![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
Vol. 16, Issue 3, 1378-1395, March 2005
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Molecular Genetics Research Laboratory, Graduate School of Science, University of Tokyo, Hongo, Bunkyo-ku, Tokyo 113-0033, Japan;
Department of Biophysics and Biochemistry, Graduate School of Science, University of Tokyo, Hongo, Bunkyo-ku, Tokyo 113-0033, Japan; and
Laboratory of Cell Regulation, Cancer Research UK, London Research Institute, Lincoln's Inn Fields Laboratories, London WC2A 3PX, United Kingdom
Submitted October 4, 2004;
Revised December 23, 2004;
Accepted December 29, 2004
Monitoring Editor: Tim Stearns
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
).
On the completion of anaphase B, elongated spindles are disassembled and replaced by cytoplasmic interphase microtubules, followed by cytokinesis, which requires the deposition of a medial actomyosin ring and its subsequent constriction. These sequential events, consisting of spindle elongation, spindle disassembly, actomyosin ring formation and its resolution, must be spatially and temporally coordinated, to achieve successful cell division. Premature or mislocated cytokinesis would result in the breakage of segregating chromosomes, whereas a delay or bypass of cytokinesis would produce binucleated cells, leading to polyploid or aneuploid daughter cells. These types of events can often be observed in various human disease and cancerous cells (Lengauer et al., 1998; Carmena and Earnshaw, 2003; Dechant and Glotzer, 2003).
As the temporal ordering of mitotic exit and cytokinesis is so vital for cell reproduction, one might imagine that a surveillance mechanism that monitors this final step of cell division would be operational. Indeed, at least in fission yeast Schizosaccharomyces pombe, the existence of such a surveillance mechanism has been reported (Le Goff et al., 1999; Liu et al., 2000). This checkpoint, called the cytokinesis or contractile ring checkpoint, monitors formation and integrity of medial actomyosin ring and septum, and thereby it delays subsequent nuclear division. The checkpoint signaling results in inhibition of Cdc2/Cdk1, leading the cell to two G2 nuclei arrest upon perturbation of cytokinetic structures (Krapp et al., 2004; Mishra et al., 2004). For this checkpoint to be functional, the Wee1 protein kinase and the SIN (septation initiation network) are indispensable. In addition protein phosphatase Clp1/Flp1, a homologue of budding yeast Cdc14, is necessary for this checkpoint signaling (Cueille et al., 2001; Trautmann et al., 2001). In budding yeast S. cerevisiae, a similar, if not identical, mechanism also exists, called the morphogenesis checkpoint, which restrains activation of Cdc28/Cdk1 via Wee1-homologue Swe1 when bud emergence is inhibited (Lew, 2003).
Ase1, first identified in budding yeast, comprises a conserved family of microtubule-associated proteins (MAPs). Depending upon how their homologues were identified and characterized, individual members from different species were designated different names. For instance, budding yeast Ase1, a founding member of this family, was identified as a mutation that displayed synthetic lethality when combined with the bik1 mutation, which is defective in a structural homologue of another MAP, CLIP-170 (Berlin et al., 1990; Pierre et al., 1992; Pellman et al., 1995). As its name represents, Ase1 is required for anaphase spindle elongation and is proposed to comprise midzone-specific spindle matrix (Schuyler et al., 2003). In fly and worm, its homologue Feo (fascetto) and SPD-1, respectively are required for central spindle formation. (Verni et al., 2004, Verbrugghe and White, 2004). In human cells, Prc1 (protein required for cytokinesis) was identified as a mitotic CDK substrate (Jiang et al., 1998) and required for spindle midzone formation and cytokinesis (Jiang et al., 1998; Mollinari et al., 2002). In addition to yeast and metazoans, the Ase1 family is also conserved in plants and called MAP65. The plant homologue also localizes to microtubule overlapping zones (Smertenko et al., 2000; Lloyd and Hussey, 2001). One member MAP65–3/PLE was recently shown to be required for cytokinetic phragmoplast function (Müller et al., 2004). What is common to this family of proteins is their involvement in microtubule bundling, and therefore these proteins are often referred to as microtubule-bundling factors.
In this study we characterize the fission yeast homologue of Ase1/Prc1, which has not been studied before in this organism. Simple rod-shaped morphology directed by tip growth makes this organism an ideal system by which to study cell morphogenesis and cytokinesis (Nurse, 1994; Chang and Verde, 2003; Wang et al., 2003; Mishra et al., 2004). Furthermore detailed knowledge of molecular networks that are involved in mitotic exit and septum formation (Simanis, 2003) is advantageous for understanding mechanisms underlying the termination of cell division. We show that fission yeast Ase1 is not essential for cell viability, but is required for genome stability and, like other members of this family, the integrity of the spindle midzone. Notably we present evidence that in addition to its structural role, fission yeast Ase1 plays a regulatory role in cytokinesis, in which it is a controlling element of the cytokinesis checkpoint, required for inhibiting nuclear division when the cell encounters perturbation of actomyosin ring formation.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
).
|
Nucleic Acids Preparation and Manipulation
Enzymes were used as recommended by the suppliers (New England Biolabs, Beverley, MA and Takara Shuzo Co., Kyoto, Japan). An S. pombe expression vector pREP1, which carried the strong thiamine-repressible promoter, was used for overexpression of ase1+. The ase1+ ORF was amplified by PCR with a pair of primers carrying an NdeI site and a BamHI site. PCR products were digested with NdeI and BamHI and cloned into pREP1.
