Derangement of microtubule arrays in interphase and mitotic PtK2 cells

treated with deuterium oxide (heavy water)

JAN LAMPRECHTt, DIETER SCHROETER and NEIDHARD PAWELETZ*

German Cancer Research Center, Institute of Cell and Tumor Biology, Im Neuenheimer Feld 280, D-6900 Heidelberg, Germany

*Author for correspondencet Present address: Department of Histology and Embryology, School of Medicine, Chalubinskiego 5, 02-004 Warsaw, Poland

Summary

The extent and pattern of the rearrangements ofmicrotubule arrays in interphase and mitotic PtK2cells treated with deuterium oxide (2H2O) wereevaluated using light, immunofluorescence and elec-tron microscopy.

Combined labelling with anti-tubulin antibodiesand staining with a DNA-specific fluorochromerevealed that 2H2O influences the reassembly of thecytoplasmic microtubule complex (CMTC) of inter-phase cells after depolymerization of microtubules(MTs) with nocodazole. In cells entering mitosis inthe presence of 75% 2H2O the conversion of theCMTC into the mitotic spindle was affected, resultingin a retardation of the prophase/prometaphasetransition. (Pro)metaphase cells did not assemble aregular mitotic spindle and the metaphase/anaphasetransition was blocked. Immunofluorescence andultrastructural studies suggest that separation ofcentrosomes, nucleation of MTs around centro-

somes, organization of MTs into the mitotic spindle,as well as the ultrastructure and positioning of themitotic poles, are affected in deuterated PtK2 cells.In comparison with control cells, a significantlyhigher proportion of multdpolar divisions was foundafter stimulation of proliferation in the presence of25-50% 2H2O or during recovery after a long-termexposure to 75 % 2H2O.

On the basis of these results we discuss themechanism of the antimitotic action of deuteriumoxide and suggest that, apart from perturbation ofMT polymerization, it could also encompass disturb-ances in MT reorganization, most probably byimpairment of the microtubule-organizing centres(MTOC).

Key words: microtubules, microtubule organizing centres,mitosis, deuterium oxide, PtK2 cells.

Introduction

Deuterium oxide (2H2O, heavy water) arrests mitosis indiverse types of animal and plant cells (Burgess andNorthcote, 1969; Gross and Spindel, 1960a,b; Lavillaureix,1961; Leonard and Mullins, 1987; Marsland and Zimmer-man, 1963; Stein and Forrester, 1963; Takahashi and Sato,1984), but the underlying mechanism remains unclear. Onthe basis of experiments performed on sea-urchin eggs andgrasshopper spermatocytes, it has been suggested thatheavy water blocks mitosis either by immobilizing themitotic spindle (Gross and Spindel, 1960a,6; Marsland andZimmerman, 1963) or by enhancing the assembly ofspindle microtubules (MTs) (Sumitro et al. 1989). It isgenerally believed that a combination of the isotope effectsof deuterium and the solvent effects of heavy water(Thomson, 1963) influences tubulin turnover by increasingpolymerization of MTs (Olmsted and Borisy, 1973; Itohand Sato, 1984) by a stabilization of already existing MTs(Itoh and Sato, 1984) or by converting the unpolymerizabletubulin into the polymerizable form (Takahashi and Sato,1984). However, previous results have shown that, invertebrate somatic cells, 2H2O does not enhance theassembly of spindle MTs (Mclntosh, 1979; Leonard and

Journal of Cell Science 98, 463-473 (1991)Printed in Great Britain © The Company of Biologists Limited 1991

Mullins, 1987; Lamprecht et al. 1989). Moreover, asalready discussed by Itoh and Sato (1984), the 2H2O-induced increase in tubulin polymerization in vitro isobserved only when the microtubule-associated proteins(MAPs) are depleted. The presence of MAPs, what mimicsto a certain extend the conditions in vivo, inhibits bothpolymerization and depolymerization of brain tubulin(Houston et al. 1974).

Our earlier studies on the influence of 2H2O on mitosisin HeLa cells (Lamprecht et al. 1989) and PtKl cells(Lamprecht et al. 1990) showed that the presence of 75 %2H2O in the medium led not only to mitotic arrest at(pro)metaphase but also to retardation of the cell cycle.This was manifested by an abrupt decline in the number ofprophase cells. However, a second generation of mitoticcells, which crossed the G2/M boundary in the presence of75% 2H2O, appeared after 6-24 h of deuteration. Thepattern and extent of rearrangements of MT arrays inthese cells, as well as in interphase cells that weredepleted of MTs by nocodazole, were assessed during thisinvestigation. The results suggest that deuterationimpairs re-formation and interconversion of MT array,probably by influencing the microtubule-organizingcentres (MTOC).

463

Materials and methods

Cell cultureThe male rat kangaroo (Potorous tridactylis) kidney cell line PtK2(CCL66), originally obtained from the American Type CultureCollection (Rockville, Md), was used throughout this investi-gation. Stock cultures were maintained in 26 cm2 polystyrenetissue culture flasks (Corning) using minimal essential medium(MEM-Earle's) supplemented with 10 % (v/v) foetal calf serum(FCS), 1% non-essential amino acids, 1% L-glutamine, 1%sodium pyruvate and antibiotics (all from Biochrom, Berlin/Germany). The cells were subcultured for 24—48 h on glasscoverslips in plastic Petri dishes (60 mm) containing 4 mlMEM-Earle's supplemented as above, at 37 °C in a humidatmosphere of 5% CO2 prior to being used for experiments.Quiescent cultures were prepared by serum deprivation; the cellswere grown for 24 h in MEM-Earle's containing 10% FCS andwere transferred into MEM-Earle's supplemented with 0.1%FCS. After 7-14 days of cultivation the cells stopped dividingwhen the monolayers were about 80-90 % confluent.

