INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Mitosis is the process by which eukaryotic cells ensure the faithful transmission of their chromosomes. Many studies have shown that proteolysis governs mitotic progression by the orderly degradation of key proteins via the ubiquitin-proteasome pathway (see reference 50 for a recent review). Substrate specificity is conferred by a ubiquitin-protein ligase called the anaphase-promoting complex (APC) and associated regulatory proteins. After maturation-promoting factor drives cells into mitosis, the APC forms a complex with the substrate-specific activator Cdc20 to trigger the proteolysis of inhibitors of anaphase (Pds1 in Saccharomyces cerevisiae and Cut2 in Schizosaccharomyces pombe), thereby allowing the separation of sister chromatids and anaphase spindle elongation. Subsequently, APC binds another substrate-specific activator, Hct1/Cdh1, and the complex directs the destruction of AseI and the mitotic cyclin, thus allowing spindle disassembly and exit from mitosis. Mitotic progression is also controlled by a surveillance system, the spindle checkpoint, which prevents anaphase onset until chromosomes are properly attached to the spindle. Studies with yeasts and vertebrates indicate that the checkpoint consists of components able to sense unattached kinetochores and a signal transduction pathway which delays the onset of anaphase through inhibition of the APC. When the correct attachment is achieved, the negative signal disappears, APC is activated, and anaphase proceeds (for recent reviews, see references 19 and 50).

Proteins ubiquitinated by the APC are degraded by the proteasome, a large 28S protein complex made of a 20S proteolytic core particle (CP) and a 19S regulatory particle (RP). The crystal structure of the 20S complex from budding yeast has been determined previously (18). It consists of two A and two B rings, each made of seven different A and seven different B subunits. The active proteolytic sites are within the CP, and substrates are thought to enter the CP through the lumens located at each end of the cylinder. The CP alone can hydrolyze only small peptides and exists predominantly in a closed state, indicating that the ends of the channel are gated. Degradation of ubiquitinated proteins necessitates the binding of RP. In budding yeast, the RP contains at least 18 subunits and can be dissociated into two subparticles, the base and the lid (15). The base contains the six ATPase subunits and connects the RP to the CP. The base activates the CP for degradation of peptides and nonubiquitinated proteins, whereas the lid is required for ubiquitin-dependent degradation. It has been proposed that the lid selects ubiquitinated substrates and that the base gates the CP channel and unfolds the substrates.

