INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The dynamic nature of the microtubule cytoskeleton is essential for its function in a variety of cellular processes such as organelle organization and chromosome segregation. Interestingly, microtubules in vivo are much more dynamic than microtubules assembled from pure tubulin. To account for the dynamics of microtubules in cells, much effort has been directed toward the identification and characterization of proteins that regulate microtubule dynamics (Desai and Mitchison, 1997; Cassimeris, 1999). Although an increasing number of proteins that influence microtubule dynamics in vitro have been identified, their relative importance and precise roles in vivo are not yet well understood.

Individual microtubules exhibit stochastic transitions between states of growth and shrinkage, a behavior known as dynamic instability (Desai and Mitchison, 1997). Transitions from growth to shrinkage are called catastrophes; transitions from shrinkage to growth are rescues. Thus, microtubule behavior can be defined by four parameters: the growth rate, the shrinkage rate, the catastrophe frequency, and the rescue frequency. Additionally, microtubules sometimes enter a paused state, in which they neither grow nor shrink. Proteins that regulate microtubule dynamics have been classified as either microtubule-stabilizing or microtubule-destabilizing factors on the basis of their net effect on microtubule polymerization. Maintaining the proper balance among these regulatory factors must be critical for controlling the dynamics of microtubules in the cell.

Previously, we described the identification of Stu2p, an essential microtubule-binding protein in yeast (Wang and Huffaker, 1997). Stu2p is a member of the protein family that includes Schizosaccharomyces pombe Dis1 (Nabeshima et al., 1995), human TOGp (Charrasse et al., 1998), Caenorhabditis elegans ZYG-9 (Matthews et al., 1998), Drosophila Msps (Cullen et al., 1999), and Xenopus XMAP215 (Tournebize et al., 2000). All of these proteins localize to the microtubule-organizing center and spindle during mitosis. dis1, zyg-9, and msps mutations cause improper spindle formation and chromosome missegregation.

XMAP215 has been shown to have a direct effect on microtubule dynamics. XMAP215 promotes the polymerization of pure tubulin in vitro by increasing the growth rate without reducing catastrophes or stimulating rescues (Vasquez et al., 1994). However, in Xenopus egg extracts, the primary function of XMAP215 is to suppress catastrophes. As XMAP215 has no intrinsic ability to suppress catastrophes, it has been proposed that the role in XMAP215 in extracts is to compete for access to microtubules with the microtubule-destabilizing protein XKCM1 (Tournebize et al., 2000). These results highlight the commonly held belief that microtubule dynamics in cells are likely to be determined by the complex interplay of regulatory factors. Unfortunately, such complex regulation of microtubule dynamics has been difficult to study in intact cells.

In this study we present the first in vivo analysis of microtubule dynamics in cells lacking a member of this protein family. With the use of stu2 mutant yeast cells expressing tubulin fused to green fluorescent protein (GFP; Carminati and Stearns, 1997; Shaw et al., 1997b; Tirnauer et al., 1999; Adames and Cooper, 2000), we show that Stu2p promotes the dynamics of cytoplasmic and spindle microtubules in Saccharomyces cerevisiae, contrary to the role of XMAP215 in Xenopus extracts. In addition, we show that the promotion of microtubule dynamics by Stu2p is required for proper spindle orientation and metaphase chromosome alignment.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Yeast Strains, Media, and Plasmids

The yeast strains used are listed in Table 1. YPD and SD media were prepared as described (Sherman, 1991). To deplete Stu2p in stu2^cu strains, cupric sulfate (CuSO₄) was added to SD media to a final concentration of 500 µM. Cells were treated with 3 µg/ml -factor or 100 mM hydroxyurea for 2 h to arrest them in G1 or S phase, respectively.

The plasmid encoding GFP-Tub1p under control of the TUB1 promoter, pAFS125 (Straight et al., 1997), was a gift from Dr. Andrew Murray (Harvard University, Cambridge, MA). The KpnI to NotI fragment of pAFS125 containing GFP-TUB1 was ligated into KpnI- and NotI-digested pRS404 to create pKAK101. pKAK101 was linearized with NsiI and transformed into CUY1147 and CUY1148 to create CUY1243 and CUY1244, respectively. The plasmid encoding Cse4-GFP, pKK1 (Chen et al., 2000), was a gift from Dr. Richard Baker (University of Massachusetts Medical School, Worcester).