Gene Disruption and Construction of C-terminally Tagged Strains
A PCR-based gene targeting method (Bähler et al., 1998) was used for constructing gene deletion or C-terminally tagged strains under the natural promoter. All the tagged strains used in this study behaved like nontagged parental strains, indicating that tagging does not interfere with protein function.
Construction of Plasmid Containing the RFP-tubulin (RFP-atb2+) Gene
The entire ORF encoding a monomeric RFP (Campbell et al., 2002) was PCR-amplified with NdeI and BamHI sites at its 5' and 3' end, respectively. The amplified fragment was cloned into NdeI-BamHI sites of pDQ105 (Ding et al., 1998) to replace GFP with the mRFP fragment. This plasmid was designated pREP1-mRFP-atb2+.
Indirect Immunofluorescence Microscopy
Cells were fixed with methanol. Primary antibody (TAT-1, 1/50, provided by Dr. Keith Gull, Oxford University, United Kingdom, or polyclonal anti-GFP antibody, 1/50, 8372-2, Clontech, Palo Alto, CA) was applied, followed by Cy3-conjugated goat anti-mouse IgG (AP124C, CHEMICON, Temecula, CA) or BODIPY-conjugated goat anti-rabbit IgG (B-2766, Molecular Probes, Eugene, OR). Fixed cells were imaged with a fluorescence microscope (Axioplan 2, Carl Zeiss, Oberkochen, Germany) equipped with a chilled CCD camera (Photometrics, Quantix, Tucson, AZ) and MetaMorph software (Universal Imaging Corporation, Downingtown, PA) or a confocal microscope (LSM5, Carl Zeiss). Captured images were processed by use of Adobe Photoshop (version 6.0; San Jose, CA).
Time-lapse Live Imaging Analysis
CFP-
-tubulin was expressed from pRL72 (nmt-CFP-atb2+ plasmid) and used as previously described (Glynn et al., 2001). To observe GFP--tubulin, atb2+ was expressed from the weak thiamine-repressible promoter on pREP81. To observed RFP--tubulin, the procedures previously described (Zimmerman et al., 2004b) were followed. Briefly, cells carrying pREP1-mRFP-atb2+ were grown on minimal plates containing thiamine (5 µg/ml) for 1–2 d and inoculated into liquid minimal culture in the presence of thiamine for 3–5 h.
For live analysis using Metamorph systems, cells were cultured in minimal medium and then mounted on a thin layer of 1% agarose containing minimal medium, which was attached to a glass slide. For analysis of cells in meiosis, homothallic haploid cells were cultured in minimal medium until the midlog phase, washed, shifted to nitrogen-free minimal medium, and incubated for 4–6 h. Live images (stained with GST-GFP-NLS) were viewed with fluorescence microscopy (Niccoli et al., 2004). For live imaging analysis using the DeltaVision microscope (Applied Precision, Issaquah, WA), cells were stuck by lectin to the slide glass at the bottom of a plastic dish, which was then filled with 2–3 ml of minimal media with appropriate supplements. Imaging and processing were performed by SoftWoRx software (Applied Precision). For cold treatment, a dish carrying the cells was chilled with ice, and time-lapse analysis was performed afterward.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
) using budding yeast Ase1 and human Prc1 as queries showed that one ORF (SPAPB1A10.09) encodes a potential homologue, which we call ase1+ hereafter. The ase1+ gene contains one intron and codes for a protein of 731 amino acid residues with stretches of several coiled-coil regions (Figure 1A). As reported previously, Ase1 is a conserved protein among all eukaryotic species (see Supplementary Figure S1), in which overall identity is 
20% (and 40% similarity). Homology is seen in the entire protein among different species, but the most homologous region is the central coiled-coil domain.
|
The Cellular Localization of Ase1
As a first step to address whether Ase1 colocalizes to microtubules, its intracellular localization was examined. For this purpose, the C-terminus of Ase1 and Cut12, a component of the SPB (spindle pole body, the centrosome equivalent; Bridge et al., 1998), was tagged with YFP and CFP, respectively, under the natural promoter. The C-terminal tagging did not interfere with Ase1 or Cut12 function, as a strain containing these tags (Ase1-YFP Cut12-CFP) behaved like wild-type cells in terms of growth properties and cell morphology (see below as to phenotypes of ase1-deleted cells). In addition, to visualize microtubules simultaneously, plasmids carrying RFP-tubulin (RFP-atb2+, see Materials and Methods) were constructed and introduced into an Ase1-YFP Cut12-CFP strain. As shown in Figure 1B, Ase1-YFP colocalized to microtubules during both interphase and mitosis. In interphase, Ase1 localized along cytoplasmic microtubule arrays (top two rows). It was observed that the middle regions of interphase microtubule filaments, which are believed to represent overlapping zones of individual microtubules (Drummond and Cross, 2000; Tran et al., 2000), were most brightly stained with Ase1-YFP. Consistent with this notion, intensities of signals from RFP-tubulin were also stronger in these regions. On entry into mitosis, Ase1-YFP partially colocalized to mitotic spindles (bottom two rows). During this stage, Ase1 localized to mainly three regions, two dots or arrays coincident with the position of the SPBs and the area of the spindle midzone.
