Members of the RAD52 Epistasis Group Contribute to Mitochondrial Homologous Recombination and Double-Strand Break Repair in Saccharomyces cerevisiae

Abstract

Mitochondria contain an independently maintained genome that encodes several proteins required for cellular respiration. Deletions in the mitochondrial genome have been identified that cause several maternally inherited diseases and are associated with certain cancers and neurological disorders. The majority of these deletions in human cells are flanked by short, repetitive sequences, suggesting that these deletions may result from recombination events. Our current understanding of the maintenance and repair of mtDNA is quite limited compared to our understanding of similar events in the nucleus. Many nuclear DNA repair proteins are now known to also localize to mitochondria, but their function and the mechanism of their action remain largely unknown. This study investigated the contribution of the nuclear double-strand break repair (DSBR) proteins Rad51p, Rad52p and Rad59p in mtDNA repair. We have determined that both Rad51p and Rad59p are localized to the matrix of the mitochondria and that Rad51p binds directly to mitochondrial DNA. In addition, a mitochondrially-targeted restriction endonuclease (mtLS-KpnI) was used to produce a unique double-strand break (DSB) in the mitochondrial genome, which allowed direct analysis of DSB repair in vivo in Saccharomyces cerevisiae. We find that loss of these three proteins significantly decreases the rate of spontaneous deletion events and the loss of Rad51p and Rad59p impairs the repair of induced mtDNA DSBs.

Fig 1. Models for the generation of deletions between directly repeated sequences.

Repair of a double-strand break (DSB) located between directly repeated sequences (grey boxes) can result in the deletion of one of the repeats and all of the intervening sequence. 5’ to 3’ resection at the DSB reveals the direct repeats. Rad52p alone or in conjunction with Rad59p, promotes the annealing of the repeats. Template switching can occur if a lesion (yellow star) is encountered during replication. At a stalled fork, the nascent strand may invade at the incorrect repeat, leading to the generation of a deletion. Unequal exchange can occur when HR occurs between misaligned repeats leading to a deletion. Both template switching and unequal exchange can happen intramolecularly or intermolecularly. Arrowheads indicate 3’ end.

Fig 2. Rad51p and Rad59p localize to the mitochondrial matrix.

(A) Intact mitochondria, mitoplasts, or Triton X-100 lysed mitochondria were treated with proteinase K and subjected to immunoblot analysis with the indicated antibodies. Por1p, outer membrane protein; Cob, intermembrane space protein; Cit1p, matrix protein. (B) Rad51p physically interacts with mtDNA. DNA samples (3 biological replicates) were immunoprecipitated with antibodies against endogenous Rad51p or V5 epitope tag (mock). Samples were evaluated by qPCR and changes in fold enrichment were compared to the mock IP. Error bar indicates SD. Asterisks indicate significant differences between the Rad51p IP and the mock IP (** = p ≤ 0.01).

Fig 3. Spontaneous nuclear and mitochondrial repeat mediated deletions.

(A) The nuclear DRMD reporter. The URA3 gene was inserted 99 bp into the TRP1 sequence, and is followed directly by the entire TRP1 gene lacking the start codon, resulting in a 96 bp direct repeat flanking the URA3 marker. Spontaneous deletions are selected on medium lacking tryptophan, and rates were determined using the method of the median. (B) The mitochondrial DRMD reporter. The ARG8 ^m gene was inserted 99 bp into the COX2 sequence, and is followed directly by the entire COX2 gene lacking the start codon, resulting in a 96 bp direct repeat flanking the ARG8 ^m marker. Spontaneous deletions are selected on medium containing glycerol as the sole carbon source, and rates were determined using the method of the median. (C) Average rates of nuclear DRMD. (D) Average rates of mitochondrial DRMD. Error bars indicate SD. Asterisks indicate significant differences between the mutant and wild-type rates (* = p ≤ 0.05, ** = p ≤ 0.01, *** = p ≤ 0.001, **** = p ≤ 0.0001).

