Mutation in the Glucose-6-phosphate Dehydrogenase Gene Leads to Inactivation of Ku DNA End Binding during Oxidative Stress*

Glucose-6-phosphate dehydrogenase (G6PD), the rate-limiting enzyme of the oxidative pentose phosphate cycle, regulates the NADPH/NADP+ ratio in eukaryotic cells. G6PD deficiency is one of the most common mutations in humans and is known to cause health problems for hundreds of millions worldwide. Although it is known that decreased G6PD functionality can result in increased susceptibility to oxidative stress, the molecular targets of this stress are not known. Using a Chinese hamster ovary G6PD-null mutant, we previously demonstrated that exposure to a thiol-specific oxidant, hydroxyethyldisulfide, caused enhanced radiation sensitivity and an inability to repair DNA double strand breaks. We now demonstrate a molecular mechanism for these observations: the direct inhibition of DNA end binding activity of the Ku heterodimer, a DNA repair protein, by oxidation of its cysteine residues. Inhibition of Ku DNA end binding was found to be reversible by treatment of the nuclear extract with dithiothreitol, suggesting that the homeostatic regulation of reduced cysteine residues in Ku is a critical function of G6PD and the oxidative pentose cycle. In summary, we have discovered a new layer of DNA damage repair, that of the functional maintenance of repair proteins themselves. In view of the rapidly escalating number of roles ascribed to Ku, these results may have widespread ramifications.

Reactive oxygen species (ROS) 1 produced during oxidative stress are widely believed to cause cell death or late effects such as cancer, aging related diseases, etc. (1). Several studies have indicated that ROS would be even more damaging if it were not for the protective role of intracellular superoxide dismutase, catalase, and glutathione (2,3). Superoxide dismutase, catalase, and GSH are the most widely investigated modifiers of oxidative damage to cells (2,3). Glutathione is believed to be involved in the maintenance of intracellular redox by scaveng-ing ROS directly and/or reducing oxidatively modified macromolecules. NADPH, a major cofactor required for the maintenance of GSH during oxidative stress, is produced by the oxidative limb of the pentose phosphate cycle, i.e. OPPC (4 -7). Glucose-6-phosphate dehydrogenase (G6PD) is the rate-limiting enzyme for OPPC (4). At low levels of oxidative stress, G6PD activity can increase several hundredfold, producing NADPH necessary to reduce/repair the oxidized molecules (4,7). Severe oxidative stress has been shown to cause the inactivation of G6PD that correlated with oxidative damage (8 -10). Therefore, NADPH, the major cofactor produced by G6PD/ OPPC, is vital to cellular defense against oxidative stress.
Genetic defects in G6PD, an X-linked gene, are extremely common, occurring in more than 400 million people throughout the world (11)(12)(13). These G6PD genetic deficiencies result in an impaired ability to deal with oxidative stress and can cause various health problems such as cancer, aging related diseases, etc. (14 -16). The best known example for the immediate effect is hemolytic anemia caused by treatment of G6PD-deficient patients with antimalarial drugs (17). In addition, several studies using cells from G6PD-deficient patients have demonstrated an increased sensitivity of these cells to oxidants, toxins, and oxidative stress (17). The effects of these agents are presumed to be due to the interaction of ROS with macromolecular targets such as DNA, lipid, and protein, but the precise molecular targets are not well defined (17)(18)(19).
A recent study has demonstrated that the majority of patients with acute non-lymphocytic leukemia have decreased G6PD levels (16). This study also reported the association of a high percentage of chromosomal abnormalities, lower survival, and higher remission rates in patients having decreased G6PD activity versus normal, respectively. Although some studies have indicated that there was no difference in cancer risk among G6PD-deficient versus G6PD-normal subjects, this study showed a significant increase in mortality from non-Hodgkin's lymphoma in G6PD-deficient subjects (14). In a related animal study, Wells et al. (15) demonstrated that G6PDdeficient mice had enhanced embryopathies, indicating a teratological role for endogenous oxidative stress caused by the absence of G6PD. In their review, Martini and Ursini (20) have suggested that the high prevalence of G6PD deficiency in various ethnic populations may allow the putative role of G6PD deficiency in carcinogenesis and aging to be ascertained.
