QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

Vol. 17, Issue 3, 1421-1435, March 2006

This Article Abstract Full Text (PDF) Supplemental Material All Versions of this Article: E05-08-0802v1 17/3/1421    most recent Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Asakawa, K. Articles by Toda, T. Search for Related Content PubMed PubMed Citation Articles by Asakawa, K. Articles by Toda, T.

The V260I Mutation in Fission Yeast -Tubulin Atb2 Affects Microtubule Dynamics and EB1-Mal3 Localization and Activates the Bub1 Branch of the Spindle Checkpoint

Kazuhide Asakawa ^*  , Kazunori Kume  , Muneyoshi Kanai , Tetsuya Goshima , Kohji Miyahara , Susheela Dhut ^*, Wee Wei Tee ^*, Dai Hirata , and Takashi Toda ^*

^* Laboratory of Cell Regulation, Cancer Research UK, London Research Institute, Lincoln's Inn Fields Laboratories, London WC2A 3PX, United Kingdom; Department of Molecular Biotechnology, Graduate School of Advanced Sciences of Matter, Hiroshima University, Higashi-Hiroshima 739-8530, Japan

Submitted August 26, 2005; Revised December 13, 2005; Accepted December 27, 2005
Monitoring Editor: J. Richard McIntosh

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
We have identified a novel temperature-sensitive mutant of fission yeast -tubulin Atb2 (atb2-983) that contains a single amino acid substitution (V260I). Atb2-983 is incorporated into the microtubules, and their overall structures are not altered noticeably, but microtubule dynamics is compromised during interphase. atb2-983 displays a high rate of chromosome missegregation and is synthetically lethal with deletions in a subset of spindle checkpoint genes including bub1, bub3, and mph1, but not with mad1, mad2, and mad3. During early mitosis in this mutant, Bub1, but not Mad2, remains for a prolonged period in the kinetochores that are situated in proximity to one of the two SPBs (spindle pole bodies). High dosage mal3⁺, encoding EB1 homologue, rescues atb2-983, suggesting that Mal3 function is compromised. Consistently, Mal3 localization and binding between Mal3 and Atb2-983 are impaired significantly, and a mal3 single mutant, such as atb2-983, displays prolonged Bub1 kinetochore localization. Furthermore in atb2-983 back-and-forth centromere oscillation during prometaphase is abolished. Intriguingly, this oscillation still occurs in the mal3 mutant, indicating that there is another defect independent of Mal3. These results show that microtubule dynamics is important for coordinated execution of mitotic events, in which Mal3 plays a vital role.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
The microtubule (MT) possesses intrinsic polarity, the dynamic plus end, and the less dynamic minus end. In many, if not all, cell types, the minus end is embedded in a specialized structure, called the microtubule organizing center (Pickett-Heaps, 1969), whereas the plus end interacts with a distinct class of MT accessory factors, collectively called plus-end-tracking proteins (Schuyler and Pellman, 2001). In both interphase and mitotic spindle MTs, it is the plus end that explores spatial cues through alternating cycles of growth and shrinkage, MT mechanics termed dynamic instability (Mitchison and Kirschner, 1986). This search-and-capture process facilitates efficient attachment of MT plus end to appropriate subcellular sites such as the cell cortex and the kinetochore, which in turn results in altering dynamics of the connected MTs (Desai and Mitchison, 1997).

The mitotic bipolar spindle is a highly organized apparatus that is required for accurate chromosome segregation. The plus end of MTs that emanate from opposite poles is thought to capture sister kinetochores in a stochastic manner during prometaphase. This mechanism ensures spindle bipolarity and chromosome biorientation, which are vital for bidirectional sister chromatid segregation at anaphase. Errors in these attachment and segregation processes result in producing aneuploid progenies, which can lead to lethality, various genetic disease, and tumorigenesis. To maintain accuracy of chromosome segregation, cells have developed a mitotic surveillance mechanism, termed the spindle assembly checkpoint. This checkpoint monitors the integrity of spindle–kinetochore interaction and delays anaphase onset until bipolar attachment of all the kinetochores to spindles is secured (Millband et al., 2002; Cleveland et al., 2003; Bharadwaj and Yu, 2004). Spindle checkpoint proteins include Mad1, Mad2, Mad3 (in yeast and BubR1 in vertebrates), Bub1, Bub3, and Mps1. These proteins localize to mitotic kinetochores when bipolar attachment is not achieved and monitor the physical status of these kinetochores. It seems that the checkpoint pathway bifurcates, Mad1/Mad2 (/Mad3) and Bub1/Bub3/Mps1(/Mad3), although the two branches may not be functionally separable (Hardwick et al., 1996; Lew and Burke, 2003). Although Mad2 monitors and localizes to unattached kinetochores in most, if not all, organisms, Bub1 localizes to not only unattached kinetochores but also to those that show weaker attachment or lack tension (Waters et al., 1998; Skoufias et al., 2001; Zhou et al., 2002). In addition, Bub1 plays a role in chromosome segregation independent of its role in the spindle checkpoint (Tang et al., 2004; Vanoosthuyse et al., 2004; Kitajima et al., 2005; Meraldi and Sorger, 2005).

Inside the cell, a number of regulatory factors are involved in MT morphogenesis and dynamism. Among these, plus-end-tracking proteins interact with and localize to the plus end of the MT and regulate MT dynamics, thereby participating in a number of cellular processes (Schuyler and Pellman, 2001). The conserved EB1 protein belongs to this MT-binding family. It was originally identified as an interacting partner for vertebrate tumor suppressor protein Adenomatous Polyposis Coli (APC) (Su et al., 1995) and is thought to be crucial, together with APC, for cell polarity and chromosome stability control (Tirnauer and Bierer, 2000; Fodde et al., 2001; Kaplan et al., 2001). It is shown that chromosome instability phenotypes (CIN) observed in APC^– tumor cells are associated with defects in MT plus end attachments (Green and Kaplan, 2003), although causal involvement of EB1 in these tumor cells remains to be determined (Smits et al., 1999).

In both fission and budding yeast, EB1 homologues (Mal3 and Bim1, respectively) play a vital role in cell polarity and chromosome stability (Beinhauer et al., 1997; Schwartz et al., 1997; Muhua et al., 1998; Tirnauer et al., 1999; Chen et al., 2000; Browning et al., 2003; Hwang et al., 2003; Busch and Brunner, 2004; Busch et al., 2004; Mayer et al., 2004). Recent work in fission yeast indicates that Mal3 interacts with the kinetochore component Spc7 (homologue of budding yeast Spc105 and human KIAA1570) during mitosis and is important for accurate chromosome segregation (Kerres et al., 2004). We have recently shown that in the absence of Mal3, mitotic delay is induced, in which the Bub1 branch of the spindle checkpoint is activated, and Bub1-localizing kinetochores are associated for a prolonged period of time with only one of two separating spindle pole bodies (SPBs) (Asakawa et al., 2005; Asakawa and Toda, 2006).

