|Home | About | Journals | Submit | Contact Us | Français|
DNA polymerase γ (pol γ) is responsible for replication and repair of mitochondrial DNA (mtDNA). Over 150 mutations in POLG (which encodes pol γ) have been discovered in patients with mitochondrial disorders including Alpers, progressive external ophthalmoplegia and ataxia-neuropathy syndrome. However, the severity and dominance of many POLG disease-associated mutations are unclear, because they have been reported in sporadic cases. To understand the consequences of pol γ disease-associated mutations in vivo, we identified dominant and recessive changes in mtDNA mutagenesis, depletion and mitochondrial dysfunction caused by 31 mutations in the conserved regions of the gene, MIP1, which encodes the Saccharomyces cerevisiae ortholog of human pol γ. Twenty mip1 mutant enzymes were shown to disrupt mtDNA replication and may be sufficient to cause disease. Previously uncharacterized sporadic mutations, Q308H, R807C, G1076V, R1096H and S1104C, caused decreased polymerase activity leading to mtDNA depletion and mitochondrial dysfunction. We present evidence showing a limited role of point mutagenesis by these POLG mutations in mitochondrial dysfunction and disease progression. Instead, most mitochondrial defective mip1 mutants displayed reduced or depleted mtDNA. We also determined that the severity of the phenotype of the mip1 mutant strain correlates with the age of onset of disease associated with the human ortholog. Finally, we demonstrated that increasing nucleotide pools by overexpression of ribonucleotide reductase (RNR1) suppressed mtDNA replication defects caused by several dominant mip1 mutations, and the orthologous human mutations revealed severe nucleotide binding defects.
Mitochondrial DNA (mtDNA) maintenance is necessary for the majority of ATP production in eukaryotic cells. Various mitochondria-related diseases such as Alpers syndrome, progressive external ophthalmoplegia (PEO) and ataxia-neuropathy syndrome are characterized by mtDNA depletion, deletions or point mutations, implicating defective mtDNA replication (1). Of the 16 known human DNA polymerases, only pol γ is known to replicate mtDNA (2–4), and over 150 mutations in the gene encoding pol γ, POLG, have been identified in mitochondrial disease patients (http://tools.niehs.nih.gov/polg/) (5,6). Pol γ-related mitochondrial diseases display a wide variety of severities. Alpers syndrome manifests in infants and young children, and these patients normally die in the first decade of life. PEO and sensory ataxia neuropathy, dysarthria and ophthalmoparesis (SANDO) often are asymptomatic until 20+ years of age (5,7).
Pol γ is a Family A DNA polymerase composed of three distinct domains: a C-terminal polymerase domain, an N-terminal 3′–5′ exonuclease domain and a linker domain. Crystal structure and molecular modeling of the polymerase domain active site demonstrate that the structure of pol γ is similar to other Family A polymerases and can be compared with a right hand (8,9). A ‘palm’ domain binds the catalytic Mg+ ions, a ‘thumb’ domain is responsible for DNA binding and a ‘fingers’ domain helps insert the correct nucleotide opposite the template base. The exonuclease domain ‘proofreads’ by removing incorrect nucleotides when the polymerase inserts an incorrect nucleotide. Loss of exonuclease activity in pol γ increases base substitution mutagenesis ~20-fold in vitro (10) and 500- to 2000-fold in mouse and yeast model systems in vivo (11–15). The 2000-fold increased mutagenesis in homozygous exonuclease-deficient mice and a 100-fold increase in large deletions may explain premature aging in these animals (16,17). Finally, the linker domain is important for binding to the accessory subunit, PolG2, and enhancing polymerase processivity (18,19).