We have recently demonstrated that the Chinese hamster ovary (CHO) G6PD-null mutant has enhanced radiation sensitivity and an inability to repair DNA double strand breaks (DSBs) after oxidative stress induced by hydroxyethyldisulfide (HEDS), a thiol-specific oxidant (21). Here we demonstrate that a reasonable molecular mechanism for its defective DNA DSB repair is the inhibition of the DNA end binding activity of the Ku heterodimer, a DNA repair protein, due to the oxidation of its cysteine residues. Our reason for investigating the Ku protein is that it plays a central role in DNA damage recognition and non-homologous DNA end joining by binding to DNA ends in mammalian cells (22)(23)(24). Ku is also believed to be involved in the regulation of apoptosis and telomere fusion (25,26). Our present results demonstrate that G6PD deficiencies cause not only susceptibility to oxidative damage in general but also the specific oxidation of proteins critical for maintaining proper cellular functions, such as DNA repair.

EXPERIMENTAL PROCEDURES
G6PD-null Mutants and Cell Transfections-The cell lines were derived from the K1 clone of CHO cells as described previously (27). Briefly, G6PD-null mutants of CHO cells were isolated after mutagenesis with a minimally toxic concentration of ethylmethanesulfonate by means of the selection technique coupled with histochemical staining for enzyme activity. Cell transfections to reintroduce the G6PD gene were carried out by the electroporation of G6PD-null mutant cells with hamster G6PD-cDNA (GenBank TM accession number AF044676) inserted into pcDNA3.1 constructs and selected in G418 medium. Target clones were identified by histochemical staining for G6PD activity and Northern analysis using G6PD-cDNA (25). The A1A cells are stably transfected with a vector resistant to G418. We have also selected several clones with differing and similar G6PD activity to normal cells (27).
HEDS Treatment and Preparation of Nuclear Extract-Cells (0.8 ϫ 10 6 ) in 4 ml of medium were plated in 60-mm dishes on the day before treatment with HEDS. HEDS exposure was effected by replenishing the cells with 1 ml of fresh medium plus 50 l of 0.1 M HEDS. The dishes were swirled every 5 min during the 1-h incubation time at 37°C in a humidified 5% CO 2 incubator as described previously (21). Cells looked morphologically normal after HEDS treatment. The cells were harvested using a Teflon cell scraper after rinsing three times with 3 ml of ice-cold PBS. They were then centrifuged, resuspended, and washed twice with 4 ml of ice-cold PBS. The washed cells were resuspended gently by flicking the sample tubes in 400 l of cold hypotonic lysis buffer (pH 7.9) containing 10 mM Hepes, 1.5 mM MgCl 2 , 10 mM KCl, leupeptin (2.5 g/ml), 0.2 mM PMSF, and pepstatin (1 g/ml) and incubated on ice for 10 min. Cells swollen by this procedure were vortexed (10 s) and then centrifuged at 14,000 rpm in a microcentrifuge for 10 s. The nuclear pellet was resuspended in 30 l of high salt extraction buffer (pH 7.9) containing 20 mM Hepes, 25% glycerol, 420 mM NaCl, 1.5 mM MgCl 2 , 0.2 mM EDTA, leupeptin (2.5 g/ml), 0.2 mM PMSF, and pepstatin (1 g/ml) and incubated on ice for 20 min. Cellular debris was removed by centrifugation at 14,000 rpm for 2 min at 4°C.
Estimation of Intracellular Reductants NADPH and NADH by High Performance Liquid Chromatography (HPLC)-Our method involves an extraction of pyridine nucleotides by 70% methanol directly added to the cells after two rinses with ice-cold PBS. Reduced pyridine nucleotides were quantified by a modified method of Noack et al. (28) using an HPLC with fluorescence detection (Waters 474) at an excitation and emission wavelength of 340 and 455 nm, respectively. NADPH and NADH were eluted with 0.1 M KH 2 PO4 (pH 6.1) using a step gradient of methanol; 5 min, 0% CH 3 OH; 6 min, 4% CH 3 OH; 5 min, 12% CH 3 OH; 14 min, 40% CH 3 OH. The data were collected and analyzed using Jasco Borwin software.
