The human mitotic checkpoint protein BubR1 regulates ......15% (±4%) for CENP-E depletion and 33% (±4%) for BubR1 depletion (Fig. 1a–d, g). Thus, depletion of Ma d2 had no chromosome

L E T T E R S

93 NATURE CELL BIOLOGY VOLUME 7 | NUMBER 1 | JANUARY 2005

The human mitotic checkpoint protein BubR1 regulates chromosome–spindle attachmentsMichael A. Lampson1,2 and Tarun M. Kapoor1

Loss or gain of whole chromosomes, the form of chromosomal instability (CIN) most commonly associated with human cancers, is expected to arise from the failure to accurately segregate chromosomes in mitosis1. The mitotic checkpoint is one pathway that prevents segregation errors by blocking the onset of anaphase until all chromosomes make proper attachments to the spindle. Another process that prevents errors is stabilization and destabilization of connections between chromosomes and spindle microtubules. An outstanding question is how these two pathways are coordinated to ensure accurate chromosome segregation. Here we show that in human cells depleted of BubR1 — a critical component of the mitotic checkpoint that can directly regulate the onset of anaphase2–4 — chromosomes do not form stable attachments to spindle microtubules. Attachments in these cells are restored by inhibition of Aurora kinase, which is known to stabilize kinetochore–microtubule attachments5–7. Loss of BubR1 function thus perturbs regulation of attachments rather than the ability of kinetochores to bind to microtubules. Consistent with this finding, depletion of BubR1 increases phosphorylation of CENP-A, a kinetochore-specific Aurora kinase substrate. We propose that BubR1 links regulation of chromosome–spindle attachment to mitotic checkpoint signalling.

To prevent chromosomal instability due to unequal segregation of chromosomes during cell division, two mechanisms must be coordi-nated. One regulates the separation of sister chromosomes through activation of the mitotic checkpoint, and the other modulates the attachment of chromosomes to spindle microtubules. Genetic muta-tions or changes in expression levels of proteins involved in check-point signalling, such as BubR1 and Mad2, are associated with CIN8–12. Regulation of chromosome–spindle attachments requires the activ-ity of Aurora kinases, currently thought to be promising therapeutic targets that are also linked to CIN13. As the molecular basis for the coordination between these pathways is not known, we tested the possibility that known mammalian checkpoint proteins might also regulate the attachment of chromosomes.

1Laboratory of Chemistry and Cell Biology, Rockefeller University, New York, NY 10021, USA.2Correspondence should be addressed to M.A.L. (e-mail: [email protected])

Published online: 12 December 2004, DOI: 10.1038/ncb1208

As an indicator of proper chromosome attachment, we examined the alignment of chromosomes at the metaphase plate after depletion of checkpoint proteins by RNA interference (RNAi) in human cells. Depletion of checkpoint proteins can cause premature mitotic exit before chromosomes are properly attached to the spindle and aligned14. To control for the premature exit, we blocked the onset of anaphase using a proteasome inhibitor. Under these conditions control cells arrest with all chromosomes aligned at the metaphase plate (Fig. 1a). To quantify the extent of chromosome alignment, we measured the distance of individ-ual kinetochores from the nearest spindle pole along the pole–pole axis (Fig. 1f). Misaligned kinetochores were defined as those for which this distance is less than 20% of the pole–pole distance. Depletion pheno-types are shown for Mad2 and BubR1, two proteins that are required for checkpoint activation15,16 and interact directly with Cdc20 to inhibit the anaphase-promoting complex2–4, and CENP-E, an essential checkpoint signalling protein17 (Fig. 1b–d). Fewer than 2% of all kinetochores were misaligned in both control and Mad2-depleted cells, compared with 15% (±4%) for CENP-E depletion and 33% (±4%) for BubR1 depletion (Fig. 1a–d, g). Thus, depletion of Mad2 had no chromosome alignment phenotype, CENP-E was intermediate and BubR1 most severe. More detailed analysis revealed a uniform distribution of kinetochores along the length of the spindle in BubR1-depleted cells (Fig. 1h).

The severe chromosome misalignment phenotype of BubR1 deple-tion suggests that this mitotic checkpoint protein might be required for proper attachment of chromosomes to the spindle. To test this directly, we used three different assays for stable kinetochore-micro-tubule attachments. First we assayed for the presence of cold-stable microtubules, as kinetochore microtubules are preferentially stabilized at 4 °C18. In control and Mad2-depleted cells (data not shown), the spin-dle remained intact after cold treatment, with microtubule fibres clearly attached to each kinetochore stained with CREST antiserum (Fig. 2a–b). In CENP-E-depleted cells cold-stable kinetochore–microtubule fibres were present on aligned chromosomes but not on chromosomes near the pole (see Supplementary Information, Fig. S1i). In BubR1-depleted cells, however, only a few cold-stable microtubules remained, and most kinetochores had no microtubules attached to them (Fig. 2c–d). We also examined proteins whose kinetochore localization depends strongly