Of the 150 disease-associated mutations known in pol γ, several have been characterized biochemically. One of the most severe dominant pol γ-related disease-associated mutations, Y955C, resides in a conserved motif of the fingers domain. The Y955C polymerase displays decreased catalytic efficiency, nucleotide binding, nucleotide discrimination against incorrect nucleotides and increased misincorporation of nucleotides opposite 8-oxodG (9,20,21). Transgenic mice overexpressing the Y955C mutant polymerase in cardiac tissue exhibited decreased mtDNA copy number and cause premature death (22). PEO mutations in the fingers domain, G923D, R943H and A957S, also display decreased polymerase activity but not as severely as Y955C (9). W748S is a linker domain Alpers mutation that reduces DNA binding and is found in cis with the common single nucleotide polymorphism E1143G (23). Finally, A467T is an Alpers mutation located in the N-terminal thumb domain that is deficient in DNA binding, nucleotide binding, catalysis and binds poorly to POLG2 (18).
To address the in vivo effects caused by these mutations, Saccharomyces cerevisiae (budding yeast) has been used as a genetic model system to more easily study mtDNA stability and mutagenesis for several reasons (24–26). First, Mip1, the yeast ortholog of human pol γ that replicates the yeast mtDNA, is 43% identical to human pol γ (27). The polymerase domain is well conserved and most of the conservation in the exonuclease domain is concentrated near motifs I, II and III. Second, measuring the petite colony formation frequency determines the percentage of cells that are unable to perform aerobic respiration, a process that requires mitochondrial functions. Although modulation of many nuclear encoded genes can increase petite colony formation frequency (28), mtDNA polymerase mutations affects only mtDNA stability or mutagenesis and results in deletions or depletions of mtDNA (24,26). Third, yeast mtDNA point mutagenesis can also easily be assayed by measuring resistance to erythromycin (14,29). Fourth, heterozygous mutants allows for determining dominance of different alleles (26). Finally, yeast genetics allows for the ability to search for genetic pathways that can suppress the phenotypes of the disease-associated mutations. One known modulator of mitochondrial phenotypes in yeast is ribonucleotide reductase 1 (RNR1) (30), which is a subunit of the enzyme complex that catalyzes the rate limiting step in de novo dNTP synthesis necessary for DNA replication (31). Previously, the overexpression of RNR1 or deletion of its inhibitor sml1 in budding yeast was shown to increase mtDNA copy number (30), increase nucleotide concentrations (32) and partially suppress the increased petite colony formation frequency in the mip1 mutant strain Y757C (orthologous to the human Y955C) (25).
In agreement with in vitro studies, yeast strains containing mip1 mutations that are orthologous to the human pol γ G923D, R943H, Y955C and A957S mutants exhibited decreased mtDNA stability (25). Baruffini et al. (25) showed that a strain containing the mip1 mutation G224A, a mutation in the exonuclease domain of the polymerase that is orthologous to human disease-associated mutation G268A, increased mutagenesis about 10-fold. Recently, five other disease-associated mutations have been created in mip1 and the resulting mutant strains display various effects of mtDNA stability and mutagenesis (33). However, there remains many uncharacterized pol γ disease-associated mutations that correspond to amino acids conserved in Mip1. To understand the polymerase defects that are caused by mutations in the conserved regions, we characterized mtDNA instability and mutagenesis in yeast strains containing Mip1 mutant enzymes in the presence or absence of wild-type Mip1. Our results revealed that several yeast strains containing conserved mutations displayed increased mutagenesis and/or abolished mtDNA replication. These results suggest that many of these conserved mutations in pol γ are sufficient to alter polymerase activity and can initiate or exacerbate mitochondrial disease.
Using an alignment of human pol γ to yeast mitochondrial polymerase Mip1, we identified 31 disease-associated mutations that had not been previously characterized. These conserved amino acids include 24 mutations in the polymerase domain, five in the exonuclease domain and two in a non-defined region (Fig. 1). Site-directed mutations were constructed in the mip1-encoded pFL39 plasmid and transformed into the E134 yeast strain that contains a chromosomal-encoded wild-type copy to form heteroallelic mutants. MtDNA replication efficiency was determined by measuring the frequency of the formation of petite colonies that are composed of cells unable to respire and grow on a non-fermentable carbon source such as glycerol. Because Mip1 has no known role other than mtDNA synthesis, the increase in mip1 mutant cells that cannot respire must result from the mtDNA perturbations. Fifteen of the 31 mip1 mutant strains significantly increased petite formation compared with the wild-type strain (Table 1). The strains containing the T654A, R656W, R656Q, H734Y and A759P mutant alleles caused the greatest increases in petite frequency, suggesting that these Mip1 mutant enzymes have polymerase defects which inhibit DNA replication, and these enzymes can compete with the wild-type polymerase during replication. We also created monoallelic strains with a plasmid-encoded mip1 mutant and a mip1 deletion in the chromosome to determine the effects of the mutant enzymes in the absence of wild-type Mip1. Of the 31 monoallelic yeast strains, 14 strains were completely unable to grow on glycerol (Table 1) and five other strains exhibited increased petite frequency. These results identify these mutant enzymes as being unable to efficiently replicate mtDNA enough to maintain mitochondrial functions.