GFP Tagging of Stu2p

pYGFP, a gift from Brendan Cormack (Johns Hopkins School of Medicine, Baltimore), contains a version of GFP with codons optimized for usage in yeast (Cormack et al., 1997). The HIS5 gene from S. pombe was cloned into the BamHI and EcoRI sites of pYGFP to create pHY181. The yGFP::HIS5 sequence was amplified from pHY181 by PCR with the use of the forward primer 5'-GAATTG-AAAAAATGAAGGCCAAATCAAGACGGGAAGGGACAACCA -GGACGATGTCTAAAGGTGAAGAAT-3' and the reverse primer 5'-TCAAGTTGAAGACTATATATTTTATTGAGTTTATGTTATGGGGAGGCTACCTGGATGGCGGCGTTAGTAT. The resulting DNA fragment contained 50 bp of sequence immediately upstream of the STU2 stop codon followed by yGFP, HIS5, and then 50 bp of STU2 sequence beginning 3 bp downstream of the stop codon. This fragment was transformed into the diploid yeast strain CUY546 to fuse GFP to the C-terminal end of the chromosomal STU2 gene. One His⁺ transformant was sporulated and dissected to generate CUY1247.

Stu2p Depletion Strain

We created a yeast strain in which, upon addition of copper, there is a simultaneous repression of STU2 mRNA synthesis and degradation of Stu2p (Moqtaderi et al., 1996). Briefly, a 404-bp fragment of the STU2 gene (beginning with ATG) was subcloned into the SpeI site of the ZM168 polylinker producing pKAK59. pKAK59 was digested with KpnI and XbaI, and the resulting P_ANB1-Ub-R-STU2 fragment was ligated into the same sites of pRS406 (Sikorski and Hieter, 1989). The resulting plasmid, pKAK60 was linearized with HindIII, transformed into ZMY60, and plated onto selective media. Integrating linearized pRS406 into ZMY60 created the control strain (CUY1148). The Ura⁺ transformants were then tested for copper sensitivity by comparing their growth with CUY1148 on SD plates containing CuSO₄ at 250 or 500 µM. One Ura⁺ transformant was selected (CUY1147) that was inviable on SD plates containing 500 µm CuSO₄, a concentration that allowed growth of CUY1148.

stu2 Temperature-Sensitive Mutants

STU2 was amplified by PCR with primers 5'-GGGACCAAATAGCATTAC-3' and 5'-GTGCAGTGTGCTTATCTC-3' under mutagenic conditions with the use of 7 mM MgCl₂ (Muhlrad et al., 1992). PCR products were digested with BamHI and SphI and ligated into the larger BamHI-SphI fragment of pS2 (stu2-1 LEU2 CEN6 ARSH4; Wang and Huffaker, 1997). To screen for conditional-lethal mutants, the library was transformed into the haploid strain CUY1087 whose viability is dependent on the expression of the plasmid-borne STU2 from the GAL1 promoter. Leu⁺ transformants were screened for conditional lethality on YPD plates at various temperatures.

Stu2p Antibody and Western Blots

The 1.16-kb KpnI fragment from pWP82 (Wang and Huffaker, 1997) was cloned into the KpnI site of pQE-31 (QIAGEN Inc., Chatsworth, CA) to make pWP86; this construct encodes a polypeptide containing the 385 carboxy-terminal amino acids of Stu2p fused to the His₆ tag. The 44-kDa fusion protein was expressed in Escherichia coli strain M15 and purified on Ni²⁺-NTA resin under denaturing conditions. Rabbit antiserum to this polypeptide was produced by the Center for Research Animal Resources at Cornell University (Ithaca, NY).

Total yeast extracts were made by grinding cells that had been rapidly frozen in liquid nitrogen (Sorger et al., 1994). Forty micrograms of total protein was resolved by SDS-PAGE and then transferred to nitrocellulose. The rabbit polyclonal anti-Stu2p antibody was used at a 1:5000 dilution. The secondary antibody, HRP-conjugated donkey anti-rabbit (Amersham Pharmacia Biotechnology, Piscataway, NJ) was used at a 1:10,000 dilution and was visualized with the use of an enhanced chemiluminescence system (Amersham Pharmacia Biotechnology).

Fluorescence Microscopy and Image Analysis

Indirect immunofluorescence was performed as described previously (Pasqualone and Huffaker, 1994). Rat monoclonal anti-yeast -tubulin antibody, YOL1/34 (Kilmartin et al., 1982), was provided by J. Kilmartin (Medical Research Council, Cambridge, United Kingdom). Fluorescein goat anti-rat secondary antibody was purchased from Cappel Research Products (Durham, NC). 4',6-diamidino-2-phenylindole (DAPI) staining of DNA was done as described previously (Sullivan and Huffaker, 1992). Laser photobleaching of GFP-Tub1p in live cells and statistical analysis of FRAP was performed as described previously (Maddox et al., 2000).

Time-lapse analysis of live cells was performed on cells grown at 23°C in SD media supplemented with adenine, lysine, and CuSO₄. Slides were prepared and images captured as described by Shaw et al. (1997a). Time lapse series were acquired at 10 or 20 s intervals. The 2 × 2 binned images were acquired in Z-series either five planes 0.75 µm apart or six planes 0.5 µm apart. At each time point, we measured the length of the microtubule in the x-y plane (A); its height in the z-axis (B) was determined by counting the number of planes containing an in-focus portion of the microtubule. The total length of the microtubule at each time point was then calculated to be C, where A² + B² = C².

