Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://cellmotility.genebee.msu.ru/html/articles/Yuko2006_gtc_990.pdf Äàòà èçìåíåíèÿ: Mon Feb 4 12:51:54 2008 Äàòà èíäåêñèðîâàíèÿ: Mon Oct 1 20:14:49 2012 Êîäèðîâêà:

Y CLASPs in mitosis Or iginal Article 8 Blackwi-Kiy lishing Ltd 11MimorCells osue et al. ? 2006 toUSA lishing Inc © 1365-2443 Genes GTC Malden, Pub Blackwell ell Pub

Mammalian CLASPs are required for mitotic spindle organization and kinetochore alignment

Yuko Mimori-Kiyosue1,*, Ilya Grigoriev2,3, Hiroyuki Sasaki1,4, Chiyuki Matsui1, Anna Akhmanova3, Shoichiro Tsukita5,6 and Ivan Vorobjev7,*
1 2

KAN Research Institute, Kyoto Research Park, Shimogyo-ku, Kyoto 600-8815, Japan Department of Cell Biology and Histology, Moscow State University, Vorobjevi Gory, Moscow, 119992, Russia 3 MGC Department of Cell Biology and Genetics, Erasmus Medical Center, 3000 DR Rotterdam, The Netherlands 4 Institute of DNA Medicine, Jikei University School of Medicine, Minato-ku, Tokyo 105-8461, Japan 5 Department of Cell Biology, Faculty of Medicine, Kyoto University, Sakyo-ku, Kyoto 606-8315, Japan 6 Solution Oriented Research for Science and Technology, Japan Science and Technology Corporation, Sakyo-ku, Kyoto 606-8501, Japan 7 Laboratory of Cell Motility, A. N. Belozersky Institute, Moscow State University,Vorobjevi Gory, Moscow, 119992, Russia

CLASP1 and CLASP2 are homologous mammalian proteins, which associate with the ends of growing microtubules, as well as the cell cortex and the kinetochores of mitotic chromosomes. Previous studies have shown that in inter phase cells CLASPs can attach microtubule plus ends to the cortex and stabilize them by repeatedly rescuing them from depolymer ization. Here we show that CLASP1 and 2 play similar and redundant roles in organizing the mitotic apparatus in HeLa cells. Simultaneous depletion of both CLASPs causes mitotic spindle defects and a significant metaphase delay, which often results in abnor mal exit from mitosis. Metaphase delay is associated with decreased kinetochore tension, increased kinetochore oscillations and more rapid microtubule growth. We show that the association of CLASP2 with the kinetochores relies on its C-ter minal domain, but is independent of microtubules or association with CLIP-170. We propose that CLASPs exhibit at the kinetochores an activity similar to that at the cortex, providing apparent stabilization of microtubules by locally reducing the amplitude of growth/shortening episodes at the microtubule ends. This local stabilization of microtubules is essential for the for mation of nor mal metaphase spindle, completion of anaphase and cytokinesis.

Introduction
The mitotic spindle is a highly dynamic specialized microtubule ar ray required for chromosome separation. Segregation of sister chromatids depends on the interaction of their kinetochores with microtubules. A growing body of evidence suggests that the chromosome­microtubule interaction dur ing mitosis involves numerous proteins associated with microtubules and/or with kinetochores (Cleveland et al. 2003). Since the kinetochores interact with microtubule plus ends, it is not sur pr ising that many proteins acting at the kinetochore-microtubule interface belong to microtubule plus-end-tracking proteins (+TIPs), a group of factors specifically associated with ends of growing microtubules (Schuyler & Pellman 2001; Maiato
Communicated by: Shuh Narumiya *Correspondence: E-mail: y-kiyosue@kan.gr.jp; ivorobjev@mail.ru

et al. 2004; Akhmanova & Hoogenraad 2005). These proteins include EB1, CLIP-170, dynein and dynactin as well as CLASPs, which are the focus of the cur rent study. CLASPs are evolutionary conserved proteins, which are represented in mammals by two homologs, CLASP1 and 2 (Inoue et al. 2000; Lemos et al. 2000). They are involved in spatial organization of microtubule networks both in inter phase and in mitosis. In mammalian fibroblasts, CLASPs stabilize microtubules directed to the leading edge (Akhmanova et al. 2001); this function probably depends on the capacity of these proteins to rescue microtubules in the vicinity of certain cortical regions (Mimor i-Kiyosue et al. 2005). The budding yeast CLASP counter par t, Stu1p, participates in the for mation of the mitotic spindle (Yin et al. 2002). The human CLASP1, its Xenopus homolog Xorbit/CLASP, the Drosophila homolog, MAST/Orbit, and the C. elegans homolog, CLASPcls-2 (R106.7) all play a role in the proper mitotic progression (Gonczy et al.
Genes to Cells (2006) 11, 845­ 857

DOI: 10.1111/j.1365-2443.2006.00990.x © 2006 The Authors Jour nal compilation © 2006 by the Molecular Biology Society of Japan/Blackwell Publishing Ltd.

845


Y Mimori-Kiyosue et al.

