Dynamic aspects of intermediate filament networks ... - Semantic Scholar

Proc. Natl. Acad. Sci. USA Vol. 86, pp. 549-553, January 1989 Cell Biology

Dynamic aspects of intermediate filament networks in BHK-21 cells (biotinylation/intermediate filament organizing center/microinjection)

KAREN L. VIKSTROM*, GARY G. BORISYt,

AND

ROBERT D. GOLDMAN*

*Department of Cell Biology and Anatomy, Northwestern University Medical School, Chicago, IL 60611; and Wisconsin, Madison, WI 53706

tLaboratory of Molecular Biology, University of

Communicated by Laszlo Lorand, October 6, 1988 (received for review July 19, 1988)

10% tryptose phosphate broth, and antibiotics. Twenty-four hours before use, cells were treated with trypsin and plated onto etched glass locator coverslips (Bellco Glass) that had been coated with a solution (1 mg/ml) of poly-L-lysine. Preparation of Vimentin. Vimentin was purified from bovine lens by a modification of published procedures (16, 17). Fresh or frozen lenses (30 g) were homogenized at 40C in 250 ml of buffer H (50 mM Tris HCl, pH 7.4/5 mM MgCI2/0.2% 2-mercaptoethanol/1 mM PhMeSO2F) and the homogenate was centrifuged for 20 min (25,000 x g; 40C). The resulting pellet was washed twice by resuspension in 250 ml of buffer H followed by centrifugation. The final pellets were extracted for 4 hr at 40C with 100 ml of buffer E (8 M urea/50 mM Tris HCI, pH 7.4/1 mM EGTA/0.2% 2mercaptoethanol/1 mM PhMeSO2F). The urea extract was clarified by centrifugation (100,000 x g; 30 min; 40C) and protein remaining in the supernatant was precipitated by the addition of solid ammonium sulfate at 40C (36 g per 100 ml). The precipitated protein was dissolved in 10 ml of HA buffer (6 M urea/8 mM sodium phosphate, pH 7.2/0.140 M NaCl/1 mM dithiothreitol/0.1 mM PhMeSO2F), desalted by passage through a Sephadex G-25 column (2.5 x 18 cm), and then applied to a hydroxylapatite column (2.5 x 9.5 cm) that had been equilibrated in HA buffer. The column was then washed with one-halfthe column volume of HA buffer and the protein was eluted with an 8-40 mM linear sodium phosphate gradient in HA buffer (450 ml). All column chromatography was performed at room temperature. The resulting fractions were assayed by NaDodSO4/PAGE on slab gels by the buffer system of Laemmli (18). Vimentin-containing fractions were pooled and dialyzed versus 500 vol of assembly buffer (6 mM sodium potassium phosphate, pH 7.4/0.170 M NaCI/3 mM KCl/0.2% 2-mercaptoethanol/0.2 mM PhMeSO2F) to assemble IF (14, 15). Suspensions of IF were frozen dropwise in liquid N2 and stored at -70°C until further use. Frozen IFs were thawed and cycled once by disassembly/reassembly (see below) prior to use. Biotinylation of IF Protein (Vimentin). IFs were centrifuged (100,000 x g; 30 min), disassembled in 8 M urea/5 mM sodium phosphate, pH 7.2/0.2% 2-mercaptoethanol/1 mM PhMeSO2F, and then dialyzed against low ionic strength buffer (5 mM sodium phosphate, pH 7.4/0.2% 2-mercaptoethanol/0.2 mM PhMeSO2F). This buffer maintained the vimentin in a depolymerized state (14). The soluble protein (2.5-3.0 mg/ml) was incubated in a buffer containing a 100:1 molar excess of N-hydroxysuccinimidobiotin (Molecular Probes), 10% dimethylformamide, and 0.170 M NaCl. The N-hydroxysuccinimidobiotin was dissolved in dimethylformamide immediately before it was added to the reaction mixture. Under these conditions, IF polymerized during the biotinylation reaction. After 90 min, this preparation was centrifuged (100,000 x g; 30 min) and the resulting IF pellet was subjected to two cycles of disassembly/reassembly using solubilization in urea buffer to disassemble the IF and dialysis