Linear regression analysis was used to obtain rates of microtubule growth and shrinkage. Growth and shrinkage events were defined by a line with an R² value of 0.85 and a net change in length of 0.5 µm. Pauses were defined as events that lasted 30 s in which no significant growth or shrinkage occurred. Catastrophes were defined as transitions to shrinkage after a growth or pause; rescues were defined as transitions to growth after a shrinkage or pause. The frequencies of catastrophe and rescue were calculated by dividing the total number of events by the total time that all microtubules spent pausing and growing or pausing and shrinking, respectively.

Variations around mean values are given as SDs. Comparisons of statistical significance were by unpaired t test.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Depletion of Stu2p Causes a Mitotic Arrest

An stu2 conditional depletion strain was constructed based on the methods first described by Moqtaderi et al. (Moqtaderi et al., 1996). Addition of copper ions to the medium causes simultaneous repression of STU2 mRNA synthesis and degradation of Stu2p; hence, we refer to this strain as stu2^cu. In this strain, ROX1 and UBR1 are under control of a tightly regulated copper-inducible promoter. UBR1 encodes the N-end recognition component of the ubiquitin degradation pathway. Rox1p is a transcriptional repressor of the ANB1 gene. The STU2 gene is replaced by a derivative that contains an N-end rule recognition signal for rapid ubiquitin-dependent degradation and is driven by the ANB1 promoter. In the absence of copper, neither the repressor nor the protein degradation system is active, and Stu2p is expressed. Addition of copper sulfate activates both systems and causes a decrease in Stu2p levels. Stu2p is undetectable in stu2^cu cells by 2 h after addition of copper (Figure 1A). Cells lacking Stu2p arrest in mitosis. After 6 h of growth at 23°C in copper-containing medium, most of the cells were large budded (74%) and contained replicated but unsegregated chromosomes (Figure 1, B and C). Immunofluorescence microscopy with the use of anti--tubulin antibody showed that most of these cells (>80%) possessed preanaphase spindles (Figure 1B).

View larger version (34K): [in this window] [in a new window]   Figure 1.   Phenotype of Stu2p-depleted cells. (A) Time course of Stu2p depletion. STU2 (CUY1148) and stu2cu (CUY1147) cells were diluted into copper-containing media for the time (h) indicated above each lane. Equal amounts of protein extract were subjected to SDS-PAGE followed by immunoblotting with the use of anti-Stu2p antibody. Note that in the construction of the stu2cu strain, an HA epitope tag is fused to the N-terminus of Stu2p. (B) stu2cu (CUY1147) cells after 4 h copper-containing medium. Left, DIC; middle, DAPI staining of DNA; right, immunofluorescence with the use of Tub2p antibody. Bar, 4 µm. (C) DNA content of stu2cu cells. CUY1243 cells were grown asynchronously at 23°C and shifted into copper-containing medium for the times indicated. The DNA content of individual cells was determined by flow cytometry (Haase and Lew, 1997).

We analyzed spindles in live STU2 and stu2^cu cells expressing GFP-Tub1p that were grown in copper-containing medium for at least 4 h (Figure 2). The average length of the preanaphase spindles in STU2 cells (Figure 2A) was 1.25 ± 0.28 µm (n = 43). The average length of preanaphase spindles in large-budded stu2^cu cells (Figure 2, B and C) was 1.08 ± 0.28 µm (n = 52), a small but statistically significant difference in spindle length (p = 0.004 by unpaired t test). In addition, 15% of stu2^cu large-budded cells contained partially elongated spindles up to 5.0 µm in length. However, none possessed fully elongated anaphase spindles, typically 8-10 µm in length in STU2 cells. In contrast to the GFP fluorescence of STU2 spindles (Figure 2D), which is relatively uniform along their lengths, the fluorescence of 3-5 µm spindles in stu2^cu cells was brightest at the poles and gradually diminished away from the poles (Figure 2, E and F). These stu2^cu spindles would occasionally bend and eventually break despite failing to elongate more than half the length of a normal anaphase spindle (Figure 2G).

View larger version (65K): [in this window] [in a new window]   Figure 2.   Analysis of spindles in stu2cu cells expressing GFP-Tub1p. (A-F) DIC (left) and GFP-Tub1p fluorescence of microtubules (right) in STU2 (CUY1244) and stu2cu cells (CUY1243). (A) Preanaphase spindle in STU2 cell; (B and C) similar short spindles in stu2cu cells. (D) Elongated ~3-µm spindle in STU2 cell; (E and F) ~3-µm spindles in stu2cu cells. (G) Time-lapse images taken at 1-min intervals of a stu2cu cell showing a spindle breaking down before reaching full length. Spindle breakage occurs in the panel labeled with the asterisk. Bars, 2 µm.

