Comparative Biochemistry and Physiology Part A 134(2003) 115–120

1095-6433/03/$ - see front matter� 2002 Elsevier Science Inc. All rights reserved.PII: S1095-6433Ž02.00225-8

Electron microscopic evidence for the thick filamentinterconnections associated with the catch state in the anterior byssal

retractor muscle ofMytilus edulis

Ichiro Takahashi , Mitsuyo Shimada , Tsuyoshi Akimoto , Teruhiko Kishi , Haruo Sugi *a b b b b,

aCentral Electron Microscopic Laboratory, School of Medicine, Teikyo University, 2-1-1 Kaga, Itabashi-ku, Tokyo 173-8605, JapanbDepartment of Physiology, School of Medicine, Teikyo University, 2-1-1 Kaga, Itabashi-ku, Tokyo 173-8605, Japan

Received 16 May 2002; received in revised form 11 July 2002; accepted 17 July 2002

Abstract

The anterior byssal retractor muscle(ABRM) of a bivalve molluscMytilus edulis is known to exhibit catch state, i.e.a prolonged tonic contraction maintained with very little energy expenditure. Two different hypotheses have been putforward concerning the catch state; one assumes actin-myosin linkages between the thick and thin filaments thatdissociate extremely slowly(linkage hypothesis), while the other postulates a load-bearing structure other than actin-myosin linkages(parallel hypothesis). We explored the possible load-bearing structure responsible for the catch state byexamining the arrangement of the thick and thin filaments within the ABRM fibers, using techniques of quick freezingand freeze substitution. No thick filament aggregation was observed in the cross-section of the fibers quickly frozen notonly in the relaxed and actively contracting states but also in the catch state. The thick filaments were, however,occasionally interconnected with each other either directly or by distinct projections in all the three states studied. Theproportion of the interconnected thick filaments relative to the total thick filaments in a given cross-sectional area wasmuch larger in the catch state than in the relaxed and actively contracting states, providing evidence that the thickfilament interconnection is responsible for the catch state.� 2002 Elsevier Science Inc. All rights reserved.

Keywords: Catch state; Molluscan catch muscle;Mytilus edulis; Anterior byssal retractor muscle; Quick-freezing; Freeze-substitution;Thick filament fusion; Electron microscopy

1. Introduction

When the anterior byssal retractor muscle(ABRM) of a bivalve molluscMytilus edulis ismade to contract by acetylcholine(ACh), it firstshows active force development and then the forceis sustained long after removal of ACh(Twarog,1976). The above tonic contraction is maintainedwith very little energy expenditure(Nauss andDavies, 1966; Baguet and Gillis, 1968), and is

*Corresponding author. Tel.:q81-339-643-593; fax:q81-353-758-789.

E-mail address: sugi@med.teikyo-u.ac.jp(H. Sugi).

called the catch state. During the catch state, theactive state is absent(Jewell, 1959), and themembrane potential is at the resting level withoutany electrical activity(Hidaka and Goto, 1973).Evidence has been presented that the catch stateis accompanied by a decrease in the intracellularfree Ca concentration below the level required2q

for activation of the contractile system(Baguet,1973; Atsumi and Sugi, 1976; Ishii et al., 1989).The ABRM in the catch state can be made to relaxby 5-hydroxytryptamine(5-HT) (Twarog, 1954).

Two different hypotheses have been put forwardto account for the catch mechanism. One is thelinkage hypothesis, in which the catch state is

116 I. Takahashi et al. / Comparative Biochemistry and Physiology Part A 134 (2003) 115–120

associated with a marked decrease in dissociationrate of actin–myosin linkages between the thickand thin filaments(Lowy and Millman, 1963).The other is the parallel hypothesis, whichassumes, in addition to actin–myosin linkagesresponsible for active force development, forma-tion of linkages interconnecting the paramyosin-containing thick filaments to produce the catchstate (Johnson et al., 1959; Heumann and Zebe,1968). The parallel hypothesis is based on theelectron microscopic evidence that a marked aggre-gation of the thick filaments takes place duringthe catch state(Heumann and Zebe, 1968; Gillo-teaux and Baguet, 1977; Hauk and Achazi, 1987),and also on the biochemical evidence that energyconsumption during the catch state is not due toactivity of actomyosin ATPase(Guth et al., 1984;Galler et al., 1999).

