mpm-2 antibody-reactive phosphorylations can be created in … · 2001-05-03 · mpm-2 antibody in...

2013Journal of Cell Science 110, 2013-2025 (1997)Printed in Great Britain © The Company of Biologists Limited 1997JCS1412

MPM-2 antibody-reactive phosphorylations can be created in detergent-

extracted cells by kinetochore-bound and soluble kinases

L. Renzi1, M. S. Gersch1, M. S. Campbell1,*, L. Wu2, S. A. Osmani2 and G. J. Gorbsky1,†

1Department of Cell Biology, Health Sciences Center, University of Virginia, Charlottesville, VA 22908, USA2Weis Center for Research, Geisinger Clinic, 100 North Academy Avenue, Danville, PA 17822, USA

*Present address: Department of Cell and Molecular Biology, Dana-Farber Cancer Institute, Boston, MA, USA†Author for correspondence (e-mail: [email protected])

The MPM-2 antibody labels mitosis-specific and cell cycle-regulated phosphoproteins. The major phosphoproteins ofmitotic chromosomes recognized by the MPM-2 antibodyare DNA topoisomerase II (topoII) αα and ββ.. In immuno-fluorescence studies of PtK1 cytoskeletons, prepared bydetergent lysis in the presence of potent phosphataseinhibitors, the MPM-2 antibody labels phosphoproteinsfound at kinetochores, chromosome arms, midbody andspindle poles of mitotic cells. In cells extracted withoutphosphatase inhibitors, labeling of the MPM-2 antibodiesat kinetochores is greatly diminished. However, incytoskeletons this epitope can be regenerated through theaction of kinases stably bound at the kinetochore. Variouskinase inhibitors were tested in order to characterize theendogenous kinase responsible for these phosphorylations.We found that the MPM-2 epitope will not rephosphory-late in the presence of the broad specificity kinase inhibitorsK-252a, staurosporine and 2-aminopurine. Several otherinhibitors had no effect on the rephosphorylation indicat-ing that the endogenous MPM-2 kinase at kinetochores isnot p34cdc2, casein kinase II, MAP kinase, protein kinase Aor protein kinase C. The addition of N-ethylmaleimide

inactivated the endogenous kinetochore kinase; thisallowed testing of several purified kinases in the kineto-chore rephosphorylation assay. Active p34cdc2-cyclin B,casein kinase II and MAP kinase could not generate theMPM-2 phosophoepitope. However, bacterially expressedNIMA from Aspergillus and ultracentrifuged mitotic HeLacell extract were able to catalyze the rephosphorylation ofthe MPM-2 epitope at kinetochores. Furthermore, frac-tionation of mitotic HeLa cell extract showed that kinasesthat create the MPM-2 epitope at kinetochores and chro-mosome arms are distinct. Our results suggest that multiplekinases (either soluble or kinetochore-bound), including ahomolog of mammalian NIMA, can create the MPM-2phosphoepitope. The kinetochore-bound kinase thatcatalyzes the formation of the MPM-2 phosphoepitope mayplay an important role in key events such as mitotic kin-etochore assembly and sister chromatid separation atanaphase.

Key words: Rephosphorylation, Kinetochore, MPM-2, Mitosis,Chromosome, NIMA

SUMMARY

INTRODUCTION

Phosphorylation is a ubiquitous mechanism for regulating cellcycle processes and has been implicated in all stages of mitosis.Antibodies that specifically recognize the phosphorylatedforms of proteins are valuable tools for studying the regulationof mitotic events. One such monoclonal antibody, termedMPM-2, was raised against mitotic HeLa extracts (Davis et al.,1983). The MPM-2 antibody binds to a subset of proteins thatbecome phosphorylated during G2 and are dephosphorylated atthe end of mitosis (Davis et al., 1983; Engle et al., 1988). Thisantibody reacts with mitotic proteins from a wide variety ofspecies, including hamster, mouse, frog, chicken, mosquito,nematode, Drosophila and onion (Davis et al., 1983; Vandre etal., 1984, 1986; Millar et al., 1987). This widespread distribu-tion of MPM-2 reactivity indicates that mitosis-specific phos-phorylations are conserved throughout evolution.

The major mouse chromosomal protein recognized by the

MPM-2 antibody in mitosis is DNA topoisomerase IIα (topoII)(Taagepera et al., 1993). Immunofluorescence studies haveshown that topoII is present in chromosome arms and concen-trated at centromeres (Taagepera et al., 1993; Gorbsky, 1994;Rattner et al., 1996). TopoII, like most MPM-2 antigens, isphosphorylated during late G2 and M phases (Wells andHickson, 1995; Kimura et al., 1996).

MPM-2 has also been shown to recognize one of the tworegulatory phosphorylations on mitogen-activated proteinkinase (MAP kinase); MPM-2 binding to this epitope resultsin a decrease in MAP kinase activity in vitro (Taagepera et al.,1994). MAP kinase belongs to a family of kinases known asextracellular signal-regulated kinases (ERKs) which playimportant roles in activation of G0 and G2 arrested cells(Ruderman, 1993). The appearance of MPM-2 reactivity in thecell cycle coincides with an increase in maturation promotingfactor (MPF) activity in progesterone stimulated Xenopuslaevis oocytes (Kuang et al., 1989). The MPM-2 antibody can

2014 L. Renzi and others

also inhibit ubiquitin-mediated destruction of cyclin B inXenopus egg extracts, by binding components of the ubiquitin-ligase complex termed the cyclosome or anaphase promotingcomplex (King et al., 1995).

Other important proteins recognized by MPM-2 includeNIMA (Ye et al., 1995), certain microtubule-associatedproteins (Tombes et al., 1991; Vandre et al., 1991; Kuang andAshorn, 1993), Xenopus CDC25 (Kuang et al., 1994), Xenopuswee1-like kinase (Mueller et al., 1995) and other mitotic phos-phoproteins still unidentified (Matsumoto-Taniura et al., 1996).Controversy exists over the exact nature of the MPM-2 epitope.It appears to recognize phosphothreonine in the context of sur-rounding amino acids (Zhao et al., 1989; Westendorf et al.,1994; Taagepera et al., 1993, 1994; Ding et al., 1997).

Several kinases appear to create MPM-2-reactive phospho-rylations. MAP kinase kinase (MEK) is responsible forcreating the MPM-2 reactive phosphorylations on MAP kinase(Taagepera et al., 1994). Westendorf (1994) used p34cdc2

kinase to generate the MPM-2 phosphoepitope on peptides invitro, and p34cdc2 kinase can also create an MPM-2 reactivephosphorylation on NIMA protein kinase (Ye et al., 1995).Kuang and Ashorn (1993) have reported two kinase activitiesthat can generate the MPM-2 phosphoepitope on unphospho-rylated substrates present in a Xenopus extract. One of themwas found to be MAP kinase. Recent data show that Plx1, aXenopus polo-like kinase, acts as an MPM-2 kinase by phos-phorylating CDC25C (Kumagai and Dunphy, 1996). However,little is still known about other kinases that create the MPM-2phosphoepitope in vivo.

In PtK1 cells extracted with detergent in the presence of phos-phatase inhibitors, the MPM-2 antibody detects phosphorylatedepitopes present at the kinetochores of mitotic chromosomes.This labeling is bright in prophase, prometaphase, and metaphasebut decreases in anaphase and telophase (Vandre et al., 1984;Taagepera et al., 1995). If cells are extracted in the absence ofphosphatase inhibitors, labeling of MPM-2 phosphoepitopesdiminishes, particularly at kinetochores (Taagepera et al., 1995).We have developed a protocol to rephosphorylate the dephos-phorylated MPM-2 epitope using the endogenous kinase boundat the kinetochores. We went on to develop methods to test ifpurified kinases can create the MPM-2 epitope on chromosomes.Active p34cdc2-cyclin B, casein kinase II and MAP kinase couldnot generate the MPM-2 phosophoepitope. However, purifiedNIMA or soluble HeLa mitotic cell extract were able to catalyzethe rephosphorylation of the MPM-2 epitope at kinetochores.Fractionation of the mitotic extract showed that kinetochores andchromosome arms can be separately rephosphorylated by twodifferent chromatographic column fractions. Our results suggestthat multiple kinases (either soluble or kinetochore-bound),including the mammalian homolog of NIMA, can create theMPM-2 phosphoepitope. Kinetochore kinases that catalyze theformation of the MPM-2 phosphoepitope may play an importantrole in such key events as mitotic kinetochore assembly and sisterchromatid separation at anaphase.

