Cytosolic energy reserves determine the effect of glycolytic sugar phosphates on sarcoplasmic reticulum Ca²⁺ release in cat ventricular myocytes

Abstract

Localization of glycolytic enzymes in close proximity to Ca²⁺ transport systems of the sarcoplasmic reticulum (SR) in cardiac cells suggests an important functional role for glycolysis in intracellular [Ca²⁺] regulation and, consequently, excitation–contraction coupling. Here, we investigated the mechanisms of regulation of SR Ca²⁺ release by glycolytic sugar phosphate intermediates in cat ventricular myocytes. Experiments with permeabilized myocytes revealed that with normal cytosolic energy reserves (mm: ATP 5, ADP 0.01, phosphocreatine (CrP) 10) fructose-1,6-bisphosphate (FBP; 1 mm) and fructose-6-phosphate (F6P; 1 mm) caused a transient increase of Ca²⁺ spark frequency by 62 and 42%, respectively. This effect of sugar phosphates was associated with a 13% decrease in SR Ca²⁺ load. Pretreatment of the cells with an inhibitor of glycolysis, iodoacetate (IAA; 0.5 mm), did not prevent the effects of FBP and F6P on Ca²⁺ sparks. Recording of single ryanodine receptor (RyR) channel activity indicated that FBP and F6P significantly increased RyR open probability. Reduction of cytosolic energy reserves decreased Ca²⁺ spark activity. Increasing [ADP] to 0.4 mm or removal of CrP ([ATP] was kept constant) caused a slowly developing decrease of Ca²⁺ spark frequency by 29 and 42%, respectively. Changing [ADP] and [CrP] simultaneously decreased Ca²⁺ spark frequency by 66%. This inhibition of Ca²⁺ sparks was associated with a 40% decrease in SR Ca²⁺ load. The subsequent addition of FBP (1 mm) partially restored Ca²⁺ spark frequency and SR Ca²⁺ load. This recovery of Ca²⁺ sparks was blocked completely by IAA. These data suggest that at physiological ATP, ADP and CrP levels accumulation of sugar phosphates from glycolysis can stimulate SR Ca²⁺ release. This effect does not require the activity of downstream glycolytic enzymes, but rather is the result of direct activation of RyRs. However, under conditions associated with depletion of cellular energy reserves (e.g. myocardial ischaemia), ATP generated from glycolysis may play an important role in maintaining myocardial Ca²⁺ homeostasis by improving SR Ca²⁺ uptake.

Despite the fact that glycolytically derived ATP contributes only a small fraction of the total ATP produced in a cardiac myocyte under normal aerobic conditions (Kobayashi & Neely, 1979; Stanley et al. 2005), an intact glycolytic pathway appears to be essential for cardiac function (Mallet et al. 1990; Jeremy et al. 1993; O'Rourke et al. 1994). Several studies indicate that glycolysis is especially important to maintain intracellular Ca²⁺ homeostasis and normal excitation–contraction (EC) coupling during ischaemia-reperfusion-related Ca²⁺ overload (Jeremy et al. 1992; Kusuoka & Marban, 1994; Aasum et al. 1998). The possible explanation for such a vital role of glycolysis is a functional compartmentalization of glycolytic enzymes in cardiac myocytes. Glycolytic enzymes have been found associated with the sarcolemmal and sarcoplasmic reticular membranes (Pierce & Philipson, 1985; Xu & Becker, 1998) where they support the activity of ion pumps and channels participating in EC coupling (Weiss & Lamp, 1987; Glitsch & Tappe, 1993; Xu et al. 1995). For example, it has been shown that addition of glycolytic substrates (sugar phosphates) and cofactors to isolated sarcoplasmic reticulum (SR) microsomes produced sufficient amounts of ATP to support SR Ca²⁺-ATPase (SERCA) activity (Xu et al. 1995). Furthermore, glycolytically produced ATP was more effective in maintaining Ca²⁺ transport into the SR than exogenously added ATP. It has been suggested therefore that ATP generated by SR-associated glycolytic enzymes may have preferential access to the SR Ca²⁺ pump in a restricted microenvironment. Thus, a functional compartmentalization of SERCA and glycolysis could be a critical factor for maintaining Ca²⁺ homeostasis and the stability of beat-to-beat Ca²⁺ cycling of the heart, especially under conditions of limited energy supply (e.g. during myocardial ischaemia).

There is also evidence that glycolysis regulates the function of the SR to accumulate and release Ca²⁺ in a more direct way. The SR Ca²⁺ release channel, the ryanodine receptor (RyR), is sensitive to intermediates and products of glycolysis. Fructose-1,6-bisphosphate, fructose-6-phosphate and glucose-6-phosphate have been reported to activate the cardiac RyR in planar lipid bilayers and to stimulate ryanodine binding to the cardiac RyR (Kermode et al. 1998a; Kockskämper et al. 2005). On the other hand, products of glycolysis, i.e. pyruvate and l-lactate, cause direct inhibition of RyR activity (Zima et al. 2003; Kockskämper et al. 2005). Therefore, alterations of glycolytic flux which occur, for example, during periods of ischaemia-reperfusion are expected to cause fluctuations of the levels of glycolytic intermediates in close proximity to the SR membrane and, thereby, modulate the activity of the RyR channel and the SR Ca²⁺ pump. This, in turn, will alter EC coupling via changes in SR Ca²⁺ load and the amplitude of the [Ca²⁺]_i transient.