Although mutations in several nuclear genes can cause petite colony formation, increased petite frequency in mip1 mutants is likely caused by mtDNA deletions, point mutations or depletion. Because mtDNA depletion is a common symptom of pol γ-related diseases, we hypothesized that many mip1 mutant strains would have decreased mtDNA synthesis and decreased mtDNA copy number. To test mtDNA copy number, we performed real-time PCR on total DNA extractions of the heteroallelic strains and monoallelic strains that exhibited increased petite frequencies. In most of the strains, large increases in petite frequency were caused by significant decreases or total depletion of mtDNA (Table 2), suggesting that mtDNA loss is characteristic of POLG-related disease. Also, these results demonstrate that petite colony formation frequency caused by mip1 mutants is a valid measurement for mtDNA replication.
Mutations in mtDNA are prevalent in tissue samples of mtDNA-related mitochondrial disease patients (34). In addition to directly decreasing or destroying enzyme function, the propensity to incorporate incorrect nucleotides may result in polymerase stalling that can lead to deletion formation. To determine if the pol γ disease mutants confer an increase in mutagenesis, we measured the frequency of point mutations in both heteroallelic and monoallelic mip1 mutant strains. Point mutagenesis was assayed by measuring resistance to erythromycin, which is conferred by a missense mutation in nucleotides 1950 (G to T or G to A), 1951 (A to T, A to G or A to C), 1952 (A to T or A to G), 3993 (C to G) or an insertion of G between nucleotide 1949 and 1950 of the 21S rRNA gene (14,29). As a positive control for mutagenesis, we constructed an exonuclease deficient mip1 (Exo−) by altering two critical amino acids, Asp171 and Glu173, in motif I in the exonuclease domain to alanine residues. These residues are equivalent to Asp198 and Glu200 in the human pol γ that displays no detectable exonuclease activity (35). Thirteen heteroallelic strains containing mip1 mutants increased the frequency of mtDNA point mutagenesis (Table 1). However, none of the mutant strains reached the mutation frequency displayed by the proofreading deficient mip1 strain. In monoallelic strains, mutagenesis could only be measured in strains that did not become 100% petite. Only the mip1 mutants, L211P, R607P and D941N showed a significant increase in mutagenesis (Table 1), and this mutagenesis was higher than observed in the heteroallelic strains. This indicates that the point mutations generated by the mutant polymerase can be proofread by the wild-type Mip1 polymerase. Although we found significant increase in point mutagenesis by several mutant polymerases, none reached the mutagenic potential of the proofreading deficient mip1 in the heteroallelic situation, suggesting a limited role of mutagenesis to disease progression.
Both early onset diseases like Alpers Syndrome and late onset diseases like PEO are linked to POLG mutations. It has been speculated that the severity of the mutation determines age of onset and severity of the disease. There are four pairs of alterations that change the same codon but are mutated to different amino acids and associated with different disease prognoses. For example, arginine 656 in yeast Mip1 is orthologous to arginine 853 in human pol γ, which is mutated to a glutamine in an Alpers patient and to a tryptophan in a PEO patient (36–38). The strains with the mutation found in Alpers patients had a greater increase in petite frequency and mutagenesis than the strains with the mutation found in PEO patients (Fig. 2). These data suggest that mutants found in Alpers commonly induce more severe phenotypes than those found in PEO patients, supporting the overall hypothesis that the severity of the biochemical defect correlates with the timing and severity of mitochondrial diseases.
