15 Pages


Course Number: MB 523, 2009

College/University: Princeton

Word Count: 10077


Document Preview

Cell, Vol. 100, 619633, March 17, 2000, Copyright 2000 by Cell Press Establishing Biorientation Occurs with Precocious Separation of the Sister Kinetochores, but Not the Arms, in the Early Spindle of Budding Yeast Gohta Goshima* and Mitsuhiro Yanagida* CREST Research Project * Department of Biophysics Graduate School of Science and Department of Gene Mechanisms Graduate School of Biostudies Kyoto University...

Unformatted Document Excerpt
Coursehero >> New Jersey >> Princeton >> MB 523

Course Hero has millions of student submitted documents similar to the one
below including study guides, practice problems, reference materials, practice exams, textbook help and tutor support.

Course Hero has millions of student submitted documents similar to the one below including study guides, practice problems, reference materials, practice exams, textbook help and tutor support.

Vol. Cell, 100, 619633, March 17, 2000, Copyright 2000 by Cell Press Establishing Biorientation Occurs with Precocious Separation of the Sister Kinetochores, but Not the Arms, in the Early Spindle of Budding Yeast Gohta Goshima* and Mitsuhiro Yanagida* CREST Research Project * Department of Biophysics Graduate School of Science and Department of Gene Mechanisms Graduate School of Biostudies Kyoto University Kitashirakawa-Oiwakecho, Sakyo-ku Kyoto 606-8502 Japan protein (GFP)-tagging method for visualizing proteins and DNA sequences has opened up a way to monitor the behavior of the kinetochores during the cell cycle of budding yeast and fission yeast, which contain small kinetochore structures difficult to observe by electron microscopy (Ding et al., 1993; Winey et al., 1995; Straight et al., 1997; Michaelis et al., 1997; Saitoh et al., 1997; Nabeshima et al., 1998; Goshima et al., 1999). During replication in a typical eukaryotic cell, the centromere sequence is duplicated along with the chromatid DNAs, forming sister centromere DNAs and sister kinetochores that are held together until anaphase. During this mitotic stage, rapid separation of sister chromatids takes place. Kinetochores associate with kinetochore microtubules that connect chromosomes with the spindle poles: one end of the microtubules is bound to the kinetochores and the other end to the spindle poles that serve as the microtubule organization center (MTOC). When the bipolar spindle is formed, sister kinetochores are bound to kinetochore microtubules that in turn are joined to one of the spindle poles. Biorientation of chromosomes can thus be established in mitotic metaphase, with kinetochore microtubules exerting a pulling force on the sister chromatids to bring them toward the opposite poles. The opposing force against pulling requires the presence of the connection between sister kinetochores. Defects in this connection may abolish correct chromosomal segregation as seen in the cases of fission yeast mis6 and mis12 mutants (Takahashi et al., 1994; Saitoh et al., 1997; Goshima et al., 1999). The kinetochore-microtubule interactions thus occur only during mitosis, and sister kinetochores are freed from microtubules in interphase. Kinetochores may receive a signal that triggers shortening of the kinetochore microtubules to pull chromosomes toward the poles (e.g., Gorbsky, 1997). We undertook a study of the budding yeast Saccharomyces cerevisiae Mtw1p (mis twelve like), as its predicted amino acid sequence resembled that of the fission yeast Schizosaccharomyces pombe kinetochore protein, Mis12 (Goshima et al., 1999). As this would be the first clear example of homologous kinetochorelocalized proteins between two evolutionarily distant yeasts, it is of considerable interest to determine how they are functionally and structurally related. Behavior of Mtw1p during the cell cycle is of particular interest, because Mis12 colocalizes with the centromere DNAs throughout the cell cycle by associating with the inner nonrepetitive regions of each centromere. Budding yeast kinetochore proteins have often been found to be located along the short spindle or near the spindle pole bodies (SPBs) (Goh and Kilmartin, 1993; Meluh et al., 1998; Hyland et al., 1999). In mis12-537 mutant cells, the metaphase spindle is considerably expanded, and this expansion is proposed to be the basis for disrupting correct segregation of mitotic chromosomes (Goshima et al., 1999). There are two enigmatic features in the differences of the kinetochores and spindles between S. pombe and S. cerevisiae. First, the functional size of S. cerevisiae Summary Sister kinetochores are bioriented toward the spindle poles in higher eukaryotic prometaphase before chromosome segregation. We show that, in budding yeast, the sister kinetochores are separated in the very early spindle, while the sister arms remain associated. Biorientation of the separated kinetochores is achieved already after replication. Mtw1p, a homolog of fission yeast Mis12 required for biorientation, locates at the centromeres in an Ndc10p-dependent manner. Mtw1p and the sequences 1.8 and 3.8 kb from CEN3 and CEN15, respectively, behave like the precociously separated kinetochores, whereas the sequences 23 and 35 kb distant from CEN3 and CEN5 previously used as the centromere markers behave like a part of the arm. Mtw1p and Ndc10p are identically located except for additional spindle localization of Ndc10p. A model explaining small centromeres and early spindle formation in budding yeast is proposed. Introduction A particular chromosomal DNA segment called the centromere serves as the site for proteinDNA and protein protein interactions to form a higher-order structure, the kinetochore, which ensures high-fidelity chromosome segregation in mitosis and meiosis (Allshire, 1997; Karpen and Allshire, 1997). Centromere DNAs have been identified in diverse eukaryotic cells, and detailed genetic analyses show that their functional sizes range from the order of one hundred base pairs in budding yeast (Fitzgerald-Hayes et al., 1982) and tens of thousands of base pairs in fission yeast (Takahashi et al., 1992) to millions of base pairs in animals (Sun et al., 1997; Ikeno et al., 1998). Behavior of centromeres during the cell cycle was investigated in animal cells that contain relatively large kinetochore structures (Rieder and Salmon, 1994). More recently, the identification of centromere DNAs and kinetochore proteins in lower eukaryotes allows the functional and structural dissection of kinetochores (e.g., Meluh and Koshland, 1997; Saitoh et al., 1997; Meluh et al., 1998; Goshima et al., 1999; Ortiz et al., 1999). Introduction of the green fluorescent To whom correspondence should be addressed (e-mail: yanagida@ kozo.biophys.kyoto-u.ac.jp). Cell 620 centromere DNAs is 100-fold smaller than that of S. pombe (Fitzgerald-Hayes et al., 1982; Takahashi et al., 1992). This is not easily explained because the genome sizes of the two yeasts are roughly the same. The average size of individual chromosomal DNAs is severalfold longer in S. pombe than S. cerevisiae, as the number of S. pombe chromosomes is three, while that of S. cerevisiae is sixteen. As the largest chromosome XII of S. cerevisiae does not greatly differ from the smallest S. pombe chromosome III, the average chromosome size difference is not sufficient to explain such a large discrepancy in the centromere DNAs. Secondly, cell cycle timing for formation of the spindle differs greatly between the two organisms. In S. pombe, the SPBs begin to have the MTOC activity for spindle formation only after cells enter mitosis (Masuda et al., 1992), whereas formation of the spindle in S. cerevisiae occurs early in the cell cycle, probably during the S phase (Byers and Goetsch, 1975). Unbudded S. cerevisiae cells in the G1 phase contain a single SPB, while very small spindles with separated SPBs can be observed in cells with a small bud so that a high population of budded cells contains spindles of variable sizes. In sharp contrast, only a small population of S. pombe growing cells contain spindles: the period of spindle formation (prophase), constant spindle length (metaphase and anaphase A), and spindle elongation lasts 1.5, 4.0, and 6.2 min, respectively, within a cell cycle length of 130 min at 36 C (Nabeshima et al., 1998). In this study, we show that