Report

Recruitment of Mec1 and Ddc1 Checkpoint Proteins to Double-Strand Breaks Through Distinct Mechanisms

See allHide authors and affiliations

Science  26 Oct 2001:
Vol. 294, Issue 5543, pp. 867-870
DOI: 10.1126/science.1063827

Abstract

In response to DNA damage, eukaryotic cells activate checkpoint pathways that arrest cell cycle progression and induce the expression of genes required for DNA repair. In budding yeast, the homothallic switching (HO) endonuclease creates a site-specific double-strand break at the mating type (MAT) locus. Continuous HOexpression results in the phosphorylation of Rad53, which is dependent on products of the ataxia telangiectasia mutated–related MEC1 gene and other checkpoint genes, including DDC1, RAD9, andRAD24. Chromatin immunoprecipitation experiments revealed that the Ddc1 protein associates with a region near the MATlocus after HO expression. Ddc1 association required Rad24 but not Mec1 or Rad9. Mec1 also associated with a region near the cleavage site after HO expression, but this association is independent of Ddc1, Rad9, and Rad24. Thus, Mec1 and Ddc1 are recruited independently to sites of DNA damage, suggesting the existence of two separate mechanisms involved in recognition of DNA damage.

When DNA is damaged, cells activate a response pathway that arrests the cell cycle and induces the transcription of genes that facilitate repair. The failure of this response results in genomic instability that may lead to cancer in higher eukaryotes (1). The systems that monitor the structure of the chromosomes and coordinate repair and cell-cycle progression are termed “checkpoints” in eukaryotic cells. In budding yeast, two essential genes, MEC1 andRAD53, form the core of the DNA damage checkpoint pathway (2, 3). Mec1 belongs to a superfamily of large protein kinases, including human ATM and ATM-Rad3-related (ATR).RAD53 encodes a protein kinase similar to human Chk2 that functions downstream of MEC1 in the response pathway (1, 3). After DNA damage, Rad53 is phosphorylated and activated by a Mec1-dependent mechanism. Phosphorylation and activation of Rad53 is also dependent on RAD9 and the genes of the RAD24 epistasis group, consisting of DDC1, MEC3,RAD17, and RAD24 (2). Homologs ofDDC1, MEC3, RAD17, andRAD24 have been identified in mammalian cells (1).

The checkpoint machinery consists of proteins that recognize DNA damage and initiate the signaling response (1). Such sensory proteins are expected to bind to the aberrant DNA structures and/or the repair machinery localized on damaged DNA. To address this possibility, we created DNA lesions at a specific sequence and tested whether checkpoint proteins bound to that region. In budding yeast, the HO endonuclease generates a single site-specific double-strand break (DSB) in the genome at the MAT locus (4). The HO cleavage site is also present at the HML andHMR loci, but these sites are normally not accessible for cutting. The HO-induced DSB at the MAT locus is usually repaired by recombination with the homologous HML andHMR loci (4). When both HMLand HMR are deleted, the HO-induced DSB becomes irreparable by recombination. This persistent DSB is sufficient to cause cells to arrest at the G2-M phase transition of the cell cycle (5–7). In the presence of bothHML and HMR, continuous expression of the HO endonuclease results in the persistence of cleaved MAT loci and also delays the cell-cycle progression before anaphase (8, 9).