MATERIALS AND METHODS

MaterialsPtK1 Potoroo tridactylis kidney cells (ATCC CCL35) and HeLa S3cells (ATCC CCL 2.2) were obtained from American Type CultureCollection (Rockville, MD). Cell culture reagents as well as tyr-

phostin, RCAM-lysozyme (5-amino-[N-2-5-dihydroxybenzyl)-N′-2-hydroxybenzyl-salicylic acid (lavendustin A) and calphostin C werefrom Gibco-BRL (Gaithersburg, MD). CHAPS detergent, ATP, DTT,nocodazole, vinblastine, heparin, pepstatin A, leupeptin, N-ethyl-maleimide (NEM), ethyleneglycol-bis-(B-aminoethyl)-N-N-N′-N′-tetraacetic acid (EGTA), ethylenediaminetetraacetic acid (EDTA)were purchased from Sigma Chemical CO (St Louis, MO). Taxol wasa generous gift from Dr Matthew Suffness, National Cancer Institute,NIH. Olomoucine was from Promega (Madison, WI). Proteaseinhibitor Pefablock was purchased from Boehringer Mannheim Corp.(Indianapolis, IN). p13suc1 beads and the MPM-2 antibody wereprovided by Upstate Biotechnology Inc. (Lake Placid, NY). 4′,5,7-tri-hydroxyisoflavone (genistein), 2-aminopurine, K-252a, staurosporine,microcystin LR and protein phosphatase 1 were from CalbiochemNovabiochem (San Diego, CA). Centricon-100, Centricon-50, Cen-triprep-10 and Centricon-10 concentrators were purchased fromAmicon Inc. (Beverly, MA). For immunoreactions, Cy3-conjugatedgoat anti-mouse, fluorescein-conjugated goat anti-human and horse-radish peroxidase-conjugated anti-mouse secondary antibodies werefrom Jackson Immunoresearch Lab. (West Grove, PA). The DNA stainYo-pro-1 (Molecular Probes, Inc., Eugene OR) and Vectashieldmounting medium (Vector Labs, Inc., Burlingame, CA) were alsoused. Immunoblots were developed with Supersignal ChemiLumi-nescence-HRP Substrate System (Pierce, Rockford, IL). Okadaic acidwas purchased from LC Laboratories (Woburn, MA). Rabbit poly-clonal anti-polo-like kinase (PLK) antibodies and recombinantProtein G-Sepharose were from Zymed Lab. Inc. (San Francisco, CA).Purified topoII was from TopoGEN (Columbus, OH). For the FPLCfractionation, the 5×5 MonoQ column was obtained from PharmaciaBiotech (Piscataway, NJ).

Cell culture and preparation of cell extractsPtK1 cells were grown in MEM medium supplemented with 10% (v/v)fetal bovine serum, 20 mM Hepes, 0.1 mM non-essential amino acids,1 mM pyruvate, 60 µg/ml penicillin and 100 µg/ml streptomycin.Cells were plated onto 18 × 18 mm glass coverslips and grown toapproximately 70% confluence in a humidified incubator with 5%CO2. In experiments in which nocodazole (16 µM), vinblastine (10µM) and taxol (5 µM) were used, PtK1 cells were incubated for 3hours in the presence of microtubule poisons before the rephospho-rylation assay was performed. HeLa S3 cells were grown in a 1 litrespinner flask at a density of 5×105 cells/ml in DMEM containing 5%(v/v) bovine calf serum, 20 µg/ml gentamicin, 20 mM Hepes, 0.1 mMnon-essential amino acids and pluronic F68 0.05% (w/v). Cells wereblocked in mitosis by adding 5 µM vinblastine 18 hours before har-vesting. This procedure resulted in a mitotic index of 80% as judgedby fluorescence labeling of DNA. To prepare S-phase cell extract,HeLa cells were blocked with 1 µg/ml aphidicolin for 18 hours beforeharvesting. Cells were centrifuged at 300 g, rinsed twice in MBS (10mM MOPS, pH 7.4, 150 mM NaCl), then rinsed once in mitoticextraction buffer (MEB): 50 mM Tris-HCl, pH 7.3, 50 mM KCl, 10mM MgCl2, 10 mM EGTA, 1 mM dithiothreitol (DTT). Cells wereresuspended in MEB (5 ml per 1 ml pelleted cells), incubated 20minutes on ice and then lysed with a Dounce homogenizer (pestle A,about 100 strokes). Homogenate was spun down 10 minutes at 1,000g. The pellet was discarded, and supernatant centrifuged for 1 hour at140,000 g. The high speed supernatant, termed S2, was collected,flash frozen and stored at −70°C.

Partial purification of the mitotic HeLa cell extract was accom-plished by precipitation with ammonium sulphate, ultrafiltrationthrough a 100 kDa cut-off membrane and chromatography with aMonoQ column.

Chromosome isolationHeLa cells were grown in suspension at the same conditions describedabove. Cells were blocked for 16-18 hours in 0.15 µg/ml colcemid.Cells were collected by centrifugation at 300 g for 5 minutes, and

2015Kinetochore phosphorylation

pellet rinsed twice in 50 ml swelling buffer (10 mM Hepes, 40 mMKCl, 5 mM EGTA, 4 mM MgSO4, pH 7.4, in the presence of proteaseinhibitors as described above). Pelleted cells were checked for mitoticindex, always above 75% as judged by fluorescence labeling of DNA.Swollen cells were lysed by repetitive pipetting in 10 mlextraction/lysis buffer (60 mM Pipes, 25 mM Hepes, 10 mM EGTA,4 mM MgSO4, pH 6.9, in the presence of 1 mM DTT, 1% CHAPSdetergent and protease inhibitors). When indicated, 200 nM micro-cystin was used. The suspension was centrifuged at 64 g for 5 minutes,and supernatant containing chromosomes was saved. In order to getthe majority of chromosomes free in the supernatants, lysis and cen-trifugation steps were repeated 4 times. The collected supernatantswere then centrifuged for 7 minutes at 200 g to pellet the fewremaining interphase nuclei and large pieces of cell debris. The super-natant was centrifuged again for 10 minutes at 1,600 g to pellet chro-mosomes. The pellet was washed three times in extraction lysis buffer,and resuspended in 7 ml of the same buffer. One aliquot of the chro-mosome suspension was treated with 10 mM NEM for 20 minutes atroom temperature to inactivate chromosome kinases, and furtherprocessed as described below. A 40/80% glycerol step-gradient wasused to purify chromosomes. The suspension above was layered ontop of the gradient and centrifuged at 2,700 g for 30 minutes. Mostother organelles and soluble molecules remain above or within the40% glycerol during centrifugation of the chromosomes throughglycerol step-gradient. Chromosomes were collected from the 40/80%interphase and within the 80% glycerol. The suspension was washedtwice in buffer, flash frozen and stored at −70°C.

Purification of NIMA protein kinase from Escherichia coli Bl21(DE3)pLYS S E. coli cells containing pET-21a-NIMA (Ye et al.,1995) were cultured in NYZ medium with 34 µg/ml chlorampheni-col and 200 µg/ml carbenicillin at 37°C. Cells were harvested by cen-trifugation when the OD600 nm of the culture reached 0.4-0.6. Bacteriawere washed by centrifugation and resuspended in fresh NYZmedium. Expression of recombinant NIMA was induced by additionof IPTG to a final concentration of 0.4 mM and shaking the cultureat 21°C for 3.5 hours. The induced cells were then harvested by cen-trifugation and frozen in liquid nitrogen. Soluble cell extract wasprepared in binding buffer from the His-BindTM buffer kit (Novagen)with addition of protease inhibitors (10 µg/ml leupeptin, 10 µg/mlsoybean chymotrypsin/trypsin protease inhibitor, 50 µg/ml PMSF and10 µg/ml aprotinin). NIMA protein kinase was purified using the His-BindTM protein purification system (Novagen). After elution, fractionscontaining NIMA were determined by kinase assay and western blot.The purified NIMA protein was washed, concentrated by centrifuga-tion in a Centricon concentrator (molecular mass cut-off: 30,000), andstored in phosphate buffer (pH 7.2 with 2.5 KCl, 10 mM MgSO4, 0.1mM EGTA, 20% glycerol and protease inhibitors).

