Nature and origin of gap filaments in striated muscle

Transcription

1 Nature and origin of gap filaments in striated muscle K. TROMBITAS, P. H. W. W. BAATSEN, M. S. Z. KELLERMAYER and G. H. POLLACK Bioengineering WD-12, University of Washington, Seattle, Washington 98195, USA Summary Immunoelectron microscopy was used to study the nature and origin of 'gap' filaments in frog semitendinosus muscle. Gap filaments are fine longitudinal filaments observable only in sarcomeres stretched beyond thick/thin filament overlap: they occupy the gap between the tips of thick and thin filaments. To test whether the gap filaments are part of the titin-filament system, we employed monoclonal antibodies to titin (T-ll, Sigma) and observed the location of the epitope at a series of sarcomere lengths. At resting sarcomere length, the epitope was positioned in the I-band approximately 50 run beyond the apparent ends of the thick filament. The location did not change perceptibly with increasing sarcomere length up to 3.6 fan. Above 3.6 fan, the span between the epitope and the end of the A-band abruptly increased, and above 4 fan, the antibodies could be seen to decorate the gap filaments. Between 5 and 6 fan, the epitope remained approximately in the middle of the gap. Even with this high degree of stretch, the label remained more or less aligned across the myofibril. The abrupt increase of span beyond 3.6 fan implies that the A-band domain of titin is pulled free of its anchor points along the thick filament, and moves toward the gap. Although this domain is functionally inextensible at physiological sarcomere length, the epitope movement in extremely stretched muscle shows that it is intrinsically elastic. Thus, the evidence confirms that gap filaments are clearly part of the titin-filament system. They are derived not only from the I-band domain of titin, but also from its A-band domain. Key words: elastic filament, titin, skeletal muscle, immunoelectron microscopy. Introduction The discovery of titin (also called connectin), a giant structural protein, led eventually to the conclusion that the sarcomere contained, in addition to thick and thin filaments, a third longitudinal filament system composed of titin (Wang, 1985; Maruyama, 1986). On the basis of immunolocalization studies, it was suggested that titin formed a continuous filament connecting the M-line to the Z-line (Wang et al ; Fiirst et al. 1988; Itoh et al. 1988; Pierobon-Bormioli et al. 1990). These studies (see also Trombitas et al ) further implied that at physiological sarcomere length only part of the titin strand was free and elastic: the I-band region could stretch with increases of sarcomere length, but the A-band domain appeared to be tethered to the thick filament. Logically, as the sarcomere is extended, the I-band domain of titin should be extended concomitantly. As the sarcomere is extended beyond thick and thin filament overlap, a segment of the titin filament should become visualizable in the gap. Although gap filaments have long been observable, there has been no direct evidence that they are made of titin. In earlier studies, the misalignment of titin epitopes that occurs at longer sarcomere length has made it impossible to obtain regular gap-filament labelling (Itoh et al. 1988; Ramirez-Mitchell et al. 1990; Wang, 1985). The present study was designed to define the nature and origin of gap filaments. To achieve this, we monitored the relative positions of a titin epitope that ordinarily lies Journal of Cell Science 100, (1991) Printed in Great Britain The Company of Biologists Limited 1991 close to the A-I junction, as the sarcomere was stretched. We confirm that the gap filament is indeed made of titin. However, the filament is derived not from titin's I-band domain exclusively: increasing tension apparently rips titin from its A-band tether, allowing a segment of the A- band domain to be translated into the gap. Materials and methods Single fibres from frog semitendinosus muscle were mechanically skinned by peeling off the sarcolemma in relaxing solution. The solution contained (HIM): calcium proprionate, 0.035; magnesium proprionate, 6; potassium proprionate 5; K 2 EGTA, 15; Mops, 117; Na 2 ATP, 4.4; Na 2 CP, 15.6; KOH, 57; pca9.2, ionic strength 0.2, ph7.0). One end of each fibre was pinned down with a minutien pin (Fine Science Tools Inc., Cat. no ) to a Sylgard elastomer dissection base on the bottom of a small Petri dish. The fibres were then stretched in small increments to the appropriate sarcomere length. Sarcomere length was set from 2.5 jjia to 6.0 fim. Then the fibres were fixed in a freshly prepared fomaldehyde/pbs fixative (3.7 % paraformaldehyde, 2.7 mm KC1, 1.5 mm KH 2 PO 4,137 DIM NaCl, 8 mm Na 2 HPO 4, ph 7.2) for 15 min at 4 C, and washed three times in PBS for 30min per wash. The unspecific binding sites were blocked using PBS/1% BSA solution for 30 min. Conventionally available monoclonal titin antibody (Til, Sigma) was used for the immunolabelling experiments. The test fibres were then incubated for 24 h in the primary antibody solution (25^gml -1 mouse anti-titin IgG in PBS/BSA) at 4 C. After washing the preparations three times with PBS/BSA for 30 min, the fibres were treated with secondary antibody solution (50 fig ml" 1 rabbit anti-mouse IgG (Sigma) in 809

