|Home | About | Journals | Submit | Contact Us | Français|
Mutations in mitochondrial DNA is an outcome of errors produced by DNA polymerase γ during replication and failure of the repair mechanism. Misincorporation of non-canonical dUTP leads to mutagenesis or apoptosis, and may contribute to the cytotoxic effects of 5′-fluorouracil chemotherapy. Tumor suppressor p53 protein in the mitochondria displays physical and functional interactions with mitochondrial DNA and polymerase γ, and by its intrinsic 3′→5′ exonuclease activity can diminish the polymerization errors. Here we demonstrate the impact of p53 on incorporation of uracil into DNA examined with mitochondrial fractions, as the source of polymerase γ. p53 in mitochondria facilitates DNA damage repair functions resulting from uracil–DNA misincorporation. Our biochemical studies revealed that the procession of U:A and mismatched U:G lesions enhances in the presence of recombinant or endogenous cytoplasmic p53. p53 in mitochondria can function as an exonuclease/proofreader for polymerase γ by either decreasing the incorporation of non-canonical dUTP into DNA or by promoting the excision of incorporated nucleotide from nascent DNA, thus expanding the spectrum of DNA damage sites exploited for proofreading as a trans-acting protein. The data suggest that p53 may contribute to defense of the cells from consequences of dUTP misincorporation in both normal and tumor cells.
A high frequency of mutations within mitochondrial DNA (mtDNA), resulting in mitochondrial dysfunctions, are an important source of various diseases including cancer and human aging [1, 2]. MtDNA damage yield inaccurate translation and synthesis of respiratory chain enzymes imperative for energy production in the mitochondrial inner membrane . A “mutator” phenotype of the mitochondrial genome is a phenomenon caused by incorporation of incorrect nucleotide, errors produced due to the presence of unbalanced dNTP concentrations and oxidative mutagenesis [2, 3]. Mitochondrial DNA polymerase γ (pol γ), a singular replicase responsible for the biogenesis of mtDNA, consists DNA polymerase and exonuclease activities . The mutagenic mechanisms are related to replication errors formed by pol γ during DNA synthesis by misinsertion of wrong nucleotide or by diminished proofreading capacity [4, 5]. Remarkably, mice expressing an exonuclease-deficient pol γ exert a 3–5-fold increase in the level of point mutations and undergo enhanced aging . MtDNA is not protected by histones and mtDNA repair is ineffective . To certify mtDNA integrity, cells contain quality control mechanisms and DNA damage response pathway(s) comprising mtDNA replication/repair preservation programs that either preclude or repair damage . Mitochondrial context is affected by cellular stresses. MtDNA integrity and replication rest on numerous nucleus-encoded DNA repair proteins normally localized in mitochondrial matrix or imported upon various stress signals . Tumor suppressor p53 protein is involved in diverse cellular processes including cell cycle arrest (thus preventing the replication of damaged DNA), apoptosis (for eliminating defective cells), or DNA-damage repair [9–10]. p53 protein executes multi-compartmental functions in the cell by either a numerous p53-regulated proteins and/or by its intrinsic biochemical activities [11, 12]. p53, by providing exonuclease activity, enables error-correction thus functioning as a guardian of genome [12, 13]. In addition to regulation of apoptosis by p53 via protein-protein interaction at the mitochondrial outer membrane, the biological routes regulated by p53 comprises mtDNA synthesis/repair for mitochondrial function after translocating to matrix . RECQL4 was detected essential for the transport of p53 to mitochondria [15, 16]. The multi-functional p53 protein promotes DNA repair both directly or indirectly through multiple mechanisms . Mitochondrial p53 levels are proportional to total p53 levels and the majority of p53 was present inside the intra-mitochondrial compartment-matrix . Co-localization of p53 with mtDNA, pol γ and various constituents of the mtDNA replication apparatus (e.g. transcription factor A-TFAM and single-stranded DNA-binding proteins), even in the absence of exogenous stress independent of apoptosis, establishes a non-apoptotic function for matrix-localized p53 which underlines an importance of p53 in mtDNA homeostasis [18–20]. Several studies illustrated the participation of p53 in mtDNA repair in a variety of systems: a)p53 enhances base excision repair through direct interaction with the repair complex in mouse liver and cancer cells . b) Intra-mitochondrial p53 provides an error-repair proofreading function for pol γ by excision of misincorporated nucleotides . c)p53 is proficient of hydrolyzing the 8-oxo-7,8-dihydro-2′-deoxy-guanosine (8-oxodG) present at the 3′-end of DNA, a well-known marker of oxidative stress . d)p53 regulates mtDNA copy number, which may impact mitochondrial and cellular functions . Apparently, the functional interaction of p53 and pol γ is significant for avoiding mtDNA mutations and mtDNA depletions that are frequently observed in human cancers and neurodegenerative diseases .
