1990; Cowan and Storey, 1999; Cowan et al., 2000; Kayes et al., 2009b). In the present study, all isolated muscles examined from aestivating C. alboguttata demonstrated substantially suppressed rates of oxygen consumption. Rates of oxygen consumption by the gastrocnemius, cruralis and sartorius muscles were, however, significantly higher in frogs aestivating at 30°C than in frogs aestivating at 24°C, suggesting that these muscles may be more susceptible to MDA as a consequence of increased ROS production (Hudson and Franklin, 2002) proportional to metabolic rate (Adelman et al., 1988; Turrens, 2003). However, aestivation temperature clearly did not influence the metabolic rates of all muscles in the same way, as the metabolic rate of the iliofibularis and, to a lesser extent, the gluteus magnus, was unaffected by aestivation temperature. This implies that the muscles examined are differentially affected by temperature and/or aestivation and this may reflect muscle-specific differences in their underlying physiochemical properties, which in turn might influence a muscles' predisposition for atrophy during aestivation.

Muscle-specific differences in their metabolic responses to temperature probably reflect underlying morphological and/or physiological differences between muscles and/or muscle-specific alterations to metabolic pathways that accompany aestivation. In the present study, the iliofibularis and sartorius muscles generally consumed significantly more oxygen (per gram of tissue) than the gastrocnemius, gluteus magnus and cruralis, and previous studies have shown that differences in the metabolic physiology of isolated skeletal muscles may reflect their functional similarity (Duellman and Trueb, 1994; Lutz et al., 1998; Mantle et al., 2009). Indeed, Mantle and colleagues found that changes in key metabolic enzymes (cytochrome c oxidase, citrate synthase and lactate dehydrogenase) were consistent within the functional classification of ‘non-jumping’ (iliofibularis and sartorius) and ‘jumping’ (gastrocnemius and cruralis) muscles, but were distinctly different from each other (i.e. between jumping and non-jumping), suggesting that muscles which perform similar functions display similar metabolic profiles and are similarly affected by aestivation (Mantle et al., 2010). However, muscle fibre-type composition is unlikely to be related to differences in the metabolic rate of muscles and their differential responses to temperature as recent work has suggested that fibre composition does not necessarily predict muscle metabolic capacity (Crockett and Peters, 2008). Similarly, anuran muscles can also show vastly different metabolic and contractile properties despite having similar fibre compositions (Putnam and Bennett, 1983).

In addition to underlying muscle-specific differences, aestivation itself may promote differential changes in muscles that alter their metabolic profiles and responses to temperature. Cellular metabolic enzyme and signal transduction pathway changes associated with dormancy have been shown to occur in a tissue-/organ-specific fashion in aestivating spadefoot toads, whereby hindlimb skeletal muscle undergoes a much more profound reorganisation than does either the brain or liver (Cowan et al., 2000). Moreover, previous studies have shown that aestivation alters the activity of various metabolic enzymes in the skeletal muscles of C. alboguttata, and that different muscles show different degrees of biochemical change (Mantle et al., 2010). Changes in the metabolic physiology of skeletal muscles may involve modification of glycolytic (reversible phosphorylation of regulatory enzymes) and/or oxidative pathways (mitochondrial substrate metabolism and oxidative phosphorylation) (for a review, see Storey, 1997) [for aestivating anurans, see studies by Cowan and colleagues (Cowan and Storey, 1999; Cowan et al., 2000)]. In addition, enzymes involved in the regulation of metabolic suppression in hibernating mammals show different kinetic responses to temperature in dormant and awake animals (see Storey, 1997), suggesting changes at the biochemical level occur in some proteins during dormancy that may affect their thermal responsiveness and alter the metabolic properties of tissues.

