Proc. Natl. Acad. Sci. USA 78 (1981)

Correction. In the article "Specific disruption of vimentin fil-ament organization in monkey kidney CV-1 cells by diphtheriatoxin, exotoxin A, and cycloheximide" by Arlene H. Sharpe,Lan Bo Chen, John R. Murphy, and Bernard N. Fields, which

appeared in the December 1980 issue of Proc. Natl. Acad. Sci .USA (77, 7267-7271), Figs. 1 and 3 were reproduced poorly.They are printed again here.

FIG.1.(gtmo

FIG. 1. (Legend appears at the- bottom of the next page.)

3284 Corrections

Dow

nloa

ded

by g

uest

on

July

8, 2

020

Dow

nloa

ded

by g

uest

on

July

8, 2

020

Dow

nloa

ded

by g

uest

on

July

8, 2

020

Dow

nloa

ded

by g

uest

on

July

8, 2

020

Dow

nloa

ded

by g

uest

on

July

8, 2

020

Dow

nloa

ded

by g

uest

on

July

8, 2

020

Dow

nloa

ded

by g

uest

on

July

8, 2

020

Proc. Natl. Acad. Sci. USA 78 (1981) 3285

FIG. 3. Effects of 1011 M diphtheria toxin neutralized by antitoxin (A), CRM197 (B), 1011 MP. aeruginosa exotoxin A (C), and 10 jtg of cy-cloheximide per ml (D) on the vimentin filament system. The effect ofcycloheximide was reversed when growth medium containing cycloheximidewas removed and replaced with fresh medium. The appearance ofvimentin filaments at 0.5 hr (E), 1 hr (F), 2 hr (G), and 4 hr (H) after the removalof cycloheximide is shown. (A, F, 0, and H, bar = 40 pam; B. C, D, and E, bar = 25 /Am.)

FIG. 1 (on preceding page). (A-C) Organization of microtubules (A), microfilaments (B), and intermediate filaments (C) in control CV-1 cells.(D-F) Organization of microtubules (D), microfilaments (E), and intermediate filaments (F) in CV-1 cells treated with 10-11 M diphtheria toxinfor 22 hr. Cells were subjected to indirect immunofluorescence microscopy with antibody against tubulin, actin, and vimentin. Note that only theintermediate filament system is affected by diphtheria toxin. (A, B, D, and E, bar = 10 Am; C, bar = 40 tm; F, bar = 30 ,AM.)

Corrections

Dow

nloa

ded

by g

uest

on

July

8, 2

020

Proc. Natl. Acad. Scd. USAVol. 77, No. 12, pp. 7267-7271, December 1980Cell Biology

Specific disruption of vimentin filament organization in monkeykidney CV-1 cells by diphtheria toxin, exotoxin A, and cycloheximide

(intermediate filaments)

ARLENE H. SHARPE*, LAN Bo CHENtt, JOHN R. MURPHY*, AND BERNARD N. FIELDS*§*Department of Microbiology and Molecular Genetics, tDepartment of Pathology, and tSidney Farber Cancer Institute, Harvard Medical School; and§Department of Medicine, Section of Infectious Disease, Peter Bent Brigham Hospital, Boston, Massachusetts 02115

Communicated by A. M. Pappenheimer, Jr., August 18, 1980

ABSTRACT We have examined the effect of diphtheriatoxin, Pseudomonas aeruginosa exotoxin A, and cycloheximideon the CV-1 cell cytoskeleton. Within a few hours after pro-ducing an inhibition of cellular protein synthesis, all theseagents specifically disrupted the organization of the vimentinfilament system with no discernable effect on microtubules ormicrofilaments during the period of observation. Furthermore,just as the inhibition of protein synthesis by cycloheximide isreversible, so was the disruption of vimentin filaments by cy-cloheximide.

