Proc. Nati. Acad. Sci. USAVol. 81, pp. 788-792, February 1984Cell Biology

Lysosomes are associated with microtubules and not withintermediate filaments in cultured fibroblasts

(cytoskeleton/vimentin/double immunofluorescence)

MICHELE COLLOT*t, DANIEL LOUVARDt, AND S. J. SINGER*§*Department of Biology, University of California at San Diego, La Jolla, CA 92093; and hInstitut Pasteur, 75742 Paris, France

Contributed by S. J. Singer, October 21, 1983

ABSTRACT Double immunofluorescent labeling experi-ments for lysosomes and either microtubules or vimentin inter-mediate filaments in cultured well-spread fibroblasts show aremarkable degree of superposition of the lysosomes and themicrotubules. Under two different sets of conditions where themicrotubules and intermediate filaments are well segregatedfrom one another, the lysosomes remain codistributed with themicrotubules. It is suggested that this specific association oflysosomes with microtubules reflects some type(s) of linkage(s)between them and that such linkages may play an importantrole in the location and intracellular transport of lysosomesinside cells.

Microtubules (MT) and intermediate filaments (IF) havebeen increasingly recognized as playing important roles inregulating the location and movement of intracellular organ-elles (cf. ref. 1). In this article, we are specifically concernedabout the relationships between lysosomes and cytoskeletalelements. Previous studies have suggested an involvementofMT or IF or both in lysosome activities and functions. Forexample, the saltatory movements of lysosomes and otherorganelles have been correlated with MT integrity (2, 3), theintracellular distribution of lysosomes in mouse macro-phages has been correlated with the degree of MT and IFextension into the cytoplasm (4), and the secretion of lyso-somal enzymes in polymorphonuclear leukocytes wasshown to be affected by the colchicine-induced disruption ofMT (5). These effects, however, only indirectly suggest anassociation of lysosomes with cytoskeletal elements, andmore direct evidence concerning a possible physical associa-tion is desirable.

In principle, electron microscopy is the technique ofchoice for the detection and characterization of any physicallinkages between specific organelles and cytoskeletal fila-ments. For cells with an amorphous internal structure, how-ever, transmission electron microscopy of random ultrathinsections may be of only limited use for this purpose, becauseit may be difficult to find linkage elements in any given fieldof a section. Under these circumstances, double immunoflu-orescence light microscopy may yield significant informationabout organelle-filament associations. Despite the low reso-lution of the method, it may reveal distinct spatial correla-tions of a specific immunolabeled organelle with a particulartype of immunolabeled cytoskeletal filament throughout anentire cell. If other filament types show no such spatial cor-relations with the organelle in question, the suggestion is thata specific association exists between the organelle and thespatially correlated filament. We have used this method pre-viously to provide evidence for an association of mitochon-dria with MT (6, 7) in a variety of cultured cells. In this pa-per, similar double, indirect immunofluorescence experi-

ments have indicated a strong spatial correlation ofimmunolabeled lysosomes with MT in cultured fibroblasts,but not with vimentin IF, suggesting that some kind(s) oflinkages exist between lysosomes and MT in these cells.

MATERIALS AND METHODSAntibody Preparations. Antisera to lysosomes (8) were

raised in rabbits to a membrane fraction of lysosomes isolat-ed from rat liver, and anti-lysosome antibodies were purifiedby extensive absorption of the sera with other cellular frac-tions (9). The anti-lysosome antibodies reacted with both pri-mary and secondary lysosomes, according to immunoper-oxidase labeling in electron microscopy, and occasionallywith some Golgi stacks and the plasma membrane (8). Theantigen recognized by these antibodies appears to be a100,000-dalton protein, according to immunolabeling proce-dures. The guinea pig antibodies to chick brain tubulin and tovimentin isolated from baby hamster kidney cells, were af-finity-purified monospecific antibodies previously character-ized (10). Affinity-purified and cross-absorbed goat antibod-ies to rabbit IgG and to guinea pig IgG and their conjugationto fluorescein and rhodamine fluorophores were as de-scribed (10).

