Mitosis in Living Budding Yeast: Anaphase A But No Metaphase Plate

See allHide authors and affiliations

Science  25 Jul 1997:
Vol. 277, Issue 5325, pp. 574-578
DOI: 10.1126/science.277.5325.574


Chromosome movements and spindle dynamics were visualized in living cells of the budding yeast Saccharomyces cerevisiae. Individual chromosomal loci were detected by expression of a protein fusion between green fluorescent protein (GFP) and the Lac repressor, which bound to an array of Lac operator binding sites integrated into the chromosome. Spindle microtubules were detected by expression of a protein fusion between GFP and Tub1, the major alpha tubulin. Spindle elongation and chromosome separation exhibited biphasic kinetics, and centromeres separated before telomeres. Budding yeast did not exhibit a conventional metaphase chromosome alignment but did show anaphase A, movement of the chromosomes to the poles.

The reproduction of eukaryotic cells depends on the ability of the mitotic spindle to segregate the replicated chromosomes into two identical sets. As cells enter mitosis they organize microtubules into a bipolar spindle. In most eukaryotes the kinetochore, a specialized region of the chromosome, binds microtubules, and in higher eukaryotes the condensed chromosomes move to a position equidistant from the spindle poles called the metaphase plate. Once all sister chromatids are properly aligned in metaphase, the linkage between the sister chromatids is dissolved, causing two changes that segregate the sisters from each other: Chromosomes move toward the spindle pole (anaphase A), and elongation of the spindle separates the spindle poles (anaphase B).

Knowledge of mitosis comes from three sources: microscopic observation and manipulation of animal and plant cells (1), studies in cell cycle extracts (2), and genetic analysis, especially in budding and fission yeast (3). To integrate the findings of these different approaches, we need to know if the basic features of mitosis are conserved among eukaryotes. Although the budding yeastSaccharomyces cerevisiae is an excellent organism for genetic analysis, its mitotic spindle is small and difficult to see in living cells. Individual mitotic yeast chromosomes are not visible by conventional light or electron microscopy and can only be seen by in situ hybridization to fixed cells (4, 5). We have simultaneously visualized the mitotic spindle and individual chromosomes in living yeast cells. These studies show that (i) budding yeast chromosomes do not align on a metaphase plate, (ii) spindle elongation and chromosome separation exhibit biphasic kinetics, (iii) yeast exhibit anaphase A chromosome-to-pole movement, (iv) chromosomes are under poleward force as they separate, and (v) centromeres separate before telomeres.

We used green fluorescent protein (GFP) (6) to follow chromosome and spindle movements in living yeast cells by fluorescence microscopy. We visualized the spindle with a protein fusion between GFP and the major alpha tubulin (Tub1) (7). Specific loci were marked by binding of a protein fusion between GFP and the Lac repressor (GFP-LacI) to an integrated array of the Lac operator (LacO) (8,9). We observed synchronized cells passing through mitosis, taking vertical stacks of images every 26 s and subjecting them to iterative deconvolution (10, 11). The top panel of Fig.1A shows a series of optical sections after deconvolution projected as a stereo pair to reveal the structure of the entire spindle and the staining of the centromeres in three dimensions, and the bottom panel shows a three-dimensional (3D) model of the same cells. Although the chromosomal LacO array and microtubules are both marked with GFP, the two structures are easily distinguished: The LacO array is a spherically symmetrical dot, whereas the spindle is a straight, rod-shaped structure that stains more intensely at its two ends. Although subsequent images are 2D projections of the 3D data, all distances are true distances between objects in 3D space.

Figure 1

Visualization of microtubules and chromosomes during mitosis. Images are 2D projections of 3D data sets. Spindle microtubule staining is the bright bar joining the chromosomal staining. (A) (Top) Stereo projection of yeast cells during anaphase. Bar, 2 μm. (Bottom) A 3D model of the top panel constructed by tracing the fluorescence intensity in the 3D data. (B) (Top) Images from a single cell passing through mitosis (arrowheads mark chromosomal staining). (Bottom) Cartoon of centromeres, spindle, and cell outline. Bar, 2 μm. (a) Early mitosis. A short, bipolar spindle has formed. (b) Onset of anaphase with sister chromatids at one spindle pole. (c) Early anaphase. Sister chromatids have separated and the spindle is elongating. (d) Late anaphase. Chromosomes are distantly separated and spindle has elongated into the daughter cell (10).