Immunodepletion of mitotic HeLa cell extractTo investigate if it was possible to deplete the kinase activity from themitotic HeLa cell extract, immunoprecipitation was performed byusing either p13suc1 beads or anti-PLK antibodies. GST-p13suc1 beadswere used to precipitate p34cdc2 out of mitotic HeLa cell extract. Themitotic extract diluted 1:10 in TM was incubated for 2 hours on arocker in the presence of 6.2 mg/ml GST-p13suc1. Immunocomplexeswere pelleted, rinsed and stored at -70°C. Supernatant was used in therephosphorylation assay to test for the presence of MPM-2 kinaseactivity.

Immunoprecipitation of PLK from mitotic HeLa cell extract wasperformed as follows. 50 µl Protein G bead slurries were washed threetimes in RIPA buffer (50 mM Tris, 150 mM NaCl, 1% NP-40, 0.5%deoxycholate, pH 8.0) and prebound to 0.2 µg/µl anti-PLK antibodiesfor 2 hours while rocking at 4°C. Beads were then washed three timesin TM, and added to 500 µl of mitotic HeLa cell extract diluted 1:10in TM in the presence of 0.5% NP-40, for 2 hours on a rocker at 4°C.Immunocomplexes were pelleted, rinsed in TM and stored at −70°C.

Supernatants were used in the rephosphorylation assay to test forMPM-2 kinase activity. The pellet and the immunodepleted super-natant were analyzed by electrophoresis and immunoblotting to verifythe complete depletion of PLK. Immunoprecipitation by using beadsalone was carried out as control.

Kinetochore rephosphorylation assayThroughout the rephosphorylation assay, buffers were maintained at37°C. PtK1 cells were rinsed briefly in 50 mM Tris-HCl, 5 mM MgSO4,pH 7.4 (TM buffer) and extracted for 4 minutes in TM buffer contain-ing 1% CHAPS detergent. The coverslips were then rinsed and incubatedin TM buffer for 7 minutes to allow dephosphorylation of kinetochoresby endogenous phosphatases. The kinetochore rephosphorylationreaction was then performed by inverting coverslips onto 50 µl drops ofTM containing 670 nM microcystin LR, 1.7 mM ATP, and 1.7 mM DTT.The coverslips were allowed to rephosphorylate for 20 minutes at 37°Cin a humidified chamber. To test whether the disassembly of micro-tubules would affect the expression of the MPM-2 phosphoepitope, theendogenous rephosphorylation assay was also carried out on PtK1 cellstreated with vinblastine for 3 hours before lysis.

Experiments to understand which protein phosphatase was respon-sible for kinetochore dephosphorylation were also carried out. PtK1cells were extracted for 4 minutes in TM containing 1% CHAPS, inthe presence of either microcystin LR or okadaic acid (1, 10, 100 and1,000 nM), and immediately fixed as described below.

In order to test kinase inhibitors, the assay was conducted verysimilarly to the rephosphorylation assay, with the inhibitors beingincluded directly in the rephosphorylation mixture. The reagents testedwere: heparin (1 µM), ethyleneglycol-bis-(B-aminoethyl)-N-N′-N′-N’-tetraacetic acid (EGTA, 10 mM), ethylenediaminetetraacetic acid(EDTA, 5 mM); 4′,5,7-trihydroxyisoflavone (genistein, 200 µM), 2-aminopurine (10 mM), K-252a (20, 50, 100 µM), staurosporine (100,1,000 nM); tyrphostin (200 µM), RCAM-lysozyme (10 µg/ml), (5-amino-[N-2-5-dihydroxybenzyl)-N′-2-hydroxybenzyl-salicylic acid(lavendustin A, 5 µM), calphostin C (1 µM) and olomoucine (100 µM).

The rephosphorylation assay was further modified to test the abilityof externally applied kinases to catalyze the rephosphorylation ofkinetochores. Immediately before the rephosphorylation step, the cov-erslips were incubated in 5 mM NEM for 10 minutes to inactivate thekinase activity bound at the kinetochore. The coverslips were thenrinsed 3 times in TM buffer, and inverted on drops of purified activatedMAP kinase (0.2 ng/µl) (a gift from M. Weber, University ofVirginia), casein kinase II (3 and 6 ng/µl) (a gift from D. Litchfield,University of Manitoba; New England Biolabs), p34cdc2-cyclin B(0.02 and 0.2 U/µl) (New England Biolabs), NIMA (41.5 ng/µl) ormitotic HeLa cell extract. Rephosphorylation buffers included 670 nMmicrocystin LR, 1.7 mM ATP, and 1.7 mM DTT. The coverslips wereallowed to rephosphorylate for 20 minutes at 37°C in a humidifiedchamber. In some cases coverslips were treated with protein phos-phatase 1 (PP1, 0.1 u/ml) before the rephosphorylation step, and thenwashed with TM containing microcystin and DTT. Phosphatasetreatment reduced the background labelling of the kinetochores andthus increased the sensitivity of the rephosphorylation assay.

Cell fixation and immunofluorescenceCoverslips were briefly rinsed in TM buffer and then fixed in freshlyprepared 1% formaldehyde in 60 mM Pipes, 25 mM Hepes, 10 mMEGTA, and 4 mM MgSO4 pH 6.9 for 12 minutes. Coverslips werethen rinsed in MBS, pH 7.4, containing 0.05% Tween-20 (MBST).MPM-2 mouse ascites fluid was used at a 1/30,000 dilution in MBSTwith 5% (v/v) normal goat serum (NGS). When applied, anti-kineto-chore antibodies (ACA, a gift from J. B. Rattner, University ofCalgary) were used at 1/2,000 dilution in MBST with 5% NGS. Cov-erslips were inverted onto 40 µl drops of the antibody solution andincubated at 37°C for 20 minutes. Afterwards, the coverslips wererinsed in 2 changes of MBST. The coverslips were then incubated at37°C for 20 minutes on 40 µl drops of Cy3-conjugated goat anti-


mouse IgG, 2.5 µg/ml in 5% NGS/MBST. When ACA was used asprimary antibodies, coverslips were incubated with fluorescein-con-jugated goat anti-human IgG, 7.5 µg/ml in 5% NGS/MBST. The cov-erslips were then rinsed with 2 changes of MBST, and stained for 1minute with the DNA stain Yo-pro-1 at 0.2 µM in water or DAPI at5 µg/ml. The coverslips were then rinsed in distilled water andmounted on slides with Vectashield mounting medium supplementedwith 10 mM MgSO4. Images were taken using a Nikon invertedmicroscope configured for epifluorescence with a planapochromat×60 oil immersion 1.4 NA objective. An image intensifier and video-rate CCD camera (Dage-MTI, Michigan City, IN) with Image-1software (Universal Imaging, Media, PA) were used to digitize theimages. Except where indicated in the figure legends, images in eachfigure were taken under identical exposure conditions and processedidentically to allow comparison of the degree of labeling. Images wereprinted on a Kodak XLS 8600 PS dye sublimation printer.

Kinase assay, electrophoresis and immunoblot analysis In order to compare the patterns of MPM-2 reactive proteins on chro-mosomes in vivo and after rephosphorylation, chromosomes extractedin the absence of phosphatase inhibitors were allowed to rephospho-rylate in vitro. Isolated chromosomes (equivalent to approximately3×103 cells), prepared in the absence of microcystin, were incubatedat 37°C for 30 minutes in the presence of a phosphorylation buffercontaining 1 mM ATP, 1 mM DTT, 200 nM microcystin and proteaseinhibitors (Pefablock, leupeptin, pepstatin A; each at 5 µg/ml). Whenindicated, NIMA (83 ng/µl) or the PLK-immunodepleted extract (2µl) was incubated with the isolated chromosomes previously treatedwith NEM to inactivate endogenous chromosome kinases. Forcontrols, equal amounts of isolated chromosomes, prepared with orwithout microcystin, were incubated in the identical buffer withoutadded kinases or cell extract.

NIMA kinase, mitotic HeLa cell extract and FPLC fractions fromHeLa extract were tested for their ability to create the MPM-2 epitope

Fig. 1. Immunofluorescence labeling of the MPM-2 phosphoepitope in pphosphatase inhibitor microcystin (MC) (A,B), or in the absence of micrwhen cells are extracted in the absence of phosphatase inhibitor (C,D). Kphosphoepitope at kinetochores in the presence of ATP and microcystin poles (arrows) are resistant to dephosphorylation (D). Cells were fixed, lcounterstained with the DNA dye Yo-Pro. Images B, D and F were takenexperiment. Bar, 5 µm.

on topoII. Purified topoII (15 ng/µl) was incubated with NIMA (83ng/µl), mitotic HeLa cell extract or the FPLC fractions at 37°C for 30minutes in the presence of a phosphorylation buffer containing 1 mMATP, 1 mM DTT, 200 nM microcystin and protease inhibitors. To testNIMA kinase activity, NIMA was incubated for 20 minutes with 100µg β-casein and topoII in the presence of 25 µCi [32P]ATP.

