Calcium waves induced by hypertonic solutions in intact frog skeletal muscle fibres

Abstract

1. Regenerative Ca²⁺ waves and oscillations indicative of calcium-induced calcium release (CICR) activity were induced in fully polarized, fluo-3-loaded, intact frog skeletal muscle fibres by exposure to hypertonic Ringer solutions.

2. The calcium waves persisted in fibres exposed to EGTA-containing solutions, during sustained depolarization of the membrane potential or following treatment with the dihydropyridine receptor (DHPR)-blocker nifedipine.

3. The waves were blocked by the ryanodine receptor (RyR)-specific agents ryanodine and tetracaine, and potentiated by caffeine.

4. In addition to these pharmacological properties, the amplitudes, frequency and velocity of such hypertonicity-induced waves closely resembled those of Ca²⁺ waves previously described in dyspedic skeletal myocytes expressing the cardiac RyR-2.

5. Quantitative transmission and freeze-fracture electronmicroscopy demonstrated a reversible cell shrinkage, transverse (T)-tubular luminal swelling and decreased T-sarcoplasmic reticular (SR) junctional gaps in fibres maintained in and then fixed using hypertonic solutions.

6. The findings are consistent with a hypothesis in which RyR-Ca²⁺ release channels can be partially liberated from their normal control by T-tubular DHPR-voltage sensors in hypertonic solutions, thereby permitting CICR to operate even in such fully polarized skeletal muscle fibres.

Ryanodine receptor (RyR)-sarcoplasmic reticular (SR) Ca²⁺-release channels in striated muscle are large homotetrameric channels that consist of 560 kDa polypeptides (Meissner, 1994; Franzini-Armstrong & Protasi, 1997). RyRs are activated to release the intracellularly stored Ca²⁺ that triggers muscle contraction by T-tubular voltage sensor-dihydropyridine receptors (DHPRs), which are voltage-gated L-type Ca²⁺ channels (Rios & Brum, 1987; Huang, 1990). Skeletal and cardiac muscles express different RyR subtypes controlled by their respective DHPRs by different mechanisms. The skeletal muscle RyR-1 is thought to be coupled to its cognate DHPR by direct mechanical gating independent of extracellular Ca²⁺ (Schneider & Chandler, 1973; Nakai et al. 1996; Huang, 1996). In contrast, cardiac RyR-2 gating is triggered by Ca²⁺ influx through cardiac DHPRs in response to electrical excitation by Ca²⁺-induced Ca²⁺ release (CICR) (Fabiato, 1985; Cannell et al. 1995).

The prevailing type of excitation-contraction (E-C) coupling in a cell critically depends on both the DHPR and RyR types expressed. The pore-forming α-subunits of skeletal and cardiac DHPRs have four homologous transmembrane domains but different cytoplasmic regions. Both E-C coupling and the slow L-type Ca²⁺ current characteristic of skeletal muscle can be recovered by expression of the skeletal DHPR α-subunit in dysgenic murine skeletal myotubes (Tanabe et al. 1988). In contrast, cardiac DHPR expression in such dysgenic myotubes supports cardiac-type E-C coupling; this depends on extracellular Ca²⁺ and is associated with the fast, large amplitude, L-type current characteristic of the cardiac DHPR (Tanabe et al. 1990b). Indeed a chimaeric cardiac DHPR containing the cytoplasmic loops of the skeletal DHPR produced a skeletal muscle-type E-C coupling, independent of extracellular Ca²⁺ and a large L-type Ca²⁺ current (Tanabe et al. 1990a).

Conversely, work on dyspedic skeletal myotubes lacking RyR-1 similarly demonstrates that the RyR subtype also determines the nature of E-C coupling. Skeletal myotubes from RyR-1-deficient mice lack E-C coupling (Takeshima et al. 1994). This is recovered by RyR-1 but not RyR-2 expression (Yamazawa et al. 1996; Nakai et al. 1997); the latter instead results in spontaneous Ca²⁺ oscillations accompanied by Ca²⁺ waves similar to those observed in Ca²⁺-overloaded cardiac myocytes (Lipp & Niggli, 1993; Cheng et al. 1996). Such spontaneous Ca²⁺ oscillations were observed only in RyR-2-transfected myotubes and never in RyR-1- or RyR-3-transfected cells (Yamazawa et al. 1996). This may reflect the greater sensitivity of RyR-2 to CICR: myotubes expressing RyR-1 similarly exhibited Ca²⁺ waves following potentiation of RyR-1 sensitivity to CICR by caffeine (Yamazawa et al. 1996; Fessenden et al. 2000).

