Changes in striated muscle fibres during contraction and growth with particular reference to myofibril splitting

Recently, the author (Goldspink, 1970) reported that during the post-natal growth of the mouse biceps brachii muscle the myofibrils within an individual muscle fibre may increase in number by as much as 15-fold. Evidence obtained by examining muscle fibres at different stages of growth strongly suggests that this increase in number is due to the longitudinal splitting of the myofibrils when they attain a certain size. One possible mechanism for the splitting is suggested by the observation that the peripheral actin filaments of the myofibril always appear to be pulled slightly obliquely to the Z-disk axis so that when the tension developed by 2 adjacent half sarcomeres is great enough, the lateral force resulting from this oblique pull would cause the Z-disk to rip. The oblique pull of the actin filament is presumably due to a discrepancy in the actin lattice spacing at the Z-disk and at A-band level. In mammalian muscles the actin lattice is known to be square or slightly rhombic at the Z-disk (Knappeis & Carlson, 1963; Reedy, 1964) and hexagonal at the A-band (Huxley, 1957). This means that there must be some displacement of most of the actin filaments as they run from the Z-disk to the A-band. Pringle (1968) has shown that it is theoretically possible to transform a rhombic lattice into a hexagonal lattice by an equidistant displacement of each actin filament (see Fig. 1, p. 128). The dimensions of the rhombic and hexagonal lattices are critical if such a transformation is to take place with the equidistant displacement of the actin filaments. However, the spacing of the hexagonal myosin filament lattice varies according to the degree of contraction of the myofibril (Huxley, 1953; Brandt, Lopez, Reuben & Grundfest, 1967; Elliott, Lowy & Worthington, 1963; Elliott, Lowy & Millman, 1967) and therefore a mismatch of the Z- and A-band lattices may occur at the shorter sarcomere lengths. The purpose of the present work was to obtain further information about the splitting process, in particular to measure the lattice dimensions of myofibrils at different ages and at different states of con traction.

The significance of myofibril splitting appears to be that it allows the transverse tubular system and sarcoplasmic reticulum to invade the myofibril mass. This is most important in the case of fast muscles which require rapid activation and relaxation of their contractile apparatus. It was therefore felt that the amount of transverse tubular system and sarcoplasmic reticulum should be measured to see whether it develops to the same extent during growth as the myofibrils. In addition to these measurements, it seems appropriate to include some further qualitative observations on myofibril splitting and myofibril shape and size in different fibres at different ages.

The animals used in this study were all homozygous normal male mice (dydy) of the I29/Re strain originally obtained from the Jackson Memorial Laboratories, Bar Harbor, U.S.A. They were maintained on a formula 41 b diet (Oxoid Ltd.), with food and water available at all times. The main muscle chosen for study was the biceps brachii muscle; however, the soleus muscle was also used for the purpose of comparison as this muscle is a slow red muscle, whereas the biceps brachii is a typical fast white muscle.

Mice of different ages were killed by dislocation of the cervical vertebrae and pinned to a cork board with their forelimbs in the outstretched position and with the limb at 90° to the body axis. The skin was removed and the biceps brachii exposed and freed from overlying connective tissue. The muscle was fixed at this limb length by pipetting 2·5 % glutaraldehyde fixative containing 0·5 % glucose in a 0·062 M phosphate buffer, pH 7·3, over the muscle for 10–15 m>n at room temperature. It was then removed from the limb and immersed in fixative maintained at 4 °C for a further 100 min. In the case of the soleus, the lower limb was pinned to a cork board with the foot at go ° to the tibia. The gastrocnemius muscle was removed to expose the soleus, which was then fixed in the same manner as the biceps brachii. However, one difficulty en-countered in this investigation arose from the fact that the younger mice have a shorter sarcomere length for a given limb position (Goldspink, 1968). Therefore it was necessary to stretch out the muscles from very young mice after dislocating the elbow or ankle joints to obtain sarcomere lengths comparable to those of the older mice. As described below, the sarcomere lengths were checked in each case from longitudinal sections of the fibres.

