Role of Ca²⁺ in the rapid cooling-induced Ca²⁺ release from sarcoplasmic reticulum in ferret cardiac muscles

Abstract

Rapid lowering of the solution temperature (rapid cooling, RC) from 24 to 3°C within 3 s releases considerable amounts of Ca²⁺ from the sarcoplasmic reticulum (SR) in mammalian cardiac muscles. In this study, we investigated the intracellular mechanism of RC-induced Ca²⁺ release, especially the role of Ca²⁺, in ferret ventricular muscle. Saponin-treated skinned trabeculae were placed in a glass capillary, and the amount of Ca²⁺ released from the SR by RC and caffeine (50 mM) was measured with fluo-3. It was estimated that in the presence of ATP about 45% of the Ca²⁺ content in the SR was released by RC. The amount of SR Ca²⁺ released by RC was unchanged by the replacement of ATP by AMP-PCP (a non-hydrolysable ATP analogue and agonist for the ryanodine receptor but not for the Ca²⁺ pump of SR), suggesting that the suppression of the Ca²⁺ pump of SR at low temperature might not be a major mechanism in RC-induced Ca²⁺ release. The free Ca²⁺ concentration of the solution used for triggering RC-induced Ca²⁺ release was estimated to be only about 20 nM with fluo-3 or aequorin. When this solution was applied to the preparation at 3°C, only a small amount of Ca²⁺ was released from SR presumably by the Ca²⁺-induced Ca²⁺ release (CICR) mechanism. Thus, in mammalian cardiac muscles, RC releases a part of the (<50%) stored Ca²⁺ contained in the SR, and the mechanism of RC-induced Ca²⁺ release may differ from that of CICR, which is thought to play a role in frog skeletal muscle fibres that express ryanodine receptors of different types.

Introduction

Rapid lowering the solution temperature is known to induce contracture in skeletal, cardiac and smooth muscles (rapid cooling contracture, RCC) [1–3]. Among these muscles, treatment of the preparation with low concentrations of caffeine is required for RCC in frog skeletal muscle [4]. In RCC of frog skeletal muscle, caffeine releases a small amount of Ca²⁺ from the sarcoplasmic reticulum (SR) by the Ca²⁺-induced Ca²⁺ release (CICR) mechanism, and low temperature inhibits Ca²⁺ uptake by SR. Thus, it was supposed that both caffeine and low temperature synergistically increase Ca²⁺ concentration around the SR and work to facilitate the CICR mechanism, leading to a massive Ca²⁺ release from SR [5].

However, in mammalian cardiac muscles, pre-treatment with caffeine is not required for RCC [3], and the intracellular Ca²⁺ concentrations before rapid cooling (RC) measured in intact ferret papillary muscles do not correlate with the amount of Ca²⁺ released by RC [6], suggesting that the mechanism of RC-induced Ca²⁺ release differs from that of CICR. Thus, the Ca²⁺ release mechanism by RC in cardiac muscle may differ from that in skeletal muscle.

In this study, we investigated the intracellular mechanism of RC-induced Ca²⁺ release in ferret ventricular muscle. We focused on whether the RC-induced Ca²⁺ release in cardiac muscle could be explained by CICR. For this purpose, quantitative measurement of Ca²⁺ concentrations around SR before RC is essential. Therefore, we used the previously established methodology with saponin-treated skinned preparations in which the SR functions are known to be well preserved [7, 8]. The saponin-skinned trabeculae were placed in a glass capillary, and various solutions with different compositions were directly applied to the SR to estimate the amount of released Ca²⁺ from the SR and Ca²⁺ content in SR.

A part of the results was presented in abstract form [9].

Materials and methods

All experiments were performed in accordance with the Guidelines on Animal Experimentation of The Jikei University School of Medicine.

Preparations

Male ferrets weighing 600–1,200 g were anaesthetised by intraperitoneal injection of pentobarbitone (150 mg kg⁻¹), and the hearts were quickly removed. The blood in the heart was washed out by retrograde perfusion of normal Tyrode’s solution at 30°C through the aorta. Then, the right ventricle wall was quickly opened, and papillary muscles or trabeculae were dissected out in normal Tyrode’s solution at 30°C. The papillary muscles or trabeculae were cut along the longitudinal axis in the relaxing solution with 10 mM EGTA and 4.6 mM ATP at 4°C. The diameter of the preparation was 180–305 μm (241 ± 5 μm, mean ± standard error of the mean, n = 38) and the length was 2.1–6.0 mm (3.8 ± 0.2 mm, n = 38). The preparation was tied at both ends to a tungsten wire (diameter 50 μm, length 12 mm) with silk thread and treated with saponin (10 μg/ml) in the relaxing solution for 30–40 min. The saponin-treated preparation was then inserted into a glass capillary tube (internal diameter 500–600 μm) and fixed inside the capillary by bending the ends of the tungsten wire. The fluorescence signal was measured in the central portion (1 mm long) of the preparation.

