Mitochondrial size and shape in equine skeletal muscleA three-dimensional reconstruction study.код для вставкиСкачать
THE ANATOMICAL RECORD 222:333-339 (1988) Mitochondria1 Size and Shape in Equine Skeletal Muscle: A Three-Dimensional Reconstruction Study S.R. KAYAR, H. HOPPELER, L. MERMOD, ANDE.R.WEIBEL Institute of Anatomy, University of Bern, CH3000 Bern, Switzerland ABSTRACT Individual mitochondria were reconstructed from ultrathin serial sections of selected muscle fibers in the M. semitendinosus of a horse, over a length of nearly two sarcomeres. Mitochondria were found to be highly variable, with size and complexity of single mitochondria increasing with the fractional part of a fiber occupied by mitochondria. In fibers with a mitochondrial volume density of less than 4%, corresponding to the mitochondrial content of fast-twitch glycolytic fibers, mitochondria were generally rather simple cylindrical shapes, oriented parallel to the myofibrils. In fibers with a mitochondrial volume density of more than 7%, corresponding to the mitochondrial content of slow-oxidative or fast-oxidative glycolytic fibers, mitochondria were generally cylindrical at the A-band and Z-plate level of the muscle sarcomeres. However, these mitochondria often had transverse extensions or interconnections that occurred a t the I-band level. Volumes of individual mitochondria ranged from as small a s a few thousandths of a pm3 up to several pm3 for the incompletely reconstructed portions of the largest mitochondria. Mitochondrial profiles that one would classify from single sections as subsarcolemmal were found to interconnect with other profiles deeper within the fiber. This suggests that it is unlikely that subsarcolemmal and interfibrillar mitochondria are two structurally distinct populations. However, we found no evidence of a reticulum completely interlinking all mitochondrial material in a muscle fiber. The three-dimensional shape and volume of individual muscle mitochondria have recently become important in estimating oxygen gradients in muscle cells (Clark and Clark, 1985; Hoofd, 1987). This has spurred interest in analyzing skeletal muscle mitochondria either by serial section reconstruction (Bakeeva et al., 1978, 19811, by high-voltage electron microscopy (Kirkwood et al., 1986, 1987), or by scanning electron microscopy (Ogata and Yamasaki, 1985)to obtain such information. It has long been considered by researchers (Gauthier and Padykula, 1966) that these mitochondria may be rather complex and variable in shape, despite the common textbook image of simple spheroids or sausages (Tzagoloff, 1982). Indeed it has been suggested that in some muscles all mitochondrial material is interlinked into a complex reticulum (Bakeeva et al., 1978; Kirkwood et al., 1986). The distinction between discrete, spheroidal mitochondria and a mitochondrial reticulum is important if one considers that extensive branching and interlinkage of mitochondria may be significant to intracellular transport of substances involved in cellular respiration. In the present study we analyzed the structure of mitochondria in four fibers of horse semitendinosus by serial section reconstruction, over a distance of approximately 2 sarcomeres, to determine mitochondrial connectivity. Horse semitendinosus was selected as a muscle well suited to this work. Since horses are highly aerobic athletes, their muscles are particularly interesting to analyze biochemically and ultrastructurally, and the semitendinosus is considered to have a typical function 0 1988 ALAN R. LISS, INC in hindlimb movement (Straub et al., 1983). The fiber type composition (Hoppeler et al., 19831, myoglobin content (Kayar et al., 1988a), and activities of several key metabolic enzymes (Essen-Gustavsson et al., 1983; Kayar et al., 1988a) have all been well characterized in horse semitendinosus. The average distribution of mitochondrial particles in cross sections of muscle fibers has also been analyzed in this muscle (Kayar et al., 1988a). The present study extends our information on muscle mitochondria to include the third dimension. MATERIALS AND METHODS Tissue Fixation One horse (Equus caballus, gelding, body