The centromere-kinetochore complex is a key element in the process of mitosis. It is defined as the nucleoprotein complex containing chromosomal factors and centromeric DNA. It fulfills mechanical functions such as sister chromatid cohesion, attachment to the spindle, and chromosome movement as well as the regulatory function of monitoring chromosome attachment to the spindle. In most eukaryotes, centromeric DNA consists of long arrays of repetitive DNA packaged into transcriptionally silent, recombination cold, late-replicating heterochromatin. The exact function of heterochromatin is unknown, but kinetochores in higher eukaryotes generally form within centromeric heterochromatin (23). In the fission yeast, S. pombe, three regions of the genome are known to be assembled into heterochromatin. These are the silent mating-type loci, telomeres, and centromeres (2, 34, 46). Centromeres consist of inner (imr/B) and outer (otr/K+L) repetitive sequences arranged in a large inverted structure around a central core (44, 45) (see Fig. 1A). As is often the case for heterochromatic regions in higher eukaryotes, recombination (33) and transcription (2, 3) are repressed within fission yeast centromeres, and they are located at the nuclear periphery in interphase (12). The link between heterochromatin formation and centromere function has been clearly established for S. pombe. Mutations in clr1, clr2, clr3, clr4, rik1, and swi6 genes were shown to affect repression at the silent mating-type loci, at centromeres, and, to a limited extent, at telomeres (3, 11, 26, 47). In particular, mutations in clr4, rik1, and swi6 strongly affect silencing within the imr and otr regions, but not the central core, of the centromere and cause an elevated rate of chromosome loss. The swi6⁺ gene encodes a protein with similarity to Drosophila melanogaster HP1 (27). HP1 localizes to pericentric heterochromatin, and mutations in HP1 suppress repression of marker genes lying in the vicinity of centromeric heterochromatin and apparently induce chromosome missegregation events (24). Both Swi6 and HP1 proteins contain a chromodomain and a chromo-shadow domain. It has been suggested that this class of proteins could act as adapter molecules to mediate heterochromatin formation (1, 38). Swi6 localizes to the three known silent domains of the fission yeast genome, centromeres, telomeres, and the mating-type region (9), and functional Clr4 and Rik1 are required for this localization (10). Clearly, heterochromatin assembly appears to be crucial for fission yeast centromere function, since the silencing defect caused by the lack of a functional clr4, rik1, or swi6 gene is correlated with defective movement of centromeres at anaphase and an elevated rate of chromosome loss (3, 9). It is very likely that many other structural components as well as regulatory factors are required to form normal centromeric heterochromatin and kinetochore assembly. Here a genetic screen for mutants which affect centromeric silencing within the central core region of S. pombe centromeres has been employed. The screen has identified a new class of mutants called cep (centromere enhancer of position effect). All cep mutants were found to be defective in mitotic chromosome segregation. They define three genes designated cep1, cep2, and cep3. The cep1⁺ and cep2⁺ genes were cloned and, surprisingly, were found to encode regulatory subunits of the proteasome complex. The cep3⁺ gene remains to be identified and characterized. The results are discussed with respect to centromeric silencing and regulation of the metaphase-anaphase transition in fission yeast.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Strains, media, transformation, and genetic techniques. All the strains used in this study are listed in Table 1. Media were essentially as described previously (2, 32). YEA refers to yeast extract medium (YE) supplemented with adenine; N/S refers to pombe minimum glutamate medium PMG containing adenine, uracil, histidine, and leucine; FOA is N/S but with the addition of FOA (5-fluoro-orotic acid) at 1 g/liter and URA is N/S medium without uracil. For transformation, a simplified version of the lithium acetate procedure was used (32). Standard genetic techniques were used (32). Comparative plating and serial dilution experiments were performed as described previously (2). Ch16 loss was measured by the half-sectoring assay method as previously described (3) with YE plates supplemented with a limiting amount (12 mg/liter) of adenine.

Isolation of cep mutants. Strain FY312 was mutagenized with 2% ethyl methanesulfonate as described previously (32). Following ethyl methanesulfonate treatment, cells were washed three times with 0.9% NaCl, resuspended in YEA at a density of 4 × 10⁶ cells/ml, and allowed to grow at 32°C for 21 h, after which time the cell density was around 4 × 10⁷ cells/ml. Cells were plated onto FOA plates at a density of 10⁵ and 10⁶ cells/plate and incubated for 4 days at 35°C. From a total of 10⁷ cells plated, approximately 1,000 FOA-resistant colonies formed. To screen for cryosensitivity (cs), the 1,000 colonies were grown at 35°C for 3 days and then replica plated onto two N/S plates containing phloxin B and incubated at 18 and 32°C for 4 days. Six strains grew very poorly and stained dark pink at 18°C. The six mutant strains were backcrossed three times. For all mutants, tetrad analysis showed that FOA resistance and cs cosegregated and segregated 2:2 with wild type.

Cytological techniques. Immunostaining and fluorescence in situ hybridization (FISH) were performed as described previously (10). Antihemagglutinin (anti-HA) monoclonal antibody 12CA5 (Babco) was used at a 1/100 dilution. Polyclonal anti-HA (Babco) was cleaned up by preabsorption three times against total S. pombe proteins from a strain lacking the HA epitope immobilized on a membrane filter by Western blotting. The anti--tubulin monoclonal antibody TAT1 (51) was a gift from Keith Gull. A Zeiss Axioplan fluorescence microscope coupled to a Photometrics camera and IP Lab Spectrum software with Digital Scientific extensions were used for the capture and analysis of images.