A procedure was developed for the conjugaABSTRACT tion of vimentin with biotin. Biotinylated vimentin was then microinjected into BHK-21 cells and the fate of the labeled protein was determined at various times postinjection by indirect immunofluorescence. Microinjected vimentin could be traced through a specific sequence of morphological changes ultimately resulting in the formation of a filamentous network. The injected protein was first detected in spots dispersed throughout the cytoplasm. Subsequently, these spots appeared to cluster near the nucleus where they merged into a diffuse "cap." This cap coincided with a concentration of endogenous intermediate filaments and eventually gave rise to a filamentous network that was coincident with the endogenous intermediate filament network as determined by double-label immunofluorescence. The results indicate that the incorporation of exogenous vimentin into a filamentous network is initiated in a perinuclear region and progresses in a polarized fashion toward the cell surface.

Much has been learned about the in vivo dynamics of two of the major cytoskeletal systems, microtubules and microfilaments, through studies of the fate of microinjected tubulin and actin (1-7). The incorporation of fluorescently labeled tubulin and actin into microtubules and microfilaments has been detected directly in living cells by low-light-sensitive video systems (1, 2, 5, 6). In addition, fluorescein- or biotin-conjugated tubulin has been detected indirectly by immunofluorescence using an antibody directed against fluorescein or biotin (3, 4, 7). These techniques have permitted investigators to study processes in vivo, such as the polarized growth of microtubules (3, 7). The apparent stability of intermediate filaments (IFs) in vivo and the insolubility of IF proteins in most nondenaturing solutions have led to the suggestion that microinjection of IF protein might be difficult or even impossible (8). Based on these considerations, more indirect approaches have been taken to study the dynamics of IF assembly in vivo, such as microinjection of keratin mRNA (8, 9) and transfection of normal and altered IF protein genes (10-12). Although these methods avoid the solubility problem associated with IF proteins, they do not permit one to examine the exchange of protein between IF and a pool of translated subunits. However, by taking advantage of its solubility in low ionic strength buffers (13-15), we have initiated studies involving the microinjection of vimentin into baby hamster kidney (BHK-21) cells. This approach should enable us to address questions concerning the exchange between polymers and subunits in the IF system in addition to determining whether there is polarity in the growth of IF in vivo.

MATERIALS AND METHODS Cell Culture. BHK-21 cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% calf serum, The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Abbreviation: IF, intermediate filament.

549

550

Proc. Natl. Acad. Sci. USA 86

Cell Biology: Vikstrom et al.

against assembly buffer to reassemble them (14, 15). IF formation was assayed after each cycle of disassembly/ reassembly by negative-stain electron microscopy with 1% aqueous uranyl acetate. Suspensions ofbiotinylated vimentin IF were frozen dropwise in liquid N2 and stored at -70'C until further use. Biotinylated vimentin stored in this manner would polymerize into IF when thawed and cycled once as described above. Bovine serum albumin was biotinylated by a similar procedure and then was dialyzed extensively against 5 mM sodium phosphate, pH 8.5/0.05% 2-mercaptoethanol. Immunoblotting Analysis. Immunoblotting analysis was conducted by the method of Towbin et al. (19). Microinijection of Biotinylated Vimentin. Frozen aliquots of biotinylated vimentin were thawed and collected by centrifugation (100,000 x g; 30 min). The resulting IF pellet was solubilized in urea buffer and the solution was exchanged to 5 mM sodium phosphate, pH 8.5/0.05% 2-mercaptoethanol by chromatography over Sephadex G-25. The resulting sample was clarified by centrifugation (100,000 x g; 30 min) immediately prior to microinjection. Protein concentrations in samples used for microinjection varied between 2 and 4 mg/ml as determined by the method of Bradford (20). Micropipettes (W-P Instruments, New Haven, CT) were prepared on an Industrial Science Associates micropipette puller (Ridgewood, NY). Microinjections were done with a Leitz Diavert phase-contrast microscope fitted with a Leitz micromanipulator according to standard techniques (21, 22). Pneumatic pressure was provided by a 10-ml glass syringe filled with a low viscosity silicon oil. Immunofluorescence. Microinjected cells were prepared for indirect immunofluorescence as described (23). The primary antibodies used were goat anti-biotin (Sigma), a monoclonal antibody against P-tubulin (Amersham), and a rabbit polyclonal antibody directed against vimentin and desmin from BHK-21 cells (23). Rhodamine-labeled phalloidin (Molecular Probes) was also used. Fluorescein-conjugated donkey anti-goat IgG, rhodamine-conjugated donkey anti-rabbit IgG (Jackson Immunoresearch), and rhodamineconjugated rabbit anti-mouse IgG (Kirkegard and Perry Laboratories, Gaithersburg, MD) were used as secondary antibodies.