After the phosphorylation reactions, 2× sodium dodecyl sulfate(SDS) sample buffer was added (1:1, v/v) to samples and heated to70°C for 5 minutes. Protein samples were then analyzed by elec-trophoresis on 5-20% gradient SDS-polyacrylamide gels and trans-ferred to polyvinyldifluoride membranes. Autoradiography wasperformed at this point for the NIMA kinase assay. For westernanalysis, blots were blocked overnight in 5% BSA/TBST (10 mMTris-HCl, pH 8, 150 mM NaCl, 0.05% Tween-20) at 4°C. MPM-2antibody was used to immunostain blots at 1/20,000 dilution of ascitesfluid in TBST for 2 hours on a rocker at room temperature. Afterrinsing 3 times for 5 minutes, blots were incubated with 30 ng/ml anti-mouse horseradish peroxidase-conjugated secondary antibody inTBST for 90 minutes on a rocker at room temperature. After rinsingas above, immunoblots were developed with Supersignal ChemiLu-minescence-HRP Substrate System.

RESULTS

Dephosphorylation and endogenousrephosphorylation of the MPM-2 phosphoepitopeMitotic PtK1 cells that were detergent extracted in the presenceof the phosphatase inhibitor microcystin and labeled withMPM-2 displayed strong kinetochore labeling (Fig. 1A,B). Ifthe cells were extracted without microcystin and allowed toincubate in TM buffer, the bright kinetochore labeling disap-peared completely (Fig. 1C,D). When the extracted, dephos-

rometaphase PtK1 cells. Cells were extracted in the presence of theocystin (C,D). The MPM-2 phosphoepitope at kinetochores is lostinetochore-bound MPM-2 kinases can recreate the MPM-2

either in untreated (E,F) or in vinblastine-treated (G,H) cells. Spindleabeled by indirect immunofluorescence with the MPM-2 antibody and under identical conditions. Image H was obtained in a separate


Fig. 2. Rephosphorylation of PtK1 kinetochores by endogenous bound kinases. Cells extracted in the presence of microcystin (A) generallyshow the typical MPM-2 labeling at kinetochores when cells are allowed to dephosphorylate and rephosphorylate by endogenous kinases in thepresence of ATP (B). Strong MPM-2 labeling is present at kinetochores in prophase, prometaphase and metaphase. However, in late anaphase,rephosphorylated cells show a more intense labeling with MPM-2 antibody compared to their counterparts extracted in the presence of MCwithout rephosphorylation. Bar, 5 µm.

phorylated cells were incubated with ATP, the reducing agentDTT, and the phosphatase inhibitor microcystin, MPM-2labeling of the kinetochores was restored (Fig. 1E,F). Cellstreated with the microtubule-disrupting agent vinblastinebefore extraction showed also robust kinetochore rephos-phorylation (Fig. 1G,H). The spindle poles were relativelyinsensitive to dephosphorylation (Fig. 1D, arrows). Moreover,they did not exhibit any noticeable increase in MPM-2 labelingafter the rephosphorylation step. Chromosome arms wereweakly labeled by the MPM-2 antibody (Fig. 1B,D,F,H), withno remarkable difference after extraction in the presence orabsence of microcystin. In order to identify the protein phos-phatase involved in the dephosphorylation of the MPM-2epitope, okadaic acid, which inhibits type 2A protein phos-phatases more effectively than type 1 protein phosphatases, andmicrocystin, which inhibits both types, were used. Results,shown in Table 1, indicate that microcystin was more effectivethan okadaic acid in inhibiting the dephosphorylation of the

MPM-2 epitope at kinetochores. The MPM-2 epitope waspreserved only in the presence of 1 µM okadaic acid, while 10nM microcystin was sufficient to achieve strong MPM-2labeling.

In most cases, the pattern of rephosphorylation of the MPM-2 epitope at kinetochores was similar to the in vivo labelingseen after extraction in the presence of microcystin (Fig.2A,B). Interestingly, in late anaphase cells the MPM-2 labelingafter rephosphorylation showed higher intensity compared toanaphase cells labeled before dephosphorylation (Fig. 2A,B).Although MPM-2 labeling at kinetochores was evident afteronly 5 minutes of rephosphorylation, robust rephosphorylationwas achieved in 20 minutes at 37°C.

The immunofluorescence data were consistent with thepattern shown by the western blot analysis on isolated chro-mosomes (Fig. 3). Chromosomes isolated in the presence ofmicrocystin showed distinct bands (Fig. 3, w/MC) that are lostwhen chromosomes are extracted in the absence of microcystin


Table 1. Effect of concentrations of phosphatase inhibitorsmicrocystin and okadaic acid to inhibit dephosphorylation

of the MPM-2 epitope at kinetochoresMicrocystin Okadaic acid

1 nM − 1 nM −10 nM + 10 nM −50 nM + 50 nM −

100 nM + 100 nM +/−1 µM + 1 µM +

(+) MPM-2 labeling; (−) no MPM-2 labeling; (+/−) partial MPM-2labeling.

Fig. 3. Isolated HeLa chromosomes can regenerate the MPM-2pattern of phosphoprotein. Immunoblot with MPM-2 antibody showschromosomes extracted in the presence (w/MC) and absence(w/oMC) of microcystin, as well as chromosomes extracted w/oMCafter rephosphorylation by endogenous kinases (rephos). Note somesimilarities between the pattern of MPM-2 reactive band inchromosomes extracted w/MC and after rephosphorylation. Sizestandards (in kDa) are shown on the right.

Table 2. Kinase inhibitors used to prevent kinetochorerephosphorylation

Kinase inhibitors Concn. Rephos Kinase inhibited

Heparin 1 µM + Casein kinase IIOlomoucine 100 µM + Cdc2, CDKs, ERKsCalphostin C 1 µM + PKC2-aminopurine 10 mM − GeneralStaurosporine 100 nM +/− Phospholipid Ca-dep

1 µM − PKC, PKG,General>100 nM

K-252a 20 µM + CAM kinase, MLC50 µM +/− kinase, PKA, PKC

100 µM − General>1 µMGenistein 200 µM + Tyrosine kinasesRCAM-lysozyme 10 µg/ml + Tyrosine kinasesTyrphostin 100 µM + Tyrosine kinasesLavendustin 5 µM + Tyrosine kinases

(+) Rephosphorylation occurred; (+/−) Partial rephosphorylation occurred;(−) Rephosphorylation inhibited.

(Fig. 3, w/oMC). When in vitro rephosphorylation of dephos-phorylated chromosomes (i.e. chromosomes extracted in theabsence of phosphatase inhibitors) was carried out, some of theMPM-2 reactive bands reappeared (Fig. 3, rephos).

The kinetochore-bound MPM-2 kinase was inhibitedonly by kinase inhibitors of broad specificityIn order to characterize the endogenous kinase that couldcreate the MPM-2 epitope at kinetochores, kinase inhibitorsand other chemicals were included in the rephosphorylationprocedure described above. Microcystin was included in therephosphorylation mixture to ensure that endogenous phos-phatases did not interfere with the assay. The results of theseexperiments are indicated in Table 2. At low concentrations,staurosporine, K-252a, calphostin C, heparin, and olomoucineall failed to inhibit rephosphorylation. Collectively at theseconcentrations these compounds should inhibit cyclindependent kinases, calcium-calmodulin dependant proteinkinases I and II, myosin light chain kinase, casein kinase II,protein kinases A, C and G, and MAP kinase. Thus, it isunlikely that any of these kinases are the endogenous kinaseresponsible for rephosphorylating the MPM-2 phosphoepitope.Additionally, the specific tyrosine kinase inhibitors genistein,lavendustin, RCAM-lysozyme, and tyrphostin all failed toinhibit the rephosphorylation supporting previous data thatMPM-2 phosphoepitope contains phosphothreonine (Zhao etal., 1989; Taagepera et al., 1994; Westendorf et al., 1994; Dinget al., 1997). The rephosphorylation reaction was blockedunder conditions of general kinase inhibition. These conditionsincluded the addition of 2-aminopurine, and high concentra-tions of staurosporine (1 µM) and K-252a (100 µM). At theseconcentrations staurosporine and K-252a lose their specificityand inhibit many kinases. Brief pretreatment of the extractedcells with 5 mM NEM and 5 mM DTNB, two sulfhydryloxidizing agents, also inhibited rephosphorylation, even afterwashing away the reagents. These results suggested that NEMand DTNB could permanently inactivate the kinetochore-bound kinase. Furthermore, the kinase was not able to use GTPas an alternative for ATP. Inclusion of 5 mM EDTA, but not10 mM EGTA, inhibited rephosphorylation, indicating thatkinase activity at kinetochores requires magnesium but notcalcium.