The mechanisms by which infectious agents and their toxicproducts damage cells are poorly understood. For example,despite extensive knowledge of the molecular biology of viralreplication, little is known about how viruses produce cellularinjury (1). Similarly, although it is well established that theprimary lesion of diphtheria intoxication is the inhibition ofcellular protein synthesis (2, 3), almost nothing is known aboutthe complex series of events arising from the initial lesion at themolecular level which eventually leads to metabolic derange-ments, morphologic damage, and cell death (4). In an attemptto clarify the structural basis of the injury produced by infec-tious agents and chemicals, we have begun an investigation ofthe effects of these agents on the three major filamentous sys-

tems in the cell cytoplasm: microtubules, microfilaments, andintermediate filaments.

For our initial studies, we have chosen to examine the effectsof diphtheria toxin, Pseudomonas aeruginosa exotoxin A (ex-otoxin A), and cycloheximide on the cytoskeleton because themolecular mechanisms of their actions are well characterized.Both toxins inhibit cellular protein synthesis by catalyzing theNAD-dependent, ADP-ribosylation of elongation factor 2 (2,3, 5). Despite this detailed biochemical knowledge, it is notknown whether these toxins affect the cytoskeleton during theprocess of cell injury. We report here that diphtheria toxin andP. aeruginosa exotoxin A specifically influence the organizationof the vimentin class of intermediate filaments but have no

discernable effect on microtubules or microfilaments. Fur-thermore, the cycloheximide-induced disruption of vimentinfilaments is readily reversible.

MATERIALS AND METHODSCells and Media. CV-1 (African green monkey kidney) cells

were grown as monolayers in roller bottles in Richter's im-proved minimal essential medium containing zinc, insulin, and10mM Hepes (Irvine Scientific, Santa Ana, CA) supplementedwith 10% (vol/vol) fetal calf serum (Sterile Systems, Logan, UT)at 37°C.

The publication costs of this article were defrayed in part by page

charge payment. This article must therefore be hereby marked "ad-vertisement" in accordance with 18 U. S. C. §1734 solely to indicatethis fact.

7267

Toxins and Antitoxins. Corynebacterium diphtheriae strainsC7(,tox+) and C7(13tox-197) and P. aeruginosa PA103 havebeen described (6, 7). C. diphtheriae strains were grown in C-Ymedium under optimal conditions for tox gene expression (8).Diphtheria tox gene products were purified from culture su-pernatant fluids as described (9). P. aeruginosa PA103 wasobtained from P.-C. Tai (Bacterial Physiology Unit, HarvardMedical School), and exotoxin A was partially purified fromculture supernatant fluids as described (10).

Anti-diphtheria toxin and anti-exotoxin A were obtainedfrom the Massachusetts Antitoxin and Vaccine Laboratory(Jamaica Plain, MA) and Barbara Iglewski (Department ofMicrobiology and Immunology, University of Oregon Schoolof Medicine, Portland, OR), respectively.

Antisera to Cytoskeletal Filament Proteins. Rabbit anti-serum prepared against tubulin (11), provided by Frank Solo-mon (Massachusetts Institute of Technology, Cambridge, MA),was used for the detection of microtubules. Antiserum to gerbilvimentin was prepared according to the procedure of Hynesand Destree (12). Rabbit antiserum against actin (13), providedby Keith Burridge (Cold Spring Harbor Laboratory, ColdSpring Harbor, NY), was used to detect microfilaments. Fluo-rescein-conjugated goat anti-rabbit IgG was obtained fromMeloy Laboratories (Springfield, VA).

Indirect Immunofluorescent Microscopy. CV-1 cells,grown at low density on 12-mm round glass coverslips (Roch-ester Scientific) in 60 X 15 mm culture dishes (2 X 105 cells perdish), were incubated with various concentrations of diphtheriatoxin, CRM197, exotoxin A, or cycloheximide for the indicatedtimes. Prior to staining of cells with antibody to tubulin orvimentin, the cells attached to coverslips were fixed by theprocedure of Osborn and Weber (14) and transferred to a hu-midified chamber. Ten microliters of an appropriate dilutionof antiserum was applied to each coverslip. The coverslips wereincubated at 37°C for 1 hr, rinsed in phosphate-buffered saline,and returned to the humidified chamber. Ten microliters ofan appropriate dilution of fluorescein-conjugated goat anti-rabbit IgG was applied to each coverslip. The coverslips wereincubated at 37°C for 40 min, rinsed in phosphate-bufferedsaline and water, and mounted on glass slides in Gelvatol(Monsanto, St. Louis, MO). Prior to being stained with antibodyto actin, the cells were fixed for 10 min in 3.5% (vol/vol)formaldehyde in phosphate-buffered saline, rinsed in phos-phate-buffered saline and water, and incubated for 1 min inacetone that had been chilled to -200C. After being rinsed inphosphate-buffered saline, the fixed cells were incubated withantisera to actin for 1 hr at 370C. The cells were rinsed withphosphate-buffered saline and then incubated with fluores-cein-conjugated goat anti-rabbit IgG as described above. Allcells were examined with a Zeiss photomicroscope III equippedwith epifluorescence and a Planapochromat objective lens (40Xand 63X). Photomicrographs were made with Kodak Tri-X(ASA 400) at Exposure Index 1600.

