Prevention of mitochondrial genomic instability in yeast by the mitochondrial recombinase Mhr1

Mitochondrial (mt) DNA encodes factors essential for cellular respiration, therefore its level and integrity are crucial. ABF2 encodes a mitochondrial DNA-binding protein and its null mutation (Δabf2) induces mtDNA instability in Saccharomyces cerevisiae. Mhr1 is a mitochondrial recombinase that mediates the predominant form of mtDNA replication and acts in mtDNA segregation and the repair of mtDNA double-stranded breaks (DSBs). However, the involvement of Mhr1 in prevention of mtDNA deletion mutagenesis is unknown. In this study we used Δabf2 mhr1-1 double-mutant cells, which lose mitochondrial function in media containing fermentable carbon sources, to investigate whether Mhr1 is a suppressor of mtDNA deletion mutagenesis. We used a suppresivity assay and Southern blot analysis to reveal that the Δabf2 mutation causes mtDNA deletions rather than an mtDNA-lacking (ρ0) phenotype, and observed that mtDNA deletions are exacerbated by an additional mhr1-1 mutation. Loss of respiratory function due to mtDNA fragmentation occurred in ∆mhr1 and ∆abf2 mhr1-1 cells. However, exogenous introduction of Mhr1 into Δabf2 mhr1-1 cells significantly rescued respiratory growth, suggesting that Mhr1-driven homologous mtDNA recombination prevents mtDNA instability.


Results
Double-mutant ∆abf2 mhr1-1 cells rapidly lose respiratory function in fermentable media. In order to examine severely compromised mtDNA maintenance, we used ∆abf2 cells (Table 1), which display a well-documented loss-of-mtDNA phenotype upon cultivation in fermentable media 8,14,15 . In order to compare the extent of respiratory function loss in these backgrounds, we first selectively pre-cultivated wild-type (WT), single-mutant ∆abf2, mhr1-1, or double-mutant ∆abf2 mhr1-1 cells in glycerol medium, a carbon source requiring mitochondrial respiration for its utilization. We then transferred the cells to synthetic complete fermentable media containing glucose (Glu) or raffinose and galactose (RGal) as carbon sources. While both Glu and RGal media are fermentable, the use of RGal allows for distinction from the transcriptional effects of glucose, which changes the global gene expression pattern 25 . We cultivated cells for nearly eight generations at 30 °C or 34 °C (Fig. 1a) and then spread equal amounts of dilute culture of each strain onto rich glucose (YPD) and rich glycerol (YPGly) plates (Fig. 1b). The proportion of colony-forming units (CFUs) that retained mitochondrial respiratory activity and were thus able to grow on YPGly, compared to the total number of CFUs on YPD, was quantified ( Fig. 1c).
We observed that WT cells retained almost all respiratory function under each condition, while ∆abf2 cells showed remarkable loss of respiratory activity in Glu media, to 59.1 ± 5.9% ρ + at 30 °C and 14.0 ± 5.3% ρ + at 34 °C. On the other hand, cultivation of ∆abf2 cells in RGal medium resulted in a large proportion of cells retaining respiratory activity, forming ρ + CFUs at rates of 103.9 ± 14.3% at 30 °C and 83.9 ± 20.7% at 34 °C. These observations appear consistent with the previously reported mtDNA instability phenotype of ∆abf2 cells in glucose 14 .
