MCB
Home Help [Feedback] [For Subscribers] [Archive] [Search] [Contents]
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Other Versions of this Article:
MCB.00774-07v1
27/19/6948    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Schwartz, D. C.
Right arrow Articles by Hochstrasser, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Schwartz, D. C.
Right arrow Articles by Hochstrasser, M.

 Previous Article  |  Next Article 

Molecular and Cellular Biology, October 2007, p. 6948-6961, Vol. 27, No. 19
0270-7306/07/$08.00+0     doi:10.1128/MCB.00774-07
Copyright © 2007, American Society for Microbiology. All Rights Reserved.

The Ulp2 SUMO Protease Is Required for Cell Division following Termination of the DNA Damage Checkpoint{triangledown}

David C. Schwartz,1,{dagger} Rachael Felberbaum,2 and Mark Hochstrasser1,2*

Departments of Molecular Biophysics & Biochemistry,1 Molecular, Cell, & Developmental Biology, Yale University, 266 Whitney Avenue, New Haven, Connecticut 06520-81142

Received 2 May 2007/ Returned for modification 15 June 2007/ Accepted 18 July 2007


   ABSTRACT
Top
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Eukaryotic genome integrity is maintained via a DNA damage checkpoint that recognizes DNA damage and halts the cell cycle at metaphase, allowing time for repair. Checkpoint signaling is eventually terminated so that the cell cycle can resume. How cells restart cell division following checkpoint termination is poorly understood. Here we show that the SUMO protease Ulp2 is required for resumption of cell division following DNA damage-induced arrest in Saccharomyces cerevisiae, although it is not required for DNA double-strand break repair. The Rad53 branch of the checkpoint pathway generates a signal countered by Ulp2 activity following DNA damage. Interestingly, unlike previously characterized adaptation mutants, ulp2{Delta} mutants do not show persistent Rad53 phosphorylation following DNA damage, suggesting checkpoint signaling has been terminated and no longer asserts an arrest in these cells. Using Cdc14 localization as a cell cycle indicator, we show that nearly half of cells lacking Ulp2 can escape a checkpoint-induced metaphase arrest despite their inability to divide again. Moreover, half of permanently arrested ulp2{Delta} cells show evidence of an aberrant mitotic spindle, suggesting that Ulp2 is required for proper spindle dynamics during cell cycle resumption following a DNA damage-induced cell cycle arrest.


   INTRODUCTION
Top
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
All eukaryotic cells must recognize and respond to various types of DNA damage and DNA replication defects that threaten genomic integrity. The signal transduction pathways that respond to these DNA lesions are often called DNA damage checkpoints (31). A common feature of these pathways is a transient block to cell division that allows time to repair the lesions. In addition to cell cycle arrest, the DNA damage checkpoint pathway leads to transcriptional up-regulation of genes encoding DNA repair proteins. The signal transduction pathway leading to a DNA damage checkpoint response has been extensively investigated, but the mechanism by which a cell resumes division after checkpoint arrest remains poorly understood.

Following DNA damage recognition, the initial damage is processed, leading to recruitment of DNA checkpoint and repair proteins (30). In the budding yeast Saccharomyces cerevisiae, recruitment of the Mec1 kinase (ATR in humans) initiates a pair of signaling cascades (see Fig. 1A) (31). In one, phosphorylation of the Rad53 kinase leads ultimately to transcriptional induction, via the Dun1 kinase, of DNA repair and replication proteins. In the other pathway, phosphorylation of the Chk1 kinase and to some extent also Rad53 leads to stabilization of the yeast securin, Pds1, resulting in a preanaphase mitotic arrest.


Figure 1
View larger version (19K):
[in this window]
[in a new window]

  		FIG. 1. Ulp2 is required for cell division following DNA damage. (A) Outline of the DNA damage checkpoint pathway. (B) The ulp2{Delta} mutant is sensitive to DNA-damaging agents. (C) The ulp2{Delta} mutant is sensitive to very low levels of HU. (D) The ulp2{Delta} mutant is sensitive to transient exposure to high HU levels. (E) ulp2{Delta} cells arrest division following DNA damage from a single HO endonuclease-derived DSB. Fivefold serial dilutions of yeast cultures were spotted onto plates for panels B to D. Cells were photographed using Nomarski optics, and all plates were grown at 24°C for 4 days.

 
After checkpoint-mediated cell cycle arrest, repair of the DNA damage leads to a process termed recovery, in which the cell terminates checkpoint signaling and resumes cell division. Alternatively, instances in which a cell is unable to repair the DNA damage lead to a process known as adaptation (34). Adaptation allows the resumption of cell division in the continued presence of DNA damage. Like recovery, adaptation requires that the checkpoint-signaling cascade be turned off to proceed with the cell cycle. One key step is thought to be dephosphorylation of Rad53, since this event correlates with the timing of adaptation and recovery and fails to occur in known adaptation/recovery mutants (6, 18-20, 32, 41). Finally, the cell must restart its cell cycle. The transition to anaphase is marked by the activation of multiple signaling pathways, including the Cdc fourteen early anaphase release (FEAR) and mitotic exit network (MEN) pathways (36). Both of these pathways result in the release of the Cdc14 phosphatase from the nucleolus throughout the mother and bud cells. The released Cdc14 promotes mitotic progression in multiple ways, including reversal of CDK phosphorylation and regulation of mitotic spindle dynamics.

The molecular mechanisms underlying the reversal of DNA damage-induced cell cycle arrest are poorly understood. Several proteins participate in adaptation to DNA damage. Srs2, Tid1, and Sae2 appear to help remove specific repair or checkpoint proteins from DNA (6, 18, 41). Removal of these complexes is necessary for adaptation and presumably also occurs following repair. The Ptc2 and Ptc3 phosphatases dephosphorylate Rad53, and their activity against Rad53 correlates with the ability of cells to adapt to DNA damage (20). Cdc5 is a polo-like kinase required for mitotic exit and has a role in adaptation that appears linked to this function (14, 40). Last, Yku70-Yku80 binds to DNA breaks and slows rates of DNA resection; in its absence, cells accumulate excessive single-stranded DNA, leading to permanent arrest (17). Persistent Rad53 phosphorylation correlates with the adaptation defect of cells with mutations in any of these proteins. This implies that DNA checkpoint signaling remains activated in these cells, presumably maintaining them in the arrested state. No mutants that permanently arrest in response to DNA damage with deactivated checkpoint signaling are currently known.

We have identified Ulp2 (Smt4) as a protein that is required for cell division following DNA damage-induced cell cycle arrest. Ulp2 is one of two known yeast SUMO proteases, which remove the ubiquitin-like covalent modifier SUMO (Smt3 in yeast) from substrate proteins (22). As is true for ubiquitin, SUMO attachment to specific substrate proteins requires a series of sequential enzyme reactions (15, 35). The C terminus of SUMO is first activated by SUMO-activating enzyme (E1), which forms an intermediate with SUMO and then transfers the SUMO to SUMO-conjugating enzyme (E2). Subsequently, E2 transfers SUMO to a substrate lysine residue, usually with the aid of a SUMO protein ligase (E3). Formation of such isopeptide (amide)-linked SUMO conjugates can alter the localization, stability, or protein-protein interactions of the substrate protein (11).

