Mre11 and Ku regulation of double-strand break repair by gene conversion and break-induced replication

Abstract

The yeast Mre11-Rad50-Xrs2 (MRX) and Ku complexes regulate single-strand resection at DNA double-strand breaks (DSB), a key early step in homologous recombination (HR). A prior plasmid gap repair study showed that mre11 mutations, which slow single-strand resection, reduce gene conversion tract lengths and the frequency of associated crossovers. Here we tested whether mre11Delta or nuclease-defective mre11 mutations reduced gene conversion tract lengths during HR between homologous chromosomes in diploid yeast. We found that mre11 mutations reduced the efficiency of HR but did not reduce tract lengths or crossovers, despite substantially reduced end-resection at the test (ura3) locus. End-resection is increased in yku70Delta, but this change also had no effect on tract lengths. Thus, heteroduplex formation and tract lengths are not regulated by the extent of end-resection during DSB repair in a chromosomal context. In a plasmid-chromosome DSB repair assay, tract lengths were again similar in wild-type and mre11Delta, but they were reduced in mre11Delta in a gap repair assay. These results indicate that tract lengths are not affected by the extent of end processing when broken ends can invade nearby sites, perhaps because MRX coordination of the two broken ends is dispensable when ends invade nearby sites. Although HR outcome was largely unaffected in mre11 mutants, break-induced replication (BIR) and chromosome loss increased, suggesting that Mre11 function in mitotic HR is limited to early HR stages. Interestingly, yku70Delta suppressed BIR in mre11 mutants. BIR is also elevated in rad51 mutants, but yku70Delta did not suppress BIR in a rad51 background. These results indicate that Mre11 functions in Rad51-independent BIR, and that Ku functions in Rad51-dependent BIR.

Fig. 1

Allelic and direct repeat HR substrates and outcomes. (A) Allelic HR substrate has ura3 alleles inactivated by frameshift mutations comprising an HO site (HO432) and X764. The chromosome carrying the HO site is marked with HIS3:telV. HR produces short- and long-tract gene conversions (GC); only the converted (recipient) chromosome is shown. Some conversions have associated crossovers and in G2 cells, half of crossovers result in LOH at HIS3:telV (His⁻ or His⁺⁺ products) and half remain heterozygous. Some G2 cells experience only a single DSB (as shown) due to limited access of HO to its site in ura3. BIR produces only His⁻ products (shown above the dotted line). His⁻ products also arise by chromosome loss (not shown) that are also Ura⁻. (B) Direct repeat substrate has the same ura3 alleles flanking pUC19 and LEU2. The four principal HR product types (with associated phenotype) are shown below.

Fig. 2

DSB-induced allelic HR product spectra. (A) Gene conversion (GC) frequencies are shown for wild-type (WT), mre11Δ, and mre11-H125N with or without yku70Δ. Values for each product class are averages ±SD for 4 determinations. His⁻ products are crossovers (CO) but His⁺ may be crossover or non-crossover (NCO) products. These values do not include BIR or chromosome loss events. (B) Total allelic HR frequencies from panel A. * indicates P < 0.05; ** indicates P < 0.01.

Fig. 3

Allelic gene conversion tract lengths do not correlate with extent of end-processing. (A) Tract lengths were estimated as the percentage of long-tract products (Ura⁻) in wild-type (+), deletion mutants of mre11 or yku70 (Δ), or mre11-H125N (HN). Values are averages ±SD for 3–8 determinations. (B) Percent crossovers estimated as His⁻ fractions among all products, or just Ura⁺ classes. (C) End-processing measured by dot-blot of native genomic DNA from wild-type, mre11Δ, and yku70Δ diploid cells isolated before and at indicated times after GALHO induction.

Fig. 4

Mre11 has little or no role in DSB-induced direct repeat HR. (A) Direct repeat HR product spectra in mre11Δ cells with wild-type complementing vector (mre11Δ/MRE11), empty vector control (mre11Δ), or vectors expressing nuclease-defective mre11 alleles. Values are averages ±SD for 4 determinations. There were no significant differences in frequencies of individual HR product classes, or total HR.

Fig. 5

Plasmid-chromosome HR. (A) HR between chromosome V and a linear, transformed plasmid carrying ura3 inactivated by an EcoRI linker in NcoI (position 432). Selected His⁺ transformants were pooled, and isolated genomic DNA was digested with EcoRI. This linearizes parental molecules, and only the circular gene conversion products efficiently transform E. coli. Ampicillin-resistant E. coli transformants were pooled and mapped to determine the ratio of long and short tract HR products. The first three lanes are controls: HindIII excises a 1.2 kbp ura3 fragment from the vector; smaller fragments result from HindIII sites in HIS3. The HindIII/EcoRI digest confirms that no parental molecules are present. The HindIII/NcoI digest confirms that all molecules arose by gene conversion (ura3 cleaved by NcoI). The HindIII/XbaI digest reveals the relative fractions of molecules with long or short tracts (ura3 cleaved by XbaI or not cleaved, respectively). (B) mre11Δ does not reduce plasmid-chromosome tract lengths. The fourth lane of each set (HindIII/XbaI) reveals similar fractions of long and short tract HR products. (C) Scanning densitometric analysis of agarose gels from two independent experiments (Expt #1 data are from gel shown in panel B). (D) Gap repair assay; symbols are described in panel A. (E) Percentage of long tracts among gap repair products, calculated as the ratio of Ura⁻ recombinants per total recombinants (Ura⁺ + Ura⁻).

Fig. 6

mre11 mutations increase BIR, chromosome loss, and DSB-induced cell killing. All values are averages ±SD for 3–4 determinations. (A) BIR in mre11 mutants is suppressed 5- to 6-fold by yku70Δ. (B) yku70Δ does not suppress chromosome loss in mre11 mutants. (C) DSB-dependent cell killing is enhanced in mre11Δ and further enhanced in this background by yku70Δ.

Fig. 7

yku70Δ does not suppress BIR in rad51KR. Data are presented as shown in Figs. 2B, 6A, and 6B.