Role of Dot1-dependent histone H3 methylation in G1 and S phase DNA damage checkpoint functions of Rad9

Abstract

We screened radiation-sensitive yeast mutants for DNA damage checkpoint defects and identified Dot1, the conserved histone H3 Lys 79 methyltransferase. DOT1 deletion mutants (dot1Delta) are G1 and intra-S phase checkpoint defective after ionizing radiation but remain competent for G2/M arrest. Mutations that affect Dot1 function such as Rad6-Bre1/Paf1 pathway gene deletions or mutation of H2B Lys 123 or H3 Lys 79 share dot1Delta checkpoint defects. Whereas dot1Delta alone confers minimal DNA damage sensitivity, combining dot1Delta with histone methyltransferase mutations set1Delta and set2Delta markedly enhances lethality. Interestingly, set1Delta and set2Delta mutants remain G1 checkpoint competent, but set1Delta displays a mild S phase checkpoint defect. In human cells, H3 Lys 79 methylation by hDOT1L likely mediates recruitment of the signaling protein 53BP1 via its paired tudor domains to double-strand breaks (DSBs). Consistent with this paradigm, loss of Dot1 prevents activation of the yeast 53BP1 ortholog Rad9 or Chk2 homolog Rad53 and decreases binding of Rad9 to DSBs after DNA damage. Mutation of Rad9 to alter tudor domain binding to methylated Lys 79 phenocopies the dot1Delta checkpoint defect and blocks Rad53 phosphorylation. These results indicate a key role for chromatin and methylation of histone H3 Lys 79 in yeast DNA damage signaling.

FIG. 1.

Histone H3 methyltransferase Dot1 is required for G₁ phase and intra-S-phase checkpoint response to IR. (A) Arrest defect in dot1Δ after DNA damage in G₁. α-Factor-arrested cells were irradiated with 300 Gy, released into fresh media, and analyzed at 15-min intervals for DNA content by flow cytometry. (B) Aliquots from the same experiment as in panel A were mixed with α-factor/nocodazole trapping media and incubated for an additional 90 min. The fraction of cells that remained in G₁ was determined by flow cytometry. (C) Intra-S-phase checkpoint defect in dot1Δ. Cells were irradiated 15 min after release from G₁ arrest, and DNA replication was monitored by flow cytometry. (D) The G₂/M checkpoint is intact in dot1Δ. Wild-type (W303-1A), dot1Δ (SKY2849), and rad9Δ (SKY2851) cells were synchronized with nocodazole before irradiation and release into fresh media. Mitotic progression was determined by the percentage of large-budded cells with separated nuclei.

FIG. 2.

DNA damage sensitivity and order-of-function between DOT1 and checkpoint genes RAD9 and RAD17. (A and B) Asynchronous (A) or G₁-arrested (B) cultures were irradiated, serially diluted, and plated on rich media. Strains used were wild type (W303-1A), dot1Δ (SKY2849), rad9Δ (SKY2851), dot1Δ rad9Δ (SKY2852), rad17Δ (SKY2862), and dot1Δ rad17Δ (SKY2863).

FIG. 3.

Histone H3 methylation and histone H2B ubiquitination are involved in G₁ DNA damage checkpoint. (A) DNA damage checkpoint dependence on Dot1 methyltransferase activity. The dot1Δ mutant (SKY2849) was transformed with an empty vector, wild-type DOT1, or catalytically defective dot1-Gly401Arg. Percentage of cells remaining in G₁ was determined quantitatively from flow cytometry of DNA content using FlowJo 6.3.2 software. The 1N population on a two-dimensional scatter plot of side scatter versus DNA content was gated, and the counts were normalized to flow cytometry at t = 0 min. (B) Checkpoint role of H3-Lys 79. Wild type or Lys79Ala mutant H3 (WZY42) and dot1Δ mutant (SKY2849) were grown to saturation to increase G₁ content, irradiated, released into fresh media, and analyzed by flow cytometry. Quantitation was performed as in panel A. (C) Scheme of Paf1/RNA polymerase II and H2B ubiquitination-dependent modifications of histone H3. (D) Dependence of G₁ checkpoint on upstream regulators. Paf1-complex mutant rtf1Δ (SKY2853), H2B ubiquitination mutants bre1Δ (SKY2854) and rad6Δ (SKY2855), and histone H2B Lys123Ala mutant (Y132) were arrested in G₁, irradiated, released into fresh media, and analyzed by α-factor/nocodazole trap assay to determine the percentage of cells that remained in G₁. Triangles and squares represent mock-treated (0 Gy) and irradiated (300 Gy) samples, respectively.

FIG. 4.

