Contractile properties of atrial and ventricular myocardium of the heart of rainbow trout Oncorhynchus mykiss: effects of thermal acclimation

In mammals, atrial systole occurs immediately before ventricular systole and thus allows atrial contraction to augment the volume of blood in the ventricles at end-diastole when ventricular volume determines the strength of the ventricular contraction. The important timing of atrial systole is made possible by the delay of impulse conduction in the atrioventricular node (Katz, 1992) and by rapid contraction kinetics of the atrial muscle. Although atrial contraction exerts significant regulatory function in ventricular contractility through the Frank–Starling mechanism, its role in ventricular filling is relatively small; most of the blood flow into the ventricles of mammalian heart occurs directly from the veins, and atrial contraction provides only approximately 30 % of the end-diastolic blood volume (Guyton, 1981). The relative significance of atrial contraction for ventricular filling and the performance of the cardiac pump is generally assumed to be much larger in the fish heart than in the mammalian heart. In fact, atrial systole is considered to be the sole determinant of ventricular filling in fish heart, whereby the preceding atrial contraction would be a necessary prerequisite for the subsequent ventricular contraction (see Johansen and Burggren, 1980; Farrell and Jones, 1992). According to this concept, atrial contractility would determine the extent of ventricle filling, the strength of ventricular contraction and hence the cardiac stroke volume in fish hearts. More recent findings suggest, however, that the contribution of the atrial systole to ventricular filling is relatively small in fish heart and that the ventricular filling mechanism is basically similar in all vertebrates (Lai et al., 1998).

In mammals, the atrial contraction is much faster than the ventricular contraction (Urthaler et al., 1975; Asgrimsson et al., 1995; Bottinelli et al., 1995). The cellular and molecular basis for this difference is complicated, involving action potential duration, kinetics of intracellular Ca²⁺ transients and properties of myofibrillar proteins (for references, see Minajeva et al., 1997). Although the importance of atrial contraction for the ventricular function is assumed to be larger in fish hearts than in other vertebrates, the contractile properties of the fish atrium are poorly known. The objective of the present study was to compare the basic characteristics of atrial and ventricular contraction in the heart of teleost fish. First, we wanted to determine whether the difference in contraction kinetics between the atrium and ventricle that is typical of the mammalian heart is similar in fish heart. Second, we tried to clarify the importance of myofibrils and sarcoplasmic reticulum (SR) in possible differences between the kinetics of atrial and ventricular contraction. Third, the effects of thermal acclimation on the contractile properties of atrial and ventricular muscle were compared. Experiments were conducted on a fish species (rainbow trout, Oncorhynchus mykiss) that remains active at low temperatures and is likely to show compensatory changes in cardiac function after acclimation to cold (Driedzic et al., 1996; Aho and Vornanen, 1998).

Rainbow trout Oncorhynchus mykiss (Walbaum) (N=61) were obtained from a local fish farm, randomly divided into two groups and acclimated either to 4 °C (CA) or to 17 °C (WA) for more than 4 weeks. The fish were of three size classes: the smallest animals (WA fish, 53.4±3.9 g, N=16; CA fish, 39.3±2.0 g, N=17; values are means ± S.E.M.) were used in contractile measurements, the mid-sized fish (WA fish, 247.2±19.0 g, N=8; CA fish 216.9±33.6 g, N=8) for Ca²⁺ uptake measurements and the biggest fish (WA fish, 566.4±62.2 g, N=6; CA fish, 599.7±65.7 g, N=6) in myofibrillar Ca²⁺/Mg²⁺-ATPase determinations. These size groups were selected to fulfil special experimental requirements and were necessary because of the large difference in size between atrial and ventricular myocardium. Small fish were used in contractile measurements to achieve dimensionally appropriate preparations for force recording, whereas bigger fish were necessary for Ca²⁺ uptake and myofibrillar ATPase measurements in order to obtain a sufficient amount of muscle tissue without pooling from a large number of animals. During acclimation, the fish were held in 500 l stainless steel tanks with a continuous circulation of aerated tap water. The fish were fed five times a week with a commercial fish food (Ewos, Turku, Finland). The photoperiod was a constant 12 h:12 h L:D cycle. Fish were stunned by a sharp blow to the head and killed by cutting the spine.

