Journal of Muscle Research and Cell Motility 13, 640-653 (1992)

Ultrastructure of sarcoballs on the surface of skinned amphibian skeletal muscle fibres TREVOR M. LEWIS,# ANGELA CAROLYN STANHOPE

F. D U L H U N T Y , *

PAULINE

R. J U N A N K A R

and

Muscle Research Group, Division of Neuroscience, John Curtin School of Medical Research, ANU, PO Box 334, Canberra City, A C T 260L Australia Received 10 September 1991; revised 19 February 1992; accepted 25 March 1992

Summary The formation of sarcoballs on the surface of skinned fibres from semitendinosus muscles of Xenopus laevis, and the sarcoplasmic reticulum content of the structures, have been studied using conventional electron microscopic techniques and immunoelectron microscopy. Examination of the fibres showed many membrane-bound blebs projecting from the surface in areas where vesicles of internal membranes (including sarcoplasmic reticulum, T4ubules and mitochondria) were clustered in interfilament spaces. The blebs varied in size from 1 ~m to 150 ~m and those with diameters > 10 ~tm are referred to as sarcoballs. Small blebs were often seen in close association with each other and might have fused during sarcoball formation. The interior of the sarcoball was filled with foam-like material made up of vesicles with diameters of i00 nm to 1.0 ~tm. The sarcoplasmic reticulum membrane content of the sarcoballs was evaluated using two monoclonal antibodies, one to the Ca2+ATPase of the sarcoplasmic reticulum and the second to ryanodine receptor calcium release channels in the junctionalface membrane. The antibodies bound to some components of the surface and interior of the sarcoball, but not to mitochondrial-like structures and tubular vesicles. The results show that a large component of the sarcoball and its surface is derived from sarcoplasmic reticulum and suggest that mitochondria and T-tubules might also contribute membranes to the structures. Our hypothesis is that (a) blebs bud out from the surface of the skinned fibre following fusion of internal vesicles that are extruded along interfilament channels during unrestrained contractures, (b) blebs grow into sarcoballs by additional fusion of internal membrane vesicles and fusion of adjacent blebs, and (c) the sarcoball is a foam-like structure composed of bathing medium and membrane lipid (containing membrane proteins).

Introduction The aims of this investigation were to examine the formation of sarcoballs and to confirm that the surface of the sarcoball contains sarcoplasmic reticulum membrane. Sarcoballs are large blebs that appear along the surface of unrestrained single muscle fibres during contractures induced after the plasmalemma has been mechanically removed. The structures are large enough, and sufficiently stable, to allow single channel recording with patch-clamp electrodes (Stein & Palade, 1988). The sarcoball membrane is thought to be of sarcoplasmic reticulum origin because the ryanodine sensitive calcium channels (Stein & Palade, 1988) and large conductance chloride channels (Hals et al., 1989; Lewis & Bretag, 1991) are also seen *To whom correspondenceshould be addressed. qzPresent address: Department of Physiology, University of Adelaide, Adelaide, S.A. 5001, Australia. 0142-4319 9

1992 Chapman & Hall

when sarcoplasmic reticulum vesicles are incorporated into lipid bilayers (Miller, 1978; Coronado & Miller, 1982; Smith et al., 1985; Hamilton et al., 1989; Rousseau & Meissner, 1989). The sarcoball technique provides a unique opportunity to study single ion channels in internal membranes without microsomal isolation procedures. Before the technique was developed, ion channels in sarcoplasmic reticulum membranes were studied by either incorporating microsomal vesicles into artificial lipid bilayers (Miller, I978; Coronado & Miller, 1982; Smith et a]., 1985; Hamilton et al., 1989; Rousseau & Meissner, 1989) or forming the vesicles into bilayers of native membrane across the tip of a patch clamp pipette (Suarez-Isla et al., 1986). The disadvantage of these methods is the considerable experimental intervention necessary before single channel activity is recorded and the use of artificial lipids to support the channel in bilayer experiments: the characteristics of the channels may be very different from

Ultrastructure and sarcoplasmic reticulum content of sarcoballs those of native channels. Results obtained using the sarcoball technique are relevant to the general problem of calcium regulation in many cells. For example some neurons contain a ryanodine receptor calcium release channel which cross-reacts with monoclonal antibodies to the ryanodine receptor calcium release channel in skeletal muscle (Ellisman et al., I990). Despite the extensive use of the sarcoball technique, little is known about the formation of the structures or the source of their membranes. We have examined the ultrastructure of skinned muscle fibres and used two monoclonal antibodies specific for the Ca2+ATPase in the sarcoplasmic reticulum (Dulhunty et al., 1987; Dulhunty, 1990; Molnar et al., 1990) or the ryanodine receptor Ca z+ release channel (Dulhunty et al., 1992), as markers for sarcoplasmic reticulum and junctional-face membrane respectively. Our observations suggest that vesicles of internal membranes fuse to form the foam-like sarcoball structure. The results showed that a large component of the surface of the sarcoball and its contents are derived from sarcoplasmic reticulum membrane. Mitochondria were present in the sarcoball and might have contributed membranes to its outer surface.

Materials and m e t h o d s

PREPARATION OF 'SARCOBALLS' The technique was modified from that of Stein and Palade (1988). The semitendinosus muscle from Xenopus Iaevis was isolated and placed in a Ringer's solution containing (mM): NaCl, 117; KC1, 4.7; CaC1z, 1.2; TES (N-tris[Hydroxymethyl]methyl-2-aminoethane-sulphonic acid), 5 (pH adjusted to 7.2 with NaOH); tonicity, 240 mosM. A bundle of muscle fibres was dissected from tendon to tendon and transferred to a solution containing (mM): CsF, 200; TES, 10, adjusted to pH 7.2. One end of a single fibre was dissected and fine forceps used to peel the sarcolemma away from the fibre. Once skinned, the unrestrained fibre slowly contracted (over several seconds) and translucent hemispheres, 10-150 I.tm across (i.e. sarcoballs), formed at its surface. The mechanism stimulating contraction in CsF is uncertain and is currently under investigation. However, anion channels in the sarcoballs formed in CsF (Lewis & Bretag, 1991) are identical to those in sarcoballs formed during contractures induced by the elevation of calcium concentrations (Hals et al., 1989). It seems likely that the mechanism of sarcoball formation depends only on the unrestrained contraction of skinned fibres, not on the mechanism used to induce the contracture. Blebs of sarcolemmal membrane form on split muscle fibres without contraction (Vivaudou et al., 1991). The sarcoballs formed in CsF solutions in the present experiments were distinctly different from sarcolemmal blebs since the sarcoballs formed only during strong unrestrained contractures of skinned fibres. PREPARATION FOR ELECTRON MICROSCOPY Skinned fibres were transferred from the CsF solution to Ringer's solution just prior to fixation for electron microscopy. The sarcoball structure was found to be stable for several hours

641

in this solution, but was less stable if maintained in CsF (T. Lewis, unpublished data).

