INTRODUCTION

Mitosis is naturally prone to error, but normally the errorsare flawlessly corrected. Early in mitosis, improper arrange-ments of kinetochore microtubules are common. For exam-ple, the microtubules associated with the two kinetochoresof a pair of chromosomes may extend to the same polerather than to opposite poles. Such a malorientation, if itpersisted, would lead to nondisjunction: chromosomeswould be distributed to the same daughter cell. Normally,however, malorientations are unstable and are soon cor-rected by reorientation. One kinetochore or the other gainsmicrotubules extending toward the opposite spindle poleand the old, misdirected microtubules are lost or redirected

(reviewed by Nicklas, 1988). By metaphase, malorient-ations are extremely rare. We have discovered that mal-orientations are extremely common in cells treated witheither of two very different protein kinase inhibitors, DMAP(6-dimethylaminopurine; Néant and Guerrier, 1988) andgenistein (Akiyama et al., 1987). These malorientations per-sist in both mitosis and meiosis and the result is wholesalenondisjunction. One of the processes required for reorient-ation might be affected by the drugs, such as the acquis-ition of new, properly directed kinetochore microtubules orthe loss or redirection of misdirected ones. In fact, neitherprocess is much affected by the drugs. The most significanteffect of DMAP is a novel kinetochore-kinetochore inter-action that it induces. Kinetochores move toward other

961Journal of Cell Science 104, 961-973 (1993)Printed in Great Britain © The Company of Biologists Limited 1993

Errors in chromosome orientation in mitosis and meio-sis are inevitable, but normally they are quickly cor-rected. We find that such errors usually are not cor-rected in cells treated with protein kinase inhibitors.Highly inaccurate chromosome distribution is the result.When grasshopper spermatocytes were treated with thekinase inhibitor 6-dimethylaminopurine (DMAP), 84%of maloriented chromosomes failed to reorient; inanaphase, both partner chromosomes were distributedto the same daughter cell. These chromosomes wereobserved for a total of over 60 h, and not a single reori-entation was seen. In contrast, in untreated cells, mal-oriented chromosomes invariably reoriented, andquickly: in 10 min, on average. A second protein kinaseinhibitor, genistein, had exactly the same effect asDMAP. DMAP affected PtK1 cells in mitosis as it didspermatocytes in meiosis: improper chromosome orien-tations persisted, leading to frequent errors in distribu-tion. We micromanipulated chromosomes in spermato-cytes treated with DMAP to learn why malorientedchromosomes often fail to reorient. Reorientationrequires the loss of improper microtubule attachmentsand the acquisition of new, properly directed kineto-chore microtubules. Micromanipulation experimentsdisclose that neither the loss of old nor the acquisitionof new microtubules is sufficiently affected by DMAP to

account for the indefinite persistence of malorientations.Drug treatment causes a novel form of chromosomemovement in which one kinetochore moves towardanother kinetochore. Two kinetochores in the samechromosome or in different chromosomes can partici-pate, producing varied, dance-like movements executedby one or two chromosomes. These kinetochore-kineto-chore interactions evidently are at the expense of kin-etochore-spindle interactions. We propose that mal-orientations persist in treated cells because thekinetochores have numerous, short microtubules with afree end that can be captured by a second kinetochore.Kinetochores capture each other’s kinetochore micro-tubules, leaving too few sites available for the efficientcapture of spindle microtubules. Since the efficient cap-ture of spindle microtubules is essential for the correc-tion of errors, failure of capture allows malorientationsto persist. Whether the effects of DMAP actually aredue to protein kinase inhibition remains to be seen. Inany case, DMAP reveals interactions of one kinetochorewith another, which, though ordinarily suppressed, haveimplications for normal mitosis.

Key words: mitosis, chromosome distribution, nondisjunction,protein kinase, DMAP (6-dimethylaminopurine)

SUMMARY

Odd chromosome movement and inaccurate chromosome distribution in

mitosis and meiosis after treatment with protein kinase inhibitors

R. Bruce Nicklas*, Lawrence E. Krawitz and Suzanne C. Ward

Department of Zoology, Duke University, Durham, NC 27706, USA

*Author for correspondence

962

kinetochores, rather than toward the spindle. Evidently thekinetochores are so bound up in interacting with eachother’s microtubules that they often fail to interact properlywith spindle microtubules.

MATERIALS AND METHODS

Grasshopper (Melanoplus sanguinipes, Fabricius) spermatocyteswere cultured at 25-26˚C, observed by phase contrast microscopy,and micromanipulated by standard procedures (Nicklas et al.,1982, and references therein). The cells were exposed to drugs bysoaking the testes in medium containing the drug for 20 min beforethe cell cultures were prepared. Mammalian cells (line PtK1 ratkangaroo kidney cells from American Type Culture Collection,Rockville, MD) were cultured in Ham’s F-12 medium supple-mented with 2% LPSR-1 (low protein serum replacement-type 1;Sigma, St. Louis, MO) and were observed at 34˚C in chambersas described by Lee (1989). The cells were exposed to inhibitorsby replacing the standard medium with one containing theinhibitor. Chromosome behavior in spermatocytes and PtK1 cellswas recorded on an optical disk recorder (model TQ-2021FCB,Panasonic Video Systems, Secaucus, NJ).

The inhibitor 6-dimethylaminopurine (DMAP; Sigma, St.Louis, MO) was added to the medium just before use to give aconcentration of 0.35-0.65 mM for spermatocytes and 0.25-0.3mM for PtK1 cells. Genistein (4′,5,7-trihydroxy-isoflavone; Ind-ofine, Somerville, NJ) was used only on spermatocytes. It was dis-solved in DMSO (dimethylsulfoxide; Mallinkrodt, Paris, KY) andadded to the spermatocyte culture medium to give a final con-centration of 40-50 µg/ml genistein and 2% DMSO; 2% DMSOby itself has no effect on spermatocyte division. For bothinhibitors, the range of useful concentrations is narrow: half thelowest concentration used has little or no effect, while twice thehighest concentration has general cytotoxic effects.

Some living cells were observed by high extinction/high reso-lution polarization microscopy, as described by Inoué (1986,1988). Nikon (Melville, NY) optics selected for freedom fromstrain were used: a rectified achromatic-aplanatic condenser witha numerical aperture (NA) of 1.4 and a 1.4NA/60× plan apo-chromatic objective. The images were acquired and processedusing an Image1 system (Universal Imaging Corp., West Chester,PA); noise was reduced by averaging, haze was removed byunsharp masking, and contrast was optimized.

