Temperature effects on Ca²⁺ cycling in scombrid cardiomyocytes: a phylogenetic comparison

One of the most important events in vertebrate evolution is the transition from ectothermy to endothermy. This transition is difficult to study because endothermy is a derived characteristic in birds and mammals. The family Scombridae includes highly active regionally endothermic (tunas) and ectothermic (mackerel and bonito) fishes that occupy diverse habitats, including cold, temperate and tropical waters (Gunn and Block, 2001). In addition, within the family Scombridae, the tribe Thunnini contains guilds of closely related tunas with recent common ancestry such as the yellowfin (tropical) and bluefin (cold temperate) tunas (Chow and Kishino, 2010). Thus, scombrid fishes provide a model lineage to investigate specialisations that evolved in response to endothermy, and those that developed in response to thermal niche expansion.

Among tunas, the bluefin tunas have evolved the highest degree of endothermy, possessing internalised slow twitch oxidative (red) muscle and countercurrent heat exchangers that conserve metabolic heat produced in the locomotor musculature, viscera, brain and eyes (Carey and Lawson, 1973; Carey et al., 1984; Linthecum and Carey, 1971). Electronic tagging studies indicate that bluefin tunas occupy a wide thermal niche. Atlantic bluefin tuna range from the frigid waters off Canada in swimming sea surface temperatures (SST) as low as 0.02°C, to the warmer spawning grounds in the Gulf of Mexico (SST as high as 30°C) (Teo et al., 2007; Walli et al., 2009; Lawson et al., 2010). Owing to a high degree of endothermy, the bluefin tuna maintains an elevated body temperature with a mean visceral temperature of 21–23°C and a maximum of over 33°C (Block et al., 2001; Block et al., 2005; Walli et al., 2009). Electronic tagging of juvenile Pacific bluefin in the eastern Pacific demonstrates occupation of waters that average 17.8°C (Kitagawa et al., 2004; Boustany et al., 2010). Maintenance of a thermal excess of 1–10°C above ambient is routine in these juvenile bluefin despite the cool California Current waters (Marcinek et al., 2001a; Kitagawa et al., 2007).

Yellowfin tunas also have internalised slow-twitch muscles but fewer countercurrent heat exchangers (Graham and Dickson, 2004). These fish have a more tropical distribution (Weng et al., 2009; Shaefer et al., 2007). Juvenile yellowfin tunas tracked simultaneously with similar sized Pacific bluefin tunas in the California Current tend to occupy warmer waters to the south (mean SST of 22°C) and are limited in their northern geographic range (Shaefer et al., 2007). The capacity to maintain a thermal excess is not present in the ectothermic sister taxa of the Thunnus, such as Scomber (mackerel) and Sarda (bonito), because they lack countercurrent heat exchangers. Thus, comparative analyses of these fishes with tuna provide an excellent opportunity to examine the evolutionary sequence of events leading to endothermy and concomitant changes in performance and niche utilisation in extant teleosts.

Despite the ability to conserve metabolic heat, tunas face unique physiological challenges; their heart is close to the gills and separate from the supply of the vascular countercurrent heat exchangers (Carey et al., 1984; Brill et al., 1994). The heart also receives cold coronary blood directly from the gills (Brill, 1987; Brill et al., 2001). Thus, as with mackerel and bonito, the hearts of even the largest tunas operate at ambient water temperatures. Therefore, the tuna heart has an extra physiological burden compared with the mackerel and the bonito, as it must be able to function over a broad thermal range while delivering blood to muscle, viscera and cranial tissues that are often warmer than ambient water temperature and have high metabolic demands.

