Excitation—contraction coupling in skeletal and caudal heart muscle of the hagfish Eptatretus burgeri Girard

Hagfishes are now considered, on both morphological and molecular grounds,to belong in the Cyclostomata as a sister-group to lampreys, the Cyclostomata itself being the sister-group to the vertebrates. They are less evolved than lampreys and so are the most primitive craniates available for study(Forey and Janvier, 1993;Mallat et al., 2001; Zardova and Meyer,2001). Consequently, most studies on these animals have addressed phylogenetic aspects (see books edited by Brodal and Fänge, 1963; Jørgensen et al., 1998; Ahlberg, 2001), but they have also attracted a variety of other studies as they are specialized animals with a unique physiology. For example, hagfish blood is iso-osmotic to seawater. Our studies on hagfish were focused on the mechanisms of excitation—contraction coupling (E—C coupling) of skeletal and caudal heart muscles. This topic is of some phylogenetic interest, since the mechanism of E—C coupling is quite different in invertebrates and vertebrates.

In the wide spectrum of invertebrate and protochordate twitch muscle fibres so far examined, influx of extracellular Ca²⁺ is necessary for E—C coupling under physiological conditions(Hagiwara et al., 1971; Inoue et al., 1994, 1996, 1997; Bone et al., 1997, 1999; Tsutsui et al., 2000). However, as the vertebrate group evolved, skeletal muscle gained a new unique function, and E—C coupling in vertebrates is different: external Ca²⁺ is not required (Armstrong et al., 1972). The fibres contract in response to electrical excitation of the muscle membrane using only intracellular Ca²⁺stored in the sarcoplasmic reticulum (SR). Depolarisation of the transverse tubules directly triggers Ca²⁺ release from the SR, known as depolarisation-induced Ca²⁺ release (DICR)(Ríos and Brum, 1987). Hence this `vertebrate type' of E—C coupling is in striking contrast to the E—C coupling of invertebrate twitch muscles.

An interesting question is where in phylogeny this change in the mechanism of skeletal muscle E—C coupling took place, i.e. when the DICR function was acquired. We have previously shown that in the skeletal muscle of the lamprey Lampetra planeri, E—C coupling is of the `vertebrate type', whilst in the acraniate Branchiostoma lanceolatum, E—C coupling is of the `invertebrate type', requiring influx of extracellular Ca²⁺ (Inoue et al.,1994). Based on these studies, we proposed that the DICR function was acquired somewhere in phylogeny between the acraniate and agnathan levels(Inoue et al., 1994). Unfortunately there is no living intermediate between the two levels; however from molecular analyses, both lampreys and hagfish should be regarded as a natural group of Agnatha, forming the sister group to the Gnathostomata. In order to test our hypothesis, we examined whether E—C coupling of the hagfish skeletal muscle is of the `vertebrate type' (no requirement for Ca<sup>2+</sup> influx) or the `invertebrate type' (requiring Ca²⁺influx), bearing in mind that the body fluid of hagfishes is iso-osmotic to seawater (Alt et al., 1981) and contains a much higher concentration of Ca²⁺ than the body fluid of any other vertebrate. We studied two different types of skeletal muscle fibres, one with white (fast) fibres and the other with red (slow) fibres.

The hagfishes possess a unique circulation system, using a secondary caudal heart. The caudal heart is a bilaterally symmetrical organ below the notochord, consisting of two chambers on each side of a median caudal cartilage. Both sides of the caudal heart muscle contract alternately, bending the caudal cartilage to the contracting side. These rhythmic movements cause the chamber of the other side to expand and flatten, so that blood is returned to the central veins (see reviews by Johansen, 1963; Lomholt and Franko-Dossar,1998). The caudal heart muscle is extrinsic to each heart chamber. The muscle fibres are long and cross-striated, and innervated from the spinal cord. The hagfish Myxine glutinosa the caudal heart muscle was examined by Retzius (1890, 1892), who found that it is derived from parietal muscle and innervated by spinal nerves terminating in similar nerve endings to those of red parietal muscle fibres. The fibres of the caudal heart muscle are smaller in diameter than any in the myotomes. Our study reveals that E—C coupling in the caudal heart muscle as well as in both white and red skeletal muscle fibres is of the `vertebrate type'.