Isometric contractile properties of excised atrial and ventricular tissues were determined at the acclimation temperatures of the animals as well as at the intermediate temperature of 11 °C. Whole atria (6.88±0.6 mg, N=23) or ventricular strips (32.5±1.03 mg, N=23, approximately 1 mm thick) were mounted on the bottom of a chamber by means of a stimulating platinum electrode while the other end of the muscle was connected to the lever arm of a force transducer (Grass FT03) by a short braided-silk suture (size USP 6–0; Davis and Geck, Great Britain). The muscles were stretched stepwise to the length where the developed force was maximal when stimulated and were then allowed to stabilize at this length for about 1 h before the experiments were started. The voltage of the 8 ms long square-wave pulse was twice the threshold value (1–5 V). The solution in the chamber was oxygenated (100 % O₂) continuously and consisted of (in mmol l⁻¹): NaCl, 140; KCl, 2.8; CaCl₂, 1.8; MgSO₄, 1.2; NaH₂PO₄, 1.2; Hepes, 10; glucose, 10; pH 7.7, at 22 °C. The temperature of the saline was maintained at 4, 11 or 17 °C with the aid of a recirculating water bath. Preparations were stimulated to contract at physiological heart rates (Graham and Farrell, 1990; Farrell et al., 1988, 1996) of 0.6 Hz (36 beats min⁻¹), 0.8 (48 beats min⁻¹) and 1.2 Hz (72 beats min⁻¹) at 4, 11 and 17 °C, respectively. The time course of isometric twitches was monitored and recorded with a computer-based system. Analog signals from the chart recorder (Grass 7D) were digitised by an AD-converter (Digidata 1200, Axon instruments) and the recordings were stored on the hard disk of the computer by data acquisition software (PClamp 6.03 or AxoScope 1.1, Axon Instruments). The maximal developed force F_max, time-to-peak force T_PF (the time from the stimulus pulse to the peak force), time to half-relaxation T_0.5R (the time from F_max to 0.5 F_max) and contraction duration T_DC (T_PF+T_0.5R) were determined off-line using the Clampfit-program of the PClamp 6.0.3 software package. The first derivative of the developed force (dF/dt_max) was sometimes calculated and is shown in some of the figures. We also studied the effects of verapamil, a sarcolemmal Ca²⁺ channel blocker, and ryanodine, an SR Ca²⁺-release channel blocker, on isometric contractile characteristics. In each case the agent was added to the bathing solution to give a final concentration of 10 μmol l⁻¹ after a 1 h equilibration period. During this initial period, there was a slight decline in F_max (see Fig. 2A). In order to allow a ‘control’ F_max to be predicted after addition of the drug, a double exponential equation, y=ae^−bt+ce^−dt was fitted to the force during the equilibration period, where t is time, a and c are the magnitudes of the two exponential components, and b and d the reciprocals of their time constants (Vornanen, 1996). By extrapolation, this function was used to predict the force at selected points during later phases of the experiment. The drug effect was obtained as the difference between the measured and the predicted value (Fig. 2A). By this means each muscle could be used as its own control. The effect of ryanodine was recorded 1 h after its addition. Due to its use-dependent properties, the effect of verapamil was recorded after a constant number of contractions, i.e. after 30 and 60 min of its application in WA and CA fish, respectively.

Mechanical restitution and rest-potentiation were determined to characterize the force–frequency relationship of trout atrial and ventricular muscle. Muscle preparations were stimulated at physiological heart rates at the acclimation temperatures of the animals or at the mean experimental temperature of 11 °C (0.6 Hz at 4 °C, 0.8 Hz at 11 °C and 1.2 Hz at 17 °C). To record mechanical restitution, the regular pacing was interrupted by a single test pulse of variable interval, beginning from an interval just shorter than the regular pacing interval and proceeding to increasingly shorter intervals, until the preparation failed to elicit any contractile response. The force of the test contraction (F_T) was normalized to the force of the preceding control contraction (F_C) and was plotted as a function of the test interval (Δt) to produce a restitution curve. The force of the post-test contraction (F_PT), elicited at a constant delay after the extrasystolic test contraction, was also recorded. Two variables were used to characterize the mechanical restitution. First, the mechanical refractory period (MRP) is the shortest stimulus interval in which the muscle was able to produce a measurable contraction after the control contraction was measured. Second, the time course of mechanical restitution was determined by the time constant (τ) of the single-exponential equation, y=a(1−e^−b/t)+c, where t is time, a is the magnitude of the exponential component, b is the reciprocal of the time constant and c is the part of force recovery that is not explained by the exponential component. This equation was used to fit the mechanical restitution curve. The time constant gives the time needed for F_max to recover to 63 % of its maximum value. To keep the cellular Ca²⁺ load constant, each restitution pulse sequence was separated by six conditioning pulses delivered at the control cycle length.

Rest-potentiation describes the force production at diastolic intervals longer than that of steady-state beating. Steady-state stimulation (as given above) was interrupted for 10 different rest periods ranging from 5 s to 300 s, after which regular pacing was resumed. In atrial preparations, spontaneous beats often limited the pause durations to less than 120 s. The force (F_max) of the first post-rest contraction was normalized to the force of the preceding steady-state beat and is a measure of rest-potentiation. The maximum potentiation of each preparation, achieved at slightly variable diastolic intervals, was recorded. When the effects of ryanodine (10 μmol l⁻¹) on restitution or rest-potentiation were studied, the drug was allowed to incubate for 60 min before the response was measured.

The ATP-dependent Ca²⁺ uptake of SR was measured fluorometrically using a Ca²⁺-selective fluorescent dye, Fura-2 (Hove-Madsen and Bers, 1993a; Kargacin and Kargacin, 1994), as described in detail elsewhere (Aho and Vornanen, 1998). Crude cardiac homogenates were prepared in an ice-cold medium containing (in mmol l⁻¹): Hepes, 20; KCl, 100; MgCl₂, 4; pH 7.0 at 22 °C, and were used within 60 min of their preparation. Ca²⁺ uptake was measured at room temperature (22±1 °C) in a buffer containing (in mmol l⁻¹): Hepes, 20; KCl, 100; MgCl₂, 4; oxalate, 10; Na₂ATP, 1.25; creatine phosphate, 1.25; pH 7.0 at 22 °C. pH 7.0 was shown to be optimal for the Ca²⁺-ATPase of fish muscle SR (Vornanen et al., 1999). Creatine phosphokinase (0.4 i.u. ml⁻¹) was included in the medium to regenerate ATP. Fura-2 was added to a final concentration of 2 μmol l⁻¹ and Ca²⁺ uptake was initiated with 5 μmol l⁻¹ CaCl₂. This gave an initial of 1 μmol l⁻¹, which is close to the optimal Ca²⁺ concentration for SR Ca²⁺ uptake rate (Kargacin and Kargacin, 1994). Ca²⁺ removal from the solution was followed for 5 min, and the maximal rate of Ca²⁺ uptake by the SR was obtained by differentiation of this recording. Ca²⁺ uptake rate was also measured at the K_m (0.4 μmol l⁻¹ Ca²⁺) of the pump, which gave qualitatively similar results as at optimal ⁠. The contribution of mitochondrial Ca²⁺ uptake to Fura-2 fluorescence is excluded, since the concentration of free Ca²⁺ in the uptake medium was less than the threshold concentration (3 μmol l⁻¹) for mitochondrial Ca²⁺ uptake (Hove-Madsen and Bers, 1993b). The mitochondrial blocker Ruthenium Red has an inhibitory effect on SR Ca²⁺ uptake (Aho and Vornanen, 1998; Kargacin and Kargacin, 1998) and was therefore not included. As is normally assumed for an SR-dependent process, fluorescence changes were completely prevented by pretreatment of cardiac homogenates with 2 μmol l⁻¹ thapsigargin (Aho and Vornanen, 1998).

