Temperature-Dependence of L-Type Ca²⁺ Channel Current in Atrial Myocytes From Rainbow Trout

Many fish inhabit eurythermal environments and must therefore be able to cope with changes in environmental temperature. With seasonal temperature changes, fish can acclimatize, often adjusting their physiological and biochemical processes in a compensatory manner. However, temperature changes can be more rapid, either as a result of diurnal cycles (Matthews and Berg, 1997) or as the fish moves vertically through the water column, often in association with feeding or escaping predation (Reid et al., 1997). Acute temperature changes are of considerable interest for cardiac physiologists because the heart is required to maintain cardiac function while it reaches a new temperature equilibrium with the surrounding environment. The temperature tolerance of eurythermal fish, therefore, raises important questions concerning the temperature-dependency of cardiac ion channel function, particularly because acute temperature changes can be cardioplegic to mammals. We were specifically interested in the ion channels involved in Ca²⁺ transport during excitation–contraction coupling of the heart.

In most fish, including rainbow trout, transsarcolemmal Ca²⁺ influx is considered to be the primary source of Ca²⁺ for activation of the myofilaments during excitation–contraction coupling (Vornanen, 1989; Tibbits et al., 1990; Keen et al., 1994; Shiels and Farrell, 1997). The majority of this Ca²⁺ enters the myocyte through the L-type Ca²⁺ channel (Vornanen, 1997, 1998; Hove-Madsen and Tort, 1998), although the contribution of the Na⁺/Ca²⁺ exchanger can also be quite significant (contributing up to 50 % of the activator Ca²⁺; Vornanen, 1999). Mammalian studies indicate that the L-type Ca²⁺ channel is temperature-dependent, with Q₁₀ values ranging from 1.8 (Herve et al., 1992; guinea pig and ground squirrel) to 2.9 (Cavalié et al., 1985; guinea pig). If the temperature-dependency of the L-type Ca²⁺ channel is similar in fish, then there must be other mechanisms in place that allow fish to maintain cardiac viability during acute temperature change. To date, however, the temperature-dependence of the L-type Ca²⁺ channel has not been investigated in fish cardiac myocytes. Thus, the aim of this study was to characterize the temperature-dependence of the L-type Ca²⁺ channel in trout atrial myocytes during physiologically relevant acute temperature changes. The temperature-dependency of the L-type Ca²⁺ channel was studied at 14 °C, the thermal preference for rainbow trout (Coutant, 1977), and after a temperature change of ±7 °C, encompassing the range of temperatures trout experience in nature (Kaya, 1978; DeVera and Priede, 1991; Farrell et al., 1996). The results indicate that the temperature-sensitivity of I_Ca for trout atrial myocytes is within the range found in most mammals, exhibiting a Q₁₀ of approximately 2. However, when cells are stimulated with an action potential waveform appropriate for the test temperature, the total charge carried by I_Ca is temperature-independent, suggesting a possible compensatory mechanism for coping with acute temperature change.

Rainbow trout Oncorhynchus mykiss (Walbaum) (175.5±13.6 g, mean ± S.E.M., N=6) of both sexes were obtained from a local fish farm (Kontiolahti, Finland) and held in the laboratory in a 500 l tank receiving aerated tap water. The tank temperature was 14±1 °C. Fish were held for a minimum of 4 weeks prior to experimentation and were fed daily on commercial trout pellets (Ewos, Turku, Finland). The photoperiod was 12 h:12 h light:dark.

A detailed description of the cell preparation has been published previously (Vornanen, 1998, 1999). Briefly, fish were stunned with a blow to the head, the spine was then cut and the heart was carefully excised. All procedures were in accordance with local animal handling protocols. A cannula was inserted through the bulbus arteriosus into the lumen of the ventricle, and the heart was perfused from a height of 50 cm first with an isolating solution (nominally Ca²⁺-free, low-[Na⁺]) solution for 8–10 min, and then with a proteolytic enzyme solution for 20 min. After enzymatic treatment, the atrium was removed from the ventricle and placed in a small dish containing the isolating solution. The atrium was cut into small pieces with scissors and then triturated through the opening of a Pasteur pipette. The isolated cells were stored in this solution at 14 °C.