a principal difference in chromosome behavior during the cell cycle between the two organisms lies in the timing for separation of sister centromere DNAs. We found that, in S. cerevisiae, sister centromere DNAs as well as sister kinetochores are already separated when the tiny spindle is formed. The distance between separated sister centromeres/kinetochores increased to nearly 1 m in the medial short spindle. The sister arms, however, remain associated during this period, perhaps generating the opposing force against pulling by kinetochore microtubules, and separate only later in regular anaphase accompanied by spindle extension. These surprising results challenge the currently held concept on timing of sister centromere separation in this organism, but our finding helps to explain a number of unique features in budding yeast spindles and kinetochores. Results Interaction of Mtw1p with Centromeres in an Ndc10p-Dependent Manner To identify the gene product, the budding yeast MTW1 gene was tagged with 8Myc at the 3 terminus and integrated into the chromosome of the S. cerevisiae wildtype strain W303-1a. Immunoblotting of cell extracts from the integrated MTW1-Myc strain using anti-myc antibodies produced the expected 53 kDa band (Figure 1A, lane 2), whereas no band was detected in the control nonintegrated W303-1a strain (lane 1). The tagged Mtw1p was functional, as plasmids carrying the MTW1Myc gene conferred the ability to rescue gene disrupted mtw1 null cells (described below). We examined whether Mtw1p was the centromerebound protein using the chromatin immunoprecipitation (CHIP) method (Meluh and Koshland, 1997; Saitoh et al., 1997). Extracts prepared from formaldehyde-fixed cells were sonicated to reduce the average DNA length to less than 1 kb and used for immunoprecipitation by anti-myc antibodies. Coprecipitated DNAs were extracted and used as templates for polymerase chain reaction (PCR): the four primers used were CEN16, CEN3, a sequence 500 bp apart from CEN16, and an AT-rich sequence located 24 kb distant from CEN3. Only the primers CEN16 and CEN3 coprecipitated with Mtw1p-8Myc (Figure 1B). The sequences 0.5 and 24 kb apart from the centromeres were not present in the precipitates. Beads only or extracts without the Myc tag did not produce any amplification of the centromere DNAs. As a positive control, the gene encoding the authentic centromere protein Ctf19p (Hyland et al., 1999) tagged with Myc at the C terminus was integrated into the chromosome and found to coprecipitate with centromere DNAs at levels similar to Mtw1p-8Myc. These results demonstrated that Mtw1p physically interacted with the centromere DNAs; the sequence 500 bp distant from the centromere was no longer capable of interacting with Mtw1p. The procedure of the above CHIP experiment was likewise performed with cell extracts of a metaphase-arrested cdc16-1 mutant (Cdc16p is a component of the APC [anaphase promoting complex]/ cyclosome complex required for polyubiquitination of mitotic cyclins and Pds1p; Irniger et al., 1995). No significant difference was observed in the results obtained from asynchronous culture (data not shown), suggesting that interaction occurs in both interphase and mitosis. To determine whether coprecipitation of the centromere DNAs with Mtw1p requires other centromere proteins, we introduced an ndc10-1 mutation (Goh and Kilmartin, 1993) into the Mtw1-8Myc integrated strain, and the CHIP experiment was again performed. It is known that budding yeast centromeres are associated with the CBF3 complex, and the degree of interaction between the cen DNAs and CBF3 components was diminished in ndc10-1 mutant at 36 C, the restrictive temperature, as Ndc10p is an essential subunit of this complex (Sorger et al., 1995). As shown in Figure 1C, the levels of both coprecipitated CEN16 and CEN3 were severely diminished in ndc10-1 cultured at 36 C for 2.5 hr. Therefore, binding of Mtw1p to the centromere DNA appears dependent on the presence of functional Ndc10p. Mtw1p Is Essential for Cell Viability To determine whether the MTW1 gene is essential for cell viability, a plasmid carrying the disrupted MTW1 gene was made and introduced into the chromosome of an S. cerevisiae diploid strain SP1/DC124 (Toda et al., 1985), replacing one of the two native MTW1 genes by one step gene replacement (Rothstein, 1983). Gene disruption in heterozygous diploid cells was verified by PCR and Southern hybridization (data not shown). Tetrad dissection of spores derived from the heterozygous diploids showed 2 :2- segregation: all the viable spores were Leu at 2236 C, showing that Mtw1p is required for viability. Preearly Centromere Separation in Budding Yeast 621 Figure 1. Interaction of Mtw1p with Centromere DNAs (A) Identification of Mtw1p by immunoblotting. The MTW1-Myc gene with native promoter integrated into the chromosome of wild-type W3031a is detected by anti-myc antibodies (lane 2). The nonintegrated control W303-1a (lane 1) does not produce any band. (B) The CHIP method reveals interaction of Mtw1p with CEN16 and CEN3. Extracts of cells integrated with the MTW1-Myc gene were prepared and were immunoprecipitated with anti-myc antibodies. PCR was applied for amplification of the coprecipitated DNAs using the primers of CEN3, CEN16, the sequences 500 bp apart from CEN16, and those 24.5 kb apart from CEN3. For a negative control, only beads were used. As a positive control, extracts of cells containing the integrated CTF19-Myc gene (Ctf19p is a nonessential centromere component [Hyland et al., 1999]) were used. The lanes for input were the PCR products from extracts before immunoprecipitation. (C) Interaction of Mtw1p with CEN16 and CEN3 depends on the presence of functional Ndc10p. CHIP was performed for extracts of wildtype and ndc10-1 mutant integrated with the MTW1-Myc gene. ndc10-1 mutant cells were cultured at 23 and 36 C for 2.5 hr. The levels of the amplified centromere DNAs from ndc10-1 cultured at 36 C were greatly reduced. Temperature-Sensitive mtw1-1 Mutant Leads to Unequal Chromosome Segregation Knowing that the fission yeast mis12-537 mutation is a G to E substitution at the 52nd position in the N-terminal region that is conserved between Mis12 and Mtw1p (Goshima et al., 1999), we constructed an mtw1-1 mutant containing a corresponding substitution at the 64th residue from Gly (GGA) to Glu (GAG). Plasmid pMTW11[trp] carrying the mutant gene G64E was used to replace the native MTW1 gene by integration using TRP1 as a marker (Experimental Procedures). The resulting mutant designated mtw1-1 was found to be temperature sensitive (ts) as mis12-537, and this ts phenotype could be rescued by transformation using pMTW1. The phenotype of mtw1-1 was investigated in the synchronous culture using factor (10 g/ml) at 26 C for 1 hr 40 min. Resulting G1-arrested mutant cells were shifted to 36 C and released after 20 min by washing out factor (time 0). Cells were collected at 20 min intervals, and cellular phenotypes as well as the DNA content and cell viability were determined. Mutant cells underwent S phase around 4060 min (Figure 2A) and then arrested with a large bud, short spindle, and a single nucleus in the neck for about 40 min (76%88% from 100 to 140 min, shown in Figure 2B). Cell viability started to decrease at 140 min in parallel with the appearance of cells with elongated spindles accompanying unequally segregated chromosomes (C and D). Mutant cells thus entered mitosis but temporally arrested with a short spindle followed by abnormal separation of sister chromatids. Examples of DAPI-stained mtw1-1 cells are shown in the left panel of (D). Unequal chromosome segregation along the spindle is also clearly seen at the late mitotic stage (180 min) in mtw1-1 cells stained by DAPI (DNA) and anti-tubulin antibody (TUB). Cells revealing aberrant segregation reached 28%34% after 160180 min (B). In asynchronous cultures at 36 C, consistent phenotypes were observed. Cell division was temporally arrested at 36 C, and cells containing a single mitotic nucleus accumulated and their frequencies peaked at 66% after 2 hr (data not shown). The phenotypes of Mtw1p-depleted cells were also examined by the shut-off experiment using the GAL promoter, and phenotypes highly resembling that of ts mutant cells were observed by Mtw1p depletion (data not shown). This unequal segregation of