|Home | About | Journals | Submit | Contact Us | Français|
Mitochondria contain their own genome, the integrity of which is required for normal cellular energy metabolism. Reactive oxygen species (ROS) produced by normal mitochondrial respiration can damage cellular macromolecules, including mitochondrial DNA (mtDNA), and have been implicated in degenerative diseases, cancer, and aging. We developed strategies to elevate mitochondrial oxidative stress by exposure to antimycin and H2O2 or utilizing mutants lacking mitochondrial superoxide dismutase (sod2Δ). Experiments were conducted with strains compromised in mitochondrial base excision repair (ntg1Δ) and oxidative damage resistance (pif1Δ) in order to delineate the relationship between these pathways. We observed enhanced ROS production, resulting in a direct increase in oxidative mtDNA damage and mutagenesis. Repair-deficient mutants exposed to oxidative stress conditions exhibited profound genomic instability. Elimination of Ntg1p and Pif1p resulted in a synergistic corruption of respiratory competency upon exposure to antimycin and H2O2. Mitochondrial genomic integrity was substantially compromised in ntg1Δ pif1Δ sod2Δ strains, since these cells exhibit a total loss of mtDNA. A stable respiration-defective strain, possessing a normal complement of mtDNA damage resistance pathways, exhibited a complete loss of mtDNA upon exposure to antimycin and H2O2. This loss was preventable by Sod2p overexpression. These results provide direct evidence that oxidative mtDNA damage can be a major contributor to mitochondrial genomic instability and demonstrate cooperation of Ntg1p and Pif1p to resist the introduction of lesions into the mitochondrial genome.
The mitochondrial genome of eukaryotic cells encodes integral components of the ATP-producing oxidative phosphorylation pathway. Maintenance of the mitochondrial genome is critical to ensure proper functioning of this organelle in mediating cellular energy demands (2). However, as a consequence of normal aerobic respiration, reactive oxygen species (ROS) are produced and create a highly oxidative environment (18). Upon exposure to ROS, oxidative modification to biomolecules, including DNA, can result (53).
Various defense mechanisms operate in mitochondria to minimize the deleterious effects of oxidative stress and maintain mitochondrial genomic integrity. Primary defenses against oxidative damage include protective proteins and small molecules that scavenge ROS or sequester metal ions. The antioxidant enzyme superoxide dismutase is involved in the conversion of superoxide anion to dioxygen and hydrogen peroxide, which is further detoxified by catalase or peroxidase (24). Saccharomyces cerevisiae possesses two genes encoding superoxide dismutase, one of which, Sod2p, specifically localizes to the mitochondrial matrix and is believed to protect mitochondria against oxidative stress (19). Additionally, secondary defenses consist of enzymes that repair the products of oxidatively damaged components (35). If oxidative stress defense mechanisms are compromised or overwhelmed, disease states may ensue. It has been proposed that mitochondrial mutations, presumably resulting from oxidative mitochondrial DNA (mtDNA) damage, contribute to neurodegenerative disorders, cancer, and the aging process in humans (12, 47, 55).
Repair of nuclear and mitochondrial oxidative DNA lesions occurs predominantly through the base excision repair (BER) pathway (37). As demonstrated by mammalian in vitro repair studies, the overall pathway for mitochondrial BER is similar to the short patch pathway of BER in the eukaryotic nucleus (39, 51). BER is initiated by the action of DNA N-glycosylases that specifically recognize and excise damaged bases from DNA. Release of the damaged base creates an apurinic/apyrimidinic (AP) site that can be further processed by an N-glycosylase-associated AP lyase or a separate, hydrolytic AP endonuclease. Following appropriate 3′ or 5′ end-trimming steps, DNA polymerase inserts the correct base and DNA ligase seals the nick, thereby repairing the damage (4). Mitochondrial BER in S. cerevisiae presumably occurs via a similar mechanism and is supported by the observation that several components of the yeast nuclear BER machinery colocalize to mitochondria, including the major AP endonuclease, Apn1p (54), as well as the N-glycosylases, Ogg1p (48) and Ntg1p (1, 56), which initiate repair of oxidatively damaged purines and pyrimidines, respectively. In addition to BER, other yeast mitochondrial repair and damage tolerance pathways have been implicated in handling oxidative mtDNA damage, such as recombination (3) as well as recombination-independent functions of proteins such as Abf2p and Pif1p (38).