Time-lapse live imaging of a strain containing Ase1-GFP (GFP tagging was performed in a manner similar to Ase1-YFP) revealed dynamic alterations of Ase1 localization upon mitotic exit. As shown in Figure 1C, at the end of anaphase B, midzone localization started to fade out and Ase1-GFP signals were seen as medial dots (8 min and see Supplementary Movie S1). Then these dots displayed sequential structural alterations, which would correspond to spindle disassembly and reconstruction of interphase microtubules as reported previously (Heitz et al., 2001; Zimmerman et al., 2004b). At this stage Ase1 appeared to colocalize to a new MTOC, called eMTOC (equatorial microtubule organizing center, 8–11 min). In summary fission yeast Ase1 preferentially localizes to overlapping zones of both interphase microtubules and mitotic spindles. In addition Ase1 is found in the regions to which the mitotic SPBs and the postmitotic eMTOC are situated.
Ase1 Is Not a Core Component of the SPB
Data of subcellular localization presented above suggested that Ase1 localizes to the mitotic SPBs, the centrosome equivalent structures. This apparent colocalization appears inconsistent with the current view that the Ase1/Prc1 family is a microtubule-bundling factor, whereas the centrosomes (and the SPBs) are believed to be involved in microtubule nucleation, but not bundling (Jiang et al., 1998; Smertenko et al., 2000; Lloyd and Hussey, 2001; Schuyler et al., 2003; Tolic-Norrelykke et al., 2004). Also budding yeast Ase1 was not copurified as SPB components (Wigge et al., 1998) and Prc1/Ase1 localization to the centrosomes has not been reported in higher eukaryotes. Given these facts, we set up two experiments to address whether Ase1 is a component of the SPBs. First experiment was depolymerization of microtubules by cold treatment. If Ase1 is a core component of the SPB, Ase1 should remain in the vicinity of the SPBs after microtubule depolymerization. Live cells containing Ase1-YFP, Cut12-CFP, and RFP-tubulin were placed on plastic dish plates precoated with lectin and liquid media were added into the dish, followed by in situ cold treatment. Alterations of localization patterns of these proteins were then examined in live cells with fluorescence microscopy. As shown in Figure 2A, it was clear that Ase1 proteins apparently colocalizing to the SPBs before cold treatment (time 0) disappeared after 20 and 45 min (the left panels, the positions of the SPBs are marked with arrows). Microtubule arrays were indeed depolymerized from these cells (the second panels). Note that SPB signals were retained during cold treatment (Cut12-CFP, the third panels). In fact not only these SPB-like spots but also almost all of other Ase1 signals including those in the spindle midzones (shown with arrowheads) and eMTOC disappeared after 45 min cold treatment. This result strongly suggested that a specific subcellular localization of Ase1 is dependent on microtubule structures and that Ase1 is not a structural component of the SPBs.
|
Second experiment performed was the usage of mutants defective in astral microtubule function. MboI/Mod20 is a nonessential component of the fission yeast 
-tubulin complex and is required for nucleation of cytoplasmic astral microtubules during mitosis (Sawin et al., 2004; Venkatram et al., 2004). As a result, in the mbo1/mod20 mutant, no astral microtubules were formed from the SPB during anaphase B. Observation of Ase1-YFP and Cut12-CFP in mbo1/mod20 mutants showed unambiguously lack of SPB localization of Ase1 during anaphase B (Figure 2B, arrows). Apparent localization of Ase1 to eMTOC is also diminished in this mutant (Sato and Toda, unpublished results). In clear contrast Ase1 was capable of localizing to the spindle midzones in the absence of MboI/Mod20. This result substantiated the notion that Ase1 is not a component of the SPB and its apparent colocalization to the mitotic SPBs is likely to correspond to overlapping astral microtubules that emanated from these structures. In summary Ase1 localizes to overlapping zones of interphase microtubules, mitotic spindles, and anaphase astral microtubules.
Spindle Localization of Ase1 Is Lessened during Midmitosis
In vertebrates, it is reported that mitotic phosphorylation, possibly performed by Cdk, down-regulates microtubule-bundling activity of the Prc1 protein (Mollinari et al., 2002). As a first step to address the cell cycle–dependent regulation of fission yeast Ase1, mitotic localization of Ase1 was examined in more detail using an Ase1-YFP Cut12-CFP strain. On mitotic entry, Ase1 first appeared around the SPB (Figure 3A, row 1), and then moved to the spindle midzone and two distinct areas, which colocalized to the separating SPBs as shown earlier (see Figures 1, B and C, and 2). Careful observation showed that during midmitosis, Ase1-YFP signals at the middle region of the mitotic spindle appeared to be reduced (row 2 in Figure 3A. Note that sister chromosomes did not separate at this stage.). On initiation of anaphase B, Ase1 signals in the spindle midzone became much stronger (row 3 in which DAPI showed two separated bodies).
|
To substantiate the notion of decreased Ase1 localization to preanaphase A spindles, we used the metaphase-arrested nuc2–663 mutant, which is defective in the APC/C ubiquitin ligase (anaphase promoting complex/cyclosome; Hirano et al., 1988; Yamashita et al., 1996; Peters, 2002). A strain was made that contained Ase1-GFP and RFP-tubulin in the nuc2-663 background and GFP and RFP signals were observed under the restrictive conditions (2 h at 36°C). As evident with RFP-tubulin patterns, more than 70% of cells showed metaphase spindles (Figure 3B, the middle panel). Under this condition, Ase1-GFP signals were substantially decreased (the left panel, arrows). Thus, like other members of this family, fission yeast Ase1 appears to be down-regulated during midmitosis, in which Ase1 targeting to the spindle midzone is significantly inhibited before anaphase onset.