Tunas that have been studied (bluefin, yellowfin and skipjack) have evolved a robust cardiovascular system. Relative to body size, yellowfin and bluefin tuna hearts are significantly larger than other teleosts, have a thicker ventricular wall and an increased percentage of compact myocardium (Farrell et al., 1992). Furthermore, the tuna heart has large myoglobin reserves (Giovane et al., 1980), a higher oxygen affinity (bluefin tunas) (Marcinek et al., 2001b), high anaerobic enzyme activities (Dickson, 1995; Driedzic and Gesser, 1994) and a highly developed coronary circulation (Farrell et al., 1992), which appears to be reduced in the bonito and almost absent in the mackerel (Davie, 1991). Previous studies have also characterised yellowfin and skipjack tunas as possessing significantly higher intrinsic maximal heart rates (up to 200 beats min^–1), which are almost double those found in most teleosts (Brill et al., 2005; Farrell et al., 1992; Keen et al., 1995; Roberts and Graham, 1979). However, recent in situ heart measurements in bluefin, yellowfin and albacore tunas (Blank et al., 2004), and in vivo measurements in bluefin tunas (Clark et al., 2008), indicate heart rates are primarily 40–130 beats min^–1, which is not exceptional among teleosts (Farrell, 1991).

Variation in scombrid cardiovascular performance may be explained in part by specialisations in excitation–contraction (E–C) coupling. It was previously thought that most teleosts cycle Ca²⁺ across the sarcolemma primarily through the L-type Ca²⁺ channel or the Na⁺/Ca²⁺ exchanger (NCX) (Keen et al., 1992; Shiels and Farrell, 1997; Shiels et al., 2006; Vornanen et al., 2002) with the sarcoplasmic reticulum (SR) contributing negligible amounts of Ca²⁺ for contraction. However, more recent evidence suggests that the role of the L-type Ca²⁺ channel and the SR in teleost E–C coupling depends on a number of factors, including temperature, frequency of contraction, phylogeny and cardiovascular capacity (Hove-Madsen, 1992; Moller-Nielsen and Gesser, 1992; Keen et al., 1994; Shiels and Farrell, 1997; Shiels et al., 1998; Aho and Vornanen, 1999; Rivaroli et al., 2006; Galli et al., 2008). With regards to scombrids, recent physiological and electron microscopy work has demonstrated that the bluefin ventricle and atrium, the bigeye tuna ventricle and the yellowfin tuna atrium have evolved a more advanced ‘mammalian-like’ form of E–C coupling, with a significant proportion (40–50%) of activator Ca²⁺ used for contraction originating in the SR (DiMaio and Block, 2008; Galli et al., 2008; Shiels et al., 1999; Shiels et al., 2011). Furthermore, biochemical studies indicate bluefin tuna have increased SR Ca²⁺ ATPase (SERCA-2) activity and expression over a range of temperatures when compared with other scombrids (Landeira-Fernandez et al., 2004; Castilho et al., 2007). Lastly, Ca²⁺ entry through the L-type Ca²⁺ channel is greater and inactivation kinetics are faster in atrial myocytes of bluefin tuna than of Pacific mackerel (Shiels et al., 2004). Collectively, these results suggest enhanced Ca²⁺ cycling through E–C coupling pathways may have evolved within the scombrid lineage and may be crucial in the evolution of increased cardiac performance, endothermy and niche expansion. Importantly, these Ca²⁺ pumps and ion channels are all known to be sensitive to acute temperature change (Shiels et al., 2000; Shiels et al., 2002a; Shiels et al., 2002b; Landeira-Fernandez et al., 2004; Castilho et al., 2007; Shiels et al., 2011).

To gain a greater understanding of the evolution of endothermy and cardiovascular performance within scombrids, it is crucial to obtain information on myocyte E–C coupling in a phylogenetic selection of scombrid species. The present study examined the effects of physiologically relevant acute temperature change on aspects of two major E–C coupling pathways [L-type Ca²⁺ channel current (I_Ca) density and SR Ca²⁺ content] in ventricular myocytes from four members of the family Scombridae: Pacific bonito (Sarda chiliensis), Pacific mackerel (Scomber japonicus), yellowfin tuna (Thunnus albacares) and Pacific bluefin tuna (Thunnus orientalis). Atrial myocytes were also examined in Pacific bluefin tuna. These species encompass both ectothermic (mackerel and bonito) and endothermic (yellowfin and bluefin tuna) species, as well as cold-tolerant (mackerel and bluefin tuna), temperate (bonito) and tropical (yellowfin) species. By using a comparative approach, we aimed to identify specialisations in E–C coupling pathways that could have permitted increased cardiac performance, endothermy and/or increased thermal tolerance.