We used an experimental system in which HO endonuclease is continuously expressed to develop an assay for the recruitment of checkpoint proteins. Rad53 is phosphorylated in response to genotoxic stress, and this phosphorylation has been shown to correlate with activation of the DNA damage checkpoint pathway (2). We first analyzed the status of Rad53 phosphorylation in cells in which HO expression was driven from the galactose-inducible GAL10 promoter (10). Cells expressing the influenza hemagglutinin (HA)–tagged Rad53 (Rad53-HA) protein were transformed with the GAL-HO plasmid or empty vector, and were grown initially in sucrose to repress HO expression (11). Cells were then transferred to medium containing galactose to induce HOexpression. Extracts were collected at time intervals and analyzed by Western blot to monitor the modification of Rad53. Rad53 phosphorylation became detectable 4 hours after cells carrying the GAL-HO plasmid were shifted to medium containing galactose (Fig. 1A). DNA blot analysis indicated that cleavage at the MAT locus occurred within 2 hours and persisted 6 hours after shifting to the galactose medium (12, 13). The lag of time between DSB formation and Rad53 phosphorylation raises a possibility that the checkpoint activation might result from certain DSB repair events. The HO-induced DSB is usually repaired by the RAD52-dependent homologous recombination pathway or in some cases by Ku-dependent nonhomologous end-joining pathway (4). However, Rad53 phosphorylation after HO expression was still observed in rad52Δ and Ku (hdf1Δ) mutants (Fig. 1B). The HO-induced DSB activates the checkpoint pathway only when it cannot be efficiently repaired (14). Continuous HO cleavage might induce DNA lesions that could not be repaired efficiently by the recombination and end-joining pathways.

Figure 1

Rad53 phosphorylation after HOexpression. (A) Wild-type cells containing YCpT-RAD53-HA (11) were transformed with pGAL-HO (10) or the YCp50 vector. Transformants were grown in sucrose medium without uracil and tryptophan, and transferred to medium containing 2% galactose. At the indicated time points, portions of the cells were harvested for Western blotting analysis (16). (B) Effect of the hdf1Δ and rad52Δ mutations on the HO-induced Rad53 phosphorylation. Cells containing YCpT-RAD53-HA and pGAL-HO or the vector were analyzed as in (A). (C) Effect of the ddc1Δ, mec3Δ,rad17Δ, rad9Δ, rad24Δ, andmec1Δ mutations on the HO-induced Rad53 phosphorylation. Cells containing YCpT-RAD53-HA and pGAL-HO were analyzed as in (A).

DNA damage-induced Rad53 phosphorylation is dependent on checkpoint genes, including MEC1, DDC1,MEC3, RAD9, RAD17, andRAD24 (2). We examined Rad53 phosphorylation status after HO expression inddc1Δ, mec3Δ, rad9Δ,rad17Δ, and rad24Δ mutants carrying the GAL-HO and RAD53-HA plasmids. Rad53 phosphorylation afterHO expression was decreased in ddc1Δ,mec3Δ, rad9Δ, rad17Δ, andrad24Δ mutants compared with that observed in wild-type cells (Fig. 1C). The sml1Δ mutation suppresses the lethality but not the checkpoint defect associated with themec1Δ mutation (15). As found in wild-type cells, Rad53 became phosphorylated afterHO expression in sml1Δ cells. In contrast, no Rad53 modification was observed inmec1Δ sml1Δ mutants (Fig. 1C). Thus, continuous HO expression induces phosphorylation of Rad53 in a manner dependent on the DDC1, MEC1,MEC3, RAD9, RAD17, and RAD24genes.

We used an HO expression system to examine the recruitment of checkpoint proteins to a region near the HO-cleavage site. Ddc1 forms a complex with Rad17 and Mec3, and this Ddc1-Mec3-Rad17 complex has been suggested to function downstream of Rad24 in the checkpoint pathway (16). We examined the association of Ddc1 with sites near the HO-induced DSB with the use of a chromatin immunoprecipitation assay (17). Cells expressing the Ddc1-HA protein were transformed with the GAL-HO plasmid or empty vector (11). After transformants were grown in sucrose, one-half of the culture was maintained in sucrose and the other half was incubated with galactose for 4 hours to induce HOexpression. Extracts prepared after formaldehyde cross-linking were subjected to immunoprecipitation with antibody to HA (anti-HA). Precipitated DNAs were extracted and amplified by polymerase chain reaction (PCR) using either primer sets (HO1 and HO2) corresponding to regions near the HO-restriction site on the MATα locus on chromosome III (Fig. 2A) or primers for the SMC2 locus containing no cleavage site on chromosome VI. Interaction of Ddc1 with the MATα locus was detected in cells carrying the GAL-HO plasmid after incubation with galactose but not in similarly treated cells carrying the control vector (Fig. 2B). This Ddc1 interaction is specific to the MATα locus because the PCR amplified from the SMC2 locus did not increase after incubation with galactose (Fig. 2B). Moreover, increase in PCR amplification of the HML or HMR locus was not detected (12), consistent with the observation that the HO endonuclease does not normally generate a DSB at the HML or HMR locus. The Ddc1 interaction was observed 4 hours after cells were exposed to galactose with a similar kinetics to Rad53 phosphorylation, suggesting a temporal correlation between the Ddc1 interaction and checkpoint activation (Figs. 1A and 2B). PCRs with the primer pairs HO1 and HO2 amplified regions 2.0 and 1.0 kb away from the cleavage site, respectively (Fig. 2A), and Ddc1 interacted with these regions in a similar manner (Fig. 2B). These results suggest that Ddc1 is recruited to sites near the HO-induced DSB.