Since the marked thick filament aggregation hasbeen observed in the ABRM fibers fixed withglutaraldehyde, which is known to crosslink lysinresidues(Haas, 1966), it has been pointed out thatthe thick filament aggregation may be an artifactresulting from glutaraldehyde fixation(Miller,1968). In fact, Bennett and Elliott(1989), whoused techniques of quick-freezing and freeze-sub-stitution, observed no thick filament aggregationin the ABRM fibers during the catch state. Recent-ly, however, Sugi et al.(1999) have presentedevidence that the catch state is maintained by aload-bearing system other than actin–myosin link-ages, based on their mechanics experiments. Thepresent electron microscopic work was undertakento explore the presence of the load-bearing systemspecialized for the catch state, using techniques ofquick-freezing and freeze-substitution. The resultsobtained strongly suggest that the catch state isassociated with thick filament interconnections,thus providing evidence in favor of the parallelhypothesis.

2. Materials and methods

2.1. Preparation

Specimen ofMytilus edulis were collected atthe Misaki Marine Biological Station, and kept inartificial sea water. The ABRM was isolated witha piece of shell attached to one end and the byssalorgan left at the other, and carefully teased toobtain a fiber bundle of approximately 0.5 mmdiameter. The ABRM fibers were completely

relaxed in the presence of 10 M 5-HT, andy6

placed in an experimental chamber filled withartificial sea water(ASW) containing 513 mMNaCl, 10 mM KCl, 10 mM CaCl and 50 mM2

MgCl (pH adjusted to 7.2 by NaHCO or 10 mM2 3

Tris maleate). Some preliminary experiments werealso made in natural sea water with similar electronmicroscopic results to those described in this paper.The shell end of the fibers was clamped, while thebyssal end was tied to the extension of a forcetransducer(U-gage, Shinko). All experiments weremade at the slack lengthL , at which the resting0

force was just barely detectable.

2.2. Quick-freezing

The ABRM fibers were frozen quickly betweentwo copper blocks precooled in liquid nitrogen.The copper blocks were attached to a slammerdevice operated by hand. The inward-facing sur-faces of the copper blocks were shaped in such away that the blocks are made in contact with theABRM fibers over an area of approximately 7=3mm. The contacting surfaces were polished andvacuum annealed to remove dusts. Solutions in theexperimental chamber(3 ml) were continuouslycirculated(at 1 mlys) with a water vacuum suctiontube. The ABRM fibers were frozen in threedifferent states, i.e.(a) the relaxed,(b) the activelycontracting and(c) the catch state. The ABRMfibers in the relaxed state were frozen in ASWcontaining 10 M 5-HT. The ABRM fibers iny6

the actively contracting state were frozen as soonas the peak force induced by 10 M ACh wasy3

reached. To produce the catch state, the ABRMfibers were first made to contract with ACh, andwhen the ACh-induced force rose to the peak,ACh was removed to put the fibers in the catchstate. The catch state, in which the fibers do notredevelop isometric force following a quick releaseand do not shorten actively, was confirmed to beestablished at 3–5 min after removal of ACh. Thecatch force at the moment of freezing was 50–70% of the peak force induced by ACh.

