The cockayne syndrome group B gene product is involved in cellular repair of 8-hydroxyadenine in DNA.

Cockayne syndrome (CS) is a human disease characterized by sensitivity to sunlight, severe neurological abnormalities, and accelerated aging. CS has two complementation groups, CS-A and CS-B. The CSB gene encodes the CSB protein with 1493 amino acids. We previously reported that the CSB protein is involved in cellular repair of 8-hydroxyguanine, an abundant lesion in oxidatively damaged DNA and that the putative helicase motif V/VI of the CSB may play a role in this process. The present study investigated the role of the CSB protein in cellular repair of 8-hydroxyadenine (8-OH-Ade), another abundant lesion in oxidatively damaged DNA. Extracts of CS-B-null cells and mutant cells with site-directed mutation in the motif VI of the putative helicase domain incised 8-hydroxyadenine in vitro less efficiently than wild type cells. Furthermore, CS-B-null and motif VI mutant cells accumulated more 8-hydroxyadenine in their genomic DNA than wild type cells after exposure to gamma-radiation at doses of 2 or 5 Gy. These results suggest that the CSB protein contributes to cellular repair of 8-OH-Ade and that the motif VI of the putative helicase domain of CSB is required for this activity.

Cockayne syndrome (CS) 1 is a human genetic disorder with diverse clinical symptoms (1). There are two complementation groups of CS, CS-A and CS-B. The CSB gene (CSB) encodes the CSB protein (CSB) with 1493 amino acids and a molecular mass of 168 kDa (2,3). An important clinical feature of CS-B is accelerated aging. Thus, CS-B may be a useful model for studying the aging process in humans. Fibroblasts from CS-B patients are hypersensitive to ultraviolet (UV) radiation and show delayed recovery of RNA synthesis after exposure (1). CS-B cells have a pronounced defect in repair of UV radiationinduced DNA damage in actively transcribed genes (4). UV radiation-induced DNA damages are mostly bulky helix-distorting lesions that are removed by nucleotide-excision repair (NER). The transcription-coupled repair (TCR) pathway specifically removes these lesions in actively transcribed genes (5). CS-B cells are deficient in TCR, and this is thought to be the molecular basis of some features of the CS phenotype (6), especially the increased sensitivity to UV radiation. However, it is unlikely that the TCR defect and the resulting UV radiation sensitivity of CS-B cells account for other clinical features of CS such as progressive neurodegeneration. Thus, CS-B cells may be deficient in other processes besides TCR of UV radiation-induced DNA damage.
All organisms are continuously exposed to reactive oxygen species, which cause cellular damage and are thought to contribute to the aging process (7). Reactive oxygen species damage DNA and generate numerous DNA lesions such as modified bases and sugar moieties, strand breaks, and DNA-protein cross-links (reviewed in Ref. 8). 8-hydroxy-7,8-dihydroguanine (8-OH-Gua) is a ubiquitous lesion in oxidatively damaged DNA that is mutagenic, causing G 3 T transversions in vitro and in vivo (9,10). 8-hydroxy-7,8-dihydroadenine (8-OH-Ade) is another common lesion formed in DNA exposed to oxidizing agents such as hydroxyl radicals in vitro and in vivo (reviewed in Ref. 8 and structure is shown in Fig. 1). 8-OH-Ade is premutagenic and induces A 3 G and A 3 C mutations in mammalian cells, although it leads to a lower mutation frequency than 8-OH-Gua (11)(12)(13)(14)(15)(16). Nevertheless, the biological consequences of 8-OH-Ade could be significant, because it is abundant in oxidatively damaged DNA (17). 8-OH-Ade is repaired in mammalian cells (18), but the mechanism by which it is repaired is not known (19).