Amphibian skeletal muscle cells expresse two RyR isoforms, α and β, homologous to RyR-1 and RyR-3, respectively (Franzini-Armstrong & Protasi, 1997). Here we show that such cells nevertheless generate Ca²⁺ waves following treatment with hypertonic solutions, even at fully polarized membrane potentials. We describe their characteristics as well as their responses to pharmacological RyR and DHPR blockers and show that the waves resemble those seen in RyR-2-expressing dyspedic myotubes (Yamazawa et al. 1996). Our anatomical studies provide a possible explanation for such phenomena in terms of a physical uncoupling of at least some of the RyRs present in skeletal muscle from their corresponding DHPRs permitting the appearance of a CICR gating mode (Huang, 1998).

METHODS

Fluorescence measurements

Lumbricalis V muscles were dissected from cold-adapted Rana temporaria killed by concussion followed by pithing (Schedule I: Animal Procedures Act, Home Office, UK). Intact muscles were loaded with the acetoxymethyl (AM) ester of fluo-3 (Molecular Probes) by incubating with 5 μm fluo-3 AM in frog Ringer solution (composition (mm): 115 NaCl, 2.5 KCl, 1.8 CaCl₂, 3.0 Hepes, pH 7.0) at 25 °C for 25 min. Muscles were subsequently washed in Ringer solution and mounted close to their resting lengths (sarcomere length 1.9–2.0 μm) between two glass coverslips held together with vacuum grease in a 300 μl perfusion chamber. Muscles were cooled by perfusing the chamber with cold Ringer solutions.

Fluo-3 fluorescence emission, excited with a 488 nm argon laser, was measured at 505–550 nm using a Zeiss SM-510 or a Leica TCS-SP-MP laser scanning confocal system with either a × 20 air objective (NA 0.5; confocal aperture 1000 μm, slice thickness < 42.4 μm) on a Zeiss Axiovert 100M inverted microscope or a × 20 water immersion objective (NA 0.7; confocal aperture 600 μm) on a Leica DMIRBE inverted microscope. Measurements were made from fibres close to the muscle edge that showed optimal dye loading using frame scans before, during and after perfusion with cooled (4–10 °C) hypertonic solutions. A series of 128 frames sampled at 786 ms per frame (512 × 512 pixels per frame) or a series of 256 frames sampled at 393 ms per frame (256 × 256 pixels per frame) monitored fluorescence changes over time. The confocal aperture was set to its maximum. Images were analysed using a modified version of the public domain NIH Image program (National Institutes of Health, Bethesda, MD, USA). Fluorescence measurements made within a 5 × 5 pixel region of interest (F) were normalized to resting fluorescence (F₀) values. For each of the n fibres studied, we calculated the frequency and peak F/F₀ values of the events through the series of acquired frames. We then averaged results through the different fibres. Line scan images (typically 500 lines × 512 pixels) were sampled at a frequency of either 30 or 3 ms per line.

Electrophysiological study

Standard 3 m KCl-filled glass capillary microelectrodes (resistance 8–20 MΩ, tip potentials less than 5 mV) separately measured resting potentials in fibres from lumbricalis muscles pinned onto a Sylgard-bottomed dish using standard electrophysiological techniques. Microelectrode penetrations were performed on successive adjacent surface fibres before, and following the replacement of isotonic solutions with cooled Ringer solutions to which different concentrations of sucrose were added. The resting potentials were read off deflections in electrical traces displayed on the screen of a calibrated digital oscilloscope (Model HM205–3, Hameg, Germany).

Electron microscopy

Sartorius muscles were dissected in isotonic Ringer solution, maintained at constant length and divided into three experimental groups for: (1) fixation by immersion in isotonic solution (250 mosmol l⁻¹) containing 0.05 m Pipes buffer at pH 7.4, 3 % glutaraldehyde and 2 mm CaCl₂; (2) incubation in 350 mm sucrose-Ringer solution (650 mosmol l⁻¹) for 30 min before fixation as before but with 0.1 m Pipes buffer and 10.2 % sucrose; (3) exposure to 350 mm sucrose-Ringer solution for 30 min and return to isotonic Ringer solution for 30 min before fixation as in (1). H₂O₂ (33 %; 100 μl (10 ml)⁻¹) was added to the fixative as an oxygen donor (Peracchia & Mittler, 1972). The muscles were osmicated, bulk stained and embedded in Spurr's resin. Transverse sections were used to estimate fibre cross-sectional area. Isotropic random planes of section (Mattfeldt et al. 1990) were used to estimate T-tubule diameter and the volume fraction of junctional SR (JSR). The barrier thickness between the JSR and the T-tubules was estimated from orthogonal intercepts (Jensen et al. 1979). Other muscles similarly treated and fixed were prepared for freeze fracture replication (Skepper, 1989).

