Proc. Nati. Acad. Sci. USAVol. 91, pp. 8329-8333, August 1994Cell Biology

Mos oncogene product associates with kinetochores in mammaliansomatic cells and disrupts mitotic progressionXIAO MIN WANG*, NELSON YEWt, JOHN G. PELOQUIN*, GEORGE F. VANDE WOUDEt,AND GARY G. BOiuSY*I*Laboratory of Molecular Biology, University of Wisconsin, Madison, WI 53706; and tAdvanced BioSciences Laboratories-Basic Research Program,National Cancer Institute-Frederick Cancer Research and Development Center, Frederick, MD 21702

Contributed by George F. Vande Woude, May 27, 1994

ABSTRACT The mos protooncogene has opposing dfectson cell cycle pgression. It is required for reitatlon ofmeiotic maturation and for meiotic progression through meta-phase II, yet it is an active component of cytostatic factor. mosis a potent oncogene in fibroblasts, but high levels ofexpionare lethal. The lethality of mnos gene expression in ma liancells could be a consequence of a blockage induced by itscytostatic factor-related activity, which may appear at highdosage in mitotic cells. We have directly tested whether ex-pression of the Mos protein can block mitosis in mammaliancells by mnicroecting a fusion protein between Esckerichiacoli maltose-binding protein and Xenopus c-Mos Into PtK1epithelial cells and analyzing the cells by video time-lapse andimmunofluorescence microscopy. Time-course analysesshowed that Mos blocked mitosis by preventing progression toa normal metaphase. Chromosomes frequently falled to attaina bipolar orientation and were found near one pole. InJectionof a kinase-deficlent mutant Mos had no effect on mitosis,indicating that the blockage of mitotic progression requiredMos kinase activity. Antitubulin immunostaining of cellsblocked by Mos showed that microtubules were present butthat spindle morphology was abnormal. Immunotalning forthe Mos fusion protein showed that both wild-type and kinasemutant proteins- localized at the kinetochores. Our resultssuggest that mitotic blockage by Mos may result from an actionof the Mos kinase on the kinetochores, thus increasing chro-mosome instability and preventing normal congression.

Early transformation studies of fibroblasts with v-mosshowed that a majority of cells acutely infected with Moloneymurine sarcoma virus rounded up, detached from the mono-layer, and failed to proliferate (1-3). This cytotoxic effect ofthe v-mos product has been proposed to be related to thecytostatic factor (CSF) activity of Mos in unfertilized eggs(4). CSF is an activity believed to be responsible for thenatural arrest of oocytes in metaphase II of meiosis and wasoriginally discovered as an activity in unfertilized eggs thatinhibited mitosis in early cleavage embryos (5).The identified cellular activities of Mos depend upon its

function as a serine/threonine protein kinase. The ability ofMos to reinitiate oocyte maturation (4, 6, 7), provide CSFactivity (8), and transform fibroblasts (3, 9) is abolished in aprotein kinase-deficient mutant of Mos with an anino acidsubstitution (Lys9- Arg) in the canonical ATP binding site (7,10).

Several candidate substrates of Mos have been suggestedfrom in vitro experiments, including cyclin B2 (11) and tubulin(12). Most recently, Mos has been shown to stimulate mito-gen-activated protein kinase (MAP kinase), entering theMAP kinase phosphorylation cascade apparently by directactivation of MAP kinase kinase (13).

Long before the mos protooncogene product was found tobe a regulator of meiosis, the ability of v-mos and its cellularhomolog, c-mos, to transform fibroblast cells indicated thatMos substrates were present in somatic cells. The cell deathcaused by high levels of expression in mos-transformedfibroblasts could have been a consequence of mitotic block-age, although no evidence was provided in the early studiesthat the acutely infected cells were in fact blocked in mitosis.To directly evaluate whether Mos could cause mitotic block-age in mammalian cells, we microinjected a bacterially ex-pressed maltose-binding protein (MBP)-Xenopus Mos(Mosxe) fusion protein or its kinase-deficient mutant (MBP-Mosxekm) into cells at defined stages of mitosis. The potorooepithelial cell line PtK1 was chosen as the recipient cellsbecause they generally remain flat during mitosis, allowing aclear visualization of chromosome and spindle mechanics.