Pif1p is a 5′-to-3′ DNA helicase that localizes to both the nucleus and mitochondria (27, 28). With respect to its role in mitochondria, Pif1p is important for maintenance, repair, and recombination of mtDNA (27). Recently, we demonstrated that the recombination-independent activities of Pif1p cooperate with BER to resist spontaneous oxidative mtDNA damage (38). Our previous analysis of the mitochondrial genomic instability phenotype of pif1Δ mutants led us to postulate a role for this unique helicase in mtDNA repair, through governing the rate of mtDNA replication or regulating mtDNA copy number (38). However, the extent to which these proteins cooperate under conditions of chronic oxidative stress is unknown.
When pathways involved in maintaining nuclear genomic integrity are corrupted, many deleterious consequences ensue, including chromosomal aberrations, cellular transformation, and death (9). Additionally, recent studies have implicated ROS in nuclear gross chromosomal rearrangements (22, 49). However, studies examining potentially similar events during the loss of mitochondrial genomic integrity resulting from oxidative stress are limited. S. cerevisiae is an ideal model system for delineating the basic molecular mechanisms underlying the maintenance of mtDNA. Yeast are facultative anaerobes; therefore, cells that are devoid of functional mitochondrial respiration (resulting from genomic instability) can still be propagated and studied when a fermentable carbon source, such as glucose, is provided (45). In contrast, respiration-deficient mammalian cells are typically inviable (43).
Although oxidative stress is thought to contribute to mitochondrial mutagenesis, to date very few studies have directly addressed molecular mechanisms by which mtDNA oxidative damage leads to genomic instability or have elucidated the involvement of proteins responsible for maintaining mitochondrial genomic integrity. We have employed strategies to selectively elevate levels of mitochondrial oxidative DNA damage through exploitation of genetic mutants lacking pathways involved in mtDNA maintenance to determine the biological consequences of mitochondrial oxidative stress and to evaluate the relative contributions of these proteins toward prevention of oxidative DNA damage in mitochondria. Our results directly demonstrate that oxidative DNA damage causes mtDNA instability and indicate that Ntg1p and Pif1p are collectively involved in maintaining mitochondrial genomic integrity under conditions of chronic oxidative stress.
All yeast strains used in this experiment are derivatives of DBY2006 (α his3Δ200 leu2-3,112 ura3-52 trp1-Δ1 ade2) with the exception of BS127 (α ade1 ade2 leu2 trp1 ura3 cyhR canR sapR supR gal). Wild type (WT) refers to the TWO1 strain (DSC0121) (α his3Δ200 leu2-3,112 ura3-52 trp1-Δ1 ade2 ntg1Δ::KanMX4 pRS316-NTG1 [URA3 CEN/ARS]), which contains a chromosomal disruption of the NTG1 locus that is covered by a plasmid containing a wild-type copy of the NTG1 gene expressed from its own promoter. This strain is functionally wild type due to the plasmid copy of the NTG1 gene complementing the chromosomal disruption. The following strains are derivatives of TWO1 constructed by standard marker disruption and/or plasmid shuffle as previously described: DSC0122 (ntg1Δ), DSC0125 (pif1Δ), and DSC0126 (pif1Δ ntg1Δ) (38).
SOD2 gene disruptions were created using a PCR-based gene deletion system as described by Longtine et al. (32). The following strains were generated: the sod2Δ strain (DSC0245) (α his3Δ200 leu2-3,112 ura3-52 trp1-Δ1 ade2 ntg1Δ::KanMX4 sod2Δ::TRP1 pRS316-NTG1 [URA3 CEN/ARS]), the ntg1Δ sod2Δ strain (DSC0246) (α his3Δ200 leu2-3,112 ura3-52 trp1-Δ1 ade2 ntg1Δ::KanMX4 sod2Δ::TRP1), the pif1Δ sod2Δ strain (DSC0247) (α his3Δ200 leu2-3,112 ura3-52 trp1-Δ1 ade2 ntg1Δ::KanMX4 pif1Δ::HIS3 sod2Δ::TRP1 pRS316-NTG1 [URA3 CEN/ARS]), and the pif1Δ ntg1Δ sod2Δ strain (DSC0248) (α his3Δ200 leu2-3,112 ura3-52 trp1-Δ1 ade2 ntg1Δ::KanMX4 pif1Δ::HIS3 sod2Δ::TRP1). During strain construction, all strains were maintained on yeast extract-peptone-glycerol (YPG) medium to sustain mitochondrial respiration competence.
To create a strain that overexpressed Sod2p from its own promoter, BS127 was transformed with 2μm plasmid p42K-TEF (DSC0250) (Dual Systems Biotech).
All strains were grown in standard synthetic dextrose (SD), synthetic glycerol, yeast extract-peptone-dextrose (YPD), or YPG as previously described (46).
