PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of molcellbPermissionsJournals.ASM.orgJournalMCB ArticleJournal InfoAuthorsReviewers
 
Mol Cell Biol. 2011 June; 31(12): 2470–2483.
PMCID: PMC3133416

S-Phase Cyclin-Dependent Kinases Promote Sister Chromatid Cohesion in Budding Yeast [down-pointing small open triangle]

Abstract

Genome stability depends on faithful chromosome segregation, which relies on maintenance of chromatid cohesion during S phase. In eukaryotes, Pds1/securin is the only known inhibitor that can prevent loss of cohesion. However, pds1Δ yeast cells and securin-null mice are viable. We sought to identify redundant mechanisms that promote cohesion within S phase in the absence of Pds1 and found that cells lacking the S-phase cyclins Clb5 and Clb6 have a cohesion defect under conditions of replication stress. Similar to the phenotype of pds1Δ cells, loss of cohesion in cells lacking Clb5 and Clb6 is dependent on Esp1. However, Pds1 phosphorylation by Cdk-cyclin is not required for cohesion. Moreover, cells lacking Clb5, Clb6, and Pds1 are inviable and lose cohesion during an unperturbed S phase, indicating that Pds1 and specific B-type cyclins promote cohesion independently of one another. Consistent with this, we find that Mcd1/Scc1 is less abundant on chromosomes in cells lacking Clb5 and Clb6 during replication stress. However, clb5Δ clb6Δ cells do accumulate Mcd1/Scc1 at centromeres upon mitotic arrest, suggesting that the cyclin-dependent mechanism is S phase specific. These data indicate that Clb5 and Clb6 promote cohesion which is then protected by Pds1 and that both mechanisms are required during replication stress.

INTRODUCTION

A temporal coupling between S phase and mitosis is essential to prevent aberrant chromosome segregation that can result in the aneuploidy associated with many cancers. One crucial regulatory process in S phase ensures that loss of chromosome cohesion does not occur until after DNA replication has been completed. In budding yeast, Pds1 is the only known protein that prevents cohesion loss, and it functions by inhibiting the Esp1 protease that cleaves the cohesin subunit Mcd1/Scc1 (9). Pds1 binds to Esp1, a yeast separase homolog, until anaphase, when Pds1 is removed by ubiquitin-dependent degradation (26). When this occurs, the Esp1 protease is able to cleave the cohesin subunit that holds sister chromatids together. This cleavage allows the separation of sister chromatids (44).

However, the role of Pds1 in cohesion maintenance appears to be limited to conditions of replication stress and has been observed only in late S phase (12). pds1-null mutants are viable and are able to prevent loss of cohesion when DNA replication is blocked in early S phase (47, 48). Mcd1/Scc1 cleavage occurs in a cell cycle-regulated manner in the absence of Pds1 (2). Moreover, pds1-null cells can maintain cohesion throughout S phase when DNA replication proceeds in the absence of stress at replication forks (3). Only when DNA replication is slowed in the presence of the ribonucleotide reductase inhibitor hydroxyurea (HU), which depletes nucleotide pools, does premature loss of cohesion occur in pds1-null cells (12). Together, these data indicate that slow or stalled replication forks put stress on the mechanisms that lead to the establishment and maintenance of cohesion. In the absence of replication stress, there may be a redundant mechanism that renders Pds1 dispensable for cohesion.

In mammalian cells the Pds1 homolog, securin, is similarly dispensable for viability (24), and mice lacking securin are viable (29). Still, the gene encoding securin is often overexpressed in aggressive tumors (49). These facts indicate the central importance of Pds1/securin anaphase inhibitors but also indicate that other mechanisms must be sufficient for cohesion in mammals as well as in yeast.

In addition to securin, cyclin B1 has been shown to prevent the onset of anaphase and loss of cohesion during mitosis in vertebrates (7, 21, 38). A nondegradable cyclin B1 mutant inhibits anaphase by activating Cdk1 kinase. Based on biochemical studies, the mechanism by which Cdk1-cyclin B1 blocks anaphase may be related to its ability to inhibit separase, although direct proof of this mechanism from a biological perspective has been difficult to obtain (18).

In budding yeast, the cell cycle is controlled by the cyclin-dependent kinase Cdc28/Cdk1. B-type cyclins, Clb1 to -6, interact with Cdc28 as stage-specific activators at different points during the cell cycle. There is some redundancy in the functions of these B-type cyclins, and this has hampered dissection of their roles in the cell cycle. Clb5 and Clb6 promote the initiation of DNA replication (28). Although other Clbs can also initiation replication (30), S phase occurs more slowly in clb5-null cells (13). Cells lacking Clb3 and Clb4 have defects in spindle pole body separation and spindle formation (37), while either Clb1 or Clb2 is required for mitosis (28, 30).

Since, at least in vitro, vertebrate Cdk/B-type cyclin complexes can inhibit separase activity, we asked if one mechanism of S-phase cohesion is mediated by Cdc28/cyclin. To investigate possible roles of B-type cyclins in S-phase control of cohesion, we surveyed the sensitivity of clb deletion strains to replication stress induced by HU. We observed a marked HU sensitivity specifically in clb5Δclb6Δ strains and present evidence that these particular cyclins are required in S phase for sister chromatid cohesion. The yeast Cdk, Cdc28, must phosphorylate Pds1 to allow efficient binding of Pds1 to Esp1. But we find that these phosphorylations are not necessary for cohesion. Therefore, the role of cyclin in promoting cohesion is independent of Pds1 phosphorylation. In agreement, there is a synthetic interaction between pds1Δ and clb5Δ clb6Δ. The triple mutant lacked the ability to maintain cohesion in S phase even in the absence of exogenous replication stress. Moreover, loss of cohesion in clb5Δ clb6Δ cells required Esp1, and similarly to vertebrates, we were able to detect Cdc28/Esp1 complexes in yeast cell extracts. However, intriguingly we found that the cohesin subunit Mcd1/Scc1 does not become enriched on chromatin in early-S-phase clb5Δ clb6Δ cells during replication stress. clb5Δ clb6Δ cells were not defective in enhancing the abundance of Mcd1/Scc1 near centromeres upon mitotic arrest. Together, the data indicate that parallel mechanisms using both Pds1 and B-type cyclins ensure cohesion between the sister chromatids in S phase.

MATERIALS AND METHODS

Yeast strain construction.