Characterization of kinases responsible for therephosphorylation of the MPM-2 phosphoepitopeThe kinetochore rephosphorylation assay described above wasfurther modified, allowing us to test exogenously applied kinasesfor their ability to create the MPM-2 phosphoepitope. As noted,

treatment with N-ethylmaleimide permanently inactivates theendogenous kinetochore kinase. This provides a dephosphory-lated cellular substrate to test purified kinases or cell extracts fortheir ability to rephosphorylate the MPM-2 epitope at kineto-chores. We found that a soluble extract prepared from mitoticHeLa cells was able to regenerate the MPM-2 phosphoepitopeat kinetochores. As shown in Fig. 4, PtK1 chromosomesappeared labeled at kinetochores. Kinetochores were brightlylabeled in prophase, prometaphase, metaphase. Labeling at kin-etochores is gone or very dim in anaphase and telophase cellsafter rephosphorylation with mitotic cell extract (Fig. 4). Cellsincubated with ATP alone in the absence of mitotic cell extractshowed no rephosphorylation (Fig. 4). Similarly, rephosphory-lation carried out with mitotic extract in the absence of ATP didnot result in rephosphorylation of kinetochores (not shown). Themidbody bound the MPM-2 antibody with no remarkable dif-ferences between control and rephosphorylated cells, the degreeof labeling varying depending upon the stage of development ofthe midbody during late anaphase/telophase. As in the rephos-phorylation assays with the kinetochore-bound kinase, inrephosphorylation assays with exogenously added mitotic


extract, spindle poles did not undergo dephosphorylation orrephosphorylation, as shown in Fig. 4. Interestingly, PtK1 cellsrephosphorylated in the presence of mitotic HeLa cell extractshowed brighter MPM-2 labeling at chromosome arms (Fig. 4,HeLa extract + ATP) with respect to control cells (Fig. 4, buffer+ ATP).

Fig. 4. Mitotic HeLa cell extract can catalyze the rephosphorylation of kinhibitors and incubated with NEM to inactivate endogenous kinases. Cethe presence of ATP. Kinetochores are rephosphorylated by soluble kinafigures, spindle poles are always intensely labeled both in treated and in occurs in treated cells with respect to controls. Bar, 5 µm.

Typically, mitotic HeLa cell extracts (prepared under theconditions described in Materials and Methods) could bediluted 1,000-fold and still catalyze kinetochore rephosphory-lation. Full strength extracts from S-phase cells showed someactivity in rephosphorylating kinetochores but this was lost ifthe S-phase extracts were diluted only 10-fold (not shown).

inetochores. PtK1 cells were detergent extracted without phosphatasells were then incubated with mitotic HeLa cell extract or in buffer in

ses present in the mitotic HeLa cell extract. As shown in previouscontrol cells. Some increase in MPM-2 labeling at chromosome arms


Fig. 5. NIMA kinaserephosphorylates MPM-2phosphoepitopes at kinetochores.Cells were detergent extractedwithout phosphatase inhibitorsand incubated with NEM toinactivate endogenous kinases.Cells were then incubated withpurified NIMA or buffer in thepresence of ATP. Kinetochoresbecome rephosphorylated byNIMA. (A,B) DNAcounterstained with DAPI;(C,D) anti-kinetochoreantibodies (ACA); (E,F) MPM-2labeling. Bar, 5 µm.

This evidence suggests that the MPM-2 kinase is less active oris present at lower concentrations in cells blocked in S-phase.PtK1 cells blocked in a prometaphase-like state by microtubuledrugs, either by microtubule-stabilizing agents such as taxol ormicrotubule-disrupting agents such as nocodazole and vin-blastine, could also be rephosphorylated with mitotic cellextract at the same extent as untreated cells (data not shown).

To determine whether known kinases were able to regener-ate the MPM-2 epitope on kinetochores, purified activatedMAP kinase, p34cdc2-cyclin B, casein kinase II and NIMAwere tested on dephosphorylated cellular substrates. MAPkinase, p34cdc2-cyclin B and casein kinase II failed to rephos-phorylate the dephosphorylated MPM-2 epitope. NIMA wasable to create the MPM-2 epitope at kinetochores whenincubated with ATP, as shown in Fig. 5. However, the extentof the rephosphorylation by NIMA at kinetochores was lessthan the MPM-2 signal after rephosphorylation by the en-dogenous kinase or by M-phase extract. Control cells, eitherincubated with ATP without NIMA or with NIMA withoutATP did not show any MPM-2 labeling at kinetochores. It hasbeen shown that hyperphosphorylated NIMA is reactive to theMPM-2 antibody (Ye et al., 1995). For this reason we wishedto determine whether the added NIMA became bound to kin-

AFig. 6. (A) Extent of immunodepletion of mitotic HeLacell extract using anti-PLK antibodies. Mitotic HeLa cellextract was incubated with protein-beads alone orimmunoprecipitated with beads and anti-PLK antibody.The anti-PLK antibody recognizes an approx. 68 kDaband specific for PLK, plus another cross reactive band ataround 46 kDa. Western blot with anti-PLK antibodiesshows mitotic HeLa cell extract (cells),immunoprecipitated materials from controls (ip cont) andfrom PLK depleted extract (ip PLK), and the respectivesupernatants (sup cont, sup PLK). PLK isimmunoprecipitated by the antibodies. Theimmunodepleted extract shows an undetectable level of PLK. (B) Dephoextract, or with PLK-immunodepleted extract plus ATP. Size standards (

etochores during the assay. After incubation with NIMA,cytoskeletons were washed and subsequently incubated withATP. Kinetochores were not able to be rephosphorylated underthese conditions, suggesting that NIMA does not bind to kin-etochores. These data indicate that NIMA kinase activitycreates the MPM-2 epitope at kinetochores, without represent-ing an MPM-2 antigen bound to kinetochores. Since it was lesseffective than mitotic extract even after extended incubation, itmay be only partly responsible for the rephosphorylation atkinetochores. Alternatively, the Aspergillus NIMA kinaseexpressed in bacteria may not efficiently recognize the PtK1cell substrate protein. In vitro rephosphorylation on isolatedchromosomes by NIMA did not show any specific patterns ofregenerated MPM-2 epitopes (see Fig. 10A).

Mitotic extract depleted of cyclin-dependent kinases(CDKs) with p13suc1-bound agarose beads did not havereduced MPM-2 kinase activity. Mitotic HeLa cell extract wasalso immunodepleted with antibodies to PLK. The anti-PLKantibodies detected a band at approx. 68 kDa corresponding toPLK, as well as a cross reactive protein of unknown identityat around 46 kDa (Fig. 6A). The immunoprecipitation with thisantibody out of mitotic HeLa cell extract left an undetectableamount of PLK in the supernatant (Fig. 6A). When the kin-

B

sphorylated chromosomes were incubated with PLK-immunodepletedin kDa) are shown on the right.


Fig. 7. Anti-PLK antibodiesimmunodeplete MPM-2 kinaseactivity. Mitotic HeLa cell extractwas incubated with protein-beadsalone or immunoprecipitated withbeads and anti-PLK antibody.Supernatants were tested forMPM-2 kinase activity. Theincubation with the PLKimmunodepleted sample showsreduced MPM-2 kinase activity atkinetochores (G, arrowhead),while incubation with controlbeads maintains full kinaseactivity (H). Rephosphorylationwith buffer and ATP alone is alsoshown (I). (A,B,C) DNA wascounterstained with DAPI;(D,E,F) anti-kinetochoreantibodies (ACA). Bar, 5 µm.

etochore rephosphorylation assay was performed in thepresence of PLK-immunodepleted extract, the MPM-2 labelingwas greatly diminished (Fig. 7). These data suggest that PLKmay be involved in the rephosphorylation of the MPM-2epitope. Isolated purified chromosomes incubated with PLK-immunodepleted extract and ATP showed no appreciable dif-ference after rephosphorylation (Fig. 6B).