The activity of the myofibrillar Ca²⁺/Mg²⁺-ATPase of both atrial and ventricular muscle tissues was measured from purified cardiac preparations (Vornanen, 1996). Briefly, the samples were homogenised three times for 5 s in 10 volumes of ice-cold buffer 1 containing (in mmol l⁻¹): KCl, 100; Tris-HCl, 10; dithiothreitol, 1; pH 7.4 at room temperature, and centrifuged at 10 000 g for 10 min. Pellets were resuspended in homogenising buffer 2 (buffer 1+1 % Triton X-100) and centrifuged again at 10 000 g for 10 min. The recovered pellets were washed three times in buffer 1 and between washings were centrifuged at 600 g for 15 min. After the last centrifugation, the pellets were resuspended in 20 volumes of low-ionic-strength buffer containing (in mmol l⁻¹): imidazole, 45; KCl, 50; dithiothreitol, 1; pH 7.0 at 22 °C. Ca²⁺/Mg²⁺-ATPase activity of purified atrial and ventricular preparations was measured at three different temperatures (5, 10 and 15 °C). The pH was allowed to freely change according to temperature and was 7.30, 7.25 and 7.11 at 5, 10 and 15 °C, respectively. These pH changes have minimal effects on myofibrillar ATPase activity of the trout heart (Churcott et al., 1994). Total ATPase activity was measured from the liberated inorganic phosphate (Atkinson et al., 1973) in a solution containing (in mmol l⁻¹): imidazole, 45 (pH 7.0 at 22 °C); KCl, 50; EGTA, 5; MgCl₂, 5; Na₂ATP, 3; CaCl₂, 5 (pCa≈4). The background Mg²⁺-ATPase activity was determined in the same solution, but without CaCl₂ (pCa>9). The Ca²⁺/Mg²⁺-ATPase activity was obtained as a difference between the two activities. The concentration of protein was determined by the method of Lowry et al. (1951).

All results are given as mean ± S.E.M. Differences between acclimation groups were compared by one-way analysis of variance (ANOVA). Statistical differences between treatments were evaluated using a paired t-test. All percentage values were compared after arcsine-transformation with Student’s t-test or ANOVA. Kruskal–Wallis ANOVA on ranks was used for populations that did not have equal variance. Post-hoc comparison of two groups were always accomplished using a Student–Newmann–Keuls test. The differences were considered to be significant at P<0.05.

Developed force was larger in atrial muscle (1.3±0.2 mN mg⁻¹, N=4, and 1.4±0.2 mN mg⁻¹, N=4, for CA and WA trout, respectively) than in ventricular muscle (0.3±0.05 mN mg⁻¹, N=4, and 0.3±0.08 mN mg⁻¹, N=4, for CA and WA fish, respectively), but there were no significant differences between acclimation groups (P>0.05). In both acclimation groups, the rate of isometric contraction was much faster (P<0.05) in the atrial than in the ventricular preparation (Fig. 1, Table 1) and the rate of the contraction increased with temperature. Furthermore, temperature acclimation changed the contraction kinetics of the rainbow trout heart. When measured at 11 °C, the isometric contraction was significantly faster (P<0.05) in the heart of the CA fish than in the heart of the WA fish, indicating a positive thermal compensation in contraction kinetics (Fig. 1; Table 2). As indicated by the somewhat slower rate of contraction in the CA fish at 4 °C relative to the contraction of the WA fish heart at 17 °C, the compensation was only partial (Fig. 1). It is notable that the effect of thermal acclimation was mainly seen in the rate of contraction and less so in the rate of isometric relaxation. Ryanodine had no effect on the time course of ventricular contraction in either WA or CA trout. In the atrium of the CA fish, ryanodine prolonged T_PF and shortened T_0.5R at 4 °C (Table 1). Significant prolongation of T_PF was also seen in the atrium of the WA fish at 11 °C (Table 2).