Strains described herein were derived from BF264-15 15DU:MATa ura3Δns ade1 his2 leu2-3,112 trp1-1a (35) and were grown at 30°C unless temperature sensitive in medium containing yeast extract and peptone (YEP) plus dextrose (YEPD) or galactose (YEPG) as a carbon source. Strain genotypes are shown in Table 1.

Table 1.
Strains used in this work

Cell cycle analysis.

Overnight cultures were grown with extra adenine to decrease background green fluorescence. The temperature-sensitive clb5Δ clb6Δ pds1Δ triple deletion strain was grown overnight at 25°C in YEPG to maintain PDS1 expression until cells were synchronized in YEPD. The Esp1-2td protein was inactivated as previously described (33). All other strains were grown at 30°C overnight in YEPD. For synchrony, cells were diluted to an optical density (OD) of 0.15 in YEPD with mating pheromone (concentrations varying from 1:1,000 to 1:5,000 of a 1-mg/ml stock). After 2 to 3 h, cells were monitored microscopically until G1 synchrony (85 to 100%) was reached. Mating pheromone was washed off with water, and cells were released under experimental conditions. For each strain, 100 to 200 cells were counted and samples were taken for FACScan analysis as described previously (11, 20) using Sytox green DNA stain (Molecular Probes, Inc., Eugene, OR).

Microscopy.

Spindle morphologies were visualized using TUB1-GFP (40) and sister chromatid segregation with the LacO/green fluorescent protein (GFP)-LacI system (39). Fluorescence and differential interference contrast (DIC) microscopy with Plan Apo 63×/1.4 and Alpha Plan Fluar 100×/1.45 objectives and a Zeiss Axio Plan II microscope were used for scoring and capturing images of live cells with a Zeiss Axiocam camera and AxioVision software.

Western blots.

Cell cultures were handled as described above except with no extra adenine addition. Collected samples were washed with ice water, and pellets were kept at −80°C until protein was extracted using glass beads in NP-40 buffer (50 mM Tris-HCl at pH 7.5, 250 mM NaCl, 0.1% NP-40, 10 mM sodium pyrophosphate, 5 mM EDTA, 0.1 mM orthovanadate, 1 mM phenylmethylsulfonyl fluoride [PMSF], 2 μg/ml aprotinin, 2 μg/ml pepstatin, and 1 μg/ml leupeptin) (similar to what was described in reference 25). At 4°C, cells were vortexed in 1.5-ml tubes for 10 min and centrifuged at 14,000 rpm for 10 min and again for 5 min to remove sediment. Equal levels of protein, as determined by 280-nm absorption, were used for SDS-PAGE. Immunostaining was performed with a 1:1,000 dilution of anti-Myc (9E10; Santa Cruz) and a 1:10,000 dilution of anti-PSTAIRE. Secondary antibody, horseradish peroxidase (HRP)-conjugated goat anti-mouse antibody (Pierce, Rockford, IL), was used at 1:5,000. Western Lightning Chemiluminescence Reagent Plus (Perkin-Elmer Life Sciences, Boston, MA) was used per the manufacturer's instructions. For the Suc1 precipitations, extracts were made in lysis buffer with 1% Triton. Cdc28 was precipitated using Suc1 beads (Upstate Biotechnology) at 4°C, washed three times with lysis buffer, and resuspended in 2× loading dye.

ChIP and real-time qPCR.

Yeast cells were collected at each designated time point after release from G1 arrest and into medium containing hydroxyurea and/or nocodazole. The cells were fixed with 1% formaldehyde. Chromatin immunoprecipitation (ChIP) analysis was performed as described previously (16) using anti-Myc antibodies (9E10; Santa Cruz Biotechnology). DNA was analyzed by real-time quantitative PCR (qPCR) using the Roche LightCycler 480 system, and the primer pairs were designed to detect Mcd1/Scc1-Myc occupancy at CEN16, TRP1, and LYS4 loci according to the genome-wide mapping of Mcd1/Scc1 (17). The primer pairs used were as follows: TRP1, 5′ TGACGAAGAGGATCTTTCCTG, 3′ CCAGCTCCACCAACTCGT; LYS4, 5′ AATGGAATTTCGCTTGTGGA, 3′ TGCCGTGCTGGGTAAGAT; CEN16, 5′ TGAGCAAACAATTTGAACAGAAA, 3′ CGCTTTAGAACCGCTACCAT.

RESULTS

Sensitivity to replication stress upon deletion of B-type cyclins.

Pds1-null yeast are viable and can maintain cohesion during S phase, indicating the existence of a second mechanism that promotes cohesion. Budding yeast synthetic genetic arrays with pds1-null mutants have revealed important information about Pds1 functions but have not identified a pathway that acts redundantly with Pds1 (36). Cyclin B1 has been implicated in inhibition of anaphase onset in mammals, but it is not known whether B-type cyclins promote cohesion in S phase. Although viable, yeast pds1-null mutants are sensitive to replication inhibitors such as hydroxyurea (HU) and lose cohesion in S phase under conditions of replication stress. To ask if B-type cyclins might participate in cohesion, we surveyed HU sensitivity in single and double deletion mutants. None of the single B-type cyclin deletions were HU sensitive, with the exception of mildly sensitive clb5Δ cells (Fig. 1A) (10). Double deletions, including the clb5Δ mutants, had similar mild sensitivities, but clb5Δ clb6Δ mutants alone were markedly sensitive to HU (Fig. 1B). This indicated a specific S-phase function shared by Clb5 and Clb6, needed for a faithful S phase under the condition of replication stress.

Fig. 1.
Genetic interactions among B-type cyclins result in hydroxyurea sensitivity. Spotting assays indicating yeast growth on rich medium or medium containing hydroxyurea at 30°C. From top to bottom of each panel, 5,000, 500, or 50 cells were plated ...

Some checkpoint mutants are sensitive to HU because of a failure to sustain nucleotide pool levels, a defect that can be corrected by upregulation of RNR1 (15, 45). However, the HU sensitivity of clb5Δ clb6Δ mutants was not suppressed by RNR1 overexpression (Fig. 1C), indicating that an alternative deficiency must account for the lethality during perturbed replication.

Premature loss of cohesion in clb5Δ clb6Δ mutants in HU.