2000; Maiato et al. 2002, 2003; Hannak & Heald 2006). Using Drosophila mutants and RNA interference (RNAi) approach, MAST/Orbit was shown to be necessary for maintaining bipolar ity of the mitotic spindle, for the microtubule-kinetochore attachment and chromosome congression (Maiato et al. 2002). CLASPs are components of the outer kinetochore layer; however, CLASP depletion does not disrupt the targeting of other proteins that might affect kinetochore-microtubule interactions, such as CLIP-170 or dynein (Maiato et al. 2002, 2003). In insect cells, MAST/Orbit is an essential component of the spindle flux machinery: it is required for the polymer ization of the microtubules attached to kinetochore (Maiato et al. 2005). In addition, analysis of the hypomor phic MAST/Orbit mutants revealed a function for this protein in cytokinesis (Inoue et al. 2004). In C. elegans, CLASPcls-2 is targeted to the kinetochore by CENP-F-like proteins HCP-1/2 and is required for sister chromatid bior ientation (Cheeseman et al. 2005). In Xenopus meiotic egg extracts, Xorbit/ CLASP is required for nor mal spindle for mation and chromosome congression, as well as microtubule stabilization dur ing anaphase (Hannak & Heald 2006). The precise role of CLASPs, especially of CLASP2, in mammalian mitosis remains to be elucidated. In this study we used an RNAi approach combined with live cell imaging and immunostaining to examine the spatial distr ibution and function of both CLASP homologs in mitotic human cells. We show that CLASP1 and 2 have similar and redundant roles in maintaining mitotic spindle. CLASP2 has the same mitotic localization as CLASP1. Its binding to kinetochores occurs through the C-ter minal CLIP-170-binding domain, but is independent of CLIP-170. Partial depletion of the two CLASPs results in a metaphase delay accompanied by increased kinetochore oscillations. We propose that dur ing mitosis CLASPs act as local modulators of microtubule dynamic instability responsible for stabilization of kinetochore fibers and generation of pulling forces applied to chromosomes dur ing metaphase.

associated with the spindle poles, microtubule plus ends within the spindle and kinetochores. In anaphase and telophase, CLASP2 accumulated at the spindle midzone, while dur ing cytokinesis it was present in the midbody. Previous analysis of CLASP1 protein has established that it binds to mitotic kinetochores through its Cter minus (Maiato et al. 2003; Hannak & Heald 2006).We have obtained similar results with the CLASP2 deletion mutants (Fig. 2A,B). GFP-CLASP2-C, but not GFPCLASP2 C associated with kinetochores, also when the microtubules were depolymer ized with nocodazole (Fig. 2B and data not shown). The C-ter minal domain of CLASPs is responsible for their association with CLIP-170 (Akhmanova et al. 2001). CLIP-170 also displays kinetochore localization, which depends on the presence of dynactin complex and can be disrupted by over-expression of the dynactin subunit dynamitin (Dujardin et al. 1998). GFP-CLASP2 could still bind to kinetochores when p50/dynamitin was over-expressed and endogenous CLIP-170 was removed from the kinetochores (Fig. 2C), while GFP-CLASP2 was displaced from kinetochores by over-expression of DsRed2CLASP2-C fusion without affecting the kinetochore localization of CLIP-170 (Fig. 2D). Therefore, we conclude that although CLASP2 associates with kinetochores through its CLIP-binding C-ter minal domain, this association is independent of CLIP-170.
CLASP1 and 2 are required for mitotic progression

Results
Mitotic localization of CLASP2

The localization of CLASP1 dur ing mitosis has been descr ibed previously (Maiato et al. 2003). We have analyzed the distr ibution of CLASP2 dur ing mitosis, by antibody staining and by expressing low levels of GFPCLASP2 in HeLa cells (Fig. 1 and data not shown). We found that CLASP2 localization was very similar to that of CLASP1: starting from prophase to anaphase, it
846
Genes to Cells (2006) 11, 845­ 857

To deplete CLASPs from human cultured cells we used siRNA-mediated knockdown. Two pairs of siRNAs, specific for CLASP1 and CLASP2 (CLASP1#A or #B and CLASP2#A or #B) were character ized in detail previously (Mimori-Kiyosue et al. 2005).Three days after transfection with these siRNAs, the levels of CLASP1 and CLASP2 were reduced by 70% (Mimor i-Kiyosue et al. 2005). In agreement with these results, all CLASP1/2specific signals in mitotic cells were highly reduced after both CLASPs were knocked down (Fig. 2E,F), confir ming the efficiency of CLASP1/2-directed siRNAs. Down-regulation of a single CLASP with CLASP1 or CLASP2-specific siRNAs had no visible effect on the mitotic progression (Fig. 3A). However simultaneous down-regulation of both CLASPs caused a considerable increase in the proportion of mitotic cells (Fig. 3A), suggesting that CLASP1 and CLASP2 act in mitosis as a common pool. When the progression of mitosis was observed by phase contrast microscopy (Fig. 3B ­E, Supplemental Movie S1), the major ity of CLASP-depleted cells was unable to exit mitosis in time (within 60 min after mitosis onset detected by the rounding up of the

© 2006 The Authors Jour nal compilation © 2006 by the Molecular Biology Society of Japan/Blackwell Publishing Ltd.


CLASPs in mitosis

Figure 1 Localization of GFP-CLASP2 in different phases of mitosis. HeLa cells, transiently expressing low levels of GFP-CLASP2, were stained for -tubulin, CENP-A and the DNA. GFP-CLASP2 was localized at the spindle and kinetochores, and dur ing late mitosis, transfer red to the central spindle (ar rows in D) and the midbody (box in E). Bar 5 µm.

cell). Instead, CLASP-depleted cells were delayed in mitosis while for ming different abnor mal figures (Fig. 3F and see below).The duration of the rounded-up stage in mitotic cells with lowered CLASP levels increased nearly tenfold (from 30 min to up to 5 h). Yet some cells were able to undergo cytokinesis, exit from mitosis and proceed into the inter phase.

Analysis of the mitotic progression after CLASP knockdown

Staining of cells for microtubules revealed bipolar spindles as well as several mitotic defects. When bipolar spindles were for med in CLASP-depleted cells, in many cases their pole-to-pole distance was significantly shorter
Genes to Cells (2006) 11, 845­ 857

© 2006 The Authors Jour nal compilation © 2006 by the Molecular Biology Society of Japan/Blackwell Publishing Ltd.