chromosomes at 36 C was consistent with the instability of centromere-containing plasmids in mtw1-1 at the semirestrictive temperature, 30 C, by the colony color assay method (data not shown). Expanded Metaphase Spindle in mtw1-1 The metaphase spindle of the fission yeast mis12-537 mutant was 60% longer than that of the wild-type (Goshima et al., 1999). To determine whether the metaphase spindle in mtw1-1 was also longer than normal, the double mutant mtw1-1 cdc16-1 was constructed and cultured at 36 C. cdc16-1 is arrested in metaphase at the restrictive temperature. The size of the metaphase spindle was determined by the distance between the SPBs visualized by GFP-tagged Tub4p ( -tubulin homolog localized at the SPBs throughout the cell cycle; e.g., Pereira and Schiebel, 1997) expressed in the single cdc16-1 and the double mtw1-1 cdc16-1 mutant cultured at 36 C for 3 hr (Figure 2E). Cells were fixed and stained with DAPI. In the double mutants, the frequency of large budded cells with a single nucleus reached 80% while that in the single mutants was approximately 50%. In these cells, the distances between the SPBs were measured (Figure 2F). The average length of the mitotic spindle in mtw1-1 cdc16-1 was found to be 2.7 0.8 m, while that of the single cdc16-1 mutant was 1.8 0.6 m. The spindle length therefore increased 50% in the background of mtw1-1, consistent with the aberrantly expanded metaphase spindle in the fission yeast mutant mis12-537. Cell 622 Figure 2. Synchronous Culture of mtw1-1 at 36 C (AC) For the synchronous culture, factor (10 g/ml) was used at 26 C for 1 hr 40 min. G1-arrested mtw1-1 cells were transferred to 36 C and released after 20 min by washing out factor (time 0). Cells collected at 20 min intervals were examined for DNA contents by FACS analysis (A), and cellular phenotypes were determined by anti-tubulin antibody and DAPI stain (B and C) and cell viability (C). Percentages of cells showing different mitotic phenotypes are tabulated in (B). Stages corresponding to S and G2/M are indicated by arrows in (C). Mutant Preearly Centromere Separation in Budding Yeast 623 Figure 3. Kinetochore Localization of Mtw1p Requires Ndc10p (A) Mtw1-GFP expressed by the chromosomally integrated gene was observed as one or two dots in wild-type cells. Hoechst 33342 was used for DNA staining. (B) Cells expressing Mtw1p were fixed and stained by anti-tubulin antibodies. Arrows indicate the position of SPBs. (C) Images of a single living cell expressing Mtw1-GFP were taken at 1 min intervals for 30 min (numbers indicate min) at 26 C. The time 0 was set when the distance between the two signals rapidly increased in anaphase. Circles illustrate the mother cell and bud. (D) Dot-like localization of Mtw1p requires presence of functional Ndc10 protein. ndc10-1 mutant cells cultured at 26 C (0 hr) showed normal localization, whereas the dot-like signals disappeared after 24 hr at 36 C. Wild-type control is shown. Bars, 10 m. (E) The level of Mtw1-GFP was maintained in ndc10-1 mutant cells at 36 C. Mtw1p was detected by anti-GFP antibodies. Lane 1, wild-type (W303-1a); lane 2, Mtw1-GFP in wild-type at 26 C; lane 3, Mtw1-GFP in ndc10-1 mutant at 26 C; lane 4, Mtw1-GFP in wild-type at 36 C for 2.5 hr; lane 5, Mtw1-GFP in ndc10-1 at 36 C for 2.5 hr. Localization of Mtw1p along the Spindle The MTW1-GFP gene was integrated onto the chromosome of wild-type cells to determine intracellular localization. The dot-like GFP signal was clearly seen in all the cells examined (Figure 3A). In unbudded G1 cells (1,2), a single dot was always found at the periphery of the nucleus, whereas in budded cells with a single nucleus (35), two dots were seen at variable distances was temporally arrested with the short spindle, followed by abnormal separation of sister chromatids. (D) Left panel, DAPI-stained mtw1-1 cells revealing unequal segregation (an example indicated by arrow). Right panel, short spindle arrested cells (120 min) and cells displaying unequal chromosome segregation at late stages with the long spindle or without (180 min). Bars, 10 m. (E and F) Spindle length is expanded in mtw1-1. Single cdc16 cells were arrested in mid mitosis at 36 C revealing the short spindle (SPBs visualized by GFP-tagged Tub4p). In the double mutant cdc16 mtw1, the metaphase spindle with longer length was made. Distribution of spindle length is shown in (F). Cell 624 Figure 4. Mtw1p Colocalizes with the Sequences 1.83.8 kb from Centromeres but Not with Those Positioned 23 and 35 kb Distant from Centromeres (A) The TetO repeat integrated at the position 35 kb apart from the CEN5 previously used as the centromere marker (Michaelis et al., 1997) does not colocalize with Mtw1-GFP. Micrographs obtained from nonfixed cells showing Mtw1-GFP and CEN5(35)-GFP taken by a confocal microscope. Preearly Centromere Separation in Budding Yeast 625 (0.40.8 m). It was surprising that two dots could already be seen even in cells with a small bud, considering that Mtw1p was shown to interact with the centromeres by the CHIP method. In late mitotic cells showing the dividing nucleus (6), one dot was observed for each daughter nucleus. Positioning of Mtw1-GFP in regard to the spindle apparatus could be assigned by simultaneously observing the images obtained by Mtw1-GFP and anti-tubulin antibody (TAT1) in fixed cells (Figure 3B). The GFP signal was closely (but not identically, see results described in following sections) situated at the SPB in the unbudded G1 cell (left), while two GFP signals were positioned near the SPBs of the short and long spindles (middle, right). The arrows indicate the positions of the SPBs so that Mtw1p seems to be located slightly inward toward the nucleus from the SPBs. The behavior of Mtw1-GFP in single living cells was monitored at intervals of 1 min (Experimental Procedures). Wild-type cells expressing Mtw1-GFP by the integrated gene were observed at 26 C (Figure 3C, the number indicates min). A budded cell initially revealing a single ellipsoidal GFP signal at 21 min showed two dots at 14 min, and their distance increased to about 1 m around 4 to 0 min, followed by rapid spindle extension (starting from 1 min). One of the two signals was seen to migrate into the bud at 3 min. Note that in our observation, separation of the GFP signals appeared to be continuous and did not apparently show any feature of breathing or reversible movements of association and dissociation. However, such breathing could take place and might be unrecognized by our light microscopy. Mtw1p Did Not Colocalize with DNA Sequences 30 kb Apart from the Centromeres We addressed the question of whether Mtw1p colocalizes with the sequences previously used as markers for the centromere-linked DNAs. For this purpose, we constructed two strains integrated with the TetO repeat (TetO: Michaelis et al., 1997) situated approximately 35 kb from CEN5 or the LacO repeat (LacO) that resides 23 kb from CEN3 (Straight et al., 1997) and also expressing the MTW1-GFP gene. We thus could observe simultaneously the signals of the marker sequence and Mtw1p in the same cells. Results obtained suggested that these sequences did not colocalize with Mtw1-GFP. In single unbudded cells of the strain TetR-GFP Mtw1GFP, two GFP signals were always observed, while three GFP signals (one intense and two relatively weak) were detected in budded cells (Figure 4A, left). These results can be explained by presuming that Mtw1-GFP produced one and two signals in unbudded and budded cells, respectively, while the TetR-GFP produced one intense signal until nuclear division in large budded cells. GFP signals in the strain integrated with only the TetRGFP are shown at right as control. Basically identical results were obtained for the LacI-GFP Mtw1-GFP strain (data not shown). We thus confirmed the behavior of the marker sequences used in the previous studies and found that these sequences apparently do not colocalize with Mtw1-GFP. The greater fraction of Mtw1-GFP might not be present in the centromere regions. Alternatively, the real centromere DNAs may be visualized differently from the positions of the TetO or LacO sequences used. These sequences were, in fact, quite distant from the actual CEN5 and CEN3 (approximately 10 m as naked DNA). Preearly Separation of the Sister Centromeres, but Not the Arms, in Cells with the Short Spindle To