Experimental treatmentsCell monolayers prepared in the above manner were placed eitherin fresh, complete MEM-Earle's (control groups) or in the mediain which 25 %, 50 % or 75 % of the 2H2O was substituted by 2H2O.To obtain 2H2O-supplemented media the concentrated (lOx)complete MEM-Earle's was diluted with Millipore-flltered heavywater (deuterium oxide, 99.95 %, Serva, Heidelberg/Germany; 3Hcontamination 74Bqml~1). Duration of the treatment varied asdescribed in the text. Since heavy water is considered to behygroscopic (Leonard and Mullins, 1987), in long-term exper-iments Petri dishes were sealed with Parafilm. To evaluate theinfluence of 2H2O on the polarity of cell division the monolayerswere: (1) cultivated in the presence of 75 % 2H2O for 1 or 2 weekswith a medium change every 3 days, then they were washed andrecovered in non-deuterated, complete medium, or (2) starved byserum deprivation, then re-fed with medium containing 10 % FCSand 25 % or 50 % 2H2O. To assess the influence of 2H2O on thereassembly of the cytoplasmic microtubule complex (CMTC) ininterphase cells, the monolayers were grown for 48 h inMEM-Earle's and then were treated for 2h with nocodazole(Janssen, Beerse, Belgium), a microtubule-depolymerizing agent.A stock solution of nocodazole in dimethyl sulphoxide (DMSO)was kept at -20°C and diluted with culture medium to a finalconcentration of 10/igmT1 (34/ai) just before use. Coverslipswere quickly rinsed in warm (37 °C) MEM-Earle's to removenocodazole, recovered in normal MEM-Earle's or in 75 % 2H2O-supplemented MEM-Earle's for 10, 20, 30 and 60 min, and thenprocessed for indirect immunofluorescence microscopy.

Mitotic indicesThe mitotic index (MI) has been expressed as the proportion ofpro-, meta-, ana- and telophases, and cytokinetic figures, per 100cells. To assess the influence of ^ j O on the incidence ofmultipolar mitoses, the percentage of multipolar anaphases,telophases and cytokineses was calculated (index of multipolarmitoses, IMM) per 100 anaphases, telophases and cytokineses.The mean value and standard deviation were calculated from sixmonolayers per control or per experimental group. The signifi-cance of the differences was tested by Student's i-test.

Immunofluorescence and electron microscopyIndirect immunofluorescence with monoclonal antibodies against/3-tubulin and electron microscopy were performed as describedpreviously (Lamprecht et al. 1989, 1990). Cell monolayersassigned for immunofluorescence were fixed in cold methanol,followed by cold acetone, then labelled with monoclonal anti-^-tubulin antibodies and rhodamine-conjugated anti-mouseIgG+IgM, and counterstained with Hoechst 33342 to identify thestage of mitosis. Cell monolayers assigned for electron microscopywere fixed with glutaraldehyde and osmium tetroxide containingK3Fe(CN)e- Following staining en bloc with uranyl acetate,

monolayers were embedded in Epoxy resin for ultrathin section-ing and postetained with uranyl acetate and lead citrate.

Results

2B.2O modifies the pattern of CMTC reassemblyfollowing treatment with nocodazoleThe dynamics of reassembly of the CMTC in the absenceand presence of heavy water after depolymerization withnocodazole were evaluated in interphase cells by immuno-fluorescence microscopy (Fig. 1 A-G). Two-hour incu-bation of cells in medium supplemented with 34 UMnocodazole resulted in a total depletion of the CMTC(Tig. IB). In most cells, there was a diffuse fluorescentbackground in the cytoplasm, which is characteristic ofdepolymerized tubulin. In addition, a single bright patchof fluorescence, usually in close proximity to the nuclearsurface, was observed that corresponded morphologicallyto the centrosome. After 10 min of recovery in MEM-Earle's, the cytoplasm was filled with a dense meshwork ofshort, randomly orientated MTs and a single small aster ofradiating MTs was seen at the juxtanuclear position(Fig. 1C). The typical, elaborate array of the CMTC wasrestored after 20-30 min of recovery (Fig. ID).

Reassembly of the CMTC after nocodazole treatmentwas different in cells recovering in the presence of 75 %2H2O. After 10 min of recovery immunostaining revealednumerous bright foci embedded in a mesh of short fibrils(Fig. IE). The foci displayed a clear tendency for acircumnuclear arrangement. After 20 min of recovery inthe presence of 2H2O, the foci largely vanished and thecytoplasm contained^ uniformly packed dense meshworkof short, disordered MTs (Fig. IF). The first signs of aradial regrowth of MT bundles around a single focus wasdetected after 30 min of recovery in the presence of 2H2O,and by 60 min the typical CMTC of interphase cells wasrestored (Fig. 1 G).

2B.2O disturbs the conversion of the CMTC into themitotic spindle and the development of the mitoticspindleThe conversion of the CMTC into the mitotic spindle aswell as the development of the mitotic spindle wereevaluated in consecutive mitotic phases of control cellsand cells grown in medium containing 75 % 2H2O.