Bluefin tuna (Thunnus orientalis, Temminck and Schlegel 1844, n=6, fish mass=11.1±0.5 kg), yellowfin tuna (Thunnus albacares, Bonnaterre 1788, n=4, fish mass=4.33±0.3 kg) and Pacific bonito (Sarda chiliensis, Girard 1858, n=4, fish mass=1.84±0.4 kg) were caught at sea, by purse or seine or by hook and line, off the coast of San Diego, CA, USA. Fish were held aboard the Shogun fishing vessel in large wells flooded by seawater, or in captive farm pens (bluefin tuna) at Mariculture Del Norte, Ensenada, Mexico. All fish were transported by truck during a 9–11 h transport to the Tuna Research and Conservation Centre (TRCC) in pacific Grove, CA, USA. Pacific mackerel (Scomber japonicus, Houttuyn 1782, n=10, fish mass=0.52±0.1 kg) were caught in Monterey Bay, CA, USA and held in captivity at Monterey Bay Aquarium and the TRCC. Fish were kept for no less than 3 months prior to experimentation in 109 m³ tanks at 20±1°C and fed a diet of krill, squid, sardines and enriched gelatine, as previously described (Farwell, 2001). All experimental procedures were conducted in accordance with Stanford University institutional animal use protocols.

Ventricular myocytes from all species were isolated by adaptation of protocols previously developed for bluefin tuna (Shiels et al., 2004; Galli et al., 2009). Because atrial tissue is known to be more SR dependent than ventricular tissue in rainbow trout (Aho and Vornanen, 1999), we also isolated atrial myocytes from bluefin tuna to test whether the same holds true for this species and to serve as a measure of the sensitivity of our analyses of SR Ca²⁺ content. Fish were killed by pithing and decapitation (for mackerel) and severing of the spinal cord (for tuna), and the heart was removed intact [mean mass (in g): bluefin tuna, 37.6±1.4; mackerel, 1.14±0.2; bonito=3.99±0.5; yellowfin tuna, 17.1±4.9]. The bulbus arteriosus was cannulated and the heart was retrogradely perfused with an oxygenated, Ca²⁺-free isolation solution to promote cardiomyocyte relaxation and clear the heart of blood. For the tuna and bonito, the coronary artery was also cannulated and perfused, whereas that of the mackerel was not, because of size constraints. Isolation solution was perfused from a height of 60 cm for ∼10 min for mackerel, ∼15 min for bonito, and ∼20 min for tuna. Proteolytic enzymes were then added to the isolation solution and run through the cannulated bulbus arteriosus for ∼20 min in mackerel and ∼30 min in tuna and bonito. After enzymatic digestion, the ventricle and atrium were dissected and placed into separate Petri dishes containing fresh isolation solution. Tissues were gently triturated with a Pasteur pipette to free cardiomyocytes, and cells were stored at 20°C for the duration of the experiment (up to 8 h).

