Mitochondrial genome maintenance: roles for nuclear nonhomologous end-joining proteins in Saccharomyces cerevisiae

Abstract

Mitochondrial DNA (mtDNA) deletions are associated with sporadic and inherited diseases and age-associated neurodegenerative disorders. Approximately 85% of mtDNA deletions identified in humans are flanked by short directly repeated sequences; however, mechanisms by which these deletions arise are unknown. A limitation in deciphering these mechanisms is the essential nature of the mitochondrial genome in most living cells. One exception is budding yeast, which are facultative anaerobes and one of the few organisms for which directed mtDNA manipulation is possible. Using this model system, we have developed a system to simultaneously monitor spontaneous direct-repeat-mediated deletions (DRMDs) in the nuclear and mitochondrial genomes. In addition, the mitochondrial DRMD reporter contains a unique KpnI restriction endonuclease recognition site that is not present in otherwise wild-type (WT) mtDNA. We have expressed KpnI fused to a mitochondrial localization signal to induce a specific mitochondrial double-strand break (mtDSB). Here we report that loss of the MRX (Mre11p, Rad50p, Xrs2p) and Ku70/80 (Ku70p, Ku80p) complexes significantly impacts the rate of spontaneous deletion events in mtDNA, and these proteins contribute to the repair of induced mtDSBs. Furthermore, our data support homologous recombination (HR) as the predominant pathway by which mtDNA deletions arise in yeast, and suggest that the MRX and Ku70/80 complexes are partially redundant in mitochondria.

Figure 1

Nuclear and mitochondrial direct-repeat mediated deletions. (A) Nuclear DRMD reporter consists of the URA3 gene inserted 99 bp into the TRP1 gene followed by the entire TRP1 gene lacking the start codon (Kalifa et al. 2009). (B) Mitochondrial reporter consists of the ARG8^m gene inserted 99 bp into the COX2 gene followed by the entire COX2 gene lacking the start codon (Phadnis et al. 2005). Both reporters contain 96 bp of directly repeated sequence shown in hashed boxes. Strain with intact reporter constructs are Ura⁺ Trp⁻ Arg⁺ and respiratory deficient. (C) Average rates of nuclear DRMD. (D) Average rates of mitochondrial DRMD plotted on a log scale. Error bars indicate the SEM. *, at least one or more of the rates used to determine the average was derived using the Luria and Delbruck (1943) method due to a median value of 0.

Figure 2

Inducing a specific mtDSB with mtLS-KpnI. (A) Schematic of probes used for detection of mtDSB. Dotted line indicates where the probe anneals. (B) Representative Southern blot of AvaII-digested DNA extracted from wild-type strains after induction of mtLS-KpnI. WT (DFS188) contains an intact COX2 gene and was grown in SD −complete media and time points were taken in Sgal −complete media. REP96 (EAS748) contains the REP96::ARG8^m::cox2 reporter and was grown in SD −Arg media and time points were taken in SGal −complete media. EV (LKY623) contains the REP96::ARG8^m::cox2 reporter and pYES2.1, and was grown in SD −Arg −Ura media with time points taken in SGal −Ura media. mtLS-KpnI (EAS812) contains the REP96::ARG8^m::cox2 reporter and pEAS100. It was grown in SD −Arg −Ura media and time points were taken in Sgal −Ura media. Ctrl is the LKY623 T = 0 sample digested with KpnI in vitro to depict where the break product B would be relative to the other COX2 bands. The 21S rRNA gene was probed to detect total mtDNA. The nuclear 25S rRNA gene was probed for total nuclear DNA. The average increases in recombinant and break product relative to time 0 for several independent experiments are given. Data were obtained by quantification of Southern blots as described in Materials and Methods. Significance of increase by 3 hr after shift to galactose was determined using paired t tests to compare the percentage of each product before and after endonuclease induction.

Figure 3

Inducing a specific mtDSB with mtLS-KpnI in MRX and KU70/80 mutant strains. (A) Induction of a specific mtDSB in mrx-Δ cells. EV (LKY642) contains the REP96::ARG8^m::cox2 reporter and pYES2.1, and mtLS-KpnI (LKY648) contains the REP96::ARG8^m::cox2 reporter and pEAS100. (B) Induction of a specific mtDSB in ku70/80-Δ cells. EV (LKY644) contains the REP96::ARG8^m::cox2 reporter and pYES2.1, and mtLS-KpnI (LKY650) contains the REP96::ARG8^m::cox2 reporter and pEAS100. A darker exposure of this blot is shown so that the break product can be seen. (C) Induction of a specific mtDSB in mrx-Δku70/80-Δ cells. EV (LKY646) contains the REP96::ARG8^m::cox2 reporter and pYES2.1, and mtLS-KpnI (LKY652) contains the REP96::ARG8^m::cox2 reporter and pEAS100. All strains were grown in SD −Arg −Ura media and time points were taken in SGal −Ura media. Total DNA was extracted, digested with AvaII, and subjected to gel electrophoresis on a 0.8% agarose gel prior to Southern blotting. Ctrl is the EV T = 0 sample digested with KpnI in vitro to depict where the break product B would be relative to the other COX2 bands. The 21S rRNA gene was probed to detect total mtDNA. The nuclear 25S rRNA gene was probed for total nuclear DNA. The average increases in deletion and break product relative to time 0 for several independent experiments are given. Data were obtained by quantification of Southern blots as described in Materials and Methods. Significance of increase by 3 hr after shift to galactose was determined using paired t tests to compare the percentage of each product before and after endonuclease induction.

Figure 4

Detection of reciprocal products of recombination. In addition to the deletion product we select, HR pathways may produce the reciprocal products shown in (A). The circular product would be produced by homologous recombination between repeats within the same mitochondrial genome, while the duplication would be produced by recombination between the first repeat on one mitochondrial genome with the second repeat on a separate mtDNA molecule. Primers indicated by the small arrows below ARG8^m were used to amplify a 489 bp fragment from the reciprocal products. In (B), the amplified fragments were run on a 1.3% agarose gel and stained with ethidium bromide. Relative concentrations of mtDNA were estimated by amplifying a fragment of comparable size from the COX1 gene in the mitochondrial genome. All reactions shown in the figure for each primer pair were amplified using the same master mix. The ratio of DNA amplified from reciprocal products and from COX1 is indicated below each lane after normalizing for staining intensity as described in Materials and Methods.

Figure 5

Analysis of KpnI resistance. Representative gel of PCR amplification and digestion with KpnI in vitro of time 0 samples grown under chronic DSB conditions. Cells were grown to saturation in SD −Arg −Ura, then subcultured twice in the same media, and total DNA was extracted for PCR. Amplification of the ARG8^m fragment from LKY196 served as the control for KpnI^R (R) and KpnI sensitive (S) populations. Quantification of independent experiments is present in Table 3.

Figure 6

Model for repair of DSB in the mitochondrial compartment. A proposed model is shown for mtDSBR when the DSB occurs between directly repeated sequences. There is competition between MRX and Ku70/80 complexes for the break substrate. When repair is promoted by MRX, the favored repair pathway is HR. This model depicts SSA in which DSBR results in a deletion product and does not produce reciprocal product; however, MRX could also promote other types of HR repair. If repair occurs via Ku-dependent repair, the primary pathway would be classical NHEJ, in which the ends of the breaks are protected and religated together restoring the original sequence. Alternatively, degradation at the break site resulting in complete loss of the intervening sequence, subsequently followed by end binding and ligation, would result in a deletion product.