print ncb1208.indd 93print ncb1208.indd 93 8/12/04 3:14:19 pm8/12/04 3:14:19 pm

Nature Publishing Group© 2005


L E T T E R S

on microtubule attachment: dynein/dynactin and Mad2 (refs 19,20). In control cells, chromosomes were aligned and both Mad2 and p150 — a subunit of the dynactin complex — were generally undetectable on kinetochores (Fig. 2e, g). In cells depleted of BubR1, however, both Mad2 and p150 were detected on kinetochores at high levels (Fig. 2f, h). Quantification of fluorescence intensities showed that average Mad2 levels at BubR1-depleted kinetochores (n = 169) were more than 20 times greater than at control kinetochores (n = 118) (data not shown). Together, the data demonstrate that BubR1-depleted cells fail to form stable kinetochore–microtubule attachments.

We considered two possible explanations for the lack of stable kineto-chore–microtubule attachments in BubR1-depleted cells. First, regula-tion of microtubule attachments may be perturbed, so that attachments are not stabilized. Second, kinetochores may not assemble properly and therefore not form stable microtubule attachments. We first tested the possibility that the BubR1-depletion phenotype is mediated by a known regulator of microtubule attachments, Aurora kinase. Inhibition of the kinase activity stabilizes attachments, whereas active Aurora kinase has a destabilizing effect6,7. To test whether the BubR1-depletion phenotype requires Aurora kinase activity, we used two different small-molecule inhibitors of Aurora kinase activity: hesperadin and AKI-1 (refs 7,21,22). To minimize any effects of Aurora kinase inhibition on spindle forma-tion, cells already arrested in mitosis were treated with an Aurora kinase inhibitor for 1 h before analysis. In these cells, depletion of BubR1 was confirmed by immunofluorescence (see Supplementary Information, Fig. S1), kinetochores were concentrated in the centre of the spindle (Fig. 3a–b) and the fraction of bipolar spindles with a recognizable met-aphase plate increased from 30% (n = 100) to 85% (n = 100) (Fig. 3c). These data show that inhibition of Aurora kinase activity suppressed the chromosome misalignment phenotype in BubR1-depleted cells.

To test if the effect of Aurora kinase inhibition was due to stabilization of kinetochore–microtubule attachments, we again used three assays for stable attachments. In BubR1-depleted cells treated with an Aurora kinase inhibitor, cold-stable microtubule fibres were clearly attached to

kinetochores (Fig. 3f–g; also see Supplementary Information, Fig. S1). In these cells, 84% (±4% s.d.) of kinetochores had cold-stable microtubule fibres attached, compared with only 5% (±2% s.d.) with Aurora kinase active (>80 kinetochores counted per cell). Chromosomes aligned at the centre of the spindle had microtubule fibres extending from sister kine-tochores towards opposite poles (Fig. 3f). For some kinetochores near the poles, syntelic orientation was observed, with both sister kineto-chores attached to the same pole (Fig. 3f). The Aurora kinase inhibitors have not been shown to be specific for Aurora A compared with Aurora B, but we expected that the stabilization of kinetochore–microtubule attachments was due to loss of function of Aurora B, which localizes to centromeres. To test this, cells were depleted of both BubR1 and Aurora B by RNAi, shown by both immunofluorescence in single cells and western blot analysis (see Supplementary Information, Fig. S2). Loss of Aurora B function was consistent with the effects of the Aurora kinase inhibitors in BubR1-depleted cells, as many chromosomes aligned at the metaphase plate with cold-stable kinetochore–microtubule fibres attached (Fig. 3h). In independent assays for microtubule attachment, we found that Mad2 and p150 no longer localized to kinetochores of BubR1-depleted cells after inhibition of Aurora kinase activity (see Supplementary Information, Fig. S3), consistent with attachment of the kinetochore to microtubules.

As an additional test of kinetochore–microtubule attachments in BubR1-depleted cells, we examined the ability of kinetochore micro-tubules to drive chromosome movements during the transition from a monopolar to bipolar spindle. Spindles are monopolar in cells treated with the reversible Eg5 inhibitor monastrol, with chromosomes positioned all around the unseparated centrosomes23. The spindle becomes bipolar after removal of the inhibitor, and chromosomes bi-orient and align at the metaphase plate, a process that requires kinetochore–microtubule-dependent transport. Spindles were monopolar in BubR1-depleted cells treated with monastrol (M.A.L. and T.M.K., unpublished observations) and became bipolar after removal of monastrol (Fig. 3j–k). This finding suggests that all spindle microtubules are not perturbed after BubR1

