PTEN/Akt Signaling Controls Mitochondrial Respiratory Capacity through 4E-BP1

Akt, a serine/threonine kinase has been shown to stimulate glycolysis in cancer cells but its role in mitochondrial respiration is unknown. Using PTEN-knockout mouse embryonic fibroblasts (MEFPTEN−/−) with hyper-activated Akt as a cell model, we observed a higher respiratory capacity in MEFPTEN−/− compared to the wildtype (MEFWT). The respiratory phenotype observed in MEFPTEN−/− was reproduced in MEFWT by gene silencing of PTEN which substantiated its role in regulating mitochondrial function. The increased activities of the respiratory complexes (RCs) I, III and IV were retained in the same relative proportions as those present in MEFWT, alluding to a possible co-ordinated regulation by PTEN/Akt. Using LY294002 (a PI3K inhibitor) and Akt inhibitor IV, we showed that the regulation of enzyme activities and protein expressions of the RCs was dependent on PI3K/Akt. There was insignificant difference in the protein expressions of mitochondrial transcription factor: peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α) and its downstream targets, the nuclear respiratory factor 1 (NRF-1) and mitochondrial transcription factor A (mtTFA) between MEFPTEN−/− and MEFWT. Similarly, mRNA levels of the same subunits of the RCs detected in Western blots were not significantly different between MEFPTEN−/− and MEFWT suggesting that the regulation by Akt on mitochondrial function was probably not via gene transcription. On the other hand, a decrease of total 4E-BP1 with a higher expression of its phosphorylated form relative to total 4E-BP1 was found in MEFPTEN−/−, which inferred that the regulation of mitochondrial respiratory activities by Akt was in part through this protein translation pathway. Notably, gene silencing of 4E-BP1 up-regulated the protein expressions of all RCs and the action of 4E-BP1 appeared to be specific to these mitochondrial proteins. In conclusion, PTEN inactivation bestowed a bioenergetic advantage to the cells by up-regulating mitochondrial respiratory capacity through the 4E-BP1-mediated protein translation pathway.


Introduction
PTEN (phosphatase and tensin homology deleted on chromosome 10) is one of the tumor suppressors frequently lost in cancers [1,2]. It has both protein and lipid phosphatase activities; the latter is associated with tumor suppression [3,4]. By dephosphorylating phosphatidylinositol-3,4,5-trisphosphate (PIP 3 ), PTEN prevents the membrane recruitment and activation of Akt which promotes cell survival, proliferation, growth and glycolysis [5,6]. The ability of Akt in stimulating cell proliferation is best illustrated in Cowden and other PTEN hamartoma tumor syndromes [7]. The benign nature of these tumors could be due to the fact that neither PTEN nor Akt alone is sufficient for the cell to become cancerous [8,9]. The signaling downstream of Akt is integrated by mTOR (mammalian target of rapamycin) and transduced to ribosomal protein kinase S6 kinase (S6K) and the eukaryotic translation initiation factor 4E (eIF4E)-binding proteins (4E-BP1, 2 and 3) to regulate protein translation [10]. 4E-BP1 is a repressor of 59capdependent mRNA translation and it is inactivated upon phosphorylation by mTORC1. The phosphorylated 4E-BP1 is then released from the eukaryotic initiation factor 4E (eIF4E) which subsequently binds eIF4G to begin protein translation [11]. 4E-BP1 is a key effector of oncogenic activation of the Akt and ERK signaling pathways that integrate their function in tumors [12]. It correlates with the clinical findings that expression of high level of phosphorylated 4E-BP1 is associated with poor prognosis in several types of tumor, independent of the alteration of upstream oncogenic signaling [13].
While the activation of Akt by defective respiration in cancer cells has been shown to stimulate aerobic glycolysis [14,15], the action of Akt on the mitochondrial respiration in cell transformation remains largely unknown. In this study, MEF WT and MEF PTEN2/2 were used as a model to examine if the PTEN/ Akt pathway has an effect on oxidative metabolism in normal and transformed cells. We compared a number of mitochondrial parameters and found that the MEF PTEN2/2 with hyper-activated Akt exhibited higher enzyme activities and protein expressions of all respiratory complexes. Pre-treatment of MEF PTEN2/2 with LY294002 (LY), an inhibitor of PI3K and mTOR [16] and Akt inhibitor IV decreased (a) the phosphorylation of Akt at both Ser 473 and Thr 308 , (b) the phosphorylation of 4E-BP1 at Thr 37/46 and (c) the expressions of RC I, III and IV. We also measured the mRNAs of the same subunits of the RCs detected in Western blots and the protein expressions of the peroxisome proliferatoractivated receptor gamma coactivator 1-alpha (PGC-1a), the nuclear respiratory factor 1 (NRF-1) and mitochondrial transcription factor (mtFTA) and showed that the regulation of mitochondrial function by Akt was probably not dependent on transcription. In conclusion, the higher proliferation in MEF PTEN2/2 was accompanied by increased mitochondrial respiratory capacity. The protein expressions of the RCs in MEF PTEN2/2 are attenuated by LY294002 and Akt inhibitor IV but are increased in MEF WT by siRNA targeting PTEN or 4E-BP1, suggesting a regulation of mitochondrial respiratory activities by the PTEN/ Akt/mTOR pathway through 4E-BP1-mediated protein translation.