Figure 2

Association of Ddc1 with sites near the HO-induced DSB. (A) The MATα locus on chromosome III (chr. III) contains a single HO endonuclease cleavage site. When the HO endonuclease is expressed, it can readily introduce a DSB at the cleavage site. The primer pairs HO1 and HO2 were designed to amplify regions near the cleavage site on the MATα locus by PCR. (B) Strains (MATα, DDC1-HA) (11) were transformed with pGAL-HO or the YCp50 vector. The resulting strains were grown in sucrose medium without uracil and were transferred to galactose to induce HO expression or maintained in sucrose to repress HO expression. Cells were then collected at the indicated times for a chromatin immunoprecipitation (IP) assay (17). PCR was done with the primers for the MATα locus shown schematically in (A) and for the control SMC2 locus. PCR products from the respective extracts (input) are shown in parallel. (C) Effect of the rad17Δ, rad24Δ,rad9Δ, and mec1Δ mutations on Ddc1 association with DSB. Cells transformed with pGAL-HO were grown in sucrose medium without uracil and were transferred to galactose to induce HO expression (+) or maintained in sucrose to repress HO expression (−) for 4 hours. Extracts were prepared from cells and subjected to chromatin immunoprecipitation assay as in (B). Strains used (11) contain MATα and DDC1-HA.

We next examined whether RAD9, RAD17,RAD24, or MEC1 is required for the recruitment of Ddc1 to the HO-induced DSB lesions. After HO expression, association of Ddc1 with the MATα locus was detected inrad9Δ mutants (Fig. 2C). Similarly, that association was detected in mec1Δ sml1Δ as well assml1Δ strains (Fig. 2C). However, interaction of Ddc1 with the MATα locus was not observed in rad17Δ orrad24Δ mutants (Fig. 2C). These results indicate that the association of Ddc1 with a region near the HO-cleavage site is dependent on RAD17 and RAD24 but not onRAD9 or MEC1. One reason that Ddc1 association with the MATα locus was not detected inrad17Δ and rad24Δ mutants could be that these mutations affect nuclear localization of the Ddc1 protein, thereby preventing association with the HO-induced DSB lesions. To address this possibility, we examined Ddc1 localization by immunofluorescence microscopic analysis (Fig. 3) (18). Control experiments with wild-type cells expressing untagged Ddc1 revealed no staining. In wild-type cells expressing Ddc1-HA, the tagged Ddc1 proteins were found to localize in the nucleus. Ddc1 localization was unaffected in rad24Δ mutants, but in rad17Δ mutants the immunofluorescence signal of Ddc1-HA was decreased in the nucleus and distributed diffusely in the cytoplasm. Expression of Ddc1-HA is similar inrad17Δ mutants and wild-type cells (16, 19). Because Ddc1 fails to form a complex with Mec3 in rad17Δ mutant cells (16,19), it appears that the assembly of Ddc1 into the Ddc1-Mec3-Rad17 complex may be required for its nuclear localization. If so, the defect in the Ddc1 association with theMATα locus in rad17Δ mutants may result from the decreased accumulation of Ddc1 in nucleus. In contrast to therad17Δ mutation, the rad24Δ mutation does not affect the interaction among Rad17, Mec3, and Ddc1 (12,19). Together, these results are consistent with a model in which Rad24 regulates the recruitment of the Ddc1-Mec3-Rad17 complex to the HO-induced DSB. In agreement with this model, Mec3 and Rad17 also associate with the MATα locus after HOexpression in a RAD24 dependent manner (Fig. 4).