RESULTS

Ca²⁺ oscillations and waves in skeletal muscle fibres in hypertonic solutions

Frog skeletal muscle fibres treated with extracellular Ringer solutions made hypertonic with sucrose reproducibly displayed both Ca²⁺ oscillations and Ca²⁺ waves. Figure 1A displays successive frames that together represent propagation of a regenerative Ca²⁺ wave from right to left in a frog lumbricalis muscle fibre treated with 350 mm sucrose-Ringer solution. The successive confocal images were obtained every 786 ms and run from top to bottom. The total distance traversed by such Ca²⁺ waves varied between individual muscle fibres. Such waves often were initiated repetitively, frequently at preferred sites along the length of the muscle, and could also be observed at higher tonicities (500 and 800 mm sucrose-Ringer solution). Figure 1B demonstrates Ca²⁺ oscillations derived from F/F₀ measurements within a 5 × 5 pixel region of interest in the centre of the region indicated by the black box in the top panel of A. These had an average frequency of 6.8 ± 0.3 min⁻¹ and produced a peak F/F₀ of 1.97 ± 0.08 (mean ±s.e.m.; n = 29 fibres). The individual events showed rise times around 0.8 s, and more prolonged decays lasting over 4.8 s. Figure 1C displays a typical line scan image (30 ms per line) from which the average wave propagation velocity of 34.0 ± 3.5 μm s⁻¹ (n = 14 fibres) was deduced. In most fibres the waves propagated with a relatively constant conduction velocity from their sites of origin, and died away around 100–400 μm away from such initiation sites. Some fibres showed non-uniform wave propagation. Often, two waves initiated from the same site travelled in opposite directions along the long axis of muscle fibres. Conversely, annihilations of events were also seen where waves approached each other and then collided within the same fibres.

Figure 1. Ca²⁺ oscillations and Ca²⁺ waves in lumbricalis muscle in 350 mm sucrose-Ringer solution.

A, series of frames representing propagation of a regenerative Ca²⁺ wave from right to left obtained at 0.786 s intervals. False colour scale from 0 (minimum) to 256 (maximum). B, F/F₀ measurements of Ca²⁺ oscillations from a 5 × 5 pixel region of interest in the centre of the region indicated by the black box in the top panel of A. C, typical higher resolution linescan image (30 ms per line) from which a constant wave propagation velocity can be deduced.

These Ca²⁺ signalling phenomena were observed for up to 40 min following introduction of the hypertonic solutions. The waves ceased within a few seconds of returning the muscle to isotonic solutions but were re-initiated by restoring 350 mm sucrose in the extracellular solution within this interval. Related events were observed at lower external sucrose concentrations which produced a variety of non-propagating Ca²⁺-release events. Figure 2A shows a muscle fibre treated with 200 mm sucrose-Ringer solution. Such fibres developed discrete foci of elevated cytosolic [Ca²⁺] between 5 and 10 min following addition of sucrose but these often failed to propagate as waves. High resolution line scan images (3 ms per line) of such fibres also revealed sparks or macrosparks within and near such regions of increased [Ca²⁺] (Fig. 2B; cf. Klein et al. 1996).

Figure 2. Spontaneous non-propagating Ca²⁺ release events.

A, discrete but often stationary foci of increased cytosolic [Ca²⁺] following treatment with 200 mm sucrose-Ringer solution followed through frame scans separated by 0.786 s time intervals. B, high resolution line scan image (3 ms per line) demonstrating sparks or macrosparks within and near such foci of increased [Ca²⁺].