The aim of this study was to investigate the mechanisms of regulation of SR Ca²⁺ release by glycolytic sugar phosphate intermediates in ventricular myocytes. In cardiac myocytes, global [Ca²⁺]_i transients result from the summation of elementary release events, Ca²⁺ sparks (Cheng et al. 1993). Therefore, to gain a better understanding of the mechanisms of regulation of [Ca²⁺]_i by glycolysis, we investigated spontaneous Ca²⁺ sparks using laser scanning confocal microscopy in permeabilized myocytes. Membrane permeabilization allows the study of SR function under controlled conditions with regard to cytoplasmic ion concentrations and energy supply. Furthermore, this technique also allows studying directly the effect of membrane-impermeable substances such as glycolytic sugar phosphates on SR Ca²⁺ release. The results revealed that the effects of the glycolytic sugar phosphates on SR function were dependent on the cellular energy status. With normal cellular energy reserves, accumulation of sugar phosphates stimulated SR Ca²⁺ release directly by activation of RyR channels. However, under conditions of a low ATP/ADP ratio and absence of CrP, which depresses SR Ca²⁺ uptake, sugar phosphates could be used by glycolytic enzymes for ATP synthesis and support of SR Ca²⁺ pump activity to normalize SR Ca²⁺ load and release.

Methods

Measurements of Ca²⁺ sparks in permeabilized cat ventricular myocytes

The procedure for cell isolation was approved by the Institutional Animal Care and Use Committee of Loyola University Chicago, Stritch School of Medicine. Adult mongrel cats of either sex were anaesthetized with thiopental sodium (30 mg kg⁻¹i.p.). Following thoracotomy hearts were quickly excised, mounted on a Langendorff apparatus, and retrogradely perfused with collagenase-containing solution at 37°C according to the method previously described (Rubenstein & Lipsius, 1995). Twenty hearts were used to isolate ventricular myocytes and SR vesicles.

The cell membrane was permeabilized by exposure to saponin (0.005% for 30 s; Zima et al. 2003). After permeabilization cells were placed in a solution composed of (mm): potassium aspartate 100; KCl 15; KH₂PO₄ 5; MgATP 5; Na₂ADP 0.01; NAD⁺ 2; EGTA 0.4; CaCl₂ 0.1; MgCl₂ 0.75; phosphocreatine (CrP) 10; Hepes 10; fluo-3 pentapotassium salt (Molecular Probes Invitrogen, Carlsbad, CA, USA) 0.04; creatine phosphokinase 5 U ml⁻¹; dextran (MW: 40 000) 8%, and pH 7.2 (KOH). Free [Ca²⁺] and [Mg²⁺] of this solution were 100 nm and 1 mm, respectively (calculated using WinMAXC 2.05, Stanford University, CA, USA). All experiments were performed at room temperature (22–25°C). Changes in [Ca²⁺]_i were measured with a laser scanning confocal microscope (Radiance 2000 MP, Bio-Rad, UK) equipped with a × 40 oil-immersion objective lens (N.A. = 1.3). The Ca²⁺ indicator fluo-3 was excited with the 488 nm line of an argon ion laser and fluorescence was measured at wavelengths > 515 nm. Images were acquired in the linescan mode (3 or 6 ms per scan; pixel size 0.1 μm). Ca²⁺ sparks were quantified in terms of amplitude (ΔF/F₀), duration (ms), spatial width (μm) and frequency (sparks s⁻¹ (100 μm)⁻¹) using an automated detection algorithm (Cheng et al. 1999; Zima et al. 2004). F₀ is the initial fluorescence recorded under steady-state conditions and ΔF = F − F₀. Duration and width of sparks were measured at half-maximal amplitude. The algorithm detects sparks as areas of elevated fluorescence (F) intensity relative to the standard deviation (s.d.) of background noise of the fluorescence image. The detection threshold was set to 3.0 ± s.d., which translated to the detection of sparks with an amplitude of ΔF/F₀ > 0.3. Since detection relies on a threshold ΔF/F₀ value changes in Ca²⁺ spark amplitude potentially affect calculated Ca²⁺ spark frequencies, such that, e.g. a general decrease of the Ca²⁺ spark amplitude (cf. for example Figs 1, 3 and 4) would tend to lead to an apparent decrease and underestimation of frequency because a larger fraction of sparks would fall below detection level. No corrections for such missed events were made in the frequency calculations.

Figure 1. Effect of fructose-1,6-bisphosphate (FBP) on Ca²⁺ sparks in permeabilized cat ventricular myocytes.

A, top: confocal linescan images of Ca²⁺ sparks under control conditions, 2 and 6 min after addition of FBP (1 mm) and following washout of FBP. Bottom: local F/F₀ profiles obtained by averaging fluo-3 fluorescence from 2 μm wide regions marked by the black boxes. B, average effect of FBP on Ca²⁺ spark frequency (values were normalized to the levels recorded under control conditions); C, spark amplitude; and D, spark width. Ca²⁺ spark amplitude and width were averaged over 6 min under control conditions and from 0 to 2 min and from 4 to 6 min following application of FBP. *P < 0.05; **P < 0.01 versus control. In this and in the following experiments (Figs 1–4) [ATP] (5 mm), [ADP] (0.01 mm) and [CrP] (10 mm) in the perfusion solution were kept constant throughout the entire experiment.

Figure 3. Effect of fructose-6-phosphate (F6P) on Ca²⁺ sparks.

A, top: confocal linescan images of Ca²⁺ sparks under control conditions, 2 and 6 min after addition of F6P (1 mm) and following washout of F6P. Bottom: local F/F₀ profiles obtained by averaging fluo-3 fluorescence from 2 μm wide regions marked by the black boxes. B, average effect of FBP on Ca²⁺ spark frequency; C, spark amplitude; and D, spark width. Ca²⁺ spark amplitude and width were averaged over 6 min under control conditions and from 0 to 2 min and from 4 to 6 min following application of F6P. *P < 0.05 versus control.

Figure 4. Effect of fructose-1,6-bisphosphate (FBP) on Ca²⁺ sparks in the presence of the inhibitor of glycolysis iodoacetate (IAA).

A, top: confocal linescan images of Ca²⁺ sparks in the presence of IAA (0.5 mm), 2 and 6 min after the subsequent addition of FBP (1 mm) and following washout of FBP. Bottom: local F/F₀ profiles obtained by averaging fluo-3 fluorescence from 2 μm wide regions marked by the black boxes. B, average effect of FBP on Ca²⁺ spark frequency; C, spark amplitude; and D, spark width. Ca²⁺ spark amplitude and width were averaged over 6 min in the presence of IAA and from 0 to 2 min and from 4 to 6 min following application of FBP. *P < 0.05 versus control.