All solutions used were adapted from those previously developed for Pacific bluefin tuna (Shiels et al., 2004). The isolation solution consisted of (in mmol l^–1) 100 NaCl, 10 KCl, 1.2 KH₂PO₄, 4 MgSO₄, 50 taurine, 20 glucose and 10 Hepes, with pH adjusted to 6.9 with KOH at 20°C. The enzymatic digestion solution comprised collagenase (type IA), trypsin (type IX) and fatty-acid-free BSA added to the isolation solution. The extracellular solution used for current recordings contained (in mmol l^–1) 150 NaCl, 5.4 CsCl, 1.2 MgSO₄, 0.4 NaH₂PO₄, 3.2 CaCl₂, 10 glucose and 10 Hepes, adjusted to pH 7.7 with CsOH. Tetrodotoxin (TTX; 1 μmol l^–1) was added to the extracellular solution just before perfusion of cells in the bath in order to block fast Na⁺ channels. Caffeine (10 mmol l^–1) was similarly added just before use to promote release of SR Ca²⁺ stores. The pipette solution consisted of (in mmol l^–1) 130 CsCl, 5 MgATP, 15 tetraethylammonium chloride, 1 MgCl₂, 5 sodium phosphocreatine, 10 Hepes and 0.025 EGTA with pH adjusted to 7.2 with CsOH. CsCl and tetraethylammonium chloride were incorporated into the solutions in order to block K⁺ currents. Thapsigargin (2 μmol l^–1) and ryanodine (10 μmol l^–1) were added to the extracellular solution to block caffeine transient currents for control purposes. Nifedipine (2.5 μmol l^–1) was similarly added to block I_Ca. Except for TTX (purchased from Tocris Bioscience, Ellisville, MO, USA), all chemicals were purchased from Sigma-Aldrich (Dorset, UK).

Procedures for obtaining and analysing I_Ca and SR Ca²⁺ characteristics were adapted from established methods for fish preparations (Shiels et al., 2004; Shiels et al., 2006; Vornanen, 1997; Vornanen, 1998). A sample of cardiomyocytes, suspended in isolation solution, were added to the recording chamber and allowed to settle and attach to the bottom for ∼5 min. The cell chamber was superperfused at 23°C with extracellular solution containing TTX at a rate of ∼2 ml min^–1. Voltage clamp experiments were conducted in the whole-cell configuration using a PC 505B amplifier (Warner Instruments, Hamden, CT, USA) with a 202B headstage (Warner Instruments). Micropipettes were made from borosilicate glass (Garner Glass, Claremont, CA, USA) using a P-87 pipette puller (Sutter Instrument Company, Novato, CA, USA) and had resistances of 2.35±0.03 MΩ when filled with pipette solution. Junction potentials were zeroed prior to seal formation, and pipette capacitance was compensated for once a gigaohm seal was achieved. Access to the cell was obtained by rupturing the membrane patch with a brief voltage pulse. Capacitive transients were compensated for by adjusting the series resistance and cell capacitance compensation circuits. Mean series resistance was 9.7±0.5 MΩ. Currents were low-pass filtered at a frequency of 2 kHz and analysed offline using Clampfit 9.2 (Axon Instruments, Sunnyvale, CA, USA).

I_Ca was activated using the protocol illustrated in Fig. 1. Current–voltage relationships were elicited by depolarizing the cell from –80 mV to +60 mV using 500 ms square pulses in 10 mV increments. The amplitude of the L-type Ca²⁺ channel was calculated as the difference between the maximal inward current deflection and the current at the end of the depolarizing pulse. I_Ca was normalised for cell size by dividing I_Ca by cell capacitance to give I_Ca density in pA pF^–1. Current–voltage curves were obtained at 15, 19 and 23°C by adjusting superfusate temperature through a temperature-controlled rapid solution changer (SC-20 in-line heater/cooler, CL-100 bipolar temperature controller; Warner Instruments). When possible, the effects of all three temperatures were measured in each cell. However, because of problems in maintaining a high quality seal (in the gigaohm range), some cardiomyocytes were only tested at one or two of the experimental temperatures. The order of temperature exposure was reversed (perfused first at 15°C and then warmed to 19 and 23°C) in some experiments, and this did not affect the results. Inactivation kinetics of I_Ca and the time constants of fast (τ_f) and slow inactivation (τ_s) were obtained by fitting a second-order (mackerel and tuna) or first-order (bonito, τ) exponential curve to the decaying portion of the I_Ca using the Chebyshev procedure (Clampfit software, Axon Instruments). Charge density of I_Ca (Q_Ca) was calculated as the integral of the peak I_Ca for each current–voltage curve and is given in pC pF^–1. This value was used to calculate the change in total cellular Ca²⁺ resulting from Ca²⁺ influx through L-type Ca²⁺ channels (Δ[Ca²⁺]) using cell volume (Vornanen, 1997). Cell volume was calculated from the measured cell capacitance and the surface:volume ratio of the cells. Cardiomyocytes were considered to be flat elliptical cylinders with an axis ratio of 1:2 for the elliptical cross section (Vornanen, 1997; Vornanen, 1998). Δ[Ca²⁺] was expressed as a function of non-mitochondrial volume (Vornanen, 1997; Vornanen, 1998).