Uracil (dU) in DNA, resulting from spontaneous cytosine deamination and/or incorporation of non-canonical dUTP during replication, leads to mutagenesis and apoptosis [24, 25]. The frequency of dU incorporation depends upon the relative intracellular pool size of dUTP and dTTP . The enzyme deoxyuridine triphosphate nucleotidohydrolase (dUTPase), which facilitates the conversion of dUTP to dUMP further utilized by thymidylate synthase (TS) for synthesis of dTMP, may avoid misincorporation of dU into DNA by decreasing the dUTP/dTTP ratio . The presence of dUTPase in nuclei and mitochondria points that keeping a low level of uracil in DNA is important for the integrity of nuclear as well as mitochondrial DNA . Remarkably, the expression of the nuclear isoform of dUTPase is predominantly cell-cycle and proliferation-dependent, whereas the mitochondrial isoform is constitutively expressed .
The misincorporation of dU, as a result of accumulation of dUTP, plays a critical role in cytotoxicity mediated by TS inhibitors, such as the commonly used anticancer drug 5-fluorouracil (5-FU) . DNA directed cytotoxicity of chemotherapeutic agents (e.g.5-FU) not only depends on accumulation of dUTP, but may also be determined by the efficiency of the DNA repair mechanisms (e.g. excision repair, mismatch repair) which preclude the incidence of the mistake . Pol γ is incapable to readily correct U:A mismatches . Of note, an augmented levels of mitochondrial p53 was reported in response to 5-FU . Given the functional cooperation between pol γ and p53 in mitochondria, we aimed to ascertain the impact of p53 during the incorporation of dU into DNA. Our biochemical studies revealed that p53 in mitochondria can function as an exonuclease/proofreader for pol γ by either decreasing the incorporation of non-canonical dUTP into DNA or by promoting the excision of incorporated dU from DNA, thus substantiating and expanding the role of p53 in DNA damage repair.
The functional interaction between p53 and pol γ during incorporation of dU was assessed during DNA synthesis catalyzed by the replication machinery of mitochondrial (mit) fractions derived from isogenic HCT116 cells, as a model system . Western blot analysis validated the location of endogenous p53 in mit(p53+/+) but not in mit(p53−/−) (Figure (Figure1A).1A). An antibody directed against cytochrome c was used to control for the enrichment of the mitochondrial fraction and to serve as a reference when comparing p53 levels between extracts. The absence of c-jun nuclear marker protein confirmed the purity of lysates, thus excluding the possibility that the polymerization or excision activity detected in mitochondrial fractions, exploited as a the source of pol γ, might be due to a contamination of the extracts with nuclear DNA polymerases and exonucleases.
We designed dsDNA substrates containing 3′-terminal correct At:T pair or At:A mispair. The sequences of the substrates were optimized to provide an AT-rich surrounding region (A:T versus G:C rich sequences) of the target site to maximize efficient proofreading in adjacent to relatively unstable DNA regions . This point assists to accentuate the capability of p53 to accomplish effective error correction. The mitochondrial lysates of (p53+/+) and (p53−/−) cells expressing equivalent DNA polymerization activity (implying similar levels of pol γ) tested for exonuclease activity showed that under non-polymerization conditions the 3′-terminal A:A mispair excision with mit(p53+/+) (Figure (Figure1B-left,1B-left, lane 3) was more efficient than with mit(p53−/−) (lane 2). Since both fractions had comparable DNA polymerization activities (lanes 4 and 5), it is apparent that the observed efficient exonuclease activity in mit(p53+/+) was not due to a high level of pol γ. Apparently, the presence of protein with an intrinsic exonuclease activity in p53-harboring extracts may account for the efficient excision activity.
