The Ku heterodimer and the metabolism of single-ended DNA double-strand breaks

Abstract

Single-ended double-strand breaks (DSBs) are a common form of spontaneous DNA break, generated when the replisome encounters a discontinuity in the DNA template. Given their prevalence, understanding the mechanisms governing the fate(s) of single-ended DSBs is important. We describe the influence of the Ku heterodimer and Mre11 nuclease activity on processing of single-ended DSBs. Separation-of-function alleles of yku70 were derived that phenocopy Ku deficiency with respect to single-ended DSBs but remain proficient for NHEJ. The Ku mutants fail to regulate Exo1 activity, and bypass the requirement for Mre11 nuclease activity in the repair of camptothecin-induced single-ended DSBs. Ku mutants exhibited reduced affinity for DNA ends, manifest as both reduced end engagement and enhanced probability of diffusing inward on linear DNA. This study reveals an interplay between Ku and Mre11 in the metabolism of single-ended DSBs that is distinct from repair pathway choice at double-ended DSBs.

Figure 1. yku70* Mutants Are Deficient for Ku-Specific S Phase Function but Proficient for NHEJ

(A) Schematic illustration of the genetic screen used to identify yKu70* separation-of-function mutants. The screen was performed using JPY5025, a yku70Δ mre11-3 yeast strain carrying GAL-HO hmlΔ hmrΔ in a W303+ background. (B) Centromeric plasmids carrying yku70* alleles were transformed in JPY5025 strain (see Extended Experimental Procedures). Exponentially growing cells of the indicated genotypes were 1:5 serially diluted and spotted on DO-TRP lactate in the presence or absence of CPT (12 µM). (C) The same yeast strains were plated onto solid media containing either glucose or galactose to repress (−HO) or induce (+HO) expression of HO endonuclease. (D) Integrity of the yKu70-yKu80 complex in yeast strain JPY5097 transformed with empty vector (vector) or vector expressing yKu70 wild-type (WT) or yKu70* mutants (yku70−). FLAG-tagged (+) or untagged (−) yKu80 was immunoprecipitated with anti-FLAG antibody and yKu70, yKu80 proteins were analyzed by western blot (WB) with anti-yKu70-yKu80 antibody. Input yKu80-FLAG and yKu70 were analyzed by WB (* indicates nonspecific band). (E) Position of human Ku70 residues that correspond to the residues mutated in yku70* alleles mapped on the Ku-DNA crystal structure (Walker et al., 2001). hKu70 and hKu80 are shown in gray and green, respectively, and the DNA is in red. The predicted positions of theyKu70 mutations are shown by colored spheres (blue spheres map inside or adjacent to the protein loops, and magenta spheres map to β sheets). See also Figures S1 and S3 and Tables 1 and S1.

Figure 2. Rescue of mre11-3 CPT Sensitivity and Telomere Shortening in yku70* Alleles Require EXO1

(A) Telomere length analysis of the indicated yku 70* mutants. Plasmids bearing yku70* alleles or control vectors were transformed into yku70Δ oryku70Δ exo1Δ strains. After 60 generations, telomeric sequences were detected by southern blot analysis of Xhol-digested genomic DNA. (B) Telomeric G-overhang assay. Genomic DNA isolated from yku70Δ oryku70Δ exo1Δ strains transformed with CEN plasmids containing the indicated yku 70* alleles were digested with Xhol restriction enzyme and separated on a 0.7% agarose gel. The gel on the left was treated as a nondenaturing gel and hybridized to the end-labeled CA-oligo. The amount of telomeric ssDNA was quantitated before the gel was denatured and probed with the same telomere probe to reveal the total amount of telomeric DNA (right). (C) Ratio of telomeric ssDNA to total telomeric DNA, normalized to WT. Error bars indicate SDs of three independent quantifications. (D) CPT sensitivity of mre11-3 yku70 DNA end-binding-defective mutants and yku80 telomere mutants. Top: serial dilutions of JPY5025 strain containing the indicated yku70 or yku80 alleles on a CEN plasmid were spotted onto DO-TRP lactate in the presence or absence of CPT. Bottom: NHEJ assays as described in Figure 1C. yku70 and yku80 alleles nomenclature: yku70-EBD (DNA end-binding deficient: yku70-R456E spores a and b; Lopez et al., 2011); yku80-tel (yku80-P437L spores a and b; Bertuch and Lundblad, 2003). See also Figure S2.