A cohesion defect in S phase could explain the HU sensitivity of clb5Δ clb6Δ cells. To determine if clb5Δ clb6Δ mutants have a cohesion defect, strains containing a LacO(TRP1)/GFP-LacI combination were used to assay cohesion (39). Its proximity to ARS1 ensures that the TRP1 locus is replicated in early S phase, but cohesion keeps these loci close together (seen as a single green fluorescent dot; Fig. 2A, left) until shortly before anaphase (when the dots separate and then rapidly segregate; Fig. 2A, middle and right). Premature loss of cohesion is indicated if cells possess separated TRP1 loci before S phase is complete. To ask if clb5Δ clb6Δ mutants can maintain TRP1 cohesion in S phase, the cells were arrested in G1 using mating pheromone and then released into the cell cycle in the presence of 75 mM HU. In the strain background used, doses of HU below 0.3 M slow replication but do not cause an S-phase arrest. Wild-type cells completed DNA replication after ~2.5 h, at which time no cells were seen with two green fluorescent dots (Fig. 2B and C). Thirty minutes later, TRP1 separation was observed in some cells, followed shortly thereafter by anaphase. In contrast, loss of cohesion at TRP1 in clb5Δ clb6Δ cells was observed before the first cells reached G2. Therefore, some cells must have failed to maintain cohesion in S phase. This can account for the sensitivity of clb5Δ clb6Δ mutant cells to HU and implicates these cyclins, presumably in association with Cdc28, in sister chromatid cohesion.

Fig. 2.
Premature loss of cohesion in clb5Δ clb6Δ mutants in hydroxyurea. Analysis of synchronous cell cycles in the presence of hydroxyurea in wild-type, clb5::ARG4 clb6::ADE1, clb5::URA3 clb6::KAN, and pds1-128 strains. (A) Micrographs of wild-type ...

We compared loss of cohesion in the clb5Δ clb6Δ mutant and a pds1 mutant that had been previously studied (Fig. 2D and E) (11, 12). Similarly to the pds1-128 mutant, loss of cohesion in clb5Δ clb6Δ cells did not occur until mid- to late S phase. We also asked if loss of cohesion was due to the lower rate of DNA replication in clb5Δ clb6Δ cells. DNA replication in wild-type cells was decreased to a similar rate by using higher doses of HU. Under these conditions, wild-type cells were capable of maintaining cohesion even when replication was as slow as in clb5Δ clb6Δ cells (Fig. 2D and E). Therefore, the reduced rate of replication per se is not responsible for premature loss of cohesion in clb5Δ clb6Δ cells.

Loss of cohesion in clb5Δ clb6Δ cells occurs before nuclear export of Mcm7 in HU.

In the above experiments, clb5Δ clb6Δ mutants lost cohesion before the cells entered G2 as judged by fluorescence-activated cell sorting (FACS) analysis of DNA content. To examine replication and cohesion simultaneously, we developed a single-cell assay to identify S-phase cells using a GFP-tagged Mcm7. The minichromosome maintenance (MCM) complex is imported into the nucleus at the end of mitosis and remains nuclear until the completion of S phase, after which it is exported into the cytoplasm (32). G1-arrested wild-type cells expressing MCM7-GFP and possessing the LacO(TRP1)/GFP-LacI combination were washed and released into the cell cycle (Fig. 3). In wild type, all of the cells in which Mcm7-GFP remained nuclear had a single fluorescent signal at the TRP1 locus. Loss of cohesion at TRP1 was observed only in cells without nuclear Mcm7-GFP fluorescence, and this category of cells was observed concomitantly with the appearance of anaphase cells (Fig. 3B). Therefore, as expected, loss of cohesion occurs only after S phase, based on observation of Mcm7-GFP nuclear export. To ask if clb5Δ clb6Δ mutants lost cohesion during S phase in HU, we repeated the above experiment and released G1-arrested wild-type and clb5Δ clb6Δ strains into the cell cycle in the presence of 75 mM HU (Fig. 3C). Cells with separated TRP1 loci were categorized into those with nuclear Mcm7-GFP or diffusely localized Mcm7-GFP. We observed that a substantial fraction of clb5Δ clb6Δ cells lost cohesion with nuclear Mcm7-GFP, indicating that these cells had not completed S phase. Thus, the data indicate that premature loss of cohesion in clb5Δ clb6Δ mutants is a consequence of failed cohesion during S phase.

Fig. 3.
A single cell assay demonstrates that clb5Δ clb6Δ cells lose cohesion during DNA replication in the presence of hydroxyurea. (A) Micrographs of wild-type cells expressing Mcm7-GFP and carrying the LacO(TRP1)/GFP-LacI combination. The top ...

clb5Δ clb6Δ deletion strains had previously been shown to have compromised G2 DNA damage checkpoint controls when grown in the presence of the alkylating agent methyl methanesulfonate (MMS) (31). Using nuclear Mcm7-GFP as an S-phase marker, we found that loss of cohesion at TRP1 occurred after the completion of S phase (data not shown), unlike the HU-treated clb5Δ clb6Δ cells, which lost cohesion while the cells were still in S phase. This suggests that depleted nucleotide pools (HU treatment) place a distinct type of stress on replication forks, compromising cohesion in the absence of Clb5/Clb6.

Pds1 phosphorylation by Cdk is not needed for viability upon replication stress.

Cdk activity is required for Pds1 phosphorylation that occurs on three sites during S phase, although the cyclin partner(s) of Cdk needed for these phosphorylations is not known (1). These phosphorylations stabilize the interaction of Pds1 with Esp1, an interaction that promotes maintenance of cohesion (1). Esp1/Pds1 complexes enter the nucleus around the time of late S phase (25), consistent with the timing of loss of cohesion in clb5Δ clb6Δ mutants in HU. A possible role of Clb5 and Clb6 in promoting cohesion is therefore the phosphorylation of Pds1. In this case, a lack of Pds1 modification at these sites should render cells HU sensitive and result in premature loss of cohesion in S phase. We therefore asked if pds1-38 cells in which all three of the Cdc28 phosphorylation sites are converted to alanines (S277A, S292A, and T304A), thus perturbing the Pds1-Esp1 interaction (1), are sensitive to HU. Using HU-sensitive pds1-128 cells as a positive control, the pds1-38 cells surprisingly had a viability indistinguishable from wild-type cells when grown on HU-containing solid medium (Fig. 4A). Thus, the HU sensitivity of clb5Δ clb6Δ cells cannot be due to a failure of Pds1 phosphorylation on these three Cdk sites. In agreement with these data, we did not observe premature loss of cohesion in pds1-38 cells grown in HU (data not shown).