In interphase cells we were unable to define anydifferences in spatial organization of the CMTC whencomparing control and 2H2O-treated cell monolayers(Figs 3 and 4, below). The dynamics of the development ofthe mitotic apparatus in conventionally grown PtKl andPtK2 cells has already been described in detail (Brinkleyet al. 1976; De Brabander et al. 1980, 1986; Vandre andBorisy, 1986; Mclntosh, 1989). In control cells, at prophase(Fig. 2A-B) two distinct asters (MT-nucleating centres)superimposed on the CMTC are seen. At prometaphase(Fig. 2C-D) cytoplasmic MTs undergo depolymerization,evoking a faint background fluorescence and the two half-spindles begin to converge into an elongated bipolarspindle. The mitotic poles are visible as bright fluorescentspots. A regular, compact spindle is established atmetaphase when the chromosome-to-pole MTs becomeshorter (Fig. 2E-F). The asters become more prominentduring anaphase (Fig. 2G-H) with elongated MTs point-ing toward the cell periphery. Numerous cytoplasmic MTsforming a new CMTC reappear at telophase (Fig. 2 I-J).Parallel bundles of the interzonal MTs in the pre-midbody

464 J. Lamprecht et al.

Fig. 1. Reconstruction of the CMTC after depolymerization with nocodazole; immunostaining for tubulin. (A) Control PtK2 cells.(B) Cells treated for 2 h with 34 ^M nocodazole; note immunostaining of centrosomes (arrows) at the juxtanuclear position. Recoveryof cells in MEM-Earle's medium for lOmin (C) and 20min (D). Recovery in 75% 2H2O-supplemented medium for lOmin (E),20min (F) and 60min (G); note multiple MT-nucleating centres after lOmin recovery in 2H2O.x750.

Influence of2H2O upon microtubule arrays 465

Fig. 2. Control mitotic PtK2 cells double-stained for DNA with Hoechst 33342 (A,C,E,G,I,K) and for tubulin with anti-beta-tubulinantibodies followed by rhodamine-conjugated IgG/IgM (B,D,F,H,J,L)- Prophase (A-B), prometaphase (C-D), metaphase (E-F),anaphase (G-H), telophase (I-J) and cytokinesis (K-L).x750.

466 J. Lamprecht et al.

region become compressed by a cytokinetic furrow. Duringcytokinesis (Fig. 2K-L) the daughter cells are conjoinedby a midbody and MTs of the CMTC are distributedthroughout the cytoplasm.

Obvious changes in MT reorganization were observed incells entering prophase after exposure to 75% 2H2O(Figs 2B and 3B,D). Immunostaining usually revealed asingle, small aster of radiating MTs against the back-ground of diffusely stained cytoplasm (Fig. 3B). In someprophase cells two such asters, located close to the nucleus,could be detected (see below). Often a 2H2O-treatedprophase cell was completely devoid of polymerized MTsand the immunostaining was restricted to a tiny, brightspot located at a juxtanuclear position against thebackground of faint uniform cytoplasmic fluorescence(Fig. 3 D).

We evaluated the incidence of single or paired MT-nucleating centres in prophase or prometaphase cells inthe control and experimental groups. In control cells, twoMT-nucleating centres were usually observed, while thegreat majority of 2H2O-treated cells displayed only oneMT-nucleating centre (Table 1).

Virtually all cells entering prometaphase after incu-bation for 6-24 h in the presence of 2H2O displayedirregular masses of polymerized tubulin, located at thecentre of a fan-shaped array of loosely clustered chromo-somes. In some prometaphase and metaphase cells, twoseparate tubulin-positive complexes could be recognized(Figs 3H and 4B). In the 2H2O-treated monolayers no(pro)metaphase cells were found that were completelydevoid of polymerized tubulin or displayed a regularmitotic spindle.

Mitotic cells are unable to progress beyond metaphasein the presence of high concentrations of 2H2O. Chromo-somes are re-enveloped in a new nuclear membrane andeither a cell with single reconstituted nucleus, or amultinucleate/micronucleate cell, arises as a consequenceof a mitotic block (Leonard and Mullins, 1987; Lamprechtet al. 1989, 1990). In cells blocked at (pro)metaphase thecompact chromatin masses that are present suggest thatthe reconstruction of the nuclear envelope has started (seeLamprecht et al. 1989, 1990). In some blocked cellsnumerous miniature aster-like structures, fortuitouslyscattered throughout the cytoplasm, were seen (Fig. 4A-B). At more advanced stages of nuclear envelopereassembly, as detected by chromatin decondensation, thecytoplasmic asters were more prominent and their raysmingled with those of neighbouring asters (Fig. 4C-D), oran irregular array of MT bundles was seen (Fig. 4E-F).

2H20 influences the ultrastructure of the centrosomePrevious analysis of PtKl cells entering prophase in thepresence of 75 % 2H2O using light and electron microscopy(Lamprecht et al. 1990) revealed an asynchrony inchromosome condensation and nuclear envelope break-

down, which resulted in fully condensed chromosomesenwrapped in a continuous nuclear envelope. In these'overmature' prophase cells the polymerized tubulin wasabsent or present in only low amounts (Fig. 3A-D).Therefore, we wished to examine more closely thecentrosomal regions of these cells. Electron microscopydisclosed a pair of unseparated or adjacent centrosomesusually located close to a nucleus (Fig. 5) and thecentrosome was composed of two centrioles in an orthogo-nal configuration with a surrounding cloud of osmiophilicpericentriolar material (Fig. 5, inset). Around the centro-some, single MTs were arranged in an irregular fashionand some MTs seemed to be anchored to the pericentriolarcloud.

