Variation in expression of calcium-handling proteins is associated with inter-individual differences in mechanical performance of rat (Rattus norvegicus) skeletal muscle

Skeletal muscle performance and locomotion underlie the non-cognitive dimension of animal behaviour and are therefore major determinants of organismal fitness (Dickinson et al., 2000). The specialization of species for long distance locomotion, such as for migration (Piersma, 1998), or for short burst activity (Vanhooydonck et al., 2001) will partly determine their biogeography and ecological relationships (Irschick and Losos, 1998). At an individual level, better sprint or endurance performance will modulate sporting success in humans (Van Damme et al., 2002), and will affect fitness in natural systems by influencing predator escape and prey capture (Jayne and Bennett, 1990; Le Galliard et al., 2004). There can be pronounced differences in locomotor performance between individuals (Wilson et al., 2002; Rupert, 2003). For example, in human athletes that excel in sprint performance, endurance performance is compromised, and vice versa (Van Damme et al., 2002). A similar trade-off exists between species of vertebrates at the level of locomotion in the whole animal (Vanhooydonck et al., 2001) or within species at the level of isolated skeletal muscle performance (Wilson et al., 2002; Wilson and James, 2004; Wilson et al., 2004). Such sprint–endurance trade-offs may reflect a typical response of the vertebrate locomotory system to conflicting demands. The implication is that, although animal fitness often relies on both sprint and endurance performance, the two cannot be maximized simultaneously.

The mechanistic basis for a sprint–endurance trade-off remains largely unexplained despite its far-reaching importance for understanding physiological constraints in animal ecology, sporting performance and medicine. Most evidence documenting a trade-off is correlational at the level of the whole animal or muscle. Our aim was to advance knowledge by investigating the potential roles of specific molecular mediators of muscle contractile performance in the rat (Rattus norvegicus) lateral gastrocnemius muscle. Our approach was to represent both Ca²⁺ handling, which underlies muscle contraction and relaxation, and metabolic capacity with multiple parameters to determine their relative importance and possible interaction. We correlated muscle kinetics with: (1) mRNA expression of candidate proteins to determine possible changes at the genome level, and (2) protein concentration and enzyme activity to determine responses at the protein level. Following correlational analyses, we determined cause and effect between molecular mechanisms and muscle performance by specific pharmacological manipulation of Ca²⁺-handling proteins.

The morphology of the skeletal system poses only a limited constraint on either sprint or endurance performance (Vanhooydonck et al., 2001), such that a major limitation to locomotion is likely to reside in muscle physiology. The most likely processes that constrain muscle function are excitation–contraction coupling via Ca²⁺ cycling (Berchtold et al., 2000), and the capacity of the muscle to produce the necessary energy (ATP) to power contraction. Ca²⁺ release into the myocyte and subsequent binding with troponin controls the interaction between actin and myosin and thereby the rate of muscle force generation and, conversely, re-sequestration of Ca²⁺ into the sarcoplasmic reticulum determines relaxation rates (Farah and Reinach, 1995; Berchtold et al., 2000). Two of the key elements of this calcium-handling process are the ryanodine receptor and sarco/endoplasmic reticulum Ca²⁺ ATPase (SERCA). The ryanodine receptor interacts with the dihydropyridine receptor during the excitation–contraction coupling process, and acts as the sarcoplasmic reticulum calcium release channel. An increased density of dihydropyridine and ryanodine receptors in muscle is correlated with increased shortening velocity and force in isolated muscle (Kandarian et al., 1992; Golden et al., 2003; Mänttäri and Järvilehto, 2005), and their density is increased following endurance training in trout (Anttila et al., 2008). SERCA is responsible for calcium reuptake from the cytoplasm to the sarcoplasmic reticulum, and its activity can therefore determine muscle relaxation rate (Fleming et al., 1990; Wilson et al., 1998). SERCA activity is regulated by phosphorylation of phospholamban, which in its unphosphorylated form inhibits SERCA activity (Verboomen et al., 1992; Periasamy and Kalyanasundaram, 2007). Phospholamban phosphorylation is under sympathetic control via β2-adrenoreceptor-mediated release of cAMP (McCormick et al., 2010). ATP is used during myosin ATPase activity, which powers the interaction between actin and myosin leading to muscle contraction, and by SERCA-mediated Ca²⁺ transport into the sarcoplasmic reticulum during muscle relaxation. Hence, ATP supply can be an important constraint for muscle function. Ca²⁺ transport and cellular energy metabolism may therefore be mechanisms underlying the functional sprint–endurance trade-off.