MATERIALS AND METHODSMicroinjection. Both MBP-Mosxe (wild type) and MBP-

Mosxekm (kinase-deficient Mos, Lys9O - Arg) fusion proteinswere expressed and purified as described (7). PtK1 ratkangaroo epithelial cells were plated onto photoetched cov-erslips (Bellco Glass) and grown for at least 2 days in Ham'sF-10 medium (GIBCO) with 10%6 fetal bovine serum. Mi-cropipettes were pulled on a vertical pipette puller (KopfInstruments, Tujunga, CA) to a tip diameter of 0.7-0.9 pm.Microinjection was carried out with a Narishige IM-200microinjector (Greenvale, NY). The temperature was main-tained at 37°C with a stage and objective heater during theperiod of monitoring cells. To prevent pH change, the cellmedium was overlaid with mineral oil. After microinjectionwith Mos protein, the cells were viewed with a x40 phase-contrast objective, n.a. 1.0, and recorded onto an optical discrecorder (Panasonic, Secaucus, NJ) for up to 3 hr postinjec-tion. For analysis, time-lapse discs were played at 300 timesthe recorded speed.

Nocodazole Treatment of Injected Cells. In some experi-ments, cells were treated with nocodazole to prolong themitotic interval and allow more time for the injected Mos toact on cellular substrates. After microinjection of Mos, themedium was aspirated and fresh F-10 medium, prewarmed to37°C and containing nocodazole at 1 ug/ml was added. Aftera 45-min incubation, the nocodazole-containing medium wasaspirated and the cells were rinsed with three changes offreshF-10 medium without nocodazole.Immu rnce. Cells were directly fixed and permeabi-

lized in methanol for 5 min at -20°C. The coverslips were thenprocessed at 37°C for immunofluorescence by the followingsteps, each of which was separated by four 10-min rinses with

Abbreviations: MBP, maltose-binding protein; Mosxe, XenopusMos; Mosxekm, kinase-deficient mutant of Mosxe; CSF, cytostaticfactor; MAP kinase, mitogen-activated protein kinase.tTo whom reprint requests should be addressed.

8329

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

Aug

ust 2

3, 2

021

Proc. Natl. Acad. Sci. USA 91 (1994)

phosphate-buffered saline (PBS): (i) 10%o normal goat serum inPBS, 45 min; (ii) rabbit anti-MBP antibody (New EnglandBiolabs) at 1:650 in 1% goat serum/PBS and/or rat YL-1/2anti-tubulin monoclonal antibody (Accurate Chemicals) at 1:500in 1% goat serum/PBS, 60 m=n; (iii) fluorescein-conjugated goatanti-rabbit IgG at 1:100 in 1% goat serum/PBS and/or TexasRed-conjugated goat anti-rat IgG at 1:100 in 1% goat serum/PBS (Jackson ImmunoResearch), 45 min. Images were re-corded using a x 100, n.a. 1.3, objective on an inverted micro-scope (Zeiss) equipped with a cooled charge-coupled-devicecamera (Photometrics, Tucson, AZ).

RESULTSMos Protein Specificafly Blocks Ceils in Mitosis. PtK1 cells

in mid to late prophase were initially selected for microin-jection. Cells at this stage normally progress to nuclearenvelope breakdown within 30 min, reach metaphase by 40min, onset of anaphase within 60 min, and begin cleavage byabout 75 min (14). Cells microinjected into the cytoplasmwith MBP-Mosxe progressed through nuclear envelopebreakdown normally but generally failed to form a metaphaseplate (Fig. 1 Left). The PtK1 cells became atypically roundedand remained blocked in a mitotic state with condensedchromosomes. All cells were monitored for at least 3 hr andsome were monitored for up to 6 hr.