At the ultrastructural level, the cells blocked at(pro)metaphase displayed an aberrant mitotic spindlewith irregularly arranged MTs and with mitotic poleslocated close to the congregated chromosomes. Analysis ofserially sectioned, abortive (pro)metaphase cells disclosedthat no more than two centrosomes with unseparatedcentrioles could be found in the mitotic spindle area (datanot shown). In many 2H2O-blocked cells the osmiophilicpericentriolar material formed small, separate clumpsaround which MTs radiated in all directions (Fig. 6A-C).

2H2O induces multipolar mitosesThe induction of multipolar mitoses by 2H2O wasevaluated in two different experiments. In the first, themonolayers were incubated for either one or two weeks ina medium supplemented with 75 % 2H2O. This treatmentabolished cell division almost completely (MI<0.01%).After recovery for 5-7 days in fresh, non-deuteratedmedium, a resumption of mitotic activity was observed.The control groups consisted of monolayers incubated forone or two weeks in non-deuterated medium, and re-fed asthe experimental groups after one or two weeks. The meanpercentages of multipolar divisions (TMMs) were 22.0%(S.D.=5.9%) and 29.3% (S.D.=9.1%) after one and twoweeks of deuteration, respectively. The differences in themeans between IMMs of experimental groups and controlgroups (*= 1.0 %, S.D.=0.6 % and x=1.3 %, S.D.=0.9 %) werestatistically significant (<r=0.01) as assessed by the £-test.

In the second experiment, the cells were made quiescent(blocked at G0/Gi) by serum deprivation (MI<0.05 %) andwere returned to the cell cycle by re-feeding with a fresh,complete medium containing 10 % FCS and 25 % or 50 %2H20. As 25 %-50 % concentrations of 2H2O are permiss-ive for both the initiation and progression of mitoticdivision in HeLa and PtK cells (Leonard and Mullins,1987; Lamprecht et al. 1989, 1990), an analysis of theinduction of multipolar mitoses in the presence of 2H2Oduring the first wave of cell divisions after starvation waspossible. The first wave of mitoses in cells grown in thepresence of 25 % 2H2O appeared 48 h after re-feeding withcomplete medium and the IMM was 3.2% (S.D.=1.0%). In

Table 1. The proportions of prophase and prometaphase cells with single or paired MT-nucleating centres(centrosomes) in control and 75% H20-treated monolayers of PtK2 cells

N, number of cellB.

ProphasePrometaphase

N

1410

Control

Single

20

Nucleatingcells

Paired

1210

centres in:

N

2827

2H2O-treated cells

Single

2625

Paired

32

Influence of2H2O upon microtubule arrays 467

Fig. 3. PtK2 cells entering mitosis after incubation for 6-24 h in the presence of 75% 2H2O. Staining for DNA (A,C,E,G) andimmunolabelling for tubulin (B,D,F,H)- 'Overmature' prophase cells (A-D); fully condensed chromosomes arranged in a roundedloose cluster. The cells display background cytoplasmic fluorescence of depolymerized MTs and single inconspicuous aster (B) or abright spot (centrosome), not nucleating MTs (D, arrow). Prometaphase cell (E and F); a fan-shaped pattern of chromosomecongression. Immunostaining for tubulin reveals scanty, irregular bundles of polymerized tubulin located close to chromosomes.Metaphase cell (G and H); chromosomes congregated to the metaphase plate. Inconspicuous immuno-positive material is locatedclose to, and on both sides of, the metaphase plate. X750.

468 J. Lamprecht et al.

Fig. 4. PtK2 cells entering mitosis during incubation for 6-24h in medium supplemented with 75% 2H2O. DNA staining (A,C,E);anti-tubulin immunostaining (B,D,F). (A and B) A blocked metaphase cell with an abortive spindle and numerous small MT-nucleating centres distributed throughout cytoplasm. (C and D) A cell with the reconstituted nucleus displaying early signs ofchromatin decondensation and two lagging chromosomes (C, arrows). In the cytoplasm, apart from the bright irregularimmunofluorescent masses overlying the reconstituted nucleus, many MT-nucleating centres with overlapping radiations could beseen (D, arrows). (E and F) A multinucleate/micronucleate cell with decondensed chromatin; cytoplasmic MTs form an irregulararray that is indistinguishable from the CMTC of control interphase cells. x750.

Influence of2H2O upon microtubule arrays 469

comparison, the control group had an EMM of 0.6%(S.D.=0.4%). After incubation in the medium containing50% 2H2O, the resumption of mitotic acitivity wasobserved after 4 days and the EMM was 6.7% (S.D.=0.6%).

The control group had an EMM of 2.2% (S.D. = 1.0%). Atboth concentrations the differences between the means ofcontrol and experimental groups are statistically signifi-cant (cr=0.01 %).

" Jf

ne

* N^Tfc

«f

V TV

Fig. 5. Ultrastructure of an 'overmature' prophase PtK2 cell entering mitosis during incubation for 12 h in 75 % 2H2O-supplemented medium, ne, nuclear envelope; c, split but unseparated centrosomes; ch, chromosomes. X8600. Inset: highermagnification of the centrosome region. Note scanty, irregularly distributed MTs. Some of the MTs appear to converge at thepericentriolar cloud, x 20 000.

Fig. 6. (A-C) Ultrastructural appearance of the centrosomal region in three different PtK2 cells blocked in (pro)metaphase byincubation for 6 h in 75% 2H2O. The centrosome (mitotic pole) is positioned close to the chromosomes. The pericentriolar cloud isirregular in outline, forming discrete patches (arrows) about which MTs radiate in an irregular fashion, x 21000.