The solution used for cell isolation contained (mmol l⁻¹): NaCl, 100; KCl, 10; KH₂PO₄, 1.2; MgSO₄, 4; taurine, 50; glucose, 20; and Hepes, 10; adjusted to pH 6.9 with KOH. For the enzymatic digestion, collagenase (Type IA), trypsin (Type III) and fatty-acid-free bovine serum albumin (BSA) were added to this solution. The physiological saline used as the extracellular solution for recording I_Ca contained (mmol l⁻¹): NaCl, 150; CsCl, 5.4; MgSO₄, 1.2; NaH₂PO₄, 0.4; CaCl₂, 1.8; glucose, 10; Hepes, 10; adjusted to pH 7.6 with CsOH. In addition, 1 µmol l⁻¹ tetrodotoxin (TTX) was added to the perfusate to block fast Na⁺ channels and 10 µmol l⁻¹ ryanodine was added to inhibit the sarcoplasmic reticulum Ca²⁺ release channel. These concentrations of TTX (Vornanen, 1998) and ryanodine (Hove-Madsen et al., 1999) have been shown to inhibit completely I_Na and sarcoplasmic reticulum function in rainbow trout myocytes, respectively. To record action potentials from the cells, the solution was modified by replacing the CsCl with equimolar KCl, pH balanced with KOH, and by omitting the TTX and ryanodine. The pipette solution for I_Ca recordings contained (mmol l⁻¹): CsCl, 130; MgATP, 5; tetraethylammonium chloride (TEA⁺), 15; MgCl₂, 1; oxaloacetate, 5; Hepes, 10; EGTA, 5; disodium GTP, 0.03, pH adjusted to 7.2 with CsOH. For action potential recordings, the pipette solution contained (mmol l⁻¹): potassium aspartate, 125; KCl, 15; MgCl₂, 1; MgATP, 5; EGTA, 0.05; disodium phosphocreatine, 5; Hepes, 10; and was adjusted to pH 7.2 with KOH. All drugs were purchased from Sigma (MO, USA) with the exception of ryanodine (Calbiochem, CA, USA) and TTX (Tocris, Bristol, UK).

The whole-cell patch-clamp technique was used to study the effects of temperature on the electrophysiological properties of the L-type Ca²⁺ channel in rainbow trout atrial myocytes. In the first series of experiments, square pulses were used to determine the effects of acute temperature change on the current–voltage relationship and the kinetics of activation and inactivation of I_Ca.

A sample of atrial cells (cell capacitance 38.9±1.88 pF, mean ± S.E.M., N=28) was added to the recording chamber (volume 0.5 ml) and allowed to settle on the bottom. The cells were then superfused at a rate of 2 ml min⁻¹with the physiological extracellular solution. 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 were pulled from borosilicate glass (Vitrex microhaematocrit tubing, Modulohm A/S, Denmark) using a vertical patch-pipette puller (List-Medical, Darmstadt, Germany) and had a resistance of 1.5–2.5 MΩ when filled with pipette solution. Junction potentials were zeroed prior to seal formation. Pipette capacitance (6–8 pF) was compensated after formation of a gigaohm seal. The membrane under the pipette was ruptured with a brief voltage pulse (zap), and capacitative transients were eliminated by iteratively adjusting the series resistance and cell capacitance compensation circuits. Mean series resistance was 7.2±0.7 MΩ. Currents were leakage-corrected using the P/N procedure of the software (Clampex, Axon Instruments) and were low-pass-filtered using the four-pole low-pass Bessel filter on the Axopatch-1D amplifier at a frequency of 2 kHz for I_Ca and 10 kHz for action potentials, and then analyzed off-line on a computer using pClamp 6.0 software (Axon Instruments). Membrane capacitance was measured using the calibrated capacity compensation circuit of the Axopatch amplifier.

Each cell was tested at three temperatures (7 °C, 14 °C and 21 °C). Experimental temperature was maintained (±0.5 °C) by adjusting the temperature of the perfusion solution. Three water baths were set to the three experimental temperatures. The inflow tube carrying the perfusion solution was heated or chilled by the water bath circuit before emptying into the recording chamber. A thermocouple was placed inside the recording chamber and positioned no less than 5 mm from the cell to determine when the desired temperature had been achieved. Each experiment began at 14 °C, the acclimation temperature of the fish. The current–voltage relationship and the steady-state activation and inactivation curves for I_Ca were obtained using a double-pulse protocol. Every 6 s, a 1000 ms prepulse depolarized the membrane from a holding potential of −70 mV to +70 mV in 10 mV steps. Each voltage step was followed by a test pulse to +10 mV for 500 ms. The prepulse and test pulse were separated by a 5 ms return to the holding potential. After the protocol had been completed at 14 °C, the perfusate temperature was changed to 21 °C. It took approximately 1 min for the chamber to stabilize at the new temperature. At this point, the double-pulse protocol was repeated. To assess for possible run-down and any side effects of the temperature change, the temperature was returned to 14 °C, and the protocol was repeated. The temperature was then changed to 7 °C, and the protocol was repeated. A fifth and final protocol was performed on the same cell after the temperature had been returned to 14 °C. The duration of each experiment was approximately 11 min. The successive 14 °C trials are referred to as 14 °C’ for the second trial and 14 °C’’ for the third trial at this temperature. The order of the temperature change was reversed in some experiments (i.e. 14 °C to 7 °C to 14 °C’ to 21 °C to 14 °C’’), but this did not influence the effect of temperature on I_Ca.