Conventional electron microscopy Skinned fibres were fixed in the Ringer's solution plus 2% glutaraldehyde for 2 h, washed in the Ringer's solution, soaked in 2% OsO4 in 0.1 M cacodylate buffer for i h, and washed and dehydrated in methanol (70-100%) for I h in each dilution at room temperature. The fibres were infiltrated with Spurs resin (1:1, v:v, in ethanol) embedded in 100% Spurs resin and set at 70~ for 8 h. Immunogold electron microscopy Skinned fibres were fixed in Ringer's solution plus 2% paraformaldehyde and 0.1% glutaraldehyde for 2 h at room temperature. The fibres were washed in the Ringer's solution for 10 min before dehydration in 90% methanol for i h. Some shrinkage was observed during fixation and dehydration, The fibres were placed in L. R. White acrylic resin (London Resin Company) overnight at 4~ and then set in resin at 60~ for at least 8 h. ANTIBODY STAINING Grey to silver sections were mounted on collodion-coated grids, blocked for 15 h with 5% skimmed milk powder (Diploma) in water and rinsed in 0.5% bovine serum albumin in phosphate buffered saline (PBS) containing in mM: NazPO 4, 7; NaH2PO 4, 2.6; NaC1, 137; pH 7. Sections were exposed for I h to the primary antibody (D12-CaZ*ATPase antibody, purified immunoglobulin filtered and diluted with PBS to a final protein concentration of 1.7 I.tg ml 1; or the ryanodine receptor antibody, 5C3, in filtered media diluted 1:10 with PBS), washed in PBS, exposed to the second antibody (5 nm gold-conjugated goat antimouse antibody: Janssen Pharmaceuticals) for lh, washed in distilled water, and stained with uranyl acetate. Sections were examined on a Hitachi H 7000 electron microscope. PREPARATION OF ANTIBODIES TO THE Ca 2+ ATPASE Swiss Outbred-Nu/+ (nude) mice were primed with pristane 2 weeks before an intraperitoneal injection of mouse hybridoma cells producing the monoclonal antibody to the CaZ+ATPase, D12 (Dulhunty et al., 1987). Ascites fluid was removed after 3 to 4 weeks, spun at 2600g and the supernatant extracted with 1,I,l-trichlorotrifloroethane to remove lipids. After precipitation in 50% NH4SO4 (pH 7.1), the immunoglobulin was further purified by HPLC on a Bakerbond ABx column. PREPARATION OF ANTIBODIES TO THE RYANODINE RECEPTOR Details of the production and characterization of the ryanodine receptor antibodies are given by Dulhunty and colleagues (1992). Briefly, Balb/C mice were immunized with junctionalface membrane which is enriched in the ryanodine receptor calcium release channel (Costello et al., 1986). Spleen cells were fused with NS-1 mouse myeloma cells and hybridomas whose media showed immunoreactivity to junctional-face membrane were cloned by limiting dilution. Media from the clones was used as a source of antibody.

642 Results Several examples of skinned muscle fibres are shown in Fig. 1. The very short sarcomere length of the fibres reflected the strong contracture that developed during mechanical skinning in the CsF solution and that was maintained after transfer to Ringer's solution. Sarcomere lengths of between 1.0 I,tm and 1.5 I-tm were commonly measured. These short sarcomere lengths can be compared with values of about 2.5 I.tm in intact muscle fibres held at rest length during fixation. Sarcoballs were always observed near areas where internal membrane vesicles were aggregated at the surface of the fibres and internal membranes had presumably been situated just below the sarcolemma before skinning. Such areas were located between bundles of myofilaments and were continuous

LEWIS, DULHUNTY, JUNANKAR and STANHOPE with the interfilament spaces throughout the fibre that contained mitochondria and sarcoplasmic reticulum. VARIETY IN THE SIZE AND SHAPE OF SARCOBALLS Membrane-bound blebs frequently protruded from the surface of the skinned fibres and assumed an enormous variety of sizes and shapes. The diameter of the blebs varied continuously from 1 I,tm to 150 lam. For the purpose of this description, blebs with diameters greater than 10 t,tm, that would normally be used for patch-clamp recording, are referred to as sarcoballs. Occasional sarcoballs were very flat, extending only 5 - I 0 t,tm away from the fibre surface, but 50-100 I,tm along its length (Fig. 1A). The sarcoball (S) shown in Fig. 1A appeared in several serial thick and thin sections and was thus also elongated around the transverse dimension of the fibre.

Fig. 1. Four examples of the variety of shapes and sizes of sarcoballs (S) on the surface of skinned muscle fibres conventionally fixed in 2% glutaraldehyde. V indicates vesicles and M indicates mitochondria. The short arrows (starred) point to areas of the fibre surface formed by a monolayer of small vesicles. Bar A 8 p.m; B 3.3 ~tm; C 2 ~m; D 6.6 I~m.