470 J. Lamprecht et al.

Discussion

It is well established that the number, length, polarity andspatial distribution of cytoplasmic MTs and the intercon-version of the microtubular assemblages, are governed bydiscrete foci termed the microtubule-organizing centres(MTOCs). The primary MTOC of interphase cells is thecentrosome, while in mitotic cells the spindle MTs areorganized by kinetochores and mitotic poles. However,little is known about the molecular composition of theMTOCs and the mechanisms controlling their activities(for reviews see Brinkley, 1985; Salmon, 1989).

The present experiments were undertaken with themain goal of determining the MTOCs activities ininterphase and mitotic cells grown in the presence ofdeuterium oxide. Under the conditions of our experiments,both polymerization and depolymerization of MTs as wellas reassembly of the CMTC progressed in 2H2O-treatedcells, but the development of the mitotic spindle wasprofoundly altered. The results obtained suggest that,apart from the postulated influence of 2H2O on thedynamic equilibrium of polymerized/depolymerized tubu-lin, an alternative mechanism involving the impairmentof activity of MTOCs should be considered.

In interphase PtK2 cells, the number of MTs and theordered assembly of the CMTC appeared unchanged aftertreatment with 75 % 2H2O for hours (Figs 3 and 4), orweeks (data not shown). In cells recovered in controlmedium, after depolymerization of MTs with nocodazole,the reconstruction of CMTC always started at a singlenucleating centre, the centrosome (De Brabander et al.1981a,6,1986; and Fig. 1C-D). Reconstruction of CMTC incells grown in medium containing 2H2O was only slightlydelayed, but the pattern of the reassembled CMTC wasnoticeably modified. Formation of the multiple MT-nucleating centres around the nuclear envelope couldresult from dispersion of a compact centrosome intomultiple assembly sites or decreased activity of acentrosome as a major MTOC. Disappearance of themultiple MT-nucleating centres with concomitant conser-vation of the cytoplasmic microtubular network 20minafter removal of nocodazole, and a rapid re-establishmentof the typical CMTC in the presence of 2H2O, favour theview that the centrosome was not dispersed. Mostprobably, the centrosome no longer served as the majorMTOC, as a result of a generalized reduction in the criticaltubulin concentration. A reduction in the critical tubulinconcentration has been shown to induce spontaneousnucleation of MTs (De Brabander et al. 19816). In this case,the nuclear envelope could compete successfully with acentrosome for tubulin subunits, playing the role of themajor microtubule organizer of the CMTC (Wick et al.1981; Tassin et al. 1985; Kronebusch and Singer, 1987).Similar formation of multiple cytasters, but dispersedrandomly in the cytoplasm, has been observed in amphib-ian (Van Assel and Brachet, 1966; Karsenti et al. 1984) andsea-urchin eggs (Mazia, 1978; Kuriyama and Borisy, 1983;Kallenbach, 1985) treated with heavy water.

During normal prophase and prometaphase, a progress-ive enhancement of the MT-nucleating activity of thealready doubled centrosomes with an accompanyingdepolymerization of cytoplasmic MTs takes place (Kur-iyama and Borisy, 1981). This culminates in conversion ofthe CMTC into the mitotic spindle (Brinkley et al. 1976; DeBrabander et al. 1986; Vandre and Borisy, 1986; Mclntosh,1989; and Fig. 2A-F). Surprisingly, in many PtK2 cells

entering mitosis in the presence of 75 % 2H2O, neither theCMTC nor the mitotic spindle was observed. It is obviousthat the CMTC has been dismantled, but not replaced bythe mitotic spindle. Therefore, it may be concluded thatthe ability of centrosomes to nucleate the astral (mitotcspindle) MTs was either greatly diminished (Fig. 3 A-B)or completely suppressed (Fig. 3 C-D). This observationsupports the conclusion of Sumitro and Sato (1989) that2H2O does not prevent disassembly of MTs, but itcontradicts the view that deuteration favours polymeriz-ation of spindle MTs.

Although it is well recognized that separation ofcentrosomes in PtK cells is often postponed until prometa-phase (Eattner and Berns, 1976; Aubin et al. 1980), anassessment of the proportions of prophase and prometa-phase cells displaying single or paired MT-nucleatingcentres (Table 1) supports the notion that the process ofcentrosome separation was delayed or persistentlyimpaired in 2H2O-treated cells. The ultrastructural analy-sis lends further support to the above conclusion showingthat in the 'overmature' prophase cells centrosomes aresplit but unseparated or positioned close to each other,with only a moderate number of MTs nucleated at thepericentriolar cloud. It seems therefore, that 2H2O, likediazepam (Andersson et al. 1981), inhibits separation ofthe centrosomes at the prophase/prometaphase tran-sition. If the separation of centrosomes depends on theformation of the intercentrosomal MTs, as suggested byRattner and Berns (1976), it can be concluded thatdeuteration either inhibits MT polymerization, or sup-presses the MT-nucleating activity of centrosomes. How-ever, the mechanism responsible for centrosome separ-ation seems to be more complex. Results have beenpublished (Gotlieb et al. 1983; Euteneuer and Schliwa,1985; Schatten et al. 1988; Buendia et al. 1990) thatsuggest that, apart from MTs, an intact actin-basedcontractile network is also required. In this context it isinteresting to note that 2H2O-induced alterations in thepolymerization and organization of the F-actin networkhave recently been reported (Zimmermann et al. 1988).