Quantitation of Intracellular and Extracellular Thiols by HPLC-Concentrations of the intracellular and extracellular thiols, glutathione, and mercaptoethanol were quantified by an HPLC with electrochemical detection (21). The mobile phase was 100 mM phosphate (pH 2.0) with 15% methanol in a reversed phase C-18 column (Alltech Altima). To quantify extracellular thiols, 0.5 ml of the extracellular medium from the dish was transferred to a microcentrifuge tube containing an equal volume of sulfosalicyclic acid (SSA) lysis buffer (100 mM SSA, 0.1 mM diethylenetriaminepentaacetic acid, 0.1 mM diethyldithiocarbamic acid, and 0.1 mM EDTA). After decanting the remaining medium, the cells were rinsed three times with ice-cold PBS. To quantify intracellular thiols, attached cells were lysed with 0.5 ml of ice-cold SSA lysis buffer and 0.5 ml of water. After 15 min on ice, the cells were scraped with a Teflon spatula. Both the extracellular and intracellular extracts were transferred to microcentrifuge tubes and centrifuged at high speed in a Fisher 59A microcentrifuge followed by analysis of the supernatant.
Estimation of Intracellular Protein Thiols-Cells attached to the dishes were incubated with HEDS for the desired time. These cells were then washed three times with ice-cold PBS and then treated with SSA as above. This treatment precipitates cellular macromolecules in place without loss of protein. The acid was removed, and dishes were washed two more times with 5 ml of SSA, which was removed by aspiration. Under these experimental conditions, only non-protein sulfhydryl groups were washed off the cells. Acid-fixed cells were then incubated with 1 ml of 100 mM phosphate buffer (pH 7.4) containing 1.5 mM 5,5Ј-dithiobis(2-nitrobenzoic acid) for 15 min at 37°C, and the absorbance was read at 412 nm. Protein thiols were then estimated using an absorption coefficient for reduced 5,5Ј-dithiobis(2-nitrobenzoic acid) of 13,600 at 412 nm. The protein thiol level in untreated cells was 0.040 Ϯ 0.002 pmol/cell. DNA Probe Preparation-A 144-base pair probe was prepared from a pUC18 plasmid using PvuII and EcoRI digestion and purified by 10% preparative gel electrophoresis (29). The appropriate 144-base pair fragment was excised and eluted from the gel into 100 mM NaCl in TE buffer (10 mM Tris (pH 8.0), 0.1 mM EDTA). The eluted DNA was extracted with phenol/chloroform/isoamyl alcohol (25:24:1), ethanolprecipitated, and resuspended in TE buffer. The 144-bp fragment was labeled with [␣-32 P]dATP (PerkinElmer Life Sciences) using the Klenow fragment of DNA polymerase I. Unincorporated nucleotide was removed by chromatography on Sephadex G-50 spin columns.
DNA Mobility Shift and Antibody Supershift Assays-Mobility shift mixtures containing 2 g of nuclear protein, 1 ng of 32 P-labeled 144-bp probe, and 1 g of pUC18 (nonspecific DNA competitor) in a final volume of 20 l of binding buffer (10 mM Tris-HCl (pH 8.0), 0.1 mM EDTA, 150 mM NaCl, 1 mM PMSF, leupeptin (2.5 g/ml) and pepstatin (1 g/ml), and 10% v/v glycerol) were incubated for 15 min at room temperature (29). For mobility supershift assays, a Ku70 antibody (29) was incubated with the nuclear extract for 30 min at the desired concentration after the 15-min incubation with the radioactive probe and pUC18. The mixtures were electrophoresed on a 6% polyacrylamide gel at 20 -25 mA in TBE buffer (45 mM Tris-HCl (pH 8.0), 45 mM boric acid, 1 mM EDTA), and the gel was dried and subjected to autoradiography or phosphorimaging and NIH image analysis for quantitation (Bio-Rad).