Ca²⁺ waves are independent of extracellular Ca²⁺

The occurrence of both the Ca²⁺ waves and oscillations induced by addition of 350 mm sucrose-Ringer solution was independent of extracellular Ca²⁺. Thus adding 3 mm EGTA to the 350 mm sucrose-Ringer solution conserved the Ca²⁺-release events once initiated; these persisted with a frequency of 5.25 ± 0.80 min⁻¹ and peak F/F₀ of 2.09 ± 0.05 (n = 8 fibres). Similarly, fibres for which isotonic EGTA-containing Ringer solution was then replaced by EGTA-containing sucrose-Ringer solution continued to initiate Ca²⁺ waves. The waves also persisted with a peak F/F₀ of 1.45 ± 0.03 and frequency of 5.04 ± 1.04 min⁻¹ (n = 5 fibres) following addition of 10 mm Co²⁺, known to block voltage-gated Ca²⁺ channels (McDonald et al. 1994), to the hypertonic extracellular solutions.

Ca²⁺ waves occur in the absence of surface membrane potential change

The gating of amphibian RyRs, similar to mammalian muscle, is normally controlled by transitions in the DHPR driven by voltage change (Schneider & Chandler, 1973; Huang, 1990, 1994, 1997). However, the Ca²⁺-release events shown here were also indifferent to the presence of even relatively high concentrations (10 μm) of extracellular tetrodotoxin (peak F/F₀ 1.8 ± 0.1; frequency 4.4 ± 1.0 min⁻¹; n = 9 fibres) or the replacement of extracellular Na⁺ by equimolar choline⁺. Furthermore, the stable resting membrane potentials in fibres in hypertonic solutions (−83.3 ± 2.0 mV; n = 12 fibres; least negative recorded value −70 mV, in 350 mm sucrose-Ringer solution: −86.0 ± 1.5 mV; n = 6; least negative recorded value −80 mV, in 500 mm sucrose-Ringer solution) were not significantly different from those of fibres in normal isotonic Ringer solution (−89.3 ± 1.2 mV; n = 9 fibres; least negative recorded value −80 mV) and closely agreed with earlier findings from fibres exposed to similar solutions (Gordon & Godt, 1970; Parker & Zhu, 1987). This excluded spontaneous or hypertonicity-induced action potential firing as a possible cause of the signalling phenomena described here.

Ca²⁺ waves persist despite DHPR-voltage sensor inactivation

Figure 3 illustrates further experiments that ruled out voltage-dependent transitions within the DHPR-voltage sensor as the primary cause of the Ca²⁺ signalling phenomena. Figure 3A plots the development of a typical series of Ca²⁺ waves following addition of sucrose before (left) and after (right) introduction of 30 mm external KCl. Independent electrophysiological measurements indicated that this significantly altered fibre resting potential (−39 ± 1.5 mV; n = 5 fibres: least negative recorded value −36 mV at 30 mm KCl); this in turn should inactivate the DHPR-voltage sensor and therefore the E-C-coupling processes that normally lead to Ca²⁺ release or tension generation (Adrian & Peres, 1979; Huang, 1994). Figure 3A (right) demonstrates that 30 mm KCl initially increased background cytosolic [Ca²⁺] but this then declined back to the original baseline, as expected for a sustained release of intracellularly stored Ca²⁺ followed by its eventual inactivation. The latter finding is consistent with an activation, followed by inactivation of those Ca²⁺-release processes still under voltage control; certainly, larger depolarizations brought about by higher (75 mm) KCl concentrations (−25.4 ± 1.8 mV; n = 5: least negative recorded value −21 mV at 75 mm KCl) evoked larger Ca²⁺ changes. However, the repetitive Ca²⁺ waves remained superimposed upon these prolonged changes with peak F/F₀ deflections (1.73 ± 0.26) and frequencies (10.80 ± 1.38 min⁻¹; n = 3 fibres) similar to those shown by controls.

Figure 3. Ca²⁺ release events persist despite voltage sensor inactivation.

A, development of a typical series of Ca²⁺ waves following addition of sucrose before (left) and following (right) addition of 30 mm extracellular KCl. B, sustained pattern of oscillating cytosolic [Ca²⁺] both before (left) and after (right) addition of 30 μm nifedipine.

Additionally, application of the DHPR antagonist nifedipine at concentrations (30 μm) well in excess of levels (10 μm) known to achieve maximal inhibition of transitions in the voltage sensor in intact fibres (Huang, 1990) spared both oscillations and waves (Fig. 3B). If anything, addition of the antagonist increased their frequency (peak F/F₀ 2.20 ± 0.22; frequency 10.0 ± 2.0 min⁻¹; n = 7 fibres). Nimodipine (20 μm) gave similar results (data not shown).

