Temperature dependence of cardiac sarcoplasmic reticulum function in rainbow trout myocytes

Eurythermal animals maintain aerobic performance throughout a range of acute temperature fluctuations. Rainbow trout (Oncorhynchus mykiss),for example, maintain swimming performance over a broad ambient temperature range (Randall and Brauner,1991; Matthews and Berg,1997; Reid et al.,1997), which implies that physiological systems that support exercise, such as the heart, also function well over a broad temperature range. The focus of the present study was to investigate how temperature influences cardiac function in trout heart and, in particular, how temperature affects Ca²⁺ flux through the sarcoplasmic reticulum (SR) in trout atrial myocytes.

In rainbow trout heart, Ca²⁺ delivery to the myofilaments during excitation—contraction (e—c) coupling primarily involves the influx of extracellular Ca²⁺ across the sarcolemmal membrane (SL)and secondarily involves the mobilization of the intracellular Ca²⁺stores of the SR (Keen et al.,1994; Aho and Vornanen,1999; Hove-Madsen et al., 1999, 2001; Harwood et al., 2000; Shiels et al., 2002a). The relative importance of the SR Ca²⁺ flux pathway in trout hearts is generally considered to be minor but can vary depending on a number of physiological parameters. For example, SR Ca²⁺ cycling appears to play a greater role in e—c coupling in (1) atrial tissue compared with ventricular tissue, (2) at low (<0.6 Hz) compared with high (>0.6 Hz)contraction frequencies, (3) after acclimation to cold temperatures(<12°C) and (4) after an acute change to warm temperatures(>18°C) (Hove-Madsen,1992; Keen et al.,1994; Shiels and Farrell,1997; Aho and Vornanen,1999; Shiels et al.,2002a).

In most adult mammalian hearts, the SR is the major source of activator Ca²⁺ during e—c coupling, with SL Ca²⁺ influx serving primarily as the trigger for SR Ca²⁺ release(Bers, 2001). The SR Ca²⁺ flux pathways are known to be temperature-sensitive in many mammals, and the inhibition of SR function at cold temperatures is a major cause of cold-induced contractile dysfunction (Wang et al., 1997, 2000). Cold temperatures increase the open probability of the mammalian SR Ca²⁺ release channels (ryanodine receptors) causing the leak of available SR Ca²⁺ into the cytoplasm(Sitsapesan et al., 1991). Cold-induced slowing of ion transport mechanisms prevents this Ca²⁺from being easily removed from the cytoplasm, contributing to a loss of Ca²⁺ homeostasis (Bers,1987; Sitsapesan et al.,1991).

Despite the fact that the fish heart continues to function at cold temperatures, early research on whole-muscle contractility in trout ventricle suggested that the teleost SR had a temperature dependency similar to that of the mammalian SR. This is because contractile force was found to be sensitive to ryanodine (an inhibitor of the SR Ca²⁺ release channel; Rousseau et al., 1987) at warm temperatures (18-25°C) but insensitive to ryanodine at cold temperatures(<15°C) (Hove-Madsen,1992; Driedzic and Gesser,1994; Keen et al.,1994). However, a study with isolated trout atrial tissue showed a 20% reduction in developed force after ryanodine treatment at 4°C(Aho and Vornanen, 1999),indicating maintained SR function in the cold and suggesting that, similar to mammals (Bers, 2001), SR involvement may be greater in fish atrium than in fish ventricle. These findings were supported in a recent cellular study that demonstrated maintained SR function at cold temperatures in rainbow trout atrial myocytes(Hove-Madsen et al., 2001). However, the experimental conditions in this cellular study favoured a large SR Ca²⁺ content, which could result in an overestimation of SR capabilities in the cold. High intracellular Na⁺ concentrations(approximately 17 mmol l^-1) in conjunction with long depolarizing pulses (>30 s) to positive membrane potentials were used in this study to Ca²⁺-load the SR via the reverse mode (Ca²⁺ in,Na⁺ out) of the sarcolemmal Na⁺/Ca²⁺exchanger (NCX; Hove-Madsen et al., 1998, 1999, 2001). Indeed, mammalian and teleost studies have shown that experimentally elevating intracellular Na⁺ concentrations to a range of 15-20 mmol l^-1(concentrations in vivo are approximately 7 mmol l^-1)induces supra-physiological Ca²⁺ influx via the NCX(Hove-Madsen et al., 2000; Bers, 2001). Thus, it is possible that previous studies with fish have overestimated the physiological SR Ca²⁺ loading and release capabilities at cold temperatures. In fact, cellular studies with permeabilized trout ventricular myocytes indicated that the trout SR Ca²⁺-ATPase, which pumps cytosolic Ca²⁺ back into the SR during relaxation, is temperature-dependent,with an average Q₁₀ of approximately 1.6 over a temperature range of 5-20°C (Aho and Vornanen,1998; Hove-Madsen et al.,1998). As trout can experience acute temperature fluctuations of as much as ±10°C either while transversing thermoclines or as a result of diurnal changes in shallow streams(Matthews and Berg, 1997; Reid et al., 1997), the temperature dependency of the SR Ca²⁺-ATPase, combined with temperature-sensitive SR Ca²⁺ release channels, could limit the utility of the SR as a source of activator Ca²⁺ during e—c coupling in trout heart.