Experimental apparatus

The experimental apparatus used was slightly modified from the original apparatus shown in Fig. 1 of Kawai and Konishi [7]. The glass capillary containing the preparation was placed on an inverted microscope (IMT2-F4, Olympus, Tokyo, Japan). One end of the glass capillary was connected to a silicon tube as a solution inlet. The tube was passed through either one of two heat exchangers set at 24 and 3°C, and was connected to a step-motor-controlled multiposition (16 channels) valve (EMT-O-CSD16UWP-HC, Valco Instruments Co. Inc., Houston, TX, USA) for the selection of perfusing solution. A multiposition valve actuator for the rotation of ports was carried out using a personal computer (PC-9801NS/A, NEC, Tokyo, Japan). The other end of the capillary (a solution outlet) was linked to a peristaltic pump (Perista Bio-minipump, ATTO Co., Tokyo, Japan) via a silicon tube. The rate of solution flow was controlled by changing the rotor speed of the pump.

Fig. 1

The relation between fluo-3 concentration and [Ca · Fluo-3] in the presence or absence of 50 mM caffeine. After Ca²⁺ loading, Ca²⁺ was released from the SR into a glass capillary lumen by rapid cooling (a) or 50 mM caffeine (b). G0M1.5F and G0RM0CafF solutions (see Table 1) were used for a and b, respectively, at 3°C. The fluorescence signal was measured with five different concentrations of fluo-3 and was calibrated in terms of Δ[Fluo-3-Ca] (see Eq. 1 in text). Each symbol represents mean ± SEM (n = 5)

Excitation light from a xenon lamp (UXL-75XB, Ushio Inc., Tokyo, Japan) was passed through a 480-nm filter (excitation filter, half bandwidth 15 nm, RDF480, Omega Optical Inc., Brattelboro, VT, USA), and the fluorescence emitted from the capillary of a wavelength longer than 530 nm was detected through a barrier filter (emission filter, OG530, Melles Griot, Carlsbad, CA, USA) with a photomultiplier (R03-RB12M, Hamamatsu Photonics). The output of the photomultiplier was fed to a photometer system (P101/P102, Nikon Co., Tokyo, Japan). Storage and analysis of the data were carried out on a personal computer (PC-9801DX, NEC, Japan). The A/D converter, pump, pinch valve and shutter for the excitation light (EC-601, Copal, Tokyo, Japan) were controlled by the stimulator pulses of a preset sequence (DPS-1300D, Dia Medical System Co., Ltd., Tokyo, Japan).

Solutions and chemicals

The normal Tyrode’s solution used for dissection was as follows (mM): Na⁺, 135; K⁺, 5; Ca²⁺, 2; Mg²⁺, 1; Cl⁻, 102; HCO₃ ⁻, 20; HPO₄ ²⁻, 1; acetate, 20; glucose, 10; insulin, 5 units l⁻¹; pH, 7.3–7.4 at 30°C when equilibrated with 5% CO₂ + 95% O₂.

The composition of the relaxing solution used for cutting out the preparations and for treating the preparations with saponin was as follows (mM): K₂ATP (adenosine 5′-triphosphate, dipotassium salt), 5.2; EGTA (ethylene glycol-bis (2-aminoethyl)-N,N,N′,N′-tetraacetic acid), 10.0; PIPES (piperazine-N-N′-bis (2-ethane sulfonic acid)), 20.0; Mg²⁺, 1.5; MgMs₂, 5.6; KMs, 90.8; NaMs, 10.0; Leupeptin, 10 μg/ml; pH, 7.0 adjusted by KOH at 4°C.