Verapamil (10 μmol l⁻¹), a blocker of L-type Ca²⁺ channels, strongly suppressed F_max in all preparations tested at the acclimation temperatures. In atrial and ventricular muscle of the WA fish, F_max dropped to 13±1 (N=4) and 19±5 % (N=4) of the control, respectively. In the ventricle of the CA fish, F_max decreased to 17±3 % (N=4) and in the atrium to 27±5 % (N=4) of the control value (Fig. 2B). The importance of Ca²⁺ entry through the channels may be due to its direct contribution to cytosolic Ca²⁺, to the maintenance of the action potential plateau, allowing Ca²⁺ entry through the reverse Na⁺/Ca²⁺ exchanger, or as a trigger for further Ca²⁺ release from the SR. The latter alternative was checked by the ryanodine sensitivity of contractility. When measured at physiological pacing frequencies and body temperatures, ryanodine had no effect on the F_max of atrial contraction in the WA fish or on the F_max of ventricular contraction in either WA or CA fish. In contrast, ryanodine had a clear (19±9 %, N=6) and consistent negative inotropic effect on the atrial contraction of the CA fish. This effect was seen in preparations paced to contract at the physiological heart rate (0.6 Hz) and acclimation temperature of the fish (4 °C) (Fig. 2B). This finding indicates that the Ca²⁺ stores of SR can contribute to the activation of contraction in the CA trout atrium under physiologically relevant conditions.

The recovery of contractility from inactivation was determined by restitution experiments as shown in Fig. 3. The mechanical refractory period and time course of restitution were determined from individual muscles. With the lengthening of the test interval, the force of contraction recovers approximately with a single-exponential time course, whereas the force of the post-extrasystolic contraction, delivered at constant delay, remains virtually unaltered. The results of the restitution experiments are shown in Fig. 4. Mechanical restitution was much faster in atrial than in ventricular muscle. In both acclimation groups, the mechanical refractory period was much shorter and the rate constant of force restitution significantly smaller in atrial than in ventricular tissue (Table 3). Acclimation to cold shortened the mechanical refractory period, as was expected on the basis of temperature-induced changes in the duration of contraction. Furthermore, the rate constant of restitution was decreased in both atrial and ventricular tissue after acclimation to cold (Fig. 4). These findings indicate that partial thermal compensation occurs in the force restitution of the rainbow trout heart. Moreover, blockade of the SR Ca²⁺-release channels with ryanodine caused a marked lengthening of the refractory period and an increase in the rate constant of restitution τ in both the atrium and ventricle of the CA fish. This finding indicates that the Ca²⁺ handling of SR is involved in the restitution process of the fish heart (Fig. 4C). In the atrium and ventricle of the WA fish, ryanodine had no effect on the restitution process (Table 3). Furthermore, the mechanical refractory period and the time course of restitution were very sensitive to experimental temperature. The high R₁₀ values (Table 4) suggest that the force restitution is not a passive diffusion-dependent process but more like an enzymatic reaction.

In mammalian hearts, the force of contraction of the post-extrasystolic beat is potentiated, which is considered to reflect the time-dependent redistribution of activator Ca²⁺ within the SR (Braveny and Kruta, 1958; Wohlfart and Noble, 1982). In the trout heart, the force of post-test contraction was not potentiated in any of the cardiac preparations. Even in the atrium of the CA trout, where ryanodine-sensitive Ca²⁺ stores contribute to steady-state force production, post-extrasystolic potentiation was absent. In this regard, fish and mammalian hearts are different.

Rest-potentiation describes the effect of long diastolic intervals on the force generation of the muscle. Prolonged diastole, measured at the acclimation temperatures of the fish, strengthened post-rest contraction in all cardiac preparations of the trout heart (Figs 5 and 6). Maximal rest-potentiation was achieved at rest durations of 40–60 s (being significantly shorter for atrial than for ventricular tissue), whereas at longer rests intervals it either remained relatively constant (atria) or gradually declined (ventricles). Rest-potentiation was significantly stronger (P<0.05) in ventricular (157±5, N=5, and 204±23 %, N=5, for WA and CA fish, respectively) than in atrial (129±1, N=5, and 148±9 %, N=5, for WA and CA fish, respectively) muscle (Fig. 7). When isometric force was used as a measure, the ventricular rest-potentiation was totally insensitive to ryanodine in both CA and WA fish, as was the atrial rest-potentiation of the WA fish (Figs 5 and 6). In contrast, in the atrium of the CA fish, rest-potentiation was completely abolished by ryanodine, as was expected on the basis of ryanodine sensitivity of the steady-state force production (Fig. 5). Again, this reinforces the importance of the SR as a source of cytosolic Ca²⁺ in the atrium of the CA trout. In contrast, if the rate of isometric force (dF/dt_max) was used as a measure, the depressive effect of ryanodine was also evident in the ventricle of the CA fish (Fig. 6).

At 11 °C, rest (60 s) increased the force to 145±7 (N=5) and 136±7 % (N=5) of the control contraction in the ventricles of WA and CA fish, respectively. Ryanodine had no effect on these responses. In the atrial tissue, rest (40 s) increased force production to 140±8 % (N=5) of the control value in both acclimation groups. Ryanodine had no effect on rest-potentiation in the atrium of the WA fish, whereas in the atrial tissue of the CA fish, rest-potentiation was depressed to 108±1 % (N=5) (P=0.025) after the ryanodine treatment.

Regardless of the experimental temperature, the ventricular rest-potentiation was always associated with the prolongation of the twitch (P<0.001) (Fig. 6C). In contrast, rest had no significant effect on the time course of atrial contraction (P=0.094) (Fig. 5).

Briefly stated, rest-potentiation and its ryanodine sensitivity indicate that the prolonged diastole increases the SR Ca²⁺ stores in the atrium and possibly also in the ventricle of the CA trout, but not in the atrial or ventricular muscle of the WA trout.

