Effect of Microtubules and Intermediate Filaments on Mitochondrial Distribution

Home , Intermediate filament, Microfilament, Microtubule

J. Cell Sd. 61, 87-105 (1983) 87 Printed in Great Britain © The Company of Biologists Limited 1983

EFFECT OF MICROTUBULES AND INTERMEDIATE FILAMENTS ON MITOCHONDRIAL DISTRIBUTION

IAN C. SUMMERHAYES, DAVID WONG AND LAN BO CHEN Dana-Farber Cancer Institute and Department of Pathology, Harvard Medical School, 44 Binney Street, Boston, MA 02115, U.SA.

SUMMARY The laser dye rhodamine 123 specifically stains mitochondria in living cells and facilitates the observation of changes in mitochondrial distribution in single cells under a variety of experimental conditions. Visualization of mitochondria in a number of cell lines followed by processing of these cells to study different cytoskeletal elements by indirect immunofluorescence, revealed good but not absolute correlation between mitochondria and microtubules or intermediate filaments. Mitochon- dria and microfilament distribution within the same cell did not show such a correlation. On the basis of observations made by various experimental approaches, we suggest that mitochondrial distribution is under the strong influence of the two systems, microtubules and intermediate filaments. Neither plays an absolute role but one seems able to play a more dominant role in the absence of the other.

INTRODUCTION Mitochondria are some of the most extensively studied organelles and have been characterized biochemically; their major role being in energy production essential for cell survival and proliferation (Lehninger, 1964; Tandler & Hoppel, 1972; Racker, 1976; Hinckle & McCarty, 1978). Earlier investigations have described the intracellular distribution of mitochondria and observed the high motility and con- siderable morphological heterogeneity displayed by these organelles (Lewis & Lewis, 1914; Palade, 1953; Gey, Shapres & Borysko, 1954; Tobioka & Biesele, 1956; Bier- ling, 1954; Mann, 1975); however, the mechanisms that determine the location and movement of such organelles are little understood. Electron-microscopic studies in various organisms, involving different organelles, have implicated microtubules as the directing influence in both cytoplasmic organization (Tilney & Porter, 1965), and organelle movement and distribution (Ledbetter & Porter, 1963; Whaley & Mollenhauer, 1963; Rudzinska, 1965; Bikle, Tilney & Por- ter, 1966; Holmes & Choppin, 1968; Murphy & Tilney, 1974; Smith, Jarlfors & Cameron, 1975; Smith, Jarlfors & Cayer, 1977). In all these studies the presence of microtubules has been demonstrated in close proximity to the organelle of interest, and movement of such intracellular packages has been observed parallel to the long axis of the microtubules. Physical connections between organelles and microtubules have been resolved in a few instances; for example, that of chromosome kinetochores with the spindle apparatus (Roth, Wilson & Chakraborty, 1966; Barnicot, 1966) and that of mitochondria in neuronal axons, by electron-dense cross-bridges between the 88 /. C. Summerhayes, D. Wong and L. B. Chen structures (Smith et al. 1975, 1977). Using anti-mitotic drugs such as colchicine, disassembly of microtubules can be shown to have profound effects on organelle movement (Holmes & Choppin, 1968) and distribution (Heggeness, Simon & Singer, 1978; Johnson, Walsh & Chen, 1980) within different cell types, providing strong evidence for the influence of microtubules on organelle distribution. With the recent discovery that rhodamine 123 and other permeant cationic fluorescent probes can be utilized for staining mitochondria in living cells (Johnson et al. 1980; Johnson, Walsh, Bockus & Chen, 1981; Johnson, Summerhayes & Chen, 1982), we have a unique opportunity to observe the distribution of this organelle without complicating fixation factors. Since the laser dye rhodamine 123 is found to be non-toxic to fibroblasts at low concentrations, we have been able to observe the mitochondrial distribution in a single cell prior to, during and after recovery from different drug treatments. In this paper we describe a possible involvement of intermediate filaments (see review by Lazarides, 1980) and microtubules in mitochondrial distribution.

MATERIALS AND METHODS Mitochondrial staining The laser dye rhodamine 123 (Eastman) was dissolved in dimethyl sulphoxide at a concentration of 1 mg/ml and subsequently diluted to 10/xg/ml in Dulbecco's modified Eagle's medium. Cells on coverslips were incubated with rhodamine 123 (10^g/ml) for lOminat 37 °C, rinsed in medium and mounted in medium supplemented with 5 % calf serum on a live-cell observation chamber (Johnson et al. 1980). Stained cells were viewed by epifluorescent illumination at 485 nm on a Zeiss photomicroscope III, and photographs were taken with a 40X Planapo objective lens using Tri-X film.