In a long term effort to characterize the various MPM-2kinase activities present in mitotic cells, we have begun to frac-tionate HeLa cell extract. Although we have not yet purifiedthe kinases, we could observe that the flowthrough fraction(fraction 1) after FPLC was able to regenerate MPM-2 labelingat kinetochores (Fig. 8). The fraction eluted at 0.5-0.6 M NaCl(fraction 4) was not able to rephosphorylate kinetochore, butgave rise to an extremely bright MPM-2 labeling on wholechromosome arms (Fig. 8).

Phosphorylation of the MPM-2 epitope ontopoisomerase IIααSince it is known that topoII is a major protein recognized bythe MPM-2 antibody on mouse chromosomes (Taagepera et al.,1993), we wanted to investigate if topoII could be rephos-phorylated by the mitotic HeLa cell extract. Western blotanalysis showed that the MPM-2 reactivity of topoII increased

after incubation with the mitotic cell extract and ATP (Fig. 9A).This suggests that mitotic HeLa cell extract contained solublekinases able to phosphorylate topoII and generate the MPM-2epitope. In order to characterize topoII kinases in the mitoticcell extract and understand if the same kinase was also able torephosphorylate kinetochore or chromosome arms in ourexogenous rephosphorylation assay, several fractions fromMonoQ chromatography of mitotic HeLa cell extract wereincubated with ATP to rephosphorylate topoII. The productswere run on a PAGE gel and blotted with MPM-2 antibody(Fig. 9B). Only one fraction was able to rephosphorylate thetopoII (fraction 4, 0.5-0.6 M NaCl eluate), and this was thesame fraction that catalyzed the appearance of MPM-2 labelingon chromosome arms in the kinetochore rephosphorylationassay. The flowthrough (fraction 1), which was able to rephos-phorylate kinetochores, did not rephosphorylate topoII insolution.

In order to identify if topoII is a target protein rephosphory-lated by NIMA, we incubated NIMA with topoII and therephosphorylation mixture. Western blotting with MPM-2 ab(Fig. 10B) showed that NIMA did not rephosphorylate topoIIin the MPM-2 site. A kinase assay using [32P]ATP (Fig. 10C)showed that NIMA was active and able to rephosphorylate β-casein as substrate, but unable to phosphorylate topoII.


Fig. 8. Kinetochores and chromosome arms are rephosphorylated bytwo distinct pools of kinases. Mitotic HeLa cell extract wasfractionated as described in Materials and Methods. Fraction 1(flowthrough) can rephosphorylate MPM-2 epitope at kinetochoresbut does not act on chromosome arms. Fraction 4 (eluted at 0.5-0.6M NaCl) is able to rephosphorylate chromosome arms but notkinetochores. Fraction 4 can also rephosphorylate topoII in solution(see Fig. 9B). Bar, 5 µm.

DISCUSSION

Phosphoepitope antibodies are useful tools to study groups ofproteins that share consensus phosphorylation domains,providing more specificity in contrast to the traditional radioac-tive incorporation of phosphate. The MPM-2 phosphoepitopeantibody labels mitosis-specific and cell cycle-regulated phos-phoproteins localized at chromosome arms, spindle poles,midbody and kinetochores (Davis et al., 1983; Vandre et al.,1984; Engle et al., 1988; Zhao et al., 1989; Hirano andMitchison, 1991; Vandre et al., 1991; Taagepera et al., 1993,1995). Kinetochores play a fundamental role in regulating thecheckpoint system that monitors the metaphase-anaphase tran-sition at mitosis (reviewed by Gorbsky, 1995). The signalingpathway prevents chromosome segregation at anaphase untilall the chromosomes have reached the metaphase plate. Riederand colleagues (1994, 1995) recently demonstrated that asingle unattached kinetochore can inhibit anaphase onset, and

B

AFig. 9. (A) Mitotic HeLa cell extract canrephosphorylate topoisomerase IIα. Theextract was incubated with purified topoIIwithout or with ATP; samples were then runon a 5-20% PAGE gel and processed forwestern blotting with MPM-2 antibody. Thetopoisomerase IIα band shows an increase inlabeling representing phosphorylated form;(B) fraction 4 (eluted at 0.5-0.6 mM NaCl)from mitotic HeLa cell extract canrephosphorylate topoII in solution. Thisfraction is the same one that gives intense labeling on chromosome armMPM-2 kinase active at kinetochores) is not able to rephosphorylate top

destruction of the last unattached kinetochore can relieve thisblock, allowing the cell to complete mitosis. Still largelyunknown is the signaling pathway that activates the kineto-chore checkpoint. Tension alters the phosphorylation status ofkinetochore proteins, such as the cell cycle-regulated dephos-phorylation of proteins recognized by the 3F3/2 antibody (Liand Nicklas, 1995; Nicklas et al., 1995). Work done in our lab-oratory with 3F3/2 antibody has shown that 3F3/2 phospho-epitope is an integral part of the metaphase-anaphase check-point (Campbell and Gorbsky, 1995). PtK1 kinetochores arebrightly labeled with 3F3/2 antibody in prophase andprometaphase, but this epitope needed to be dephosphorylatedfor anaphase onset to occur. The fact that MPM-2 labeling alsoundergoes a dramatic change as the cell progresses throughmitosis suggests that phosphorylation and dephosphorylationof the MPM-2 phosphoepitope may also play an important rolein regulation of mitosis or meiosis.

We have developed a unique protocol for creating the MPM-2 phosphoepitope by rephosphorylating the dephosphorylatedepitope present at the kinetochores of mitotic PtK1 cells. SinceMPM-2 only recognizes phosphorylated proteins, the absenceof kinetochore labeling when cells are extracted without phos-phatase inhibitors would suggest that the MPM-2 phosphoepi-tope has been either dephosphorylated by endogenous phos-phatases or detached from kinetochores. The reappearance ofkinetochore labeling in rephosphorylation conditions suggeststhat protein(s) on which the MPM-2 phosphorylation occur areretained at kinetochores. Furthermore, this rephosphorylationoccurs in cells that have been detergent extracted and washedseveral times, suggesting that endogenous MPM-2 kinasesmust also be bound at the kinetochore. Western blots indeedconfirm these data, showing that dephosphorylated isolatedchromosomes can regenerate the MPM-2 patterns of phospho-protein after rephosphorylation in vitro.

The rephosphorylation and appearance of the MPM-2epitope provides some clues about the regulation of kineto-chore-bound kinases. Kinetochore kinases are bound and ableto rephosphorylate kinetochores from prophase throughanaphase. How the kinases are turned off in late anaphase isunknown. The immunofluorescence data and the western blotanalysis demonstrate the striking similarity between the MPM-2 labeling in vivo and after rephosphorylation, supporting theidea that the endogenous kinase that is phosphorylating theMPM-2 epitope also creates MPM-2 reactive phosphorylationsin vivo.

Mitotic HeLa cell extract contains highly active kinases thatcatalyze the phosphorylation of the MPM-2 epitope at kineto-

s in the rephosphorylation assay (see Fig. 8). Fraction 1 (containingoII.


Fig. 10. Rephosphorylation by NIMA.(A) Isolated chromosomes were incubatedwith NIMA or with NIMA in the presence ofATP. The extent of the MPM-2 pattern ofphosphoprotein is shown. Size standards (kDa)are shown on the right. (B) NIMA is not ableto rephosphorylate topoII and create the MPM-2 epitope. Rephosphorylation conditions andwestern blot analysis are the same as describedabove. (C) NIMA is active and phosphorylatesβ casein, but does not phosphorylate topoII.NIMA was incubated with topoII with orwithout β casein in the presence of [32P]ATP.Immobilized proteins were exposed for 5minutes (bottom) or for 5 hours (top; sizestandards in kDa are shown on the left);autophosphorylation of NIMA can be seen atlong exposures.

chores. We do not know yet if the kinetochore-bound kinasesand the soluble kinases present in the mitotic HeLa cell aresimilar molecules. One explanation is that the soluble andinsoluble activities represent two distinct pools of the samekinase, one cytoplasmic and the other bound at kinetochores.The soluble kinetochore kinases are cell cycle-regulated, sincethe highest MPM-2 kinase activity at kinetochores wasrecovered in cell extracts from HeLa cells blocked in mitosis.On the other hand we find low levels of the kinase activity inS-phase cells. This suggests that MPM-2 kinases need furtherprocessing or production after S phase to be fully active inmitosis.