In the present study, we were interested in assessing the physiological efficacy of the trout atrial SR at different temperatures. We acknowledge that it is difficult to be truly physiological using isolated voltage-clamped myocytes. However, to better approach the physiological situation, we included 10 mmol l^-1 Na⁺ in our pipette solutions and, in addition to stimulating cells with conventional square (SQ) voltage-clamp pulses of constant dimensions, we applied physiological action potentials(APs) whose shape and rate of firing varied dynamically with changes in temperature. We assessed SR Ca²⁺ accumulation using rapid caffeine application, and SR Ca²⁺ release, by examining the effect of SR Ca²⁺ content on the inactivation kinetics of I_Ca (L-type Ca²⁺ channel current). We find that the shape of the AP and the rate of AP firing may be critical for ensuring adequate atrial SR Ca²⁺ cycling during temperature change in rainbow trout in vivo.

Rainbow trout Oncorhynchus mykiss Walbaum (female, mean mass 263.6±25.5 g, N=22) were obtained from a local fish farm(Kontiolahti, Finland) and held in the laboratory in a 5001 tank receiving aerated tapwater. The tank temperature was 14±1°C. Fish were held for a minimum of four weeks prior to experimentation and were fed daily on commercial trout pellets (Biomar, Brande, Denmark). The photoperiod was a 15 h:9 h L:D cycle. All procedures were in accordance with local animal handling protocols.

A detailed description of the myocyte preparation has been previously published (Shiels et al.,2000; Vornanen,1998). Briefly, fish were stunned with a blow to the head, the spine was cut just behind the brain and the heart was excised. The heart was then perfused first with the isolating solution for 8-10 min, and then with the proteolytic enzyme solution for 20 min. After enzymatic treatment, the atrium was removed from the ventricle and placed in a small dish containing isolating solution. The atrium was cut into small pieces with scissors and then triturated through the opening of a Pasteur pipette. The isolated myocytes were stored in fresh isolating solution at 14°C.