Proc. Nati. Acad. Sci. USA 77 (1980)

b.-.~~ .:.~,w.iD

FIG. 1. (A-C) Organization of microtubules (A), microfilaments (B), and intermediate filaments (C) in control CV-1 cells. (D-F) Organizationof microtubules (D), microfilaments (E), and intermediate filaments (F) in CV-1 cells treated with 10-11 M diphtheria toxin for 22 hr. Cellswere subjected to indirect immunofluorescence microscopy with antibody against tubulin, actin, and vimentin. Note that only the intermediatefilament system is affected by diphtheria toxin. (A, B, D, and E, bar = 10 ,um; C, bar = 40 ,m; F, bar = 30.um.)

Determination of Rate of Cellular Protein Synthesis. CV-1cells were transferred from roller bottles to 35-mm plasticFalcon dishes (2 X 105 cells per dish). To measure the effectsof diphtheria toxin, CRM197, exotoxin A, or cycloheximide onprotein synthesis, we added various concentrations of theseagents to CV-1 cells. At the times indicated, the growth medium

was removed from the CV-1 cultures and replaced with Eagle'sminimal essential medium without methionine (GIBCO)containing 1 ,uCi of [(aSImethionine per ml (Amersham/Searle;>400 Ci/mmol, 1 Ci = 3.7 X 1010 becquerels). The cultureswere incubated for an additional hour at 370C. Incorporationof radioactive methionine was stopped by the removal of the

7268 Cell Biology: Sharpe et al.

Proc. Natl. Acad. Sci. USA 77 (1980) 7269

20-

10

5

0 4 8 12 24Time, hr

FIG. 2. Effect of diphtheria toxin on the rate of CV-1 cell proteinsynthesis. Various concentrations of diphtheria toxin were added toCV-1 cells. At 2, 5, 8, 11, and 24 hr after addition of toxin, the growthmedium was removed and replaced with Eagle's minimal essentialmedium without methionine containing 1 ACi of [35S]niethionine per

ml. The cultures were incubated for 1 hr at 370C, and incorporationof radioactive methionine was determined. 0, Control; *, 10-10 Mtoxin; *, 10-11 M toxin.

radioactive medium and the addition of phosphate-bufferedsaline containing 0.2% EDTA. The cells were removed fromthe dishes by agitation with a pasteur pipette, transferred todisposable glass tubes, and sonically treated. The cell suspensionswere adjusted to a final concentration of 5% (wt/vol) trichlo-roacetic acid, incubated at 40C for 1 hr, and filtered ontoWhatman GF/c filter paper. The filters were washed withethanol, dried, and placed in scintillation vials. A Triton/toluenescintillation fluid (Yorktown Research, Hackensack, NJ) was

added to each vial, and radioactivity was measured in a Beck-man liquid scintillation counter.

RESULTSCytoskeleton of CV-1 Cells. After examining several types

of cells, we found CV-1 cells to be ideal for these studies becauseof their high sensitivity to diphtheria toxin and ability to growas well-spread and flattened cells in monolayers. Anti-tubulinstaining revealed an extensive network of microtubules ema-

nating from the nucleus (Fig. 1A). Antibodies to actin showedtypical bundles of microfilaments in the cytoplasm (Fig. 1B).Staining of intermediate filaments with vimentin antibodyrevealed a cytoplasmic filamentous system emanating or ter-minating at one or more focal points surrounding the nucleus(Fig. 1C).