Cells were scored for mitotic inhibition according to theseverity of the phenotype. A strong (terminal) phenotype wasdefined as failure of all chromosomes to form a stablemetaphase plate, extreme cell rounding, and failure of thecells to divide. A weak (unstable) phenotype was defined asfailure of some chromosomes to arrive at the metaphaseplate, some rounding, abnormal anaphase (failure of sisterchromosome segregation or nondisjunction), followed byeither failure of cleavage or uneven cleavage that alteredploidy. By these criteria, microinjection of Mos at a needleconcentration of 0.5 mg/ml inhibited mitosis and produced aterminal phenotype in 70%6 of the cells (Fig. 1 Left; Fig. 2),although 20%6 of the cells passed mitosis with some delay.The specificity of action of Mos on mitotic cells was

evaluated in several ways. First, the possibility of nonspecificinjury to cells due to microinjection was assessed by injectionof buffer alone. Cells were injected at various stages frommid-prophase to metaphase. All 34 cells injected dividednormally (Fig. 2), indicating no detectable effect of microin-jection per se. Second, the action of Mos was dependent onconcentration, with a lower concentration giving a lowerproportion of the strong phenotype. When cells were injectedwith Mos at 0.25-mg/ml needle concentration, althoughmitotic blockage was still produced, a higher proportion ofcells reached metaphase, and many cells displayed the weakphenotype (data not shown). Analysis of the time-lapse videorecords indicated that chromosomes frequently failed todisplay normal mitotic behavior. Either they failed to con-gress to the metaphase plate or they congressed and thensubsequently became irregularly arranged. Resultinganaphases were abnormal in that sister chromosomes failedto segregate.Although some rounding is normal for mitosis, Mos-

injected mitotic cells became extremely rounded. To evaluatewhether this effect was a nonspecific cytotoxic reaction onthe PtK1 cells, interphase cells were also microinjected andobserved for 3 hr. Microinjected interphase cells remainedwell spread and normal in appearance (data not shown).Therefore, the rounding effect was specific to mitotic cells,and we conclude that Mos is not generally cytotoxic to PtK1epithelial cells.

Next, we determined whether the cellular effect of Moswas dependent on its kinase activity. The kinase-deficientmutant MBP-Mosxekm does not induce oocyte maturation or

MBP- Mosx- MBP---Mos"eL.. ... Xtie......

I

FIG. 1. Kinase dependence of mitotic blockage by Mos. Mitoticprogression was evaluated for PtK1 cells microinjected with eitherMBP-Mosxe or MBp-Mosxekm. (Left) The cell was injected inprophase with MBPMosxe at 0.5 mg/ml; the first image shown is 5min postinjection, just after nuclear envelope breakdown. Up to 1 hrafter nuclear envelope breakdown, chromosomes failed to organizea metaphase plate and remained in a pseudoprometaphase config-uration (last image). Three hours postinjection, the cell had still notdivided and appeared very rounded (not shown). (Right) The cell wasinjected with MBP-Mosxekm in prophase 5 min before recording thefirst image shown and monitored as Left. The first image shown isshortly after nuclear envelope breakdown. The cell progressedthrough prometaphase (second image), to metaphase (third image)within 30 min of nuclear envelope breakdown, and chromosomesegregation occurred in a normal anaphase (last image). PM,prometaphase; M, metaphase; P-PM, pseudoprometaphase; A,anaphase. (Bar = 10 pm.)

possess CSF or transformation activity (7, 10). Injection ofcells with MBp-MoSxekm at the same concentration as wild-type Mos failed to inhibit mitosis (Fig. 1 Right; Fig. 2). Cellsprogressed to metaphase, initiated anaphase, and cleaved, allwith normal kinetics. This result indicates that the blockageof mitotic progression caused by Mos is specifically depen-dent on its kinase activity.

8330 Cell Biology: Wang et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

3, 2

021

Proc. Natl. Acad. Sci. USA 91 (1994) 8331

1oo

9o

80

- 70

0 60

E 50

E 40soc. 30

20

10 T~~~~~~oWild type Mutant

ES weak phenotypestrong phenotype

ce

Buffer

FIG. 2. Quantitation of mitosis blockage by Mos. Bar chart showsthe response of PtK1 cells injected at prophase with MBP-Mosxe(wild type) or MBP-Mosxekm (mutant) at 0.5 mg/ml (needle concen-tration) or with buffer. Cells were observed and scored 3 hr afterinjection. Strong phenotype (filled bar), all chromosomes failed toalign on the metaphase plate and cells did not divide; weak phenotype(hatched bar), most but not all chromosomes arrived at the meta-phase plate, followed by delayed and abnormal anaphase.

