FOXO3 Transcription Factor Is Essential for Protecting Hematopoietic Stem and Progenitor Cells from Oxidative DNA Damage*

Accumulation of damaged DNA in hematopoietic stem cells (HSC) is associated with chromosomal abnormalities, genomic instability, and HSC aging and might promote hematological malignancies with age. Despite this, the regulatory pathways implicated in the HSC DNA damage response have not been fully elucidated. One of the sources of DNA damage is reactive oxygen species (ROS) generated by both exogenous and endogenous insults. Balancing ROS levels in HSC requires FOXO3, which is an essential transcription factor for HSC maintenance implicated in HSC aging. Elevated ROS levels result in defective Foxo3−/− HSC cycling, among many other deficiencies. Here, we show that loss of FOXO3 leads to the accumulation of DNA damage in primitive hematopoietic stem and progenitor cells (HSPC), associated specifically with reduced expression of genes implicated in the repair of oxidative DNA damage. We provide further evidence that Foxo3−/− HSPC are defective in DNA damage repair. Specifically, we show that the base excision repair pathway, the main pathway utilized for the repair of oxidative DNA damage, is compromised in Foxo3−/− primitive hematopoietic cells. Treating mice in vivo with N-acetylcysteine reduces ROS levels, rescues HSC cycling defects, and partially mitigates HSPC DNA damage. These results indicate that DNA damage accrued as a result of elevated ROS in Foxo3−/− mutant HSPC is at least partially reversible. Collectively, our findings suggest that FOXO3 serves as a protector of HSC genomic stability and health.

The accumulation of damaged DNA compromises the genomic stability of hematopoietic stem cells (HSC) 5 (1) and may promote hematological malignancies, specifically with age (2,3). Reactive oxygen species (ROS) are one of the major insults that induce DNA damage (4), including in normal hematopoietic and leukemic stem cells. ROS are generated both endogenously by cell metabolism or expression of oncoproteins and exogenously by ionizing radiation and genotoxic drugs (1,5,6). ROS induce single and double strand DNA breaks as well as various species of oxidized nucleotides and are implicated in the pathophysiology of hematological malignancies (7). Thus, modulations of ROS may be used therapeutically (7,8). Cells, with distinct kinetics, constantly repair oxidative damage using mainly the base excision repair (BER) pathway (9). The relatively low levels of endogenous ROS in HSC restrict damage to DNA (10); however, the impact of increasing ROS levels on HSC DNA remains relatively unexplored (11). Importantly, the mechanism of oxidative DNA damage repair in primitive hematopoietic cells remains poorly understood.
The transcription factor FOXO3 of the Forkhead family with four (FOXO1, FOXO3, FOXO4, and FOXO6) related members maintains HSC quiescence by ensuring low levels of ROS (12,13). FOXOs belong to an evolutionarily conserved family of transcription factors that exert critical functions in the regulation of aging and longevity, including humans (14,15). FOXOs are phosphorylated and inhibited mainly by AKT kinase downstream of the PI3K-signaling pathway. In addition to phosphorylation by AKT and other kinases, FOXOs are modified by a variety of post-translational modifications that together determine FOXOs' functional output (14). Notably, in hematopoietic stem and progenitor cells and in embryonic stem cells, AKT is not the dominant regulator of FOXO function (16 -20). FOXOs are involved in an array of fundamental biological pro-cesses, including the regulation of metabolism and oxidative stress, cell cycle, apoptosis, and DNA repair (14,15).