Genome Instability in ase1-deleted Mutants
To clarify a physiological role for Ase1, the whole ase1+ ORF was deleted. ase1+ was not an essential gene, but cells deleted for ase1+ showed a number of defective phenotypes as described below. First of all we realized that during propagation of ase1 mutant strains on plates containing phloxine B, the red dye removed from diploid cells less efficiently than haploid cells (Moreno et al., 1991), two types of colonies appeared, one was pink and the other red (Figure 4A). Flow cytometry analysis of the DNA content of cells derived from a single colony displaying each color showed that pink colonies consisted of haploid cells (2C), whereas red colonies were diploid (4C, Figure 4B). Thus the ase1-deleted fission yeast cells are prone to diploidization, which is a rare and unfavorable event in this organism (Broek et al., 1991; Kominami and Toda, 1997). This suggested that ase1 mutants suffer genome instability. To assess the fidelity of chromosome segregation, a strain was constructed in the ase1 deletion background that contained chromosome III–derived nonessential minichromosome Ch16 (Niwa et al., 1989). It was found that in addition to spontaneous diploidization, ase1 mutants displayed a high rate of minichromosome loss, as red or sectored colonies appeared at a high frequency (indicative of minichromosome loss, 3.2%; Figure 4C). These results suggest that although Ase1 is dispensable for cell viability, it is important for maintenance of genome ploidy and stability.
|
ase1 Mutants Are Defective in Cytokinesis and Septation
Microscopic observation of ase1-deleted cells hinted at the reason for diploidization and genome instability. As shown in Figure 4D, many of septated ase1 cells displayed abnormal septa, multiple and/or misoriented (right panels). DNA staining with DAPI showed that separated chromosomes in these septated cells often existed in only one of the two compartments, whereas the other part was anucleated, or the septa appeared to cut through the middle of the chromosomes (top cell). Quantitative measurement of these abnormal cytokinesis and nuclear division showed that in the absence of Ase1, more than 50% of septated cells displayed defects. It is most likely that this defect in septum formation and cytokinesis leads to the appearance of cells that contain two pairs of separated chromosome sets (diploids) or unequal sets (aneuploids). Consistent with this notion, we often found that cultures started originally from haploid cells contain a population (20–30%) of larger cells, both cell length and width, which would correspond to diploidized ase1 mutant cells (bottom two cells in Figure 4D, right). Ase1, therefore, plays an important part in genome stability via spatial and temporal regulation of septation and cytokinesis and thereby it ensures equal partition of segregating sister chromatids.
ase1 Mutants Are not Defective in Progression of Early Mitosis, but Display Abrupt Collapse of Elongating Anaphase B Spindles
Mutants defective in chromosome segregation often show hypersensitivity to microtubule-depolymerizing drugs. Given the severe defects in sister chromatid segregation in the absence of Ase1, we next examined sensitivity of this mutant to MBC (methyl benzimidazole-2-yl carbamate). As shown in Figure 5A, it was found that ase1 mutant cells were hypersensitive to this drug. The sensitivity to microtubule-depolymerizing drugs is sometimes ascribable to failure of or compromised attachment of the kinetochore to the mitotic spindle, which is monitored by the spindle assembly checkpoint (Shah and Cleveland, 2000; Millband et al., 2002). To examine the potential effect derived from simultaneous elimination of Ase1 and the spindle checkpoint, double mutants were constructed between deletions of ase1+ and mad2+, which encodes a conserved component of the spindle checkpoint (Shah and Cleveland, 2000; Millband et al., 2002) and their growth characteristics was assessed. No obvious reduction of colony forming abilities was observed between ase1 single and ase1mad2 double mutants (Figure 5B). This suggested that Ase1 is not involved in progression of early mitosis, the process that ensures bipolar attachment of the kinetochore to the spindle (Waters et al., 1998). Further genetic analysis indicated that, as is the case for budding yeast ase1 mutants (Pellman et al., 1995), fission yeast ase1 deletion strains were synthetically lethal with mutations in cold-sensitive 
-tubulin nda3-311 (Hiraoka et al., 1984). On the other hand, unlike budding yeast, in fission yeast, the double deletions of ase1+ and CLIP-170 homologue-encoding tip1+ (Brunner and Nurse, 2000) were viable. These results implied that some mitotic steps other than early stages are impaired in ase1 mutants. Also it is suggested that Ase1 function is similar, but not identical, between these two yeasts.
|
|
). Consistent with previous results, we did not see any significant delay during early mitosis in ase1 mutants (Table 2). Ase1 is, therefore, required for structural maintenance of elongating anaphase B spindles.