Higher rates of tissue oxygen consumption during aestivation at high temperatures have the potential to enhance rates of MDA in C. alboguttata as a result of the relationship between metabolic rate and ROS production, and ROS production and MDA (Powers et al., 2005; Powers et al., 2007). Although the hypometabolic state during aestivation probably affords some protection against MDA by lowering the production of ROS (Hudson and Franklin, 2002), previous studies have shown that tissues of aestivating anurans still do accumulate ROS damage (Grundy and Storey, 1998). Aestivating land snails, Otala lactea, show an increase in oxidative damage to lipids when the rate of oxygen consumption increases (Hermes-Lima and Storey, 1995). The elevated oxygen consumption rates found in some skeletal muscles of aestivating C. alboguttata lends credence to the hypothesis that skeletal muscles may be differentially susceptible to disuse atrophy on the basis of rates of oxygen consumption. Indeed, Mantle and colleagues determined that in C. alboguttata the iliofibularis and sartorius muscles underwent substantially more atrophy during aestivation than either the gastrocnemius or cruralis muscles (Mantle et al., 2009), and based on the results of the present study this is consistent with the aforementioned hypothesis. Given that exposure to high temperatures accelerates ROS production in skeletal muscles of mammals (Edwards et al., 2007; Muller et al., 2007; van der Poel et al., 2007), it might be expected that the accumulation of oxidative damage during aestivation would also be exacerbated by higher aestivation temperatures. It might be predicted that the muscles from C. alboguttata aestivating at 30°C that had metabolic rates in excess of those observed in frogs aestivating at 24°C (i.e. the gastrocnemius, sartorius and cruralis) would be susceptible to increased oxidative damage and increased rates of MDA during aestivation at this higher temperature. Further investigations should focus on clarifying this proposition.

Given the importance of skeletal muscle immediately post-aestivation, an increase in susceptibility of some muscles to disuse atrophy at higher temperatures may have serious consequences for locomotor performance and survival, unless compensatory processes are up-regulated. Muscle protection against ROS damage may be afforded by an increase in the production of antioxidants (Gutteridge, 1995). In C. alboguttata aestivating at approximately 24°C, antioxidant production, normalised to metabolic rate, is up-regulated ∼3.5-fold (Hudson et al., 2006). If the rate of antioxidant production is indeed related to metabolic rate in aestivating frogs, then an increase in metabolic rate at 30°C may also result in the up-regulation of compensatory antioxidants. In addition, heat-shock proteins (HSP) have a significant and highly conserved role in the protection of skeletal muscle tissue (Liu et al., 2006) and are suggested to protect tissues against MDA in rats exposed to high temperature (Naito et al., 2000). Enhanced production of antioxidants and HSP during aestivation in C. alboguttata would probably pose an additional energy cost and, consequently, the higher metabolic rate of muscles from frogs aestivating at 30°C may in part be due to increased energy costs associated with enhanced mobilisation of the underlying mechanisms that protect against MDA. An ability to protect muscle function despite a high ambient temperature may mean that effects on locomotor performance following aestivation are minimised.

Metabolic depression across all levels of biological organisation is a vital strategy allowing organisms to survive for extended periods in environments that become unfavourable. Higher temperatures during the aestivation period of burrowing frogs present challenges to this strategy. At both the whole-animal and tissue levels, significantly higher oxygen consumption rates during aestivation at 30°C would accelerate the use of endogenous fuel stores and would also expose frogs to a higher rate of ROS insult relative to frogs aestivating at lower ambient temperatures. At the tissue level, increases in the rate of oxygen consumption by muscles from aestivating frogs at higher aestivation temperatures and the associated increased ROS production have implications for muscle integrity. The muscle-specific differences in metabolic rate and temperature sensitivity indicate that future investigations should focus on both better elucidating the mechanisms resulting in muscle-specific effects of temperature and examining the fitness implications that this may have for organisms aestivating for prolonged periods at high temperatures.

The authors wish to thank four anonymous reviewers for their helpful comments on an earlier version of this manuscript which resulted in its improvement.