The composition of the solutions used for measuring fluorescence is shown in Table 1. Ionic constituents were computed by solving multi-equilibrium equations using binding constants compiled by Martell and Smith [10]. All solutions contained 20 mM PIPES, 5 μM CCCP (carbonyl cyanide m-chlorophenylhydrazone) and 10 μM DCB (2′,4′-dichlorobenzamil hydrochloride). Ionic strength was 0.2 M. Loading and assay solutions contained 10–20 μg/ml Leupeptin. pH was adjusted to 7.0 by KOH at 3 or 24°C. K₂ATP, AMP (adenosine 5′-monophosphate), K₂AMP-PCP (β,γ-methyleneadenosine-5′-triphosphate), saponin and CCCP were obtained from Sigma Chemical Co. (St. Louis, MO, USA). DCB and fluo-3 (pentapotassium salt) were purchased from Invitrogen (OR, USA). PIPES and caffeine were obtained from Nacalai Tesque, Inc. (Kyoto, Japan), EGTA from Fluka Chemie AG of Sigma-Aldrich Co. and Leupeptin from Peptide Institute, Inc. (Osaka, Japan). All chemicals used were of analytical grade. Aequorin was purchased from Dr. J.R. Blinks.

Table 1 Composition of solutions

Estimation of total Ca²⁺ released from the SR

After RC, Ca²⁺ was released from the SR to the capillary lumen filled with the “assay” solution containing 20 μM fluo-3 (G0M1.5F, Table 1). The fluorescence intensity of fluo-3 in the assay solution was low in the absence of Ca²⁺ release from the SR (F ₀), but it rose to higher levels (F) after the onset of Ca²⁺ release from the SR. To quantify the amount of released Ca²⁺, we used the method previously established in our laboratory [7]. At the end of each experiment, the assay solution plus 1 mM Ca²⁺ was introduced into the capillary that contained the muscle preparation to estimate the fluorescence of Ca²⁺-bound fluo-3 (F _max). The fluorescence of Ca²⁺-free fluo-3 (F _min) was measured by introducing the assay solution plus 1 mM EGTA. F _min was slightly lower than F ₀ because of slight contamination of Ca²⁺ in the assay solution (1.3 μM estimated with fluo-3). The change in fluo-3 fluorescence after Ca²⁺ release (F − F ₀), could be calibrated in terms of the change in concentration of Ca²⁺-bound fluo-3 (Δ[Fluo-3 · Ca]) by the following equation:

$$ \Updelta [{\text{Fluo - 3}} \cdot {\text{Ca}}] = [{\text{Fluo - 3}}]_{\text{tot}} (F - F_{0} )/(F_{\max } - F_{\min } ), $$

(1)

where [Fluo-3]_tot was the total fluo-3 concentration in the assay solution and was 20 μM in the present study.

The amount of Ca²⁺ released from the SR, defined as the change in total Ca²⁺ concentration in the capillary lumen (Δ[Ca]_tot), could be expressed as the sum of changes in concentrations of free Ca²⁺ (Δ[Ca²⁺]), Ca²⁺ bound to fluo-3 (Δ[Fluo-3 · Ca]) and Ca²⁺ bound to other buffers (e.g., intrinsic intracellular buffers and ATP in the solution) (Δ[B · Ca]):

$$ \Updelta [{\text{Ca}}]_{\text{tot}} = \Updelta [{\text{Ca}}^{{ 2 { + }}} ] + \Updelta [{\text{Fluo - 3}} \cdot {\text{Ca}}] + \Updelta [{\text{B}} \cdot {\text{Ca}}]. $$

(2)

When the fluo-3 concentration is high enough to predominantly bind released Ca²⁺ (i.e., Δ[Fluo-3 · Ca] ≫ Δ[Ca²⁺] + Δ[B · Ca]), Δ[Ca]_tot could be approximated with Δ[Fluo-3 · Ca]. Figure 1a shows the relation between the fluo-3 concentration in the assay solution (G0M1.5F, Table 1) and Δ[Fluo-3 · Ca]. After the SR was fully loaded with Ca²⁺, Ca²⁺ release was induced by RC. The calibrated Δ[Fluo-3 · Ca] increased as the fluo-3 concentration was increased up to 200 μM, but was saturated at higher fluo-3 concentrations (200–400 μM). The result suggested that the fluo-3 concentration of 200 μM was high enough to predominantly bind released Ca²⁺ from the SR and that Δ[Fluo-3 · Ca] obtained at high fluo-3 concentrations (≥200 μM) could be an index of the amount of Ca²⁺ released from the SR (i.e., Δ[Ca]_tot expressed in units of μM in the capillary space). With 20 μM fluo-3 primarily used in the present study for economic reasons, on the other hand, Δ[Fluo-3 · Ca] had to be scaled to attain the value that would be obtained if the fluo-3 concentration was sufficiently high (≥200 μM). The scaling factor of 1.80 estimated from Fig. 1a was used in the following sections.