The two yeast SUMO isopeptidases, Ulp1 and Ulp2, have distinct substrate specificities inferred from the phenotypic differences of the respective mutants and their different patterns of accumulated sumoylated proteins (22). Ulp2 localizes to the nucleus and is associated with chromatin (22, 38). Deletion of the ULP2 gene results in a pleiotropic phenotype: slow growth, sensitivity to high temperature and to both DNA and spindle damage, sporulation and germination defects, and high rates of chromosome and plasmid loss (22). Here we show that Ulp2 is required for cell division following DNA damage, although it is not required for DNA double-strand break (DSB) repair. Following induction of a nonrepairable DSB, ulp2{Delta} mutants permanently arrest in a Mec1- and Rad53-dependent manner, indicating that Ulp2 activity is required for resumption of cell division after DNA checkpoint arrest. Interestingly, Ulp2 does not affect the dephosphorylation of Rad53, implying that checkpoint signaling has been terminated and no longer asserts an arrest in ulp2{Delta} cells despite their inability to resume cell division. Closer examination reveals that ulp2{Delta} mutants, like wild-type cells, initially arrest at metaphase upon DSB induction and either remain permanently arrested in metaphase or progress past metaphase before terminally halting division. In addition, approximately half of ulp2{Delta} cells show evidence of an aberrant mitotic spindle after arresting their cell cycle in response to DNA damage. We propose that one or more proteins is sumoylated following DNA damage-induced checkpoint activation, and this substrate(s) must be desumoylated by Ulp2 for proper cell cycle resumption.


   MATERIALS AND METHODS
Top
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Yeast strains and plasmids. Yeast strains used are listed in Table 1. Strains were grown in standard media. All Gal-HO microcolony assay strains were derived from RMY169 (gift from R. Michelson [28]). The Gal-HO SSA strains were derived from YMV2 (gift from J. Haber [41]). All deletions were created using EUROSCARF gene deletion plasmids (7-10) and confirmed using PCR analysis. The plasmid pRad53-FLAG, used for integrative FLAG-Rad53 epitope tagging, was obtained from R. Michelson. Matched ULP2 and ulp2-H531A plasmids were obtained from A. Strunnikov (38). The TUB1-GFP allele was introduced into the URA3 locus using the pAFS125 integrating plasmid (gift from K. Tatchell) (37) cut with StuI. A PCR-based strategy was used to introduce a green fluorescent protein (GFP) tag at the C terminus of the CDC14 chromosomal locus (42). When necessary, plasmids were maintained by growth of yeast in selective medium.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Yeast strains used in this study

 
SSA repair assay. Cells were grown overnight in either yeast extract-peptone-dextrose (YPD) or raffinose-containing medium at 24°C. On the following day, genomic DNA from 2 optical-density-at-600-nm (OD600) cell equivalents grown in YPD was prepared. Cells were resuspended in equal parts buffer (1% sodium dodecyl sulfate, 2% Triton X-100, 100 mM NaCl, 10 mM Tris [pH 8.0], 1 mM EDTA) and phenol:chloroform:isoamyl alcohol (25:24:1) and lysed by shearing with glass beads for 2 min. This was followed by chloroform extraction and ethanol precipitation. The DNA pellet was resuspended in 200 µl Tris-EDTA buffer. On the same day, 2 OD600 cell equivalents grown in raffinose were transferred to galactose-containing medium and incubated overnight at 24°C. The following day, genomic DNA from 2 OD600 equivalents of cells was prepared according to the protocol described above. PCR analysis was performed using 1 µl genomic DNA and the following primers: P1 (GCCAGGTGACCACGTTGGTCAAG), P2 (GGATGATGCATTAGCCCATTCTTCC), P3 (GCTACATATAAGGAACGTGCTGCTAC), Control Fwd (GGTCGTTGGCCCTCAACAGCATTGG), and Control Rev (GGCAAAGCGTTTGTATACATAAGAAATG).

Gal-HO microcolony assays. Cells were grown overnight in raffinose-containing medium at 24°C. On the following day, 1 OD600 equivalent of cells was transferred to galactose-containing medium (~1:500 dilution) and incubated for 1 h at 24°C. Cells were micromanipulated onto complete plates containing galactose and examined microscopically for cell division over the next 4 days. For Cdc14-GFP or Tub1-GFP visualization, a coverslip was placed on top of the plate and cells were viewed directly with a Zeiss Axioskop microscope (Carl Zeiss MicroImaging, Inc., Thornwood, NY) with a Plan-Apochromat 100x NA1.4 objective lens at room temperature using Fluka immersion oil 10976 (Sigma). For GFP, the XF100-2 filter set was used (Omega Optical, Brattleboro, VT). Pictures were taken on a Zeiss Axiocam camera using a Uniblitz shutter driver (model VMM-D1; Vincent Associates, Rochester, NY) and the program Open Lab 3.1.5 (Improvision, Lexington, MA). Note that all microcolony assays were performed at 24°C rather than the more commonly used 30°C because of improved growth of ulp2{Delta} strains at the lower temperature.

HU microcolony assay. Cells were grown overnight in YPD at 24°C. The next day, 1 OD600 equivalent of cells was resuspended in either 2 ml YPD plus 0.2 M hydroxyurea (HU) or water and incubated for 8 h at 24°C. Cells were micromanipulated onto YPD plates and examined microscopically for cell division over the next 2 days.

Rad53 phosphorylation. For assaying Flag-Rad53 phosphorylation, cells were grown in raffinose-containing medium until mid-log phase; 2% galactose was added to the medium for 4 h to induce DNA damage and Rad53 phosphorylation. The cells were harvested and resuspended in 2% dextrose-containing medium to allow adaptation. Approximately 1 OD600 equivalent of cells was taken for each time point, and proteins were prepared using a trichloroacetic acid precipitation technique (24). To detect Flag-Rad53, a monoclonal Flag antibody (Sigma) was used at a 1:5,000 dilution followed by a goat antimouse secondary antibody (1:5,000; Roche).


   RESULTS
Top
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Response of ulp2{Delta} cells to DNA damage and replication delay. Loss of Ulp2 confers sensitivity to many different DNA-damaging and replication-delaying agents, including methyl methanesulfonate, 4-nitroquinoline 1-oxide, UV light, and HU (Fig. 1B) (22), suggesting that this SUMO protease is required either for the DNA damage checkpoint or for DNA repair. To help distinguish between these possibilities, we measured the sensitivity of the ulp2{Delta} mutant to low levels of HU. HU depletes the deoxynucleoside triphosphate (dNTP) pools available for DNA replication. Low HU levels cause mild retardation of DNA replication and infrequent DSBs, whereas prolonged high HU levels cause a strong block to replication and many DSBs (27). Most DNA checkpoint mutants are sensitive even to low HU levels due to continued replication in the absence of sufficient dNTPs. In contrast, cells defective for DNA repair usually survive low HU levels because they still have an intact DNA checkpoint, allowing them time to replenish dNTPs; however, they will die in high levels of HU due to their inability to repair the DNA damage. Upon exposure to low HU levels (10 mM), the ulp2{Delta} mutant failed to grow into colonies, as was observed for a mec1{Delta} DNA checkpoint mutant (Fig. 1C). (Note that in all strains containing either mec1{Delta} or rad53{Delta}, SML1 has also been deleted to suppress the lethality of the single deletions [43]). In contrast, deletion of the DNA repair gene RAD52 caused no growth defect at this HU concentration, reflecting the absence of significant numbers of DSBs.

To confirm this result, cells were transiently exposed to high HU levels (0.2 M). The 8-h exposure was shorter than that required to produce significant DSBs. Cells with an intact DNA checkpoint survive this treatment, whereas checkpoint mutants often die (2). As shown in Fig. 1D, wild-type (WT) cells and the rad52{Delta} DNA repair mutant were resistant to transient HU exposure, but the mec1{Delta} DNA checkpoint mutant was highly sensitive. Although not as sensitive as mec1{Delta}, ulp2{Delta} cells also grew relatively poorly, suggesting that Ulp2 function is linked to the DNA checkpoint pathway and not DNA repair. However, unlike most checkpoint mutants, which fail to arrest following DNA damage, ulp2{Delta} cells arrested normally as large-budded cells (Fig. 1E). This result is consistent with previous data (22) and indicates that Ulp2 is most likely affecting cell cycle restart following checkpoint arrest (see below).