examine the above possibilities, we constructed strains integrated with the LacO repeats closer to the actual centromere DNAs. One strain CEN3 (3.8)-GFP was constructed with the 10 kb LacO repeat integrated at a position 3.8 kb from CEN3 with the marker gene URA3, while another strain CEN15(1.8)-GFP contained integration of the LacO repeat at 1.8 kb distant from CEN15 (Experimental Procedures). We first examined the behavior of CEN3 (3.8)-GFP during the cell cycle in comparison with that of CEN3 (23)-GFP 23 kb apart from CEN3 in a strain in which two LacO repeats are integrated: one repeat 23 kb distant from CEN3 and the other repeat 3.8 kb from CEN3 so that the physical Ndc10p-Dependent Localization of Mtw1p The interaction of Mtw1p with the centromere DNAs revealed by the CHIP method requires Ndc10p. We therefore examined whether intracellular localization of Mtw1-GFP was dependent on the presence of functional Ndc10p. For this purpose, an ndc10-1 mutant containing the integrated MTW1-GFP gene was constructed and observed at 36 C (Figure 3D). Localization was normal at 0 hr but no longer visualized after 2 and 4 hr at 36 C. In control wild-type cells expressing Mtw1-GFP, normal signal was found after 4 hr at 36 C. Disappearance of the Mtw1-GFP signal was not due to degradation of Mtw1-GFP, as the protein level did not decrease in ndc10-1 mutant cells after 2.5 hr at 36 C as shown in Figure 3E. Normal intracellular localization of Mtw1p hence requires functional Ndc10p. (B) The LacO repeat 23 kb distant from CEN3 previously used as the centromere marker (Straight et al., 1997) does not colocalize with the other LacO repeat at 3.8 kb from CEN3 made in this study (the integrated strain made is illustrated at the top). Cells were fixed with ethanol and observed by a fluorescence microscope. (C) The sequence CEN15 (1.8)-GFP only 1.8 kb apart from CEN15 is similarly localized with Mtw1-GFP (Figure 3B). Positioning of the integrated LacO repeat integrated 1.8 kb distant from CEN15 in relation to the spindle was determined by anti-tubulin (TUB) antibody staining. (D) The LacO repeat integrated 3.8 kb distant from CEN3 is also similarly localized with Mtw1-GFP. In budded cells, two signals along the short spindle were observed, while the signal was single in unbudded. (E) The LacO repeat integrated 23 kb from CEN3 localizes differently from Mtw1-GFP. The GFP signal of CEN3(23)-GFP was clearly single in budded cells with short spindle, and splitting of the signals occurred only in the long spindle. In (C)(E), cells were formaldehyde fixed and stained by anti-GFP and anti-tubulin (TAT1) antibodies. DNA was stained by DAPI. Bars, 10 m. Cell 626 Figure 5. Distinct Localization of Centromere and Arm Signals in the Spindle and mtw1-1 Mutant (A) Clustering of the two LacO repeats CEN15(1.8)-GFP and CEN3(3.8)-GFP integrated in the same strain is clearly seen in unbudded G1 cells (panels 1 and 2). Clustering may persist, as only two signals were seen in the short spindle (panels 35). Three signals were only occasionally seen in late nuclear division. (B) The signals of Tub4-GFP (Tub4 is a -tubulin homolog located at the SPBs) and CEN15(1.8)-GFP visualized in the same cells are closely connected. Intense Tub4-GFP signals are indicated by the arrows. Separated sister centromere signals are on the inner side along the spindle. (C) Double labeling experiment was done in wild-type cells expressing CEN15(1.8)-GFP and Mtw1-Myc. GFP and Myc were visualized in the same cells by FITC and rhodamine, respectively. Their signals were nearly identical, but the signal of CEN15(1.8)-GFP is often off center of the signals derived from Mtw1p-Myc (see text). (D and E) In wild-type cells containing the short spindle, positions of CEN [CEN15(1.8)-GFP] and ARM [CEN3(23)-GFP] signals were determined by the coordinates (x/L and y/L) and plotted in (D) and (E), respectively. Spindle length is L, with an average of 1.4 0.2 m for CEN15(1.8) and 1.5 0.2 m for CEN3(23). x and y are defined in the text. Each pair of split CEN signals shown by letters (D) is situated closely to the spindle and the spindle poles. However, the arm signals are always seen as single dots (E) and centrally distributed but away from the spindle axis. (F) mtw1-1 mutant cells containing CEN15 (1.8)-GFP were cultured at 36 C for 23 hr and observed after anti-tubulin antibody staining. Two examples of mutant cells exhibiting the spindle with abnormal sister centromeres are shown (arrow indicates the single signal, see text). Bars, 10 m. distance of these two repeats was 27 kb. As shown in Figure 4B, two GFP signals were seen in most unbudded G1 cells, while in budded cells, three and infrequently four GFP signals were observed. These results strongly suggested that the two LacO repeats behaved differently in the cell cycle. In the strain CEN15(1.8)-GFP containing the singly integrated LacO repeat, the GFP signal was indistinguishable from that of Mtw1-GFP (data not shown). To determine precisely the position of the GFP signal in regard to the spindle, CEN15(1.8)-GFP cells were observed after anti-tubulin (TUB) antibody staining (Figure 4C). In unbudded G1 cells, the GFP signal was single, locating near the SPB. In budded cells with short spindles, two signals were clearly observed along the short spindle. Even in tiny spindles, the GFP signals were split into two within the resolution of light microscopy ( 0.3 m). These results suggested the preearly separation of sister centromeres in the short spindle. In anaphase cells containing elongated spindles, the signals were Preearly Centromere Separation in Budding Yeast 627 located near the end of the spindle. These results are identical to the localization of Mtw1p. Similar results were obtained for the strain integrated with CEN3 (3.8)GFP (Figure 4D). As a control, cells containing the LacO repeat 23 kb apart from CEN3 are shown in Figure 4E. The GFP signal was still single in budded cells with the short spindle, and separation occurred only in the long spindle. The same results were obtained for the other integrated repeat 35 kb apart from CEN5 (data not shown). We interpret these results to mean that the regions 23 or 35 kb apart from CEN3 and CEN5, respectively, behave actually as parts of the arms, which remain associated until spindle extension, rather than the centromeric DNAs. We further constructed strains that contained a reduced LacO repeat length from the initial 10 to 4 kb and also of only 1 kb in the CEN3(3.8)-GFP strain. Behavior of the signals in these strains with 4 and 1 kb LacO length during mitotic cell cycle was identical to that of 10 kb LacO strain, although the intensity of the GFP signals was progressively weakened (data not shown). This result completely eliminates the possibility that precocious sister centromere separation is caused by the insertion of the long 10 kb LacO repeats near the centromeres. Furthermore, chromosome segregation of those strains integrated with the LacO repeats was examined by tetrad dissection and found to produce normal segregation pattern (data not shown). Clustering of Centromere and Mtw1p Signals and Their Localization Distinct from the SPBs To determine mutual localization patterns of different centromere DNAs, a strain containing two LacO repeats CEN15(1.8)-GFP and CEN3(3.8)-GFP was constructed and observed. One or two closely situated dots were seen in unbudded cells (Figure 5A, panels 1 and 2), indicating that the two cen signals were closely situated or formed a cluster. Only two signals were found in many budded cells (panels 35), suggesting that clustering of centromere DNAs took place even along the short spindle. More than two signals were seen in cells at the late stage of nuclear division (panel 6). We then examined how close the centromere DNA signals were located to the SPBs. For this purpose, cells integrated with the LacO repeat of CEN15 (1.8)-GFP and expressing GFP-tagged -tubulin (Tub4p, an essential component of the SPBs) were used. The signals of Tub4GFP and CEN15(1.8)-GFP were closely but never identically located within the resolution of light microscopy, suggesting that the distance between the two signals exceeded 0.3 m. For cells containing the short spindle, the signals derived from the CEN15(1.8)-GFP were clearly located apart from the Tub4-containing SPBs (indicated by the arrows), residing