In this study, we screened for temperature-sensitive mutants that are defective in chromosome segregation, in which the Bub1-dependent spindle checkpoint is activated. One of mutants identified was a novel allele of -tubulin gene (atb2-983) that contained a missense mutation in the highly conserved valine residue (V261I). We isolated the mal3⁺ gene as a multicopy suppressor of atb2-983. Consistent with this, Mal3 loading to the MT is compromised in this mutant. Characterization of atb2-983 and mal3 mutants provides a novel insight into roles of microtubule dynamics in chromosome stability and spindle checkpoint control.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
Strains, Media, and Genetic Methods
Strains used in this study are listed in Table 1. The standard methods were followed as described previously (Moreno et al., 1991; Bähler et al., 1998). The kan^r-nmtP3-GFP cassette (Bähler et al., 1998) was integrated in front of the initiator methionine to create a strain containing nmtP3-GFP-atb2-983 (GFP-Atb2-983, KZ176). The C-terminal tagging of Mal3 with Myc epitope was performed using the nourseothricin-resistance gene as a selectable marker (mal3⁺-myc-nat^r) as described previously (Sato et al., 2005).

Isolation of cin Mutants
A parental strain (KZ2; Table 1), which contains minichromosome Ch16 (Niwa et al., 1989) and Bub1-GFP was mutagenized with nitrosoguanidine as described previously (Radcliffe et al., 1998), spread on rich YE5S plates and incubated at 27°C. Colonies were replica-plated onto YE5S at 36°C and YE plates (rich medium lacking exogenously added supplements) at 32°C, on which cells that had lost Ch16 turned red because of the chromosomal ade6-210 mutation (Niwa et al., 1989). On this procedure, colonies that were temperature sensitive (ts) and also displayed red or sectored appearance on YE plates were selected; 221 isolates out of 500,000 mutagenized colonies showed both minichromosome-loss and ts phenotypes. On backcrossing with a marker strain (containing Ch16), 39 mutants showed cosegregation of minichromosome loss and ts phenotypes. These 39 mutants were subsequently classified into 31 complementation groups. Among these isolates, mutations in 10 complementation groups (cin1–cin10) showed hypersensitivity to thiabendazole (TBZ) compared with a wild-type strain (Wang et al., 2005; Supplemental Table 1).

Nucleic Acids Preparation and Manipulation
Enzymes were used as recommended by the suppliers (New England Biolabs, Beverly, MA; Takara Shuzo, Kyoto, Japan; and Stratagene, La Jolla, CA).

Identification of cin3 as -Tubulin-encoding atb2
The genomic library was introduced into ts cin3–983 cells, and Ts⁺ transformants were isolated by replica plating. Whole DNAs were prepared from these transformants, and plasmid DNAs were recovered via transformation into Escherichia coli, which yielded in total 97 plasmids. DNA sequencing and restriction enzyme mapping of these plasmids revealed that seven different genes were able to suppress a ts phenotype of cin3-983. Two of them were -tubulin-encoding genes, nda2⁺ (28 plasmids) and atb2⁺ (36 plasmids), respectively (Toda et al., 1984; Adachi et al., 1986). The third gene (25 plasmids) was mal3⁺. Cloning and characterization of the other four genes, and their products will be described elsewhere.

Genetic linkage analysis showed that the cin3-983 mutation is tightly linked to the atb2 locus. Identity between cin3 and atb2 was confirmed, because nucleotide sequencing of the atb2 gene prepared from a cin3-983 strain shows that it contains a point mutation, which results in amino acid replacement at position 260 from valine to isoleucine (see Figure 2A).

View larger version (61K): [in this window] [in a new window]   Figure 2. The cin3-983 mutant is allelic to atb2. (A) The mutation site of Atb2-983. The valine 260 residue (white letters in black boxes) was substituted with isoleucine (red) in the atb2-983 mutant. The regions around the valine 260 in -tubulins from various species were aligned. The corresponding regions in fission yeast -tubulin (Nda3) and -tubulin (Tug1/Gtb1) are also shown. Sp, Schizosaccharomyces pombe; Hs, Homo sapiens; Mm, Mus muscalas; Dm, Drosophila melanogaster; At, Arabidopsis thaliana; Ce, Caenorhabditis elegans; and Sc, Saccharomyces cerevisiae. (B) Predicted microtubule structure. The structure of fission yeast /-tubulin dimers was predicted by RasMol software and aligned longitudinally based on bovine /-tubulin protofilament (Nogales et al., 1998). Structure is oriented with the plus end upward (+), the minus end downward (–), the inside leftward (i), and the outside rightward (lat). The location of the atb2-983 mutation was indicated by arrow with red circle in a magnified view of the interdimer interface (right). (C) The atb2-V260I mutation confers temperature sensitivity. p(atb2+), p(atb2-V260I) or vector plasmids were introduced into an atb2-deletion strain (KZ115, top). A strain containing the integrated atb2-V260I allele at its genomic locus (KZ217, middle), wild-type diploids (atb2+/atb2+, bottom) and heterozygous diploids (atb2-983/atb2+) were also tested for temperature sensitivity. Each strain was spotted on rich medium and incubated at 27 or 36°C for 3 d. (D) Incorporation of GFP-Atb2-983 into the microtubules. A strain containing wild-type GFP-atb2+ (P3-GFP-atb2) or mutant GFP-atb2-983 (KZ176) was grown at 27°C in the liquid YE5S medium, and GFP signals were recorded. Images of interphase (top) and mitosis (bottom) are shown. Bar, 10 µm.

Replacement of the atb2⁺ Gene with the atb2-V260I Allele at Its Genomic Locus
The atb2-V260I allele was integrated into the genomic atb2 locus in the following manner. First, 90% of the atb2⁺ open reading frame (ORF) (corresponding to 27–422 amino acid residues) was deleted and replaced with the ura4⁺ gene (atb2::ura4⁺, KZ115). Then, the 2-kb fragments containing the entire atb2-V260I ORF and each 300 base pairs of 5' and 3' flanking sequences were PCR-amplified from the p(atb2-V260I) plasmid DNA and used to transform KZ115. Uracil auxotroph colonies were selected on 5-fluoroorotic acid-containing plates. Correct replacement was verified with colony PCR (KZ217).

Immunochemistry
The following antibodies were used as primary antibodies for immunofluorescence microscopy: mouse monoclonal anti--tubulin antibody (TAT-1; provided by Dr. Keith Gull, The University of Oxford, Oxford, United Kingdom) and affinity-purified rabbit polyclonal anti-Sad1 antibody (gift from Dr. Mizuki Shimanuki, Kazusa DNA Research Institute, Chiba, Japan). For immunoblotting, mouse monoclonal anti-Myc antibody (9E10; BAbCO, Richmond, CA), anti-green fluorescent protein (GFP) antibody (1814 460; Roche Diagnostics, Indianapolis, IN), anti--tubulin antibody (TAT-1), and anti-Cdc2 antibody (PSTAIRE, sc-53; Santa Cruz Biotechnology, Santa Cruz, CA) were used. For immunoprecipitation of GFP-Atb2 (–983) and Mal3-Myc or Mal3-GFP and Atb2 (–983), mouse monoclonal anti-Myc antibody or rabbit polyclonal anti-GFP antibody (A-11122; Molecular Probes, Eugene, OR) was used for a primary antibody, respectively, followed by addition of protein G-coupled Dynabeads (100.04; Dynal Biotech, Lake Success, NY). Immunoprecipitates were run on SDS-PAGE, transferred to Immobilon-P filter paper (IPVH00010; Millipore, Billerica, MA), and immunoblotted with rabbit polyclonal anti-Myc antibody (PRB-150C; BAbCO), mouse monoclonal anti-GFP antibody (8372-2; Clonetech, Mountain View, CA), or TAT-1 antibody. Two micrograms of total protein extracts was used.