2 PBS/BSA) for 24h, at 4 C. Control fibres were incubated only in the secondary antibody solution under the same conditions as the test fibres. In addition, the overstretched fibres were labelled either with polyclonal tropomyosin antibody (Lemanski, 1979) as described earlier (Trombitas et al. 1990a), or with monoclonal troponin-t antibody (Sigma), in a manner similar to that used with antititin. The unbound secondary antibody was removed from the fibres by washing in PBS, for 30 min per wash, three times. Then the PBS was replaced with Mops buffer solution (20 mm K-Mops, 5mM MgCl2, ph6.8), and the fibres were fixed in 2.5% glutaraldehyde, 0.2% tannic acid, 10 mm MgCl2, 5mM EGTAand 20 mm K-Mops at ph6.8 for 30min at 4 C, according to the method of Reedy and Reedy (1985). Subsequently, the fibres were washed in Mops buffer and in 100 mm potassium phosphate buffer. Then they were postfixed in 1% OsO4, 100 mm potassium phosphate buffer at ph 6.0 for 30 min at 0 C, washed in the same buffer twice for 15 min each, and in water in the same way. Fibres were stained en bloc with 2 % aqueous uranyl acetate, dehydrated in a graded ethanol series, and embedded in Araldite. Ultrathin sections were cut with an LKB Ultratome III, stained with potassium permanganate and lead citrate, and observed and photographed with a Philips 420 electron microscope. Results In muscle stretched above a sarcomere length (SL) of 3.6 nm, a gap appears between the thick and thin filaments. In such a case, sarcomere continuity is maintained by the gap filaments (Fig. 1). Gap filaments associate with thick filaments in the A-band (arrows) and run across the gap into the I-band. They are obviously different from the thin filaments, having much smaller diameter. A fine periodicity of about nm is occasionally observable along the gap filament both in the gap itself, and in the I-band (Fig. 1, arrowheads). Furthermore, immunolabelling experiments show that although both the anti-tropomyosin antibodies (Fig. 1) and the antitroponin-t antibodies (Fig. 2) label the I-segment, there is no labelling of the gap filaments. Thus, the gap filaments are not displaced thin filaments. To study the nature of the gap filaments further, we performed immunolabelling experiments using a monoclonal titin antibody that was reported to label chickenbreast myofibrils near the A-I junction. Although this T11 antibody was reported to cross-react with muscles from several different species, frog striated muscle was not checked. We therefore carried out pilot studies on isolated myofibrils from frog leg muscle using immunofluorescence microscopy. The results (data not shown) revealed labelling at the A-I junction, the same region in which chickenbreast muscle was labelled. Experiments performed using immunoelectron microscopy showed excellent antibody labelling (Fig. 3B). The sarcomeres visualized in this field are at a relatively short sarcomere length, 2.5 /an. The epitope was distinctly labelled, forming a stripe perpendicular to the fibre axis, situated some 50 nm from the visually detectable end of the thick filament. The 50-nm separation is the same as originally reported in chicken muscle (Fiirst et al. 1988). Unexpectedly, the 50-nm separation did not change as the unactivated sarcomere was stretched, at least moderately. Fig. 3C shows the labelling in a sarcomere fixed at 3.2 /on, while Fig. 3D shows labelling in a sarcomere stretched to the verge of no overlap. In both cases the span between the two epitopes straddling the A-band was approximately 1.7 /an, the same as at the shorter lengths. Note that the label pattern remains approximately normal to the filament axis in spite of the large degree of stretch. At sarcomere lengths beyond 3.6/on, where the gap Fig. 1. Anti-tropomyosin-labelled overstretched fibres; SL (sarcomere length)=4.0/un. Labelling is restricted to the I-segment. Each thick filament continues in a single gap filament (arrows). Fine periodicity can occasionally be detected along the gap filaments (arrowheads). Bar, 0.5 fan. Fig. 2. Anti-troponin-T-labelled highly stretched fibre (SL=4.2 fan). I-segments are heavily labelled, while gap filaments remain unlabelled. Bar, 0.5 /an. 810 K. Trombitds et al.