2.3. Freeze-substitution and embedding

The frozen specimen was transferred into a tubecontaining precooled acetone at the temperature ofsolid CO , and kept for a couple of days to allow2

the water in the specimen to be substituted by

117I. Takahashi et al. / Comparative Biochemistry and Physiology Part A 134 (2003) 115–120

Fig. 1. Cross-sections of the ABRM fibers quickly frozen inthe relaxed state(a), in the actively contracting state(b), andin the catch state(c). The thick filaments are occasionallyinterconnected either directly or by projections as indicated byarrows=65 000.

acetone. The specimen was further transferred intoanother tube containing acetone and 5% glutaral-dehyde(added in the form of a 50% water solu-tion), and then put into a brass box also at thetemperature of solid CO containing 50% ethanol2

in water (in the form of ice). The box, placed ina polystyrene foam container, was kept at 48C togradually warm up the specimen for two days.After the above freeze-substitution procedures, thespecimen were embedded in Epoxy lesin to besectioned. The thin sections(thickness, 80–100nm) were stained with 10% uranyl acetate, andobserved in a JEOL 100CX electron microscope.Only the fibers without ice damage were used forelectron microscopic observation.

3. Results

3.1. Absence of thick filament aggregation in thecatch state

Fig. 1 shows typical cross-sections of the ABRMfibers in the relaxed(a), actively contracting(b),and catch states(c). In agreement with the reportof Bennett and Elliott(1989), who used quickfreezing and freeze substitution techniques, nothick filament aggregation was observed not onlyin the relaxed and the contracting states but alsoin the catch state. This indicates that the thickfilament aggregation associated with the catch state(Heumann and Zebe, 1968; Gilloteaux and Baguet,1977; Hauk and Achazi, 1987) may be an artifactarising from crosslinking of lysine residues duringglutaraldehyde fixation(Fig. 1).

3.2. Thick filament interconnections

As can also be seen in Fig. 1, the thick filamentswere occasionally interconnected with each other,either directly or by distinct projections. In thecross-section, the projections interconnecting thethick filaments were 13.2"5.1 nm (mean"S.D.,ns30) in length, and 18.5"3.4 nm (ns30) indiameter(Fig. 2b), and these values were larger(t-test,P-0.01) than the corresponding values ofthe projections between the thick and thin filamentlinkages (diameter 6.0"0.5 nm, ns30; length5.2"0.8 nm, ns30) (Fig. 2a). Although it wasvery difficult to prepare the fiber longitudinalsections showing the thick filament interconnec-tions, we could obtain a few such longitudinal

sections, in which the projections interconnectingthe thick filaments could be seen(Fig. 3b) as wellas the linkages between the thick and thin fila-ments(Fig. 3a). As can be seen in Fig. 3b, thethick filaments were interconnected by the projec-tions at many points along their length.

118 I. Takahashi et al. / Comparative Biochemistry and Physiology Part A 134 (2003) 115–120

Fig. 2. Cross-sections at higher magnification showing(a) the linkage between the thick and thin filaments(arrows), and (b) theinterconnection between the thick filaments by a projection(arrows) =120 000. The ABRM fibers were frozen in the actively con-tracting state in a, and in the catch state in c.

Fig. 3. Longitudinal sections showing(a) the linkages between the thick and thin filaments(arrows), and(b) the projections intercon-necting the thick filaments(arrows) =100 000. The contrast of the electron micrograghs is reversed to clearly show the thick-to-thinlinkages and the thick-to-thin interconnections=100 000. The ABRM fibers were frozen in the catch state.

3.3. Change in proportion of the interconnectedthick filaments

To explore the possibility that the thick filamentinterconnections are associated with the catch state,we measured the proportion of the interconnectedthick filaments relative to the total thick filamentsin the ABRM fiber cross-sections frozen in thethree different states. For each state, eight differentcross-sections(2=2 mm), prepared from fourdifferent ABRM fiber preparations, were used tomeasure the proportion of the interconnected thickfilaments relative to the whole thick filamentswithin each fiber cross-section, containing each

120–160 thick filaments. The results obtained aresummarized in Fig. 4. In both the relaxed andactively contracting states, the proportion of theinterconnected thick filaments was less than 20%of the total thick filaments, being 16.6"2.4%(mean"S.D., ns8) in the relaxed state and14.5"3.3%(ns8) in the actively contracting state(Fig. 4).