Western Blot Analysis-A total cellular Ku70 protein was quantified using Western blot and NIH image analysis. Cells were rinsed with PBS three times and mixed with 200 l of room temperature-cooled, preboiled (2 min) Laemmli buffer containing 10 mM dithiothreitol (DTT). Cells in Laemmli buffer were removed using a Teflon scraper and transferred to an Eppendorf tube using a micropipette. The cells were homogenized using a 1-ml syringe (18-gauge needle), and samples were stored at Ϫ80°C. Protein extracts (15 g/ml) in sample buffer were electrophoresed on a 7.5% precast gel from Bio-Rad at 150 V for 75 min in TBE buffer (45 mM Tris-HCl (pH 7.8), 45 mM boric acid, 1 mM EDTA). The proteins were transferred to nitrocellulose by electrophoresis at 100 V for 60 min. The nitrocellulose paper was incubated in 10 ml of blocking buffer for 1.5 h at room temperature on a rocker and stored in the cold room overnight. The nitrocellulose paper was washed five times with TBST (20 mM Tris base (pH 7.6), 137 mM NaCl, 0.05% Tween 20) and incubated with 38 l of primary Ku (p70) antibody (Ab-4, clone N3H10, Neomarkers) in 15 ml of blocking buffer for 2 h at room temperature. The nitrocellulose paper was washed 4 times with TBST before incubation with the secondary antibody for 1 h per the manufacturer's instruction (Amersham Biosciences, Inc.). The bands were detected and quantified using a standard ECL kit and NIH image analysis, respectively.
To measure bioreductive capacity, we quantified the extracellular mercaptoethanol produced by cells treated with 5 mM HEDS immediately after a 1-h incubation (Fig. 3). The wildtype K1 cells released 522 nmol of ME into the extracellular medium. G6PD-deficient E89 cells released 53 nmol of ME, which is a 10-fold decrease in bioreductive capacity compared with K1 cells. The G6PD-transfected A1A cells released 815 nmol of ME, indicating the recovery of bioreductive capacity due to the transfection of the wild-type G6PD gene into the null mutant cells. The intracellular GSH level was almost the same in untreated wild-type (13 nmol), null mutant (18 nmol), and transfectant (16 nmol) cells (Fig. 4). The incubation of HEDS for 1 h did not significantly affect the GSH levels in wild-type (12 nmol) or transfectant (17 nmol) cells. However, the GSH levels in G6PD-null mutant cells decreased from 18 to 2.0 nmol (9-fold decrease) after a 1-h incubation with HEDS. The effect of HEDS on protein thiols was measured by using Ellman's reagent (Fig.  5). All three untreated cell lines had similar amounts of intracellular PSH (ϳ42 nmol/10 6 cells). G6PD-null mutants treated with 5 mM HEDS showed a significant decrease (30%) in PSH compared with untreated E89 cells. However, wild-type K1 and A1A transfectants showed an insignificant decrease (7%) in protein thiols after HEDS treatment (Fig. 5).
To test whether cellular redox changes caused by HEDS treatment (Figs. [1][2][3][4][5] affect Ku binding to DNA ends, we used a modified method of the standard mobility gel shift procedure (29). This assay contains a 32 P-labeled 144-bp DNA probe, a nuclear extract, and an excess of circular plasmid DNA. Under these assay conditions, proteins that bind to internal DNA sequences will be bound by the excess unlabeled plasmid DNA that does not contain ends. Thus, only proteins that specifically bind to DNA ends will bind to the labeled probe. DTT, a strong reducing agent, is normally added to protein extracts to protect against spontaneous oxidation. To measure the effect of oxidation on end binding activity in vivo, it was necessary to determine whether DTT could be eliminated from the assay because its presence would be likely to reverse such oxidation. Therefore, we compared DNA end binding activity with and without DTT for nuclear proteins extracted from wild-type (K1) cells, G6PD-null mutant (E89) cells, and G6PD-null mutant cells stably transfected with a cDNA encoding the wild-type hamster G6PD (A1A). The NIH image analysis of labeled bands (Fig. 6,  plots 2-7) shows that similar amounts of DNA end binding activity (Ku-DNA complex) were observed in nuclear extracts isolated from cells in the presence or absence of DTT. Authentication of Ku as the DNA end binding species was made by demonstrating a supershift of the bound probe upon addition of the anti-Ku70 antibody to the reaction mixture (Fig. 6, lanes  8 -10).