Immunofluorescence staining of microtubules was carried outas described earlier (Nicklas et al., 1989), except for the use of amonoclonal anti-tubulin antibody (TU-27, generously provided byLester Binder, University of Alabama, Birmingham, AL) followedby a Cy3-labeled secondary antibody (Jackson ImmunoResearchLaboratories, West Grove, PA); Cy3 has outstandingly brightfluorescence. Several antibodies in addition to anti-tubulin werealso tried, as noted in Results; all were used with a Cy3-labeledsecondary antibody. The stained cells were viewed on a confocalmicroscope (model MRC-600, Bio-Rad, Cambridge, MA); imageswere averaged to reduce noise, sharpened by a high-pass filter,and contrast was optimized.

RESULTS

DMAP effects on spermatocytesCells treated with DMAP are like no others ever described.Maloriented chromosomes, normally rare, are common incells at all stages of division. In the cell shown in Fig. 1,four bivalents out of eleven were maloriented. One of these

eventually reoriented (arrow, 162- and 202-min prints), andthe partner chromosomes segregated normally in anaphase.This malorientation was stable for over 2.5 h before re-orientation occurred. The other three maloriented chromo-somes persisted in their misguided orientation, and the part-ner chromosomes segregated to the same spindle pole inanaphase (425.8- and 426-min prints). These malorient-ations were stable for 421 min, or 7 h, measured from thetime they were first observed until the beginning ofanaphase; the time spent without reorientation in anaphasewas not counted because reorientation never occurs inanaphase even in untreated cells. We studied a total of 25maloriented chromosomes in 16 cells treated with DMAP;4 (16%) of them reoriented while 21 (84%) persisted, lead-ing to nondisjunction in anaphase. The chromosomes thatdid not reorient were observed for a total of 3,691 min,over 60 h, in totally stable malorientations. The averagetime before reorientation of the four chromosomes that didreorient was 87 min (range: 30-152 min).

Untreated control cells are very different. Reorientationalways follows soon after a chromosome is induced to mal-orient (Nicklas and Koch, 1969). In the present experi-ments, the average time before reorientation in a sample of10 chromosomes was 9.9 min (range: 2 to 19.5 min).

Maloriented chromosomes in cells treated with DMAPare seen as soon as the preparations are made, after 20 minin DMAP. Some of these are due to the usual errors in earlyprometaphase, but others arise in cells at later stages, pre-sumably because of temporary spindle disorganizationcaused by mechanical abuse of the cells during culturepreparation. We never saw a properly oriented chromosomebecome maloriented in the presence of DMAP. Hence,DMAP does not induce malorientations, but it makes thosethat arise from other causes abnormally stable. At mid-prometaphase, the majority of DMAP-treated cells have oneor more maloriented chromosomes. In contrast, in untreatedcells at the same stage, the frequency is so low as to dis-courage counting; certainly not more than one cell in a thou-sand has a maloriented chromosome.

We used micromanipulation to characterize the malori-entations induced by DMAP. When a maloriented chromo-some is pulled toward the opposite pole with a micro-needle, both partner chromosomes are stretched (Fig. 2,0- and 26-min prints). Experiments on an additional 15 mal-oriented chromosomes gave the same result. Thus, both part-ner chromosomes of maloriented chromosomes are attachedto one spindle pole. Electron microscopy of untreated cellsdisclosed that in such malorientations the kinetochore micro-tubules of both partners extend to the pole to which theyare attached (Ault and Nicklas, 1989). Micromanipulationpermits malorientations to be induced (Nicklas and Koch,1969). In Fig. 2, 29-min print, a properly oriented chromo-some was detached from the spindle and bent to face thesame pole as its partner. Orientation to that pole promptlyfollowed, as shown by stretching the chromosome with themicromanipulation needle - both partners were attached tothe lower pole (Fig. 2, 33 min). After the new orientationwas stabilized by applying tension for 3 min, the chromo-some was released from the needle (Fig. 2, 34 min). Itmoved to the lower spindle pole (Fig. 2, 51 min) where itremained in a perfectly stable malorientation. Both partner

R. B. Nicklas, L. E. Krawitz and S. C. Ward

963Mitosis and kinase inhibitors

chromosomes segregated to the same pole in anaphase, likethe naturally occurring, maloriented chromosome at theother pole (Fig. 2, 275 min). The naturally occurring mal-orientation was stable for over 271 min and the induced mal-orientation was stable for 237 min (from the time of itsrelease from the needle to the beginning of anaphase).

In all, 9 malorientations induced by micromanipulationwere studied in 9 cells treated with DMAP. Not one of themaloriented chromosomes reoriented, and collectively theywere stable for a total of 1,184 min: almost 20 h withouta single reorientation.

DMAP at the concentration used affects mitoticprocesses in addition to chromosome distribution, thoughnot so dramatically. Chromosome congression to ametaphase position at the spindle equator sometimes failsin treated cells (Fig. 1), and such failure is rare in untreatedcells. However, most chromosomes congress normally inthe presence of DMAP (Figs 1 and 2). Chromosome move-ment in anaphase looks normal (Figs 1 and 2), but the rateis greatly decreased - from an average of 1.04 µm/min inuntreated cells to 0.38 µm/min in treated cells (the rates areaverages for 3-4 chromosomes in each of 4 untreated and

Fig. 1. Persistent malorientations and nondisjunction in a spermatocyte treated with 0.55 mM DMAP. The time in min is given on eachprint. Four maloriented chromosomes were present initially (arrows, 0-min print). One of these subsequently reoriented (arrow, 162 and202 min), but the other three did not, and their partner chromosomes were distributed to the same pole in anaphase rather than to oppositepoles (arrows, 425.8- and 426-min prints; arrowhead: the unpaired X-chromosome, segregating to one pole, which is normal).Chromosome movement was generally normal in this cell, but one chromosome (lower left, 162-425.8 min) failed to congress to aposition midway between the poles. Bar, 10 µm.

Fig. 2. Micromanipulation experiments in a spermatocyte treated with 0.55 mMDMAP. The time in min is given on each print. A maloriented chromosome(arrow, 0-min print) was pulled downward with a microneedle (26 min). Bothpartner chromosomes were stretched (arrows at kinetochores, 26 min), revealingthat both were attached to the same (upper) spindle pole. In a second experiment,a properly oriented chromosome was induced to malorient (arrow, 29 min). Onepartner chromosome was detached from the spindle and bent, so that thekinetochores of both partners faced the same (lower) pole. Soon, both partnershad attached to the lower pole, as shown by stretching the chromosome (arrowsat kinetochores, 33 min) with the microneedle. After release from the needle, themaloriented chromosome moved to the lower pole (34-51 min), and bothpartners were distributed to that pole in anaphase (arrows, 275 min). The originalmalorientation (upper pole, 0-275 min) also persisted, and nondisjunctionfollowed in anaphase (275 min). Bar, 10 µm.