The isolating solution contained: NaCl, 100 mmol l^-1; KCl, 10 mmol l^-1; KH₂PO₄, 1.2 mmol l^-1;MgSO₄, 4 mmol l^-1; taurine, 50 mmol l^-1;glucose, 20 mmol l^-1; and Hepes, 10 mmol l^-1; adjusted to pH 6.9 with KOH. For enzymatic digestion, collagenase (Type IA), trypsin(Type IX) and fatty-acid-free bovine serum albumin (BSA) were added to this solution. The physiological saline used as the extracellular solution for recording Ca²⁺ currents (I_Ca) and NCX currents(I_NCX) contained: NaCl, 150 mmol l^-1; CsCl, 5.4 mmol l^-1; MgSO₄, 1.2 mmol l^-1;NaH₂PO₄, 0.4 mmol l^-1; CaCl₂, 1.8 mmol l^-1; glucose, 10 mmol l^-1; and Hepes, 10 mmol l^-1; adjusted to pH 7.7 with CsOH. Additionally, 1 μmol l^-1 tetrodotoxin (TTX) was added to the perfusate to block fast Na⁺ channels. Caffeine (10 mmol l^-1) was added to the extracellular solution to release the SR Ca²⁺ stores. For AP recordings, the extracellular solution was modified by replacing the CsCl with equimolar KCl, pH balancing with KOH and by omitting the TTX. The pipette solution for I_Ca and I_NCX recordings contained: CsCl,130 mmol l^-1; MgATP, 5 mmol l^-1; tetraethylammonium chloride (TEA), 15 mmol l^-1; MgCl₂, 1 mmol l^-1; oxaloacetate, 5 mmol l^-1;Na₂-phosphocreatine, 5 mmol l^-1; Hepes, 10 mmol l^-1; EGTA, 0.025 mmol l^-1; and Na₂GTP, 0.03 mmol l^-1; pH was adjusted to 7.2 with CsOH. For action potential recordings, the pipette solution contained: K-aspartate, 125 mmol l^-1; KCl, 15 mmol l^-1; MgCl₂, 1 mmol l^-1; MgATP, 5 mmol l^-1; EGTA, 0.05 mmol l^-1;Na₂-phosphocreatine, 5 mmol l^-1; and Hepes, 10 mmol l^-1; adjusted to pH 7.2 with KOH. All drugs, with the exception of TTX (Tocris, Bristol, UK), were purchased from Sigma (St Louis, MO, USA).

Myocytes (average cell capacitance 36.2±0.92 pF, N=92) were superfused at a rate of 2 ml min^-1 with extracellular solution at 7°C, 14°C or 21°C. Each cell was tested at one experimental temperature only. The extracellular solution and the solution flowing from a rapid solution changing device (RS200, Biologic, Claix, France) was heated or chilled by water bath circuits before emptying into the recording chamber(RC-26, Warner Instrument Corp, Brunswick, USA; volume 150 μl). A thermocouple was placed inside the recording chamber and positioned no less than 5 mm from the cell to ensure it was experiencing the desired temperature. Rapid application (approximately 50 ms) of caffeine was achieved by switching between two barrels of the rapid solution changer; one containing control extracellular solution and the other containing the same solution plus caffeine, both flowing at approximately 0.5 ml min^-1.

Whole-cell voltage-clamp experiments were performed using an Axopatch 1D amplifier with a CV-4 1/100 headstage (Axon Instruments, Foster City, CA,USA). Pipettes had a resistance of 2.4±0.11 MΩ when filled with pipette solution. Junction potentials were zeroed prior to seal formation. Pipette capacitance (9.7±0.5 pF) was compensated after formation of a GΩ seal. Mean series resistance was 6.1±0.2 MΩ. Membrane capacitance was measured using the calibrated capacity compensation circuit of the Axopatch amplifier. Signals were low-pass filtered using the 4-pole lowpass Bessel filter on the Axopatch-1D amplifier at a frequency of 2 kHz for I_Ca and 10 kHz for action potentials, and were then analyzed off-line using pClamp 6.0 software (Axon Instruments).

Rainbow trout atrial myocytes were stimulated to elicit APs at 7°C,14°C or 21°C at a frequency that corresponded to the resting heart rate of rainbow trout in vivo at each of these temperatures. The temperature/frequency pairs were as follows: 0.6 Hz at 7°C, 1.0 Hz at 14°C and 1.4 Hz at 21°C (Aho and Vornanen, 2001; Tuurala et al., 1982; Farrell et al.,1996). APs were elicited by the minimum voltage pulse able to trigger a self-sustained AP (1 ms, 0.8 nA). The resultant AP waveforms(Fig. 1) were subsequently used to provide a more physiologically relevant indication of how temperature and frequency affected SR Ca²⁺ accumulation and release in rainbow trout atrial myocytes. In all experiments, the capacitive transients resulting from the AP waveform were compensated for by the amplifier.