Western blotting. Total proteins were extracted in radioimmunoprecipitation assay buffer as described previously (32). Samples containing 50 µg of total protein were electrophoresed on duplicate sodium dodecyl sulfate-10% polyacrylamide gels. One gel was Coomassie blue stained to check for equal loading, and the other was electroblotted onto a polyvinylidene difluoride membrane (Bio-Rad). The filter was incubated with a 1:1,000 dilution of the anti-HA monoclonal antibody 16B12 (Babco). Bound antibody was detected with alkaline phosphatase-conjugated sheep anti-mouse immunoglobulin G with 5-bromo-4-chloro-3-indolyl phosphate-nitroblue tetrazolium as a substrate.

Plasmids. The S. pombe cDNA library was provided by C. J. Norbury and B. Edgar (Imperial Cancer Research Fund, Cell Cycle Group, Oxford, United Kingdom). In this library, cDNAs are directionally cloned into pREP3X, an S. pombe vector in which expression is under the control of the thiamine-repressible nmt1 promoter (29). A fission yeast genomic library in pUR19 with the S. pombe ura4⁺ gene as the selection marker was provided by A. M. Carr (Medical Research Council Cell Mutation Unit, Brighton, United Kingdom).

Isolation of the cep1⁺ and cep2⁺ genes. The cep1⁺ and cep2⁺ genes were isolated by complementation of the cryosensitive phenotype of the mutant strains. Strain FY1008 (cep1-1) was transformed with the S. pombe cDNA library. Transformants were selected on minimal medium without thiamine at 32°C, and colonies were replica plated onto the same medium and incubated at 18°C for 7 days. Two colonies formed at 18°C. Plasmid DNA was recovered in Escherichia coli DH5 by the extraction procedure previously described (5). Sequence data from the ends of the inserts showed that both plasmids contained the same cDNA. Database searches with the BLAST program (4) showed that the cep1-1-complementing cDNA was derived from the previously cloned pad1⁺ gene (41). The complementing plasmid (pREPcep1⁺) linearized within cDNA sequences at the single NruI site was integrated into FY947 (cep1⁺). Integration by homologous recombination was checked by Southern blotting of genomic DNA extracted from the transformants (data not shown). Genetic analysis of crosses between integrant strains and FY1007 (cep1-1) showed that the LEU2 marker from the plasmid was tightly linked (<1 centimorgan [cM]) to the cep1 locus.

Generation of HA-tagged cep1 alleles. The cep1 locus of strain FY1521 carries a cep1⁺ allele extended with three copies of the HA sequence. The construction of FY1521 is described elsewhere (37). The same procedure was used to tag the cep1-1 allele. Plasmid pUC19ura4-cep1-3HA (37) was linearized at the single SmaI site lying within cep1 sequences and integrated into FY1008 (cep1-1) to give FY1644 (cep1^cs-HA).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation of enhancer-of-centromere-position-effect mutants. The overall organization of DNA sequences within fission yeast centromere 1 (cen1) is shown in Fig. 1A. As previously demonstrated, insertion of the marker gene ura4⁺ within centromeric sequences results in repression of ura4⁺ gene expression (2, 3). This can be visualized by assaying the ability of the marked strains to grow on selective (URA) and counterselective (FOA) plates (see Fig. 3). Whatever the site of ura4⁺ insertion within cen1, the reporter gene is always transcriptionally repressed. However, repression is never complete, since all integrant strains eventually form colonies on medium lacking uracil. The level of repression is dependent on the site of insertion, therefore defining domains within the centromere (3). For example, ura4⁺ is more tightly repressed when inserted within the otr and imr repeats than when inserted within the central core cnt1. In addition, expression of ura4⁺ from within the central core is temperature sensitive: increased temperature allows more growth on URA selective medium, but growth on FOA is inhibited. These growth characteristics are indicative of increased expression from the central core at higher temperatures. This property was exploited to perform a screen for mutations which enhance transcriptional repression (silencing) within the central core of fission yeast centromeres.

View larger version (12K): [in this window] [in a new window]   FIG. 1.   Strategy utilized to screen for mutants which enhance transcriptional repression within the central domain of S. pombe centromeres. (A) Schematic representation of S. pombe cen1. (B) Screening procedure. Details are given in the text. wt, wild type; EMS, ethyl methanesulfonate.