Loss of FOXO3 results in oxidative stress-mediated myeloproliferation that does not progress, at least not rapidly enough to have been detected, toward leukemia (12,13,21). FOXO3 is intimately involved in hematopoietic malignancies, as FOXO3 is found in chromosomal translocations of human acute myeloid leukemia (22) and is inhibited in malignant hematopoietic cells (18,(23)(24)(25)(26)(27)(28). Despite these observations, FOXO3 is also required for the maintenance of both mouse and human leukemic stem cells (29 -31). In addition, FOXO3 is implicated in HSC aging (32)(33)(34) and in stem cell pluripotency (16). Given the importance of DNA damage response for aging and malignancies of stem cells, we sought to explore the potential FOXO3 function in the regulation of DNA damage response in HSC.
in the primitive hematopoietic cell DNA damage response, specifically in base excision repair, and it protects HSPC from oxidative DNA damage under homeostasis. These findings raise the possibility that DNA damage accrual as a result of loss of FOXO3 function, as may occur with age, promotes HSC aging (32)(33)(34), predisposes HSPCs to premature aging, and/or contributes to hematopoietic stem cell malignant transformation (18,23,29,30).
Under homeostatic conditions, a significantly higher fraction of freshly isolated Foxo3 Ϫ/Ϫ LSK cells exhibited enhanced phosphorylation of histone H2AX variant (␥H2AX) (Fig. 1B), a sensor of DNA double strand breaks (37). A highly elevated level of damaged DNA was detected in Foxo3 Ϫ/Ϫ HSPC by both ␥H2AX immunofluorescence staining, in which cells with more than six nuclear foci were considered as positive, and the more sensitive flow cytometry assay, which enabled the quantification of the amount of damage (Fig. 1B, n Ͼ 40 cells analyzed per condition and Fig. 1C; n ϭ 5, * p Ͻ 0.05). Damaged DNA accumulated in Foxo3 Ϫ/Ϫ HSPC subpopulations, including long term repopulating HSC (LT-HSC; LSK Flk2 Ϫ CD34 Ϫ ) and c-Kit ϩ (Lin Ϫ Sca1 Ϫ c-Kit ϩ ) multipotent progenitor cells but not in Foxo3 Ϫ/Ϫ total bone marrow (BM) control cells (Fig. 1,C for gating strategy, and D; n ϭ 4 per genotype); although ␥H2AX was relatively increased in lineage negative cells depleted of mature blood cells and enriched for hematopoietic stem and progenitor cells, it did not reach significance in the samples evaluated. Using alkaline single-cell gel electrophoresis (comet assay), we further visualized the damage to single strand DNA and quantified an approximate 3-fold increased damage level in FACS-sorted Foxo3 Ϫ/Ϫ versus wild type (WT) LSK cells by the use of (38) % of DNA in Tail and Olive tail moment parameters ( Fig. 1E; n ϭ 6). These results confirmed the increased amount of damaged DNA in Foxo3 Ϫ/Ϫ HSPC.
To evaluate whether ROS were involved in the accumulation of DNA damage in Foxo3 Ϫ/Ϫ HSPC, we used the FLARE hOGG1 comet assay that specifically detects oxidative DNA damage. This approach showed that oxidative DNA damage is significantly increased in freshly isolated Foxo3 Ϫ/Ϫ versus WT HSPC ( Fig. 2A and B, representative comet images and quantification). Using a specific probe that detects the main DNA oxidation lesion, 8-hydroxyguanosine (8-OHdG), by flow cytometry, we noted 8-OHdG levels were increased in Foxo3 Ϫ/Ϫ LSK cells as compared with WT cells (Fig. 2C), although the difference did not reach significance in the replicates analyzed (n ϭ 3). Altogether, these results indicate that Foxo3 Ϫ/Ϫ HSPC DNA accumulates high levels of oxidative insults.