0 00.5

Control Mad2 CENP-EBubR1

a b c de

f g h

Control Mad2 CENP-E BubR1

Tubulin

BubR1

Tubulin

Mad2

CENP-E

Tubulin

RNAi Control

010

2030

4050

60

0 0.1 0.2 0.3 0.4 0.5

Per

cent

age

ofki

neto

chor

es

Control

BubR1RNAi

Fractional distance along spindle axis

Kinetochore positionalong spindle axis

Una

ligne

d k

inet

ocho

res

(per

cent

age)

0

10

20

30

40

Figure 1 Depletion of BubR1 causes severe chromosome misalignment. (a–d) Hela cells were mock transfected (control) or tranfected with Mad2, BubR1 or CENP-E siRNA, and after 24 or 48 h, arrested in mitosis for 2 h (MG132, 10 µM) and processed for immunostaining (tubulin, green; kinetochores (CREST), red). Scale bar represents 3 µm. (e) Western blot analysis of knockdown. (f) Kinetochore positions were measured by the

distance from the nearest pole along the pole–pole axis, and normalized for pole–pole distance. (g) Fraction of unaligned kinetochores (position <0.2) was calculated for each RNAi. The average value (± s.e.m.) over multiple cells (n ≥ 5, >400 kinetochores) is shown. (h) Profiles of kinetochore position after control and BubR1 RNAi. Positions were binned in increments of 0.1.



NATURE CELL BIOLOGY VOLUME 7 | NUMBER 1 | JANUARY 2005 95

L E T T E R S

depletion, and that interpolar and astral microtubules are most probably functional and can drive centrosome separation. Chromosomes failed to align on bipolar spindles in BubR1-depleted cells (Fig. 3j), consistent with the absence of functional kinetochore–microtubule attachments. If spindles bipolarized in these cells in the presence of an Aurora kinase inhibitor, however, many chromosomes aligned at the metaphase plate, with a few chromosomes remaining near the pole with syntelic attach-ments (Fig. 3k). This result indicates that if both BubR1 and Aurora kinase functions are blocked, kinetochore–microtubule attachments form that allow chromosome movements. We also found that CENP-A and CENP-E localized to the kinetochore in BubR1-depleted cells, whereas Aurora B and MCAK localized to the inner centromere (Fig. 4e; also see Supplementary Information, Fig. S4), consistent with previous observations14,16. Together, these data indicate that kinetochore structure is intact and competent to form functional attachments capable of driv-ing chromosome movement after BubR1 depletion.

Suppression by Aurora kinase inhibition suggested that the kineto-chore–microtubule destabilization phenotype of BubR1 depletion might be caused by increased activity of Aurora kinase at kinetochores. To test this prediction, we selected a known kinetochore substrate to use as a reporter for kinetochore-specific Aurora kinase activity. MCAK is one known substrate, but its localization at the kinetochore and inner cen-tromere is regulated by Aurora kinase activity24, making a quantitative analysis of its phosphorylation state at kinetochores difficult. CENP-A is also a known substrate and a core kinetochore component whose

localization does not depend on phosphorylation by Aurora kinases25,26. We measured CENP-A phosphorylation at kinetochores by calculating the ratio of phospho-CENP-A to CREST antigens, which include CENP-A27. To control for the effects of unattached kinetochores in BubR1-depleted cells, experiments were performed both with and without taxol, which stabilizes microtubules28. In BubR1-depleted cells treated with taxol, kinetochores had associated microtubules (Fig. 4b), but sister kinetochores were not bi-oriented as in control cells. Taxol treatment also changed the distribution of chromosomes in BubR1-depleted cells so that kinetochores were positioned near each pole and excluded from the centre of the spindle (Fig. 4c). CENP-A phosphorylation in BubR1-depleted cells was increased by 60% (±4%) over control cells, and was not changed by the addition of taxol (Fig. 4d–f). Inhibition of Aurora kinase activity reduced CENP-A phosphorylation by >75% in both control and BubR1-depleted cells (data not shown). These data indicate that the balance of kinase to phosphatase activities at kinetochores is shifted in BubR1-depleted cells to favour CENP-A phosphorylation.

As an additional test of the effect of BubR1 depletion on CENP-A phosphorylation, we reduced phospho-CENP-A to low levels by inhi-bition of Aurora kinase activity. Removal of the inhibitor, which is reversible, activates Aurora kinase7. Thirty minutes after removal of the inhibitor, CENP-A phosphorylation was still low in control cells but markedly higher (5.9 ± 1.3-fold) in BubR1-depleted cells (Fig. 4g–k). Thus, phosphorylation of a known Aurora kinase substrate is increased at kinetochores in BubR1-depleted cells.