View larger version (85K): [in this window] [in a new window]   FIG. 2.   Cold sensitivity and MBC resistance of cep mutants. Cells grown at 32°C were harvested, and about 103 cells were spotted onto complete medium (YEA) with or without MBC (20 mg/liter). Incubation times were 7 days at 18°C and 3 days at 32°C. WT, wild type.

View larger version (92K): [in this window] [in a new window]   FIG. 3.   (Top and middle) Mutations in cep1, cep2, and cep3 genes enhance repression of the ura4 gene when placed within the central domain of cen1. (Bottom) cep mutations also enhance silencing within the outer (otr) and inner (imr) repeats of centromere 1. Cells grown at 32°C in N/S medium were serially diluted (one-fifth); spotted onto URA, FOA, and N/S plates; and incubated at 32 or 36°C. About 2 × 104 cells were plated in the highest-density spots. Strains and location of the ura4 gene are as indicated. wt, wild type.

Mutations in the cep genes alter the fidelity of chromosome transmission. We have previously shown that clr4, rik1, and swi6 mutants affect centromeric silencing and the fidelity of transmission of chromosomes, pointing out the link between centromeric silencing and centromere function (3, 9, 10). As cep mutations modify centromeric silencing, we asked whether chromosome stability was also affected. The rate of chromosome loss was estimated by scoring the mitotic loss rate of the 530-kb Ch16 linear minichromosome (28) at 32 and 36°C (Table 2) by the half-sectoring assay method (3). In wild-type cells, the Ch16 loss rate is less than 0.1% of cell divisions (Table 2) (3, 35). As shown in Table 2, the minichromosome was lost at an elevated rate in all cep mutants. The most dramatic increase in loss rate is seen in cep1 and cep3 mutants, where the loss rate per division is up to 80- to 100-fold higher than that of wild type; a similar elevated rate was observed for rik1, clr4, and swi6 mutants (3). For cep2 mutants, the loss rate is moderate but still significantly higher than that in wild type (5- to 15-fold increase). As previously observed for suppressors of centromeric silencing, enhancers of centromeric repression within the fission yeast centromere also cause an elevated rate of chromosome loss, suggesting a role for cep genes in centromere function.

Mutations in all three cep genes cause defective mitosis at restrictive temperature. The cep mutants lose chromosomes at elevated rates at the permissive temperature and are cs for growth. It is possible that the cs phenotype results from a failure in centromere function during mitosis. To investigate this possibility, we performed cytological analysis of mutant cells upon shift to the restrictive temperature. Cells were grown to early log phase at the permissive temperature (32°C) and shifted to 18°C. At a 1.5-h interval, samples were removed for determination of survival over time and cytological analysis. 4',6-Diamidino-2-phenylindole (DAPI) and antitubulin antibody staining was used to visualize nuclear chromatin and microtubules, respectively. Approximately 200 to 300 cells were observed for each time point. Based on the absence or presence of the mitotic spindle, cells were classified to be in interphase, metaphase and early anaphase (spindle < 5 µm) or late anaphase (spindle > 5 µm). When shifted from 32 to 18°C, wild-type and mutant cells nearly stopped dividing, as shown by the very small proportion of metaphase cells (Fig. 4A), but after 4.5 h, the population was composed of about 12 to 16% metaphase cells (Fig. 4A and D, 4.5 h). Therefore, the shift from 32 to 18°C appeared to synchronize cells to some extent. For wild type, the fraction of metaphase cells decreased after 4.5 h but remained high until 6 h for all three mutants, indicating that mutant cells spent more time in metaphase than did wild type.

