Normal conduction of surface action potentials in detubulated amphibian skeletal muscle fibres

Abstract

1. The influence of the transverse (T) tubules on surface action potential conduction was investigated by comparing electrophysiological and confocal microscopic assessments of tubular changes in osmotically shocked and control fibres from frog sartorius muscle.

2. The membrane-impermeant fluorescent dye, di-8-ANEPPs spread readily from the bathing extracellular solution into the tubular membranes in control, intact fibres. Prior exposure of muscles to a hypertonic glycerol-Ringer solution, its replacement by an isotonic Ca²⁺-Mg²⁺ Ringer solution and cooling sharply reduced such access. In contrast, dye application in the course of this osmotic shock procedure stained the large tubular vacuoles hitherto associated with successful muscle detubulation.

3. Conduction velocities in intact, control fibres (1.91 ± 0.048 m s⁻¹, mean ±s.e.m., n = 32 fibres) agreed with earlier values reported at room temperature (18-21 °C) and were unaffected by prior episodes of steady cooling to 8-10 °C (1.91 ± 0.043 m s⁻¹, n = 30).

4. Cooling to 11.5 °C reduced these velocities (1.47 ± 0.081 m s⁻¹, n = 25) but action potential waveforms still included early overshoots and the delayed after-depolarizations associated with tubular electrical activity.

5. In contrast, action potentials from cooled, superficial fibres in osmotically shocked muscles lacked after-depolarization phases implying tubular detachment. Their mean conduction velocities (1.62 ± 0.169 m s⁻¹, n = 25) were not significantly altered from values obtained in untreated controls or in intact fibres in muscle similarly treated with glycerol, in direct contrast to earlier results.

6. Cooling produced similar reductions in maximum rates of voltage change dV/dt in action potentials from all fibre groups with lower rates of change shown by detubulated fibres.

7. Use of an antibody to a conserved epitope of the α-subunit of voltage-gated sodium channels suggested a concentration of sodium channels close to the mouths of the T tubules.

8. These electrophysiological and anatomical findings are consistent with a partial independence of electrical events in the transverse tubules from those responsible for the rapid conduction of surface regenerative activity.

9. The findings are discussed in terms of a partial separation of the electrical activity propagated over the surface membrane, from the initiation of propagated activity within the T tubules, by the triggering of the sodium channels clustered selectively around the mouths of the T tubules.

Synchronous electrical excitation of skeletal muscle requires rapid action potential propagation along the cell surface. This regenerative surface activity then propagates into an extensive, normally anatomically communicating transverse tubular system (T system) to ensure fast and synchronous contractile triggering throughout the fibre cross section (Huxley & Taylor, 1958; Adrian et al. 1969). The delayed after-depolarization that follows the initial overshoot has been attributed to such tubular action potential initiation and propagation (Adrian & Peachey, 1973). Thus it is abolished following tubular detachment by the osmotic shock produced by the introduction and withdrawal of extracellular hypertonic glycerol solution (Gage & Eisenberg 1969a, Gage 1969a; Gage 1969b;Howell, 1969; Nakajima et al. 1973; Huang & Peachey, 1989; Nik-Zainal et al. 1999; Khan et al. 2000). These changes accompany large decreases in total low frequency membrane capacitance (Gage & Eisenberg 1969a, Gage & Eisenberg 1969b; Zachar et al. 1972; Dulhunty & Gage, 1973) and extensive transverse tubular vacuolation (Krolenko, 1969; Krolenko et al. 1995; Fraser et al. 1998)

Conduction velocity is affected by the rate at which local circuit currents discharge membrane capacitance ahead of the action potential wave. At least part of the T system would be expected to contribute to this capacitative load. Conversely, detubulation would exclude such electrical contributions and would therefore be expected to markedly increase conduction velocity along the fibre surface in accordance with an earlier report (Hodgkin & Nakajima 1972a, Hodgkin & Nakajima 1972b). However, Adrian & Peachey (1973) could reproduce the rapid, early deflection followed by a prolonged after-depolarization, shown by skeletal muscle action potentials, only by introducing a partial independence between tubular events and electrical changes propagating along the surface membrane. This might be achieved by an access resistance that might reflect tubular luminal narrowings near the fibre edge or a sparser tubule distribution at the fibre periphery relative to that in its axial regions. Such functional separations between tubules and surface membrane would reduce the effect of detubulation on conduction velocity

The present study examined the influence of the transverse tubules upon conduction velocity using a recently established osmotic shock procedure known to produce a more reliable tubular detachment in more viable preparations than previous methods (Koutsis et al. 1995; Khan et al. 2000). In addition, detubulation was confirmed using both electrophysiological and microscopic criteria. The chosen experimental configuration made it possible to measure conduction velocity over substantially greater lengths of statistically larger numbers of muscle fibres systematically, and gave measurements that agreed with earlier reports from intact muscle fibres studied at the same temperature (Hodgkin & Nakajima 1972a, Hodgkin & Nakajima 1972b). Conduction velocities in detubulated fibres were then compared with those in intact fibres that had been spared glycerol treatment as well as in intact fibres that had been exposed to glycerol. Finally, the basis of the findings was explored through an assessment of sodium channel distribution using an antibody to a conserved epitope of the α-subunit of voltage-gated sodium channels. We report (a) electrophysiological results that directly suggest a significant partial isolation of the transverse tubules from the surface membrane. However, (b) anatomical findings using two-photon confocal microscopy did not demonstrate the basis for a ‘passive’ electrical isolation in the form of the tubular constrictions suggested by Adrian & Peachey (1973). (c) Nevertheless the antibody-binding studies offered evidence for an ‘active’ electrical isolation that might take place through the triggering of preferentially located sodium channels.