Cells and Immunofluorescence Labeling. All experimentswere carried out with normal rat kidney (NRK) cells infectedwith a temperature-sensitive mutant (LA23) of Rous sarco-ma virus. The infected cells were from stocks kindly sup-plied by P. K. Vogt of the University of Southern California.LA23-infected NRK cells were grown on coverslips either at39°C, where they exhibit the normal phenotype, or at 33°C,where they are transformed. In some experiments, cyclo-heximide (20 ,g/ml) was added to the medium 15 hr beforethe end of the growth period; in others, nocodazole (10,g/ml) was added 6 hr before. After a minimum of 2 days ofgrowth, the cells were fixed for 20 min with 3% formalde-hyde and then permeabilized by a 5-min exposure to 0.2%oTriton X-100. Immunolabeling was carried out as described(7, 10). In some of the experiments, in order to visualizemore clearly the spatial relationships of lysosomes and MT,optical leak-through of the rhodamine immunofluorescencefor tubulin into the fluorescein immunofluorescence for lyso-somes was deliberately generated. This was achieved by ap-propriately increasing the fluorescence intensity of tubulinlabeling and by longer photographic exposures, so as to pro-duce the desired effects. This required calibration with eachdifferent batch of primary and secondary antibodies used.Immunolabeling was observed by epifluorescence with aZeiss Photoscope-III instrument with the filters CZ487710for fluorescein fluorescence and CZ487714 for rhodamine.

Abbreviations: MT, microtubules; IF, intermediate filaments; NRKcells, normal rat kidney cells.tPresent address: Departement de Biologie Cellulaire, Facultes Uni-versitaires de Namur, 5000 Namur, Belgium.§To whom reprint requests should be addressed.

788

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

July

14,

202

0

Proc. Natl. Acad. Sci. USA 81 (1984) 789

RESULTSIn LA23-infected NRK cells grown at 390C and exhibitingthe normal phenotype, double immunolabeling for lyso-somes (Fig. 1A) and for microtubules (Fig. 1B) in the samecells revealed a high concentration of lysosomes in the vicin-ity of the microtubule-organizing center (arrows) near thecell nucleus. The exposure in Fig. 1A does not allow clearvisualization of the many lysomes out in the cell periphery,which are shown in Fig. 1C. In this field, the immunolabeledlysosomes are the bright dots of fluorescein fluorescence,and the background filamentous labeling is due to the opticalleak-through of the rhodamine immunolabeled microtubulesthat also were recorded separately in Fig. 1D through therhodamine optical filter. More than 80% of the labeled lyso-somes in the cell periphery were superimposed on labeledMT, and this number was increased under different photo-graphic conditions that revealed other labeled MT not appar-ent in Fig. 1C.

In such well-spread cells exhibiting the normal phenotype,it is known that MT and IF show substantial codistributions

in double immunofluorescence experiments (10, 11). There-fore, in order to discriminate whether the lysosome associa-tion was primarily with MT or IF, two sets of conditionswere used that result in an intracellular dissociation of thetwo filament types (7). One of these conditions involvedgrowth of the LA23-infected NRK cells at 330C, at whichtemperature the cells exhibited the transformed phenotype,and the vimentin IF were retracted around the nucleus (Fig.2D), while the MT remained extended to the cell periphery(Fig. 2B). The labeled lysosomes in such cells with retractedIF remained distributed out in the cell periphery (Fig. 2C), toa large extent superimposed on the labeled MT (Fig. 2 A andB). The second condition involved growth of the cells at 390Cin the presence of cycloheximide, which is known (12) tocause a retraction of the vimentin IF around the nucleus(Fig. 2H). Under these circumstances, the lysosomes werestill found in relatively normal amounts out in the cell periph-ery (Fig. 2G), where they were largely superimposed on MT(Fig. 2 E and F).

In the presence of nocodazole in LA23-infected NRK cells

IFIG. 1. LA23-infected NRK cells grown at 39°C (normal phenotype). Double indirect immunofluorescent labeling for lysosomes (bright dots

in A and C) and MT (B and D). Treatment with a mixture of rabbit anti-lysosome antibodies (40 Ag/ml, see text) and affinity-purified guinea piganti-tubulin antibodies was followed by treatment with a mixture of cross-adsorbed rhodamine- or fluorescein-conjugated affinity-purified goatantibodies to rabbit and guinea pig IgG. (A and D) Rhodamine labeling. (B and C) Fluorescein labeling. Note the high concentration of lyso-somes (arrow in A) in the vicinity of the MT-organizing center (arrow in B). Out in the cell periphery, most of the lysosomes (C) are closelyassociated with individual MT (C and D). The background filamentous labeling in C is due to the rhodamine fluorescence of labeled MT that wasallowed to leak through the fluorescein filter, producing an effect of superimposing the two patterns. The longer time of exposure ofC comparedto D allows the visualization of some additional labeled MT. All fields are at the same magnification. (Bar in B = 20 Am.)