Fig. 4.
clb5Δ clb6Δ pds1Δ-null cells lose cohesion in S phase and become aneuploid. (A) Spotting assay indicating yeast growth at either 26°C or 30°C on rich medium or medium containing hydroxyurea. From left to right, ...

clb5Δ clb6Δ pds1Δ-null cells lose cohesion in S phase and become aneuploid.

That Clb5 and Clb6 might promote cohesion through a different mechanism, independent of Pds1 phosphorylation, predicts that a combined lack of Pds1 and these cyclins would enhance the cohesion defect. We therefore crossed strains containing clb5Δ clb6Δ and pds1Δ to obtain a triple mutant. Tetrad dissection revealed a synthetic interaction among these mutant combinations—viable triple mutant spores were not recovered (Fig. 4B and data not shown). Inviable spores germinated and then divided once or twice, indicating a possible cell division catastrophe (Fig. 4B, lower panel).

To analyze this phenotype further, we produced triple deletion strains in which viability was restored using a GAL1-driven PDS1 gene (GAL1-PDS1{3xHA} clb5Δ clb6Δ pds1Δ). Such strains were viable on medium containing galactose but not on glucose-containing medium, which represses GAL1-driven transcription (although after several days some colonies grew out on the glucose plates, most likely from single cells that had achieved GAL1-PDS1 derepression) (Fig. 4C and data not shown). Since Pds1 is degraded in G1 phase, we could achieve a mating pheromone-arrested population lacking Pds1 protein. After release from G1 in rich medium containing glucose, GAL1-PDS1 clb5Δ clb6Δ pds1Δ cells initiated DNA replication with a delay compared to wild type (Fig. 4D). While replication was complete in wild type cells by ~40 min, it took until ~70 min for the GAL1-PDS1(3XHA) clb5Δ clb6Δ pds1Δ cells to finish bulk replication. Loss of cohesion and anaphase onset began after 50 min in wild-type cells, but in the triple mutant, loss of cohesion was observed before cells reached G2. Thus, these mutant cells were unable to maintain cohesion during S phase even in the absence of replication stress. After several cell cycles in the presence of glucose, GAL1-PDS1{3xHA} clb5Δ clb6Δ pds1Δ cells became aneuploid (Fig. 4E), consistent with loss of cohesion in S phase, resulting in missegregation of chromosomes during anaphase. This synthetic interaction between pds1Δ and clb5Δ clb6Δ indicates that independent mechanisms enforce cohesion, one relying on Pds1 function and the other on an S-phase cyclin-dependent function.

HU sensitivity of cks1 and cdc28 mutants and synthetic lethality with pds1 mutants.

We next investigated if the Clb5/Clb6 function in promoting cohesion depends on Cdc28 kinase. We observed that cdc28-1N mutant cells are sensitive to HU (Fig. 5A). This allele is a temperature-sensitive allele of CDC28 that retains kinase activity at the restrictive temperature but fails to bind to Cks1 and therefore fails to act upon a subset of its substrates (42). We found that like cdc28-1N and cks1Δ mutants were as sensitive to HU as clb5Δ clb6Δ mutants (Fig. 5A). If complexes of Cks1-Cdc28-Clb5 or Cks1-Cdc28-Clb6 contribute to the Pds1-independent function that promotes cohesion in S phase, we would expect that cdc28-1N and cks1 mutants would be synthetically lethal with pds1 mutants. We therefore crossed cdc28-1N and two temperature-sensitive cks1 mutants to the pds1Δ and pds1-128 strains. All of the mutants either were inviable or showed synthetic sickness in combination with the pds1 mutants (Table 2 ). These data support the hypothesis that Cks1-bound Cdc28/Clb5 or Cdc28/Clb6 complexes promote cohesion independently of Pds1.

Fig. 5.
Cdk and Pds1 maintain cohesion in S phase. (A) Spotting assays indicating yeast growth on rich (YPD) medium or medium containing 100 mM hydroxyurea. From left to right, 5,000, 500, or 50 cells were plated per spot for each strain. (B and C) Analysis of ...
Table 2.
Results of crosses between mutant strains

cdc28-1N pds1Δ strains show premature loss of cohesion.

To ask if cdc28-1N is synthetically lethal with pds1Δ due to failed cohesion during S phase, we constructed a cdc28-1N pds1Δ strain kept alive with an analog-sensitive CDC28 allele, cdc28-as1 (43). Inhibition of Cdc28-as1 kinase by the ATP analog 1-NM-PP1 has been previously characterized (43). We monitored cohesion at TRP1 in the presence or absence of 1-NM-PP1 after synchronizing cells in G1 (Fig. 5B and C). After releasing the cells from the G1 arrest, the strains were shifted to 30°C, at which temperature Cdc28-1N is severely defective for binding Cks1 (6). We found that cdc28-1N cdc28-as1 pds1Δ cells lose cohesion during S phase in the presence of 1-NM-PP1 but not in the absence of inhibitor, thereby suggesting that Cdc28 functions to promote cohesion during S phase and that the interaction with Cks1 is required for this function.

Clb5/Clb6 are required for the association of Esp1 with Cdc28/cyclin complexes.

In human cells, Cdk1/cyclin B1/hsCks1 complexes bind and inhibit Esp1/separase (18). To determine if Cks1/Cdc28/cyclin complexes bind to Esp1, we precipitated Cdc28 using Suc1 Sepharose beads. (Suc1 is the highly conserved Schizosaccharomyces pombe homolog of Cks1.) Precipitations were performed with extracts from wild-type and clb5Δ clb6Δ strains that expressed a Myc epitope-tagged allele of ESP1. The reaction mixtures were then run on SDS-PAGE gels and Western blotted. The blots were probed with anti-Myc antibodies to detect Esp1 and PSTAIRE antibodies to detect Cdc28. We found that in wild-type strains, Esp1 and Cdc28 coprecipitated with Suc1 beads but did not precipitate with protein G-Sepharose (PGS) beads (Fig. 5D). This result suggests that Esp1 can bind to Cdc28 complexes. We were unable to precipitate Esp1 using extract from the clb5Δ clb6Δ strain, indicating that Esp1 may not bind directly to Suc1 beads or Cdc28 associated with the beads and that the interaction is likely to depend directly on Clb5 or Clb6. In human cells, less than 5% of separase protein was estimated to be associated with Cdk1/cyclin B1/hsCks1 (18). Similarly, the corresponding complexes in yeast were not efficiently copurified. Nevertheless, these data leave open the possibility that Esp1 may be directly inhibited by Cdk-cyclin complexes in yeast.