Fluorescence Microscopy, Time-Lapse Live Analysis, and Image Processing
For GFP-fluorescence observation, cells were fixed with 3.7% formaldehyde. For indirect immunofluorescence with TAT-1, cells were fixed with methanol at –80°C, and then TAT-1 (1/50 dilution) was applied, followed by Cy3-conjugated goat anti-mouse IgG (C2128; Sigma-Aldrich, St. Louis, MO). Still images were viewed with a Zeiss Axioplan equipped with a chilled video charge-coupled device camera (C4742-95; Hamamatsu Photonics, Shizuoka, Japan) and a PC computer containing kinetic image AQM software (Kinetic Imaging, Durham, NC) and processed by use of Adobe Photoshop version 7.0 (Adobe Systems, Mountain View, CA).

For time-lapse live imaging, a 35-mm glass-bottomed culture dish (P35G-1.5-10-C; MatTek, Ashland, MA) was coated with 200 µg/ml soybean lectin (L1395; Sigma-Aldrich). The culture of logarithmically growing cells (100 µl) was deposited in the well for a couple of minutes and then removed. The dish was filled with 3 ml of YE5S medium, and the cells that were attached to the bottom of the well were subjected to microscopic analysis at the room temperature. Images were taken using an Olympus IX70 wide-field inverted epifluorescence microscope with an Olympus PlanApo 100x, numerical aperture 1.4, oil immersion objective. DeltaVision image acquisition software (softWoRx 3.3.0; Applied Precision, Issaquah, WA) equipped with CoolSNAP HQ (Roper Scientific, Tucson, AZ) was used for capture of live images, which were processed with Adobe Photoshop version 7.0. The sections of images at each time point were compressed into a two-dimensional (2D) projection using the DeltaVision maximum intensity algorithm. Kymograph pictures derived from 2D-projected time-lapse images were constructed using softWoRx 3.3.0. All time-lapse analyses were performed at the room temperature (22 or 25°C).

Imaging Conditions
Still images were reconstructed from 10 to 14 sections of Z-stacks (z = 0.3 µms), which included the whole cell volume, by Lucida software. For the simultaneous observation of Sad1-red fluorescent protein (RFP) and cen2-GFP (Figures 4, B and C, and 9), eight to 10 sections of images (z = 0.4 µm) were taken at every 30 or 60 s. For the simultaneous observation of Mal3-GFP and Sad1-RFP (Figure 8B), four sections of images (z = 0.4 µm) were taken at every 10 s. The SPBs were included in these four sections at every time point. For the simultaneous observation of Sad1-RFP and Bub1-GFP (Figures 6A and 8F), 8–10 s of images (z = 0.4 µm) were taken every 30- or 60-s interval.

View larger version (53K): [in this window] [in a new window]   Figure 4. Mitosis is slowed down in atb2-983 cells. (A) Mitotic delay in atb2-983 cells. Wild-type (513) and atb2-983 cells (KZ94) were incubated at 36°C for 4 h, and MTs were visualized with immunofluorescence as in Figure 3. The population of cells with short (<2 µm) or long spindle (>2 µm) was counted from >300 cells. (B) Passage of mitotic phases in live atb2-983 cells. Wild-type (KZ85) and atb2-983 cells (KZ139) containing cen2-GFP and Sad1-RFP were grown, and time-lapse live analysis was performed at the room temperature (n = 10). Typical examples for the two strains were shown. The timing of anaphase A onset and completion is marked with yellow and white circles, respectively. (C) Kymograph images of the cells shown in B. Kymograph images were produced every 1-min interval. Yellow and white arrows indicate the duration of prophase and prometaphase/metaphase, respectively. Vertical bar, 5 µm.

View larger version (63K): [in this window] [in a new window]   Figure 9. Lack of sister centromere oscillation in the atb2-983 mutation. (A) Visualization of sister centromere movement in live cells. Kymograph images at high resolution (10-s intervals) are shown (cen2-GFP in green and Sad1-RFP in red) in wild-type (top; n = 32) and atb2-983 cells (bottom; n = 30). cen2-GFP signals that were persistently associated with one SPB like Bub1-GFP (see Figure 6A), which probably represented syntelic attachment, were not analyzed in this assay. Yellow bidirectional arrows show the duration of anaphase A. Solid and dotted arrows indicate the duration of prometaphase and anaphase, respectively. Arrowhead indicates the timing of metaphase. (B) Prometaphase movement of cen2-GFP. The positions of cen2-GFP (green circles) are plotted relative to the two SPBs (red squares) every 20-s interval. The central position of the two SPBs is shown with purple lines (diamonds). When cen2-GFP signals seemed split, the middle position is plotted. The representative images are shown in wild-type (n = 12) and atb2-983 (n = 10). (C) Visualization of sister centromere movement in mal3 cells. Kymograph image in the mal3 mutant is constructed as described in A. Vertical bars, 5 µm.

View larger version (57K): [in this window] [in a new window]   Figure 8. Mal3 loading onto the spindle and binding to Atb2–983 are reduced in the atb2-983 mutant. (A) Reduced Mal3 localization. Wild-type (MA145) and ab2-983 cells containing Mal3-GFP (KZ165) were grown at 27°C and GFP signals were taken. (B) Time-lapse analysis of mitotic Mal3-GFP. The live signals of mitotic Mal3-GFP in wild-type (KZ194; left) or atb2-983 (KZ195; right) cells were obtained and converted to kymograph pictures (Mal3-GFP in green and Sad1-RFP in red). The length of the white arrow and the black bar represents 1 min and 2 µm, respectively. (C) Protein levels of Mal3 in atb2-983. Extracts were prepared from exponentially growing wild-type (lane 2) or atb2-983 cells containing Mal3-GFP (lane 3) and immunoblotted with indicated antibodies. An untagged wild-type strain (lane 1) was used as a control. (D) Suppression of compromised spindle localization of Mal3. atb2-983 mutants carrying various multicopy plasmids (an empty vector, atb2+, nda2+, or mal3+) were grown at 27°C and shifted to 36°C for 4 h, and Mal3-GFP signals were taken. Representative Mal3-GFP signals localizing to the spindle are shown (n = 20). The exposure time of these pictures are the same (2 s). (E) Immunoprecipitation between GFP-Atb2 (-983) and Mal3-Myc. Wild-type (KZ214; lanes 3, 6, and 8) or atb2-983 mutants containing GFP-Atb2 (-983) and Mal3-GFP (KZ216; lanes 4, 7, and 9) were grown at 25°C (lanes 2, 3, 6, and 7) and shifted to 36°C (lanes 8 and 9). Total cell extracts were prepared at 0 (lanes 3 and 4, 10 mg of extracts was loaded) and 4 h (lanes 5–8), and immunoprecipitation was performed using monoclonal anti-Myc antibody as a primary antibody. Immunoprecipitates were run on SDS-PAGE, and immunoblotting was performed with polyclonal anti-Myc antibody (top) or monoclonal anti-GFP antibody (bottom). As a control, immunoprecipitation was performed from extracts prepared from a strain containing only GFP-Atb2 (lanes 2 and 5). Extracts prepared from an untagged wild-type strain was also run as another control (lane 1). Precipitated bands were marked with arrowheads, and nonspecific bands, probably corresponding to IgG heavy chains, were shown with asterisk. (F) Prolonged localization of Bub1 blob during early mitosis in mal3 cells. Time-lapse images (1-min intervals) of Bub1-GFP and Sad1-RFP localization during mitosis were recorded and converted to a kymograph. For patterns of wild-type and atb2-983 cells, see Figure 6A. Vertical bars, 2 µm.