Scavenging ROS by NAC Decreases Foxo3
Ϫ/Ϫ HSPC DNA Damage and Corrects the G 2 /M Delay-To investigate whether ROS have any functional role in the accumulation of DNA damage, mice were treated in vivo with NAC (100 mg/kg/day), a source of glutathione for 14 days (Fig. 3A, schematic). As anticipated, NAC treatment normalized ROS levels in Foxo3 Ϫ/Ϫ LSK cells without any significant impact on WT cells (Fig. 3B) (13). Notably, NAC treatment led to a 2-fold decrease of ␥H2AX levels in Foxo3 Ϫ/Ϫ -treated as compared with nontreated LSK cells (n ϭ 3 per group; p Ͻ 0.05) (Fig. 3C). In addition, the levels of DNA breaks in Foxo3 Ϫ/Ϫ LSK cells, as quantified by the % Tail DNA and the Olive tail moment, were significantly reduced (2-fold) in response to NAC (Fig. 3D, quantification in the right panel). Despite normalized ROS levels (Fig. 3B), the % Tail DNA (p Ͻ 0.0007) and the Olive tail moment (p Ͻ 0.00001) remained significantly higher in NACtreated Foxo3 Ϫ/Ϫ LSK cells as compared with WT controls (Fig.  3D), raising the possibility that not all damage to Foxo3 Ϫ/Ϫ LSK DNA resulted from elevated ROS.
In addition to reducing DNA damage, NAC treatment normalized the defective Foxo3 Ϫ/Ϫ cell cycle parameters (Fig. 4).  were comparable (Fig. 4, C and D). These combined results suggest that the block of cell cycle progression, G 2 /M delay, in Foxo3 Ϫ/Ϫ HSPC is due to ROS-mediated DNA damage.

Increased numbers of LSK cells isolated from NAC-treated
Defective DNA Repair Machinery in Foxo3 Ϫ/Ϫ HSPC Contributes to DNA Damage Accumulation-A number of key genes of BER, including DNA polymerase ␤ (polB), X-ray repair cross-complementing protein 1 (Xrcc1), and DNA Ligase 1 (Lig1), were significantly down-regulated in Foxo3 Ϫ/Ϫ LSK cells (Fig. 5A). Other genes implicated in nucleotide excision repair (NER), were also significantly reduced in Foxo3 Ϫ/Ϫ LSK cells (Fig. 5A). Altogether, these results raise the possibility that the key pathways implicated in oxidative DNA damage repair may also be compromised in Foxo3 Ϫ/Ϫ HSPC.
To address whether DNA in LSK cells that lack FOXO3 were preferentially susceptible to ROS elevation, we exposed these cells to 100 M hydrogen peroxide for 1 h (Fig. 5B, schematic). This treatment led to significantly higher DNA breaks in Foxo3deficient LSK cells as compared with their WT counterparts. The results were relatively similar when analyzed by either the standard or the FLARE hOGG1 comet assays (Figs. 5, C and D); the FLARE assay detected a more exacerbated DNA damage (Fig. 5D), suggesting increased sensitivity to removal of base damage such as those removed by the OGG1 DNA glycosylase. These findings indicate that the repair of oxidative DNA insults  might be compromised in Foxo3 Ϫ/Ϫ LSK cells. Despite the increase in oxidative stress-mediated accumulation of damaged DNA (Fig. 2) and consistent with mitochondrial defects observed in Foxo3-null HSPC (36), apoptosis was increased only mildly but significantly in in vitro hydrogen peroxidetreated Foxo3 Ϫ/Ϫ LSK cells, suggesting that Foxo3 Ϫ/Ϫ LSK cells might exhibit some resistance to oxidative stress-mediated apoptosis (Fig. 5E).
To address a possible DNA repair defect, we used a recently developed (39) BER molecular beacon assay to quantitatively evaluate APE1 endonuclease activity and OGG1-mediated gly-cosylase activity for removal of the 8-oxo-dG lesion in an 8-oxo-dG/A base pair (Fig. 5F, schematic). Using this fluorescent real time quantitative assay, we found no difference in the APE1 endonuclease activity of WT and Foxo3 Ϫ/Ϫ primitive hematopoietic cells (data not shown). However, compared with WT controls, the glycosylase activity of OGG1 in Foxo3 Ϫ/Ϫ primitive hematopoietic cells was significantly reduced (Fig. 5G, n ϭ 6, p ϭ 0.003).