weight 516 kg) was sedated, anesthetized, and euthanized by cardiac arrest. Within less than 1 hour of death, the M. semitendinosus of one leg was exposed, and the limb was held in a position that fully extended this muscle. Using a fine needle, 10 ml of a fixative solution (6.25% glutaraldehyde in 0.01 M sodium cacodylate buffer, adjusted with NaCl to 1,100 mOsm, pH 7.4) were injected into the superficial portion of this muscle over a n area of approximately 3 x 1cm and a depth of 1.5 cm. After 10-15 minutes, the injection-fixed portion of muscle was cut out, immersed in glutaraldehyde fixative, and furReceived January 22, 1988; accepted June 22, 1988. Address reprint requests to S.R. Kayar, Institute of Anatomy, Uni. versity of Bern, Biihlstrasse 26, CH3000 Bern, Switzerland. 334 S.R. KAYAR ET AL. ther processed for electron microscopy as described in detail elsewhere (Hoppeler et al., 1981; Mermod et al., 1988). This procedure was followed to ensure that the muscle fibers were fixed a t a n extended sarcomere length. Mitochondrial Reconstruction From one tissue block of the semitendinosus, four individual muscle fibers were selected for analysis of the size and three-dimensional shape of the mitochondria within them. These four fibers were chosen because their small size ( < 120 pm2 in cross-sectional area) permitted a n entire fiber profile to appear in one electron micrograph of 16,000-20,000 final magnification. For each fiber, 50-70 ultrathin (90-nm) serial sections, which spanned 1complete sarcomere and all or part of a second sarcomere (sarcomere length = 3.00 pm) were cut with an LKB ultramicrotome. A series of eight to ten sections was then cut a t 50-pm intervals, and another series of eight to ten sections was cut at 200-pm intervals. The serial reconstruction was performed using the following computer system configuration: a microcomputer (DEC, PDP-l1/23) with a terminal (DEC, VT220) and a color graphic display (Barco, CD233), with the program SSRCON (Perkins and Green, 1984). Mitochondrial volume density in each of these fibers was calculated in a subsample of 25 micrographs of all the serial sections. The injection fixation technique, which we employed to fix muscles a t an extended sarcomere length, rendered it impossible to perform definitive fiber typing; this requires fresh-frozen tissue (Hoppeler et al., 1983). However, in horse semitendinosus it has been demonstrated that there is a distinction between the mitochondrial volume density of type IIb fast glycolytic fibers, versus the type I slow-oxidative and type IIa fast-oxidative glycolytic fibers. The type IIb fibers have mitochondrial volume densities of 1-5%, whereas the type I and IIa fibers have mitochondrial volume densities from 5% to more than 8%. There is no significant difference in the mitochondrial content of type I and IIb fibers. Thus while we cannot unequivocally demonstrate the histochemical types of fibers analyzed in this study, the range of mitochondrial volume densities in the selected fibers makes it probable that two were type IIb and two were either type I or IIa. RESULTS The four individual fibers of semitendinosus selected for three-dimensional reconstruction were of very small cross-sectional area [a(f); Table 11, to make it practical to reconstruct all the mitochondrial profiles found in the serial sections. These fibers were also selected to span a range of mitochondrial volume density [Vv(mt,fl;Table 11 from less than 2% to 996, and therefore to include at least two fiber types (Hoppeler et al., 1983). An electron micrograph of one section of fiber A appears in Figure l a , along with the fiber and mitochondrial profiles outlined for the serial reconstruction (Fig. lb). This is probably either a type I or 1Ia fiber, given its high volume density of mitochondria. Individual mitochondrial profiles appear to be either circular or relatively simple elongate ovoids. When this fiber was reconstructed from 50 serial sections (spanning a vertical distance of 4.5 pm), a general cross-sectional view shows many interconnections between the mitochon- TABLE 1. Mean fiber cross-sectional area [a(D] and volume density of