Cells Lacking Stu2p Contain Fewer Cytoplasmic Microtubules

We counted the number of cytoplasmic microtubules in live STU2 and stu2^cu cells expressing GFP-Tub1p that were grown in copper-containing medium for at least 4 h. stu2^cu cells contained on average just over half the number of cytoplasmic microtubules as STU2 cells (Figure 3A). In unbudded G1 cells that contain a single SPB, STU2 and stu2^cu cells had an average of 2.6 ± 0.8 and 1.5 ± 1.2 cytoplasmic microtubules per cell, respectively (n = 63, 44; p < 0.001). In cells that contained preanaphase spindles and two SPBs, the STU2 and stu2^cu cells had an average of 1.9 ± 0.7 and 1.1 ± 1.0 microtubules per cell, respectively (n = 44, 55; p < 0.001). Significantly, 18% of unbudded stu2^cu cells and 31% of stu2^cu cells with preanaphase spindles contained no cytoplasmic microtubules. In contrast, every STU2 cell we examined contained at least one cytoplasmic microtubule.

View larger version (30K): [in this window] [in a new window]   Figure 3.   Cytoplasmic microtubule number and spindle orientation. (A) Asynchronous STU2 (CUY1244) and stu2cu (CUY1243) cells were shifted into copper-containing media for 4 h. Cytoplasmic microtubules were visualized by GFP-Tub1p fluorescence. (B) Asynchronous stu2cu (CUY1243) cells and hydroxyurea-arrested STU2 (CUY1244) cells were shifted into copper-containing media for 4 h. GFP-Tub1p was visualized in live cells, and images were collected of budded cells containing short spindles. The spindle angle was calculated as the angle between a straight line drawn through the mother-bud neck axis and a second line drawn through the spindle. Spindle distance is the distance between the bud-proximal end of the spindle and the bud neck.

We did not find a statistically significant difference between cytoplasmic microtubule lengths in STU2 and stu2^cu cells. The average length of cytoplasmic microtubules in unbudded cells was 1.3 ± 0.6 µm for STU2 cells and 1.4 ± 0.8 µm for stu2^cu cells (n = 73, 61). For cells that contained spindles, the average cytoplasmic microtubule length was 1.3 ± 0.7 µm for STU2 cells and 1.4 ± 1.1 µm for stu2^cu cells (n = 62, 61).

Cytoplasmic Microtubules Are Less Dynamic in Stu2p-depleted Cells

Cytoplasmic microtubule dynamics were measured in live STU2 and stu2^cu cells expressing GFP-Tub1p after growth on copper-containing medium for at least 3 h (Figure 4A). We analyzed both G1 cells that lacked spindles and preanaphase cells that contained short bipolar spindles. Because the stu2^cu population lacked cells with long anaphase spindles, we did not analyze anaphase cells in the STU2 population. Representative microtubule life history plots are shown in Figure 4B, and the data described in this section are listed in Table 2.

View larger version (29K): [in this window] [in a new window]   Figure 4.   Cytoplasmic microtubule dynamics in STU2 (CUY1244) and stu2cu (CUY1243) cells expressing GFP-Tub1p. (A) Cytoplasmic microtubules in a single STU2 G1 cell. Images show two-dimensional projections of Z-focal plane series at 20-s intervals. One microtubule (arrow) grew throughout this time course. A second (arrowhead) grew and then shrank back to the SPB. The SPB is located in the lower left corner. Bar, 1 µm. (B) Plots were constructed from measurements of cytoplasmic microtubule length (see MATERIALS AND METHODS) versus time for G1 (left panels) and preanaphase (right panels) cells.

The dynamic properties of cytoplasmic microtubules in STU2 cells were similar in G1 and preanaphase cells. There was no statistically significant difference in the growth or shrinkage rates of microtubules between G1 and preanaphase cells. Both the catastrophe and rescue frequencies were slightly higher in preanaphase cells.

Stu2p depletion did not have a dramatic effect on the growth or shrinkage rates of cytoplasmic microtubules. The only statistically significant rate difference between STU2 and stu2^cu cells was in the shrinkage rates observed in G1 (p = 0.014). This latter rate was 1.8-fold lower in stu2^cu cells (0.85 vs. 1.55 µm/min in STU2 cells). Similarly, the average duration of growth and shrinkage events in STU2 and stu2^cu cells were not significantly different, with the exception again of shrinkage events in G1 cells (p = 0.011). The average duration of these latter events was 1.8-fold greater in stu2^cu cells (1.90 min vs. 1.05 min in STU2 cells). This increase in duration compensated for the decrease in rate so that average length change of shrinkage events in G1 was nearly identical in STU2 and stu2^cu cells (1.63 vs. 1.62 µm, respectively).

