Mitochondrial Transcription Factor A Induction by Redox Activation of Nuclear Respiratory Factor 1*

The nuclear expression of mitochondrial transcription factor A (Tfam), which is required for mitochondrial DNA (mtDNA) transcription and replication, must be linked to cellular energy needs. Because respiration generates reactive oxygen species as a side-product, we tested the idea that reactive oxygen species regulate Tfam expression through phosphorylation of nuclear respiratory factor (NRF-1) and binding to the Tfam promoter. In mitochondriarich rat hepatoma cells that overexpress NRF-1, basal and oxidant-induced increases were found in Tfam expression and mtDNA content. Specific binding of NRF-1 to Tfam promoter was demonstrated by electrophoretic mobility shift assay and chromatin immunoprecipitation. NRF-1-Tfam binding was augmented under pro-oxidant conditions. NRF-1 gene silencing produced 1:1 knockdown of Tfam expression and decreased mtDNA content. To evaluate oxidation-reduction (redox) regulation of NRF-1 in Tfam expression, blockade of upstream phosphatidylinositol 3-kinase was used to demonstrate loss of oxidant stimulation of NRF-1 phosphorylation and Tfam expression. The oxidant response was also abrogated by specific inhibition of Akt/protein kinase B. Examination of the NRF-1 amino acid sequence revealed an Akt phosphorylation consensus at which site-directed mutagenesis abolished NRF-1 phosphorylation by Akt. Finally, Akt phosphorylation and NRF-1 translocation predictably lacked oxidant regulation in a cancer line having no PTEN tumor suppressor (HCC1937 cells). This study discloses novel redox regulation of NRF-1 phosphorylation and nuclear translocation by phosphatidylinositol 3,4,5-triphosphate kinase/Akt signaling in controlling Tfam induction by an anti-oxidant pro-survival network.

Mitochondrial transcription factor A (Tfam 2 ; mtTFA) is a 25-kDa nuclear-encoded protein with a 42-amino acid pro-sequence that is removed after mitochondrial importation (1). The functional protein has an amino-terminal high mobility group domain, a basic linker region, a second high mobility group domain, and a basic carboxylterminal tail (2,3). Biochemical studies have demonstrated that human Tfam, like other high mobility group-domain proteins, binds, unwinds, and bends DNA without respect to sequence specificity (4). The mammalian Tfam yield after purification is ϳ1 molecule/10 -20 bp of mtDNA (5,6), as in yeast and avian cells (7)(8)(9), and a high molar ratio may signify that Tfam protects mtDNA. Immunochemical studies of Tfam also support the idea that it is involved in nucleoid formation, i.e. packaging mtDNA into protein-DNA aggregates (5,10).
Promoter alignment studies indicate that mouse and rat Tfam promoters contain conserved Sp1 and NRF-2 recognition sites, but a standard NRF-1 consensus binding site has not been reported for either species (3). Genomic footprinting of tumor cells has shown that high levels of Sp1 binding to the Tfam promoter up-regulate the mRNA expression (18). Although these controls on Tfam activity have been identified, much information is lacking about retrograde signaling from mitochondria to nucleus and about factors that regulate the protein response to various cellular stressors.
Communication between mitochondria and the nucleus is essential to energy homeostasis, and close coordination is especially important because the electron transport chain produces reactive oxygen species (ROS) (19), which are damaging to mtDNA and loss of oxidative phosphorylation proteins. ROS leakage increases from damaged mitochondria, and low levels of ROS have been proposed as possible retrograde mediators of signaling from mitochondria to nuclei (20). Specific pathways have not been elucidated, but we found previously in rat liver cells that the lipophilic oxidant, tertiary butyl hydroperoxide (t-BOOH), at low concentrations up-regulates NRF-1 and Tfam (15). This response was interpreted in the context of inflammation and recovery from mitochondrial genomic damage by an apparent redox signal for nuclearencoded mitochondrial gene expression. To investigate whether ROS participate in retrograde signaling from mitochondria to nucleus, we tested the hypothesis in rat cells that peroxide regulates Tfam expression through NRF-1 phosphorylation and Tfam promoter binding to service mtDNA maintenance.