3 Fig. 3. Immunoelectron microscopical localization of T-11 titin epitope in sarcomeres of different length. (A) Control muscle, unlabelled. (B) Antibody-labelled fibre at short sarcomere length (SL=2.5/OTI). The distance between the epitope and the end of the A-band is about 0.05/.an. (C) SL=3.2/an. The epitope position is unchanged relative to B. (D) SL=3.6. Epitope position is again unchanged. (E) SL=3.9/an. The gap between the thick and thin filaments has just appeared. Distance between label and edge of A-band is abruptly increased. Note that the epitope distribution has become slightly irregular. (F) SL=5.4^m. At sarcomere lengths between 5 and 6 /an, the label is found in the middle of the gap, /an from the end of the thick filaments. (G) Overstretched control muscle. Bar, 0.5 fan. between thick and thin filaments just appears, the span between epitopes no longer remains constant; it increases abruptly (Fig. 3E). In this Figure, which shows sarcomeres at 3.9 /on, not only is the span increased (to 1.9 /an), but it is also slightly less regular than at shorter lengths. In this case the epitope is not in the gap; it is localized approximately at the thin filament tips. Fig. 3F shows the labelling pattern well beyond overlap. Here, the label is clearly within the gap. Even with this extreme degree of stretch, the label remains more or less aligned across the myofibril. The epitope-epitope span increased with increasing sarcomere length. The label remained approximately in the middle of the gap. At sarcomere lengths between 5 and 6/an, the epitope is found from 0.4 to 0.6 /an from the edge of the A-band. In instances in which the myofibril was skewed, the label followed the angle of skew of the A-I junction or the Z-line (Fig. 4). This shows that the antibodies are not deposited Gap filaments 811

4 Fig. 4. Labelling pattern in highly stretched sarcomere. The label follows the skew of the striated pattern. d, 3rd A Sarcomere length(^m) Fig. 5. Pattern of stretch of titin epitope T-ll. Continuous line represents the epitope-epitope distance across the A-band (d A ); broken line, the epitope-epitope distance across the Z-line (d r ). Bar, 0.5/an. accidentally. Furthermore, since the epitope lies near the A-I junction at physiological sarcomere lengths, in overstretched muscle it evidently separates titin's A-band and I-band domains. Thus, both domains are seen within the gap. The results obtained at a series of sarcomere lengths are shown in Fig. 5. Data are averaged from 25 fibres. Continuous lines represent the epitope-epitope span across the A-band, broken lines the epitope epitope span across the Z-line. The former does not change until the sarcomere is stretched to approximately the point of no overlap (3.6 \im); then it changes linearly with stretch. The latter increases steadily with stretch, first with steep slope, and later, as the stretch is shared by A-band-freed titin, with shallower slope. Discussion A now classical method of demonstrating that a continuous filament runs effectively from Z-line to Z-line is to stretch the sarcomere beyond thick and thin filament overlap and examine the structure in the electron microscope. A gap then appears between the tips of thick and thin filaments. However, the gap is not empty. Very thin, so-called 'gap filaments' run longitudinally across 812 K. Trombitds et al.

5 the gap (Huxley and Peachey, 1961; Sjostrand, 1962; Carlsen et al. 1965; Locker and Leet, 1975). Such morphological experiments have not yet revealed the full nature of these filaments. The gap filaments have been shown to be distinct from thin filaments (Magid et al. 1984; Trombitas et al. 1990a) and this study also presents both morphological and immunological evidence showing that the gap filaments are not thin filaments that had been drawn out of the I-band (Figs 1 and 2). The filaments' thickness, fine periodicity and immunological properties reveal a third type of filament system. A reasonable postulate has been that the gap filament is a segment of the titin (also known as connectin) filament (Maruyama et al. 1984; Locker, 1984; Wang, 1985). Earlier tests of this postulate have proved supportive but less than conclusive: although anti-titin labelling studies have revealed label in the gap near the A-band-gap junction (Itoh et al. 1988), near the thin filament tips (Wang, 1985; Ramirez-Mitchell et al. 1990), or in the gap (La Salle et al. 1983; Gassner, 1986), the labelling pattern was generally sparse and semi-random. In this study we present direct evidence for a titin epitope in the gap. Labelling was consistent, and, considering the degree of stretch, reasonably regular across the myofibril. The