A New Mitochondrial Redox Regulator: Human AP endonuclease/Redox Factor APE1/Ref-1 modulates mitochondrial function after Oxidative Stress by Regulating Transcriptional Activity of

Free Radical Biology and Medicine 53 (2012) 237–248

Contents lists available at SciVerse ScienceDirect

Free Radical Biology and Medicine

0891-58

http://d

Abbre

immuno

electrop

chondri

oxidase

NRF1, n

recepto

peroxid

outer mn Corr

E-m

journal homepage: www.elsevier.com/locate/freeradbiomed

Original Contribution

Human AP endonuclease/redox factor APE1/ref-1modulates mitochondrialfunction after oxidative stress by regulating the transcriptional activityof NRF1

Mengxia Li a, Carlo Vascotto b, Shangcheng Xu c, Nan Dai a, Yi Qing a, Zhaoyang Zhong a,Gianluca Tell b, Dong Wang a,n

a Cancer Center and Department of Pathology, Research Institute of Surgery, Daping Hospital, Third Military Medical University, Chongqing 400042, People’s Republic of Chinab Department of Medical and Biological Sciences, University of Udine, 33100 Udine, Italyc Department of Occupational Health, Third Military Medical University, Chongqing 400038, People’s Republic of China

a r t i c l e i n f o

Article history:

Received 4 January 2012

Received in revised form

6 March 2012

Accepted 6 April 2012Available online 11 May 2012

Keywords:

NRF1

APE1

Mitochondrial function

Oxidative stress

Free radicals

49/$ - see front matter & 2012 Elsevier Inc. A

x.doi.org/10.1016/j.freeradbiomed.2012.04.00

viations: APE1, apurinic/apyrimidinic endo

precipitation; COX, cytochrome c oxidase; D

horetic mobility-shift assay; ETC, electron tr

al membrane potential; MT-CO1, mitochondr

1; mtDNA, mitochondrial DNA; MTS, mitoch

uclear respiration factor 1; PGC-1a, peroxiso

r g coactivator 1a; ROS, reactive oxygen spec

e; TFAM, mitochondrial transcriptional factor

itochondrial membrane

esponding author. Fax: þ86 23 68757151.

ail address: [email protected] (D. W

a b s t r a c t

Maintenance of mitochondrial functionality largely depends on nuclear transcription because most

mitochondrial proteins are encoded by the nuclear genome and transported to the mitochondria.

Nuclear respiration factor 1 (NRF1) plays a crucial role in regulating the expression of a broad range of

mitochondrial genes in the nucleus in response to cellular oxidative stress. However, little is known

about the redox regulatory mechanism of the transcriptional activity of NRF1. In this study, we show

that the human apurinic/apyrimidinic endonuclease/redox factor (APE1/Ref-1) is involved in mito-

chondrial function regulation by modulating the DNA-binding activity of NRF1. Our results show that

both APE1 expression level and its redox activity are essential for maintenance of the mitochondrial

function after tert-butylhydroperoxide-induced oxidative stress. Upon knocking down or redox

mutation of APE1, NRF1 DNA-binding activity was impaired and, consequently, the expression of its

downstream genes, including Tfam, Cox6c, and Tomm22, was significantly reduced. NRF1 knockdown

blocked the restoration of mitochondrial function by APE1 overexpression, which further suggests APE1

regulates mitochondrial function through an NRF1-dependent pathway. Taken together, our results

reveal APE1 as a new coactivator of NRF1, which highlights an additional regulatory role of APE1 in

maintenance of mitochondrial functionality.

& 2012 Elsevier Inc. All rights reserved.

Introduction

Human apurinic/apyrimidinic endonuclease/redox factor (APE1/Ref-1,hereafter referred to as APE1)1 is a dual-function protein,which plays two distinct yet critical roles in many important cellularbiological processes. On one hand, APE1 is essential in DNA baseexcision repair by acting as an AP-site endonuclease; on the otherhand, APE1 exerts its reduction–oxidation (redox) modificationactivity on some transcriptional factors, thus regulating their

ll rights reserved.

2

nuclease; ChIP, chromatin

OX, doxycycline; EMSA,

ansport chain; MMP, mito-

ial-encoded cytochrome c

ondrial targeting sequence;

me proliferator-activated

ies; TBHP, tert-butylhydro-

A; TOMM, translocase of

ang).

DNA-binding activity (reviewed in [1]). APE1 stimulates the DNA-binding activity of those transcription factors by switching theadjacent cysteine between disulfide bond and thiol groups, which ismodulated by the intracellular redox status. The biological relevanceof APE1 is highlighted by studies that reported the impossibility ofestablishing a homozygous deficient APE1 mouse or knockoutmammalian cell lines [2,3]. Recently, a growing body of evidencehas revealed additional functional activities for APE1 protein, such asits role in RNA metabolism [4,5] and its transcriptional regulatoryfunction modulated through specific acetylation of the protein itself[6,7]. APE1 is now considered an important player in many physio-logical and pathological processes, including tumorigenesis [8], aging[9], angiogenesis [10], and oxidative stress signaling [11].

The mitochondrion is the major source of energy for eukaryoticcells via oxidative phosphorylation through the electron transporta-tion chain (etc). When challenged with oxidative stress, endogenousreactive oxygen species (ROS) are mainly (495%) generated by theetc as by-products, which makes the mitochondrion the mainredox-sensitive subcellular organelle [12]. Additionally, because ofthe features of the mitochondrial genome, including the lack of

M. Li et al. / Free Radical Biology and Medicine 53 (2012) 237–248238

protection by histones, its high gene density, and a limited range ofDNA damage repair activities, mitochondrial DNA (mtDNA) is moresusceptible to damage produced by ROS with respect to nuclearDNA [13]. ROS promote accumulation of mtDNA oxidative lesions,which are reported to be 10- to 15-fold higher than in the nucleargenome, thus leading to downregulation of mtDNA-encoded genes,therefore impairing the mitochondrial membrane potential [14,15].This loss of mitochondrial membrane integrity produces release ofsome mitochondrial proapoptotic molecules, such as cytochrome c,

which in turn triggers apoptosis [16]. It is noteworthy that somestudies reported a regulatory role of APE1 on mitochondrial func-tion. Pioneering studies suggested that APE1 translocates intomitochondria under oxidative stress in lymphocytes and humanthyroid cells. They showed that accumulation of APE1 withinmitochondria, as a consequence of oxidative stress, may act as aprotective mechanism facilitating mitochondrial genome mainte-nance and preventing apoptosis [17,18].

APE1 is predominantly localized in the nucleus of most cell types,and mitochondrial APE1 is basically scarce and its mitochondrialtranslocation is largely conditional [19]. On the other hand, Singhet al. [8] found that mtDNA depletion promotes nuclear genomicinstability and that the APE1 expression level is controlled bymtDNA copy number. That study introduced the idea that APE1may mediate the intergenomic cross talk between nucleus andmitochondria [8]. One recent study by Tell and colleagues [20]showed that the stable knockdown of APE1 in HeLa cells increasedthe release of proapoptotic factors, including cytochrome c andapoptosis-inducing factor, from mitochondria to cytosol. Data pro-vided by comparing proteomic and genomic analyses indicated alsothat APE1 loss of expression causes downregulation of some anti-oxidant genes, implying that the role of APE1 in regulating mitochon-drial functions may also be indirect [20]. Our unpublished data,obtained from the same inducible APE1 knockdown HeLa cell lines,also indicated that after hydrogen peroxide-induced oxidative stress,a number of mitochondrial-related genes are downregulated inAPE1-deficient cells. Interestingly, most of the affected mitochondrialgenes are encoded by the nuclear genome and their protein productsare transported to mitochondria through mitochondrial outer mem-brane translocase receptors. We therefore hypothesized that nuclearAPE1 may modulate mitochondrial function by regulating nuclear-encoded mitochondrial genes at the transcriptional level.