964

3 treated cells). DMAP has more drastic effects on the lesshealthy cells that inevitably occur in spermatocyte cultures(Nicklas et al., 1979). In particular, hypotonicity causesshort, ill-organized spindles in untreated cells, and DMAPexacerbates this effect so that the spindles quickly collapse.

Prometaphase-metaphase lasts a long and variable time(4-8 h) in both untreated and treated cells. The onset ofanaphase may be somewhat delayed in treated cells whenseveral maloriented chromosomes are present (Fig. 1), butcertainly the cells are not blocked in metaphase. Nor is theentry into mitosis blocked: spindle formation andprometaphase are initiated in cells after 5 h or more of expo-sure to DMAP.

Genistein effects on spermatocytesIn cells treated with genistein, maloriented chromosomesare indefinitely stable and nondisjunction often is the con-sequence (Fig. 3), just as with DMAP. Experiments withgenistein were limited to confirming in 8 cells the moreextensive observations on cells treated with DMAP. Sixmaloriented chromosomes reoriented but only after an aver-age of 40.7 min (range: 25 to 60 min). Three chromosomesdid not reorient; they spent a total of 255 min in completelystable malorientations and showed nondisjunction inanaphase. Genistein has a less specific effect on chromo-some distribution than does DMAP. Even at concentrationsthat cause fewer persistent malorientations than DMAP,genistein inhibits chromosome movement in anaphase (Fig.3), and the spindles at all stages are shorter and subject tocollapse.

DMAP effects on PtK1 cellsDMAP stabilizes improper orientations in PtK1 cells, andthereby causes nondisjunction in anaphase (Fig. 4). Eightimproperly oriented chromosomes in 5 cells spent a totaltime of 580 min, over 9 h, without a single correction -nondisjunction of the sister chromatids invariably occurred.Metaphase is prolonged in these cells, with an average dura-tion of 91 min, as compared to 25-40 min in controls. Chro-mosome movement to the spindle equator in prometaphaseand to the poles in anaphase is not noticeably affected by0.25 mM DMAP, but spindle elongation in anaphase isoften inhibited, as in the cell in Fig. 4.

The effects of DMAP on PtK1 cells are swiftly reversible(Fig. 5). Soon after the DMAP-medium was replaced withnormal medium, the maloriented chromosomes orientedproperly, as signaled by movement toward the equator (32-and 35-min prints), and a normal anaphase followed (73min). Two additional experiments gave the same result; onaverage the maloriented chromosome oriented properly 9.2min after the DMAP was removed, and anaphase followed20.0 min later.

Why do malorientations persist in treated cells?The cause of persistent malorientations was investigated inDMAP-treated spermatocytes. The correction of malorient-ation by reorientation requires the formation of a new,proper kinetochore attachment and the loss of an old,improper one.

Attachment testsWe tested the possibility that the formation of new attach-ments is defective in treated spermatocytes by detachingchromosomes from the spindle by micromanipulation. Inuntreated cells, detached chromosomes initially lack kineto-chore microtubules, but invariably acquire new ones andbegin to move again (Nicklas and Kubai, 1985). Therefore,the lag time between detachment and the beginning ofmovement is a measure of the time required for reattach-ment to the spindle by the acquisition of new kinetochoremicrotubules. We detached relatively small chromosomesand placed them at one pole of the spindle. One kineto-chore faced the farther pole, nearly as far from that pole asa kinetochore in a maloriented chromosome. The averagetime required for the reattachment of that kinetochore, sig-naled by movement toward the farther pole, was 1.6 minin untreated cells (range: 0.2-4.7 min; 13 experiments) and4.4 min in cells treated with DMAP (range: 1.8-11.8 min;13 experiments). The delay in reattachment in cells treatedwith DMAP is statistically significant (t-test, P = 0.01), buta delay of a few minutes cannot be the reason that mal-orientations often persist for several hours.

Tether testsMicromanipulation was also used to test the possibility thatthe old, improper attachment persists in treated cells, and

R. B. Nicklas, L. E. Krawitz and S. C. Ward

Fig. 3. A second inhibitor, genistein, also causes persistent malorientation and nondisjunction. This spermatocyte was treated with 40µg/ml genistein. The time in min is given on each print. A maloriented chromosome (arrows) was stable for 60 min, and both partnerswere distributed to the same pole in anaphase (61 and 65 min prints). Bar, 10 µm.

965Mitosis and kinase inhibitors

therefore keeps the kinetochores tethered to the same pole.A maloriented chromosome was produced by micromanip-ulation, as described above (Fig. 2). After some time, aneedle was inserted near a kinetochore and gently movedtoward the opposite pole, to mimic the movement follow-ing the capture of a microtubule extending toward that pole.Sometimes the tug of the needle was resisted, and the kin-etochore region was pulled out (Fig. 6, upper row). Obvi-ously the kinetochore was still tethered to the spindle, andtherefore if a microtubule from the opposite pole had beencaptured by the kinetochore, the associated motors wouldnot have been able to shift the kinetochore toward the oppo-site pole. Hence, capture at that time probably would nothave led to reorientation. In contrast, sometimes the kin-etochore could be moved by the microneedle without resis-tance (Fig. 6, lower row). At such a time, a properlydirected microtubule, if captured by the kinetochore, couldhave led to reorientation.

The experiment must be done somewhat differently incontrol and treated cells. The time before reorientation ishighly variable in control cells, so we obtained a randomsample by waiting until one of the two kinetochores of amaloriented bivalent began to reorient and then probed theattachment of the other kinetochore. This would not workwith cells treated with DMAP, simply because waiting forreorientation often would be futile. We reasoned that if

DMAP prevents reorientation by stabilizing old spindleattachments, it must do so when reorientation normallyoccurs in untreated cells, at an average of 10 min after mal-orientation. Hence, we probed for attachment in DMAP-treated cells at 10 min after malorientation. We found thatimproper attachments commonly persist in both treated anduntreated cells, but there is not much difference betweenthem. About half of the kinetochores were tethered to thepole, 53% in untreated cells (15 experiments) and 63% incells treated with DMAP (24 experiments).