We next explored the effect of extended cultivation time on loss of cellular respiratory function. We pre-cultivated the four strains in YPGly medium, then transferred the cells to RGal medium and cultivated for approximately 20 generations (Fig. 1d). Finally, we spread the cells on YPD and subsequently replicated the colonies onto YPGly plates (Fig. 1e), or simultaneously spread equal amounts of dilute culture onto YPD and YPGly plates (Fig. 1f). In contrast to WT and mhr1-1 single-mutant cells, which remained almost entirely ρ + , only 27.0 ± 18.4% of the ∆abf2 cells were able to form ρ + colonies on glycerol plates following simultaneous spreading. Remarkably, none of the ∆abf2 mhr1-1 double-mutant cells were able to grow on YPGly plates after growth for approximately 20 generations in RGal (Fig. 1g). Therefore, extended cultivation of the ∆abf2 single-mutant in   www.nature.com/scientificreports www.nature.com/scientificreports/ fermentable RGal media increases loss of cellular respiratory function, while the additional loss of Mhr1 function causes rapid and complete loss of cellular respiratory function 1,14 . Nucleoid numbers are significantly reduced in ∆abf2 mhr1-1 cells. Next, we investigated the relative abundance of mtDNA nucleoids in WT, Δabf2 or mhr1-1 single-mutant, or Δabf2 mhr1-1 double-mutant cells selectively grown in YPGly media, or grown in YPD media at 30 °C or 34 °C for more than 20 generations (Fig. 2). In Δabf2 single-mutants, approximately 27% of CFUs remained ρ + after cultivation in glucose (Fig. 1g), yet all ∆abf2 mother cells we observed displayed mtDNA-derived nucleoid signals after cultivation in YPD at 30 °C (Fig. 2b), suggesting that loss of respiratory function in ∆abf2 cells may be due to mtDNA deletion mutagenesis. In addition, double-mutant ∆abf2 mhr1-1 cells showed the lowest number of mitochondrial nucleoid signals after cultivation in YPD media (Fig. 2a,b), and nucleoid signals were absent from 44.4% of ∆abf2 mhr1-1 mother cells following cultivation at 34 °C, compared to 0% of WT and mhr1-1 cells and only 10.0% of ∆abf2 mother cells. These results indicate that mtDNA maintenance is defective without Abf2 and fully functional Mhr1. Since we observed that respiratory function and nucleoid abundance generally decrease in ∆abf2 mhr1-1 cells relative to ∆abf2 cells, it is likely Mhr1 plays a role in protecting mtDNA genomic integrity. the mhr1-null mutation (Δmhr1) causes mtDNA fragmentation. In order to demonstrate that Mhr1 is required to maintain mtDNA integrity, we introduced mhr1::LEU DNA fragments into WT/WT diploid cells to disrupt one of the two MHR1 alleles, thereby creating WT/Δmhr1 haploinsufficient diploid cells (see: Table 1). We then conducted tetrad analysis to determine whether haploid spores with the nuclear genotype mhr1::LEU (Δmhr1) retain respiratory function. All leucine prototrophic spores were unable to grow on YPGly plates, confirming that the Δmhr1 mutation causes complete loss of respiratory function (Fig. 3a). We further cultivated WT cells and seven Δmhr1 spores ( Fig. 3b,a-g) in glucose medium, after selecting cells that still displayed mtDNA signals upon DAPI-staining. ApaI-digests of mtDNA from wild-type cells gave rise to many discrete bands (Fig. 3b), while ApaI-digests of mtDNA from Δmhr1 cells gave only a few discrete bands, indicating mtDNA deletions occur in Δmhr1 cells with some observable amounts of mtDNA remaining.

MtDNA deletion mutagenesis in ∆abf2 mhr1-1 cells.
To determine whether mtDNA deletion mutagenesis or the complete loss of mtDNA occurred in ∆abf2 single-or ∆abf2 mhr1-1 double-mutant cells, we analyzed suppressiveness according to previously established methods 24 . The degree of suppressiveness is determined by: (1) The length of the remnant mtDNA molecule after undergoing a deletion and (2) the preservation of an active ori sequence (Fig. 4a). Small mtDNA deletions result in relatively large remnant molecules. Therefore, when crossed with ρ + haploid cells, small mtDNA deletion-bearing cells give rise to diploid populations that are in a range of 10% to 90% ρ + . In contrast, when mitochondrial genomes undergo a large deletion event but retain at least one active replication origin, crosses of these haploid cells with ρ + haploid cells of the opposite mating type will give rise to <5% ρ + diploid progeny, a phenotype termed "hypersuppressive" 26 . Finally, crossing haploid cells without mtDNA (ρ 0 ) with ρ + haploid cells will result in 100% ρ + diploid progeny, a phenotype termed "non-suppressive" (Fig. 4a).