Characteristics of Ca²⁺ transport and energy metabolism are commonly used for muscle fibre typing, which is variously based on the relative proportions of protein isoforms of myosin heavy chain (Bottinelli and Reggiani, 2000; Plomgaard et al., 2006), SERCA and troponin (Bottinelli and Reggiani, 2000; Berchtold et al., 2000). Differences in glycolytic and oxidative metabolic capacities are also used as the basis of muscle fibre classification, where slow, endurance fibres are more oxidatively poised and fast fibres have greater glycolytic capacity (Bottinelli and Reggiani, 2000). However, the molecular traits used to categorize muscle fibre types often do not correspond to the expected patterns of muscle performance and sprint–endurance trade-offs (Gibb and Dickson, 2002; James et al., 2005; Plomgaard et al., 2006; Seebacher and James, 2008). This may be because actual muscle contractile performance is determined by interactions between multiple molecular traits, so that deducing a direct relationship between fibre type categories and muscle performance is difficult. We therefore investigated individual traits separately without any preconceived definitions of fibre types.

We tested the following hypotheses. (1) There is a trade-off between maximal force generated and fatigue resistance of force generation in isolated lateral gastrocnemius. (2) Muscle oxidative metabolic capacity is associated with fatigue resistance and muscle glycolytic metabolic capacity is associated with maximal muscle activity. Specifically, muscles with a shorter contraction time will have higher lactate dehydrogenase activity, while muscle fatigue resistance will be positively related to citrate synthase and cytochrome c oxidase activities. (3) Ca²⁺-handling dynamics are associated with muscle performance during sustained and maximal activities. Specifically, muscles with shorter contraction times will have higher ryanodine receptor mRNA concentrations, and variations in SERCA isoform mRNA concentration, SERCA protein concentration and total SERCA activity are associated with muscle relaxation times; additionally, the relative proportion of troponin I isoforms will be associated with variation in maximal (sprint) and sustained muscle performance.

All experimental and animal-handling procedures were approved by the University of Sydney Animal Ethics Committee (approval number L04/6-2008/1/4865). We obtained rats [Rattus norvegicus (Berkenhout); N=20, body mass 130±9 g] from three different breeding backgrounds to increase inter-individual variation. Rats were kept in standard rat cages at 22°C with a 14 h light:10 h dark cycle, and with access to ad libitum food and water. Rats were killed by an injection of dilute sodium pentabarbitone (150 mg kg^–1 body mass) immediately before experimentation. Body mass was measured to the nearest 0.1 g using an electronic balance. The skin was removed from the hindlimbs, then the hindlimbs were removed and placed in aerated Krebs solution (composition, in mmol l^–1: 118 NaCl, 4.75 KCl, 1.18 MgSO₄, 24.8 NaHCO₃, 1.18 KH₂PO₄, 10 glucose, 2.54 CaCl₂; pH 7.5 at room temperature prior to oxygenation) kept on ice. Muscle tissue samples (approximately 0.05 g) were removed from one lateral gastrocnemius head in each leg and either placed in RNAlater for real-time PCR assays or immediately frozen in liquid nitrogen for enzyme activity assays. The other lateral gastrocnemius muscle head in each leg was removed, leaving tendons attached to the muscle and a small section of bone, for use in mechanical measurements of muscle performance. This approach makes the assumption that the gastrocnemius muscle taken from the left leg will be structurally, mechanically and functionally similar to that taken from the right leg of the same individual.