It is known that the transcriptional regulation of severalmitochondrial constitutive members, including etc members,translocases, and transcriptional factors, is critical to mitochondrialfunction. Because the mitochondrial genome is small, encodingonly 13 polypeptides, most mitochondrial proteins are encoded bythe nuclear genome and transported to the mitochondria. Hence,mitochondrial function largely depends on nuclear transcription[21]. Nuclear respiration factor 1 (NRF1) is a nuclear transcriptionfactor, which regulates the expression of a broad range of down-stream target genes. A subset of NRF1 downstream target genes isrelated to mitochondrial membrane transportation, etc function,and mitochondrial transcription [22]. The transcriptional activity ofNRF1 was reported to be enhanced after hydrogen peroxide-induced oxidative stress and its downstream mitochondrial targetgenes, including subunits of cytochrome c oxidase (COX) [23] andmitochondrial transcriptional factor A (TFAM) [24], were upregu-lated. These results strongly suggested the existence of a direct linkbetween the transcriptional activity of NRF1 and the cellularredox state. APE1 is an important transcriptional coactivator,which maintains the reduced status of the DNA binding domainof several transcription factors, including NF-kB and p53, whichare stimulated in response to oxidative stress [25]. In this study,we explored the regulatory role of APE1 on mitochondrial functionby measuring the mitochondrial membrane potential and tested thecorrelation existing between APE1 deficiency and the expression of

nuclear-encoded mitochondrial genes. Additionally, we analyzed theactivation of NRF1 DNA-binding activity by APE1. The resultsprovided by our study demonstrate that nuclear APE1 regulatesnuclear-encoded mitochondrial-related genes by modulating NRF1transcriptional activity in a redox-dependent manner.

Materials and methods

Materials

Minimal essential medium a, Opti-MEM I reduced-serum med-ium, fetal bovine serum, Lipofectamine 2000 transfection reagent,TRIzol RNA isolation reagent, and primers were from Invitrogen(Grand Island, NY, USA). Hydrogen peroxide, tert-butylhydroperoxide(TBHP), E3330, CRT0044876, mitochondria staining kit, and syntheticsiRNA against NRF1 were from Sigma–Aldrich (St. Louis, MO, USA).Tetrahydrofuran site-containing oligonucleotides were from Takara(Dalian, China). pGL4.17 vector, Steady-Glo luciferase assay kit, T4polynucleotide kinase, T4 ligase, restriction endonucleases, and high-fidelity Pfu DNA polymerase were from Promega (Madison, WI, USA).Halt protease inhibitor cocktail, protein A/G–agarose beads, GSTprotein interaction pull-down kit, LightShift chemiluminescenceEMSA kit, Super Signal West Pico chemiluminescence reagents, andhorseradish peroxidase (HRP)-conjugated anti-mouse or anti-rabbitIgG antibodies were from Pierce (Rockford, IL, USA).

Cell lines used

All these procedures have been already described [5,20]. APE1siRNA sequences and scrambled siRNA sequences were cloned intoBglII and HindIII restriction sites of the pTer vector, which contains adoxycycline (DOX)-responsive promoter to generate the pTer/APE1vector and pTer/Scr, respectively. HeLa cells were first transfectedwith pcDNA6/TR to generate stable Tet-repressor-expressing cellclones, which were further selected for the acquired resistance byincubation with blasticidin (5 mg/ml; Invitrogen) for 14 day. Indivi-dual cell clones expressing higher levels of Tet repressor wereisolated and then selected for transfection with the pTer/APE1vector previously linearized with Bst1107I (Fermentas, St. LeonRot, UK) and subjected to selection with zeocin (200 ng/ml; Invitro-gen) for 14–21 day. Single clones were isolated and analyzed forAPE1 expression by immunoblotting after 10 day of doxycycline(Sigma) inducement at the final concentration of 1 mg/ml. As acontrol, we used cell clones transfected with the pTer/Scr vector. Forgeneration of APE1 knock-in cell lines, an APE1 expression vectorwas generated by cloning an EcoRI–BamHI fragment into thep3XFLAG-CMV-14 vector (Sigma). To avoid the degradation of theectopic APE1 mRNA by the specific siRNA sequence, two nucleotidesof the APE1-cDNA coding sequence were mutated using a site-directed mutagenesis kit (Stratagene), leaving unaffected the APE1amino acid sequence. The site-directed mutagenesis kit was used togenerate the APE1C65S mutant, which was confirmed by DNAsequencing (MWG, Ebersberg, Germany). Then, the APE1 siRNAclone was transfected with p3XFLAG-CMV/APE1WT and mutants,previously digested with ScaI (Fermentas). Forty-eight hours aftertransfection, the cells were subjected to selection with Geneticin(Invitrogen) for 14 day and selected for the acquired resistance.Individual cell clones were isolated and after 10 day of doxycyclinetreatment, whole-cell lysates were analyzed for APE1 expression byWestern blot.

Mitochondrial membrane potential assay

The mitochondrial membrane potential (MMP) was assessedusing the JC-1 mitochondria staining kit for flow cytometry,

M. Li et al. / Free Radical Biology and Medicine 53 (2012) 237–248 239

following the manufacturer’s recommendations. Cells were trea-ted with various doses of hydrogen peroxide or TBHP for theindicated time and then incubated in medium containing JC-1probe at the working concentration of 2.5–5 mg/ml for 30 min in a37 1C incubator. After the cells were washed with ice-cold JC-1binding buffer twice, MMP was measured immediately by usingflow cytometry. JC-1 is a mitochondrial-selective sensor andaggregates in normal and highly polarized mitochondria, result-ing in a red emission of 590 nm after excitation at 490 nm. Upondepolarization of the mitochondrial membrane, JC-1 forms greenfluorescent monomers and determines an increase in the green/red (FL1/FL2) fluorescence intensity ratio. Thus, the loss of JC-1aggregates directly correlates with changes in the MMP. Treat-ment with valinomycin, representing mitochondrial gradientdissipation, was performed as a positive control.

Cytochrome c release assay

To measure the cytochrome c release from mitochondria tocytoplasm after oxidative stress, mitochondria-free cytosolicfractions were subjected to Western blot using an antibodyagainst cytochrome c. Mitochondria-free cytosolic fraction wasprepared by using the protocol described before. Briefly, 5�107

cells were washed once with 10 ml of grinding buffer (250 mMsucrose, 2 mM EDTA, 1 mg/ml bovine serum albumin, pH 7.4) andcollected by centrifugation at 800 g for 5 min, at 4 1C. The pelletwas resuspended in 200 ml of grinding buffer and sonicated on iceat 30 W for 15 s. Approximately 50–70% cell lysis was ensured bychecking under a microscope. The supernatant was immediatelycentrifuged at 8500 g for 20 min, at 4 1C, to pellet the mitochon-dria. The obtained supernatant contained the cytosolic fraction.

Western blot assay and antibodies

Equal amounts of nuclear or cytosolic extract or whole-celllysate, obtained from HeLa and U2OS cells, were electrophoresedby 10% SDS–PAGE. Proteins were then transferred onto polyviny-lidene difluoride membranes (Bio-Rad, Hercules, CA, USA). Afterbeing blocked in TBST (50 mM Tris–HCl, pH 7.5, 150 mM NaCl,and 0.1% (v/v) Tween 20) containing 5% (w/v) nonfat dry milkfor 1 h at room temperature, membranes were incubated withthe specific primary antibody. After three washes with TBST,the membranes were incubated for 1 h at room temperature withthe appropriate peroxidase-conjugated secondary antibodies.Then, the membranes were washed five times with TBST and theblots were reacted with chemiluminescence reagents and revealedwith Biomax-Light films (Kodak, Rochester, NY, USA). Band inten-sities were analyzed using the Gel Doc 2000 apparatus and software(Quantity One; Bio-Rad).