Adult hagfishes Eptatretus burgeri Girard (30-50 cm long) were caught with net traps at the bottom of the Wakasa Bay, off the coast of Ine,Kyoto, Japan. Animals were maintained in a laboratory tank filled with filtered seawater at 18°C. Each animal was killed by decapitation, the skin was removed, and muscle tissue used for experiments was dissected out in nominally Ca²⁺-free artificial seawater (0Ca²⁺-ASW) at 15°C.

White skeletal muscle fibres were dissected from the musculus tubulatus that protrudes the dental plate cartilage from the mouth. Approximately ten fibres (approx. 400 μm diameter and 3-5 cm long) were dissected under a binocular microscope. The fibres were separated (except for a portion at both ends) using a pair of fine needles. The tendon at one end of the fibres was pinned down to the Sylgard base of a small chamber. The other end of the fibres was connected to a strain gauge (SensoNor N801, Norway) via a small dissecting pin (Seirin Hinaishin No 2, Japan) as described below. The volume of the chamber was 4 cm³, and the bathing solution continuously flowed at a rate of 10 ml min^-1. A sheet of red skeletal muscle fibres (100-150 μm thick, approx. 3 mm wide and 7 mm long) was sliced off from the parietal muscle using a vibrotome(Campden Instruments Vibroslice, UK), and was pinned down to the Sylgard base of the chamber under the binocular microscope.

Caudal heart tissues were dissected using a small pair of scissors. The caudal heart muscle was carefully dissected out from each heart and pinned down to the Sylgard base of the chamber under the binocular microscope. Epithelial tissue and connective tissue were removed. The caudal muscle fibres were 40-50 μm diameter and 5-7 mm long.

Artificial seawater (ASW) consisted of 450 mmol l^-1 NaCl, 9 mmol l^-1 KCl, 10 mmol l^-1 CaCl₂, 50 mmol l^-1 MgCl₂ and 15 mmol l^-1 Hepes-Na buffer, pH 7.8. Nominally Ca²⁺-free ASW (0Ca²⁺-ASW) was made by replacing CaCl₂ with MgCl₂. Na⁺-free ASW(0Na⁺-ASW) consisted of 450 mmol l^-1 choline chloride, 9 mmol l^-1 KCl, 10 mmol l^-1 CaCl₂, 50 mmol l^-1 MgCl₂ and 15 mmol l^-1 Tris-HCl buffer, pH 7.8.

Tetrodotoxin (TTX) (Sankyo, Japan), acetylcholine (ACh) (Sigma, USA) and d-tubocurarine (d-TC) (Sigma, USA) were dissolved in ASW to make 1 mmol l^-1, 100 mmol l^-1 and 1 mmol l^-1 stock solutions, respectively. Each stock solution was kept at 4°C in the dark,and used by adding appropriate amounts to the bathing solutions. CoCl₂ was added to bathing solutions to block ionic currents through calcium channels.

Fibres of either skeletal muscle or caudal heart muscle were stimulated electrically using a pair of electrodes connected to the isolating unit of an electric stimulator (Nihon Kohden SEM3201, Japan). One of the electrodes was a suction electrode made of fine polyethylene tubing (100 μm inner diameter). A microelectrode (10-20 MΩ) filled with a 3 mol l^-1 KCl solution connected to a head stage of an electrometer (World Precision Instruments Duo773, USA) was inserted in a fibre whose twitches had been observed in response to electrical stimuli. For direct stimulation of a caudal heart muscle fibre, a microelectrode was inserted to the fibre and transmembrane current was applied to the fibre through the microelectrode using the current injection circuit of the electrometer.

Twitches were monitored with a CCD camera and recorded on videotape (Video recorder: Sony FS10, Japan). Records of the twitches were obtained subsequently by placing a CdS photocell on the video monitor screen on the twitching muscles. The photocell signal was recorded through a low-pass filter(<33 Hz). Force generation of skeletal muscle fibres in response to electrical stimulation was measured with a strain gauge. A dissecting pin fixed to the gauge was inserted in the tendon at one end of the fibres.