The conditions of the dichlorofluorescein diacetate (DCFHDA) assay were adapted from the protocol of Brennan and Schiestl (8). This assay utilizes the oxidant-sensitive probe DCFHDA to assess intracellular ROS levels. Individual colonies of various strains were used to inoculate 10 ml YPG. Cultures were grown at 30°C in a roller drum until an optical density at 60 nm (OD600) of 0.5 was achieved. Cell counts were taken for each sample to normalize reported relative fluorescence values. Cells were washed twice, resuspended in 10 ml distilled H2O, and divided into 1.5-ml aliquots. When applicable, cells were treated with various concentrations of antimycin and H2O2 for 1 h at 30°C. Next, DCFHDA was added from a 5 mM stock in ethanol to a final concentration of 10 μM. Cells were incubated at 30°C for 30 min. Following treatments, an aliquot of cells was removed, washed, diluted, and plated to determine the fraction of survival. Next, each sample was washed twice and resuspended in 1.5 ml of a solution containing 1% sodium dodecyl sulfate, 2% Triton X-100, 100 mM NaCl, 10 mM Tris, pH 8.0, and 1 mM EDTA. After addition of 0.3 g acid-washed glass beads, samples were vortexed for 10 min. Cells were incubated at room temperature for 10 min and then pelleted. Fluorescence of 200 μl of the supernatant was measured using a Packard Fluoro-count fluorescence multiwell plate reader with a fluorescence excitation of 485 nm and emission at 520 nm. Reported average results were obtained from at least four independent cultures.
Cells were grown in YPG and isolated during log phase. Approximately 2 × 109 cells were harvested by centrifugation and resuspended in 0.8 ml SorEDTA (0.9 M sorbitol, 0.1 M EDTA [pH 7.4]) containing 25 μl dithiothreitol (1 M) and 100 μl zymolyase 20T (10 mg/ml). Cells were incubated overnight at 37°C. Spheroblasts were resuspended in 0.5 ml Tris-EDTA (50 mM Tris-HCl [pH 7.4], 20 mM EDTA) containing 50 μl of 20-mg/ml proteinase K and incubated at 55°C for 1 h. One hundred microliters of 10% sodium dodecyl sulfate was added, and samples were incubated for 20 min at 65°C. After addition of 300 μl of 5 M potassium acetate, samples were incubated on ice for 30 min. Supernatants from each sample were twice extracted with chloroform. DNA was precipitated with isopropanol, resuspended in TE buffer (10 mM Tris, 1 mM EDTA [pH 8]) and digested with 5 μl RNase A (10 mg/ml) at 37°C for 1 h. Samples were precipitated with ethanol, washed twice, and resuspended in TE buffer. DNA was quantified by fluorometry using the fluorescent dye bisbenzamide (Hoechst 33258) (Sigma DNA quantitation kit, product no. DNA-QF). Fluorometry was performed according to the manufacturer's recommendations.
This electrophoretic assay determines the levels of Ntg1p recognizable lesions present within a coding region of the mitochondrial genome and is an adaptation of a similar method previously used by our group for measuring oxidative nuclear DNA damage (15). Ten-microgram aliquots of genomic DNA were digested with NdeI for 4 h at 37°C. Digestion with this enzyme yields a 4.4-kb fragment containing the mitochondrially encoded COB1 gene. After digestion, DNA was precipitated, pooled, and quantitated. Ten micrograms of NdeI-digested genomic DNA was subsequently incubated with 5 μg recombinant glutathione S-transferase-tagged Ntg1p in a 25-μl reaction mixture containing 15 mM KH2PO4, pH 6.8, 10 mM EDTA, 10 mM β-mercaptoethanol, and 40 mM KCl. The reaction was performed at 37°C for 30 min (56). Reactions were terminated by heating to 60°C for 5 min. The recombinant Ntg1p used in these experiments was N-terminally tagged with glutathione S-transferase and expressed in Escherichia coli as described previously (34). Ntg1p was purified by glutathione-agarose chromatography followed by Mono-S fast-performance liquid chromatography as previously described (34).