Kinetochore-kinetochore interactionsMore by accident than design, we discovered that DMAPcauses a unique form of chromosome movement that isbased on kinetochore-kinetochore interactions. A mal-oriented chromosome (arrows, Fig. 7, 0-min print) wasdetached from the spindle and moved to the cytoplasm (13min). It remained at rest for several min, and then was sud-denly jerked toward the spindle (21 min). The movementdid not continue, but instead, the two kinetochoric ends ofthe chromosome curled toward each other (22 min): the endnearer the spindle at 21 min actually moved away from thespindle, toward the other kinetochore. This movement, too,did not persist; the chromosome uncurled (29 min) only tocurl again (31 min). In this curled condition, the chromo-some finally began to move back toward the spindle (34

Fig. 4. DMAP causes nondisjunction in a mammalian cell in mitosis. This PtK1 cell was treated with 0.25 mM DMAP. The time in min isgiven on each print. Two chromosomes (arrows, 0-min print) lying near the upper spindle pole remained there, and their sister chromatidswere distributed to that pole in anaphase (41 and 54 min) rather than to opposite poles. Bar, 10 µm.

Fig. 5. The effects of DMAP are reversible. The time in min is given on each print. Two chromosomes (arrows, 30-min print) in this PtK1cell remained maloriented for 30 min while in 0.28 mM DMAP (one of the two chromosomes is obscured by the other in the 0-min print).At 30 min, the DMAP was washed out, and first one chromosome (lower arrow, 32 min) and then the other (35 min) oriented properlyand moved to the metaphase plate. A normal anaphase followed (73 min). Bar, 10 µm.

966

min) and reached the spindle with its kinetochores together,maloriented. The malorientation (70-min print) persisted for42 min, but eventually the chromosome reoriented andmoved toward the equator (105 min).

These kinetochore-kinetochore interactions are not lim-ited to maloriented chromosomes, and more than one chro-mosome can join in the dance. In the cell in Fig. 8, twochromosomes, one maloriented chromosome and one nor-mally oriented, were detached from the spindle and placednear one another in the cytoplasm, far from the spindle (Fig.8, 0 min). A kinetochore of one chromosome approacheda kinetochore of the other chromosome (1- and 7-minprints) until the two kinetochores touched. The duo becameattached to the spindle and moved to it, still kinetochore-to-kinetochore (12- and 20-min prints). Eventually, bothoriented properly on the spindle (30 min).

The average velocity of the kinetochore-to-kinetochoremovements in these two cells was 2.1 µm/min (n = 5). Thisvalue is indistinguishable from the velocity of the chromo-some-to-spindle movements that follow chromosomedetachment from the spindle and for ordinary chromosome-spindle movement early in prometaphase (0.7-2.3 µm/min:Nicklas, 1967).

The kinetochores that move toward one another aremechanically connected, very likely by microtubules. Themechanical linkage was revealed by micromanipulation.For instance, the chromosomes in Fig. 8 were both movedcloser to the spindle (7- and 12-min prints) by pulling onjust one of the chromosomes. Occasionally, presumptivemicrotubules between interacting kinetochores can be seenby polarization microscopy (Fig. 9A-C); they look just likemicrotubules in spindles (Fig. 9D). Attempts to demonstratekinetochore-to-kinetochore microtubules by immunostain-

ing yielded no really satisfactory evidence. This is prob-ably because the interactions are so transitory (Fig. 7) and/or because the microtubules are hard to preserve.

Kinetochore-kinetochore interactions have never beenseen in hundreds of detachment experiments in untreatedcells. In contrast, one or more kinetochore-kinetochoreinteractions were seen in 56% (15 of 27) of experimentson cells treated with DMAP. Interactions were not seenwhen the detached chromosome or chromosomes soonmoved back to the spindle, and hence there was little timefor kinetochore-kinetochore interactions. The comparisonbetween normal and DMAP-treated cells is imperfect, ofcourse, because kinetochores in untreated cells usually haverelatively little time to interact with each other before theybecome attached to the spindle. We did a set of detachmentexperiments in untreated cells in which detached chromo-somes were placed very far from the spindle, as in theexperiments on DMAP-treated cells such as those in Figs7 and 8. In 16 experiments, the chromosome began a sus-tained movement back toward the spindle in an average of1.3 min (range, 0.2-6 min). Thus, in these cells, there isvery little time for kinetochore-kinetochore interactionbefore interactions with the spindle begin. In 15 compar-able experiments in DMAP-treated cells, the first movementtoward the spindle occurred after an average of 5.7 min(range, 0.2-18 min). This is not very different from thevalue for controls, but often the movement was not sus-tained. The time from detachment until sustained movement(to be compared with the average of 1.3 min in untreatedcells) was 28 min, with a range of 0.2-180 min.

Labile initial attachmentsThe difference between the time for initial movement

R. B. Nicklas, L. E. Krawitz and S. C. Ward

Fig. 6. Tether tests on aspermatocyte treated with 0.6mM DMAP. The time in minis given on each print. Achromosome was induced bymicromanipulation tomalorient to the upper poleand then released from theneedle. 10 min later, onepartner chromosome waspulled gently toward the lowerpole with a microneedle (toosmall to be visible) insertednear its kinetochore (arrows,0-0.5-min prints). Thekinetochore region was pulledout (0.3 and 0.5 min),revealing that the kinetochorewas firmly tethered to theupper pole. In contrast, thekinetochore of the otherpartner chromosome wasfreely movable (arrow, 1.6-2.5min): it was not tethered andcould have reoriented andmoved to the lower pole if ithad captured a microtubulefrom that pole. Bar, 10 min.

967Mitosis and kinase inhibitors

versus sustained movement in the above experiments is dueto the lability of the initial attachment of chromosomes inDMAP-treated cells. A good example is the movement ofone kinetochore in Fig. 7. Between 13 and 21 min, the kin-etochore moved several µm toward the spindle, and then,1 min later, it moved in the opposite direction. Whendetached chromosomes are placed far from the spindle,twitches of one kinetochore or the other toward the spin-dle occur and recur but movement is rarely sustained. Inconsequence, chromosomes often remain for many minuteswithout stable interactions with the spindle, allowing timefor kinetochore-kinetochore interactions to occur.

Immunostaining observationsSpindle microtubule organization in DMAP-treated cellscannot be distinguished from that in untreated spermato-

cytes (Fig. 10) or PtK1 cells (not illustrated). Note that thereare numerous microtubules in the cytoplasm in both treatedand untreated cells that detached chromosomes might cap-ture (Fig. 10). Consistent with a plentiful supply of cyto-plasmic microtubules is the short time required for theinitial reattachment of detached chromosomes in bothuntreated and treated cells.