Ultrastructure and sarcoplasmic reticulum content of sarcoballs Other blebs were elongated at 45 ~ to 90 ~ to the fibre axis and some were composed of linear aggregations of smaller blebs which projected into the bathing solution (Fig. 1B). Larger sarcoballs had a roughly hemispherical cross-section (S, Fig. 1C and D) and were similar to the example shown by Stein and Palade (1988). Sarcoballs were formed during the initial contracture of the mechanically skinned fibres. The contracture normally developed over a period of seconds and sarcoballs were seen under the light microscope to appear and increase rapidly in size during this time. Neighbouring sarcoballs aggregated more slowly in the following few minutes, before and during fixation. The variety of sarcoball sizes seen under the electron microscope presumably reflects: (1) the population of sarcoballs initially formed, (2) the dynamics of sarcoball aggregation during the slower phase, and (3) any mechanical disruption that may have distorted sarcoballs when the fibres were transferred to the fixative solution. CONTENTS OF SARCOBALLS

Material fixed for conventional electron microscopy The interior of the large sarcoballs in Fig. 1B-D was full of cell debris including filamentous material. This structure was seen only with conventional fixation and standard dehydration in 70-100% methanol with i h in each dilution at room temperature. Sarcoballs fixed for immunoelectron microscopy (see Fig. 2), or conventionally fixed and dehydrated in acetone (70-100%, at 0~ for < 1 min at each dilution) were filled with vesicular material (Lewis, unpublished data). Although the appearance of the sarcoball seen in Fig. 1 may be an artefact of fixation (see Discussion), it is useful to consider the structure since it shows (a) that the vesicular interior of the sarcoball is more sensitive to dehydration procedures than the vesicles within the fibre, or than isolated sarcoplasmic reticulum vesicles, and (b) that the transition to this more fragile material occurs within a few micrometers of the fibre surface. Vesicles with diameters of 100 nm to 800 nm formed the interface between the fibre and the sarcoball in Fig. 1C and D; larger vesicles with an irregular outline and dimensions of i ~m or more projected from the clustered organelles into the sarcoball. The sarcoball contents became disrupted within 2 or 3 ~m of the fibre surface in Fig. 1C and D. Blebs and smaller sarcoballs (Fig. 1A) were not disrupted after slow dehydration in methanol and were filled with rounded vesicles with the same dimensions as those at the surface of the fibre. Small vesicles (most with diameters between 400 and 500 nm) were packed between the larger mitochondria which had diameters of 600 nm to 800 nm and contained membranous laminae. About 10% of sarcoballs were flattened and did not extend more than 5 ~m from the skinned fibre (Fig. 1A). These sarcoballs did not show any internal disruption

643

and, like small blebs, contained vesicles and mitochondria. Disruption of the sarcoball interior (Fig. 1B-D) apparently only occurred when the sarcoball extended further than 5 ~tm from the fibre surface. The dense clusters of vesicles on the fibre surface were often 4-5 ~tm thick and it is difficult to tell, in Fig. 1A for example, exactly where the fibre surface ends and the sarcoball begins.

Material fixed for immunoelectron microscopy In contrast to the conventionally-fixed methanoldehydrated material, the interior of sarcoballs that had been lightly fixed for immunoelectron microscopy was tightly packed with vesicle-like structures having diameters of 100 nm to 1 ~m (Fig. 2A and B). The overall appearance was similar to that of a foam-structure (see Discussion). The micrographs in Fig. 2 are of the surface and interior of large sarcoballs similar to that shown in Fig. 1D. The bundle of filaments (F), close to the external surface of the sarcoball in Fig. 2A, may have been pulled away from the fibre during skinning. Other sections of the same sarcoball are shown in Figs 4A, 5A and 5B. The vesicle-like structures in the sarcoball differed from those within the fibre (described below), and on the fibre surface, in that neither triads, T-tubules or terminal cistemae could be differentiated from longitudinal sarcoplasmic reticulum vesicles. Some tubular structures (T) were seen. Vesicles containing amorphous electron-dense material and internal membranes (M) were observed and thought to be from mitochondria. In contrast to the conventionally fixed material shown in Fig. 1, the surface of the sarcoball in the lightly fixed material often appeared to be made up of abutting vesicles, rather than a continuous membrane (Fig. 2A). In other places, the surface did appear to be formed by a continuous membrane, although this membrane was still closely associated with the underlying vesicle-like material (Fig. 2C). As the definition of membranes in the lightly fixed material was poor it was difficult to tell whether the vesicles that made up the surface and interior of the sarcoball were fused (i.e. separated by a single bilayer) as would be expected for a true foam structure, or whether the vesicles were tightly aggregated and each surrounded by a bilayer (with double membranes where two of the structures come into contact). The structure of vesicles within the fibre (Fig. 2D) and on the surface of the fibre was not disrupted by slow dehydration in methanol. A triad within a skinned fibre fixed for immunoelectron microscopy is shown in Fig. 2D and is similar to triads in the conventionally fixed material shown in Fig. 3. The small electron-dense particles in Fig. 2D are 5 nm gold particles marking calcium ATPase sites (see below). The definition of the membranes is clearly better after fixation with 2% glutaraldehyde followed by 2% OsO4 (Fig. 3). As seen previously (Dulhunty et al., 1987), T-tubules and terminal cisternae are more swollen after light fixation than after conventional fixation. The fact that slow methanol dehydration

644

LEWIS, DULHUNTY, JUNANKAR and STANHOPE

Fig. 2. Structure of the surface and interior of sarcoballs (A-C) and skinned muscle fibres (D) following fixation for immunoelectron microscopy. The vesicular nature of the surface and interior of sarcoballs is shown in A-C. The letter F is located on a bundle of filaments at the surface of the sarcoball in A and M indicates mitochondria. T indicates tubular vesicles in B or transverse tubules in D. The letters TC are located in the terminal cisternae. The section shown in D has been stained with primary antibody to the calcium ATPase and secondary antibody conjugated to 5 nm colloidal gold which can be seen as small electron-dense particles. Bar A 0.4 ftm; B 0.54 [.tm; C 1.6 p.m; D 0.46 ~m. did not disrupt the vesicles within the fibre suggests that these vesicles are more robust than the vesicle-like interior of the sarcoball and further that there are physical differences between the vesicles within the fibre and those within the sarcoball. The vesicles within the fibres are more reminiscent of microsomal vesicles which survive dehydration in ethanol (Saito et al., 1978).