We measured the activities of the mitochondrial enzymes cytochrome c oxidase (COX) and citrate synthase (CS) as indicators of aerobic metabolic capacity, and the activity of lactate dehydrogenase (LDH) as an indicator of anaerobic metabolic capacity (St-Pierre et al., 1998). Frozen muscle samples (0.05–0.1 g) were homogenized immediately in 9 volumes of ice-cold extraction buffer (50 mmol l^–1 imidazole, 2 mmol l^–1 MgCl₂, 5 mmol l^–1 EDTA, 0.1% Triton and 1 mmol l^–1 glutathione, pH 7.5 at 0°C) using a rotor-stator homogenizer (Pro200, Pro Scientific, Oxford, CT, USA). All assays were conducted according to published protocols (Seebacher et al., 2003), and activities were expressed as μmol of substrate converted min^–1 g^–1 tissue.

Total SERCA activity was determined according to a previously described ATPase assay (Walter and Seebacher, 2009), except that the assay medium consisted of (mmol l^–1): imidazole 25, CaCl₂ 0.2, KCl 0.6, MgCl₂ 0.048 (Ramnanan and Storey, 2008). Inorganic phosphate concentration was quantified by the ammonium molybdate assay (Bonting et al., 1961). Total SERCA activity was calculated as the difference in inorganic phosphate liberated in the presence and absence of 1 μmol l^–1 thaspsigargin (Ramnanan and Storey, 2008). All enzyme assays were conducted in duplicate at 30°C in a spectrophotometer with a temperature-controlled cuvette holder (Ultrospec 2100 pro UV; Amersham Pharmacia, Sydney, Australia).

RNA was extracted from 40–100 mg of muscle samples using TRIreagent (Molecular Research Centre, Cincinnati, OH, USA), following the manufacturer's instructions. RNA quality and concentration were verified using a Bioanalyzer (Agilent Biotechnologies, Santa Clara, CA, USA). Total RNA (1 μg) was treated with DNase I (Sigma-Aldrich, Castle Hill, NSW, Australia) and reverse-transcribed using RNase H^– MMLV reverse transcriptase (Bioscript, Bioline, Taunton, MA, USA) and random hexamer primers (Bioline). Pre-validated Taqman probes (Applied Biosystems, Scoresby, Victoria, Australia) were used to quantify troponin I type 1 (TNNI1; catalogue number Rn00567843_m1), troponin I type 2 (TNNI2; Rn00437157_g1), Ca²⁺-transporting ATPase 1 (SERCA1; Rn00589545_m1), Ca²⁺-transporting ATPase 2 (SERCA2; Rn00568762_m1) and ryanodine receptor 1 (RYR1; Rn01545082_g1) expression and values were normalized to expression of eukaryotic translation elongation factor 1 α2 (Eef1a2; Rn00561973_m1). Quantitative real-time PCR runs were performed on an Applied Biosystems 7500 qRT-PCR machine using Taqman Gene Expression Mastermix with the standard PCR protocol as recommended by the manufacturer. Relative gene expression of the five target genes was calculated according to Pfaffl (Pfaffl, 2001), and normalized to that of Eef1a2. The grand mean of all tissues was used as a control.