Esp1 is required for loss of cohesion in HU-treated clb5Δ clb6Δ cells.

The data described above indicate that B-type cyclins Clb5 and Clb6 together with Cdc28 promote cohesion during S phase in parallel with Pds1. If the cohesion defect in clb5Δ clb6Δ cells is due to misregulation of cohesin complexes, the defect ought to be suppressed in the absence of Esp1. We constructed clb5Δ clb6Δ mutant strains with a heat-inducible degron allele of the esp1-2 temperature-sensitive allele of ESP1 (esp1-2td). Additionally, the strain carried a galactose-inducible UBR1 gene to enhance degradation of Esp1-2td at the restrictive temperature (37°C). To determine if loss of Esp1 prevented the premature loss of cohesion at the TRP1 locus, wild-type, esp1-2td, clb5Δ clb6Δ, and esp1-2td clb5Δ clb6Δ strains were synchronized in G1 with mating pheromone. G1-arrested cells were washed and released into medium containing 75 mM HU at 37°C (Fig. 6A and B). As previously observed, clb5Δ clb6Δ strains displayed premature loss of cohesion in S phase. esp1-2td strains maintained cohesion in S phase, as would be predicted. Furthermore, esp1-2td clb5Δ clb6Δ strains maintained cohesion during S phase. Therefore, the data indicate that the cohesion function of Clb5 and Clb6 is mechanistically directed at regulation of the cohesin complex.

Fig. 6.
Esp1 is required for loss of cohesion in clb5Δ clb6Δ cells. (A and B) Analysis of synchronous cell cycles in the presence of 75 mM hydroxyurea in wild-type, clb5Δ clb6Δ, esp1-2td, and clb5Δ clb6Δ esp1-2 ...

Because Esp1 cleaves the Mcd1/Scc1 subunit of cohesin to allow sister chromatid separation at anaphase, we predicted that premature Mcd1/Scc1 cleavage would occur in clb5Δ clb6Δ cells. Surprisingly, under normal growth conditions, the cleaved forms of Mcd1/Scc1 were less abundant than in wild-type cells (Fig. 6C). Furthermore, we were not able to detect the cleaved forms in clb5Δ clb6Δ cells under conditions of replication stress (Fig. 6D). One possible explanation of these data is a more rapid degradation of the cleavage products. Alternatively, Clb5 and Clb6 might not regulate cohesin via its cleavage.

The cohesion defect in clb5Δ clb6Δ cells treated with HU is observed cytologically only in the presence of spindle tension.

The suppression of loss of cohesion in esp1-2td clb5Δ clb6Δ indicates that TRP1 locus separation is mechanistically due to misregulation of the cohesin complex. However, an additional factor that could promote premature loss of cohesion is prolonged spindle forces that might pull the TRP1 locus apart. A recent study by Renshaw et al. used the LacO/GFP-LacI reporter system to observe the centromere, the HIS3 locus, and a telomere on the same chromosome, and their data showed that unlike centromeric regions, which separate at the onset of anaphase, chromosome arms do not separate until midanaphase upon further spindle extension (34). They showed that complete cohesin removal requires sufficient spindle pulling force in addition to Esp1 activity. We therefore asked if attachment of chromosomes to the spindle is required for TRP1 separation. Wild-type and clb5Δ clb6Δ strains were arrested in G1, washed, and then released into rich medium with 50 mM HU and in the presence or absence of nocodazole. Separation of the TRP1 locus was monitored through the cell cycle as described for Fig. 2. In the presence of nocodazole, wild-type and clb5Δ clb6Δ cells did not show any loss of cohesion at TRP1 (Fig. 7A). This result suggested that spindle tension is required for loss of cohesion to be observed in clb5Δ clb6Δ cells.

Fig. 7.
TRP1 locus separation in clb5Δ clb6Δ mutant requires spindle tension. (A) Separation of the TRP1 locus depends on spindle tension in clb5Δ clb6Δ mutants. Analysis of synchronous cell cycles in the presence of 50 mM hydroxyurea ...

Considering these data, then we asked if clb5Δ clb6Δ mutants prematurely elongate their spindles in HU, coordinately with loss of cohesion. We synchronized clb5Δ clb6Δ and wild-type cells, expressing GFP-TUB1, in G1 and then released them into the cell cycle in the presence of 100 mM HU. Wild-type cells reached G2 ~3.5 h after release from G1, and spindle elongation took place ~30 min later (Fig. 7B). In clb5Δ clb6Δ mutants, like wild-type cells, spindle elongation was delayed until bulk DNA replication was complete (Fig. 7B). Therefore, we found no evidence that the cohesion defect in clb5Δ clb6Δ cells is due to premature spindle elongation.

We next examined cohesion at loci unaffected by spindle tension: LYS4 and a locus adjacent to TEL4. Using the conditions in Fig. 2, we observed no loss of cohesion in wild-type cells or in clb5Δ clb6Δ cells during replication in the presence of HU (Fig. 8). We also noted that unlike TRP1, these loci separated coincident with anaphase in wild-type cells. These data are therefore consistent with the findings of Renshaw et al. that anaphase spindle forces are required to fully remove cohesin and separate chromosome arm loci (34). Our failure to observe premature loss of cohesion at LYS4 and TEL4 loci, as well as at the TRP1 locus in the presence of nocodazole, in clb5Δ clb6Δ cells is presumably due to the lack of preanaphase spindle-pulling forces.

Fig. 8.
Timing of loss of cohesion at the LYS4 and TEL4 loci in hydroxyurea. Analysis of synchronous cell cycles in the presence of 75 mM hydroxyurea in wild-type and clb5Δ clb6Δ strains. After release from G1, samples were taken for scoring loss ...

Reduced abundance of chromosomal Mcd1/Scc1-Myc in clb5Δ clb6Δ cells treated with HU.

The above experiments showed that preanaphase spindle tension is required for premature TRP1 locus separation in clb5Δ clb6Δ cells. Importantly, however, wild-type cells were able to endure prolonged spindle tension under the same conditions of replication stress (Fig. 2D): the TRP1 locus did not separate prematurely and the cells were resistant to HU, indicating faithful chromosome segregation. Together these data indicated that a bona fide cohesion defect in clb5Δ clb6Δ cells is compounded by prolonged spindle tension to result in premature TRP1 locus separation. To test this hypothesis directly, we sought to compare the abundance of cohesin on chromatin in the wild-type and clb5Δ clb6Δ cells using chromatin immunoprecipitates of Mcd1/Scc1-Myc and real-time quantitative PCR (qPCR).