847


Y Mimori-Kiyosue et al.

Figure 2 CLASP association with the kinetochores. (A) Schematic representation of CLASP2 and the relevant deletion mutants. (B­ D) Kinetochore labeling in nocodazole-treated Xenopus A6 cells. B. Cells were transiently transfected with GFP-CLASP2-C and stained with antibodies against p150Glued. (C,D) Cells, stably expressing low levels of GFP-CLASP2, were transiently transfected with DsRedp50/dynamitin (C) or DsRed-CLASP2-C (D), and stained for CLIP-170. Cells expressing high levels of DsRed fusion proteins are shown. Note that the kinetochore localizations of DsRed-CLASP2-C and DsRed-p50/dynamitin were undetectable due to their highlevel expression in the cytoplasm. Bars, 5 µm. (E,F) HeLa cells, transfected either with the control siRNA or the CLASP1+2#B siRNAs, were treated with 10 µM nocodazole for 1 h, and stained with Hoechst 33342, anti-centromere antibodies (CREST), anti-p150 and antibody #402 against CLASP1. Bar (E) 5 µm and (F) 1 µm.

than that in control cells (Fig. 4A,B). Besides, multipolar and severely disorganized spindles were observed (Fig. 4C­I). Staining of cells for DNA showed that after knockdown of both CLASPs a larger number of cells were car r ying scattered chromosomes in a prometaphase-like manner (Fig. 4J).
848
Genes to Cells (2006) 11, 845­ 857

To observe directly the for mation of mitotic spindles and centromere movements, we have followed mitotic division in live HeLa cells, which were stably expressing GFPtubulin or GFP-Aurora B, a centromere protein, three days after transfection with either control siRNA or the CLASP1+2#B siRNA combination (Fig. 5,

© 2006 The Authors Jour nal compilation © 2006 by the Molecular Biology Society of Japan/Blackwell Publishing Ltd.


CLASPs in mitosis

Figure 3 Mitotic abnor malities after CLASP knockdown. (A) Percentage of mitotic cells calculated using phase contrast images of live HeLa cultures, 72 h after transfection with the indicated siRNAs. (B­E) Examples of mitotic progression in control (B) and CLASP depleted cells (C­E) observed by phase contrast microscopy. In control cells, mitosis was completed in 38.5 ± 17.0 min (n = 122). In CLASP depleted cells, delay in mitotic progression was frequently observed (C) (> 90 min, on average 123.8 ± 187.3 min (n = 150)). In addition, some cells exited from mitosis without detectable cytokinesis (D) and some daughter cells fused immediately after cytokinesis (E). Occasionally, cells undergoing abnor mal mitosis died dur ing mitosis or shortly after the exit from mitosis. The frequencies of the observed phenotypes are plotted in F. 122 control cells and 150 CLASP knockdown cells were analyzed. Bar 10 µm.

Supplemental Movies S2, S3). In agreement with previously reported observations (Maiato et al. 2002) some spindles had collapsed after initial for mation and lost bipolar ity, while others remained bipolar, yet were shorter than metaphase spindles in control cells (Supplemental Movie S2). Some spindles underwent a transition from an initially bipolar configuration to a multipolar state (Fig. 5C, Supplemental Movie S2). Consistently, in many cells with reduced CLASPs, the centromeres visualized with GFP-AuroraB were almost aligned around the equator ial region, but later disorganized again dur ing a prolonged ar rest (Fig. 5D, Supplemental Movie S3).

Based on these observations, the mitotic progression in CLASP-depleted cells was classified into the following types: (i) hypomor phic phenotype; character ized by bipolar or ientation of chromosomes, prolonged metaphase, and division into two nor mal daughter cells, (ii) altered mitosis; bipolar or ientation of chromosomes was apparently lost, however, after a significant delay the cell entered cytokinesis. These cells occasionally failed to divide nor mally, (iii) strongly altered mitosis; metaphase spindle had collapsed and chromosomes were gathered in the center of the cell or scattered. These failures were often preceded by rear rangement of the spindle into a
Genes to Cells (2006) 11, 845­ 857

© 2006 The Authors Jour nal compilation © 2006 by the Molecular Biology Society of Japan/Blackwell Publishing Ltd.

849


Y Mimori-Kiyosue et al.

Figure 4 Abnor mal configuration of mitotic spindles after CLASP knockdown. (A,B) Pole-to-pole distance in bipolar spindles of HeLa cells, stained for -tubulin 72 h after transfection with the indicated siRNAs (A). In (B) the mean is indicated with a red line, mean ± SEM with a green box, and whiskers delineate 99% confidence interval. The difference between the control and CLASP knockdown values is statistically significant (P < 0.001). (C­F) Spindle structures observed in CLASP depleted cells. (C) Bipolar, (D) multipolar, (E) monopolar, which includes nor mal early prometaphase cells, (F) diffuse, abnor mal configuration without detectable polar structure. Cells were stained for DNA (blue), -tubulin (green) and CENP-A (red). (G,H) Example of a multipolar spindle in a CLASP-depleted cell, stained for -tubulin (G; red in H), DNA (blue) and -tubulin (green). Bar 5 µm. (I) Relative proportion of the mitotic spindles, shown in C-F, in HeLa cells 24, 48 or 72 h after transfection with the indicated siRNAs. Cells in nor mal early prometaphase and cells having abnor mal monopolar configuration were included in the "monopolar" category. ( J) Relative proportion of mitotic HeLa cells, 24, 48 or 72 h after transfection with the indicated siRNAs, in different phases of mitosis, based on DNA staining. Cells with scattered chromosomes were included in the "prometaphase-like" category.

multipolar configuration with gamma-tubulin located in numerous mitotic poles (Fig. 4G,H). Cells with strongly altered phenotype never exit mitosis in a nor mal way. The cells of type (ii) and (iii) often underwent asymmetr ic divisions or died dur ing or soon after the exit
850
Genes to Cells (2006) 11, 845­ 857

from mitosis (Supplemental Movie S1, Fig. 3F). Besides, telophase cells often failed to cleave or fused back immediately after an incomplete division (Fig. 3E­F), possibly due to the presence of lagging chromosomes. Alter natively, since CLASPs localize to the spindle midzone

© 2006 The Authors Jour nal compilation © 2006 by the Molecular Biology Society of Japan/Blackwell Publishing Ltd.