The effects of temperature on the kinetics of inactivation of I_Ca were assessed by fitting the decay of I_Ca with a second-order exponential function using the Chebyshev transformation procedure from the Clampfit software (Axon Instruments). From this, the time constants for the fast (τ_f) and slow (τ_s) components of inactivation were derived.

The effects of temperature on steady-state activation and inactivation were studied by fitting the curves with Boltzmann functions. The methods for calculating the voltage-dependence of steady-state activation and inactivation have been reported elsewhere (Vornanen, 1998). Briefly, steady-state inactivation was calculated by dividing the amplitude of the test current by the maximal current elicited. The voltage-dependence of activation was determined as the normalized Ca²⁺ conductance (g_Ca/g_max), where g_max is the maximum value for I_Ca conductance. The voltage-dependence of the peak conductance of Ca²⁺ channels was calculated from the equation: g_Ca=I_Ca/(V−V_rev), where g_Ca is the membrane Ca²⁺ conductance, I_Ca is the peak Ca²⁺ current at the given potential (V) and V_rev is the apparent reversal potential obtained from the current–voltage relationship, or by extrapolating the ascending portion of the current–voltage relationship to zero current. Steady-state kinetic parameters were obtained by fitting the data to Boltzmann equations to determine the half-activating or half-inactivating potential (V_h) and the slope of activation and inactivation (k) (for details, see Vornanen, 1998).

The contribution of I_Ca to total cellular [Ca²⁺] was calculated from the transferred charges and cell volume. Charge transfer through L-type Ca²⁺ channels was determined by integrating the inactivating portion of the Ca²⁺ current obtained from the prepulse depolarization to +10 mV. Cell volume was calculated from the measured cell capacitance and the surface-to-volume ratio of the cell (see Vornanen, 1998). The myocytes were considered to be flat elliptical cylinders with an axis ratio of 1:2 for the elliptical cross section (Vornanen, 1997, 1998). The change in total cellular [Ca²⁺] due to Ca²⁺ influx through L-type Ca²⁺ channels was expressed as a function of myofibrillar volume (40 %; Vornanen, 1998).

To provide a more physiologically relevant indication of how temperature affects I_Ca, a second series of experiments was conducted. Action potentials from rainbow trout atrial cells were elicited by a short (2 ms) 1 V pulse at 7, 14 or 21 °C. Representative action potential recordings for each temperature were then used as stimulus waveforms. The shape of the action potential varied significantly with temperature, so each myocyte was tested with the three action potential waveforms and with a square pulse waveform as in the first series of experiments. The protocol was as follows. Starting at 14 °C, the cell was voltage-clamped first with the 14 °C action potential waveform, second with the 21 °C action potential waveform, third with the 7 °C action potential waveform, and finally with a 500 ms square pulse from a holding potential of −70 to +10 mV. Each of these waveforms was applied three times before the next waveform was applied. The average of the three recordings for each individual waveform was used to give N=1 for each waveform for that cell. The bath temperature was then changed to 21 °C, and the four stimulation waveforms were repeated. As with the first series of experiments, the stimulation protocol was then repeated at 14 °C’, at 7 °C and finally at 14 °C’’.

In general, comparisons among temperatures were made using one-way repeated-measures analysis of variance (ANOVA) with SigmaStat 1.0 software (Jandel Scientific). In some cases, the data did not fit a normal distribution, and a Friedman repeated-measures ANOVA on ranks was then conducted. Differences were considered significant when P<0.05 and were assessed using a Student–Newman–Keuls post-hoc analysis. Details of statistical procedures for each data set are found in the corresponding figure legend. The three tests at 14 °C (14 °C, 14 °C’ and 14 °C’’) were tested as individual groups to ensure that there were no statistical differences between them before the values were pooled to make an average 14 °C data point. For this reason, the N value for the 14 °C data is three times that for the 7 °C and 21 °C groups.