Restricted DNA samples were resolved on a 1.1% denaturing, alkaline agarose gel (30 mM NaOH, 1 mM EDTA) using alkaline electrophoresis buffer (30 mM NaOH, 2 mM EDTA) and denaturing loading dye (50 mM NaOH, 1 mM EDTA). Electrophoresis was performed for 30 min at 40 V followed by 15 h at 17 V (6). Additionally, an internal standard (loading control) corresponding to a 400-bp segment of the COB1 gene, generated by PCR amplification, was added to each sample after Ntg1p incubation at a concentration of 0.15 ng per μg of genomic DNA. Transfer and Southern hybridization were performed according to the method of Sambrook and Russell (42) with the following modifications. Transfer was performed utilizing alkaline transfer buffer (20× SSC [1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate], 0.4 N NaOH). After 24 h of transfer, DNA was UV cross-linked to nylon membranes and hybridized at 65°C in rapid-hyb buffer (Amersham Biosciences). A 400-bp PCR-generated fragment of the COB1 gene was labeled with [α-32P]dATP using randomly primed synthesis. Band intensity was determined by phosphorimager analysis and normalized to the COB1 loading control. The number of Ntg1p-recognized lesions were determined as previously described (5). Average values reported were obtained from at least four independent experiments.
The conditions of the erythromycin resistance assay were adapted from the protocol of Chi and Kolodner (10). Five milliliters of YPG medium was inoculated with individual colonies from the indicated strains, and the cultures were grown at 30°C until an OD600 of 2 to 4 was reached. Cultures were diluted and plated onto YPG grown at 30°C for 4 to 5 days to obtain single colonies. Ten to fifteen individual colonies were used to inoculate separate 10-ml cultures of YPG grown at 30°C for 48 h. Samples were plated onto YPG containing 1 mg/ml erythromycin and grown at 30°C for 6 to 7 days. Additionally, a small sample of each culture was removed to determine the total number of respiration-competent cells by plating onto YPG. Mutation frequencies were calculated as follows: number of erythromycin-resistant colonies/total number of colonies. Potential jackpot cultures were statistically identified as outliers by using the Grubbs test (21). Reported average frequencies are calculated from 30 to 45 independent cultures, and standard error of the mean (SEM) is indicated.
Individual colonies were used to inoculate 5 ml of YPD or SD where indicated. One-milliliter aliquots were removed during late-logarithmic phase growth (OD600, ~0.5 to 0.6) for microscopic evaluation. Cells were washed twice and resuspended in 70% ethanol. One microliter of 1-mg/ml diaminophenylindole (DAPI) was added, and cells were incubated at room temperature for 5 min. Cultures were washed twice before examination by microscopy at magnification ×100.
The basis for the petite-mutant assay is that growth of yeast cells in YPG (glycerol-containing) medium requires mitochondrial respiration, whereas growth on YPD (glucose-containing) medium can occur in the absence of mitochondrial respiration. To determine the rate of spontaneous petite formation over time, cultures of various yeast strains were grown to near-saturation (stationary phase) in 5 ml of YPG medium to maintain respiration competence. The culture was then diluted to an OD600 of 0.005 into 100 ml SD medium with appropriate nutritional supplements as required by each strain. Samples from the experimental cultures were removed immediately (time zero) and every 24 h thereafter during growth at 30°C with agitation (200 rpm) on a rotary shaker. Identical samples from each culture were diluted and plated onto YPD and YPG plates and allowed to grow at 30°C for 2 to 5 days until colonies formed. The number of colonies on each plate was determined (100 to 1,000 CFU/plate in all experiments), and the percentage of respiration-competent cells was calculated as follows: (number of colonies on YPG/number of colonies on YPD) × 100.
Aliquots of cells were streaked onto YPD plates at each time point to assess cell viability. Plates were allowed to grow at 30°C for 24 h and then viewed at ×40 magnification. Approximately 300 colonies were counted and assessed as viable (formed colony of greater than five cells) or nonviable (failed to form colony of greater than five cells). Additionally, cells were analyzed by DAPI staining and fluorescence microscopy at each time point as described above.
Indicated strains were grown to mid-logarithmic phase (OD600 of ~0.5). Cells were washed twice and resuspended in deionized H2O together with the indicated concentrations of antimycin and H2O2. Cells were incubated at 30°C for 1 h. After treatment, cells were washed twice then resuspended in deionized H2O. Cells were subsequently analyzed for petite formation, viability, and mitochondrial DNA status as appropriate.