Each muscle bundle was pinned down to a silicone pad in a culture dish. The fibres were loaded with the Ca²⁺ indicator dye fluo-3 by incubating them in each test solution containing 2 μmol l^-1 fluo-3-AM(Dojin, Japan) for 1.5 h at 14°C. After washing off the fluo-3-AM, the chamber was placed on the inverted microscope fitted with an epi-illumination apparatus, and a laboratory-built fluorescence photometer system(Bone et al., 1997). A small area of the specimen (100×100 μm²) was illuminated by excitation wavelength light (470 nm with half-bandwidth of 20 nm), and the intensity of emitted light (520-600 nm wavelength) was measured with a photomultiplier tube (Hamamatsu Photonics R268, Japan). The detecting area(10×10 μm²) was determined by a slit, and background level measured on the area off the muscle.

Signals from the photocell, the strain gauge, and the photomultiplier were recorded with a VHS video system via a laboratory-modified PCM converter (Sony PCM501ES, Japan) at a sampling frequency of 15 kHz. Data were analysed offline using an IBM/PC compatible computer and pCLAMP software.

After decapitation of the animals, blood was collected (1-2 ml from each),and centrifuged at 500 revs min^-1 for 5 min. The supernatant was taken and centrifuged at 3,000 revs min^-1 for 10 min. 250 μl of the supernatant was used for measurements of the osmolarity with an osmometer(Fiske Mark 3, USA), and the remainder for measurements of the electrolyte concentrations with a biochemical analyser (Fuji Film FDC3500, Japan).

Numerical values are presented as means ± S.D. Experiments were carried out at a room temperature of 21±1°C.

Table 1 shows the osmolarity and electrolyte concentrations of the blood plasma of the hagfish Eptatretus burgeri. Each value is similar to that of the blood plasma of the Atlantic hagfish Myxine glutinosa(Alt et al., 1981), and close to that of seawater in the laboratory tank (993 mosmol kg^-1), being mainly composed of electrolytes.

Membrane potentials were recorded from white muscle fibres of the musculus tubulatus using microelectrodes. The resting potentials in ASW recorded from 16 fibres were -55 to -80 mV, mean -67.4±7.0 mV. This is similar to that recorded from the white muscle fibres of the musculus longitudinalislinguae in Myxine glutinosa, which is used for retracting the dental plate into the mouth(Andersen et al., 1963; Nicolaysen, 1966; Flood, 1998).

Electrical stimulation with a suction electrode at the tendon evoked a propagating action potential and a muscle twitch. A propagating action potential recorded from one of those twitching fibres overshot to +40 mV,followed by undershoot beyond resting potential(Fig. 1A). Externally applied TTX (1 μmol l^-1) blocked both action potential and twitch (not shown). In contrast, both action potential and twitch were maintained after switching the external ASW to 0Ca²⁺-ASW that contained 10 mmol l^-1 Co²⁺ (N=6). Electrical stimulation in ASW,using a bundle with seven fibres, evoked a transient force reaching 3×10^-5 N (Fig. 1Bi). The same force was evoked even 10 min after switching the external ASW to 0Ca²⁺-ASW containing 10 mmol l^-1Co²⁺ (Fig. 1Bii). Hence, external Ca²⁺ was not involved in E—C coupling.

Although we did not measure the membrane potential of the red fibres, we assumed that, like the red fibres of Myxine glutinosa, they do not propagate overshooting action potentials(Jansen and Andersen, 1963)and, like them, operate by slow and graded depolarisations. We mimicked this effect with K⁺ depolarisation. To test the mechanism of E—C coupling, bundles of red fibres were pretreated with 0Ca²⁺-ASW containing 10 mmol l^-1 Co²⁺ for 1.5 h, then pinned down in the chamber filled with the same solution. Contraction of red fibres was observed under the binocular microscope when the external K⁺concentration was increased to 100 mmol l^-1, to be sure that the function of DICR by intracellular Ca²⁺ release (see Fig. 5) is definitively present in the red fibres. The force generated by the K⁺ depolarisation was measured with a strain gauge, and an example of the results obtained from six specimens is shown in Fig. 2.

Membrane potentials were recorded from caudal heart muscle fibres with microelectrodes. The resting potential was between -30 and -60 mV, mean-43.6±8.9 mV (N=53).

