ultrastructural observations and - Semantic Scholar

J. Cell Sci. 46, 129-147 (1980) Printed in Great Britain © Company of Biologists Limited igSo

J2Q

RELATIONSHIPS BETWEEN MITOCHONDRIAL OUTER MEMBRANES AND AGRANULAR RETICULUM IN NERVOUS TISSUE: ULTRASTRUCTURAL OBSERVATIONS AND A NEW INTERPRETATION j . SPACEK Department of Pathology, Charles University Hospital, 500 36 Hradec Krddovi, Czechoslovakia, and

A. R. LIEBERMAN Department of Anatomy, University College London, Gotuer Street, London WC1E6BT, England

SUMMARY

This study is concerned with extensions of the outer membrane of mitochondria in cells of nervous tissue, and with possible relationships between the extensions and the agranular reticulum. A variety of preparative techniques was applied to a large number of different central nervous tissues (CNS) and peripheral nervous tissues (PNS), using conventional thin sections, thicker sections (100 nm or more) and 3-dimensional reconstructions of serial thin sections. Extensions were commonly observed, particularly from the ends of longitudinally oriented mitochondria in axons and dendrites. Often these had the appearance of, and could be traced into apparent continuity with, adjacent elements of the agranular reticulum. In addition to these apical tubular extensions, we also observed and reconstructed short lateral tubular or sac-like extensions and vesicular protrusions of the outer mitochondrial membrane. We discuss and discount the possibility that the extensions are artefacts, consider the structural and biochemical similarities between the outer mitochondrial membrane and the agranular reticulum and propose that the outer mitochondrial membrane is part of the agranular reticulum (or a specialized portion of the agranular reticulum). We suggest that the translocation of mitochondria in nerve cells, and probably in other cells as well, involves movement of the inner mitochondrial membrane and the enclosed matrix (mitoplast) within channels of agranular reticulum in continuity, or in transient continuity, with the outer mitochondrial membrane.

INTRODUCTION Tubular or cisternal extensions of the outer mitochondrial membrane (OMM), similar in appearance to components of the endoplasmic reticulum, particularly the smooth endoplasmic reticulum (agranular reticulum) have been described in a variety of plant and animal tissue (e.g. Robertson, i960; Diers, 1966; Ruby, Dyer & Skalho, 1969; Bracker & Grove, 1971; Franke & Kartenbeck, 1971; Morre\ Merritt & Lembi, 1971; Volk, 1971; Weakley, 1977), and the concept that the OMM is continuous with the endoplasmic reticulum and thus a part of the cells endomembrane system has Correspondence and offprint requests to: Dr A. R. Lieberman at University College London, as shown above.

130

J. Spatek and A. R. Lieberman

stemmed from such observations (e.g. Bracker & Grove, 1971). Similar extensions of the OMM have also been described in vertebrate nervous tissue (Hamori & Dyachkova, 1964; Reiter, 1966; Sandborn, 1966; Grainger & James, 1969; Laatsch, 1969; Lieberman, 1971; Batty & Millhouse, 1976; Spa6ek & Lieberman, 1977) and in many other papers describing ultrastructural features of axons or dendrites, extensions of the OMM have been illustrated, although not commented upon (e.g. Hinds, 1970, fig. 3; Ghetti, Horoupian & Wisniewski, 1972, fig. 13; Saito, 1972, fig. 11; Chan-Palay, 1973, fig. 3a; Peracchia, 1973, fig. 1; Grofova& Rinvik, 1974, figs. 17, 20; Rinvik& Grofova, 1974, fig. 10; Palay & Chan-Palay, 1973, figs. 131, 164; Rustioni& Sotelo, 1974, fig. 1; Peters, Palay & Webster, 1970, figs. 3-10; Spencer & Schaumburg, 1977, fig. 23). It is surprising that despite these numerous reports, the concept of continuity between the OMM and the endoplasmic reticulum is not as well established and accepted for nervous tissue as it appears to be for other tissues, and few authors have considered the nature and consistency of the relationship between OMM and endoplasmic reticulum or what its significance might be (e.g. Lieberman, 1971). Even in recent reference works on the organization of nerve fibres, in which, because of the alignment of organelles parallel to the long axis of the fibre, such relationships may be most readily observed, this aspect of axonal ultrastructure is not described (Webster, 1974; Landon & Hall, 1976; Peters et al. 1970; Berthold, 1978; Hirano & Dembitzer, 1978). We present here observations on relationships between the OMM and the agranular reticulum (AR) in central and peripheral nervous tissue, and propose a new concept of the relationship between mitochondria and the AR.

MATERIALS AND METHODS The observations summarized here, which derive from the examination of many hundreds of relevant micrographs, were made in the course of ultrastructural studies of peripheral nerves and sensory ganglia, cerebral and cerebellar cortices, corpus callosum, thalamic nuclei, optic tectum and spinal cord of laboratory mammals (mice, rats, rabbits) and frogs. A variety of standard fixatives, of immersion and perfusion techniques and of embedding materials has been used in these studies. The animals were anaesthetized with ether, pentobarbital or urethane (frogs). For immersion fixation, small tissue blocks were placed in 1 or 2 % phosphate-buffered osmium tetroxide or o-6 % veronal-acetate-buffered potassium permanganate (Luft, 1956). Most of the material studied, however, derived from anaesthetized animals killed by transcardiac perfusion with phosphate- or cacodylate-buffered aldehyde solutions of variable composition and concentration. Tissue blocks were postfixed in phosphate- or cacodylate-buffered 1 or 2 % osmium tetroxide, dehydrated through ethanol (or occasionally acetone), stained in block with uranyl acetate, passed through propylene oxide and embedded in Vestopal W, Araldite, Durcupan ACM or Epon-Durcupan mixture. Thin sections stained with uranyl acetate and/or lead citrate were examined in Tesla BS500 or Philips EM 300 or 301 electron microscopes. The figure legends indicate the sources of individual micrographs. Three-dimensional reconstructions from serial sections were prepared as described previously (Spa£ek& Lieberman, 1974).