Increasing dNTP pools by overexpressing RNR1 partially suppresses the increased petite colony formation frequency in the mip1 mutants Y755C (25). To determine whether increased dNTP concentration suppressed the increased frequency of petite colony formation due to mip1 mutations, we transformed a multicopy vector containing RNR1 and its promoter into the six heteroallelic mip1 mutants that exhibit increased petite colony formation frequency. Overexpression of RNR1 reduced petite colony formation frequencies in the T654A, R656W, H734Y and R745H mip1 mutant strains (Fig. 3) but had no effect on strains containing the Q264H, R656Q and A759P (Fig. 3). These results suggest that increasing nucleotide concentration suppresses mip1 mutants with specific polymerase defects. In addition, RNR1 overexpression did not significantly alter random point mutagenesis (Fig. 4) and therefore did not result in a potentially mutagenic dNTP pool imbalance.
We hypothesize that mip1 mutants which are suppressed by RNR1 overexpression have a defect in nucleotide binding and is relieved by an overabundance of dNTPs. Thus, suppression of yeast mip1 mutant phenotypes by RNR1 overexpression could predict the affinity of nucleotide binding of the POLG human ortholog. In particular, petite formation frequency in H734Y, which is orthologous to POLG H932Y mutation found in PEO and SANDO (sensory ataxic neuropathy, dysarthria and ophthalmoparesis) (39,40), was dramatically reduced due to RNR1 overexpression (Fig. 3). To test H932Y nucleotide binding in vitro, we purified recombinant human wild-type and H932Y pol γ to perform a pre-steady state kinetic analysis which determines the relative contributions of the catalytic rate and the binding affinity on the overall polymerase activity of H932Y (Fig. 5). As shown in Table 3, H932Y displayed ~79% of the maximum rate of polymerization (kpol) of the wild-type enzyme; however, binding affinity (Kd) of H932Y to the incoming nucleotides was reduced over 200-fold compared with wild-type. Therefore, increasing dNTP concentration restores H932Y to near wild-type polymerase activity in vitro similar to the suppression of mip1 mutant H734Y by increased dNTP concentration in vivo. Previously published kinetic constants of the human orthologs to the other mutations reported in Figure 4 shows a striking correlation between decrease in petite frequency and decrease in dNTP binding affinity, suggesting that the RNR1 rescue assay can be used to predict dNTP binding.
We used the budding yeast as a genetic model system to study 31 previously uncharacterized mutations in yeast MIP1 that are orthologous to disease-associated mutations in humans (Fig. 1 and Table 1). Although there have been previous studies that use petite frequency and point mutagenesis to draw conclusions about mitochondrial diseases, this study is important because we sampled a large variety of mip1 mutations to draw more general conclusions about POLG-related disease. First, several novel disease mutations with no other evidence of disease causation were shown to cause mtDNA depletion. Second, we identified several mutations with dominant or strictly recessive phenotypes. Third, we determined increased point mutagenesis alone is rarely characteristic of disease-associated mutations, suggesting a limited role of increased random mtDNA point mutations on disease state. Fourth, many of the disease associated mutations caused severe mtDNA depletion. Fifth, the severity of the polymerase mutation correlates with the age of onset and severity of the disease. Sixth, mtDNA dysfunction caused by mutated polymerases with decreased dNTP binding affinity was suppressed by increasing dNTP pools. Finally, the correlation between known data about the human Pol γ and yeast Mip1 demonstrates that further searches for genetic suppressors and antagonists of mtDNA depletion may provide several novel insights in POLG replication and mitochondrial disease that will be useful in future studies, diagnoses and therapies.