First, we determined that Mcd1/Scc1-Myc was present at equivalent levels in wild-type and clb5Δ clb6Δ cells and could be immunoprecipitated with similar efficiencies (Fig. 9A). In order to establish the efficiency with which we could detect chromatin-associated cohesin, we used extracts from cells arrested in G1 with mating pheromone or in mitosis with nocodazole. The rationale for this was that cohesin had been previously shown to become enriched at pericentric regions by de novo loading when cells are arrested in mitosis with microtubule-depolymerizing drugs (14). The mechanism of de novo loading appears to be different from that which establishes cohesion concurrent with DNA replication (14, 41, 46). Therefore, we predicated that clb5Δ clb6Δ cells would be competent to load cohesin at pericentric regions during mitotic arrest. After arrest in G1, or upon arrest in the presence of nocodazole, samples were taken to compare wild-type and clb5Δ clb6Δ cells, using ChIP and real-time qPCR. The abundance of Mcd1/Scc1-Myc near a centromere locus (CEN16) was comparable in the two strains, and as described previously (46), cohesin was enriched in the mitosis-arrested cells (Fig. 9B). These data therefore establish that the de novo loading of cohesin upon mitotic arrest occurs in both wild-type and clb5Δ clb6Δ cells.

Fig. 9.
Relative abundance of Mcd1/Scc1-Myc at CEN16, TRP1, and LYS4 during S phase in hydroxyurea or after mitotic arrest in nocodazole. (A) Western blotting of Mcd1/Scc1-Myc in wild-type and clb5Δ clb6Δ cells in whole-cell extracts (WCE) or ...

We then prepared extracts from cells progressing synchronously through S phase in the presence of HU (using the same conditions as for the cytological assays in Fig. 2). We compared samples in which wild-type and clb5Δ clb6Δ cells were in early S, mid-S, and late S, as well as postmitotic cells, to ask how abundant Mcd1/Scc1-Myc was at TRP1 and LYS4 (Fig. 9C). In wild-type cells, as previously described (4, 5), the abundance of Mcd1/Scc1-Myc at TRP1 and LYS4 peaked in early/mid-S phase and dropped as the cell population completed bulk replication. However, in clb5Δ clb6Δ cells, Mcd1/Scc1-Myc did not peak in abundance at either locus. Therefore, regardless of the presence (TRP1) or absence (LYS4) of spindle tension, clb5Δ clb6Δ cells have reduced levels of cohesin on chromatin under the conditions of replication stress that result from reduced nucleotide pool levels. Together with the cytological assays for locus separation, these data indicate that clb5Δ clb6Δ cells have a cohesion defect at TRP1 and LYS4 and that premature separation of the replicated loci requires spindle tension applied at that locus. The combination of spindle tension during a prolonged S phase and reduced cohesin abundance on chromatin presumably leads to TRP1 separation and the aneuploidy observed in clb5Δ clb6Δ cells following replication stress.

DISCUSSION

Identification of the anaphase inhibitor Pds1/securin led to a greater understanding of how chromosome cohesion is maintained until anaphase in eukaryotes (47, 48). However, the viability of pds1-null yeast and securin-null mammals (23, 24, 47, 48) indicated that cohesion is maintained in S phase independently of Pds1. Pds1 inhibits the protease Esp1, which cleaves the Mcd1/Scc1 component of cohesin. Maintenance of cohesion in S-phase pds1Δ cells might be a fortuitous result of the reduced Esp1 activity (8) and compromised nuclear import of Esp1 (25) when Pds1 is not present. However, loss of cohesion occurs during S phase in pds1-null cells under conditions of replication stress. Therefore, Pds1-independent mechanisms are required for cohesion when replication is perturbed.

Here we have described evidence that, in budding yeast, the S-phase cyclins Clb5 and Clb6 are important regulators of cohesion. A cohesion defect in clb5Δ clb6Δ cells was observed cytologically (separation of the newly replicated TRP1 loci) and biochemically (reduced abundance of Mcd1/Scc1-Myc at TRP1 and LYS4). Under conditions of replication stress (in HU), either a lack of Pds1 or a lack of Clb5 and Clb6 results in a cohesion defect. Thus, neither of these factors is sufficient when replication fork progression is perturbed. When both Pds1 and the cyclins are absent, cells lose cohesion in S phase even in the absence of the replication stress induced by HU. The consequences are aneuploidy and cell death. We also find that Cdc28-dependent phosphorylation of Pds1 is not required for cohesion in S phase. These data indicate that the function of Clb5 and Clb6 in promoting cohesion does not depend on Pds1 but rather constitutes a distinct mechanism.

Interestingly, we did not observe premature loss of cohesion at the TRP1 locus in clb5Δ clb6Δ cells treated with HU and nocodazole. We also failed to observe premature loss of cohesion at the LYS4 locus and near the telomere of chromosome IV in clb5Δ clb6Δ cells treated with HU alone. These experiments raised the possibility that prolonged spindle forces during S phase might pull the centromeric regions further apart, giving rise to the observed separation of the TRP1 loci. Counter to this possibility, in clb5Δ clb6Δ mutants treated with HU, the mitotic spindle did not elongate until after S phase was completed, indicating that anaphase spindle forces are not prematurely initiated. More importantly, we observed comparably reduced abundances of chromosomal Mcd1/Scc1-Myc at TRP1 and LYS4 in clb5Δ clb6Δ cells treated with HU. Spindle forces are therefore not required for the cohesin deficiency in clb5Δ clb6Δ cells because LYS4 is not subject to tension during S phase. Presumably, the lack of TRP1 separation in clb5Δ clb6Δ cells treated with nocodazole and HU demonstrates that this cytological assay requires spindle force in order for the cohesion defect to be observed. The reduced abundance of Mcd1/Scc1-Myc on chromosomes in S-phase clb5Δ clb6Δ cells is intriguing given a recent study which demonstrated that reduced levels of cohesin on replicating DNA slow down the replication forks (19). This suggests that the prolonged DNA replication observed in clb5Δ clb6Δ mutants might be a consequence of reduced chromosomal cohesin loading.