In contrast with the above parameters, there was a consistent difference in the transition frequencies of cytoplasmic microtubules in STU2 and stu2^cu cells. During G1, the catastrophe frequency was 2-fold lower in stu2^cu cells (0.0035/s vs. 0.0071/s in STU2 cells) and the rescue frequency 2.1-fold lower (0.0030/s vs. 0.0063/s in STU2 cells). In preanaphase stu2^cu cells, the catastrophe frequency was 1.6-fold lower (0.0054 vs. 0.0085 in STU2 cells) and the rescue frequency 2.4-fold lower (0.0038 vs. 0.0092 in STU2 cells). Microtubules in stu2^cu cells also exhibited a marked increase in the amount of time spent in the paused state. In G1, microtubules in stu2^cu cells spent 36% of the time pausing compared with 19% for microtubules in STU2 cells. During preanaphase, microtubules in stu2^cu cells spent 40% of the time pausing compared with 12% for microtubules in STU2 cells.

Dynamicity is a measure of the mean rate of tubulin exchange on microtubules and is equivalent to the number of tubulin dimers gained or lost per second (Toso et al., 1993). Differences in dynamicity between microtubules in STU2 and stu2^cu cells reflects differences in their growth and shrinkage rates as well as differences in the fraction of time spent growing and shrinking. During G1, the dynamicity of cytoplasmic microtubules in stu2^cu cells was 10 dimers/s, 1.8-fold lower than the STU2 rate of 18 dimers/s. In preanaphase cells, the dynamicity of cytoplasmic microtubules in stu2^cu cells was 16 dimers/s, 1.5-fold lower than the STU2 rate of 24 dimers/s.

In summary, Stu2p depletion produces microtubules that are significantly less dynamic, with less frequent catastrophes and rescues, and more time spent in pauses. Taken together, these results indicate that Stu2p promotes dynamics of cytoplasmic microtubules in vivo.

Stu2p Depletion Causes a Spindle Orientation Defect

In yeast, cytoplasmic microtubules are required to orient the spindle in a two-step process (Stearns, 1997; Bloom, 2000). Before anaphase, the spindle moves to the bud neck and becomes oriented along the mother-bud axis, a process involving actin cables, Myo2p, Kar9p, Bim1p, and Kip3. At anaphase onset, the spindle traverses the bud neck in a second process involving the dynein heavy chain protein, Dhc1p. These steps are partially redundant; either one is sufficient for cell viability, but loss of both results in lethality.

As cells lacking Stu2p contain fewer and less dynamic cytoplasmic microtubules, we asked whether these changes affect the ability of cytoplasmic microtubules to orient spindles. We grew stu2^cu cells expressing GFP-Tub1p in copper-containing medium for at least 4 h and obtained random images of spindles in live cells. As a control, STU2 cells expressing GFP-Tub1p were treated with hydroxyurea to arrest them with preanaphase spindles. We measured the distance of the closest end of the spindle to the mother-bud neck and the angle of the spindle relative to the mother-bud axis (Figure 3B). The position and angle of preanaphase spindles appeared random in stu2^cu cells. The mean spindle-to-bud neck distance was 2.2 ± 1.1 µm in stu2^cu cells vs. 1.0 ± 0.5 µm in STU2 cells (n = 98, 117; p < 0.001). The mean spindle orientation angle was 53 ± 25° in stu2^cu cells vs. 22 ± 19° in STU2 cells (n = 98, 117; p < 0.001).

The defect observed in stu2^cu cells is indicative of a block in the early pathway of spindle orientation. Genetic experiments, with the use of two temperature-sensitive alleles of STU2, stu2-10 and stu2-13, were consistent with this interpretation. stu2-10 and stu2-13 strains were crossed to strains containing either a deletion of KAR9 (kar9) or DHC1 (dhc1), and the double mutants were analyzed. Both stu2-10 kar9 and stu2-13 kar9 double mutants grew well. However, a large fraction of the stu2-10 dhc1 and stu2-13 dhc1 double mutant spores were inviable (30% and 60%, respectively) and those that were viable grew slowly. Because only cells compromised in both pathways display growth defects, these results are consistent with Stu2p playing a role in the early Kar9p-dependent pathway for spindle orientation.

Kinetochore Microtubules Are Less Dynamic in Stu2p-depleted Cells

The haploid S. cerevisiae spindle contains 16 kinetochore microtubules originating from each SPB, one for each of the 16 sets of duplicated chromosomes. Each SPB also nucleates ~4 polar microtubules that interdigitate with their counterparts from the opposite pole (Winey et al., 1995). This high density of microtubules makes it impossible to observe the dynamics of individual microtubules in the spindle. Instead, we have used fluorescence redistribution after photobleaching (FRAP) as a measure of spindle microtubule dynamics. In a previous study, Maddox et al. (2000) provided evidence that FRAP of yeast preanaphase spindles is a measure of kinetochore microtubule dynamics. Kinetochore microtubule turnover is likely due to the growth and shortening of these microtubules, coupled to oscillations of the chromosomes along the length of the spindle.