View larger version (25K): [in this window] [in a new window]   FIG. 4.   Phenotypes of cep1, cep2, and cep3 mutants at the restrictive temperature. (A to C) Wild-type, cep1-1, cep2-12, and cep3-16 cells were grown to early log phase at 32°C and transferred at 18°C, and samples were taken for analysis every 1.5 h as described in the text. When shifted to the cold, cells from wild-type and cep mutants stopped dividing (A), but after 4.5 h, growth resumed, and a wave of mitosis was observed. Unlike the wild-type control, cep mutant cells divided only once at 18°C (B) and lost viability (C). (D) Cytological analysis by DAPI (red, pseudocolor) and antitubulin staining (green). Row 1 shows that early mitotic cells (spindle length < 5 µm) were first observed 4.5 h after transfer at 18°C. Rows 2 and 3 show that wild-type cells proceeded normally through mitosis (leftmost panels) but that all three cep mutants experienced defective chromosome segregation. After 6 h (row 2), about 50% of anaphase cells displayed lagging and/or aberrantly segregated chromosomes, and later (9 h [row 3]) abnormal telophase and septated cells were observed. After 12 h (row 4), cep1 and cep2 mutants eventually arrested as interphase-like cells while cep3-16 cells attempted a second mitosis but arrested as metaphase-like cells (rightmost panel). (E) Centromere FISH analysis of anaphase cells. The leftmost panel shows a wild-type cell with the centromeres (red signal) separated at both ends of the anaphase spindle (green). In most cases, the FISH signals in cep mutant cells were clustered at only one pole or in the midzone of the anaphase spindle, indicating a defect in sister chromatid separation. (F) rDNA FISH analysis of anaphase cells showing the absence of separation of chromosome III sister chromatids in cep mutants. Bars, 5 µm.

The cep1 and cep2 genes encode regulatory subunits of the proteasome. The cep1⁺ and cep2⁺ alleles were cloned by complementation of the cs phenotype with a cDNA library and a genomic library, respectively. Strains in which the complementing plasmid was integrated by homologous recombination within insert sequences were made. Genetic analyses showed that homologous recombination had occurred at cep1 and cep2 loci, thus demonstrating that cloned sequences contained the cep1⁺ and cep2⁺ genes.

Cellular localization of Cep1. The cep1 locus of strain FY1521 carries cep1⁺-HA, a fully functional cep1 allele tagged with three copies of the HA epitope at the C terminus (37). The cs allele of cep1-1 was also tagged with the three-HA tag (cep1^cs-HA, strain FY1644). As shown in Fig. 5, anti-HA antibodies readily detect a protein of the expected size (39 kDa) in extracts from FY1521 and FY1644, whereas no signal was detected in an untagged cep1⁺ strain. The intensity of the signal appears lower in cep1^cs-HA. A time course experiment after shifting to the restrictive temperature showed that the level of Cep1^cs-HA remained constant with time, showing that the cs does not result from the loss of the protein (data not shown).

View larger version (97K): [in this window] [in a new window]   FIG. 5.   Western blot detection of HA-tagged Cep1 and Cep1cs proteins. Equal amounts of total proteins from the indicated strains were loaded in the lanes. The blot was probed with monoclonal anti-HA antibody.

View larger version (50K): [in this window] [in a new window]   FIG. 6.   Cellular localization of Cep1. Cells from strain FY1521 (cep1+-HA) were grown to early log phase and then fixed and stained with anti-HA (red), antitubulin (green), and DAPI (blue). (A) Montage showing the localization of Cep1 through the cell cycle. For each cell, the corresponding panels are placed one below the other. From left to right, an example is shown of cells at various stages of the cell cycle from interphase to cytokinesis. (B) Anti-HA signal alone. Bar, 5 µm.