Immunofluorescence staining Three different preparative fixation techniques were used, depending on the filament system to be stained. Microtubules. Cells growing on coverslips were washed for 30 s at room temperature with stabilization buffer (01 M-piperazine-N,./V'-bis-2-ethanesulphonic acid, sodium salt adjusted to pH6-9 with KOH, 1 mM-ethylene glycol bisQS-aminoethyl ether)-AW-tetraacetic acid, 2-5 mM- GTP and 4 % polyethylene glycol 6000) and then incubated for 5 min at room temperature in the same buffer containing 05 % Triton X-100 (Sigma). After this treatment cells were washed twice with stabilization buffer and then fixed in cold methanol (-20°C) for 5 min (Osborn & Weber, 1977). Intermediate filaments. Coverslips were rinsed in phosphate-buffered saline (PBS) and fixed for 5 min in cold methanol ( — 20°C), then transferred back to PBS and washed in three changes during a 10 min period. Microfilaments. Cells were fixed in 3-7 % formaldehyde for 20 min, washed in PBS for 2-5 min and then permeabilized in cold acetone ( — 20°C) for 5 min. Cells were rinsed thoroughly in PBS before processing for immunofluorescence. After fixation cells were processed in the same manner. All cells were washed thoroughly in PBS after fixation, drained, overlaid with the appropriate cytoskeletal antiserum and incubated for 30 min at 37 °C in a humidified chamber. Coverslips were then rinsed thoroughly in PBS and overlaid with rhodamine-conjugated goat anti-rabbit immunoglobulin G (IgG; Miles) at a dilution of 1/10 and incubated for a further 30min at 37°C. After rinsing again, coverslips were mounted in gelatin/glycerol. Antibodies to cytoskeletal elements were generous gifts from Dr T. T. Sun (keratin) (Sun & Green, 1978), Dr F. Solomon (tubulin) (Solomon, Magendantz & Salzman, Mitochondrial distribution and cytoskeleton 89 1979), Dr K. Burridge (actin) (Burridge, 1976) and Dr R. O. Hynes (vimentin) (Hynes & Destree, 1978). Microinjection Mouse monoclonal antibody (JLB-7) (Lin, 1981), generously provided by Dr Jim J. C. Lin of Cold Spring Harbor Laboratory, was microinjected into cells grown on glass coverslips with a glass capillary needle drawn out to a tip of less than 0-5 jUrn in diameter by a Narishige PN-3 puller (Narishige Scientific Instrument, Japan). A Leitz micromanipulator equipped with a vacuum and pressure device was used for micromanipulation. A Leitz Diavert phase-contrast microscope equipped with an RCA Newvicon camera, a Panasonic television monitor and a Panasonic video recorder was used for visualization of cells during the course of the microinjection process. All procedures were essentially as described by Graessman & Graessman (1976), Feramisco (1979) and Johnson et al. (1982). Before fixation for immunofluorescence, microinjected cells were stained with rhodamine 123 and photographed as described above.

RESULTS Live gerbil fibroma cells (CCL146) incubated with rhodamine 123 (lO^ig/ml) for 10 min and viewed using epifluorescent microscopy, display a discontinuous filamentous array of cytoplasmic structures previously shown to be mitochondria (Johnson et al. 1980, 1981). All mitochondria of the cell take up the dye and show uniform fluorescent intensity within a single cell. Observation of the mitochondria in a par- ticular cell and relocation of the same cell after processing for immunofluorescence was used for the study of mitochondrial distribution and its correlation with different cytoskeletal elements.