Suppliers of and incubation conditions for antibodies used forWestern blot were as follows: anti-APE1 monoclonal (Novus), 1 hat 37 1C, dilution 1:5000; anti-Flag (M2; Sigma) monoclonal, 1 hat 37 1C, dilution 1:2000; anti-NRF1 monoclonal (Santa Cruz),overnight at 4 1C, dilution 1:400; anti-b-actin monoclonal (Sigma),1 h at 37 1C, dilution 1:2000; anti-cytochrome c monoclonal (SantaCruz) overnight at 4 1C, dilution 1:500; anti-Sp1 polyclonal (SantaCruz) overnight at 4 1C, dilution 1:500; anti-peroxisome proliferator-activated receptor g coactivator 1a (PGC-1a) monoclonal (Sigma)overnight at 4 1C, dilution 1:2000.

Quantitative RT-PCR

Total RNA was isolated using TRIzol reagent and chloroform/isoamyl alcohol precipitation according to the manufacturer’sinstructions. RNA concentrations were determined by spectro-photometer (Eppendorf AG, Hamburg, Germany). Subsequently,

cDNA was synthesized based on 1 mg of total RNA using theReverTra Ace reversal transcription kit (Toyobo, Osaka, Japan).Quantitative RT-PCR was performed using SYBR Premix Ex Taq(Takara) in a LightCycler 480 real-time PCR system (Roche,Indianapolis, IN, USA). Primer pairs for genes, including Nrf1,

Tfam, Cox6c, Tomm22, Apex1, and b-actin, were designed to yield100- to 300-bp amplicons, which are suitable for real-timequantitation. Sequences of the primers are available upon request.

Coimmunoprecipitation assay

Cells were harvested by scraping and washed once with ice-cold phosphate-buffered saline (PBS). The cell pellet was resus-pended and incubated in immunoprecipitation (IP) lysis buffer(Beyotime Institute of Biotechnology, Jiangsu, China) supplemen-ted with protease inhibitor cocktail (Pierce) at a cell density of107 cells/ml on ice for 30 min. After centrifugation at 12,000g for10 min at 4 1C, the supernatant was collected as total cell lysate.Protein concentration was determined by using the Bradfordassay (Bio-Rad). Samples were precleared by incubating withprotein A/G–agarose resin for 30 min on ice and then coimmuno-precipitated for 3 h using anti-Flag M2 antibody, APE1 antibody,NRF1 antibody, or PGC-1a antibody following the manufacturer’sinstructions. After incubation, protein A/G–agarose resin was addedand incubated for 1 h at 4 1C. After three washes with PBS contain-ing protease inhibitor, the pellet containing agarose beads togetherwith binding proteins was mixed with sample buffer and incubatedat 100 1C for 5 min. The samples were then stored at �80 1C orsubjected to Western blotting analysis immediately.

DNA affinity precipitation

DNA affinity precipitation analyses were performed as describedwith minor modifications [26]. Ten picomoles of gel-purified, biotin-labeled 55-mer oligonucleotide corresponding to a minimal func-tionally active segment of the NRF1 binding sites in the promoterregion of the Tfam gene, 50–CCGGGGTACGCTCTCCCGCGCCTGCGC-CAATTCCGCCCCGCCCCGCCCCCATCTA-30 (bold sequence representsNRF1 binding site and italic sequence represents Sp1 binding sites),was incubated with 120 ml of streptavidin–agarose resin (60 mlsettled resin; Pierce) for 30 min at room temperature. The biotiny-lated oligonucleotide bound to the streptavidin–agarose wascollected by centrifuging at 500 g for 1 min. Bead-associatedoligonucleotide was washed two times with Tris–EDTA andequilibrated with PBS. Twenty-five micrograms of nuclear pro-teins (25 mg) from HeLa or U2OS cells was then added to theoligonucleotide–streptavidin–agarose pellet and incubated on icefor 2 h. After incubation, the pellet was washed twice with PBS toremove unbound proteins. Protein–DNA complexes were thenresuspended in sample buffer and incubated at 100 1C for 5 minto elute the bound proteins. The samples were then subjected toWestern blot analysis.

Electrophoretic mobility-shift assay (EMSA)

EMSA was accomplished according to the user’s instructionmanual of the LightShift chemiluminescence EMSA kit with minormodifications. Briefly, 5 mg of nuclear extracts was incubated with30-biotin-labeled and purified double-stranded oligonucleotideprobes. The probes containing the NF-kB consensus (NF-kBF, 50–AGTTGAGGGGACTTTCCCAGGC-30, and NF-kBR, 30–TCAACTCCCCT-GAAAGGGTCCG-50) or the NRF1 consensus in the Tfam promoter(NRF1/TFAMF, 50–CGCTCTCCCGCGCCTGCGCCAATT-30, and NRF1/TFAMR, 50–GGGCGCGGACGCGGTTAAGGCGGG-30) were synthe-sized (Invitrogen, Shanghai, China). After incubation, the sampleswere separated on a prerun 5% polyacrylamide gel at 100 V for

M. Li et al. / Free Radical Biology and Medicine 53 (2012) 237–248240

90 min and then transferred to a Zeta-Probe GT nylon membrane(Bio-Rad). The probes were detected by HRP-conjugated strepta-vidin (1:300) and the bands were visualized by ECL reagentsprovided with the kit. The resultant bands were quantified usingQuantity One imaging software (Bio-Rad).

Chromatin immunoprecipitation (ChIP)

ChIP assay were performed to analyze the DNA binding affinityof NRF1 to a specific promoter region of its downstream genesusing a ChIP kit (Millipore) following the manufacturer’s instruc-tions. Briefly, cells plated in 100-mm petri dishes and subjected tovarious TBHP treatments were collected and crosslinked withformaldehyde. The cells were then lysed, the chromatin wasbroken down through ultrasonication, and the lengths of DNAfragments between 400 and 600 bp were confirmed by agarosegel electrophoresis. Immunoprecipitation was performed usingNRF1 antibody (Santa Cruz) to fractionate NRF1 protein–DNAcomplexes. Quantitative PCR was then performed to amplify the Tfam

gene promoter region (containing potential specific NRF1 bindingsites) using primers as follows: TFAMP forward, 50–TCTACCG-ACCGGATGTTAGC-30, and TFAMP reverse, 50–CTTCCCAGGGCACT-CAGC-30. The predicted sizes of the products were 215 and 200 bp,respectively. Preimmunoprecipitation lysates were also included asinput controls.

Luciferase reporter gene assay

The transcriptional activity of the Tfam promoter was assessedby luciferase reporter gene assay (Promega) following the man-ufacturer’s instructions. The luciferase vectors with 314 bp (�205to þ145) of the Tfam promoter containing NRF1, NRF2, andoverlapping Sp1 consensus binding sites were constructed basedon the pGL4.17-Neo vector (Promega). Additionally, vectorscarrying mutations at the critical sites for either NRF1 or NRF2binding were also created. HeLa cells with various APE1 statuseswere plated in six-well plates 1 day before transfection after 5 dayof induction with doxycycline, at a density of 400,000 cells/well.Then, the cells were transfected with 4 mg of luciferase reportervector or mutant vectors using Lipofectamine 2000 reagent(Invitrogen). Cells were harvested at 24 h after transfection andlysed using the Promega cell culture lysis reagent. The luciferaseassay was performed using the Promega luciferase assay systemand a GloMax 96 luminometer (Promega). Samples were normal-ized for total protein content using the Bradford protein assay.