Studies suggest that oxidative damage to DNA may contribute to the premature aging phenotype associated with progeroid syndromes (20). Thus, it is possible that the DNA repair defect in CS plays an important role in the course and symptoms of this disorder. Whole cell extracts (WCEs) from primary CS-B cells incise 8-OH-Gua at a lower rate than normal cell lines, and this deficiency is complemented by transfection of the cells with wild type CSB (21). In addition, CS-B-null cells and CS-B motif V/VI mutants are more sensitive to ␥-radiation than wild type cells (22). WCEs from transfected CS-B-null cells and motif V/VI mutants also incise 8-OH-Gua-containing oligonucleotides less efficiently than wild type cells. Furthermore, 8-OH-Gua accumulates to a greater level in CS-B-null cells and motif VI mutants following ␥-irradiation than in wild type cells (22).
In this study, we investigated the role of CSB in cellular repair of 8-OH-Ade. CSB mutants were expressed in CS-Bnull cells and in vitro incision of 8-OH-Ade was analyzed using WCEs from these cell lines. In addition, the nucleoside form of 8-OH-Ade, 8-hydroxy-2Ј-deoxyadenosine (8-OH-dAdo) was measured in genomic DNA of these cells by means of a recently developed assay that uses liquid chromatography/ mass spectrometry (LC/MS) with the isotope dilution technique (17).

EXPERIMENTAL PROCEDURES
Cell Lines and Culture Conditions-Cell lines were derived from CS1AN.S3.G2, a SV40-transformed human fibroblast in which CSB is disrupted. These cells were described previously (23). CS1AN.S3.G2 was transfected with pcDNA3.1 carrying the wild type CSB, mutant CSB altered in the putative helicase motif VI (Q942E and R946A) or the vector pcDNA3.1 (pc3.1) (Invitrogen) (Fig. 2). The reason we applied Q942E and R946A is that these cell lines are the most deficient ones in repair of 8-OH-Gua among the 8 stably transfected cell lines established in this laboratory with site-directed mutation(s), which are distributed in various motifs of the helicase domain (22). Construction of the mutants and cell lines were described previously (22,24).
Exposure was carried out as follows. Cells attached on 10-cm 2 dishes were washed with phosphate-buffered saline and then irradiated at the indicated doses with Gammacell 40 Exactor 137 Cs ␥-source (Nordian International Inc., Kanata, Ontario, Canada) followed by incubation in the complete medium at 37°C for 30 min. Genomic DNA was extracted as described (22).
In Vitro Incision of an 8-OH-Ade-containing Oligonucleotide-An oligonucleotide with a single 8-OH-Ade at position 11 was the substrate for the incision assay. The sequences of the oligonucleotide used in the incision assay are listed in Table I. The oligonucleotide with the lesion was 32 P-5Ј-end-labeled and annealed with a complementary strand containing a T opposite 8-OH-Ade. We also used a double-stranded substrate containing C opposite the lesion because this structure is the optimal substrate for nick-forming activity at position of 8-OH-Ade (25,26). Incision reactions were carried out in a volume of 20 l containing 100 fmol of oligonucleotide duplex, 1 g of poly(dIdC) competitor, 20 mM Hepes-KOH, pH 7.8, 100 mM KCl, 5 mM dithiothreitol, 5 mM EDTA, 2 mM MgCl 2 , and 40 g of WCE protein prepared as described (22,27). Reactions were incubated at 37°C for 3 h, terminated by the addition of 0.8 l of 10% SDS and 0.8 l of 5 mg/ml proteinase K, and incubated for 10 min at 55°C. DNA was precipitated with 2 l of 5 mg/ml glycogen (Ambion), 4 l of 11 M ammonium acetate, and 70 l of cold ethanol overnight at Ϫ20°C. Samples were centrifuged for 1 h at 4°C, washed with 200 l of 70% ethanol, and collected by centrifugation at 12,000 ϫ g for 10 min. The pellet was dried in a speed-vacuum and resuspended in 10 l of formamide-loading dye (5% EDTA, 0.02% bromphenol blue, 0.02% xylene cyanol in 95% formamide). Samples were separated on a denaturing 20% polyacrylamide gel (containing 7 M urea, 89 mM Tris borate, pH 8.0, and 2 mM EDTA). The reaction products were visualized by autoradiography and quantified on a PhosphorImager (Molecular Dynamics). Oligonucleotides containing a single uracil or 5-hydroxycytosine, which were known to have no differences in incision activity of our CS-B wild type, null, or mutant cell lines, were used as the WCE control (22).