By rapidly changing between ports on the rapid solution changer, caffeine (10 mmol l^–1) was briefly (∼1 s) administered to induce SR Ca²⁺ release. Ca²⁺ released by the SR is extruded from the cell via the NCX, resulting in an inward Na⁺ current (I_NCX). This Na⁺ current is proportional (3:1) to the Ca²⁺ released by the SR, and thus the time integral of the I_NCX can be used to calculate the Ca²⁺ released (in pC) by the SR (Varro et al., 1993). This value was adjusted for cell size by dividing it by the cell capacitance (pC pF^–1). Integrating the I_NCX and cell volume gave total Ca²⁺ released by the SR (in μmol l^–1). Cell volumes were calculated by extrapolating surface area measurements from cell capacitance (in pF) and using previously developed calculations for surface:volume ratios and specific membrane capacitance (Vornanen, 1997). Preliminary studies in each species and at each temperature showed that at least 25 pulses from –80 mV to 0 mV (500 ms) were necessary to achieve a steady-state SR Ca²⁺ load (ssSR_load). Therefore, 25 pulses were used in all subsequent experiments when measuring ssSR_load. Maximal SR Ca²⁺ load (maxSR_load) was calculated as the caffeine-induced I_NCX after a square wave pulse from –80 to +50 mV for 6–9 s. ssSR_load and maxSR_load Ca²⁺ loads were measured in ventricular myocytes from all species at each test temperature (15, 19 and 23°C, as described in the previous section). Additionally, as SR dependence is known to be higher in atrial than in ventricular muscle, we measured ssSR_load and maxSR_load in atrial myocytes from one of the scombrid species, the bluefin tuna. This allowed assessment of the sensitivity of our SR_load measurements.

With the exception of original traces and voltage protocols, data are given as means ± s.e.m. N-values are the number of cells and are given in each figure legend. A one-way ANOVA with the Student–Newman–Keuls test was used to evaluate the effects of temperature on maximal and steady-state SR Ca²⁺ content, I_Ca density and I_Ca inactivation kinetics. Data were considered significantly different if P<0.05.

The size of scombrid cardiomyocytes was compared by measuring the cell capacitance. Bluefin and yellowfin ventricular cardiomyocytes had a significantly smaller cell capacitance than the bonito and the mackerel (Table 1).

In all species a voltage step from –80 mV to 0 mV in the presence of TTX gave rise to a slow activating and inactivating I_Ca that varied in amplitude and kinetics according to temperature (Fig. 1A). To confirm this current originated from the L-type Ca²⁺ channel we applied nifedipine (2.5 μmol l^–1), a specific L-type Ca²⁺ channel inhibitor, that completely abolished I_Ca (Fig. 1A dotted line).

The temperature dependence of the I_Ca–voltage relationships for each scombrid species tested are shown in Fig. 1B–E. Activation of I_Ca began at –40 mV, peaked at ∼0 mV and reversed at approximately –50 mV. Peak I_Ca amplitude was similar between mackerel and yellowfin tuna at each temperature tested (Fig. 1B,D). I_Ca inactivation kinetics for these species were also similar (Table 1), and as a result, Q_Ca (the charge carried by I_Ca) was not significantly different between the two species over the range of temperatures tested (Table 1). Peak I_Ca density and Q_Ca in bonito myocytes was significantly smaller than in bluefin tuna, yellowfin tuna and mackerel myocytes at all temperatures tested (Fig. 1B–D and Table 1). The decay of bonito I_Ca was better fit by a single exponential equation. Bluefin tuna peak I_Ca was larger than that of bonito, but significantly smaller than the mackerel and yellowfin tuna peaks (Fig. 1E). Furthermore, bluefin tuna I_Ca inactivation kinetics were generally slower than those of mackerel and yellowfin tuna, leading to a smaller Q_Ca at 19 and 15°C (Fig. 1B–D, Table 1).