The average peak I_Ca at 21 °C ranged from −5.43±0.78 pA pF⁻¹ in the first series of experiments (see Table 1) to −1.77±0.21 pA pF⁻¹ in the second series of experiments (see Table 3). This range is within that reported previously for peak current amplitude in rainbow trout atrial (Hove-Madsen and Tort, 1998) and ventricular (Vornanen, 1998) myocytes at room temperature (19–21 °C) using slightly different experimental conditions. Similarly, the charge carried by I_Ca at 21 °C in our experiments (−0.129±0.02 to −0.075±0.03 pC pF⁻¹) was within the range found previously for trout myocytes (Hove-Madsen and Tort, 1998; Vornanen, 1998).

The effects of temperature on the current–voltage relationship for I_Ca in trout atrial myocytes are shown in Fig. 1. I_Ca amplitude was positively correlated with temperature (Fig. 1; Table 1). When temperature was increased from 14 to 21 °C, peak current density increased from −3.45±0.34 to −5.43±0.78 pA pF⁻¹, which corresponds to a Q₁₀ of 1.9.

Similarly, peak current density decreased from −3.45±0.34 to −2.05±0.43 pA pF⁻¹ on cooling from 14 to 7 °C, yielding a Q₁₀ of 2.1. The shape of the current–voltage relationship was unchanged by temperature. There was no significant rundown of I_Ca over the course of the experiment, as shown by the inset of the three 14 °C trials in Fig. 1.

In contrast to peak current amplitude, the charge carried by I_Ca (at peak current; +10 mV) was negatively correlated with temperature (Table 1). As temperature was decreased from 21 °C through 14 °C to 7 °C, charge density increased from −0.129±0.02 to −0.194±0.017 and −0.213±0.04 pC pF⁻¹.

Similarly, the change in total cellular [Ca²⁺] increased from 30.5±4.8 µmol l⁻¹ at 21 °C to 46.1±3.9 µmol l⁻¹ at 14 °C and to 50.6±9.7 µmol l⁻¹ at 7 °C. Thus, although the peak current amplitude decreased with temperature, the amount of Ca²⁺ entering the atrial myocyte increased. These results are explained by a temperature-induced change in the kinetics of I_Ca.

Fig. 2 shows the effect of temperature on the kinetics of I_Ca. The time constants for the fast (τ_f) and slow (τ_s) components of inactivation were obtained by fitting a double exponential function to the decay of I_Ca. Over a range of membrane potentials (−10, 0, +10 mV), both τ_f and τ_s increased as temperature decreased (Fig. 2). These longer time constants explain why the charge carried by I_Ca increased while the amplitude decreased. At peak current (+10 mV), τ_f increased from 22.9±2.1 ms at 21 °C to 39.6±3.1 ms at 14 °C and further to 60.9±7.3 at 7 °C, a 2.7-fold increase in the time required for fast inactivation. The effect of temperature on the time constant for the slow inactivation was a 3.4-fold increase. At peak current, τ_s increased from 161.9±18.6 at 21 °C and 213.9±17.2 at 14 °C to 543.9±98.3 at 7 °C. The effect of temperature on τ_f and τ_s was greater between 14 and 7 °C than between 14 and 21 °C.

The temperature-induced changes in the kinetics of I_Ca revealed by the fast and slow components of inactivation are reinforced by the Boltzmann fits for steady-state activation and inactivation (Fig. 3; Table 2). It was possible to superimpose the steady-state activation curves at 21 and 14 °C (Fig. 3). However, at 7 °C, the activation curve was significantly shifted such that half-maximal activation (V_h,act) occurred at more depolarized voltages (see Fig. 3; Table 2), probably indicating slower channel activation at the low temperature. This is supported by the more gradual slope of the Boltzmann fit for activation (k_act) at 7 °C. There was no difference in V_h for inactivation between temperatures; however, the slope of inactivation was also greater at 7 °C (Table 2), suggesting that the L-type Ca²⁺channel makes the transition from the activated to the inactivated state more gradually at the colder temperature. In addition, there was a clear difference in the steady-state inactivation curves at positive membrane potentials. Between +30 and +70 mV, the attenuation of inactivation of I_Ca was significantly greater at 7 °C than at 21 °C (Fig. 3), which can probably be explained by a reduction in Ca²⁺-dependent inactivation at cold temperatures, where the density of I_Ca is significantly reduced. Furthermore, because the activation and inactivation curves overlap, there is a small non-inactivating component of I_Ca, known as a window current, which peaks at approximately −10 mV (see Fig. 3 inset). The size of the window current is greatest at 7 °C, which again is probably due to the lower current density found at cold temperatures in trout atrial myocytes.