METHODS

Sartorius muscles were dissected from cold-adapted frogs (Rana temporaria) killed by concussion followed by pithing (Schedule I: Animal Procedures Act, Home Office, UK). The tendon of insertion was ligated and the muscle dissected free from adjacent tissue up to and including its origin at the pelvic girdle and the acetabulum. The muscle was secured in a methacrylate polymer (Perspex) chamber by pinning the acetabulum and the ligature such that the muscle was stretched to between 1.4 and 1.5 times its in situ length, dorsal surface uppermost, over a 1 mm supporting ramp. The recording chamber was then divided into two watertight, and therefore electrically isolated, compartments by a methacrylate polymer (Perspex) partition lined with Vaseline, with the muscle running through a notch across which a brief voltage stimulus could be applied using two platinum electrodes (cf. Luttgau, 1965). The bath solution could be cooled by circulating water (between 0 and 16 °C) from an ice bath, through a glass coil placed in the chamber, using a Minipuls 3 peristaltic pump (Gilson, France). The temperature of the solution near the muscle was monitored using a digital thermometer (J. Bibby Science Products, UK) previously calibrated against a platinum film resistor. The following solutions, titrated to pH 7.0, were used in the osmotic shock protocol described below.

Solution A (isotonic): normal Ringer solution (mm): 115 NaCl, 2.5 KCl, 1.8 CaCl₂, 3.0 Hepes.

Solution B (isotonic): Ca²⁺-Mg²⁺ Ringer solution (mm): 115 NaCl, 2.5 KCl, 5.0 CaCl₂, 5.0 MgCl₂, 3.0 Hepes.

Solution C (hypertonic): glycerol-Ringer solution (mm): 115 NaCl, 2.5 KCl, 1.8 CaCl₂, 3.0 Hepes, 400 glycerol.

Osmotic shock procedure

Detubulation was achieved with the osmotic shock procedure used by Khan et al. (2000), itself an optimization of previously established procedures (Howell & Jenden, 1967; Caputo, 1968; Howell, 1969; Koutsis et al. 1995; Fraser et al. 1998; Nik-Zainal et al. 1999). The mounted muscle was first exposed to hypertonic (400 mm) glycerol-containing Ringer solution (solution C) for 18-20 min. This was then replaced by the Ca²⁺-Mg²⁺ Ringer solution (solution B) for 30 min and the preparation then steadily cooled from room temperature (18-21 °C) to 9-10 °C in the same solution over 30 min. The bath solution was then replaced with cold Ringer (solution A) at 9-10 °C and maintained at 9-13 °C for electrophysiological assessment. The control muscles were maintained in Ringer solution for 50 min and not subject to osmotic shock. They were then similarly cooled to 9-10 °C. Room temperature controls were also carried out with and without such a prior episode of steady cooling to confirm that the present conditions gave similar results to those obtained in earlier studies.

Electrophysiological study

KCl (3 m) glass capillary microelectrodes (resistance 8-20 MΩ, tip potentials less than 5 mV; Adrian, 1956) followed muscle resting and action potentials. The microelectrode amplifier output recording membrane potential was fed directly into input (1), driving the first beam of a calibrated digital oscilloscope (Model HM205-3, Hameg, Germany) that accordingly provided a representation of the membrane potential, V. The latter input was also connected to the input driving the second oscilloscope beam through a 100 pF capacitance C (Fig. 1). This second input, (2), in turn was connected to earth through a 100 kΩ resistor. In the resulting differentiating circuit, the current flowing across the capacitance C equals that flowing across the resistor R by Kirchoff's first law, thus: C d(V - V′)/dt =V′/R, where V′ is the voltage at the second input. This would give V′∼R C (dV/dt) where V > V′, with the result that the second oscilloscope channel would effectively measure the first derivative dV/dt of the voltage V.

Figure 1. Differentiating circuit used to derive rates of voltage change, dV/dt.

The amplifier output, V, is fed to input (1) driving the first oscilloscope channel; both are connected to input (2) by capacitance C. Input (2) is separated from earth by resistance R. In the resulting differentiating circuit, C d(V - V′)/dt =V′/R, where V′ is the voltage at the second input. If V > V′, V′∼R C (dV/dt) and the second oscilloscope channel effectively measures the first derivative dV/dt of the voltage V.

Action potentials were elicited by direct stimulation across the Perspex partition using a just suprathreshold amplitude for the fibre from which recordings were obtained initially over a timebase giving a record length of 100 ms. Fibre microelectrode impalements were made close to the fibre midpoints. Brief (250 μs) electrical stimuli of gradually increasing amplitude were then applied in order selectively to attain the threshold for excitation for the impaled fibre as far as possible and minimize the number of fibres in which electrical activity was initiated. These precautions together minimized mechanical artefacts resulting from muscle contraction. Their latencies were measured from the stimulus artefacts and peaks of the dV/dt trace using a fast time base giving a 20 ms record length. The distance between the site of action potential initiation and the recording microelectrode was ascertained from vernier scales built into the electrode holder. Sets of adjacent muscle fibres were systematically studied at successively varying distances from the edge of the partition and fibres in deeper muscle layers also examined. The fibres were categorized as detubulated or intact according to the presence or absence of an after-depolarization phase in their action potential waveforms (Zachar et al. 1972; Adrian & Peachey, 1973; Koutsis et al. 1995; Nik-Zainal et al. 1999; Khan et al. 2000).