Effect of Diphtheria Toxin on CV-1 Cytoskeleton. CV-1cells are exquisitely sensitive to diphtheria toxin (15, 16). Tomeasure the rate of inactivation of cellular protein synthesis bydiphtheria toxin, we incubated CV-1 cells with [35S]methioninefor 1 hr at intervals after the addition of toxin. After the addition

fdtoxin there was a characteristic lag period followed by aninhibition of protein synthesis (Fig. 2). As the concentration oftoxin was increased from 10-12 M to 10-10 M, the lag perioddecreased and the rate of inhibition of protein synthesis in-creased. In these studies the effects of 10-11 M diphtheria toxinon the cytoskeleton were examined after a 22-hr exposure totoxin. Diphtheria toxin had little or no effect on microtubules(Fig. 1D) or microfilaments (Fig. 1E). On the other hand, theorganization of intermediate filaments was disrupted. Wavyfilaments with no apparent organization were seen in thecytoplasm (Fig. 1F). This disruption of intermediate filamentswas a specific effect of diphtheria toxin; antitoxin neutralizedthe effect (Fig. 3A).

Effect of CRM197 on CV-1 Cytoskeleton. In order to de-termine whether the enzymatic activity of diphtheria toxin wasessential for the disruption of vimentin filaments, a mutantdiphtheria toxin, CRM197, was examined for its effects on thecytoskeleton. CRM197 has a molecular weight identical to thatof diphtheria toxin but is enzymatically inactive (6). Ittlesonand Gill (17) have demonstrated that CRM197 is a competitiveinhibitor of diphtheria toxin with a Ki of approximately 10-8M. It is presumed that, after binding to a eukaryotic cell re-ceptor, CRM197 fragment A is transported normally into thecytosol. CRM197 added to cells at a concentration of 10-11 Mdid not disrupt vimentin filaments (Fig. 3B). In addition,CRM197 had no effect on microtubules and microfilaments(data not shown). Thus, the enzymatic activity of diphtheriatoxin appears to be necessary for disruption of vimentin fila-ments to occur.

Effect of P. aeruginosa Exotoxin A on CV-1 Cytoskeleton.P. aeruginosa exotoxin A inhibits protein synthesis in a manneridentical to that of diphtheria toxin (5). To determine whetherexotoxin A would also disrupt vimentin filaments, we addedexotoxin A to CV-1 cells at a concentration of approximatelyi0-11 M. At this concentration, exotoxin A inhibited proteinsynthesis but had no effect on microtubules or microfilaments(data not shown). Vimentin filaments were disrupted in amanner identical to that of diphtheria toxin (Fig. 3C); theperinuclear organization sites disappeared and wavy fibers werevisible in the cytoplasm. No effect on any filament system wasobserved when antibody to exotoxin A was preincubated withtoxin prior to the addition of toxin to CV-1 cells.

Effects of Cycloheximide on CV-1 Cytoskeleton. Bothdiphtheria toxin and exotoxin A inhibit protein synthesis by theNAD-dependent ADP-ribosylation of elongation factor 2. Todetermine whether an antibiotic that inhibits eukaryotic proteinsynthesis would also disrupt vimentin filaments, we studied theeffect of cycloheximide on the CV-1 cytoskeleton. Cyclohexi-mide (10 ,g/ml) inhibited protein synthesis by 80% within 5min of its addition to cells. The intermediate filament systembegan to appear altered by 75 min. The filaments that had ra-diated from the perinuclear organization sites were no longerpresent at 3 hr, and only wavy fibers that differed in organi-zation from vimentin filaments were visible (Fig. 3D). Mi-crotubules or microfilaments were not affected (data notshown).The inhibition of protein synthesis induced by cycloheximide

is reversible (18). Protein synthesis resumed within minuteswhen growth medium containing cycloheximide was removedfrom CV-1 cells and replaced with fresh medium. Vimentinfibers began to reappear (Fig. 3 E and F) by 1 hr after re-placement of medium. Multiple focal sites were present by 2hr (Fig. 3G). The intermediate filament system resumed itsnormal appearance (Fig. 3H) by 4 hr after the removal of cy-cloheximide.