Despite using three different commercial (goat, rabbit, and mouse) anti-mouse OGG1 antibodies probing wild type and OGG1-deficient mouse embryonic fibroblasts (MEFs), we were unable to confirm their specific binding to OGG1 protein (data  FEBRUARY 17, 2017 • VOLUME 292 • NUMBER 7 not shown). Using the same approach, we confirmed that anti-DNA Pol␤ and anti-XRCC1 antibodies specifically bind to their targets in wild type MEF but not in pol ␤ Ϫ/Ϫ and XRCC1 Ϫ/Ϫ MEFs, respectively (Fig. 6A). We further showed that whereas the expression of pol ␤ is increased, the expression of XRCC1 is significantly reduced in HSPC in the absence of Foxo3 (Fig. 6, B-C). XRCC1 is an essential element of base excision repair (40). Loss of FOXO3 may lead to discrepancies between mRNA and protein expressions (41,42). Given the critical OGG1-XRCC1 interaction for BER, it is likely that reduced expression of XRRC1 mediates the defective OGG1 activity (43)(44)(45). Together with XRCC1 requirement for recruiting pol ␤ to damaged DNA (46), these results support the notion that BER-mediated DNA repair in HSPC is dependent on FOXO3.

FOXO3 Is Key to HSC Oxidative DNA Damage Response
Foxo3 Ϫ/Ϫ Hematopoietic Stem and Progenitor Cells Respond Normally to Ionizing Radiation Insult-To investigate Foxo3 Ϫ/Ϫ HSPC response to additional insults, we used ionizing radiation (IR). LSK cells freshly isolated from WT and Foxo3 Ϫ/Ϫ mice and kept in vitro displayed similar numbers of ␥H2AX nuclear foci 2 h after 10 Gy IR (Fig. 7A). In addition, WT and Foxo3 Ϫ/Ϫ c-Kit ϩ hematopoietic progenitor cells responded similarly to 4 Gy IR, a dose to which HSPC are relatively tolerant (Fig. 7B) (47). More precise quantitative analysis by FACS 2 h after LSK and c-Kit ϩ multipotent progenitor cells were submitted to a 4 Gy IR dose confirmed that the capacity of Foxo3 Ϫ/Ϫ HSPC in accumulating ␥H2AX is similar to WT cells (Fig. 7B). Similar results were obtained by in vivo total body irradiation (TBI). WT and Foxo3 Ϫ/Ϫ LSK cells isolated 6 or 24 h after a 4 Gy TBI showed similar levels of DNA breaks (Fig. 7, C and D). Interestingly, although the expression of genes involved in homologous recombination (HR), such as Brca1, Brca2, and Rad51, was not modulated by the loss of FOXO3 in HSPC, genes implicated in the error-prone non-homologous end-joining pathway, which is the main repair mechanism of damaged DNA in HSPC (47) were significantly up-regulated in Foxo3 Ϫ/Ϫ HSPC (Fig. 7E). Although the increased expression of XRCC6 (Ku70) (and possibly XRCC5 (Ku80)) might be part of a compensatory response (48), the source of this increase is unclear. These results suggest that loss of FOXO3 does not exacerbate ionizing radiation-induced DNA damage or alternatively does not compromise the DNA damage response machinery in response to ionizing radiation.
Altogether, our data identify FOXO3 as a regulator of DNA damage repair in HSPC under homeostasis and suggest that the observed DNA breaks (Figs. 1, B, D, and E, 2, 3, and 5G) are due to both an increase in endogenous ROS levels and a deficiency in oxidative DNA repair machinery in Foxo3 Ϫ/Ϫ HSPCs (Fig. 8, Model).

Discussion
We showed here that Foxo3 Ϫ/Ϫ HSPCs accumulate damaged DNA under homeostatic conditions. We also showed that the damaged DNA in homeostatic Foxo3 Ϫ/Ϫ HSPCs is mediated by both elevated endogenous ROS and a defective base excision DNA repair program. Additionally, our data suggest that the Foxo3 Ϫ/Ϫ HSC G 2 /M delay is mediated primarily by elevated ROS. These results are consistent with and extend the scope of known FOXO3 (FOXO) functions in DNA damage response pathways (14, 48 -50).