Preanaphase spindles labeled with GFP-Tub1p in the stu2^cu strain were photobleached before and after Stu2p depletion. Control cells exhibited fluorescence recovery of 64 ± 9% of the bleached region after 266 ± 7 s (n = 4; Figure 5, A and C) with a half-life of ~60 s, similar to previous experiments with the use of wild-type cells (Maddox et al., 2000). The 36% incomplete recovery in these cells is due to photobleaching of a significant fraction of the cellular tubulin. Interestingly, spindles bleached in cells lacking Stu2p recovered fluorescence to only 10 ± 7% after 234 ± 49 s (n = 9; Figure 5, B and C). Because of the low percentage of recovery, a half time for recovery could not be accurately calculated. These data indicate that spindle microtubule dynamics are substantially attenuated in the absence of Stu2p.

View larger version (62K): [in this window] [in a new window]   Figure 5.   FRAP of stu2cu spindles. (A) FRAP of a control cell (CUY1243 grown without copper). (B) FRAP of a cell lacking Stu2p (CUY1243 grown with copper). In both A and B, the 0 s time point shows the spindle before laser exposure. Time point 5 s shows the same spindle for each condition after a laser pulse. The cell was followed by 3D time-lapse imaging for a total of 5 min. The control spindle shows fluorescence recovery of the targeted region of the spindle (A, black arrows). However, the bleached region of the Stu2p-depleted spindle (B, black arrows) displays attenuated recovery. (C) Quantitation of the fluorescence recovery of a cell lacking Stu2p (dotted lines) and a control cell (solid lines). Bleached (squares) and unbleached (triangles) half spindles of a control cell. Bleached (crosses) and unbleached (stars) half spindles of a Stu2p-depleted cell.

Stu2p Depletion Inhibits Metaphase Chromosome Alignment

Kinetochore microtubule dynamics are believed to be essential for chromosome capture and alignment. Because cells lacking Stu2p contain less dynamic kinetochore microtubules, we tested whether this decrease in dynamics affects chromosome alignment. In S. cerevisiae chromatin arms remain associated until anaphase, but the centromere-proximal chromatin is stretched apart into two distinct clusters during metaphase (Goshima and Yanagida, 2000; He et al., 2000; Tanaka et al., 2000; Pearson et al., 2001). This stretched chromatin indicates the presence of poleward pulling forces on the centromeres and provides strong evidence for the existence of tension along the length of the spindle.

We determined the localization of centromeres in stu2^cu cells expressing Cse4-GFP, a histone H3-like protein that is specifically present at all centromeres (Meluh et al., 1998; Chen et al., 2000). Before depletion of Stu2p, 50% of the large-budded cells exhibited metaphase centromere conformations, characterized by two distinctly separated but clustered foci representing the separated sister centromeres (Figure 6). Forty-two percent exhibited anaphase configurations characterized by two more widely separated (>2 µm) centromere foci.

View larger version (22K): [in this window] [in a new window]   Figure 6.   Centromere localization in stu2cu cells. Population counts of Cse4-GFP fluorescence morphologies in control cells (KBY2660.1 grown without copper) and cells lacking Stu2p (KBY2660.1 cells grown with copper for 5 h) determined by acquiring single time-point optical Z-series images. A representative image of each phenotype observed is displayed in the figure, and the percentages of cells in each category are indicated. Cell counts represent the average of four separate experiments. Bar, 2 µm.

After Stu2p depletion most of the large-budded stu2^cu cells contained preanaphase spindles (see above), but only 48% contained two distinguishable centromere foci (Figure 6). Thirty-one percent contained a single focused cluster, and 21% contained a single elongated cluster, defined as a linear array of ~1 µm with signals of variable intensity along its length. These single cluster morphology classes are distinct from the normal metaphase conformation of two distinguishable foci and indicate either a lack of chromosome attachment or attachment in the absence of force to stretch the centromeres apart. The normal metaphase conformation was observed in only 29% of the stu2^cu cells. Approximately 18% of stu2^cu cells had clusters of fluorescence separated by 2-4 µm, corresponding well to the fraction of partially elongated spindles (15%) observed with the use of GFP-Tub1p fluorescence. Only 1% of the cells had foci separated by >4 µm. In summary, Stu2p-depletion causes a significant increase in metaphase chromosome misalignment.

Stu2p Localizes to the Spindles, SPBs, and Cytoplasmic Microtubules

We previously described the localization of Stu2-GFP in living cells (Wang and Huffaker, 1997). We reexamined this localization with more sensitive image analysis tools and with the use of a haploid strain in which the endogenous STU2 gene had been replaced with a single copy of STU2-GFP under control of the STU2 promoter. Consistent with our previous results, Stu2-GFP fluorescence was concentrated at the spindle pole in all cells at every stage in the cell cycle (Figure 7, A-F). With the more sensitive techniques, Stu2p fluorescence was also seen in a discontinuous manner along the lengths of fibers, presumably cytoplasmic microtubules, in 90% of the unbudded cells. Often the most distal spot was located near the cell cortex and likely represented the tip of the cytoplasmic microtubule, equivalent to the plus end (Figure 7, A and B). In cells with spindles, <50% contained any cytoplasmic staining; this usually occurred as a single dot near the SPB (Figure 7C). Stu2p was located mostly at the poles of short bipolar spindles; in a few cases an additional single dot was seen between the poles (Figure 7, C and D). As spindles elongated, we occasionally observed Stu2p along their lengths. In these cases, Stu2p localized in a discontinuous manner along the spindle or in one or two dots at the midzone, the location of the plus-ends of overlapping polar microtubules (Figure 7, E and F).