Ulp2 is not required for DNA DSB repair. In the presence of HU, ulp2{Delta} more closely resembled a DNA checkpoint mutant (mec1{Delta}) than a DNA repair mutant (rad52{Delta}). However, the possibility remains that it plays a role in both pathways. To analyze further whether Ulp2 is involved in DNA repair, we utilized a single-strand annealing (SSA)-based repair assay that has previously been described (41) (Fig. 2A). An HO endonuclease cleavage site was inserted into the LEU2 locus, and the normal HO sites in the mating loci were deleted. Upon induction with galactose, the HO endonuclease is expressed and cleaves leu2, creating a DSB. This is followed by extensive 5'-to-3' DNA resection, which activates the DNA damage checkpoint and eventually reveals a region of homology to leu2 inserted 30 kb away. SSA repair occurs when this homology is used to anneal the two DNA strands, resulting in a deletion of the intervening DNA.


Figure 2
View larger version (18K):
[in this window]
[in a new window]

  		FIG. 2. Ulp2 is not required for DNA repair by SSA. (A) Diagram of SSA repair assay. Galactose-induced HO cuts at a cleavage site inserted in leu2. 5'-to-3' resection eventually reaches homologous DNA 30 kb away (depicted as a shaded rectangle). Repair occurs in a Rad52-dependent manner and results in a deletion of the intervening DNA. (B) ulp2{Delta} cells can repair an HO-induced DSB by SSA, but rad52{Delta} cells cannot. Control cells were grown overnight in YPD. Cells induced with galactose were first grown overnight in yeast extract-peptone (YEP) raffinose and then transferred to YEP galactose medium for a second night. All samples were grown at 24°C. Genomic DNA was used as a template for PCR analysis. P1, P2, and P3 are primers used for PCR (annealing sites are depicted in panel A). (C) ulp2{Delta} cells have a growth defect in response to a repairable DSB. Tenfold serial dilutions of yeast cultures were spotted onto glucose or galactose plates and grown for 4 days at 24°C.

 
Similar to WT cells, ulp2{Delta} cells were able to repair the HO-induced DSB, as indicated by PCR analysis (Fig. 2B). Primers P2 and P3 anneal to DNA that is ~30 kb apart in undamaged cells, but SSA repair brings these regions within 3 kb of each other. Thus, a 3-kb PCR product is generated only upon successful completion of DNA repair. In contrast, Rad52 is required for SSA and rad52{Delta} cells do not produce a repair product. In all cases, a DSB was created in response to galactose induction as indicated by the loss of PCR product from primers P1 and P2, located on either side of the HO cleavage site (Fig. 2B). The ulp2{Delta} cells had a growth defect despite their ability to complete DSB repair (Fig. 2C), again suggesting that Ulp2 function is linked to the DNA checkpoint pathway and not DNA repair. This result also agrees with previous reports of ulp2{Delta} sensitivity to repairable DNA damage and implies that ulp2{Delta} is defective not only in resuming cell division after adaptation to damage but also following recovery (3, 22; D. C. Schwartz and M. Hochstrasser, unpublished data).

Ulp2 is required for division following a nonrepairable DSB. We hypothesized that Ulp2 is required either for proper maintenance of DNA checkpoint arrest or for resumption of the cell cycle following such an arrest. As noted earlier, yeast cells can adapt to persistent DNA damage and resume cell division, allowing them to attempt repair in a subsequent cell cycle, although without repair most will ultimately die. To determine whether the ulp2{Delta} mutant is capable of resuming cell division following nonrepairable DNA damage, we utilized a galactose-inducible HO endonuclease to generate a single DSB on the left arm of chromosome VII (28). Because this is the only copy of the HO-cut site and there are no homologous sequences in the genome in this assay, this creates a DSB that cannot be repaired by homologous recombination, the major mechanism of DSB repair in yeast, or by SSA, as was the case for the repair assay described above.

Upon induction of the HO endonuclease and subsequent DNA cleavage, a WT cell will arrest in metaphase for approximately 24 h at 24°C, at which point it undergoes adaptation and resumes cell division. Cells will divide in the presence of the DSB until they die from the loss of essential genetic information on chromosome VII. As shown in Fig. 3A, 82% of WT cells were capable of resuming cell division, defined as a cell that has divided and formed a microcolony that contains four or more cells. When ulp2{Delta} cells suffered the same DSB, they arrested as large-budded cells (counted as "two cells") with similar kinetics, indicating the presence of an intact DNA checkpoint response (Fig. 1E). Further evaluation confirmed that like WT cells, ulp2{Delta} cells also initially arrested in metaphase (see Fig. 8). However, unlike WT cells, only 8% of ulp2{Delta} cells resumed cell division, and most of those that did went through only a single division (Fig. 3A). Those that did not divide either remained in metaphase or rearrested in late mitosis, as shown below. This suggested that ulp2{Delta} cells were defective in their ability to resume cell division following DNA damage-induced cell-cycle arrest, similar to the behavior of known adaptation mutants (for example, the srs2{Delta} mutant) (41) (Fig. 7B). By contrast, cells lacking the Rad51 or Rad52 repair protein adapted to the DSB and grew similarly to WT cells (Fig. 3B), confirming that an inability to repair DNA damage does not lead to permanent cell cycle arrest.


Figure 3
View larger version (26K):
[in this window]
[in a new window]

  		FIG. 3. Ulp2 is required for cell division following induction of a nonrepairable DSB. Cells were grown in liquid medium containing galactose for 1 h to induce the HO endonuclease and then micromanipulated onto plates containing galactose. Black bars represent the percentage of cells in each microcolony size category after 24 h with DNA damage, and the gray bars represent the percentage of cells in each category at the end of the experiment (72 h). The percentage arrested after 24 h is noted at left, and the percentage noted at right is the percentage of microcolonies containing >3 cells after 72 h. (A) Following transient cell cycle arrest, WT cells adapt to DNA damage and undergo multiple cell divisions (see inset; 72 h). The ulp2{Delta} cells permanently arrest as large-budded cells following damage. (B) rad51{Delta} and rad52{Delta} cells adapt to and divide after DNA damage. (C) mec1{Delta} and mec1{Delta} ulp2{Delta} cells show defective arrest following DNA damage. Note that in the double mutant, which grows very poorly, a large fraction likely fails to divide further as a result of this general growth defect.

 

Figure 8
View larger version (23K):
[in this window]
[in a new window]

  		FIG. 8. ulp2{Delta} cells can escape DNA damage-induced metaphase arrest. Cells were grown at 24°C in liquid medium containing galactose for 1 h to induce HO and then micromanipulated onto plates containing galactose. Cdc14-GFP was observed by fluorescence microscopy after either 1 or 2 days (d) (1 day = 20 to 31 h; 2 days = 44.5 to 49.5 h). (A) Representative arrested WT, ulp2{Delta}, or srs2{Delta} cells 1 or 2 days after galactose induction. WT cells divide after 1 day. Arrows indicate Cdc14-GFP. (B) Forty-two percent of ulp2{Delta} cells (and 23% of srs2{Delta} cells) show Cdc14 in both cells by 2 days. Pictures were scored in groups of 40, and 120 cells of each genotype were counted. Error bars represent standard deviations.

 

Figure 7
View larger version (23K):
[in this window]
[in a new window]

  		FIG. 7. Rad53 is dephosphorylated in ulp2{Delta} cells as in WT cells, indicating termination of DNA checkpoint signaling. (A) WT and ulp2{Delta} strains show similar Rad53 phosphorylation and dephosphorylation kinetics following nonrepairable DNA damage. Cells were incubated with galactose for 4 h to induce DNA damage and then placed into glucose-containing medium (times shown are from the addition of galactose). (B) Ptc2 overexpression does not suppress ulp2{Delta} cell cycle arrest, in contrast to the case with srs2{Delta} cells. Bars are as in Fig. 3.