Results

The MMP is correlated with the expression level of APE1 after

oxidative stress

We first examined the relationship between the expressionlevel of APE1 and the overall mitochondrial function. To eliminatethe bias in all single-knockdown strategies, we employed twobiological models for APE1 expression deficiency in different celltypes. The first is based on HeLa cells stably transfected withinducible siAPE1 to exert knockdown of APE1 expression by DOXinduction, already described [5] and reported in SupplementaryFig. S1A. The second approach was performed by using thesynthetic siRNA to transiently knockdown APE1 expression inU2OS cells. It is known that MMP loss is an early sign of impairedmitochondrial function after oxidative stress. We used a flowcytometry JC-1 staining assay to measure MMP in cells withvarious APE1 expression levels. As shown in Fig. 1A, in APE1stable-knockdown HeLa cells (shRNA), MMP is significantly lower

in the controls or upon TBHP-induced oxidative stress comparedto control cells (Scr-1 cell clone, which was stably transfectedwith scramble shRNA). To further exclude that this relationshipbetween APE1 protein level and MMP is due to the clonal effects,a synthetic siRNA against APE1 was used to transiently knockdown APE1 expression in U2OS cells. We found that 24–72 h aftertwo rounds of transfection represents the ideal time framein which to test the effects by APE1 deficiency (SupplementaryFig. S1B). We then treated cells with TBHP for 6 h and measuredthe consequent MMP loss (Fig. 1B). The results confirmed thatreduced APE1 protein levels correlated with lower MMP afteroxidative stress. To further address the functional meaning of theMMP loss on the activation of apoptosis, we performed acytochrome c release assay, which measures the initiation of themitochondrial-specific apoptotic pathway by releasing the mito-chondrial proapoptotic agents (Fig. 1C). At 6 h after TBHP treat-ment, the cytosolic cytochrome c level increased in shRNA cellsand upon siRNA transfection of U2OS cells compared to theirappropriate controls.

APE1 regulates MMP through its redox activity

APE1 is known as a dual-function protein possessing twodistinct activities of DNA repair and redox regulation of transcrip-tional factors. Previous reports showed that APE1 localizes intomitochondria and participates in mtDNA repair through its APendonuclease activity [18]. However, the majority of APE1 proteinaccumulates within the nuclear compartment and there are fewreports showing an effect of nuclear APE1 on mitochondrialfunction [20]. We then employed two strategies to evaluate theeffects of the redox activity of APE1 on MMP after oxidative stress.First, based on the inducible APE1 knockdown cell shRNA, bothwild-type and redox-deficient mutant APE1 reconstituted cell lines,APE1WT and APE1C65S, were used (Supplementary Fig. S1A). Theoligonucleotide incision assay and EMSA were performed to verifythe activation status of APE1 (Supplementary Figs. S2A and S2C). InAPE1C65S cells, the redox activity is significantly reduced with aslight repair activity loss [28]. Compared with the APE1WT-expres-sing cells, MMP was significantly reduced after hydrogen peroxide-induced oxidative stress in the APE1C65S cells (po0.01; Fig. 2A).Second, we used the APE1 redox inhibitor E3330 to suppress theAPE1 redox activity [29]. APE1 activities were tested by using theaforesaid methods and are shown in Supplementary Figs. S2B andS2D. In agreement with the APE1 reconstituted cell model, theresults showed that MMP was significantly reduced in the E3330-treated group at 6 h after TBHP-induced oxidative stress (Fig. 2B).Additionally, cytochrome c release was assayed and the resultsshown in Fig. 2C indicated that the mitochondrial-mediated proa-poptotic molecule release was increased in the APE1C65S cells and inthe E3330-treated U2OS cells. The activity assays indicate that bothC65S mutation and E3330 treatment robustly suppress the redoxactivity of APE1, suggesting that the redox activity of APE1 plays acritical role in regulating mitochondrial function after oxidativestress.

Mitochondrial-related genes are downregulated

in APE1-deficient cells

To explore the mechanism responsible for the mitochondrialregulatory role of APE1 and based on the already published datashowing that expression of some mitochondrial genes may beaffected by APE1 silencing under normal conditions [20], weevaluated whether mitochondrial-related genes may be directlyregulated by APE1 also after oxidative stress. We then analyzedmRNA expression of four nuclear-encoded genes, Nrf1, Tfam,

Cox6c, and Tomm22, by quantitative RT-PCR. NRF1 is an important

Fig. 1. Mitochondrial membrane potential is downregulated in APE1-deficient cells. Mitochondrial function was evaluated by MMP and cytochrome c release post-TBHP

treatment. (A) MMP was measured at 6 h after 100, 200, and 400 mM TBHP treatment (left graph) or at 3, 6, and 12 h after 200 mM TBHP treatment (right graph) in both

Scr-1 and shRNA cells. The scatter plots show the MMP alteration at 6 h after 200 mM TBHP treatment. (B) MMP alteration in U2OS cells transfected with siRNA against

APE1 or scramble RNA after 100 and 200 mM TBHP treatment (left graph) or at 3, 6, and 12 h after 200 mM TBHP treatment (right graph). The scatter plots show the MMP

alteration after 200 mM treatment. (C) Mitochondrial apoptosis initiation was assayed by cytosolic cytochrome c level at 6 h after 100 mM TBHP treatment using Western

blot. The cytochrome c levels were normalized to the level of subcellular markers and are shown in bar graphs. All data in bar graphs are from three independent

experiments. *po0.01 according to the statistics process.

M. Li et al. / Free Radical Biology and Medicine 53 (2012) 237–248 241

nuclear transcription factor that is closely related to mitochon-drial regulation [30]. TFAM is a nuclear-encoded mitochondrialprotein that is a critical factor for mtDNA transcription [24].COX6C is one of the components of the etc encoded by the nucleargenome [23]. TOMM22 is a nuclear-encoded mitochondrial outermembrane receptor, which is important for translocation of mostnuclear-encoded mitochondrial proteins [31]. The expressionlevels of Nrf1, Tfam, Cox6c, and Tomm22 were reported to beregulated by NRF1. As shown in Fig. 3A, in the untreated groups,the mRNA expression levels of Tfam, Cox6c, and Tomm22 wereslightly lower in APE1-deficient cells and APE1C65S reconstitutedcells, compared to Scr-1 and APE1WT cell clones separately.At 12 h after TBHP-induced oxidative stress, Tfam, Cox6c, andTomm22 gene expression was significantly induced in Scr-1 andAPE1WT cells but remained at basal levels in APE1-deficient cellsand APE1C65S reconstituted cells. Nrf1 gene expression level wasunaffected by APE1 status at least in the first 6 h of oxidativestress.

To further confirm the functional loss of those genes, a seriesof further tests was also performed. First, the overall mitochon-drial morphological changes in the various groups were visualizedunder the microscope by MitoTracker staining (Fig. 3B). Inthe control cells Scr-1 and APE1WT, there is an extensive and

well-organized network of mitochondria surrounding the nucleus.In contrast, in APE1-deficient cells, including shRNA and APE1C65S,the mitochondrial fluorescence was greatly decreased upon TBHPtreatment. Because TFAM contains a typical MTS, and it targetsto the mitochondrion through a TOMM20/TOMM22-dependentmechanism [32], we tested the mitochondrial transmembranetransportation efficiency by measuring the TFAM mitochondrialdistribution level using both morphological and biochemical ana-lyses (Fig. 3C). EGFP-tagged TFAM was observed mainly localizedin the mitochondria of Scr-1 and APE1WT cells. However, inthe APE1C65S and shRNA cells, a considerable level of TFAM isdistributed in the cytosolic fraction, suggesting downregulationof mitochondrial cross-membrane transportation. Then, TFAMtranscriptional activity was assayed by EMSA (Fig. 3D). Whole-celllysates were prepared and incubated with a biotin-conjugatedprobe containing the nucleotide sequence of the heavy-strandpromoter. The results indicated that TFAM DNA-binding activitywas significantly decreased in APE1-deficient cells. Furthermore,we measured the mRNA expression of the mtDNA-encoded genecytochrome c oxidase 1 (MT-CO1), which is a downstream gene ofTFAM. The mRNA level of MT-CO1 was reduced in APE1-deficientcells or redox function-impaired cells upon TBHP-induced oxida-tive stress (Fig. 3E).