The shape of the action potential recorded from trout atrial myocytes was significantly altered by temperature (see Discussion for an explanation). Representative action potential recordings at each temperature from one cell were selected and converted into a series of stimulus waveforms for subsequent activation of the L-type Ca²⁺ channel (Fig. 4, inset). The action potential plateau duration at 50 % repolarization from peak increased from 70 ms at 21 °C to 134 ms at 14 °C and to 288 ms at 7 °C. The size of the hyperpolarizing phase decreased (from resting potential) with decreasing temperature from −24 mV at 21 °C, to −19 mV at 14 °C and to −11 mV at 7 °C. In addition, the resting potential became progressively depolarized as temperatures were increased from −55 mV at 7 °C, to −41 mV at 14 °C and to −38 mV at 21 °C. Given such profound changes in the action potential, it was necessary to examine I_Ca with action potential waveforms for different temperatures and test whether these temperature-induced modifications of the action potential influenced I_Ca.

This series of experiments demonstrated that I_Ca is sensitive to the shape of the stimulus waveform. At 7 and 14 °C, but not at 21 °C, peak current density elicited from the square-pulse stimulus waveform was greater than that elicited by the action potential waveform (Table 3; compare values vertically). This suggests that the typical square pulse (usually to 0 or +10 mV) used to study I_Ca may overestimate current amplitude because the driving force for Ca²⁺ entry is greater at these voltages than with the more depolarized action potential overshoot (+20 to +40 mV). Even so, and in accordance with the first series of experiments, the peak amplitude of I_Ca elicited from the action potential waveforms was significantly temperature-dependent, with a Q₁₀ of approximately 2 (Fig. 4; Table 3). Furthermore, there was a trend towards increasing charge transfer with decreasing temperature using the square-pulse waveforms (compare charge densities at the three test temperatures with the square pulse in Table 3) but, unlike in the first series of experiments, statistical significance could not be established. In contrast, a physiologically relevant action potential clamp did not result in an increase in charge transfer with decreasing temperature (Table 3).

Table 3 summarizes the effects of the three action potential waveforms at the three test temperature on the peak density of I_Ca and the charge carried by I_Ca. While there were significant differences among these values, the most physiologically important comparison is for the three action potential waveforms corresponding to the comparable test temperature. The three I_Ca curves are compared in Fig. 4. While peak current density was temperature-dependent, the charges carried by these currents were not significantly different (Fig. 4; Table 3). As a result, provided that the cell is stimulated with an action potential waveform appropriate for the test temperature, the cell maintains a relatively constant Ca²⁺ influx via the L-type Ca²⁺ channel over the temperature range 7–21 °C. This finding has important physiological implications when trying to understand how trout maintain cardiac function during acute temperature change.

The hearts of fish continue to function at temperatures that are cold enough to be cardioplegic to most mammals. Possible explanations for this temperature tolerance include (i) changes in myocyte metabolic rate, although this form of adaptation usually occurs with seasonal acclimatization; (ii) changes in protein–protein interactions during excitation–contraction coupling; (iii) changes in Ca²⁺ uptake and release from the sarcoplasmic reticulum and (iv) changes in sarcolemmal ion channel function. Because contractility is dependent on both the rate and magnitude of Ca²⁺ delivery to the myofilaments, and because Ca²⁺ influx through the L-type Ca²⁺ channel is vital both for activating the myofilaments directly and for triggering the possible release of Ca²⁺ from the sarcoplasmic reticulum, we studied the effects of temperature on the electrophysiological properties of the L-type Ca²⁺ channel in rainbow trout atrial myocytes.