location of the epitope could be followed at each of a series of sarcomere lengths, conferring reassurance that the labelling could not have been of a non-specific variety. Thus, the gap filament appears to contain titin. The titin origin is supported also by the gap filaments' fine periodicity. At a sarcomere length of 4 fan, the I-band domain of the titin filament is elongated to about three times its resting length. Since the titin axial periodicity was reported to be nm in isolated unstrained strands (Trinick et al. 1984; Wang et al. 1984), at a sarcomere length of 4 fan, the observed nm titin periodicity is expected (Fig. 1). One surprise was the sarcomere-length-dependent behavior of the titin epitope. Our initial hypothesis was that the gap filament was a stretched segment of titin's I- band domain. We therefore selected from among antibodies available to us one that lodged in the I-band, very near to the tip of the thick filament, in the hope that this locus would manifest itself in the gap when the sarcomere was overextended. Thus, the T-ll antibody (Furst et al. 1988) appeared suitable. However, we were surprised to find that with moderate stretch, up to 3.6 fan, this epitope failed to move away from the A-band (Fig. 3). We interpret this in one of two ways. The first possibility is that a short segment of titin emanating from the end of the thick filament is stiff, unlike the rest of the filament. A second and simpler possibility is that the epitope does not really lie in the I- band proper, but lies in the zone of the thick filament's tapered ends, where the lower electron density confers an I-band appearance. If the epitope remains bound to the thick filament during stretch, the absence of translation would be explained. Above 3.6 fim, this epitope translated abruptly from the thick filament, toward the Z-line (Fig. 3E,F). There was no perceptible length change in the A-band, implying that the ends of the thick filament are not disassembled during extension (Wang and Wright, 1988). The abruptness of the translation implies that some or all of the A-band titin becomes 'unglued' from the thick filament, and is pulled into the gap. Presumably, the unbinding is a consequence of the increasing tension associated with stretch. The gap filament, then, is not only a segment of I-band titin; it consists, at least in part, of A-band titin pulled from the thick filament. On the other hand, the fact that the epitope - ordinarily at the A-I junction - remains approximately in the center of the gap (Fig. 3F,G) implies that the gap contains a section of I-band titin as well. The antibody is convenient in that it separates A- and I-band domains of titin in the gap. The stretch-dependent translation of the epitope reveals some clues as to the relative stiffness of the two zones of titin. One clear implication is that the segment that comes unbound is not rigid (Maruyama et al. 1989; Wang and Wright, 1988). The finding that the epitope remains near the middle of the gap as the sarcomere is stretched implies that zones on either side of the epitope are intrinsically stretchable. More quantitative conclusions are assumption-dependent. If it is assumed, for example, that: (1) it is the entire A-band domain of titin that comes unbound, leaving a tether only at the M-line (Pierobon-Bormioli et al. 1990); (2) the slack length of the I-band domain of titin is about 0.3,um (assuming the resting sarcomere length is 2.2 fan); and (3) the A-band domain of titin is bound while in the unstrained condition. At sarcomere length of 5.2 fan (Fig. 5) the A-band domain is 1.2 fan long, or 1.5 times its original length, while the I-band domain is 1.4 fan, or about five times its original length. Thus, the A-band domain of titin would appear to be stiffer than the I-band domain (Itoh et al. 1988). On the other hand, if one of the above assumptions were erroneous, this conclusion might not hold. For example, if only a distal section of A-band titin comes loose rather than the entire strand (Wang and Wright, 1988), so that the freed segment has shorter slack length, the observed epitope behavior might imply a similar compliance for this segment as for the I-band segment. More definitive conclusions can be drawn through the use of antibodies bound simultaneously to several epitopes along the titin strand. Uncertainties notwithstanding, the results demonstrate several points unambiguously. The gap filament clearly contains titin - as has previously been suggested. The titin in the gap is recruited from two zones, the edge of the A- band and the adjoining edge of the I-band - not purely from the I-band, as generally believed. Finally, the A-band segment of titin, which is ordinarily bound along the thick filament, is not rigid; it is at least moderately compliant. References CARLSEN, F., KNAPPEIS, G. G. AND BUCHTAL, F. (1961). Ultrastructure of the resting and contracted striated muscle fibre at different degrees of stretch. J. biophys. biochem. Cytol. 