Assessment of detubulation

Confocal microscopy was used to visualize transverse tubular anatomy following osmotic shock. Three types of experiment were performed. First, the membrane-impermeant fluorescent dye, di-8-ANEPPs (2.0 μm; Molecular Probes, CA, USA), which stains membrane (Rohr et al. 1994), was added to the bathing extracellular solution after the muscles had undergone the detubulation procedure. The second type of experiment added the dye to muscle that had been spared osmotic shock but had otherwise been similarly cooled. The final dye-trapping experiments added 0.2 μm dye to solutions B and C used in the detubulation protocol. The dye is known to enter the transverse tubular lumina rapidly and so any subsequent tubular detachment from the surface membrane should then trap fluorescent dye in the resulting vacuoles and permit their visualization. The muscles were then mounted in a chamber, dorsal surface uppermost, and gently flattened beneath a glass coverslip. Individual fibres were then viewed using a Leica TCS-SP-MP confocal microscope. Dye excitation was achieved either by using an argon laser, with an excitation wavelength of 488 nm or by two-photon excitation using a Tsunami infrared laser (Spectra Physics, USA) tuned to a wavelength of 780 nm. The emitted light was captured using a spectrophotometer detector set between 550 and 750 nm. Once optimal conditions for the pinhole aperture were obtained, usually close to one Airy disk equivalent for the objective lens used, photomultiplier gain and laser power were kept constant. Forty to fifty fibres were viewed and a series of 16 optical sections were captured with a 0.35 μm interval between sections.

Western blotting

Membranes were isolated from frog sartorius muscle according to Barry et al. (1995). Briefly, 2 g of frog sartorius muscle was homogenized in a cooled 20 mm Tris-HCl (pH 7), 1 mm EDTA buffer (TE) containing protease inhibitors (1 μg ml⁻¹ pepstatin A, 1 μg ml⁻¹ leupeptin, 1 μg ml⁻¹ aprotinin, 8 μg ml⁻¹ calpain inhibitor I, 8 μg ml⁻¹ calpain inhibitor II, and 0.2 mm pefebloc). All solutions and centrifugations were at 4 °C. After a brief centrifugation at 1000 g to remove nuclei and debris the supernatant was centrifuged at 40 000 g for 10 min. The resulting pellet was resuspended in TE-containing 0.6 m KI and protease inhibitors, incubated on ice for 10 min, and centrifuged at 40 000 g. After resuspending the pellet in TE containing protease inhibitors to wash out the KI, the sample was centrifuged again at 40 000 g for 10 min. Membrane proteins were solubilized by incubating the pellet with 2 % Triton X-100 in TE with protease inhibitors on ice for 1 h. This was followed by a brief centrifugation at 13 000 g to remove insoluble material. Membrane proteins were denatured in SDS sample buffer (100 mm Tris-HCl: pH 6.8, 2 % SDS, 10 % glycerol, 10 mm DTT, 0.1 % bromophenol blue) for 10 min at 80 °C and separated on a 3-8 % SDS polyacrylamide gel. Proteins were blotted onto PVDF membranes and blots were blocked for 2 h at room temperature by incubating with 5 % bovine serum albumin in Tris-buffered saline (TBS; pH 7.4) containing 0.1 % Triton X-100 (blocking solution). Blots were probed with the sodium channel antibody SP19 (Alomone Labs), diluted 1:1000 in blocking solution at 4 °C overnight, washed three times with TBS containing 0.1 % Triton X-100 (TBST) and incubated for 1 h at room temperature with 1:1000 diluted horseradish peroxidase-conjugated anti-rabbit swine antibody (DAKO) in blocking solution. After three washes in TBST, the signals were detected using chemiluminescence with ECL detection reagents from Amersham (UK) according to the instructions of the manufacturer.

Immunohistochemistry

Frog sartorius muscles stretched to 1.4-1.5 times in situ length on standard glass microscope slides were fixed in 2 % formaldehyde solution made from paraformaldehyde, for 30 min. Cryostat sections (10 μm) were mounted on superfrost slides. They were blocked for 10 min in TBS (pH 7.4) + 0.5 % calf serum (blocking solution) and then incubated in SP19 antibody, diluted 1:100 in blocking solution overnight at room temperature in a humidified chamber. After four washes in blocking solution, sections were incubated for 1 h at room temperature in fluorescein isothiocyanate-conjugated goat anti-rabbit antibody diluted 1:200 in blocking solution. Sections were subsequently washed four times in blocking solution followed by two rinses in de-ionized water. Sections were mounted in vectashield (Vector Laboratories, Inc.) and viewed on a Leica TCS-SP-MP laser scanning confocal microscope (excitation wavelength 488 nm; detection wavelengths 510-560 nm).