Our findings show that FOXO3 modulates a gene network related to BER and NER oxidative DNA repair, because a number of related genes are down-regulated in Foxo3 Ϫ/Ϫ LSK cells (Fig. 5A). We were able to reveal oxidative DNA damage in Foxo3 Ϫ/Ϫ hematopoietic stem and progenitor cells using the FLARE hOGG1 assay, in which the human OGG1 glycosylase is introduced into the comet assay to induce DNA breaks at locations of oxidative base lesions (Fig. 2). Furthermore, the hypersensitive BER molecular beacon assay that we recently developed (39) enabled us to show that OGG1-mediated glycosylase activity, which mediates the removal of the 8-oxo-dG lesion, is reduced in a population of Foxo3 Ϫ/Ϫ hematopoietic cells enriched for stem and progenitor cells (Fig. 5, F and G). The reduction in OGG1 glycosylase activity in Foxo3 Ϫ/Ϫ LSK was despite similar Ogg1 transcript expression in Foxo3 Ϫ/Ϫ as compared with WT LSK cells (Fig. 5A). Furthermore, we found that the expression of the XRCC1 protein, which is critical for BER, is highly reduced in Foxo3 Ϫ/Ϫ Lin Ϫ cells (Fig. 6, B and C). Interestingly, despite the reduction in transcript expression of both XRCC1 and Pol␤, only the XRCC1 protein was reduced in Foxo3 Ϫ/Ϫ LSK cells (Fig. 6). As we had noted previously, the transcript expression in primary Foxo3 Ϫ/Ϫ hematopoietic cells may not always fully correlate with the protein expression (41,42). These results together implicate FOXO3 in the regulation of BER in hematopoietic stem and progenitor cells. These combined findings raise the possibility that compromised FOXO3 function, as it might occur in aging stem cells (33,34) or in the context of stem cell malignancies, is likely to sustain damaged DNA and mitochondrial defects (36) and further contribute to stem cell aging and/or malignancy (18,23,29,30,(32)(33)(34)36). In agreement with this, FOXO3 inactivation is proposed to be one of the early events in the evolution of myeloid and perhaps other malignancies (18,51). One of the implications of these combined results (36) is that elevated ROS, as a result of both mitochondrial defects and reduced anti-oxidant enzyme expression, contribute significantly to some, specifically to enhanced myeloproliferation, but not all of the main Foxo3 Ϫ/Ϫ HSC defects. They also suggest that ROS-mediated Foxo3 Ϫ/Ϫ HSC DNA damage may constitute a partially reversible phase in this process, because NAC treatment decreased DNA break levels and rescued HSPC from the cell cycle G 2 /M delay (Figs. 3 and 4). However, potential clinical applications of these findings warrant careful consideration. As FOXO3 loss negatively affects the BER pathway (Fig. 5), the normalization of HSC cycling and G 2 /M correction that followed antioxidant therapy (Fig. 4) might constitute only a transient response, in agreement with our recent report showing that NAC treatment is unable to rescue the long term reconstitution ability of Foxo3 Ϫ/Ϫ HSC (36).