EXPERIMENTAL PROCEDURES
Materials-All reagents were purchased commercially unless indicated otherwise. [␥-32 P]ATP (3000 Ci/mmol) was from Amersham Biosciences, and phospho-Akt (Ser 473 ) and Akt substrate antibodies were from Cell Signaling Technology (Beverly, MA). Rabbit ␣Akt/protein kinase B and NRF-1 antibody and ␣PTEN monoclonal antibody were from Santa Cruz Biotechnology (Santa Cruz, CA). MitoTracker Green FM and MitoTracker Red CM-H 2 XRos were from Invitrogen. LY294002 was from Alexis Corp. (San Diego, CA). pcDNA 3.1/V5-His TOPO TA expression kit was purchased from Invitrogen, and the Chip It kit was from Active Motif (Carlsbad, CA). All other chemicals and reagents, including RPMI and Dulbecco's-modified Eagle's medium with 25 mM Hepes and 4.5 g/liter glucose, were from Sigma.
Cell Culture-H4IIE rat hepatoma cells (ATTCC, Manassas, VA) were grown in stationary culture in Dulbecco's-modified Eagle's medium supplemented with 10% fetal calf serum in 5% CO 2 at 37°C. Control hepatocytes were obtained from primary cultures of healthy liver of SD rats. HCC1937 cells (ATTCC) were grown and maintained in RPMI 1640 medium with 10% fetal bovine serum, 2.5 g/liter glucose, and 1 mM sodium pyruvate. Cell growth and viability were measured using the Cell Titer-Blue Assay (Promega, Madison WI). Where indicated, cells were preincubated with t-BOOH at 0 -135 mM for 1, 3, or 5 h and without or with 50 M LY294002 (phosphatidylinositol 3,4,5triphosphate kinase (PI3K) inhibitor) 30 min before oxidant exposure. For specific inhibition of Akt isoenzymes, cells were preincubated with Akt inhibitor VIII (50 M, Calbiochem) for 30 min before oxidant exposure.
Mitochondrial Integrity-Mitochondrial function was assessed by the conversion of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) dye to formazan in situ by the "succinate-tetrazolium reductase" system of the mitochondrial respiratory chain, which is active only in metabolically intact cells (21). H4IIE cells were plated at a density of 5 ϫ 10 4 cells/well in a 96-well plate and exposed to 10 M t-BOOH at 37°C under 5% CO 2 . Cells were stained with MTT (0.5 mg/ml) for 4 h after incubation with t-BOOH. The resultant formazan crystals were dissolved in 200 l of Me 2 SO, and absorbance was measured at 540 nm on an automatic microplate assay reader (Genius, Tecan Systems, Inc. San Jose, CA). MTT reduction was defined relative to untreated control cells, i.e. 100 ϫ (absorbance of treated/absorbance of control sample).
Rat Tfam mRNA Expression-RNA was extracted with TRIzol reagent (Invitrogen), and 1 g of each sample was reverse-transcribed (20-l volume) using Moloney murine leukemia virus reverse transcriptase (Promega) in a buffer of random hexamer primers, dNTPs, and ribonuclease inhibitor (RNasin, Promega). Gene transcripts were amplified in triplicate by RT-PCR using specific primers Tfam sense, 5Ј-GCTTCCAGGAGGCTAAGGAT-3Ј, and Tfam antisense, 5Ј-CCC-AATCCCAATGACAACTC-3Ј. 18 S rRNA was used to control for variation in efficiency of RNA extraction, reverse transcription, and PCR for nuclear RNA expression.
Immunoprecipitation-WT or mutant NRF-1 recombinant vectors were transfected transiently into H4IIE cells. Non-transfected and transfected cells were treated with t-BOOH (10 M) without or with LY294002 (50 M) 30 min before oxidant exposure. After incubation for 3 h, the cells were washed with ice-cold phosphate-buffered saline and lysed for 20 min in lysis buffer (20 mM Tris, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM Na 3 VO 4 , 1 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride). Cell lysates or nuclear extracts (100 g) were incubated with phospho-(Ser/Thr) Akt substrate antibody or with ␣NRF-1 antibody overnight at 4°C and protein A-agarose beads (Santa Cruz Biotechnology) for 2 h. Beads were washed twice in lysis buffer and twice in phosphate-buffered saline, and proteins were eluted with SDS sample buffer for Western analysis using anti-phospho-Ser/Thr monoclonal antibody (BD Biosciences Pharmingen) or NRF-1 primary antibody (1:2000).
Cloning of Rat Hepatoma Tfam-The 5Ј region encompassing the promoter and 5Ј region of Tfam was cloned by PCR with Pfu polymerase (Stratagene). Primers were designed from the published Tfam sequence (GenBank TM accession number AF264733), and rat genomic DNA was used as a template. The amplified DNA was cloned into TOPO vector (Invitrogen), and both strands were sequenced using a PerkinElmer Life Sciences Dye Terminator Cycle Sequencing system with AmpliTaq DNA Polymerase combined with ABI 3730 and 3100 PRISM DNAsequencing instruments.