3

12

2

1

3

1

2

3

a

1

2

3

Control BubR1 RNAi

1

2

3

321

b c d

2

1

3

CREST Mad2

Control

e

3

1

2

CREST Mad2

BubR1 RNAi

f

3

1

2

CREST p150

3

1

2

Control

g

2

1

3

3

21

CREST p150

BubR1 RNAi

h

Figure 2 Depletion of BubR1 results in the loss of kinetochore–microtubule attachments. (a–d) Analysis of cold-stable microtubules in control (a, b) or BubR1-depleted (c, d) cells stained for tubulin (green), CREST (red) and DNA (blue). (b, d) Tubulin staining alone. Insets (1–3) detail kinetochore–microtubule interactions. (e, f) Mad2 localization in control (e) or BubR1-

depleted (f) cells (tubulin, green; Mad2, red; CREST, blue). Insets (1–3) show kinetochores without (e) or with (f) Mad2. (g–h) p150 localization in control (g) or BubR1-depleted (h) cells (p150, green; CREST, red; DNA, blue). Insets (1–3) show kinetochores without (g) or with (h) p150. Scale bar represents 3 µm. Insets show optical sections, 200% magnification.




L E T T E R S

Our data show that unlike other proteins known to interact directly with Cdc20 to inhibit the anaphase promoting complex, BubR1 is required for regulation of kinetochore–microtubule attachments as well as activation of the mitotic checkpoint (Fig. 4l). Changes in the normal function of proteins involved in checkpoint signalling, such as reduction in BubR1 expression9,10, are associated with human cancer, and small-molecule inhibitors of protein kinases that regulate attachment are being validated in animal models as promising therapeutic agents for cancer treatment13. Understanding the interplay between these pathways, which are directly linked through BubR1, has important consequences for developing new strategies to treat human disease.

METHODSsiRNA transfection and antibodies. Hela cells were cultured in growth medium, Dulbecco’s modified Eagle’s medium (DMEM; Invitrogen, Carlsbad, CA) with 10% fetal bovine serum (Sigma, St Louis, MO) and penicillin–streptomcyin (100 U ml–1

and 100 µg ml–1 respectively; Invitrogen), at 37 °C in a humidified atmosphere with 5% CO2. Cells were plated on glass coverslips in six-well dishes, and medium was changed to Opti-MEM I (Invitrogen) before transfection, performed using Oligofectamine following the manufacturers instructions (Invitrogen). siRNA duplexes targeting Mad2 (5′-AAGAGUCGGGACCACAGUUUA-3′), BubR1 (5′-AACGGGCAUUUGAAUAUGAAA-3′), CENP-E (SMARTpool, containing four pooled siRNA duplexes) or Aurora B (5′- AACGCGGCACUUCACAAUUGA-3′) were from Dharmacon Research (Lafeyette, CO). Buffer alone was used for mock transfection (control). Results for BubR1 depletion were confirmed using a different oligo (5′-AAAGAUCCUGGCUAACUGUUC-3′). Twenty-four hours (CENP-E) or 48 h (Mad2 and BubR1) after transfection, cells were either used for experiments or lysed for western blot analysis.

Antibodies used for western blots were: polyclonal antibodies against Mad2 (a gift from D. Compton, Dartmouth Medical School, Hanover, NH), BubR1 (a gift from W. Dai, New York Medical College, Valhalla, NY) or CENP-E (HX-1; a gift from T. Yen, Fox Chase Cancer Center, Philadelphia, PA); a monoclonal antibody against α-tubulin (DM1A; Sigma); and a monoclonal antibody against Aurora B (BD Transduction Laboratories, San Diego, CA). The same antibodies

1

2

1

32

0

10

20

30

40

50

0.0 0.1 0.2 0.3 0.4 0.5

Per

cent

age

ofki

neto

chor

es

Auora kinaseactiveAurora kinaseinhibited

1

2

34

31 421 2 43

Aurora kinase active

31

4

2

a b c

d e f gAurora kinase inhibited

0

20

40

60

80

100

Bip

olar

sp

ind

les

with

met

apha

se p

late

(per

cent

age)

Control

BubR1 RNAi

BubR1 RNAi,


Aurora kinase active Aurora kinase inhibited

1 2

h

i

j k1

3

2

Aurora kinaseinhibited

BubR1 and Aurora BRNAi

Monastrolarrest

Recovery Spindlepole

Chromosome pair

Microtubulefibres

Figure 3 Inhibition of Aurora kinase activity suppresses the BubR1-depletion phenotype. (a) A BubR1-depleted cell treated with hesperadin (20 nM, 1 h) and stained for tubulin (green), CREST (red) and DNA (blue). (b) Profiles of kinetochore positions along the spindle axis (calculated as in Fig. 1). (c) The fraction of cells containing bipolar spindles with a recognizable metaphase plate was determined. Bars show the average of two experiments. (d–h) Cold-stable microtubules in BubR1-depleted cells treated without (d, e) or with (f, g) hesperadin or depleted of Aurora B simultaneously (h), stained for tubulin (green), CREST (red) and DNA (blue). (e, g) Tubulin alone. Insets (1–4) show

individual kinetochores without (d) or with (f, h) cold-stable microtubules, some with syntelic orientation (f, insets 1, 4). Similar results were obtained with another Aurora kinase inhibitor (see Supplementary Information, Fig. S1). (i) Schematic of spindle bipolarization as cells recover from monastrol treatment. (j, k) After removal of monastrol, BubR1-depleted cells were incubated for 1 h without (j) or with (k) hesperadin and then processed for immunofluorescence. Tubulin (green), CREST (red) and DNA (blue). Insets (1–3) show individual kinetochores with syntelic (1, 3) or bi-oriented (2) attachments. Scale bar represents 3 µm. Insets show optical sections, 200% magnification.