Ryanodine receptor single channel recordings

SR vesicles isolated from cat ventricle were incorporated into planar lipid bilayers and RyR channel activity was measured as previously described (Zima et al. 2003). For recording of Cs⁺ currents through RyRs, the cis- and trans-chambers contained (mm): CsCH₃SO₃ 400; CaCl₂ 0.1; Hepes 20; pH 7.3. The cis- and trans-chambers corresponded to the cytosolic and luminal side of the RyR channel, respectively. Free [Ca²⁺] in the cis-chamber was adjusted to 3 μm by adding an appropriate amount of EGTA. Single channel currents were recorded with an Axopatch 200B amplifier (Axon Instruments, Foster City, CA, USA) at a holding potential of −20 mV. Currents were filtered at 1 kHz and sampled at 5 kHz. The mean open probabilities (P_o) of the channels were used as an index of activity. P_o values were calculated from the 50% threshold analysis using pCLAMP software (Axon Instruments).

Drugs

Glucose-6-phosphate (G6P) mono-potassium salt, fructose-6-phosphate (F6P) di-potassium salt, fructose-1,6-bisphosphate (FBP) di-potassium salt, iodoacetic acid (IAA) mono-sodium salt and NAD⁺ mono-sodium salt were purchased from Sigma-Aldrich. Addition of glycolytic intermediates at the concentrations used in this study (0.2–3 mm) did not change the free [Ca²⁺] in the experimental solutions for Ca²⁺ sparks measurements and RyR channel recordings (verified with a Ca²⁺-sensitive electrode; Orion Research Incorporated) or the fluorescent properties of fluo-3.

Statistics

Data are presented as mean ± s.e.m. of n measurements. Statistical comparisons between groups were performed with the Student's t test. Differences were considered statistically significant at P < 0.05.

Results

Effects of glycolytic sugar phosphates on Ca²⁺ sparks at physiological cellular energy reserves

We studied the effect of glycolytic sugar phosphate intermediates on spontaneous Ca²⁺ sparks in saponin-permeabilized ventricular myocytes at a physiological cytosolic ATP/ADP ratio ([ATP] = 5 mm, [ADP] = 0.01 mm) and CrP content ([CrP] = 10 mm). Under these conditions, Ca²⁺ sparks were detected with a stable frequency of 3.0 ± 0.5 sparks s⁻¹ (100 μm)⁻¹ (n = 13 cells). Figure 1A shows representative confocal linescan images of Ca²⁺ sparks and plots of F/F₀ from selected subcellular regions under control conditions, 2 and 6 min after addition of fructose-1,6-bisphosphate (FBP) and following washout of FBP. Average data of FBP effects on Ca²⁺ spark frequency indicate that the sugar phosphate caused a transient increase in Ca²⁺ spark activity (Fig. 1B). Specifically, within 2 min after addition of FBP (1 mm) Ca²⁺ spark frequency increased by 40 ± 9% (n = 13; P < 0.05). After 4–6 min the frequency almost returned to the control level (4 ± 20%). Upon washout of FBP, the frequency of Ca²⁺ sparks decreased to 69 ± 9% of control (n = 13; P < 0.05 versus control). During the initial stimulation of SR Ca²⁺ release by FBP, the spatio-temporal properties of Ca²⁺ sparks were not affected by the sugar phosphate. However, during the later phase of FBP action (after > 4 min of FBP application), spark amplitude and width (but not duration) significantly decreased (Fig. 1C and D). In control conditions, Ca²⁺ spark amplitude, duration and width averaged 0.88 ± 0.04 (ΔF/F₀), 48.5 ± 2.7 ms and 1.75 ± 0.04 μm, respectively. After 4 min of FBP (1 mm) exposure, these parameters decreased to 0.64 ± 0.03 (ΔF/F₀), 45.0 ± 2.3 ms and 1.51 ± 0.06 μm, respectively (n = 13 cells; number of sparks analysed in control conditions and in the presence of FBP was 798 and 592, respectively).

FBP did not cause any detectable changes in [Ca²⁺]_i when RyR channels were inhibited by ryanodine (10 μm) or tetracaine (1 mm) indicating that the sugar phosphate enhanced specifically RyR-mediated SR Ca²⁺ release (data not shown).

Ca²⁺ spark frequency is highly dependent on SR Ca²⁺ load (Lukyanenko et al. 1996; Satoh et al. 1997). Therefore, we examined changes in SR Ca²⁺ load during application of FBP. SR Ca²⁺ load was estimated from the [Ca²⁺]_i transient amplitude during fast application of caffeine (20 mm). While FBP (1 mm) did not cause any changes in SR Ca²⁺ load during the first minutes, a significant decrease in the amplitude of caffeine-induced [Ca²⁺]_i transient was observed after 6 min of FBP application (Fig. 2A). On average, the amplitude of the caffeine response in the presence of FBP was reduced to 87 ± 2% (n = 6; P < 0.05) of control (Fig. 2B).

Figure 2. Effect of fructose-1,6-bisphosphate (FBP) on SR Ca²⁺ load.

A, Ca²⁺ release induced by application of 20 mm caffeine under control conditions, after addition of 1 mm FBP (at 2 and 6 min) and following washout of FBP. B, average amplitudes of caffeine-induced [Ca²⁺]_i transients under control conditions, in the presence of FBP and after washout of FBP. *P < 0.05 versus control.

Another glycolytic intermediate, fructose-6-phosphate (F6P), caused qualitatively and quantitatively similar effects on SR Ca²⁺ release. F6P (1 mm) transiently increased Ca²⁺ spark frequency by 43 ± 13% (n = 6; P < 0.05; Fig. 3A and B). After the initial stimulatory phase, Ca²⁺ spark activity returned to the control level. During the later phase, the F6P effect was associated with decreases in spark amplitude and width (Fig. 3C and D). Upon washout of F6P, the frequency of Ca²⁺ sparks decreased to 54 ± 13% of control (n = 6; P < 0.01 versus control).