OBSERVATIONS

Extensions of the OMM were observed in neuronal cell bodies, dendrites, axons and axon terminals, and in astrocyte and ependymoglial cell bodies and processes. The

Outer mitochondrial membrane extensions

131

extensions fall into 3 broad categories: apical tubular extensions; lateral tubular or sac-like extensions; vesicular protrusions.

Apical tubular extensions (Figs. 1-4, 9-13) were commonly observed when elongate mitochondria were sectioned parallel to their long axes, and were particularly frequent in axons (Figs, i, 10-13), where they could be followed for up to 2/(m in a single section. Occasionally, extensions were apparent at both ends of a longitudinally sectioned mitochondrion (Fig. 10). More rarely, OMM extensions linking 2 mitochondria in a chain were observed (Figs. 2, 13). Some apical extensions, particularly in axons and dendrites appeared to be rather wide (as wide as the mitochondrion itself in some cases), devoid of content, and suggestive of artefact (Figs. 1, 9-11). Others were narrower, sometimes very narrow (Fig. 6), and similar in appearance to adjacent elements of the AR, as in the broad process of an astrocyte illustrated in Fig. 3. It is, however, extremely difficult to demonstrate convincingly a continuity between OMM extensions and the general AR in single thin sections. In attempts to clarify the relationship between AR and OMM we also studied sections of greater than usual thickness (i.e. 100 nm plus) and made 3-dimensional reconstructions from series of very thin sections. In thick sections, the incidence of observations of OMM extensions was higher than in thin sections and it was possible to follow direct continuities between OMMs and channels of AR, which in parts were of about the same diameter as microtubules (Figs. 11, 12). Also observed in this material were wide, tubular apical extensions of OMMs interconnected by very narrow tubules (Fig. 13). Analysis of serial sections parallel to the long axes of axonal mitochondria were inconclusive with respect to the demonstration of continuities between AR and OMMs. Although thin AR tubules were observed in the vicinity of mitochondrial apices and OMM extensions (Fig. 17, p. 137), continuity with the latter could not be established with certainty because the diameter of the AR tubules (20-40 nm) was generally less than the thickness of the sections (minimum thickness 50 nm) and image overlap precluded unequivocal demonstration of continuity. One situation in which continuity between apical extensions of OMMs and AR was frequently and convincingly observed was in the organelle accumulations within crushed axons just above the site of crush (Figs. 4, 5). Lateral tubular or sac-like extensions. These were seen only occasionally in longitudinal sections of mitochondria (Fig. 7). Reconstructions of such protrusions from serial sections perpendicular to the long axes of 2 axonal mitochondria are shown in Figs. I8A, B, and some of the micrographs from the series on which Fig. I8A was based are shown in Fig. I6A-D. In both cases, the lateral extensions appeared to be blind-ending and irregularly shaped. Vesicular protrusions. Small, smooth surfaced vesicle-like protrusions of the OMM, approximately 40-100 nm in diameter, were observed in all parts of neurons and in astrocyte processes (Figs. 7, 8). Three-dimensional reconstructions confirmed the smooth, spherical form of these protrusions (Fig. 19).

J. Spacek and A. R. Lieberman


133

Dependence of OMM extensions on method of fixation

A variety offixativetechniques was used in an attempt to determine whether or not there is a relationship between the type of fixative and the occurrence of OMM extensions. Apical tubular extensions and vesicular protrusions were found in all samples of nervous tissue from animals perfused with aldehydes (glutaraldehyde or paraformaldehyde alone or mixtures of both at various concentrations) before postfixation with osmium tetroxide. Vesicular protrusions were also observed in material fixed by immersion in osmium tetroxide solution, but apical tubular extensions were very seldom seen, perhaps because of the well known tendency of plasma and endocellular membranes to break down into vesicles under such fixation conditions (Fawcett, 1966). In material fixed by immersion in potassium permanganate, mitochondria were swollen and AR cisterns generally widened: nevertheless it was possible to observe and trace tubular extensions of the OMM for short distances. Two artefacts of fixation, occasionally seen in this material, and relevant to the interpretation of extensions of OMMs are illustrated in Figs. 14 and 15. The 'blebs' protruding from the plasma membranes of the thin perineurial cell laminae in Fig. 14 display an indistinct membrane, thinner than, and without the unit membrane characteristics of the plasma membrane of the cellular processes (e.g. at arrowheads) or of the plasmalemmal vesicles (crossed arrows): these blebs we interpret as aldehyde fixation artefact. Fig. 15 shows an extreme example of a rather different artefact occasionally seen in poorly fixed tissue, which involves irregular sac-like protrusions of both mitochondrial membranes. These protrusions seem to be devoid of content, Fig. 1. Broad apical tubular extension of outer mitochondrial membrane (OMM) in a myelinated axon of rabbit corpus callosum./, neurofilament; t, microtubule, 3 % glutaraldehyde. Bar, o-i fim. Fig. 2. Tubular connexion (arrow) between the OMMs of 2 mitochondria in a longitudinally sectioned Purkinje cell dendrite of mouse cerebellar cortex. 1 % paraformaldehyde: 1 % glutaraldehyde. Bar, o-2/tm. Fig. 3. Sac-like extension of the OMM (arrow) in a large astrocyte process of mouse cerebellar cortex. Note the adjacent elements of agranular reticulum (ar). 1 % paraformaldehyde : 1 % glutaraldehyde. Bar, 02 fim. Fig. 4. Apical tubular extensions (arrows) from the outer membrane of mitochondria accumulated with other organelles in a swollen axon of rabbit spinal cord just above the site of a crush i h previously, ar, vesicles and cisterns of agranular reticulum; /, neurofilaments. 4 % paraformaldehydc: 0-5 % glutaraldehyde. Bar, 02 fim. Fig. 5. Lateral tubular extension of the OMM from similar material to that shown in Fig. 4. Here, the wide extension is continuous with irregularly shaped narrow elements of the agranular reticulum (ar). 4 % paraformaldehyde: 0-5% glutaraldehyde. Bar, 0-2

fim.