 
To determine whether loss of Ulp2 was preventing cell division in response to DNA damage checkpoint arrest or was halting cell division or killing cells for some other reason, ulp2{Delta} was combined with mec1{Delta}, which completely abolishes DNA damage checkpoint activation. When mec1{Delta} cells experience a DSB, they continue to divide as if there were no damage. This phenotype was best observed 24 h after break induction. Whereas 91% of WT cells had arrested at the two- to three-cell stage (Fig. 3A), 80% of mec1{Delta} cells continued to divide into microcolonies of four or more cells (Fig. 3C). If Ulp2 was required for cell division following a Mec1-mediated arrest, then the inability of ulp2{Delta} cells to resume division should be overcome by deleting MEC1. Indeed, after 72 h of persistent DNA damage, mec1{Delta} ulp2{Delta} cells behaved similarly to mec1{Delta} cells, with 56% progressing beyond the two- to three-cell stage compared to only 8% of the ulp2{Delta} single mutant cells. These data indicate that activation of the DNA damage checkpoint is required for permanent ulp2{Delta} cell cycle arrest.

Two further controls were performed to show that the permanent arrest seen with ulp2{Delta} was dependent on the damage derived from the DSB. First, if the ulp2{Delta} cells lacked the HO endonuclease cut site, 57% formed microcolonies of four or more cells, as opposed to only 8% of congenic ulp2{Delta} cells with the HO cut site (Fig. 4A). Second, if the HO enzyme was not induced in cells bearing the HO cut site, 71% of ulp2{Delta} cells formed microcolonies (Fig. 4B). These results show that the permanent arrest of ulp2{Delta} cells depended on the DSB.


Figure 4
View larger version (16K):
[in this window]
[in a new window]

  		FIG. 4. ulp2{Delta} cells without DNA damage were able to form viable colonies. Bars for panels A and B are as in Fig. 3. (A) ulp2{Delta} strains lacking an HO cut site. (B) ulp2{Delta} cells with an HO cut site that were not induced for HO expression with galactose. (C) ulp2{Delta} cells treated for 8 h with either HU (0.2 M) or water can form microcolonies.

 
Finally, to determine whether the permanent arrest seen with ulp2{Delta} was specific to a DSB-induced checkpoint arrest or was a more general response to arresting the cell cycle at any stage, ulp2{Delta} cells were treated for 8 h with either HU (0.2 M) or water. By 38 h, 39% of HU-treated cells formed microcolonies of four or more cells, compared to 43% of water-treated cells (Fig. 4C). (Note that a significant portion of ulp2{Delta} cells never passed two to three cells, but this was most likely due to micromanipulation conditions and did not differ significantly upon HU or water treatment.) Although ulp2{Delta} cells are mildly sensitive to transient HU treatment, they do not permanently arrest as they do in response to a DSB.

Ulp2 is required downstream of the Rad53 branch of the checkpoint pathway. Next, we determined which components of the Mec1 signaling pathway were required for the permanent DSB-induced ulp2{Delta} arrest. Multiple proteins are involved in sensing DNA damage and transmitting the checkpoint signal. Upon Mec1 activation, Mec1 phosphorylates both Rad53, which leads ultimately to transcription of DNA repair genes and has a primary role in preventing mitotic exit, and Chk1, which leads to a preanaphase cell cycle arrest (Fig. 1A). We therefore asked if ulp2{Delta} permanent arrest is dependent on the Rad53 or Chk1 branch of the checkpoint pathway.

In the Gal-HO microcolony assay, rad53{Delta} behaved similarly to mec1{Delta}. By 24 h after DNA damage induction, when WT cells had arrested, 52% of the rad53{Delta} cells progressed beyond the two- to three-cell stage, indicating a defect in checkpoint arrest (Fig. 5A). A strain lacking both Rad53 and Ulp2 had a similar phenotype: after a 24-h exposure to a DSB, 51% of the cells were past the two- to three-cell stage, and by 72 h, 60% of the rad53{Delta} ulp2{Delta} cells had formed microcolonies. Therefore, ulp2{Delta} arrest is dependent on Rad53.


Figure 5
View larger version (25K):
[in this window]
[in a new window]

  		FIG. 5. Ulp2 is required downstream of the Rad53 branch of the checkpoint pathway for cell division following nonrepairable DNA damage. Bars are as in Fig. 3. (A) Both rad53{Delta} and rad53{Delta} ulp2{Delta} cells show impaired arrest and undergo multiple divisions in the presence of DNA damage. (B) The chk1{Delta} mutant shows impaired arrest, but the chk1{Delta} ulp2{Delta} mutant is permanently arrested following DNA damage. (C) The dun1{Delta} and dun1{Delta} ulp2{Delta} mutants continue to divide in the presence of DNA damage. (D) The pds1{Delta} and pds1{Delta} ulp2{Delta} mutants show impaired arrest following DNA damage.

 
Similar to both rad53{Delta} and mec1{Delta} cells, chk1{Delta} cells were unable to arrest properly following DNA damage; 24 h after HO induction, 40% of chk1{Delta} cells had gone beyond the two- to three-cell stage (Fig. 5B). In contrast to mec1{Delta} ulp2{Delta} and rad53{Delta} ulp2{Delta} cells, however, 96% of the chk1{Delta} ulp2{Delta} cells were large budded (i.e., two cells) at 24 h, and only 11% ever resumed cell division. This indicates that persistent ulp2{Delta} arrest does not require Chk1. These data are consistent with Ulp2 functioning downstream of the Rad53 branch, but not the Chk1 branch, of the checkpoint pathway.

Dun1 and Pds1 function downstream of Rad53 and Chk1, respectively (Fig. 1A). Ulp2 function was also analyzed in cells lacking these proteins. Both dun1{Delta} and pds1{Delta} mutants fail to arrest cell division in response to DNA damage. In the Gal-HO microcolony assay, 77% of dun1{Delta} cells had divided beyond the two- to three-cell stage by 24 h (Fig. 5C), and 56% of pds1{Delta} cells were past the two- to three-cell stage (Fig. 5D), showing the absence of a proper DNA checkpoint in both cases. As predicted from the position of Dun1 downstream of Rad53, dun1{Delta} ulp2{Delta} cells no longer displayed the terminal arrest seen in ulp2{Delta} single mutants; eventually 46% of the double mutant cells divided beyond the two- to three-cell stage (Fig. 5C). (After 24 h, 79% of the cells were still at the two- to three-cell stage, suggestive of an arrest, but this was most likely due to the much slower growth of the dun1{Delta} ulp2{Delta} strain [not shown]. Slow growth was also seen with pds1{Delta} ulp2{Delta} cells, described below.) Unexpectedly, Pds1 was also required for the ulp2{Delta} arrest: 55% of the pds1{Delta} ulp2{Delta} cells eventually divided beyond the two- to three-cell stage (Fig. 5D). This was surprising, because the chk1{Delta} ulp2{Delta} strain retained the ulp2{Delta} arrest in response to DNA damage, and Pds1 is phosphorylated directly by Chk1. One explanation for this result is that Pds1 may be regulated by both Chk1 and Rad53. In this scenario, Pds1 would be required for Ulp2 function, but checkpoint signaling to Pds1 would originate in the Rad53 branch of the checkpoint pathway (see Discussion).