Real-time DNA PCR-DNA primers were designed to detect COII and ␤-actin at a maximum amplicon length of 150 bp: ␤-actin forward, 5Ј-TGTTCCCTTCCACAGGGTGT-3Ј, and reverse, 5Ј-TCCCAGTT-GGTAACAATGCCA-3Ј; COII forward, 5Ј-TGAGCCATCCCTTCA-CTAGG-3Ј, and reverse, 5Ј-TGAGCCGCAAATTTCAGAG-3Ј. The PCR reaction mixture contained 1ϫ platinum SYBR green qPCR SuperMix UDG (Invitrogen), 500 nM each primer, and ϳ10 ng of total genomic DNA or mtDNA. Real-time PCR conditions were 2 min at 50°C and 10 min at 95°C followed by 40 cycles of 15 s at 95°C and 60 s at 60°C. Fluorescence intensities during PCR were recorded and analyzed in an ABI Prism 7000 sequence detector system (Applied Biosystems 7000 SDS software). The threshold cycle (C T ) is the cycle at which the first significant increase in the fluorescent signal is detected. C T values in the linear exponential phase were used to measure the original DNA template copy numbers from a standard curve generated from five 10-fold dilutions of either pure mtDNA (COII) or pure nuclear DNA (␤-actin). Relative values for COII and ␤-actin within samples were used to obtain a ratio of mtDNA to nuclear DNA in that sample.
NRF-1 Gene Silencing-Silencing experiments were performed using small interfering NRF-1 (siNRF-1) duplexes to target sequences in the open reading frame of NRF-1 mRNA. Multiple nucleotide sense and antisense siRNAs were synthesized and obtained in annealed form from Ambion. The siRNA target sequences were submitted to BLAST searches against other rat genome sequences to ensure specificity. After preliminary studies, one pair of siRNA sequences was selected and transfected at a concentration of 80 nM into H4IIE cells using Oligofectamine TM (Invitrogen) sense, 5Ј-GGAGGUUAAUUCAGAGCUG-3Ј, and antisense, 5Ј-CAGCUCUGAAUUAACCUCC-3Ј. A scrambled negative control siRNA from Ambion was also used. The effect of siRNA on NRF-1 and Tfam mRNA expression was established by both conventional and quantitative real time RT-PCR. For real-time PCR expression analysis, cells transfected with siNRF-1 or scrambled siRNA were washed in phosphate-buffered saline, and total RNA was extracted using Trizol reagent (Invitrogen). RNA samples were then DNasetreated with a DNA-free kit (Ambion). Treated RNA (1 g) was reversetranscribed in 20-l reaction mixtures with random hexamer primers and the TaqMan reverse transcription reagents kit (Applied Biosystems, Foster City, CA) according to the manufacturer's protocol. Two microliters of undiluted cDNA were used per reaction; a standard curve for NRF-1 and glyceraldehyde-3-phosphate dehydrogenase was generated from serial dilutions of positively expressed cDNA. The probe and primer sets were "pre-designed assay on demand" probes (Applied Biosystems) tested and standardized for reproducible expression analysis.
Primer and cDNA were added to TaqMan universal master mix (Applied Biosystems) containing all the reagents for PCR. Relative quantification of the PCR products was carried out using the ABI Prism 7000 (Applied Biosystems) and a method that compares the amount of target gene amplification normalized to an endogenous reference (glyceraldehyde-3-phosphate dehydrogenase). The formula was 2 Ϫ⌬⌬CT , representing the n-fold differential gene expression in a treated sample compared with the control, where C T is the mean threshold cycle (cycle at which the product is detected initially). ⌬C t was the difference in the C T values for the target gene and the reference gene, glyceraldehyde-3phosphate dehydrogenase (in each sample assayed), and ⌬⌬C t represents the difference between C T of control and sample.
ChIP Assay-Cells were transfected with NRF-1, WT, or mutant vectors and 48 h later were treated with 50 M LY294002 30 min before 10 M t-BOOH. After 3 h, 37% formaldehyde was placed directly into the medium of 4.5 ϫ 10 7 cells to achieve a 1% final concentration. After 30 min at 37°C, formaldehyde was quenched with 0.125 M glycine. Cells were washed with phosphate-buffered saline, harvested, and processed for immunoprecipitation using a kit (ChIP It assay kit, Active Motif) and NRF-1 antibody or V5 antibody following the manufacturer's protocol. After ethanol precipitation, DNA was resuspended in 200 l/10 7 cells, and 2-5 l were used as the template for each PCR. Input samples represent 1% of total DNA and were diluted 1:5, and immunoprecipitated fractions were diluted 1:2. PCR was carried out on 1 l of each sample using sense primer, 5Ј-GGC AGTTTGCTGCTGGGT-3Ј, and antisense primer, 5Ј-GGCACTGTGGGAGGCCCA-3Ј, to amplify a 331-bp segment of the rat Tfam gene at Ϫ359 to Ϫ28 from the transcription start site. PCR products were analyzed on 2% ethidium bromide-stained agarose gels.