Cell culture
The MEF WT and MEF PTEN2/2 cell lines [6] were kindly provided by Dr. Tak W. Mak of University of Toronto. They were cultured in DMEM (Sigma, D1152) with 100 U/ml penicillin, 100 mg/ml streptomycin, 2 mM glutamine and 10% fetal bovine serum in 5% CO 2 at 37uC.

Respiratory flux
The measurement of respiratory flux in intact MEF cells in a high-resolution respiratory system (Oroboros, Oxygraph-2k) was carried out according to the protocol described [17]. Cellular routine (R) state was established with the presence of cells in 2 ml of culture medium in the chamber at 37uC with constant stirring for 15-30 min. In the absence of exogenous substrates, the R (routine) respiration is dependent on endogenous substrates and substrates present in the culture medium. Following the stabilization of R, oxygen consumption is measured in the presence of 2 mg/ml oligomycin to obtain the Leak (L) state. In this nonphysiological state, L respiration reflects intrinsic uncoupling of the mitochondria. Subsequently, a maximal electron transport system capacity (E) of the mitochondria is achieved by titrating sequentially with repeated doses of 1 ml from 1 mM FCCP until a maximal response is achieved. The net Routine/E is then calculated by applying the formula: Net routine/E = (R2L)/E. This parameter therefore represents the net oxygen consumption coupled to ATP biosynthesis, taking into account the leak component.

Isolation of mitochondria
The procedure described in [18] was used with slight modifications. Cells were trypsinized, centrifuged at 1000 g for 5 min and the pellet was re-suspended in 1 ml of medium containing 250 mM sucrose, 20 mM Hepes, 10 mM KCl, 1.5 mM MgCl 2 , 1 mM EDTA and 1 mM EGTA, pH 7.4. This was followed by homogenization with 40 up-and-down strokes using a hand-held glass pestle. The mixture was then centrifuged at 800 g for 10 min and the supernatant obtained was subjected to another step of centrifugation at 14,000 g for 15 min. The final pellet was re-suspended in the same medium and was used immediately for ATP biosynthesis and measurement of MMP or kept at 280uC as freeze-thawed mitochondrial extracts for other enzyme assays.

ATP biosynthesis
ATP production in 5 min was measured from the oxidation of the following substrates 6 inhibitor: 10 mM succinate+5 mM rotenone, 5 mM each malate/glutamate, or 5 mM ascorbate and 1 mM tetramethylphenylenediamine (TMPD)+2 mg/ml antimycin A. 125 mM ADP and 30 mg mitochondrial protein were used in the reaction which was stopped by placing the tubes in a heating block at 100uC for 3 min, followed by centrifugation at 14,000 g for 15 min. ATP formed was measured by the luciferin/luciferase reaction and the luminescence produced was read in a luminometer (Victor 3, Perkin-Elmer) as described previously [19,20].

Measurement of mitochondrial membrane potential (MMP)
The membrane potential was measured in isolated mitochondria using JC-1, a specific mitochondrial probe, by a change in the ratio of red fluorescence of the J-aggregates to that of the green fluorescence of its monomers at Ex/Em wavelengths of 485/ 595 nm and 485/535 nm, respectively. Routinely, 0.2 mM JC-1, 125 mM ADP and 0.5 mg of mitochondrial protein were used with the following added sequentially: 5 mM each glutamate/malate, 5 mM rotenone; 10 mM succinate and 10 mM malonate.

Activities of tricarboxylic acid cycle (TCA) enzymes
Assay conditions for measuring citrate synthase (CS) malate dehydrogenase (MDH) and glutamate dehydrogenase (GDH) in the freezed-thawed mitochondrial extracts have been described previously [21].