NATURE CELL BIOLOGY VOLUME 7 | NUMBER 1 | JANUARY 2005 97

L E T T E R S

against BubR1, Aurora B and α-tubulin were used for immunofluorescence, as well as a polyclonal anti-XMad2, a monoclonal anti-p150glued (BD Transduction Laboratories) and a polyclonal anti-MCAK (a gift from C. Walczak, Indiana University, Bloomington, IN). Human CREST antiserum (a gift from W. Brinkley, Baylor College of Medicine, Houston, TX) was used to stain kinetochores. A polyclonal antibody against CENP-A phosphorylated at Ser-7 was a gift from D. Allis (Rockefeller University, New York, NY). Secondary antibodies conjugated to fluoroscein isothyocyante (FITC), rhodamine, Texas red or Cy5 were from Jackson ImmunoResearch (West Grove, PA).

Immunofluorescence. All cells were incubated for 2 h in growth medium with the proteasome inhibitor MG132 (10 µM; Sigma) before fixation, unless oth-erwise noted. To inhibit Aurora kinase activity, hesperadin (20 nM) or AKI-1 (1 µM) were added for the last hour (for details of these inhibitors, see ref. 7). For analysis of cold-stable microtubules, cells were incubated for 10 min on ice in L-15 medium (Invitrogen) with 20 mM Hepes at pH 7.3, then fixed for 10 min at room temperature with 3.7% formaldehyde in 100 mM Pipes at pH 6.8, 10 mM EGTA, 1 mM MgCl2 and 0.2% Triton X-100. For all other experiments, cells were fixed for 10 min in methanol at –20 °C. After fixation, cells were blocked in TBS with 0.1% Triton X-100 and 2% bovine serum albumin (BSA), and antibodies were diluted in the same medium. DNA was labelled with Hoechst 33342 (Sigma).

For the monastrol-washout experiment, cells were incubated for 2 h with MG132 and monastrol (100 µM). The medium was exchanged to remove

monastrol, and cells were incubated with MG132 with or without hesperadin for 1 h before fixation with methanol.

All images were acquired as Z-stacks with 0.2-µm spacing using a ×100, 1.35 NA objective on a DeltaVision Image Restoration Microscope (Applied Precision Instruments, Issaquah, WA and Olympus, Melville, NY), and proc-essed by iterative constrained deconvolution (SoftWoRx, Applied Precision Instruments). Maximal intensity projections of the entire Z-stack are shown, and optical sections show individual kinetochores more clearly (insets). Image analysis was performed using either SoftWoRx or Metamorph (Universal Imaging, Downington, PA) software.

For comparison of Mad2 and p150 at kinetochores, image processing for the insets showing individual kinetochores was identical for control and BubR1-depleted cells. Mad2 fluorescence intensity was quantified for individual kinetochores (data not shown), selected manually by CREST staining, in single deconvolved Z-sections. After subtraction of background intensity — estimated from regions of the cell without kinetochores — intensities were averaged over multiple kinetochores.

To count kinetochores with cold-stable microtubules attached, CREST staining was used to identify kinetochores in deconvolved image Z-stacks. Each kineto-chore was characterized as ‘attached’ or ‘unattached’, depending on whether a microtubule fibre ended at the kinetochore. Kinetochores for which the attach-ment state could not be clearly determined (that is, the microtubule density was too high) were not counted. An average of 97 kinetochore were counted per cell, with three cells analysed in each case.