Fig. 6. Narrow tubular extension of the OMM (arrow) close to the tip of a mitochondrion in an axon terminal (mouse cerebellar cortex). 1 % paraformaldehyde: 1 % glutaraldehyde. Bar, 01 fim. Fig. 7. Lateral sac-like extensions (arrows) of the OMMs of 2 adjacent mitochondria in a longitudinally sectioned myelinated nerve fibre of rat trigeminal nerve, ar, peripherally located elements of the agranular reticulum;/, neurofilaments. 1 % paraformaldehyde: 1 % glutaraldehyde. Bar, o-2 fim.

J. Spalek and A. R. Lieberman


135

their membranes sometimes discontinuous, and we interpret them as a consequence of anoxia during tissue preparation.

DISCUSSION

The question of artefact

Interpretive difficulties associated with possible fixation artefacts are a general problem in cytology, nowhere more so than in the cytology of nervous tissue. Aldehyde fixation often results in mitochondrial swelling, poor preservation of lipids, and vesiculation of membranes (Buckley, 1973; Glauert, 1975; Scott, 1976). Furthermore, fixation by almost any technique is commonly associated with the presence of at least some swollen and apparently 'exploded' mitochondria, even in tissue which appears to be generally well fixed. Most of the material analysed in the present study was fixed with purified glutaraldehyde to avoid the more familiar glutaraldehyde artefacts (but see Robertson & Schultz, 1970, who consider that unpurified commercial glutaraldehyde is a better fixative for CNS tissue than distilled and purified glutaraldehyde). Even so, some unequivocally artefactual changes in mitochondria were observed, such as the double-walled wide tubular formations depicted in Fig. 15. Against this background it is difficult to rule out the possibility of an artefactual basis for the formation of the vesicular protrusions and possibly also of the tubular or sac-like lateral extensions. However, it is even more difficult to invoke artefact to explain the tubular extensions of the OMM at the apices of longitudinally sectioned mitochondria. These extensions appear consistently in this location and not at other points on the mitochondrial surface, and are particularly common at the apices of elongate mitochondria in axons and dendrites. It would be very surprising if a Fig. 8. Vesicular protrusion (arrow) of the OMM of an astrocytic mitochondrion (mouse cerebellar cortex). 1 % paraformaldehyde: 1 % glutaraldehyde. Bar, o-i fim. Fig. 9. Long apical tubular extension of the OMM (arrow) in a longitudinally sectioned Purkinje cell dendrite (mouse cerebellar cortex). Numerous other longitudinal elements of the agranular reticulum lie adjacent to this and another mitochondrion. at, axon terminal. 1 % paraformaldehyde: 1 % glutaraldehyde. Bar, 0-2 fim. Fig. 10. Extensions of the OMM at both ends of a longitudinally sectioned mitochondrion in a myelinated axon of rat medial lemniscus. /, neurofilaments. 4 % paraformaldehyde: 0-5 % glutaraldehyde. 150-nm section. Bar, o-a fim. Fig. 11. Probable continuity between an apical tubular extension (arrow) and a longitudinal element of the agranular reticulum (ar), in a myelinated axon of rat medial lemniscus. 4 % paraformaldehyde: 0-5% glutaraldehyde. 150-nm section. Bar, 0-2 fim. Fig. 12. Apical tubular extension of the OMM in continuity (arrowheads) with a longitudinal element of the agranular reticulum. Note that the reticulum element narrows down (between open arrowheads) to a diameter even less than that of micro tubules (t). Myelinated axon of rabbit corpus callosum. 3 % glutaraldehyde. 150-nm section. Bar, 0'2/im. Fig. 13. Apical tubular extensions of OMMs linking 2 mitochondria at what is interpreted as a very narrow tubular connexion (arrow). Myelinated axon of rabbit corpus callosum. 3 % glutaraldehyde. 150-nm section. Bar, o-i fim.

J. Spaiek and A. R. Lieberman


137

fixation artefact were to affect only this portion of the OMM and improbable that channels 2 fim long in continuity with the OMM, could appear as a result of differential shrinkage of the inner compartment of the mitochondrion, particularly as only the apical portions of the OMM are involved and no evidence of shrinkage of the inner mitochondrial compartment was apparent in the relevant micrographs. An additional argument against an artefactual basis for this or the other types of OMM extension is that whereas presumptively artefactual protrusions (blebs) of some membranes in this material appeared to be thinner and less electron-opaque than the membranes from which they protruded (e.g. Fig. 14), the OMM extensions were all of normal unit

mt

Fig. 17. Reconstruction of an apical tubular extension of the OMM, based on serial longitudinal sections of a myelinated axon in rabbit corpus callosum. The arrows indicate regions where continuities between narrow tubular longitudinal elements of the agranular reticulum and the OMM (at left) and the extension of the OMM (at right) were suspected but could not be demonstrated for technical reasons (see text). Idcv, large dense-cored vesicle; mt, microtubules; nf, neurofilament.

membrane appearance and thickness. Finally the existence of an extensive literature containing observations to the effect that AR membranes are commonly continuous with the OMM (see references in Introduction) and with the limiting membranes of other intracellular membraneous organelles also favours the interpretation that the extensions of the OMM, particularly in the apical regions of elongate mitochondria, are not artefactually induced during tissue preparation.