These results suggest that at a physiological cytosolic ATP/ADP ratio and CrP levels the glycolytic sugar phosphates FBP and F6P stimulate SR Ca²⁺ release through RyR channels, thereby leading to a reduction of SR Ca²⁺ load.

The effect of glycolytic sugar phosphates on Ca²⁺ sparks during inhibition of glycolysis

The stimulatory effect of the glycolytic intermediates FBP and F6P on SR Ca²⁺ release could be mediated indirectly through accumulation of glycolytic products (e.g. ATP) as a result of increased activities of downstream glycolytic enzymes. Therefore, we tested the effects of FBP on Ca²⁺ sparks in the presence of iodoacetate (IAA). IAA at submillimolar concentrations selectively inhibits glycolysis downstream of FBP production at the stage of the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) reaction (Carlson & Siger, 1959). Pretreatment of cells with IAA (0.5 mm) for 6 min did not significantly affect Ca²⁺ spark frequency (3.0 ± 0.5 sparks s⁻¹ (100 μm)⁻¹ in control versus 4.0 ± 0.7 sparks s⁻¹ (100 μm)⁻¹ in the presence of (IAA). Furthermore, inhibition of glycolysis by IAA did not prevent the stimulatory effects of FBP on SR Ca²⁺ release at the normal ATP/ADP ratio. In the presence of IAA, FBP (1 mm) increased Ca²⁺ spark frequency by 71 ± 12% (n = 7; P < 0.01). Figure 4A shows confocal images of Ca²⁺ sparks and selected plots of subcellular changes of F/F₀ in the presence of IAA (0.5 mm), 2 and 6 min after the subsequent addition of FBP (1 mm), and following washout of FBP. Similar to control conditions (Fig. 1), the effect of FBP on Ca²⁺ spark frequency in the presence of IAA reached a maximum within the first minutes, followed by recovery to near-control level (Fig. 4B). After 4 min of exposure to FBP the spark frequency remained 23 ± 13% above the initial control value. This degree of stimulation, however, was not significantly different from the effect of IAA alone. Furthermore, in the presence of IAA (0.5 mm), Ca²⁺ spark amplitude and width were reduced after 4 min exposure to FBP from ΔF/F₀ = 0.80 ± 0.03 to 0.71 ± 0.04 (P < 0.05) and from and 1.66 ± 0.06 to 1.44 ± 0.05 μm (P < 0.05), respectively. Spark duration remained unchanged (40.9 ± 1.5 versus 40.3 ± 1.8 ms, P = N.S.; n = 7 cells; number of sparks analysed in control conditions and in the presence of FBP was 538 and 445, respectively). Similar results were obtained with F6P (data not shown).

Therefore, the stimulatory effect of glycolytic sugar phosphate intermediates on SR Ca²⁺ release was not dependent on an intact glycolytic pathway, but rather appeared to be the result of a direct interaction with the RyR channel.

Effects of glycolytic sugar phosphates on single RyR channel activity

To test further the hypothesis that stimulation of SR Ca²⁺ release by sugar phosphates was caused by direct stimulation of the RyR, we studied the effects of sugar phosphates on the activity of isolated RyR channels. Heavy SR microsomes were fused with planar lipid bilayers and single channel currents through the RyR were recorded using Cs⁺ as the charge carrier. The channels were activated by 3 μm cytosolic [Ca²⁺] (no ADP, ATP and Mg²⁺). Figure 5A shows effects of various concentrations of F6P and FBP (added to the cis-chamber) on the activity of RyR channel currents recorded at −20 mV. Both sugar phosphates increased the activity of RyR channels in a concentration-dependent manner, acting on the cytosolic side of the channel. This effect was fully established in less than 1 min, stable throughout the recording (2–6 min) and completely reversible upon washout. Figure 5B summarizes the effects of various concentrations of three sugar phosphates on the open probability (P_o) of the RyR. The results indicate that FBP was most effective in stimulating RyR channel activity. P_o increased from 0.038 ± 0.016 under control conditions (cis[Ca²⁺] = 3 μm) to 0.132 ± 0.032 (n = 5) in the presence of 3 mm FBP, i.e. reflecting an activation by 340 ± 49% (n = 5; P < 0.01). The same concentration of F6P increased the activity (P_o) of the RyR from 0.024 ± 0.009 to 0.039 ± 0.011 (n = 4) or by 69 ± 11% (n = 4; P < 0.05), whereas glucose-6-phosphate (G6P; 3 mm) had no significant effect on the P_o (change from 0.033 ± 0.012 to 0.038 ± 0.016 or by 14 ± 12%; n = 4).

Figure 5. Effect of sugar phosphates on ryanodine receptor (RyR) single channel activity.

A, original single channel recordings of RyR activity before (Control) and after the addition of various concentrations of fructose-6-phosphate (F6P) and fructose-1,6-bisphosphate (FBP) to the cytosolic (cis) side of the channel. B, percentage changes of RyR open probability (P_o) induced by glucose-6-phosphate (G6P), F6P and FBP relative to control (P_o in the absence of sugar phosphates). C, RyR single channel recordings in the presence of cytosolic (cis) ATP (5 mm) and Mg²⁺ (5.75 mm; [Mg²⁺]_free = 1 mm). Control activity (left) and after addition of FBP (1 mm; right). All recordings were made at a holding potential of −20 mV and with 3 μmcis [Ca²⁺]. *P < 0.05; **P < 0.01.