RESULTS Preparation of Biotinylated Vimentin. Vimentin was purified from bovine lens and the resulting preparation was homogeneous on NaDodSO4/polyacrylamide gels as judged by Coomassie blue staining (Fig. 1). The protein preparation was biotinylated by using an N-hydroxylsuccinimidyl ester of biotin and purified by two cycles of disassembly/reassembly by using solubilization in urea and then dialysis against vimentin assembly buffer. The yield of the biotinylation reaction was typically 60-70%. When examined by negativestain electron microscopy, the labeled protein formed smooth-walled IFs, which were morphologically indistinguishable from unlabeled IFs. The labeling process did not cause appreciable degradation of the protein but the biotinylated vimentin exhibited a retarded mobility on NaDodSO4/ PAGE (Fig. 1). The biotin-vimentin was recognized by both anti-biotin and anti-vimentin antibody preparations as determined by immunoblotting analysis (Fig. 1). Injection of Biotin-Vimentin into BHK-21 Fibroblasts. Biotin-vimentin IFs were pelleted by centrifugation and disassembled in urea buffer. The urea buffer then was exchanged to 5 mM sodium phosphate, pH 8.5/0.05% 2-mercaptoethanol. Under these conditions, biotin-vimentin remained soluble at concentrations of 2-4 mg/ml. After clarification by centrifugation, the sample was injected into BHK-21 cells growing on etched grid (locator) coverslips.

(1989)

b

aow

1

2

3

4

5

6

FIG. 1. (a) Negative-stain electron micrographs of biotin-vimentin IFs. (Bar = 200 nm.) (b) Coomassie blue-stained NaDodSO4/ PAGE of biotin-vimentin (lane 1) and unlabeled bovine lens vimentin (lane 2). Immunoblots of biotin-vimentin (lanes 3 and 5) and unlabeled bovine lens vimentin (lanes 4 and 6) processed with anti-biotin (lanes 3 and 4) and anti-vimentin (lanes 5 and 6) antibody preparations.

Fate of the Inijected Protein. Cells were fixed at various time intervals after microinjection and processed for indirect immunofluorescence using an anti-biotin antibody preparation to recognize the labeled protein. Immediately after microinjection, biotinylated vimentin was present as numerous discrete bright spots of various sizes in one region of the cell, presumably near the site of injection. Within 30 min of microinjection, these spots were distributed more extensively throughout the cytoplasm (Fig. 2 a and b). After 30 min to 1 hr, the biotin-vimentin spots had become very closely associated with each other and were concentrated in a juxtanuclear region. Frequently, they had merged into more diffuse patterns with the appearance of a typical nuclearassociated IF cap (Fig. 2 c and d; refs. 15, 24, and 25). In most cells 1-2 hr postinjection, the spots were no longer detectable and the region of diffuse fluorescence was larger, with fibrous projections extending toward the cell borders (Figs. 2 c and d and 3c). Within 3-4 hr after microinjection, the biotinvimentin was located in a fibrous network that resembled a typical interphase IF pattern (Figs. 2 e and f and 3e). The Effect of Biotin-Vimnentin on the Endogenous IF Network. We were concerned that the bright spots of biotinylated vimentin seen within 30 min of injection might indicate a disruption in the endogenous IF network. Therefore, we processed injected cells for double-label indirect immuno-

Cell

;r.7W

Proc. Natl. Acad. Sci. USA 86 (1989)

Biology: Vikstrom et al.