Cell Biology: Collot et aL

Dow

nloa

ded

by g

uest

on

July

14,

202

0

Proc. NatL. Acad Sci. USA 81 (1984)

FIG. 2. (Legend appears at the bottom of the next page.)

790 Cell Biology: Collot et aL

Dow

nloa

ded

by g

uest

on

July

14,

202

0

Proc. NatL Acad Sci. USA 81 (1984) 791

FIG . 3. LA23-infected NRK cells grown at 39.C (normal pheno-type). Double indirect immunofluorescent labeling for lysosomes(A) and vimentin IF (B) after 6 hr of treatment with nocodazole at 10A.g/ml. For rhodamine labeling of lysosomes and simultaneous fluo-rescein labeling of IF, see Figs. 1 and 2, respectively, for proce-dures. Note that lysosomes are randomly distributed in the cyto-plasm, to some extent apparently in patches (large bright dots in A),but show no spatial correlation with the IF collapsed near the nucle-us (B). MT are completely disrupted under these conditions (notshown). (C) Same field in Nomarski optics. Cell edges are indicatedby arrowheads. (Bar = 20lttm.)

grown at 390C, the MT were disrupted and vimentin IF werecondensed to a perinuclear distribution, as has been ob-served by others in similar systems (cf. ref. 13). In such no-codazole-treated cells, the lysosomes appeared to be aggre-gated to some extent (large bright spots in Fig. 3A) but other-wise to be distributed randomly in the cytoplasm andshowed no spatial correlation with the collapsed IF (Fig.3B). In particular, there was no longer a concentration oflysosomes near the nucleus, as observed in Fig. 1A, in thesecells in which the microtubule-organizing center was disrupt-ed.

DISCUSSIONWe have provided evidence by double immunofluorescencelight microscopy that lysosomes are to a remarkable extentcodistributed with MT inside cultured fibroblasts. The anti-lysosome antibodies used in this study (8) did not discrimi-nate between primary and secondary lysosomes, so that thisconclusion presumably applies to both types. The codistri-bution of lysosomes with MT is manifested in several cir-cumstances.

(i) A concentration of lysosomes is spatially coincidentwith the concentration ofMT that exists in the vicinity of themicrotubule-organizing center (Fig. 1 A and B).

(ii) A large fraction of the lysosomes that are located outin the periphery of the cell are closely associated with thedispersed and convoluted MT that are also present there(Fig. 1 C and D).

(iii) Because the interpretation of this last observation iscomplicated by the fact that MT and vimentin IF are codis-tributed with one another in LA23-infected NRK cells grownat 390C (10), it is especially interesting that under two quitedifferent sets of conditions where the vimentin IF are retract-ed to a perinuclear distribution away from the MT, a largenumber of lysosomes remain out in the cell periphery codis-tributed with the MT still located there. This is true forLA23-infected NRK cells grown at 33TC exhibiting the trans-formed phenotype (Fig. 2 A-D) and for the cells grown at390C in the presence of cycloheximide exhibiting the normalphenotype (Fig. 2 E-H).

(iv) When the MT are disrupted with nocodazole and theIF become aggregated into perinuclear bundles (Fig. 3B), thelysosomes appear to be randomly dispersed (14pand no long-er tethered (Fig. 3A).These results indicate that lysosomes are extensively as-

sociated with MT but not with vimentin IF inside fibroblasts.Because there is much evidence that the distribution of actinmicrofilaments is not coordinated with either MT or IF inthese cells (cf. refs. 15 and 16), no significantly spatial corre-lation of lysosomes with microfilaments is indicated. There-fore, our results are not consistent with the suggestion ofMoore et al. (17) that there is an association of lysosomeswith microfilaments, as studied with polymorphonuclear leu-kocytes. The specificity and extent of the lysosome-MT co-distribution in the fibroblasts suggest that some kind or kindsof direct or indirect linkages exist between lysosomes andMT in these interphase cells that involve most of the lyso-somes at any given time. Such linkages could play importantroles in lysosomal activities, including the saltatory move-