Figure 1B shows mitosis in a living yeast cell. Even in the shortest spindles, the GFP-tagged centromeres were close to the spindle's long axis (Fig. 1B). The centromeric staining was not restricted to the central region of the spindle and was often found near one of the spindle poles. Once anaphase began, sister chromatids moved rapidly toward opposite spindle poles along the spindle axis. Finally, spindle elongation distributed one sister chromatid to the mother cell and its partner to the daughter cell.

In animal and plant cells, once the two kinetochores of a pair of sister chromatids have attached to opposite spindle poles, the chromatid pair moves to a point equidistant from the two poles (12, 13). This congression is thought to be achieved by a balance between opposing forces on chromosomes: a poleward force applied at the kinetochore, a polar ejection force or the “polar wind” acting along the length of the chromosome, and forces acting on the attached sister kinetochore (12). We examined yeast cells to see if congression occurred in these very small spindles, which appear to lack the nonkinetochore microtubules that have been postulated to be responsible for the polar wind (14). We observed multiple cells (6 out of 14) in which the centromeric staining moved back and forth along the long axis of the spindle (Fig.2, A and B). The oscillations spanned the length of the spindle (1.5 to 2.0 μm) and continued until anaphase began (t = 0 s). Animal and plant chromosomes also oscillate along the spindle axis, and the absolute magnitude of the oscillations is similar to that seen in yeast. Unlike yeast spindles, however, the relative amplitude of these movements is much less than the length of the spindle so that the oscillations do not bring chromosomes near the poles. Centromere oscillation along the spindle axis implies that either kinetochores detach from and reattach to spindle microtubules or that microtubules lengthen and shorten while attached to the kinetochore, as occurs in higher eukaryotes (15). Detachment and reattachment of kinetochores to microtubules seems unlikely because yeast appear to have a single microtubule per kinetochore (16, 17) and unattached kinetochores would activate the spindle assembly checkpoint (18). Thus, chromosome oscillations suggest that an individual microtubule can polymerize and depolymerize while maintaining attachment to a yeast kinetochore.

Figure 2

Chromosomes oscillating along the spindle axis. (A) Images from a single sequence showing centromere oscillations. Arrowhead indicates the centromere and time is in seconds relative to centromere separation. Bar, 1 μm. (B) Distances during metaphase oscillation. Chromosome 1–chromosome 2, distance between the two centromeres; chromosome 1–SPB1, distance between the chromosome pair and one spindle pole body; chromosome 2–SPB2, distance between the chromosome pair and the other spindle pole body.

Yeast cells often initiate anaphase when the centromeres are closer to one pole than the other. Figure 2A shows a chromosomal spot abutting one spindle pole immediately before anaphase chromosome separation. Ten out of 17 cells (59%) initiated anaphase when the centromere was within the terminal 25% of the spindle, and 7 (41%) did so when the centromere was between 25 and 75% of the distance along the spindle. This observation suggests that budding yeast lack a conventional metaphase plate and that chromosomal alignment on a metaphase plate is not required for anaphase chromosome separation in S. cerevisiae. This result is consistent with fixed cell studies that show that chromosomes are distributed along the length of the metaphase spindle and that presumptive kinetochore microtubules do not exhibit a metaphase arrangement (4, 17, 19, 20). Although the chromosomes often separated when the sister chromatid pair was near one of the spindle poles, both sister chromatids never moved to the same pole. Thus, the sister kinetochores on all the observed chromosomes were attached to opposite spindle poles before the onset of anaphase, satisfying the cell cycle checkpoint that senses kinetochore attachment to the spindle microtubules (18, 21).