Stage Dependence of Mos Effect on Mitosis. The effect onmitosis of microinjection of MBP-Mosxe was more dramaticwhen cells were injected earlier in mitosis. As stated before,injection of Mos at 0.5 mg/ml into prophase cells preventednormal chromosome congression. In contrast, all cells in-jected at metaphase progressed through mitosis, and cellsinjected at prometaphase gave intermediate results, with 36%of the cells showing the strong phenotype (Fig. 3).

It has been reported by Yew et al. (7) that bacteriallyexpressed MBP MoSxe fusion protein efficiently induces ger-minal vesicle breakdown and the activation of maturation-promoting factor in the absence of protein synthesis, althougha posttranslational modification is apparently required for ac-tivation of its kinase activity. Since activation of Mos may beslow compared to mitotic progression, we extended the time ofinteraction of Mos with putative cellular activators by blockingcells in mitosis with the microtubule-depolymerizing drug no-codazole. Nocodazole blocks cells by depolymerizing micro-tubules, resulting in a pseudoprometaphase state. Cells wereinjected with Mos at metaphase or prometaphase, treated withnocodazole at 1 Mg/ml for 45 min, and then washed and placed

100*

90 -

80'

° 70-

co 60'

E 50,

E 40,.So.0 30'

20-

10'

0 weak phenotypestrong phenotype

Prophase Prometaphase MetaphaseStage at time of microinjection

FIG. 3. Stage dependence of Mos effect on mitosis. Cells wereinjected at different stages of mitosis with MBP-Mosxe at 0.5 mg/ml.Cells were observed and scored 3 hr after injection. Phenotypes areas explained in Fig. 2.

in fresh medium to release them from the nocodazole block.With this treatment by nocodazole, all the microtubules aredepolymerized, including stable kinetochore microtubules (15).After release from nocodazole, the metaphase injected cellswere delayed somewhat but reached anaphase in 20-40 min anddivided normally. In contrast, cells injected in prometaphasedid not progress to metaphase, became abnormally rounded,and failed to divide. Cells injected with either buffer or thekinase-deficient mutant in prometaphase and blocked withnocodazole divided normally after releasing. Thus, the failure ofMos to block mitosis after injection at metaphase is not due toinsufficient time for Mos activation. This result suggests that forMos to exert its inhibitory effect in PtK1 cells, it must act earlyin mitosis, implying that either its substrate is inaccessible (16)or its putative activator is rendered inactive by metaphase.Mose Fusion Protein Targets the Mitotic Apparatus, Espe-

cially the Kinetochores. To identify the possible cellular target(s)ofMos, we performed immunofluorescence analysis to localizeMos after microinjection. Cells were microinjected in prophasewith either MBP-Mosxe or MBP-MosxeIm, fixed 25 min later,and double-labeled by immunofluorescence to determine thedistribution ofMos and microtubules. Cells injected in prophasewith wild-type Mos at 0.5 mg/ml generally formed spindleslacking normal bipolar morphology, and chromosomes wereirregularly arranged. However, with Mos at 0.3 mg/ml, cellsprogressed to a nearly metaphase configuration, with only afewchromosomes remaining off the plate (Fig. 4 A-C). Theseintermediate phenotypes were instructive as to the nature oftheeffect produced by Mos. Anti-tubulin staining showed thatspindle microtubules were present but that some kinetochorefibers pointed away from the metaphase plate. Mos was local-ized by staining with an antibody to the MBP moiety. Theimmunofluorescence localization showed paired dots and polestaining. Superposition of the immunofluorescent image ontothe phase image showed that the paired dots corresponded tothe spindle attachment sites of chromosomes (Fig. 4 D-G),which indicates Mos localization on kinetochores as well asspindle poles. However, many of the chromosomes, as indi-cated by the positions of the paired dots, failed to congress tothe equator. Detailed comparison with the anti-tubulin imagessuggests that the nonequatorial chromosomes were either non-attached or unipolar in attachment. Chromosomes attached toboth poles can be readily identified by a kinetochore-axis test.A line drawn through the sister kinetochores will be approxi-mately parallel to the spindle axis. As is evident in Fig. 4B,several of the chromosomes fail this test, the line connectingtheir kinetochore dots pointing away from the spindle axis.Thus, at low Mos concentrations, although a bipolar spindlewas formed, significant levels of malorientation were seen.Terminally arrested cells (high Mos concentration) failed toform a metaphase plate (see Fig. 1), which is an indicator ofextreme malorientation in which most or all of the chromo-somes failed to attach stably to both poles.