In order to assess the biological endpoints of chronic oxidative stress in mtDNA repair-deficient mutants, we developed a system to selectively enhance mitochondrial oxidative stress utilizing endogenous and exogenous sources of ROS. To induce conditions of oxidative stress, cells were exposed to antimycin in combination with H2O2. Antimycin, a drug that selectively inhibits respiratory complex III, has been previously shown to enhance mitochondrial superoxide production in isolated organelles (7) as well as in vivo (41). To enhance ROS levels via endogenous mechanisms, mutants with a deleted mitochondrial superoxide dismutase (SOD2) gene were created. Through its enzymatic scavenging of superoxide anion radicals, Sod2p mediates protection of yeast against oxygen toxicity in the mitochondria (20). To determine the extent to which the levels of intracellular ROS could be manipulated, the oxidant-sensitive fluorescent probe DCFHDA was utilized (57). Both sod2Δ and ntg1Δ sod2Δ strains displayed an approximately threefold increase in the amount of fluorescence over that of their isogenic parental strains (Fig. (Fig.1),1), indicating that sod2 mutants display a significant increase in intracellular ROS levels. The level observed in sod2Δ deletion strains was approximately equivalent to treatment with 10 μg/ml antimycin, a concentration that results in about 80% survival (data not shown). When antimycin treatment was used in combination with H2O2 exposure, there was a synergistic increase in ROS levels compared to treatment with either agent alone (Fig. (Fig.1).1). This result suggests that the mitochondrial specificity of antimycin can be exploited to elevate the level of mitochondrial ROS when used in combination with H2O2, resulting in a model system that can be used to study mitochondrial oxidative stress. However, it should be noted that the proportion of ROS present in the mitochondria is unknown.
To investigate the relationship between oxidative stress and oxidative mtDNA damage, we utilized a mitochondrial gene-specific DNA damage detection assay that measures the amount of Ntg1p-recognizable lesions in a 4.4-kb fragment of the mitochondrial genome. Total cellular DNA was obtained using a method that minimizes introduction of adventitious oxidative DNA damage (15). DNA was subsequently treated with Ntg1p, a BER DNA glycosylase with associated AP lyase activity that primarily recognizes and cleaves DNA containing oxidative pyrimidine base damage and abasic sites (56). Under the conditions of denaturing agarose gel electrophoresis, such cleavage products produce DNA single-strand breaks at sites of unrepaired base damage recognized by Ntg1p (15). When probed with the mitochondrial COB1 gene, the extent of hybridization signal decrease is proportional to the amount of Ntg1p-recognizable lesions present within this locus and can be extrapolated to estimate the number of lesions per mitochondrial genome (5).
Experiments were conducted with strains previously shown to be compromised in mitochondrial BER (ntg1Δ) and oxidative mtDNA damage resistance (pif1Δ) under conditions of chronic oxidative stress achieved by elimination of Sod2p, a major mitochondrial ROS-scavenging enzyme. A small, but reproducible increase (~2-fold) in Ntg1p-recognizable lesion levels was observed in the ntg1Δ null mutant strain (Table (Table1).1). In contrast, higher levels (sevenfold) of Ntg1p-recognizable lesions were detected in the pif1Δ mutant strain, and this increase was further enhanced (greater than additive) in the pif1Δ ntg1Δ double mutant strain (Table (Table1).1). Such increases in the levels of Ntg1p-recognizable lesions when considered with previously published reports (38) demonstrating a similar pattern of increase in mtDNA mutagenesis and petite induction in these strains suggest that Ntg1p and Pif1p are important mediators for the prevention of oxidative mtDNA damage and mutagenesis. In addition, simultaneous elimination of Ntg1p and Sod2p (ntg1Δ sod2Δ mutant) results in a greater than additive increase in the number of oxidative lesions present within the mitochondrial genome (Table (Table1).1). It is important to note these experiments cannot be performed with the pif1Δ sod2Δ double mutant and the ntg1Δ pif1Δ sod2Δ triple mutant strain because such petite mutants either possess corrupt mtDNA genomes (rho−) or lack mtDNA altogether (rho0). Together, these results indicate that the mitochondrial genome is a major target of oxidative stress resulting in the generation of deleterious oxidative DNA lesions.
Erythromycin is an antibiotic that specifically inhibits mitochondrial translation (11). Resistance to the drug can be acquired through specific point mutations in the mitochondrial rib2 and rib3 loci, which encode the two mitochondrial rRNAs (14, 50). Therefore, acquisition of erythromycin resistance is a convenient, direct measurement of mtDNA point mutagenesis (10). Using this assay, we determined mtDNA mutation frequencies in ntg1Δ and sod2Δ single-mutant strains as well as in the corresponding double mutant. The ntg1Δ mutant displays a slight increase in mutation frequency (Table (Table2)2) as previously reported (38). The sod2Δ mutant also exhibits an increased mitochondrial mutation frequency (~5-fold) that increases synergistically when NTG1 is also deleted (~15-fold). This result strongly supports an in vivo biological function for the repair of mutagenic, oxidative mtDNA damage by Ntg1p. Additionally, it indicates a direct relationship between enhanced mitochondrial oxidative stress, oxidative DNA damage, and mtDNA mutagenesis.