Having demonstrated that Ku binding was not affected by the absence of DTT in this assay, we examined the effect of HEDS treatment on Ku binding in the same three cell lines. HEDS treatment inhibited Ku DNA end binding activity by 70% in the G6PD-null mutant E89 cells (Fig. 7, lane 5). In contrast, HEDS had no effect on Ku DNA end binding activity in the wild-type cell line K1 or the G6PD ϩ transfectant A1A (Fig. 7, lanes 3 and 7).
To ensure that the inhibition of Ku DNA end binding observed in HEDS-treated cells was not due to a decrease in the Ku protein, we measured the amount of Ku70 by Western blot analysis using a monoclonal anti-Ku70 antibody. A single ϳ70-kDa immunoreactive band of equal intensity was observed in total cell lysates extracted from all three cell lines treated with and without HEDS (Fig. 8). This indicates that treatment with 5 mM HEDS had no measurable effect on the amount of Ku70 in any of the CHO cell lines. These results support the hypothesis that HEDS-induced oxidation of Ku leading to reduced end binding activity was responsible for the previously observed increase in radiation sensitivity and reduction in DSB repair in G6PD-null mutant cells (21).
To determine whether the effect of HEDS on Ku extracted from G6PD-null mutant cells was a direct consequence of thiol oxidation, we treated the nuclear extract with the thiol-specific reductant, DTT. Ku DNA end binding activity was fully restored by treating the lysates from HEDS-treated G6PD-null mutant cells with 5 mM DTT for 10 min (Fig. 9, lane 5). Ku DNA end binding activity in lysates from G6PD-null mutant cells that had not been exposed to HEDS was not affected by DTT. Taken together, the results indicate that HEDS blocks Ku DNA end binding activity in vivo through the oxidation of cysteine residues in the Ku heterodimer without affecting the protein level because this oxidation can be reversed by reduction of cysteine residues by DTT. The fact that Ku was not affected by HEDS treatment in G6PD-proficient cells suggests that homeostatic regulation of cysteine residues in Ku is a critical function of G6PD and the oxidative pentose cycle. DISCUSSION A major biochemical function of G6PD is that it acts as the rate-limiting enzyme of the oxidative pentose phosphate cycle. The enzymes of the OPPC are highly conserved and ubiquitously expressed in prokaryotic and eukaryotic cells, indicating the importance of this pathway for cell growth and function (5, 8 -10). The oxidative limb of the pentose cycle catalyzes the oxidation of G6P to NADPH, ribulose-5-phosphate, and CO 2 (7). NADPH is utilized in the reductive biosynthesis of fatty acids, cholesterol, steroids, and the prenyl groups of mem-  2, 5, and 8), G6PD-null mutant E89 (lanes 3, 6, and 9), and G6PD-transfected A1A (lanes 4, 7, and 10) cells in the absence (lanes 2-7) and presence (lanes 8, 9, and 10 brane-bound proteins (30). NADPH is also utilized in the reduction of nucleotide diphosphates to deoxyribonucleic acids and is therefore necessary for de novo synthesis of DNA (30).
G6PD-deficient and null cells remain viable because of the availability of NADPH produced by pathways such as malic enzyme, isocitrate dehydrogenase, and NADH/NADPH transhydrogenase (30). In addition, the ribose necessary for DNA and RNA synthesis is available from the transketolase and transaldolase enzymes of the nonoxidative limb of the pentose cycle (31,32). Furthermore, many of the nucleotides can also be scavenged from the matrix (33). Although cells can survive in the absence of G6PD by using reducing equivalents produced by G6PD-independent pathways, G6PD becomes more important for cell survival during oxidative stress. In wild-type cells, the oxidative limb of the pentose cycle is activated by oxidative stress, producing the reducing equivalents to maintain the redox status of the cells (7). As the capability of the protective function declines, cells become progressively susceptible to oxidative stress (10,34,35). We showed that cells deficient in G6PD were sensitive to a variety of toxins thought to act by induction of oxidative damage (27,34,36). Cells deficient in G6PD become progressively susceptible to radiation-induced apoptosis (27). In addition, a slightly slower DNA repair kinetics was observed for G6PD-deficient cells compared with wild-type cells after radiation damage (21). Our unpublished results 2 indicated that malic enzyme and isocitrate dehydrogenase were not modified in these cell lines. In view of the sensitivity of G6PD-null cells to oxidative stress, this suggests that G6PD plays the major role in the cellular response to oxidative damage.