Fig. 2. Mitochondrial membrane potential is affected by the redox activity of APE1. Mitochondrial function was evaluated by MMP and cytochrome c release after TBHP

treatment. (A) MMP was measured at 6 h after 100, 200, and 400 mM TBHP treatment (left) or at 3, 6, and 12 h after 200 mM TBHP treatment (right) in both APE1WT and

APE1C65S cells. The scatter plots show the MMP alteration at 6 h after 200 mM treatment. (B) MMP alteration in U2OS cells treated with various doses of the redox inhibitor

E3330 at 6 h after TBHP treatment (left) or at 3, 6, and 12 h after 10 mM E3330 and 100 mM TBHP treatment (right). The scatter plots show the MMP alteration after 10 mM

E3330 pretreatment. (C) Mitochondrial apoptosis initiation was assayed by cytosolic cytochrome c level at 6 h after 100 mM TBHP treatment using Western blot. The

cytochrome c levels were normalized to the levels of subcellular markers and are shown in bar graphs. All data in bar graphs are from three independent experiments.

*po0.01 according to the statistics process.

M. Li et al. / Free Radical Biology and Medicine 53 (2012) 237–248242

APE1 protein is associated with the NRF1-transcriptional complex

and enhances the NRF1/PGC-1a interaction

NF-kB is an important transcription factor involved in thecellular response to oxidative stress, which is known to beregulated by the redox function of APE1. Based on the geneexpression profile obtained at 6 h after oxidative stress (datanot shown), the expression levels of NRF1, which is transcription-ally regulated by NF-kB, were comparable in both APE1-proficientand APE1-deficient cells. However, the downstream genes ofNRF1 were significantly induced after oxidative stress only inAPE1-expressing cells. We therefore hypothesized that activationof NRF1 may occur at the posttranscriptional level, probablythrough a redox modification by APE1. To test this hypothesis,we first measured the level of protein–protein interaction occur-ring between APE1 and NRF1 or its coactivator, PGC-1a, afteroxidative stress using coimmunoprecipitation assays. As shownin Fig. 4A–C, when APE1 was precipitated, NRF1 was detectedas part of the immune complex. When the immunoprecipita-tion experiments were performed with antibodies against NRF1,APE1 was also detected in the complex. Additionally, APE1 binds

to NRF1 with much higher affinity in the nuclear extracts ofTBHP-treated cells. In contrast, there is no significant interac-tion between APE1 and PGC-1a according to the co-IP assay.In addition, we used anti-FLAG antibody to immunoprecipitateexogenous APE1 in APE1WT and APE1C65S reconstituted cells.Notably, our results showed that the C65S mutation significantlydecreased the apparent affinity of APE1 for NRF1, suggestingthat the redox activity is critical for the stability of the APE1/NRF1complex formation (Fig. 4D). To confirm the presence of APE1in the NRF1 transcriptional complex, DNA affinity precipita-tion was used. An oligonucleotide containing the previouslyreported active NRF1 binding site within the Tfam promoter waslabeled with biotin and used to precipitate the potential tran-scriptional components bound to it. According to the Westernblot results, shown in Fig. 4E, NRF1 and its coactivator, PGC-1a,bind to the canonical DNA binding site in the Tfam promoterat 3 h post-TBHP treatment. Notably, Sp1 constantly binds tothis region under either basal or oxidative stress conditions.Most significantly, APE1 was detected in the complex afteroxidative stress, suggesting that it may also act as a coactivatorof NRF1.

Fig. 3. The expression of mitochondrial-related genes and its functional alteration affected by redox activity of APE1. (A) The expression of four nuclear genome-encoded

genes, Nrf1, Tfam, Cox6c, and Tomm22, was analyzed in Scr-1/shRNA/APE1WT/APE1C65S cells or after 200 mM TBHP treatment by quantitative RT-PCR. The expression levels

were statistically processed from three independent experiments. *po0.01 according to the statistics process. (B) The mitochondrial morphological changes in TBHP-

treated Scr-1/shRNA/APE1WT/APE1C65S cells were visualized by MitoTracker red staining. Scale bar, 25 nm. (C) Mitochondrial transmembrane transportation was assayed

by TFAM subcellular translocation. EGFP-tagged TFAM expression vector was transfected into HeLa cells and its subcellular location was observed by Western blot.

(D) TFAM transcriptional activity was assayed by its DNA-binding activity using EMSA. (E) Additionally, the expression of its downstream gene MT-CO1, which is a

mitochondrial genome-encoded gene, was analyzed using quantitative RT-PCR. *po0.05 according to the statistics process.

M. Li et al. / Free Radical Biology and Medicine 53 (2012) 237–248 243

Fig. 4. APE1 is present in the NRF1 transcriptional complex and enhances the NRF1/PGC-1a interaction. (A–C) To detect the presence of APE1 in the NRF1 transcriptional

complex, co-IP assays were performed in HeLa cell nuclear extracts after 1 h of 100 mM TBHP treatment. Secondary antibody alone precipitation was used as a negative

control for specificity of the immunoprecipitation by primary antibodies. Ten percent of the pre-precipitation nuclear extracts was also loaded as an input control.

(D) Precipitation of the exogenously expressed proteins using anti-FLAG antibody in both APE1WT and APE1C65S nuclear extracts after 1 h of 100 mM TBHP treatment. FLAG

precipitation was also performed in the FLAG-free Scr-1 cells to indicate the specificity. (E) The presence of NRF1, PGC-1a, Sp1, and APE1 on the double-stranded DNA

probe containing the NRF1 binding site is shown through DNA affinity precipitation.

M. Li et al. / Free Radical Biology and Medicine 53 (2012) 237–248244

APE1 regulates NRF1 transcriptional activity through its redox

function

The effect of APE1 deficiency on the DNA-binding activity ofNRF1 from the nuclear extract was then assayed by EMSA and ChIPanalysis. For EMSA, synthetic oligonucleotide DNA probes containingthe putative NRF1 binding site in the Tfam gene promoter was used.We first verified the specificity of the DNA-binding activity of NRF1and the correlation with the mixture redox state. As shown inFig. 5A, the retarded bound complexes were specifically bound byNRF1 antibody, which further formed a ‘‘supershift.’’ Also, theretarded bound complexes were competitively bound by unlabeledprobe but unaffected by mutated probe. Moreover, the DNA-bindingactivity of NRF1 was elevated in the DTT-containing mixture butreduced in the hydrogen peroxide-containing mixture. Additionalresults indicate that the affinity of NRF1 for the promoter regions ofdownstream genes was reduced in APE1-knockdown cells (Fig. 5B)and redox-activity-deficient cells (i.e., APE1C65S reconstituted cells;Fig. 5C). When the EMSA reaction mixture was treated with theredox inhibitor E3330, the NRF1 DNA-binding activity to the probeswas also reduced (Fig. 5C). The results of a ChIP assay furtherindicated that the binding of NRF1 to the Tfam promoter, induced byoxidative stress, was compromised in shRNA- and APE1C65S-expres-sing cells compared to the Scr-1 and APE1WT, respectively (Fig. 5D).

As an additional means of assessing the transcriptional activity ofNRF1 downstream genes a reporter assay was used. To this aim, a

luciferase reporter construct containing most of the Tfam promoterregion, termed pGL-TP, was created. At 12 h post-pGL-TP transfec-tion, the fluorescence intensity, which is a measure of promoteractivity, significantly increased after a 6-h TBHP–hydrogen peroxidetreatment. However, the fluorescence readings were dramaticallyreduced in APE1 shRNA cells, and a significant drop was alsoobserved in APE1C65S cells (Fig. 5E). Altogether, these results newlyestablish a regulatory role for APE1 in DNA-binding and transcrip-tional activity of NRF1.