In contrast, the proportion of the interconnectedthick filaments was found to increase by approxi-mately threefold in the catch state compared to therelaxed and actively contracting states(t-test,P-0.01); the proportion of the interconnected thickfilaments was 48.3"4.0% (ns8). The marked

119I. Takahashi et al. / Comparative Biochemistry and Physiology Part A 134 (2003) 115–120

Fig. 4. Proportion of the interconnected thick filaments asmeasured in cross-sections of the ABRM fibers frozen in therelaxed state, in the actively contracting state, and in the catchstate.

increase in proportion of the interconnected thickfilaments in the catch state strongly suggests thepossibility that the thick filament contacts andinterconnections are responsible for the catch state,supporting the parallel hypothesis that the catchstate is maintained by the load-bearing systemother than the actin–myosin linkages.

4. Discussion

Using the techniques of quick freezing andfreeze substitution, we have shown that the thickfilaments are occasionally interconnected witheach other in the relaxed, actively contracting andcatch states(Figs. 1–3), and that the proportionof the interconnected thick filaments in the fibercross-section increased markedly in the catch statecompared to the relaxed and actively contractingstates(Fig. 4). In fact, the thick filament directinterconnections can also be seen in the ABRMcross-sections in the report of Bennett and Elliott(1989), although they paid no attention to them.

In the fiber longitudinal sections, the same pairof the thick filaments are interconnected by theprojections at many points while they run inparallel with each other(Fig. 3b). Since the thick

filaments are not straight but somewhat wavy, theytend to be very close to each other(Fig. 3b). Inthe fiber cross-sections, the interconnected pair ofthick filaments may therefore appear either asinterconnected by the projections or interconnecteddirectly (Fig. 1), depending on the level at whichthe extensive thick filaments were cut transversely.The decrease in lateral spacing between the thickfilaments resulting from the interconnectionsbetween them would give an explanation for thethick filament aggregation in the catch state, sincethe markedly increased proportion of the intercon-nected thick filaments would greatly increase theprobability of the thick filaments to be cross-linkedduring glutaraldehyde fixation.

In various invertebrate muscles, a class of myo-sin-binding proteins belonging to the immunoglob-ulin superfamily are expected to play structuraland regulatory roles in muscle functions. One ofthe myosin-binding proteins is named twitchin,and has been shown to exist inMytilus ABRM(Siegman et al., 1997). It is a high molecularprotein (molecular weight,;600 kDa), and itsphosphorylation controls the catch state in theABRM (Butler et al., 1996, 2001). Although thearrangement of twitchin in the thick filament ofthe ABRM is unknown, it seems possible that itis twitchin that interconnects the thick filament toproduce the catch state. To clarify the arrangementof twitchin in the ABRM thick filament, we arecurrently performing high-resolution scanningelection microscopy of the thick filaments inquickly frozen ABRM fibers.

References

Atsumi, S., Sugi, H., 1976. Localization of calcium-accumu-lating structures in the anterior byssal retractor muscle ofMytilus edulis and their role in the regulation of active andcatch contractions. J. Physiol. 257, 549–560.

Baguet, F., 1973. The catch state in glycerol extracted fibersfrom a lamellibranch smooth muscle. Pflug. Arch. Eur. J.Physiol. 340, 19–34.

Baguet, F., Gillis, J.M., 1968. Energy cost of tonic contractionin a lamellibranch catch muscle. J. Physiol. 198, 127–143.

Bennett, P.M., Elliott, A., 1989. The ‘catch’ mechanism inmolluscan muscle: an electron microscopy study of freeze-substituted anterior byssus retractor muscle ofMytilus edu-lis. J. Muscle Res. Cell Motil. 10, 297–311.