We tried several antibodies specific for various phos-phoproteins and other proteins that might be affected by theinhibitors: (1) monoclonal anti-phosphotyrosine antibodies:number 5-321 (Upstate Biotechnology, Inc., Lake Placid,NY), PY-20 (ICN Biomedicals, Inc., Costa Mesa, CA) andPY-54 (Zymed Laboratories, South San Francisco, CA); (2)polyclonal anti-phosphorylation antibodies: #702 (JeanWang, University of California, San Diego, CA); (3) anti-sarc (David Morgan, University of California, San Fran-

Fig. 7. A solo dance. Kinetochore-kinetochore movements within a single chromosome in a spermatocyte treated with 0.35 mM DMAP.The time in min is given on each print. A maloriented chromosome (arrows at kinetochores, 0-min print) was detached from the spindlewith a microneedle and placed far out in the cytoplasm (arrows, 13 min). After darting a short distance toward the spindle (21 min), thechromosome stopped, and one kinetochore moved toward the other so that the chromosome curled on itself (arrows, 22 min). Thechromosome relaxed (29 min) but then curled again (31 min). With kinetochore pressed to kinetochore, the chromosome moved backtoward the spindle (34 min), and both kinetochores oriented to the upper pole (70 min). Eventually, the chromosome oriented properlyand moved toward the equator (105 min). Bar, 10 µm.

968

cisco, CA); (4) anti-γ-tubulin (Berl Oakley, Ohio State Uni-versity, Columbus, OH) and (5) MPM-2, an antibody thatreacts with several phosphoproteins prominent in mitoticcells (Potu Rao, M. D. Anderson Cancer Center, Houston,TX). We found no significant difference between treatedand untreated cells with any of these antibodies. This neg-ative result is disappointing but not meaningful. Generalphosphoprotein antibodies such as MPM-2 may or may notreact with particular proteins, e.g. those phosphorylated ontyrosine, and the staining with anti-phosphotyrosine anti-bodies is so low in intensity and so diffuse in control cellsthat important differences between controls and treated cellswould be undetectable even if present.

DISCUSSION

Two protein kinase inhibitors, DMAP and genistein, causeinaccurate chromosome segregation at meiosis in grasshop-per spermatocytes. Reorientation, the normal process oferror correction, often fails to occur, and improper chro-mosome orientations persist for hours. In sharp contrast, inuntreated cells, on average only 10 min pass before mal-oriented chromosomes reorient. DMAP also causes mal-

orientations to persist in mammalian (PtK1) cells in mito-sis and again, inaccurate chromosome distribution to thedaughter cells is the result. In PtK1 cells, the effects ofDMAP are completely reversible. Reversibility of drugeffects in grasshopper spermatocytes cannot be tested in ourpreparations because the cells are not attached to the glassand hence the medium around the cells cannot be replaced.Many other treatments and conditions are known to pro-duce abnormal chromosome complements (Dellarco et al.,1985), including inhibitors of certain protein phosphatases(Vandré and Wills, 1992) and kinases (Andreassen andMargolis, 1991). DMAP and genistein, however, are theonly agents that are known to do so by stabilizing mal-orientations.

The inhibitors we studied affect other processes in addi-tion to chromosome distribution. Chromosome congressionsometimes fails. Chromosome movement in anaphase isreduced in extent and speed. Especially with genistein, thespindles sometimes collapse, halting cell division. Theonset of anaphase may be delayed in spermatocytes andcertainly is delayed in PtK1 cells. Multiple effects are nosurprise in view of the numerous processes affected byDMAP, including progression through the cell cycle (e.g.see Rebhun et al., 1973; Néant and Guerrier, 1988; Luca

R. B. Nicklas, L. E. Krawitz and S. C. Ward

Fig. 8. A pas de deux. Kinetochore-kinetochore movements of two chromosomes in a spermatocyte treated with 0.35 mM DMAP. Thetime in min is given on each print. Two chromosomes were detached and placed in the cytoplasm (arrows at two kinetochores, 0-minprint). One kinetochore of the upper chromosome approached and touched a kinetochore of the other chromosome (1-7 min). The twochromosomes were mechanically connected, as shown by pulling one of them toward the spindle with a micromanipulation needle - theother chromosome was towed along (12 min). The two chromosomes moved together to the spindle, but later they oriented properly andseparated (20-30 min). Bar, 10 µm.

969Mitosis and kinase inhibitors

and Ruderman, 1989; Jessus et al., 1991) and spindle org-anization (Dufresne et al., 1991). What is noteworthy is thatproper chromosome distribution is more sensitive to thedrugs than most other processes in the cells we studied.

DMAP and genistein are structurally dissimilar com-pounds yet both are potent and specific inhibitors of par-ticular protein kinases, tyrosine kinases, in vitro (Néantand Guerrier, 1988; Akiyama et al., 1987). In living cells,s p e c i fic inhibition of protein tyrosine phosphorylation bythese drugs has been directly demonstrated in a variety ofcells. For example, DMAP inhibits the phosphorylation ofparticular proteins in echinoderm, X e n o p u s and mouseoocytes (Néant et al., 1989; Jessus et al., 1991; Rime etal., 1989). As for genistein, the list includes A431 humancarcinoma cells, mouse fibroblasts, human platelets andlymphocytes (Akiyama et al., 1987; Hill et al., 1990;Gaudette and Holub, 1990; Lane et al., 1991). Compara-bly direct evidence that the inhibitors affect protein phos-phorylation in spermatocytes and PtK1 cells is lacking atpresent. We note, however, that we used DMAP andgenistein at the same concentrations that specific a l l y

inhibit tyrosine kinase in a great variety of other livingcells and that the two very different inhibitors have iden-tical effects on chromosome distribution in spermatocytes.Thus, reduced protein phosphorylation may be the causeof inaccurate chromosome distribution in cells treated withthe inhibitors, but other possibilities certainly are not ruledout. Our major focus in the present report is the uniqueeffects of the drugs on chromosome movement and dis-tribution, leaving the molecular mechanism an open ques-tion.

Whatever the inhibitors do to molecules, how does thisresult in abnormally stable malorientations and inaccuratechromosome distribution? Two conditions are necessary forreorientation: a new chromosome-spindle attachment mustform, and an old attachment must be lost. Either of theseprocesses might be affected by the drugs.