APE1 regulates mitochondrial-related genes through an

NRF1-dependent pathway

Because some other nuclear transcription factors may alsoparticipate in mitochondrial regulation, it was necessary to confirmthat APE1 specifically regulates mitochondrial function through anNRF1-dependent pathway. We combined inducible APE1 knockdownand reconstituted cells and performed the NRF1 siRNA transfectionafter 1 day of DOX induction. During this time, endogenous APE1expression was slightly reduced down to 30–50% of wild-type-expressing cells, and exogenous APE1 reexpression in APE1WT cellsreconstituted the original APE1 expression levels. After NRF1 siRNAtransfection, NRF1 expression levels in cells bearing various APE1expression levels were significantly reduced according to Westernblot assays (Fig. 6A). We then tested Tfam expression, representingall NRF1-targeted genes, by using quantitative RT-PCR as the readout.

Fig. 5. APE1 regulates NRF1 transcriptional activity through its redox activity. The DNA-binding activity of NRF1 to its downstream gene Tfam promoter region was

measured by EMSA. (A) Anti-NRF1 antibody, 10-fold unlabeled ‘‘cold’’ probe, and mutated ‘‘cold’’ probe were employed to estimate the specificities of both the NRF1

binding and the probe. DTT and H2O2 were added in the mixture to measure the impact of the redox status on the DNA-binding activity of NRF1. (B) DNA-binding activity

of NRF1 was affected by APE1 level, shown in Scr-1/APE1shRNA and APE1 siRNA-transfected U2OS cells. (C) DNA-binding activity of NRF1 was affected by APE1 redox

activity, shown in APE1WT/APE1C65S and E3330-treated U2OS cells. (D) ChIP assay using NRF1 antibody to precipitate the DNA regions that bind to NRF1 in Scr-1, shRNA,

APE1WT, and APE1C65S cells at 1 h after 100 mM TBHP treatment. The pre-precipitation lysates were included as an input control. The Tfam promoter regions were detected

and quantified using real-time PCR. The PCR amplicon levels were normalized to input amplicons and values are from three independent experiments. (E) NRF1

transcriptional activity was then assayed by luciferase reporter assay. The luciferase reporter vector containing the wild-type Tfam promoter region or mutant in critical

sites of the NRF1 or NRF2 binding sequence was transfected into Scr-1, shRNA, APE1WT, and APE1C65S cells. Data were normalized to protein concentration and those

shown in the bar graph are from three independent experiments. *po0.01 and **po0.05 according to the statistics process.

M. Li et al. / Free Radical Biology and Medicine 53 (2012) 237–248 245

The results indicated that APE1 reexpression restored the Tfam

expression levels compared to the shRNA cells. However, NRF1knockdown resulted in a dramatic decrease in Tfam expression inAPE1WT cells, indicating that APE1 regulates Tfam mRNA expressionin an NRF1-dependent manner (Fig. 6B). Consequently, as shown in

Fig. 6C, mitochondrial functionality, which is impaired by APE1deficiency, was restored by reexpressing APE1. As expected, NRF1knockdown blunted the MMP functional recovery in APE1-reexpres-sion cells but not in the APE1-deficient cells, confirming the leadingrole of the APE1–NRF1 axis in mitochondrial functionality.

Fig. 6. APE1 regulates mitochondria-related genes through an NRF1-dependent pathway. Both APE1 knockdown and knock-in cells were induced by DOX for 1 day and

then siRNA transfection was performed. (A) APE1 and NRF1 expression levels at 48 h post-siRNA transfection were tested by Western blot. (B) The Tfam gene expression

was quantified at 12 h after 100 mM TBHP treatment by real-time RT-PCR. Data shown in the bar graph are from three independent experiments. *po0.01 between siGFP-

transfected APE1 WT group and other groups according to the statistics process. (C) MMP was measured at 6 h after TBHP treatment. Data shown in the bar graph are from

three independent experiments. *po0.05 between siGFP-transfected APE1 WT group and other groups according to the statistics process.

M. Li et al. / Free Radical Biology and Medicine 53 (2012) 237–248246

Discussion

Modulation of mitochondrial function after oxidative stress isa critical cellular process given its close relationship with cell fatedetermination. However, the maintenance of normal mitochon-drial function largely depends on the nuclear transcriptionbecause the vast majority of functional mitochondrial proteinsare encoded by the nuclear genome and transported to themitochondria. In this study, we report APE1 as a new player inthe nuclear control of mitochondrial function. The resultsobtained indicate that APE1 regulates the expression of somenuclear-encoded mitochondrial constitutive genes and conse-quently modulates mitochondrial functions in response to oxida-tive stress. Those functions, possibly affected by APE1, include,but are not limited to, mitochondrial transcription, transmem-brane transportation, etc capacity. As a result, a deficiency ofAPE1, or the loss-of-function mutation of cysteine 65 to serine,caused a dramatic MMP functional loss and activation of themitochondrial-mediated apoptotic pathway after oxidative stress.We also provide a possible mechanism for this regulation througha functional interaction of APE1 with the nuclear transcriptionfactor NRF1.

Mechanisms of transcriptional control of mitochondrial long-term biogenesis or transient energy demands in the nucleus arenow well established. A previous study recognized the conservedmotifs at the promoters of a subset of nuclear-encoded mitochon-drial genes, leading to the identification of nuclear transcriptionfactors NRF1 and NRF2, also known as GA-binding proteins [33].Additionally, PGC-1a, initially identified as a coactivator ofperoxisome proliferator-activated receptor g, is a well-establishedcoactivator of NRF1 and NRF2 [34]. The biological significance ofPGC-1a is to integrate physiological signals and enhance mito-chondrial biogenesis and oxidative function. According to theprevious research, the expression of genes downstream of NRF1or NRF2 is modulated in response to intracellular redox status, aswell as the DNA-binding activities of NRF1 and NRF2. We showedhere that the interaction between APE1 and NRF1 or the DNA-binding activity of NRF1 is also dependent on a redox-basedmechanism in the in vitro reaction system. The common featureof the transcription factors regulated by APE1 is the presence ofthe disulfide bond or the adjacent thiol groups at their DNAbinding domain [35]. We searched the amino acid sequence of theDNA binding domain of NRF1 in silico and found that C134, C229,and C269 are in dimensional proximity and could represent the

M. Li et al. / Free Radical Biology and Medicine 53 (2012) 237–248 247

possible regulatory sites for APE1 redox activity. Additional studieswill be required to address this issue. In the NRF1-centeredtranscriptional complex, PGC-1a is considered the most importantcoactivator [30]. However, we failed to observe any alteration in theinteraction between NRF1 and PGC-1a in response to the redoxstatus in vitro, which suggests that stimulation of NRF1 by PGC-1ais redox-independent. In contrast, after TBHP treatment, the inter-action between NRF1 and PGC-1a seemed to be enhanced inAPE1-proficient cells. Our results suggested that APE1 could beeither the redox-dependent coactivator of NRF1 or the possibleredox-dependent mediator between NRF1 and PGC-1a.