ATP-dependent Ca²⁺ uptake of SR was measured with Fura-2 from crude atrial and ventricular homogenates. The rate of Ca²⁺ uptake of atrial preparations was significantly higher than that of ventricular preparations in both acclimation groups (Table 5). Furthermore, acclimation to cold caused thermal compensation in the rate of SR Ca²⁺ uptake, which was evident in both atrial and ventricular tissue (Table 5). When the measured Ca²⁺ uptake rates at 22 °C are transformed to the body temperatures of the fish, using the previously determined temperature coefficient (Q₁₀) of 1.64 (Aho and Vornanen, 1998), the SR Ca²⁺ uptake rate was highest in the atrium of the WA trout (100 %), and the relative activities of other preparations were 68, 66 and 52 % for CA atrium, WA ventricle and CA ventricle, respectively. These results indicate that thermal compensation of Ca²⁺ uptake rate in the atrial tissue was only partial, whereas compensation was more complete in the ventricular tissue.

The activity of myofibrillar ATPase was determined at three different temperatures. For both acclimation groups, the myofibrillar ATPase activity was higher in atrial than in ventricular preparations (P<0.05). Furthermore, in both acclimation groups, the temperature-dependence of the ATPase activity (Q₁₀) was higher for atrial than for ventricular preparations (P<0.05) (Table 5). Acclimation to cold increased the activity of myofibrillar Ca²⁺/Mg²⁺-ATPase of the trout atrial tissue (P<0.05), whereas no temperature-induced changes were found in the activity of ventricular ATPase. The Q₁₀ values of the CA fish were lower than those of the WA fish (P<0.05, Table 5), indicating temperature-induced changes in myofibrillar function.

The current concept of fish heart function assumes that filling of the ventricle is primarily determined by atrial contraction (see Johansen and Burggren, 1980; Farrell and Jones, 1992), in contrast to the mammalian heart in which atrial systole has only a relatively small regulatory function. The fast contraction of the mammalian atrium (Urthaler et al., 1975; Asgrimsson et al., 1995; Bottinelli et al., 1995; Minajeva et al., 1997), together with the atrioventricular delay in impulse conduction, is considered to provide accurate timing for the atrial ‘kick’ to augment end-diastolic volume before ventricular systole (Katz, 1992). If atrial function in fish heart is decisively different from that of mammalian heart, it might be assumed that the contractile properties of the fish atrium would also differ from its mammalian counterpart. Specifically, it might be anticipated that contractile differences between atrial and ventricular muscle would be minor if the atrium and ventricle were working as two separate pumps in series. The present results indicate, however, that the rainbow trout heart atrial contraction is much faster than ventricular contraction, as has been observed in mammalian hearts. This suggests that the demands on atrial contractility are not that different in fishes and mammals. This agrees with the recent findings of Lai et al. (1998), indicating that a major part of ventricular filling in fish hearts, as in mammals and other vertebrates, occurs during ventricular relaxation and well before atrial contraction. In this scheme, the fast contraction of atrial cells generates a relatively short and accurately timed atrial systole, which adjusts the end-diastolic volume of ventricle according to the circulatory demands.

The rate of isometric contraction in atrial tissue of the trout heart was approximately double the contraction rate of ventricular tissue (Fig. 1; Table 1). Similar contraction (T_PF) and relaxation (T_0.5R) times to those presented here have also been reported by Ask et al. (1981) for rainbow trout atrium. The fast contraction kinetics of trout atrial tissue was associated with significantly higher myofibrillar ATPase activity relative to the ATPase activity of ventricular tissue, which can explain most of the atrioventricular differences in the rate of isometric contraction (Bárány, 1967). Our values for the specific activity of trout ventricular ATPase (Table 5) are similar to those reported by Churcott et al. (1994), and the difference in myofibrillar ATPase activity between the atrium and ventricle agrees with the findings of Deng and Gesser (1997). In addition to faster myofibrillar ATPase, atrial tissue also had a higher rate of Ca²⁺ uptake into the SR than ventricular tissue (Table 5). It could therefore be assumed that the greater participation of SR in the contractile activation might make the intracellular Ca²⁺ transient faster and thus contribute to faster contraction kinetics in the atrial cells. The slight lengthening of contraction duration in atrial tissue after ryanodine treatment supports this assumption (Table 1). In addition to the function of myofibrils and SR, differences in action potential duration could contribute to atrioventricular differences in the kinetics of contraction (Asgrimsson et al., 1995), although this aspect was not studied systematically in the present work.