An electrical stimulus supplied by a suction electrode evoked a twitch of several fibres simultaneously. Subthreshold stimuli evoked graded depolarising responses that lasted >50 ms. However, the subthreshold stimuli did not evoke muscle twitches. When stimulation exceeded the threshold, a fibre generated an action potential in an all-or-none manner(Fig. 3A), and the fibre twitched.

When a train of subthreshold stimuli at the same intensity was applied through a suction electrode, subthreshold responses were intensified until they reached a constant height (Fig. 3Bi). The amplitude of the subthreshold responses was different between fibres. These subthreshold responses were reversibly suppressed by addition of a number of agents, such as TTX (1 μmol l^-1, N=6), d-TC (20 μmol l^-1, N=8), or Co²⁺ (5 mmol l^-1, N=7), and by a substitution of Na⁺ with choline⁺ (N=1), or Ca²⁺with Ba²⁺ (N=4). Fig. 3Bii,iii shows the reversible suppression of subthreshold depolarisations by d-TC. Depolarisation was also induced by extracellular application of ACh (N=3, not shown). The subthreshold depolarisations represent junction potentials responding to neuromuscular transmitters released from nerve endings by electrical stimulation of the nerves innervating the muscle fibre. Hence, action potential and associated twitch were evoked when the junction potential exceeded the threshold for the Na⁺ spike.

In order to stimulate a muscle fibre directly, electric current was applied to a fibre through a microelectrode inserted to the fibre. Application of a 0.2 ms outward-directed current pulse evoked twitches in an all-or-none manner when the pulse intensity exceeded a threshold level (N=14). No other fibres twitched, indicating that there is no electrical coupling between fibres. The force generated by one fibre in the bundle of fibres was insufficient to be detected by the strain gauge. We detected muscle twitches from a video monitor screen using a CdS photocell. Fig. 4 shows the effects of 0Ca²⁺-ASW containing 10 mmol l^-1 Co²⁺ and of 0Na⁺-ASW on twitches in response to electrical stimulation applied every 2s. Twitches were maintained in 0Ca²⁺-ASW containing 10 mmol l^-1 Co²⁺ (Fig. 4B). In contrast, twitches declined gradually in the 0Na⁺-ASW, and eventually disappeared at 15 min after the Na⁺ substitution (Fig. 4C). The twitches recovered to near the original level within a short time (5 min) after replacing the external 0Na⁺-ASW with ASW(Fig. 4D). These results indicate that the twitch does not require influx of extracellular Ca²⁺, but does require Na⁺ for action potential generation.

To examine the presence of DICR further, changes in the intracellular Ca²⁺ level in response to depolarisation by increased external K⁺ concentration were measured in fibres loaded with fluo-3(N=4). The fibres were bathed in 0Ca²⁺-ASW containing 30 mmol l^-1 Co²⁺ for 30 min before fluorescence measurements were done. On increasing the external K⁺ concentration to 100 mmol l^-1, the fluorescence signal increased(Fig. 5A), indicating that the K⁺ depolarisation had induced intracellular Ca²⁺ release without the influx of extracellular Ca²⁺.

Fig. 5B shows the effect of external K⁺ concentration on the membrane potential in caudal heart muscle fibres. Each point indicates mean ± S.D. obtained from 10 fibres. Between 9 and 200 mmol l^-1 K⁺, the membrane potential changed linearly against log[K⁺]. The slope was approximately 35 mV per tenfold change, indicating that the membrane does not behave as a pure K⁺ electrode, and permits permeation of other ions. Therefore, the intracellular K⁺ concentration could not be estimated in this way.

The present experiments revealed that in the hagfish Eptatretus burgeri, E—C coupling with depolarisation-induced Ca²⁺release (DICR) is present in white and red skeletal muscle, and in the caudal heart muscle. This supports our hypothesis that evolution to acquire the function of DICR for muscle E—C coupling occurred in phylogeny between the acraniate and agnathan levels (Inoue et al., 1994).