View larger version (71K): [in this window] [in a new window]   Figure 6. The Bub1 spindle checkpoint prevents unequal chromosome segregation in atb2-983 cells. (A) Prolonged localization of Bub1 blob during mitosis. Time-lapse images of Bub1-GFP and Sad1-RFP localization during mitosis were recorded and converted to a kymograph every 1-min interval (top, wild type, KZ181; and bottom, atb2-983, KZ182). Vertical bar, 5 µm. (B) Unequal chromosome segregation phenotype of atb2-983bub1 cells. atb2-983bub1 cells containing cen2-GFP were kept viable by expressing episomal atb2+, starved for nitrogen for 12 h, and released into the rich medium (26°C) to induce plasmid loss. At 8 h after release, patterns of chromosome segregation were observed with cen2-GFP (green) and DAPI staining (red). Representative images displaying 2:0 distribution of cen2-GFP (marked with arrows) are shown. Bar, 10 µm. (C) The percentage of unequal chromosome segregation in atb2-983 and atb2-983bub1 cells.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

View larger version (102K): [in this window] [in a new window]   Figure 1. Isolation of the cin3-983 mutant. (A) Temperature and TBZ sensitivity. Wild-type (513; Table 1) or cin3 (atb2)-983 mutants (KZ94) were spotted on rich medium (5 x 103 cells in the far-left spots for each plate and then diluted 10-fold in each subsequent spot rightward) in the presence (middle plate) or absence (left and right) of 10 µg/ml TBZ and incubated at 27°C (left and middle) or at 36°C (right) for 3 d. (B) Localization of Bub1-GFP. nda3-1828 (KZ35) or cin3 (atb2)-983 cells (ts983) were incubated at 36°C for 4 h and Bub1-GFP and DNA were visualized. The top panels show Bub1-GFP signals, whereas the bottom panels show merged pictures (Bub1-GFP in green and DAPI in orange). Bar, 10 µm.

In wild-type cells, Bub1 is recruited to the kinetochore at early mitosis for a short time such that in asynchronously growing cultures, only a small population of cells (<2%) displays Bub1 localization to the kinetochore as a single nuclear dot (Bernard et al., 1998; Toyoda et al., 2002). On the other hand, in the ts -tubulin mutant (nda3-1828), which loses the MT structure at the restrictive temperature (Radcliffe et al., 1998), and is therefore expected to contain unattached kinetochores, 60% of cells showed Bub1 at the kinetochore at 36°C and even at the permissive temperature (26°C), Bub1 localized to the kinetochore in a higher proportion of cells (6.2%; Figure 1B). In the case of the rad21-K1 mutant defective in cohesion, which is also required for bipolar attachment of the kinetochore to the spindle (Tatebayashi et al., 1998; Tanaka et al., 2000), a higher value of Bub1 dots (3.0% at 26°C and 5.6% at 36°C) was obtained. Having established the experimental system in which to assess the state of the spindle–kinetochore interaction, Bub1 localization in cin1–cin10 mutants was investigated at both permissive and restrictive temperatures. It was found that all the cin mutants displayed increased Bub1 localization to the kinetochore (Figure 1B and Supplemental Table 1).

cin3 Is Allelic to atb2, a Locus Encoding a Nonessential 2-Tubulin
We cloned the cin1⁺-cin4⁺ genes and identified their corresponding gene products (Supplemental Table 1). Among these, cin3-983 was of our particular interest, because it was suppressed by plasmids containing the atb2⁺ gene and tightly linked to atb2, which encodes 2-tubulin (Toda et al., 1984). Fission yeast contains two -tubulin-encoding genes, atb2⁺ and nda2⁺, in which atb2⁺ is nonessential, whereas nda2⁺ is essential for cell viability (Adachi et al., 1986). This implied that if cin3-983 were allelic to atb2, this mutation must be dominant over chromosomal nda2⁺. To confirm identity between cin3 and atb2, nucleotide sequence of the atb2 gene was determined using genomic DNA prepared from a cin3-983 strain. This analysis demonstrated that cin3-983 indeed contains a point mutation in atb2 at position +778 from guanine to adenine (adenine for initiator ATG is denoted as +1), which results in amino acid substitution at position 260 from valine (GTA) to isoleucine (ATA). Comparison of amino acid residues among -tubulin molecules from different species indicated that V260 is invariant (Figure 2A). Close inspection of the three-dimensional structure of the -tubulin heterodimer (Nogales et al., 1998) shows that V260 is situated in a linker region between helix 8 (H8) and -sheet 7 (B7), which locates in proximity to the interacting interface between - and -tubulin (Figure 2B). This residue also faces toward the outer surface of MT protofilaments.

To verify the V260I mutation results in temperature sensitivity of atb2-983, plasmids carrying atb2-V260I that contain the V260I mutation were introduced into an atb2-deleted strain. As shown in Figure 2C (top), this transformed strain indeed became ts, and even at the low temperature (27°C), introduction of the Atb2-V260I mutation compromised cell growth. Furthermore, to confirm that ts phenotypes of cin3-983 are ascribable to an atb2-V260I single mutation, fragments carrying the mutated atb2-V260I gene were integrated into the corresponding atb2 locus, which resulted in replacement of wild-type atb2⁺ with this mutation at its endogenous locus (see Materials and Methods). This newly constructed strain indeed displayed ts growth (Figure 2C, middle). Heterozygous diploids containing atb2-983 and wild-type atb2⁺ alleles could form colonies at 36°C (Figure 2C, bottom), indicating that strictly speaking, atb2-983 is not a dominant mutation. From these results, we concluded that cin3 is allelic to atb2, in which its ts phenotype is derived from a single amino acid substitution (V260I) of Atb2 (hereafter, cin3-983 is referred to as atb2-983).