Discerning pathological mutations from neutral polymorphisms in patients with mitochondrial disease is critical for the molecular diagnosis of POLG-related disorders. However, in many cases, the human genetics is lacking because the mutation is reported in only one patient with limited family genetic history. Furthermore, many patients are compound heterozygotes, meaning that they have two or more POLG mutations in cis and/or in trans and it is unclear which mutation is disease causing. Therefore, classification of POLG mutations as deleterious or neutral is paramount for clinical understanding. We identified mtDNA replication defects in 20 of the 31 mutants studied, which expands the number of disease-associated mutations that clearly decrease polymerase activity. Of these 20 mutations, only POLG mutations L304R and R1096C have been identified in homozygous individuals (41), whereas the other mutations have only been identified in compound heterozygotes. Eleven of the 20 mutations have no reported genotypic family history to assist in identifying the important disease causing mutations. Seven mutations (Q308H, R807C, R853Q, G1076V, R1096H, S1104C and V1106I) have been reported in only one patient (36,42–45), and only R853Q has been previously shown to decrease polymerase activity (33). Therefore, this study provides the first evidence that POLG mutations Q308H, R807C, G1076V, R1096H and S1104C encode mutant enzymes with decreased polymerase activity leading to mtDNA depletion, mitochondrial dysfunction and mitochondrial disease. Although it is possible that the other 11 mutants are neutral polymorphisms that do not decrease polymerase activity, we cannot exclude the possibility that these mutations affect a function specific to the human enzyme, such as species specific protein–protein interaction.
Because most of the mutations are identified in compound heterozygotes, it is important to understand how dominance affects mutant phenotypes. The collection of mutants allowed us to identify five mip1 mutants (T654A, R656W, R656Q, H734Yand A759P) that had large increases in petite colony formation frequency (Table 1) and mtDNA depletion (Table 2) in heteroallelic backgrounds which includes a wild-type MIP1 allele. These data suggest that the mip1 genotypes led to a dominant negative phenotype. Although there have been no documented cases of these mutations causing dominant mitochondrial disease, family histories of patients with the human ortholog to T654A and H734Y (human T851A and H932Y) suggest that there is low penetrance of neurological problems and migraines in simple heterozygous family members (40,46). It is possible that the lower mtDNA copy number causes yeast (20–30 copies per yeast cell) to be more sensitive to mtDNA perturbations than humans (1000–10000 copies per human cell). Because of the lack of genetic data and mtDNA analyses of the parental carriers of these alleles, we cannot exclude the possibility that these mutations have dominant effects on mtDNA that are unreported. Therefore, the novel identification of possible dominant effects as shown in this study warrants further study into future patients with these genotypes.
In addition to identifying dominant mtDNA replication phenotypes, it is also important to identify mip1 mutant strains that cause recessive phenotypes. Recessive phenotypes imply that mutant Mip1 is unable to compete with the wild-type enzyme at the replication fork. This could be caused by a mutation that renders the polymerase catalytically inactive, unstable, misfolded or with a severe DNA binding defect. The strains containing the G651S, N667S, G832V and D941N mutants had no significant mitochondrial effect in the heteroallelic strains; however, the mutant monoallelic strains showed mtDNA depletion and increased petite colony formation frequency. These data suggest that the human orthologs may act essentially as a null mutation. Families of patients reported to have the orthlogous mutation to G651S, N667S and D941N (human G848S, N864S and D1184N), indicated that the mutations were recessive based on asymptomatic heterozygotes (47,48). Furthermore, Kasiviswanathan et al. showed that human G848S mutant enzyme has only 0.1% polymerase activity compared with wild-type and 5-fold reduction in DNA binding activity, which corroborates the recessive nature of the in vivo human and yeast data. If the other mutants also have severe polymerase defects as implied by this study, then the disease state in humans would be determined by the genotype of the other POLG allele. This is very useful information in determining the relative severity of the other POLG alleles that have been identified in these compound heterozygotes.