mitochondria [V,(mt,D] in four selected fibers of horse M. semitendinosus Fiber A B C D am (urn2) V,(mt,D 120 + 8 51 + _ 4 88 +_ 3 74 + 7 9.03 k 0.53 7.91 +_ 0.29 1.74 k 0.23 3.42 + 0.37 (5%) drial profiles (Fig. lc). Two mitochondria in this fiber have been selected for detailed reconstruction (Figs. Id and 2). The first mitochondrion (Fig. 2a) is located near the sarcolemma. In its lowest portion, which corresponds to the A-band level of the muscle sarcomere, it consists of unbranched cylinders. Farther up, corresponding to the I-band level of the muscle, transverse connections between cylinders can be seen not only along the sarcolemma, but also deeper into the fiber. Note the brief interruption in the horizontal connections, which occurs at the Z-disk level. At the top of the figure, corresponding to the next A-band of the muscle, the mitochondrion returns to its cylindrical pattern. All profiles in Figure 2a appear interconnected and can be assumed to extend beyond the upper and lower bounds of the series of sections. The volume of the mitochondrion that is included in these sections is 6.73 pm3, with a n unknown additional volume beyond the serial sections. It is particularly noteworthy in this reconstruction that a number of mitochondrial profiles that would be counted as subsarcolemmal mitochondria are connected to what would appear on single cross sections to be interfibrillar mitochondria. The second mitochondrion reconstructed from fiber A (Fig. 2b) is strictly interfibrillar. The sarcomeres in this fiber, a s expected (Peachey and Eisenberg, 19781, were staggered across the fiber profile. Thus the second reconstruction happens to include I-bands in the upper and lower portions of the figure, whereas the first mitochondrion (Fig. 2a) includes a n I-band in the center of the reconstruction. Just as in the preceding mitochondrion, a t the I-band level of the muscle there is extensive transverse branching, with a short interruption in the branching at the Z-disk. When it reaches the A-band level, the network separates and forms cylindrical columns running parallel to the muscle fiber axis. The estimated volume of the reconstructed portion of this mitochondrion is 2.70 pm3. Fig. 1 , Fiber A, horse M. semitendinosus. a: Electron micrograph of a single ultrathin (90 nm) section. b: Profiles of fiber and mitochondrial particles from this section, for computer reconstruction. Profiles marked in red and blue are shown reconstructed in subsequent figures. c: Computer reconstruction of 50 ultrathin serial sections (4.5 p m ) , viewed from above. All mitochondrial profiles shown in red or blue are interconnected with each other. Outlined in white are all other mitochondria] profiles on the first and every tenth section. d: Same computer reconstruction, viewed from a -60" rotation of the x axis. 336 S.R. KAYAR ET AL. In fiber B (Fig. 3a), a general cross-sectional view shows fewer transverse connections than in fiber A, and mitochondrial volume density is slightly lower (Table 1). It is probably a type I or IIa fiber (Hoppeler et al., 1983). Two mitochondria, one subsarcolemmal and one interfibrillar, are shown reconstructed from 66 serial sections (a vertical span of 6 pm; Fig. 3b). The subsarcolemmal mitochondrion is a simple cylinder, with only a slight enlargement a t the level of the I-band. The volume of the portion illustrated is 0.361 pm3, but this mitochondrion presumably extends beyond the upper and lower planes of sectioning. The interfibrillar mitochondrion is composed of two cylindrical portions that are parallel to the fiber axis in the A-band regions of the muscle. They are interconnected in the center of the figure a t the level of the I-bands within the muscle by two transverse portions, one otcurring just above and the other just below the Z-disk. There is another transverse interconnection at the top of the diagram, corresponding to the next Iband region. The volume included within these sections is 1.56 pm3. Fibers C and D have a very low mitochondrial content (Table 1)and are therefore probably type IIb fibers (Hoppeler et al., 1983). There are very few connections between mitochondria in these fibers. In