RyR blocking agents abolish Ca²⁺ waves

The regenerative Ca²⁺ changes described above are consistent with a CICR process through RyR-Ca²⁺ release channels autonomous of the surface membrane potential and prompted investigations of the effects of the known RyR-specific inhibitors tetracaine and ryanodine. Skeletal muscle may operate both voltage-induced and Ca²⁺-induced Ca²⁺-release processes (Jacquemond et al. 1991) that are completely blocked by local anaesthetics at millimolar concentrations (Bull & Marengo, 1993; Huang, 1997). Figure 4A demonstrates that 2 mm tetracaine completely blocked an established series of Ca²⁺ oscillations despite relatively small changes in background fluorescence.

Figure 4. RyR antagonists block Ca²⁺ oscillations.

A, typical, complete, block of Ca²⁺ oscillations (left) following addition of 2 mm tetracaine (right). B, hypertonicity-induced Ca²⁺ waves progressively blocked following introduction of ryanodine (100 μm)-containing hypertonic solution (bar over trace) as monitored by Sulphorhodamine B fluorescence.

Ryanodine, particularly at higher (0.1–500 μm) concentrations, inactivates Ca²⁺ fluxes through junctional SR as well as isolated RyR in artificial membranes (Rouseau et al. 1987). Figure 4B demonstrates the results of an experiment that first induced Ca²⁺ waves by introducing hypertonic extracellular solutions. Ryanodine (100 μm)-containing hypertonic solution was then added during this response. The bar over the trace indicates its arrival in the bathing solution as monitored by fluorescence from Sulphorhodamine B added to the perfusing solution. Figure 4B shows a rising followed by a falling background cytosolic [Ca²⁺] consistent with an initial opening followed by block of RyR-Ca²⁺ release channels (cf. Rouseau et al. 1987). This accompanied complete inactivation of the Ca²⁺ oscillations. In contrast, addition of 100 μm (2-aminoethoxy)diphenylborane was without effect (data not shown). This compound inhibits the inositol-trisphosphate receptor (Maruyama et al. 1997); it may also be called 2-APB or 2-aminoethyl diphenylborinate (and has misleadingly been called 2-aminoethoxydiphenylborate).

Caffeine potentiates hypertonicity-induced Ca²⁺ waves

Figure 5A demonstrates that the CICR potentiator caffeine (e.g. Jacquemond et al. 1991) applied at 4 mm also induced Ca²⁺ waves and oscillations in fibres in isotonic Ringer solution; these otherwise did not show such phenomena. Their frequency (5.80 ± 1.34 min⁻¹), peak deflections (1.490 ± 0.027; n = 6 fibres) and propagation velocity (50.0 ± 1.2 μm s⁻¹; n = 3 fibres) were close to those observed in 350 mm sucrose-Ringer solution. Lower caffeine concentrations (2–3 mm) induced sarcomeric oscillations (cf. Kumbarachi & Nastuk, 1982; Herrmann-Frank, 1989) accompanied by non-propagating Ca²⁺-release events similar to those seen with 200 mm sucrose (data not shown). Finally 0.5 mm caffeine, which did not itself induce Ca²⁺ waves or non-propagating Ca²⁺-release events, increased their frequency (from 4.4 ± 1.05 to 14.6 ± 3.3 min⁻¹; n = 3 fibres; F/F₀ from 1.96 ± 0.28 to 2.3 ± 0.16; n = 3 fibres) when these were first initiated by hypertonic solutions (Fig. 5B).

Figure 5. Effect of caffeine on signalling phenomena.

A, typical Ca²⁺ oscillations in a fibre in isotonic Ringer solution induced by addition of 4 mm extracellular caffeine. B, Ca²⁺ oscillations in a fibre in 350 mm sucrose-Ringer solution before and following addition (horizontal bar) of 0.5 mm caffeine which did not by itself induce calcium release events.

Hypertonicity alters triad anatomy

The preceding experiments demonstrated that SR-Ca²⁺ gating was at least partially liberated from a normal restraint by the voltage-sensing processes within the tubular membranes (Huang, 1990, 1998). This might arise if interactions between RyR, DHPR and other associated proteins were modified by increased intracellular ionic strength produced by altered extracellular tonicity (Gordon & Godt, 1970). Such maneouvres additionally could alter T-tubular morphology (Franzini-Armstrong et al. 1978). We accordingly compared fibre, T-tubular and junctional SR (JSR) morphology in muscles fixed in isotonic solution (Fig. 6A), with fibres exposed to and fixed in hypertonic solutions (Fig. 6B) and fibres exposed to hypertonic solutions but returned to isotonic solutions prior to fixation (Fig. 6C).