NADH dehydrogenase or RCI
The enzyme activity of RCI was measured by a decrease in fluorescence of NADH in 5 min at excitation wavelength 352 nm and emission wavelength 464 nm using decylubiquinone as an electron acceptor [22]. A mitochondrial extract containing 20 mg protein was added to a respiratory buffer of pH 7.2 containing 25 mM KH 2 PO 4 , 2 mM NaN 3 , 5 mM MgCl 2 , 64 mM decylubiquinone, 2 mg/ml albumin and 25 mM NADH.

Succinate dehydrogense or RCII
The assay was based on a decrease in absorbance of 2,6dichlorophenolindolephenol (DCPIP) at 600 nm monitored for 5 min, with decylubiquinone as an acceptor [23]. A mitochondrial extract containing 30 mg protein was employed in the presence of 1 mM rotenone, 2 mg/ml antimycin A and 2 mM NaN 3 to inhibit RCI, III and IV, respectively. A pre-incubation with succinate was carried out at 37uC for 5 min with the mitochondrial extract to overcome the known inhibitory effect of oxaloacetate on RCII.

Ubiquinol dehydrogenase or RCIII
This activity was measured by a coupled assay using reduced decylubiquinone and cytochrome c (cyt c) as described [24]. The assay was started by adding 50 mM cyt c to a buffer containing 250 mM sucrose, 1 mM EDTA and 50 mM Tris-HCl, pH 7.4 and 30 mg mitochondrial protein. The absorbance of reduced cyt c was followed for 5 min at 550 nm.

Cytochrome c oxidase or RCIV
The assay measured the rate of decrease of reduced cyt c as reported previously [25]. Cyt c was first reduced by adding 4 mg ascorbate to 10 mg/ml cyt c. The reaction was started by adding 80 mg mitochondrial protein to a solution containing 4 mM KH 2 PO 4 and reduced cyt c. The decrease in absorbance of reduced cyt c was followed at 550 nm for 20 sec.

F o F 1 -ATPase or RCV
This was measured by coupling the ADP formed to pyruvate kinase using phosphoenolpyruvate as the substrate. Pyruvate formed was reduced to lactate by lactate dehydrogenase (LDH) as described previously [25,26]. The decrease in NADH was monitored for 10 min at 340 nm in an absorbance microplate reader.

Determination of mitochondrial protein
The protein concentrations of the mitochondrial fraction and the whole cell lysate prepared from MEF WT and MEF PTEN2/2 were determined by the bicinchoninic acid assay [27]. The purity of the mitochondrial fraction was examined by Western blot analysis using the mitochondrial and cytosolic markers, voltage dependent anion channel (VDAC) and Cu-Zn superoxide dismutase (SOD-1), respectively.

Western blot analysis
Aliquots of the mitochondrial or cytosolic fractions or total cell lysates containing 50 mg protein were subjected to electrophoresis on sodium dodecyl sulfate-polyacrylamide gel and electro-blotted onto a nitrocellulose membrane. After blocking with 10% non-fat milk in TBS with 0.1% Tween-20, the membranes were probed with the respective primary antibodies of PTEN, Akt, p-Ser 473 Akt, p-Thr 308 Akt, 4E-BP1, p-Thr 37/46 4E-BP1, B-cell lymphoma 2 (Bcl-2) (Cell Signaling Technology, Beverly, MA, USA), VDAC (Calbiochem, San Diego, CA, USA) malic enzyme 1 (ME 1), hexokinase II (HK II), pan-actin (product no. sc-1615), PGC-1 a, NRF-1, mtTFA, SOD-1 (Santa Cruz Biotechnology, CA, USA), lactate dehydrogenase isoenzyme V (LDH-V) (Abcam, Cambridge, UK), uncoupling protein 2 (UCP2) (Alpha Diagnostic, Texas, USA), a-tubulin (Sigma-aldrich, St. Louis, MO, USA) and aconitase 2 (ACO2) (Abgent, San Diego, CA USA). Western blot analysis of the respiratory complexes was carried out using a cocktail containing antibodies against the following subunits of the five RCs: complex I subunit NDU FB8, complex II subunit 30 kDa (SdhB), complex III subunit Core 2 (RC III core 2), complex IV subunit I (CO1) and ATP synthase a-subunit (ATP5 F 1 a1) (MitoSciences, Oregon, USA). All primary antibodies were used at a dilution of 1:1000 except for a-tubulin and pan-actin, the loading controls, which were diluted 1:20000. Anti-goat pan-actin was used as the loading control for the Western blot analysis of the respiratory complexes because the a-subunit of ATP synthase and a-tubulin of same species (mouse) have similar molecular weight and therefore could not be resolved.