FIG. 2 (on preceding page). LA23-infected NRK cells grown at 33°C (transformed phenotype) (A-D) and grown at 39°C (normal phenotype)(E-H) after 15 hr of treatment with cycloheximide at 20 pg/ml. Double, indirect immunofluorescent labeling for lysosomes (bright spots in Aand E) and MT (B and F) or for lysosomes (C and G) and vimentin IF (D and H). In the pairs A/B and E/F, lysosomes are fluorescein labeledand MT are rhodamine labeled. The rhodamine fluorescence for tubulin was allowed to leak through the fluorescein filter in A and E. In the pairsCID and G/H, lysosomes are rhodamine labeled. For IF labeling (D and H), affinity-purified guinea pig antibodies to vimentin were used,followed by fluorescein-conjugated affinity-purified goat antibodies to guinea pig IgG. Note in each case the coordinate distribution of lyso-somes (A and E) and MT (A/B and E/F). Both are concentrated in a perinuclear area, and lysosomes can be seen out in the cell peripherysuperimposed on individual MT (small arrows, for example). In this region (large arrows in CID and G/H) lysosomes are not associated withIF, which have retracted around the nucleus as indicated by the absence of vimentin labeling. Cell edges are indicated by arrowheads in E andG. All fields are at the same magnification. (Bar in D = 20 ,um.)

Cell Biology: Collot et aL

.,.4.:7. ..l.

A.

Dow

nloa

ded

by g

uest

on

July

14,

202

0

Proc. NatL Acad. Sci. USA 81 (1984)

ments of lysosomes inside cells that have been shown to stopupon disruption of microtubules (2, 3) and the secretoryfunctions of lysosomes, which also appear to depend uponthe integrity of microtubules (5).The evidence presented here for a preferential association

of lysosomes with MT does not exclude the possibility that,in other cells and under other metabolic conditions, lyso-somes may associate with IF. By means of immunofluores-cence observations similar to those presented here, we earli-er demonstrated (6, 7) an association of mitochondria withMT and not with IF in cultured cells. However, althoughthere is evidence from several studies that mitochondrial-MT associations are preferred in certain systems (1, 7, 18,19), there is also evidence for mitochondrial-IF interactionsin other cases (20-23).The present results make a specific contribution to the

growing body of evidence that most, if not all, intracellularorganelles may form direct or indirect linkages to one ormore cytoskeletal filaments in most eukaryotic cells. Instructurally well-organized nerve axons, regularly arrayedlinkages between MT or IF and various membrane-boundedorganelles have been observed by transmission electron mi-croscopy (20, 24, 25), although specific reference to linkagesto primary or secondary lysosomes has not often been madein these studies. It also is not evident that special cases suchas the nerve axon are of general relevance; the axon is spe-cialized for the long-distance transport of membrane-bound-ed organelles, and may exhibit cytoskeletal-organelle con-nections for the purpose that are not necessarily characteris-tic of other cells. However, the immunofluorescenceobservations of lysosome-MT association reported in thispaper and similar evidence for mitochondrion-MT (6, 7, 18)and for Golgi apparatus-MT (26) associations in amorphouscultured cells as well as electron microscopic evidence forassociations of nuclei with IF in striated muscle (22, 23, 27,28), of nuclei with MT/IF in other cells (29), and of secretorygranules with MT in cultured pancreatic beta cells (30) areamong the kinds of evidence for the existence of a widerange of organelle-cytoskeletal filament interactions in eu-karyotic cells, which very likely serve to locate these organ-elles and to direct their movements inside cells.The molecular bases for such organelle-cytoskeletal inter-

actions are only poorly understood. Whether different mem-brane-bound organelles that interact with a particular cyto-skeletal filament do so via a common recognition mechanismor via specific mechanisms for different organelles is notclear. In cases where MT are involved, the linkage elementsmay include the high molecular weight MT-associated pro-teins (MAPs) (31, 32), which project from microtubules (33)and which appear also to associate with IF (34-36). In vitroevidence implicating MT-associated proteins has been re-ported in the case of MT-secretory granule interactions (37),but further investigations along these lines are required witha variety of organelle-cytoskeletal filament systems.We are grateful to Dr. Adrienne Rogalski for very helpful advice

and discussions. We thank Mrs. Margie Adams and Mrs. BrachaYacobsen for their excellent technical assistance. M.C. was sup-ported in part by a grant from the International Federation of Uni-versity Women. These studies were funded by National Institutes ofHealth Grant GM-15971 to S.J.S., who is an American Cancer Soci-ety Research Professor.

1. Hayden, J. H., Allen, R. D. & Goldman, R. D. (1983) Cell Mo-til. 3, 1-19.

2. Freed, J. J. & Lebowitz, M. M. (1970) J. Cell Biol. 45, 334-354.

3. Wang, E. & Goldman, R. D. (1978) J. Cell Biol. 79, 708-726.4. Phaire-Washington, L., Silverstein, S. C. & Wang, E. (1980) J.