Ase1 Is a Dose-dependent Microtubule Bundling Factor
We next examined the degree of microtubule bundling in the absence of Ase1 or under Ase1 overproduction conditions. Microtubules staining showed that interphase microtubules in ase1 mutants showed overall weaker staining than in wild type and the number of microtubule filaments per cell was increased (Figure 6, A and B). In wild-type cells, we saw on average 3 microtubules per cell, whereas in ase1 mutants the number was increased to 5 or more (Figure 6B). Given the localization of Ase1 to medial microtubule overlapping regions (see Figures 1B and 2), this suggests that Ase1 plays a connecting role in bundling cytoplasmic microtubules. In contrast to deletion phenotypes, ectopic overproduction of Ase1 was lethal (Figure 6C) and resulted in apparent hyper-bundling of microtubules as shown in Figure 6D (1–2 per cell, top right panels). Furthermore these cells showed branched cell morphology, which is a hallmark of compromised cytoplasmic microtubules (Toda et al., 1983; Mata and Nurse, 1997; Brunner and Nurse, 2000). Ase1 overproduction also resulted in mitotic defects, as mitotic spindles looked shorter and thicker (bottom two panels), suggesting that the excess amount of Ase1 blocks spindle elongation. It is of note that these mitotic cells often displayed chromosome segregation defects. Taken together, Ase1 is required for microtubule bundling in both interphase and mitosis and its deletion or overproduction leads to bundling defects in an opposed manner. Thus it appears that Ase1 is a dosage-dependent microtubule-bundling factor.
|
Ase1 Is Not Involved in Cytoplasmic MTOC Function
Given the requirement of Ase1 for microtubule bundling, we next asked whether Ase1 is also involved in nucleation of interphase cytoplasmic microtubules. It is known that in fission yeast, two non-SPB MTOCs, eMTOC, and iMTOC (interphase MTOC), play a role in morphogenesis of interphase microtubules (Heitz et al., 2001; Zimmerman et al., 2004b). eMTOC plays a vital role in reorganization of cytoplasmic microtubules after mitotic exit, whereas iMTOC is required for nucleation of interphase microtubules that emanate from the nuclear periphery (Heitz et al., 2001; Sato and Toda, 2004; Sawin et al., 2004; Zimmerman et al., 2004b). Wild-type and ase1 strains were constructed that contained either MboI/Mod20-GFP or Alp4-GFP, an essential component of the -tubulin complex homologous to budding yeast Spc97 and human Gcp2 (Vardy and Toda, 2000; Fujita et al., 2002; Sawin et al., 2004). To show the position of SPBs, the SPB marker Sad1 protein (Hagan and Yanagida, 1995) was also tagged with dsRed in these strains. It is known that eMTOC resides in the medial region of postanaphase cells before cytokinesis as either dot or ring-like structures to which MboI/Mod20 and Alp4 localize (Vardy and Toda, 2000; Heitz et al., 2001; Sawin et al., 2004; Venkatram et al., 2004). Close examination of postanaphase cells in wild-type and ase1 mutants showed that both MboI/Mod20 and Alp4 proteins exist in a similar pattern, and no obvious differences were observed between these two strains (Figure 7A, dots and rings are marked with arrows and arrowheads, respectively. Quantification data are shown in the right-hand side).
|
To confirm the functional integrity of eMTOC in the absence of Ase1, time-lapse live analysis was performed in strains containing MboI/Mod20-GFP, CFP-tubulin, and Sad1-dsRed. It was found that, like wild-type cells, ase1 mutants displayed microtubule-nucleating activities originated from eMTOC in postanaphase cells, characterized by elongation of reorganized microtubules (arrowheads in Figure 7B). Note that, as shown earlier (see Figure 6A), in ase1 mutants (lower three rows), the number of microtubule filaments, which were thinner in width than those in wild-type cells (upper three rows), was increased. We also examined microtubule nucleation and elongation from iMTOC in ase1 mutants using the same strains described above. As is the case for eMTOC, ase1 mutants appeared to show functional iMTOC activities, as cytoplasmic microtubules emanated from MboI/Mod20 spots elongated toward the cell cortex, though again as in postanaphase cells, cytoplasmic microtubules in ase1 mutants were increased in number (Supplementary Figure S2). Taking these results together, Ase1 is required for anaphase spindle elongation and proper cytokinesis, thereby maintaining genome ploidy, but this protein appears to play no major role in eMTOC and iMTOC functions that involve nucleation and elongation of cytoplasmic microtubules from non-SPB MTOCs.
Ase1 Is Required for Oscillatory Nuclear Movement during Meiotic Prophase and for Production of Viable Spores
Microtubules play an important role in nuclear positioning in many eukaryotes including fission yeast (Reinsch and Gönczy, 1998; Tran et al., 2001). In fission yeast, the nucleus displays dynamic back and forth oscillatory movement during meiotic prophase, which is often referred to as horse tail movement (Robinow, 1977; Chikashige et al., 1994). This oscillatory movement is driven by astral microtubules and believed to ensure pairing of homologous chromosomes before meiotic recombination (Ding et al., 1998). Dynein and dynactin proteins are absolutely required for this oscillatory movement (Yamamoto and Hiraoka, 2001; Niccoli et al., 2004); however, whether other microtubule binding regulatory proteins also play a role in this process remains to be determined. Given the requirement of Ase1 for mitotic microtubule bundling and anaphase completion, we addressed the role of this protein in meiotic nuclear oscillation. Homothallic wild-type, dhc1- and ase1-deleted cells were allowed to proceed into mating and meiosis by nutrient starvation, nuclear movement was followed using GFP fused to a nuclear localization signal. Nuclear movement was substantially compromised in the absence of Ase1, though, unlike dhc1 mutants defective in dynein heavy chain (Yamamoto et al., 1999), some marginal oscillation was still observed. One representative example of live images in each strain is shown in Figure 8A (see Supplementary Movies S5 and S6). We traced the edges of the nucleus, which highlighted compromised nuclear movement in the ase1 mutant (Figure 8B).