Data presented here, combined with published work (33,34,36,52,53), depict FOXO3 as a molecular node that wires together mitochondrial metabolism (36), ROS signaling, and DNA damage repair mechanisms for the maintenance of healthy HSPC. These collective findings join growing evidence in support of the notion that FOXO3 serves as a barrier to genomic instability in HSPC (52, 53). FIGURE 7. Foxo3 ؊/؊ hematopoietic stem and progenitor cells respond normally to ionizing radiation insult. A, ␥H2AX staining on WT and Foxo3 Ϫ/Ϫ LSK cells that were submitted to 10 Gy of ionizing radiation, kept in culture, and analyzed after 2 h. B, quantification of ␥H2AX levels by flow cytometry in WT and Foxo3 Ϫ/Ϫ c-Kit ϩ and LSK cells 2 h after 4 Gy ionizing radiation dose (n Ն3 per group). Plot presents geometric mean values of FITC fluorescence that were normalized to basal WT levels. C, representative comet assay pictures of LSK cells isolated from control (C) or 4 Gy irradiated (IR) WT and Foxo3 Ϫ/Ϫ animals after 6 or 24 h and comet assay quantification (D). E, QRT-PCR analysis of DNA breaks repair genes involved in HR (Brca1, Brca2, and Rad51) or non-homologous end-joining (Xrcc5 and Xrcc6) in Foxo3 Ϫ/Ϫ c-Kit ϩ or LSK cells under homeostatic conditions. Actnb was used as an internal control, and expression was normalized to that of WT samples. Data are expressed as mean Ϯ S.D. Student's t test. *, p Ͻ 0.05; **, p Ͻ 0.001 ; ***, p Ͻ 0.0002; ns, not significant. FEBRUARY 17, 2017 • VOLUME 292 • NUMBER 7

Experimental Procedures
Mice-All mice were from the C57BL/6 genetic background and were 10 -12 weeks old (13). Protocols were approved by the Institutional Animal Care and Use Committee of the Icahn School of Medicine at Mount Sinai. NAC treatment was performed as described previously (13,21).
Flow Cytometry and Hematopoietic Stem Cell Isolation-Antibody staining and bone marrow cell preparation for FACS sorting were performed as described previously (13,21,33). Briefly, for isolation of LSK and c-Kit ϩ cells, freshly isolated bone marrow cells were stained with biotinylated hematopoietic multiple lineage monoclonal antibody mixture (Stem Cell Technologies), PE-Sca-1, APC-c-Kit (BD Biosciences), and incubated with Pacific Blue TM streptavidin secondary antibody. In addition to LSK staining and to isolate the long term HSC (LSK-Flk2 Ϫ CD34 Ϫ ) , total bone marrow cells were stained with FITC-CD34 (eBioscience) and PE-Cy5-Flk2 (BD Biosciences) antibodies. Analyses and FACS sorting were performed at the Flow Cytometry Core at Icahn School of Medicine at Mount Sinai.
QRT-PCR-RNA from FACS-sorted c-Kit ϩ or LSK cells was extracted using the EasySep MicroPlus RNA extraction kit (Qiagen). RNA was retro-transcribed using SuperscriptII (Invitrogen) and cDNA corresponding to 300 cells was used per well for QRT-PCRs with specific primers (Table 1). QRT-PCRs were run on an ABI7900 thermal cycler (Applied Biosystems).
In Vitro Hydrogen Peroxide (H 2 O 2 ) Treatment of LSK Cells-FACS-sorted WT and Foxo3 Ϫ/Ϫ LSK cells were incubated in StemSpan SFEM (Stem Cell Technology) and treated with or without 100 M of H 2 O 2 , washed extensively after 1 h, and either analyzed immediately (1 h) or cultured for further analyses at the indicated (2 and 4 h) time points by comet assay.
TBI-Mice were submitted to total body irradiation (4 Gy) (Icahn School of Medicine at Mount Sinai Irradiator Shared Resource Facility). Mice were sacrificed, and bone marrow cells were collected from non-irradiated controls or after 6 or 24 h post-TBI, and live LSK cells were FACS sorted and submitted immediately to comet assay.
Apoptosis Assay-Freshly FACS-sorted control and/or 1 h of H 2 O 2 -treated LSK cells were suspended in 1ϫ annexin V-binding buffer containing 2.5 l of annexin V-APC. Samples were co-stained with 7-AAD, following the manufacturer's instructions (BD Biosciences), and analyzed immediately by FACS.
␥H2AX Analysis by Immunofluorescence Staining and Flow Cytometry-WT and Foxo3 Ϫ/Ϫ FACS-sorted LSK cells were cytospun onto glass slides. ␥H2AX nuclear foci were analyzed by immunofluorescence staining using a rabbit polyclonal antiphospho H2AX (Ser-139) (Millipore) and imaged on a Leica DMRA2 fluorescence microscope using ϫ400 oil immersion objective. Freshly isolated bone marrow cells stained for HSPC were fixed in 2% paraformaldehyde and stained overnight at 4°C with 1:100 mouse monoclonal anti-H2AX pS139 FITC conjugate (Millipore) (or anti-H2AX pS139 APC conjugate, Biolegend) in BLOCK9 solution as described previously (54). The samples were next diluted into PBS, 2% FBS and analyzed by FACS.