We used the other assay solution that contained 50 mM caffeine and 25 mM AMP (G0RM0CafF, Table 1) to release all Ca²⁺ in the SR. High concentrations of caffeine and AMP should maximise Ca²⁺ release, and the Ca²⁺ pump was expected to be inactive in the absence of ATP. Because caffeine significantly influences fluo-3 fluorescence and affinities of intracellular buffer sites for Ca²⁺, we carried out a separate set of calibrations of fluo-3 fluorescence signals; at the end of each experiment, F _max and F _min were estimated by the introduction of the G0RM0CafF assay solution plus 1 mM Ca²⁺ and the assay solution plus 1 mM EGTA, respectively. The relation between the fluo-3 concentration and Δ[Fluo-3 · Ca] was constructed in the G0RM0CafF assay solution (Fig. 1b). Then Δ[Fluo-3 · Ca] was calculated by Eq. 1 and was scaled by a factor of 2.78 (estimated from Fig. 1b) to yield Δ[Ca]_tot.

Measurement of free Ca²⁺ concentration in the solutions for rapid cooling

The free Ca²⁺ concentration of the solution for RC (G0M1.5F, Table 1) was measured with two different methods. The first method used fluo-3 fluorescence measured in the glass capillary that contained the muscle preparation without loading SR with Ca²⁺, as described below in the next section (protocol D1, Fig. 3). We calculated [Ca²⁺] with

$$ [{\text{Ca}}^{2 + } ] = K_{\text{d}} \, [(F_{0} - F_{\min } )/(F_{\max } - F_{0} )], $$

(3)

where K _d denotes a dissociation constant of fluo-3 for Ca²⁺. For the value of K _d, we used 0.476 μM determined previously at 22°C [7]. Note that little temperature dependence of the K _d value has been reported for Ca²⁺-fluo-3 bindings [11].

The second method used a Ca²⁺ photoprotein, aequorin; aequorin luminescence of the G0M1.5F solution before application to the preparation was measured in vitro at 3°C in the absence of the preparation. Aequorin was dissolved in the solution containing 150 mM KCl and 10 mM PIPES at pH 7.0 to a final concentration of 150 μM. The luminescence was converted to [Ca²⁺] using the following equation [12];

$$ [{\text{Ca}}^{2 + } ] = (L/L_{ \max } )^{1/n} + (L/L_{\max } )^{1/n} \times (K_{\text{TR}} - 1)/\left( {K_{\text{R}} (1 - (L/L_{\max } )^{1/n} )} \right) $$

(4)

where L is the aequorin light obtained from the solutions. L _max is the maximal light at saturating Ca²⁺. For constants n, K _R and K _TR, we used, respectively, 3.11, 3.62 × 10⁶ and 85.62 obtained previously at 3°C (see Methods in [6]).

Experimental protocol

SR was loaded with Ca²⁺ by incubation of the preparation at pCa 6.5 in the loading solution (CaG1M1.5, Table 1) for 2 min at 24°C, the condition optimised to achieve nearly full loading of the SR as in preliminary experiments (data not shown).

• 1. Measurements of the fractional Ca²⁺ release by RC

Protocol A After loading SR with Ca²⁺ and washout of the Ca²⁺ in the preparation with the Ca²⁺-free solution (G1R) at 24°C for 1 min, fluo-3 was introduced (G0RM1.5F), and the bathing solution temperature was lowered from 24 to 3°C within 3 s (G0M1.5F, Table 1). Ca²⁺ was released into a glass capillary (RC-induced Ca²⁺ release), and the fluorescent Ca²⁺ signal was measured (A in Figs. 2, 3).

Fig. 2

Overview of protocols. After loading SR with Ca²⁺ at pCa 6.5 in the solution containing 4.1 mM ATP (CaG1M1.5, see Table 1) for 2 min at 24°C, Ca²⁺ release was triggered by rapid cooling (protocol A), 50 mM caffeine (protocol B), and rapid cooling followed by 50 mM caffeine (protocol C). Ca²⁺ released from the SR was bound to fluo-3 in a glass capillary lumen, and the resultant fluorescence signal was measured as described in the text. Protocol A Temperature of a solution was lowered from 24 to 3°C within 3 s with simultaneous application of 4.6 mM ATP and 20 μM fluo-3 (G0M1.5F, see Table 1). Protocol B A solution that contained 50 mM caffeine, 25 mM AMP and 20 μM fluo-3 (G0RM0CafF, see Table 1) was applied at 3°C. Protocol C Rapid cooling (temperature of G0M1.5F lowered to 3°C) and 50 mM caffeine (G0RM0CafF) were sequentially applied. The meanings of dark and shaded trapezoids are full Ca²⁺ and decreased Ca²⁺ in SR, respectively

Fig. 3

Timetable for solution exchange. Protocol A–C Correspond to, respectively, protocol A–C in Fig. 2. Protocol D1 and D2, which lack the Ca²⁺ loading step, provide the “background” for the RC and caffeine assays, respectively

Protocol B After Ca²⁺-loading and washing the preparation in the same manner as in protocol A, fluo-3 was introduced (G0RM0F, Table 1), and the total Ca²⁺ content in SR was released by 50 mM caffeine at 3°C (B in Figs. 2, 3).