We used either SQ pulses (-80 mV to +10 mV, 200 ms, interpulse holding potential of -80 mV) or AP pulses (see Fig. 1) and varied both temperature and pulse frequency to examine their effect on SR Ca²⁺ accumulation. SR Ca²⁺accumulation was assessed by the application of caffeine, which induces the release of Ca²⁺ from the SR. This Ca²⁺ is then extruded from the cell via the NCX generating an inward current(Fig. 2) that is directly proportional to the Ca²⁺ released from the SR(Varro et al., 1993). The time integral of the caffeine-induced I_NCX current was used to calculate the SR Ca²⁺ content (in pC) at the time of caffeine application. This value was expressed per unit capacitance (pC pF^-1). The SR Ca²⁺ content was also expressed in μmol l^-1 and was calculated from the integral of I_NCX and the cell volume. Cell volume was calculated from cell surface area [obtained by measurements of cell capacitance (pF) and assuming a specific membrane capacitance of 1.59 μF cm^-2] and the surface:volume ratio of 1.15 [determined experimentally in previous studies; see Vornanen(1997)]. Finally, SR Ca²⁺ content was expressed as a function of myofibrillar volume(40%), as determined previously (Vornanen,1998).

Possible interaction between extracellular Ca²⁺ influx via the L-type Ca²⁺ channel (I_Ca) and intracellular Ca²⁺ release through the SR Ca²⁺ release channel was examined by investigating the effect of SR Ca²⁺ content on the inactivation of I_Ca. Previous work has shown that if there is interaction between these Ca²⁺ fluxes, I_Cainactivation is faster due to SR Ca²⁺-dependent inactivation(Lipp et al., 1992; Sham, 1997). To test whether this mechanism occurs in trout atrium, we fitted either a double(τ_f and τ_s; SQ) or single (τ; AP)exponential function to the inactivating portion of I_Ca immediately after the SR Ca²⁺ stores were depleted with caffeine (loading pulse 1) and then again after SR Ca²⁺ content had reached steady-state(after 25 pulses). In some cells, the effect of progressive SR Ca²⁺accumulation on I_Ca with each pulse was examined between the first and 100th pulse. Single and double exponential fits were calculated using the standard exponential fitting procedure with the Chebyshev transformation with Clampfit 6.0 software (Axon Instruments).

One-way repeated measures analysis of variance (RM ANOVA) was used to compare the effect of pulse number on SR Ca²⁺ content. One-way ANOVA was used to test the effects of temperature, frequency and shape of the stimulation pulse on SR Ca²⁺ accumulation. Significant differences(P<0.05) were assessed with Student—Newman—Keuls (SNK) post-hoc analysis. The effects of steady-state SR Ca²⁺content on the inactivation kinetics of I_Ca were tested using paired Student's t-tests.

Our first experiments set out to investigate the effect of temperature on steady-state SR Ca²⁺ accumulation in rainbow trout atrial myocytes. At the onset of each experiment, caffeine was applied to the cell to release SR Ca²⁺ stores. The cell was then stimulated with 25 SQ-voltage pulses at 1.0 Hz and caffeine was re-applied. Application of caffeine caused the myocytes to contract strongly and induced a large inward current(I_NCX), as shown in Fig. 2. The time integral of this current was used to assess the SR Ca²⁺ content at the time of caffeine application (see inset in Fig. 2). Myocytes were then stimulated with 50 SQ-voltage pulses at 1.0 Hz and caffeine was again applied to assess SR Ca²⁺ accumulation. Finally, the cell was stimulated with 100 pulses at 1.0 Hz and SR Ca²⁺ accumulation was again assessed. The mean results for all cells at each temperature are given in Fig. 3A. At both 14°C and 7°C, the trout SR Ca²⁺ content had reached a steady-state of approximately 1 pC pF^-1 by 25 pulses (equivalent to 474±75μmol l^-1 Ca²⁺ and 365±43 μmol l^-1 Ca²⁺ at 14°C and 7°C, respectively, N=8-16). SR Ca²⁺ content did not increase significantly with additional pulses between 25 and 100(Fig. 3, and also see Fig. 4).