Immunocytochemistry and Laser Scanning Confocal Microscopy-Rat H4IIE cells or human HCC1937 cells were grown in one-or twowell LabTek chamber slides and treated with t-BOOH (10 M). The Tfam mRNA expression in normal rat hepatocytes and rat hepatoma cells (H4IIE) by RT-PCR. PCR products electrophoresed on 1.5% agarose and stained with Gel Star. As an internal control, RT-PCR for rat 18 S was performed using the same amount of cDNA. B, immunoblots of control rat hepatocytes and hepatoma cells for NRF-1 and NRF-2. Cell protein extracts (20 g) were resolved on 12% SDS-polyacrylamide gels and transferred to polyvinylidene difluoride membranes, and blots were probed with ␣-NRF-1 or ␣-NRF-2. The same membranes washed and blotted with anti-tubulin show equal loading of protein. The protein signal from the ECL detection was quantified by densitometry, and the ratio of NRF-1/2/tubulin was calculated. Conclusions derive from three independent experiments (graph not shown; see "Results"). cells were then washed with Hanks' solution and incubated in normal media with MitoTracker green (200 nM for H4IIE cells) or Mito-Tracker red (100 nM for HCC1937 cells), incubated for 20 min, and rinsed with Hanks' balanced salt solution. H4IIE cells were fixed with cold 70:30 acetone/ethanol (v/v) for 10 min and blocked for 1 h with 10% fetal bovine serum in Hanks' solution at room temperature. The cells were incubated for 1 h with rabbit ant-NRF-1 primary antibodies followed by FluoroLink Cy3-labeled goat anti-rabbit IgG for 1 h. After extensive washing to remove unbound antibodies, the cells were mounted using the Slow Fade kit (Invitrogen). HCC1937 cells were mounted with Slow Fade immediately after fixation. Specific immunostaining was assessed by using one or both secondary antibodies without primary antibody. Multiple dilutions of each primary and secondary antibody were tested and optimized to minimize nonspecific adsorption of fluorescent antibodies, ensure separation of fluorescent signals, and optimize fluorophore concentration.
Laser scanning confocal microscopy was performed on an LSM 410 microscope (Carl Zeiss MicroImaging, Inc.). Fluorescence images were collected using appropriate band-pass filters at excitation wavelengths of 488 and 568 and emission wavelengths of 520 and 590 nm.
Statistics-Group values were expressed as the means Ϯ S.D. Significance was tested by Student's unpaired t test, and a p value of Ͻ0.05 was accepted as significant. The n values provided refer to independent samples.

RESULTS
Tfam and mtDNA Expression in H4IIE Rat Hepatoma Cells-Tfam mRNA levels were compared by RT-PCR in H411E rat hepatoma cells and normal rat hepatocytes. Tfam transcript analysis indicated a 5-fold higher expression level in hepatoma relative to normal hepatocytes (Fig.  1A). On this basis it was anticipated that transcription factors involved in Tfam gene expression would be up-regulated in rat hepatoma cells. Comparison of steady-state expression levels of NRF-1 and NRF-2 protein by Western blot analysis in normal rat hepatocytes and hepatoma cells (Fig. 1B) indicated that NRF-1 and NRF-2 protein levels were 4and 3-fold higher, respectively, in rat hepatoma cells compared with normal hepatocytes (p Ͻ 0.05 for n ϭ 4, graph not shown). For unknown reasons, these differences were only associated with 25% more mtDNA per H4IIE cell by real time PCR compared with normal rat hepatocytes (not shown).
Because of the uncertainty that NRF-1 regulated Tfam gene expression in rat hepatoma cells, the 5Ј upstream region of Tfam was cloned first. The nucleotide sequences for the 5Ј-upstream region and exon one of rat hepatoma Tfam are shown in Fig. 2A. The cloned region showed 99.4% sequence homology with the normal upstream region of rat Tfam (GenBank TM accession number AF264733). The transcription start site is numbered ϩ1 according to the rat cDNA sequence (22). A computerassisted sequence analysis of the 5Ј promoter region identified a putative NRF-1 binding site in the Tfam promoter region (Fig. 2A). This motif is a near-perfect match for the NRF-1 consensus and contains all the invariant nucleotides in functional NRF-1 recognition sites (23)(24)(25). The binding site showed high homology with the human Tfam sequence (Fig. 2B) but overlapped with the Sp1 site, indicating the complex organization of the rat promoter.