in the interior side toward the center of the spindle (Figure 5B). GFP signals in the strain singly integrated with the Tub4-GFP are shown at the bottom as control. These results suggested that the centromere DNAs were connected with the SPBs by kinetochore microtubules. To demonstrate that Mtw1p is colocalized with CEN DNA, double labeling was performed in the wildtype cells by immunofluorescence microscopy using CEN15(1.8)-GFP and Mtw1-Myc. GFP and Myc were visualized in the same cells FITC- by and rhodamineconjugated antibodies. Their signals were nearly identical (Figure 5C). Note that Mtw1p binds to all 16 pairs of sister centromeres, whereas CEN15(1.8)-GFP reveals only one centromere DNA, so that the signal of CEN15 (1.8)-GFP is often off center of the signals derived from Mtw1p-Myc. Distinct Localization of Sister Centromeres and Arms in Mitotic Cells To determine how sister centromeres and arms are positioned relative to the spindle, CEN15(1.8)-GFP or CEN3(23)-GFP signals were measured in wild-type cells stained by anti-tubulin antibodies. The following distances were measured: spindle length (L), distances between the GFP signals and the vertical line running across the spindle poles (x), and distances between signals and the spindle axis (y). Thus, x/L and y/L determine the relative positions of signals in the spindle matrix (5D and 5E). For CEN15(1.8)-GFP, 92% of cells showed two signals, whereas 96% of cells containing CEN3(23)-GFP produced only one signal. The cen signals (5D) were positioned near the spindle axis and the spindle poles, absent from the central region of the spindle. The arm signals (5E), however, were centrally distributed but away from the spindle axis. These results are consistent with the notion that the centromeres are always separated while the arms are continually associated. Two broad peaks of frequencies of the CEN signals were approximately 0.30.4 m distant from the vertical line running across the spindle poles, suggesting that the separated centromeres are not directly associated with the SPBs. Established biorientation of sister centromeres in the short spindle was frequently abolished in mtw1-1 mutant cells after 2 hr at 36 C. Mutant cells containing CEN15 (1.8)-GFP were cultured at 36 C for 23 hr, and the number of GFP signals along the spindle was counted. Initially, 90% cells showed two separated signals, but that value decreased to 57% after 3 hr. A single signal on the spindle was seen in 42% of the cells (an example shown in the upper panel of 5F). Another example of aberrant cells exhibiting abnormal segregation of separated sister centromeres is shown in the lower panel. Effect of Nocodazole and Hydroxyurea on Sister Centromere Separation To determine whether sister centromeres remained separated when the spindle is damaged, the strain CEN15(1.8)-GFP was observed after addition of nocodazole (15 g/ml). Ninety six percent of cells having a large bud showed single dots in the presence of nocodazole (Figure 6A). The FACS analysis indicated that the cells contained 2C. Similar results were obtained in cells containing the GFP signal from the arm CEN3(23)-GFP. Disruption of the spindle structure thus appeared to induce reassociation of the sister centromeres. Consistently, the once separated SPBs were pulled so close to each other by depolymerization of the microtubules in the presence of nocodazole that they could not be resolved by light microscopy (Jacobs et al., 1988). To confirm that sister centromeres reassociated in the presence of nocodazole, cells were first uniformly arrested in metaphase in cdc16-1 mutants at 36 C for 3 hr, and then nocodazole was added. As shown in Figure 6B, two CEN15-GFP dots seen in 85% of metaphase-arrested cells remained in only 5% of cells after the addition of Cell 628 Figure 6. Behavior of CEN DNAs under Spindle and Replication Checkpoint Control (A) Two strains integrated with the LacO repeat at the position 1.8 kb from CEN15, CEN15(1.8)-GFP, or 23 kb from CEN3, CEN3(23)-GFP were blocked by the addition of nocodazole for 2.5 hr at 26 C. The GFP signals observed became single in both strains when cells were blocked with a large bud. (B) cdc16-1 mutant cells containing CEN15(1.8)-GFP were grown exponentially at 20 C and then arrested at metaphase at 36 C for 3 hr in the absence of nocodazole, followed by further culture at 36 C for 1.5 hr in the absence or the presence of nocodazole (15 g/ml). Cells were collected and fixed by ethanol, followed by DAPI staining. Typical example cells of different treatments are shown. (C) The CEN15(1.8)-GFP strain was blocked by hydroxyurea, and the GFP signals were observed. Bars, 10 m. nocodazole. Turning off spindle force thus led to reassociation of sister centromeres. We then tested whether sister centromeres were separated in the presence of hydroxyurea (HU) using the CEN15(1.8)-GFP strain as above. The FACS pattern indicated the 1C DNA peak in the presence of hydroxyurea (100 mM) at 26 C for 2.5 hr. Twenty-nine percent of HUblocked cells showed two CEN15(1.8)-GFP dots (71% were single), whereas only 7% of the CEN3(23)-GFP strain cells showed two signals (Figure 6C). These results were confirmed in two independent experiments. This might be due to early timing of replication in the centromeres. Double Color Labeling of Ndc10 and Mtw1 To compare the localization of Mtw1p with that of Ndc10p, an authentic kinetochore protein, we attempted to visualize Mtw1p and Ndc10p by different colors in the same cells. For this end, Ndc10-GFP and Mtw1CFP (cyan fluorescent protein) were constructed and integrated into the chromosome. First, single color Ndc10-GFP was observed in concert with anti-tubulin staining (Figure 7A). Ndc10-GFP was seen as a single dot in cells without buds and two dots in budded cells as reported previously (Goh and Kilmartin, 1993). Note that the GFP signals in the short spindle are always toward the interior of the spindle as is Mtw1p, being discrete from the localization of the SPBs. Additional localization along the spindle was found for Ndc10-GFP, particularly intense in late anaphase. In cells coexpressing Ndc10-GFP and Mtw1-CFP (Figure 7A, right), these two proteins were observed using appropriate filters and found to be localized identically except for the additional presence of Ndc10-GFP in the spindle between sister Preearly Centromere Separation in Budding Yeast 629 Figure 7. Behavior of Centromeres/Kinetochores during the Cell Cycle of Budding Yeast (A) Left panel, cells integrated with Ndc10-GFP were observed after staining by DAPI and anti-tubulin antibody. Right panel, cells expressing integrated Ndc10-GFP and Mtw1-CFP were photographed after ethanol fixation. See text for explanation. Bars, 10 m. (B) Sister centromeres (red circle) are separated at an early stage in cells containing the very short spindle during or after the S phase. The sister arms (cross) are separated later during anaphase spindle extension. Short kinetochore microtubules (0.3 m long; Winey et al., 1995 and OToole et al., 1999) appear to connect the spindle poles with the precociously separated sister centromeres in the short spindle. Localization of Mtw1p (yellow) and Ndc10p (green) is also shown. centromeres (dim fluorescence along the short spindle but intense in the long spindle). These results confirm that Mtw1p is a kinetochore protein and that CEN DNA is colocalized with two kinetochore proteins, Ndc10p and Mtw1p. Ndc10p may interact with pole to pole microtubules as suggested previously (Goh and Kilmartin, 1993). Discussion We show that a novel budding yeast kinetochore protein Mtw1p resembling fission yeast inner centromere protein Mis12 is essential for cell viability and that its specific interaction with the centromeres requires Ndc10p, a component of the CBF3 kinetochore protein complex. The ts mtw1-1 mutant induces the formation of longer metaphase spindles and leads to unequal segregation of sister chromatids. These results are similar to those obtained for Mis12 (Goshima et al., 1999). Mis12/Mtw1plike proteins are also found in the genome of filamentous fungi, suggesting that this family of proteins is conserved in higher eukaryotes. An emerging concept from the present study is that establishment of the biorientated sister kinetochores is a crucial aspect leading to correct sister chomatid separation, and this may be