Our previous study also provided a plausible explanation forthe mechanism of mitochondrial regulation by APE1, especiallythe functional significance of the cysteine 65 amino acid in thisprocess. The replacement of cysteine 65 with serine impairs themitochondrial accumulation efficiency of APE1 and consequentlyaffects mitochondrial function after oxidative stress, possiblythrough a mechanism influencing the in vivo folding of themolecule [28]. We also observed the occurrence of subcellulartargeting alteration in APE1 after C65S mutation or E3330 treat-ment, as well as the inhibition of APE1 redox activity in the C65Scellular nuclear extracts. Because of the dual-functional featuresof this unique protein, we hypothesized that APE1 may alsoparticipate in mitochondrial modulation through its redox activ-ity. In this study, we focused on the role of APE1 in nucleartranscriptional regulation of some mitochondrial targeting pro-teins. Consistent with their data, we also obtained significantlydifferent gene expression profiles between Scr-1 and APE1 siRNAcells at 6 h after hydrogen peroxide treatment, which include asubset of mitochondrial-related genes. After careful analysis andconfirmation of the differential gene expression array by quanti-tative PCR, we found that the loss of expression of those genesdepended at least partially, if not completely, on the redoxactivity of APE1 (unpublished data). Our data also suggested thatthe reduced level of those mitochondrial-related genes should bethe reason, rather than the result of, mitochondrial functional lossin APE1-deficient cells. Therefore, our present results furtherextend the model that was proposed by Tell’s group initiallyand provide a supplemental explanation for the mitochondrialfunction regulation by C65S mutation of APE1. Taking bothmodels together, in response to oxidative stress, nuclear APE1affected nuclear-encoded mitochondrial gene expression by redoxcontrol of NRF1, and in the meantime, elevated mitochondrialAPE1 levels enhance mtDNA repair.

We also observed that some mitochondrial genome-encodedpolypeptides such as MT-CO1 were downregulated at the mRNAlevel in APE1-deficient cells. The mitochondrial transcription iscontrolled by the common transcriptional initiation complex,which comprises TFAM, TFBM, and mitochondrial RNA polymer-ase, at the D-loop regulatory region, which contains bidirectionalpromoters for transcribing H and L strands [36,37]. Because in ourstudy the expression of TFAM is suppressed in APE1-deficientcells, we concluded that that is the most plausible reason for themitochondrial-encoded gene downregulation. One may askwhether APE1 exerts redox activity after mitochondrial transloca-tion and regulates mitochondrial transcription. Actually, accord-ing to the present observations, we cannot exclude this possibilitybecause we failed to exam the direct redox functions of APE1 inmitochondria. A previous study characterized mitochondrial APE1in bovine liver cells as a truncation form of 33 amino acids at theN-terminus [38]. Because of this deletion, they claimed thatmtAPE1 lacks the K6/K7-dependent acetylation, leaving the C65-dependent redox activity unaffected [38]. Unfortunately, theseauthors failed to identify and confirm the redox activity of mtAPE1.Moreover, we found mtAPE1 in HeLa cells as a full-length protein,suggesting that mtAPE1 structurally has full potential to possess

both activities in the mitochondrion. Thus, APE1 may participate inmitochondrial biogenesis and maintenance through both nuclearand mitochondrial networks, which further raises a theoreticalmodel of an APE1–NRF1–TFAM axis in regulating mitochondrialfunction. Additional work should be done to settle this hypothesis.

In conclusion, this study enhances our understanding of APE1biological function and its mechanism in mitochondrial regula-tion after oxidative stress. We elucidated the mechanism of APE1in the redox regulation of nuclear genome-encoded mitochondrialgenes, which partially explains the involvement of Cys65 residuesof APE1 in this process. Additionally, we have demonstrated thepresence of APE1 as a redox regulatory factor in the NRF1transcriptional complex, which may provide a possible mechan-ism for transcriptional regulation of NRF1 downstream mitochon-drial-related genes in response to intracellular redox alteration.

Acknowledgments

This work was supported by a grant from the National NaturalScience Foundation of China (30900553) and a Third MilitaryMedical University Young Scientists grant to M.L. Professor GianlucaTell is supported by grants from the MIUR (FIRB_RBRN07BMCT andPRIN2008_CCPKRP_003) and AIRC (IG10269). The authors are grate-ful to Dr. David M. Wilson III, Dr. Zhenzhou Yang, Dr. Jiayin Xie, andDr. Ge Wang for critically reading the manuscript; Dr. Quanzhen Lifor data analysis; Dr. Hualiang Xiao, Dr. Zengpeng Li, Ms. YuxinYang, Mr. Yi Cheng, and Mr. Linli Zeng for their technical support;Ms. Wei Sun for her assistance in laser confocal microscopy; andDr. Shichuan Zhang for his assistance in flow cytometry analysis.

Appendix A. Supplementary material

Supplementary data associated with this article can be foundin the online version at http://dx.doi.org/10.1016/j.freeradbiomed.2012.04.002.

References

[1] Tell, G.; Fantini, D.; Quadrifoglio, F. Understanding different functions ofmammalian AP endonuclease (APE1) as a promising tool for cancer treat-ment. Cell. Mol. Life Sci 67:3589–3608; 2010.

[2] Ludwig, D. L.; MacInnes, M. A.; Takiguchi, Y.; Purtymun, P. E.; Henrie, M.;Flannery, M.; Meneses, J.; Pedersen, R. A.; Chen, D. J. A murine AP-endonu-clease gene-targeted deficiency with post-implantation embryonic progres-sion and ionizing radiation sensitivity. Mutat. Res. 409:17–29; 1998.

[3] Izumi, T.; Brown, D. B.; Naidu, C. V.; Bhakat, K. K.; Macinnes, M. A.; Saito, H.;Chen, D. J.; Mitra, S. Two essential but distinct functions of the mammalianabasic endonuclease. Proc. Natl. Acad. Sci. USA 102:5739–5743; 2005.

[4] Tell, G.; Wilson 3rd D. M.; Lee, C. H. Intrusion of a DNA repair protein in theRNome world: is this the beginning of a new era? Mol. Cell. Biol 30:366–371;2010.

[5] Vascotto, C.; Fantini, D.; Romanello, M.; Cesaratto, L.; Deganuto, M.; Leonardi,A.; Radicella, J. P.; Kelley, M. R.; D’Ambrosio, C.; Scaloni, A.; Quadrifoglio, F.;Tell, G. APE1/Ref-1 interacts with NPM1 within nucleoli and plays a role inthe rRNA quality control process. Mol. Cell. Biol. 29:1834–1854; 2009.

[6] Sengupta, S.; Mantha, A. K.; Mitra, S.; Bhakat, K. K. Human AP endonuclease(APE1/Ref-1) and its acetylation regulate YB-1-p300 recruitment and RNApolymerase II loading in the drug-induced activation of multidrug resistancegene MDR1. Oncogene 30:482–493; 2011.

[7] Yamamori, T.; DeRicco, J.; Naqvi, A.; Hoffman, T. A.; Mattagajasingh, I.;Kasuno, K.; Jung, S. B.; Kim, C. S. Irani, K. SIRT1 deacetylates APE1 andregulates cellular base excision repair. Nucleic Acids Res. 38:832–845; 2010.

[8] Singh, K. K.; Kulawiec, M.; Still, I.; Desouki, M. M.; Geradts, J.; Matsui, S. Inter-genomic cross talk between mitochondria and the nucleus plays an impor-tant role in tumorigenesis. Gene 354:140–146; 2005.

[9] Wilson 3rd D. M.; Bohr, V. A. The mechanics of base excision repair, and itsrelationship to aging and disease. DNA Repair (Amsterdam) 6:544–559; 2007.

[10] Wang, D.; Zhong, Z. Y.; Li, M. X.; Xiang, D. B.; Li, Z. P. Vector-based Ape1 smallinterfering RNA enhances the sensitivity of human osteosarcoma cells toendostatin in vivo. Cancer Sci. 98:1993–2001; 2007.