Skeletal muscles in vertebrates are large and multinucleated. The SR is coupled in a regular way to the transverse (T)-tubular membrane. The couplings comprise end-feet consisting of dihydropyridine (DHP) receptors in the T-tubular membrane and ryanodine receptors in the SR membrane(Inui et al., 1987; Block et al., 1988; Bers and Stiffel, 1993). In the E—C coupling of skeletal muscle, depolarisation of the T-tubular membrane is sensed by a voltage sensor in the DHP receptor, and directly transferred to the ryanodine receptor (the intracellular Ca²⁺release channel) via the end-feet to trigger Ca²⁺ release(Schneider and Chandler, 1973; Huang 1989; Ríos and Pizarro,1991). Hence E—C coupling does not require influx of external Ca²⁺ (Armstrong et al.,1972). The voltage sensor is thought to be a part of the α1 subunit of the DHP receptor, a superfamily of L-type calcium channels(Tanabe et al., 1987; Brum et al., 1988). Thus the DHP receptor of skeletal muscle possesses a dual function: as a voltage sensor for E—C coupling and as an L-type calcium channel.

So far as is known, influx of external Ca²⁺ is indispensable for physiological E—C coupling in invertebrates and protochordates including the acraniate Branchiostoma lanceolatum(Hagiwara et al., 1971; Inoue et al., 1994, 1996, 1997; Bone et al., 1997, 1999; Tsutsui et al., 2000). Ca²⁺ enters the fibres through voltage gated L-type calcium channels (Melzer, 1982; Inoue et al., 1994, 1996, 1997; Bone et al., 1997, 1999; Tsutsui et al., 2000). The activation of those L-type calcium channels in protochordates and invertebrates is fast (<1 ms), whereas that of vertebrate skeletal muscle is much slower (>50 ms) (e.g. Inoue et al., 1994). This suggests that the evolution of E—C coupling may be accompanied by a molecular evolution of the DHP receptor to enhance its function as a voltage sensor. However, at present little is known about the molecular mechanism of the phylogeny of E—C coupling. Future studies from the point of molecular phylogeny may reveal the steps involved in the striking change in the mechanism of skeletal muscle E—C coupling.

Our experiments revealed that the caudal heart muscle fibres possess the DICR function. Therefore, the caudal heart is a `vertebrate type' skeletal muscle, perhaps unsurprisingly since it is derived from parietal muscle and similarly innervated.

Na⁺ spikes were evoked in an all-or-none manner when the stimulation intensity exceeded a threshold level(Fig. 3). Twitches also occurred in an all-or-none manner when the stimulating current intensity delivered by a microelectrode inserted into the fibre exceeded a threshold level (Fig. 4). These suggest that the sodium spike propagates along the fibre, although there is no direct experimental evidence. Although the resting potential of the caudal heart muscle fibres (-43.6±8.9 mV) is similar to that of red muscle fibres of Myxine glutinosa (Andersen et al.,1963; Flood,1998), the electrical properties of the caudal heart are different from those of the slow fibres. In red fibres of Myxine glutinosa,membrane depolarisation is produced by neuromuscular junction potentials,never overshoots, and does not propagate along the fibres(Andersen et al., 1963; Nicolaysen, 1966; Flood, 1998). The electrical responses of the caudal heart muscle fibres rather resemble intermediate fibres of parietal muscle (Flood,1998). The resting potential of the intermediate fibres is approximately -60 mV, and action potentials overshoot and propagate(Andersen et al., 1963; Flood, 1998). In the dogfish,multiply-innervated red fibres propagate action potentials, though these do not overshoot (Bone et al.,1994).

In living animals, at rest, the rate of contraction of the caudal heart is slow (0.1-0.5 Hz), and fatigue-resistant. However, rather strong and fast contractions appear within a few minutes after swimming. Contractions of the caudal heart stop when they are swimming, as has been reported by Greene(1900). Surprisingly these caudal heart activities are maintained even 24 h after decapitation. The swimming movements of the trunk muscle are also maintained after decapitation;they begin responding to touch stimulation to the skin and continue for 10-30 s. These observations suggest that the caudal heart activities are linked with the swimming movements of the trunk muscle and are principally regulated by cycles of spinal chord reflexes.

We thank Dr K. Sasaki, the president of The National Institute for Physiological Sciences, for allowing us to use the facilities of the Ine Marine Laboratory. Thanks are also due to Mr. S. Mukai for collection of hagfish and Dr N. Ogiso for measurements of ionic composition of the blood plasma.