Injection of cells earlier in prophase allowed us to evaluateMos distribution before nuclear envelope breakdown. Muchof the Mos remained cytoplasmic, but some entered thenucleus and showed the paired dot staining characteristic ofkinetochores. Cells injected with kinase-deficient Mos alsoshowed nuclear entry and paired dot staining which weinterpret to be kinetochores (Fig. 4H and I). The kinetochorelocalization of the mutant Mos indicates that binding to thekinetochores does not require kinase activity. Since injectionof the mutant Mos failed to block mitotic progression, theresults suggest that kinetochore binding in the absence ofkinase activity is not sufficient to inhibit mitosis. To deter-mine whether the MBP portion of the fusion protein contrib-uted to the kinetochore localization of MBP-Mos, MBP wasinjected alone into prophase cells and analyzed as for thefusion protein. MBP alone failed to enter the nucleus atdetectable levels in prophase cells and failed to show any

Cell Biology: Wang et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

3, 2

021

Proc. Natl. Acad. Sci. USA 91 (1994)

FIG. 4. Immunolocalization ofinjected Mos fusion protein. (A-C)Cell was injected with MBP-Mosxe at 0.3 mg/ml in prophase,allowed to progress 25 min, fixed,and prepared for immunofluores-cence. The phase-contrast image(A) shows that the nuclear enve-lope has broken down but that notall chromosomes have- aligned atthe metaphase plate. Staining forMos (B) shows paired dots thatcorrespond to sister kinetochores,several of which point away fromthe metaphase plate. Tubulinstaining (C) shows that microtu-bules are intact. (D-G) Magnifiedimages showing paired dots (ar-rows inE and F) on achromosomeconnected to kinetochore fibers ofthe spindle (arrows in D and G):phase contrast (D), Mos staining(E), superposition of phase con-trast and Mos (F), and tubulinstaining (G). (H and I) Cell wasijected with MBP-1Mosxk at 0.5mg/ml in prophase and fixed be-fore nuclear envelope breakdown.Notice the Mos staining of paireddots in the nucleus (I). (J and K)Cell was injected with MBP at 1.7mg/ml in early prophase and fixedbefore nuclear envelope break-down. No nuclear and only diffusecytoplasmic fluorescence wasseen. [A, D, H, and J, phase con-trast; B, E, I, and K, rabbit poly-clonal antibody against MBP; Cand G, rat monoclonal anti-tubu-lin. Bars = 10 AM (A<; J and K)or 5 pm (D-G; H and I).]

paired dot staining characteristic ofkinetochore binding (Fig.4 J and K). We conclude that potential cellular targets forMos are components of the kinetochore and the spindle pole.

DISCUSSIONOur results show that Mos injection prior to metaphaseinhibits chromosome congression, that Mos is localized to thekinetochores and spindle poles in arrested cells, that Mos canenter the nucleus and bind to the kinetochore in prophase,and that the inhibitory activity of Mos depends upon its

kinase activity. Therefore, we suggest that the mitotic block-age produced by Mos results from the alteration of kineto-chore structure or activity early in mitosis, perhaps by Moskinase phosphorylation of some kinetochore component(s)involved in stable attachment of spindle microtubules. Alter-natively, the effect of Mos on kinetochore behavior may beindirect, resulting from interaction with a signaling pathwayinvolved in mitotic progression.The nonkinetochore microtubules ofthe mitotic spindle are

highly dynamic whereas the kinetochore microtubules aremore stable, presumably by virtue oftheir binding interaction