Ulp2 SUMO protease activity is required for cell division following DNA damage. Ulp2 is a member of a subclass of cysteine proteases and is a SUMO-specific protease (22, 38). All members of this family require a conserved histidine in the catalytic site. To determine if Ulp2 desumoylation activity was required for the resumption of cell division following nonrepairable DNA damage, a version of Ulp2 lacking the catalytic histidine, ulp2-H531A, was used in the Gal-HO microcolony assay (38). Transformation of ulp2{Delta} cells with a WT ULP2 plasmid suppressed the ulp2{Delta} permanent arrest phenotype, with 68% of the cells resuming cell division following DNA damage (Fig. 6A). In contrast, only 11% of ulp2{Delta} cells harboring the ulp2-H531A plasmid got past the two- to three-cell stage. Neither WT ULP2 nor the ulp2-H531A plasmid showed any effect on arrest or cell cycle resumption in a WT background (not shown). These results demonstrate that Ulp2 SUMO protease activity is required for cell division following nonrepairable DNA damage and suggest that a substrate(s) must be desumoylated by Ulp2 for the cell cycle to resume.


Figure 6
View larger version (17K):
[in this window]
[in a new window]

  		FIG. 6. SUMO protease activity of Ulp2 is required for cell division following DNA damage, but only when sumoylation is unimpaired. Bars are as in Fig. 3. (A) A WT copy of ULP2 on a plasmid rescues ulp2{Delta} permanent arrest, whereas a plasmid-borne ulp2-H531A catalytically inactive allele fails to do so. (B) Sumoylation is not required for DNA damage checkpoint arrest or the resumption of cell division following arrest. Cells were shifted to 37°C, the restrictive temperature for ubc9-1 cells, for 4 h. Galactose was added to the medium to induce HO, and cells were incubated at 37°C for 1 h longer and then manipulated onto galactose plates and incubated at 24°C for 4 days. Both WT and ubc9-1 strains arrest following DNA damage and subsequently adapt and divide. (C) The siz1{Delta} siz2{Delta} strain is able to arrest following damage and then adapt and divide (experiment done at 24°C). (D) The permanent arrest of ulp2{Delta} cells following DNA damage is suppressed by ubc9-1.

 
Sumoylation is not required for DNA checkpoint arrest or cell cycle resumption. The fact that the desumoylating enzyme Ulp2 was required for cell division following a DNA checkpoint arrest raised the question of whether protein sumoylation might be required for the initial arrest, although this need not be the case. Ubc9 is the sole SUMO-conjugating enzyme (E2), so we first tested mutations in this protein for effects on cell cycle arrest following DNA damage or subsequent cell cycle restart. Ubc9 is essential for viability, so the microcolony experiments were done with the ubc9-1 temperature-sensitive allele. At restrictive temperature, the mutant ubc9-1 protein is rapidly degraded and sumoylation is virtually eliminated (4). The Gal-HO microcolony experiments were performed two different ways: ubc9-1 inactivation followed by DSB induction or HO cleavage followed by ubc9-1 inactivation. Both methods gave similar results, so only the former is presented. After a temperature shift and DSB induction, 93% of WT cells were able to resume cell division (four or more cells in the microcolony) (Fig. 6B). The time points used in this assay were similar to those in the previous assays, but since both the duration of DNA checkpoint arrest and the cell cycle period decreased at 37°C, the majority of the cells (76%) had already adapted and divided by 24 h, even though they had initially arrested. Cells with the ubc9-1 mutation were DNA checkpoint proficient, since 49% of the cells remained at the two- to three-cell stage after 24 h, which is an even greater percentage than in the WT strain. The mutant cells also could resume division in the presence of persistent DNA damage (73% division versus 94% in WT), although a minor role in this process cannot be entirely discounted.

Loss of the Siz1 and Siz2 SUMO E3s blocks the majority of SUMO conjugation in yeast (16). Use of a siz1{Delta} siz2{Delta} double mutant in the microcolony assay eliminated the requirement for the temperature manipulations used with ubc9-1. Loss of Siz1/Siz2-dependent SUMO conjugation failed to cause any defect in either arrest following DNA damage (96% arrest) or cell division following the checkpoint arrest (75% division) (Fig. 6C). Although residual SUMO conjugation in the siz1{Delta} siz2{Delta} cells might be enough for DNA damage checkpoint activation (or for cell cycle resumption), the combined results for the siz1{Delta} siz2{Delta} and ubc9-1 mutants suggest that SUMO conjugation is unlikely to be necessary for checkpoint arrest.

Because SUMO ligation has little or no effect on checkpoint activation or resumption of cell division following checkpoint-induced arrest, we asked if reducing SUMO ligation could suppress the permanent ulp2{Delta} arrest. When the ulp2{Delta} strain was shifted to high temperature for 4 h, followed by induction of a nonrepairable DSB, the response of the mutant was similar to its response at 24°C: cells arrested as large-budded cells, and only 20% got past this arrested state by the end of the experiment (Fig. 6D). In contrast, 50% of the ubc9-1 ulp2{Delta} double mutant cells had resumed cell division by 72 h. This implies that a sumoylated substrate(s) acts as a signal for maintaining arrest or blocking cell division following arrest. If sumoylation is prevented, this block is removed and the requirement for Ulp2-mediated desumoylation is bypassed.

DNA checkpoint signaling is terminated in ulp2{Delta} cells. Termination of the checkpoint-signaling cascade is essential for cell division to resume following DNA damage-induced arrest. Current evidence suggests that Rad53 is a key target for checkpoint termination; in order for cells to resume their cell cycle, Rad53 must be dephosphorylated (20). Two phosphatases have been identified, Ptc2 and Ptc3, which specifically dephosphorylate Rad53. In mutants lacking either or both of these phosphatases, cell cycle resumption from checkpoint arrest is strongly impaired. We used the Gal-HO microcolony assay strains to determine whether ulp2{Delta} cells were also impaired in dephosphorylating Rad53 following checkpoint arrest. Creation of a nonrepairable DSB in WT cells by inducing HO for 4 h in galactose led to readily detected Rad53 phosphorylation (Fig. 7A). As expected, the modification required the Mec1 kinase, which directly phosphorylates Rad53. Eight hours after damage induction, Rad53 phosphorylation decreased, and by 24 h, it was largely gone. Deletion of PTC2 strongly inhibited Rad53 dephosphorylation, as expected. In contrast, Rad53 was phosphorylated and dephosphorylated with similar kinetics in WT and ulp2{Delta} strains. This result suggests that checkpoint signaling is terminated in ulp2{Delta} cells, yet they are still unable to resume cell division.

The Srs2 helicase is another protein reported to be involved in cell cycle restart following DNA damage-induced arrest, but unlike ulp2{Delta} cells, srs2{Delta} cells permanently arrest with hyperphosphorylated Rad53 (41). Ptc2 overexpression partially suppressed the srs2{Delta} defect in resuming cell division (Fig. 7B), as was previously reported for yku70{Delta} and cdc5-ad mutants (20), implying that termination of checkpoint signaling was sufficient to promote cell cycle resumption in these mutants. However, Ptc2 overexpression did not suppress the permanent arrest of ulp2{Delta} cells (Fig. 7B), supporting the inference that Ulp2 acts downstream of Rad53 dephosphorylation in the resumption of cell division.

Progression of ulp2{Delta} cells past metaphase following DNA damage-induced arrest. The finding that Rad53 was dephosphorylated in terminally arrested ulp2{Delta} cells prompted us to ask whether these cells were permanently arrested in metaphase, the stage of initial arrest, or whether they resumed the cell cycle before terminally halting prior to cytokinesis. To address this question, we fused a sequence encoding GFP in-frame and downstream of the chromosomal CDC14 gene in the Gal-HO microcolony assay strains. The Cdc14 phosphatase activates both the FEAR and MEN pathways, which are required for mitotic exit (36). Cdc14 is retained in the nucleolus through most of the cell cycle, but when cells enter anaphase, it is released. Cdc14 is thus a useful marker of cell cycle progression beyond metaphase; it is found only in the mother cell at metaphase but is present in both the mother cell and the daughter bud postmetaphase.