CLASPs in mitosis

Figure 5 Analysis of mitotic progression in live HeLa cells. Time-lapse images of HeLa cells, stably expressing GFP--tubulin or GFPAurora B, 72 h after transfection with control (A,B) or CLASP1+2#A siRNA oligos (C,D). Cell cycle phase is indicated on the lower r ight of the images (i, inter phase; pm, prometaphase; m, metaphase; ml, metaphase-like; t, telophase; ck, cytokinesis; ar, ar rested). Lines indicate the outlines of cells. Time (min) is indicated on the upper r ight of each image. Bar 10 µm.

(Fig. 1D, Maiato et al. 2003), completion of cytokinesis may be compromised due to defects in for mation of stable microtubule bundles at the central spindle, as observed in MAST/Orbit-mutated Drosophila sper matocytes (Inoue et al. 2004).
CLASP depletion affects the mitotic microtubule dynamics and tension at the bi-or iented metaphase kinetochores

To understand the reason for metaphase delay after knockdown of CLASPs we further analyzed cells displaying hypomor phic phenotype. Electron microscopy analysis of the siRNA-treated mitotic cells showed nor mal bipolar or ientation of the chromosomes after CLASP knockdown (Fig. 6A,B and data not shown). Kinetochores kept their tr ilaminar structure and were associated with microtubule ends in the nor mal way (Fig. 6C­F). The number of microtubules attached to a kinetochore observed within a 0.1 µm section, was almost the same in control cells (4.5 ± 1.3 per kinetochore, 68 kinetochores in 24 cells) and CLASP-depleted cells (3.7 ± 1.0 per kinetochore, 55 kinetochores, 20 cells). Thus the spatial organization of the kinetochoremicrotubule interaction appeared to be unchanged.

However, in such CLASP-depleted cells chromosomes failed to align properly; instead they were scattered around the equator ial plane (Fig. 6B, ar row). To get an insight into the temporal organization of the spindle we have analyzed the dynamics of microtubule plus ends and centromeres using HeLa cells expressing monomer ic red fluorescent protein (mRFP)-tagged Aurora B and EB3-GFP, a +TIP protein (Stepanova et al. 2003). High time resolution two color videos have shown that after the knockdown of both CLASPs microtubules and centromeres remained highly dynamic (Supplemental Movie S4, Fig. 7A,B). Transient accumulation of EB3-GFP at the kinetochores, moving away from the spindle poles, similar to that descr ibed for EB1 in PtK1 cells (Tir nauer et al. 2002), could still be observed after CLASP knockdown (data not shown). The rate of microtubule growth in mitotic cells (measured mainly from the astral microtubules) increased from 21.0 ± 7.1 µm/min in control cells (451 microtubule ends in 5 cells) to 26.5 ± 6.7 µm/min (217 microtubule ends in 4 cells) after the knockdown with CLASP1 +2#B siRNAs. This difference was statistically significant (P < 0.001, Kolmogorov­Smir nov two sample test). The increase in microtubule growth rate dur ing mitosis was fully in line with our obser vations in
Genes to Cells (2006) 11, 845­ 857

© 2006 The Authors Jour nal compilation © 2006 by the Molecular Biology Society of Japan/Blackwell Publishing Ltd.

851


Y Mimori-Kiyosue et al.

Figure 6 Kinetochore-microtubule attachment after CLASP knockdown. (A,B) Electron micrographs of the metaphase plates in HeLa cells 72 h after transfection with (A) the control or (B) CLASP1+2#B siRNAs. Bundles of kinetochore microtubules are visible on both sides of the chromosomes. Ar row in (B) points a chromosome shifted from the metaphase plate. (C­F) Electron micrographs, showing kinetochore-microtubule attachment in HeLa cells, 72 h after transfection with the control (C,D) or CLASP1+2#B (E,F) siRNAs. (D,F) enlarged portions of the boxed regions in (C,E) respectively. Bars 1 µm (A,B,C,E) and 200 nm (D,F).

inter phase cells, where the diminished microtubule stability, caused by the lack of CLASPs, resulted in an increased pool of soluble tubulin and a more rapid microtubule polymer ization (Mimor i-Kiyosue et al. 2005). The amplitude and the speed of centromere oscillations around the equator ial plate along the pole-to-pole axis after CLASP knockdown significantly increased (Fig. 7C,D). Since kinetochore movements within the spindle are coupled to microtubule polymer ization and depolymer ization (for review see Inoue & Salmon 1995), these results point to a change in the microtubule dynamics at the tips of the kinetochore fiber s. Immunofluorescent staining of kinetochores in CLASP knockdown cells indicated that after a metaphaselike ar rangement of chromosomes was achieved, the distances between bi-or iented kinetochores of the two sister chromatids were shorter than in control metaphase cells (Fig. 7E­H), indicating a decreased tension on the kinetochores (Waters et al. 1996). BubRI, a spindle
852
Genes to Cells (2006) 11, 845­ 857

checkpoint protein (for reviews see Cleveland et al. 2003; Lew & Burke 2003), remained accumulated at the kinetochores of metaphase-like cells after CLASP knockdown (Fig. 7G). This observation is in agreement with the data obtained in insect cells (Maiato et al. 2002), indicating that the lack of CLASPs does not affect the maintenance of the spindle checkpoint.