8-332 Ceff Biblogy: Wang et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

3, 2

021

Proc. Natl. Acad. Sci. USA 91 (1994) 8333

with the kinetochore (17-20). Stable, bipolar attachment ofchromosomes to poles is required for normal chromosomecongression (21). The range of phenotypes and mitotic block-age produced by Mos can be readily accounted for in termsof a weakened attachment of kinetochores to spindle micro-tubules.One important feature of prometaphase is that upon at-

tachment to the spindle, the chromosomes undergo conspic-uous oscillatory movements toward and away from the pole.The chromosome movements may be controlled by phos-phorylation of kinetochore proteins, including motor mole-cules (22). Both dynein and kinesin-like proteins have beenlocalized at kinetochores (23-26). The localization ofMos atthe kinetochore suggests that the Mos kinase could alter thenormal phosphorylation pattern of the kinetochore proteins.A phosphorylated epitope in mammalian tissue culture cellshas been found to be differentially expressed at the kineto-chores of chromosomes in mitotic cells (27). Duringprophase, label appeared at the kinetochores, and in earlyprometaphase, all the kinetochores were brightly labeled bythe antibody. With progression through prometaphase, het-erogeneity of labeling among the kinetochores became ap-parent. Chromosomes that were very near the poles werestrongly labeled on both kinetochores. Those that were at themetaphase plate tended to be weakly labeled on one or bothkinetochores. Chromosomes situated between the pole andthe metaphase plate often exhibited asymmetric labeling oftheir two kinetochores, with greater labeling on the kineto-chore toward the metaphase plate. Kinetochores in meta-phase and anaphase cells exhibited no labeling with theantibody. Phosphorylation of the epitope may identify aregulatory event that controls chromosome movement to themetaphase plate. Dephosphorylation of the kinetochore epi-tope is correlated with acquisition of an attachment to a poleand subsequent movement to that pole. The introduction ofMos kinase into mitotic cells may phosphorylate some com-ponent(s) normally dephosphorylated in prometaphasewhich, in turn, could interfere with chromosome attachmentand therefore mitotic progression.

Rattner et al. (28) showed that a portion of cellular p34cdc2was localized to several distinct domains within the mitoticapparatus, including the kinetochore, spindle pole, and ki-netochore microtubules. The detection of Cdc2 at specificsites within the mitotic apparatus may reflect the distributionof its substrate(s). Since both Cdc2 and Mos are localized atkinetochores and spindle poles and since Mos coimmuno-precipitates with p34cdc2 (29), it is possible that the effect ofMos on the kinetochore may be mediated through an inter-action or competition with cdc2. MAP kinase has beenlocalized at spindle poles in mouse oocytes (30), and Posadaet al. (13) found that Mos activated MAP kinase kinase.Therefore, the effect of Mos on mitosis may be mediatedthrough MAP kinase activity, as has been suggested inmeiosis (31). It remains an open question whether the Mosfusion protein is targeted to the same substrate(s) in mitosisas CSF/Mos is in meiosis II.According to Yew et al. (32), the level of Mos protein

expression in c-mosxe-transformed mouse cells is =0.2% ofthe total protein. We estimate the concentration of Mosfusion protein required to produce the strong phenotype inour experiments to be about 2-fold higher than Mos proteinlevels in transformed cells. This result is consistent with thequantitation by Yew et al. (32), because 2-fold dilution of theMos fusion protein injected into cells reduced the effect,producing primarily the weak phenotype, and 3-fold dilutedMos did not cause any abnormal mitotic phenotype.

If Mos is superimposing a meiotic process on the mitoticprogram of somatic cells, it is not surprising that significantconsequences for chromosome segregation will result. This,

in turn, will result in genetic instability, which is consideredto be the major component leading to tumor progression. Ourstudies demonstrate that the Mos oncogene product cannegatively influence chromosome stability and mitotic pro-gression. The conservation of the chromosome segregationmechanism suggests that the molecular identification of theMos target(s) will prove worthwhile.

We thank Drs. P. Kronebush and Y. Zhai for expert assistance inimage processing and Dr. W. Matten for helpful comments on themanuscript. This work was supported by National Institutes ofHealth Grant GM25062 (G.G.B.) and the National Cancer Instituteunder Contract N01-CO-74101 with Advanced BioSciences Labora-tories-Basic Research Program.

1. Fischinger, P. J. & Haapala, D. K. (1971) J. Gen. Virol. 13,203-214.

2. Papkoff, J., Verma, I. M. & Hunter, T. (1982) Cell 29, 417-426.3. Yew, N., Strobel, M. & Vande Woude, G. F. (1993) Curr.