Tissue lysates were prepared in homogenization buffer (65 mmol l^–1 Tris HCl, 150 mmol l^–1 NaCl, 5 mmol l^–1 EDTA pH 7.4, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 10% glycerol) with PhosStop phoshatase inhibitors and complete protease inhibitors (Roche Diagnostics, Basel, Switzerland), and centrifuged for 15 min at 12,000 g. Protein content of supernatants was quantified by the bicinchoninic acid assay (BCA, Sigma-Aldrich) according to the manufacturer's instructions. Aliquots, diluted 1:1 in 2× Laemmli buffer and denatured at 95°C, were resolved on 10% precast SDS-PAGE gels (BioRad, Gladesville, NSW, Australia) (Laemmli, 1970). Proteins were transferred to a nitrocellulose membrane (BioRad) and stained with Ponceau red dye to ensure accurate protein loading and transfer. Membranes were blocked for 1 h with 5% skimmed milk powder in Tris-buffered saline with 0.1% Tween-20. Blots were probed with anti-SERCA1 ATPase (MA3-912; Affinity Bioreagents, Golden, CO, USA) at 1:2500 and with anti-α-tubulin (T5168; Sigma-Aldrich) at 1:8000 in blocking buffer as a loading control. Secondary anti-mouse horseradish peroxidase (Abnova, Waterloo, NSW, Australia) was used at 1:10,000. Horseradish peroxidase localization was detected with Amersham ECL Plus (GE Healthcare, Rydalmere, NSW, Australia) according to the manufacturer's instructions and visualized by the G:Box gel documentation system. Relative band intensity was quantified using Alpha Innotech Software (Alpha Innotech Corp., Santa Clara, CA, USA) and normalized with α-tubulin.

Muscle preparations were transferred to a bath of aerated circulating Krebs solution maintained at 30.0±0.5°C. The bone at one end of the lateral gastrocnemius was attached via a crocodile clip to a strain gauge (model UF1, Pioden Controls Ltd, Canterbury, Kent, UK) and the bone at the other end was attached via a crocodile clip to a servomotor (V201, Ling Dynamic Systems, Royston, Hertfordshire, UK). Electrical stimulation was delivered to the muscle in 1.5 ms pulses via parallel platinum electrodes. Muscle stimulation was controlled using custom-written software (CEC Testpoint version 7, Measurement Computing, Norton, MA, USA) via a D/A board (KPCI3108, Keithley Instruments, Cleveland, OH, USA). Force data were sampled at a rate of 2 kHz. A series of isometric twitches were used to optimize stimulation amplitude and muscle length to yield the maximum isometric twitch force. An isometric tetanic response was elicited by subjecting the muscle to a 200 ms train of stimulation delivered at 160 Hz stimulation frequency. Maximal force generated, time to half-peak tetanic force (defined as ‘contraction time’) and time from the last stimulus to 50% tetanus relaxation (defined as ‘relaxation time’) were measured. A 5 min recovery period was allowed between each tetanic response.

Additionally, to test for the importance of ryanodine receptors in determining muscle force production and fatigue resistance, we incubated one muscle preparation from nine of the rats in 10 μmol l^–1 dantrolene (Sigma-Aldrich) for 10 min (Van Winkle, 1976; Fruen et al., 1997) while subjecting the contralateral preparation from each of these rats to control Krebs solution for the same time duration. Similarly, to test the role of SERCA, we incubated one muscle preparation from nine of the rats in 10 μmol l^–1 thapsigargin (Sigma-Aldrich) for 5 min (Kurebayashi and Ogawa, 2001; Galli et al., 2006), using the contralateral preparation from each of these rats as a control.

To measure fatigue resistance of muscle force generation (endurance), we subjected each control muscle preparation to a fatigue run consisting of a series of 25 tetani, where a 200 ms train of stimulation was delivered once per second. Peak tetanic force was measured in the 1st and the 25th tetanus of the fatigue run. Therefore, every muscle that was subjected to a fatigue resistance test was also tested for maximal force generation.