As described above, WT cells will remain arrested at metaphase for approximately 1 day at 24°C following induction of a DSB prior to adapting and resuming cell division. Accordingly, 92% of WT cells showed Cdc14-GFP in one cell 1 day following DSB induction (Fig. 8). Similarly, 89% of ulp2{Delta} cells and 93% of srs2{Delta} cells showed Cdc14-GFP in only one cell at 1 day, indicating that like the WT, both mutants had an intact checkpoint and arrested preanaphase. By 2 days following DSB induction, WT cells had adapted and divided. Of the ulp2{Delta} cells that remained as single large-budded cells at 2 days, 42% showed Cdc14-GFP in both the mother cell and daughter bud, suggesting that these cells had progressed past metaphase. The remaining 58% showed Cdc14-GFP in only one cell, indicating a sustained preanaphase arrest. Interestingly, 23% of permanently arrested srs2{Delta} cells also showed Cdc14-GFP in both cells at 2 days, suggesting that this mutant also has a limited ability to break through the initial metaphase arrest.

Mitotic spindle abnormalities in ulp2{Delta} cells following DNA damage-induced arrest. We also directly examined mitotic spindles in ulp2{Delta} cells to determine if any had progressed beyond their initial arrest point. TUB1-GFP was integrated into the URA3 locus of the Gal-HO test strains to follow {alpha}-tubulin distribution in living cells. The cell cycle stage was scored as metaphase/early anaphase if the spindle was short and was at the bud neck or slightly protruding into the bud. Cells were counted as being in telophase if the spindle stretched from one end of the mother to the opposite end of the bud. One day following DSB induction, 97% of WT cells showed a metaphase/early anaphase spindle (Fig. 9). Similarly, 92% of ulp2{Delta} cells and 97% of srs2{Delta} cells also had short spindles at 1 day, demonstrating that they had arrested normally. By 2 days, WT cells had again adapted and resumed cell division. Surprisingly, of the terminally arrested ulp2{Delta} cells, 50% showed signs of a broken or otherwise aberrant spindle at 2 days. These spindles varied in appearance but generally showed Tub1-GFP fluorescence in both cells, with fluorescence signals that did not seem connected even when stepping through different focal planes (Fig. 9A). This spindle phenotype was not simply a response to prolonged cell cycle arrest, since only 18% of spindles appeared abnormal in permanently arrested srs2{Delta} cells. This result suggests that Ulp2 is required for proper spindle function when cells restart the cell cycle after a DNA damage-induced arrest.


Figure 9
View larger version (16K):
[in this window]
[in a new window]

  		FIG. 9. Aberrant spindles in ulp2{Delta} cells following DNA damage-induced arrest. Cells were grown as described in the legend to Fig. 8. Tub1-GFP was observed by fluorescence microscopy after either 1 or 2 days(d) (1 day = 19.5 to 28.5 h; 2 days = 48 to 52 h). (A) Pictures of representative WT, ulp2{Delta}, or srs2{Delta} cells 1 or 2 days after galactose induction. (B) Fifty percent of ulp2{Delta} cells (and 18% of srs2{Delta} cells) show noncontiguous Tub1-GFP fluorescence at 2 days. Pictures were scored in groups of 40, and at least 120 cells of each genotype were counted. Between 1 and 3 focal planes per cell were examined to ensure the entire GFP signal was scored. Error bars represent standard deviations. Cells were scored as being in metaphase/early anaphase if the spindle was short and either at the bud neck or slightly protruded into the bud, in telophase if the spindle stretched from one end of the mother to the opposite end of the bud, or as having an aberrant spindle if the spindle appeared as disconnected segments.

 

   DISCUSSION
Top
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
DNA damage checkpoint and DNA repair mechanisms are needed for cells to deal with genetic insults and ensure accurate transmission of genetic information. How a cell recommences cell division following DNA damage-induced arrest remains a major question. Here we show that the Ulp2 SUMO protease is required for this process following induction of a nonrepairable DSB. Our data suggest that in the presence of a DSB, ulp2{Delta} mutants activate and maintain a normal DNA damage checkpoint arrest and then terminate checkpoint signaling. These mutants then either maintain a sustained preanaphase arrest or progress past metaphase prior to permanent cell cycle arrest. A few proteins have previously been shown to be defective in cell division in the presence of nonrepairable DNA damage (Cdc5, Srs2, Ptc2/3, Sae2, Tid1, Rad51, Yku70-80, and CKII). These mutants, however, show persistent Rad53 phosphorylation, which is characteristic of a defect in adaptation. Normally, dephosphorylation of Rad53 accompanies termination of the DNA damage checkpoint, and persistent Rad53 phosphorylation following DNA damage implies that a checkpoint signal most likely continues to be transmitted in the mutant cells. The data with ulp2{Delta} suggest that it is not adaptation defective per se but rather is defective in the resumption of cell division following turn-off of the checkpoint signal.

Our data indicate that ulp2{Delta} cells are competent for at least one type of DNA DSB repair (Fig. 2B), supporting the inference that the permanent arrest of ulp2{Delta} cells is a response to activation of the DNA damage checkpoint rather than an inability to repair DNA damage (Fig. 1). In order for ulp2{Delta} cells to complete SSA repair, multiple repair genes, such as RAD52, RAD1, and RAD10, must be functional, implying that these proteins are active in the ulp2{Delta} mutant (12, 13). Ulp2-deleted cells also more closely resembled checkpoint mutants than repair mutants in response to HU exposure (Fig. 1C and D), again suggesting that Ulp2 function is linked to the DNA damage checkpoint response rather than to DNA repair. Based on these results, it is likely that ulp2{Delta} cells are generally proficient in DNA repair, including other repair mechanisms, such as homologous recombination, which utilizes at least one of the same repair proteins as SSA, although an additional role in non-SSA DNA repair cannot be ruled out.

Epistasis analysis revealed that the permanent DNA damage-induced arrest of ulp2{Delta} cells requires a functional DNA damage checkpoint. Deletion of the Mec1 kinase abolishes checkpoint signaling and suppressed the ulp2{Delta} arrest (Fig. 3C). A similar suppressive effect is observed when the RAD17 checkpoint gene is deleted from ulp2{Delta} cells (not shown). The ulp2{Delta} arrest is also suppressed if the Rad53 kinase is deleted (Fig. 5A) but not when Chk1 is deleted (Fig. 5B). This finding suggests either that Ulp2 counters a signal generated specifically by the Rad53 branch of the pathway or that ulp2{Delta} cells are sensitive to a minimal threshold of checkpoint signaling that still is reached when Chk1 is deleted. The DNA damage checkpoint has a dual role in inhibiting both anaphase and mitotic exit, and the Rad53 branch is thought to be more important than Chk1 for preventing mitotic exit (23). Therefore, the fact that ulp2{Delta} is epistatic to chk1{Delta} but not rad53{Delta} is consistent with Ulp2 having an important role in cell cycle restart after a block to mitotic exit (imposed by Rad53). Our finding that a large fraction of ulp2{Delta} cells arrest postmetaphase supports this interpretation (Fig. 8).

Unexpectedly, the cell cycle arrest of ulp2{Delta} cells depends on both Pds1 (securin) and Dun1, even though Pds1 is thought to be regulated primarily by the Chk1 kinase (Fig. 5C and D). That Pds1 is required for ulp2{Delta} arrest but Chk1 is not suggests cross talk between the two branches of the checkpoint pathway (Fig. 1A). In fact, both Chk1 and Rad53 were shown to play a role in inhibiting Pds1 ubiquitination, thereby stabilizing Pds1 and preventing anaphase (1, 33).