The Atb2-983 Protein Is Incorporated into the Microtubule
To examine the cellular localization of Atb2-983, the GFP cassette (under the control of the thiamine-repressible nmtP3 promoter) (Bähler et al., 1998) was inserted in front of the initiator methionine of the atb2-983 gene (GFP-Atb2-983). Then, localization patterns of GFP-Atb2-983 were examined in exponentially growing cells under repressed conditions. As shown in Figure 2D, like wild-type GFP-Atb2, GFP-Atb2-983 could localize to both interphase MTs (top) and mitotic spindles (bottom). Although Atb2-983 is incorporated into the MTs, it is possible that the atb2-983 mutation results in down-regulation of essential Nda2, which might, at least in part, account for its ts phenotypes. However, protein levels of Nad2 and Atb2 (-983) did not alter significantly between wild type and atb2-983 mutants (Supplemental Figure S1), suggesting that this mutation does not affect the quantity of total -tubulin levels or a ratio between Nda2 and Atb2. Thus the Atb2-983 protein is incorporated into MTs, which might be directly related to the apparent dominance of atb2-983 over the endogenous nda2⁺ gene.

atb2-983 Cells Display Less Dynamic Microtubules and a High Rate of Unequal Chromosome Segregation
As a first step to characterize the defective phenotypes of atb2-983, the MT structure was examined with indirect immunofluorescence microscopy using anti--tubulin antibody (TAT-1). We found that, unlike other ts atb2 mutants previously isolated, which show reduced staining resulting from destabilized MTs (Radcliffe et al., 1998), atb2-983 cells retained apparently intact MT structures during both interphase and mitosis at the restrictive temperature (Figure 3A). Interphase atb2-983 cells showed a filamentous MT array that extended along the cell axis (top, wild type in the left and atb2-983 in the right), and mitotic cells also showed seemingly normal spindle structure with distinct astral MTs like wild-type cells (second and third panels). Furthermore, telophase cells displayed long spindles with a medial ring (fourth panels), and postmitotic cells showed characteristic postanaphase array of interphase MT structures (bottom). This result suggested that MTs are not destabilized in the atb2-983 mutant. We then measured MT dynamics using GFP-Atb2 and GFP-Atb2-983 described previously. Compared with wild type, both the growth and shrinkage rates were decreased down to 33 and 60%, respectively, in the atb2-983 mutant (Table 2). Thus, it seems that the V260I mutation of Atb2 renders MTs less dynamic. It is likely thatthe efficiency of MT polymerization/depolymerization is somehow compromised by the V260I mutation in Atb2.

View larger version (41K): [in this window] [in a new window]   Figure 3. The atb2-983 mutant shows unequal chromosome segregation. (A) Microtubule structure in atb2-983 cells during the cell cycle. Wild-type (513) and atb2-983 cells (KZ94) were incubated at 36°C for 4 h, fixed, and processed for immunofluorescence microscopy with anti--tubulin antibody (TAT-1; green) and DAPI (red). Representative images of wild-type (left) and atb2-983 cells (right) in each cell cycle stage are shown (from the top, interphase, early mitosis, late anaphase, telophase, and postmitosis). (B) Unequal chromosome segregation in atb2-983. Wild-type (top) or atb2-983 cells (bottom three panels) were processed for immunofluorescence microscopy as in described in A. Representative cells displaying chromosome missegregation in atb2-983 are shown. Bars, 10 µm. (C) Quantification of chromosome missegregation. Wild-type and atb2-983 cells were grown at 25°C and shifted to 36°C. After 4 and 6 h, aliquots were taken, fixed, stained with DAPI, and the percentage of chromosome missegregation was counted (n = 100). We did not find any chromosome missegregation in wild-type cells under this condition.

Despite normal-looking MT architectures in fixed samples, consistent with minichromosome loss phenotypes, we often observed mitotic cells that undergo unequal chromosome segregation at the restrictive temperature (Figure 3B). Even at the permissive temperature, 18% of binucleate cells showed chromosome missegregation, and the value was increased to 60% upon temperature shift-up (Figure 3C). This suggested that less dynamic MTs in the atb2-983 mutant impose chromosome missegregation without impairing overall MT structural integrity.

Prolonged Mitosis in the atb2-983 Mutant
Quantification of mitotic cells in fluorescence microscopy showed that in the atb2-983 mutant, the population of mitotic cells was significantly augmented compared with wild-type cells (n = 200; Figure 4A). Simple immunofluorescence, however, did not show which mitotic stage was prolonged in the atb2-983 mutant, because the percentage of mitotic cells containing either short (<2 µm) or long (>2 µm) spindles was increased in atb2-983 (13 or 6% for atb2-983 and 5 or 1% for wild type, respectively; Figure 4A).

To determine precisely the mitotic phases in which the atb2-983 mutant shows delay, time-lapse imaging analysis was performed. To monitor the length of the mitotic spindle in each live cell, the SPB was visualized by RFP-tagged Sad1 (Sad1-RFP), which is a constitutive component of the SPB (Hagan and Yanagida, 1995). In addition, to visualize centromere movement, the centromere of chromosome 2 was marked by the GFP-Lac repressor (GFP-LacI)/lacO system (cen2-GFP), in which the lacO array was integrated within 5 kb of cen2 in a strain carrying GFP-LacI (Straight et al., 1996; Nabeshima et al., 1998; Ding et al., 2004). This assay system enabled us to distinguish each mitotic stage, including the initial stage of SPB separation (prophase equivalent), prometaphase, metaphase, and the subsequent later stages, including anaphase A and B. Observation and quantification of a number of independent mitotic cells showed that the atb2-983 mutant was delayed at three mitotic stages. First, the duration between initiation of SPB separation and onset of anaphase A was extended by 40% (16.4 ± 3.7 min in atb2-983 and 11.6 ± 1.3 min in wild type; Table 3). Representative patterns of kinetics of mitotic phases are shown in Figure 4B). It is of note that this 40% increase in atb2-983 is ascribable to the delay in both prophase and prometaphase (kymograph images with the duration of these phases are shown in Figure 4C). In addition, the period of anaphase B was also expanded twofold (31.3 ± 9.5 min in atb2-983 and 15.6 ± 1.3 in wild type). This extension of anaphase B is largely attributable to the decrease in the velocity of spindle elongation (0.36 ± 0.12 µm/min in atb2-983 and 0.65 ± 0.11 µm/min in wild type). Therefore, the atb2-983 mutation results in slower passage in three distinct stages, prophase, prometaphase, and anaphase B, suggesting that in addition to during interphase, MT dynamics during mitosis is also impeded in this -tubulin mutation.

Bub1, but Not Mad2, Spindle Checkpoint Is Activated in the atb2-983 Mutant
Next, we addressed a relationship between the spindle checkpoint and the ab2-983 mutant. First, kinetochore localization of Bub1 was confirmed by colocalization between Bub1-GFP and Nuf2-CFP, which is a core component of the kinetochore (Nabetani et al., 2001) (Figure 5A). To further investigate the state of spindle checkpoint activation, we examined the localization of another spindle checkpoint protein Mad2. Interestingly, unlike Bub1, Mad2 was not recruited to the kinetochore during mitosis in the atb2-983 mutant (1%; Figure 5, B and C), suggesting that Bub1, but not Mad2, detects the defect in the spindle–kinetochore interaction in the atb2-983 mutant.