Mitochondrial distribution and cytoskeletal elements Staining of gerbil fibroma cells with rhodamine 123 revealed filamentous mitochondria throughout the cytoplasm (Figs 1A, 2A, 3A), with variable morphology and distribution within and between cells. The non-uniform distribution of this organelle throughout the cytoplasm, where mitochondria often appeared in parallel array (Figs 1A, 2A, 3A), suggested a directive element present within the cellular environment. Photographic visualization of mitochondria followed by processing of the cells for indirect immunofluorescence, using antibodies to different cytoskeletal elements, revealed a close correlation between the mitochondrial distribution and microtubules within the same cell (Fig. 1A, B), as has been previously described by Heggeness et al. (1978). Cells studied in the same manner and stained for vimentin (the mesenchymal intermediate filament type) also showed a strong correlation in every cell, with mitochondria distributed along major intermediate filament pathways within the cell (Fig. 2A, B). In both cases the distribution and orientation of mitochondria within a single cell corresponded with the major filamentous networks visible after staining with tubulin or vimentin antibody, suggesting a possible channelling of mitochondria along these filaments. In contrast to these observations, processing of cells to visualize mitochondria and the cytoplasmic microfilament system recognized by actin antibody revealed no consistent correlation with respect to distribution. Major mitochondrial channels often displayed actin filaments running along their length; however, other /. C. Summerhayes, D. Wong and L. B. Chen

Fig. 1. Mitochondrial staining in live gerbil fibroma cell (A) and the corresponding microtubule (tubulin) staining in the same cell (B). Bar, 25 ^m. Mitochondrial distribution and cytoskeleton 91

Fig. 2. Mitochondrial staining in live gerbil fibroma cell (A) and the corresponding intermediate filament (vimentin) staining in the same cell (B). Bar, 30/im. Fig. 3. Mitochondrial staining in a portion of live gerbil fibroma cell (A) and the corresponding microfilament (actin) staining in the same cell (B). Bar, Mitochondrial distribution and cytoskeleton 93 areas within the same cell showed mitochondria and microfilaments orientated at 90 ° to one another (Fig. 3A, B), suggesting the absence of actin involvement in mitochondrial distribution.

Mitochondrial distribution and the effects of colchicine and eolcemid Colchicine (10//g/ml) and eolcemid (1 ^zg/ml), which depolymerize microtubules but not intermediate filaments (Goldman, 1971; Goldman & Knipe, 1972), enabled us to study the relative importance of microtubules and intermediate filaments in mitochondrial distribution. Staining of cells before and after a 3 h incubation period with colchicine, when microtubules were shown to be absent, demonstrates the effect of microtubule disruption on mitochondrial distribution. At this time in all cells studied mitochondria lost their extended filamentous form, appearing more wavy in contour. Staining of cells with tubulin antibody, after exposure to colchicine, revealed a complete absence of filamentous staining in most cells. Vimentin staining in cells after colchicine treatment for 3 h showed retraction of intermediate filaments from peripheral cell regions with a bundling of fibres (Fig. 4B) (Goldman, 1971; Goldman & Knipe, 1972) en route to the perinuclear region to form the characteristic nuclear whorl. Superimposing negatives of mitochondrial distribution and vimentin filaments in the same cell after colchicine treatment showed a strong correlation between the distribution of these two cytoplasmic elements (Fig. 4A, B), although the mitochondrial orientation was not always parallel to the intermediate filaments as was observed prior to drug treatment. Prolonged exposure of cells to colchicine (24-48 h) evoked a gradual retraction of vimentin filaments from peripheral regions of the cell, resulting in a coiled bundle of filaments surrounding the nucleus (Fig. 4D). Studies of mitochondrial distribution during this period revealed a temporally similar with- drawal (Fig. 4c) with mitochondria and vimentin filaments displaying a close correlation in distribution (Fig. 4c, D). Similar events were observed with cells exposed to eolcemid. However, none of the above effects were detected in lumicolchicine-treated cells.

Mitochondrial distribution after the removal of eolcemid Since interpretation of the previous data could possibly be accounted for by coin- cident distribution of vimentin and mitochondria due to entrapment in cytoplasmic areas after the removal of microtubules, we studied the distribution of mitochondria after the removal of eolcemid before extensive regrowth of microtubules. Cells exposed to eolcemid for 90min were stained with rhodamine 123, photographed, and then returned to culture in the absence of the drug. Coverslips were restained with the rhodamine dye at different times after eolcemid removal and then processed for indirect immunofluorescent studies. As before, the same cell was relocated at each of the different stages. In all colcemid-treated cells mitochondria appeared more com- pacted and closer to the nucleus (Figs 5A, B, 6A), as seen previously in colchicine- exposed cells. Staining of cells 30min after recovery from eolcemid showed few changes in mitochondrial distribution as assessed by fluorescence microscopy. Sixty minutes after removal of the drug, mitochondria had begun to be relocated in the cell /. C. Summerhayes, D. Wong and L. B. Chen

Fig. 4. Mitochondrial staining in live gerbil fibroma cells exposed to colchicine for 3 h (A) and 24 h (c). B and D. Vimentin staining of cells in A and c, respectively. Bar, 30 fim.