The first condition for reorientation is satisfied when akinetochore captures microtubules growing outward from apole (Mitchison and Kirschner, 1985b; Rieder and Alexan-der, 1990; Hayden et al., 1990). This requires a plentifulsupply of microtubules extending from each pole

Fig. 9. (A-C) Polarization microscopy of chromosome-chromosome movement in a spermatocyte treated with 0.5 mM DMAP. Fibersrunning vertically appear in maximum dark contrast. (A) Two chromosomes (arrows at kinetochores) were detached from the spindle aminute earlier and moved to a cytoplasmic region above one spindle pole. A dark fiber extends between the kinetochores. (B,C) Thechromosomes move together. The fiber seen in A is scarcely visible 5 s later (B), but this is commonly the case, because the fibers areconstantly in motion and the depth of focus is very shallow. The elapsed time between A and C is 20 s. (D) Very early prometaphase inanother living spermatocyte to show the appearance of microtubules when viewed by polarization optics. One chromosome (arrowheadsat kinetochores) is shown. The spindle poles lie out of view, above and just below the region shown. Several fibers, one of whichterminates in a kinetochore, are identified by arrows. At early prometaphase, electron microscopy discloses that microtubules are sparse,so the fibers probably are composed of one or a very few microtubules. Bar, 5 µm.

970

nearly to the opposite pole and kinetochores that arecompetent to capture them.

A deficient supply of long microtubules probablyexplains the persistence of maloriented chromosomes inP t K1 cells treated with DMAP. “Centrophilic chromo-somes” are common in untreated cells in mitosis. They lienear a pole to which one sister kinetochore is oriented,while the second kinetochore faces away from the poleand usually is not oriented at all - it lacks kinetochoremicrotubules (Rieder, 1990; Ault and Rieder 1992). Aftera few minutes, the second kinetochore orients to the oppo-site pole and normal anaphase follows. When cells areexposed to DMAP, however, orientation to the oppositepole is indefinitely delayed and nondisjunction oftenresults. The kinetochores are competent to capture micro-tubules, since those facing a nearby pole do so. The un-oriented kinetochore, however, faces a pole much fartheraway and hence a shortage of long microtubules wouldaccount for its failure to attach to the spindle. Certain pro-tein kinases are known to affect microtubule length dis-tribution (Verde et al., 1992 and references therein), andDMAP affects those kinases, if indirectly (Luca and Rud-erman, 1989). These studies provide no direct evidencethat DMAP treatment reduces the number of long micro-

R. B. Nicklas, L. E. Krawitz and S. C. Ward

Fig. 10. Microtubule immunostainingof two untreated spermatocytes (A andB) and two spermatocytes treated withDMAP (C and D; 0.55 mM and 0.35mM DMAP, respectively). Neither thespindle nor cytoplasmic microtubulesof treated cells can be distinguishedfrom untreated cells; the spindle of thecell in C is shorter than the others inthis sample but not outside the normalrange. Both the treated cells hadmaloriented chromosomes though theyare not visible in these images (cell Chad one and cell D had two). Bar, 10µm.

Fig. 11. A diagram of one model for kinetochore-kinetochoremovement in the presence of DMAP. Kinetochores arerepresented by black circles and microtubules by thin lines; in (B),the polarity of the ends of two microtubules is indicated by plus orminus signs and the direction of movement by the nearby arrow.(A) Kinetochores have short microtubules attached to them, as aresult of either kinetochore nucleation of microtubules or ofcapture of microtubule fragments nucleated elsewhere. (B) Thefree end of a microtubule attached to one kinetochore can becaptured by another kinetochore, regardless of microtubulepolarity. Movement results from minus-end directed kinetochoremotors.

971Mitosis and kinase inhibitors

tubules in mitosis, but at least they show that such aneffect is plausible.

The situation is different in grasshopper spermatocytes,however, and the remainder of the discussion centers onthem. In spermatocytes, the partner kinetochores of a mal-oriented chromosome are both attached to the same pole(Fig. 2), and their kinetochore microtubules extend towardthat spindle pole (Ault and Nicklas, 1989). Reorientationrequires the formation of a new spindle attachment via thecapture of microtubules from the opposite pole by one kin-etochore or the other. We determined the time required forthe formation of new attachments. A longer time wasrequired in cells treated with DMAP, but still was only 4.4min on average. Obviously, a delay of a few minutes inmaking a new attachment cannot explain why malorient-ations generally persist for several hours.

Formation of a new attachment is not sufficient for re-orientation. A second requirement is that the old, misdirectedkinetochore microtubules must be lost or must lose theiranchorage to the spindle. Otherwise, the reorienting kin-etochore is not free to move toward the opposite pole (Nick-las and Kubai, 1985; Ault and Nicklas, 1989). We directlytested for loss of anchorage in maloriented chromosomes byusing a micromanipulation needle to mimic a microtubuleand the motors associated with it, asking whether or not thekinetochore was free to move at a given moment. Cellstreated with DMAP had a somewhat smaller fraction offreely movable kinetochores, but in both treated anduntreated cells, only a minority of kinetochores was free tomove at a given time. Thus, anchorage to the pole appearsto be a significant impediment to reorientation even inuntreated cells. While the impediment may be a little greaterin cells treated with DMAP, it is an unlikely source of theenormous difference in how long malorientations persist.

We conclude that in spermatocytes, the processes neces-sary for reorientation are not much affected by DMAP.Hence the explanation for persistent malorientations mustlie elsewhere, probably in the novel kinetochore-kineto-chore interactions that DMAP causes. Often, a kinetochoremoves toward another kinetochore, rather than toward thespindle. The kinetochores of one chromosome (Fig. 7) orof two chromosomes (Fig. 8) can move together. The veloc-ity of kinetochore-kinetochore movement is typical of ord-inary, microtubule-based, chromosome movement. Theinteracting kinetochores are mechanically linked together(Fig. 8) by connections that have the appearance of spin-dle microtubules as seen by polarization microscopy (Fig.9). The possibility that the connections are chromatin fiberscan be dismissed: while chromosomes sometimes sticktogether in hypertonic cells (whether or not DMAP is pre-sent), such connections once formed are stable and neverlead to dances such as those in Figs 7 and 8. A strikingexample of the impact that kinetochore-kinetochore associ-ations can have on connection to the spindle is seen in Fig.7, from 13 to 22 min. A kinetochore achieves a connectionto the spindle, but that connection lapses. The kinetochorethen becomes connected to its partner’s kinetochore andmoves toward it, actually moving away from the spindle.