Chromosome segregation in other mutants with defective proteasome function. The mts2-1, mts3-1, and mts5-1 (pad1-1) temperature-sensitive (dead at 36°C) mutations in components of the proteasome have been described previously (16, 17, 37). To test if elevated rates of chromosome loss are a general property of defective proteasome function, the minichromosome Ch16 was crossed into mts2-1, mts3-1, and pad1-1 backgrounds, and the rate of minichromosome loss was assessed at the permissive temperature (25°C) and semipermissive temperature (30°C). As can be seen from Table 4, Ch16 stability is not affected at the permissive temperature and the loss rate remains low at 30°C even though colony growth was affected. All fission yeast proteasome mutants isolated confer MBC resistance at permissive temperatures (32°C for cep mutants and 25°C for mts mutants), indicating that the gene products are not fully functional even under permissive conditions. However, only cep alleles show elevated chromosome loss rates in these conditions. Thus, chromosome loss is not a general phenotype associated with defective proteasome function but appears to be specific to the pad1 (cep1-1) and mts2 (cep2) alleles isolated in this study.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Two cep genes encode structural components of the proteasome. In this study, we report the results of a screen for mutations which enhance silencing from within the central domain of fission yeast centromeres. The primary screen with silencing as a sole selection gave a large number of putative mutations. An additional screen for cold sensitivity reduced the number to six cep (centromere enhancer of position effect) mutations. They define three genes, two of which were cloned and identified by sequence analyses. The cep1⁺ gene was found to be identical to pad1⁺, encoding an essential regulatory subunit of the proteasome (37, 41). Similarly, cep2⁺ is identical to mts2⁺, encoding another essential component of the 19S regulatory cap of the 26S proteasome (16). In addition to the phenotypes described, the cep1, cep2, and cep3 mutants were also found to be resistant to the microtubule poison MBC. This property, MBC resistance, is also shared with fission yeast strains bearing mutations in genes encoding other components of the proteasome such as mts3 and mts4 (16, 17, 37, 49), and overexpression of the human pad1 homolog, POH1, results in similar phenotypes (43). The similarity in phenotypes of all the cep mutants and defects in other proteasome subunits strongly suggest that cep3⁺ might encode another component of the proteasome. This hypothesis awaits the isolation of the cep3⁺ gene. However, attempts to clone cep3⁺ by the normal approach of complementing cep3-16 cs have so far failed, suggesting that multicopy cep3⁺ could be deleterious to the cell.

Mutations in 19S cap proteasome subunits impair chromosome segregation. The proteasome is known to degrade critical regulatory proteins involved in distinct cellular processes. In fission yeast, temperature-sensitive mutations in mts2⁺, mts3⁺, mts4⁺, and pad1⁺ genes cause cells to be transiently arrested at the metaphase stage of mitosis. At the restrictive temperature, cells accumulate with a short spindle and undivided nucleus. The spindle never elongates to the length of an anaphase spindle but rather is degraded. Septation occurs without nuclear division, forming an anucleate cell (16, 17, 37, 49). It has been shown that cells with defective Mts2 are unable to degrade ubiquitinated proteins, consistent with the idea that the metaphase arrest results from a failure to hydrolyze APC substrates (16). The cs alleles of mts2⁺ and pad1⁺ isolated in this study show somewhat different mitotic phenotypes. Unlike the heat-sensitive alleles, cep1-1 and cep2-12 cells did not accumulate with a short spindle but instead showed a rather normal spindle elongation process and went through anaphase, telophase, and cytokinesis and eventually arrested in interphase. However, the segregation of chromosomes was defective. In roughly half of anaphase cells, chromosomes were found to lag on the spindle, a situation which is rarely seen with wild-type cells. FISH analyses revealed that aberrantly segregating chromosomes had often failed to separate their sister chromatids even in mid- to late anaphase, indicating that cep mutants are defective in sister chromatid separation. Similarly, cells expressing nondegradable Cut2 block sister chromatid separation, but importantly, spindle elongation is also inhibited (13, 14, 25). As spindle elongation is clearly not affected in cep1 and cep2 mutants, it seems likely that Cut2 proteolysis occurs but that another, unknown target (perhaps centromeric) is not degraded, leading to the chromatid separation defect observed in late anaphase (see below). Detection of changes in Cut2 protein levels is clearly difficult with mutants which are not completely defective in proteasome function and might just slightly alter Cut2 dynamics.

Centromeric silencing and the ubiquitin-proteasome pathway. The mutants isolated in this study display different phenotypes depending on the temperature. At the restrictive temperature, sister chromatid separation is defective. At the permissive temperature of 36°C, centromeric silencing is enhanced and the Ch16 minichromosome is lost at an elevated rate. As cep1 and cep2 mutations affect components of the proteasome, it seems reasonable to assume that the mutant phenotypes result from defective proteolysis. A key question is whether the cep1 and cep2 chromosome segregation and silencing phenotypes result from the altered degradation of a single protein or whether there are other protein targets which are not properly processed and cause the enhancement of silencing and elevated chromosome loss.