RESULTS

Action potential waveforms following osmotic shock

The electrophysiological experiments determined resting potentials, action potential waveforms, spike heights, conduction velocities and the maximum rates of rise of the action potential (dV/dt). Measurements were performed in cooled (∼11.5 °C) muscles with the exceptions indicated below: this ensured fibre viability particularly in osmotically shocked muscle (Koutsis et al. 1995). It was significantly more difficult to perform such systematic studies on detubulated muscle fibres at room temperature without a significant element of selection for fibres viable enough for electrophysiological analysis; this may partly explain some of the discrepancies between the present and earlier results (Hodgkin & Nakajima 1972a, Hodgkin & Nakajima 1972b). Control muscles were cooled over 0.5 h to 8-10 °C prior to electrophysiological study; this procedure was integral to the detubulation protocol applied to osmotically shocked muscle. Only data from fibres with stable resting potentials more negative than -60 mV with positive action potential overshoots were analysed to ensure that comparable data sets of viable muscle fibres were examined (cf. Eisenberg et al. 1971).

Figure 2 displays typical action potentials following direct electrical stimulation in both control (A) and osmotically shocked muscle (B and C). Control fibres produced an early, rapid overshoot followed by the distinct after-depolarization that is considered to reflect successful tubular propagation of regenerative activity (Fig. 2A;Gage & Eisenberg, 1969b; Zachar et al. 1972; Koutsis et al. 1995). Such delayed waveforms were lost in at least 75 % of the superficial fibres in osmotically shocked muscle (Fig. 2Ba-d); they were then frequently replaced by an actual after-hyperpolarization (Fig. 2Bc and d). Such findings have previously been interpreted in terms of a tubular detachment from the surface membrane (‘detubulation’: Zachar et al. 1972; Gallagher & Huang, 1997). The after-depolarization wave thus persisted only in the minority of superficial fibres of muscles exposed to osmotic shock (Fig. 2Cc and d). As reported earlier, after-depolarization components persisted in all the deeper fibres from which recordings were obtained by electrodes advanced beyond the surface fibres after their initial impalement under micromanipulator guidance (Fig. 2Ca and b). The latter finding was consistent with a more gradual osmotic effect of glycerol withdrawal in the muscle interior leaving intact tubules (Koutsis et al. 1995).

Figure 2. Action potentials from muscles with and without prior exposure to the osmotic shock procedure.

A, typical action potentials from superficial fibres of muscle spared osmotic shock. Note the clear after-depolarization lasting well beyond 20 ms, consistent with an intact T system. B, action potentials from superficial fibres of osmotically shocked muscle showing loss of after-depolarization and its replacement by after-hyperpolarization (fibres c and d), consistent with detubulation. C, action potentials from intact fibres in osmotically shocked muscle. a and b are from deeper layer fibres; c and d are from superficial layer fibres.

Osmotically shocked muscle contained fibres that fired repetitively following a single stimulus. This was observed in 41 % of those surface fibres with persistent after-depolarizations suggesting an intact T system. Figure 3A shows such a succession of progressively diminishing spikes probably reflecting a development of sodium channel refractoriness following tubular depolarization consequent on tubular K⁺ accumulation (Freygang et al. 1964). Figure 3B demonstrates the persistent after-depolarization phases, particularly at the end of such action potential trains. In contrast, only 16 % of detubulated fibres and only 5.9 % of the deeper fibres exhibited this behaviour. Finally, only one out of the 87 fibres from muscles that were spared any exposure to glycerol exhibited repetitive firing. These findings suggest that such re-entrant activity requires a persistent, even if compromised, tubular system.

Figure 3. Repetitive action potentials from superficial fibres of muscle previously exposed to the standard osmotic shock procedure.

Note the diminishing amplitude of the successive spikes in the action potential train in A and the clear after-depolarization after the last of a train of spikes, consistent with an intact T system in B.

Consequences of osmotic shock for action potential conduction

Figure 4 shows action potential (left) and dV/dt (right) traces used for conduction velocity measurements. Action potential latencies were measured from the distinct stimulus artefact to peak dV/dt (right traces). Both control fibres (A) and intact fibres from osmotically shocked muscle (C) continued to exhibit after-depolarizations in contrast to the detubulated fibres (B), although this was less clear at the faster timebases used here. A one-way ANOVA tested for significant variation between data sets of action potential parameters and gave F = 2.842; P = 0.017 for the conduction velocity measurements and F = 37.08, P < 0.0001 for the dV/dt measurements. It was followed by a post hoc Duncan's multiple range test to identify homogeneous subsets. By considering all data collectively it was possible to avoid the problem of multiplicity and the forced imposition of theoretical hypotheses on data, which might have occurred with the selection of inappropriate t tests. This provided a protected statistical analysis of both the conduction velocity and the peak dV/dt values in all the experimental groups taken together to a 95 % least significant difference interval level.

Figure 4. Electrophysiological results used to measure conduction velocities of muscles with and without prior exposure to the standard osmotic shock procedure.