Effect of Exogenous Oxidant on Tfam Expression-To determine whether oxidants mediated Tfam gene expression, hepatoma cells were exposed to low concentrations of t-BOOH for 1 to 5 h. These exposures led to ϳ10-fold increases in cellular Tfam gene expression (Fig. 3A) and were associated with stimulation of cell growth by 24h (Fig. 3B) and comparable increases in MTT reduction and mtDNA content by 48 h (Fig. 3C).
Functional Characterization of the 5Ј-Flanking Region of Rat Tfam-To determine whether the putative NRF-1 consensus in rat Tfam could act as a cis element and bind NRF-1, EMSAs were performed in rat hepatoma cell nuclear extracts with oligonucleotides containing the NRF-1 recognition site of human Tfam. NRF-1 produced a slow-migrating band that was lost after the addition of a 100-fold excess of unlabeled oligonucleotide or PCR products containing the full rat Tfam promoter (Ϫ462/ϩ79) or the GC-rich oligonucleotides Ϫ205 to Ϫ185 (Fig. 4A). Binding was checked and confirmed for recombinant NRF-1 protein alone. The results indicated that the cloned rat Tfam promoter  contains a cis-acting element that binds NRF-1 specifically in the region between Ϫ205 and Ϫ185 (Fig. 4A).
To assess NRF-1 activation in t-BOOH-induced Tfam expression and to demonstrate functionality of the Ϫ205 to Ϫ185 5Ј-region, an EMSA was performed on nuclear extract of hepatoma cells treated with or without 3h t-BOOH and using a 32 P-labeled oligonucleotide containing Ϫ205 to Ϫ185 5Ј-region (Fig. 4B). A large DNA-protein complex was found after oxidant treatment, and its specificity for NRF-1 was demonstrated by supershifting the complex with ␣NRF-1 antibody. Extracts pre-absorbed with ␣AP1 or ␣Sp1 showed no supershift. The addition of 100-fold excess unlabeled rat oligonucleotide to the nuclear extracts completely blocked DNA-protein complex formation, but neither excess unlabeled AP-1 nor Sp1 oligonucleotide inhibited DNA binding. Thus, specific NRF-1 binding to the Ϫ205 to Ϫ185 region is demonstrated in the hepatoma Tfam promoter.
To check the effect of t-BOOH on NRF-1 activation in intact cells, NRF-1 nuclear translocation was examined by confocal microscopy (Fig. 5). Mitotracker Green FM was used to define mitochondrial distribution before (Fig. 5, A-C) and 1-2 h after 3 h of exposure of the cells to growth-stimulating levels of t-BOOH (10 M) (Fig. 5, D-F). The oxidant caused high intensity punctate green fluorescence in the cytoplasm in a greater proportion of treated than untreated cells. This staining pattern was accompanied by prominent translocation of NRF-1 protein to the nucleus relative to untreated H4IIE cells. To examine whether PI3K and Akt were involved, the cells were preincubated for 30 min with specific Akt or a PI3K inhibitors. Akt inhibitor VIII (Fig. 5, G-I) or PI3K inhibitor (LY294002, Fig. 5, J-L) ablated the nuclear translocation of NRF-1 by the oxidant, thus indicating interdependence of NRF-1/Akt.

Oxidant-induced Akt/Protein Kinase B Activation and NRF-1 Phosphorylation in H4IIE Cells-The interference of PI3K and Akt blockade
with NRF-1 nuclear translocation and the reported phosphorylation of NRF-1 suggested that oxidant activation of Akt might cause NRF-1 phosphorylation. Therefore, the effect of the oxidant on NRF-1 phosphorylation and Tfam mRNA expression was checked before and after PI3K or Akt inhibition (Fig. 6). Akt phosphorylation after t-BOOH exposure (10 M) was readily demonstrated in H4IIE cells (Fig. 6A). Then NRF-1 phosphorylation in response to the oxidant and its effect on Tfam expression were investigated. Fig. 6B shows that t-BOOH stimulates serine/threonine phosphorylation of NRF-1, which was inhibited by Akt inhibition with LY294002 and with Akt-VIII, placing NRF-1 phosphorylation downstream of the oxidant activation step. These inhibitors also blocked Tfam mRNA expression (Fig. 6B). The rat NRF-1 amino acid sequence was examined for potential phosphorylation sites because one serine has been reported to enhance NRF-1/DNA binding activity (26). A putative canonical Akt phosphorylation motif (RXRXXT) in the NRF-1 sequence was identified at Thr-109. To determine the specificity of Thr-109 phosphorylation in NRF-1 activation, a mutant NRF-1 was constructed by substituting Thr-109 with an Ala residue (Fig. 6C). This mutant and a WT construct were then expressed comparably in H4IIE cells (Fig. 6D). To check phosphorylation of NRF-1 by Akt, H4IIE cells were co-transfected with His-V5-tagged NRF-1 vector (WT or mutant), and the recovered NRF-1 proteins were subjected to in vitro phosphorylation by active recombinant Akt. Fig. 6E shows that Akt phosphorylates WT-NRF-1 (Fig. 6E, first and second  lanes). When Thr-109 was mutated to Ala, phosphorylation was diminished (Fig. 6E, third and fourth lanes), indicating a site for Akt-dependent phosphorylation of NRF-1.