Inhibition by LY294002 and Akt inhibitor IV
Briefly, MEF PTEN2/2 grown to 70% confluence were preincubated with either 25 mM LY for 4 h or 1 mM Akt inhibitor IV for 8 h prior to enzymatic assays and Western blot analyses.

RNA extraction and reverse transcription
Total RNA from MEF WT and MEF PTEN2/2 was isolated using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. 500 ng RNA was reverse transcribed using 1 ml of ImPromII reverse transcriptase and 0.5 mg random hexamer for 60 min at 42uC and terminated at 70uC for 5 min according to the instruction from the manufacturer (Promega, Madison, WI, USA).

Primers and real-time PCR
The mRNAs of the same five RC subunits detected in Western blot analyses were quantified by real-time PCR by using their respective forward and reverse primers as shown in Table S1. The mRNA of 4E-BP1 was detected using forward primer: 59-GGACCAGCCGTAGGAC-39, reverse primer: 59-TGAGT-GAGGAGCAGGAC-39. Mouse alpha 1A tubulin (forward

Statistical Analysis
Data were presented as means 6 SD, and were analyzed by the Student's t-test where a p value of ,0.05 was considered significant.

Characterization of MEF PTEN2/2
We first confirmed the PTEN knockout phenotype in MEF PTEN2/2 by immunoblotting the protein expressions of PTEN, total and phospho-Akts in the whole cell lysate. The absence of PTEN causing Akt hyperactivation was evident by the higher expressions of p-Ser 473 Akt and p-Thr 308 Akt without a significant change in the total Akt (Fig. 1A). In the absence of PTEN, a significantly higher rate of proliferation was observed in MEF PTEN2/2 from day 1 to 4 ( Fig. 1B)  Oxidative phenotype in MEF PTEN2/2 ATP biosynthesis. The overall function of mitochondria in energy production was measured by the rate of ATP biosynthesis in isolated intact mitochondria upon the addition of respiratory substrates. As shown in Fig. 2A, the rate of ATP biosynthesis from the oxidation of malate/glutamate for RCI, succinate for RCII and TMPD/ascorbate for RCIV was significantly higher in MEF PTEN2/2 compared to the MEF WT .
Mitochondrial Membrane potential (MMP). The generation of mitochondrial membrane potential is often used as an indicator to assess mitochondrial respiratory function. We measured the MMP of intact mitochondria using the JC-1 dye with serial addition of substrates and inhibitors. The uptake of JC-1, a cationic dye, depends essentially on the MMP. The oxidation of respiratory substrates such as malate/glutamate and succinate added to mitochondria induced the generation MMP, resulting in an increased entry of JC-1 monomers which exhibited a green fluorescence. With enhanced accumulation, the monomers combined to form J-aggregates which showed a red fluorescence. The ratio of red:green fluorescence measured the magnitude of the MMP. The MMP induced by malate/glutamate and succinate was higher in the MEF PTEN2/2 compared to MEF WT as shown in Fig. 2B. Inhibition of RCI by rotenone and of RCII by malonate confirmed the measurement and functionality of RCI and RCII.
Respiratory Flux. To investigate the effect of PTEN/Akt on mitochondrial function, oxygen consumption was measured with intact cells in normal culture medium. As compared to isolated mitochondria, the intact cell analysis allowed an integration of regulatory events by which cytoplasmic and nuclear factors influenced mitochondrial function. Net R/E (where R = Routine and E = ETS) combined basal oxygen consumption (Routine), oligomycin-inhibited respiration (Leak) and maximal capacity of electron transport system (E) to provide a parameter R/E which measured the proportion of respiratory capacity coupled to oxidative phosphorylation. As shown in a representative set of data for MEF PTEN2/2 and MEF WT (Fig. 2C), the former had higher Net Routine/E, which inferred that a higher proportion of oxygen consumption of the electron transport system was coupled to ATP synthesis. This could be contributed by a lower degree of leak respiration, expressed as Leak/E as shown in Fig. 2C inset, which was obtained from three independent experiments.