In this and subsequent sections, Q₁₀ values have been calculated across the range of 23 to 15°C. In all species tested, acute reductions in temperature from 23 to 19 to 15°C led to a significant decrease in peak I_Ca density (Fig. 1). Although the temperature sensitivity of I_Ca in mackerel and yellowfin tuna was relatively high (Q₁₀=3.1 and 2.7, respectively), a cold-induced slowing of I_Ca kinetics (τ_s and τ_f) allowed Q_Ca to remain unaffected by acute temperature reductions (Q₁₀=1.1 and 1.2, respectively; Table 1). In bluefin tuna, the cold-induced slowing of I_Ca kinetics was not significant, and as a result, reductions in temperature led to both a significant decrease in peak I_Ca (Q₁₀=2.6) and Q_Ca (Q₁₀=1.8; Table 1). In bonito, although τ was significantly slower at cold temperatures, the dramatic effect of cold temperature on peak I_Ca (Q₁₀=11.0) caused Q_Ca to decline significantly (Q₁₀=4.23; Table 1).

Atrial cardiomyocytes are known to more SR dependent than ventricular cardiomyocytes, relying more strongly on SR Ca²⁺ contributions. We assessed ssSR_load and maxSR_load in both ventricular and atrial cardiomyocytes from the bluefin tuna. Application of caffeine caused the cells to contract strongly and induced a large inward current (I_NCX), as shown in Fig. 2A (ventricular myocyte) and B (atrial myocyte). This current was completely abolished by pre-incubation with SR inhibitors (10 μmol l^–1 ryanodine and 2 μmol l^–1 thapsigargin) for 5 min (not shown). The time integral of this current, as shown in Fig. 2C (ventricular myocyte) and D (atrial myocyte), was used to assess SR Ca²⁺ content at the time of caffeine application at each temperature.

Under stead-state conditions, bluefin tuna atrial myocytes had significantly larger SR_load values compared with ventricular myocytes (Fig. 2E,G), consistent with increased SR content in atrial cells. This result was also apparent under maximal conditions, but only significant at 23°C (Fig. 2F,H).

Having verified our technique, we went on to measure SR_load in the other scombrid cardiomyocytes. ssSR_load and maxSR_load was greater in mackerel and bluefin ventricular myocytes compared with yellowfin tuna and bonito (Figs 2 and 3). This difference was found to be statistically significant at all temperatures tested, except at 19°C, where bluefin ssSR_load was not significantly different from that of bonito (P>0.05). No significant differences in ssSR_load or maxSR_load were observed between mackerel and bluefin tuna, or between yellowfin tuna and bonito, at any temperature tested (Figs 2 and 3).

The ssSR_load in ventricular and atrial cardiomyocytes from the bluefin tuna decreased significantly when the temperature was reduced from 21 to 15°C (Q₁₀=2.1 and 2.2, respectively), and was more temperature sensitive within the upper temperature range (19–21°C; Fig. 2E,G). Ventricular maxSR_load was significantly decreased between 21 and 15°C (Q₁₀=1.8), whereas atrial maxSR_load was only temperature sensitive at the lower temperature range (15–19°C; Fig. 2F,H). For the other scombrids, ssSR_load in ventricular myocytes was only sensitive to temperature in bonito (Q₁₀=1.3), whereas maxSR_load decreased significantly with a temperature reduction from 21 to 15°C in the bonito and yellowfin (Q₁₀=1.7 and 1.8, respectively) and was more temperature sensitive in the upper thermal range (Fig. 3).

The major findings from our study are, (1) the magnitude of ventricular I_Ca density does not correlate with endothermy, cardiovascular capacity or thermal niche expansion, (2) SR Ca²⁺ content is greater in the more cold-tolerant species (bluefin tuna and mackerel), and (3) bonito utilises a different E–C coupling strategy from mackerel and tunas.