We tested further whether this stimulatory effect of FBP was mediated by an interaction with the adenine nucleotide binding site(s) of the RyR. For these experiments, the solution in the cis-chamber contained 5 mm ATP and 5.75 mm Mg²⁺ (free [Mg²⁺] = 1 mm). These concentrations were chosen to activate fully the ATP binding sites of the RyR and to mimic the concentrations of these compounds used in experiments with permeabilized myocytes. Figure 5C shows the effect of 1 mm FBP on RyR channel activity in the presence of ATP and Mg²⁺. Under these conditions, FBP (1 mm) increased the P_o from 0.028 ± 0.010 to 0.090 ± 0.022 or by 307 ± 89% (n = 3), which was not significantly different from the stimulatory effect of FBP observed in the absence of ATP and Mg²⁺ (265 ± 60%; n = 9; Fig. 5B).

These results indicate that the glycolytic sugar phosphate intermediates F6P and FBP investigated here increased the activity of RyR channels, but with different degrees of efficiency. Furthermore, this direct stimulatory effect of the sugar phosphates on RyR-mediated SR Ca²⁺ release is largely independent of the cytosolic ATP level.

Effects of glycolytic sugar phosphates on Ca²⁺ sparks under low cellular energy reserves

Next, we tested how glycolytic sugar phosphates affect SR Ca²⁺ release and load under conditions of depleted cytosolic energy reserves. At the early stage of ischaemia, the intracellular concentration of ATP remains relatively constant, whereas CrP concentration falls dramatically and the level of ADP rises (for review see Allen & Orchard, 1987). In the first set of experiments, we studied effect of CrP and ADP on Ca²⁺ sparks in permeabilized ventricular myocytes separately. Withdrawal of CrP from the experimental solution ([ATP] and [ADP] were kept constant throughout the experiment) decreased Ca²⁺ spark frequency by 29 ± 5% of control (n = 6; P < 0.05; Fig. 6A and C). Removal of CrP caused a reduction of Ca²⁺ spark amplitude from ΔF/F₀ = 0.76 ± 0.03 to 0.60 ± 0.03 (P < 0.01; n = 6 cells; number of sparks analysed in control conditions and in the absence of CrP was 797 and 396, respectively; Fig. 6D). Increasing the cytosolic [ADP] caused qualitatively similar effects on SR Ca²⁺ release. Increasing [ADP] from 0.01 to 0.4 mm ([CrP] was kept constant at 10 mm) decreased Ca²⁺ spark frequency by 42 ± 7% (n = 7; P < 0.01; Fig. 6B and C), concomitant with a decrease in spark amplitude (Fig. 6D). In both cases, the inhibition of Ca²⁺ sparks required 7–8 min to reach a maximum.

Figure 6. Effects of CrP and ADP on Ca²⁺ sparks.

A, top: confocal linescan images of Ca²⁺ sparks under control conditions and 8 min after withdrawal of CrP (−CrP) from the experimental solution. Bottom: local F/F₀ profiles obtained by averaging fluo-3 fluorescence from 2 μm wide regions marked by the black boxes. B, top: confocal linescan images of Ca²⁺ sparks under control conditions and 8 min after increasing ADP from 0.01 to 0.4 mm (+ADP). Bottom: local F/F₀ profiles obtained by averaging fluo-3 fluorescence from 2 μm wide regions marked by the black boxes. Average effects of CrP removal and increase of [ADP] on Ca²⁺ spark frequency (C), and Ca²⁺ spark amplitude (D). The amplitude of Ca²⁺ sparks averaged over 5 min under control conditions and from 7 to 10 min after changing to [CrP] = 0 or [ADP] = 0.4 mm. **P < 0.01 versus control.

Since ischaemia is associated with changes of both [CrP] and [ADP], in the following experiments we studied the effects of FBP on Ca²⁺ sparks under conditions where the ATP/ADP ratio was lowered by increasing [ADP] from 0.01 to 0.4 mm and at the same time CrP was omitted from the experimental solution (termed ‘low energy reserve solution’). Figure 7A shows confocal images of Ca²⁺ sparks and selected plots of subcellular changes of F/F₀ at a normal ATP/ADP and CrP levels (mm: ATP 5; ADP 0.01, CrP 10), after switching to the low energy reserve solution, and 2 and 8 min after the subsequent addition of 1 mm FBP. Exposure of the cells to the experimental solution with depleted cytosolic energy reserves caused a slowly developing (∼8 min) decrease of Ca²⁺ spark frequency to 34 ± 7% of the initial control level (n = 7; P < 0.01). The inhibition of spark activity was accompanied by a decrease of Ca²⁺ spark amplitude from ΔF/F₀ = 0.80 ± 0.04 to 0.67 ± 0.03 (P < 0.01; n = 7 cells; number of sparks analysed in control conditions and in low energy reserve solution was 513 and 211, respectively; Fig. 7C). The subsequent addition of FBP (1 mm) led to a gradual recovery of Ca²⁺ spark activity. After 8 min of FBP application the spark frequency was restored to 84 ± 7% of the control level (Fig. 7B; filled circles). Furthermore, in the presence of FBP the spark amplitude increased to ΔF/F₀ = 0.71 ± 0.03 (n = 7 cells; 291 sparks analysed) which was significantly different (P < 0.05) from the amplitude recorded in low energy reserve solution alone.

Figure 7. Effect of fructose-1,6-bisphosphate (FBP) on Ca²⁺ sparks under conditions of low cytosolic energy reserves.

A, top: confocal linescan images of Ca²⁺ sparks in control conditions ([ATP] = 5 mm; [ADP] = 0.01 mm; [CrP] = 10 mm), in low cytosolic energy reserve solution ([ATP] = 5 mm; [ADP] = 0.4 mm; [CrP] = 0) and 2 and 8 min after subsequent addition of FBP (1 mm). Bottom: local F/F₀ profiles obtained by averaging fluo-3 fluorescence from 2 μm wide regions marked by black boxes. B, effect of FBP on Ca²⁺ spark frequency under conditions of low cytosolic energy reserves in the presence (○) and absence of iodoacetate (IAA; •). Effect of FBP on Ca²⁺ spark amplitude in low cytosolic energy reserve solution in the absence (C) and presence of IAA (D). The amplitude of Ca²⁺ sparks averaged over 2 min under control conditions, from 7 to 10 min after changing to the low cytosolic energy reserve solution and from 6 to 8 min following subsequent application of FBP. *P < 0.05; **P < 0.01 versus control.