A metaphase cell is unable to develop a regular mitoticspindle in the presence of 75 % 2H2O. Immunofiuorescencehas shown that in cells blocked at (pro)metaphase, a small,abortive spindle is assembled, and that these cells are notdevoid of polymerized tubulin, as was observed in the'overmature' prophase cells. This could mean that kineto-chores, which are accessible after nuclear envelopebreakdown and anchoring MTs in the presence of heavywater (Lamprecht et al. 1990), may nucleate new MTs (DeBrabander et al. 1981a,6) or may assist in the stabilizationof existing MTs. Alternatively, polymerizaton of MTs maybe potentiated in the vicinity of condensed chromosomes,as suggested by Karsenti et al. (1984) and Nicklas andGordon (1985).

In the conventionally grown PtK2 cells at lateanaphase/early telophase the conversion of the mitoticspindle into the CMTC begins with elongation of the astralMTs that focus on centrosomes (Fig. 2G-H). Conversely,in the presence of 2H2O, the re-formation of the nuclearenvelope around reconstituted nuclei or around multiplekaryomeres is accompanied by the reappearance ofcytoplasmic MTs in the form of multiple small asters(Fig. 4 A-B). This indicates that in the presence of 75 %2H2O the CMTC is reassembled either on discrete,dissociated patches of the centrosomal material, orspontaneously, as a consequence of a reduction in the

Influence of2H2O upon microtubule arrays 471

critical tubulin concentration and diminished MT-nucleat-ing acitivity of centrosomes.

The high proportion of multipolar divisions observedafter prolonged treatment with 75 % 2H2O probably arisesfrom at least two different processes. Incubation in 75 %2H2O produces some polyploid cells, as a result of nuclearreconstruction in the blocked (pro)metaphase cells. Thesecells may form multipolar spindles during the next mitosisas doubled centrosomes are not partitioned betweendaughter cells. However, as shown by the accompanyingexperiment with the starved PtK2 cells, 2H2O inducesmultipolar divisions in cells with a normal amount ofcentrosomal material. Starvation arrests cells at theGi/Go phases of the cell cycle, i.e. quiescent cells possessonly one centrosome (Tucker et al. 1979). Re-feeding withmedium containing 10 % FCS and 25-50 % 2H2O inducesre-entry of cells into the cell cycle and the formation of astatistically significant increase in the number of multi-polar divisions. This raises the possibility that deuterationimpairs bipolarization of cell division by influencing thecentrosomal structure, most probably through the disper-sion of the pericentriolar material as observed aftercolchicine treatment (Sellitto and Kuriyama, 1988), or byprevention of recruitment of cytoplasmic componentscontaining MT-organizing activity (Leslie, 1990).

The results presented in this paper are inconsistent withthe postulated mechanism of the antimitotic acitivity ofdeuterium oxide (Gross and Spindel, 19606; Marsland andZimmerman, 1963; Sumitro et al. 1989). The phenomenaobserved cannot be explained sufficiently by an over-stabilization of the mitotic apparatus and enhancedpolymerization of MTs. Incorporation of deuterium intobiopolymers and intracellular water is rapid and pro-portional to the concentration of 2H2O in the medium(Unno et al. 1988). Most of the incorporated deuterium isnonexchangeable (Unno et al. 1988), inducing confor-mational and functional modifications to diverse macro-molecules (Thomson, 1963; Lemm and Wenzel, 1981).Because tubulin accounts for only 20 % of the total proteinin a mitotic spindle (Salmon, 1982), and because themaximum anti-MT effect is achieved at 45 % 2H2O, whichpermits unperturbed mitosis in marine eggs flnoue' andSato, 1967), it can be assumed that the observedphenomena cannot arise from the direct influence of 2H2Oon MT polymerization. Complex modifications of diversespindle constituents or factors governing the formation ofthe mitotic spindle, like the microtubule-associated pro-teins (MAPs) (Sloboda et al. 1976) or cdc2 kinase (Verde etal. 1990), could be also involved. Although not proven byour results, it seems reasonable to speculate that, at leastin somatic vertebrate cells grown in vitro, the antimitoticeffect of deuterium oxide is in part the result of influences:(1) on the centrosomal integrity; (2) on the MT-nucleating,MT-stabilizing and MT-organizing activities of the centro-some; (3) on its migratory behaviour and; (4) on the bipolarpartitioning of the centrosomal components between themitotic poles. However, the question still remains open asto which of the observed perturbations (and to whatextent) result from a direct action of 2H2O on thecentrosome components.

We are indebted to Dr C. S. Potten and Dr C. de Vac for criticalreading of the manuscript. We thank Mrs U.-L. Kiesewetter forher expert technical assistance and Mr G.E. Withers Jr for hisphotographic work and Ms C. Kamp for typing the manuscript. J.Lamprecht acknowledges a research grant for these studies fromthe German Cancer Research Center.

References

ANDERSSON, L. C, LETHO, V.-P., STENMAN, S , BADLEY, R. A. ANDVIRTANEN, I. (1981). Diazepam induces mitotic arrest at prometaphaseby inhibiting centriolar separation. Nature 291, 247-248.

AUBIN, J E., OSBORN, M. AND WEBER, K. (1980). Variations in thedistribution and migration of centriole duplexes in mitotic PtK2 cellsstudied by immunofluorescence microscopy. J. Cell Sci. 43, 177-194.

BRINKLBY, B. R. (1985). Microtubule organizing centers. A. Rev. CellBiol. 1, 146-172.

BRINKLEY, B. R., FULLER, G. M. AND HIGHFIELD, D. P. (1976). Tubulinantibodies as probes for microtubules in dividing and nondividingmammalian cells. Cold Spring Harb. Symp. quant. Biol. 3, 435-456.