STRUCTURE OF INTERNAL MEMBRANES CONTRACTED SKINNED MUSCLE FIBRES

IN

The longitudinal elements of sarcoplasmic reticulum appear to be very different in skinned and intact fibres. Areas that are usually occupied by longitudinal tubules of sarcoplasmic reticulum (i.e. areas aligned with the A and

Fig. 3, Structure of membrane systems deep within skinned muscle fibres following conventional fixation. The letter V is placed in vesicles of longitudinal sarcoplasmic reticulum; TC is placed in the swollen terminal cistemae and T is in a T-tubule. The short arrows in (A) and (B) point to triad junctions containing electron-dense foot structures. Bar A 0.54 p.m; B 1.0 p.m.

Ultrastructure and sarcoplasmic reticulum content of sarcoballs I bands) now contained rounded vesicles with diameters of 300-400 nm (V, Fig. 3A and B), that were slightly smaller than the 400-500 nm of the majority of surface vesicles. T-tubules remained in register with Z lines and the triads retained their three component structure. The triad junction (short arrows) was clearly marked by regular junctional feet. The T-tubules and terminal cisternae were less swollen after conventional fixation than after fixation for immunoelectron microscopy, but were more swollen in skinned fibres than in intact fibres. The short axis of the T-tubule in intact frog fibres is between 20 nm and 30 nm (Dulhunty, 1984) and terminal cisternae extend longitudinally 80-100 nm from the T-tubule. In contrast, the short axis of T-tubules in Fig. 3A and B is between 100 nm and 200 nm and the terminal cisternae extend 200-300 nm into the I band. The osmolarity of the bathing solution and fixativecontaining solutions was the same in this study and in that of Dulhunty (1984). Thus the different structure of the sarcoplasmic reticulum and T-tubules can be attributed either to the skinning procedure, the unrestrained contracture, or the presence of fluoride ions. SARCOPLASMIC RETICULUM MEMBRANE IN SARCOBALLS The electron-dense 5 nm gold particles, seen above in Fig. 2D and in Fig. 4 (arrows), were conjugated to the second antibody and mark the approximate sites of Ca2+ATPase molecules. No gold particles were seen in control experiments, where sections were stained with 5 nm gold-conjugated second antibody, in the absence of the primary DI2 antibody. Non-specific activity was occasionally seen in sections stained with primary and secondary antibodies if blocking was not successful or if the sections were contaminated. Areas of the section outside the fibre and the sarcoball (i.e. the extracellular space) were routinely checked to make sure that nonspecific labelling was low before the sections were examined further. Gold particles were seen only in the sarcoplasmic reticulum, they were not seen in mitochondria or within the myofilament lattice (Figs 4E and 2D), of fibres in sections with little or no labelling in the extracellular space (Fig. 4A-C). The labelling of antigenic sites in thin sections is always less than that expected from the amount of membrane in the section because antibodies bind only to antigens in membrane on the surface of the section. However, the density of labelling is proportional to the density of antigenic sites (Dulhunty et al., 1987). The membrane over the filaments in Fig. 4A is on the surface of the large sarcoball shown in Fig. 2A and Fig. 5A and several Ca2+ATPase sites are marked by gold particles. Similarly, the membrane in Fig. 4B and C is on the outer surface of another large sarcoball, removed from the fibre by 20-30 gm, and in an area that could be used for patch clamp recording. Gold can be seen on the surfacr of the sarcoball and some particles were located

645

in the extracellular space, slightly away from the surface of the sarcoball (Fig. 4B). Some of these apparently displaced particles could be on membrane that is either parallel to the plane of the section and thus not visible as an electron-dense line, or on fragments of membrane that are sheared away from the membrane during sectioning. On the other hand, particles may be displaced from binding sites by the dimensions of the primary and secondary antibody which separates gold labels from antigenic sites by as much as 30 nm (Dulhunty et al., 1992). Gold particles were not associated with the membranes of all the vesicles within the sarcoball. Some vesicles that lacked gold particles, and thus Ca2+ATPase, had the appearance of mitochondria (M, Fig. 4D) which do not contain the Mg z+ activated CaZ+ATPase of the sarcoplasmic reticulum. Areas of membrane on the surface and interior of sarcoballs that lacked gold particles may have been derived from organelles other than sarcoplasmic reticulum. However, it is also possible that the antigenic sites in the membrane were not on the surface of the section and therefore not available for antibody binding. It was concluded that sarcoplasmic reticulum membranes contributed to the sarcoball and that a fraction of the surface of the sarcoball may have been derived from other membrane systems. JUNCTIONAL-FACE MEMBRANE IN SARCOBALLS The antibody to the ryanodine receptor calcium release channel binds to fewer than 1% of potential ryanodine receptor sites in intact fibres in material processed in the same way as that shown in Fig. 5 (Dulhunty et al., 1992). Thus a low density of labelling was expected. The presence of gold particles does indicate that junctionalface membrane and ryanodine receptor proteins were present: it was unlikely that the particles in Fig. 5 represented non-specific labelling. No gold was seen in the extracellular space of uncontaminated 5Cdlabelled sections. As with D12-1abelling, sections showing non-specific activity in the extracellular space were rejected: all gold particles observed in sarcoballs and in the interior of the fibres in the sections that were examined were associated with membranes- they were not seen in the empty interior of the vesicle-like structures, or in the myofilament lattice- and were seen only in areas of the section that would be expected to contain junctional-face membrane. No gold labelling was seen in control sections stained with secondary antibody alone. Gold particles in Fig. 5 (thick arrows), in this case marking the site of ryanodine receptor protein, were associated with the surface of large sarcoballs and small blebs. In Fig. 5B, the ryanodine receptor site is located on vesicles on the outer surface of the large myofilamentcontaining sarcoball in Figs 2A and 4A. The surface of the sarcoball is shown in Fig. 5A, where the arrow points to the area shown at higher magnification in Fig. 5B.

646

LEWIS, DULHUNTY, JUNANKAR and STANHOPE

Fig. 4. Distribution of Ca2+ATPase molecules in sarcoballs fixed in 0.1% glutaraldehyde and 2% paraformaldehyde. The section was stained with a monoclonal antibody to the Ca2+ATPase. The small black spots indicated by the arrows are 5 nm gold particles conjugated to the second antibody. F, filaments dislocated from the fibre and sitting at the surface of a sarcoball; M, mitochondria. Bar A-C and E 0.13 ~m; D 0.2 p,m.