Species-specific variability in the relative contribution of I_Ca to force development has previously been demonstrated in scombrids and supports the hypothesis that differences in cardiac performance may be related to differences in I_Ca density (Shiels et al., 2004). However, in the present study I_Ca and Q_Ca density were greatest in mackerel and yellowfin at all temperatures tested. This result suggests the magnitude of I_Ca density does not correlate with the evolution of cardiovascular capacity, endothermy or thermal niche expansion.

Interestingly, bonito peak I_Ca and Q_Ca was dramatically lower than the other scombrids at all temperatures tested. In fact, peak I_Ca density in bonito ventricular myocytes tested at 23°C was three times and six times lower than that recorded in ventricular myocytes of trout (tested at 21°C and acclimated to 17°C) and carp (tested at 21°C and acclimated to 23°C) (Vornanen, 1997; Vornanen, 1998). Similar results on L-type Ca²⁺ channel characteristics found in bonito have been reported in burbot ventricular myocytes (Shiels et al., 2006). Although direct comparisons are difficult because of different experimental conditions, low I_Ca and Q_Ca density in burbot was related to a smaller density of sarcolemmal L-type Ca²⁺ channels (Tiitu and Vornanen, 2003), whereas I_NCX density was robust and thought to be the primary pathway for sarcolemmal Ca²⁺ influx (Shiels et al., 2006). We propose that the NCX might be the primary pathway for sarcolemmal Ca²⁺ influx in the bonito, and might provide the majority of activator Ca²⁺ to the myofilaments (see SR discussion below).

All scombrid fishes showed significant reductions in I_Ca with decreasing temperature. In the mackerel and yellowfin, the slowing of inactivation kinetics of I_Ca offset the negative effects of temperature, resulting in a temperature-insensitive Ca²⁺ flux (Q₁₀=1.2 for mackerel, Q₁₀=1.1 for yellowfin). This effect has been observed in most other teleosts examined to date (Shiels et al., 2000), and may be a common protective mechanism against the depressive effects of the cold. In contrast, the Ca²⁺ carried by I_Ca in the bluefin tuna and mackerel appears to be temperature sensitive with distinctly different affects (Q₁₀=1.7 and 4.2, respectively). It is possible that temperature-induced alterations in heart rate and action potential shape and/or duration in these species may offset the negative effects of cold temperature on I_Ca (see Limitations of the study and future directions).

The ssSR_load in scombrid ventricular myocytes at 23°C ranged between 229 and 578 μmol l^–1 Ca²⁺. These values are similar to recently published data from warm-acclimated (18°C) freshwater teleosts: ssSR_load in ventricular myocytes from the rainbow trout, burbot and crucian carp ranged between 332 and 540 μmol l^–1 at 18°C (Shiels et al., 2002b; Haverinen and Vornanen, 2009). Although these results suggest scombrid ssSR_load is not exceptional among teleosts, the values presented here are three- to ninefold larger than those reported for rabbit (87 μmol l^–1 at 22°C) and rat (64 μmol l^–1 at 23°C) (Walden et al., 2009) ventricular myocytes. Furthermore, our values for maxSR_load suggest scombrid ventricular myocytes can store up to 628 μmol l^–1 of Ca²⁺ in the SR, which is tenfold larger than mammalian SR_load under steady-state conditions. Thus, bluefin tunas and cold-tolerant active teleosts in general, appear to have an exceptional capacity to store Ca²⁺ in the SR.