Given their ability to maintain cardiac function at low temperatures and during acute temperature changes above and below their acclimatization temperature, we had anticipated that the temperature-sensitivity of the trout L-type Ca²⁺ channel would be less than that reported for mammals. Indeed, the Q₁₀ of approximately 2 is similar to the lowest of Q₁₀ values reported for mammals (1.8, Herve et al., 1992; guinea pig and ground squirrel ventricle; 2.9, Cavalié et al., 1985; guinea pig ventricle; 2.3, Kiyosue et al., 1993, guinea pig ventricle; 2.7, Shimoni and Banno, 1993, rabbit ventricle; 5.8, Allen and Mikala, 1998, human L-type Ca²⁺ channel expressed in Xenopus laevis oocytes conducting Ba²⁺; 1.9, Rosen, 1996, GH₃ cells).

Although the temperature-dependence of peak current through the L-type Ca²⁺ channel is not as great as that found in most mammals, this alone can hardly account for the temperature tolerance observed in fish in nature. Thus, other mechanisms regulating contractility must help to confer the relative temperature tolerance observed in rainbow trout. Here, we have established that the shape of the action potential is critical in determining Ca²⁺ delivery across the sarcolemma. In fact, provided that we used an appropriate action potential waveform for a given test temperature, Ca²⁺ delivery was apparently independent of temperature. Therefore, this observation helps explain the normal temperature tolerance observed in rainbow trout hearts.

In addition to the sarcolemma, the other primary organelle involved in handling Ca²⁺ during excitation–contraction coupling is the sarcoplasmic reticulum. On the basis of studies with multicellular preparations, it has been proposed that the Ca²⁺ contribution made by the sarcoplasmic reticulum is temperature-dependent in rainbow trout, contributing more significantly at warmer temperatures (Hove-Madsen, 1992; Møller-Nielsen and Gesser, 1992; Shiels and Farrell, 1997). Furthermore, using cardiac homogenates or suspensions of ventricular myocytes from rainbow trout, it has been demonstrated that Ca²⁺ uptake by the sarcoplasmic reticulum is also temperature-sensitive, with faster uptake rates at higher temperatures (Aho and Vornanen, 1998; Hove-Madsen and Tort, 1998). It is expected, therefore, that the Ca²⁺ contribution of the sarcoplasmic reticulum may be significantly altered by acute temperature change. Indeed, if sarcoplasmic reticulum Ca²⁺ release changes with temperature, it may also significantly affect the Ca²⁺-dependent inactivation of the L-type Ca²⁺ channel (Hove-Madsen et al., 1999; Harwood et al., 2000). However, to prevent this confounding effect of possible feedback from the sarcoplasmic reticulum, we included 10 µmol l⁻¹ ryanodine in all experiments when examining the properties of the trout L-type Ca²⁺ channel. Thus, for the purpose of this paper, we must examine other possible mechanisms that may affect Ca²⁺ delivery to the myofilaments under conditions of changing ambient temperature.

The results of this study suggest that the kinetics of I_Ca plays an important role in maintaining adequate Ca²⁺ influx during temperature change in rainbow trout. The slow component of inactivation (time constant τ_s) was especially sensitive, demonstrating a 3.4-fold increase in the duration of inactivation with a 14 °C decrease in temperature. This powerful temperature-dependence of the inactivation time constants is a novel finding. Our slow time constants for an acute temperature change to 21 °C (161.9±18.9 ms) are longer than those reported in a previous study performed only at room temperature using ventricular myocytes from rainbow trout (101 ms for fish acclimated to 18 °C and 73 ms for fish acclimated to 4 °C; Vornanen, 1998). However, our values for fast inactivation (22.9±2.1 ms) are within the range reported for trout ventricular myocytes (27 ms for fish acclimated to 18 °C and 18 ms for fish acclimated to 4 °C; Vornanen, 1998) and for atrial myocytes (approximately 13 ms; Hove-Madsen et al., 1999) at room temperature.

Recent studies with caffeine blockade of sarcoplasmic reticulum Ca²⁺ release in rainbow trout (Hove-Madsen et al., 1999, atrial cells at room temperature; Harwood et al., 2000, ventricular cells at 15 °C) indicate that τ_f may be related to the sarcoplasmic reticulum Ca²⁺-dependent component of inactivation of the L-type Ca²⁺ current in rainbow trout. We observed slower inactivation of I_Ca at low temperatures despite the the absence of sarcoplasmic reticulum Ca²⁺-dependent inactivation (due to ryanodine). This observation could be partly related to Ca²⁺-dependent inactivation from Ca²⁺ entering through the L-type Ca²⁺ channel itself, a mechanism already established for mammalian myocytes (McDonald et al., 1994). This idea is supported by the observation that the recovery from steady-state inactivation was greatly facilitated between +30 and +70 mV at cold temperature (see Fig. 3) when the density of I_Ca was small and inactivation of I_Ca was slow. Because the density of I_Ca at 7 °C is greatly reduced, Ca²⁺-dependent inactivation can also be expected to be reduced.