Cell Biol. 86, 641-655.5. Hoffstein, S., Goldstein, I. M. & Weissman, G. (1977) J. Cell

Biol. 73, 245-256.6. Heggeness, M. H., Simon, M. & Singer, S. J. (1978) Proc.

Natl. Acad. Sci. USA 75, 3863-3866.7. Ball, E. H. & Singer, S. J. (1982) Proc. Natl. Acad. Sci. USA

79, 123-126.8. Reggio, H., Harms, E., Bainton, D., Coudrier, E. & Louvard,

D. (1982) J. Cell Biol. 95, 413a (abstr.).9. Louvard, D., Reggio, H. & Warren, G. (1982) J. Cell Biol. 92,

92-107.10. Ball, E. H. & Singer, S. J. (1981) Proc. Natl. Acad. Sci. USA

78, 6986-6990.11. Geiger, B. & Singer, S. J. (1980) Proc. Natl. Acad. Sci. USA

77, 4769-4773.12. Sharpe, A. H., Chen, L. B., Murphy, J. R. & Fields, B. N.

(1980) Proc. Natl. Acad. Sci. USA 77, 7267-7271.13. Goldman, R. D. & Knipe, D. M. (1972) Cold Spring Harbor

Symp. Quant. Biol. 37, 523-534.14. Robbins, E. & Gonatas, N. K. (1964) J. Histochem. Cyto-

chem. 12, 704-711.15. Heggeness, M. H., Wang, K. & Singer, S. J. (1977) Proc.

Natl. Acad. Sci. USA 74, 3883-3887.16. Gotlieb, A. I., Heggeness, M. H., Ash, J. F. & Singer, S. J.

(1979) J. Cell. Physiol. 100, 563-578.17. Moore, P. L., Bank, H. L., Brissie, N. T. & Spicer, S. S.

(1976) J. Cell Biol. 71, 659-666.18. Couchman, J. R. & Rees, D. A. (1982) Eur. J. Cell Biol. 27,

47-54.19. Papasozomenos, S. C., Yoon, M., Crane, R., Autilio-Gam-

betti, L. & Gambetti, P. (1982) J. Cell Biol. 95, 672-675.20. Hirokawa, N. (1982) J. Cell Biol. 94, 129-142.21. David-Ferreira, K. L. & David-Ferriera, J. F. (1980) Cell Biol.

Int. Rep. 4, 655-662.22. Tokuyasu, K. T., Dutton, A. H. & Singer, S. J. (1983) J. Cell

Biol. 96, 1727-1735.23. Tokuyasu, K. T., Dutton, A. H. & Singer, S. J. (1983) J. Cell

Biol. 96, 1736-1742.24. Smith, D. S., Jarlfors, U. & Cager, M. L. (1977) J. Cell Sci.

27, 235-272.25. Ellisman, M. H. & Porter, K. R. (1980) J. Cell Biol. 87, 464-

479.26. Rogalski, A. A., Bergmann, J. E. & Singer, S. J. (1982) J. Cell

Biol. 95, 337a (abstr.).27. Ferrans, V. J. & Roberts, W. C. (1973) J. Mol. Cell. Cardiol.

5, 247-257.28. Behrendt, H. (1977) Cell Tissue Res. 180, 303-315.29. Wang, E., Cross, R. K. & Choppin, P. W. (1979) J. Cell Biol.

83, 320-337.30. Orci, L., Like, A. A., Amherdt, M., Blondel, B., Kanazawa,

Y., Marliss, E. B., Lambert, A. E., Wollheim, C. B. & Ren-old, A. E. (1973) J. Ultrastruct. Res. 43, 270-297.

31. Murphy, D. B. & Borisy, G. (1975) Proc. Natl. Acad. Sci.USA 72, 2696-2700.

32. Dentler, W. L., Granett, S. & Rosenbaum, J. L. (1975) J. CellBiol. 65, 237-241.

33. Kim, H., Binder, L. I. & Rosenbaum, J. L. (1979) J. Cell Biol.80, 266-276.

34. Pytela, R. & Wyche, G. (1980) Proc. Natl. Acad. Sci. USA 77,4808-4812.

35. LeTerrier, J.-F., Leim, R. K. H. & Shelanski, M. L. (1982) J.Cell Biol. 95, 982-986.

36. Bloom, G. S. & Vallee, R. B. (1983) J. Cell Biol. 96, 1523-1531.

37. Suprenant, K. A. & Dentler, W. L. (1982) J. Cell Biol. 93, 167-174.

792 Cell Biology: Collot et aL

Dow

nloa

ded

by g

uest

on

July

14,

202

0