Figure 6. Effects of extracellular tonicity on T-tubular and junctional sarcoplasmic reticular morphology.

Tubular and junctional SR morphology in control fibres left in isotonic Ringer solution (A), fibres exposed to and fixed in hypertonic solutions (B) and fibres exposed to hypertonic solutions but subsequently returned to isotonic solutions prior to fixation (C). P-face fractures through the plasma membrane (i) demonstrate diameters of T-tubular apertures (indicated by arrows) at the surface membrane. Freeze fracture replicas through the cytoplasm (ii) show the JSR and T-tubules (arrowed). Direction of shadowing indicated by arrows at bottom left of panels A–C of i and ii. Thin sections (iii) show T-SR junctions with visible ‘end-feet’ traversing the gap between JSR (J) and T-system membranes (arrowed) in A and C and dilated tubular lumina in B.

Adding 350 mm sucrose reduced mean fibre cross-sectional area from 4903 ± 264 (mean ±s.e.m.; n = 144) to 1618 ± 86 μm² (n = 177); restoring normal tonicity reversed this to 3392 ± 166 μm² (n = 171). In contrast changes in sarcomere lengths were relatively small (1.93 ± 0.002 μm in isotonic, 1.75 ± 0.003 μm in hypertonic and 1.99 ± 0.004 μm in hypertonic returned to isotonic; n = 200 sarcomeres). Fracture planes through the plasma membrane show significant (Fig. 6Ai and Bi) but reversible (Fig. 6Ci) increases in the diameters of T-tubular apertures. Such changes also occurred in fracture planes through JSR (J) and T-tubules (arrowed) before (Fig. 6Aii), during (B) and after exposure to hypertonic solutions (C).Figure 6iii shows T-SR junctions in thin sections with the classic ‘end-feet’ and gaps, identical to those previously reported (Franzini-Armstrong & Nunzi, 1983), in fibres either left in (A), or returned to isotonic extracellular solutions (C). In contrast, fibres exposed to 350 mm sucrose-Ringer solution (B) showed significant morphological changes. Hypertonicity increased T-tubular diameters from 25.70 ± 0.60 to 85.02 ± 4.70 nm, a change reversed to 24.40 ± 0.96 nm (n = 12) following return to isotonicity. In contrast the volume fraction of muscle occupied by JSR decreased from 2.72 ± 0.32 to 1.68 ± 0.20 with hypertonic exposure but reversed to 2.75 ± 0.40 (n = 8) with return to isotonicity. The gap between T and SR membranes fell from 9.90 ± 0.24 nm, in agreement with earlier reports (Franzini-Armstrong & Nunzi, 1983), to 6.60 ± 0.41 nm with hypertonic exposure, but was reversed to 10.10 ± 0.44 nm (n = 12) on return to isotonicity.

DISCUSSION

The present findings demonstrate for the first time that extracellular hypertonicity drives skeletal muscle into an extreme form of CICR manifest as propagated Ca²⁺ waves in intact amphibian muscle. Thus, the addition of 200 mm sucrose resulted in discrete foci of elevated cytosolic [Ca²⁺] that did not propagate and resembled sparks or macrosparks reported on earlier occasions (see Klein et al. 1996). Further elevations of extracellular sucrose concentration to 350 mm elicited regenerative, propagated Ca²⁺ waves. These Ca²⁺-signalling phenomena closely resembled those observed in dyspedic skeletal myotubes expressing RyR-2 (Yamazawa et al. 1996; Nakai et al. 1997). Both were independent of either extracellular Ca²⁺ or surface electrical activity (Yamazawa et al. 1996). Both persisted despite voltage sensor inactivation (Yamazawa et al. 1996) but were blocked by RyR inhibitors and enhanced by the RyR potentiator caffeine (Yamazawa et al. 1996; Nakai et al. 1997). They co-existed with the activation and inactivation of baseline [Ca²⁺] brought about by the prolonged depolarization following increased extracellular [K⁺] that would reflect the voltage-operated Ca²⁺-release processes expected in skeletal muscle.