Figure 5C shows another gold label further along the surface of the same sarcoball. The gold particles in Fig. 5B and C were again on areas of membrane that would be used for patch clamp recording. Gold particles were also located on membranes in the vesicle-like interior of sarcoballs (Fig. 5D). DO SARCOBALLS CONTAIN MITOCHONDRIAL MEMBRANES? Although we did not have a specific mitochondrial marker, the organdies were identified by their laminated membranes and, as described above, were mixed together with other vesicles associated with blebs and sarcoballs. Some of the more striking examples of mitochondria in these structures are shown in Fig. 6. Some mitochondria at the fibre surface remained intact (M, Fig. 6A), some swollen mitochondria developed vacuoles (MV), while others appeared to be disintegrating (DM). Figure 6B

shows a dense cluster of intact, vacuolated and disintegrating mitochondria at the interface between the fibre surface and a sarcoball. The organdies were located at the fibre surface and within the sarcoballs in Fig. 6C and D. If the membranes within the sarcoball contribute to its surface as it expands, then it is possible that parts of the outer surface also contain membrane of mitochondrial origin.

FORMATION OF SARCOBALLS The apparently obligatory presence of clustered vesicles between the fibre and the sarcoball, and presence of larger vesicle-like structures within the sarcoball, suggested the first hypothesis that small vesicles of internal membranes fuse at the surface of the fibre to form larger vesicles which contribute to the interior of the sarcoball and to its surface.

Ultrastructure and sarcoplasmic reticulum content of sarcoballs

647

Fig. 5. Identification of ryanodine receptor molecules in sarcoballs fixed in 0.1% glutaraldehyde. The section was stained with a monoclonal antibody to the ryanodine receptor. The small black spots indicated by the arrows in B-D are 5 nm gold particles conjugated to the second antibody. The arrow in A points to the area of the sarcoball surface magnified in B. Bar A 0.58 I.tm; B-D 0.I ~tm.

Surface of skinned fibres

Membranous structure on the surface of skinned fibres

Examination of the surface of the fibres in areas not occupied by large sarcoballs provided a clearer indication of the mechanisms that might contribute to sarcoball formation. Bare myofilaments were never seen. A single layer of vesicles (short arrows, Fig. 1A and D) covered the myofilaments in some areas. In other areas a thicker (0.5-1.5 lam) layer of vesicles (V) and mitochondria (M), similar to the clusters of organelles associated with sarcoballs (Fig. 1), formed the surface of the fibre (Fig. 7A). Triads were often observed in interfilament spaces approaching the surface of the fibre (thick arrows, Fig. 7A), but were less often seen in the surface organelles. Tubular vesicles within blebs (TV, Fig. 7C and D) may have been of T-tubule origin. It seemed probable that the clusters of organelles at the surface of the fibres contained triads, but that these were infrequently identified because their normal orientation was lost.

A membrane-like structure covered the surface vesicles and was continuous with the membrane surrounding the sarcoball in conventionally fixed material (SM, Fig. 7A-C). This membrane was not purely residual plasmalemma as we have shown that the surface of the sarcoballs contain the Ca2+ATPase found only in sarcoplasmic reticulum (Fig. 4). We concluded that the membrane was formed from internal membranes of sarcoplasmic reticulum origin. A contribution from T-tubule vesicles and mitochondria could not be excluded.

Surface vesicles aggregate to form small blebs The membrane covering the surface organelles formed occasional small blebs, with diameters of 0.5-1 ~m, which projected from the fibre surface into the external solution (Fig. 7A and B). Vesicles and mitochondria were seen at the interface between the fibre and the blebs (Fig. 7A-D)

648

LEWIS, DULHUNTY, JUNANKAR and STANHOPE

Fig. 6. Examples of mitochondria that are intimately associated with sarcoballs in material fixed in 2% glutaraldehyde. M, intact mitochondria; MV, enlarged mitochondria containing vacuoles; DM, apparently disintegrating mitochondria. Bar A 1.0 ~tm; B 0.8 p.m; C 0.85 I~m; D 0.5 ~m.

and large vesicles can be seen within the larger blebs shown in Fig. 7C and D. These observations, coupled with the immunogold labelling of sarcoplasmic reticulum vesicles in the fibre and membranes on the surface and interior of sarcoballs, suggested that the surface vesicles aggregated and that blebs formed as this material moved away, or was pushed from the surface of the fibre during the unrestrained contracture.

Aggregation of small blebs to form sarcoballs A second mechanism for expansion of blebs to form sarcoballs is indicated by the micrographs in Fig. 8: small blebs came into close proximity and contact was seen between large areas of their surfaces. The original outlines of the small blebs in Fig. 8 can still be seen and the dark lines (indicated by long arrows) are double membranes formed where the surface of the structures touch. The boundaries of individual small blebs could not be identified in larger sarcoballs, suggesting that the structures eventually fused. Close contact between blebs was also seen in lightly fixed material and examples of contacts between small blebs on the surface of a larger sarcoball are shown in Fig. 5B and C above. Aggregation of smaller blebs to form larger blebs and sarcoballs was also observed under the light microscope following the initial contracture of muscle fibres before fixation.