11, FURST, D. 0., OSBORN, M., NAVE, R AND WEBER, K. (1988). The organization of titin filaments in the half sarcomere revealed by monoclonal antibodies in immunoelectron microscopy a map of ten nonrepetitive epitopes starting at the Z-line extends close to the M- line J. Cell Biol. 106, GASSNER, D. (1986). Myofibrillar interaction of blot immunoaffinitypunfied antibodies against native titin as studied by direct immunofluorescence and immunogold staining Eur J. Cell Biol. 40, HUXLEY, A. F AND PEAOHEY, L. D. (1961). The maximum length for contraction in vertebrate striated muscle. J. Physwl. 156, ITOH, Y., SUZUKI, T., KIMURA, S., OHASHI, K., HIGUCHI, H., SAWADA, H., SHIMIZU, T., SHIBATA, M. AND MARUYAMA, K. (1988). Extensible and less-extensible domains of connectin filaments in stretched vertebrate skeletal muscle as detected by immunofluorescence and immunoelectron microscopy using monoclonal antibodies. J. Biochem. (Tokyo) 104, Gap filaments 813

6 LA SALLE, F., ROBSON, R. M., LUSBY, M. L., PARRISH, F. C, STROMER, M. H. AND HUIATT, T. W. (1983). Localization of titin in bovine skeletal muscle. J. Cell Biol. 97, 258a. LEMANSKI, L. F. (1979). Role of tropomyosin in actin filament formation in embryonic salamander heart cells. J. Cell Biol. 82, LOOKER, R H. (1984). The role of gap filaments in muscle and meat. Food Microstructure, vol. 3, pp SEM Inc., Chicago. LOCKER, R. H. AND LEET, N. G. (1975). Histology of highly-stretched single fibres. J. Ultrastruct. Res. 52, MAGID, A., TING-BEALL, H. P., CARVELL, M., KONTIS, T. AND LUCAVECHE, C. (1984). Connecting filaments, core filaments and side-struts, a proposal to add three new load-bearing structures to the sliding filament model. In Contractile Mechanism in Muscle (ed. Pollack, G. H. and Sugi, H.), pp Plenum Press, New York. MARUYAMA, K. (1986). Connectin, an elastic filamentous protein of striated muscle. Int. Rev Cytol. 104, MARUYAMA, K, MATSUNO, A., HIGUCHI, H., SHIMAOKA, S., KIMURA, S. AND SHIMIZU, T. (1989). Behavior of connectin (titin) and nebulin in skinned muscle fibres released after extreme stretch as revealed by immunoelectron microscopy. J. Muscle Res. Cell Motil. 10, MARUYAMA, K., SAWADA, H., KIMURA, S., OHASHI, K., HIGUCHI, H. AND UMAZUME, Y. (1984). Connectin filaments in stretched skinned fibres of frog skeletal muscle. J. Cell Biol. 99, PIEROBON-BORMIOLI, S., BETTO, R. AND SALVIATI, G. (1990). The organization of titin (connectin) and nebulin in the sarcomere: an immunolocalization study. J. Muscle Res. Cell Motil. 10, RAMIREZ-MITCHELL, R., WRIGHT, J. AND WANG, K. (1990). Image processing of immunogold-labelled titin filaments in skeletal muscle. Proc. Int. Congr. Electron Microsc. XII vol. 3, pp San Francisco Press, San Francisco REEDY, M. K. AND REEDY, M. C. (1985). Rigor cross-bridge structure in tilted single filament layers and flared-x formations from insect flight muscle. J. molec. Biol. 185, SJOSTRAND, F. (1962). The connections between A- and I-band filaments in striated frog muscle. J Ultrastruct. Res. 7, TRINICK, J. A., KNIGHT, P. AND WHITING, A. (1984). Purification and properties of native titin. J. molec. Biol. 180, TROMBITAS, K., BAATSEN, P. H. W. W., LIN, J J. C, LEMANSKI, L. F. AND POLLACK, G. H. (1990a). Immunoelectron microscopic observations on tropomyosin localization in striated muscle. J. Muscle Res. Cell Motil. 11, TROMBITAS, K., POLLACK, G. H., WRIGHT, J. AND WANG, K (19906). Elastic behavior and arrangement of titin filaments in the I-band. Proc. Int. Congr. Electron Microsc. XII. vol. 3, pp San Francisco Press, San Francisco. WANG, K. (1985). Sarcomere associated cytoskeletal lattices in striated muscle. In Cell and Muscle Motility (ed. by Shay, J. W.), pp New York, Plenum Press. WANG, K., RAMIREZ-MITCHELL, R. AND PALTER, D. (1984a) Titin is an extraordinarily long, flexible and slender myofibrillar protein. Proc. natn. Acad. Sci. U.S.A. 81, WANG, K. AND WRIGHT, J. (1988). Sarcomere matrix of skeletal muscle: The role of thick filaments in the segmental extensibility of elastic titin filaments. Biophys. J. 53, 25a. WANG, K., WRIGHT, J. AND RAMIREZ-MITCHELL, R. (19846). Architecture of the titin/nebulin containing cytoskeletal lattice of the striated muscle sarcomere. Evidence of elastic and inelastic domains of the bipolar filaments. J. Cell Biol. 99, 435a (Received 13 May Accepted 19 August 1991) 814 K. Trombitds et al.