To determine whether NRF-1 phosphorylation increases DNA binding activity, EMSAs were done on the nuclear extracts of cells transfected with WT or mutant NRF-1 vectors. The assays showed that phosphorylation markedly stimulated NRF-1 DNA binding activity (Fig. 7B). WT-NRF-1 DNA binding was diminished by preincubating the cells with low concentrations of PI3K inhibitor, LY294002 (compare Fig. 7B,  firstandthirdlanes),demonstratingNRF-1phosphorylationafteroxidantdependent Akt activation. Cell transfection with mut-NRF-1 prevented phosphorylation-reduced DNA-protein complex formation (ϳ40% of WT), but residual DNA binding implicated other contributions to activation of the complex. To confirm NRF-1 binding to the Tfam promoter and that Akt phosphorylation of Thr-109 increased NRF-1 transcriptional activity, HisV5-tagged WT-NRF-1 and T109A mutant were transfected into cells, and ChIP assays were performed. The assay shown in Fig. 7C demonstrates recruitment of HisV5-tagged WT NRF-1 to the Tfam promoter after stimulation by t-BOOH, abolition of effect by LY294002, and abrogation by Thr-109 mutation (Fig. 7C). Immunoprecipitation with ␣NRF-1 revealed that cells treated with t-BOOH increase endogenous NRF-1 binding to Tfam promoter, which is attenuated by LY294002 (Fig. 7C). In the absence of antibody, promoter fragments in the immunoprecipitate were not detected. Thus, Akt-mediated phosphorylation of Thr-109 enhances NRF-1 transcriptional activity by increasing its recruitment to the Tfam promoter.
To confirm that oxidant stimulation of NRF-1 led to Tfam expression in H4IIE cells, NRF-1-specific siRNA transfection was used to knock down Tfam mRNA levels. Transfection of NRF-1 siRNA in H4IIE cells reduced NRF-1 and Tfam mRNA expression equally (ϳ75%) as determined by two techniques, conventional RT-PCR and real-time PCR. Both methods showed that Tfam mRNA levels were much lower in cells transfected with NRF-1-siRNA than with scrambled siRNA (Fig. 8, A  and B). The fall in Tfam mRNA levels also resulted in a decline in total mtDNA content after 48 h (Fig. 8C). Therefore, loss of NRF-1 expression had a major negative impact on Tfam expression and mtDNA content in these rat cells.
Phospho-Akt and NRF-1 Expression in PTEN-deficient Cells-Because oxidant signaling through PI3K/Akt involves inhibition of regulatory phosphatase activity, the oxidant effect on NRF-1 was checked in a second cell line, a human breast cancer cell (HCC1937) that genetically lacks PTEN. PTEN, a lipid phosphatase and tumor suppressor, regulates PI3K and physiologically inhibits Akt transduction. HCC1937 cells showed constitutive Akt hyperphosphorylation that did not stimulate with t-BOOH but did abrogate after PI3K inhibition with LY294002 (Fig. 9A). A fall in phospho-Akt content in HCC1937 cells of ϳone-third after t-BOOH exposure may be nonspecific, but there was still an absence of the oxidant stimulation of NRF-1 nuclear localization (Fig. 9B, left).