We also find that loss of cohesion does not occur in clb5Δ clb6Δ esp1-2td cells. However, we did not observe premature Mcd1/Scc1 cleavage in clb5Δ clb6Δ cells, indicating that the cohesion defect is not due to premature Esp1 activation. Indeed, the Mcd1/Scc1 cleavage products were less abundant in clb5Δ clb6Δ cells compared to wild-type cells. Mcd1/Scc1 cleavage occurs preferentially on the chromatin-associated pool of cohesin complexes (22, 27). The data are therefore consistent with our observation that Mcd1/Scc1 is less abundant on chromatin during S phase in clb5Δ clb6Δ cells than in wild-type cells. Together the experiments suggest that the cohesin defect in clb5Δ clb6Δ cells, and the reduced level of cleavage products, is due to a cohesin loading defect in early S phase. The cytological assay, observation of TRP1 locus separation, indicated a cohesion defect in the latter half of S phase. Since loss of cohesion required Esp1 and spindle tension, a model to explain these data is that the low level of chromosomal Mcd1/Scc1-Myc in clb5Δ clb6Δ cells is sufficient for cohesion until later in S phase when Esp1 activity is increased and spindle forces contribute to pulling the locus apart. However, we cannot rule out the possibility that S-phase cyclins also act directly on Esp1, given that we were able to coprecipitate Esp1 in complexes with Cdc28 and Cks1.

In summary, we describe evidence that specific yeast S-phase cyclins promote loading of cohesin complexes onto chromatin in early S phase (Fig. 9D). In an unperturbed S phase, this mechanism acts in parallel with Pds1 to ensure cohesion is sufficient to hold sister chromatids together. When replication fork progression is stressed, both efficient cohesin loading (Clb5 and Clb6 function) and protection (Pds1 function) are required for cohesion.

ACKNOWLEDGMENTS

We thank Steve Haase for strains and helpful discussion of the data, Orna Cohen-Fix for the pds1-38 mutant, Kevan Shokat for the cdc28-as1 allele and the ATP analog 1-NM-PP1, Joachim Li for the MCM7-GFP plasmids, and Frank Uhlmann for the esp1-2td strain. We are grateful to Judy Berman for help with qPCR. We also thank the anonymous reviewers of the manuscript for their insightful comments.

This work was supported by NIH grant CA099033 (D.J.C.) and NSF grant MCB-0842157 (D.J.C.).

Footnotes

[down-pointing small open triangle]Published ahead of print on 25 April 2011.