Cellular processes that are responsible for the initiation of cardiac contraction need a finite time to recover from inactivation. Mechanical restitution describes this recovery process at short diastolic intervals and reflects the basic cellular mechanisms underlying excitation–contraction coupling (Wohlfart and Noble, 1982). The rate constant of restitution was much shorter in the atrial than in the ventricular muscle of the trout heart (Table 3), indicating that the processes that regulate the recovery of contraction from inactivation are faster in atrial cells than in ventricular myocytes. This kind of atrioventricular difference is also typical of the mammalian heart (Asgrimsson et al., 1995). In mammals, cardiac contraction is initiated by sarcolemmal Ca²⁺ influx, mainly through L-type Ca²⁺ channels, which triggers a large Ca²⁺ release from the SR (Fabiato, 1983; Cheng et al., 1996). Accordingly, mechanical restitution might be explained in terms of either the trigger Ca²⁺ or the SR Ca²⁺ store. Sarcolemmal Ca²⁺ influx is the first step in excitation–contraction coupling, and the force restitution could therefore represent recovery of the Ca²⁺ channels from inactivation. In the function of the release store, at least two different mechanisms could contribute to mechanical restitution: slow replenishment of the releasable Ca²⁺ in the SR, or time-dependent recovery of the SR release channels from inactivation (for references, see Vornanen and Shepherd, 1997). Quantitative analysis of the sarcolemmal Ca²⁺ current has shown that, in trout ventricular myocytes, L-type Ca²⁺ channels can provide a significant proportion of the activator Ca²⁺ (Vornanen, 1998). Therefore, it might be assumed that the time course of restitution in the fish ventricle would, to a relatively large extent, reflect the time- and voltage-dependent recovery of sarcolemmal Ca²⁺ channels from inactivation. Our preliminary findings on isolated ventricular myocytes suggest that, at the physiological resting potential, the time-dependent recovery of Ca²⁺ current from inactivation is significantly faster than the rate of mechanical restitution in multicellular preparations. For example, in ventricular myocytes of the WA trout, the recovery of L-type Ca²⁺ current from inactivation occurred with a time constant of 52±7 ms (holding potential −70 mV, 19 °C, N=4), which is approximately twice as fast as the rate of force recovery (112±3 ms, N=5, at 17 °C) in the ventricular strips. Nor can the atrio-ventricular difference in the rate of force restitution be ascribed to the kinetic properties of the L-type Ca²⁺ current, because the recovery of Ca²⁺ current from inactivation in the atrial cells of the WA trout occurred with a time constant of 41.6±7 ms (N=3), which is not significantly different (P=0.37) from that of ventricular cells. The recovery of Ca²⁺ channels from inactivation is much slower at more depolarized membrane potentials, and it is therefore plausible that the time course of force restitution also partly reflects frequency-dependent changes in the shape of the action potential and hence in the voltage-dependent recovery of the L-type Ca²⁺ current from inactivation. In any case, the atrio-ventricular differences in the time course of mechanical restitution were not abolished by ryanodine, which excludes SR as a major determinant for frequency-dependent differences between the two cardiac compartments. Clearly, the connection between electrical and mechanical restitution in fish cardiac cells needs to be studied in detail. The shorter mechanical refractory period of atrial tissue is probably explained by the shorter contraction and the shorter action potential (Asgrimsson et al., 1995; M. Vornanen, unpublished observations) compared with ventricular muscle.

The force recovery from inactivation at long diastolic intervals is expressed in the potentiation of the post-rest contraction, which was present in both atrial and ventricular tissue of the rainbow trout heart (Figs 5, 6 and 7). The maximum potentiation was smaller and it was achieved at shorter rest intervals in atrial than in ventricular muscle. In this regard, fish and mammalian hearts are again quite similar (Minajeva et al., 1997). The smaller size of the rest-potentiation in atrial tissue suggests that, because of the faster recovery process, the inactivation of force development is reduced in atrial compared with ventricular tissue at physiological heart rates. This means that, with increasing heart rate, the atrial tissue will be able to maintain its force production better than the ventricular muscle. The fast recovery of atrial contraction from inactivation may be significant for cardiac function in fishes, in as much as the strong atrial contraction will maintain adequate ventricular filling and thereby efficient cardiac output at high heart rates.

The present results show that there is partial thermal compensation in the contractile properties of the rainbow trout cardiac muscle. In both cardiac compartments, the effects of thermal adaptation are seen in the kinetics of isometric contraction, in the time course of restitution, in the mechanical refractory period and in rest-potentiation (Fig. 1; Tables 1 and 2). In the subcellular systems, positive thermal compensation was expressed in the Ca²⁺ uptake rate of SR in both the atrial and the ventricular myocardium and in the myofibrillar ATPase activity of the atrial tissue. In fact, the only site where the compensation was clearly absent was the myofibrillar ATPase of the ventricular muscle (Table 5). These temperature-induced changes in contraction kinetics provide partial compensation for the effects of low ambient temperature in the cardiovascular function of the cold-active rainbow trout.

We have demonstrated that acclimation to cold increases the rate of isometric contraction in the rainbow trout heart. A decrease in relaxation time has been found in the heart of the yellow perch (Perca flavescens) (Bailey and Driedzic, 1990), suggesting that this kind of response might be common for teleost fish that remain active in the cold. The kinetics of isometric contraction are determined by the attachment and detachment rates of cross bridges (i.e. it is a function of myofibrillar ATPase activity), and by the rate of activation induced by Ca²⁺ and by Ca²⁺ removal (i.e. the rate of the Ca²⁺ transient) (Rossmanith et al., 1986; Hoh et al., 1988; Ruegg, 1990). The faster kinetics of atrial contraction after cold-acclimation can be explained, to large extent, by positive thermal compensation in the activity of myofibrillar ATPase. We do not know whether this change in myofibrillar function is associated with alterations in myosin heavy chain or in myosin light chain composition, both of which can affect myosin ATPase activity (Bottinelli et al., 1995). In another teleost fish, the crucian carp (Carassius carassius L.), temperature-induced changes in myofibrillar ATPase activity are associated with altered myosin heavy chain expression (Vornanen, 1994). It is notable that positive thermal compensation was absent in the ventricular myofibrils of the rainbow trout. Thus far, we have no explanation for this. The temperature-dependence of ATPase activity was, however, weaker in the CA trout, which suggests that the myofibrillar function of the ventricular myocardium is also modified by the thermal history of the animal.