Opin. Genet. Dev. 3, 19-25.4. Sagata, N., Daar, I., Oskarsson, M., Showalter, S. & Vande

Woude, G. F. (1989) Science 245, 643-646.5. Masui, Y. & Markert, C. L. (1971) J. Exp. Zool. 177,129-146.6. Sagata, N., Oskarsson, M., Copeland, T., Brumbaugh, J. &

Vande Woude, G. F. (1988) Nature (London) 335, 519-525.7. Yew, N., Mellini, M. L. & Vande Woude, G. F. (1992) Nature

(London) 335, 649-652.8. Sagata, N., Watanabe, N., Vande Woude, G. F. & Ikawa, Y.

(1989) Nature (London) 342, 505-511.9. Singh, B., Hannink, M., Donoghue, D. J. & Arlinghaus, R. B.

(1986) J. Virol. 60, 1148-1152.10. Freeman, R. S., Pickham, K. M., Kanki, J. P., Lee, B. A.,

Pena, S. V. & Donoghue, D. J. (1989) Proc. Nat!. Acad. Sci.USA 86, 5805-5809.

11. Roy, L. M., Singh, B., Gautier, J., Alinghaus, R. B., Nordeen,S. K. & Mailer, J. L. (1990) Cell 61, 825-831.

12. Zhou, R., Oskarsson, M., Paules, R. S., Schulz, N., Cleveland,D. & Vande Woude, G. F. (1991) Science 251, 671-675.

13. Posada, J., Yew, N., Ahn, N. G., Vande Woude, G. F. &Cooper, J. A. (1993) Mol. Biol. Cell 13, 2546-2553.

14. Vandenbunder, B. & Borisy, G. G. (1986) Cell Motil. Cytoskel-eton 6, 570-579.

15. Vandr6, D. D. & Borisy, G. G. (1989) J. Cell Sci. 94, 245-258.16. Simerly, C., Balczon, R., Brinkley, B. R. & Schatten, G. (1990)

J. Cell Biol. 111, 1491-1504.17. Salmon, E. D., Leslie, R. J., Saxton, W. M., Karow, M. L. &

McIntosh, J. R. (1984) J. Cell Biol. 99, 2164-2174.18. Mitchison, T. J., Evans, L., Schulze, E. & Kirschner, M. W.

(1986) Cell 45, 515-527.19. Gorbsky, G. J. & Borisy, G. G. (1989) J. Cell Biol. 109,

653-664.20. Mitchison, T. J. (1989) J. Cell Biol. 109, 637-652.21. Nicklas, R. B. (1971) Adv. Cell Biol. 2, 225-297.22. Hyman, A. A. & Mitchison, T. J. (1991) Nature (London) 351,

206-211.23. Pfarr, C. M., Coue, M., Grissom, P. M., Hays, T. S., Porter,

M. E. & McIntosh, J. R. (1990) Nature (London) 345, 263-265.24. Steuer, E. R., Wordeman, L., Schreer, T. A. & Sheetz, M. P.

(1990) Nature (London) 345, 266-268.25. Sawin, K. E., Mitchison, T. J. & Wordeman, L. G. (1992) J.

Cell Sci. 101, 303-313.26. Yen, T. J., Li, G., Schaar, B. T., Szilak, I. & Cleveland, D. W.

(1992) Nature (London) 359, 536-539.27. Gorbsky, G. J. & Ricketts, W. A. (1993) J. Cell Biol. 122,

1311-1321.28. Rattner, J. B., Lew, J. & Wang, J. H. (1990) Cell Motil.

Cytoskeleton 17, 227-235.29. Zhou, R., Daar, I., Ferris, D. K., White, G., Paules, R. S. &

Vande Woude, G. F. (1992) Mol. Cell. Biol. 12, 3583-3587.30. Verlhac, M. H., De Pennaart, H., Maro, B., Cobb, M. H. &

Clarke, H. (1993) Dev. Biol. 158, 330-340.31. Haccard, O., Sarcevic, B., Lewellyn, A., Hartley, R., Roy, L.,

Izumi, T., Erikson, E. & Maller, J. L. (1993) Science 262,1262-1265.

32. Yew, N., Oskarsson, M., Daar, I., Blair, D. G. & VandeWoude, G. F. (1991) Mol. Cell. Biol. 11, 604-610.

Cell Biology: Wang et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

3, 2

021