At 21°C, SR Ca²⁺ accumulation was more variable, as only four out of eight cells demonstrated SR Ca²⁺ accumulation under these experimental conditions. Those cells that accumulated Ca²⁺into the SR also reached a steady-state of approximately 1 pC pF^-1after 25 pulses (equivalent to 460±76 μmol l^-1Ca²⁺) and did not change significantly with an additional number of pulses up to 100. Only the cells that were able to accumulate Ca²⁺into the SR are shown in Fig. 3A (broken line).

Because the frequency of cell depolarization should affect the Ca²⁺ available for uptake into the SR, our next experiment examined SR Ca²⁺ content at each temperature when the 200 ms SQ pulses were applied at a frequency that corresponds to the heart rate of the rainbow trout in vivo at each temperature. These results are given in Fig. 3B. In contrast to the results with 1.0 Hz stimulation, when frequency was increased to 1.4 Hz at 21°C, all cells displayed significant SR Ca²⁺ accumulation and reached a steady-state of 0.78±0.09 pC pF^-1 (equivalent to 367±42 μmol l^-1 Ca²⁺) after 25 pulses(Fig. 3B). At 14°C, the physiologically relevant frequency was 1.0 Hz and the results were no different from those presented in Fig. 3A. At 7°C, a physiologically relevant stimulation frequency was 0.6 Hz, and the SR Ca²⁺ content after 25 pulses and 50 pulses was similar to that at 1.0 Hz but increased significantly (P<0.05, N=8) after 100 pulses (Fig. 3B). As a result, SR Ca²⁺ content was 2.1±0.3 pC pF^-1 or 972±142 μmol l^-1 Ca²⁺ after 100 pulses at 0.6 Hz, which was almost double the steady-state amount observed after stimulation with 100 SQ pulses at 1.0 Hz (550±90 μmol l^-1 Ca²⁺) (Fig. 3A).

The experiments with SQ-voltage-clamp pulses indicated that SR Ca²⁺ content can vary as a function of temperature and frequency. However, in vivo, extracellular Ca²⁺ entry into the myocyte does not occur during a SQ-voltage pulse but rather during an AP. As both the shape and rate of APs vary considerably with temperature(Coyne et al., 2000; Harwood et al., 2000; Shiels et al., 2000), studies with AP clamp may be essential for understanding the physiological capacity of the SR as a function of temperature in fish hearts. Indeed, we observed a pronounced effect of acute temperature change on AP duration, which is clearly illustrated in Fig. 1. AP duration at 50% repolarization varied by more than 50% among temperatures(increasing from approximately 90 ms at 21°C to 160 ms at 14°C and to 290 ms at 7°C; Fig. 1). Thus, our next experiments set out to examine the Ca²⁺-accumulating abilities of the trout SR during AP stimulation.

When myocytes were stimulated at a physiologically relevant frequency, with the AP waveform specific for the test temperature, SR Ca²⁺ content reached a steady-state after 25 pulses at all three temperatures(2.8±0.3 pC pF^-1, 2.7±0.3 pC pF^-1 and 2.3±0.2 pC pF^-1 at 7°C, 14°C and 21°C,respectively; Fig. 3C) and the steady-state temperature and frequency-dependent changes in SR Ca²⁺content observed with SQ pulses were no longer evident. The Ca²⁺content of the SR was greater (1043±189 μmol l^-1Ca²⁺, 1138±173 μmol l^-1 Ca²⁺ and 1095±142 μmol l^-1 Ca²⁺ at 7°C, 14°C and 21°C, respectively) when cells were stimulated with AP pulses compared with 200 ms SQ pulses (664±180 μmol l^-1 Ca²⁺,474±75 μmol l^-1 Ca²⁺ and 367±42 μmol l^-1 Ca²⁺ at 7°C, 14°C and 21°C,respectively) applied at the same frequency(Fig. 3B).