Figure 3

Intracelluar localization of Ddc1. Cells were grown to mid-log phase and were subjected to immunofluorescence microscopic analysis (18). Ddc1 and nucleus were visualized using anti-HA antibody and DAPI (4′,6′-diamidino-2-phenylindole), respectively. Strains used (11) are control (MATα, DDC1), wild-type (MATα, DDC1-HA), rad17Δ (MATα, DDC1-HA, rad17Δ),rad24Δ (MATα, DDC1-HA,rad24Δ).

Figure 4

Association of Mec3 and Rad17 with sites near the HO-induced DSB. Cells transformed with pGAL-HO or the control vector were processed and subjected to a chromatin immunoprecipitation assay, as in Fig. 2C. Strains used (11) contain MATα and MEC3-HA (A) orMATα and RAD17-HA(B).

Mec1 has a central role in the DNA damage response and has been suggested to function downstream of the Ddc1-Mec3-Rad17 complex. Therefore, we examined the interaction of Mec1 with the HO-induced DSB lesions in cells expressing Mec1-HA and carrying the GAL-HO plasmid or empty vector (11). Mec1 association with theMATα locus was detected only in cells expressing HO, and no association with the SMC2,HML, or HMR loci was detected (Fig. 5) (12). The Mec1 association with the MATα locus became detectable 4 hours afterHO expression, as did the Ddc1 interaction (Figs. 2B and5A). However, the Mec1 association was not dependent on eitherRAD9 or RAD24 (Fig. 5B). Consistent with the observation that DDC1 functions in the same checkpoint pathway as RAD24, the ddc1Δ mutation did not affect the Mec1 association (Fig. 5B). Our results suggest that Mec1 and Ddc1 are recruited to sites near the HO-induced DSB through distinct mechanisms.

Figure 5

Association of Mec1 with sites near the HO-induced DSB. (A) Kinetics of Mec1 association with DSB. (B) Effect of the rad9Δ, rad24Δ, and ddc1Δ mutations on Mec1 association with DSB. Cells transformed with pGAL-HO or the control vector were processed and subjected to a chromatin immunoprecipitation assay, as in Fig. 2B (A) and Fig. 2C (B). Strains used (11) contain MATα and MEC1-HA.

Rad24 interacts with the small subunits of the replication factor C (RFC) complex (Rfc2, Rfc3, Rfc4, and Rfc5) and forms a complex related to the RFC complex (20–22). Ddc1, Mec3, and Rad17 are structurally related to the proliferating cell nuclear antigen (PCNA), which forms a doughnut-like homotrimer (23–25). During DNA replication, the RFC complex binds to PCNA and loads itonto primed DNA (26). Given these structural similarities, our results indicate that the Rad24 complex may recognize DNA damage and recruit the Ddc1-Mec3-Rad17 complex to such sites. Ddc1 is phosphorylated after DNA damage in a manner dependent onRAD24 and MEC1 (19). Because Mec1 is also recruited to DSB lesions, phosphorylation of Ddc1 may be the result from its localization with Mec1 on damaged DNA. What regulates the association of Mec1 with DNA lesions remains to be determined. ATR and hRad9 are human homologs of Mec1 and Ddc1, respectively (1), and localize to nuclear foci that may contain regions of damaged DNA (27, 28). Because homologs of the checkpoint genes have been identified in other species, it is likely that the mechanism described here is conserved in the checkpoint control among eukaryotic cells.

  • * These authors contributed equally to this work.

  • To whom correspondence should be addressed. E-mail: j46036a{at}nucc.cc.nagoya-u.ac.jp

REFERENCES AND NOTES

View Abstract

Navigate This Article