To investigate the effect of the degree of contraction or stretch, some muscles were fixed at different initial lengths. To obtain shortened muscles the animals were first anaesthetized with nembutal (Pilgrim & De Ome, 1955) and the muscles made to contract by stimulating them electrically. One tendon of the muscle was attached to an isotonic myograph apparatus loaded with a 2-g weight and the contraction recorded on a physiograph pen recorder (E. & M. Ltd., Texas, U.S.A.). The muscle was then fixed when fully shortened by pipetting glutaraldehyde fixative straight on to it. The glutaraldehyde appeared to penetrate and fix the muscle very rapidly as no change in the length of the muscle was registered on the pen recorder. To obtain muscles of length greater than the normal resting length, the animals were first killed and the muscles stretched by pulling out the limb after dislocating the elbow or ankle joint. They were then fixed in this extended position. In both cases the muscles were dissected from the limb after the initial 15-min fixation and immersed in fixative maintained at 4 °C for a further 100 min.

After fixation the specimens were washed in cold (4 °C) 0·1 M phosphate buffer at pH 7·3 for several hours and post-fixed in cold 2 % OsO₄ (Palade’s fixative) for 2 h. They were dehydrated in ethanol, partly teased apart in 50% acetone/50% Araldite mixture and transferred to 100% Araldite. Single fibres were then separated from the muscle-fibre mass, cut into 2 lengths and embedded in the same plastic container; after the Araldite had polymerized it was cut so that one of the fibre lengths could be sectioned transversely and the other longitudinally. The longitudinal sections were cut with the microtome knife edge parallel to the long axis of the fibre. Occasionally bundles of fibres were embedded and sectioned as these were useful for deriving information about the variation in myofibril arrangement of fibres found within the same fibre bundle, i.e. large and small fibres. Sections of the Araldite-embedded material were cut at a thickness of 60–100 nm using a Reichert ultramicrotome and mounted on Celloidin-coated grids. After staining with uranyl acetate and lead citrate the sections were examined with a JEOL JEM 7 A electron microscope. Electron micrographs were taken at certain set magnifications. The microscope was calibrated at these magnifications using a carbon replica of a 2160 lines/mm diffraction grating.

In the study of the ultrastructural dimensions of fibres fixed at different initial lengths, the mean sarcomere length, the angle by which the actin filaments are displaced from a plane horizontal to the Z-disk (angle of pull) and the dimensions of the Z- and M-lines were measured from longitudinal sections. The cross-sectional areas of the myofibrils at the Z-disk and M-band levels were calculated assuming a circular cross-section and expressed as a ratio. The filament lattice dimensions were measured from transverse sections of the same muscle fibre from enlarged photographic prints. The aclin-filament lattice was measured at the Z-disk or very close to the Z-disk and the myosin-filament lattice was measured both at the edge of the A-band and at the centre or M-region of the A-band. All the electron micrographs of the lattice dimensions of a particular fibre were taken at the same time to avoid any slight inaccuracy that might arise when resetting the magnification of the electron microscope. About 80 measurements were made from several myofibrils for each fibre by measuring from filament centre to filament centre using micrometer calipers. Lattice measurements were only carried out on myofibrils in the centre of the photograph images to avoid any errors due to spherical aberration and distortion caused by the electron microscope and photographic enlarger optics. Also measurements were only carried out on regions of the myofibrils where the filament lattice appeared to be free from any distortion due to compression during sectioning or to the myofibril being sectioned slightly obliquely. The standard errors of the measurements were determined by the analysis of variance method.