View larger version (59K): [in this window] [in a new window]   Figure 5. Bub1-, but not Mad2-, checkpoint is activated in atb2-983 cells. (A) Kinetochore localization of Bub1 in atb2-983 cells. Characteristic mitotic atb2-983 cells that display Bub1-GFP dots (left) at the kinetochores (Nuf2-CFP; middle) are shown (KZ144; 4 h at 36°C). In the merged image (right), Bub1-GFP (green), Nuf2-CFP (red), and DAPI (blue) are shown. (B) Mad2 recruitment to the kinetochore is not enhanced by the atb2-983 mutation. Mad2-GFP and DNA were visualized in nda3-1828 cells (KZ54; left) or atb2-983 cells (KZ156; right) after a 4-h incubation at 36°C. Bars, 10 µm. (C) Quantification of Mad2-GFP and Bub1-GFP dots. The percentage of wild-type and atb2-983 mutant cells (grown at 27°C) that display Bub1 or Mad2 dots (n > 350) is shown. (D) Synthetic lethality of an atb2-983bub1 strain. Tetrads from genetic cross between an atb2-983 strain (KZ102) and bub1::ura4+ strain (393) were dissected and incubated at 27°C for 4 d. Segregants corresponding to atb2-983bub1::ura4+ double mutants (open square) failed to form colonies. (E) Growth properties of atb2-983mad double mutants. Strains with indicated genotypes were spotted on rich medium (5 x 104 cells in the far-left spots for each plate and then diluted 10-fold in each subsequent spot rightward) and incubated at 27°C (left), 32°C (middle), or 36°C (right) for 3 d.

It is of note that the protein kinase activity of Bub1 seems not to be required for keeping viability of atb2-983, because, unlike bub1-deletion, double mutants between atb2-983 and bub1 mutants, which contain either amino acid substitution (K762M, bub1-KD) (Vanoosthuyse et al., 2004) or a whole deletion in the protein kinase domain (bub1-kinase) (Yamaguchi et al., 2003), are viable (Table 4). It is known that Bub1 plays checkpoint-dependent and -independent roles in chromosome segregation, in which its protein kinase activity is not involved in spindle checkpoint signaling (Vanoosthuyse et al., 2004; Kadura et al., 2005). The fact that atb2-982 is viable in bub1 kinase-dead mutant suggested that it is a Bub1-checkpoint function that is required for survival of atb2-983. Dependence of atb2-983 viability on the presence of other checkpoint proteins such as Bub3 and Mph1 also supported this notion. Together, these results suggest that the atb2-983 mutant has a specific defect in the spindle–kinetochore interaction, which activates only a subset of check-point proteins, including Bub1, Bub3, and Mph1, and these checkpoint proteins play an essential role in keeping atb2-983 viable at the permissive temperature.

Bub1 Localizes in Proximity to Only One of the Two SPBs
Given the essential role for Bub1 in atb2-983, we performed time-lapse imaging of Bub1-GFP in this mutant (a strain was also tagged with Sad1-RFP). In wild-type cells, Bub1-GFP signals first occurred as a blob during very early mitosis in the vicinity of the SPB before its separation and were associated with one body of separating SPBs for a short time (6 min). Then, signals disappeared some time during early mitosis (n = 20; Figure 6A, top). In atb2-983 cells, on the other hand, the duration of the Bub1-GFP blob was extended three- to fivefold, in which Bub1-GFP was retained in the vicinity of one SPB (n = 10; Figure 6A, bottom). It is important to note that in both wild-type and atb2-983 cells, Bub1-GFP signals were associated with only one SPB, and we have never observed strong Bub1 dots that were dissociated from the SPB or located between the two SPBs. Because Mad2 does not localize to the kinetochore in atb2-983 cells, Bub1-localizing kinetochores are likely to be attached to the spindle, but not in a bipolar manner.

The Bub1 Checkpoint Prevents atb2-983 Cells from Lethal Chromosome Missegregation
We sought to examine the lethal phenotype of atb2-983bub1 cells. To this end, an atb2-983bub1 strain, which contains cen2-GFP and plasmids carrying the atb2⁺ gene, was constructed. This plasmid is mitotically stable because double mutants are lethal without the episomal atb2⁺ gene. To induce plasmid loss, this strain was starved for nitrogen for 12 h and then allowed to grow again in rich medium. Nitrogen starvation, which results in growth arrest at G₁, is known to induce loss of episomal plasmids (Garcia et al., 2002a). After 8-h incubation upon release into rich medium at 26°C, atb2-983bub1 cells displayed massive chromosome missegregation. As shown in Figure 6, B and C, >70% of binucleate cells (n = 200) showed unequal chromosome segregation, which was evident by staining of cen2-GFP and 4,6-diamidino-2-phenylindole (DAPI). Under the same conditions, a single atb2-983 strain showed modest chromosome missegregation (7%). It should be noted that we hardly ever observed "cut" phenotypes, in which a septum is formed in the middle of unseparated chromosomes. This suggests that atb2-983bub1 cells are capable of segregating chromosomes by the mitotic spindle, albeit unequally. This analysis, therefore, established a role for the Bub1 checkpoint in preventing atb2-983 cells from lethal missegregation of sister chromatids.

EB1 Homologue-encoding mal3⁺ Gene Is a High-Dosage Suppressor
During screening for multicopy suppressor genes of atb2-983, we isolated the mal3⁺ gene (Figure 7A), which encodes the fission yeast homologue of MT-binding protein EB1 (Su et al., 1995; Beinhauer et al., 1997). The mal3 deletion is not lethal (Beinhauer et al., 1997), but when combined with atb2-983, the double mutant grew significantly slower even at the low temperature (27°C) than wild type or each single mutant (Figure 7B, boxed) and displayed noticeable chromosome missegregation phenotypes (Figure 7, C and F). In addition to suppression of temperature-sensitivity of the atb2-983 single mutant (Figure 7A), increased dosage of mal3⁺ rescued synthetic lethality of atb2-983bub1 double mutants (Figure 7D). Nonetheless, suppression of ts atb2-983 or lethal atb2983bub1 cells by high-dosage mal3⁺ was incomplete, because chromosome missegregation was still observed (Figure 7, E and F). Also, atb2-983bub1 mutants containing multicopy mal3⁺ grew slowly (Figure 7D). These results suggested a close functional linkage between Atb2 (-983) and Mal3.