Discussion
In this study we have used RNAi approach to reduce the levels of CLASP1 and CLASP2 in HeLa cells. Significant mitotic defects were only present when combinations of siRNAs, specific for the two CLASPs, were transfected simultaneously, and not when cells were treated separately with CLASP1 or CLASP2-directed or control duplexes. The two CLASP isofor ms, expressed in HeLa cells (CLASP1 and CLASP2), are approximately 77% similar and the two proteins display a virtually identical

© 2006 The Authors Jour nal compilation © 2006 by the Molecular Biology Society of Japan/Blackwell Publishing Ltd.


CLASPs in mitosis

Figure 7 Centromere dynamics and chromosome attachment defects after CLASP knockdown. (A,B) Live imaging of metaphase spindles of HeLa cells, expressing EB3-GFP (green) and mRFP-AuroraB (red) 72 h after transfection with the control or CLASP1+2#B siRNAs. Single image is shown on the left, and a kymograph of the indicated portion of the image (position vs. time, time interval 3 s) on the r ight. Bar 5 µm. (C,D) Distr ibution of distances from kinetochores to the equator ial plate and instantaneous displacements of kinetochores, relative to the equator ial plate, based on time-lapse imaging of EB3-GFP and mRFP-AuroraB-expressing HeLa cells, 72 h after transfection with the control or CLASP1+2#A and B siRNAs. Eight control cells (40 kinetochores, n = 536) and seven CLASP knockdown cells (35 kinetochores, n = 410) were analyzed; the differences between mean distances and mean instantaneous displacements are statistically significant (P < 0.001). (E­G) Mitotic spindles of HeLa cells, 72 h after transfection with the control or CLASP1+2#B siRNAs stained with ACA (red), anti-BubRI (green) and anti--tubulin (blue). Inset is 2â magnified image of the boxed area. Bar 5 µm. (H) Distances between kinetochores of the sister chromatids, measured from the positions of CREST dots, in early prometaphase control cells (n = 65, 3 cells), metaphase control cells (n = 214, 6 cells) and metaphase-like CLASP-depleted cells (n = 200, 9 cells). Plots are constructed in the same way, as in Figure 4B.The difference between the values for the control metaphase cells and the CLASP knockdown metaphase-like cells is statistically significant (P < 0.001).

mitotic localization (Fig. 1 and Maiato et al. 2003). Our previous study in inter phase cells pointed to a strong functional similar ity and redundancy of these two proteins (Mimor i-Kiyosue et al. 2005). All our data indicate that in HeLa cells the two CLASPs display very similar

localizations and functions both in inter phase and in mitosis. Our observations suggest that in HeLa cells the two CLASPs act as a common pool, the depletion of which below a certain level has severe consequences for the cell division. In agreement with this idea, a genetic
Genes to Cells (2006) 11, 845­ 857

© 2006 The Authors Jour nal compilation © 2006 by the Molecular Biology Society of Japan/Blackwell Publishing Ltd.

853


Y Mimori-Kiyosue et al.

knockout of CLASP2 in mouse fibroblasts does not ar rest cell proliferation, indicating that CLASP1 alone can support cell division (H. Maiato, personal communication). However, CLASP2 knockout fibroblasts display multiple mitotic abnor malities (H. Maiato, personal communication), suggesting that mammalian mitosis is exquisitely sensitive to CLASP dosage. Similar to their invertebrate homologs, CLASPs are needed for proper mitotic progression. CLASPs, as well as their Drosophila homolog MAST/Orbit, localize to the spindle poles, the mitotic spindle and the kinetochores and are necessary for maintenance of spindle bipolar ity and the metaphase chromosome alignment (Maiato et al. 2002). In agreement with the results of CLASP depletion in insect cells and in Xenopus extracts (Maiato et al. 2002, 2005; Goshima et al. 2005; Hannak & Heald 2006), bipolar spindles observed after simultaneous CLASP1/CLASP2 knock down in HeLa cells were short and displayed unstable chromosome attachment. In inter phase cells, mammalian CLASPs act as rescue factors, which promote microtubule stability (Mimor i-Kiyosue et al. 2005). Therefore, after CLASP knock down microtubule density is decreased, while the pool of soluble tubulin and microtubule growth rate are increased. We also observed an increase in microtubule growth rate in CLASP-depleted mitotic cells, suggesting that also in mitosis CLASPs stabilize microtubules. This conclusion is supported by the short pole-to-pole distance in bipolar spindles of CLASP-depleted cells, since microtubule-stabilizing factors have a positive effect on the spindle length (Goshima et al. 2005). In contrast to the observations in insect cells (Maiato et al. 2002), reduction of CLASP levels in HeLa cells more often lead to the for mation of multipolar spindles with chromosomes attached to the microtubules, rather than to monopolar spindles with unattached chromosomes (although it is possible that in insect cells chromosomes were actually attached to very short microtubules). In Drosophila cells, MAST/Orbit is required for maintaining fluxing kinetochore fibers by stimulating subunit incor poration at their plus ends (Maiato et al. 2005). The absence of CLASPs therefore causes gradual shortening and collapse of the bipolar spindle. The situation may be somewhat different in mammalian cells, where flux may play a less important role for metaphase chromosome dynamics (Mitchison & Salmon 1992), explaining the difference in the results obtained in the two systems. Alter natively, CLASP depletion achieved in HeLa cells is less profound than that obtained by RNAi in Drosophila cells, which might explain the divergent phenotypes. Multiple spindle poles were also observed in hypomor phic MAST/Orbit
854
Genes to Cells (2006) 11, 845­ 857