Protocol C When 50 mM caffeine was applied to the preparation following RC-induced Ca²⁺ release, residual Ca²⁺ in SR was measured (C in Figs. 2, 3). In protocol C, the preparation was washed for 1.5 min (by G10R for 1 min and by G1R for 30 s) between RC and the application of 50 mM caffeine.

Protocols D1 and D2 lack SR Ca²⁺ loading steps, but otherwise were very similar to protocols A and B, respectively (Fig. 3).

For estimation of the amount of Ca²⁺ released by RC or caffeine, the fluo-3 fluorescence signals obtained without SR Ca²⁺ loading (protocol D1 for RC, and protocol D2 for caffeine) were subtracted from the fluorescence signals obtained with SR Ca²⁺ loading (protocol A or C for RC; protocol B or C for caffeine). We counted an average of 5 s around the peak of the subtracted signal.

The fractional SR Ca²⁺ release by RC was calculated using the values obtained from protocol A and B.

• 2. Effect of adenine nucleotides on Ca²⁺ release by RC

Protocol A was modified to study effects of ATP on RC-induced Ca²⁺ release. For removal of ATP, the assay solution of protocol A (G0M1.5, Table 1) was replaced by G0RM1.5 that did not contain ATP. When AMP-PCP, a non-hydrolysable ATP analogue, was used, 4.6 mM K₂ATP in the assay solution (G0M1.5F) was simply substituted with 4.6 mM K₂AMP-PCP.

• 3. Measurement of the amount of Ca²⁺ released by the CICR mechanism at 3°C

The protocol was identical to protocol A (above), except for modifications of the solution temperatures. Following Ca²⁺ loading in SR at 24°C for 2 min, the preparation in a capillary was washed with the Ca²⁺-free solution (G1R, Table 1), which contains neither ATP nor Ca²⁺, for 1 min. During this period, the temperature of the solution was slowly lowered from 24 to 3°C. After introduction of fluo-3 at 3°C (G0RM1.5F, Table 1), the solution for RC containing ATP and fluo-3 (G0M1.5F, Table 1) was applied at 3°C to the pre-cooled preparation. The change of fluo-3 fluorescence in a glass capillary was measured as described above for protocol A.

Statistical analysis

We used two-tailed Student’s t test, and a significant difference of p < 0.05 was verified. Statistical values were given as mean ± SEM.

Results

Free Ca²⁺ concentration in the solution for rapid cooling

We measured the contaminated free Ca²⁺ concentration in the solution used for RC containing 4.6 mM ATP and 20 μM fluo-3 with either fluo-3 or aequorin. The average free Ca²⁺ concentration was estimated to be 19 ± 2 nM (n = 11) measured with fluo-3 (protocol D1 in Fig. 3) and 20 ± 2 nM (n = 3) with aequorin at 3°C. There was no significant difference between the two values measured with fluo-3 and that with aequorin. Thus, the free Ca²⁺ concentration around SR was about 20 nM (pCa 7.7) before Ca²⁺ release was induced by RC.

Fractional Ca²⁺ release induced by rapid cooling

Figure 4a shows typical traces of the fluorescence measurements with protocols C (upper) and B (lower) (Figs. 2, 3) obtained from the same preparation. After Ca²⁺ loading of the SR, RC induced a large increase in the [Fluo-3-Ca] in the capillary space (a). The preparation was briefly washed with Ca²⁺-free solutions (G10R and G1R, see protocol C in Fig. 3), and a subsequent application of 50 mM caffeine (plus 25 mM AMP) caused a substantial increase in [Fluo-3-Ca] (b), indicating that the SR still contained a significant amount of Ca²⁺ after RC. When the SR was reloaded with Ca²⁺ and the caffeine assay (50 mM caffeine plus 25 mM AMP) was repeated without RC in the same preparation, an increase in [Fluo-3-Ca] was greater by about a factor of two than that induced by RC (c).