Our results suggest that one or more proteins become sumoylated following checkpoint activation and in their sumoylated state interfere with the successful completion of mitosis. This model is supported by the results with the ulp2-H531A catalytic-site mutant, which, like ulp2{Delta}, permanently arrests after DNA damage. Thus, the desumoylating activity of Ulp2 is required for resumption of cell division. Also consistent with this model is the ability of ubc9-1, which strongly impairs SUMO conjugation (21), to suppress permanent DNA damage-induced arrest in ulp2{Delta} cells. Our data suggest that protein sumoylation does not play a major role in activating the DNA damage checkpoint. Sumoylation of a substrate(s) following DNA checkpoint activation may serve as a redundant mechanism for ensuring that cells do not proceed in the cell cycle until the DNA is repaired.

Intriguingly, we found that nearly half of DNA-damaged ulp2{Delta} mutants could restart the cell cycle and progress past metaphase before terminally arresting, whereas the remainder maintained a preanaphase arrest (Fig. 8). This heterogeneous phenotype suggests that desumoylation of a substrate(s) may be required at metaphase following DNA checkpoint termination, but in some cases this requirement can be bypassed, leading to the postmetaphase arrest. What causes these cells to subsequently stop dividing? One possibility is that a second desumoylation event is required later in mitosis for the cell to proceed with division. An alternative but not mutually exclusive scenario is that the persistent sumoylation of the metaphase substrate may interfere with subsequent mitotic progression. Along these lines, examination of Tub1-GFP in ulp2{Delta} cells unexpectedly revealed that roughly half the cells have an aberrant mitotic spindle (Fig. 9). In these cells, the Tub1-GFP signal usually appears in both the mother and bud cells without an apparent connection between them. This could reflect either a broken or severely distorted spindle. Normal spindles at all mitotic stages were observed with Tub1-GFP in both WT and mutant cells (Fig. 9; also not shown), indicating that the tagged TUB1 allele is not causing the spindle aberrations. Interestingly, ulp2{Delta} mutants are also sensitive to transient exposure to microtubule-depolymerizing drugs, which might reflect defective spindle assembly after drug withdrawal (22).

What could cause ulp2{Delta} spindles to be compromised to such an extent? One possibility is that a substrate(s) that is sumoylated in response to DNA damage-induced cell cycle arrest plays a role in spindle regulation. In this view, Ulp2-mediated desumoylation following checkpoint termination would allow successful mitotic spindle development. In ulp2{Delta} cells, persistent sumoylation might interfere with this process, causing spindle damage. Ulp2, SUMO, and anaphase-promoting proteins have been linked previously to spindle regulation. Both ULP2 and SMT3 (yeast SUMO) were identified as high-copy-number suppressors of the mif2-3 kinetochore mutant (26). This same mutant also exhibits broken spindles at restrictive temperature (5). Multiple kinetochore proteins are sumoylated in yeast, and preventing sumoylation of one of them, Ndc10, leads to mislocalization from the spindle midzone and abnormally long anaphase spindles (29). Finally, one of the key outcomes of the DNA damage checkpoint pathway is stabilization of the yeast securin Pds1, whose ubiquitin-dependent degradation is required for activation of the yeast separase Esp1. Activated Esp1 cleaves the cohesin complex, allowing sister chromatid separation during anaphase, but plays other roles in mitotic exit as well (39). Interestingly, like mif2-3 cells, mutant esp1-1 cells have morphologically aberrant spindles (25). Thus, both kinetochore proteins and proteins that promote progression through anaphase are potential substrates that might have to be desumoylated by Ulp2 following checkpoint termination.

Besides being sensitive to activation of the DNA damage checkpoint, ulp2{Delta} cells are also sensitive to activation of the replication checkpoint, spindle checkpoint, or an arrest in anaphase-promoting complex mutants (3, 22). Thus, one possibility is that ulp2{Delta} mutants are generally unable to survive a prolonged cell cycle arrest. However, we have shown that ulp2{Delta} cells can divide following an 8-h HU-induced arrest (Fig. 4C), unlike their response to a DSB. Moreover, we have previously shown that ulp2{Delta} mutants can be synchronized in G1 by {alpha}-factor and form microcolonies upon release (22). Although these arrests are for a shorter time period than the 24-h-induced DSB arrest presented here, we have observed a similar permanent arrest phenotype for ulp2{Delta} cells when repairable DNA damage is induced for just 4 h (unpublished data). Consequently, it is unlikely that the ulp2{Delta} permanent arrest that occurs following induction of a DSB is a general outcome of halting the cell cycle at any stage. Despite this finding, the possibility still remains that ulp2{Delta} mutants may be generally unable to tolerate a prolonged metaphase arrest. Previously it was shown that arrested ulp2{Delta} cells show a loss of centromeric cohesion due to the misregulation of DNA topoisomerase II sumoylation, providing one possible reason why ulp2{Delta} cells might not survive a sustained metaphase (3). Expression of an unsumoylatable topoisomerase II allele, top2-SNM, could partially suppress this defect. When we expressed top2-SNM in our Gal-HO microcolony assay, it did not suppress the permanent ulp2{Delta} arrest (not shown), implying that one or more additional substrates contribute to this process. Here we have shown that ulp2{Delta} cells can maintain a metaphase arrest, and the defect in subsequent cell division occurs after termination of the DNA damage checkpoint. Because the different metaphase-inducing checkpoints employ much of the same machinery to halt cell division, it is likely that the inability of ulp2{Delta} cells to desumoylate a substrate(s) common to these pathways is at least partly responsible for the shared phenotype.


   ACKNOWLEDGMENTS
 
We thank R. Michelson, A. Lewis, D. Stern, and J. Ma for comments on the manuscript and W. Hankey and D. Su for technical support. Colleagues who provided strains or plasmids are noted in Materials and Methods.

D.C.S. was supported in part by a Leukemia & Lymphoma Society postdoctoral fellowship and R.F. by NIH training grant GM007223. This work was supported by NIH grant GM053756.


   FOOTNOTES
 
* Corresponding author. Mailing address: Departments of Molecular Biophysics & Biochemistry, Yale University, 266 Whitney Avenue, New Haven, CT 06520-8114. Phone: (203) 432-5101. Fax: (203) 432-5158. E-mail: mark.hochstrasser{at}yale.edu Back

{triangledown} Published ahead of print on 30 July 2007. Back

{dagger} Present address: Bureau of Western Hemisphere Affairs, Department of State, 2201 C Street NW, Washington, DC 20520-2810. Back


   REFERENCES
Top
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
1. Agarwal, R., Z. Tang, H. Yu, and O. Cohen-Fix. 2003. Two distinct pathways for inhibiting Pds1 ubiquitination in response to DNA damage. J. Biol. Chem. 278:45027-45033.[Abstract/Free Full Text]

2. Allen, J. B., Z. Zhou, W. Siede, E. C. Friedberg, and S. J. Elledge. 1994. The Sad1/Rad53 protein kinase controls multiple checkpoints and DNA damage-induced transcription in yeast. Genes Dev. 8:2401-2415.[Abstract/Free Full Text]

3. Bachant, J., A. Alcasabas, Y. Blat, N. Kleckner, and S. J. Elledge. 2002. The SUMO-1 isopeptidase Smt4 is linked to centromeric cohesion through SUMO-1 modification of DNA topoisomerase II. Mol. Cell 9:1169-1182.[CrossRef][Medline]

4. Betting, J., and W. Seufert. 1996. A yeast Ubc9 mutant protein with temperature-sensitive in vivo function is subject to conditional proteolysis by a ubiquitin- and proteasome-dependent pathway. J. Biol. Chem. 271:25790-25796.[Abstract/Free Full Text]

5. Brown, M. T., L. Goetsch, and L. H. Hartwell. 1993. MIF2 is required for mitotic spindle integrity during anaphase spindle elongation in Saccharomyces cerevisiae. J. Cell Biol. 123:387-403.[Abstract/Free Full Text]

6. Clerici, M., D. Mantiero, G. Lucchini, and M. P. Lohghese. 2006. The Saccharomyces cerevisiae Sae2 protein negatively regulates DNA damage checkpoint signalling. EMBO Rep. 7:212-218.[CrossRef][Medline]

7. Goldstein, A. L., and J. H. McCusker. 1999. Three new dominant drug resistance cassettes for gene disruption in Saccharomyces cerevisiae. Yeast 15:1541-1553.[CrossRef][Medline]

8. Goldstein, A. L., X. Pan, and J. H. McCusker. 1999. Heterologous URA3MX cassettes for gene replacement in Saccharomyces cerevisiae. Yeast 15:506-511.