Many of the POLG mutations found in mitochondrial disease patients are located in the exonuclease domain, and there has been no biochemical and very little genetic data for the disease-associated mutations in the exonuclease domain. Therefore, it is unknown whether any or all of these exonuclease domain mutations lead to a lack of proofreading function or whether humans can survive without proofreading activity in the mitochondrial polymerase. In yeast and mice, respiratory proficient mitochondria are present in strains with proofreading deficient mitochondrial polymerases, but there is a significant effect on the overall phenotype of the organisms. In yeast, petite frequency increases to 10–40% (Table 1), whereas mice exhibit the effects of premature aging. Five human disease mutations in the exonuclease domain alter residues conserved in yeast MIP1 (yeast L211P, L260R, Q264H, R265L and R265H). However, none of these mutants exhibited mutant frequencies similar to the exonuclease-deficient mip1 mutant (Exo− in Table 1), suggesting that the mutants retain some proofreading activity. Although the L211P and Q264H mutants displayed a modest increase in point mutagenesis (Table 1), it is difficult to imagine that this small mutator effect contributed to disease. For example, heterozygous exonuclease-deficient POLG mice showed no obvious phenotypes despite a 500-fold increase in mitochondrial point mutations (12). Q264H, L260R and R265L mutant strains showed increased petite frequency despite little effect on mutagenesis (Table 1), suggesting that the exonuclease domain disease associated mutations potentially cause disease by changes other than loss of exonucleolytic function, such as protein stability, subunit interaction or polymerase activity. Because the majority of the exonuclease domain is not well conserved between yeast and humans, biochemical characterizations of several exonuclease domain mutant pol γ will be necessary to determine causal role of proofreading and mutator phenotypes on mtDNA diseases.
mip1 Mutant H734Y displayed the largest decrease in petite colony formation frequency caused by overexpression of RNR1. The pol γ homolog H932Y caused a 200-fold decrease in nucleotide binding affinity but only 1.3-fold decrease in catalysis when compared with wild-type enzyme (Table 3). These results can be explained by the molecular model of the ‘closed’ form of the active site of human pol γ (9) and the crystal structure of the ‘open’ form (8), which suggests that histidine 932 is within 3 Å of the modeled β-phosphoryl oxygen of the dNTP (Fig. 6) and close enough for hydrogen bonding. His932 is modeled in between Arg943 and Tyr951, residues that are important for nucleotide selectivity and dideoxynucleotide discrimination, respectively, suggesting that mutating H932Y may cause a steric hindrance to those important dNTP interactions. In addition to H734Y, the suppression by overexpression of RNR1 predicts that R656W has little effect on catalysis but reduces affinity for dNTP binding. This conclusion contrasts from the major catalytic defect suggested by mip1 R656Q and its pol γ ortholog R853Q (Fig. 3 and Table 4). Although we have no biochemical data, based on the location of Arg853 within the active site near the incoming nucleotide triphosphate and a catalytic magnesium ion, it could be important in both nucleotide binding affinity and catalysis.
Most mutants with increased petite colony formation frequency exhibited decreased mtDNA copy number, suggesting that mtDNA depletion is a major phenotype of these disease-associated mutants. Although the mechanism for mtDNA depletion is unclear, the mutant polymerases with the most severe biochemical defects are likely to cause polymerase stalling, which could result in mtDNA depletion in several hypothetical ways. First, delayed mtDNA synthesis may decrease mtDNA copy number if mitochondrial turnover is faster. Second, incomplete replication of mtDNA could target nuclease degradation. Third, polymerase stalling may increase deletions or mutagenesis, which could result in dysfunctional depolarized mitochondria that would be targeted for degradation by autophagosomes (49). Stalled polymerases may increase exposure of ssDNA which increases the susceptibility of the template for mutagenic lesions. (50). In support of this hypothesis, mip1 mutants T654A and R656Q have the most increased petite colony formation frequency and have the highest increases in point mutagenesis (Table 1), but there is little effect on the fidelity of the orthologous human pol γ T851A and R853Q, in vitro. If polymerase stalling increases the mutagenic potential of the template, mtDNA lesions would be more prevalent in these strains. mip1 mutant Y757C was shown to increase petite frequency, mtDNA base damage and mutagenesis (24,25) but the human orthologous, polymerase is intrinsically unfaithful (21), so it cannot be determined whether base damage caused mutagenesis. Future biochemical and genetic studies will be needed to develop models that explain mitochondrial dynamics that cause mtDNA depletion, which will be important in determining the causes for the tissue specificity and the differences in age of onset among mitochondrial diseases.