Mitochondrial morphology evaluated in HCC1937 cells by confocal microscopy using cell-permeate MitoTracker Red showed red fluorescence localized to highly fluorescent, elongate, and branching mitochondria distributed throughout the cytoplasm. Exposure to low dose were incubated with t-BOOH (10 M) for 3 h, and the cells were fractionated for Western analysis of total and nuclear protein, or total protein was immunoprecipitated (IP) with anti-NRF-1 antibody to probe for phospho-Ser/Thr by Western analysis. In 3 experiments, the low dose oxidant increased phospho-/total NRF-1 ϳ4-fold and nuclear/total NRF-1 content ϳ1.8-fold. NRF-1 phosphorylation depended strongly on Akt activation since the oxidant effect was completely blocked by inhibition of either Akt (Akt-VIII) or PI3K (LY924002). Simultaneous stimulation of Akt phosphorylation and Tfam mRNA expression by the oxidant is shown for comparison (lower panels). C, Akt motif in wild type NRF-1 protein and the substitution of Thr residue with Ala in the mutant protein. M is methionine at the start site. D, Western blot of wild type and mutant NRF-1 vectors transfected into H4IIE cells shows similar overexpression compared with ␤-actin. E, active Akt phosphorylates WT-NRF-1 but not mutant NRF-1 protein in vitro. WT and mutant NRF-1 protein for the Akt activity assay was isolated using ␣V5 antibody, and 10% of the extract was run on a gel and stained with Coomassie Blue to ensure equal protein input. Data represent one of three experiments.
t-BOOH for 3-5 h diminished the focal fluorescence, and the remaining visible mitochondria appeared round and swollen (not shown). Of note, t-BOOH levels (10 -15 M) that promoted growth of H4IIE cells actually inhibited growth of HCC1937 cells (Fig. 9B, right). Thus, oxidant growth stimulation was lost with Akt signal contingency.

DISCUSSION
Tfam regulation is highly important to metabolically active cells, and the role of NRF-1 in transcriptional control of Tfam expression, especially with respect to mitochondrial biogenesis, is not fully understood. Previous studies inferred NRF-1 binding to the rat Tfam promoter region without documentation of a 5Ј NRF-1 binding consensus (18,27). Therefore, the Tfam 5Ј region was cloned and sequenced in rat hepatoma cells, and a region was identified at Ϫ205 to Ϫ185 of high homology with three 8-mer motif clusters (GCCTGCGC, CGCCT-GCG, and GCGCCTGC) for NRF-1 recognition by human Tfam (22). This sequence differs from an optimal NRF-1 consensus (CGCVT-GCG) by 1 base, which is not strictly required for NRF-1 protein-DNA binding (23,26). A "perfect" consensus would be optimal, but the evidence overall does merit assignment of an NRF-1 consensus to the rat Tfam promoter since several other genes known to be regulated by NRF-1 do not show perfect stringency by binding sequence analysis. A role for NRF-1 in rat Tfam transcriptional activity is reinforced by the  association of NRF-1 with the Tfam promoter and mRNA expression levels in cells and tissue (15) and by the present siRNA data.
NRF-1 may be important in addition to regulating a number of respiratory genes (28,29), in other signal-transduction pathways (30), and in cell synthetic and growth functions by up-regulating gene transcription of eukaryotic initiation factors 2␣ and 2␤ (31). Moreover, NRF-1 sites are found in genes involved in cell cycle regulation (cdc2 and guanine-nucleotide exchange factor RCC1) or regulated by cell growth (DNA polymerase-␣, ornithine decarboxylase, and GADD153, growtharrest and DNA-damage-inducible protein 153) (32). Despite the inference that NRF-1 regulates metabolic processes including biogenesis, except for DNA binding enhancement by phosphorylation, little has been known about regulation of this activity (27,33).
This work establishes for the first time that low oxidant levels mediate NRF-1 phosphorylation through Akt activation, which promotes its nuclear translocation and Tfam binding. This sequence is typical of reversible catalytic inactivation of regulatory phosphatases by activesite cysteine disulfide formation (34,35). Oxidants that activate receptor PI3K inactivate reciprocating phosphatases (e.g. PTEN and PTP1B), which limit the Akt activity in cells. PI3K generates phosphorylated phosphoinositides at the plasma membrane to which Akt binds, is phosphorylated, and activates downstream pro-survival signaling pathways. New evidence is offered here in rat cells for a nuclear-mitochondrial interaction that is involved in this highly conserved survival pathway.
PI3K is involved in mitogen activation of Akt in other cell types (36), and PI3K-independent activation mechanisms have also been found. Akt/protein kinase B activation by ␤ 3 -adrenergic receptor (37), cyclic AMP (38), and cell stressors such as heat shock and hyperosmolarity (39) are insensitive to PI3K blockade. But here, a specific PI3K inhibitor (LY294002) and a specific Akt inhibitor both prevent Akt phosphorylation and block NRF-1 nuclear translocation, placing PI3K and Akt in a scheme for redox activation of NRF-1.