Activities and protein expressions of respiratory complexes. The higher MMP and rate of ATP biosynthesis in MEF PTEN2/2 observed above suggested a possible increase in the electron transport system coupled to the phosphorylation of ADP. This was reflected in the data of respiratory flux above. Indeed, extracts of mitochondria isolated from MEF PTEN2/2 showed higher activities of all five respiratory complexes compared to the MEF WT (Fig. 2D). Interestingly, the ratios of RC III/I, IV/ I, and IV/III in the MEF PTEN2/2 were similar in magnitude to malate/glutamate (M/G, 5 mM each) and succinate (10 mM). Rotenone (5 mM) and malonate (10 mM), inhibitors of RCI and II, respectively, decreased the MMP as reflected by an attenuation of the ratio of red:green fluorescence. Mitochondrial protein used was 0.5 mg. Data presented is representative of 3 independent experiments. (C) Oxygen consumption The NetR/E (where R = Routine and E = maximal electron transport system capacity) was obtained as described under ''Experimental procedures'' using an Oxygraph 2k (Oroboros). A higher NetR/E was observed in MEF PTEN2/2 compared to MEF WT as shown in a representative set of data. On the other hand, Leak/E was always lower in MEF PTEN2/2 . A significant difference in NetR/E between the two cell lines showed that the electron transport system was more efficient in driving ATP synthesis in MEF PTEN2/2 compared to MEF WT as shown in Fig. 2C inset; the latter was based on three independent sets of experiments with ** p,0.01 for n = 3. (D) Enzyme activities of the RCs These were expressed in nmol/min/mg protein. They were measured as described in the ''Experimental procedures''. Ratios of RC III/I, IV/I and IV/III are shown in the inset. * p,0.05, # p,0.005 for n = 3. (E) Western blot analyses of respiratory complexes Whole cell lysates of MEF WT and MEF PTEN2/2 were probed with a cocktail containing antibodies against various subunits of the five RCs as described in the text. Quantification of the bands by the ImageJ software showed % increase in protein expressions of RCs in MEF PTEN2/2 relative to MEF WT . A higher degree of increase in RCI, III and IV, as compared to RCII and V was apparent (Fig. 2E inset). (F) Activities of TCA cycle enzymes Glutamate dehydrogenase (GDH), malate dehydrogenase (MDH) and citrate synthase (CS) were measured in freeze-thawed mitochondrial extracts as described in [21]. # p,0.005. for n = 3. doi:10.1371/journal.pone.0045806.g002 PTEN/Akt/4E-BP1 and Mitochondrial Respiration PLOS ONE | www.plosone.org those in the MEF WT (Fig. 2D, inset). This observation alluded to a coordinated regulation of these RCs via transcription, translation and/or post-translation. At the protein level, we observed higher expressions of all five RCs in the MEF PTEN2/2 compared to the MEF WT (Fig. 2E). Quantification of the bands using the ImageJ software showed a significant difference based on triplicate measurements and a higher magnitude of increase in RCI, III and IV compared to RCII and V (Fig. 2E inset).
Tricarboxylic acid (TCA) cycle enzymes. To sustain a higher MMP and rate of ATP biosynthesis, more reducing equivalents must be fed into the electron transport system from the TCA cycle. We found the activities of glutamate dehydrogenase (GDH) and malate dehydrogenase (MDH), two NAD-linked enzymes as well as citrate synthase (CS), a mitochondrial marker enzyme, were significantly higher in the MEF PTEN2/2 as compared to the MEF WT (Fig. 2F).
Enhanced mitochondrial function in MEF PTEN2/2 (a) Akt hyperactivation. A significantly higher protein expression of all RCs was found in MEF WT with siRNA knockdown of PTEN ( Fig. 3A and inset). The PTEN-knockdown cells exhibited higher Net R/E with lower Leak/E (Fig. 3B), suggesting that PTEN aberration could be responsible for the enhanced coupling between oxidation and phosphorylation in mitochondria isolated from MEF PTEN2/2 (shown previously as NetR/E in Fig. 2C). To further explore if the higher mitochondrial respiration in MEF PTEN2/2 was attributed to hyperactivated Akt, inhibition of the PI3K/Akt signaling pathway was attempted with LY294002, known to act on PI3K and mTOR [16,28,29] and Akt Inhibitor IV, an ATP-competitive inhibitor of Akt autokinase activity [30]. Both inhibitors reduced the phosphorylation at Ser 473 and Thr 308 of Akt without affecting the total Akt ( Fig. 4A and 4B). The reduction in Akt phosphorylation by these two inhibitors correlated with (a) decreased enzyme activities of all four electron-transport complexes (I to IV) in MEF PTEN2/2 ( Table 1) and (b) repression of the protein expressions of RCI, III, and IV (these are involved in proton pumping) in MEF PTEN2/2 ( Fig. 4C and 4D). The protein expressions of RCII and V were not or minimally affected.