For the study of the ultrastructural changes during growth only fibres with a sarcomere length between 2·6 and 3·0 μm were used. After the sarcomere length had been obtained from the longitudinal sections the rest of the measurements were taken from the transverse sections. The cross-sectional area of the fibre was determined from micrographs using a planimeter. With the larger fibres, difficulty was experienced in finding sections which were not partly obscured by grid bars and also in photographing the fibre at the necessarily low magnifications on the electron microscope. Therefore 1-μm sections were cut from the same block and these were photographed under the light microscope after staining in a toluidine blue/Best’s carmine mixture. In order to estimate the percentage area and dimensions of the myfibrils and tubular system (sarcoplasmic reticulum plus transverse tubular system) electron micrographs of trans verse sections at a magnification of 30 000 were subjected to linear analysis (Smith & Guttman, 1953; Weibel, 1969) using a 10 × 10 cm grid divided into 100 squares on transparent plastic. The number of intersections and the length of the intersections made by the grid lines on the myofibrillar and tubular components (transverse tubular and sarcoplasmic reticulum) were measured for 3 different areas of the same fibre. The percentage length of the intersections to the total length of the grid lines (220 cm) gave the percentage area of the component in the fibre. The area to surface ratio for the component was calculated using the formula of Smith & Guttman (1953). This method of analysis was checked by drawing circles and squares of known size on to graph paper and analysing these in the same way as the electron micrographs were analysed. The measured dimensions usually agreed with the known dimensions to within 1 or 2%. It was concluded that linear analysis as applied in this case was not only very convenient but also reasonably accurate.

The measurements of the ultrastructural changes associated with changes in sarcomere length are. shown in Table 1. As far as the splitting process is concerned one of the most relevant measurements is the angle of pull of the actin filaments. If the actin filaments are pulled obliquely with the sarcomere at the normal resting length then it would be expected that the angle will increase as the sarcomere shortens because the myosin filaments get farther apart and also because the distance between the Z-disk and A-band is reduced, this is assuming of course that the Z-lattice does not change. As shown in Table 1, the calculated angle of pull based on the cross-sectional area ratio of the Z:M line should increase from about 7·5 to 60·0° as the sarcomere shortens from 2·9 to 2·0 /tm. These calculations are based on the assumption that the myofibril contracts as a constant volume system except for the Z-disk. However, the actual measurements showed that the angle of pull did not increase during contraction. It did, however, decrease as a result of stretching. Although the myosin filaments do move apart at the M-line and cause the M-line to increase in girth, they are in some way pulled or held together at the edge of the A-band, resulting in the bowing of myosin filaments (see Fig. 8). When the sarcomeres are stretched the bowing decreases until the filaments are almost straight. The Z-disk as well as the M-band lattice opens up during contraction, but the lattice at the edge of the A-band appears to remain fairly constant irrespective of the sarcomere length except when the muscles are very stretched. Although the Z-lattice opens up only very slightly during contraction, the increase occurs in 2 dimensions and consequently this reduces the angle of pull of the actin filaments to a significant extent. In the case of the biceps brachii muscle the increase in the Z-lattice spacing is apparently of equal importance to the bowing of the myosin filaments in compensating for the swelling of the myosin lattice during contraction.

It is interesting to note that the Z to M lattice ratio is about 1:2·0 at the relaxed sarcomere length. For a perfect transformation from a rhombic lattice to hexagonal lattice the Z:M ratio should be 1:1·51 as derived by Pringle (1968) (see Fig. 1). However, it appears that the only condition under which the ratio is attained or nearly attained is if the sarcomeres are pulled out to such an extent that there is no overlap of the actin and myosin filaments; as shown by the data for the stretched soleus muscle given in Table 1.

The results of the ultrastructural measurements carried out on muscle fibres at different ages and stages of growth are shown in Table 2. In general the results are self-explanatory; however, some of the data do require brief comment. The area of the muscle fibre which is occupied by myofibrils is seen to vary between 60 and 85 % (except in some of the fibres from the newborn mouse in which it is slightly lower). The percentage area of the fibre that is occupied by the tubular system (sarcoplasmic reticulum plus transverse tubular system) varies between 7 and 18% (except for some of the fibres from newborn mice) and there is no discernible increase or decrease during growth. When the results are expressed as the increase in the total myofibril cross-sectional area and the increase in total surface of the tubular system (calculated using the surface to volume ratio and the percentage area measurements) then it is again seen that the tubular system does in fact develop at about the same rate as the myofibrillar material. Also there were no detectable changes in the Z- or M-lattice spacings during growth.