We crossed a ρ + strain with haploid WT ρ + , ρ − , HS ρ − or ρ 0 cells as controls (Fig. 4b, top four panels) and ∆abf2 or ∆abf2 mhr1-1 cells, and analyzed the respiratory phenotypes of the resulting diploid progeny. In crosses with the ∆abf2 background, 20~80% of diploid colonies were ρ + , with a single distribution centered at around 55% ρ + (Fig. 4b, second panel from bottom). In contrast to this moderately suppressive phenotype, crossing ∆abf2 mhr1-1 cells yielded a bimodal distribution in which resulting diploid cells either displayed a hypersuppressive, or moderate to non-suppressive phenotype, ranging from 0~20% and 50~100% ρ + , respectively (Fig. 4b, bottom panel). These results indicate that large-scale mtDNA deletions or the complete loss of mtDNA occurs in ∆abf2 mhr1-1 cells, suggesting that the increased production of ρ − progeny (Fig. 1c,g) is due to a deficiency of Mhr1-driven mtDNA recombination.
Next, we analyzed mtDNA from ρ − ∆abf2 and ∆abf2 mhr1-1 colonies by Southern blot analysis. Compared to ρ + WT and ∆abf2 controls (Fig. 5a), we observed that ρ − ∆abf2 mtDNA generally lacked several mtDNA-specific bands that were present in ρ + mtDNA (Fig. 5b). Similarly, ∆abf2 mhr1-1 double-mutant cells generally lacked many mtDNA-specific bands and some samples lacked mtDNA signals altogether, compared to the mtDNA signals derived from ρ + WT and mhr1-1 cells (Fig. 5a,b). Also, to our surprise we observed several more small, mtDNA-specific bands in the mhr1-1 control compared to the ρ + WT and ∆abf2 controls. One explanation is that mtDNA deletions caused by mhr1-1 mutation result in heteroplasmy, which would be stable since the mhr1-1 mutation delays mitochondrial allele segregation 20 . Collectively, these results indicate that mtDNA deletions occur in ρ − ∆abf2 cells and that large-scale mtDNA deletion or the complete loss of mtDNA occurs in ρ − or ρ 0 ∆abf2 mhr1-1 cells.

Mhr1 overproduction prevents mtDNA deletion mutagenesis.
To verify that Abf2 and Mhr1 are required for mtDNA maintenance, we analyzed mtDNA levels relative to nuclear DNA using quantitative PCR 27 . We observed that the mtDNA level in ∆abf2 cells was less than half (46.3 ± 8.6%) that of WT cells grown in YPGly medium (Fig. 6a). Consistent with our previous results 17 , a large proportion (83.9 ± 15.3%) of mtDNA content was retained in mhr1-1 cells grown in glycerol medium, while we observed no additive effect on the depletion of mtDNA in ∆abf2 mhr1-1 double-mutant cells (51.4 ± 8.8%), suggesting Abf2 is dispensable for Mhr1-driven mtDNA replication (Fig. 6a).

Discussion
In this study, we found that increasing Mhr1 protein level prevents loss of respiratory function in cells lacking Abf2 and functional Mhr1, which display an mtDNA-instability phenotype similar to several other nuclear mutations in yeast. Our results provide further support for the notion that Mhr1 has a pivotal role in the maintenance of mitochondrial genomic integrity 30 , and that DSB-induced mtDNA replication by Mhr1 is the predominant form of mtDNA replication in ρ + yeast cells 23 . We previously reported that Mhr1-dependent mtDNA replication and homologous recombination are crucial for repair of mtDNA DSBs 22 . Collectively, our results here suggest that mitochondrial homologous DNA recombination may have utility in preventing the spontaneous generation of mtDNA deletions in a variety of circumstances (Fig. 7).