Petite mutants caused by mtDNA dysfunction can be distinguished into two phenotypic subclasses: rho− and rho0. rho− mutants possess mtDNA, albeit in an altered form that results from large deletions followed by amplification of remaining sequences. Such mutants are incapable of carrying out normal mitochondrial respiratory function (40). rho− and rho0 subclasses are distinguishable from each other by microscopic analysis of DAPI-stained, respiration-compromised cells revealing the presence (rho−) or absence (rho0) of mtDNA.
We sought to determine the rho status of mutants compromised in DNA damage resistance mechanisms in order to ascertain the relative contribution of oxidative DNA damage to mitochondrial genomic stability. Wild-type, ntg1Δ, sod2Δ, and ntg1Δ sod2Δ mutants were all characterized as being rho positive (rho+) under late-log-phase growth conditions (Fig. (Fig.2).2). Both pif1Δ and ntg1Δ pif1Δ were characterized as a mixed population of rho+ and rho− cells due to the fact that these strains exhibit a high rate of spontaneous petite formation as previously reported (38). The pif1Δ sod2Δ double mutant was characterized as rho− due to the presence of mtDNA and a complete inability to grow in YPG (Fig. (Fig.2).2). The loss of mtDNA genomic integrity is even more severe in the ntg1Δ pif1Δ sod2Δ triple mutant, since these cells exhibit a rapid and total loss of mtDNA (Fig. (Fig.2).2). These data, when taken together with the assessment of oxidative mtDNA damage levels, suggest that oxidative mtDNA damage directly contributes to cellular petite formation and complete loss of mtDNA.
We have previously reported that eliminating both Ntg1p and Pif1p results in a synergistic increase in spontaneous petite formation (38). Therefore, it was of interest to determine whether a similar synergistic effect occurs in response to induced, acute oxidative stress. For this purpose, WT, ntg1Δ, pif1Δ, and ntg1Δ pif1Δ mutants were exposed to antimycin and H2O2, a combination of agents that enhances mitochondrial oxidative stress. Following such treatment, the percentage of petite mutants observed in both ntg1Δ and pif1Δ single mutants exposed to oxidative stress does not significantly change (Fig. (Fig.3).3). In contrast, the ntg1Δ pif1Δ double-mutant strain displayed a synergistic petite induction in response to oxidative stress (Fig. (Fig.3),3), indicating that Ntg1p and Pif1p together provide resistance to induced oxidative DNA damage. Additionally, these data suggest that the ability of oxidative mtDNA damage to cause genetic instability can be enhanced under conditions of induced mitochondrial oxidative stress by exposure to exogenous agents such as antimycin and H2O2.
The above results indicate that Ntg1p-mediated mtDNA excision repair, Sod2p-mediated ROS scavenging, and Pif1p-mediated functions are all important events involved in maintaining mitochondrial genomic integrity. In order to assess the relative contribution of each of these components in preserving mtDNA stability, we assessed the length of time required for a yeast cell with these systems compromised to change from a respiration-competent (rho+) phenotype to a respiration-compromised (petite) phenotype. Mitochondrial respiration incompetence resulting from mtDNA damage can be considered an early, measurable event in the processes leading to a genome-wide corruption of mtDNA. Experiments were conducted in stationary-phase cultures in order to maximize the relative contribution of Sod2p, which has been shown to be induced under stationary-phase conditions, to counteract increases in ROS (13). Additionally, stationary-phase cells are a particularly useful model system for the study of damage that occurs during oxidative stress and disease states because this allows for time-dependent damage accumulation in these nondividing cells (30, 31).
WT cells and the ntg1Δ single mutant maintained mitochondrial genomic integrity throughout the 12-day course of the experiment (Fig. (Fig.4).4). The ntg1Δ sod2Δ double mutant converted into a predominantly petite population more rapidly than the sod2Δ single mutant alone (Fig. (Fig.4,4, day 10 versus day 12, respectively). The pif1Δ mutant displayed a rapid loss of mitochondrial genomic integrity, which was further enhanced when simultaneously deleted for NTG1. Strikingly, the ntg1Δ pif1Δ double mutant completely converted into a respiration-incompetent population within 4 days (Fig. (Fig.4).4). The rapid loss of respiration competency in cells compromised for Pif1p function reveals its important contribution to the maintenance of mitochondrial genomic stability.