fiber C (Fig. 4), approximately 30 individual mitochondria were identified from the reconstruction of 61 serial sections (vertical span of 5.5 pm). Most of them appear as cylinders of varying diameter, some of which begin andlor end within the section series. Only eight mitochondria show transverse linkages, which occur invariably a t the level of the I-band. For the single mitochondrion that begins and ends within these sections, the estimated volume is 0.183 pm3. In fiber D, while individual sections show numerous small mitochondrial profiles as usual (Fig. 5a), there were only ten mitochondria that actually began and ended within the reconstruction of 61 sections (5.5 pm vertical span; Fig. 5b). Volumes of these discrete mitochondria were between 0.004 and 0.083 pm3. DISCUSSION Skeletal muscle mitochondria seem to be functionally similar among a wide variety of mammalian species and muscles (Hoppeler et al., 1981).Biochemical studies have indicated that the respiratory complexes and ATP synthetase enzymes are integrated in the mitochondrial inner membrane a t a density set by the crystalline dimensions of the enzyme molecules and the concentration of lipid in the membranes (Tzagoloff, 1982). The respiratory capacity of the mitochondria should therefore be. set by the surface density of the inner membranes. Morphometric analysis of mitochondria from a variety of mammalian muscles has shown that this density of membranes is approximately constant (Hoppeler et al., 1981). While horses have a high maximal oxygen consumption rate during exercise, this rate can be accounted for by the high total volume of mitochondria in the muscles (Kayar et al., 198813). Thus we expect that this study of mitochondrial structure in horse muscle will be applicable to other mammalian muscles as well. Our primary finding in this study is that mitochondrial size and shape are far more variable and complex than present models of muscle energetics allow. Despite the variabilitv. there are some consistent features of mitochondrial" shape relative to the muscle sarcomeres adjacent to them; despite the complexity, discrete mitochondria can be found. In particular, we found that the mitochondrial material in the muscle fibers we analyzed did not exist as a continuous mitochondrial reticulum, but that the degree of mitochondrial connectivity appeared to be related to mitochondrial volume density. By using serial section reconstruction, we were able to estimate the volume of individual mitochondria. The conventional view of muscle mitochondrial structure is that they are discrete spheres or ellipsoids. Such a view is easily reinforced by examining single micrographs of muscle in transverse or longitudinal section, where circular and ovoid profiles are common. However, in the muscle fibers examined in our study, only one roughly spheroid mitochondrion was found (fiber D, Fig. 51, with a volume (0.004 pm3) the smallest of all those we measured. At low volume density, i.e., low fractional part of a fiber occupied by mitochondria, most mitochondria were discrete, unbranched cylindrical bodies, oriented parallel to the muscle fibers. As volume density increased, many mitochondria became larger and more complex in shape, with cylindrical portions a t the Aband and Z-disk levels of the sarcomeres, which intricately interconnected transversely at the I-band level of the muscle. Many studies of mitochondria in muscle cross sections have noted that there are large accumulations of mitochondrial particles immediately beneath the sarcolemma. In some experiments, as for example with endurance exercise training, it can be demonstrated that there are greater changes in the volume density of these subsarcolemmal mitochondria than in the mitochondrial profiles deeper within the fiber (Hoppeler, 1986). However, in our reconstructions we found that in a fiber with relatively high mitochondrial volume density, there are numerous connections between mitochondrial particles that one would classify from single micrographs as subsarcolemmal or interfibrillar (Fig. 1). A similar observation was made by Kirkwood et al. (1986, 1987) in rat muscle. It does not seem to be