One type of molecular damage that can be expected from G6PD deficiency during oxidative stress is the oxidation of protein thiols because G6PD-deficient cells lack production of reducing equivalents necessary to maintain protein cysteine residues in their reduced form. Cysteine residues in proteins are important for the structure and functions of proteins (37,38). Although it is clear that a change in the normal redox status of protein cysteine thiols can alter function, protein redox is well regulated except when cells are subjected to oxidative stress or potentially toxic oxidative agents such as hydrogen peroxide, menadione, etc. (39 -43). However, these agents are so nonspecific that it is impossible to determine whether protein thiol oxidation plays a major role in the cytotoxic response. Indeed, it has not been possible to identify the critical targets of damage (e.g. membrane, mitochondria, DNA, protein) using these agents (44). Part of the reason for such nonspecificity is that cells massively resist oxidative attack, utilizing the reducing capacity provided by the oxidative limb of the pentose cycle (NADPH) or respiration (NADH) (34,45,46). Therefore, the amount of oxidant required must be strong enough to overwhelm the protective defenses of the cell, thereby causing massive damage to multiple cellular targets (41,46).
To better define the molecular nature of thiol-specific oxidative damage in G6PD-null mutant cells, we used the disulfide HEDS, which is relatively non-reactive with any molecule other than thiols. The fast bioreduction of HEDS (ϳ500 nmol of ME per million cells in 1 h) by wild-type cells suggests that HEDS may be reduced by intracellular reductants to produce ME as shown in Reaction 1.
2 XH ϩ RSSR (HEDS) 7 XH ϩ RSH ϩ RSSX 7 2 RSH (ME) ϩ XSSX REACTION 1 Although the G6PD-null mutant cells have almost the same initial amounts of GSH, PSH, NADPH, and NADH, the conversion of HEDS into ME is much lower than that of G6PD normal cells. HEDS may be reduced by any one or all of the intracellular reductants such as GSH, cysteine, and even protein thiols either directly or through enzymatic reduction using NADPH and NADH as cofactors. The loss of NADPH in both G6PD-null mutant and control cells indicates that HEDS consumes NADPH (Fig. 1). This suggests that cells require OPPC to produce enough reducing equivalents (NADPH) to reduce 5 mM HEDS. A significant amount of NADPH remaining in wildtype cells during HEDS treatment suggests that G6PD maintains NADPH through recycling of NADP ϩ by the oxidative pentose phosphate cycle, which is shown in Reaction 2. This is consistent with the bioreductive capacity of E89 cells, which is 6-and 12-fold lower than that of the K1 and A1A cells, respectively.
The ability of wild-type and transfectant cells to maintain normal levels of GSH, PSH, and NADH despite a 10-fold reduction in steady state NADPH attests to the adaptability of this pathway and cellular bioreductive capacity. However, for the G6PD-null mutant cells, the near total loss of NADPH results in a cascade of other oxidations, including the loss of NADH, GSH, and PSH (Figs. 1-5).
Our results showed that of 54 nmol of ME produced by E89 cells, only 28 nmol is accounted for by GSH (16 nmol), PSH (12 nmol), G6PD-independent NADPH (0.06nmol), and NADH (0.6 nmol). The remaining 26 nmol of reductants involved in the reduction of HEDS in E89 cells may have come from NADH or G6PD-independent NADPH cycling. It has been shown that NADH produced by respiration can reduce thiols, albeit at a much slower rate compared with NADPH (47).
The simplest explanation for the loss of GSH, NADPH, and NADH in G6PD-null mutant cells would be that HEDS may oxidize GSH and NAD(P)H independently, either by chemical or enzymatic reduction. However, it is more likely that all of these oxidations are highly interdependent. It has been demonstrated that the reduction of HEDS is facilitated by thiol transferase-linked reactions (48 -50). The thioltransferase enzyme utilizes GSH in the reduction of disulfides to stimulate the second step of Reaction 3 (51).