(Fig. 5B), with little microtubule regrowth (Fig. 5c). The small number of microtubules shown in Fig. 5c appear to be colcemid-resistant. Mitochondria were often localized in areas where no microtubules were found. At 90 min after the removal of colcemid, similar results were obtained (Fig. 5D, E, F). Similar studies 90 min after recovery from colcemid (Fig. 6A, B), followed by immunofluorescent staining to visualize vimentin filaments, revealed a good correlation between the redistributed mitochondrial pattern and vimentin filaments (Fig. 6B, C). However, mitochondria at this stage still appeared wavy in contour, lacking the more extended appearance of mitochondria in untreated cells. It appears that full recovery of the microtubule network is essential for the reappearance of normal mitochondrial morphology. Mitochondrial distribution and cytoskeleton 95

Fig. 5. Mitochondrial staining in live colchicine-treated gerbil fibroma cells (A, D) and the same cells 60 min (B) and 90 min (E) after the removal of colchicine. Microtubule staining of cells in B and E is shown in c and F, respectively. Bar, 30jUm. 96 /. C. Summerhayes, D. Wong and L. B. Chen

Fig. 6. Mitochondrial staining in live colchicine-treated gerbil fibroma cell (A) and the same cell 90 min after the removal of colchicine (B). Vimentin staining of the same cell is shown in c. Bar, 30 pirn. Mitochondrial distribution and cytoskeleton 97

Mitochondrial distribution and disruption of vimentin filaments Although the above experiments suggest that, in the absence of an intact microtubule system, mitochondrial distribution seems to be correlated with vimentin filament distribution, it is still not known how mitochondria would behave if microtubules were kept intact but vimentin distribution were altered. Until recently it has not been possible to alter vimentin organization without affecting microtubules. A recent report by Sharpe, Chen, Murphy & Fields (1980) describes one of the methods in monkey kidney CV-1 cells, where vimentin filament distribution was disrupted after protein synthesis was inhibited by cycloheximide treatment. The other method, involving microinjection of antibody against vimentin-filament-associated protein has been described by Lin & Feramisco (1981). Microtubule and microfilament arrange- ments appeared unaffected by either treatment, as assessed by indirect immunofluorescence. CV-1 cells show the same correlation of mitochondrial distribution with microtubules or intermediate filaments as observed in the gerbil fibroma cell line. Exposure of these cells to cycloheximide (lOjUg/ml) blocks 80% of protein synthesis within 5 min and results in disruption of the mitochondrial distribution after 3—5 h, when the vimentin filament pattern appears disorganized. Mitochondria in the cycloheximide- treated cells aggregate around the nucleus and often appear isolated at the cell periphery (Fig. 7B), lacking their usual orientation (Fig. 7A). Staining of vimentin in the same cell after cycloheximide treatment reveals disruption of intermediate filaments into disorganized arrays (Fig. 7c), but not depolymerization of the vimentin filaments. Staining of cells with tubulin antibodies, after the same period of cycloheximide exposure, shows a complete, extended, microtubular network apparently unaffected by drug treatment and similar to those reported by Sharpe et al. (1980). Recovery from these treatments results in reorganization of both mitochondria and vimentin filaments after 24 h.

Fig. 7. Mitochondrial staining in live CV-1 cell before (A) and after cycloheximide treatment (B). Vimentin staining of the same cell is shown in c. Bar, 30|Um. 98 /. C. Summerhayes, D. Wong and L. B. Chen Gerbil fibroma cells (CCL146) were microinjected with monoclonal antibody (JLB7) against vimentin-filament-associated protein (Lin, 1981). Fig. 8 shows that vimentin filaments gradually retract toward the perinuclear region upon microinjection, as described by Lin & Feramisco (1981). However, mitochondrial distribution is clearly unaffected by the altered vimentin organization. Unlike those in the cycloheximide experiment, these results suggest that an intact intermediate filament system is not an absolute requirement for normal mitochondrial distribution.