Fig 4. Inducing a specific mitochondrial DSB.

(A) The mitochondrial DRMD reporter, with restriction recognition sequences indicated. Horizontal line beneath COX2 indicates where the COX2 probe anneals. (B) Representative Southern blot of AvaII-digested DNA extracted from EAS748 containing pEAS115 (mtLS-KpnI-intein^dead) or pEAS114 (mtLS-KpnI-intein^ts). Both strains were grown in SRaffinose-Arg-Ura until OD₆₀₀ = 0.150. Galactose was added to a final concentration of 2%, and the strains were shifted to 20°C to induce the either mtLS-KpnI-intein^ts or mtLS-KpnI-intein^dead. Genomic DNA in lanes labeled P are from preinduced cultures. Cultures were incubated for 16–18 hours at 20°C and samples were taken for the time 0 timepoint. The cultures then were shifted to 30°C to allow repair. Lane I contains DNA from the t = 0 timepoint of the mtLS-KpnI-intein^ts-containing strain digested in vitro with KpnI and AvaII to demonstrate the migration of DNA with a DSB. The 21S rRNA gene was probed to detect total mtDNA. The 25S rRNA gene was probed to detect total nuclear DNA. (C) Relative mtDNA content was determined by taking the ratio of the 21S rRNA signal and the 25S rRNA signal, then normalizing to the t = 0 sample for each strain. (D) The intensity of the COX2-hybridizing bands were measured using Image Lab software (www.biorad.com) The proportions of the total COX2 signal present in the DSB and deletion products were calculated as a percent of the total COX2 signal.

Fig 5. Induced mitochondrial repeat mediated deletions.

Strains were grown in SRaffinose-Arg-Ura until OD₆₀₀ = 0.150. Galactose was added to a final concentration of 2%, and the strains were shifted to 20°C for 16–18 hours in order to induce DSBs via mtLS-KpnI. Appropriate dilutions were plated on YPD and allowed to grow for 3 days at 30°C. The colonies were then replica plated to YPG and SD-Arg. Growth on YPG was scored in order to determine the percent of colonies that had at least some cells that had undergone a deletion after non-selective growth. Error bars indicate SD. Asterisks indicate significant differences between the mutant and wild-type rates (* = p ≤ 0.05, **** = p ≤ 0.0001).

Fig 6. Repair of induced mitochondrial DSBs in HR single mutants.

All samples were obtained after 16–18 hours of mtLS-KpnI induction. DNA was extracted from the appropriate strains and digested with AvaII. Samples in lanes labeled P are from pre-induced cultures. Lane I is the t = 0 sample from the mtLS-KpnI-intein^ts expressing strain digested with KpnI and AvaII to demonstrate the migration of the reporter with a DSB in relation to the other COX2 bands. (A) Representative Southern blot of rad51-Δ strains ASY114 (intein^dead) and ASY113 (intein^ts). (B) Representative Southern blot of rad52-Δ strains ASY116 (intein^dead) and ASY115 (intein^ts). (C) Representative Southern blot of rad59-Δ strains ASY118 (intein^dead) and ASY117 (intein^ts). (D) A representative Southern blot of rad51-Δ rad52-Δ strains, ASY120 (intein^dead) and ASY119 (intein^ts).

Fig 7. Conversion of DSBs to deletion product is impaired in HR mutants.

Experiments were performed as in Fig 4. The ratios of the amount of COX2 signal in the cut band at t = 0 and the amount of COX2 signal in the deletion product at t = 8 were calculated. The intensity of the COX2-hybridizing bands was measured using Image Lab software (www.biorad.com). The proportions of the total COX2 signal present in the DSB and deletion products were calculated as a percent of the total COX2 signal. Error bars indicate SD. Asterisks indicate significant differences between the mutant and wild-type rates (* = p ≤ 0.05, *** = p ≤ 0.001).