Fig. 14. Plasma membrane ' blebs' (arrows) interpreted as aldehyde fixation artefact. Compare the indistinct 'membrane' of the bleb with the distinct unit membranes of the cell surface (arrowheads) and plasmalemmal vesicles (crossed arrows). Perineurial cells of rat trigeminal nerve. 1 % paraformaldehyde: 1 % glutaraldehyde. Bar, o-i fim. Fig. 15. Wide, double-walled structures (asterisks) with empty appearance interpreted as derived from damaged mitochondria. Neuronal perikaryon, rat ventrobasal thalamus. 2-5 % glutaraldehyde. Bar, 0-2 fim. Fig. 16. Sections 1, 2, 3 and 6 (A, B, C and D respectively) of a longer series on which the reconstruction in Fig. I8A was based. A lateral sac-like extension of the OMM in a transversely sectioned myelinated axon of rat trigeminal nerve is shown (arrows). Note the close relationship of the OMM to adjacent microtubules (solid arrowheads) and neurofilaments (open arrowheads). 1 % paraformaldehyde: 1 % glutaraldehyde. Bar, o-i fim. 10

CEL

46

138


Are the extensions of the OMM in continuity with the AR ? Although, as we have stated above, the extensions of the OMM particularly the apical tubular extensions, are likely to be genuine features of the organization of mitochondria in nervous tissue, firm evidence for the existence of extensive and

O1 jim

Fig. 18 A. For legend see facing page.

permanent continuities between the OMM and the AR was not obtained in the present study. Nevertheless, the indications that there are structural relationships between OMMs and the AR are strong and come from several previously published studies in addition to the present study (references in Introduction; see particularly Reiter, 1966; Lieberman, 1971; Batty & Millhouse, 1976; Spacek & Lieberman, 1977). It should be said, however, that in several studies of the arrangement of AR in nerve cells, notably that of Ducros (1974) employing serial-section analysis of salivary nerves of Octopus, that of Teichberg & Holtzman (1973) using electron microscopy and ultracytochemistry on cultured chick embryo sympathetic neurons and those of Droz,


Rambourg & Koenig (1975) and Tsukita & Ishikawa (1976) employing high-voltage electron microscopy to examine metal-impregnated AR in peripheral nerve fibres, no connexions of any kind were reported between mitochondria and the AR. There are, however, inherent technical limitations with respect to the demonstration of continuities between the OM and AR in the ultrastructural approaches employed mt

nf

Fig. 18A, B. Reconstructions of lateral sac-like OMM extension based on serial transverse sections of a myelinated axon (A) and an unmyelinated axon (B) in rat trigeminal nerve. The arrow in Fig. 18 A indicates the level in the series illustrated in Fig. 16c. mt, microtubules; nf, neurofilaments.

in the present study and in those cited in the preceding paragraph. The problems associated with reconstructing tortuous tubular structures with diameters less than the thickness of the sections have already been mentioned. There are also considerable problems in resolving such continuities in thick sections of metal-impregnated tissue (Droz et al. 1975; Tsukita & Ishikawa, 1976) even when stereo-pairs of pictures are analysed. There may also be a tendency for long tubular continuities between AR and OMMs to break down into vesicles or short tubules, particularly under conditions of

140


suboptimal tissue preservation (e.g. Fawcett, 1966; Buckley, 1973; Bird & James, 1977). Finally, of course, it may be that the continuities are of a transient, dynamic nature and thus not readily amenable to demonstration by ultrastructural methods. In our opinion, and notwithstanding the negative observations and the equivocal nature of some of the positive observations, there is, on balance, sufficient evidence to support the proposal that continuities exist between the AR and the OMMs (or rather between AR and the OMMs of some mitochondria at a particular time) in axons and dendrites and probably in other components of central and peripheral nervous tissue.

Fig. 19. Reconstruction of a lateral vesicular protrusion of the outer membrane of a mitochondrion in an astrocyte process (rat cerebellar cortex). The protrusion has a short, narrow neck and a larger smooth-surfaced, spherical head.

The structural evidence is much strengthened by biochemical evidence which suggests on the one hand that the OMM and the inner membrane are fundamentally dissimilar and on the other hand that the OMM and AR share certain important characteristics. For example, it is well established that the outer and inner mitochondrial membranes differ in ultrastructure, chemical composition, permeability and enzyme content (Ernster & Kuylenstierna, 1969, 1970; Bracker & Grove, 1971; Lehninger, 1975) and the concept that the mitochondria, in the functional sense, are constituted by the inner membrane, its cristae and the enclosed matrix ('mitoplast') is not a new one in cell biology (Ernster & Kuylenstierna, 1969, 1970). Furthermore, there are numerous indications in the literature that the OMM may be, or may be part of, a specialized portion of the endoplasmic reticulum. For example, there are similarities in protein and lipid composition, staining characteristics and turnover properties of AR and OMMs (e.g. Ernster & Kuylenstierna, 1970). There are also some apparent differences between the OMM and the general AR. For example, cytochrome P ^ is found in the microsomal fraction but not in the OMM fraction, whereas monoamine oxidase characterizes the latter but not the microsomal fraction,


141

and there are also differences between phospholipases in OMM and microsomal fractions (Ernster & Kuylenstierna, 1970). Such differences are, however, minor by comparison with the differences between the OMM and the inner mitochondrial membrane, and in relation to the similarities between OMM and AR. Furthermore, such differences are based in large part on comparisons between relatively pure OMM

Fig. 20. Diagram incorporating the observations made in the present study and illustrating the hypothesis of translocation of the inner membrane components of mitochondria (mitoplasts) within channels of agranular reticulum. ar, agranular reticulum; b, beaded agranular reticulum; v, vesiculated agranular reticulum; 1-5, mitochondria: 1, with a vesicular protrusion of the OMM; 2, in a continuous channel of agranular reticulum (moving phase ?); 3, 4, sharing the same tubule of reticulum (moving in sequence along the same channel ?); 5, classical view of a mitochondrion (resting phase ?). Note also the lateral connexions between AR elements and the lateral connexion between mitochondrion 3 and a longitudinal element, which might represent channels for lateral movement.