Single Cell Gel Electrophoresis (Comet Assay) and FLARE TM Assay-FACS-sorted LSK cells were submitted to alkaline comet assay using the Trevigen CometAssay kit following the manufacturer's instructions. Briefly, cells were mixed in melted agarose, placed in glass slides, and allowed to jellify at 4°C for 30 min. After that, slides were subsequently immersed in lysis and unwinding solutions and immediately submitted to electrophoresis at 4°C in an alkaline electrophoresis solution, pH 13. Finally, slides were dehydrated in 70% ethanol and allowed to dry before analysis.
For the FLARE hOGG1 assay, prior to the alkaline electrophoresis step, cells in agarose were incubated with a 1:2 dilution of human OGG1 glycosylase for 1 h at 37°C to convert all oxidized bases into DNA breaks. Slides were analyzed on a Leica DMRA2 fluorescence microscope (ϫ100 objective). Comet parameters were quantified using the CometScore software (TriTek Corp). The parameter percentage (%) of DNA in tail corresponds to the amount of pixels in the comet tail (migrated DNA), whereas the Olive tail moment (Olive et al., (55)) corresponds to the product of the tail length and the fraction of DNA in the tail (intensity of the DNA in the tail).
DNA Oxidation Analysis-FACS-sorted LSK cells were incubated in the kit staining solution containing 1:20 dilution of a specific FITC-conjugated probe to 8-OHdG, in accordance with the manufacturer's instructions (OxyDNA assay kit, Calbiochem).
Cell Pellet and Lysate Preparation for BER Molecular Beacon Assay-Approximately 6 ϫ 10 6 lineage-negative (LinϪ) cells from WT and Foxo3 Ϫ/Ϫ mice (six mice per group) were collected and pelleted at 228 ϫ g for 5 min. The cell pellets were washed once with PBS and then flash-frozen and kept at Ϫ80°C. Whole cell lysates were prepared by a freeze-thaw method. Briefly, cells were resuspended in 150 l of the BER molecular beacon reaction buffer (HEPES 25 mM, pH 7.8, KCl 150 mM, EDTA 0.5 mM, glycerol 1%, DTT 0.5 mM, 1ϫ protease inhibitor (Pierce, catalog no. 539131)). The cell suspension was then frozen on dry ice for 5 min and then thawed in a 37°C water bath for 5 min followed by vortexing at the maximum speed for 30 s. Cells were frozen and thawed for three cycles and then centrifuged at 16,400 rpm for 5 min to remove cellular debris. The protein concentration of each cell lysate was measured. Lysates were diluted with BER molecular beacon reaction buffer to a final concentration of 0.4 mg/ml and immediately used for the activity assay as detailed below (39).
BER Molecular Beacon Assay and Data Analysis-OGG1mediated glycosylase activity was measured by a BER molecular beacon assay, essentially as we have described previously (39). Briefly, an OGG1 substrate (the 8-oxo-dG/A beacon, 5 l, 200 nM) was added to the whole cell lysate (20 l) that was prepared in BER molecular beacon reaction buffer (above). The reaction was performed at 37°C using a StepOnePlus QRT-PCR machine (Applied Biosystems). Fluorescence was measured for three technical replicates every 20 s for 1 h and normalized to the signal from the completely denatured beacon within each well, as in previously published methods (39).
Author Contributions-C. L. B. and S. G. conceived and designed the study, analyzed the data, and wrote the manuscript. C. L. B. performed the experiments with assistance from R. L. and P. R. J. L. performed and J. L. and R. W. S. designed and analyzed the experiments in Figs. 5G and 6 and contributed to the manuscript preparation.