The ability of FBP to maintain normal Ca²⁺ spark activity at low cellular energy reserves was dependent on an intact glycolytic pathway. Figure 7B presents average data of the effect of FBP (1 mm) on Ca²⁺ spark frequency at depleted cytosolic energy reserves in the presence (open circles) and absence (filled circles) of IAA. In contrast to the experiments conducted in the presence of an intact glycolytic pathway, FBP failed to restore Ca²⁺ spark frequency (Fig. 7B) and amplitude (Fig. 7D) when glycolysis was inhibited by 0.5 mm IAA. After 8 min exposure to low energy reserve solution Ca²⁺ spark frequency decreased to 30 ± 11% of control (n = 4; P < 0.01) and remained unaltered after the subsequent addition of FBP (22 ± 12%; n = 4). Initially FBP caused a small and transient increase of Ca²⁺ spark frequency (from 30 ± 11 to 39 ± 10% after 3 min; n = 4); however, this effect was not statistically significant. Ca²⁺ spark amplitude decreased from ΔF/F₀ = 0.83 ± 0.01 to 0.61 ± 0.01 in low energy reserve solution (n = 4 cells; number of sparks analysed in control conditions and in the presence of the low energy reserve solution was 217 and 93, respectively). Addition of FBP caused a further decrease to ΔF/F₀ = 0.51 ± 0.04 (61 sparks analysed).

The changes in Ca²⁺ spark frequency and amplitude observed with depleted cytosolic energy reserves could be the result of a decrease in SR Ca²⁺ content. Figure 8A shows a representative experiment in which caffeine (20 mm) was applied under control conditions, in low energy reserve solution and at 2 and 8 min after the subsequent addition of FBP (1 mm). Average results from six myocytes are presented in Fig. 8B. Reduction of the ATP/ADP ratio and withdrawal of CrP decreased the amplitude of caffeine-induced [Ca²⁺]_i transient to 62 ± 7% (n = 6; P < 0.01) of control. After the first 2 min, the subsequent exposure of FBP (1 mm) did not affect SR Ca²⁺ load. After 8 min, however, FBP caused a recovery of SR Ca²⁺ loading to 88 ± 7% of control (n = 6; P < 0.05 versus low energy reserve solution). Similar to the Ca²⁺ spark experiments, this effect of FBP on SR Ca²⁺ load in the presence of the solution with depleted cytosolic energy reserves was sensitive to IAA, i.e. in the presence of IAA FBP failed to rescue SR Ca²⁺ load (Fig. 8C). SR load under control conditions with and without IAA present was not significantly different.

Figure 8. Effect of fructose-1,6-bisphosphate (FBP) on SR Ca²⁺ load under conditions of low cytosolic energy reserves in the presence and absence of IAA.

A, Ca²⁺ release induced by application of 20 mm caffeine under control conditions ([ATP] = 5 mm; [ADP] = 0.01 mm; [CrP] = 10 mm), in low cytosolic energy reserve solution ([ATP] = 5 mm; [ADP] = 0.4 mm; [CrP] = 0) and after the subsequent addition of 1 mm FBP. B, average amplitudes of caffeine-induced [Ca²⁺]_i transients under control conditions (Ctrl), at conditions of low cytosolic energy reserves, and in the additional presence of FBP (2 and 8 min). *P < 0.05; **P < 0.01. C, same experiment as in A, but in the presence of IAA (1 mm).

These results indicate that in permeabilized ventricular myocytes a reduction of the ATP/ADP ratio and a concomitant withdrawal of CrP leads to inhibition of Ca²⁺ release from the SR. This effect might be mediated by depression of the activity of the SR Ca²⁺-ATPase with subsequent depletion of the SR Ca²⁺ content. The addition of glycolytic sugar phosphate (FBP) partially restores SR Ca²⁺ release. The likely mechanism underlying this effect is a stimulation of glycolytic ATP production (re-phosphorylation of ADP) and thus an increase in the cytosolic ATP/ADP ratio. This will improve SR Ca²⁺ uptake and normalize SR Ca²⁺ load and release.

Discussion

In the heart, Ca²⁺ release from the SR is affected by a number of cytosolic factors, including Ca²⁺, Mg²⁺, ATP and pH. In the present work we have shown that sugar phosphates metabolized during glycolysis also modulate local SR Ca²⁺ release in cat ventricular myocytes. The effect of sugar phosphates on SR Ca²⁺ release was mediated by altering the activity of two major Ca²⁺ transport systems of the SR, the RyR and the SR Ca²⁺-ATPase. First, glycolytic sugar phosphates directly stimulated RyR channel activity and, thereby, caused a transient increase of Ca²⁺ release from the SR. Second, energy accumulated in phosphate bonds of sugar phosphates could be used by glycolysis for ATP production. High ATP/ADP levels, in turn, can improve SR Ca²⁺ uptake and increase SR Ca²⁺ turnover. The relative contribution of these two mechanisms, however, is highly dependent on the cellular energy status.