Another mechanism might involve greater intracellular Ca²⁺ buffering at cold temperatures, which could further decrease the degree of Ca²⁺-dependent inactivation of I_Ca. The myofilaments are one of the largest Ca²⁺ sinks, and the Ca²⁺ affinity of trout cardiac troponin C changes significantly with a change in temperature from 21 to 7 °C (Gillis et al., 2000). Furthermore, we observed a more prominent window current at cold temperatures (see inset in Fig. 3), a result often described under conditions where Ca²⁺-dependent inactivation is reduced (i.e. rabbit versus rat, neonate versus adult, Bers, 1991). Indeed, the potential physiological significance of a larger window current at colder temperatures can be appreciated when one considers the role of the window current during an action potential. At the onset of the action potential, depolarization initiates opening of L-type Ca²⁺ channels. The subsequent repolarization will facilitate channel inactivation. However, during the action potential plateau (which usually corresponds to voltages in the window current range; McDonald et al., 1994), some of the inactivated channels recover and re-open (Hirano et al., 1992), providing the non-inactivating current.

The recovery from inactivation is kinetically slow, so lengthening the time that the action plateau remains in the window current voltage range increases the probability of Ca²⁺ channels reopening within the window (Hirano et al., 1992). Indeed, when the duration of the action potential ‘plateau’ was held constant (i.e. with the square pulse protocols in the first series of experiments), we observed both an increase in recovery from inactivation at positive potentials and an increase in the window current at 7 °C compared with 14 and 21 °C. This may have a profound effect on transsarcolemmal Ca²⁺ influx at cold temperatures (see Perspectives) and may help to account for the larger charge transfer observed at low temperature in our study. Furthermore, when we recorded action potentials from myocytes at different temperatures in the second series of experiments, we observed a significant increase in the action potential plateau as temperature decreased. Thus, it is possible that, in addition to temperature-induced changes in K⁺ currents (see below), there are temperature-induced changes in the I_Ca window current which, along with the slower decay of I_Ca, may also contribute to the lengthening of the action potential plateau at lower temperatures.

Recent studies on the temperature-dependency of the two primary K⁺ currents involved in shaping the action potential (I_Kr) and maintaining the resting potential (I_K1) shed more light on the temperature-induced changes in the action potential we observed in trout atrial myocytes. First, the density of the inwardly rectifying K⁺ current (I_K1) was very low in trout atrial myocytes, resulting in a depolarized and relatively unstable resting membrane potential (A. Nurmi and M. Vornanen, personal communication). Indeed, the resting membrane potential in our study was significantly more depolarized than that found in most mammalian cells and in trout ventricular cells (Harwood et al., 2000). In contrast to the inward rectifier, the density of the rapid component of the delayed rectifier (I_Kr) was fairly high in atrial cells (A. Nurmi and M. Vornanen, personal communication) and contributed not only to the plateau and hyperpolarizing phases of the action potential but also to the maintenance of the diastolic resting membrane potential because specific blockade of this channel resulted in almost complete loss of resting membrane potential. In addition, A. Nurmi and M. Vornanen (personal communication) observed a higher density and faster kinetics of I_Kr at warmer temperatures, which may explain the faster repolarization and transient hyperpolarization of the trout atrial action potentials recorded at 21 °C in our study. In short, when the temperature-induced changes in the K⁺ currents are combined with our results concerning the temperature sensitivity of the L-type Ca²⁺ channel, the changes in the shape of the action potential appear to culminate in the maintenance of electrical excitability in trout cardiac muscle over a range of temperatures. Indeed, when the cells were stimulated with an appropriate action potential waveform for the applied test temperature, the charge carried by I_Ca did not differ significantly among temperatures. Thus, factors that change the shape of the action potential appear to have a large impact on I_Ca because they may help to compensate for the temperature-dependence of peak I_Ca.