generally true for all organisms. Timing of sister kinetochore separation is probably variable among species, and budding yeast may belong to the earliest, as demonstrated in this study. In fission yeast, transient splitting of centromere-linked sequences occurs only in prometa and metaphase (Nabeshima et al., 1998). Similarly, the centromeres in newt lung cells under tension are stretched in prometaphase and metaphase (Waters et al., 1996). In budding yeast, nearly full sister centromere separation in the medial-sized spindle seems to be allowed by association of the sister arms until anaphase. The arm association likely generates an opposing force against pulling by the spindle. Our results strongly suggested that precocious centromere separation was permanent or continuous after replication, but very Cell 630 fast breathing by spindle dynamics might not be recognized due to limitation of light microscopy. In the early stage of this investigation, the unexpected localization behavior of Mtw1-GFP during the cell cycle raised the question whether Mtw1-GFP really colocalized with centromeres, as its pattern strikingly differed from that of the previously used centromere-linked LacO or TetO sequences (Michaelis et al., 1997; Straight et al., 1997): the reported signals were single until the stage of sister chromatid separation, suddenly splitting into two almost simultaneously with anaphase spindle extension. In sharp contrast, the signals of Mtw1-GFP were already separated and localized near SPBs in cells containing the short spindle. The distance between the continually separated signals of Mtw1p was initially tiny but increased to nearly 1 m as the spindle reached medial size in the mother cells (illustrated in Figure 7B). This apparent discrepancy in localization patterns was solved when we employed other strains containing LacO repeat sequences that were integrated at positions much closer to the centromeres than those used before (1.8 3.8 kb apart from CEN15 and CEN3 in this study instead of 2335 kb from CEN3 and CEN5 in the previous study). We found that these closer sequences colocalized with Mtw1p and Ndc10p. These and other results led us to conclude that the signals in the previous reports represented behavior of the arm sequences rather than the centromeres and that budding yeast sister kinetochores are precociously separated. A principal feature of budding yeast chromosomes during the cell cycle may thus be the distinct timing of separation between sister centromeres/kinetochores and sister arms. Precocious separation of sister kinetochores may be necessary for biorientating them toward the spindle poles. This result is consistent with a recent report showing that centromere-containing plasmids separate precociously (Tanaka et al., 1999). Our finding implies that once biorientation of the sister kinetochores is established, precocious kinetochore separation does not cause any harm in anaphase. When are sister centromeres actually separated in budding yeast? They appear to be separated in the early stage of cell cycle, probably during the S phase, because a significant fraction of cells arrested by hydroxyurea contain split centromeres. Separated sister kinetochores were also observed in cells with a very small bud or on the very short spindle. It is known that timing of replication for centromere DNAs is early in the S phase (McCarroll and Fangman, 1988). Replication and separation of centromeres might be temporally linked. Byers and Goetsch (1975) clearly showed that the SPB is duplicated at the time of bud emergence that coincides with the onset of chromosome replication. The formation of the short spindle is nearly coincident with the completion of replication. We show that separated sister kinetochores are positioned near, but distinct from, the separated SPBs, but sister arms remain associated until spindle elongation. Budding yeast kinetochores/centromeres are always closely positioned to, though they may not be directly associated with, the SPBs. How then do kinetochore microtubules function in budding yeast? Kinetochore microtubules in higher eukaryotic cells associate with chromosomal kinetochores only during mitosis and are shortened to pull the separated sister kinetochores toward the opposite poles in anaphase. In metaphase, there are paired kinetochore microtubules that are associated with nonseparated sister kinetochores of one chromosome (Ding et al., 1993). Our results predict that, in budding yeast, such paired kinetochore microtubules may be lost in the short spindle, as the sister kinetochores are separated and situated near the poles. Thin-section electron microscopy (Winey et al., 1995) indeed produced microtubule images consistent with this. The postulated kinetochore microtubules are only 0.30.4 m long in the short spindle present in the mother cells. A sufficient number of such polar microtubules can be counted to allow for one to associate with each kinetochore and several to form an interpolar spindle. The presence of such 0.30.4 m kinetochore microtubules is also consistent with the localization of sister centromeres visualized in this study (Figure 5D). Note that electron microscopy fails to reveal kinetochore structures in budding yeast so that identification of kinetochore microtubules by electron microscopy is not unambiguous. The stage to increase the pole to pole distances called anaphase B perhaps occurs when sister arms are separated. Are kinetochore microtubules shortened, and, if so, when are they shortened? OToole et al. (1999) proposed that shortening might occur after spindle elongation, suggesting that the decrease in the distance between the kinetochores and the poles (anaphase A) might take place in telophase after full spindle elongation. A similar conclusion was obtained for another budding yeast, Candida albicans (Chibana and Tanaka, 1996). Rapid trafficking of a set of separated sister chromatids through the narrow neck in anaphase B might require avoidance of their clustering around the poles, so that each might be associated with kinetochore microtubules for successive trafficking of separated individual chromatids through the neck, delaying the timing of an anaphase Alike event until telophase. The above kinetochore/centromere behavior is strikingly different from those of typical mammals (Rieder and Salmon, 1994) and that of the fission yeast S. pombe (Funabiki et al., 1993; Saitoh et al., 1997; Nabeshima et al., 1998; Goshima et al., 1999). An explanation is thus required for why the budding yeast kinetochores are separated early in the cell cycle. We would like to propose that duplicated sister centromere DNAs are bound to kinetochore microtubules that may pull apart the sister kinetochores. One possible reason for such early association between spindle microtubules and kinetochores may be the tiny size of the centromere sequences in budding yeast, which may not be easily recognized or captured by microtubules if kinetochores move rapidly along the prometaphase spindle. The large size of centromere DNAs may be needed for capture by microtubules during the rapid movement of kinetochores along the spindle. If this movement of sister chromatids before separation is unavoidable due to dynamic properties of the spindle, budding yeast might have devised a way to prepare the linkage between the sister centromeres and the spindle poles long before the entry into mitosis to ensure high-fidelity segregation of chromosomes. The centromeres of budding yeast might be bound to kinetochore microtubules throughout the cell cycle. Preearly Centromere Separation in Budding Yeast 631 The effect of nocodazole on separated kinetochores and SPBs may explain why cells arrested by the spindle checkpoint resume mitosis upon removal of the drug. In the presence of nocodazole, sister centromeres were reassociated and the SPBs were close together, probably due to the disappearance of the spindle, leading to the apparent reversal of centromeres and SPBs to the early stage in mitotic apparatus formation. This would mechanically ensure that cells can reinitiate spindle structure formation after removal of the drug. Based on the finding of reassociation of sister kinetochore DNAs in the absence of the spindle, colocalization of CEN DNA with Ndc10p, and visualization of only 1 kb long LacO, it is very unlikely that