l

BubR1

Mad2

Aurora BKinetochore

4

5

1

2

3

d

e

g i

jh

a b

0

10

20

30

40

50

0 0.1 0.2 0.3 0.4 0.5

Per

cent

age

ofal

l kin

etoc

hore

s Control

Control + taxol

BubR1 RNAi

BubR1 RNAi+ taxol

c

f

0.0

0.5

1.0

1.5

2.0

P-C

EN

P-A

/ c

rest

Control BubR1RNAi

- Taxol + Taxol

k

0.00.20.40.60.81.0 Control

BubR1RNAi

Control + taxol BubR1 RNAi + taxol


4

5

1

2

3P

-CE

NP

-A /

cre

st

Aurorakinase

inhibitor

Inhibitorremoved

Microtubulefibre

Anaphase promotingcomplex/Cyclosome

Destabilizeattachment

Figure 4 BubR1 depletion increases Aurora kinase-dependent CENP-A phosphorylation at kinetochores. (a, b) Control (a) and BubR1-depleted (b) cells were treated with taxol (10 µM, 1 h) and stained for tubulin (green) and CREST (red). Insets (optical sections, magnification 300%) show kinetochore-associated microtubules in BubR1-depleted cells. (c) Profile of kinetochore positions along the spindle axis, calculated as in Fig. 1, for cells treated with or without taxol. (d, e) Control (d) and BubR1-depleted (e) cells were stained for phospho-CENP-A (red) and CREST (blue). (f) Fluorescence intensities of phospho-CENP-A and CREST at kinetochores were summed over an entire cell, and the phospho-CENP-A/CREST ratio

was averaged over multiple cells (n = 5). The average values of at least two independent experiments are shown. (g, h) Control (g) and BubR1-depleted (h) cells were treated with hesperadin for 1 h and stained for phospho-CENP-A (red) and CREST (blue). (i, j) Control (i) and BubR1-depleted (j) cells were treated as in (g, h), except that hesperadin was removed after 1 h, and cells were allowed to recover for 30 min before fixation. (k) The phospho-CENP-A/CREST ratio was calculated for cells treated as in (g–j). (l) Schematic representation, highlighting the role of BubR1 as a link between the regulation of kinetochore–microtubule attachment and checkpoint activation. Scale bar represents 3 µm.




L E T T E R S

Analysis of kinetochore positions. Kinetochore positions were determined from maximal intensity projections of deconvolved images, using an automated analy-sis. Based on CREST staining, objects (representing kinetochores) were defined as regions of contiguous pixels with intensity above a threshold. For each object, an internal threshold of 75% of the maximum for that object was used to separate overlapping objects. An average of 90 kinetochores per cell were selected by this method. The position of each kinetochore was recorded as the object centroid, as well as the positions of the two spindle poles, determined from tubulin staining. For each kinetochore, the nearest point on the pole–pole line was determined, and the distance from that point to the nearest pole was calculated and normalized by the pole–pole distance. Misaligned kinetochores were defined as those with normalized distance <0.2. At least five cells (>400 kinetochores) were analysed for each RNAi experiment.

Quantification of phosphorylated CENP-A. Cells were incubated for 2 h with MG132 (10 µM), with taxol (10 µM) added for the last hour as indicated, then fixed and stained for phospho-CENP-A, CREST and tubulin. Cells for analysis were selected without knowledge of phospho-CENP-A staining. For each plane in a deconvolved Z-stack, CREST staining was used to define kinetochore regions, using an intensity threshold. The CREST and phospho-CENP-A intensities in these regions were summed over the entire cell. Background fluorescence, estimated from regions of the cell without kinetochores, was subtracted. The phospho-CENP-A/CREST ratio was calculated and averaged over multiple cells (typically, n = 5). To compare between independent experiments, the average phospho-CENP-A/CREST ratio for each experimental condition was normalized to control cells without taxol. Averages of at least two independent experiments are presented.

For experiments with taxol, individual kinetochores visibly associated with microtubules were also selected manually. The phospho-CENP-A/CREST ratio was calculated for each kinetochore and averaged over multiple kinetochores. A 50% increase in BubR1-depleted cells was measured (data not shown), consistent with the analysis of all kinetochores combined.

For experiments with hesperadin (Fig. 4g–k), cells were fixed after one of the following treatments: first, 1 h with MG132; second, same as the first, then 1 h with MG132 and hesperadin (20 nM); and third, same as the second, then removal of hesperadin and 30 min treatment with MG132 alone. The phospho-CENP-A/CREST ratio was calculated as described above and normalized to control cells with MG132 alone (treatment 1). The averages of two independent experiments are presented.

Note: Supplementary Information is available on the Nature Cell Biology website.

ACKNOWLEDGEMENTSWe thank A. North and the Rockefeller University Bioimaging facility. We thank D. Compton, W. Dai, T. Yen, C. Walczak, W. Brinkley and D. Allis for gifts of antibodies, and H. Funabiki for discussions. This work was supported by National Institutes of Health grant GM65933 (T.M.K.). M.A.L. is a Goelet fellow.

COMPETING FINANCIAL INTERESTSThe authors declare that they have no competing financial interests.

Received 05 October 2004; accepted 18 November 2004Published online at http://www.nature.com/naturecellbiology.

1. Lengauer, C., Kinzler, K. W. & Vogelstein, B. Genetic instabilities in human cancers. Nature 396, 643–649 (1998).

2. Sudakin, V., Chan, G. K. & Yen, T. J. Checkpoint inhibition of the APC/C in HeLa cells is mediated by a complex of BUBR1, BUB3, CDC20, and MAD2. J. Cell Biol. 154, 925–936 (2001).