Fig. 8. Microinjection of gerbil fibroma cell with monoclonal antibody JBL7. Mitochon- dria staining (A) and vimentin staining of the same cell (B). Bar, 30/xm. Mitochondria! distribution and cytoskeleton 99

Mitochondrial distribution and intermediate filaments in epithelial cells Since intermediate filaments seem to be involved in mitochondrial distribution, it is of interest to examine the established epithelial cell lines in which the presence of two intermediate filament systems has been shown (Franke et al. 1979a; Franke, Schmid, Weber & Osborn, 19796). Keratin has been shown to be the major intermediate filament system of epithelial cells in culture (Franke et al. 19786; Sun & Green, 1978), but is absent from all the mesenchymal cells studied. In contrast, the occurrence of both the epithelial specific intermediate filament, keratin, and the mesenchymal-derived filament, vimentin, has been reported in a number of established epithelial cell lines. In some cases staining with antibody to either intermediate filament type reveals an extensive filamentous array (Franke et al. 1979a), facilitating the study of mitochondria and these cytoskeletal elements. The kangaroo kidney epithelial cell line, PtK2, is one such line and was used in this study. Incubation of PtK2 cells with rhodamine 123 revealed a mitochondrial distribution different from that observed with the gerbil cell line, as mitochondria were less filamentous, with more even distribution throughout the cytoplasm (Fig. 9A, C). Staining of cells with the mitochondrial dye, followed by processing of these cells to observe vimentin intermediate filaments, revealed a less clear-cut correlation of these two cytoplasmic elements within the same cell, compared with observations on mesenchymal cells (Fig. 9c, D). Similar experiments with these cells, to look at keratin distribution, revealed an elaborate array of keratin filaments throughout the cell, making interpretation difficult (Fig. 9A, B). Overlay of negatives of mitochondrial pattern and the corresponding keratin filaments in the cell showed mitochondria present with keratin in all regions.

Effect of colchicine on mitochondrial distribution in PtK2 cells In order to differentiate between the two intermediate filament systems expressed in PtK2 cells we exposed cells to colchicine, as it has been previously reported that vimentin will retract to the perinuclear region as in mesenchymal cells, but keratin will be little affected (Sun & Green, 1978; Franke et al. 19796). A 24-h exposure of cells to colchicine led to a more granular mitochondrial form accompanied by retraction of cells. Corresponding staining of retracted PtK2 cells for keratin showed keratin filaments present with mitochondria in all regions. In contrast, vimentin filaments in PtK2 cells clustered in the perinuclear region after 24 h treatment with colchicine (Fig. 10B) and did not appear to be correlated with mitochondrial distribution (Fig. 10A). The differential fluorescent intensities among different PtK2 cells seen in Fig. 10A reflected the different dye accumulation in mitochondria among these cells at the time of staining, as reported previously (Johnson et al. 1981, 1982).

Mitochondrial distribution in primary bladder epithelial cell cultures after colchicine treatment We have reported the absence of vimentin filaments in primary bladder epithelial cell cultures using immunofluorescent techniques and two-dimensional gel analysis, 100 /. C. Summerhayes, D. Wong and L. B. Chen

Fig. 9. Mitochondrial staining in live PtK-2 cells (A, C). Keratin staining of A is shown in B, and vimentin staining of c is shown in D, respectively. Bar, Mitochondria! distribution and cytoskeleton 101

Fig. 10. Mitochondrial staining in live PtK-2 cells treated with colchicine (A) ; and vimentin staining of the same cells (B). Bar, 20fitn. whereas an extensive keratin cytoskeleton is present in these cells (Summerhayes, Cheng, Sun & Chen, 1981). Keratin cytoskeletons in these cells are little affected by colchicine, and it is of interest to discover the factor controlling mitochondrial distribution in cells where vimentin is absent. Staining of epithelial outgrowths with rhodamine 123 reveals mitochondria distributed throughout the cytoplasm in filamentous array, similar to the staining observed in mesenchymal cells within the same culture. A 24-h exposure of cells to colchicine resulted in little change in staining in epithelial cells, in which mitochondria were still filamentous and extending out to the cell periphery. In contrast, mesenchymal cells at the edge of the same outgrowth displayed retraction of mitochondria toward the nucleus, consistent with our previous observations. Epithelial outgrowths displaying an unaffected mitochondrial distribution showed an extensive keratin cytoskeleton after immunofluorescence staining, which extended to all areas of the cell where mitochondria were present.