REFERENCES

1. Agarwal R., Cohen-Fix O. 2002. Phosphorylation of the mitotic regulator Pds1/securin by Cdc28 is required for efficient nuclear localization of Esp1/separase. Genes Dev. 16: 1371–1382 [PubMed]
2. Alexandru G., Uhlmann F., Mechtler K., Poupart M. A., Nasmyth K. 2001. Phosphorylation of the cohesin subunit Scc1 by Polo/Cdc5 kinase regulates sister chromatid separation in yeast. Cell 105: 459–472 [PubMed]
3. Alexandru G., Zachariae W., Schleiffer A., Nasmyth K. 1999. Sister chromatid separation and chromosome re-duplication are regulated by different mechanisms in response to spindle damage. EMBO J. 18: 2707–2721 [PubMed]
4. Beckouet F., et al. 2010. An Smc3 acetylation cycle is essential for establishment of sister chromatid cohesion. Mol. Cell 39: 689–699 [PubMed]
5. Blat Y., Kleckner N. 1999. Cohesins bind to preferential sites along yeast chromosome III, with differential regulation along arms versus the centric region. Cell 98: 249–259 [PubMed]
6. Bourne Y., et al. 2000. Crystal structure and mutational analysis of the Saccharomyces cerevisiae cell cycle regulatory protein Cks1: implications for domain swapping, anion binding and protein interactions. Structure 8: 841–850 [PubMed]
7. Chang D. C., Xu N., Luo K. Q. 2003. Degradation of cyclin B is required for the onset of anaphase in mammalian cells. J. Biol. Chem. 278: 37865–37873 [PubMed]
8. Ciosk R., et al. 1998. An ESP1/PDS1 complex regulates loss of sister chromatid cohesion at the metaphase to anaphase transition in yeast. Cell 93: 1067–1076 [PubMed]
9. Clarke D. J., Gimenez-Abian J. F. 2000. Checkpoints controlling mitosis. Bioessays 22: 351–363 [PubMed]
10. Clarke D. J., et al. 2003. S-phase checkpoint controls mitosis via an APC-independent Cdc20p function. Nat. Cell Biol. 21: 21 [PubMed]
11. Clarke D. J., Segal M., Jensen S., Reed S. I. 2001. Mec1p regulates Pds1p levels in S phase: complex coordination of DNA replication and mitosis. Nat. Cell Biol. 3: 619–627 [PubMed]
12. Clarke D. J., Segal M., Mondesert G., Reed S. I. 1999. The Pds1 anaphase inhibitor and Mec1 kinase define distinct checkpoints coupling S phase with mitosis in budding yeast. Curr. Biol. 9: 365–368 [PubMed]
13. Donaldson A. D., et al. 1998. CLB5-dependent activation of late replication origins in S. cerevisiae. Mol. Cell 2: 173–182 [PubMed]
14. Eckert C. A., Gravdahl D. J., Megee P. C. 2007. The enhancement of pericentromeric cohesin association by conserved kinetochore components promotes high-fidelity chromosome segregation and is sensitive to microtubule-based tension. Genes Dev. 21: 278–291 [PubMed]
15. Elledge S. J., Zhou Z., Allen J. B., Navas T. A. 1993. DNA damage and cell cycle regulation of ribonucleotide reductase. Bioessays 15: 333–339 [PubMed]
16. Ezhkova E., Tansey W. P. 2006. Chromatin immunoprecipitation to study protein-DNA interactions in budding yeast. Methods Mol. Biol. 313: 225–244 [PubMed]
17. Glynn E. F., et al. 2004. Genome-wide mapping of the cohesin complex in the yeast Saccharomyces cerevisiae. PLoS Biol. 2: E259. [PMC free article] [PubMed]
18. Gorr I. H., Boos D., Stemmann O. 2005. Mutual inhibition of separase and cdk1 by two-step complex formation. Mol. Cell 19: 135–141 [PubMed]
19. Guillou E., et al. 2010. Cohesin organizes chromatin loops at DNA replication factories. Genes Dev. 24: 2812–2822 [PubMed]
20. Haase S. B., Reed S. I. 2002. Improved flow cytometric analysis of the budding yeast cell cycle. Cell Cycle 1: 132–136 [PubMed]
21. Hagting A., et al. 2002. Human securin proteolysis is controlled by the spindle checkpoint and reveals when the APC/C switches from activation by Cdc20 to Cdh1. J. Cell Biol. 157: 1125–1137 [PMC free article] [PubMed]
22. Hornig N. C., Uhlmann F. 2004. Preferential cleavage of chromatin-bound cohesin after targeted phosphorylation by Polo-like kinase. EMBO J. 23: 3144–3153 [PubMed]
23. Huang X., Hatcher R., York J. P., Zhang P. 2005. Securin and separase phosphorylation act redundantly to maintain sister chromatid cohesion in mammalian cells. Mol. Biol. Cell 10: 4725–4732 [PMC free article] [PubMed]
24. Jallepalli P. V., et al. 2001. Securin is required for chromosomal stability in human cells. Cell 105: 445–457 [PubMed]
25. Jensen S., Segal M., Clarke D. J., Reed S. I. 2001. A novel role of the budding yeast separin Esp1 in anaphase spindle elongation: evidence that proper spindle association of Esp1 is regulated by Pds1. J. Cell Biol. 152: 27–40 [PMC free article] [PubMed]
26. King R. W., Deshaies R. J., Peters J. M., Kirschner M. W. 1996. How proteolysis drives the cell cycle. Science 274: 1652–1659 [PubMed]
27. Kucej M., Zou H. 2010. DNA-dependent cohesin cleavage by separase. Nucleus 1: 4–7 [PMC free article] [PubMed]
28. Kuntzel H., Schulz A., Ehbrecht I. M. 1996. Cell cycle control and initiation of DNA replication in Saccharomyces cerevisiae. Biol. Chem. 377: 481–487 [PubMed]
29. Mei J., Huang X., Zhang P. 2001. Securin is not required for cellular viability, but is required for normal growth of mouse embryonic fibroblasts. Curr. Biol. 11: 1197–1201 [PubMed]
30. Mendenhall M. D., Hodge A. E. 1998. Regulation of Cdc28 cyclin-dependent protein kinase activity during the cell cycle of the yeast Saccharomyces cerevisiae. Microbiol. Mol. Biol. Rev. 62: 1191–1243 [PMC free article] [PubMed]
31. Meyn M. A., III, Holloway S. L. 2000. S-phase cyclins are required for a stable arrest at metaphase. Curr. Biol. 10: 1599–1602 [PubMed]
32. Nguyen V. Q., Co C., Irie K., Li J. J. 2000. Clb/Cdc28 kinases promote nuclear export of the replication initiator proteins Mcm2-7. Curr. Biol. 10: 195–205 [PubMed]
33. Queralt E., Lehane C., Novak B., Uhlmann F. 2006. Downregulation of PP2A(Cdc55) phosphatase by separase initiates mitotic exit in budding yeast. Cell 125: 719–732 [PubMed]
34. Renshaw M. J., et al. 2010. Condensins promote chromosome recoiling during early anaphase to complete sister chromatid separation. Dev. Cell 19: 232–244 [PMC free article] [PubMed]
35. Richardson H. E., Wittenberg C., Cross F. R., Reed S. I. 1989. An essential G1 function for cyclin-like proteins in yeast. Cell 59: 1127–1133 [PubMed]
36. Sarin S., et al. 2004. Uncovering novel cell cycle players through the inactivation of securin in budding yeast. Genetics 168: 1763–1771 [PubMed]
37. Segal M., et al. 2000. Coordinated spindle assembly and orientation requires Clb5p-dependent kinase in budding yeast. J. Cell Biol. 148: 441–452 [PMC free article] [PubMed]
38. Stemmann O., Zou H., Gerber S. A., Gygi S. P., Kirschner M. W. 2001. Dual inhibition of sister chromatid separation at metaphase. Cell 107: 715–726 [PubMed]
39. Straight A. F., Belmont A. S., Robinett C. C., Murray A. W. 1996. GFP tagging of budding yeast chromosomes reveals that protein-protein interactions can mediate sister chromatid cohesion. Curr. Biol. 6: 1599–1608 [PubMed]
40. Straight A. F., Marshall W. F., Sedat J. W., Murray A. W. 1997. Mitosis in living budding yeast: anaphase A but no metaphase plate. Science 277: 574–578 [PubMed]
41. Strom L., Lindroos H. B., Shirahige K., Sjogren C. 2004. Postreplicative recruitment of cohesin to double-strand breaks is required for DNA repair. Mol. Cell 16: 1003–1015 [PubMed]
42. Surana U., et al. 1991. The role of CDC28 and cyclins during mitosis in the budding yeast S. cerevisiae. Cell 65: 145–161 [PubMed]
43. Ubersax J. A., et al. 2003. Targets of the cyclin-dependent kinase Cdk1. Nature 425: 859–864 [PubMed]
44. Uhlmann F., Lottspeich F., Nasmyth K. 1999. Sister-chromatid separation at anaphase onset is promoted by cleavage of the cohesin subunit Scc1. Nature 400: 37–42 [PubMed]
45. Vallen E. A., Cross F. R. 1999. Interaction between the MEC1-dependent DNA synthesis checkpoint and G1 cyclin function in Saccharomyces cerevisiae. Genetics 151: 459–471 [PubMed]
46. Weber S. A., et al. 2004. The kinetochore is an enhancer of pericentric cohesin binding. PLoS Biol. 2: E260. [PMC free article] [PubMed]
47. Yamamoto A., Guacci V., Koshland D. 1996. Pds1p is required for faithful execution of anaphase in the yeast, Saccharomyces cerevisiae. J. Cell Biol. 133: 85–97 [PMC free article] [PubMed]
48. Yamamoto A., Guacci V., Koshland D. 1996. Pds1p, an inhibitor of anaphase in budding yeast, plays a critical role in the APC and checkpoint pathway(s). J. Cell Biol. 133: 99–110 [PMC free article] [PubMed]
49. Zou H., McGarry T. J., Bernal T., Kirschner M. W. 1999. Identification of a vertebrate sister-chromatid separation inhibitor involved in transformation and tumorigenesis. Science 285: 418–422 [PubMed]

Articles from Molecular and Cellular Biology are provided here courtesy of American Society for Microbiology (ASM)