mutants, where they were associated with polyploidy (Inoue et al. 2000). It is likely that some of the multipolar spindles, which we observed, are also due to a cytokinesis failure in the previous round of cell division. However, in addition we frequently monitored an initial for mation of a bipolar spindle, which later collapsed into a multipolar configuration (Fig. 5C), suggesting that CLASPs may be involved in maintaining spindle pole integr ity. Partial knockdown of the two CLASPs, achieved by us in HeLa cells, may be phenotypically compared to a hypomor phic mutant. In accord with this idea, some cells do proceed beyond metaphase and complete cell division. In agreement with the observations on effects of a hypomor phic MAST/Orbit allele in fly testis (Inoue et al. 2004), reduced CLASP levels also lead to cytokinesis defects, as cells often failed to undergo cleavage or fused back after un incomplete division (Fig. 3). Reduced levels of CLASPs did not inhibit spindle dynamics or kinetochore movements, in contrast to the previous observations in monopolar spindles, obtained after the microinjection of CLASP1 antibodies in HeLa cells (Maiato et al. 2003). On the contrary, the rate of microtubule growth was slightly increased, and centromere oscillations were enhanced. This difference might have several explanations. First, it is possible that the CLASP1 antibodies, used in that study, bound to other proteins in addition to CLASP1, since the authors show that the antibody preferentially reacted with an endogenous HeLa cell protein of 212 kDa, which significantly exceeded in size the largest known CLASP1 isofor m (160 kDa). In our hands, the CLASP1 antibodies raised by Maiato et al. (2003) did recognize endogenous CLASP1 of 160 kDa (A. Akhmanova, unpublished observations).This raises the possibility that antibody injection caused a dominant-negative effect on the spindle dynamics. Our data on increased kinetochore movements are in line with those obtained in C. elegans embryos, showing dramatic chromosome oscillations after CLASPcls-2 depletion (Cheeseman et al. 2005) and with the chromosome behavior in Drosophila cells after RNAi-mediated CLASP knockdown (Maiato et al. 2002). In agreement with previous data (Maiato et al. 2003) mammalian CLASPs are not cr itical for the kinetochoremicrotubule end-on attachment. Partial depletion of CLASPs causes metaphase-like ar rangement of chromosomes, which is maintained for prolonged time. Bipolar spindles ar rested in metaphase were also observed after treatment with colcemid or nocodazole at low concentration (Kleinfeld & Sisken 1966; Alieva & Vorobjev 1991). However, in contrast to these phar macological

© 2006 The Authors Jour nal compilation © 2006 by the Molecular Biology Society of Japan/Blackwell Publishing Ltd.


CLASPs in mitosis

treatments, CLASP depletion enhances, rather than inhibits microtubule dynamics. Increased oscillations of centromeres and reduced kinetochore tension, in the presence of an intact spindle checkpoint, indicate that CLASPs probably help maintain proper interactions between the tips of dynamic microtubules and the kinetochores. It is likely that CLASPs locally regulate microtubule dynamics in the vicinity of kinetochores, promoting their growth. Microtubule growth-promoting activity probably depends on microtubule tip recognition through the middle part of the protein (Mimor i-Kiyosue et al. 2005) and on the microtubule-independent association of CLASPs with kinetochores through their Cter minal domains. Therefore, the action of CLASP at the kinetochore is probably similar to that at the cell cortex, to which CLASPs also bind in a microtubuleindependent fashion (Mimor i-Kiyosue et al. 2005). It appears that CLASPs can be regarded as local regulators of microtubule plus end dynamics at certain cellular structures, which can for m a boundary for microtubule growth.

UCSD, La Jolla, CA USA (Campbell et al. 2002)) to generate red fluorescent fusions.

Antibodies and immunofluorescent staining
We used mouse monoclonal antibodies (mAbs) against EB1, p150Glued and BubRI (BD Biosciences), CENP-A (MBL, Japan), - and -tubulin and acetylated tubulin (Sigma); rat mAb against -tubulin (YL1/2, Abcam); rabbit polyclonal antibodies (pAbs) against CLASP1 (Mimor i-Kiyosue et al. 2005), CLASP2 (Akhmanova et al. 2001), CLIP-170 (Coquelle et al. 2002) and human CREST serum (kindly provided by DrY. Muro, Nagoya University); chicken pAb against GFP (Chemicon). For secondary antibodies, Cy2-conjugated anti-mouse IgG and anti-rabbit IgG pAbs, FITC-conjugated anti-human IgG pAb, TexasRedconjugated anti-mouse IgG, anti-rat IgG and anti-rabbit IgG pAb, Cy5-conjugated anti-mouse IgG, anti-rat IgG and anti-rabbit IgG pAbs were purchased from Jackson. Cell fixation and staining were perfor med as previously descr ibed (Mimor i-Kiyosue et al. 2005).