View larger version (93K): [in this window] [in a new window]   Figure 7.   Stu2-GFP localization in vivo. (A-F) Images of GFP fluorescence in living CUY1247 cells expressing Stu2-GFP from the endogenous promoter. These images are either single focal planes (A-D) or compilations created from optical Z-series images (E and F). Bar, 1 µm. (G) Immunoelectron micrograph of Stu2-GFP with the use of the Nanogold-silver intensification technique (Adams and Kilmartin, 1999). Arrowheads indicate the SPBs. Bar, 0.1 µm.

Immunoelectron microscopy was carried out with the use of cells that expressed Stu2-GFP. The GFP antigen was detected on serial sections with the use of the Nanogold-silver intensification method (Adams and Kilmartin, 1999). Stu2p was found near both faces of the SPB indicating that the bright dots observed by Stu2-GFP fluorescence are due to both nuclear and cytoplasmic Stu2p (Figure 7G). In addition, Stu2p was associated with microtubules all along the length of the spindle.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The Role of Stu2p in Cytoplasmic Microtubule Function

We have shown that Stu2p plays a prominent role in determining the assembly properties of cytoplasmic microtubules. Depletion of Stu2p led to less dynamic cytoplasmic microtubules in both G1 and preanaphase cells. The decrease in microtubule dynamics is due primarily to twofold decreases in both the catastrophe and rescue frequencies, and a two- to threefold increase in the fraction of time microtubules spend pausing. In G1 cells, the shrinkage rate decreased by nearly twofold as well. On the basis of the phenotype of cells lacking Stu2p, we conclude that Stu2p belongs to the class of microtubule-binding proteins that promote microtubule dynamics.

Although microtubule dynamics decrease in Stu2p-depleted cells, the average cytoplasmic microtubule length determined from static images is not altered. This finding is in agreement with the measured dynamic properties of these microtubules. The average microtubule length change for growth and shrinkage events is very similar in wild-type and Stu2p-depleted cells (Table 2). Thus, Stu2p promotes cytoplasmic microtubule dynamics without affecting overall microtubule length.

Stu2p is localized both to the SPBs and to the distal tips of cytoplasmic microtubules that correspond to the minus and plus ends of microtubules, respectively. There appears to be considerably more Stu2p at the SPB, which might be taken as evidence that Stu2p affects microtubule dynamics primarily at the minus ends of microtubules. However, a recent analysis of fluorescent speckle marks on cytoplasmic microtubules in yeast reveals that there is no detectable assembly or disassembly at minus ends (Maddox et al., 2000). Thus, we favor the interpretation that Stu2p localized at the plus ends of microtubules acts to promote microtubule dynamics.

Stu2p also plays a role in determining cytoplasmic microtubule number. Stu2p-depleted cells contain about half the number of cytoplasmic microtubules as wild-type cells. Cytoplasmic microtubules < 0.5 µm in length cannot be reliably detected by fluorescence microscopy. However, we have no reason to expect that Stu2p-depleted cells possess a significant fraction of very short microtubules, given that the length distributions of the microtubules that we could observe are similar in STU2 and stu2^cu cells. Therefore, it seems likely that Stu2p plays a role in controlling microtubule number that is distinct from its role in controlling microtubule dynamics. One possibility is that Stu2p influences microtubule nucleation. Microtubule assembly is initiated by the -tubulin complex that localizes to centrosomes in higher eukaryotes and to SPBs in fungi (Pereira and Schiebel, 1997; Wiese and Zheng, 1999). Localization of the S. cerevisiae -tubulin complex to the outer plaque or cytoplasmic face of the SPB is mediated by its interaction with the SPB component Spc72p (Knop and Schiebel, 1998). Spc72p also interacts with Stu2p (Chen et al., 1998), raising the intriguing prospect that Stu2p influences -tubulin complex activity through Spc72p.

Current models for spindle orientation propose a search-and-capture mechanism whereby dynamic cytoplasmic microtubules find and interact with specific cytoplasmic components. Do the decreases in cytoplasmic microtubule dynamics and number associated with Stu2p-depletion affect their cellular function? Our measurements of spindle orientation demonstrate that they do; in stu2 mutants, preanaphase spindles are positioned randomly in the mother cell. This phenotype and our genetic data suggest that Stu2p plays a role in the early pathway of spindle orientation. These results likely reflect a greater reliance of the early pathway on the cytoplasmic microtubule dynamics promoted by Stu2p. However, the recent demonstration that Stu2p interacts physically with Kar9p (Miller et al., 2000), a component of the early pathway, suggests that Stu2p may play a more direct role in spindle orientation as well.