Discussion

Sarcoplasmic reticulum membrane content of sarcoballs The results show conclusively that the sarcoball contained sarcoplasmic reticulum membrane as immunogold labelling of Ca~+ATPase was seen in many areas of the surface and the interior of the sarcoball. Some junctionalface membrane was labelled with an antibody to the ryanodine receptor calcium release channel. The results of immunoelectron microscopic studies must be treated with caution. Three disadvantages of the technique are: (i) labelling only of antigens on the surface of the section (Nakane, 1982), (ii) poor definition of membranes, and (iii) the separation of the gold label by as much as 30 nm from the antigenic site (Dulhunty et al., I992). However, the significant amount of specific immunolabelling with the Ca2+ATPase antibody both inside the sarcoball as well as on the surface, suggest that a large component of the structure is derived from sarcoplasmic reticulum. The presence of mitochondria at the interface between the sarcoball and the fibre, and within the sarcoball, suggested that the surface might contain mitochondrial membrane. The extent of the contribution of T-tubule and mitochondrial membranes is difficult to evaluate as we did not have specific markers for these membranes. If all internal membranes are included in the surface of the sarcoball, then the likely relative contribution of each

Ultrastructure and sarcoplasmic reticulum content of sarcoballs

649

Fig. 7. Surface of skinned muscle fibres and the formation of sarcoballs seen in conventionally fixed fibres. B, blebs; V, empty vesicles; M, mitochondria; SM, membrane covering the surface of the skinned fibres and sarcoballs; TV, tubular vesicles within sarcoballs. The long thick arrow in A points to a triadic structure. Bar A, B 0.66 ~tm; C 2 Ixm; D 0.8 txm.

type of membrane in amphibia is 73% sarcoplasmic reticulum (including terminal cisternae), 19% mitochondria and 7.9% T-tubule, based on previous measurements of membrane content in amphibian muscle fibres (Mobley & Eisenberg, 1975). A surface area of mitochondria of O.137 txm2 per ~tm 3 w a s calculated from a volume fraction of 1.6%, assuming a spherical radius of 350 nm (Fig. 5). The area of membrane in the laminae is 0.397 }xm2 per ].tm 3, assuming that there are seven disks of double membranes (Fig. 5). Therefore the total mitochondrial membrane area (surface plus laminae) is 0.534~tm 2 per ~ m 3. Despite the apparent involvement of mitochondria in sarcoball formation, definitive evidence for mitochondrial channels on the sarcoball surface has not been reported. Calcium channels in sarcoball membranes (Stein & Palade, 1988) resemble channels seen when sarcoplasmic reticulum vesicles are incorporated into lipid bilayers. However, the strong voltage dependence of the open

probability of chloride channels in sarcoballs (Hals et al., 1989; Lewis & Bretag, 1991) is not seen in chloride channels in sarcoplasmic reticulum vesicles (Hals et al., 1989; Kourie & Dulhunty, unpublished data). Although the open probability of the mitochondrial VDAC channel is strongly voltage dependent, other of its characteristics suggest that it is not the sarcoball chloride channel (see Hals et al., 1989 for discussion).

Different appearance of conventionally prepared and lightly fixed material The definition of membranes in the lightly fixed material was generally less clear than after conventional fixation. However, the vesicles within the fibre and at its surface had a generally similar appearance with all types of fixation. The most striking difference between the procedures was seen in the interior of the sarcoball. The disrupted appearance after slow dehydration in methanol can also be seen in the sarcoball shown in Stein and

650

LEWIS, DULHUNTY, JUNANKAR and STANHOPE

Fig. 8. Enlargement of blebs and sarcoballs by fusion? Examples are shown of neighbouring blebs and sarcoballs in close association, whose membranes could fuse to form a larger structure. The fibres were conventionally fixed and processed. The arrows point to dark lines seen where the membranes of adjacent discrete blebs are closely apposed. Bar A, B 3.3 I,tm; C 2.7/.tin. Palade (I988). A vesicle-like structure was seen in the lightly fixed material or when conventionally fixed material was rapidly dehydrated in acetone at 0~ Dehydration in methanol at room temperature has been shown to extract membrane lipids from lamella bodies of rat lung (Hallman et al., I976). The differences in appearance of the lightly and conventionally fixed sarcoballs shown in this

study may be accounted for by the differences in the length of time and temperature of the methanol dehydration procedure. The structure of microsomal vesicles of sarcoplasmic reticulum membrane, like the vesicles within the fibre and at its surface, is not disrupted by conventional fixation (Saito et al., 1978). This suggests that the vesicle-like material within the sarcoball might

Ultrastructure and sarcoplasmic reticulum content of sarcoballs have physical properties that are different from those of microsomal vesicles or vesicles within the fibre. Other evidence that the structure of the interior of the sarcoball is very different from that of microsomal vesicles is that isolated sarcoplasmic reticulum does not aggregate into blebs or sarcoballs with the disrupted internal structure seen after conventional fixation and dehydration with methanol. It is likely that the conditions that promote vesicle fusion (see below) and sarcoball formation are not present in isolated vesicle systems. The structure of membranes within the contracted skinned fibres was similar after both types of fixation. A vesicular appearance of the longitudinal sarcoplasmic reticulum has been described in relaxed mechanically skinned frog fibres (Ford & Surdyk, 1978) and in chemically skinned fibres, stored at -20~ for 1 week (Sorenson et al., 1980). The terminal cisternae only are swollen immediately after chemical skinning (Eastwood et al., 1979; Sorenson et al., 1980). It is not clear whether the use of a CsF solution influenced the final structure of the sarcoplasmic reticulum. Sarcoball as a foam-like structure

The interior of the sarcoball in lightly fixed material is reminiscent of a two phase foam structure with a dispersed liquid phase (the bathing solution) and a lipid medium (the internal membranes). We recognize that the effects of fixation on the structure of the fibre and the sarcoball are difficult to evaluate. However, the following hypothesis for sarcoball formation assumes (1) that the foam-like structure was closest to the true structure because it was seen after both light fixation and conventional fixation with dehydration in acetone, and (2) that vesiculation of the sarcoplasmic reticulum precedes, or occurs at the same time, as sarcoball formation (an alternative hypothesis is discussed later). The foam-like structure of the sarcoball could be formed at the fibre surface following extrusion of internal membrane vesicles. Neighbouring vesicles would aggregate because of the strong contracture of the fibre, and other forces such as capillary attractive forces (Bikerman, 1973) and Van der Waals forces (Wilschut & Hoekstra, I984), which contribute to the formation of foams and might also contribute to aggregation of vesicles. This aggregation of vesicles would extend approximately 2-3 Dm from the fibre surface where there is evidence for fusion of neighbouring vesicles in conventionally fixed material and a transition to the stable foam structure in lightly-fixed material. The continued extrusion of vesicles from the fibre into the region of aggregation allows fusion to continue and the sarcoball to grow as long as the contracture continues. The forces driving the extrusion of vesicles presumably depend on the contracture producing a pressure gradient between the inside and outside of the fibre. In general the diameter of vesicles within the fibre was about 400 nm compared with a diameter of 500 nm at the surface of the fibre. If the vesicles at the surface are from the same