DISCUSSION The experiments reported in this paper indicate a possible role for intermediate filaments and microtubules in determining the distribution of mitochondria in a number of different cell types in culture. The correlation in distribution observed between mitochondria and different cytoskeletal elements suggests intermediate filaments and microtubules as possible candidates in determining mitochondrial organization and indicates that there is a close distribution of both filamentous networks within the cytoplasm (Geiger & Singer, 1980). In contrast, microfilament organiza^ tion in cells showed no consistent correlation with mitochondrial distribution, suggesting a lesser role for actin filaments in determining mitochondrial location. How- ever, the possibility still remained that these observations may only reflect the fact that microtubules and intermediate filaments share a major subcompartment of the cytoplasm with mitochondria, without specific interaction among them. 102 /. C. Summerhayes, D. Wong and L. B. Chen The rearrangement of mitochondria observed in cells exposed to colchicine was characterized by a compacting of these organelles and their retraction from the cell periphery. The depolymerization of microtubules did not result in a complete disor- ganization of mitochondrial distribution, but rather an apparent reorganization that was correlated closely with the movement of intermediate filaments toward the perinuclear region. It is not clear whether the redistribution and reorientation of mitochondria under these conditions is a direct result of disruption of a microtubule- mitochondrion linkage or an effect mediated through the rearrangement of intermediate filaments associated with the disassembly of microtubules (Goldman, 1971; Goldman & Knipe, 1972). A role for intermediate filaments in mitochondrial distribution was demonstrated by the disruption of vimentin filaments in CV-1 cells after exposure to cycloheximide. Consistent with a previous report by Sharpe et al. (1980), we observed disorganiza- tion of vimentin filaments in CV-1 cells after exposure to cycloheximide, where immunofluorescent staining revealed a complete loss of the organization that had been apparent prior to drug treatment. Concomitant with the disruption of vimentin filaments we observed a rearrangement of mitochondrial distribution into disorganized aggregates around the perinuclear region, despite the fact that microtubules and microfilaments in these cells appeared unperturbed by the drug treatment. Whereas these results suggest that in the presence of intact microtubules in CV-1 cells it is possible to observe alterations in distribution of both mitochondria and vimentin, microinjection of monoclonal antibody against protein associated with vimentin filaments shows that vimentin organization can be induced to change without altering mitochondrial distribution. Thus, an intact vimentin filament system is not essential for the normal distribution of mitochondria. It appears that in a microinjected cell, as long as the microtubule system is intact, normal mitochondrial distribution can be maintained. It is likely that mitochondrial distribution is strongly influenced by the two systems, microtubules and intermediate filaments. Neither plays an absolute role but one seems able to play a more dominant role in the absence of the other. It is impossible to attribute mitochondrial behaviour completely to either intermediate filaments or microtubules and it is clear that both play a major role in determining mitochondrial distribution. It seems likely that intermediate filaments and microtubules represent a multicomponent system involved in mitochondrial distribution and orientation (Wang & Goldman, 1978). Since intermediate filaments appear to have a role in mitochondrial distribution, established epithelial cell lines that express both keratin and vimentin intermediate filament types appear to have two possible alternative mitochondrial associations. In such cell lines (Franke et al. 1979a) mitochondrial distribution was not found to be correlated with vimentin filaments, even in the presence of colchicine when vimentin filaments had coiled to the perinuclear region. In primary epithelial cultures, where vimentin filaments are absent (Summerhayes et al. 1981), mitochondrial distribution seems totally insensitive to colchicine or colcemid. The maintenance of mitochondrial distribution under these conditions is contrary to observations in mesenchymal cells, but is consistent with the maintenance of an extended keratin Mitochondrial distribution and cytoskeleton 103 cytoskeleton under the same conditions (Sun & Green, 1978; Franke et al. 197%). Previous literature concerning organelle movement in cells of various organisms has provided convincing morphological evidence implicating microtubules as the directing influence in this phenomenon. Although this is likely to be the case in a number of cells from lower organisms, metazoan cells present a more complex picture, since intermediate filaments in these cells are often closely associated with microtubules (Wang & Goldman, 1978; Franke, Grund, Osborn & Weber, 1978a) and have been shown to have a distribution dependent on intact microtubule structures (Goldman, 1971). As a result, depolymerization of microtubules by anti-mitotic drugs greatly affects intermediate filament distribution in these cells. The correlation observed between mitochondrial distribution and intermediate filaments or microtubules in untreated cells further emphasizes the close distribution of these cytoskeletal elements, but does not necessarily suggest any physical association. In fact, disruption of vimentin filaments in CV-1 cells by cycloheximide argues strongly against any physical association, since microtubules remained unaffected in these cells.