Next, the incorporation of dU opposite template A was examined with standing-start dsDNA substrate (wherein the target template residue immediately follows the 3′-terminal end of the primer)  (Figure (Figure1B-middle).1B-middle). The production of 17- and 18 mer products following the generation of the At:U pairs with mit(p53−/−) was related to the increase the time of incubation (lanes 1–2), the amount of extracts (lanes 5–6), or the concentration of dUTP (lanes 9–11). The results exhibited that incorporation of dU with mit(p53+/+) (lanes 3–4, 7–8, 12–14) was lower compared to that with mit(p53−/−) (lanes 1–2, 5–6, 9–11), as reflected by the decrease in the intensity of 17- and 18 mer products. Four separate preparations of mitochondrial extracts each tested for the markers establishing no contamination of nuclear proteins, yielding the similar results. Quantitative analysis further validates the lower (about 8.5-fold) insertion of dU into DNA with mit(p53+/+) related to that with mit(p53−/−) (Figure (Figure1B1B-right).
To ensure the contribution of endogenous p53 to error correction in mit(p53+/+) we explored immuneprecipitation experiments from repair-proficient lysates (Figure (Figure1C-left).1C-left). The inhibitory effect perceived with mit(p53+/+) was significantly attenuated by the specific depletion of p53 by Do-1 anti-p53 antibodies (Figure (Figure1C-right,1C-right, lane 1) while being un-affected by non-specific depletion by the anti-IgG antibodies, that did not affect neither p53 expression nor the extent of decrease in At:U pair production noticed with mit(p53+/+) (lane 2). Notably, specific depletion of p53 did not affect polymerization activity (data not shown). Based on observed correlation between the existence of p53 in mitochondria, augmented excision activity and low efficiency of At:U pair generation, p53 may be a plausible candidate for processing of this type of DNA lesion.
DNA pol γ can utilize either dTTP or dUTP as a precursor for DNA synthesis . We have assessed DNA synthesis with mit(p53−/−) in reactions where constant level (50μM) of dTTP or dUTP was supplemented individually into separate reactions. There was no significant variance in the efficiency of incorporation between dTTP and dUTP (Figure (Figure2A,2A, Seq I) (lanes 1–6). In contrast, the incorporation of “correct” dUTP opposite template A was lower (about 8-fold) than of dTTP with endogenous p53-harboring fractions (Figure (Figure2A,2A, lanes 7–12 and 2B).
Uracil may reside in DNA through the incorporation of dUTP instead of dCTP during replication, resulting in Gt:U mispairs . The insertion of correct dCTP or wrong dUTP opposite the template G was evaluated with mit(p53−/−) (Figure (Figure2C,2C, Seq II). The 17 mer product (about 4.5%) (Figure (Figure2D)2D) was generated after the production of correct C:G pair with no further extension with the time (Figure (Figure2C,2C, lanes 1–3). Noticeable, incorporation of dU and production of Gt:U mispair was achieved at much higher nucleotide concentration of dUTP (1 mM) than the correct insertion of dCTP (50 μM) (Figure (Figure2C,2C, lanes 4–6). The experiments described in Figure Figure2C2C depicted that under the same experimental conditions a reduction occurs in the incorporation of the incorrect dU opposite the template G with mit(p53+/+) (lanes 10–12) relative to that detected with the incorporation of correct dCTP (lanes 7–8). A quantitative analysis reflects that the generation of Gt:U mispair is about 8 times lower with mit(p53+/+) than with mit(p53−/−) (Figure (Figure2D).2D). Biochemical studies comparing the misincorporation of dU into DNA with p53-deficient mitochondrial lysates with those from isogenic p53-positive lysates, Showed that p53 may affect the pol γ selectivity for the incorporation of dU and production of both "correct" At:U and “incorrect” Gt:U pairs.
The next challenge was to study the impact of the endogenous p53 on the excision of incorporated dU using a sequential reaction experiment, employed by us previously . This strategy let to follow after the excision of the At:U pair produced by pol γ in mit(p53−/−) in the absence or presence of either recombinant or endogenous p53 provided by cytoplasmic extracts of HCT116 cells (Figure (Figure3).3). We took advantage of our recent observations that endogenous p53 in cytoplasmic extracts displays a high level of 3′→5′ exonuclease activity with dsDNA substrates . After 15 min incubation, pol γ in mit(p53−/−) exerts dU insertion, thus generating 17- and 18 mer substrates for the excision step (Figure (Figure3-left,3-left, lane 1). The accumulation of these products after additional 30 min incubation indicates that the At:U pair was not efficiently corrected by an intrinsic excision activity of pol γ (lane 2). Conversely, following the misincorporation event the significant removal of the polymerization errors was witnessed in the presence of recombinant p53-GST fusion protein (lane 4), or endogenous p53-harboring cytoplasmic extracts [cyt(p53+/+)] (lane 6), as reflected in the decrease of the intensity of 17- and 18 mer products. In control experiments we verified no detectable editing of the wrong nucleotide after the supplementation of GST protein (lane 3) or cyt(p53−/−) (lane 5). We complemented this study by further characterizing the excision of incorrect Gt:U pair in the presence of either recombinant (Figure (Figure3-right,3-right, lane 3) or endogenous p53 (lane 5). These results further support the notion that the presence of p53 provides exonuclease activity enabling error-repair. It is highly likely that the error removal takes place by a direct excision mode implemented by external p53 exonuclease in mitochondria for processing of At:U or Gt:U pair.