Kinetochore-kinetochore interactions via microtubulesprovide a natural explanation for the failure of reorient-ation. Each kinetochore very likely has a definite, limited

number of sites for capturing and binding microtubules(Nicklas, 1988; Zinkowski et al., 1991). We suggest thatthe capture of microtubules from other kinetochoresreduces the number of spindle microtubules the kineto-chore can capture: the maloriented kinetochores of DMAP-treated spermatocytes are literally too attached to oneanother to attach properly to the spindle. Notice that anyreduction in the number of unoccupied capture sites bykinetochore-kinetochore interactions is particularly apt toaffect maloriented chromosomes. Their kinetochores facean abundance of easily captured microtubules from thenearby pole (e.g. see Fig. 3) and do not directly face themicrotubules growing from the opposite pole. Even inuntreated cells, the initiation of reorientation by the cap-ture of a microtubule from the opposite pole is an improb-able event (Nicklas, 1988). Hence reorientation may wellbe inhibited if the number of available capture sites forspindle microtubules is reduced by the capture of kineto-chore microtubules.

The impact of kinetochore-kinetochore interactions onreorientation is probably enhanced by the lability of theinitial attachment of chromosomes to the spindle. In treatedcells, the initial movements of detached chromosomestoward the spindle often do not continue and may evenreverse (Fig. 7, 13-22 min). Apparently, the first micro-tubules to be captured are unstable themselves or haveunstable connections at the kinetochore or pole. The fre-quent failure of the initial spindle attachments to persistallows more time for kinetochore-kinetochore attachmentsto form in treated cells. It should be noted that unstablekinetochore-spindle interactions are not unique to cellstreated with DMAP. The initial movements of chromo-somes in untreated cells are intermittent and jerky (Alexan-der and Rieder, 1991; Nicklas, 1967), though pauses orreversals of the sort common in treated cells have not beenseen.

Assuming that microtubules are the basis of kinetochore-kinetochore movements, where do those microtubules comefrom and how does movement result? There are several pos-sibilities, but the following is perhaps the most likely. Sup-pose that DMAP treatment leads to kinetochores havingshort microtubules attached to them (Fig. 11A). Thosemicrotubules might arise because the drug increases thenumber of short microtubules not associated with a polewhich then are captured by kinetochores (normally, suchmicrotubule fragments probably are rare, to judge from thefew microtubules with unexpected polarity: McIntosh andEuteneuer, 1984). Alternatively, the drug might make kin-etochores more efficient in nucleating microtubules, so thatthey grow their own microtubules (kinetochores fromuntreated cells are inefficient nucleators in vitro: Mitchisonand Kirschner, 1985a). Whatever the origin of short kin-etochore microtubules might be, the free ends of thesemicrotubules could be captured by a second kinetochore(Fig. 11B). Kinetochores capture microtubule ends of eitherpolarity, at least in vitro (Mitchison and Kirschner, 1985b).Single microtubules would connect two kinetochores, withthe plus end at one kinetochore and the minus end at theother (Fig. 11B). The two kinetochores could then movetoward each other by the minus-end-directed kinetochoremotors thought to produce normal chromosome movement

972

in anaphase (reviewed by McIntosh and Pfarr, 1991; Rieder,1991).

There is some electron microscopic evidence that kine-tochore-kinetochore microtubules can arise under condi-tions in which kinetochores nucleate microtubules (McGilland Brinkley, 1975; Witt et al., 1980) and perhaps also, ifvery rarely, in normal mitosis (Bajer and Molè-Bajer, 1969;Luykx, 1970). As all of these authors imply, however, it isnot certain from the electron microscopic images that themicrotubules observed actually terminated in the kineto-chores and were attached to them.

In conclusion, a protein kinase inhibitor, DMAP, pro-duces a sort of chromosome movement that has never beenreported before, which is based on kinetochore-kinetochoreinteractions. Evidently these interactions occur at theexpense of normal kinetochore-spindle interactions. Theoutcome is persistent malorientation and inaccurate chro-mosome distribution.

The existence of frequent kinetochore-kinetochore inter-actions in treated cells raises the question of how they areavoided in normal mitosis and meiosis. Microtubule dynam-ics and kinetochore-microtubule encounters must be pre-cisely regulated to achieve a balance between unworkablestability (no reorientation) and chaotic change (continuousreorientation) so that errors in chromosome attachment tothe spindle can be corrected. The normal absence of inter-actions between kinetochores very likely is a consequenceof the essential regulation of mitotic encounters in general.If so, the interaction of one kinetochore with another,though aberrant, may reveal important features of regula-tion in normal cells.

We thank Robin Fulmer and Danette Miller for expert techni-cal help, Dahong Zhang for assistance with confocal and polar-ization microscopy and Donna Maroni for unsparing editorialreview. For generous gifts of antibodies, we are grateful to LesterBinder, David Morgan, Berl Oakley, Potu Rao and Jean Wang.This investigation was supported in part by grant GM 13745 fromthe Institute of General Medical Sciences, National Institutes ofHealth.

REFERENCES

Akiyama, T., Ishida, J., Nakagawa, S., Ogawara, H., Watanabe, S., Itoh,N., Shibuya, M. and Fukami, Y. (1987). Genistein, a specific inhibitorof tyrosine-specific protein kinases. J. Biol. Chem. 262, 5592-5595.

Alexander, S.P. and Rieder, C.L. (1991). Chromosome motion duringattachment to the vertebrate spindle: initial saltatory-like behavior ofchromosomes and quantitative analysis of force production by nascentkinetochore fibers. J. Cell Biol. 113, 805-815.

Andreassen, P.R. and Margolis, R.L. (1991). Induction of partial mitosisin BHK cells by 2-aminopurine. J. Cell Sci. 100, 299-310.

Ault, J.G. and Nicklas, R.B. (1989). Tension, microtubule rearrangements,and the proper distribution of chromosomes in mitosis. Chromosoma 98,33-39.

Ault, J.G. and Rieder, C.L. (1992). Chromosome mal-orientation andreorientation during mitosis. Cell Motil. Cytoskel. 22, 155-159.

Bajer, A. and Molè-Bajer, J. (1969). Formation of spindle fibers,kinetochore orientation, and behavior of the nuclear envelope duringmitosis in endosperm: fine structural and in vitro studies. Chromosoma27, 448-484.

Dellarco, V.L., Voytek, P. E. and Hollaender, A., ed. (1985). Aneuploidy:Etiology and Mechanisms. New York: Plenum Press.

Dufresne, L., Néant, I., St-Pierre, J., Dubé, F. and Guerrier, P. (1991).

Effects of 6-dimethylaminopurine on microtubules and putativeintermediate filaments in sea urchin embryos. J. Cell Sci. 99, 721-730.

Gaudette, D.C. and Holub, B.J. (1990). Effect of genistein, a tyrosinekinase inhibitor, on U46619-induced phosphoinositide phosphorylationin human platelets. Biochem. Biophys. Res. Commun.170, 238-242.