Fluorescence microscopy and image analysis
The cells were observed with a DeltaVision optical sectioning system (Ver. 2.5, Applied Precision Inc.) equipped with an Olympus IX70 inverted microscope and a cooled CCD camera (Ser ies300 CH350, Photometr ics), an Aquacosmos system (Hamamatsu photonics) an Olympus IX70 inverted microscope and a cooled CCD camera (ORCA ER, Hamamatsu Photonics), or a LSM510 confocal laser scanning microscope (Ver. 2.3, Carl Zeiss). Quantitative analysis of fluorescent confocal images was perfor med using MetaMor ph software (Universal imaging). To evaluate mitotic abnormalities (Fig. 4I,J), > 200 images of mitotic cells fixed and stained for DNA and tubulin were automatically collected by scanning the specimens using the panel collection function of DeltaVision (UPlanApo 20â objective), and the mitotic phase and spindle mor phology were identified on the basis of their chromosome and spindle configurations, respectively. For live imaging, cells were cultured on glass-bottomed dishes with No.1S coverslips (Iwaki, Japan). Images of cells were collected with a DeltaVision optical sectioning system using PlanApo 100x/1.40 NA oil, PlanApo 60x/1.40 NA oil ph3 or UPlanApo 20x/0.70 NA dry objectives (Olympus). Images were acquired with a cooled CCD camera (Ser ies300 CH350, Photometr ics) with an appropr iate ND filter, binning of pixels, exposure time, and time intervals. Fluorescence signals were visualized using a quad filter set (86000, Chroma) for multiple-color imaging, or Endow GFP bandpass emission filter set (41017, Chroma) for GFP imaging. The out-of-focus signals in the set of optical sections were then removed using the deconvolution technique with the DeltaVision system. The quantification of the microtubule dynamics was perfor med, using MetaMor ph software (Universal Imaging Cor p.). Statistical analysis was perfor med using SigmaPlot (SPSS Inc.) and Statistica for Windows (StatSoft Inc.). Unless stated differently, the statistical significance of the observed differences was evaluated using Kolmogorov-Smir nov two-sample test.

Experimental procedures
Cell lines and transfection of plasmids and siRNAs
HeLa cells were grown at 37 °C, in 5% CO2 atmosphere, in DMEM (Invitrogen) supplemented with 10% fetal bovine serum and antibiotics. The Xenopus kidney epithelial cell line A6 was grown at 23 °C in 50% Leibovitz's L-15 medium (Invitrogen) supplemented with 10% fetal bovine serum and antibiotics. Effectene or Superfect transfection reagents (Qiagen) were used for the transfection of plasmids into HeLa or A6 cells, according to the manufacturer's protocols. Stable clones were selected in the presence of 0.6 ­ 0.8 mg/mL or 0.3 ­ 0.4 mg/mL G418 sulfate for A6 or HeLa cells, respectively (Calbiochem), by detecting GFP fluorescence. Synthetic siRNAs (Proligo) were previously descr ibed (Mimor i-Kiyosue et al. 2005). Ten percent confluent HeLa cells were transfected, using Oligofectamine (Invitrogen) with siRNAs at the minimal effective concentration (10 nm for CLASP1#A and CLASP1#B, 200 nm for CLASP2#A and CLASP2#B), while the control siRNA was used at 200 nm concentration.

Expression constructs
We used the previously descr ibed expression vectors for GFPCLASP2 (Akhmanova et al. 2001), GFP-CLASP2, GFPCLASP2 C, GFP-CLASP2-C (Mimor i-Kiyosue et al. 2005), EB3-GFP (Stepanova et al. 2003), GFP--tubulin (Clontech). GFP-dynamitin/p50 was a kind gift of Dr T. Schroer. AuroraB cDNA was amplified from HeLa cell cDNA library (HeLa largeInsert, Clontech) and inserted into pEGFP-C2 vector. GFP was substituted for DsRed2 (Clontech) or mRFP (gift of Dr R.Tsien,

© 2006 The Authors Jour nal compilation © 2006 by the Molecular Biology Society of Japan/Blackwell Publishing Ltd.

Genes to Cells (2006) 11, 845­ 857

855


Y Mimori-Kiyosue et al.

Electron microscopy
After 72 h from siRNA transfection, mitotic HeLa cells were selectively removed from culture dishes by pipetting and collected by centrifugation. Cells were doubly fixed with 1.2% glutaraldehyde in 0.1 m phosphate buffer and 1% osmium tetroxide in 0.1 m phosphate buffer, and dehydrated with a graded ser ies of ethanol. Cells were then embedded in epoxy resin. One hundred nanometer thin sections were stained with uranyl acetate and lead citrate and observed with a transmission electron microscope (Hitachi, H-7500).

Acknowledgements
We thank Dr H. Maiato for comments on this manuscr ipt and for communicating results pr ior to publication. We are g rateful to Dr H. Yamauchi (KAN Research Institute) for continuous encouragement and to Dr Y. Muro and Dr R.Tsien for providing mater ials. This study was suppor ted by Russian Foundation for Basic Research, g rant no. 05-04-49847 to IAV and by the Netherlands Organization for Scientific Research grants to AA.

References
Akhmanova, A. & Hoogenraad, C.C. (2005) Microtubule plus-end-tracking proteins: mechanisms and functions. Curr. Opin. Cell Biol. 17, 47­ 54. Akhmanova, A., Hoogenraad, C.C., Drabek, K., et al. (2001) Clasps are CLIP-115 and -170 associating proteins involved in the regional regulation of microtubule dynamics in motile fibroblasts. Cell 104, 923 ­ 935. Alieva, I.B. & Vorobjev, I.A. (1991) Induction of multipolar mitoses in cultured cells: decay and restructur ing of the mitotic apparatus and distr ibution of centr ioles. Chromosoma 100, 532 ­ 542. Campbell, R.E., Tour, O., Palmer, A.E., et al. (2002) A monomer ic red fluorescent protein. Proc. Natl. Acad. Sci. USA 99, 7877­7882. Cheeseman, I.M., MacLeod, I., Yates, J.R., 3rd, Oegema, K. & Desai, A. (2005) The CENP-F-like proteins HCP-1 and HCP2 target CLASP to kinetochores to mediate chromosome segregation. Curr. Biol. 15, 771­777. Cleveland, D.W., Mao, Y. & Sullivan, K.F. (2003) Centromeres and kinetochores: from epigenetics to mitotic checkpoint signaling. Cell 112, 407 ­ 421. Coquelle, F.M., Caspi, M., Cordelieres, F.P., et al. (2002) LIS1, CLIP-170's key to the dynein/dynactin pathway. Mol. Cell Biol. 22, 3089 ­3102. Dujardin, D., Wacker, U.I., Moreau, A., Schroer, T.A., Rickard, J.E. & De Mey, J.R. (1998) Evidence for a role of CLIP-170 in the establishment of metaphase chromosome alignment. J. Cell Biol. 141, 849 ­ 862. Gonczy, P., Echever r i, C., Oegema, K., et al. (2000) Functional genomic analysis of cell division in C. elegans using RNAi of genes on chromosome III. Nature 408, 331­336. Goshima, G., Wollman, R., Stuur man, N., Scholey, J.M. & Vale, R.D. (2005) Length control of the metaphase spindle. Curr. Biol. 15, 1979­1988.