Lateral gastrocnemius muscle mass was determined to the nearest 1.0 mg at the end of the experiments using an electronic balance, after blotting the muscle on absorbent paper to remove excess Krebs solution. Mean muscle cross-sectional area was calculated from mean muscle fibre length, muscle mass and an assumed muscle density of 1060 kg m^–3 (Méndez and Keys, 1960). Muscle stress was calculated as force divided by mean muscle cross-sectional area. The muscle mechanics variables indicative of maximal (sprint) muscle performance were considered to be all twitch and tetanus times, and twitch and tetanus stress measurements from the maximal twitch and tetanus for each individual. A measure of fatigue resistance was determined by calculating peak tetanic force in the 25th tetanus of the fatigue run as a percentage of the peak tetanic force in the 1st tetanus of the fatigue run. For each of the 20 rats the mechanics data from the lateral gastrocnemius muscle sample from the left leg were used for analysis of the inter-individual relationship between maximal (sprint) and sustained (endurance) muscle performance.

The effects of incubation in drug or control treatments on force production were analysed by calculating peak tetanus force, activation and relaxation times in the first tetanus of the fatigue run (during drug incubation) as a percentage of peak tetanus force prior to incubation.

All statistical analyses were performed using SPSS version 17. Pearson's product moment correlation was used to analyse correlations (Table 1) between the variables. Student's paired t-tests were performed to compare the effect of dantrolene and thapsigargin treatment with their respective controls. The truncated product method (Zaykin et al., 2002) was used to combine all the P-values in this study to determine whether there was a bias from multiple hypothesis testing. The truncated product method P-value was <0.001, showing that the results were not biased. Significance was taken at the level of P<0.05.

Maximum force produced (maximum tetanus stress) was negatively correlated with fatigue resistance in lateral gastrocnemius muscle (r=–0.496, P=0.013). Those muscles that produced high maximal tetanic stress were more likely to have a poor fatigue resistance, and vice versa, indicating a trade-off between maximal (sprint) and sustained (endurance) performance within skeletal muscle, consistent with hypothesis 1.

Contraction time (time to half-peak tetanus) was significantly negatively correlated with lactate dehydrogenase activity (Fig. 1; Table 1; r=–0.487, P=0.015). This finding demonstrates that muscles with more rapid tetanus force production have higher anaerobic capacity. This is in agreement with hypothesis 2, where we postulated that the rate of force generation would be related to glycolytic capacity. In contrast muscle fatigue resistance was not significantly correlated with any measures of metabolic capacity (Table 1). Therefore, contrary to hypothesis 2, oxidative capacity was not correlated with muscle endurance mechanics.

Contraction time was significantly negatively correlated with ryanodine receptor 1 mRNA concentration (Fig. 2; Table 1; r=–0.452, P=0.023). This finding demonstrates that faster lateral gastrocnemius muscles had greater expression of ryanodine receptor 1, in agreement with hypothesis 3.

Muscle relaxation time (time from last stimulus to half tetanus relaxation) was significantly positively correlated with SERCA2 (the slow isoform of SERCA) mRNA concentration (Fig. 3; Table 1; r=0.507, P=0.011) and significantly negatively correlated with troponin I type 2 (the fast isoform of troponin I) mRNA concentration (Fig. 4; Table 1; r=–0.413, P=0.035). These findings are in agreement with hypothesis 3. Maximum force produced (peak tetanic force) decreased between the 1st and 25th tetanus (Tables 2 and 3) in each experimental treatment. Inhibition of SERCA with 10 μmol l^–1 thapsigargin significantly decreased sustained muscle performance (fatigue resistance) compared with control treatments (P<0.001; Table 2). This result is linked to the finding that there was a tendency for fatigue resistance to be positively correlated with total SERCA activity (Table 1; r=0.376, P=0.051). However, thapsigargin treatment did not affect acute peak tetanic force production significantly (P=0.92; Table 2), rate of tetanus force (stress) production in the first tetanus after incubation (P=0.21; Table 2), rate of tetanus force (stress) production in the 25th tetanus of the fatigue run (P=0.78) or rate of tetanus force (stress) relaxation in the 1st tetanus after incubation (P=0.61; Table 2). These findings support hypothesis 3.