justified on structural grounds to divide mitochondria into two truly separate populations, subsarcolemmal and interfibrillar, if their outer membranes and hence presumably their inner membranes and matrix space are continuous (Matlib et al., 1981). Mitochondria1 reticula in muscle have been described previously. However, we did not find any evidence of a continuous reticulum linking all mitochondrial material, as described by Bakeeva et al. (1978, 1981) and Kirkwood et al. (1986, 1987) in rat muscle. The latter group has suggested that mitochondria that appear to be discrete particles within a limited range of sections are in fact interconnected with the reticulum in another plane. This does not appear to be the case in the muscle fibers analyzed in our study, since in the fibers with low mitochondrial volume density we found individual mitochondria that definitely began and terminated within less than 2 sarcomeres. In the fibers with high mitochondrial volume density and branching mitochondria, these did not all interconnect within the fiber portion reconstructed. There is little reason to think that these complex mitochondria are completely interconnected with each other beyond the reconstructed region, since further sectioning of the same fibers from 50 pm up to 2,000 pm beyond the original region shows virtually identical mitochondrial locations. One reason for the different observations of our studv versus those of Bakeeva et al. (1978) and Kirkwoo$ et al. (1986) may be that the horse muscle fibers of our study had a relatively MITOCHONDRIAL SIZE AND SHAPE 337 Fig. 2. Detailed reconstruction of two portions of mitochondria as viewed from 50 ultrathin serial sections of fiber A, horse M. semitendinosus. a: Primarily subsarcolemmal mitochondrion (also shown in red in Fig. 1).b: Interfibrillar mitochondrion (also shown in blue in Fig. 1).All profiles that appeared interconnected to those shown in color in Figure l b are included in the illustrations, but both mitochondria extend beyond the first and last sections of this series. Approximate locations of the Z-band in the muscle sarcomeres are indicated by the white arrows. Fig. 3. Fiber B, horse M. semitendinosus. a: Overview of 66 ultrathin serial sections (6 pm). Marked in green and yellow are two portions of mitochondria selected for detailed reconstruction. Outlined in white are the profiles of all other mitochondrial particles on the first and every tenth section. b: Same reconstruction, viewed from a -60" rotation of the x axis. Green and yellow mitochondria illustrate all profiles appearing interconnected to each other within these sections, but both mitochondria extend beyond the first and last sections of this,series. Outlined in white are all other mitochondrial profiles in the topmost section. White arrows indicate approximate locations of Z-bands in the muscle sarcomeres. Fig. 4. Fiber C , horse M. semitendinosus. Reconstruction of 61 ultra- within these sections, only one (illustrated in red near the center of thin serial sections (5.5 Gm), viewed from a -60" rotation of the x axis. the fiber) began and ended within this fiber volume. Of the approximately 30 discrete portions of mitochondria appearing 338 S.R. KAYAR ET AL. Fig. 5. Fiber D, horse M. semitendinosus. a: Single sections, numbers 1and 60, viewed from a -60" rotation of the x axis. All mitochondria1 profiles on these two sections are outlined. b: Reconstruction of 61 ultrathin serial sections (5.5 pni) viewed from a -60" rotation of the x axis. Only the mitochondria that began and ended within this tissue volume are illustrated. Note the yellow profile in the upper right corner of the fiber, appearing on only one tissue section. It is the only mitochondrion identified in this study that appears to have a spherical form. low volume density of mitochondria compared with rat muscles. Bakeeva et al. (1978) have reported that in the ontogeny of the diaphragm in rat, embryonic rats have discrete mitochondria that merge into a continuous reticulum within the first 2 postnatal months as mitochondrial volume density increases. Mitochondrial volume density in an adult rat diaphragm is reported to be