3. Tang, Z., Bharadwaj, R., Li, B. & Yu, H. Mad2-independent inhibition of APCCdc20 by the mitotic checkpoint protein BubR1. Dev. Cell 1, 227–237 (2001).

4. Fang, G. Checkpoint protein BubR1 acts synergistically with Mad2 to inhibit anaphase-promoting complex. Mol. Biol. Cell 13, 755–766 (2002).

5. Biggins, S. & Murray, A. W. The budding yeast protein kinase Ipl1/Aurora allows the absence of tension to activate the spindle checkpoint. Genes Dev. 15, 3118–3129 (2001).

6. Tanaka, T. U. et al. Evidence that the Ipl1–Sli15 (Aurora kinase-INCENP) complex promotes chromosome bi-orientation by altering kinetochore-spindle pole connections. Cell 108, 317–329 (2002).

7. Lampson, M. A., Renduchitala, K., Khodjakov, A. & Kapoor, T. M. Correcting improper chro-mosome-spindle attachments during cell division. Nature Cell Biol. 6, 232–237 (2004).

8. Cahill, D. P. et al. Mutations of mitotic checkpoint genes in human cancers. Nature 392, 300–303 (1998).

9. Shin, H. J. et al. Dual roles of human BubR1, a mitotic checkpoint kinase, in the monitoring of chromosomal instability. Cancer Cell 4, 483–497 (2003).

10. Shichiri, M., Yoshinaga, K., Hisatomi, H., Sugihara, K. & Hirata, Y. Genetic and epige-netic inactivation of mitotic checkpoint genes hBUB1 and hBUBR1 and their relation-ship to survival. Cancer Res. 62, 13–17 (2002).

11. Wang, X. et al. Significance of MAD2 expression to mitotic checkpoint control in ovarian cancer cells. Cancer Res. 62, 1662–1668 (2002).

12. Hanks, S. et al. Constitutional aneuploidy and cancer predisposition caused by biallelic mutations in BUB1B. Nature Genet. 36, 1159–1161 (2004).

13. Harrington, E. A. et al. VX-680, a potent and selective small-molecule inhibitor of the Aurora kinases, suppresses tumor growth in vivo. Nature Med. 10, 262–267 (2004).

14. Kops, G. J., Foltz, D. R. & Cleveland, D. W. Lethality to human cancer cells through massive chromosome loss by inhibition of the mitotic checkpoint. Proc. Natl Acad. Sci. USA 101, 8699–8704 (2004).

15. Li, Y. & Benezra, R. Identification of a human mitotic checkpoint gene: hsMAD2. Science 274, 246–248 (1996).

16. Chan, G. K., Jablonski, S. A., Sudakin, V., Hittle, J. C. & Yen, T. J. Human BUBR1 is a mitotic checkpoint kinase that monitors CENP-E functions at kinetochores and binds the cyclosome/APC. J. Cell Biol. 146, 941–954 (1999).

17. Weaver, B. A. et al. Centromere-associated protein-E is essential for the mammalian mitotic checkpoint to prevent aneuploidy due to single chromosome loss. J. Cell Biol. 162, 551–563 (2003).

18. Rieder, C. L. The structure of the cold-stable kinetochore fiber in metaphase PtK1 cells. Chromosoma 84, 145–158 (1981).

19. King, J. M., Hays, T. S. & Nicklas, R. B. Dynein is a transient kinetochore component whose binding is regulated by microtubule attachment, not tension. J. Cell Biol. 151, 739–748 (2000).

20. Waters, J. C., Chen, R. H., Murray, A. W. & Salmon, E. D. Localization of Mad2 to kinetochores depends on microtubule attachment, not tension. J. Cell Biol. 141, 1181–1191 (1998).

21. Ditchfield, C. et al. Aurora B couples chromosome alignment with anaphase by targeting BubR1, Mad2, and Cenp-E to kinetochores. J. Cell Biol. 161, 267–280 (2003).

22. Hauf, S. et al. The small molecule Hesperadin reveals a role for Aurora B in correcting kinetochore-microtubule attachment and in maintaining the spindle assembly check-point. J. Cell Biol. 161, 281–294 (2003).

23. Mayer, T. U. et al. Small molecule inhibitor of mitotic spindle bipolarity identified in a phenotype-based screen. Science 286, 971–974 (1999).

24. Andrews, P. D. et al. Aurora B regulates MCAK at the mitotic centromere. Dev. Cell 6, 253–268 (2004).

25. Zeitlin, S. G., Shelby, R. D. & Sullivan, K. F. CENP-A is phosphorylated by Aurora B kinase and plays an unexpected role in completion of cytokinesis. J. Cell Biol. 155, 1147–1157 (2001).

26. Kunitoku, N. et al. CENP-A phosphorylation by Aurora-A in prophase is required for enrichment of Aurora-B at inner centromeres and for kinetochore function. Dev. Cell 5, 853–864 (2003).