Hayden, J.H., Bowser, S.S. and Rieder, C.L. (1990). Kinetochorescapture astral microtubules during chromosome attachment to the mitoticspindle: direct visualization in live newt lung cells. J. Cell Biol. 111,1039-1045.

Hill, T.D., Dean, N.M., Mordan, L.J., Lau, A.F., Kanemitsu, M.Y. andBoynton, A.L. (1990). PDGF-induced activation of phospholipase C isnot required for induction of DNA synthesis. Science 248, 1660-1663.

Inoué, S. (1986). Video Microscopy. New York: Plenum Press.Inoué, S. (1988). Progress in video microscopy. Cell Motil. Cytoskel. 10,

13-17.Jessus, C., Rime, H., Haccard, O., Van Lint, J., Goris, J., Merlevede, W.

and Ozon, R. (1991). Tyrosine phosphorylation of p34cdc2 and p42during meiotic maturation of Xenopus oocyte. Development 111, 813-820.

Lane, P.J.L., Ledbetter, J.A., McConnell, F.M., Draves, K., Deans, J.,Schieven, G.L. and Clark, E.A. (1991). The role of tyrosinephosphorylation in signal transduction through surface Ig in human Bcells. J. Immunol. 146, 715-722.

Lee, G.M. (1989). Characterization of mitotic motors by their relativesensitivity to AMP-PNP. J. Cell Sci. 94, 425-441.

Luca, F.C. and Ruderman, J.V. (1989). Control of programmed cyclindestruction in a cell-free system. J. Cell Biol. 109, 1895-1909.

Luykx, P. (1970). Cellular Mechanisms of Chromosome Distribution.International Review of Cytology.Suppl. 2. New York: Academic Press.

McGill, M. and Brinkley, B.R. (1975). Human chromosomes andcentrioles as nucleating sites for the in vitro assembly of microtubulesfrom bovine brain tubulin. J. Cell Biol.67, 189-199.

McIntosh, J.R. and Euteneuer, U. (1984). Tubulin hooks as probes formicrotubule polarity: an analysis of the method and an evaluation of dataon microtubule polarity in the mitotic spindle. J. Cell Biol. 98, 525-533.

McIntosh, J.R. and Pfarr, C.M. (1991). Mitotic motors. J. Cell Biol. 115,577-585.

Mitchison, T.J. and Kirschner, M.W. (1985a). Properties of thekinetochore in vitro. I. Microtubule nucleation and tubulin binding. J.Cell Biol. 101, 755-765.

Mitchison, T.J. and Kirschner, M.W. (1985b). Properties of thekinetochore in vitro. II. Microtubule capture and ATP-dependenttranslocation. J. Cell Biol.101, 766-777.

Néant, I., Charbonneau, M. and Guerrier, P. (1989). A requirement forprotein phosphorylation in regulating the meiotic and mitotic cell cyclesin echinoderms. Dev. Biol. 132, 304-314.

Néant, I. and Guerrier, P. (1988). 6-Dimethylaminopurine blocks starfishoocyte maturation by inhibiting a relevant protein kinase activity. Exp.Cell Res. 176, 68-79.

Nicklas, R.B. (1967). Chromosome micromanipulation. II. Inducedreorientation and the experimental control of segregation in meiosis.Chromosoma 21, 17-50.

Nicklas, R.B. (1988). Chance encounters and precision in mitosis. J. CellSci. 89, 283-285.

Nicklas, R.B., Brinkley, B.R., Pepper, D.A., Kubai, D.F. and Rickards,G.K. (1979). Electron microscopy of spermatocytes previously studied inlife: methods and some observations on micromanipulated chromosomes.J. Cell Sci. 35, 87-104.

Nicklas, R.B. and Koch, C.A. (1969). Chromosome micromanipulation.III. Spindle fiber tension and the reorientation of mal-orientedchromosomes. J. Cell Biol. 43, 40-50.

Nicklas, R.B. and Kubai, D.F. (1985). Microtubules, chromosomemovement, and reorientation after chromosomes are detached from thespindle by micromanipulation. Chromosoma 92, 313-324.

Nicklas, R.B., Kubai, D.F. and Hays, T.S. (1982). Spindle microtubulesand their mechanical associations after micromanipulation in anaphase. J.Cell Biol. 95, 91-104.

Nicklas, R.B., Lee, G.M., Rieder, C.L. and Rupp, G. (1989).Mechanically cut mitotic spindles: clean cuts and stable microtubules. J.Cell Sci. 94, 415-423.

Rebhun, L.I., White, D., Sander, G. and Ivy, N. (1973). Cleavageinhibition in marine eggs by puromycin and 6-dimethylaminopurine. Exp.Cell Res. 77, 312-318.

Rieder, C.L. (1990). Formation of the astral mitotic spindle: ultrastructural

R. B. Nicklas, L. E. Krawitz and S. C. Ward

973Mitosis and kinase inhibitors

basis for the centrosome-kinetochore interaction. Electron Microsc. Rev.3, 269-300.

Rieder, C.L. (1991). Mitosis: towards a molecular understanding ofchromosome behavior. Curr. Opin. Cell Biol.3, 59-66.

Rieder, C.L. and Alexander, S.P. (1990). Kinetochores are transportedpoleward along a single astral microtubule during chromosomeattachment to the spindle in newt lung cells. J. Cell Biol.110, 81-95.

Rime, H., Néant, I., Guerrier, P. and Ozon, R. (1989). 6-Dimethylaminopurine (6-DMAP), a reversible inhibitor of the transitionto metaphase during the first meiotic cell division of the mouse oocyte.Dev. Biol. 133, 169-179.

Vandré, D.D. and Wills, V.L. (1992). Inhibition of mitosis by okadaic acid:

possible involvement of a protein phosphatase 2A in the transition frommetaphase to anaphase. J. Cell Sci. 101, 79-91.

Verde, F., Dogterom, M., Stelzer, E., Karsenti, E. and Leibler, S. (1992).Control of microtubule dynamics and length by cyclin A- and cyclin B-dependent kinases in Xenopus egg extracts. J. Cell Biol. 118, 1097-1108.

Witt, P.L., Ris, H. and Borisy, G.G. (1980). Origin of kinetochoremicrotubules in Chinese hamster ovary cells. Chromosoma 81, 483-505.

Zinkowski, R.P., Meyne, J. and Brinkley, B.R. (1991). The centromere-kinetochore complex: a repeat subunit model. J. Cell Biol. 113, 1091-1110.

(Received 12 November 1992 - Accepted 5 January 1993)