The Role of Stu2p in Kinetochore Microtubule Function

The yeast preanaphase spindle contains ~40 microtubules, making it impossible to observe the dynamics of any single spindle microtubule. Instead, we have used FRAP to measure the dynamics of kinetochore microtubules (Maddox et al., 2000). We found that FRAP is substantially reduced in Stu2p-depleted cells. As we did not observe any differences in morphology or fluorescence intensity between preanaphase spindles in wild-type and Stu2p-depleted cells, we do not believe this result is due to a significant decrease in kinetochore microtubule number in the Stu2p-depleted cells. Instead, the lack of FRAP in the absence of Stu2p likely indicates a dramatic decrease in kinetochore microtubule turnover. Extrapolating from studies demonstrating that astral and polar microtubule minus ends do not grow or shorten (Maddox et al., 2000), we conclude that Stu2p plays a key role in promoting plus-end dynamics of kinetochore microtubules.

Dynamic microtubules are thought to be essential for the search-and-capture mechanism proposed for the attachment of microtubules to kinetochores. In addition, dynamic microtubules are likely to be required to generate tension across the spindle. To investigate whether the loss of kinetochore microtubule dynamics affects chromosome alignment, we localized the centromere protein, Cse4-GFP, in Stu2p-depleted cells. Over half of the large-budded Stu2p-depleted cells contain only a single cluster of Cse4p staining, indicating that the sister centromeres are not separated in these cells. Single clusters of Cse4p are rarely seen in wild-type cells and could reflect deficiency in either bipolar attachment or microtubule-mediated tension in the absence of Stu2p. Forty percent of the single clusters in Stu2p-depleted cells are elongated, a pattern that appears to be intermediate between the single focus and the metaphase arrangement of two foci. Therefore, elongated clusters may represent prometaphase arrangements of centromeres in yeast. Only 30% of the Stu2p-depleted cells contain two foci of Cse4p typical of metaphase. Given the high percentage of Stu2p-depleted cells that show a cell cycle arrest, it seems likely that even some of these cells with two Cse4p foci contain defective spindles that continue to activate the spindle assembly checkpoint. A single unattached kinetochore is sufficient to activate the checkpoint in vertebrates (Rieder et al., 1995) but would be difficult to detect as aberrant Cse4p localization. In summary, our data provide in vivo evidence that kinetochore microtubule dynamics are essential for efficient kinetochore attachment and/or spindle tension.

The Role of Stu2p in Polar Microtubule Function

It has recently been reported that a temperature-sensitive allele of STU2 blocks anaphase B spindle elongation (Severin et al., 2001). Elimination of the spindle checkpoint does allow cells to progress through the cell cycle but does not abrogate the spindle elongation defect, indicating that Stu2p plays a structural role in spindle elongation. Stu2p-depleted cells also fail to elongate spindles. A few Stu2p-depleted cells begin spindle elongation, but these spindles break at ~3-5 µm in length. This breakage could be due to a weak arrangement of polar microtubules or to a defect in elongating these microtubules once anaphase begins. We have not been able to measure the effect of Stu2p depletion on the dynamics of polar microtubules. But given that Stu2p promotes the dynamics of cytoplasmic and kinetochore microtubules, it seems reasonable to propose that Stu2p may promote the dynamics of polar microtubules as well. Consistent with this hypothesis, Stu2p localizes to the midregion of spindles, where the plus-ends of the polar microtubules reside. As polar microtubules must find their counterparts from the opposite pole, decreasing microtubule dynamics could impede this process and lead to the assembly of defective spindles that are unable to elongate. Alternatively, Stu2p may be involved in the switch that allows polar microtubules to lengthen at anaphase B onset. Although Stu2p depletion does not alter the length of cytoplasmic microtubules, it may have a different effect on polar microtubules. The nucleus and the cytoplasm likely contain distinct sets of proteins capable of influencing microtubule dynamics, so the relative contribution of Stu2p to microtubule assembly may differ in these two compartments.

Stu2p and XMAP215

Contrary to the role of XMAP215 that dampens microtubule dynamics in Xenopus extracts, Stu2p promotes microtubule dynamics in yeast. These opposite effects may reflect differences in the intrinsic activities of these two proteins, a property that has not yet been determined for Stu2p. However, it seems likely that this discrepancy results, at least in part, from differences in growing yeast cells and Xenopus egg extracts. Microtubules in yeast are considerably less dynamic than microtubules in Xenopus extracts and animal cells in general (Carminati and Stearns, 1997; Shaw et al., 1997b). This is probably due in part to the fact that tubulin from yeast is inherently less dynamic than tubulin from animal cells (Davis et al., 1993) and in part to variations in the arrays of proteins that regulate microtubule dynamics in these two systems. Thus, the differences in the roles of Stu2p and XMAP215 may reflect differences in their intrinsic properties or in the environment in which they operate or both.