Such phenomena can also occur in cardiac myocytes, which express the RyR-2 isoform, when CICR is potentiated owing to an overloading of their SR with Ca²⁺ (Lipp & Niggli, 1993; Cheng et al. 1996). Our observed signalling phenomena are independent of extracellular Ca²⁺; taken together these findings suggest that the skeletal muscle SR is maximally Ca²⁺ loaded even under normal conditions. Thus, cardiac E-C coupling primarily involves CICR (Fabiato, 1985; Cannell et al. 1995) rather than the direct coupling mechanisms suggested for skeletal muscle by anatomical (Franzini-Armstrong & Protasi, 1997), electrophysiological (Huang, 1996, 1997, 1998) and cell biological evidence (Yamazawa et al. 1996; Nakai et al. 1997).

However, amphibian skeletal muscle expresses RyR-α and RyR-β, homologous to mammalian RyR-1 and RyR-3, respectively, rather than the RyR-2 isoform (Franzini-Armstrong & Protasi, 1997). The RyR-α isoform may normally be directly coupled to and regulated by the DHPR-voltage sensor in contrast to the cardiac RyR-2 (Olivares et al. 1991; Murayama & Ogawa, 1992; Huang, 1996, 1997, 1998). Nevertheless it could possibly be capable of exhibiting the CICR that constitutes the major gating mechanism adopted by RyR-2 (Fessenden et al. 2000). In contrast, the uncoupled RyR-βs could be regulated either indirectly by a CICR initiated by Ca²⁺ release through the RyR-αs in response to applied voltage clamp steps (Jacquemond et al. 1991) or by their direct allosteric interactions with such coupled RyR-αs (Pape et al. 1995; Jong et al. 1995; Huang, 2001). The present findings nevertheless implicate either or both amphibian RyR subtypes in the signalling phenomena induced by extracellular tonicity as described here (cf. Bull & Marengo, 1993).

The present experiments demonstrated that treatment with hypertonic solutions evoked Ca²⁺ waves and oscillations in amphibian skeletal muscle. Earlier reports suggest that such solutions alter the physiological properties of muscle cells through alterations in intracellular ionic strength (Gordon & Godt, 1970). This might influence protein-protein interactions such as those involved in associations of RyRs with DHPR-voltage sensors or their accessory proteins such as FKBP12 (Franzini-Armstrong & Protasi, 1997) and calmodulin (Meissner, 1994), in turn altering channel gating. Hypertonicity certainly leads to the generation of both Ca²⁺ and tension transients in isolated ventricular muscle (Allen & Smith, 1987). Increased tonicity also alters T-tubular anatomy. The present experiments demonstrate marked, reversible increases in tubular diameter and diminished T-SR membrane spacing that could well alter such DHPR-RyR associations thus corroborating and extending earlier reports (Franzini-Armstrong et al. 1978). Release of intracellularly stored Ca²⁺ by either mechanism could explain the increased resting metabolic rates classically associated with extracellular hypertonicity (Yamada, 1970), partial depolarization produced by elevated external [K⁺] (Solandt, 1936; Hill & Howarth, 1957) or application of subcontracture caffeine concentrations (Hartree & Hill, 1924). Finally, both Gordon & Godt (1970) and Lannergren & Noth (1973) reported that hypertonic solutions could cause small, graded and sustained tension increases; these might conceivably have resulted from release of Ca²⁺ from internal stores.

In all events, the present findings demonstrate that fully polarized intact skeletal muscle is capable of showing CICR activity in hypertonic solutions in the absence of exogenous RyR agonists. Such CICR activity may result from altered gating of either or both RyR-α and RyR-β. Murayama & Ogawa (2001) recently suggested that the CICR activity of RyR-α is suppressed by an association between RyR-α and the DHPR. In the present experiments hypertonicity could well cause an uncoupling of at least some of the direct associations between RyR-α and DHPR. The possibility of a receptor uncoupling has also been suggested by independent electrophysiological studies of intramembrane charge movement using pharmacological agents directed at the RyR (Huang, 1996, 1997, 1998). Further investigations of such structural, biophysical and molecular changes induced by hypertonicity may provide further information on the mechanisms controlling RyR gating. Defective RyR gating has been associated with mutations in either the DHPR or the RyR-1 and leads to pathological responses associated with diseases such as malignant hyperthermia (Mickelson & Louis, 1996; Maclennan, 2000).