142


subfractions and decidedly impure and heterogeneous microsomal fractions (which probably include derivatives of AR, granular reticulum and even plasma membranes) and derive largely from studies on hepatocytes in which the AR is highly specialized and certainly different in at least some of its functions from axonal AR, and in which the mitochondria are much less motile than in nerve cells. Finally, it is entirely possible that the relationships between OMMs and AR are not only different in different cell types, but that they may differ in different parts of the neuron (see Gray, Jones & Barron, 1979). In some cell types there is also evidence that the OMM may be continuous with the granular endoplasmic reticulum, for example in yeast cells (Keyhani, 1973; Kellems, Allison & Butow, 1974) or in mammalian ovarian cells (Ruby et al. 1969). It is also established that the limiting membranes of cytoplasmic organelles other than mitochondria may be continuous with the endoplasmic reticulum, for example peroxisomes (Hicks & Fahimi, 1977), microperoxisomes (Novikoff, 1976; Gulyas & Yuan, 1977), and multivesicular bodies (Teichberg, Holtzman, Crain & Peterson, 1975; personal observations). On the basis of the evidence reviewed above it is difficult to take issue with the view that the OMM is part of the cell's endomembrane system (Bracker & Grove, 1971). Interpretations and hypothesis

The present and previous ultrastructural observations, together with biochemical data on the nature and affinities of the OMM, provide, we believe, a satisfactory basis for the conclusion that the OMM is continuous with, or is capable of becoming transiently continuous with, the AR. Perhaps, more correctly, it is with a specialized compartment of the AR that these continuities are established, since there is no reason to assume that the AR comprises a functionally homogeneous system (see, for example Novikoff, 1976; Holtzman, Schacher, Evans & Teichberg, 1977). These observations and interpretations are depicted diagrammatically in Fig. 20. It is not immediately obvious what the significance of this relationship might be. The possibility that connexions between OMMs and AR are concerned in the distribution of materials to and from mitochondria has been discussed (Grainger & James, 1969; Lieberman, 1971; Holtzman, 1977): so far as transport away from the mitochondria is concerned, one possibility would be distribution of ATP to energyrequiring sites (Lieberman, 1971). Direct continuities with the endoplasmic reticulum might also be involved in transport to the mitoplast of the many proteins, such as cytochrome c, which are contained within the organdie but synthesized outside it (Baxter, 1971). Another possibility, that they are involved in the sequestration of, and thus in regulation of the concentration of, intracellular and/or extracellular calcium ions has been discussed by Lieberman (1971). There has also been discussion of the possibility that the vesicle-like protrusions seen in this and other studies (Hamori & Dyachkova, 1964; Spriggs, Lever & Graham, 1967; Giildner, 1976; Carmichael & Smith, 1978; and see also Gray et al. 1979), might be concerned with the formation of synaptic vesicles. However, whereas the possibility that the OMM occasionally buds off vesicles cannot be discounted, the fact that such vesicular protrusions were found


143

in the present study to occur in non-synaptic as well as in presynaptic sites makes it unlikely that these protrusions are concerned with synaptic vesicle formation. Here, we would like to focus attention on the suggestion that the longitudinally oriented extensions of the OMM could be interpreted as evidence that the familiar phenomenon of mitochondrial movement within neuronal processes (and perhaps within cells in general) involves movement of the inner membrane component (i.e. the mitoplast) within channels constituted by a specialized portion of the AR (5pa6ek & Lieberman, 1977)- In other words, we suggest that the inner mitochondrial component moves through the cell within channels of AR, rather like a cylindrical container passing through a pneumatic post system (although in employing this analogy we do not intend to imply anything about the mechanism(s) for mitoplast movement within the AR channels, and have no specific suggestions as to what these mechanism(s) might be). This hypothesis need not necessarily imply that the system of channels through which the mitoplasts move is permanently interconnected and patent, although there is some evidence that the AR system in axons may be continuous over very long lengths and possibly from soma to axon terminals (Ducros, 1974; Droz et al. 1975; Tsukita & Ishikawa, 1976; Holtzman, 1977; Holtzman et al. 1977)- It may be rather, that continuities between the AR system and the OMM are established in the van of a moving mitoplast and are broken again in its wake. Nor does the hypothesis imply that the AR and OMM are stationary, with the mitoplasts as the only motile components of the system. The AR-OMM membranes may also be moving through the cell, but at rates different from the rate of movement of the mitoplast, and not necessarily at slower rates, since there is evidence that some components of the AR system move along axons in the orthograde direction at extremely fast rates (Grafstein, 1977; Holtzman, 1977). It is improbable that the system of AR-OMM channels serves only to transmit the mitoplasts, and probable that these membrane systems play other specific roles in cell function. In the absence of satisfactory data, however, it is difficult at present to suggest what these might be. One possibility is that the AR system related to the OMM is particularly concerned in the control of intracellular calcium: if so, one of the factors influencing mitoplast movement within the cell may be regional and possibly varying energy requirements of AR systems involved in the sequestration of calcium ions and in other aspects of intracellular calcium regulation. The hypothesis we put forward in the preceding paragraph is consonant not only with the structural and biochemical observations reported and reviewed above, but also with other data, particularly with observations on the movements of mitochondria in nerve fibres grown in vitro or acutely isolated for direct microscopic examination. For example, mitochondria (more strictly, particles some of which are probably mitochondria), commonly follow one another at intervals along what appear to be tracks or channels through the axoplasm (Pomerat, Hendelman, Raiborn & Massey, 1967; Kirkpatrick, Bray & Palmer, 1972; Cooper & Smith, 1974; Forman, Padjen & Siggins, 1977), and when nerves or individual nerve fibres are interrupted or compressed, the mitochondria (in particular, but other organelles as well) that accumulate on either side of the constriction, tend to do so in longitudinal chains (Zelena, Lubinska & Gutmann, 1968; Weiss & Mayr, 1971; Smith, 1980). Also, although mitochondria in