(b) Regulation of gene transcription. The peroxisomal proliferator activated receptor c coactivator-1a (PGC-1a) and its downstream targets: nuclear respiratory factor-1 (NRF-1) and mitochondrial DNA transcription factor A (mtTFA) are known to enhance respiratory function by acting on several nuclear genes encoding subunits of the respiratory complexes [31,32]. The level of these proteins could provide some indication if transcriptional regulation played a role in the Akt-mediated up-regulation of mitochondrial activities observed in MEF PTEN2/2 . No significant difference was observed in their protein expressions ( Fig. 5A and  inset). Similarly, the mRNA levels of the five subunits of respiratory RCs used in our Western blot analysis were not significantly different between MEF PTEN2/2 and MEF WT (Fig. 5B and inset). Together, these sets of data suggested that the increased expressions of the RCs observed in MEF PTEN2/2 were not likely to be dependent on gene transcription.
(c) 59 cap-dependent protein translation. We next explored if the Akt-mediated mitochondrial function was regulated by protein translation. As mentioned previously, 4E-BP1 has a central role in converging the upstream signals to protein translation [12,13]. Comparatively, MEF PTEN2/2 had lower 4E-BP1 (Fig. 6A) and a higher proportion of p-Thr 37/46 4E-BP1 relative to total 4E-BP1 when compared to MEF WT (Fig. 6A,  inset). The mRNA levels of 4E-BP1 in the two cell types were not significantly different (p = 0.148) as shown in Fig. 6B. In addition, the repression of activities and expressions of the RCs by LY294002 and Akt inhibitor IV could also be mediated by 4E-BP1 as incubation with LY294001 resulted in a band shift downwards in total and p-4E-BP1 (shown by arrows in Fig. 6C), a phenomenon which indicated hypo-phosphorylation [11]. Although there was no band shift with Akt inhibitor IV, there was evidence of a more hypo-phosphorylated p-Thr 37/46 4E-BP1 and total 4E-BP1 (shown by arrows of lower bands of greater intensity in Fig. 6D). In general, hypo-phosphorylation of 4E-BP1 and/or p-Thr 37/46 4E-BP1 would lead to decreased protein translation which was evident in RC I, III and IV expressions in MEF PTEN2/ 2 following inhibition by LY and Akt inhibitor IV (Figs. 4C and 4D). Gene silencing of 4E-BP1 in MEF WT and MEF PTEN2/2 upregulated the expressions of all RCs (Fig. 6E and 6F, respectively). The fact that this was also observed in MEF PTEN2/2 suggested that there was residual 4E-BP1 in these cells which was demonstrated in Fig. 6A. The higher degree of increase in RC expressions in the MEF WT by si-4E-BP1 (Fig. 6E) could be explained by their higher basal 4E-BP1 (Fig. 6A). A number of other assorted proteins were also measured following si-4E-BP1, but their protein expressions were not significantly changed, except for Bcl-2 which was down regulated ( Figure S1).    Fig. 6G inset. However, when 4E-BP1 was intact (Lane 4), inhibition of Akt down-regulated RC expressions. This positive control showed the efficiency of the Akt inhibitor IV, which has previously been demonstrated in MEF PTEN2/2 (Fig. 4D). Together, the results suggested that regulation of RC expressions by Akt was dependent on 4E-BP1.
(e) Increased mitochondrial protein relative to total protein. In view of the mitochondrial location of the RCs, it was conceivable that their up-regulation could contribute to more proteins in this organelle relative to the total cellular protein in MEF PTEN2/2 . Indeed, we observed a three-fold increase in the ratio of mitochondrial protein relative to total cellular protein in MEF PTEN2/2 compared to MEF WT (Fig. 6H). The purity of the mitochondrial preparation was confirmed by the use of SOD-1 and VDAC, the cytosolic and mitochondrial markers respectively (Fig. 6H, inset).