Because oxidative mtDNA damage appears to be a major contributor to mitochondrial genomic instability, it was of particular interest to determine whether ROS could cause a stable rho− cell line, with its complement of mtDNA damage resistance pathways intact, to undergo a complete loss of mtDNA. The strain BS127 is able to stably maintain a nonfunctional (rho−) mitochondrial genome comprised of multiple ori5 repeats (33). When BS127 cells were exposed to 1 μg/ml antimycin and 10 mM H2O2, the cells converted to a rho0 phenotype after 8 to 10 generations of outgrowth as determined by microscopic analysis of DAPI-stained mtDNA (Fig. (Fig.5B).5B). Isogenic cells overexpressing Sod2p from a 2μm plasmid maintained their mitochondrial genomes despite receiving equivalent exposures to mitochondrial oxidative stress (Fig. (Fig.5D).5D). These results indicate that the loss of mitochondrial genomic integrity can be directly attributed to ROS present in mitochondria.
To gain insight into how eukaryotic cells resist damage to the mitochondrial genome, we utilized two strategies to elevate mitochondrial oxidative stress using both endogenous and exogenous sources of ROS. In response to conditions of mitochondrial oxidative stress, eukaryotic cells display a range of measurable, biologically relevant endpoints, including oxidative mtDNA damage and mutagenesis, which contribute to corruption of mitochondrial genome integrity. Strikingly, strains simultaneously missing three components of pathways involved in preventing and repairing oxidative mtDNA damage exhibit a rapid and total loss of mtDNA, revealing the requirement for maintaining mitochondrial genome integrity in the context of oxidative stress. Furthermore, to directly demonstrate whether mitochondrial oxidative stress causes loss of mitochondrial genomic stability, a stable rho− strain completely lost mtDNA (i.e., converted to rho0) upon exposure to antimycin and H2O2 (Fig. (Fig.55).
In addition to demonstrating that oxidative mtDNA damage directly contributes to loss of mitochondrial genome integrity, we assessed the relative contributions of three pathways involved in maintenance of mitochondrial genome integrity: Sod2p-mediated ROS scavenging, Ntg1p-mediated mitochondrial BER, and Pif1p-mediated oxidative mtDNA damage resistance. The data presented here suggest that these three pathways operate in a delicate balance to ensure mitochondrial genomic integrity (Fig. (Fig.6).6). Sod2p maintains mitochondrial genomic integrity by preventing oxidative mtDNA damage through its antioxidant catalytic properties. However, if mitochondrial ROS scavenging systems are not operating at maximum capacity, ROS have the potential to damage mtDNA. Oxidative mtDNA base damage can be repaired by mtDNA excision repair activities mediated by Ntg1p, thereby restoring mitochondrial genomic integrity. Pif1p is a 5′-to-3′ DNA helicase that has been implicated in mtDNA maintenance, repair, and recombination (17). In addition to its reported mitochondrial functions, the nuclear functions of Pif1p include inhibition of bidirectional rRNA gene replication (23), telomere length regulation (44), and prevention of gross chromosomal rearrangements (36). Based on its known nuclear activities, it has been postulated that Pif1p may govern the rate of mtDNA replication (3, 38). Additionally, because of the multifunctional nature of Pif1p, it may be involved in several aspects of mitochondrial genomic maintenance. In the context of mitochondrial oxidative stress and subsequent mtDNA damage, Pif1p may inhibit replication progression in order to allow time for repair to occur. In the absence of such a “checkpoint,” replication may proceed through unrepaired damage-containing DNA templates, resulting in mutagenesis and stimulating recombination, both of which constitute genomic instability. Pif1p may also promote tolerance of oxidative mtDNA damage through its recombinational activities or other novel mechanisms. Thus, if the balance of these mtDNA damage management pathways is disrupted, deleterious biological endpoints occur, including mtDNA oxidative damage, mutagenesis, and extensive mitochondrial genomic corruption. Additionally, other factors, such as the mismatch repair protein, Msh1p, have been shown to play a role in maintaining mitochondrial genomic stability (10).
By exploiting genetic analysis in S. cerevisiae, it is possible to assess the relative contributions of each pathway to maintenance of mitochondrial genome integrity. A range of phenotypes was observed that were dependent upon the combination of DNA damage resistance pathway components eliminated. WT and single pathway-deficient strains display low levels of oxidative mtDNA damage (Table (Table1)1) and mtDNA mutagenesis (Table (Table2)2) and no overt growth defects and maintain a rho+ mitochondrial phenotype (Fig. (Fig.2).2). These results suggest that Ntg1p, Pif1p, and Sod2p are part of an interactive network of mtDNA damage resistance pathways. In the absence of one pathway, the remaining pathways serve as a buffer to maintain mitochondrial genomic integrity. However, when multiple pathways are compromised, as in the pif1Δ sod2Δ double mutant and the ntg1Δ pif1Δ sod2Δ triple mutant, genomic instability occurs substantially more rapidly (Fig. (Fig.22).