651

population as those within the fibre, the internal vesicles may be smaller because they are under pressure. A pressure gradient between the inside and the outside of the fibre might initially cause vesiculation of the longitudinal sarcoplasmic reticulum (although vesiculation also occurs without a contracture; Ford & Surdyk, 1978) and force the vesicles to migrate to the surface of the fibre. It is energetically favourable for vesicles to fuse, although an initial perturbation is often required before fusion occurs (see below). The Laplace-Young equation shows that the pressure difference, P across a spherical bilayer increases as the radius of the vesicle, r, decreases so that, P = 4F/r

where F is the bifacial tension of the lipid bilayer. A reduction in free energy, providing a thermodynamic driving force for fusion, would occur (1) as vesicles migrated to the exterior of the fibre and expanded, and (2) with fusion of surface vesicles to form the large 1 ~tm vesicles observed at the interface between the fibre and the sarcoball and within the sarcoball. The foam structure hypothesis is supported by many of the properties of the sarcoball. The structure is plastic and fragile: sarcoballs dimple when approached by microelectrodes with positive internal pressures (Lewis, unpublished data). Membrane patches are not easily 'ripped off' the sarcoball: 'tethers' of cellular material form between the patch and the sarcoball and need to be physically removed before single channel recording proceeds (Lewis, unpublished data). Finally, sarcoballs collapse into large diameter patch-clamp electrodes with negative internal pressures (Nooney & Dulhunty, unpublished data). The hypothesis suggests that the surface of the sarcoball is made up of aggregated or fused vesicles rather than a continuous membrane. A puzzling observation was the continuous membranelike structure on the surface of the skinned fibres. The interior of skinned muscle fibres which are used to record tension or to investigate Caz+ release and uptake by sarcoplasmic reticulum is readily accessible to the bathing solution. Thus the fibres cannot be surrounded by an ion selective membrane. If a membrane is present it must be very leaky. The continuous membrane was not as obvious in lightly fixed material (Fig. 2A, 4 and 5). A speculative hypothesis is that the membrane is an artefact produced when vesicles along the surface of the fibre and at the outer edge of the sarcoball collapse during fixation and processing for electron microscopy. Fusion of vesicles to form sarcoballs

Fusion of lipid membranes is central to our hypothesis of sarcoball formation. Small liposomes ( < 50 nm) can fuse spontaneously (Schullery et al., 1980), but larger liposomes and membrane vesicles must first be perturbed in some way. Aggregation of vesicles is an essential step before fusion can occur (Wilschut & Hoekstra, 1984). Aggregation is followed by membrane destabilization,

652 formation of non-bilayer structures and finally fusion. Several types of perturbation have been used to induce fusion of cells or vesicles. Dehydration of lipid headgroups to induce aggregation and then rehydration to initiate fusion is a c o m m o n l y used technique to fuse cells (Ahkong et al., 1975). Giant liposomes (suitable for patch clamping) can be formed from membrane vesicles b y dehydration of a drop of vesicles on a slide (Criado & Keller, 1987), or by freeze-thaw procedures (Tomlins & Williams, 1986). Millimolar calcium or polycations cause aggregation of liposomes containing acidic phospholipids (Papahadjopoulos et aL, 1990). The presence of calcium, and an osmotic gradient, allows membrane vesicles to fuse with planar lipid bilayers (Miller & Racker, 1976). Perturbations that m a y contribute to vesicle fusion during sarcoball formation include the force of the unrestrained contracture and calcium ions that are released to initiate contraction. Although our hypothesis suggests that sarcobalis form as a result of aggregation and fusion of sarcoplasmic reticulum vesicles, it has not been established whether vesiculation of the sarcoplasmic reticulum actually precedes sarcoball formation when fibres are skinned in a CsF solution. An alternative, although less likely hypothesis, is that units of terminal cisternae and longitudinal sarcoplasmic reticulum from each sarcomere are extruded as intact structures which aggregate and fuse at the fibre surface to form the sarcoball. If this is the case, vesiculation could occur following sarcoball formation. In conclusion, the results of this study suggest that sarcoballs are formed by fusion of vesicles of internal membrane that are extruded during unrestrained contractures of skinned muscle fibres. The results show that much of the sarcoba]l membrane is derived from sarcoplasmic reticulum and do not rule out the possibility that a fraction of the membrane is derived from T-tubules and mitochondria.

Acknowledgements The authors are grateful to Leslie Maxwell and Sue Bell of the JCSMR Electron Microscope Facility for their support and expert assistance, to Stuart Butterworth for photography, and to Suzanne Curtis for technical assistance.

References AHKONG, Q. F., FISHER, D., TAMPION, W. & LUCY, J. A. (1975) Mechanisms of cell fusion. Nature 253, 194-5. BIKERMAN,J. J. (i973) Formation and structure. In Foams. New York: Springer. CORONADO, R. & MILLER,C. (I9,52) Conduction and block by organic cations in a K+-selective channel from sarcoplasmic reticulum incorporated into planar phospholipid bilayers. ]. Gen. Physiol. 79, 529-47.