Fatigue resistance was not affected significantly by incubation in 10 μmol l^–1 of the ryanodine receptor blocker dantrolene (P=0.12; Table 3) when compared with control muscles. However, dantrolene treatment significantly decreased acute peak tetanus force production (P<0.001), and the rate of tetanus force production (P<0.001) and tetanus force relaxation (P<0.001) compared with controls (Table 3). These findings support hypothesis 3, that the ryanodine receptor is associated with determining rates of tetanus force (stress) production and relaxation.

Here, we show that lateral gastrocnemius muscles with relatively high maximal (sprint) performance have lower endurance performance (fatigue resistance), and that this trade-off between sprint and endurance performance is associated with a combination of variation in glycolytic metabolic capacity and variation in Ca²⁺ handling in the muscle. Interestingly, inter-individual variation in the capacity of mitochondria to produce ATP was not associated with inter-individual variation in fatigue resistance. Specifically, maximal (sprint) lateral gastrocnemius muscle performance was positively associated with both high anaerobic ATP production capacity and rapid Ca²⁺ release from the sarcoplasmic reticulum stores by ryanodine receptors. Conversely, increased SERCA2 mRNA concentration was associated with longer muscle relaxation times, and inhibition of total SERCA activity significantly decreased fatigue resistance.

In agreement with our first hypothesis, there was a trade-off between sprint and endurance performance. Previous studies have demonstrated a similar trade-off in skeletal muscle isolated from frog (Xenopus laevis) peroneus muscle (Wilson et al., 2002), toad (Bufo viridis) gastrocnemius muscle (Wilson et al., 2004) and mouse (Mus musculus) extensor digitorum longus muscle (Wilson and James, 2004). These previous studies have used the work loop technique to demonstrate that individuals with lower muscle fatigue resistance (measured as maintenance of work loop force or power) had higher work loop maximum force and power output. The consequence of such a skeletal muscle sprint–endurance trade-off is that vertebrate skeletal muscle cannot be optimized to produce both high force (or power) and high fatigue resistance. Therefore, in different species, specific muscle groups are likely to be tuned to greater sprint or endurance performance (different points on a sprint–endurance continuum) depending on the locomotory demands of the ecological niche of the species (Vanhooydonck et al., 2001). For example, species that rely on brief sprints to capture prey and to escape from predation are more likely to have muscle better suited for sprint performance in the relevant muscle groups. In contrast, species that rely on endurance performance, for example for migration, are more likely to have muscle better suited to endurance in the relevant muscle groups.

We hypothesized that variation in metabolic capacity between muscles would be associated with both variation in endurance muscle performance and variation in maximal force production. However, the associations between metabolic capacity and muscle mechanics were less clear than stated in our original hypotheses. Although variation in lactate dehydrogenase activity between muscles was significantly associated with variation in maximal force generation, variation in oxidative capacity was not correlated with variation in sustained muscle performance (fatigue resistance). The most parsimonious explanation is that oxidative ATP production was not limiting in our experimental conditions. It is possible, however, that oxidative ATP production can become limiting under natural conditions. In particular, failure to meet nutritional requirements and changes in thermal conditions could compromise mitochondrial function and thereby muscle performance (Seebacher et al., 2010). An experimental test of these relationships involving dietary and thermal treatments in live animals followed by analysis of muscle mechanics and biochemistry would significantly advance understanding of the relationship between energy metabolism and muscle physiology. In any case, our findings demonstrate that the widely used systems for classification of muscle fibre type may be too simplistic to predict muscle performance. The basic systems for classification of muscle fibre types are based upon differences in glycolytic and oxidative metabolic capacities, where slow, endurance fibres are thought to be more oxidatively poised and fast fibres have greater glycolytic capacity (Bottinelli and Reggiani, 2000). Nonetheless, the molecular traits used to categorize muscle fibre types often do not correspond to the expected patterns of muscle performance and sprint–endurance trade-offs (Gibb and Dickson, 2002; James et al., 2005; Plomgaard et al., 2006; Seebacher and James, 2008). Our data indicate that the reason for this may be that actual muscle contractile performance is determined not only by metabolic capacity but also by interactions between ecological factors and multiple molecular traits.