A consensus phosphorylation motif for Akt found in NRF-1 at amino acids 103-109 was used to demonstrate NRF-1 phosphorylation by active Akt. Disruption of NRF-1 phosphorylation by a single substitution of Thr-109 to Ala identified this as a relevant Akt phosphorylation site and supports our other evidence for NRF-1 phosphorylation by PI3K/Akt in intact cells. This included abrogation of NRF-1 phosphorylation, nuclear translocation, and DNA binding by pharmacological Akt or PI3K blockade. Some residual binding activity was detected after mut-NRF-1 cell transfection, which may reflect residual endogenous signal; nonetheless, Akt phosphorylates NRF-1 at Thr-109, which increases NRF-1 translocation and Tfam DNA binding activity. NRF-1 activity studies demonstrated that the increase in Tfam promoter binding was inhibited in situ by PI3K/Akt inhibition or T109A mutation. Thus, regulation of NRF-1 phosphorylation by Akt supports Tfam and mtDNA content; this pathway may also link mitochondrial biogenesis to cell survival. The mutagenesis experiments also establish a role for NRF-1 in Tfam promoter function. The latter is substantiated by ChIP assay, which clearly shows Tfam promoter occupancy by NRF-1 in vivo.
Other oxidant-sensitive pathways that might stimulate NRF-1 translocation have not been investigated, such as those involving focal adhesion kinase and Ras protein (40). Focal adhesion kinase mediates signal transduction by integrins and operates upstream of PI3K-Akt in H 2 O 2induced apoptosis (41). ROS generation stimulates growth in some cells (42), shown here after oxidant activation of Akt/protein kinase B in a PI3K-dependent manner. After phosphorylation, NRF-1 undergoes nuclear translocation and promotes the Tfam expression necessary for mtDNA replication. Thus, NRF-1 phosphorylation by PI3K/Akt is identified with redox signaling that supports mtDNA content. This finding has novel implications for extracellular oxidant-based signaling of biogenesis, but whether it simulates an effect of endogenous mitochondrial ROS production remains to be determined. A recent study suggests that MnSOD induction, by the superoxide catalysis to peroxide, may provide such a signal (43).
To further evaluate the effect of the PI3K/Akt system on NRF-1 activation, we used HCC1937 cells, a human breast cancer line deficient in PTEN tumor suppressor protein. PTEN, a dual-specificity phosphatase, primarily dephosphorylates inositol at the D3 position, thereby antagonizing PI3K and phosphatidylinositol 3-phosphate signaling (41,44,45). Unstimulated HCC1937 cells show Akt hyperphosphorylation and express NRF-1 constitutively, but neither was promulgated by the oxi- . HCC1937 cells exposed to t-BOOH showed stable (or decreased) total and decreased nuclear NRF-1 protein expression (tubulin control). Data represent one of duplicate studies. Right panel, effect of t-BOOH on cell growth compared for H4IIE and HCC1937 cells. Control cells (untreated) or those treated with for 5 h 15 M t-BOOH were then incubated for 4 h with CellTiter-Blue (Promega), and percent differences were determined for the two cell lines. Growth responses were opposite and statistically different in the two lines (*, p Ͻ 0.05).
dant. Akt phosphorylation was inhibitable by PI3K blockade and HCC1937 cells, unlike the hepatoma line, were exquisitely sensitive to damage by t-BOOH, implying that even mild oxidative stress compromises an adaptation contingent on this pro-survival pathway.
A final point is that HCC1937 cells, besides being PTEN-deficient, contain mutant BRCA1 (46). BRCA1 regulates cell cycle progression, apoptosis, and DNA repair, and loss of BRCA1 expression is linked to cell proliferation and transformation (47). BRCA1 overexpression confers resistance and BRCA1 deficiency sensitivity to oxidative stress, and BRCA1 may stimulate antioxidant response element-driven transcriptional activity (48). This exemplifies a class effect that could also contribute to greater susceptibility of HCC1937 cells to t-BOOH than H4IIE cells, which do express BRCA1.
In conclusion, this study provides novel evidence that stimulation of PI3K by exogenous oxidant activates Akt and promotes NRF-1 phosphorylation and nuclear translocation. This is in accordance with the known sensitivity to oxidant inactivation of cysteine-dependent phosphatases, e.g. PTEN, that are reciprocal regulators of PI3K. Phosphorylation increases the capacity of NRF-1 to stimulate transcription of Tfam, a downstream nuclear-encoded gene for a mitochondrial protein required for mtDNA transcription and replication. This NRF-1 activity leads to an increase in mtDNA content that may have implications for mitochondrial biogenesis. In contrast, low oxidant levels inhibit the growth of a tumor cell type that has already recruited Akt activity for proliferation. And by implication, oxidant damage to mitochondria without NRF-1 up-regulation will disrupt energy metabolism by decoupling demand from respiratory gene activation.