9. Güldener, U., S. Heck, T. Fiedler, J. Beinhauer, and J. H. Hegemann. 1996. A new efficient gene disruption cassette for repeated use in budding yeast. Nucleic Acids Res. 24:2519-2524.[Abstract/Free Full Text]

10. Gueldener, U., J. Heinisch, G. J. Kohler, D. Voss, and J. H. Hegemann. 2002. A second set of loxP marker cassettes for Cre-mediated multiple gene knockouts in budding yeast. Nucleic Acids Res. 30:e23.[Abstract/Free Full Text]

11. Hay, R. T. 2005. SUMO: a history of modification. Mol. Cell 18:1-12.[CrossRef][Medline]

12. Ivanov, E. L., and J. E. Haber. 1995. RAD1 and RAD10, but not other excision repair genes, are required for double-strand break-induced recombination in Saccharomyces cerevisiae. Mol. Cell. Biol. 15:2245-2251.[Abstract]

13. Ivanov, E. L., N. Sugawara, J. Fishman-Lobell, and J. E. Haber. 1996. Genetic requirements for the single-strand annealing pathway of double-strand break repair in Saccharomyces cerevisiae. Genetics 142:693-704.[Abstract]

14. Jin, F., and Y. Wang. 2006. Budding yeast DNA damage adaptation mutants exhibit defects in mitotic exit. Cell Cycle 5:2914-2919.[Medline]

15. Johnson, E. S. 2004. Protein modification by SUMO. Annu. Rev. Biochem. 73:355-382.[CrossRef][Medline]

16. Johnson, E. S., and A. A. Gupta. 2001. An E3-like factor that promotes SUMO conjugation to the yeast septins. Cell 106:735-744.[CrossRef][Medline]

17. Lee, S. E., J. K. Moore, A. Holmes, K. Umezu, R. D. Kolodner, and J. E. Haber. 1998. Saccharomyces Ku70, Mre11/Rad50, and RPA proteins regulate adaptation to G2/M arrest after DNA damage. Cell 94:399-409.[CrossRef][Medline]

18. Lee, S. E., A. Pellicioli, A. Malkova, M. Foiani, and J. E. Haber. 2001. The Saccharomyces recombination protein Tid1p is required for adaptation from G2/M arrest induced by a double-strand break. Curr. Biol. 11:1053-1057.[CrossRef][Medline]

19. Lee, S. E., A. Pellicioli, M. B. Vaze, N. Sugawara, A. Malkova, M. Foiani, and J. E. Haber. 2003. Yeast Rad52 and Rad51 recombination proteins define a second pathway of DNA damage assessment in response to a single double-strand break. Mol. Cell. Biol. 23:8913-8923.[Abstract/Free Full Text]

20. Leroy, C., S. E. Lee, M. B. Vaze, F. Ochsenbien, R. Guerois, J. E. Haber, and M. Marsolier-Kergoat. 2003. PP2C phosphatases Ptc2 and Ptc3 are required for DNA checkpoint inactivation after a double-strand break. Mol. Cell 11:827-835.[CrossRef][Medline]

21. Li, S., and M. Hochstrasser. 1999. A new protease required for cell-cycle progression in yeast. Nature 398:246-251.[CrossRef][Medline]

22. Li, S., and M. Hochstrasser. 2000. The yeast ULP2 (SMT4) gene encodes a novel protease specific for the ubiquitin-like Smt3 protein. Mol. Cell. Biol. 20:2367-2377.[Abstract/Free Full Text]

23. Liang, F., and Y. Wang. 2007. DNA damage checkpoints inhibit mitotic exit by two different mechanisms. Mol. Cell. Biol. 27:5067-5078.[Abstract/Free Full Text]

24. Longhese, M. P., V. Paciotti, R. Fraschini, R. Zaccarini, P. Plevani, and G. Lucchini. 1997. The novel DNA damage checkpoint protein ddc1p is phosphorylated periodically during the cell cycle and in response to DNA damage in budding yeast. EMBO J. 16:5216-5226.[CrossRef][Medline]

25. McGrew, J. T., L. Goetsch, B. Byers, and P. Baum. 1992. Requirement for ESP1 in the nuclear division of Saccharomyces cerevisiae. Mol. Biol. Cell 3:1443-1454.[Abstract]

26. Meluh, P. B., and D. Koshland. 1995. Evidence that the MIF2 gene of Saccharomyces cerevisiae encodes a centromere protein with homology to the mammalian centromere protein CENP-C. Mol. Biol. Cell 6:793-807.[Abstract]

27. Merrill, B. J., and C. Holm. 1999. A requirement for recombinational repair in Saccharomyces cerevisiae is caused by DNA replication defects of mec1 mutants. Genetics 153:595-605.[Abstract/Free Full Text]

28. Michelson, R. J., S. Rosenstein, and T. Weinert. 2005. A telomeric repeat sequence adjacent to a DNA double-stranded break produces an anticheckpoint. Genes Dev. 19:2546-2559.[Abstract/Free Full Text]

29. Montpetit, B., T. R. Hazbun, S. Fields, and P. Hieter. 2006. Sumoylation of the budding yeast kinetochore protein Ndc10 is required for Ndc10 spindle localization and regulation of anaphase spindle elongation. J. Cell Biol. 174:653-663.[Abstract/Free Full Text]

30. Nakada, D., Y. Hirano, and K. Sugimoto. 2004. Requirement of the Mre11 complex and exonuclease 1 for activation of the Mec1 signaling pathway. Mol. Cell. Biol. 24:10016-10025.[Abstract/Free Full Text]

31. Nyberg, K. A., R. J. Michelson, C. W. Putnam, and T. A. Weinert. 2002. Toward maintaining the genome: DNA damage and replication checkpoints. Annu. Rev. Genet. 36:617-656.[CrossRef][Medline]

32. Pellicioli, A., S. E. Lee, C. Lucca, M. Foiani, and J. E. Haber. 2001. Regulation of Saccharomyces Rad53 checkpoint kinase during adaptation from DNA damage-induced G2/M arrest. Mol. Cell 7:293-300.[CrossRef][Medline]

33. Sanchez, Y., J. Bachant, H. Wang, F. Hu, D. Liu, M. Tetzlaff, and S. J. Elledge. 1999. Control of the DNA damage checkpoint by Chk1 and Rad53 protein kinases through distinct mechanisms. Science 286:1166-1171.[Abstract/Free Full Text]

34. Sandell, L. L., and V. A. Zakian. 1993. Loss of a yeast telomere: arrest, recovery, and chromosome loss. Cell 75:729-739.[CrossRef][Medline]

35. Schwartz, D. C., and M. Hochstrasser. 2003. A superfamily of protein tags: ubiquitin, SUMO and related modifiers. Trends Biochem. Sci. 28:321-328.[CrossRef][Medline]

36. Stegmeier, F., and A. Amon. 2004. Closing mitosis: the functions of the Cdc14 phosphatase and its regulation. Annu. Rev. Genet. 38:203-232.[CrossRef][Medline]

37. Straight, A. F., W. F. Marshall, J. W. Sedat, and A. W. Murray. 1997. Mitosis in living budding yeast: anaphase A but no metaphase plate. Science 277:574-578.[Abstract/Free Full Text]

38. Strunnikov, A. V., L. Aravind, and E. V. Koonin. 2001. Saccharomyces cerevisiae SMT4 enc