|
To examine the morphology of astral microtubules in meiotic prophase cells, GFP-Atb2 was expressed and costained with DNA. As shown in Figure 8C, although wild-type cells mostly contain a single astral microtubule that traverses through the nucleus (left), in the absence of Ase1 many cells (31%) displayed astral microtubules consisting of multiple arrays (right). This suggested that Ase1 plays a role in bundling of meiotic astral microtubules, which is necessary for microtubule-dependent nuclear oscillation during prophase. Consistent with these meiotic defects, we found that spore viability derived from homozygous ase1 deletions was low, because 97% viability for wild-type and 33% viability for ase1 mutants were obtained. Thus Ase1 is necessary for microtubule bundling not only in mitotic cycles but also during meiosis.
Midzone Localization of the Aurora Kinase Is Dependent on Ase1
In vertebrates, it is known that the conserved protein kinase Aurora is required for the completion of cytokinesis (Severson et al., 2000; Speliotes et al., 2000). Although it has not been established whether the fission yeast Aurora kinase (Ark1) is involved in cytokinesis (Petersen and Hagan, 2003), Ark1, like all the members of Aurora kinases in other species, localizes to the spindle midzone (Morishita et al., 2001; Petersen et al., 2001). Accordingly we asked whether Ase1 is involved in Ark1 localization at this site during anaphase. A strain, which contained C-terminally GFP tagged Ark1 (Ark1-GFP) and episomally expressed CFP-Atb2, was constructed in the ase1 mutant and its localization during the cell cycle was examined. Although Ark1 still localized to the spindle in the absence of Ase1, we noticed that the pattern of localization during anaphase is altered in the ase1 mutant compared with wild-type. Instead of localizing to the middle region of anaphase spindles in wild-type cells (Figure 9A), in the absence of Ase1, Ark1 localized to either the region displaced from the center or multiple discrete regions along the spindle (Figure 9B). This result indicates that Ase1 is required for midzone integrity, such that in the absence of Ase1 the midzone protein Ark1 is no longer capable of localizing properly to this region.
|
Ase1 Is Required for Cytokinesis Checkpoint Signaling
Having established the role for Ase1 in microtubule bundling and cytokinesis, we next sought to identify a functional network in which Ase1 is involved during late mitosis and mitotic exit. We crossed an ase1 deletion strain with various mutants defective in septation and cytokinesis. During this genetic analysis, we found that ase1 shows a strong genetic interaction with temperature-sensitive cdc4-8 and cps1-191, defective in myosin light chain and (1,3)-D-glucan synthase, respectively (McCollum et al., 1995; Ishiguro et al., 1997; Cortes et al., 2002). As shown in Figure 10A, the ase1cdc4-8 or ase1cps1-191 double mutant did not form colonies at 32°C, the semirestrictive temperature at which a single cdc4-8 or cps1-191 mutant could grow. It is known that mutants, defective in cytokinesis, including cdc4-8 and cps1-191, trigger activation of the cytokinesis checkpoint (Trautmann et al., 2001; Mishra et al., 2004). To substantiate this genetic interaction, we also tested synthetic growth properties in double mutants between ase1 and cdc3-124 or myo2-E1, defective in profilin and myosin II heavy chain, respectively (Balasubramanian et al., 1994, 1998; Kitayama et al., 1997). These two mutations are also known to activate the cytokinesis checkpoint (Mishra et al., 2004). As is the case for cdc4-8 and cps-191, ase1cdc3-124 double mutants could not grow at 32°C (Figure 10A). In the case of ase1myo2-E1, we could manage to obtain colonies of double mutants; however, these strains grew extremely slowly even at 26°C and fast growing revertants accumulated during propagation of the double mutant, which hampered further analysis. Taken together, ase1 deletions display synthetic growth defects when combined with mutants in the pathway leading to actomyosin ring formation and cytokinesis.