BUENDIA, B., BRE, M.-H., GRIFFITHS, G. AND KARSENTI, E. (1990).Cytoskeletal control of centrioles movement during the establishmentof polarity in Madin-Darby canine kidney cells. J. Cell Biol. 110,1123-1135.

BURGESS, J. AND NORTHCOTE, D. H. (1969) Action of colchicine andheavy water on the polymerization of microtubules in wheat rootmeristem. J. Cell Sci. 5, 433-451.

DE BRABANDER, M., GEUENS, G , DE MEY, J. AND JONIAN, M. (1981a).Nucleated assembly of microtubules in living PtK2 cells after releasefrom nocodazole treatment. Cell Motil 1, 469-483.

DE BRABANDER, M., GEUENS, G., NUYDENS, R., WILLEBRORDS, R., AERTS,F., DE MBY, J. AND MCINTOSH, J. R (1986). Microtubule dynamicsduring the cell cycle: The effects of taxol and nocodazole on themicrotubule system of PtK2 cells at different stages of the mitoticcycle. Int Rev. Cytol. 101, 215-274.

DE BRABANDBR, M., GEUBNS, G., NUYDENS, R., WILLEBRORDS, R AND DEMEY, J. (1980). The microtubule nucleating and organizing activity ofkinetochores and centrosomes in living PtK2-cells. In Microtubulesand Microtubule Inhibitors 1980 (ed. M. De Brabander and J. DeMey), pp. 255-268. Amsterdam: Elsevier/North Holland BiomedicalPress.

DE BRABANDER, M., GEUBNS, G., NUYDENS, R., WILLEBRORDS, R. AND DEMEY, J. (19816). Microtubule assembly in living cells after releasefrom nocodazole block: The effects of metabolic inhibitors, taxol andpH Cell Biol. Int. Rep. 5, 913-920.

EUTENEUER, U. AND SCHUWA, M. (1985). Evidence for an involvement ofactin in the positioning and motility of centrosomes. J. Cell Biol. 101,96-103.

GOTIJEB, A . I., SUBRAHMANYAN, L. AND K A L N I N S , V. (1983) .Microtubule-organizing centers and cell migration: Effect of inhibitionof migration and microtubule disruption in endothelial cells. J. CellBiol 96, 1266-1272.

GROSS, P. R. AND SPINDEL, W. (1960a). Mitotic arrest by deuteriumoxide. Science 131, 37-39.

GROSS, P. R. AND SPINDEL, W. (19606). Heavy water inhibition of celldivision: An approach to mechanism. Ann. N.Y. Acad. Sci. 90,500-522.

HOUSTON, L. L., ODELL, J., LEE, Y. C. AND HIMBS, R. H. (1974). Solventisotope effect on microtubule polymerization and depolymerizationJ. molec. Biol. 87, 141-146.

INOUS, S. AND SATO, H (1967). Cell motility by labile association ofmolecules. The nature of mitotic spindle fibers and their role inchromosome movement. J. gen. Physiol. SO, 269-292.

ITOH, T. J. AND SATO, H. (1984). The effects of deuterium oxide (̂ HaO)on the polymerization of tubulin in uitro. Biochim. biophys. Ada 800,21-27.

KALLENBACH, R. J. (1985). Ultrastructural analysis of the initiation anddevelopment of cytasters in sea-urchin eggs. J. Cell Sci. 73, 261-278

KABSENTI, E., NEWPORT, J., HUBBLE, R. AND KIRSCHNER, M. (1984).Interconversion of metaphase and interphase microtubule arrays, asstudied by the injection of centrosomes and nuclei into Xenopus eggs.J. Cell Biol. 98, 1730-1745.

KHONEBUSCH, P. J. AND SINGER, S. J. (1987). The microtubule-organizingcomplex and the Golgi apparatus are co-localized around the entirenuclear envelope of interphase cardiac myocytes. J. Cell Sci. 88,25-34.

KURIYAMA, R. AND BORISY, G. G. (1981). Microtubule-nucleating activityof centrosomes in Chinese hamster ovary cells is independent of thecentriole cycle but coupled to the mitotic cycle. J. Cell Biol. 91,822-826.

KURIYAMA, R. AND BORISY, G. G. (1983). Cytasters induced withinunfertilized sea-urchin eggs. J. Cell Sci. 61, 175-189.

LAMPRBCHT, J., SCHROETER, D. AND PAWELETZ, N. (1989).Disorganization of mitosis in HeLa cells by deuterium oxide. Ear.J. Cell Bwl 50, 360-369.

LAMPRECHT, J., SCHROETER, D. AND PAWELETZ, N. (1990). Mitosisarrested by deuterium oxide. Light microscopic, immunofluorescenceand ultrastructural characterization. Eur. J. Cell Biol. 51, 303-312.

LAVTLLAURBIX, J. (1961). Action de l'eau lourde sur la division nucleairede cellules en culture. Cr. hebd. Seanc. Acad. Set., Paris 252,622-623.

472 J. Lamprecht et al.

LEMM, U. AND WENZEL, M. (1981). Stabilization of enzymes and antiseraby heavy water Eur. J. Biochem. 116, 441-446.

LEONARD, P. J. AND MULLINS, J. M. (1987). D20 induced alterations ofmitosis in PtKl cells. Expl Cell Res. 172, 204-211.

LESLIE, R. J. (1990). Recruitment: The ins and outs of spindle poleformation. Cell Motil. Cytoskel. 16, 225-228.