LEWIS, D U L H U N T Y , J U N A N K A R and S T A N H O P E COSTELLO, B., CHADWICK,C., SAITO, A., CHU, A., MAURER, A. & FLEISCHER,S. (1986) Characterization of the junctional face membrane from terminal cisternae of sarcoplasmic reticulum. J. Cell Biol. 103, 741-53. CRIADO, M. & KELLER,B. U. (1987) A membrane fusion strategy for single-channel recordings of membranes usually nonaccessible to patch-clamp pipette electrodes. FEBS Left. 224, 172-6. DULHUNTY,A. F. (1984) Heterogeneity of T-tubule geometry in vertebrate skeletal muscle fibres. ]. Muscle Res. Cell MotiI. 5, 333-47. DULHUNTY,A. (1990) The rate of tetanic relaxation is correlated with the density of calcium ATPase in the terminal cisternae of thyrotoxic skeletal muscle. Pfldgers Arch. 415, 433-9. DULHUNTY,A. F., BANYARD,M. R. C. & MEDVECZKY,C. J. (I987) Distribution of calcium ATPase in the sarcoplasmic reticulum of fast- and slow-twitch muscles determined with monoclonal antibodies. ]. Membr, Biol. 99, 79-92. DULHUNTY,A. F., JUNANKAR,P. R. & STANHOPE,C. (1992) Extrajunctional ryanodine receptors in the terminal cisternae of mammalian skeletal muscle fibres. Proc. R. Soc. Lond. (Biol.) 247, 69-75. EASTWOOD, A. B., WOOD, D. S., BOCK, K. L. & SORENSON, M. M. (I979) Chemically skinned mammalian skeletal muscle 1. The structure of skinned rabbit psoas. Tissue Cell 11, 553-66. ELLISMAN,M. H., DEERINCK,T, J., OUYANG,Y., BECK,C. F., TANKSLEY, S. J., WALTON, P. D., AIREY, J. A. & SUTKO, J, L. (1990) Identification and localization of ryanodine binding proteins in the avian central nervous system. Neuron 8, 135-46. FORD, L. E. & SURDYK,M. F. (1978) Electron microscopy of skinned muscle cells. ]. Gen. Physiol. 72, 5a. HALLMAN,M., MIYAL,K. & WAGNER,R. M. (I976) Isolated lamellar bodies from rat lung. Lab. Invest. 35, 79-86. HALS, G. D., STEIN, P. G. & PALADE,P. T. (1989) Single channel characteristics of a high conductance anion channel in 'sarcoballs'. ]. Gen. Physiol. 93, 385-410. HAMILTON,S. L., ALVAREZ,R. M., FILL,M., HAWKES,M. J., BRUSH,K. L., SCHILLING,W, P. & STEFANI,E. (1989) [3H]PN200-110 and [3H]Ryanodine binding and reconstitution of ion channel activity with skeletal muscle membranes. Anal. Biochem. 183, 31-41. LEWIS,T. M. & BRETAG,A. H. (1991) A large conductance anion channel in a sarcoplasmic reticulum preparation from the semif-endenosus muscle of the cane toad. Proc. Aust. Physiol. Pharmacol. Soc. 22, 19P. MILLER, C. (1978) Voltage-gated cation conductance channel from fragmented sarcoplasmic reticulum: steady-state electrical properties. J. Membr. Biol. 40, 1-23. MILLER, C. & RACKER, E. (1976) Ca2+-induced fusion of fragmented sarcoplasmic reticulum with artificial planar bilayers. J. Membr. Biol. 30, 283-300. MOBLEY, B. A. & EISENBERG,B. R. (1975) Sizes of components in frog skeletal muscle measured by methods of stereology. ]. Gen. Physiol. 66, 31-45. MOLNAR, E., SEIDLER,N. W., JONA, I. & MARTONOSI,A. N. (1990) The binding of monoclonal and polyclonal antibodies to the Ca2+-ATPase of sarcoplasmic reticulum: effects on interactions between ATPase molecules. Biochim. Biophys. Acta 1023, 147-67.

Ultrastructure a n d sarcoplasmic reticulum c o n t e n t of sarcoballs NAKANE, P. K. (1982) Immunoelectron microscopy. Methods in Cancer Research I 0 , 183-212. PAPAHADJOPOULOS, D., NIR, S. & DUZGUNES,N. (1990) Molecular mechanisms of calcium-induced membrane fusion. ]. Bioenerg. Biomembr. 22, 157-79. ROUSSEAU, E. & MEISSNER,G. (1989) Single cardiac sarcoplasmic reticulum Ca~+-release channel: activation by caffeine. Am. ]. Physiol. 256, H328-H333. SAITO,A., WANG,C. T. &FLEISCHER,S. (1978) Membrane asymmetry and enhanced ultrastructural detail of sarcoplasmic reticulum revealed with use of tannic acid. J. Ceil Biol. 79, 601-16. SCHULLERY, S. E., SCHMIDT, C. F., FELGNER, P., TILLACK, T. W. & THOMPSON, T. E. (1980) Fusion of dipalmitoylphosphatidylcholine vesicles. Biochemistry 19, 3919-23. SMITH, J. S., CORONADO, R. & MEISSNER,G. (1985) Sarcoplasmic reticulum contains adenine nucleotide-activated calcium channels. Nature 316, 446-9. SORENSON, M. M., REUBEN,J. P., EASTWOOD,A. B., ORENTLICHER,M. & KATZ, G. M. (1980) Functional heterogeneity of the

653

sarcoplasmic reticulum within sarcomeres of skinned muscle fibres. J. Membr. Biol. 53, 1-17. STEIN, P. & PALADE, P. (1988) Sarcoballs: direct access to sarcoplasmic reticulum Ca 2+ channels in skinned frog muscle fibres. Biophys. J. 54, 357-63. SUAREZ-ISLA, B. A., OROZCO, C., HELLER, P. F. & FROEHLICH, J. P. (1986) Single calcium channels in native sarcoplasmic reticulum membranes from skeletal muscle. Proc. Natl. Acad. Sci. USA 83, 7741-5. TOMLINS, B. & WILLIAMS,A. J. (1986) Solubilisation and reconstitution of the rabbit skeletal muscle sarcoplasmic reticulum K § channel into liposomes suitable for patch clamp studies. Pfldgers Arch. 407, 341-7. VIVAUDOU, M. B., ARNOULT, C. & VILLAZ, M. (1991) Skeletal muscle ATP-sensitive K + channels recorded from sarcolemmal blebs of split fibres: ATP inhibition is reduced by magnesium and ADP. J. Membr. Biol. 122, 165-75. WILSCHUT, J. & HOEKSTRA, D. (1984) Membrane fusion: from liposomes to biological membranes. Trends Biochem. Sci. 9, 479-83.

Ultrastructure of sarcoballs on the surface of skinned amphibian skeletal muscle fibres.

The formation of sarcoballs on the surface of skinned fibres from semitendinosus muscles of Xenopus laevis, and the sarcoplasmic reticulum content of ...
6MB Sizes 0 Downloads 0 Views