551

'416

, 4~~~~~~~~~~~~W

d

p1.

*

.

*0

. 'W.

,0A.i

,

e'

FIG. 2. Immunofluorescence (a, c, and e) and phase-contrast (b, d, andf) micrographs of cells microinjected with biotin-vimentin and fixed after various time intervals. Cells were processed for indirect immunofltiorescence using an anti-biotin antibody. a and b, 30 min; c and d, 1 hr; e andf, 4 hr. (a-d, x1065; e andf, x640.)

fluorescence with an anti-biotin antibody and an antibody raised against the IF proteins of BHK-21 cells. Our results demonstrated that the endogenous network was not disrupted 30 min postinjection when the bright spots of biotinvimentin are apparent (Fig. 3 a and b). To begin to determine whether the microinjected protein can be incorporated into the existing IF system, or whether it forms a new IF system, cells were processed for doublelabel indirect immunofluorescence at longer time intervals after injection. The diffuse juxtanuclear accumulation of biotinylated vimentin present 1-2 hr postinjection coincided with a region containing a perinuclear concentration of endogenous IF protein that is typical of interphase fibroblasts (Fig. 3 c and d; refs. 26 and 27). Furthermore, the fibrous pattern of biotin-vimentin gradually became similar to the endogenous IF network and by 4 hr postinjection the two patterns were almost completely coincident (Fig. 3 e andf). As a further control, double-label immunofluorescence was used to examine the overall patterns of the two other

major cytoskeletal systems in BHK cells: the microfilament and microtubule systems. Fig. 4 demonstrates that the microfilament system was not disturbed. The microtubule system also appeared to be normal (not shown). In addition, when biotinylated bovine serum albumin was microinjected in the same buffer used for biotin-vimentin, cells exhibited only diffuse cytoplasmic staining for several hours after injection (not shown).

DISCUSSION Previous studies of IF dynamics have focused primarily on organizational changes of IF polymers under different physiological conditions, such as lymphocyte activation, muscle differentiation, and cultured cell attachment and spreading (24, 25, 28, 29). However, little is known about the molecular mechanisms by which these changes are effected. Various reports have supported the idea that IFs are primarily regulated at the transcriptional level (30) and that, once

552

Cell Biology: Vikstrom et al.

Proc. Natl. Acad Sci. USA 86 (1989)

FIG. 3. Double-label immunofluorescence micrographs. Cells were processed for double-label indirect immunofluorescence using an anti-biotin antibody (a, c, and e) and an antibody directed against vimentin and desmin (b, d, and f). a and b, 30 min; c and d, 1.5 hr; e and f, 4 hr. (a-d, x890; e andf, x780.)

translated, their constituent structural proteins are stably incorporated into an IF network (31). The insolubility of most types of IF in nondenaturing solutions (29) and the lack of a sizable pool of soluble IF subunits in the cell (32) have led to the widely held idea that, at the molecular level, IF polymers are static structures that do not exchange their subunits. Recent studies involving the transfection of mutant keratin genes have suggested that IFs may be affected by the presence of newly synthesized IF protein (12). However, the relationship between IF subunits and polymers in vivo has not been established in a direct fashion. The experiments described in this paper were aimed at determining whether the response of a cell to the "pool" of excess vimentin formed by microinjection can shed light on the interplay between various forms of IF protein in vivo. The results of these microinjection experiments indicate that microinjected vimentin is incorporated into an extensive IF network within several hours. The resulting array of biotin-

vimentin is very similar to the endogenous IF network, which may indicate incorporation into preexisting IF. However, it also could be the result of the formation of a new IF network that is coincident with the endogenous network. Since these two possibilities cannot be distinguished at the light microscope level, electron microscope localization of microinjected vimentin is under way to address this issue. In either case, the formation of a filamentous network from microinjected vimentin is the result of a reproducible sequence of cytoplasmic events, which suggests that endogenous IFs interact in a specific fashion with exogenous subunits. IFs have been reported to interact very closely with the nuclear surface (26, 27, 33), which also has been suggested as the site of initiation of IF assembly (26, 27, 34). From our studies, the incorporation of exogenous vimentin into an IF network appears to initiate in a region adjacent to the nucleus. Initially, the discrete spots of biotin-vimentin seen after microinjection are scattered throughout the cytoplasm,

Cell

Biology: Vikstrom et al.