integration of LacO and TetO repeats on the centromere regions leads to artificial dissociation of sister centromeres. We advise that interpretation and generalization of the budding yeast centromere/kinetochore behavior and cell cycle regulation related to the control of kinetochore behavior should be made cautiously. Of particular interest is the implication for the role of cohesin that is enriched in the centromeres even in metaphase (Blat and Kleckner, 1999; Megee et al., 1999; Tanaka et al., 1999), during which sister centromeres are already separated. Cohesin, for example, might play a catalytic role rather than function as a structural link so that it may accumulate on the preseparated centromeres while its substrate might be later bound or inactivated. Some budding yeast kinetochore proteins may be required for the specific role of early separation or maintenance of separated sister kinetochores until spindle elongation, and some kinetochore proteins such as Ndc10p may also exist along the spindle between the separated centromeres to ensure separation of sister kinetochores. Experimental Procedures Strains, Media, and Culture Media The complete YPD (1% yeast extract, 2% polypeptone, and 2% glucose) and minimal SD (0.67% yeast nitrogen base w/o amino acids, and 2% glucose) media were used for culturing of S. cerevisiae. Strains used for gene disruption of MTW1 and integration of MTW1 tagged with GFP were previously described (Toda et al., 1985). All other experiments used strains with W303 background (derivatives of strain W303-1a; Mat a ade2-101 ura3-1 leu2-3,112 trp1-1 his3-11 can1-100). LB (0.5% yeast extract, 1% polypeptone, and 1% NaCl at pH 7.5) medium was used to grow Escherichia coli MM294 and STBL2. Epitope Tagging A NotI site was introduced at the termination codon of the MTW1 gene by PCR mutagenesis, and the resulting EcoRV-NotI fragment was cloned into the integration vector pYC11 (Chikashige et al., 1989). The NotI-SacI cassette containing GFP (or 8xMyc, ECFP [Clontech]) and nmt1-polyA-tail was inserted at the end of the MTW1 gene. The resulting plasmid was introduced into the haploid wildtype strain at the endogenous MTW1 locus after linearization with SnaBI. The integrants of CTF19-Myc, NDC10-GFP, and TUB4-GFP were similarly constructed. These strains grew normally, indicating that the epitope-tagged genes were fully functional. Gene Disruption One-step gene replacement (Rothstein, 1983) was employed. The EcoRV fragment (20 bp) in the coding region of the MTW1 gene was replaced with the S. cerevisiae LEU2 gene, and this disrupted MTW1 DNA fragment was used for transformation of the diploid strain, SP1/DC124 (Mat a/ leu2/leu2 ura3/ura3 trp1/trp1 his3/his4 ade8/ ade8 can1/ ). LEU heterozygous transformant cells were dissected by tetrad analysis. Gene disruption was verified by Southern blotting (data not shown). Only two spores of the tetrad were viable, and those spores were leu , indicating that MTW1 was essential for cell viability. Isolation of mtw1-1 Mutant The fission yeast mis12-537 gene, which had been amplified by PCR from genomic DNA of S. pombe h mis12-537 strain, was sequenced. A substitution at the 52nd position from Gly to Glu was found. The MTW1-1 N fragment, in which 64Gly65Val (ggagtt) was changed to 64Glu65Leu (gagctc) by PCR mutagenesis and the N-terminal 81 bp of MTW1 gene was deleted, was cloned into TRP1 integration plasmid made in this study, and ADH1 transcription termination sequences were added. The resulting pMTW1-1[trp] was linearized with NcoI and introduced into the haploid wild-type strain at the endogenous MTW1 locus. Integration via homologous recombination and introduction of mutations were verified by Southern blotting and PCR amplification. The mtw1-1 strain thus obtained formed colonies at 26 C but not at 36 C. Introduction of LacO Repeat into Centromere Proximal Region The centromere proximal regions (bp 116801117791, 3 kb to the right of CEN3; bp 325382326387, 1 kb to the left of CEN15), containing a SacI site at the CEN proximal terminus and a KpnI site at the CEN distal terminus, were amplified by PCR from S. cerevisiae genomic DNA. This fragment and the LacO repeat (Straight et al., 1996) were both cloned into yeast integration plasmid YIplac211 containing URA3 gene. Plasmids were transformed into LacI-GFP expressing ura cells after linearization with BglII (CEN3) or SacII (CEN15). Southern blotting was done on the genomic DNAs of the CEN3(3.8)-GFP strain digested with SacI/KpnI or SacI/SmaI. The probe used was a PCR-amplified CEN fragment (bp 116801 117791). For the CEN15(1.8)-GFP strain, genomic DNAs were digested with SacI or BglII/EcoRI, and Southern blotting was done using PCR-amplified CEN region (bp 325382326387) as the probe. Detection of the expected bands (data not shown) confirmed the integration, via homologous recombination, at the centromere proximal regions. Native genes located near CEN3 and CEN15 were not disrupted by integration of LacO sequences. The URA3 marker of the CEN3(3.8)-GFP strain was linked to the LEU2 ( 23 kb to the left of CEN3) locus (PD:TT:NPD 26:7:0). For the experiments at 36 C, thermo-stable LacI-GFP (pAFS144 ; Straight et al., 1998) was used. Microscopy Immunofluorescence microscopy using anti-tubulin and anti-GFP antibodies was done after formaldehyde fixation as described by Pringle et al. (1991). Cells were fixed by 3.7% formaldehyde for 2.5 hr and then stained by anti-tubulin TAT1 (1:50 dilution; Woods et al., 1989), anti-Myc monoclonal 9E10 (Calbiochem; 1:100 dilution), and anti-GFP polyclonal (Clontech; 1:500 or 1:200 dilution) antibodies after methanol/acetone treatment. For immunofluorescence microscopy to visualize microtubules in the MTW1-GFP integrant, the methanol fixation method (Hagan and Hyams, 1988) was used in order to maintain the signals of GFP. Ethanol fixation also allowed direct observation of the LacI-GFP signals without the use of antiGFP antibodies. DAPI was used for the staining of DNA after fixation, and Hoechst 33342 to stain DNA of nonfixed cells. The DAPI staining of mtw1-1 cells was done after 2.5% glutaraldehyde fixation (Adachi and Yanagida, 1989). For observing living cells, wide-field deconvolution 3D microscopy system, DeltaVision program (Applied Precision, Inc., Seattle, WA), was used without optical fiber illumination as described in Haraguchi et al. (1999). Every 60 s, 4 optical sections (0.1 s exposure time) were collected (vertical separation 0.4 m) at 26 C. Chromatin Immunoprecipitation and Immunological Methods Chromatin immunoprecipitation (CHIP) and immunological methods employed the procedure described by Saitoh et al. (1997). For ndc10-1 mutant, cells cultured at 36 C for 2.5 hr were used for formaldehyde fixation. Anti-Myc antibody 9E10 (Calbiochem) was used for immunoprecipitation, and portions of immunoprecipitated DNAs (1/60) were used as PCR template. The primer sequences of Cell 632 CEN3, CEN16 and CEN3 proximal regions were described by Meluh and Koshland (1997). CEN16 proximal primers (5 -AGGTCCTTC AATTCGTCTGC-3 and 5 -CGCCATCAATACCTGTATTAG-3 ) were synthesized. PCR products were separated on 2.5% agarose gels and visualized by ethidium bromide. For immunoblotting, cell extracts were prepared using HB buffer (Moreno et al., 1989). Hydroxyurea and Nocodazole Treatment Hydroxyurea and nocodazole were added to 100 mM and 15 g/ml to mid-log YPD cultures at 26 C, respectively. After 2.5 hr incubation, cells were fixed with ethanol and observed for fluorescence microscopy or used for FACS analysis. Acknowledgments We are greatly indebted to Yasushi Hiraoka for time lapse observation of Mtw1-GFP and Andrew Murray and Kim Nasmyth for strains to visualize the centromere-linked LacO and TetO repeat sequences, and thank Hiroyuki Araki, Katsuhiko Shirahige, Eiko Tsuchiya, Takashi Toda, and Keith Gull for strains, plasmids, and antibodies, and Dr. Ruth T. Yu for editing the manuscript. We thank all the laboratory members for helpful discussions and moral support. The present study was supported by the Core Research for Evolutional Science and Technology (CREST) research project of Japan Science and Technology Corporation and a Human Frontier Science Project Organization research grant. Received September 13, 1999; revised February 14, 2000. References Adachi, Y., and Yanagida, M. (1989). Higher order chromosome structure is affected by cold-sensitive mutations in a Schizosaccharomyces pombe gene crm1 which encodes a 115- kD protein preferentially localized in the nucleus and its periphery. J. Cell Biol. 108, 11951207. Allshire, R.C. (1997). Centromeres, checkpoints and chromatid cohesion. Curr. Opin. Genet. Dev. 7, 264273. Blat, Y., and Kleckner, N. (1999). Cohesins bind to preferential sites along yeast chromosome III, with differential regulation along arms versus the centric region. Cell 98, 249259. Byers, B., and Goetsch, L. (1975). Behavior of spindles and spindle plaques in the cell cycle and conjugation of Saccharomyces cerevisiae. J. Bacteriol. 