144

J- Spa/ek and A. R. Lieberman

axons are generally stationary for long periods, they are capable of very rapid ' saltatory' movements (Kirkpatrick et al. 1972; Cooper & Smith, 1974; Forman et al. 1977) and commonly make radial 'jumps', as though changing 'channels' before continuing movement in the longitudinal direction (Kirkpatrick et al. 1972; Breuer, Christian, Henkart & Nelson, 1975; Forman et al. 1977). The most widely accepted explanation for the movement of mitochondria and other small particles along defined channels or roads through axoplasm is the presence in the channels of bundles of microtubules, which are thought to be involved in some manner in the guidance and movement of such particles (Breuer et al. 1975; Grafstein, 1977; Holtzman, 1977). Another possibility, less widely discussed, is that the channels represent regions of the axon not occupied by the fairly closely packed bundles of neurofilaments, which therefore offer less resistance to the longitudinal movement of other axonal constituents. It is, however, equally likely that the 'channels' for longitudinal movement are the longitudinally oriented AR elements and those for radial movement the frequent transverse interconnexions between the latter. Furthermore, the intermittent phases of rapid movement may well be associated with periods during which AR channels are in continuity with OMMs and the stationary phases with periods during which the OMM is isolated from the AR. Mitochondria in the latter phase would be represented by those - the majority - without evident continuities with the AR in serial thin sections. Finally, the almost complete independence of OMM and mitoplast implied by this hypothesis is condordant with the proposal that mitochondria (and plastids) in eukaryotic cells arose in the course of evolution as the result of an endosymbiotic association with prokaryotic microorganisms (e.g. Nass, Nass & Afzelius, 1965; Baxter, 1971; Schnepf & Brown, 1971; Ebringer, 1972; Munn, 1974). We thank the British Council for enabling J. Spaiek to visit University College London in 1969, the Wellcome Trust for financial support, and I. Bramborova and V. Kubesova for technical assistance. We also thank Dr I. M. Hais and Dr J. Zelena (Academy of Sciences, Prague) for helpful discussions, and Dr Zelena for sharing her similar thoughts on mitochondrial translocation with us as long ago as 1972, and for making us realize that it would be a worthwhile task to seek publication of these observations and interpretations. REFERENCES H. K. & MILLHOUSE, O. E. (1976). Ultrastructure of the Gunn rat substantia nigra. II. Mitochondrial changes. Acta neuropath. 34, 7-19. BAXTER, R. (1971). Origin and continuity of mitochondria. In Origin and Continuity of Cell Organelles (ed. J. Reinert & H. Ursprung), pp. 46-64. Berlin, Heidelberg, New York: Springer. BERTHOLD, C.-H. (1978). Morphology of normal peripheral axons. In Physiology and Pathobiology of Axons (ed. S. G. Waxman), pp. 3-63. New York: Raven Press. BIRD, M. M. & JAMES, D. W. (1977). The development and ultrastructure of previously dissociated foetal human cerebral cortical cells in vitro. Cell Tiss. Res. 183, 403-417. BRACKER, C. E. & GROVE, S. N. (1971). Continuity between cytoplasmic endomembranes and outer mitochondrial membranes in fungi. Protoplasma 73, 15—34. BREUER, A. C , CHRISTIAN, C. N., HENKART, M. & NELSON, P. G. (1975). Computer analysis of organelle translocation in primary neuronal cultures and continuous cell lines. J. Cell Biol. 65, 562-576. BATTY,


145

BUCKLEY, I. K. (1973). Studies in fixation for electron microscopy using cultured cells. Lab. Invest. 29, 398-410. CARMICHAEL, S. W. & SMITH, D. J. (1978). High-voltage electron microscopy of adrenal medulla: Direct connections between mitochondria and catecholamine-storage vesicles. Experientia 34, 391-392. CHAN-PALAY, V. (1973). On the identification of the afferent axon terminals in the nucleus lateralis of the cerebellum. An electron microscope study. Z. Anat. EntwGesch. 142, 149— 186. COOPER, P. D. & SMITH, R. S. (1974). The movement of optically detectable organelles in myelinated axons of Xenopus laevis.J. Physiol., Lond. 242, 77-97. DIERS, L. (1966). On the plastids, mitochondria, and other cell constituents during oogenesis of a plant. J . CellBiol. 28, 527-543. DROZ, B., RAMBOURG, A. & KOENIG, H. L. (1975). The smooth endoplasmic rericulum:

Structure and role in the renewal of axonal membrane and synaptic vesicles by fast axonal transport. Brain Res. 93, 1-13. DUCROS, C. (1974). Ultrastructural study of the organization of axonal agranular reticulum in Octopus nerve. J. Neurocytol. 3, 513-523. EBRINGER, L. (1972). Are plastids derived from prokaryotic micro-organisms ? Action of antiboitics on chloroplasts olEuglenagracilis.J.gen. Microbiol. 71, 35-52. ERNSTER, L. & KUYLENSTIERNA, B. (1969). Structure, composition and function of mitochondrial membranes. In Mitochondria: Structure and Function (ed. L. Ernster & Z. Drahota). London, New York: Academic Press. ERNSTER, L. & KUYLENSTIERNA, B. (1970). Outer membrane of mitochondria. In Membranes of Mitochondria and Chloroplasts (ed. E. Racker), pp. 172-212. New York, Cincinnati, Toronto, London, Melbourne. Van Nostrand Reinhold. FAWCETT, D . W. (1966). The Cell: Its Organelles and Inclusions. Philadelphia: Saunders. FORMAN, D. S., PADJEN, A. L.& SIGGINS, G. R. (1977). Axonal transport of organelles visualized by light microscopy: cinemicrographic and computer analysis. Brain Res. 136, 197-213. FRANKE, W. W. & KARTENBECK, J. (1971). Outer mitochondrial membrane continuous with endoplasmic reticulum. Protoplasma 73, 35-41. GHETTI, B., HOROUPIAN, D. S. & WISNIEWSKI, H. M. (1972). Transsynaptic response of the