View larger version (46K): [in this window] [in a new window]   Figure 7. Suppression of atb2-983 by multicopy mal3+ gene encoding the EB1 homolog. (A) Suppression of the ts atb2-983 mutation by multicopy plasmids containing mal3+. wt, wild-type cells. (B) Genetic interaction between atb2-983 and mal3. Tetrads from genetic cross between an atb2-983 strain (KZ137) and a mal3-deleted strain (YP350) were dissected and incubated at 27°C for 4 d. Segregants corresponding to atb2-983mal3 double mutants are marked with open squares. (C) Chromosome missegregation of atb2-983mal3 cells. The atb2-983mal3 double mutant (KZ141) was grown at 27°C and stained with DAPI. Cells displaying chromosome missegregation are pointed by arrows. (D) Suppression of the lethality of atb2-983bub1 by a high dosage of mal3+. atb2-983bub1 (pmal3+-LEU2) (KZ135) was obtained from the cross between an atb2-983 strain carrying plasmids containing mal3+ (pmal3+-LEU2) and a bub1::ura4+ strain. (E) Chromosome missegregation of atb2-983bub1 carrying multicopy mal3+. Bar, 10 µm. (F) Quantification of chromosome missegregation. Indicated strains were grown in minimal liquid media lacking leucine at 27°C and shifted to 36°C. At 0-, 4-, and 6-h time points, aliquots were taken and stained with DAPI. The percentage of binucleate cells displaying chromosome missegregation was quantified (n = 100).

To further address a physical interaction between Mal3 and Atb2 (-983), coimmunoprecipitation analysis was performed in wild-type and atb2-983 cells. For this purpose, the chromosomal mal3⁺ gene was tagged with myc epitope (Mal3-Myc), which was crossed with strains containing GFP-Atb2 (-983). Immunoprecipitation using anti-Myc antibody as a primary antibody, followed by immunoblotting with individual antibodies showed that a binding between Mal3 and Atb2-983 is compromised in the mutant at 36°C (Figure 8E, lane 9). In addition, immunoprecipitation using wild-type and atb2-983 mutants containing Mal3-GFP also showed that binding between Mal3-GFP and Atb2-983 is impaired at 36°C (Supplemental Figure S2).

We also found that in a mal3 deletion strain, Bub1-GFP localized, like atb2-983, to the vicinity of one SPB for a prolonged period (Asakawa et al., 2005) (representative Bub1-GFP singles were shown in Figure 8F and see Figure 6A for wild-type control). Furthermore, mal3 mutants display extension of mitotic prophase, prometaphase, and anaphase B and show synthetic genetic interaction with bub1, bub3, or mph1 but not mad1 or mad2 (Asakawa et al., 2005), genetic interactions very similar to those of atb2-983 shown earlier (see Table 4). Together, these results suggested that temperature sensitivity and mitotic defects observed in atb2-983 might stem from, at least in part, compromised Mal3 localization to the mitotic spindle.

Sister Chromatid Oscillation during Prometaphase Is Less Dynamic in the atb2-983 Mutant Independent of Mal3 Dysfunction
It is shown that during mid-mitosis, fission yeast kinetochores display, like animal cells, back-and-forth oscillation between the two poles, which is likely to correspond to the prometaphase state (Saitoh et al., 1997; Garcia et al., 2002b). We addressed whether the atb2-983 mutation affects this phase using cen2-GFP and Sad1-RFP described earlier (see Figure 4). To follow cen2-GFP behavior in more detail, kymograph imaging was performed every 10 s (note that images in Figure 4C were processed every 1-min interval). We observed >30 samples of wild-type mitoses and created kymograph pictures. This analysis showed that cen2-GFP signals indeed displayed rapid and complex patterns of oscillation between the two SPBs (a representative image is shown in Figure 9A, top, and the period of prometaphase is marked with solid arrow at the bottom). Back-and-forth movement of cen2-GFP became more evident when the distance between cen2-GFP and the two SPBs were measured and plotted against time (Figure 9B, left, representative images are shown; n = 12).

In addition to dynamic oscillation, cen2-GFP signals seemed to be often split into two before anaphase onset (Figure 9A, top). This suggested that sister kinetochores are captured by the spindle MTs from both poles at this stage, thereby oscillating between the SPBs via opposing poleward forces imposed at each sister kinetochore. This oscillation period was followed by segregation of cen2-GFP signals toward each pole (anaphase A, shown with bidirectional yellow arrows). As reported previously (Tournier et al., 2004), there was a transient pause (1–2 min) between prometaphase and onset of anaphase A, in which cen2-GFP signals tended to be situated in the middle of the two SPBs, likely corresponding to fission yeast metaphase (marked by black arrowhead).

We then observed atb2-983 mutant cells. Despite a prolonged delay during early mitosis as shown earlier (see Figure 4B), cen2-GFP signals were segregated equally under live imaging conditions (at the room temperature). However, unlike wild type, atb2-983 cells did not display dynamic oscillation of cen2-GFP. Instead, in this strain, cen2-GFP stayed near the equator throughout mid-mitotic phase without back-and-forth movement (n = 10; Figure 9A, bottom; also see Figure 4C). Plotting of cen2-GFP positions relative to the two SPBs clearly indicated less dynamic behavior of cen2-GFP compared with wild-type cells (Figure 9B, right, representative patterns are shown). These observations suggest that in the atb2-983 mutant, antagonizing forces normally generated at the sister kinetochores during prometaphase are not functional, which renders centromere oscillatory movement more static.

We next asked whether static prometaphase behavior of cen2-GFP in atb2-983 is ascribable to Mal3 malfunction. We, however, found that unlike atb2-983, sister centromeres still oscillate in the mal3 mutant, although a mild reduction of oscillation seemed to occur (n = 10; a representative kymograph image is shown in Figure 9C). Together, our results suggest that, in addition to compromised Mal3 localization and function, there are Mal3-independent deficiencies in atb2-983 mutants, including the failure of sister centromere oscillation during prometaphase.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
In this study, we have presented characterization of a novel mutant allele of the fission yeast -tubulin encoding atb2⁺ gene. The atb2-983 mutant displays a number of intriguing and somewhat unexpected phenotypes. Through our analysis, we have uncovered novel aspects of fission yeast mitosis, such as dynamic behavior of sister centromeres, surveillance of the mitotic kinetochore by the Bub1-branch checkpoint, and a role for MT dynamics in bipolar microtubule attachment. We think that the results obtained in this study would be of direct relevance toward understanding of chromosome segregation in higher eukaryotes.

Reduced Mal3 Localization in the atb2-983 Mutation
Unlike other ts atb2 mutants previously isolated previously (Yaffe et al., 1996; Radcliffe et al., 1998), in the atb2-983 allele (V260I) overall MT structures are not impaired noticeably, but instead the growth and shrinkage rates of interphase MTs are reduced. Also, mitosis is slowed down, including the velocity of anaphase spindle elongation. It is, therefore, possible that MT dynamics is in general impaired in this mutant. We show that increased dosage of mal3⁺ rescues atb2-983 and Mal3 loading onto the MT is significantly compromised. Consistently, whereas Mal3 and Atb2 interact in wild-type cells, a binding between Mal3 and Atb2–983 is hampered. This suggests that the region around V260, which is situated close to the interphase of the /-tubulin heterodimer and also faces toward the lattice of MT protofilaments, is critical for an interaction of Atb2 with Mal3. Budding yeast Mal3 homologue Bim1 is also shown to interact with -tubulin (Tub1), but the mutations analogous to V260I have not been analyzed in this organism (Schwartz et al., 1997; Richards et al., 2000).