We are grateful to Dr Marcia L. Walsh for her contribution to the initial phase of this project. This work has been supported by grants from National Cancer Institute, American Cancer Society and Muscular Dystrophy Association to L.B.C., who is the recipient of an American Cancer Society Faculty Research Award.

REFERENCES BARNICOT, N. A. (1966). A note on the structure of spindle fibres. J. Cell Sd. 1(2), 217-222. BIERLING, R. (1954). Beobachtungen zum Verhalten der Mitochondrien von Zellen in vitro. Z. Krebsforsh. 60, 31-46. BIKLE, D., TILNEY, L. G. & PORTER, K. R. (1966). Microtubules and pigment migration in the melanophores of Fundulus heteroclitus L. Protoplasma 61, 322-345. BURRIDGE, K. (1976). Changes in cellular glycoproteins after transformation: identification of specific glycoproteins and antigens in SDS gels. Proc. natn. Acad. Sa. U.SA. 73, 4457-4461. FERAMISCO, J. R. (1979). Microinjection of fluorescently labeled a-actinin into living fibroblasts. Proc. natn. Acad. Sd. U.SA. 76, 3967-3971. FRANKE, W. W., GRUND, C, OSBORN, M. & WEBER, K. (1978a). The intermediate-sized filaments in rat kangaroo PtK2 cells. I. Morphology in situ. Cytobiologie 17, 365-391. FRANKE, W. W., SCHMID, E., BREITKREUTZ, D., LUEDER, M., BOUKAMP, P., FUSENIG, N. E., OSBORN, M. & WEBER, K. (1979a). Simultaneous expression of two different types of intermediate sized filaments in mouse keratinocytes proliferating in vitro. Differentiation 14, 35-50. FRANKE, W. W., SCHMID, E., WEBER, K. & OSBORN, M. (19796). HeLa cells contain intermediate-sized filaments of the prekeratin type. Expl Cell Res. 118, 95-109. FRANKE, W. W., WEBER, K., OSBORN, M., SCHMID, E. & FREUDENSTEIN, C. (19786). Antibody to prekeratin: decoration of tonofilament-like arrays in various cells of epithelial character. Expl Cell Res. 116, 429-445. GEIGER, B. & SINGER, S. J. (1980). Association of microtubules and intermediate filaments in chicken gizzard cells as detected by double immunofluorescence. Proc. natn. Acad. Sd. U.SA. 77, 4769-4773. GEY, G. O., SHAPRES, P. & BORYSKO, E. (1954). Activities and responses of living cells and their components as recorded by cinephase microscopy and electron microscopy. Ann. N.Y. Acad. Sci. 58, 1089-1109. GOLDMAN, R. (1971). The role of three cytoplasmic fibers in BHK-21 cell motility. I. Microtubules and the effect of colchicine.J. Cell Biol. 51, 752-762. GOLDMAN, R. & KNIPE, D. (1972). Functions of cytoplasmic fibers in non-muscle cell motility. Cold Spring Harbor Symp. Quant. Biol. 37, 523-534. 104 /. C. Summerhayes, D. Wong and L. B. Chen GRAESMAN, M. & GRAESMAN, A. (1976). Early simian virus 40-specific RNA contains information for turnover antigen formation and chromatin replication. Pmc. natn. Acad. Sci. U.S.A. 73, 336-370. HEGGENESS, M. H., SIMON, M. & SINGER, S. J. (1978). Association of mitochondria with microtubules in cultured cells. Proc. natn. Acad. Sci. U.S.A. 75, 3863-3866. HINCKLE, P. C. & MCCARTY, R. E. (1978). How cells make ATP. Scient. Am. 283, 104-123. HOLMES, K. V. & CHOPPIN, P. W. (1968). On the role of microtubules in movement and alignment of nuclei in virus-induced syncytia.J. Cell Biol. 39, 526-543. HYNES, R. O. & DESTREE, A. T. (1978). 