Action potential waveforms are shown on the left-hand side, with their corresponding traces for rate of potential change on the right-hand side. Latencies were measured from the clear stimulus artefact to peak dV/dt. A, typical traces from superficial fibres of muscle spared osmotic shock. B, typical traces from superficial fibres of muscle previously exposed to the standard osmotic shock procedure. Note the clear lack of the action potential after-depolarization, consistent with detubulation. C, typical traces from fibres of muscle previously exposed to the standard osmotic shock procedure that retained their after-depolarization, consistent with an intact T system. Panels a and c are from superficial fibres; b and d are from deeper fibres.

Table 1 summarizes the information from all groups of cooled fibres studied. Glycerol exposure produced a small and statistically insignificant increase in conduction velocity (to ≈1.6 m s⁻¹) when these were compared with controls left in Ringer solution (1.47 ± 0.081 m s⁻¹; n = 25 fibres). In muscles exposed to glycerol introduction and withdrawal, the conduction velocities in detubulated fibres in which after-depolarization components were abolished (1.62 ± 0.169 m s⁻¹; n = 25) were similar to those in the intact fibres in which the after-depolarization phases were preserved (deeper fibres: 1.63 ± 0.171 m s⁻¹; n = 17; surface fibres: 1.68 ± 0.094 m s⁻¹; n = 37). The Duncan's multiple range test showed that all four fibre groups described here constituted statistically homogeneous populations.

Table 1.

Conduction parameters from cooled fibres subject to different osmotic conditions

	Fibre number (n)	Conduction velocity (m s−1)	dV/dt (103 V s−1)	Resting potential (mV)
No glycerol exposure	25	1.47 ± 0.081*	1.94 ± 0.148**	76.8 ± 1.12
Detubulated fibres	25	1.62 ± 0.169*	1.52 ± 0.106	75.7 ± 1.08
Intact superficial fibres	37	1.68 ± 0.094*	1.92 ± 0.083**	84.0 ± 1.30
Intact deep fibres	17	1.63 ± 0.171*	2.09 ± 0.080**	84.2 ± 2.06
		*P = 0.201	**P = 0.267

^*Symbols

^**denote homogeneous groups as judged using 95 % least significant difference intervals with Duncan's multiple range test.

In addition, cooled but intact fibres showed similar maximum dV/dt values (≈2 × 10³ V s⁻¹). Detubulated fibres showed significantly smaller values (≈1.52 × 10³ V s⁻¹) than intact fibres. Resting potentials in the cooled, untreated control fibres and detubulated fibres also were similar but both were lower than those in the glycerol-treated but intact fibres. Nevertheless, correlation coefficients between conduction velocity (≤ 0.023) or maximum dV/dt and resting potential (≤ 0.33) were insignificant suggesting that resting potential did not influence either parameter. Taken together, these findings suggest that, at least in the cooled fibre preparations studied here, tubular detachment does not significantly influence conduction of surface action potentials. The present findings thus sharply contrast with earlier reports (Hodgkin & Nakajima, 1972b).

Table 2 summarizes the results from the further control fibre group that was spared glycerol treatment and was studied at room temperature. It compares these results to earlier findings computed from individual fibre values provided by Hodgkin & Nakajima (1972b). A second group was additionally cooled to 8-10 °C over 30 min as performed in the detubulation procedures outlined above before a return to room temperature for electrophysiological recording. Both fibre sets showed indistinguishable conduction velocities (1.91 ± 0.048 m s⁻¹, n = 32; and 1.91 ± 0.043 m s⁻¹; n = 30 fibres, respectively; means ±s.e.m.). They showed similar resting potentials (-84.0 ± 1.06 mV and -85.8 ± 0.96 mV). Their peak dV/dt values ((2.78 ± 0.073) × 10³ V s⁻¹; (3.04 ± 0.064) × 10³ V s⁻¹) formed a homogeneous statistical population distinct from the data sets obtained at the lower temperatures, confirming that cooling did not irreversibly influence action potential propagation. Furthermore, such conduction velocity measurements at room temperature under the present stimulation and recording arrangements differed by only 8.39 % and 8.72 % from the values reported in intact fibres by Hodgkin & Nakajima (1972b) (2.09 ± 0.065 m s⁻¹; n = 20). The experimental arrangement adopted here additionally followed conduction over greater distances and thereby offered more accurate computations of conduction velocity in larger fibre numbers in a single preparation.

Table 2.

Comparison of electrophysiological parameters at room temperature and in cooled fibres spared osmotic shock

	Fibre (number) (n)	Conduction velocity (m s 1)	dV/dt (103 V s−1)	Resting potential (mV)
Room temperature	32	1.91 ± 0.048	2.78 ± 0.073	84.0 ± 1.06
Room temperature with previous cooling	30	1.91 ± 0.043	3.04 ± 0.064	85.8 ± 0.96
Hodgkin & Nakajima (1972b): intact fibres	20	2.09 ± 0.065	NA	NA
Hodgkin & Nakajima (1972b): detubulated fibres	11	3.41 ± 0.208	NA	NA
Cooled fibres	25	1.47 ± 0.081	1.94 ± 0.148	76.8 ± 1.12

NA, data not available.