Proc. Natl. Acad. Sci. USA 86 (1989)

553

FIG. 4. Double-label fluorescence micrographs. Thirty minutes after microinjection of biotin-vimentin, cells were processed for immunofluorescence with an anti-biotin antibody preparation (a) and stained with rhodamine-conjugated phalloidin (b). (x780.)

but later they accumulate next to the nucleus. In addition, this is the first location in which microinjected vimentin appears filamentous. This juxtanuclear region usually coincides with a large accumulation of endogenous IF and suggests the presence of an IF organizing center (26, 27, 34). Similar juxtanuclear accumulations (IF caps) are seen in BHK-21 cells in the early stages of spreading and after colchicine treatment (15, 27). Furthermore, the progression of fibers out of the cap toward the cell periphery mimics the behavior of vimentin IF in fibroblasts spreading onto solid substrates (24, 27). This directed movement of IFs may reflect an intrinsic polarity in the behavior of the IF system. However, it is not clear that this polarity exists in individual filaments since it has not yet been determined whether IF polymers themselves are polar (35), as is the case for microtubules and microfilaments. The apparent rate of incorporation of vimentin into filamentous structures in vivo appears to be much slower than that seen with actin and tubulin. However, the slower time course of incorporation seen with microinjected vimentin should prove to be advantageous in studying the complex series of changes we see in this system. Examining the progression of biotin-vimentin from nonfilamentous accumulations into IF-like structures may be a convenient way to study other similar changes in IF organization. For example, similar nonfilamentous accumulations of IF protein have been described in epithelial cells that have had their keratin networks disrupted by microinjection of anti-keratin antibodies (36, 37) or expression of mutant keratin genes (12). In both of these cases, the endogenous IF system is disrupted. However, cells microinjected with biotin-vimentin have discrete spots of this protein in the presence of an intact IF network and the spots appear to reorganize into a filamentous network. It is intriguing to note that the spots of biotinylated vimentin, seen early after microinjection, are reminiscent of the nonfilamentous accumulations of IF protein seen in some cultured cells during mitosis (33, 38, 39). These experiments provide evidence that IFs are not static structures but are dynamic cellular elements. The function behind changes in the IF system will be better understood once the mechanisms that control IF dynamics are determined. The changes occurring at the molecular level when IF subunits incorporate into polymers in the cell may give us insight into the function and regulation of this complex cytoskeletal system. The authors wish to express their thanks to Dr. Gary Gorbsky for his help and instruction. This work has been supported by grants from the National Cancer Institute and National Institute of General Medical Sciences.