It should be pointed out that amino acid substitution from valine to isoleucine, which results in addition of another methyl group to the side chain, is usually considered as a conservative replacement. However, the fact that this valine is invariant among all -tubulin species in various organisms underscores the importance of valine at this site. The V260I mutation would induce subtle but deleterious conformational alterations in -tubulin, thereby compromising an interaction with Mal3 without affecting either /-heterodimer formation or incorporation of these heterodimers into MTs. Indeed, we show that Atb2-983 is capable of being incorporated into MTs. Compromised Mal3 binding combined with the incorporation-competent properties of Atb2-983 would contribute to the apparent dominant nature of this mutant protein over essential -tubulin isoform Nda2.

It is of note that in the atb2-983 mutant, compared with its mitotic defects, interphase deficiencies are less noticeable. The mal3-deleted mutant displays, on the other hand, a number of interphase phenotypes such as shortened cytoplasmic MTs and resulting bent or branched cell morphology (Beinhauer et al., 1997). It is thus possible that an interaction between Mal3 and its binding partners or Mal3 function is subject to cell cycle-dependent regulation. Mal3 is a phosphoprotein (Busch and Brunner, 2004) and analysis of Mal3's structural domains and modification during the cell cycle would address this question.

Phenotypic Similarities between mal3 Mutants and atb2-983
Reduced localization of Mal3 to the spindle in atb2-983 suggests that some of mitotic defects in this mutant are derived from compromised Mal3 function. Consistently, prolonged passage through distinct mitotic stages and longer lasting Bub1 kinetochore localization seem to be ascribed to Mal3 malfunction, because a mal3 deletion strain also shows very similar phenotypes (Asakawa et al., 2005; Asakawa and Toda, 2006). It should be noted that Mad2 localization to the kinetochores is not prolonged in either mal3 or atb2-983. In both vertebrates and yeast, Mad2 localizes to, mostly, if not exclusively, unattached mitotic kinetochores (Waters et al., 1998; Kapoor et al., 2000; Skoufias et al., 2001; Garcia et al., 2002a; Ikui et al., 2002; Gillett et al., 2004), suggesting that the kinetochores of mitotic atb2-983 or mal3 cells are captured by the spindle. However, subsequent establishment of bipolar attachment is substantially delayed or defective, in which Bub1 is retained in the kinetochores that are associated with only one of the two SPBs.

We have recently shown that most, if not all, of mitotic delay in the mal3 mutant is ascribable to prolonged syntelic attachment of the kinetochore to the spindle, in which spindles emanated from one pole interact with both of the sister kinetochores (Asakawa et al., 2005; Asakawa and Toda, 2006). Given the phenotypic similarities between mal3 and atb2-983 strains and the suppression of atb2-983 by high dosage of mal3⁺, we envisage that atb2-983 cells also display prolonged syntelic attachment during early mitotic stage, in which the Bub1 spindle checkpoint plays a role in preventing chromosome missegregation. It should be pointed out that a specific requirement of the Bub1 branch spindle check-point is analogous to that reported for the spindle orientation checkpoint (Rajagopalan et al., 2004; Tournier et al., 2004). atb2-983 mutants might be defective in the positioning or rotation of the mitotic spindle, which would require spindle dynamics.

Sister Centromere Oscillation during Prometaphase
Mal3 dysfunction is not likely to be the sole reason for the atb2-983 phenotype, because the mal3 deletion is not ts. Time-lapse live analysis has revealed that whereas wild type or mal3 mutant cells display dynamic sister centromere oscillation during prometaphase, atb2-983 cells show sister centromeres situated in the middle of the spindle with very little oscillatory movement. This implies that mitotic spindles in atb2-983 are defective in generation of pulling and/or pushing forces at the kinetochore during prometaphase, the stage in which Mal3 function is not involved. Disappearance of Mal3 from the tips of mitotic spindles during mid-mitosis in wild-type cells is in line with this notion (Asakawa et al., 2005). Because neither Bub1 nor Mad2 localizes to these kinetochores during prometaphase in atb2-983, it is possible that amphitelic attachment has already been established at this stage, nonetheless anaphase onset being delayed mechanically by compromised spindle dynamics.

Another possibility is that the kinetochore attachment is merotelic, in which a single kinetochore is attached to the spindles emanating from both poles (Cimini et al., 2001). It is shown that merotelic attachment occurs frequently during early mitosis in mammalian tissue cells (Cimini et al., 2003). A single kinetochore is capable of interacting with multiple MTs in fission yeast as in animal systems (Ding et al., 1993); therefore, merotelic configuration could occur in this organism. Indeed, lagging chromosomes that are frequently observed in a subset of chromosome segregation mutants are reported to represent merotelic attachment (Pidoux et al., 2000). It is possible that the atb2-983 mutant is hampered in correcting this attachment during prometaphase, possibly because of compromised dynamics of mitotic spindles. As in animal cells, the process of kinetochore–spindle attachment during fission yeast prometaphase is likely to be stochastic, which requires the dynamic nature of spindle MTs.

In conclusion, we show that in fission yeast dynamism of the spindle MT is crucial for accurate sister chromatid segregation. In this process, Mal3-EB1 interacts with -tubulin and localizes to mitotic spindles, thereby ensuring establishment of stable bipolar MT attachment possibly by enhancing spindle dynamics. In the absence of Mal3 function, syntelic attachment is prolonged, which results in activation of the Bub1/Bub3/Mph1 checkpoint. In addition, MT dynamics is involved in sister chromatid oscillation during prometaphase, which is largely independent of Mal3 function. Fission yeast mitosis is, thus, remarkably similar to that in higher eukaryotes and further analysis will contribute to elucidation of the unified view on MT dynamics and accurate chromosome segregation at the molecular levels.

	ACKNOWLEDGMENTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION ACKNOWLEDGMENTS REFERENCES

 
We thank Drs. Damian Brunner, Ursula Freig, Keith Gull, Kevin Hardwick, Yasushi Hiraoka, Jean-Paul Javerzat, Tomohiro Matsumoto, Osami Niwa, Paul Nurse, Shelley Sazer, Mizuki Shimanuki, Vincent Vanoosthuyse, Satoko Yamaguchi, and Ayumu Yamamoto for providing materials used in this study. We are grateful to Dr. Jacqueline Hayles for critical reading of the manuscript and useful suggestions. This work is supported by grants from the Ministry of Education, Science and Culture of Japan (to D. H.) and by the Cancer Research UK and the Human Frontier Science Program research grant (to T. T.).

	Footnotes

 
This article was published online ahead of print in MBC in Press (http://www.molbiolcell.org/cgi/doi/10.1091/mbc.E05–08–0802) on January 4, 2006.

Abbreviations used: MT, microtubule; SPB, spindle pole body; ts, temperature sensitive.

The online version of this article contains supplemental material at MBC Online (http://www.molbiolcell.org).

These authors contributed equally to this work.

Present address: Division of Molecular and Developmental Biology, Department of Developmental Genetics, National Institute of Genetics, Yata 1111, Mishima, Shizuoka 411-8540, Japan.

Address correspondence to: Takashi Toda (toda{at}cancer.org.uk).