Demonstration by confocal microscopy of significant tubular isolation after osmotic shock

Confocal microscopy studies further corroborated the electrophysiological evidence for successful detubulation as reflected in the loss of the after-depolarization wave in the surface fibres following the particular osmotic shock protocol used here. The initial control experiments mounted muscles in isotonic Ringer solution, cooled the preparations to 8-10 °C, then added the membrane-impermeant fluorescent dye, di-8-ANEPPs (2 μm) to the extracellular solution. It would be expected that the dye would only appreciably penetrate tubular lumina that were readily accessible from the extracellular space. Figure 5a shows a maximum intensity projection two-photon confocal image of a typical fibre in such untreated muscle; the maximum intensity projections overlay stacks of images from single optical sections that had been obtained at regularly spaced imaging planes in order to display the tubular lumina optimally. The tubular lumina stained extensively within minutes of adding dye. The projection demonstrated bifurcations in the tubules derived from the surface membrane into the more extensive tubular network. Tubular vacuoles were absent. Such findings suggest an intact T system whose lumina were continuous with the surface membrane. However, there was no evidence of the tubular constrictions that might have provided an anatomical basis for an isolation of tubular from surface membranes (cf. Adrian & Peachey, 1973).

Figure 5. Morphological study of Rana temporaria sartorius muscle fibres using the fluorescent membrane-impermeant dye, di-8-ANEPPs, visualized by confocal microscopy.

The dye was added to the bathing extracellular solution of the muscle that had been previously cooled but spared osmotic shock. Dye distribution was observed by two-photon excitation. a, a maximum intensity projection of a typical fibre obtained by overlaying 15 images from serial optical sections that had been obtained at regularly spaced imaging planes, to provide an optimal display of the tubular lumina. Note the regular, intact T system. Bifurcations near the mouths of the tubules are also visible. b, an image of a single optical section from a typical fibre of the same preparation. Single sections gave superior spatial resolution and were therefore preferred for visualizing the membranous structures. Scale bar, 10 μm.

Figure 5b and Figure 6a show that images of single optical sections that were obtained by double and single photon excitation respectively from the same preparation gave appearances in agreement with the maximum intensity projections. However, the single optical sections gave a superior spatial resolution when demonstrating tubular changes within a single confocal plane, and were therefore preferred for visualizing the membranous as opposed to luminal structures. The subsequent experiments therefore compared single optical sections using single photon excitation by the argon laser at a wavelength of 488 nm.

Figure 6. Morphological study of the effects of the standard osmotic shock procedure on muscle fibres using the fluorescent membrane-impermeant dye, di-8-ANEPPs, visualized by confocal microscopy.

All images are of single optical sections with single photon dye excitation. The dye was added to the bathing extracellular solution of the muscle preparation in each case. a, a typical fibre from muscle that had been previously cooled but spared osmotic shock. Note the intensity of staining of the intact T system. b, a typical fibre from muscle previously exposed to the osmotic shock procedure. Dye was added along with the solutions B and C in the osmotic shock protocol. The dye has been trapped in tubular vacuoles. c and d, typical fibres from muscle previously exposed to the osmotic shock procedure. This time dye was added after the osmotic shock procedure. c shows a typical fibre 15 min after dye addition. d shows a typical fibre after 45 min when the dye had equilibrated throughout the muscle. Note the decreased intensity of staining of the internal membrane structures compared to the control preparation (a) even though dye concentration was the same, suggesting a markedly reduced accessibility of the tubular lumina from the extracellular space. Such limited dye that did enter the T system again revealed the presence of large vacuoles. Scale bar, 10 μm.

Figure 6b shows a typical fibre from a muscle to which 0.2 μm dye had been added to solutions B and C in the detubulation protocol and which then had been prepared for confocal microscopy as before. This permitted a rapid initial dye entry into the transverse tubular lumina; the subsequent tubular detachment from the surface membrane would then be expected to trap fluorescent dye in the resulting vacuoles and stain their membranes. Figure 6b thus shows extensive staining of large vacuoles that have been associated with such detubulation (cf. Krolenko et al. 1995; Fraser et al. 1998; Nik-Zainal et al. 1999); some residual tubular membrane appeared to be stained, but less strongly.

In the final preparation, the osmotic shock and cooling were applied before the addition of fluorescent dye to the bathing solution at the same concentration (2 μm) as in the initial muscle. Successive confocal microscope images were then obtained over an interval of 1 h following addition of the dye. Figure 6c shows that even after the first 15 min there was no significant dye penetration into the transverse tubules. Figure 6d shows a typical confocal image after a further 45 min, after which time the relatively attenuated signal still showed no further increase in intensity suggesting that the dye had by then fully equilibrated throughout the muscle. Compared to the control preparation, there was little staining of the internal membrane structures suggesting a markedly reduced accessibility of the tubular lumina from the extracellular space.

Membrane-impermeant dye would thus stain tubular membrane whether or not fibres were exposed to osmotic shock provided dye could access tubular lumina prior to their detachment from the surface. In contrast, dye added following an osmotic shock that produced detubulation would not gain access to the detached tubules and would produce much more limited staining. Nevertheless, such staining that did occur confirmed the presence of large vacuoles that have been associated with tubular detachment in earlier reports. The microscopy findings thus corroborate the electrophysiological results in suggesting that the osmotic shock detached the T system from the surface.