Microcystin-LR, a cyclic heptapeptide isolated from fresh-water cyanobacteria, is a potent inhibitor of both PP1 and PP2aphosphatases (MacKintosh et al., 1990). In our experiments,microcystin was used to preserve the MPM-2 epitope fromdephosphorylation, and we show that it effectively inhibited thedephosphorylation of MPM-2 epitope at very low concentra-tions. Okadaic acid was not qually effective in preventing kin-etochore dephosphorylation. PP1 is located in chromosomes inmitotic cells (Fernandez et al., 1992), and our recent immuno-labeling data using an anti-PP1 antibody shows that this phos-phatase is concentrated at kinetochores (D. L. Brautigan andG. J. Gorbsky, unpublished data). Together, these resultssuggest that PP1 is likely the phosphatase responsible for kin-etochore dephosphorylation on cytoskeletons of PtK1 cells.

Ultrastructural immunolabeling studies indicate that MPM-2 labeling is concentrated in the inner and the outer denseplaque of kinetochores (Taagepera et al., 1995). Our experi-ments in the presence of microtubule-stabilizing and -disrupt-ing agents indicate that rephosphorylation of the MPM-2epitope occurs in cell with disrupted microtubules. This kin-etochore-bound kinases do not require interaction with micro-tubules for catalizing expression of the MPM-2 phosphoepi-tope.

MPM-2 also labels midbody and spindle poles, as reportedalso by Vandre et al. (1984). Epitopes present at the spindlepoles and midbody do not dephosphorylate when cells are

detergent extracted without phosphatase inhibitors. Further-more, these structures do not show any difference in MPM-2labeling when rephosphorylation occurs. These results suggestthat phosphorylations of midbody and spindle poles may beassociated with different proteins compared to the phosphory-lations at the kinetochore, or that they are less accessible toendogenous phosphatases. Chromosome arms also showMPM-2 labeling (Taagepera et al., 1993, 1995). We report herethat the intensity of labeling at chromosome arms increaseswith respect to the control cells when mitotic HeLa cell extractis used in the exogenous rephosphorylation assay. In fact, thedephosphorylation of the cellular substrate achieved using PP1treatment in this assay allows a better rephosphorylation atchromosome arms by kinases present in the incubationreaction.

The rephosphorylation protocol functions as a sensitiveassay for learning about the kinases that phosphorylate theMPM-2 phosphoepitope. The results of this study indicate thatthe kinase is a magnesium-dependent serine-threonine kinase,and argue against the possibility that the phosphorylations arecaused by purified MAP kinase, casein kinase II or CDC2kinase. Additionally, this study suggests that calcium-calmod-ulin dependant protein kinases I and II, myosin light chainkinase, casein kinase II, or protein kinases A or C are unlikelyto be responsible for the kinetochore rephosphorylation.Another possibility is that some of these kinases requireaccessory proteins or belong to a complex to be fully activeand able to rephosphorylate kinetochores.

Interestingly, we found that anti-PLK antibodies canpartially deplete MPM-2 activity from the mitotic HeLa cellextract. Recent data show that Plx1, a Xenopus polo-likekinase, directly phosphorylates CDC25C (Kumagai andDunphy, 1996). Since CDC25C is an MPM-2 antigen, wesuspect that PLK plays a role in the phosphorylation of theMPM-2 epitope at kinetochores.

NIMA is a protein kinase isolated from Aspergillus nidulansrequired for mitotic progression (Osmani et al., 1991a). NIMAkinase activity is low in G1 but increases in S and G2 to reach


a maximal level in mitosis (Osmani et al., 1991b). The proteinis then degraded before exit from mitosis (Pu and Osmani,1995). NIMA can be phosphorylated and activated by p34cdc2

(Ye et al., 1995). When hyperphosphorylated, NIMA showssome MPM-2 reactivity (Ye et al., 1995). NIMA inducespremature mitosis in HeLa cells (Lu and Hunter, 1995), andNIMA-related kinases or Neks have been isolated in mammals(Letwin et al., 1992; Schultz and Nigg, 1993, 1994; Lu andHunter, 1995). We show here that NIMA from Aspergillus canact as an MPM-2 kinase at kinetochores in PtK1 cytoskeletons,although we do not know yet if a mammalian NIMA homologis involved in the kinetochore rephosphorylation. Our experi-ments testing the ability of NIMA to bind to kinetochores showthat it is washed away after a quick rinse in buffer. This MPM-2 labeling at kinetochores is not due to NIMA binding to thekinetochores, but rather to phosphorylation of an endogenouskinetochore protein by the added NIMA kinase. Substrates ofNIMA kinase are still unknown. So far we have no evidencethat topoII is a NIMA substrate.

We show here that mitotic HeLa cell extract contains solubleMPM-2 kinases. The kinases that create the epitope at kineto-chores and chromosome arms are distinct and can be separatedby ion exchange chromatography. Further fractionation isongoing. Nevertheless, we show here that one FPLC fraction,namely the flowthrough, can rephosphorylate kinetochores onlysed PtK1 cells but is unable to phosphorylate topoII insolution. This indicates that the MPM-2 kinase at kinetochoresis not a topoII kinase. This kinase could be a NIMA-relatedkinase in mammals. In contrast, the eluted fraction couldrephosphorylate chromosome arms in cytoskeletons and topoIIin solution, but is not able to rephosphorylate kinetochores. Lit-erature data show that casein kinase II is a topoII kinase in vitro(Cardenas et al., 1993; Wells and Hickson, 1995), and in yeastit is associated with topoII (Cardenas et al., 1993). Other datareport that protein kinase C (Wells et al., 1995), p34cdc2 kinaseand MAP kinase (Wells and Hickson, 1995) are able to phos-phorylate topoII in vitro. In our exogenous rephosphorylationassay casein kinase II, p34cdc2 or MAP kinase are unable toregenerate the MPM-2 phosphoepitope on kinetochores orchromosome arms. Our study with kinase inhibitors would alsoexclude a role for protein kinase C in rephosphorylating theMPM-2 epitope.

On the whole, the pattern of MPM-2 phosphorylation wedescribed, together with other literature data, strongly suggestthat the kinetochore complex is transiently activated/phos-phorylated in a cell cycle-regulated manner. The identificationand characterization of kinetochore-associated kinases andphosphatases should give us insight into the signal transduc-tion pathway that controls mitotic chromosome assembly andsegregation.

We thank Dr M. Weber for providing MAP kinase, Dr D. Litchfieldfor casein kinase II and Dr Matthew Suffness for taxol. We acknowl-edge other members of the Gorbsky laboratory, expecially J. R. Daumfor his technical assistance. L.R. was supported by a fellowship fromthe University of Padua, Italy. This work was supported by AmericanCancer Society grants to G.J.G.

REFERENCES

Campbell, M. S. and Gorbsky, G. J. (1995). Microinjection of mitotic cells

with 3F3/2 antiphosphoepitope antibody delays the onset of anaphase. J. CellBiol. 129, 1-10.

Cardenas, M. E., Walter, R., Hanna, D. and Gasser, S. M. (1993). Caseinkinase II copurifies with yeast DNA topoisomerase II and re-activates thedephosphorylated enzyme. J. Cell Sci. 104, 533-543.

Davis, F. M., Tsao, T. Y., Fowler, S. K. and Rao, P. N. (1983). Monoclonalantibodies to mitotic cells. Proc. Nat. Acad. Sci. USA 80, 2926-2930.

Ding, M., Feng, Y. and Vandre, D. D. (1997). Partial characterization of theMPM-2 phosphoepitope. Exp. Cell Res. 231, 3-13.

Engle, D. B., Doonan, J. H. and Morris, R. N. (1988). Cell-cycle modulationof MPM-2-specific spindle pole body phosphorylation in Aspergillusnidulans. Cell Motil. Cytoskel. 10, 432-437.

Fernandez, A., Brautigan, D. L. and Lamb, N. J. C. (1992). Proteinphosphatase type 1 in mammalian cell mitosis: chromosomal localizationand involvement in mitotic exit. J. Cell Biol. 116, 1421-1430.

Gorbsky, G. J. (1994). Cell cycle progression and chromosome segregation inmammalian cells cultured in the presence of the topoisomerase II inhibitorsICRF-187 [(+)-1,2-bis(3,5-dioxopiperazinyl-1-yl)propane; ADR-529] andICRF-159 (Razoxane). Cancer Res. 54, 1042-1048.

Gorbsky, G. J. (1995). Kinetochores, microtubules and the metaphasecheckpoint. Trends Cell Biol. 5, 143-148.

Hirano, T. and Mitchison, T. J. (1991). Cell cycle control of higher-orderchromatin assembly around naked DNA in vitro. J. Cell Biol. 115, 1479-1489.