Cell Biology: Sharpe et al.

Proc. Natl. Acad. Sci. USA 77 (1980)

~- .C

FIG. 3. Effects of 10-11 M diphtheria toxin neutralized by antitoxin (A), CRM197 (B), 10-11 M P. aeruginosa exotoxin A (C), and 10 ,ugof cycloheximide per ml (D) on the vimentin filament system. The effect of cycloheximide was reversed when growth medium containing cyclo-heximide was removed and replaced with fresh medium. The appearance of vimentin filaments at 0.5 hr (E), 1 hr (F), " hr (G), and 4 hr (H)after the removal of cycloheximide is shown. (A, F, G, and H, bar = 40 Mm; B, C, D, and E, bar = 25 Am.)

DISCUSSIONThe addition of diphtheria toxin, P. aeruginosa exotoxin A, or

cycloheximide to CV-1 cells leads to a disruption of the orga-nization of cytoplasmic vimentin filaments. This effect occursin every treated CV-1 cell and is specific; the alteration doesnot occur when diphtheria toxin and exotoxin A are incubatedwith their respective antitoxins prior to addition to cells. Thedisruption is not simply related to the binding of toxin to thecell surface and the translocation of fragment A into the cytosol

because CRM197, a missense mutant of diphtheria toxin thatbinds normally to diphtheria toxin receptors but is devoid ofenzymatic activity, does not disrupt vimentin filaments.

Diphtheria toxin, exotoxin A, and cycloheximide disrupt theorganization of vimentin filaments but not that of microtubuleor microfilament bundles within a few hours after inhibitingcellular protein synthesis. Thus, intermediate filaments appearto be more sensitive than microtubules or microfilament bun-dles to the cellular injury induced by certain bacterial exotoxins.

7270 Cell Biology: Sharpe et al.

Proc. Natl. Acad. Sci. USA 77 (1980) 7271

The disruption of intermediate filaments after the inhibitionof protein synthesis may thus play a significant role in thepathogenesis of exotoxin-mediated cellular injury.The cell cytoplasm is a highly organized three-dimensional

lattice of chemically distinct filaments in which cytoplasmicinclusions and organelles are positioned. Intermediate filamentsare thought to function in the spatial organization of the con-tents of the cell cytoplasm (19). For example, intermediatefilaments may play a role in the anchoring of the polysomes tothe cytoskeleton (20, 21) as well as in influencing the distribu-tion and orientation of mitochondria (unpublished results). Theresults reported here indicate that vimentin filaments respondvery quickly to alterations in cellular metabolism.Why might inhibition of protein synthesis lead to the dis-

ruption of vimentin filaments? Rapid turnover of vimentinprotein would seem to be an unlikely explanation becausevimentin protein does not disappear, but rather its organizationis altered. We have used Penman's extraction procedure (21)to isolate the cytoskeleton, to resolve the components in atwo-dimensional gel with isoelectric focusing and Na-DodSO4/polyacrylamide gel electrophoresis, and to quantitatevimentin. When untreated CV-1 cells or cells treated withdiphtheria toxin or cycloheximide were compared, the amount,isoelectric point, and molecular weight of vimentin were verysimilar. Although there was no significant reduction in the othermajor cytoskeletal components (such as actin, tropomyosin, anda-actinin), we did observe the disappearance, or decrease, ofseveral polypeptides normally associated with cytoskeleton asestablished by Penman's procedure in both the diphtheria toxin-and cycloheximide-treated cells. It is possible that rapidlyturning over regulatory proteins may exist to assemble vimentinfilaments within their proper orientation. Inhibition of proteinsynthesis could lead to the depletion of such "assembly" pro-teins, resulting in a disorganized vimentin filament system.