1. Keith, C. H., Feramisco, J. R. & Shelanski, M. (1981) J. Cell Biol. 88, 234-240. 2. Saxton, W. M., Stemple, D. L., Leslie, R. J., Salmon, E. D., Zavortink, M. & McIntosh, J. R. (1984) J. Cell Biol. 99, 2175-2186. 3. Soltys, B. J. & Borisy, G. G. (1985) J. Cell Biol. 100, 1682-1689. 4. Sammak, P. J., Gorbsky, G. J. & Borisy, G. G. (1987) J. Cell Biol. 104, 395-405. 5. Taylor, D. L. & Wang, Y.-L. (1978) Proc. Natl. Acad. Sci. USA 75, 857861. 6. Kreis, T. E., Winterhaltar, K. H. & Birchmeier, W. (1979) Proc. Natl. Acad. Sci. USA 76, 3814-3818. 7. Schulze, E. & Kirschner, M. (1986) J. Cell Biol. 102, 1020- 1031. 8. Kreis, T. E., Geiger, B., Schmid, E., Jorcano, J. L. & Franke, W. W. (1983) Cell 32, 1125-1137. 9. Franke, W. W., Schmid, E., Mittnacht, S., Grund, C. & Jorcano, J. (1984) Cell 36, 813-825. 10. Quax, W., van den Broek, L., Vree Egberts, W., Raemaekers, F. & Bloemendal, H. (1985) Cell 43, 327-338. 11. van den Hevel, R. M. M., van Eys, G. J. J. M., Raemaekers, F. C. S., Quax, W. J., Vree Egberts, W. T. M., Schaart, G., Cuypers, H. T. M. & Bloemendal, H. (1987) J. Cell Sci. 88, 475-482. 12. Albers, K. & Fuchs, E. (1987) J. Cell Biol. 105, 791-806. 13. Ip, W., Hartzer, M. K., Pang, Y.-Y. S. & Robson, R. M. (1985) J. Mol. Biol. 183, 365-375. 14. Zackroff, R. V. & Goldman, R. D. (1979) Proc. Natl. Acad. Sci. USA 76, 6226-6230. 15. Starger, J. M., Brown, W. E., Goldman, A. E. & Goldman, R. D. (1978) J. Cell Biol. 78, 93-109. 16. Lieska, N., Chen, J., Maisel, H. & Romero-Herrara, A. E. (1980) Biochim. Biophys. Acta 626, 136-153. 17. Lehto, V.-P. & Virtanen, I. (1983) J. Cell Biol. 96, 703-716. 18. Laemmli, U. K. (1970) Nature (London) 227, 680-685. 19. Towbin, M., Staehelin, T. & Gordon, J. (1979) Proc. Natl. Acad. Sci. USA 76, 4350-4354. 20. Bradford, M. (1976) Anal. Biochem. 72, 248-254. 21. Graessmann, M. & Graessmann, A. (1976) Proc. Natl. Acad. Sci. USA 73, 366-370. 22. Kreis, T. E. & Birchmeier, W. (1982) Int. Rev. Cytol. 75, 209-227. 23. Yang, H., Lieska, N., Goldman, A. & Goldman, R. D. (1985) J. Cell Biol. 100, 620-631. 24. Goldman, R. & Follett, E. (1970) Science 169, 286-288. 25. Goldman, R. D. & Knipe, D. (1973) Cold Spring Harbor Symp. Quant. Biol. 37, 523-534. 26. Goldman, R., Goldman, A., Green, K., Jones, J., Lieska, N. & Yang, H. (1985) Ann. N. Y. Acad. Sci. 455, 1-17. 27. Goldman, R. D., Goldman, A. E., Green, K. J., Jones, J. C. R., Jones, S. M. & Yang, H. (1986) J. Cell Sci. Suppl. 5, 69-97. 28. Paulin-Levasseur, M. & Brown, D. L. (1987) Cell Motil. Cytol. 8, 227-237. 29. Lazarides, E. (1980) Nature (London) 283, 249-256. 30. Ngai, J., Capetanaki, Y. G. & Lazarides, E. (1984)J. Cell Biol. 99, 306-314. 31. Blikstad, I. & Lazarides, E. (1983) J. Cell Biol. 96, 1803-1808. 32. Soellner, P., Quinlan, R. A. & Franke, W. W. (1985) Proc. Natl. Acad. Sci. USA 82, 7929-7933. 33. Jones, J. C. R., Goldman, A. E., Yang, H.-Y. & Goldman, R. D. (1985) J. Cell Biol. 100, 93-102. 34. Eckert, B. S., Daley, R. A. & Parysek, L. M. (1982) J. Cell Biol. 92, 575578. 35. Steinert, P. M., Steven, A. C. & Roop, D. R. (1985) Cell 42, 411-419. 36. Klymkowsky, M. W., Miller, R. H. & Lane, E. B. (1983) J. Cell Biol. 96, 494-509. 37. Tolle, H.-G., Weber, K. & Osborn, M. (1985) Eur. J. Cell Biol. 38, 234-244. 38. Franke, W. W., Schmid, E. & Grund, C. (1982) Cell 30, 103-113. 39. Horwitz, B. H., Kupfer, H., Eishar, Z. & Geiger, B. (1981) Exp. Cell Res. 143, 281-290.