MARSLAND, D. AND ZIMMERMAN, A. M. (1963). Cell division: Differentialeffects of heavy water upon the mechanisms of cytokinesis andkaryokinesis in the eggs of Arbacia punctulata. Bxpl Cell Res. 30,23-35.

MAZIA, D. (1978). Origin of twoneas in cell reproduction. In CellReproduction: In honor of Daniel Mazia (ed. E. R. Dirksen, D. M.Prescott and C. F. Fox), pp. 1-14. New York, Academic Press.

MCINTOSH, J. R. (1979). Cell division. In Microtubules (ed. K. Robertsand J. S. Hyams), pp. 382-441. London, Academic Press.

MCINTOSH, J. R. (1989). Assembly and disassembly of mitotic spindlemicrotubules. In Cell Movement (ed. F. D. Warner and J. R.McIntosh), pp. 403-419. New York, Alan R. Liss, Inc.

NICKLAS, R. B. AND GORDON, G. W. (1985). The total length of spindlemicrotubules depends on the number of chromosomes present. J. CellBiol. 100, 1-7.

OLMSTKD, J. B. AND BORISY, G. G. (1973). Characterization ofmicrotubule assembly in porcine brain extracts by viscometry.Biochemistry 12, 4282-4289.

RATTNER, J. B. AND BERNS, M. W. (1976). Centriole behavior in earlymitosis of rat kangaroo cells (PtK2). Chromosoma 54, 387-395.

SALMON, E. D. (1982). Mitotic spindles isolated from sea urchin eggswith EGTA lysis buffers. Meth. Cell Biol. 25, 69-105.

SALMON, E. D. (1989). Microtubule dynamics and chromosomemovement. In Mitosis: Molecules and Mechanisms (ed. J. S. Hyamsand B. R. Brinkley), pp. 119-181. London: Academic Press.

SCHATTEN, H., WALTER, M., BIESBMANN, H. AND SCHATTEN, G. (1988).Microtubules are required for centrosome expansion and positioningwhile micronlaments are required for centrosome separation in seaurchin eggs during fertilization and mitosis. Cell Motil. Cytoskel. 11,248-259.

SELLITTO, C AND KURIYAMA, R. (1988). Distribution of pericentriolarmaterial in multipolar spindles induced by colcemid treatment inChinese hamster ovary cells. J. Cell Sci. 89, 57-65.

SLOBODA, R. D., DENTLER, W. L. AND ROSENBAUM, J. L. (1976).Microtubule-associated proteins and the stimulation of tubulinassembly in vitro. Biochemistry 15, 4497—4505.

STEIN, O. L. AND FORRESTER, G. M. (1963). The effect of highconcentrations of heavy water on root development in Pisum sativumand Zea mays. Plania 60, 349-359.

SUMITRO, S. B., IZUTSU, K. AND SATO, H. (1989). Effects of D2O on themovement of chromosomes and the shortening of kinetochore spindlefibers in anaphase in dividing spermatocytes of the grasshopper,Mongolotettix japonicus. Cell Struct. Funct. 14, 345-352.

SUMTTRO, S. B. AND SATO, H. (1989). The isotopic effect of D2O indeveloping sea urchin eggs. Cell Struct. Funct. 14, 96-111.

TAKAHASHI, T. C. AND SATO, H. (1984). Yields of tubulin paracrystals,vinblastine-crystals, induced in unfertilized and fertilized sea urchineggs in the presence of D2O. Cell Struct. Funct. 9, 45-52.

TASSIN, A.-M., MARO, B. AND BORNENS, M. (1985). Fate of microtubule-organizing centers during myogenesis in vitro. J. Cell Biol. 100,35-46

THOMSON, J. F. (1963). Biological Effects of Deuterium. Oxford:Pergamon Press.

TUCKER R. W., PABDEE, A. B. AND FUJIWARA, K. (1979). Centrioleciliation is related to quiescence and DNA synthesis in 3T3 cells. Cell17, 527-535.

UNNO, K., BUSUJIMA, H., SHIMBA, S., NABJTA, K. AND OKADA, S. (1988).Characteristics of growth and deuterium incorporation in Chlorellaellipsoidea grown in deuterium oxide. Chem. pharm. Bull., Tokyo 36,1828-1833.

VAN ASSEL, S. AND BRACHET, J. (1966). Formation de cytasters dans lesoeufs de batraciens sous Faction de l'eau lourde. J. Embryol. exp.Morph. 15, 143-151.

VANDRE, D. D AND BORISY, G. G. (1986) The lnterphase-mitosistransformation of the microtubule network in mammalian cells InCell Motility: Mechanism and Regulation (ed. H. Ishikawa, S. Hatanoand H. Sato), pp. 389-401. New York: Alan R. Liss, Inc.

VERDE, F., LABBE, J.-J., DOREE, M. AND KARSKNTI, E. (1990). Regulationof microtubule dynamics by cdc2 protein kinase in cell-free extracts ofXenopus eggs. Nature 343, 233-238.

WICK, S. M., SEAGULL, R. W., OSBORN, M., WEBER, K. AND GUNNING, B.E. S. (1981). Immunofluorescence microscopy of organized microtubulearrays in structurally-stabilized meristematic plant cells. J. Cell Biol.89, 685-690.

ZIMMEHMANN, A., KELLER, H.-U. AND COTTIER, H. (1988). Heavy water(DjOMnduced shape changes, movements and F-actin redistributionin human neutrophil granulocytes. Eur. J. Cell Biol. 47, 320-326.

(Received 19 June 1990 - Accepted, in revised form, 14 January 1991)

Influence of2H2O upon microtubule arrays 473