We hypothesized that inter-individual variation in the dynamics of calcium handling would be associated with variation in muscle mechanics. We found that muscles with greater SERCA2 mRNA concentrations had longer relaxation times. However, total SERCA activity did not influence muscle mechanics, such that SERCA2 mRNA concentrations were not indicative of total SERCA activity. A possible reason for this is that SERCA activity is regulated by phospholamban, and unphosphorylated phospholamban reduces SERCA activity (Verboomen et al., 1992). Phosphorylation of phospholamban by protein kinase A or calmodulin kinase is therefore a principal pathway for regulation of SERCA activity (East, 2000; Gustavsson et al., 2011). Phosphorylation of phospholamban is at least partly under sympathetic control via increases in cAMP concentration following stimulation of β₂-adrenoreceptors (McCormick et al., 2010). Hence, different levels of phospholamban phosphorylation will influence SERCA activity beyond protein and mRNA concentrations. More definite support for a role of SERCA in muscle performance and sprint–endurance trade-offs comes from our pharmacological studies: SERCA inhibition with thapsigargin reduced fatigue resistance, but did not influence tetanic force production. These findings were paralleled by the near-significant correlation between SERCA activity and fatigue resistance. Interestingly, inhibition of ryanodine receptors with dantrolene had the opposite effect; that is, it reduced force production but did not affect fatigue resistance. These different roles of Ca²⁺ release and re-sequestration mechanisms may partly explain the trade-off between muscle endurance and sprint performance, particularly because there is evidence from cardiac muscle that increases in SERCA1 expression cause concomitant decreases in ryanodine receptor and L-type Ca²⁺ channel density and activity (Lalli et al., 2001; Periasamy and Kalyanasundaram, 2007). Additionally, high ryanodine receptor activity is advantageous for fast muscle contraction, but it also leads to Ca²⁺ store depletion, which can accelerate muscle fatigue (Allen et al., 2008). High ryanodine receptor activity would require increased rates of Ca²⁺ re-sequestration into the sarcoplasmic reticulum to decrease relaxation times and thereby maintain muscle performance over time. High ryanodine receptor activity may therefore be incompatible with high SERCA2 concentrations, because our results and those of others (Sumbilla et al., 1999) indicate that elevated SERCA2 concentrations slow muscle relaxation. Hence, there is a functional trade-off between ryanodine receptor activity and both SERCA 1 and 2 isoforms.

The results from our dantrolene treatment are consistent with the negative correlation between ryanodine receptor mRNA concentration and contraction times, indicating that in this case mRNA concentration is representative of receptor activity. An increased density of ryanodine receptors in muscle has previously been correlated with increased shortening velocity and force in isolated muscle (Kandarian et al., 1992; Mänttäri and Järvilehto, 2005), and has been associated with training-induced improvements in trout endurance swimming performance (Anttila et al., 2008). Use of dantrolene to block ryanodine receptors in rat lateral gastrocnemius muscle demonstrated the importance of the calcium release channel in affecting rates of tetanus stress production. Dantrolene is used clinically in humans to inhibit calcium release via ryanodine receptors from the sarcoplasmic reticulum, for example to reduce the likelihood of death of malignant hyperthermia patients during anaesthesia (Mackrill, 2010). Dantrolene is also used clinically to reduce the incidence of the skeletal muscle syndrome exertional rhabdomyolysis in racehorses (Edwards et al., 2003). Therefore, these drug treatments highlight the importance of SERCA activity and ryanodine receptors in influencing inter-individual differences in sprint and endurance performance. The next step is to determine the significance of these mechanisms for the sprint–endurance trade off at the level of the whole animal, which could provide the mechanistic explanation for sporting success and inter-individual variation in ecologically important measures of performance.

Thanks to two anonymous referees whose comments helped to improve this manuscript.