lateral geniculate nucleus and the pattern of degeneration of the nerve terminals in the rhesus monkey after eye enuncleation. Brain Res. 45, 31-48. GLAUERT, A. M. (1975). Fixation, dehydration and embedding of biological specimens. In Practical Methods in Electron Microscopy, vol. 3, part I (ed. A. M. Glauert). Amsterdam,. Oxford, North-Holland, New York: American Elsevier Publishing Co. GRAFSTEIN, B. (1977). Axonal transport: the intracellular traffic of the neuron. In Cellular Biology of Neurons, part I, Handbook of Physiology, section I, The Nervous System, vol. I (ed. E. R. Kandel), pp. 691-717. Bethesda: Am. Physiol. Soc. GRAINGER, F. & JAMES, D. W. (1969). Mitochondrial extensions associated with microtubules in outgrowing process from chick spinal cord in vitro. J. Cell Sci. 4, 729—737. GRAY, E. G., JONES, D. H.& BARRON, J. (1979). Aggregations of synaptic vesicles on the exposed inner membrane of presynaptic mitochondria in brain..7. Neurocytol. 8, 675-685. GROFOVA, I. & RINVIK, E. (1974). Cortical and pallidal projections to the nucleus ventralisthalami. Electon microscopical studies in the cat. Anat. Embryol. 146, 113-132. GULYAS, B. J. & YUAN, L. C. (1977). Association of microperoxisomes with the endoplasmic reticulum in the granulosa lutein cells of the rhesus monkey (Macaca mullata). Cell Tiss. Res. 179. 357-366. GOLDNER, F.-H. (1976). Synaptology of the rat suprachiasmatic nucleus. Cell Tiss. Res. 165,. 509-544HAMORI, J. & DYACHKOVA, L. N. (1964). Electron microscope studies on developmental differentiation of ciliary ganglion synapses in the chick. Ada biol. hung. 15, 213-230. HICKS, L. & FAHIMI, H. D. (1977). Peroxisomes (microbodies) in the myocardium of rodents and primates. A comparative ultrastructural cytochemical study. Cell Tiss. Res. 175, 467— 481. HINDS, J. W. (1970). Reciprocal and serial dendro-dendritic synapses in the glomerular layer of the rat olfactory bulb. Brain Res. 17, 530-534.

146


A. & DEMBITZER, H. M. (1978). Morphology of normal central myelinated axons. In Physiology and Pathobiology of Axons (ed. S. G. Waxman), pp. 65-82. New York: Raven Press. HOLTZMAN, E. (1977). The origin and fate of secretory packages, especially synaptic vesicles. Neuroscience 2, 327-355. HOLTZMAN, E., SCHACHER, S., EVANS, J. & TEICHBERG, S. (1977). Origin and fate of the membranes of secretion granules and synaptic vesicles: membrane circulation in neurons, gland cells and retinal photoreceptors. In Cell Surface Reviews, vol. 4. The Synthesis, Assembly and Turnover of Cell Surface Components (ed. G. Poste & G. L. Nicholson), pp. 165-246. Amsterdam, New York: Elsevier, North-Holland. KELLEMS, R. E., ALLISON, V. F. & BUTOW, R. A. (1974). Cytoplastic type 80S ribosomes associated with yeast mitochondria. II. Evidence for the association of cytoplasmic ribosomes with the outer mitochondrial membrane in situ.J. biol. Chent. 249, 3297-3303. KEYHANI, E. (1973). Ribosomal granules associated with outer mitochondrial membrane in aerobic yeast cells. J. Cell Biol. 58, 480-484.. KIRKPATRICK, J. B., BRAY, J. J. & PALMER, S. M. (1972). Visualization of axoplasmicflowin vitro by Nomarski microscopy. Comparison to rapid flow of radioactive proteins. Brain Res. 43, HIRANO,

1-10.