We performed two experiments to illustrate the proofreading potential of p53 as a trans-acting protein during error-correction process via dissociating-intermolecular mechanism (i.e. following the release of the DNA from the enzyme) in the mitochondria .
To define if the excision of the newly produced At:U pair includes the dissociation of dU-DNA from the DNA polymerase, we used a sequential reaction experiment. The reaction mixture following the initial incorporation of dU into DNA with mit(p53−/−) (Figure (Figure4A,4A, lane 1) was further incubated without (lane 2) or with cyt(p53+/+) in the absence (lane 3) or presence of unlabeled ssDNA (lane 4) or ssRNA (lane 5) trap. While incubation with cyt(p53+/+) alone results in the decrease in the level of 17- and 18 mer products (lane 3), the negative effect was decreased in the presence of ssDNA (lane 4) or ssRNA trap (lane 5). Obviously, the insertion of dU leads to the dissociation of the DNA-pol γ complex, thus assisting the dsDNA containing 3′-terminal At:U pair to be reachable for editing by trans-acting exonuclease.
Proofreading is a multi-step process requiring the efficient excision of the wrong terminal nucleotide, the transfer of the primer to the polymerase for correct replacement of the incorrect nucleotide, and normal polymerization–primer elongation by DNA polymerase [5, 33]. The mode of action of p53 in error correction may be based on its direct interaction with DNA and biochemical (e.g. exonuclease) function. Since the p53 protein is a sequence-independent DNA-binding protein , it is highly likely that the physical binding of p53 to DNA may be a relevant event in the biological utility of the protein in DNA repair. Our recent studies, by analyzing the combined products of a DNA binding and polymerization reactions using sequential reaction experiment, revealed that within the context of error-correction events p53, by recognition and excision of 3′-mismatched nucleotides may be involved in DNA repair, thereby increasing the accuracy of DNA synthesis by DNA polymerases (e.g. HIV-1 RT) . Here, we studied the functional interaction between p53 and pol γ, which is also a DNA binding protein. In sequential reaction experiment after the incorporation of dU by pol γ in mit(p53−/−) (Figure (Figure4B,4B, lane 1), no further elongation products were detected after the addition of dNTP at equimolecular concentrations (2.5 μM) (lane 2), indicating the inability of pol γ to further extend At:U pair without correction of the error. However, the extension of mispaired substrates occurred by mit(p53−/−)/dNTP (2.5 μM) complex in the presence of cyt(p53+/+) resulting in the production of 19–22 mer products (lane 4). Particularly, the efficient excision of the incorporated At:U mispair happens in the presence of cyt(p53+/+) alone (lane 3). We performed parallel experiments supplemented with other DNA polymerase, e.g. exonuclease-deficient murine leukemia virus (MLV) reverse transcriptase-RT. Indeed, the substrate was elongated by the MLV RT in the presence of 1.0 μM dNTP following the addition of cyt(p53+/+) giving rise to 19–27 mer products (lane 7). The fact that in the presence of 1.0 μM dNTP there was no detectable extension of the substrate by MLV RT alone (lane 6) or by pol γ even in the presence of cyt(p53+/+) (lane 5) further substantiate a functional cooperation of p53 exonuclease and polymerase in a coordinated manner during DNA synthesis . The DNA elongation by either pol γ in mit(p53−/−) (lane 4) or by MLV RT (lane 7) occurs after initial correction of the terminal At:U pair from un-extended substrate by p53, dissociation of the DNA from the protein, transfer to the polymerase (e.g. pol γ or MLV RT), and subsequent extension from the correctly paired 3′-terminus. Hence, we may conclude that p53 trans-acting protein in the mitochondria may contribute an exonuclease/proofreading function for removal of 3′-terminal dU by dissociating-intermolecular pathway during ongoing DNA synthesis.
Several approaches were undertaken to evaluate the importance of endogenous p53 during the incorporation of dU.