M. Li et al. / Free Radical Biology and Medicine 53 (2012) 237–248248

[11] Unnikrishnan, A.; Raffoul, J. J.; Patel, H. V.; Prychitko, T. M.; Anyangwe, N.;Meira, L. B.; Friedberg, E. C.; Cabelof, D. C.; Heydari, A. R. Oxidative stressalters base excision repair pathway and increases apoptotic response inapurinic/apyrimidinic endonuclease 1/redox factor-1 haploinsufficient mice.Free Radic. Biol. Med. 46:1488–1499; 2009.

[12] Liu, Y.; Fiskum, G.; Schubert, D. Generation of reactive oxygen species by themitochondrial electron transport chain. J. Neurochem. 80:780–787; 2002.

[13] Yakes, F. M.; Van Houten, B. Mitochondrial DNA damage is more extensiveand persists longer than nuclear DNA damage in human cells followingoxidative stress. Proc. Natl. Acad. Sci. USA 94:514–519; 1997.

[14] LeDoux, S. P.; Druzhyna, N. M.; Hollensworth, S. B.; Harrison, J. F.; Wilson, G.L. Mitochondrial DNA repair: a critical player in the response of cells of theCNS to genotoxic insults. Neuroscience 145:1249–1259; 2007.

[15] Santos, J. H.; Hunakova, L.; Chen, Y.; Bortner, C.; Van Houten, B. Cell sortingexperiments link persistent mitochondrial DNA damage with loss of mito-chondrial membrane potential and apoptotic cell death. J. Biol. Chem.278:1728–1734; 2003.

[16] Ruchko, M.; Gorodnya, O.; LeDoux, S. P.; Alexeyev, M. F.; Al-Mehdi, A. B.;Gillespie, M. N. Mitochondrial DNA damage triggers mitochondrial dysfunc-tion and apoptosis in oxidant-challenged lung endothelial cells. Am. J. Physiol.Lung Cell. Mol. Physiol 288:L530–L535; 2005.

[17] Frossi, B.; Tell, G.; Spessotto, P.; Colombatti, A.; Vitale, G.; Pucillo, C. H2O2

induces translocation of APE/Ref-1 to mitochondria in the Raji B-cell line.J. Cell. Physiol. 193:180–186; 2002.

[18] Li, M. X.; Wang, D.; Zhong, Z. Y.; Xiang, D. B.; Li, Z. P.; Xie, J. Y.; Yang, Z. Z.; Jin,F.; Qing, Y. Targeting truncated APE1 in mitochondria enhances cell survivalafter oxidative stress. Free Radic. Biol. Med. 45:592–601; 2008.

[19] Li, M.; Zhong, Z.; Zhu, J.; Xiang, D.; Dai, N.; Cao, X.; Qing, Y.; Yang, Z.; Xie, J.; Li,Z.; Baugh, L.; Wang, G.; Wang, D. Identification and characterization ofmitochondrial targeting sequence of human apurinic/apyrimidinic endonu-clease 1. J. Biol. Chem. 285:14871–14881; 2010.

[20] Vascotto, C.; Cesaratto, L.; Zeef, L. A.; Deganuto, M.; D’Ambrosio, C.; Scaloni,A.; Romanello, M.; Damante, G.; Taglialatela, G.; Delneri, D.; Kelley, M. R.;Mitra, S.; Quadrifoglio, F.; Tell, G. Genome-wide analysis and proteomicstudies reveal APE1/Ref-1 multifunctional role in mammalian cells. Proteo-mics 9:1058–1074; 2009.

[21] Kelly, D. P.; Scarpulla, R. C. Transcriptional regulatory circuits controllingmitochondrial biogenesis and function. Genes Dev. 18:357–368; 2004.

[22] Scarpulla, R. C. Transcriptional paradigms in mammalian mitochondrialbiogenesis and function. Physiol. Rev. 88:611–638; 2008.

[23] Dhar, S. S.; Ongwijitwat, S.; Wong-Riley, M. T. Nuclear respiratory factor1 regulates all ten nuclear-encoded subunits of cytochrome c oxidase inneurons. J. Biol. Chem. 283:3120–3129; 2008.

[24] Piantadosi, C. A.; Suliman, H. B. Mitochondrial transcription factor A induc-tion by redox activation of nuclear respiratory factor 1. J. Biol. Chem.281:324–333; 2006.

[25] Evans, A. R.; Limp-Foster, M.; Kelley, M. R. Going APE over ref-1. Mutat. Res.

461:83–108; 2000.[26] Breit, J. F.; Ault-Ziel, K.; Al-Mehdi, A. B.; Gillespie, M. N. Nuclear protein-

induced bending and flexing of the hypoxic response element of the ratvascular endothelial growth factor promoter. FASEB J. 22:19–29; 2008.

[28] Vascotto, C.; Bisetto, E.; Li, M.; Zeef, L. A.; D’Ambrosio, C.; Domenis, R.;Comelli, M.; Delneri, D.; Scaloni, A.; Altieri, F.; Mavelli, I.; Quadrifoglio, F.;

Kelley, M. R.; Tell, G. Knock-in reconstitution studies reveal an unexpectedrole of Cys-65 in regulating APE1/Ref-1 subcellular trafficking and function.Mol. Biol. Cell 22:3887–3901; 2011.

[29] Kelley, M. R.; Luo, M.; Reed, A.; Su, D.; Delaplane, S.; Borch, R. F.; Nyland, R. L.2nd; Gross, M. L.; Georgiadis, M. M. Functional analysis of novel analogues of

E3330 that block the redox signaling activity of the multifunctional APendonuclease/redox signaling enzyme APE1/Ref-1. Antioxid. Redox Signaling

14:1387–1401; 2011.[30] Scarpulla, R. C. Nuclear activators and coactivators in mammalian mitochon-

drial biogenesis. Biochim. Biophys. Acta 1576:1–14; 2002.[31] Blesa, J. R.; Prieto-Ruiz, J. A.; Hernandez, J. M.; Hernandez-Yago, J. NRF-2

transcription factor is required for human TOMM20 gene expression. Gene

391:198–208; 2007.[32] Grey, J. Y.; Connor, M. K.; Gordon, J. W.; Yano, M.; Mori, M.; Hood, D. A.

Tom20-mediated mitochondrial protein import in muscle cells duringdifferentiation. Am. J. Physiol. Cell Physiol 279:C1393–C1400; 2000.

[33] Lenka, N.; Vijayasarathy, C.; Mullick, J.; Avadhani, N. G. Structural organiza-tion and transcription regulation of nuclear genes encoding the mammaliancytochrome c oxidase complex. Prog. Nucleic Acid Res. Mol. Biol. 61:309–344;

1998.[34] Ventura-Clapier, R.; Garnier, A.; Veksler, V. Transcriptional control of mito-

chondrial biogenesis: the central role of PGC-1a. Cardiovasc. Res. 79:208–217;2008.

[35] Abate, C.; Patel, L.; Rauscher 3rd F. J.; Curran, T. Redox regulation of fos andjun DNA-binding activity in vitro. Science 249:1157–1161; 1990.

[36] Takamatsu, C.; Umeda, S.; Ohsato, T.; Ohno, T.; Abe, Y.; Fukuoh, A.;Shinagawa, H.; Hamasaki, N.; Kang, D. Regulation of mitochondrial D-loopsby transcription factor A and single-stranded DNA-binding protein. EMBO Rep

3:451–456; 2002.[37] Falkenberg, M.; Gaspari, M.; Rantanen, A.; Trifunovic, A.; Larsson, N. G.;

Gustafsson, C. M. Mitochondrial transcription factors B1 and B2 activatetranscription of human mtDNA. Nat. Genet. 31:289–294; 2002.

[38] Chattopadhyay, R.; Wiederhold, L.; Szczesny, B.; Boldogh, I.; Hazra, T. K.;Izumi, T.; Mitra, S. Identification and characterization of mitochondrial abasic(AP)-endonuclease in mammalian cells. Nucleic Acids Res. 34:2067–2076;

2006.