The requirement for Abf2 and Mhr1 in mtDNA stability raises the question of how increasing the amount of Mhr1 alone prevents generation of deleted mtDNA and helps to sustain cellular respiratory function. The answer likely comes from one of the mechanisms of mtDNA deletion formation proposed by Krishnan et al., in which mtDNA deletions occur during repair of DSB-induced mtDNA damage, rather than during replication 31 . We inferred that a lack of Abf2 might weaken DSB repair by homologous recombination and increase the accumulation of deleted mtDNA molecules. Since DSB repair can be accomplished by homologous DNA recombination [32][33][34] , an increased amount of Mhr1 may enhance the number of homologous DNA recombination events to repair DSBs 22 . This interpretation implies that mtDNA deletions are prevented, mtDNA integrity is maintained, and a large proportion of cells with the ∆abf2 mutation sustain respiratory function upon overexpression of MHR1 (Fig. 6). On the other hand, since mtDNA deletion mutagenesis still occurs upon Mhr1 overproduction in ∆abf2 cells, the stabilization of mtDNA recombination intermediates by Abf2 14 is important. In addition, we found that a fraction of the population of ∆abf2 mhr1-1 cells contains HS ρ − mtDNA molecules (less than 20%; Fig. 4b), however this proportion is likely large enough to reduce the beneficial effect of Mhr1 overexpression on preventing deleterious mtDNA mutations. Due to the presence of HS ρ − mtDNA mutant molecules, Mhr1 overexpression likely increases the amounts of both wild type and HS ρ − mtDNA. Thus, it is very unlikely that the observed rescue of ∆abf2 mhr1-1 cells upon MHR1 overexpression (Fig. 6) is due only to increased replication and selection for wild-type mtDNA over deleterious mtDNA mutations. These results therefore indicate that increased amounts of Mhr1 can prevent mitochondrial genomic instability. www.nature.com/scientificreports www.nature.com/scientificreports/  18,19 , while Abf2 packages mtDNA and is the main component of the nucleoid 7,35 . Both of these molecular functions resemble that of RecA in Escherichia coli. RecA forms a nucleoprotein complex that both ensures the formation of heteroduplex DNA and protects DNA from nuclease degradation 36 . Since these roles are performed by discrete proteins in S. cerevisiae mitochondria, whether mtDNA-Mhr1 nucleoprotein formation occurs prior to packaging of concatemeric mtDNA molecules by Abf2 or by a contrary process, remains for further investigation.
TFAM, the mammalian ortholog of Abf2, contains an additional C-terminal domain that has been predicted to be essential for transcription 13 . TFAM can complement Abf2 in yeast by rescuing the loss-of-mtDNA phenotype of yeast ∆abf2 cells 37 . In contrast to Abf2/TFAM, the existence of a metazoan ortholog of Mhr1 remains an open issue 38 . To date, there have been several lines of evidence to indicate that human mtDNA recombination may occur [39][40][41] . For example, we demonstrated that ROS stimulate mitochondrial allele segregation from heteroplasmy towards homoplasmy in human fibroblasts 42 , a result consistent with the stimulatory effect of ROS and mechanism of recombination-driven mtDNA replication in yeast mitochondria 21,43 . Consistent with the results reported here, a method to stimulate recombination function in human mitochondria may similarly prevent human mtDNA instability and present numerous other health benefits.

Materials and Methods
Yeast strains and media. Yeast strains used in this study are listed in Table 1. General genetic techniques used in this study are described in 44 . Yeast transformation was carried out using the lithium-acetate method 45 . The assay for hypersuppressiveness was carried out according to a procedure previously described 24,46 . Overexpression of MHR1 was achieved using a yeast multiple-copy plasmid with a constitutive promoter (pVT100U) 47 . The ORF (open reading frame) of MHR1 was amplified by PCR with addition of Sac1 and Xba1 restriction sites at the 5′ and 3′ ends respectively, and inserted into pVT100U to generate pVT100U-MHR1. Immunoblot analysis for Mhr1 detection was performed using rabbit antiserum against Mhr1, according a previously established procedure 22 .