In part, the faster kinetics of contraction after cold acclimation could be ascribed to enhanced Ca²⁺ handling by the SR, which was evident in the Ca²⁺ uptake rate of SR in both atrial and ventricular muscle. The enhanced activity of the SR Ca²⁺ pump in the heart of the CA trout is in agreement with our previous findings on ventricular muscle of the rainbow trout (Aho and Vornanen, 1998) and extends it to the atrial tissue. A small SR Ca²⁺ leakage was noted when thapsigargin (TG) was applied to Ca-loaded vesicles. Although we have not determined it quantitatively, the leakage is apparently relatively constant in different cardiac preparations since the rate of TG-sensitive Ca²⁺ uptake and the activity of SR Ca-ATPase (not affected by Ca²⁺ leakage) are linearly correlated (Aho and Vornanen, 1998). Therefore, we assume that the temperature-induced changes in TG-sensitive Ca²⁺ uptake of cardiac homogenates reflect true enhancement of SR Ca²⁺ handling after acclimation to cold and not variable Ca²⁺ leakage of different preparations. The proliferation of cardiac SR as a consequence of cold-acclimation has been suggested to occur in the perch (Perca fluviatilis) heart (Bowler and Tirri, 1990), and an increase in the ryanodine sensitivity of contraction has been demonstrated in the ventricle of CA rainbow trout (Keen et al., 1994). Thus, it seems clear that the enhanced activity of SR is part of the thermal adaptation of cardiac function in cold-active fish species.

The ryanodine sensitivity of force (Fig. 2B) and the time course (Table 2) of the atrial contraction suggest that the SR is involved in the kinetics of atrial contraction. Although the acceleration of atrial contraction after cold acclimation can be explained largely by changes in myofibrillar ATPase activity and SR Ca²⁺ uptake rate, these mechanisms are not valid explanations for the kinetic changes in ventricular tissue where ryanodine had no effect on either force or time course of contraction. In trout ventricle, which on the basis of ryanodine sensitivity of force generation (Hove-Madsen and Gesser, 1989; Møller-Nielsen and Gesser, 1992; Shiels and Farrell, 1997; present study) is almost exclusively dependent on extracellular Ca²⁺, cold adaptation may alter membrane excitation, which would lead to shorter action potential duration or faster recovery of Ca²⁺ channels from inactivation. In fact, we have shown that the inactivation kinetics of L-type Ca²⁺ channels is increased after acclimation to cold (Vornanen, 1998). Whether this also applies to recovery from inactivation has not yet been studied.

Adaptation to cold increased the rate of force restitution in the trout heart (Table 3). In both atrial and ventricular tissue the compensation was partial. Furthermore, in the heart of the CA trout, but not in the heart of the WA trout, the time constant of restitution was significantly suppressed by ryanodine, whereas ryanodine had no effect on the force of the post-extrasystolic beat or on the force of the steady-state twitch (ventricle). Thus, while SR may not affect the steady-state force of contraction, it appears to alter the frequency response of the heart in a manner that allows better force generation and greater pumping capacity at high contraction frequencies (Shiels and Farrell, 1997). A similar rightward shift of the restitution curve by ryanodine, as was noted here in the atrium and ventricle of the CA trout, has been found in the rainbow trout heart in response to adrenaline stimulation (Gesser, 1996). In contrast to the trout heart, ryanodine increases the rate of cardiac force restitution (leftward shift of the curve) in mammals, suggesting that some function in the SR forms the rate-limiting step for the restitution process of the mammalian heart (Cooper and Fry, 1990). On the basis of the opposite effects of ryanodine on the time course of restitution in fish and mammalian hearts, it might be anticipated that the effect of the SR on the restitution process of fish heart is indirect; the SR may not provide activator Ca²⁺ to the myofilaments, but it might modify subsarcolemmal Ca²⁺ concentration locally and, thereby, membrane excitation (e.g. by affecting K⁺ and Ca²⁺ currents of the sarcolemma). The lengthening of the refractory period after ryanodine treatment in the atrium and ventricle of the CA trout is in agreement with this hypothesis. Alternatively, it is possible that, in effect, the SR makes a small contribution to cytosolic Ca²⁺ but that, after the blockade of the SR release channels, this is compensated for by enhanced sarcolemmal Ca²⁺ influx, which obscures the inhibitory effect of ryanodine on the steady-state twitch force. The latter hypothesis is based on mutual inhibition between SR and sarcolemmal (SL) Ca²⁺ channels (Adachi-Akahane et al., 1996) and could explain the ryanodine-induced slowing of the restitution rate which would occur if SL Ca²⁺ influx were unable to compensate for the missing Ca²⁺ release from SR due to the incomplete recovery of SL Ca²⁺ influx from inactivation. Regardless of the mechanism, the ryanodine sensitivity of the restitution rate in the CA trout indicates that temperature-induced enhancement of the SR function underlies the increased rate of recovery from inactivation. The positive thermal compensation in the recovery of contractility from inactivation would make room for positive compensation in heart rate, thereby counteracting the temperature-dependent reduction in cardiac output. Thus, temperature-induced changes in the kinetics of contraction will probably be very important for the function of the intact heart at low ambient temperatures.

Rest-potentiation has been suggested to result from the increase in Ca²⁺ release from the SR as a consequence of either the increase in SR Ca²⁺ content or the increase in fractional SR Ca²⁺ release (Bouchard and Bose, 1989; Bers et al., 1993; Bassani et al., 1995). One important finding in the present study is that rest-potentiation can, in different parts of the heart of the same species, be either ryanodine-resistant or ryanodine-sensitive (i.e. rest-potentiation is not necessarily an expression of the SR Ca²⁺ release in fish heart). Indeed, the present findings suggest that, in the majority of cases, rest-potentiation reflects the lengthening of contraction duration (active state), and only in the atrium of the CA trout can it be ascribed solely to the function of SR. It is likely that prolonged rest causes a redistribution of intracellular Ca²⁺, so that sarcolemmal Ca²⁺ entry during the first, long action potential after a rest is enhanced, resulting in potentiated contraction. In addition to the atrium of the CA fish, in the ventricle of the CA trout SR also seems to be affecting the post-rest contraction. However, in this case the force (F_max) of post-rest contraction is not affected by ryanodine, only the rate of force generation (dF/dt_max) is altered. This would again suggest that, under control conditions, a small part of the cytosolic Ca²⁺ is released from the SR and that in the presence of ryanodine this Ca²⁺ fraction is compensated for by a larger sarcolemmal Ca²⁺ influx. This could, in principle, also explain some of the earlier findings where the contraction of fish heart is apparently ryanodine-insensitive even though the SR is powerful. The effect of ryanodine would not be apparent if maximal force was measured but could be evident in its first derivative.