Saccharomyces cerevisiae were grown at 30°C in YP (yeast extract 1%, peptone 2%) with 2% glucose or glycerol as carbon sources or synthetic complete media. Escherichia coli were grown in standard LB media at 37°C. When appropriate, gentimicin and ampicillin were added to YPD (0.2 mg/ml) and LB (0.1 mg/ml), respectively.
PFL39 which contains mip1 on a centromeric plasmid was previously described (13). Site-directed mutagenesis of plasmid-encoded mip1 was performed using the QuikChange® Site-Directed Mutagenesis Kit (Invitrogen™). The absence of any other mutations in the entire mip1 gene for every plasmid was verified by sequencing. PFL39 with exonuclease deficient mip1 was previously described (51). Saccharomyces cerevisiae gene RNR1 was cloned into pRS425 (obtained by Rodney Rothstein) to make plasmid pRNR1. The RNR1 gene was amplified using RNR1fw (5′-GGAAGGGATCCGGGTGTTGAATAGAGGACGCG-3′) and RNR1rv (5′-GCGCGCTCGAGGAACAATGTTGCCTAGACCCC-3′), PCR product was digested with BamHI and XhoI (New England Biolabs®, Inc.), and was ligated into BamHI- and XhoI-digested pRS425. The resultant pRNR1 was sequenced verified and included 322 nucleotides upstream and 283 nucleotides downstream of RNR1. Plasmids containing ACT1, COX1 and COX2 fragments, used for mtDNA quantitation, were created by cloning amplified PCR fragments into TOPO2.1 plasmid (Invitrogen™) and were sequence verified.
All S. cerevisiae strains were derived from E134 (MATα ade2 trp1-289) and YH747 which is the MATa isogenic strain to E134 (24) (kindly provided by Dmitry Gordenin). Heteroallelic mip1 strains were made by transforming pFL39-mip1 or a mutant derivative into E134 by selecting for TRP. To create monoallelic strains, first, a mip1::kanMX insertion–deletion was made in YH747. To make the insertion–deletion, the kan gene was PCR amplified using oligonucleotides mipkanfw (5′-CGGTTCTAAAGAAGAGGTCGAGATGGGGATTATATGTAGTCGTACGCTGCAGGTCGAC-3′) and mipkanrv (5′-CTAGTACTCTCTAGAAATAGTAATGTCCCTTTCCAGCTCAACATCGATGAATTCGAGCTCG-3′), and the product was transformed into YH747, selecting for G418 resistance and screened for the inability to grow on media containing glycerol. Second, the resulting strain was mated with E134 to create strain JSY23, selecting for a diploid strain that was G418 resistant and could grow on YPG. Third, pFL39-mip1 plasmids or mutant derivatives were transformed into JSY23 selecting for TRP. Fourth, each strain was sporulated, and the tetrads were dissected into four spores. Finally, the spores were incubated at 30°C for 2 days and were checked for resistance to G418, ability to grow on synthetic complete media without tryptophan, and growth on YPG. For the RNR1 experiments, pFL39-mip1 and pRNR or pRS425 were cotransformed simultaneously, and the transformants were selected on synthetic complete media lacking leucine and tryptophan.
Petite frequencies are the frequency of rho− cells (petites) in the total population. Rho− cells are devoid of mitochondrial functions but are not necessarily devoid of mtDNA (rho0). To determine petite frequency in heteroallelic strains, at least 12 fresh transformants per strain were diluted in water and between 200 and 1000 colonies were plated onto YPD. For monoallelic strains, rho+ cells derived from several tetrad dissections were single colony purified on YPD and then assayed for petite frequency. Cells were incubated at 30°C for 2 days. Rho+ cells were identified either by the accumulation of red pigment as a result of mutations in the ade biosynthetic pathway (52) or by the inability to grow on YPG. Using either or both method, at least 300 colonies per plate were counted, and petite colonies were identified. Frequencies were determined for each plate, and the median number of the frequencies was calculated; 95% confidence levels were determined by using the method of the median (53). For monoallelic strains determined to be 100% petite, no rho+ cells could be isolated from cells derived from 10 different haploid spores.