roughly 20% (Hoppeler et al., 1981). For comparison, human skeletal muscle fibers are generally on the order of 2 4 % mitochondria (Hoppeler, 1986). One serious limitation to any structural study of mitochondria by three-dimensional reconstruction is that the resolution is set by the thickness of the tissue sections. Mitochondrial projections that are vertically aligned and approach each other within the plane of sectioning may be incorrectly viewed as being continuous if the distance between their outer membranes is less than the section thickness. Our sections were - 90 nm; those of Bakeeva et al. (1978,1981) were 60-70 nm, and those of Kirkwood et al. (1986) were 500 nm. Technical limitations make it difficult to cut serial muscle sections significantly thinner than 60-70 nm. On the other hand, morphometry of particles the size of mitochondria is not possible €or sections of several hundred 339 MITOCHONDRIAL SIZE AND SHAPE nm thickness because of overprojection (Weibel, 1979). The effect of overprojection of irregularly shaped particles in thicksections is readily apparent in a n overview such as in Figure lc, spanning 4,500 nm, or 50 thin sections, compared with a single thin section (Fig. la,b). The “thick section” view gives a n impression of mitochondrial content that is fallaciously high and cannot be corrected. Examination of these figures also makes clear the futility of trying to count mitochondrial number or size from individual sections, regardless of section thickness. There is a difference in the interpretation of the structural findings of Kirkwood et al. (1986) versus those of our study, which is important for considering substrate fluxes within muscle fibers. Our reconstructions show that the primary axis of interconnections within mitochondria is parallel to muscle fibers and capillaries, with a lesser degree of interconnections radially from capillaries, whereas Kirkwood et al. (1986, Fig. 5) report the contrary. It has been hypothesized (Longmuir, 1981) that oxygen, given its 8 times greater solubility in lipids than in cytosol, will diffuse through a cell more rapidly by following a biological membrane than by traveling freely through the cytosol. If mitochondria are completely transversely interconnected, one could envision this as facilitating transport of oxygen and perhaps other metabolically important substances radially from capillaries to mitochondria throughout the fiber (Kirkwood et al., 1986). If mitochondria are relatively little interconnected transversely, it is not clear what effect longitudinal interconnections of mitochondria would have on substrate flux. The complexity and variability of mitochondrial shape and size within a single muscle fiber suggests a definite word of caution to modelers of oxygen transport to tissues: individual muscle mitochondria can differ in volume by at least three orders of magnitude. Any model that is highly sensitive to this parameter must bracket a n enormous range. A model that requires only mitochondrial volume density or total mitochondrial volume in a muscle and average radial distribution will be subject to far smaller variance. A recent study in human muscle (Kawai et al., 1987) has offered the provocative observation that myoglobin is present in I-bands but not A-bands, and seems to have a n affinity for mitochondrial outer membranes. The localization of myoglobin in regions of muscle with relatively high concentrations of mitochondria could seriously affect the modeling of oxygen transport in muscle. Thus we need more information, particularly including a quantitative estimate of mitochondrial connectivity, before it is possible to construct a comprehensive model for cellular level aerobic respiration. ACKNOWLEDGMENTS The animal used in this study was from a series of experiments performed in the Department of Veterinary Medicine of the Swedish School of Agricultural Sciences, Uppsala, Sweden. We thank their faculty and staff for support and facilities. We are especially grateful for the excellent technical assistance of Ms. H. Claassen and Mr. M. Diezi. This work was supported by Swiss National Science Foundation Grant 3.036.84. LITERATURE