Hannak, E. & Heald, R. (2006) Xorbit/CLASP links dynamic microtubules to chromosomes in the Xenopus meiotic spindle. J. Cell Biol. 172, 19 ­25. Inoue, Y.H., do Car mo Avides, M., Shiraki, M., et al. (2000) Orbit, a novel microtubule-associated protein essential for mitosis in Drosophila melanogaster. J. Cell Biol. 149, 153 ­166. Inoue, S. & Salmon, E.D. (1995) Force generation by microtubule assembly/disassembly in mitosis and related movements. Mol. Biol. Cell 6, 1619 ­1640. Inoue, Y.H., Savoian, M.S., Suzuki, T., Mathe, E., Yamamoto, M.T. & Glover, D.M. (2004) Mutations in orbit/mast reveal that the central spindle is compr ised of two microtubule populations, those that initiate cleavage and those that propagate fur row ingression. J. Cell Biol. 166, 49­ 60. Kleinfeld, R.G. & Sisken, J.E. (1966) Mor phological and kinetic aspects of mitotic ar rest by and recovery from colcemid. J. Cell Biol. 31, 369 ­ 379. Lemos, C.L., Sampaio, P., Maiato, H., et al. (2000) Mast, a conserved microtubule-associated protein required for bipolar mitotic spindle organization. EMBO J. 19, 3668 ­ 3682. Lew, D.J. & Burke, D.J. (2003) The spindle assembly and spindle position checkpoints. Annu. Rev. Genet. 37, 251­282. Maiato, H., DeLuca, J., Salmon, E.D. & Ear nshaw, W.C. (2004) The dynamic kinetochore-microtubule interface. J. Cell Sci. 117, 5461­5477. Maiato, H., Fairley, E.A., Rieder, C.L., Swedlow, J.R., Sunkel, C.E. & Ear nshaw, W.C. (2003) Human CLASP1 is an outer kinetochore component that regulates spindle microtubule dynamics. Cell 113, 891­904. Maiato, H., Khodjakov, A. & Rieder, C.L. (2005) Drosophila CLASP is required for the incor poration of microtubule subunits into fluxing kinetochore fibres. Nat. Cell Biol. 7, 42 ­ 47. Maiato, H., Sampaio, P., Lemos, C.L., et al. (2002) MAST/Orbit has a role in microtubule-kinetochore attachment and is essential for chromosome alignment and maintenance of spindle bipolar ity. J. Cell Biol. 157, 749 ­760. Mimor i-Kiyosue, Y., Gr igor iev, I., Lansbergen, G., et al. (2005) CLASP1 and CLASP2 bind to EB1 and regulate microtubule plus-end dynamics at the cell cortex. J. Cell Biol. 168, 141­153. Mitchison, T.J. & Salmon, E.D. (1992) Poleward kinetochore fiber movement occurs dur ing both metaphase and anaphase-A in newt lung cell mitosis. J. Cell Biol. 119, 569 ­ 582. Schuyler, S.C. & Pellman, D. (2001) Microtubule "plus-endtracking proteins": The end is just the beginning. Cell 105, 421­ 424. Stepanova, T., Slemmer, J., Hoogenraad, C.C., et al. (2003) Visualization of microtubule growth in cultured neurons via the use of EB3-GFP (end-binding protein 3-green fluorescent protein). J. Neurosci. 23, 2655 ­ 2664. Tir nauer, J.S., Canman, J.C., Salmon, E.D. & Mitchison, T.J. (2002) EB1 targets to kinetochores with attached, polymer izing microtubules. Mol. Biol. Cell 13, 4308 ­ 4316. Waters, J.C., Skibbens, R.V. & Salmon, E.D. (1996) Oscillating mitotic newt lung cell kinetochores are, on average, under tension and rarely push. J. Cell Sci. 109, 2823 ­ 2831.

856

Genes to Cells (2006) 11, 845­ 857

© 2006 The Authors Jour nal compilation © 2006 by the Molecular Biology Society of Japan/Blackwell Publishing Ltd.


CLASPs in mitosis Yin, H., You, L., Pasqualone, D., Kopski, K.M. & Huffaker, T.C. (2002) Stu1p is physically associated with beta-tubulin and is required for structural integr ity of the mitotic spindle. Mol. Biol. Cell 13, 1881­1892. Received: 1 March 2006 Accepted: 16 Apr il 2006 Movie S1 Cell cycle progression of control and CLASPdepleted cells. Movie S2 Microtubule dynamics in control and CLASPdepleted cells dur ing mitosis. Movie S3 GFP-AuroraB dynamics in control and CLASPdepleted cells dur ing mitosis. Movie S4 Simultaneous imaging of kinetochores and growing microtubule ends in metaphase-like control and CLASP-depleted cells.

Supplementary material
The following supplementary mater ial is available for this article online:

© 2006 The Authors Jour nal compilation © 2006 by the Molecular Biology Society of Japan/Blackwell Publishing Ltd.

Genes to Cells (2006) 11, 845­ 857

857