Resistance to erythromycin is conferred by one of several missense mutation in the 21S rRNA gene in mtDNA (14,29,54). To measure mutant frequency, 20–40 independent colonies from each strain were used to inoculate into 4 ml of synthetic media without tryptophan, and these cultures were incubated to saturation (for 2 days) at 30°C. The cells were plated on YPEG (1.7% ethanol, 2% glycerol) with 4 g/l of erythromycin. A small aliquot of 5–10 cultures were used to titer the number of rho+ cells by plating 10−5 dilutions on YPG. Erythromycin resistant colonies were counted after 6 days of incubation at 30°C. The mutant frequency was the median number of erythromycin colonies per 108 rho+ cells plated. 95% confidence levels were determined using the method of the median (53).
mtDNA copy number was quantified relative to nuclear DNA copy number using real-time PCR. Primers and probes were designed to specifically amplify within the mitochondrial-encoded cox1 and cox2 gene and the nuclear-encoded act1 gene. Real-time PCR reactions using Taqman® Universal PCR Master Mix (Applied Biosystems™) were performed at 40 cycles of 95 for 30 s and 50° for 30 s. Agarose gel electrophoresis confirmed that a single amplicon resulted from the real-time PCR reaction (data not shown). Plasmid molecules containing cox1, cox2 and act1 were quantified, and real-time PCR was performed on at least five different dilutions to determine a logarithmic equation of a curve (R2 values > 0.97) that represents numbers of molecules as a function of the critical threshold of every reaction. Every reaction was done in triplicate, and four replicates were tested per strain. Data represent the average ratio of the number of cox1 or cox2 molecules to the number of act1 molecules. Copy numbers were indistinguishable when measuring cox1 or cox2 copy number (data not shown). Error bars are the SEM.
dCTP was purchased from GE Life Sciences. Oligonucleotides D23 (5′-GCCTCGCAGCCGTCCAACCAAC-3′) and D45 (5′-GGACGGCATTGGATCGAGGTTGAGTTGGTTGGACGGCTGCGAGGC-3′) were synthesized at the Keck DNA synthesis facility (Yale University) and gel purified using 20% denaturing gel electrophoresis. D23 was 5′-32P-labeled using T4 polynucleotide kinase (New England Biolabs®, Inc.) and [γ-32P] ATP (GE Life Sciences) and annealed to the D45-mer template as described previously (55).
Exonuclease deficient wild-type and H932Y recombinant HIS-tagged pol γ catalytic subunit was purified in a baculovirus-infected Sf9 host as previously described (9,35). Wild-type accessory subunit was expressed in E. coli as previously described (56).
Experiments were performed on a KinTek Instruments model RQF-3 rapid quench-flow apparatus. One hundred nanomolar pol γ catalytic subunits (determined by pre-steady state active site titration) were incubated with 500 nm accessory subunits on ice for a minimum of 15 min, followed by mixing with 300 nm 5′-32P-labeled primer-template and incubating for an additional 15 min on ice. Pre-steady state bursts were carried out by rapid mixing of a solution containing the pre-incubated pol γ holoenzyme and primer-template with a solution of 2.5 mm MgCl2 and varying concentrations of dCTP in reaction buffer (50 mm Tris, 100 mm NaCl, pH 7.8 at 37°C). Polymerization was quenched at various time points by the addition of 0.3 m EDTA. All concentrations represent final concentrations after mixing. Products were separated on a 20% denaturing polyacrylamide gel and quantitated on a Bio-Rad Molecular Imager FX. Data were quantitated and fit as described previously (57).
This work was supported by intramural funds from the National Institute of Environmental Health Sciences National Institutes of Health (NIH) [ES 065078 to W.C.C.]; NIH [GM49551 to K.S.A.] and the National Institutes of Health Summers of Discovery Program.
We would like to thank Rodney Rothstein for providing plasmid pRS425 and Maria A. Graziewicz for generation of the human H932Y mutant pol γ. Also, we would like to thank Dmitry Gordenin, Maggie Humble, Gary Pittman and Gregory Stuart for technical assistance and to Amy Abdulovic and Scott Lujan for critical reading and review of the manuscript.
Conflict of Interest statement. None declared.