Kimura, K., Nozaki, N., Enomoto, T., Tanaka, M. and Kikuchi, A. (1996).Analysis of M phase-specific phosphorylation of DNA topoisomerase II. J.Biol. Chem. 271, 21439-21445.

King, R. A., Peters, J., Tugendreich, S., Rolfe, M., Hieter, P. and Kirschner,M. W. (1995). A 20S complex containing CDC27 and CDC16 catalyzes themitosis-specific conjugation of ubiquitin to cyclin B. Cell 81, 279-288.

Kuang, J., Zhao, J.-Y., Wright, D. A., Saunders, G. F. and Rao, P. N. (1989).Mitosis-specific monoclonal antibody MPM-2 inhibits Xenopus oocytematuration and depletes maturation-promoting activity. Proc. Nat. Acad. Sci.USA 86, 4982-4986.

Kuang, J. and Ashorn, C. L. (1993). At least two kinases phosphorylate theMPM-2 epitope during Xenopus oocyte maturation. J. Cell Biol. 123, 859-868.

Kuang, J., Ashorn, C. L., Gonzalez-Kuyvenhoven, M. and Penkala, J. E.(1994). CDC25 is one of the MPM-2 antigens involved in the activation ofmaturation-promoting factor. Mol. Biol. Cell 5, 135-145.

Kumagai, A. and Dunphy, W. G. (1996). Purification and molecular cloningof Plx1, a CDC25-regulatory kinase from Xenopus egg extracts. Science 273,1377-1380.

Letwin, K., Mizzen, L., Motro, B., Ben-David, Y., Bernstein, A. andPawson, T. (1992). A mammalian dual specificity protein kinase, Nek 1, isrelated to the NIMA cell cycle regulator and highly expresses in meiotic germcells. EMBO J. 11, 3521-3531.

Li, X. and Nicklas, R. B. (1995). Mitotic forces control a cell cycle checkpoint.Nature 373, 630-632.

Lu, X. P. and Hunter, T. (1995). Evidence for a NIMA-like mitotic pathway invertebrate cells. Cell 81, 413-424.

Matsumoto-Taniura, N., Pirollet, F., Monroe, R., Gerace, L. andWestendorf, J. M. (1996). Identification of novel M phase phosphoproteinby expression cloning. Mol. Biol. Cell 7, 1455-1469.

MacKintosh, C., Beattie, K. A., Klumpp, S., Cohen, P. and Codd, G. A.(1990). Cyanobacterial microcystin-LR is a potent and specific inhibitor ofprotein phosphatase 1 and 2A from both mammals and higher plants. FEBSLett. 264, 187-192.

Millar, S. E., Freeman, M. and Glover, D. M. (1987). The distribution of a‘mitosis-specific’ antigen during Drosophila development. J. Cell Sci. 87,95-104.

Mueller, P. R., Coleman, T. R. and Dunphy, W. G. (1995). Cell cycleregulation of a Xenopus Wee1-like kinase. Mol. Biol. Cell 6, 119-134.

Nicklas, R. B., Ward, S. C. and Gorbsky, G. J. (1995). Kinetochore chemistryis sensitive to tension and may link mitotic forces to a cell cycle checkpoint.J. Cell Biol. 130, 929-939.

Osmani, A. H., McGuire, S. L. and Osmani, S. A. (1991a). Parallel activationof the NIMA and p34cdc2 cell cycle-regulated protein kinases is required toinitiate mitosis in Aspergillus nidulans. Cell 67, 283-291.

Osmani, A. H., O’Donnel, K., Pu, R. T. and Osmani, S. A. (1991b).Activation of the nimA protein kinase plays a unique role during mitosis thatcannot be bypassed by absence of the bimE checkpoint. EMBO J. 10, 2669-2679.

Pu, R. T. and Osmani, S. A. (1995). Mitotic destruction of the cell cycle


regulated NIMA protein kinase of Aspergillus nidulans is required formitotic exit. EMBO J. 14, 995-1003.

Rattner, J. B., Hendzel, M. J., Sommer Furbee, C., Muller, M. T. andBazett-Jones, D. P. (1996). Topoisomerase IIα is associated with themammalian centromere in a cell cycle- and species-specific manner and isrequired for proper centromere/kinetochore structure. J. Cell Biol. 134, 1097-1107.

Rieder, C. L., Schultz, A., Cole, R. and Sluder, G. (1994). Anaphase onset invertebrate somatic cells is controlled by a checkpoint that monitors sisterkinetochore attachment to the spindle. J. Cell Biol. 127, 1301-1310.

Rieder, C. L., Cole, R. W., Khodjakov, A. and Sluder, G. (1995). Thecheckpoint delaying anaphase in response to chromosome monoorientationis mediated by an inhibitory signal produced by unattached kinetochores. J.Cell Biol. 130, 941-948.

Ruderman, J. V. (1993). MAP kinase and the activation of quiescent cells.Curr. Opin. Cell Biol. 5, 207-213.

Schultz, S. J. and Nigg, E. A. (1993). Identification of 21 novel human proteinkinases, including 3 members of a family related to the cell cycle regulatornimA of Aspergillus nidulans. Cell Growth Differ. 4, 821-830.

Schultz, S. J. and Nigg, E. A. (1994). Cell cycle-dependent expression ofNek2, a novel human protein kinase related to the NIMA mitotic regulator ofAspergillus nidulans. Cell Growth Differ. 5, 625-635.

Taagepera, S., Rao, P. N., Drake, F. H. and Gorbsky, G. J. (1993). DNAtopoisomerase IIα is the major chromosome protein recognized by themitotic phosphoprotein antibody MPM-2. Proc. Nat. Acad. Sci. USA 90,8407-8411.

Taagepera, S., Dent, P., Her, J.-H., Sturgill, T. W. and Gorbsky, G. J. (1994).The MPM-2 antibody inhibits mitogen-activated protein kinase activity bybinding to an epitope containing phosphothreonine-183. Mol. Biol. Cell 5,1243-1251.

Taagepera, S., Campbell, M. S. and Gorbsky, G. J. (1995). Cell-cycle-regulated localization of tyrosine and threonine phosphoepitopes at thekinetochores of mitotic chromosomes. Exp. Cell Res. 221, 249-260.

Tombes, R. M., Peloquin, J. G. and Borisy, G. G. (1991). Specific associationof an M-phase kinase with isolated mitotic spindles and identification of twoof its substrates as MAP4 and MAP1B. Cell Regul. 2, 861-874.

Vandre, D. D., Davis, F. M., Rao, P. N. and Borisy, G. G. (1984).Phosphoproteins are components of mitotic microtubule organizing centers.Proc. Nat. Acad. Sci. USA 81, 4439-4443.

Vandre, D. D., Davis, F. M., Rao, P. N. and Borisy, G. G. (1986). Distributionof cytoskeletal protein sharing a conserved phosphorylated epitope. Eur. J.Cell Biol. 41, 72-81.

Vandre, D. D., Centonze, V. E., Peloquin, J., Tombes, R. M. and Borisy, G.G. (1991). Proteins of the mammalian mitotic spindle:phosphorylation/dephosphorylation of MAP-4 during mitosis. J. Cell Sci. 98,577-588.

Wells, N. J. and Hickson, I. D. (1995). Human topoisomerase IIα isphosphorylated in a cell-cycle phase-dependent manner by a proline-directedkinase. Eur. J. Biochem. 231, 491-497.

Wells, N. J., Fry, A. M., Guano, F., Norbury, C. and Hickson, I. D. (1995).Cell cycle phase-specific phosphorylation of human topoisomerase IIα. J.Biol. Chem. 270, 28357-28363.

Westendorf, J. M., Rao, P. N. and Gerace, L. (1994). Cloning of cDNAs forM-phase phosphoproteins recognized by the MPM-2 monoclonal antibodyand determination of the phosphorylated epitope. Proc. Nat. Acad. Sci. USA91, 714-718.

Ye, X. S., Xu, G., Pu, R. T., Fincher, R. R., McGuire, S. L., Osmani, A. H.and Osmani, S. A. (1995). The NIMA protein kinase ishyperphosphorylated and activated downstream of p34cdc2/cyclin B:coordination of two mitosis promoting kinases. EMBO J. 14, 986-994.

Zhao, J., Kuang, J., Adlakha, R. C. and Rao, P. N. (1989). Threoninephosphorylation is associated with mitosis in HeLa cells. FEBS Lett. 249,389-395.

(Received 29 January 1997 – Accepted, in revised form,23 June 1997)

mpm-2 antibody-reactive phosphorylations can be created in … · 2001-05-03 · mpm-2 antibody in...

Documents