Successive images of anaphase in a single cell are shown in Fig.3A. We measured all the pairwise distances between the spindle poles and the kinetochores of the sister chromatids as cells proceeded through mitosis and aligned the records from various cells by defining t = 0 as the moment of sister chromatid separation. Figure 3B shows the average distance between the spindle poles and the distance between the sister centromeres over time. The kinetics of anaphase were highly reproducible from cell to cell. The bipolar mitotic spindle was between 1.5 and 2.0 μm in length for many minutes before the initiation of anaphase. Two minutes before sister chromatid separation, the rate of spindle elongation gradually increased to 0.54 μm/min, and the spindle was between 2.5 and 3.0 μm long when chromosome separation began (Fig. 3B). In contrast to this smooth increase, the separation of the centromeres showed a jump of 1.8 μm (average) in less than 26 s, followed by a smooth increase in separation at a rate (0.64 μm/min) similar to that of the spindle poles. The initial jump in sister separation accomplished most of anaphase A, reducing the average centromere-to-pole distance from 1.5 to 0.7 μm in less than a minute. The abrupt separation of the sister chromatids suggests that either their linkage to the poles is under poleward force before sister separation or that there is some repulsive force between the chromosomes that propels them apart once the linkage between them is dissolved. The rapid phase of elongation and separation continued until the spindle length increased to 5.0 to 6.0 μm (Fig. 3B). At this point both spindle elongation and sister separation slowed to about 0.2 μm/min, reaching a final extent of 10 to 11 μm (Fig. 3B). However, the chromosomes continued to separate with the spindle, indicating that microtubule attachment of the chromosome to the spindle pole was still intact. These results are consistent with rates of spindle elongation measured by differential interference contrast microscopy in another wild-type strain (22). In most of the cells the rapid phases of both sister separation and spindle elongation occurred through the bud neck. In 2 out of 17 mitoses, however, the initial phase of separation occurred entirely within the mother cell and was followed by reorientation of the spindle and its slow elongation through the bud neck.

Figure 3

Anaphase spindle elongation and chromosome separation. (A) Anaphase video sequence. Arrowheads mark centromeres, time is in seconds relative to sister chromatid separation, cartoon shows cell outline. Bar, 2 μm. (B) Kinetics of anaphase. Distances were measured between spindle pole bodies and chromosomal spots (25) in six data sets aligned by defining the time of chromosome separation ast = 0 (arrow on the pole separation trace). Arrowhead, time when chromosomes reached the same separation as the spindle poles. Rates of movement were derived by linear regression, and all correlation coefficients exceed 0.99. (C) Observation of anaphase A chromosome movement. Chromosome-to-pole distance was calculated for cells that exhibit biased anaphase initiation (n = 3).

The unusual property that chromosomes can separate when they are much nearer one spindle pole than the other allowed us to investigate anaphase A in yeast. Figure 3A shows images from a cell that started anaphase when chromosomes were close to one spindle pole body. Att = 0 the chromatid pair was still joined but was very close to one spindle pole body (Fig. 3A). At the next time point (t = 26 s) the chromatid pair had clearly separated, and one chromatid had moved toward the more distant spindle pole. After this chromosome-to-pole movement, the spindle elongated and separated the sister chromatids into the mother and daughter cells (Fig. 3A). Figure 3C shows the distance between the sister chromatids, the distance between the chromatid that began anaphase near its spindle pole and its spindle pole, and the distance between the sister chromatid that began anaphase far from its spindle pole and its spindle pole, with respect to time. Once sister chromatid separation began, the chromatid that was far away from its spindle pole moved rapidly toward that pole at a rate of at least 1.33 μm/min. The chromatid that was near its spindle pole moved little if at all.

Our analysis of chromosome movement to the spindle pole indicates that, in yeast, proximity of the sister chromatid pair to the spindle pole is not the activating signal for the mitotic checkpoint (23). Loss of tension on a kinetochore activates the spindle checkpoint in insect spermatocytes (24). Our observation that chromatid pairs close to one spindle pole appear to satisfy the spindle assembly checkpoint suggests that either there is sufficient tension on both kinetochores to prevent the activation of the checkpoint, or the yeast checkpoint monitors a parameter other than tension. Yeast are responsive to excess kinetochores (21), and in higher eukaryotic cells unattached kinetochores activate the spindle assembly checkpoint (18), but it is still unclear how yeast cells sense defects at the kinetochore.