I_Ca records initiated immediately after depletion of SR Ca²⁺ by caffeine allowed us to monitor the effects of SR Ca²⁺ re-accumulation on I_Ca. Inactivation kinetics of I_Ca were significantly faster when the SR was loaded with Ca²⁺, and this observation is consistent with a greater SR Ca²⁺ release resulting in greater SR-Ca²⁺-dependent inactivation. Fig. 4 shows a representative recording of the progressive quickening of inactivation as SR Ca²⁺ content increases with each pulse between 1 and 25 at 21°C. After 25 pulses, inactivation kinetics did not change significantly with additional pulses, supporting the finding that SR Ca²⁺ content had reached a steady-state. Fig. 5 shows representative I_Ca currents with both SQ and AP pulses with the SR empty and at a steady-state at 14°C. The mean inactivation kinetics of I_Ca for each condition and temperature are given in Tables 1, 2. At 21°C, fast inactivation kinetics (τ_f) and slow inactivation kinetics(τ_s) were significantly faster than at 14°C and were significantly faster under the SR-loaded condition(Table 1). At 7°C, the inactivation of I_Ca was a better fit with a single exponential function. Inactivation of I_Ca at 7°C was faster as the SR refilled (Table 1). This suggests that, even with the reduced amplitude of I_Ca at 7°C(see Shiels et al., 2000),I_Ca elicited by a SQ stimulation pulse is still of sufficient magnitude to cause Ca²⁺-dependent inactivation of SL Ca²⁺ influx.

The Ca²⁺ current is more complex during AP clamp(Fig. 5B). The effect of SR Ca²⁺ content on inactivation kinetics was examined by fitting a single exponential to the first part of the inactivation of I_Ca[first 200 ms after peak current at 14°C (see Fig. 5B) and first 100 ms after peak at 21°C (not shown)]. The AP waveform could not be consistently modelled at 7°C and so inactivation kinetics at 7°C are excluded from this part of the analysis. At both 14°C and 21°C, the inactivation of I_Ca elicited by AP pulses was faster with a steady-state Ca²⁺ SR content (Table 2).

Because the voltage waveform (either SQ or AP) was the same for each pulse and the amplitude of I_Ca was not significantly different, the change in inactivation kinetics observed in these experiments is unlikely to result from voltage-dependent inactivation or Ca²⁺-dependent inactivation of Ca²⁺ entry through the channel. Thus, we conclude that there is SR Ca²⁺-dependent inactivation of the SL I_Ca in rainbow trout atrial myocytes and that it is independent of temperature between 7°C and 21°C.

The conventional SQ-voltage-clamp pulse helps characterize the kinetic behaviour of specific ion channels and separates the components of macroscopic current recordings but it does not emulate the change in membrane potential that occurs during an AP. This is a critical shortcoming in studies where the effects of temperature are being examined, as the shape of the cardiac AP(Coyne et al., 2000; Shiels et al., 2000) and the frequency of AP firing (Harwood et al.,2000) can both change as a function of temperature. The present study clearly shows that SR Ca²⁺ accumulation in rainbow trout atrial myocytes is neither temperature- nor frequency-dependent when cells are stimulated with AP pulses. This is a novel finding and is in contrast to the variability observed with 200 ms SQ stimulation pulses in the present study and with SQ pulses in other studies (Hove-Madsen et al., 1998, 1999). Our use of 200 ms SQ pulses and lower intracellular Na⁺ concentrations resulted in a lower SR Ca²⁺ content than that reported in previous studies, which showed high and constant SR Ca²⁺ content at 7°C and 21°C in rainbow trout atrial myocytes [approximately 1568 μmol l^-1Ca²⁺ (Hove-Madsen et al.,2001) versus approximately 475 μmol l^-1Ca²⁺ (present study)]. However, when AP stimulation pulses were used in the present study, SR Ca²⁺ content more than doubled and there was no temperature dependence. These findings suggest that temperature-and frequency-dependent modulation of the AP may help to ensure a large and temperature-independent SR Ca²⁺ content in rainbow trout atrial myocytes in vivo.