Direct effect of glycolytic sugar phosphates on RyR channel activity

It seems likely that under normal aerobic conditions when cellular energy levels are high, direct stimulation of RyR channels would be the predominant mechanism of modulation of SR Ca²⁺ release by glycolytic intermediates. We found that at a physiological ATP/ADP ratio and CrP content, two fructose phosphates, FBP and F6P, caused a transient increase in Ca²⁺ spark frequency, which was accompanied by decreases in amplitude and spatial width of Ca²⁺ sparks (Figs 1 and 3). Potentially, decreases in amplitude and width can be the result of smaller Ca²⁺ efflux from the SR due to smaller SR Ca²⁺ content. Experiments with caffeine confirmed that SR Ca²⁺ load eventually decreased after FBP application (Fig. 2). Therefore, despite the fact that glycolytic sugar phosphates are substrates for glycolytic ATP production and can potentially improve the efficiency of the SR Ca²⁺ATPase, at a physiological ATP/ADP ratio SR Ca²⁺ leak through the RyR channels seems to be more sensitive to sugar phosphates than SR Ca²⁺ uptake. Experiments with an inhibitor of glycolysis (IAA; Fig. 4) indicated that this stimulatory effect of sugar phosphates on SR Ca²⁺ release was not dependent on an intact glycolytic pathway (and therefore ATP production), but rather the result of a direct stimulation of RyR channels. This conclusion was drawn from the following observations. First, the transient stimulation of Ca²⁺ sparks by sugar phosphates is reminiscent of the effect of low doses of the RyR agonist caffeine (Lukyanenko et al. 2001). Sensitization of RyRs with caffeine results in brief increases in Ca²⁺ spark frequency with consequent decreases in SR Ca²⁺ content. Presumably, the high Ca²⁺ leak would gradually deplete SR Ca²⁺ stores and subsequently reduce RyR activity from the luminal side of the channel by a Ca²⁺-dependent mechanism (Eisner et al. 1998). Second, single-channel experiments revealed that glycolytic sugar phosphate intermediates stimulated the activity of cardiac RyR channels incorporated into lipid bilayers (Fig. 5). Similar effects of sugar phosphates were previously described for sheep cardiac RyR channels (Kermode et al. 1998a). From all glycolytic intermediates studied, FBP was most effective in stimulating the RyR. We found, for example, that FBP was about five times more potent in stimulating RyR channel activity than F6P (+340% by FBP versus +70% by F6P). The large difference between FBP and F6P effects on RyR channel activity observed in lipid bilayer experiments appears to be at variance with the almost identical effects of these compounds on Ca²⁺ sparks in permeabilized myocytes. In these experiments, FBP and F6P caused very similar increases in Ca²⁺ spark frequency of 60 and 40%, respectively. This apparent inconsistency could be explained by the fact that in permeabilized cells the glycolytic enzyme phosphofructokinase can metabolize F6P in the presence of all required cofactors leading to accumulation of FBP and therefore comparable effects of F6P and FBP on Ca²⁺ sparks.

Studies of isolated RyR channels also revealed that the stimulatory effect of FBP was preserved when the RyR was fully activated by cytosolic ATP. This finding indicates that an interaction of sugar phosphates with the RyR involves sites different from the adenine nucleotide binding site. It is not clear yet, however, whether sugar phosphates interact with other known binding sites or with a novel site on the RyR or a closely associated accessory protein. It has been demonstrated that inorganic phosphate (P_i) binds to a distinct domain on the cardiac RyR channel to produce an increase in P_o (Kermode et al. 1998b). Since all glycolytic intermediates studied here contain at least one phosphate group, it seems possible that they can bind to the same site as P_i. Even though the site and molecular mechanism of fructose phosphate action remain to be determined, the present results clearly suggest that the stimulatory effect of glycolytic sugar phosphates on RyR channels per se is independent of the metabolic state of the cell.

Glycolytic sugar phosphates improve SR Ca²⁺ uptake under conditions of low cytosolic energy reserves

A different situation might arise at low cytosolic ATP/ADP ratio and reduced CrP levels, as may occur during periods of ischaemia/hypoxia. During the first minutes of ischaemia the intracellular [ATP] remains relatively constant, while [ADP] rises and [CrP] falls (Allen & Orchard, 1987). We found that both of these factors affect SR Ca²⁺ release. At constant [ATP], the increase of [ADP] or the removal of CrP caused a slowly developing decrease of Ca²⁺ spark frequency, which was accompanied by a significant decrease in spark amplitude (Fig. 6). Similar effects on Ca²⁺ spark activity by CrP were previously observed in rat cardiomyocytes (Yang & Steele, 2002). It seems likely that the effects of changing ATP/ADP ratio and [CrP] were mediated by the same mechanism. As intracellularly the ATP/ADP ratio is effectively controlled by the CrP level through the creatine kinase reaction, CrP removal enhances local ADP accumulation. The simultaneous increase of [ADP] and removal of CrP, which is observed during ischaemia, caused more significant (and nearly additive) inhibition of Ca²⁺ sparks compared with the effect of each factor alone (Figs 6 and 7). The slow onset of this inhibition together with the significant depletion of SR Ca²⁺ load is indicative of a reduction of Ca²⁺ uptake by the SR Ca²⁺-ATPase. This observation is consistent with previous work showing that SR Ca²⁺ release and load are critically dependent on the ratio of [ATP] to [ADP] (Yang & Steele, 2002; Zima et al. 2003). The cytosolic ATP/ADP ratio is known to affect Ca²⁺-ATPase activity not only through the free energy accumulated in the phosphate bonds of ATP, but also through accumulation of ADP. Elevation of ADP markedly reduces the ability of the SR to store Ca²⁺ by decreasing the rate of the SR Ca²⁺ pump (Macdonald & Stephenson, 2001). It seems possible that increases in cytosolic [ADP] could also directly affect SR Ca²⁺ release. It has been shown that cytosolic ADP can potentiate Ca²⁺-induced Ca²⁺ release in rat ventricular trabeculae (Xiang & Kentish, 1995). Single-channel studies revealed that ADP can act as a partial agonist at the adenine nucleotide binding site of the cardiac RyR (Kermode et al. 1998b). However, when the adenine nucleotide binding site of the RyR was fully activated by the more specific agonist ATP (like in our study where we used 5 mm ATP), ADP could decrease RyR activity by antagonizing the effect of ATP (Kermode et al. 1998b). Therefore, the inhibition of Ca²⁺ spark frequency by high [ADP] (Fig. 7) could be due to a direct inhibition of SR Ca²⁺ release through the RyR. It has been shown, however, that interventions that inhibit RyR, typically cause only a transient inhibition of SR Ca²⁺ release (Eisner et al. 1998; Lukyanenko et al. 2001). This can be explained by a concomitant increase in SR Ca²⁺ content with subsequent Ca²⁺-dependent activation of RyR from the luminal side of the channel. In contrast to RyR inhibitors, in our study high [ADP] caused a slowly developing, however, non-transient decrease in Ca²⁺ spark frequency which was accompanied by a reduction in SR Ca²⁺ load. Thus, the inhibitory effect of high cytosolic [ADP] on Ca²⁺ sparks could not be explained solely based on the depression of RyR activity. Therefore, the observed decrease in Ca²⁺ spark frequency and amplitude with low cytosolic energy reserves are likely to be secondary to a decrease in the SR Ca²⁺-ATPase activity and depletion in SR Ca²⁺ content.