Investigations of sodium channel distribution

The above experiments demonstrated that conduction processes in the surface membrane were at least partially independent from the corresponding tubular events. However, the two-photon maximum intensity projections failed to demonstrate the in vivo anatomical features, such as the constrictions suggested on earlier occasions, that could offer such a passive electrical restriction to tubular access (cf. Adrian & Peachey, 1973). This prompted preliminary experiments to investigate whether such phenomena could, alternatively, arise through the detailed distributions of excitable sodium channels themselves. The possibility both of different sodium channel densities (Jaimovitch et al. 1976; Stuhmer & Almers, 1982; Almers et al. 1983) or even subtypes with different binding (Jaimovitch et al. 1983) or electrophysiological properties (Arispe et al. 1988) has been considered on earlier occasions.

The final experiments, accordingly, investigated the distribution of sodium channels on the plasmalemma and T tubules of sartorius muscle fibres using an antibody to a conserved epitope located in the intracellular loop between domains III and IV of the α-subunit of voltage-gated sodium channels (VGSCs). A search of sequences in the EMBL databank revealed that, although the frog VGSC has not been sequenced, the epitope is conserved in the VGSC of the amphibian, Cynops pyrrhogaster (EMBL accession number: AF123593). Because the antibody was raised against the epitope from rat voltage-gated sodium channels corresponding to residues 1491-1508, we first tested its ability to recognize frog sodium channels by Western blot analysis (Fig. 7). In a membrane preparation of sartorius muscle, the SP19 antibody recognized a single band running at approximately 220 kDa. The Western blot shown in Fig. 7d shows frog membrane proteins run in duplicate. Subsequent immunostaining of fibres is shown in Fig. 7a-c. The longitudinal section visualized under confocal microscopy (Fig. 7a) revealed an intense signal at the mouth of the T tubules. Moving the plane of focus to the fibre surface revealed a punctate staining pattern (Fig. 7b). The confocal images of transverse muscle sections similarly stained confirmed this preferential distribution of label close to the perimeter of the tubular system (Fig. 7c). In such sections, preincubation with control antigen reduced, but did not completely abolish, the surface staining (data not shown). A preferential distribution of sodium channels around the mouths of the T system might potentially give rise to a system in which tubular excitation is relatively independent of local circuit depolarization arising from the remaining surface membrane. This might occur if surface electrical activity is separated from the initiation of tubular excitation by a triggering of all-or-none activity in the elevated density of sodium channels around the tubule mouths. Such a mechanism would reduce the amount of local circuit current in the remaining surface membrane required to initiate action potential activity in the T tubules. This may give rise to the present observations that detubulation has no effect on conduction velocity.

Figure 7. Investigations of sodium channel distributions on the plasmalemma and T tubules of sartorius muscle fibres using an antibody to a conserved epitope of the α-subunit of voltage-gated sodium channels.

a, immunostaining of fibres in longitudinal section visualized under confocal microscopy. Note the intense signal at the mouth of the T tubules indicated by the arrowheads. Scale bar, 20 μm. b, a punctate staining pattern is seen at the fibre surface. A cluster of staining foci is highlighted in the box. Scale bar, 20 μm. c, confocal images of transverse muscle sections confirming preferential distribution of label close to the perimeter of the tubular system. Scale bar, 20 μm. d, SP19 antibody recognizes a single band indicated by the arrow on a Western blot of frog membrane proteins in duplicate lanes.

DISCUSSION

The tubular system makes a large contribution to the steady-state skeletal muscle membrane capacitance and could therefore influence the maximum rate of potential change brought about by local circuit currents driven by surface sodium channel activation during regenerative activity. This in turn would influence the conduction velocity of the surface action potential to an extent dependent upon the access gained by such local currents to the tubules. Thus, Hodgkin & Nakajima (1972b) reported that tubular detachment (‘detubulation’) increased surface conduction velocity by 60 %. However, Adrian & Peachey (1973) could reproduce separate early spikes and prolonged after-depolarizations in action potential waveforms only by partially isolating the respective surface and tubular electrical events by an access resistance. An all-or-none propagation of tubular excitation then persisted, provided sodium channels occurred both within the tubules and the surface membrane. Such regenerative tubular excitation was demonstrated by the exaggeration, followed by the total failure, of the late tubular wave in action potentials studied under progressively increasing temperatures (Padmanabhan & Huang, 1990). However, any partial electrical isolation of tubules from surface events would reduce the extent to which tubular detachment would influence conduction velocity along the surface membrane.

The present study tested the effect of detubulation upon conduction of the surface action potential as well as its maximum rate of voltage change, dV/dt, initiated by direct electrical stimulation. Experimental procedures were optimized to accomplish systematic studies in larger numbers of detubulated but viable fibres, and confirmed the presence or otherwise of tubular isolation by independent electrophysiological and morphological means. These modifications might partly explain the differences between the present and earlier results (Hodgkin & Nakajima1972a, Hodgkin & Nakajima 1972b). The experimental configuration adopted followed surface conduction over considerably longer distances (5-14 mm) than in the earlier report through larger numbers of muscle fibres. Its conduction velocity estimates from control muscles spared osmotic shock closely agreed with those obtained by Hodgkin & Nakajima (1972b) when recorded under similar, room temperature conditions. Furthermore, cooling episodes, as similarly applied in the detubulation protocol itself, did not irreversibly alter muscle conduction. Cooled preparations (to ≈11.5 °C) showed slowed conduction velocities and reduced dV/dt values as expected but otherwise preserved both initial and after-hyperpolarization contributions to the action potential waveform. Such cooled conditions additionally proved favourable for prolonged studies of viable fibres after osmotic shock and accordingly were adopted for the present study (Koutsis et al. 1995; Nik-Zainal et al. 1999; Khan et al. 2000).