Our results on trout ventricle differ from earlier findings which suggested that rest-potentiation is not associated with lengthening of contraction duration and that the potentiation was completely abolished by ryanodine (Hove-Madsen, 1992). These discrepancies may involve differences in acclimation and experimental temperatures as well as in pacing protocols. In Hove-Madsen’s (1992) experiments, pacing rate was low (0.2 Hz), which might have abolished the difference in time course between steady-state and post-rest contractions, elicited after a 5 min rest. However, the SR Ca²⁺ uptake in trout ventricular myocytes may be so slow that the effect of ryanodine on the force of post-rest contraction would appear only after several minutes rest (Møller-Nielsen and Gesser, 1992), and thus would not be evident in our experiments with a rest duration of 1 min. Contrary to this assumption, under our experimental conditions, ryanodine did not abolish rest-potentiation in trout ventricle even at a rest duration of 5 min.

The present work shows that the SR is functionally important in the atrial tissue of rainbow trout, whereas its physiological role in ventricular contraction is less important and more concealed. We measured the function of the SR by its Ca²⁺ uptake rate and ryanodine sensitivity of contraction. When transformed to the physiological body temperatures of the fish, the rank order for the rate of SR Ca²⁺ uptake was WA atrium (100 %), CA atrium (68 %), WA ventricle (66 %) and CA ventricle (52 %). If this reflects the functional capacity of the SR, it would suggest that the inotropic effects of ryanodine would also follow the same order and would be more evident in the tissues of WA trout than CA trout. However, contrary to this assumption, the effectiveness of ryanodine was more evident in the tissues of the CA fish and weak or nonexistent in the heart of the WA trout. Several putative explanations exist for this apparent discrepancy. First, it is possible that the number of SR Ca²⁺ release channels is lower in WA than CA fish, resulting in less extensive Ca²⁺ release. Unfortunately, there are no data available about the effects of thermal acclimation on the number of ryanodine receptors in fish myocardium. Second, the trigger might be less effective in WA trout than in CA trout. The density of L-type Ca²⁺ current is, however, virtually the same in the ventricular myocytes of WA and CA trout, suggesting that there are no great differences in the trigger (Vornanen, 1998). Third, it is possible that the SR membrane is less leaky to Ca²⁺ in CA fish than in WA fish or at low rather than high experimental temperatures. Our data are consistent with the latter hypothesis, which can explain the finding that, in spite of the lower absolute activity of SR Ca²⁺ uptake in the cold, the ryanodine sensitivity of the force is higher in CA than in WA fish. Indeed, the ryanodine sensitivity of the atrial contraction of the CA fish indicates that the SR Ca²⁺ release channels of the trout heart are not very leaky at low ambient temperatures. This is in contrast to the general idea that, as a result of the Ca²⁺ leakiness of the SR membrane, the contraction of fish cardiac muscle is insensitive to ryanodine at cold temperatures and the contribution of SR Ca²⁺ stores to contractile force is seen only at high experimental temperatures. This assumption is based on observations on sheep cardiac SR, where cooling produces an increase in Ca²⁺ efflux from the SR owing to a cold-induced increase in the open probability of the release channel (Sitsapesan et al., 1991). Our findings, especially for the atrium of the CA trout, clearly demonstrate that the SR of the trout heart is not leaking very much Ca²⁺, otherwise the Ca²⁺ pump of SR could not maintain the Ca²⁺ load that is expressed in the ryanodine sensitivity of the force of the steady-state contraction at physiological heart rates at 4 °C. Therefore, the present results indicate that the Ca²⁺ handling by the cardiac SR is physiologically important for the contractile function of the fish heart, and this is especially characteristic for fish that are physiologically adapted to low ambient temperature. The ability to retain adequate SR Ca²⁺ load in the cold is not limited to the fish heart, as shown by experiments on hibernating mammals (Zhou et al., 1991; Belke et al., 1991). It should be remembered, however, that even if the SR is functionally important in the trout heart, it is still a minor source of cytosolic Ca²⁺, the major part of the Ca²⁺ coming from the extracellular space through the sarcolemma (Tibbits et al., 1990; Møller-Nielsen and Gesser, 1992; Vornanen, 1998).

In summary, the present data indicate that there are marked differences in the kinetics of isometric contraction between atrial and ventricular myocardium, as well as between the hearts of WA and CA rainbow trout. Moreover, it is shown that, in the atrial tissue of the CA trout, the SR contributes cytosolic Ca²⁺ directly, and in the atrium and ventricle of the CA trout the SR significantly increases the rate of force recovery from inactivation. The fast kinetics of contraction and the rapid recovery from inactivation allow relatively high heart rates and therefore adequate cardiac outputs at low environmental temperatures in the cold-active rainbow trout.

This work was supported by The Academy of Finland (project No. 7641). We would like to thank the Kontiolahti fish farm for supplying the trout.