We know from our studies and those of others that NADPH is required to recycle GSSG back to GSH, as shown in Reaction 4. GSSG ϩ NADPH ϩ H ϩ 7 2 GSH ϩ NADP ϩ REACTION 4 GSSG produced as shown in Reaction 3 may in turn oxidize protein thiol, as shown in Reaction 5.
The complexity of these reactions warrants considerable additional study to identify the enzymes that use NADPH as cofac-tors to reduce oxidized PSH.
We have previously reported that DNA DSB rejoining is inhibited by HEDS in G6PD-null mutant cells (21). Because G6PD-null mutant cells have a higher level of oxidized protein thiols after HEDS treatment, it is possible that the inhibition of DNA DSB rejoining may have resulted from the oxidation of cysteine residues in one or more of the proteins involved in DSB repair. Biochemical and genetic studies using null mutant rodent cell lines sensitive to ionizing radiation have identified at least four genes, XRCC4, XRCC5, XRCC6, and XRCC7, that are required for non-homologous end joining (NHEJ) (22)(23)(24). XRCC4 encodes a 38-kDa nuclear phosphoprotein that binds strongly to DNA ligase IV (23,24). XRCC5, XRCC6, and XRCC7 are components of a DNA-dependent protein kinase (DNA-PK). XRCC5 and XRCC6 encode 86-and 70-kDa subunits of Ku autoantigen, a DNA end-binding protein and regulatory subunit of DNA-PK (22)(23)(24). XRCC7 encodes a 460-kDa catalytic subunit (DNA-PKcs) of DNA-PK (22)(23)(24). The protein kinase activity of DNA-PK is believed to be stimulated by its association with DNA end-bound Ku (22)(23)(24). A Ku heterodimer binds to the ends of DNA in a non-sequence-dependent manner (22)(23)(24). The DNA binding activity of Ku requires reduced sulfhydryl groups in cell-free systems (52). Although Ku contains 14 cysteine residues, 5 of these are located in Ku70, which is in contact with the DNA binding site, suggesting that the function of Ku may be regulated through oxidationreduction of at least 5 of its 14 cysteine residues (52). Here we studied the role of OPPC in redox control of Ku protein in vivo because Ku has multiple functions in mammalian cells (22). In particular, Ku is essential for NHEJ by DNA-PKcs, which is required in V(D)J recombination, class switch recombination, and DNA DSB repair (22)(23)(24).
Our results show that Ku is susceptible to oxidative damage and must be protected by reducing equivalents produced by the pentose cycle. The reversibility of Ku function in the nuclear extract of HEDS-treated G6PD-null mutant cells by DTT directly implicates Ku thiol oxidation. We deliberately treated the nuclear extract with DTT and not the cells because addition of DTT at such a high concentration to the cells during or after HEDS treatment might have additional effects that could complicate the interpretation of the data. For example, DTT at such a high concentration may directly eliminate the HEDS effect by direct chemical reaction with HEDS rather than reduction of oxidized Ku protein. Our data demonstrate for the first time that lower concentrations of reductants such as NADPH may be sufficient to reduce oxidized Ku in cells via recycling of NADP. Although G6PD is responsible for providing the essential cofactor NADPH, the enzyme(s) responsible for the direct reduction of Ku is unknown. An enzyme in the thioredoxin family is a reasonable possibility (37). It is highly likely that the functions of other proteins may also be affected in G6PD-deficient cells during oxidative stress. To determine whether the effect of HEDS is specific for the Ku-mediated DNA repair pathway, we compared the effects of glucose depletion on wild-type K1-and Ku-deficient XRS5 cells. As expected, our preliminary results (data not shown) demonstrated that depletion of glucose, a substrate for the OPPC, resulted in an 80% decrease in bioreduction, an 80% depletion in GSH, and 30% protein thiol oxidation in XRS5 and K1 cells exposed to HEDS. Under these conditions, HEDS radiosensitized the K1 cells with little effect on XRS5 cells (data not shown), suggesting that of the two DNA DSB repair pathways (homologous recombination and non-homologous end joining), Ku-mediated NHEJ is more specifically affected by oxidative stress. This could also mean that Ku-mediated NHEJ is the only or major pathway involved in the repair of DNA DSB induced by ␥ radiation; no further sensitization occurred in Ku-deficient cells by HEDS even though 30% of the total protein thiols were oxidized.