27. Earnshaw, W. C. & Rothfield, N. Identification of a family of human centromere proteins using autoimmune sera from patients with scleroderma. Chromosoma 91, 313–321 (1985).

28. McEwen, B. F., Heagle, A. B., Cassels, G. O., Buttle, K. F. & Rieder, C. L. Kinetochore fiber maturation in PtK1 cells and its implications for the mechanisms of chromosome congression and anaphase onset. J. Cell Biol. 137, 1567–1580 (1997).



S U P P L E M E N TA RY I N F O R M AT I O N

WWW.NATURE.COM/NATURECELLBIOLOGY 1

31

2

12

3

BubR1 RNAi BubR1 RNAi + AKI-1Control

1

2

3

CREST BubR1

2

3

1

2

3

1

CREST BubR11

2

1

2

3

3

CREST BubR1

31 2

BubR1 RNAi BubR1 RNAi + AKI-1

A B C

G H

1

2

3

4

1 2

3

4

CENP-E RNAi

I

D E F

merge

Figure S1 Inhibition of Aurora kinase activity with AKI-1 suppresses the BubR1 depletion phenotype. (A-F) Control (A, D) or BubR1-depleted (B-C, E-F) cells were treated without (A-B, D-E) or with (C, F) AKI-1 (1 µM, 1 hr) and stained for tubulin (green), CREST (red), and BubR1 (blue). (D-F) BubR1 staining alone. BubR1 staining is apparent in control (A, D) but not BubR1-depleted (B-C, E-F) cells, both cytoplasmic and on kinetochores, shown in insets (1-3). (G-H) Cold-stable microtubules in BubR1-depleted cells

treated without (G) or with (H) AKI-1, stained for tubulin (green), CREST (red) and DNA (blue). (H) Insets (1-3) show individual kinetochores with cold-stable microtubules. (I) Cold-stable microtubules in a CENP-E depleted cell, stained for tubulin (green), CREST (red) and DNA (blue). Insets show aligned (2, 3) and polar (1, 4) chromosomes with and without cold-stable kinetochore-microtubules, respectively. Insets, magnification 200%. Scale bar, 3 µm.

© 2005 Nature Publishing Group


2 WWW.NATURE.COM/NATURECELLBIOLOGY

GB C

D E F

AControl BubR1 RNAi BubR1 and Aurora B RNAi

BubR1

Aurora B

Tubulin

BubR1Aurora B

- + ++-

-- +

siRNA

Figure S2 Depletion of BubR1 and Aurora B by siRNA. (A-F) Cells were mock transfected (A, D) or with BubR1 siRNA (B, E) or BubR1 and Aurora B siRNA together (C, F). After 48 hrs cells were arrested in mitosis for 2 hrs

and processed for immunofluorescence. (A-C) Tubulin (green), DNA (blue), and Aurora B (red) staining. (D-F) BubR1 staining in the same cells. Scale bar, 3 µm. (G) Western blot analysis of knockdown.

1

2

3

3

21

CREST Mad21

2

3

Aurora kinase active

CREST Mad2

2

3

1

2

3

1

2

1

3

CREST p1501

3

2

1

3

2

CREST p150

Aurora kinase inhibited Aurora kinase active Aurora kinase inhibitedA B C D

Figure S3 Inhibition of Aurora kinase activity results in loss of Mad2 and p150 from kinetochores in BubR1-depleted cells. BubR1-depleted cells were treated without (A, C) or with (B, D) hesperadin (20 nM, 1 hr). (A-B) Mad2 localization (tubulin: green; Mad2:red; CREST:blue). Insets (1-3) show

individual kinetochores with (A) or without (B) Mad2. (C-D) p150 localization (p150:green; CREST:red; DNA, blue). Insets (1-3) show individual kinetochores with (C) or without (D) p150. Insets, magnification 200%. Scale bar, 3 µm.



WWW.NATURE.COM/NATURECELLBIOLOGY 3

12

3

BubR1 RNAiControl

1

3

2

1

2

3

1

23

1

2

3

12

3

Auo

ra B

CE

NP

-EM

CA

K

A B

E F

C D

Figure S4 Aurora B, MCAK, and CENP-E localization in BubR1-depleted cells. Control (A, C, E) or BubR1-depleted (B, D, F) cells were stained for either Aurora B (A-B), MCAK (C-D), or CENP-E (E-F), shown in green, with

CREST (red) and DNA (blue). Insets (1-3) show individual kinetochores with Aurora B (B) or MCAK (D) at the inner centromere, or CENP-E (F) at the kinetochore. Insets, magnification 200%. Scale bar, 3 µm.


Documents

The human mitotic checkpoint protein BubR1 regulates ......15% (±4%) for CENP-E depletion and 33% (±4%) for BubR1 depletion (Fig. 1a–d, g). Thus, depletion of Ma d2 had no chromosome