Fig. 4

The fraction of SR Ca²⁺ released by rapid cooling. A Examples of fluorescence traces obtained with protocols C (upper) and B (lower). Fluo-3 fluorescence signals have been calibrated in terms of [Fluo-3-Ca]. With protocol C (upper traces), rapid cooling caused a rise of [Fluo-3-Ca] (a). After a brief wash with Ca²⁺-free solutions, subsequent application of 50 mM caffeine also caused an increase of [Fluo-3-Ca] (b) that reflected the release of residual Ca²⁺ in SR. With protocol B (lower trace), a large increase in [Fluo-3-Ca] was observed with the application of the solution containing 50 mM caffeine (c). At the end of assay periods in a–c (indicated by horizontal bars), rapid decreases of [Fluo-3-Ca] were due to closing an optical shutter, which terminated the fluorescence recordings. B Columns a–c summarise the Ca²⁺ release from SR (Δ[Fluo-3-Ca]) obtained from the type of experiment shown in a–c, respectively, in A. Columns show mean ± SEM of five preparations

From pooled data shown in Fig. 4b, the amount of SR Ca²⁺ released by RC (a) was, on average, 3.16 ± 0.46 μM. On the other hand, the residual SR Ca²⁺ after RC (b) was, on average, 2.39 ± 0.61 μM. This value was significantly smaller (p < 0.01) than the 4.19 ± 0.49 μM predicted by the total amount of releasable Ca²⁺ (c = 7.34 ± 0.93 μM) minus the amount Ca²⁺ released by RC (a). This difference is likely due to Ca²⁺ leakage from the SR into solutions that contained EGTA during 1.5 min washing periods inserted between RC and caffeine applications (see Fig. 3, protocol C). In saponin-treated cardiac muscles of mice, Morimoto et al. [13] reported a substantial Ca²⁺ leakage from the SR; about 50% of the SR Ca²⁺ content was lost when the preparation was perfused with the solution that contained 1 mM EGTA for 1.5 min at 22°C.

Thus, the fraction of Ca²⁺ released by RC was, on average, 44.7 ± 1.7% of the total SR Ca²⁺ content in 16 preparations.

Rapid cooling with ATP opens the considerable SR Ca²⁺ release channels

With the standard solution containing 4.6 mM ATP (G0M1.5F, Table 1), a considerable amount of Ca²⁺ was released from the SR by RC (Fig. 5a), whereas in the absence of ATP (G0RM1.5F, Table 1) RC caused a much a smaller Δ[Fluo-3-Ca] signal (Fig. 5b), suggesting that Ca²⁺ release channels of SR hardly opened. The amount of Ca²⁺ released by RC in the presence and absence of ATP were 2.06 ± 0.27 and 0.41 ± 0.09 μM, respectively (Fig. 5c).

Fig. 5

Rapid cooling-induced Ca²⁺ release in the presence or absence of ATP. a, b Traces show changes in solution temperature (top) and [Fluo-3-Ca] (bottom) in the presence (a) or absence (b) of ATP. The fluorescence recording was terminated by closing an optical shutter at the times indicated by arrows. c Summary of RC-induced Ca²⁺ release (Δ[Fluo-3-Ca]) estimated in the type of experiments shown in a and b. Left and right columns represent mean ± SEM of data obtained, respectively, in the presence (n = 16) and in the absence (n = 7) of ATP

Rapid cooling-induced Ca²⁺ release in the presence of ATP or AMP-PCP

The SR Ca²⁺ pump still works (albeit slowly) in the presence of ATP at 3°C [14]. Therefore, SR could reuptake a considerable part of the Ca²⁺ released by RC and might decrease the free Ca²⁺ concentration in the glass capillary lumen. Inhibition of the SR Ca²⁺ pump could lead to an increase in the free Ca²⁺ concentration around SR and might trigger the CICR. In order to examine a possible role of CICR in the RC-induced Ca²⁺ release, we compared the fraction of SR Ca²⁺ released by RC in the presence of ATP and AMP-PCP, a non-hydrolysable ATP analogue. The fraction of Ca²⁺ released by RC was 44.7 ± 1.7% with 4.6 mM ATP (see above) and was 45.6 ± 8.2% with 4.6 mM AMP-PCP (Fig. 6). There was no difference between these two values. The present results suggest that the RC-induced Ca²⁺ release is not influenced by hydrolysis of ATP, in other words, by the activity of SR Ca²⁺ pump.