10 nm filaments in normal and transformed cells. Cell 13, 151-163. JOHNSON, L. V., WALSH, M. L., BOCKUS, B. J. & CHEN, L. B. (1981). Monitoring of relative mitochondrial membrane potential in living cells by fluorescence microscopy. J. Cell Biol. 88, 526-535. JOHNSON, L. V., WALSH, M. L. & CHEN, L. B. (1980). Localization of mitochondria in living cells with rhodamine 123. Pmc. natn. Acad. Sci. U.SA. 77, 990-994. JOHNSON, L. V., SUMMERHAYES, I. C. & CHEN, L. B. (1982). Decreased uptake and retention of rhodamine 123 by mitochondria in feline sarcoma virus-transformed mink cells. Cell 28, 7-14. LAZARIDES, E. (1980). Intermediate filaments as mechanical integrators of cellular space. Nature, Land. 283, 249-256. LEDBETTER, M. C. & PORTER, K. R. (1963). A 'microtubule' in plant cell fine structure. J. Cell Biol. 19, 239-250. LEHNINGER, A. L. (1964). The Mitochondrion, pp. 1-70. New York: Benjamin. LEWIS, M. R. & LEWIS, W. H. (1914). Mitochondria (and other cytoplasmic structures) in tissue cultures. Am. J.Anat. 17, 339-401. LIN, J. J. C. (1981). Monoclonal antibodies against myofibrillar components of rat skeletal muscle decorate the intermediate filaments of cultured cells. Proc. natn. Acad. Sci. U.SA. 78, 2335-2339. LIN, J. J. C. & FERAMISCO, J. R. (1981). Disruption of the in vivo distribution of the intermediate filaments in fibroblasts through the microinjection of a specific monoclonal antibody. Cell 24, 185-193. MANN, E. A., ed. (1975). The Structure of Mitochondria, pp. 1-80. New York: Academic Press. MURPHY, D. B. & TILNEY, L. B. (1974). The role of microtubules in the movement of pigment granules in teleost melanophores. J. Cell Biol. 61, 757-779. OSBORN, M. & WEBER, K. (1977). The display of microtubules in transformed cells. Cell 12, 561-571. PALADE, G. E. (1953). An electron microscope study of the mitochondrial structure.,?. Histochem. Cytochem. 1, 188-211. RACKER, E. (1976). A New Look at Mechanisms in Bioenergetics, pp. 1-80. New York: Academic Press. ROTH, L. E., WILSON, H. J. & CHAKRABORTY, J. (1966). Anaphase structure in mitotic cells typified by spindle elongation.,7. Ultrastruct. Res. 14, 460-483. RUDZINSKA, M. A. (1965). The fine structure and function of the tentacle in Tokophyra infusorium. J. Cell Biol. 25, 459-477. SHARPE, A. H., CHEN, L. B., MURPHY, J. R. & FIELDS, B. N. (1980). Specific disruption of the vimentin filament system in CV-1 cells by inhibiting of protein synthesis. Proc. natn. Acad. Sci. U.SA. 77, 7267-7271. SMITH, D. S., JARLFORS, U. & CAMERON, B. F. (1975). Morphological evidence for the participa- tion of microtubules in axonal transport. Ann. N.Y. Acad. Sci. 253, 472-506. SMITH, D. S., JARLFORS, U. & CAYER, M. L. (1977). Structural cross-bridges between microtubules in central axons of an insect (Periplaneta americana). J. Cell Sci. 27, 255-272. SOLOMON, F., MAGENDANTZ, M. & SALZMAN, A. (1979). Identification with cellular microtubules of one of the co-assemblying microtubule-associated proteins. Cell 18, 431-438. SUMMERHAYES, I. C, CHENG, Y. S., SUN, T. T. & CHEN, L. B. (1981). Expression of keratin and vimentin intermediate-sized filaments in rabbit bladder epithelium at different stages of neoplas- tic progression. J. Cell Biol. 90, 63-69. SUN, T. T. & GREEN, H. (1978). Immunofluorescent staining of keratin fibers in cultured cells. Cell 14, 469-476. Mitochondrial distribution and cytoskeleton 105 TANDLER, B. & HOPPEL, C. L. (1972). Mitochondria, pp. 1-59. New York: Academic Press. TILNEY, L. G. & PORTER, K. R. (1965). Studies on microtubules in Heliozoa, I. Pmtoplasma 60, 317-344. TOBIOKA, M. & BIESELE, J. J. (1956). Mitochondria in living cells. An analysis of movements..?. biophys. biochem. cytol. 2 (suppl.), 319-324. WANG, E. & GOLDMAN, R. D. (1978). Functions of cytoplasmic fibers in intracellular movements in BHK-21 cells. J. CellBiol. 79, 708-726. WHALEY, W. G. & MOLLENHAUER, M. H. (1963). The Golgi apparatus and cell plate formation - A postulate. J. Cell Biol. 17, 216-225.

(Received 1 October 1982-Accepted 19 November 1982)