Under conditions of depleted cytosolic energy reserves glycolytic sugar phosphates caused a completely different type of modulation of SR Ca²⁺ release than at a normal physiological ATP/ADP ratio and CrP levels. First, we did not observe any stimulation of SR Ca²⁺ release within the first minutes of FBP application (Fig. 7B, filled symbols). As mentioned before, stimulation of RyR channel activity by sugar phosphates was not mediated by an interaction with the adenine nucleotide binding site of the RyR. Therefore, increased cytosolic [ADP] will not significantly affect the ability of FBP to stimulate the RyR channel. The lack of the initial stimulation of Ca²⁺ spark frequency by FBP under conditions of low cytosolic energy reserves can be explained by a significant depletion of SR Ca²⁺ content (Fig. 8) and the inability of sugar phosphates to increase Ca²⁺ spark frequency via stimulation of the RyR after depletion of SR Ca²⁺ load. In contrast to control conditions, at a low cytosolic ATP/ADP ratio and low [CrP], FBP caused a slowly developing increase in Ca²⁺ spark frequency (Fig. 7). We suggest that the slow time course reflects the gradual accumulation of Ca²⁺ in the SR due to stimulation of the SR Ca²⁺-ATPase. It is well accepted that increased SR Ca²⁺ load can potentiate SR Ca²⁺ release via luminal Ca²⁺-dependent activation of RyRs (Eisner et al. 1998; Lukyanenko et al. 2001). Indeed, we found that augmentation of Ca²⁺ spark frequency was associated with higher SR Ca²⁺ content (Fig. 8). Experiments with IAA indicated that these effects of sugar phosphates on SR Ca²⁺ release and load were dependent on an intact glycolytic pathway. Therefore, at low cytosolic energy reserves sugar phosphates can be utilized by glycolytic enzymes to increase ATP production. Considering that glycolytic enzymes are associated with the SR membrane (Pierce & Philipson, 1985; Xu & Becker, 1998), stimulation of the glycolytic pathway by providing substrate (sugar phosphates) for ATP synthesis will maintain a high ATP/ADP ratio in close proximity to the SR Ca²⁺-ATPase. A locally high ATP/ADP ratio will normalize the activity of the SR Ca²⁺ pump and, thereby, eventually increase SR Ca²⁺ load and release.

Physiological and pathophysiological significance

In the normally oxygenated heart, glycolysis is partially inhibited at the stage of phosphofructokinase (PFK) by high levels of ATP and citrate (Opie, 1991) and does not work at full capacity. As a result of PFK inhibition, levels of FBP (product of PFK reaction) are relatively low under aerobic conditions. The concentration of FBP in the working heart ranges from ∼8 to 60 nmol (g wet weight)⁻¹ (Depre et al. 1993). Assuming that intracellular water space amounts to ∼50% of wet weight, this translates into ∼16–120 μm FBP. These values are in the lower range of concentrations of FBP that can stimulate the cardiac RyR. Glycolytic enzymes are associated with the SR membrane (Pierce & Philipson, 1985) and, hence, are in close proximity to the RyR. Furthermore, the junctional space between T-tubular membrane and terminal SR represents a microcompartment within the cardiac myocyte characterized by restricted diffusion. Therefore, the local concentration of FBP in the junctional space near the RyR is expected to be much higher than the estimates for whole heart tissue mentioned above. Thus, it is likely that concentrations of FBP (and possibly other glycolytic intermediates) are sufficiently high to regulate RyR channel activity and SR Ca²⁺ release under physiological conditions.

Depending on factors such as workload, extracellular glucose concentration and insulin, concentrations of FBP can vary by a factor of ∼7 (Depre et al. 1993). For F6P, these variations can be even higher, i.e. up to a factor of ∼45 (Depre et al. 1993). Interestingly, oscillations of glycolytic intermediates have been observed even during the cardiac cycle (Wikman-Coffelt et al. 1983). Furthermore, glycolysis is enhanced by ischaemia/hypoxia or adrenergic stimulation (Depre et al. 1998). Mild ischaemia significantly enhances glycolysis by stimulation of PFK activity. Glycolytic ATP production is especially important to maintain intracellular Ca²⁺ homeostasis during ischaemia (Jeremy et al. 1992). Severe ischaemia, however, is associated with accumulation of intracellular lactate, NADH and protons. All these factors inhibit GAPDH (Opie, 1991) and markedly increase FBP levels (substrate for GAPDH reaction). In our study we found that FBP was more effective in stimulating the RyR than other sugar phosphates. Therefore, we suggest that conditions of severe ischaemia can potentially lead to augmentation of SR Ca²⁺ leak as a result of stimulation of RyR by glycolytic sugar phosphates. It has been shown that accumulation of sugar phosphates due to inhibition of glycolysis caused intracellular Ca²⁺ overload (Kusuoka & Marban, 1994). Together with a low rate of ATP production, ischaemia would eventually deplete SR Ca²⁺ content and depress cardiac performance. Thus, FBP and other glycolytic intermediates are important physiologically relevant regulators of RyR-mediated SR Ca²⁺ release, and therefore cardiac EC coupling.