|
Activation of the cytokinesis checkpoint leads to inhibition of successive nuclear division when the completion of cytokinesis is perturbed, which results in the appearance of cells containing two G2 nuclei at the restrictive temperature with compromised actomyosin ring. Intriguingly DAPI staining of ase1cdc4-8 double mutants showed that these cells, instead of arresting as binucleates like the single cdc4-8 mutant, nuclear division continued at the restrictive temperature (Figure 10B) and 70% of cells showed more than 4 nuclei after 4-h incubation at 36°C (Figure 10C). To follow the kinetics of nuclear division in these mutants, the number of the nuclei was counted at each time point upon temperature shift-up. As a control, clp1 single and clp1cdc4-8 double mutants were included in this assay. As shown in Figure 10C, it was found that nuclear division continued in ase1cdc4-8 mutants and furthermore the kinetics of nuclear division was indistinguishable between ase1cdc4-8 and clp1cdc4-8 mutants. This continual nuclear division was reminiscent of cytokinesis checkpoint-deficient cells, as evident in clp1cdc4-8 cells. In addition to ase1cdc4-8, the ase1cps-191 mutant also displayed the increased population of 4 nuclei-cells after 4-h incubation at 36°C (Figure 10D). These results suggested that Ase1 has a regulatory function in inhibiting nuclear division when actomyosin ring formation is perturbed. Checkpoint-deficient mutant cells also include combinations of cytokinesis mutants with wee1 or mutations in the SIN (Le Goff et al., 1999; Liu et al., 2000; Cueille et al., 2001; Trautmann et al., 2001). We then constructed double mutants between ase1 and cdc7-24, in which the SIN signaling is defective, thereby the cytokinesis checkpoint being inactivated (Cueille et al., 2001; Trautmann et al., 2001). In clear contrast to cytokinesis mutations shown above, no reduction of the restrictive temperature was observed in ase1cdc7-24 double mutants (Figure 10A). Taken together analysis described here shows that Ase1 is, in addition to its structural role in microtubule morphology and cytokinesis, an integral component of cytokinesis checkpoint signaling.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Role for Ase1 in Microtubule Bundling
During both interphase and mitosis, the predominant localization site of Ase1 is at the microtubule overlapping zones. In fission yeast during interphase, medial overlapping regions of cytoplasmic microtubule arrays consist of minus ends with the plus ends extending out toward the cell tip, whereas during mitosis the spindle midzone comprises the overlapping plus ends (Tran et al., 2001). This apparent polarity-independent association of Ase1 with microtubules is in clear contrast to other MBPs, such as plus-end tracking proteins including conserved EB1 and CLIP-170 proteins (Schuyler and Pellman, 2001) and conventional MAPs that bind microtubules along longitudinal lattices (Drewes et al., 1998). Thus Ase1 constitutes a unique family of microtubule-bundling proteins that specifically localize to microtubule overlapping regions irrespective of microtubule directionality.
In budding yeast Ase1 overproduction results in forced spindle elongation under conditions of HU treatment. Usually HU addition leads to cell cycle arrest at early S phase with short spindles. This suggests that in addition to a microtubule-bundling activity, Ase1 may play a direct role in promoting spindle elongation (Schuyler et al., 2003). Consistently it is known that elongation of anaphase spindles in fission yeast, as in other systems, takes place at the overlapping zones by polymerization and subsequent sliding (Masuda and Cande, 1987; Masuda et al., 1990; Mallavarapu et al., 1999), pointing toward the possibility that fission yeast Ase1 also promotes spindle elongation at the midzone. Despite this notion, overproduction of fission yeast Ase1 results in bundling and thickening of both interphase microtubules and mitotic spindles and clearly inhibits microtubule elongation instead. The appearance of branched (T-shaped) cell morphology and chromosome segregation defects support this notion, as these phenotypes are reminiscent of malfunctioning of interphase microtubules and mitotic spindles, respectively (Radcliffe et al., 1998; Sawin and Nurse, 1998). Conversely the absence of Ase1 leads to unbundling of interphase microtubules and a physical collapse of anaphase spindles without affecting the velocity of spindle elongation. All these data suggest that fission yeast Ase1's major, if not sole, function lies in a structural connector at microtubule overlapping zones and it is not directly involved in promotion of spindle elongation reaction per se. It is worth pointing out that phenotypic consequences caused by either deletion or overproduction of Ase1 are in fact very similar to those reported for plant MAP65, fly Feo, and human Prc1 (Mollinari et al., 2002). This suggests a conservation of protein function between fission yeast and higher eukaryote Ase1/Prc1.
Mitotic Localization of Ase1
Ase1 shows dynamic cell-cycle–dependent localization. Particularly the appearance of Ase1 signals in the vicinity of the SPBs upon initiation of anaphase B is intriguing. Recent analysis indicates that cytoplasmic astral microtubules are formed at the SPBs coincident with anaphase B onset (Zimmerman et al., 2004a). Importantly these cytoplasmic astral microtubules grow toward both directions. Given that Ase1 localizes to the microtubule overlapping regions including the spindle midzones at this stage, it is most likely that cytoplasmic astral microtubules make contact at the SPBs in an interdigitating manner. Almost complete disappearance of Ase1 from these sites upon microtubule depolymerization supports this notion. At present, a physiological significance of Ase1 localization to these sites remains to be determined, as astral microtubules appear to be formed in the absence of Ase1, though more careful cytological characterization would be needed.
Another interesting point is reduced localization of Ase1 to the spindle during midmitosis (metaphase) when Cdk activity remains high. Mammalian Prc1 and budding yeast Ase1 are CDK substrates and it is shown that Cdk-phosphorylation of Prc1 suppresses microtubule-bundling activity of this protein (Jiang et al., 1998; Mollinari et al., 2002; Ubersax et al., 2003). Given the reduction of Ase1 levels to metaphase spindles, it appears that, at least in fission yeast, bundling of pole-to-pole microtubules via Ase1 during metaphase may not be required for structural maintenance of bipolar spindles, and instead hyper-bundling might be deleterious for formation and establishment of bipolar spindles. Fission yeast Ase1 contains four consensus Cdk phosphorylation sites at the C-terminal region, and we have been currently pursuing a significance of these sites (see below).