Discussion
The immortalized MEF PTEN2/2 [6] provided a cell model to study any aberrant metabolism resulting from the oncogenic action of Akt. Following its activation, Akt has been reported to translocate to the mitochondria to act on its downstream targets [33]. Thus, there is evidence of a cross-talk between Akt and mitochondria. This study explores if Akt hyperactivation has an impact on mitochondrial metabolism. It is conceivable that accelerated aerobic glycolysis from Akt signaling [14] produces increased pyruvate which is channeled concomitantly to lactate and the TCA cycle. The reducing equivalents, NADH and FADH 2 generated in the mitochondrial matrix are finally oxidized by the active electron transport system consisting of up-regulated RCs in MEF PTEN2/2 . These could contribute to the higher MMP and ATP biosynthesis in MEF PTEN2/2 and increased oxygen consumption reported in Akt-transfected MEFs [34]. Our study also showed consistently higher net routine oxygen consumption in MEF PTEN2/2 which was reproducible by silencing PTEN in the wild type. In both instances, the lower degree of leak respiration may lead to their greater efficiency in OXPHOS. Lower leak is possibly the result of their higher cholesterol content as we had noted consistently that 2-3 times more digitonin was required for cell permeabilization of MEF PTEN2/2 . It has been reported that cholesterol enhances the rigidness of mitochondrial membranes leading to a lower leak [35]. In addition, higher enzyme activities of the RCs could be another contributing factor. Of particular interest is that the MEF PTEN2/2 retain similar relative proportions of activities of RCI, III and IV as those present in the wild type which infers that they could be co-ordinately regulated. These three RCs form supercomplexes known as respirasomes [36,37] which may subject them to collective regulation. Our ratio of the enzyme activities of 1:2 for RCIII: RCIV is coincidentally similar to that deduced in the model of the ''respiratory string'' which is made up of several respirasomes [37].
Two inhibitors of the Akt/mTOR pathway were used to examine if the higher RC activities and expressions were attributed to hyperactivated Akt from PTEN aberration in MEF PTEN2/2 . They are LY294002 (LY) which targets both PI3K and mTOR [16,28,29] and Akt inhibitor IV which has previously been reported to block Akt autokinase activity [30]. Both inhibitors decreased RC enzyme activities and protein expressions of RC I, III, and IV, with little to no change in RCII and RCV, an observation which further supports the concept of organization of these three proton pumping RCs in supercomplexes of respirasomes discussed above [37]. In addition to their co-ordinated regulation, their ease of change could be explained by their different values of DG of folding energy and the length of the 59UTRs [38]. RC IV with a DG of 27Kcal/mol and short 59UTRs could render it more amendable to inhibition and possibly transcription and/or translation in contrast to RCV with a DG of 246Kcal/mol which would be more resistant to change as towards LY and Akt inhibitor in this study. Overall, LY seems to exert a greater influence on these reactions as it is an inhibitor of both PI3K and mTOR [16,28,29]. Inhibition of mTORC2 by LY would compromise the phosphorylation of Ser 473 and the subsequent phosphorylation at Thr 308 of Akt necessary for its full activation [39]. In addition, its inhibition of mTORC1 would have a more direct impact on 4E-BP1 as compared to Akt inhibitor IV which acts upstream. However, the greater effect of LY did not result in a more pronounced inhibition of RC enzyme activities. This may suggest that RC activities are more directly affected by Akt kinase since the Akt inhibitor IV has been reported to compromise the Akt autokinase activity more significantly than wortmanin, another PI3K inhibitor [30]. No attempt is made to correlate the degree of inhibition of enzyme activities with protein expressions of the RCs as Western blot analysis depends on only one of several subunits present in the RC, whereas enzyme activity measures the functionality of the composite RC. A differential mechanism of action of the two inhibitors was also inferred from the distinct band shifts in 4E-BP1 and its phosphorylated form by LY which was not apparent with Akt inhibitor IV.