CITED Bakeeva, L.E., Y.S. Chentsov, and V.P. Skulachev 1978 Mitochondrial framework (reticulum mitochondriale) in rat diaphragm muscle. Biochim. Biophys. Acta, 501:349-369. Bakeeva, L.E., Y.S. Chentsov, and V.P. Skulachev 1981 Ontogenesis of mitochondrial reticulum in rat diaphragm muscle. Eur. J. Cell Biol., 25t175-181. Clark Jr, A., and P.A.A. Clark 1985 Local oxygen gradients near isolated mitochondria. Biophys. J., 48.931-938. Essen-Custavsson, B., A. Lindholm, D. McMiken, S.G.B. Persson, and J. Thornton 1983 Skeletal muscle characteristics of young standardbreds in relation to growth and early training. In: Equine Exercise Physiology. D.H. Snow, S.G.B. Persson, and R.J. Rose, eds. Burlington Press, Cambridge, pp, 200-210. Gauthier, G.F., and H.A. Padykula 1966 Cytological studies of fiber types in skeletal muscle. A comparative study of mammalian diaphragm. J. Cell. Biol., 28t333-354. Hoofd, L. 1987 Facilitated diffusion of oxygen in tissue and model systems. Dissertation, Catholic University of Nijmegen, Netherlands, 170 pp. Hoppeler, H. 1986 Exercise-induced ultrastructural changes in skeletal muscle. Int. J. Sport Med., 7t185-202. Hoppeler, H., 0. Mathieu, R. Krauer, H. Claassen, R.B. Armstrong, and E.R. Weibel 1981 Design of the mammalian respiratory system. VI. Distribution of mitochondria and capillaries in various muscles. Respir. Physiol., 44237-111. Hoppeler, H., H. Claassen, H. Howald, and R. Straub 1983 Correlated histochemistry and morphometry in equine skeletal muscle. In: Equine Exercise Physiology. D.H. Snow, S.G.B. Pcrsson, and R.J. Rose, cds. Burlington Press, Cambridge, pp. 184-192. Kawai, H., H. Nishino, Y. Nishida, K. Masuda, and S. Saito 1987 Localization of myoglobin in human muscle cells by immunoelectron microscopy. Muscle Nerve, 10:144-149. Kayar, S.R., H. Hoppeler, B. Essen-Gustavsson, and K. Schwerzmann 1988a The similarity of mitochondrial distribution in equine skeletal muscles of differing oxidative capacity. J. Exp. Biol. 137t253263. Kayar, S.R., H. Hoppeler, S.L. Lindstedt, H. Claassen, J.H. Jones, 3. Essen-Gustavsson, and C.R. Taylor 1988b Total muscle mitochondrial volume in relation to aerobic capacity of horses and steers. Miigers Arch. (in press). Kirkwood, S.P., E.A. Munn, and G.A. Brooks 1986 Mitochondrial reticulum in limb skeletal muscle. Am. J. Physiol., 251:C395-C402. Kirkwood, S.P., L. Packer, and G.A. Brooks 1987 Effects of endurance training on a mitochondrial reticulum in limb skeletal muscle. Arch. Biochem. Biophys., 255:80-88. Longmuir, I.S. 1981 Channels of oxygen transport from blood to mitochondria. In: Advances in Physiological Science, Vol. 25. Oxygen Transport to Tissue. A.G.B. Kovach, E. Dora, M. Kessler, and LA. Silver, eds., pp. 19-22. Pergamon Press, Budapest. Matlib, M.A., D. Rebmann, M. Ashraf, W. Rouslin, and A. Schwartz 1981 Differential activities of putative subsarcolemmal and interfibrillar mitochondria from cardiac muscle. J. Mol. Cell. Cardiol., 13:163-170. Mermod, L., H. Hoppeler, S.R. Kayar, R. Straub, and E.R. Weibel 1988 Variability of fiber size, capillary density and capillary length related to muscle fixation procedures. Anat. Rec. (in press). Qgata, T., and Y. Yamasaki 1985 Scanning electron-microscopic studies on the three-dimensional structure of mitochondria in the mammalian red, white and intermediate muscle fibers. Cell Tissue Res., 241:251-256. Peachey, L.D., and B.R. Eisenberg 1978 Helicoids in the T system and striations of frog skeletal muscle fibers seen by high voltage electron microscopy. Biophys. J., 22t145-154. Perkins W.J., and R.J. Green 1982 Three-dimensional reconstruction of biological sections. J. Biomed. Eng., 4:37-43. Straub, R., M. Dettwiler, H. Hoppeler, and H. Claassen 1983 The use of morphometry and enzyme activity measurements in skeletal muscles for the assessment of the working capacity of horses. In: Equine Exercise Physiology. D.H. Snow, S.G.B. Persson and R.J. Rose, eds. Burlington Press, Cambridge, pp. 193-199. Tzagoloff, A. 1982 Mitochondria. Plenum Press, New York, London. Weibel, E.R. 1979 Stereological Methods. Vol. I: Practical Methods for Biological Morphometry. Academic Press, London, chapters 4 and 6.