Fig. 6

Fraction of SR Ca²⁺ released by rapid cooling in the presence of either ATP or AMP-PCP. Ca²⁺ release was induced by rapid cooling in the presence of 4.6 mM ATP (left, n = 16) or in the presence of 4.6 mM AMP-PCP (right, n = 4). The amount of Ca²⁺ released by rapid cooling was normalised to the total amount of releasable Ca²⁺ in the SR (ordinate). Each bar represents the mean ± SEM

The Ca²⁺-induced Ca²⁺ release at 3°C

After Ca²⁺ loading in the SR, the temperature of the solution for the wash (G1R), which contains neither ATP nor Ca²⁺, was slowly lowered from 24 to 3°C for 1 min, and then the solution for RC (containing 4.6 mM ATP and ~20 nM free Ca²⁺) was applied at 3°C to the preparation (slow cooling). Then, we measured the change of Ca²⁺ concentration in a glass capillary. The typical data are shown in Fig. 7, and the mean value was 0.23 ± 0.10 μM (n = 5).

Fig. 7

Ca²⁺ release from the SR induced by the solution for rapid cooling at 3°C. After Ca²⁺ loading of the SR, the temperature of the solution for the wash (G1R, Table 1) and the pre-assay solution (G0RM1.5F, Table 1) was gradually lowered from 24 to 3°C, and then the assay solution that contained ATP and fluo-3 (G0M1.5F) was applied at 3°C to the pre-cooled preparation as indicated at the top. Fluo-3 fluorescence signals were measured and were calibrated in terms of [Fluo-3-Ca]. An optical shutter was closed at a time indicated by an arrow

Because the temperature around the preparation was kept constant at approximately 3°C, the small Ca²⁺ release induced by the solution exchange was not due to cooling. Rather, it was thought that Ca²⁺ was released from the SR by changes in ATP and free Ca²⁺ concentrations (i.e., CICR). Aside from the initial rise, the [Fluo-3-Ca] signal also showed a gradual rise that started about 10 s after the solution exchange (Fig. 7). We might speculate that the CICR mechanism was gradually accelerated as the free Ca²⁺ concentration around the SR increased.

Discussion

Role of Ca²⁺ in the rapid cooling-induced Ca²⁺ release

In frog skeletal muscle, a slight increase in the intracellular Ca²⁺ concentration around the SR before RC was required for the RC-induced Ca²⁺ release. To increase the intracellular Ca²⁺ concentration before RC, a low concentration of caffeine was employed because the RC itself could not sufficiently increase the intracellular Ca²⁺ concentration for triggering a considerable Ca²⁺ release in skeletal muscles [4]. Therefore, the RC-induced Ca²⁺ release in skeletal muscles was assumed to be due to the enhancement of the CICR mechanism [5]. During RCC in frog skeletal muscle, the change in the aequorin light signal showing the intracellular Ca²⁺ concentration occurred in three phases. The first phase was a transient change of intracellular Ca²⁺ concentration accompanying slight tension. During the second phase, the light signal slowly increased as cooling produced maximum tension development. The third phase was an additional large light signal induced after the second phase, even though the tension was saturated. The second and third phases were more sensitive to low concentrations of procaine, an effective inhibitor for CICR, than the first phase [4].

On the other hand, pretreatment of the preparation with caffeine is not required for the initiation of RCC in mammalian cardiac and smooth muscles [2, 3]. In contrast to RCC in skeletal muscles, the amount of Ca²⁺ released by RC in ferret ventricular muscles was not influenced by the intracellular Ca²⁺ concentration before RC [6].

In the present study, in skinned cardiac muscles, the free Ca²⁺ concentration in the solution used for RC (G0M1.5F, see Table 1) was about 20 nM when estimated with either fluo-3 or aequorin, which seems lower than the level required to activate the CICR. It has been reported that the CICR is hardly activated at Ca²⁺ concentrations below 100 nM (i.e., pCa < 7) [15–17]. Thus, the contribution of CICR may be minor, if it exists at all, for the initiation of Ca²⁺ release induced by RC. This is supported by two lines of evidence. (1) When the preparation was perfused with the solution containing 20 nM free-Ca²⁺ and 4.6 mM ATP (G0M1.5F, Table 1) with the temperature set constant at 3°C (Fig. 6), the [Fluo-3-Ca] signal rose slowly with the delayed onset (~5 s after the solution exchange), and the [Fluo-3-Ca] signal was, on average, 0.23 ± 0.10 μM. This value was only about 10% of that achieved by RC (2.06 ± 0.27 μM). (2) The SR Ca²⁺ pump still works, albeit slowly, in the presence of ATP even at lower temperatures [14]. When the SR Ca²⁺ pump is completely inhibited (by replacement of ATP with AMP-PCP), the CICR may be enhanced because of the local increase in the free Ca²⁺ concentration around the SR. The fractional Ca²⁺ release induced by RC was unchanged even when ATP was replaced by AMP-PCP. Overall, the results suggest that CICR might not play an important role in the RC-induced Ca²⁺ release.