Novel Role for Ase1 in Meiotic Nuclear Oscillation
We show that in the absence of Ase1, meiotic nuclear movement is significantly compromised. Meiotic nuclear oscillation is thought to be orientated by the SPB and telomeres and driven by the force generated by dynein-dynactin motor complex (Yamamoto and Hiraoka, 2001). The requirement of microtubule-bundling factor Ase1 for this process implies that these astral microtubules consist of a multitude of overlapping microtubules rather than a single microtubule bundle. In ase1 mutant, horse tail movement is impaired, but not as severely as in dynein or dynactin mutants, which display complete abolishment of oscillatory movement. This suggests that dynein motors retain the activity to drive the nuclear oscillation even in the absence of Ase1 but do not exert its power satisfactorily due to unbundled microtubules. We also show that Ase1 is required to produce fully viable spores. The impairment of horse tail movement probably accounts for the low spore viability, as this oscillatory movement promotes homologous chromosome pairing, thereby ensuring subsequent chromosome segregation (Yamamoto and Hiraoka, 2001). In addition to this, in ase1 mutants precocious breakdown of meiotic spindles might occur, as in mitotic anaphase, which might also contribute to the reduction of spore viability. It is worth noting that Ase1 is the first MAP (except for dynein motors) that is essential for nuclear oscillation during fission yeast meiosis. Other MAPs that are involved in this process still await further analysis.
Ase1 as a Regulatory Component of the Cytokinesis Checkpoint
One of the unexpected findings in this study is the involvement of Ase1 in the cytokinesis checkpoint. This checkpoint has evolved in fission yeast to maintain viability when the cell division apparatus is perturbed (Mishra et al., 2004). This checkpoint acts through down-regulation of ubiquitous Cdc2/CDK activity and requires the Clp1/Flp1 phosphatase (Le Goff et al., 1999; Liu et al., 2000). Budding yeast Cdc14, a homologue of Clp1/Flp1, dephosphorylates CDK substrates, thereby antagonizing CDK activity (Visintin et al., 1998). Interestingly in addition to the involvement of cytokinesis checkpoint, there are other phenotypic similarities between deletions of Clp1/Flp1 and Ase1. These include impairment of cytokinesis and septation, spontaneous diploidization and the requirement of the SIN for phenotypic appearance (Cueille et al., 2001; Trautmann et al., 2001). The simplest interpretation, which explains the involvement of Ase1 in cytokinesis checkpoint signaling, is that Ase1 is a downstream CDK substrate, critical for inhibiting nuclear division and that dephosphorylation of Ase1 is required for initiating and/or maintaining checkpoint activation. As shown in this study, spindle localization of Ase1 is substantially suppressed under the conditions of high Cdk activities. This suggests that Cdk phosphorylation of Ase1 results in reduced activities of microtubule bundling, which somehow results in inhibition of the checkpoint signaling. We have been addressing this point experimentally by mutating potential Cdk sites to nonphosphorylatable or phosphorylation-mimicking residues (Yamashita and Toda, unpublished results).
The alternate possibility, albeit not mutually exclusive with the first possibility, is that Ase1 is required for localization of other proteins that are involved in this checkpoint signaling. An obvious primary candidate would be Clp1/Flp1. However, we have found that this phosphatase is still capable of localizing to central spindles in the absence of Ase1 (Sato and Toda, unpublished results). Instead we have shown that Ark1 localization to the spindle midzone is altered, despite not abolished, in ase1 mutants. It should be mentioned that recent analysis indicates that, in addition to its localization to the mitotic spindles, Clp1/Flp1 physically binds Ark1 (Trautmann et al., 2004), pointing toward the possibility that Clp1/Flp1, Ark1, and Ase1 constitute a functional network involved in mitotic exit and the cytokinesis checkpoint. Although at this stage whether or not Ark1 is implicated in the cytokinesis checkpoint remains to be determined, it is relevant to note that in vertebrates, Prc1-knockdown results in, as in fission yeast cells, the altered discrete localization of Aurora B along spindles, instead of accumulating at the spindle midzone (Kurasawa et al., 2004).
Link between the Microtubule and the Actomyosin upon Mitotic Exit
Microtubule and actomyosin pathways display a mutual interplay during mitosis and cytokinesis in many systems (Ban et al., 2004; Mishima et al., 2004; Yasuda et al., 2004). In fission yeast a contractile actin ring is required to form eMTOC-dependent postanaphase microtubule arrays, which are in turn required for equatorial retention of the contractile ring (Heitz et al., 2001; Pardo and Nurse, 2003). Given the defective cytokinesis accompanied with the displacement of actomyosin ring in ase1 mutants, Ase1 may be a key player that links the microtubule cytoskeleton to actomyosin ring function at the end of the cell cycle. More work is needed to explore the point as to how Ase1 activity and localization are regulated and the underlying mechanism coupling anaphase completion to cell division.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
|
Footnotes |
|---|
The online version of this article contains supplemental material at MBC Online (http://www.molbiolcell.org). 
These authors contributed equally to this work.