R. H. (1969). Aggregates of irregular axoplasmic tubules in early Wallerian degeneration of guinea-pig optic nerve. Brain Res. 14, 745-748. LANDON, D. N. & HALL, S. (1976). The myelinated nerve fibre. In The Peripheral Nerve (ed. D. N. Landon), pp. 1-105. London: Chapman & Hall. LEHNINGER, A. L. (1975). Biochemistry. 2nd Edition. New York: Worth Publishers. LIEBERMAN, A. R. (1971). Microtubule-associated smooth endoplasmic reticulum in the frog's brain. Z. Zellforsch. mikrosk. Anat. 116, 564-577. LUFT, J. H. (1956). Permanganate - a new fixative for electron microscopy. J. biophys. biochem. Cytol. 2, 799-802. MORRE, D. J., MERRITT, W. D. & LEMBI, C. A. (1971). Connections between mitochondria and endoplasmic reticulum in rat liver and onion stem. Protoplasma 73, 43-49. MUNN, E. A. (1974). The Structure of Mitochondria. London, New York: Academic Press. NASS, M. M. K., NASS, S. & AFZELIUS, B. A. (1965). The general occurrence of mitochondrial DNA. ExplCellRes. 37, 516-539. NOVIKOFF, A. B. (1976). The endoplasmic reticulum: A cytochemist's view (A Review). Proc. natn. Acad. Sci. U.S.A. 73, 2781-2787. PALAY, S. L. & CHAN-PALAY, V. (1973). Cerebellar Cortex. Cytology and Organisation. Berlin, Heidelberg, New York: Springer. PERACCHIA, C. (1973). Low resistance junctions in crayfish. I. Two arrays of globules in junctional membranes. J. Cell Biol. 57, 54-65. PETERS, A., PALAY, S. L. & WEBSTER, H. DE F. (1970). The Fine Structure of the Nervous System: The Nervous and Supporting Cells. Philadelphia, London, Toronto: Saunders. POMERAT, C. M., HENDELMAN, W. J., RAIBORN, C. W. & MASSEY, J. F. (1967). Dynamic activities of nervous tissue in vitro. In The Neuron (ed. H. Hyd6n), pp. 119-178. Amsterdam: Elsevier. REITER, W. (1966). Ober das Raumsystem des endoplasmatischen RetikulumsvonHautnervenfasern. Untersuchungen an Serienschnitten. Z. Zellforsch. mikrosk. Anat. 72, 446-461. RINVIK, E. & GROFOVX, I. (1974). Cerebral projections to the nuclei ventralis lateralis and ventralis anterior thalami. Experimental electron microscopical and light microscopical studies in the cat. Anat. Embryol. 146, 95-111. ROBERTSON, J. D. (i960). Cell membranes and the origin of mitochondria. In Regional Neurochemistry; Regional Chemistry, Physiology and Pharmacology of the Nervous System (ed. S. S. Kety & J. Elkes), pp. 497-530. New York, London: Pergamon. ROBERTSON, E. A. & SCHULTZ, R. L. (1970). The impurities in commercial glutaraldehyde and their effect on the fixation of brain. J. Ultrastruct. Res. 30, 275-287. RUBY, J. R., DYER, R. F. & SKALHO, R. G. (1969). Continuities between mitochondria and endoplasmic reticulum in the mammalian ovary. Z. Zellforsch. mikrosk. Anat. 97, 30—37. RUSTIONI, A. & SOTELO, C. (1974). Some effects of chronic deafferentation on the ultrastructure of the nucleus gracilis of the cat. Brain Res. 73, 527-533. LAATSCH,


147

K. (1972). Electron microscopic observations on terminal boutons and synaptic structures in the anterior horn of the spinal cord in the adult cat. Okajimas Folia anat. jap. 48, 361-412. SANDBORN, E. B. (1966). Electron microscopy of the neuron membrane systems and filaments. Can.J. Physiol. Pkarmac. 44, 329-338. SCHNEPF, E. & BROWN, R. M. JR. (1971). On relationships between endosymbiosis and the origin of plastids and mitochondria. In Origin and Continuity of Cell Organelles (ed. J. Reinert & H. Ursprung), pp. 299-322. Berlin, Heidelberg, New York: Springer. SCOTT, R. E. (1976). Plasma membrane vesiculation: A new technique for isolation of plasma membranes. Science, N. Y. 194, 743-745. SMITH, R. S. (1980). The short term accumulation of axonally transported organelles in the region of localized lesions of single myelinated axons. J. Neurocytol. 9, 39—65. SPACEK, J. & LIEBERMAN, A. R. (1974). Three dimensional reconstruction in electron microscopy of the central nervous system. Sb. vid. Prace lek. Fak. Hradci Krdlove" 17, 203-222. SPACEK, J. & LIEBERMAN, A. R. (1977). Continuity between outer mitochondrial membranes and smooth endoplasmic reticulum in nervous tissue. In Proc. XVth Czechoslovak Conf. Electron Microsc. (ed. V. Viklicky & J. Ludv(k) CSAV, Vol. A, pp. 347-348. August 22-26, Prague. SPENCER, R. S. & SCHAUMBURG, H. H. (1977). Ultrastructural studies of the dying-back process. IV. Different vulnerability of PNS and CNS fibres in experimental central-peripheral distal axonopathies. J. Neuropath, exp. Neurol. 36, 300-320. SPRICGS, T. L. B., LEVER, J. D. & GRAHAM, J. D. P. (1967). Membrane connections at adrenergic terminals. J. Microscopie 6, 425-430. TEICHBERC, S. & HOLTZMAN, E. (1973). Axonal agranular reticulum and synaptic vesicles in cultured embryonic chick sympathetic neurons. J. Cell Biol. 57, 88-108. TEICHBERG, S., HOLTZMAN, E., CRAIN, S. M. & PETERSON, E. R. (1975). Circulation and turnover of synaptic vesicle membrane in cultured fetal mammalian spinal cord neurons. J. Cell Biol. 67, 215-230. TSUKITA, S. & ISHIKAWA, H. (1976). Three-dimensional distribution of smooth endoplasmic reticulum in myelinated axons. J. Electron Microsc. 25, 141-149. VOLK, T. L. (1971). Mitochondrial-tubular membrane interconnections in the rat adrenal cortex. Lab. Invest. 25, 349-355WEAKLEY, B. S. (1977). Specialization of the outer mitochondrial membrane during oocyte development. Cell Tiss. Res. 180, 515-528. WEBSTER, H. DE F. (1974). Peripheral nerve structure. In The Peripheral Nervous System (ed. J. I. Hubbard), pp. 3-26. New York, London: Plenum Press. WEISS, P. A. & MAYR, R. (1971). Neuronal organelles in neuroplasmic ('axonal') flow. I. Mitochondria. Acta neuropath., Suppl. V, 187-197. ZELENA, J., LUBINSKA, L. & GUTMANN, E. (1968). Accumulation of organelles at the ends of interrupted axons. Z. Zellforsch. mikrosk. Anat. 91, 200-219. SAITO,

{Received 3 April 1980)