Media were prepared as previously described 17,18,24 . Selective pre-cultivation of cells for respiratory function assays was conducted either in rich glycerol (YPGly), synthetic glycerol (Gly) or synthetic glycerol minus uracil (Gly-U) media. Fermentable media, used to promote loss of respiratory function, was rich glucose (YPD), synthetic glucose (Glu), synthetic glucose minus uracil (Glu-U), synthetic raffinose-galactose (RGal), or synthetic raffinose-galactose minus uracil (RGal-U). Glycerol, glucose, raffinose and galactose concentrations used were 3% v/v, 2%, 2% and 2%, respectively. Selection for diploid cells following crossing with ∆abf2 or ∆abf2 mhr1-1 cells was carried out by cultivating cells in liquid synthetic minimal (SD) or synthetic minimal plus tryptophan (SD + W) medium containing 2% glucose at 30 °C overnight and then spreading cells on agar plates containing the same nutrients. For control crosses, diploids were selected by spreading mated cells directly onto synthetic minimal media agar plates containing leucine plus uracil (SD + LU) or tryptophan (SD + W) and 2% glucose.
Mitochondrial nucleoid analysis. WT, mhr1-1, ∆abf2 and ∆abf2 mhr1-1 cells were cultivated in YPGly media to early log-phase at 30 °C and stained with DAPI. Cells were subsequently transferred to YPD media and cultivated for two consecutive rounds at 30 °C or 34 °C, stained with DAPI and observed again. DAPI staining was performed by transferring aliquots to fresh media containing 1 µg/ml DAPI and incubating for 15 min. Cells were then washed and resuspended in fresh media and 1% low-melting point agarose at a 1:1 ratio. Cells were then mounted on glass slides and analyzed with a Deltavision fluorescence microscopy system (Applied Precision, Inc.) equipped with an Olympus IX71 microscope. DAPI foci in mother cells were counted using ImageJ software. tetrad analysis. We used an Olympus micromanipulation system for tetrad dissection to separate the four ascus-encapsulated spores derived from diploid WT/Δmhr1 cells. Spores were separated and placed on synthetic defined (SD) complete plates supplemented with adenine (A), leucine (L), uracil (U), histidine (H) and tryptophan (W), and cultivated at 30 °C for 7 days. Colonies were then replica-plated onto SD complete plates lacking L and YPGly plates.
Purification of yeast mtDNA and analysis by restriction enzyme digestion. MtDNA was purified from Δmhr1 spore-derived yeast cells using cesium chloride density-gradient centrifugation 48 . The purified mtDNA was digested with ApaI and run on a 1% agarose gel alongside DNA size markers. The DNA fragments were photographed under ultraviolet irradiation after staining the gel with 0.5 μg/ml ethidium bromide. southern blot analysis. ∆abf2 and ∆abf2 mhr1-1 cells were cultivated in YPGly media, then transferred to RGal complete media and cultivated at 30 °C for two days. Total cellular DNA was then prepared and digested by (i) ρ + CFU formation rate based on n = 3 independent experiments described in (g). (j) Scheme of extended respiratory function assay. Cells were selectively grown in Gly-U media, then 10 5 cells were transferred to RGal-U media and cultivated for two consecutive 48-hour rounds (approximately 20 generations) at 30 °C. Cells were then spread onto RGal-U plates, grown for four to seven days, and then replica-plated onto YPGly plates. (k) Representative plate images for n = 3 independent experiments as described in (j). (l) Scheme of extended respiratory function assay. Cells were selectively grown in Gly-U media, then 10 5 cells were transferred to Glu-U media and cultivated for two consecutive 48-hour rounds (approximately 20 generations) at 30 °C. Cells were then spread onto RGal-U plates, grown for four to seven days, and then replica-plated onto YPGly plates. (m) Representative plate images for n = 2 independent experiments as described in (l). All error bars indicate ± SD. Figure 7. Model for the prevention of mtDNA deletion mutagenesis by Mhr1-driven recombination and mtDNA replication. MtDNA deletions threatening cellular respiratory function arising from genetic backgrounds such as ∆abf2 mhr1-1 can be prevented by increased homologous recombination via overproduction of Mhr1 recombinase.