Detubulation involved a recently introduced procedure (Koutsis et al. 1995; Khan et al. 2000) that replaced simple extracellular glycerol additions and withdrawals (Hodgkin & Nakajima 1972a, Hodgkin & Nakajima 1972b) with timed exposures to 400 mm glycerol-containing solution, high [Ca²⁺] and [Mg²⁺] solutions and steady cooling. This method achieves a more consistent tubular detachment and vacuolation in the superficial muscle layer (Fraser et al. 1998) while leaving fibres with stable resting potentials that permitted prolonged electrophysiological study (Koutsis et al. 1995). Thus, fibres in all experimental groups were clearly viable with positive action potential overshoots and conduction velocities which were not influenced by resting potentials that in turn were consistently more negative than -69 mV. Successive adjacent fibres could be studied systematically, working from the muscle edge.

Action potentials in superficial fibres in such glycerol-treated muscle showed either clearly persistent or abolished after-depolarizations reflecting the presence or absence of T tubules capable of such all-or-none excitation. Such detubulation involved most (75 %) of the superficial fibres but spared fibres in deeper muscle layers. This confirms reports using this detubulation protocol (Koutsis et al. 1995) as well as the persistence of tubular capacitance contributions observed in fibres exposed to gradual rather than abrupt glycerol withdrawals (Huang & Peachey, 1989). Furthermore, some fibres in such glycerol-treated muscle showed a tendency to repetitive firing. This was less common in detubulated fibres (16 % of fibres) than in intact fibres in the same superficial muscle layer (40 %), but was rare in the intact fibres in deeper layers (5.9 %) for which the osmotic shock might have been less abrupt. This was consistent with such a re-entrant excitation requiring a persistent, even if partially compromised, T system whose known lack of chloride conductance would promote such electrical instability (Gage & Eisenberg, 1969b).

Finally, confocal microscopy allowed direct visualization of detubulation, thus corroborating the electrophysiological evidence discussed above. Osmotic shock greatly reduced the tubular access of the membrane-impermeant dye, di-8-ANEPPs, from the extracellular space, whereas dye access occurred readily in untreated fibres. However, dye introduction during the detubulation process itself, which permitted dye access before tubular detachment, achieved successful staining of both the residual tubules and the large vacuoles that had been associated with tubular detachment on earlier occasions (Krolenko, 1969; Krolenko et al. 1995; Fraser et al. 1998). Thus the morphological findings confirmed the electrophysiological findings suggesting tubular detachment following osmotic shock in the fibres in the superficial muscle layer.

Taken together, these controls made it possible to establish that glycerol exposure produced a small but statistically insignificant increase in conduction velocity (to ≈1.6 m s⁻¹) when compared with controls left in Ringer solution (≈1.47 m s⁻¹). However, conduction velocities in fibres from the muscles exposed to glycerol were similar whether in detubulated or intact fibres and whether the latter were from deeper or superficial fibre layers, in direct contrast to earlier findings (Hodgkin & Nakajima 1972a, Hodgkin & Nakajima 1972b). In any case, Duncan's multiple range test showed that all these fibre groups studied under cooled conditions formed a single statistically homogeneous population. Similar observations applied to peak dV/dt, which was, if anything, slightly depressed in the detubulated fibres.

These results are consistent with a partial isolation of tubular from surface electrical events and would be compatible with the constrictions at the mouth of the transverse tubules that were suggested by Adrian & Peachey (1973). However, these constrictions were not obvious in the maximum intensity projections of optical series of the confocal images. This prompted a preliminary exploration as to whether sodium channel distribution rather than a passive access resistance might also contribute to the present observations. Previous studies have considered such differential channel densities (Jaimovitch et al. 1976; Stuhmer & Almers, 1982; Almers et al. 1983) or even selective localizations of different channel subtypes between tubular and surface membranes (Jaimovitch et al. 1983; Arispe et al. 1988). The final experiments accordingly investigated detailed channel distributions in different regions of surface and tubular membranes using an antibody to a conserved epitope of the α-subunit of voltage-gated sodium channels. This produced selective immunostaining at the mouth of the T tubules in longitudinal section that reflected a punctate membrane-staining pattern when the plane of focus was moved to the fibre surface. This relatively superficial staining pattern was confirmed in stained images of transverse fibre sections.

These findings suggest that a hypothesis in which tubular events are independently triggered from sodium channels preferentially located around the T system perimeter, rather than continuously through local circuit currents generated from the surface membrane, might be worth further exploration and testing. Thus, the initiation of tubular electrical activity might be separated from surface excitation by a triggering of all-or-none activity in the sodium channels concentrated around the tubule mouths. This would offer an ‘active’, intermediate triggering of tubular excitation. Such events might either replace or supplement the ‘passive’ tubular membrane charging by the surface membrane through the series resistance offered by constrictions close to the T tubule apertures suggested by Adrian & Peachey (1973). Such a scheme would maximize the velocity with which the surface wave would reach the ends of the muscle, ensure synchronous tubular excitation, and minimize the effects of tubular detachment upon surface action potential conduction, as reported here.