To investigate the role of microtubule-dependent forces in chromosome separation, we observed mitosis in cells that lacked a spindle but still progressed through mitosis. The spindle assembly checkpoint mutant mad1 does not sense spindle depolymerization and proceeds through mitosis even though it cannot accurately segregate its sister chromatids. We constructed a mad1 yeast strain that contained the GFP-LacI tag at the centromere of chromosome III and followed mitosis in the presence of nocodazole (15 μg/ml, a concentration that depolymerizes all microtubules). The first detectable instance of sister centromere separation occurred att = 26 s (Fig. 4A), but in the absence of a spindle there was no efficient segregation of the chromosomes. Instead, the GFP-marked LacO arrays repeatedly separated into two separate dots and then approached each other so closely that they could no longer be resolved and appeared as a single dot. Sister chromatid separation in mad1 mutants in the presence of nocodazole was compared to that in wild-type cells in the absence of nocodazole (Fig. 4B). In mad1 cells, sister chromatids never separated more than 1.5 to 2.0 μm from each other, whereas chromatids in wild-type cells separated 2 to 2.5 μm within 52 s after disjunction (Fig. 4B), demonstrating that chromosome segregation requires the microtubule-dependent forces exerted by an intact spindle.

Figure 4

Sister chromatid separation without a spindle. (A) Successive images of the mad1mutant strain AFS386 (8) in nocodazole. Centromeres are marked by arrowheads, time is in seconds relative to onset of anaphase. Bar, 2 μm. (B) Distance of chromosome separation in the presence and absence of nocodazole. WT −Noc, chromatid separation in the presence of a spindle; mad1 +Noc, distances from four separate video sequences in the presence of nocodazole.

During anaphase in higher eukaryotes, chromosomes move toward the poles with their centromeres leading and their telomeres lagging behind. To visualize the centromere and telomere of the same chromosome in yeast, we integrated the LacO array both at the centromere and near the telomere of the long arm of chromosome IV, 1000 kb apart. A fortuitous recombination event reduced the size of the telomeric LacO array, allowing us to differentiate the weaker fluorescence of the telomeric array from the centromeric one. Before separation, the centromere and telomere stained as two discernible spots (Fig.5A), and in all cells the centromeres separated 1 to 3 min before the telomeres. The kinetics of centromere and telomere separation were measured as an average of five video sequences (Fig. 5B). The centromeres of chromosome IV separated with kinetics nearly identical to those of the centromeres of chromosome III (Fig. 5B). The telomeres lagged in the kinetics of their separation. The distance between centromere 2 and telomere 2 increased then decreased, suggesting that the chromatin strand was stretched as the centromere was dragged into the daughter cell (Fig. 5B) and that this stretch was relieved as the telomere approached the centromere. Although the centromeres separated smoothly, the rate of telomere separation varied during anaphase. This difference may result from the active pulling on the centromere region by the spindle as opposed to the elastic stretching and recoiling of the chromatin that links the centromere and telomere.

Figure 5

Centromere and telomere separation. (A) Successive images showing the separation of the centromeres and telomeres of chromosome IV in yeast strain AFS412 (26). Arrowheads, centromeres; barbed arrowheads, telomeres. Bar, 2 μm. Time is in seconds relative to centromere separation. (B) Distance between centromeres and telomeres. Shown is the average of five video sequences of centromere and telomere separation. The centromere and telomere that remained in the mother cell are labeled centromere 1 and telomere 1, and the centromere and telomere that were distributed to the daughter cell are labeled centromere 2 and telomere 2.

The relative importance of various mechanisms of chromosome movement in mitosis has not been established and may vary between cell types. The ability to see individual chromosome movements in living yeast should allow detailed analysis of the phenotypes of mutations in microtubule motors and components of the mitotic apparatus and contribute to understanding of the dynamics and mechanism of mitosis.


View Abstract

Navigate This Article