The increase in AP duration at cold temperatures, and the increase in AP firing rate at warm temperatures, may underlie the temperature independence observed in the present study (see Fig. 1). Indeed, the duration of the AP plateau will profoundly affect SL Ca²⁺ influx and thus the Ca²⁺ available for uptake into the SR. Additionally, changes in stimulation frequency affect diastolic Ca²⁺ load in trout atrial myocytes (see Shiels et al., 2002b), which could also influence the activity of the SR Ca²⁺ ATPase. Thus, we suggest that temperature-dependent modulation of the AP shape and firing rate may help to offset the known temperature sensitivity of the SR Ca²⁺ATPase (Q₁₀ of approximately 1.6; Aho and Vornanen, 1998; Hove-Madsen et al., 1998).

Our results show greater SR Ca²⁺ accumulation with AP stimulation than with SQ pulses in trout atrial myocytes. This may be explained by the fact that I_Ca generated by an AP has a sustained phase during the AP plateau. The sustained phase is a result of an increased driving force for Ca²⁺ as a result of repolarization (to approximately 0 mV) after the peak and the reactivation of L-type Ca²⁺ channels during the AP plateau (window current)(Arreola et al., 1991). Indeed,the relatively depolarized resting membrane potential in the AP experiments(approximately -50 mV) may have increased Ca²⁺ influx viathe window current under AP conditions. A -50 mV resting membrane potential is typical for isolated rainbow trout atrial myocytes(Shiels et al., 2000) and results from the loss of cholinergic tonus and also reflects the lower density of inward rectifier current (I_K1) in fish atrial cells compared with ventricular cells (Vornanen et al.,2002a). However, recent measurements in vivo suggest that the resting membrane potential in rainbow trout atrial cells is approximately-65 mV (Vornanen et al.,2002a). Therefore, future studies could consider adding tonic levels of acetylcholine to solutions when recording APs from isolated myocytes.

In previous studies (Shiels et al.,2000), we have measured the I_Ca window current in trout atrial cells at 7°C, 14°C and 21°C and found a peak Ca²⁺ contribution at -10 mV at all temperatures. However, we also found that the relative Ca²⁺ contribution from the window current was greater (0.08 relative units) and had a larger voltage window (-40 mV to+30 mV) at 7°C compared with at 14°C and 21°C (0.055 relative units, -30 mV to +20 mV). A larger window current at 7°C may explain, in part, the large SR Ca²⁺ content observed at 7°C with the SQ pulses at 0.6 Hz (Fig. 3B). Additionally, because kinetics are slower at 7°C (see Shiels et al., 2000),prolongation of the diastolic period at 0.6 Hz could allow for more complete recovery of I_Ca, thereby increasing Ca²⁺ influx and SR loading.

We report a high steady-state SR Ca²⁺ content during AP stimulation (approximately 1250 μmol l^-1). However, when trout atrial myocytes are depolarized to +50 mV for 32 s, the SR can accumulate even larger quantities of Ca²⁺ (approximately 2340 μmol l^-1; Hove-Madsen et al., 1998, 1999). In the adult mammalian heart, SR Ca²⁺ release predominates over SL Ca²⁺delivery during e-c coupling and yet the steady-state SR Ca²⁺content in mammalian myocytes (100-150 μmol l^-1; Bassani et al., 1995; Negretti et al., 1995; Díaz et al., 1997) is an order of magnitude less than in fish. An intriguing question for future study emerges: why does the SR of rainbow trout have such a large maximal and steady-state capacity to store Ca²⁺ when isometric muscle studies indicate that Ca²⁺ release from the SR contributes a rather small amount of the Ca²⁺ involved in contraction?