Mammalian cells have established a diverse defense network to safeguard genomic integrity . The aim of the current study was to test the hypothesis that p53 could reduce the incidence of production mutagenic uracil damage in mtDNA. Our biochemical studies demonstrate that mitochondrial p53 possesses the potential to diminish the incorporation of non-canonical dUTP into DNA by pol γ. Evidence that enhanced exonuclease/proofreading function for dU damage repair detected in mitochondrial extracts is assigned to p53 are supported by numerous experimental evidences. 1) Endogenous p53 has an impact on the replication accuracy of pol γ during the incorporation of dU into DNA according to two criteria: a) The presence of p53 reduced the efficiency of dU incorporation into DNA, although p53 status did not affect the insertion of correct nucleotide. The significance of p53 is supported by the immunodepletion of p53 from mitochondrial extracts of repair-proficient p53-harboring cells diminishing this negative effect. b) The procession of “correct” At:U and “mismatched” Gt:U lesions enhances in the presence of recombinant or endogenous wtp53. 2) Within the context of error-correction events, p53 as a DNA binding protein, contributes an external proofreading function; upon excision of the dU, the p53 dissociates, thus letting the transfer of the substrate with the correct 3′-terminus to DNA polymerase and renewal of DNA synthesis. 3) Increased removal of incorporated dU was witnessed in mit(p53−/−) by complementation of exonuclease activity executed by cyt(p53−/−) overexpressing wtp53, but not exonuclease-deficient mutant p53-R175H. 4) Augmented abundance of p53 in IR-mit(p53+/+) or nutlin-treated mit(p53+/+) caused in a decreased dU incorporation. Altogether, the data corroborate the significance of p53 as an exonuclease for the excision of non-canonical dUTP, thus expanding the spectrum of DNA damage sites exploited for proofreading as a trans-acting protein. The fact that p53 may improve the accuracy of DNA synthesis by the decline of dU incorporation capacity plus excision from the nascent strand, infers that p53 potentially may have an additional editing activity in error-correction pathways thus determining replication accuracy and preservation of genome integrity.
p53 exhibits the functional heterogeneity in its basal (non-induced) state and under various p53 inducible circumstances [9, 10]. p53 unveils pro-apoptotic and pro-survival functions thereby triggering mitochondrial-directed apoptosis and mtDNA repair [13, 14]. The dichotomy in p53 action is dictated by its level of expression and activity, and the availability of its interacting partners. The data attained raise questions about the possible functional implication of p53-exonucleolytic proofreading activity in the mitochondria of normal and tumor cells.
In sum, the current results suggest that p53 in mitochondria may facilitate mtDNA damage repair functions resulting from uracil–DNA misincorporation.
The colorectal cancer cells isogenic for p53, HCT116(p53+/+) and HCT116(p53−/−), were grown in McCoy's medium supplemented with 10% FBS. HepG2 cells were grown in EMEM (ATCC, Manassas, VA, USA) with 10% FBS (ATCC) and penicillin/streptomycin (Invitrogen). Cells were maintained at 37°C in a humidified 95% air, 5% CO2 atmosphere.
Mitochondria were prepared as described .
Equal amounts of total protein of mitochondrial, nuclear or cytoplasmic fractions were subjected to WB analysis as described . After transferring proteins onto nitrocellulose, the blots were probed with the indicated antibodies, and protein bands were developed by using chemiluminescence. The following antibodies were used: monoclonal antibodies for p53 (Do-1) (Oncogene), for cytochrome-c (Oncogene), for VDAC (Santa Cruz Biotechnology) and polyclonal antibody for c-jun (Santa Cruz Biotechnology).
The DNA primer extension assay was used allowing simultaneous detection of both, degradation and extension. The sequence of the template-primers used for the experiments are depicted in figures. The primers end labeled at the 5′-end with T4 polynucleotide kinase (Fermentas) and [γ-32P] ATP were annealed to the template DNA as described . The incubation mixture (10 μl) contained 50 mM Tris HCl (pH 7.5), 5 mM MgCl2, 1 mM DTT, 0.1 mg/ml BSA, 5′-end labeled substrates, nucleotides and mitochondrial protein extracts. The reaction products (polymerization or excision) were analyzed by electrophoresis through 16% polyacrylamide gel electrophoresis (PAGE) and detected by autoradiography, as described . The extended and excised products were quantified by densitometric scanning of gel autoradiographs and the percentage of the total amounts of primers extended or excised was calculated.
We cordially thank Dr. Orit Uziel for critically reading the manuscript. This research was supported by grant from Israel Cancer Association.
CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interest.