LC/MS and Preparation of the Stable Isotope-labeled Analog of 8-OH-dAdo-
The measurement of 8-OH-dAdo in enzymic hydrolysates of DNA samples was performed by liquid chromatography/isotope-dilution mass spectrometry (LC/IDMS) as previously described except for the stable isotope-labeled internal standard, which was used for quantification by IDMS (17). In the present work, a stable isotope-labeled analog of 8-OH-dAdo was prepared and isolated in pure form. Commercially available [ 15 N 5 ]dATP (Medical Isotopes, Inc. Pelham, NH) was used as the starting material. An aqueous solution of [ 15 N 5 ]dATP (5 mg/100 ml) was bubbled with N 2 O and exposed to ionizing radiation in a 60 Co ␥-source at a dose of 400 Gy (30 Gy/min). This treatment was expected to produce 8-OH-[ 15 N 5 ]dATP from [ 15 N 5 ]dATP among other products, because it is well known that the exposure of aqueous solutions of dAdo to ionizing radiation generates 8-OH-dAdo (reviewed in Refs. 28  The solvents and the elution gradient were as previously described (17), except that a flow rate of 2 ml/min was used. The column was kept at room temperature. Under the experimental conditions used, 8-OH-[ 15 N 5 ]dAdo eluting at 18.3 min was completely separated from [ 15 N 5 ]dAdo, which eluted at 17.3 min. This is the same elution order previously described for the unlabeled analogues of these compounds using an analytical LC column (17). The fractions corresponding to 8-OH-[ 15 N 5 ]dAdo were collected. At least 30 injections of 100 l were performed. Collected fractions were combined, dried in a SpeedVac under vacuum, and then dissolved in 200 l of water. The absorption spectrum of the solution was recorded between the wavelengths of 210 and 350 nm. The spectrum was identical to the absorption spectrum of authentic 8-OH-dAdo (30).
Analytical LC/MS was carried out to confirm the identity of 8-OH-[ 15 N 5 ]dAdo and its purity. The isolated compound was pure and did not contain any detectable unlabeled 8-OH-dAdo. The elution time of 8-OH-[ 15 N 5 ]dAdo was the same as that of 8-OH-dAdo, and its mass spectrum was similar to that of 8-OH-dAdo (17). As expected, however, the masses of the typical ions of 8-OH-[ 15 N 5 ]dAdo were shifted by 5 Da to greater masses, i.e. m/z 157 (the protonated base ion (BH 2 ϩ )), 273 (the protonated molecular ion (MH ϩ )), and 295 (the sodium adduct ion (MNa ϩ )). The concentration of the solution of 8-OH-[ 15 N 5 ]dAdo was determined by UV spectrophotometer using the absorption coefficient of 12764 M Ϫ1 cm Ϫ1 at 270 nm (30)  Statistics-Groups were compared using one-way analysis of variance tests. DuncanЈs multiple range test was used for post-hoc comparison of means. Differences were considered significant when p Ͻ 0.05.

RESULTS
In this study, we investigated the possibility that CSB is involved in cellular repair of 8-OH-Ade, a major lesion in oxidatively damaged DNA. 8-OH-Ade repair was assessed and compared in wild type, CS-B-null, and putative CS-B helicase motif VI mutant cells. Glycosylase/apurinic lyase activity in CS-B cell lines was quantified by measuring incision of an oligonucleotide with a single 8-OH-Ade residue. Fig. 3A shows

FIG. 2. Structure of CSB and the location of designed mutants.