Figure 3. Mutations in yKu70 Loops Reduce the Affinity of the yKu Heterodimer for DNA Ends

(A) yKu70 wild-type and −5 and −10 mutants were purified to homogeneity in complex with yKu80. The purified complexes were resolved by SDS-PAGE gel, analyzed by WB, and stained with Coomassie blue. (B and C) Analysis of DNA-yKu interaction by filter binding assay. Increasing amounts of the indicated recombinant yKu complexes were incubated overnight with 0.1 nM of 25 bp (B) and 465 bp (C) 5′-labeled dsDNA. Radiolabeled DNA bound to protein was retained on nitrocellulose filter and quantified by scintillation counter. Denatured boiled yKu70 was used as control. Error bars represent the SD of three independent experiments and are smaller than the symbols when not evident. R² is the coefficient of determination. (D) Gel mobility shift assay of yKu binding to 80 bp dsDNA blunt ends (dsBE). The DNA concentration was 2.5 nM in all binding reactions, whereas Ku concentrations are represented by the Ku to DNA molar ratios. Bands representing free DNA or one, two, and three yKu molecules in complex with DNA are indicated. (E) Quantification of the gel in (D). Error bars represent the SD of three independent experiments and are smaller than the symbols when not evident. (F) NHEJ repair kinetics after acute (15 min) HO induction were determined by qPCR in yKU70 (JPY5726), yku70-5 (JPY5730), and yku 70-10 (JPY5728) integrated in the endogenousyKU70 locus in a GAL-HO hmlΔ hmrΔ strain W303+ background. yku70Δ (JPY5735) was used as negative control. Cells were cultivated for another 4 or 6 hr in glucose medium after HO induction, and repair was monitored with primer flanking the HO site. See also Figure S3.

Figure 4. yKu70* Mutants Fail to Block Exo1-Dependent Resection In Vitro

(A) Representation of the DNA substrate used in the resection assay in (B). 5′ biotinylated 80-mer oligonucleotide was annealed to an equally long internally labeled oligonucleotide that would release free AMP upon Exo1 resection (* indicates the position of ³²P on DNA). B, biotin; SV, streptavidin. (B) Biotinylated DNA substrate (1 nM) was preincubated with streptavidin, followed by the presence or absence (control) of the indicated yKu complexes (5 nM). The kinetics of resection by Exo1 (3 nM) was determined at the times indicated by separation of DNA on denaturing gel, and the resection products were visualized with the PhosphorImager system. An 80 bp oligo 5′ labeled on the first adenine was digested with Phosphodiesterase I (PI) and the AMP generated was used as the control for Exo1 resection. (C) The percentage of generated AMP from the experiment shown in (B) is plotted as a function of time. A magnification of the graph in the first 0.5 min is represented to show the curve slope, and the full graph is reported in small scale. Error bars represent the SD from three independent experiments and are smaller than the symbols when not evident. See also Figure S4.

Figure 5. Interaction of yKu70-yKu80 Heterodimers with Linear Duplex DNA Molecules

Nucleoprotein complexes formed between blunt-ended linear DNA (1,821 bp) and the indicated yKu70-yKu80 heterodimers were visualized by SFM as described in Experimental Procedures. (A) Representative SFM images. Left panel: DNA-yKu70; central panel: DNA-yKu70-5; right panel: DNA-yKu70-10. The percentages of the total yKu complexes bound to DNA at either internal sites or DNA ends are reported in Table S2. (B) Graph representing the percentages of the total yKu complexes bound to DNA at either internal sites or DNA ends. Fisher’s exact test was performed on the number of yKu molecules bound to DNA, at DNA ends and internal sites, between yKu70 and yKu70-5, and yKu70 and yKu70-10. In all cases, the p value for a two-sided test is <0.0001 (***). Percentages of the total yKu complexes bound to DNA at either internal sites or DNA ends are shown in the graph. See also Figure S5.

Figure 6. Model Depicting the Effect of Mre11/Sae2 and the yKu Complex in Coordinating Exo1-Dependent Processing and Repair of Single-Ended DSBs

We propose that in the presence of replication-induced single-ended DSBs, Mre11 and Sae2 repress Ku binding at DNA ends, thereby promoting Exo1 recruitment and resection. To explain the molecular basis of yKu70 suppression of single-ended DSB resection, we postulate two nonexclusive scenarios responsible for a reduction of Ku affinity to DNA ends. In our model, a stable binding of the wild-type yKu heterodimer (dark-and light-blue oval structure) to single-ended DSB ends occurs as consequence of an equilibrium maintained between two Ku-DNA binding modes (K₁ and K₂). On this basis, reduced yKu affinity for DNA ends may take place as a consequence of mutations that increase the yKu off-rate and/or the ability of Ku to diffuse inward on the DNA molecule (Ku*: pink and purple oval structure). The reduction in Ku-DNA end-binding stability can be explained by a structural analysis of the Ku mutants. The use of a software (Rosetta Common) to predict the effect(s) of single-point mutations on a protein structure highlighted a shift in the overall backbone conformation (data not shown) of the Ku heterodimer caused by mutation in Ku70. A consequent reduction in the persistency of Ku at DNA ends translates to a failure to protect DNA ends from Exo1 exonuclease activity. We cannot test Sgs1’s role in this model because of the synthetic interaction with mre11-3 (Foster et al., 2011). K₁ and K₂ represent the equilibrium constant of Ku-DNA binding modes.