Our results revealed important differences with respect to the relative contributions of each pathway in the maintenance of mitochondrial genomic integrity, with Pif1p mediating a major role. The pif1Δ single mutant displays elevated levels of oxidative mtDNA damage (Table (Table1)1) and mutagenesis (38); levels that often exceed those found in other double pathway-deficient strains (Table (Table2,2, Fig. Fig.4)4) (38). The most severe phenotypes observed in this study always occurred in the context of a pif1Δ mutant with one additional pathway eliminated. For example, substantial levels of oxidative mtDNA damage and elevated mutagenesis are observed in the ntg1Δ pif1Δ double mutant strain (Table (Table1)1) (38). Furthermore, pif1Δ sod2Δ double mutants exhibit extensive mitochondrial genome instability (Fig. (Fig.2).2). The extreme phenotypic severity of pif1Δ mutants may reflect the multifunctional role that Pif1p mediates in response to mitochondrial oxidative stress, including governing replication rates, controlling recombination, and other potential activities, such as influencing the accessibility of mtDNA to repair proteins. The data presented here are consistent with the notion that Pif1p is a major contributor to mitochondrial genomic stability.
By comparing the endpoints of genomic stability in the stains analyzed, a sequence of events is revealed, including a progression from rho+ to rho− to rho0 that is dependent upon the combination of pathways eliminated (Fig. (Fig.2).2). For example, the pif1Δ sod2Δ double mutant maintains its mitochondrial genome, albeit in a compromised (rho−) form. However, when an additional mtDNA damage resistance pathway is eliminated (ntg1Δ) and additional mtDNA damage is introduced into the genome, these cells exhibit a rapid and total loss of mtDNA (Tables (Tables11 and and2).2). One interpretation of these results is that a putative DNA damage threshold level has been surpassed in these strains, leading to progressive genomic instability. While the exact sequence and chronology of events involved in the loss of mitochondrial genomic stability induced by oxidative stress is not entirely clear, several speculations can be made. At moderate levels of oxidative DNA damage and mutation frequencies, such as those observed in the ntg1Δ and pif1Δ single-mutant strains, it is likely that mtDNA excision repair has reached a maximum capacity (Tables (Tables11 and and2).2). As a compensatory response, recombination may be stimulated in response to collapsed replication forks in order to tolerate increased levels of oxidative DNA damage. This model is supported by the fact that recombination is involved to some degree for tolerance of oxidative damage in yeast mitochondria (29), as well as the demonstration that recombination rates are substantially increased in the absence of base excision repair in the nuclei of yeast cells (52). It can be postulated that increased recombination of a moderately damaged genome (such as those lacking components of one or more mtDNA damage resistance pathways) could promote illegitimate recombination, thereby resulting in a rearranged genome typical of those present within a rho− population of cells (16). Once a cell becomes rho− and enters a state of chronic oxidative stress with increasing, progressive oxidative DNA damage, an eventual complete loss of mtDNA occurs when mtDNA polymerase is unable to extend past a lesion on the template strand. Thus, total loss of mtDNA within the context of a rho− phenotype can be achieved through the inability of mtDNA polymerase to replicate due to the presence of extensive, unrepairable mtDNA damage. It has also been postulated that extensively damaged mtDNA can be selectively targeted for degradation, which would lead to a rho0 phenotype (26).
To date, very few studies have systematically addressed the contribution of oxidative mtDNA damage to the loss of mitochondrial genomic integrity resulting in defective respiration competency (rho−) and complete loss (rho0) of the mtDNA genome. Yeast provides an attractive eukaryotic model system for addressing the consequences of oxidative stress with regard to mtDNA stability, since mtDNA is dispensable for cell growth. In contrast, loss of mitochondrial genomic stability in human cells is a lethal event; therefore, studies with yeast offer a great advantage over similar studies with mammalian cells. Although petite formation is not a biological outcome associated with mitochondrial defects in humans, it provides an observational window to assess the consequences of mtDNA damage and provides a direct indication of mtDNA functionality. As an increasing number of human diseases are being identified that are associated with mtDNA-linked mutations (55), it is important to understand processes involved in maintaining mitochondrial genomic integrity. Knowledge regarding mechanisms of mitochondrial genome maintenance may provide valuable insights into the generation of mitochondrium-based diseases as well as understanding the contribution of mitochondrial dysfunction in cancer and aging.
This work was supported by NIH grants ES11163 and CA07341.
We appreciate Thomas O'Rourke for his contribution to strain construction.