Other investigators have reported that cycloheximide doesnot alter the cytoskeleton (22). We have examined several typesof cells for their response to the addition of this antibiotic.Disruption of vimentin filaments was observed in kangaroo ratepithelial cells (PtK2 cells) but not in gerbil fibroma cells (CCL146) or in BALB/c 3T3 cells. Possibly, cells that have a morelabile "assembly" protein are sensitive to this effect.We have also found that reovirus infection of CV-1 cells leads

to the destruction of vimentin filaments although other changeswere observed as well. We do not know if the disruption ofvimentin filaments is related to a virally mediated inhibitionof protein synthesis in CV-1 cells. Thus, viruses as well as toxinsmay exert their pathologic effects, at least in part, not onlythrough interference with intermediary metabolism and thebiosynthesis of macromolecules but also through disuption ofvimentin filaments.

This study demonstrates a means to specifically disrupt theorganization of vimentin filaments but not that of microtubulesor microfilaments. Other agents that alter the cytoskeleton

affect more than one filament system. Colchicine, for example,causes the breakdown of microtubules and induces the for-mation of juxtanuclear aggregates of vimentin-containingfilaments (2t3). Because diphtheria toxin, exotoxin A, and cy-cloheximide disrupt the organization of vimentin filamentsonly, these agents may help to elucidate the precise biologicfunctions of vimentin filaments.

We thank Karen Byers for excellent technical assistance. A.H.S. isa National Institutes of Health Predoctoral Trainee (5t32 GMO 7196).J.R.M. is a recipient of Cancer Development Award K04 AI-00146from the National Institute of Allergy and Infectious Diseases. Thisinvestigation was supported by Grants Al 13178 from the NationalInstitutes of Health (to B.N.F.), P.O. 1 CA 22427 from the NationalCancer Institute (to L.B.C.), and AI-12500 from the National Instituteof Allergy and Infectious Diseases (to J.R.M.).

1. Tamm, I. (1975) Am. J. Pathol. 81, 163-178.2. Honjo, T., Nishizuka, Y., Hayaishi, 0. & Kato, I. (1968) J. Biol.

Chem. 243,3533-3535.3. Gill, D. M., Pappenheimer, A. M., Jr., Brown, R. & Kurnick, J.

T. (1969) J. Exp. Med. 129, 1-21.4. Pappenheimer, A. M., Jr. (1977) Annu. Rev. Biochem. 46, 69-

94.5. Iglewski, B. H. & Kabat, D. (1977) Infect. Immun. 5, 138-

144.6. Uchida, T., Pappenheimer, A. M., Jr. & Greany, R. (1973) J. Biol.

Chem. 248, 3838-3844.7. Liu, P. V. (1966) J. Infect. Dis. 116,481-489.8. Bacha, P. & Murphy, J. R. (1978) J. Bacteriol. 136, 1135-

1142.9. Pappenheimer, A. M., Jr., Uchida, T. & Harper, A. H. (1972)

Immunochemistry 9,891-906.10. Liu, P. V. (1973) J. Infect. Dis. 128, 506-513.11. Solomon, F., Magendantz, M. & Salzman, A. (1979) Cell 18,

431-438.12. Hynes, R. 0. & Destree, A. T. (1978) Cell 13, 151-163.13. Burridge, K. (1976) Proc. Natl. Acad. Sci. USA 73, 4457-

4461.14. Osborn, M. & Weber, K. (1977) Cell 12,561-571.15. Lennox, E. S. & Kaplan, A. S. (1957) Proc. Soc. Exp. Biol. Med.

95,700-702.16. Middlebrook, J. L., Dorland, R. B. & Leppla, S. H. (1978) J. Biol.

Chem. 253, 7325-7330.17. Ittleson, T. R. & Gill, D. M. (1973) Nature (London) 242,

330-332.18. Colombo, B., Felicetti, L. & Baglioni, C. (1966) Biochim. Biophys.

Acta 119, 109-119.19. Lazarides, E. (1980) Nature (London) 283,249-256.20. Wolesewick, J. J. & Porter, K. R. (1975) Anat. Rec. 181, 511-

512.21. Fulton, A. B., Wan, K. M. & Penman, S. (1980) Cell 20, 849-

857.22. Wang, E., Cross, R. K. & Choppin, P. W. (1979) J. Cell Biol. 83,

320-337.23. Starger, J. M., Brown, W. E., Goldman, A. E. & Goldman, R. D.

(1978) J. Cell Biol. 78, 93-109.

Cell Biology: Sharpe et al.