124, 511523. Chibana, H., and Tanaka, K. (1996). Analysis of the cell cycle in the budding yeast Candida albicans by positioning of chromosomes by fluorescence in situ hybridization (FISH) with repetitive sequences. Genes Cells 1, 727740. Chikashige, Y., Kinoshita, N., Nakaseko, Y., Matsumoto, T., Murakami, S., Niwa, O., and Yanagida, M. (1989). Composite motifs and repeat symmetry in S. pombe centromeres: direct analysis by integration of NotI restriction sites. Cell 57, 739751. Ding, R., McDonald, K.L., and McIntosh, J.R. (1993). Three-dimensional reconstruction and analysis of mitotic spindles from the yeast, Schizosaccharomyces pombe. J. Cell Biol. 120, 141151. Fitzgerald-Hayes, M., Clarke, L., and Carbon, J. (1982). Nucleotide sequence comparisons and functional analysis of yeast centromere DNAs. Cell 29, 235244. Funabiki, H., Hagan, I., Uzawa, S., and Yanagida, M. (1993). Cell cycle-dependent specific positioning and clustering of centromeres and telomeres in fission yeast. J. Cell Biol. 121, 961976. Goh, P.Y., and Kilmartin, J.V. (1993). NDC10: a gene involved in chromosome segregation in Saccharomyces cerevisiae. J. Cell Biol. 121, 503512. Gorbsky, G.J. (1997). Cell cycle checkpoints: arresting progress in mitosis. Bioessays 19, 193197. Goshima, G., Saitoh, S., and Yanagida, M. (1999). Proper metaphase spindle length is determined by centromere proteins mis12 and mis6 required for faithful chromosome segregation. Genes Dev. 13, 16641677. Hagan, I.M., and Hyams, J.S. (1988). The use of cell division cycle mutants to investigate the control of microtubule distribution in the fission yeast Schizosaccharomyces pombe. J. Cell Sci. 89, 343357. Haraguchi, T., Ding, D.-Q., Yamamoto, A., Kaneda, T., Koujin, T., and Hiraoka, Y. (1999). Multiple-color fluorescence imaging of chromosomes and microtubules in living cells. Cell Struct. Funct. 24, 294298. Hyland, K.M., Kingsbury, J., Koshland, D., and Hieter, P. (1999). Ctf19p: a novel kinetochore protein in Saccharomyces cerevisiae and a potential link between the kinetochore and mitotic spindle. J. Cell Biol. 145, 1528. Ikeno, M., Grimes, B., Okazaki, T., Nakano, M., Saitoh, K., Hoshino, H., McGill, N.I., Cooke, H., and Masumoto, H. (1998). Construction of YAC-based mammalian artificial chromosomes. Nat. Biotechnol. 16, 431439. Irniger, S., Piatti, S., Michaelis, C., and Nasmyth, K. (1995). Genes involved in sister chromatid separation are needed for B-type cyclin proteolysis in budding yeast. Cell 81, 269278. Jacobs, C.W., Adams, A.E., Szaniszlo, P.J., and Pringle, J.R. (1988). Functions of microtubules in the Saccharomyces cerevisiae cell cycle. J. Cell Biol. 107, 14091426. Karpen, G.H., and Allshire, R.C. (1997). The case for epigenetic effects on centromere identity and function. Trends Genet. 13, 489496. Masuda, H., Sevik, M., and Cande, W.Z. (1992). In vitro microtubulenucleating activity of spindle pole bodies in fission yeast Schizosaccharomyces pombe: cell cycle-dependent activation in Xenopus cell-free extracts. J. Cell Biol. 117, 10551066. McCarroll, R.M., and Fangman, W.L. (1988). Time of replication of yeast centromeres and telomeres. Cell 54, 505513. Megee, P.C., Mistrot, C., Guacci, V., and Koshland, D. (1999). The centromeric sister chromatid cohesion site directs Mcd1p binding to adjacent sequences. Mol. Cell 4, 445450. Meluh, P.B., and Koshland, D. (1997). Budding yeast centromere composition and assembly as revealed by in vivo cross-linking. Genes Dev. 11, 34013412. Meluh, P.B., Yang, P., Glowczewski, L., Koshland, D., and Smith, M.M. (1998). Cse4p is a component of the core centromere of Saccharomyces cerevisiae. Cell 94, 607613. Michaelis, C., Ciosk, R., and Nasmyth, K. (1997). Cohesins: chromosomal proteins that prevent premature separation of sister chromatids. Cell 91, 3545. Moreno, S., Hayles, J., and Nurse, P. (1989). Regulation of p34cdc2 protein kinase during mitosis. Cell 58, 361372. Nabeshima, K., Nakagawa, T., Straight, A.F., Murray, A., Chikashige, Y., Yamashita, Y.M., Hiraoka, Y., and Yanagida, M. (1998). Dynamics of centromeres during metaphase-anaphase transition in fission yeast: Dis1 is implicated in force balance in metaphase bipolar spindle. Mol. Biol. Cell 9, 32113225. Ortiz, J., Stemmann, O., Rank, S., and Lechner, J. (1999). A putative protein complex consisting of Ctf19, Mcm21, and Okp1 represents a missing link in the budding yeast kinetochore. Genes Dev. 13, 11401155. OToole, E.T., Winey, M., and McIntosh, J.R. (1999). High-voltage electron tomography of spindle pole bodies and early mitotic spindles in the yeast Saccharomyces cerevisiae. Mol. Biol. Cell 10, 2017 2031. Pereira, G., and Schiebel, E. (1997). Centrosome-microtubule nucleation. J. Cell Sci. 110, 295300. Pringle, J.R., Adams, A.E., Drubin, D.G., and Haarer, B.K. (1991). Immunofluorescence methods for yeast. Methods Enzymol. 194, 565602. Rieder, C.L., and Salmon, E.D. (1994). Motile kinetochores and polar ejection forces dictate chromosome position on the vertebrate mitotic spindle. J. Cell Biol. 124, 223233. Rothstein, R.J. (1983). One-step gene disruption in yeast. Methods Enzymol. 101, 202211. Saitoh, S., Takahashi, K., and Yanagida, M. (1997). Mis6, a fission yeast inner centromere protein, acts during G1/S and forms specialized chromatin required for equal segregation. Cell 90, 131143. Preearly Centromere Separation in Budding Yeast 633 Sorger, P.K., Doheny, K.F., Hieter, P., Kopski, K.M., Huffaker, T.C., and Hyman, A.A. (1995). Two genes required for the binding of an essential Saccharomyces cerevisiae kinetochore complex to DNA. Proc. Natl. Acad. Sci. USA 92, 1202612030. Straight, A.F., Belmont, A.S., Robinett, C.C., and Murray, A.W. (1996). GFP tagging of budding yeast chromosomes reveals that proteinprotein interactions can mediate sister chromatid cohesion. Curr. Biol. 6, 15991608. Straight, A.F., Marshall, W.F., Sedat, J.W., and Murray, A.W. (1997). Mitosis in living budding yeast: anaphase A but no metaphase plate. Science 277, 574678. Straight, A.F., Sedat, J.W., and Murray, A.W. (1998). Time-lapse microscopy reveals unique roles for kinesins during anaphase in budding yeast. J. Cell Biol. 143, 687694. Sun, X., Wahlstrom, J., and Karpen, G. (1997). Molecular structure of a functional Drosophila centromere. Cell 91, 10071019. Takahashi, K., Murakami, S., Chikashige, Y., Funabiki, H., Niwa, O., and Yanagida, M. (1992). A low copy number central sequence with strict symmetry and unusual chromatin structure in fission yeast centromere. Mol. Biol. Cell 3, 819835. Takahashi, K., Yamada, H., and Yanagida, M. (1994). Fission yeast minichromosome loss mutants mis cause lethal aneuploidy and replication abnormality. Mol. Biol. Cell 5, 11451158. Tanaka, T., Cosma, M.P., Wirth, K., and Nasmyth, K. (1999). Identification of cohesin association sites at centromeres and along chromosome arms. Cell 98, 847858. Toda, T., Uno, I., Ishikawa, T., Powers, S., Kataoka, T., Broek, D., Cameron, S., Broach, J., Matsumoto, K., and Wigler, M. (1985). In yeast, RAS proteins are controlling elements of adenylate cyclase. Cell 40, 2736. Waters, J.C., Skibbens, R.V., and Salmon, E.D. (1996). Oscillating mitotic newt lung cell kinetochores are, on average, under tension and rarely push. J. Cell Sci. 109, 28232831. Winey, M., Mamay, C.L., OToole, E.T., Mastronarde, D.N., Giddings, T.H., Jr., McDonald, K.L., and McIntosh, J.R. (1995). Three-dimensional ultrastructural analysis of the Saccharomyces cerevisiae mitotic spindle. J. Cell Biol. 129, 16011615. Woods, A., Sherwin, T., Sasse, R., MacRae, T.H., Baines, A.J., and Gull, K. (1989). Definition of individual components within the cytoskeleton of Trypanosoma brucei by a library of monoclonal antibodies. J. Cell Sci. 93, 491500.