Set1 contributes to intra-S but not to G₁ or G₂/M checkpoint after DNA damage. (A and B) The set1Δ mutant shows intact G₁ checkpoint arrest (A) but a defective intra-S-phase checkpoint response (B). Cells arrested with α-factor were either irradiated with 300 Gy for G₁ checkpoint analysis by α-factor/nocodazole trap assay (A) or released into 0.03% MMS and analyzed by flow cytometry (B). Arrows indicate the first time point at which a 2N population, consistent with completion of replication, is observed. (C) Analysis of G₂/M checkpoint after IR. Cells synchronized in G₂/M with nocodazole, treated with IR, and released into fresh media were analyzed for mitotic progression by microscopy. (D) Sensitivity to IR in combinations of dot1Δ, set1Δ, and set2Δ mutations. Asynchronous cultures were irradiated, serially diluted and plated on rich media. For panels A to D, the strains used were wild type (W303-1A), dot1Δ (SKY2849), set1Δ (SKY2856), set2Δ (SKY2857), set1Δ set2Δ (SKY2858), dot1Δ set1Δ (SKY2859), dot1Δ set2Δ (SKY2860) and dot1Δ set1Δ set2Δ (SKY2861).

FIG. 5.

G₁-specific loss of Rad9 and Rad53 phosphorylation in dot1Δ cells. Wild-type (SKY2864 and SKY2866) and dot1Δ (SKY2865 and SKY2867) strains expressing Rad9-13Myc or Rad53-13Myc were synchronized in either G₁ with α-factor (A) or G₂/M with nocodazole (B) and irradiated, and arrest was maintained for the duration of the experiment. Aliquots were subjected to Western analysis to detect mobility shifts of Rad9 and Rad53.

FIG. 6.

Ddc2-Rad53 fusion protein restores the integrity of intra-S but not G₁ checkpoint in cells lacking DOT1 or RAD9 after IR. Wild-type (W303-1A), dot1Δ (SKY2849), and rad9Δ (SKY2850) cells expressing DDC2-RAD53 or RAD53 were synchronized in G₁, irradiated, released, and analyzed by flow cytometry (A) and α-factor/nocodazole trap assay (B).

FIG. 7.

A rad9-Tyr798Gln tudor domain mutation phenocopies dot1Δ defects in G₁ and intra-S checkpoint after IR. (A) Entry into S phase and DNA replication are not delayed in rad9Δ (SKY2850) cells expressing the rad9-Tyr798Gln allele in response to IR, whereas G₂/M checkpoint remains intact. The integrity of G₁, intra-S, and G₂/M checkpoints in the rad9Δ transformants containing empty plasmid pRS416, pRS416-RAD9, or pRS416-rad9-Tyr798Gln was determined. (B) Failure of rad9-Tyr798Gln mutant to promote Rad53 phosphorylation after IR in G₁. The rad9Δ strain expressing the Myc-tagged Rad53 was transformed with the wild-type RAD9 and the rad9-Tyr798Gln allele. Transformants were synchronized in G₁ and treated with 300 Gy of IR to induce checkpoint response. Cells were harvested 30 min after IR for protein extracts and subjected to Western analysis with anti-Myc antibodies. (C) Effect of rad9-Tyr798Gln mutation on survival after IR. Aliquots of irradiated and mock-treated transformants were spotted onto YPD plates to determine the rate of survival after IR.

FIG. 8.

Dot1-dependent association of Rad9 with the HO-induced DSB. (A) Association of Rad9 near the HO-induced DSB was determined for chromosomal sites 60bp (HO site), 0.5kb (HO + 0.5 kb), 1.5kb (HO + 1.5 kb), and/or 10kb (HO + 10 kb) from a HO-cutting site using the indicated pairs of primers and for a large intergenic region on chromosome V (control locus). (B) Recruitment of Rad9 to the HO-induced DSB is profoundly decreased in dot1Δ (QY368) cells despite the high level of histone H2A phosphorylation at Ser 129 near DSB. In addition, level of histone H3 methylation at Lys 79 is not altered near the HO-induced DSB in wild-type cells (QY363). ChIP assays were performed on asynchronous growing cells after formation of HO-induced DSB (180 min in YPGal medium) or without expression of HO endonuclease (YPD medium). (C) Kinetics of Rad9-HA recruitment to the HO-induced DSB in G₁-arrested cells. The time course experiment was performed by incubating G₁-synchronized wild-type (QY363) and dot1Δ (QY368) cells in YPGal medium for 0, 20, 40, 60, or 120 min before cross-link-ChIP and real-time PCR. The efficiency of HO cleavage at these time points was 0, 66, 96, 98, and 99% in wild-type cells, whereas 0, 75, 98, 99, and 99% in dot1Δ cells (data not shown). (D) Association of Rad9-HA with the HO-induced DSB is higher in G₁- than in G₂-arrested cells but is equally reduced in dot1Δ mutant cells. Wild-type (QY364) and dot1Δ (QY367) cultures were synchronized in either G₁ or G₂/M and incubated in galactose-containing medium for 180 min to induce DSB. (B to D) Data are presented as occupancy at specific loci based on the immunoprecipitation/input ratio obtained by real-time PCR (duplicate) after correction for efficiency of the specific pairs of primers over the range of PCR cycles used. All experiments were performed in JKM179 background with integrated GAL1,10:HO cassette and deleted HMR/HML loci (64).