Previous studies (Hove-Madsen, 1992; Keen et al., 1994; Shiels and Farrell, 1997) have suggested a degree of interplay between Ca²⁺ flux across the sarcolemma and Ca²⁺ flux through the sarcoplasmic reticulum at different temperatures in rainbow trout. The results of the present study may help to explain some of the potential mechanisms underlying this interplay between the sarcoplasmic reticulum and sarcolemma. Studies with isolated ventricular muscle preparations from rainbow trout have revealed a greater ryanodine sensitivity at warm temperatures. The effect of ryanodine is often abolished at temperatures below 15 °C (El-Sayed and Gesser, 1989; Hove-Madsen, 1992; Møller-Nielsen and Gesser, 1992; Keen et al., 1994; Gesser, 1996; Shiels and Farrell, 1997). Studies with isolated atrial tissue have suggested that the ryanodine sensitivity may be less warm-temperature-dependent than in ventricular tissue (trout, Gesser, 1996; yellowfin tuna, Shiels et al., 1999). If the initial magnitude of I_Ca is the trigger for sarcoplasmic reticulum Ca²⁺ release, as suggested by Fabiato (1983, 1985), then at warm temperatures where the density of I_Ca in trout cells is high, it is quite feasible that, under physiological conditions, a sizeable Ca²⁺-induced Ca²⁺ release (CICR) could occur. Indeed, a recent study at room temperature with trout atrial myocytes has provided evidence for CICR and has even shown that it may contribute to Ca²⁺-dependent inactivation of I_Ca (Hove-Madsen et al., 1999). The pumping capabilities of the sarcoplasmic reticulum Ca²⁺-ATPase are also significantly greater at warm temperatures (Aho and Vornanen, 1998; Hove-Madsen et al., 1998), so sarcoplasmic reticulum Ca²⁺-loading could be faster. Taken together, the faster and larger I_Ca and the increased sarcoplasmic reticulum Ca²⁺-ATPase activity may enable the sarcoplasmic reticulum to contribute Ca²⁺ on a beat-to-beat basis at warm temperatures.

At cold temperatures, when the initial magnitude of I_Ca is small, it may be impossible to trigger sarcoplasmic reticulum Ca²⁺ release under physiological conditions. However, because of (i) the slower decay of I_Ca and (ii) the increased recovery from inactivation during window current voltages at low temperatures, there may be sufficient Ca²⁺ influx across the sarcolemma to meet the demands of the myofilaments. Correspondingly, we report an increase in the total Ca²⁺ influx through the L-type Ca²⁺ channel at low temperatures. This increased Ca²⁺ influx at colder temperatures could result in increased sarcoplasmic reticulum loading. Moreover, if the slow component of inactivation (τ_s) is the sarcoplasmic reticulum loading component of I_Ca, as suggested by Fabiato (1983, 1985), then the drastically slower τ_s observed at 7 °C in the present study would greatly facilitate sarcoplasmic reticulum loading. This contention is supported by the recent discovery that, unlike the sarcoplasmic reticulum of mammals, the sarcoplasmic reticulum of rainbow trout does not leak Ca²⁺ at low temperature (Hove-Madsen et al., 1998). Taken together, these observations may help to explain why rainbow trout appear to have a large sarcoplasmic reticulum Ca²⁺ load at low temperatures, but generally do not seem to release it under physiological conditions (for an exception in trout atrium acclimated to 4 °C, see Aho and Vornanen, 1999). Clearly, a complete study of the effects of temperature on sarcoplasmic reticulum loading and release in trout myocardium is needed to elucidate these relationships properly.

In summary, I_Ca plays many roles in maintaining contractility in rainbow trout myocardium. First, it provides a transsarcolemmal Ca²⁺ influx that can directly activate the myofilaments. Second, it can function as the trigger for the release of Ca²⁺ from the intracellular stores of the sarcoplasmic reticulum, and lastly it can serve to increase sarcoplasmic reticulum Ca²⁺ loading. All these potential roles are modulated by temperature. The results of this study indicate that the temperature-sensitivity (Q₁₀) of I_Ca for trout atrial myocytes is within the lower range of those found for most mammals. This could compromise Ca²⁺ cycling at the myofilaments and affect cardiac performance during an acute temperature change. However, we also discovered that, when the cells were stimulated with an appropriate action potential waveform for the applied test temperature, the charge carried by I_Ca did not differ significantly among temperatures. This suggests that the temperature-induced modifications in the action potential, and the resultant changes in the density and kinetics of I_Ca, may help to ensure adequate Ca²⁺ cycling for the maintenance of cardiac performance during acute temperature change in rainbow trout.

This work was supported by a Journal of Experimental Biology travel fellowship and a Simon Fraser University Graduate Fellowship to H.A.S., an NSERC operating grant to A.P.F. and the Academy of Finland (project no. 63090) to M.V.