Butler, T.M., Mooers, S.U., Navayan, S., Siegman, M.J., 1996.Regulation of catch muscle by twitch phosphorylation:Effects on force, ATPase and shortening. Biophys. J. 75,1904–1914.

Butler, T.M., Narayan, S.R., Mooers, S.U., Hartshorne, D.J.,Siegman, M.J., 2001. The myosin cross-bridge cycle and its

120 I. Takahashi et al. / Comparative Biochemistry and Physiology Part A 134 (2003) 115–120

control by twitch phosphorylation in catch muscle. Biophys.J. 80, 415–426.

Galler, S., Kogler, H., Ivemeyer, M., Ruegg, J.C., 1999. Forceresponses of skinned molluscan catch muscle followingphotoliberation of ATP. Pflug. Arch. Eur. J. Physiol. 438,525–530.

Gilloteaux, J., Baguet, F., 1977. Contractile filaments organi-zation in functional states of the anterior byssus retractormuscle (ABRM) of Mytilus edulis. L. Cytobiol. 15,192–220.

Guth, K., Gangelmann, M., Ruegg, J.C., 1984. Skinned smoothmuscle: time course of force and ATPase activity duringcontraction cycle. Experientia 40, 174–176.

Haas, D.J., 1966. Studies of cross-linked lysozyme crystals.Acta Crystallogr. 21, A159–160.

Hauk, R., Achazi, R.K., 1987. The ultrastructure of a mollus-can catch muscle during a contraction-catch-relaxation cycle.Eur. J. Cell Biol. 45, 30–35.

Heumann, H.G., Zebe, E., 1968. ber die Funktionsweise glatterMuskelfasern. Electronenmikroskopishe Untersuchungen amByssusretraktor(ABRM) von Mytilus edulis. Z. Zellforsch.85, 534–551.

Hidaka, T., Goto, M., 1973. On the relationship betweenmembrane potential and tension inMytilus smooth muscle.J. Cell. Comp. Physiol. 82, 357–364.

Ishii, N., Simpson, A.W.M., Ashley, C.C., 1989. Free calciumat rest and during ‘catch’ in single smooth muscle cells.Science 243, 1367–1368.

Jewell, B.R., 1959. The nature of the phasic and tonicresponses of the anterior byssal retractor muscle ofMytilus.J. Physiol. 149, 154–177.

Johnson, W.H., Kahn, J.S., Szent-Gyorgyi, A.G., 1959. Para-myosin and contraction of ‘catch’ muscles. Science 130,160–161.

Lowy, J., Millman, B.M., 1963. The contractile mechanism ofthe anterior byssus retractor muscle(ABRM) of Mytilusedulis. Phil. Trans R. Soc Lond. B. 246, 105–148.

Miller, A., 1968. A short periodicity in the thick filaments ofthe anterior byssus retractor muscle ofMytilus edulis. J.Mol. Biol. 32, 687–688.

Nauss, K.M., Davies, R.E., 1966. Changes in inorganic phos-phate and arginine during the development, maintemanceand loss of tension in the anterior byssus retractor muscleof Mytilus edulis. Biochem. Z. 345, 173–187.

Siegman, M.J., Mooers, S.U., Li, C., et al., 1997. Phosphory-lation of a high molecular weight(600 kDa) protein regu-lates catch in invertebrafe smooth muscle. J. Muscle Res.Cell Motil. 18, 655–670.

Sugi, H., Iwamoto, H., Shimo, M., Shirakawa, I., 1999.Evidence for load-bearing structures specialized for the catchstate inMytilus smooth muscle. Comp. Biochem. Physiol.A. 122A, 347–353.

Twarog, B.M., 1954. Responses of a molluscan smooth muscleto acetylcholine and 5- hydroxytryptamine. J. Cell. Comp.Physiol. 44, 141–163.

Twarog, B.M., 1976. Aspects of smooth muscle function inmolluscan catch muscle. Physiol. Rev. 56, 829–838.