The Ca²⁺-storing capacity of the SR is determined by the volume fraction of the SR in the myocyte, the concentration of low-affinity Ca²⁺ buffers inside the SR and the concentration gradient of free Ca²⁺ across the SR membrane(Shannon and Bers, 1997). It is unclear which of these factors contributes to the enhanced Ca²⁺-storing capacity of the trout SR. It is possible that functional differences in the rainbow trout SR Ca²⁺ release channels, possibly at the luminal side of the channel, or differences in SR Ca²⁺ buffers may account for the difference in maximal SR Ca²⁺ content. However, to date there is no information available to accept or reject this idea. In mammals, it is difficult to increase steady-state SR Ca²⁺ above approximately 100 μmol l^-1without spontaneous release (Bassani et al., 1995), probably because of a direct effect of luminal Ca²⁺ on the SR Ca²⁺ release channel(Lipp et al., 1992; Sitsapesan and Williams, 1994; Bassani et al., 1995). One possibility is that the SR Ca²⁺ release channels in fish heart are approximately 10-fold less sensitive to luminal Ca²⁺ than those in mammals. Indeed, experiments with carp (Cyprinus carpio) suggest that their SR Ca²⁺ release channels may be an order of magnitude less sensitive to Ca²⁺ than mammals(Chugun et al., 1999). Furthermore, indirect studies of SR function do suggest striking differences between fish and mammalian SR Ca²⁺ release channels with respect to cold sensitivity. The results of the present study confirm earlier work(Bowler and Tirri, 1990; Hove-Madsen et al., 1998; Aho and Vornanen, 1999; Tiitu and Vornanen, 2002)indicating that the teleost SR Ca²⁺ release channels remain functional at cold temperatures whereas the mammalian SR Ca²⁺release channels open such that Ca²⁺ leaks out of the SR(Bers, 1987; Sitsapesan et al., 1991).

One speculation as to why rainbow trout store large amounts of Ca²⁺ in their SR may be related to their burst mode of swimming. Burst swimming has been shown to decrease blood pH by as much as 0.5 pH units,and an extracellular pH change of such magnitude will significantly inhibit cardiac performance (Milligan and Farrell,1986; Churcott et al.,1994). Although there is no information at present, a large SR Ca²⁺ store, if releasable, could serve as a safety factor during acidosis to override decreased myofilament Ca²⁺ sensitivity. Indeed, we know that adrenergic stimulation can protect force development in the acidotic rainbow trout myocardium(Farrell et al., 1985; Farrell and Milligan, 1986)although the mechanisms have not yet been defined. However, under normoxic conditions, a 30-60% release of the steady-state SR Ca²⁺ content in trout (as reported for mammals; Bassani et al., 1995) could cause severe contracture and could damage the contractile machinery. It may be that, regardless of SR Ca²⁺content, there is an upper limit to the amount of Ca²⁺ that is releasable via Ca²⁺-induced Ca²⁺ release, as the absolute amount of Ca²⁺ released in fish and mammalian cardiomyocytes at 21°C is comparable: 35-60 μmol l^-1Ca²⁺ (Bassani et al.,1995; Janczewski et al.,1995; Hove-Madsen et al.,1998).

In summary, the present study has demonstrated that the rainbow trout atrial SR Ca²⁺ content reaches a steady-state after approximately 25 stimulation pulses over a physiological range of temperatures and heart rates during AP stimulation. Furthermore, SR Ca²⁺ accumulation and release are not compromised by temperature change, indicating significant differences between rainbow trout and mammalian SR. Finally, because the temperature changes used in the present study represent a realistic physiological challenge for the rainbow trout heart, our results suggest that the plasticity of the rainbow trout contractile machinery (see Vornanen et al., 2002b for a recent review) may help to maintain SR Ca²⁺ content during temperature change in vivo.

We acknowledge Kontiolahti fish farm for donating the rainbow trout and Virpi Tiitu for her help in transporting them to the University of Joensuu. We thank Dr Leif Hove-Madsen for his critical reading of earlier versions of this paper. The study was supported by NSERC to A.P.F., Simon Fraser University Steele Memorial Fellowship to H.A.S., and the Academy of Finland to M.V.