Although our data of LY and Akt inhibitor IV described above provide a plausible relationship between Akt, 4E-BP1 and RC expressions, siRNA targeting 4E-BP1 is deemed necessary to establish unequivocally a direct association between 4E-BP1 and RC expressions. In eukaryotes, protein synthesis is mainly regulated by 4E-BP1, a repressor of 59cap-dependent translation [40]. 4E-BP1, a downstream target of mTORC1, binds eIF4E to inhibit protein translation. On phosphorylation, it detaches from eIF4E to initiate protein translation [11]. With gene silencing of 4E-BP1, expressions of RC I to V were up-regulated in MEF WT . Likewise, PTEN gene silencing showed a similar effect. The first suggestive evidence of an involvement of translational regulation was the observation of a drastic reduction of this translational repressor in MEF PTEN2/2 . To examine if up-regulation is peculiar to the RCs, we measured the expressions of assorted proteins (namely cytosolic proteins: ME 1, LDH-V and mitochondrial proteins: HK II, UCP2, ACO2, Bcl-2 and VDAC) besides RCI to V, following silencing of 4E-BP1. Of these, only the RCs were up-    Fig. 6A. The fact that RC up-regulation was also observed in MEF PTEN2/2 suggested that there was residual 4E-BP1 in the knockout which was indeed shown in Fig. 6A. (G) Gene silencing of 4E-BP1 with and without Akt Inhibitor IV on RC expressions in MEF WT RC expressions were significantly higher with si-4E-BP1 compared with untreated control (Lane 1) in MEF WT in the absence or presence of Akt inhibitor IV as shown in Lanes 2 and 3, respectively (*p,0.05, # p,0.005 relative to untreated controls). Thus, it was evident that in the absence of 4E-BP1, Akt did not affect RC expressions significantly with p values .0.05 (Fig. 6G inset, comparing Lanes 2 and 3). In contrast, in the presence of an Akt inhibitor (AktIV), RC expressions were down-regulated when 4E-BP1 was intact (Lane 4) showing that the action of regulated which appeared to suggest some degree of specificity of 4E-BP1 regulation on the electron-transport proteins Generally, the selection of mRNAs for translation is exerted at the initiation stage with binding of elF4F complex to 59-end of each mRNA and interaction between 39UTR binding proteins and 4E-BPs [41]. The binding of the mRNA binding proteins of the Pumilio (PUF) family to the 39 UTR can control mRNA stability and translational repression [42] to provide some measure of selectivity. Of all the Pufs indentified in yeast, Puf3p has been shown to bind nearly exclusively to nuclear encoded mitochondrial mRNAs [43,44]. Mitochondrial mRNAs, which essentially code for the assembly factors of respiratory chain complexes and the mitochondrial translation machinery are highly dependent on the presence of Puf3 [45]. Taken together, we concluded that it is not impossible for 4E-BP1 to regulate expressions of mitochondrial RCs in a selective manner. However, post-translational modification of enzyme activities cannot be ruled out as several subunits of RCI were predicted from proteomic studies to be potentially capable of phosphorylation, e.g. tyrosine phosphorylation of RCI and RCV [46,47]. Transcriptional control was deemed unlikely as there was no significant difference in the mRNAs of several subunits of the RCs between MEF PTEN2/2 and MEF WT . Furthermore, the protein expressions of the transcription factor PGC-1a and its associated downstream targets, NRF-1 and mtTFA were similar between MEF PTEN2/2 and MEF WT . Our conclusion concurred with the report that the regulation of mitochondrial function by mTOR is independent of transcription [48].
In conclusion, the hyperactivated Akt confers on MEF PTEN2/2 a bioenergetic advantage by up-regulating their oxidative mitochondrial functions. We propose the enhanced mitochondrial capacity in MEF PTEN2/2 is used to support their higher rate of proliferation upon substrate availability. Increased OXPHOS has been shown in H-Ras transformed fibroblast [49] and this bestows a proliferative advantage to K-Ras-mediated tumorigenecity [50] and provides energy to transformed human mesenchymal stem cells [51]. Our hypothesis that mitochondrial respiratory capacity is regulated by the Akt/mTOR/4E-BP1 cascade as summarized in Fig. 7. Supporting evidence was provided using known small molecules as inhibitors of this signaling pathway and siRNA to knockdown PTEN or 4E-BP1 to demonstrate the link between PTEN/Akt/mTOR/4E-BP1 signaling and protein expressions of the electron transport respiratory complexes. Our study has highlighted the role of 4E-BP1 in mitochondrial function in the early-transformed cells with PTEN/Akt aberration. Figure S1 Gene silencing of 4E-BP1 on other proteins. There was little to no effect on other proteins examined with siRNA on 4E-BP1 in MEF WT except for Bcl-2 which was decreased, p,0.05. In contrast, there was up-regulation of all respiratory complexes as shown in Figure 6E and 6F. (TIF)

Supporting Information
Table S1 Primers used in Real-time PCR. mRNAs of subunits (encoded by mitochondrial and nuclear DNAs) of respiratory complexes were measured using the respective primers as shown. These mRNAs are of the same RC subunits detected by Western blots. (DOC)