Mechanisms of opening the Ca²⁺ release channels by rapid cooling

Global conformational changes upon binding ligands were observed with the gap junction protein [18] and with the nicotinic acetylcholine receptor [19]. Orlova et al. [20] demonstrated the three-dimensional structure of the rabbit skeletal muscle Ca²⁺ release channel in an open state using electron cryomicroscopy and angular reconstitution. In contrast to its closed state, in the open state reconstruction, a central cavity was revealed in the transmembrane region of the channel in the presence of Ca²⁺ and ryanodine. The opening of the channel is associated with a 4° rotation of its transmembrane region with respect to its cytoplasmic region and with significant mass translocations within the entire cytoplasmic region of the channel tetramer.

RC alone hardly opens the SR Ca²⁺ release channels. In the present study, RC could release considerable Ca²⁺ when ATP or AMP-PCP co-existed (Figs. 5, 6). Sitsapesan et al. [21] also confirmed that the SR Ca²⁺-release channels in the artificial lipid bilayer were not activated at low temperatures in the absence of cytosolic Ca²⁺ or a cardiotonic agent, sulmazole. Another study revealed that the presence of 10 μM cytosolic Ca²⁺ and 100 μM cytosolic ATP increased the mean open probability from 0.052 to 0.284 in sheep cardiac ryanodine receptor channels incorporated into planar phospholipid bilayers [22].

At low temperature and in the presence of Ca²⁺ [21], ryanodine, caffeine and adenine compounds [22], the Ca²⁺ release channels seem to be stable in an open state compared with that in closed state. In other words, low temperature is considered one of the so-called opening factors on the SR Ca²⁺ release channels. In cardiac, skeletal and smooth muscles, in the presence of one or more of those opening factors, the Ca²⁺ release channels might be stable in an open state. In addition, it was considered that the rapid lowering of the temperature may be important to synchronize the opening of a large population of Ca²⁺ release channels on SR membranes.

Fractional amount of Ca²⁺ release by rapid cooling

Bers et al. measured intracellular Ca²⁺ transients during RCCs in guinea pig ventricular myocytes using the fluorescent Ca²⁺ indicator, Indo-1. They demonstrated that RC of myocytes from 22 to 0–1°C could release a large amount of Ca²⁺ ([Ca²⁺]_i > 10 μM) from the SR and suggested the use of RCCs as a useful means of assessing the SR Ca²⁺ content in intact cardiac muscle cells [23].

In our previous study, we estimated the fraction of SR Ca²⁺ released by the RC in intact papillary muscles of ferrets [6]. The intracellular free Ca²⁺ concentration was measured with aequorin, and Ca²⁺ release was first induced by RC (from 30 to 4°C) and subsequently by 15 mM caffeine. This experimental protocol was somewhat similar to that employed in the present study (Fig. 4a, upper). Both RC and subsequent application of caffeine caused transient rises in [Ca²⁺]_i, and the peak changes in [Ca²⁺]_i induced by RC and caffeine averaged, respectively, 1.59 and 0.90 μM [6]. From these results, the fraction of SR Ca²⁺ released by RC was calculated to be 64% [=1.59/(1.59 + 0.9) × 100], which was higher than that estimated in the present study (44.7%; see “Results”). The difference may be attributed to the experimental methods employed in the previous study, in which peaks of Δ[Ca]_i measured with or without 15 mM caffeine were directly compared. The increase in apparent Ca²⁺ sensitivity was produced by caffeine in skinned cardiac and skeletal muscle fibres of the rat [24]. Furthermore, the complex of troponin (Tn) C with Tn I and Tn T resulted in an increase in the affinity of Tn C for Ca²⁺ in the presence of caffeine in the bovine heart [25]. Therefore, the amount of Ca bound to Tn C should be greater at any given [Ca²⁺]_i. It follows that the use of Δ[Ca]_i measured during caffeine application would erroneously underestimate the total amount of Ca²⁺ released from the SR (Δ[Ca]_tot), unless an appropriate correction was made. Consequently, a fraction of Δ[Ca]_tot induced by RC would have been overestimated in the previous study. In the present study, on the other hand, Δ[Ca]_tot was deduced by scaling Δ[Fluo-3 · Ca]_tot with factors separately estimated in the presence (2.78) and in the absence (1.80) of 50 mM caffeine (Fig. 1). Thus, we consider that the revised value in the present study (44.7%) is more relevant as an estimate of the fraction of SR Ca²⁺ released by RC.