The large SR Ca²⁺ storage capacity of fish compared with mammalian cardiomyocytes does not appear to relate to differences in total SR volume (DiMaio and Block, 2008; Shiels et al., 2011; Bowler and Tirri, 1990). Rather, Haverinen and Vornanen suggest functional differences in Ca²⁺ management within the SR results in a higher total Ca²⁺ capacity (Haverinen and Vornanen, 2009). For example, the fish ryanodine receptor appears to have a lower affinity for Ca²⁺ binding than mammals, which would stabilise SR Ca²⁺ content (Vornanen, 2006). Potentially the cooler ambient temperatures experienced by fish cardiomyocytes (compared with mammals) requires higher Ca²⁺ content to initiate and maintain Ca²⁺-release events (Haverinen and Vornanen, 2009). Support for this contention comes from recent imaging of intracellular Ca²⁺ in bluefin tuna myocytes, which shows both steady-state SR Ca²⁺ utilisation during E–C coupling and increased SR Ca²⁺ utilisation in the cold (Shiels et al., 2011).

Species-specific differences in SR Ca²⁺ content do show significant variations within the scombrid lineage, which provides some evidence for linkage between thermal tolerance and niche expansion. Ventricular steady-state and maximal SR Ca²⁺ loads decrease in the order of bluefin tuna>mackerel>yellowfin tuna>bonito, with no statistical difference between bluefin tuna and mackerel. As far as the bluefin tuna is concerned, our finding of a comparatively large SR Ca²⁺ storage capacity is consistent with previous studies indicating an enhanced role for the SR in E–C coupling. Ultrastructural studies have revealed a significant SR density in bluefin tuna cardiomyocytes (DiMaio and Block, 2008), and protein work has demonstrated enhanced expression of SERCA-2 compared with other scombrids (Landeira-Fernandez et al., 2004; Landeira-Fernandez et al., 2007). Most recently, fluorescent microscopy has confirmed that the bluefin tuna SR releases Ca²⁺ on a beat-to-beat basis, contributing activator Ca²⁺ for contraction (Shiels et al., 2011). Thus, the bluefin tuna SR has an enhanced capacity to store, release and reuptake Ca²⁺ from the cytosol, which contributes to the generation of increased cardiac force and frequencies of contraction in this species.

Also consistent with previous studies on the significance of the SR in scombrids, the yellowfin tuna had significantly smaller SR_load values than the bluefin tuna. Isolated ventricular muscle from the yellowfin tuna has been shown to be unresponsive to SR inhibition with ryanodine and thapsigargin (Galli et al., 2008). Furthermore, SERCA-2 protein expression in the yellowfin tuna ventricle is significantly less than in the bluefin tuna (Landeira-Fernandez et al., 2004). Thus, the geographical distribution in warm warmer waters of the yellowfin tuna may not require the specialisations in E–C coupling that are seen in the bluefin tuna heart.

The equally large SR Ca²⁺ storage capacity of the mackerel and bluefin tuna was a somewhat surprising result in our study. Prior work suggested the Pacific mackerel has a lower expression of SERCA-2 than yellowfin, albacore and bluefin tuna (Landeira-Fernandez et al., 2004). However, ryanodine receptors have been identified in the mackerel, using western blot analysis (Castilho et al., 2007), and ryanodine has been shown to inhibit ventricular force of contraction by 20% (Shiels and Farrell, 2000). Although our results suggest SR Ca²⁺ loading was more temperature sensitive in the bluefin tuna and mackerel (higher Q₁₀ values), ssSR_load and maxSR_load were larger at any given temperature than in the more tropical species, the bonito and yellowfin. A large and maintained capacity to store Ca²⁺ in the SR at cold temperatures may allow these species to maintain Ca²⁺ cycling and cardiac function during exploitation of cold waters.

Our study aimed at identifying differences between scombrids in the capacity of their cardiomyocytes to move Ca²⁺ through the L-type Ca²⁺ channel and store Ca²⁺ in the SR. By using square-wave pulses at a single frequency to measure I_Ca, we were able to isolate these differences without interspecific variation in action potential shape or frequency. However, temperature independently affects the frequency of contraction and action potential shape and/or duration, which in turn can have profound affects on I_Ca, (Shiels et al., 2000) and SR loading (Shiels et al., 2002b). Thus, future experiments should be aimed at assessing I_Ca properties with physiologically relevant frequencies of contraction and action potentials as the command waveform.