Some of the general observations relating to myofibril splitting are illustrated by Figs. 3–9. Figs. 3, 4 show 2 myofibrils in the process of splitting. Fig. 5 shows a myo fibril that is apparently just commencing to split as only one Z-disk has split at this stage. Figs. 6, 7 show transverse sections of Z-disks which are also apparently just commencing to split, and it can be seen that there is a small hole in the centre of the Z-disk. This is what would be expected if the peripheral actin filaments are pulled obliquely as the greatest strain would be developed at the centre of the disk. Fig. 8 is a longitudinal section of a contracted biceps brachii muscle showing the bowing of the myosin filaments.

Several observations were made regarding myofibril shape in fibres at different ages and stages of development. In general, myofibrils are polygonal and not circular in section. This shape is presumably due to the repeated splitting; however, the myo–fibrils in the small fibres tend to be more irregular in shape than those of the large fibres (Fig. 9).

The results presented in Table 2 show that the amount of contractile material (total myofibril cross-sectional area) in an individual muscle fibre of the mouse may increase by as much as 40-fold during the period from birth to maturity. If the myofibrillar material was built up as one uninterrupted mass this would present problems as far as the activation of the contractile apparatus is concerned. However, the evidence presented in this paper strongly suggests that as the contractile filament mass of the original myofibrils increases, it is fragmented into many more myofibrils because of a mismatch between the filament lattice of the Z-disk and the M-line. This mismatch causes the actin filaments to be pulled obliquely to the Z-disk axis and therefore when the filament mass attains a certain size the oblique pull is sufficient to cause the Z-disk to rip. The rip apparently commences as a small hole in the centre of the disk (Figs. 6, 7) as it is at this point that the greatest stress is developed. The rip then presumably extends across the disk with the direction of the weave of the lattice. Once a rip has occurred in one Z-disk, then the adjacent Z-disks split in like manner until the filament mass has divided longitudinally into 2 halves. This process is repeated many times to yield many myofibrils of between 0-5 and i-o/tm in diameter.

The significance of the splitting process is that it allows the sarcoplasmic reticulum and transverse tubular system to develop at the same rate as the contractile apparatus and to invade the myofibril mass. This is very important in fast muscles as each myo–fibril is surrounded by an elaborate envelope of sarcoplasmic reticulum which is responsible for the rapid activation and relaxation of the contractile apparatus. The splitting of the myofibrils also permits the mitochondria to become interspersed throughout the myofibrillar mass instead of being restricted to the periphery of the muscle fibre. This is, of course, an important factor in the availability of energy for the contraction of the myofibrils.

The main muscle used in this study, the biceps brachii muscle of the mouse, is one of the fastest muscles found in the animal kingdom. (In general the short muscles of the smaller mammals have a very high rate of shortening per sarcomere.) The myo–fibrils in slow muscles and the small fibres of fast muscles tend to be less discrete and more irregular in cross-section than in the large fibres of fast muscles (Kruger, 1950; Hess, 1961; Shear & Goldspink, 1971). It seems possible that in slow muscles the splitting of the myofibrils is less frequent and often incomplete. The reason may be that it is not just the total tension developed by 2 adjacent half sarcomeres but the rate at which tension is developed that is important in the ripping of the Z-disk. To demonstrate this effect a model experiment was set up using a Levin–Wyman ergo meter (Palmer Ltd.) and some high-grade plaited silk thread. A short length of the thread was weighted with a plastic bottle containing lead shot and the Levin–Wyman ergometer was used to lift the weight at a given rate using rates of pull over the physiological range. The weight was increased by adding more lead shot until the silk thread eventually broke. The procedure was repeated for different rates of pull and the results plotted in Fig. 2. The sigmoid curve obtained clearly demonstrates that above a certain rate of pull the stress necessary to break the thread is, in effect, greatly reduced. In other words, there is a snatch effect due to the inertia of the load and also due to there being insufficient time for the force applied to be dissipated along the entire length of the thread. The initial rate of shortening of many fast muscles, e.g. mouse extensor digitorium longus (Close, 1965), chick posterior latissimus dorsi, the hamster biceps brachii, and hamster diaphragm (Goldspink, Larson & Davies, 1970) does in fact exceed 70 mm/s; i.e. the point above which the stress necessary to break the thread drops very rapidly. Therefore it seems likely that this is the reason why the myofibrils of fast muscles are more regular than the myofibrils of the slow muscles. The more frequent and more complete longitudinal splitting of the myofibrils in turn enables the extensive development of the sarcoplasmic reticulum which is so important in fast muscles.

The use of the plaited silk thread in the model experiment was not foituitous as it can be calculated that the stress per square micron on the Z-disk filaments is about the same as that which has to be applied to break the plaited silk thread (16 mg/μm²). This calculation is based on data derived from previous work which showed that the mature biceps brachii of the mouse has approximately 1600 muscles fibres of about 960 μm² mean cross-sectional area (Rowe & Goldspink, 1969). As shown in Table 2, the per centage of the fibre occupied by myofibrils is approximately 70 % and therefore the total myofibril cross-sectional area of the muscle is about 1·07×10⁸ μm². The maximum isometric tension that the muscle can develop is approximately 15 g, thus the tension developed by the myofibril is approximately 15 μg/μm². In a previous paper (Goldspink, 1970) it was shown that myofibrils split when they attain a size greater than 1 fim. This means that the oblique force necessary to split the Z-disk must be approximately 15 μg. The resultant stress from this would be very approximately 2·0 μg, assuming a mean angle of pull of 8° (by constructing simple force diagrams). The Z-filaments are about 1 × 10^-4/tm² in cross-sectional area and therefore the breaking stress is about 20 mg/μm², which is slightly greater than that of the silk thread used (16 mg/μm²).

The ultrastructural measurement carried out on muscle fibres of different ages did not reveal any changes in the myosin or actin lattices. Although the lattice spacings do not change during growth it is apparent that at all ages there is a mismatch between the actin-filament lattice at the level of the Z-disk and the myosin-filament lattice at the M-region of each sarcomere. The measurements taken from fibres at different states of contraction revealed that the mismatch occurs at all sarcomere lengths, except perhaps when they are pulled out to such an extent that there is no overlap of the filaments. All the measurements reported here were made on tissue fixed in isotonic glutaraldehyde solution and it may be argued that the measurements do not represent the true in vivo dimensions. Even if the dimensions are not absolutely accurate it is apparent that the myosin lattice expands more than the Z-disk lattice during the shortening of the myofibril and therefore a mismatch must occur at certain sarcomere lengths.

It is interesting to note that in blowfly flight muscle the myofibrils do not split longitudinally after the early stage of formation in spite of much addition of filaments and they thus grow to a very large size, – up to 5 μm in diameter (Auber, 1969). In insect muscle, however, the Z-filament lattice matches the A-lattice exactly (Ashurst, 1967). In contrast to this, vertebrate muscle apparently has a built-in mechanical mismatch of the lattices which causes the myofibrils to split once they attain a certain size. One of the important consequences of splitting is that it allows the sarcoplasmic reticulum and transverse tubular system to develop to the same extent as the myosin and actin filament mass. However, there may also be some biochemical link between the splitting process and the synthesis of more contractile and more membrane proteins. This possibility is being investigated further.

This work was supported by research grants from the Agricultural Research Council and the Muscular Dystrophy Group of Great Britain. The author wishes to express his gratitude to Mrs Lesley Clapison and Miss Janice Haider for expert technical assistance, to Professor J. W. S. Pringle for reading the manuscript and for his helpful comments and also to Professor J. G. Phillips for his constant encouragement.