The protein contains the seven conserved helicase motifs, a highly acidic region, two nuclear localization signals (NLS), and a nucleotide binding domain (NTB). CSBQ942E is constructed by replacing a glutamine residue highly conserved in the SNF2 family with a negatively charged glutamic acid. CSBR946A is constructed by replacing an arginine residue highly conserved in the SNF2 family with alanine. the incision activity of wild type, CS-B-null, and two CS-B mutant cell lines, and the results are summarized in Fig. 3B. Fig. 4, A and B show similar results, the only difference being that we used the substrate with C opposite the 8-OH-Ade lesion to possibly generate maximal nick-forming activity (25,26). CS-B-null and mutant cell lines incise 8-OH-Ade less efficiently than wild type cells in both situations. The activity of CS-B-null cells was ϳ3-fold lower than wild type cells (p Ͻ 0.05), and the activity of motif VI mutants (CSBQ942E and CSBR946A) was ϳ2-fold lower than wild type cells (p Ͻ 0.05). There were no differences in uracil or 5-hydroxycytocine incision of WCE of the tested cell lines (data not shown) (22).
The role of CSB in repairing 8-OH-Ade was also examined by measuring the level of 8-OH-Ade in DNA of cells exposed to ␥-radiation at doses of 2 or 5 Gy. 8-OH-Ade was measured as its nucleoside 8-OH-dAdo in wild type, CS-B-null, motif VI mutant cells using LC/IDMS, a recently developed assay for identification and quantification of oxidatively modified DNA nucleosides (17). DNA samples isolated from cells were hydrolyzed to nucleosides by endo-and exonucleases. Prior to hydrolysis, an aliquot of 8-OH-[ 15 N 5 ]dAdo was added as internal standard to the DNA samples. LC/IDMS was carried out in SIM mode to monitor the characteristic ions of 8-OH-dAdo and 8-OH-[ 15 N 5 ]dAdo at the appropriate retention time period, when these compounds eluted. The BH 2 ϩ , MH ϩ , and MNa ϩ ions of both compounds were simultaneously recorded. As expected, no difference between the retention times of these analogues was observed.   pair of this lesion. The time of complete repair within 30 min is in agreement with the recently reported repair kinetics of 8-OH-Ade in human cells (18). In contrast, significantly greater levels of 8-OH-dAdo were observed in ␥-irradiated CS-B-null and motif VI mutant cells. In motif VI mutants, exposure to ␥-radiation at 5 Gy followed by a 30-min incubation resulted in a higher level of 8-OH-dAdo than exposure to ␥-radiation at 2 Gy. These results demonstrate that 8-OH-dAdo accumulates in a dose-dependent manner in irradiated mutant cells but does not accumulate in irradiated wild type cells. DISCUSSION Previous studies have shown that mutations in CSB cause a deficiency in cellular repair of 8-OH-Gua (22). The present study shows that CS-B-null and motif VI mutant cells are also deficient in incision of 8-OH-Ade. Consistent with this observation, 8-OH-Ade accumulates more in genomic DNA of CS-Bnull and motif VI mutant cells than in wild type cells following exposure of cells to ␥-radiation at low doses. These results suggest that CS-B-null and motif VI mutant cells are deficient in cellular repair of 8-OH-Ade. This in turn indicates that CSB plays an important role in cellular repair of 8-OH-Ade and that the putative helicase motif VI of CSB is important for this DNA repair function.
CSB is highly homologous to proteins of the SWI/SNF2 family (2,3,31). SWI/SNF proteins participate in a wide variety of cellular functions including DNA repair, regulation of transcription, maintenance of chromosome stability, and chromatin remodeling (5,(31)(32)(33). SWI/SNF2 proteins contain seven highly conserved motifs for DNA or RNA helicase activity (2, 32), but so far, none has been shown to have this function. Previous experiments were carried out to characterize the cellular response of various CS-B mutant cell lines to different challenges (Table II) (22,24,34). We have also mapped the CSB putative helicase motifs and determined their role in cellular repair of 8-OH-Gua (Table II). Those experiments showed that motifs V and VI are essential for the cellular response to oxidative stress and for cellular repair of 8-OH-Gua in genomic DNA. The results presented here show that CSB motif VI also plays an important role in cellular repair of 8-OH-Ade.
It is not clear what mechanism underlies the role of CSB in the repair of 8-OH-Gua and 8-OH-Ade. The enzymes involved specifically in repair of 8-OH-Ade have not yet been identified (reviewed in Ref. 19), but some evidence suggests that repair of 8-OH-Ade is mechanistically different from repair of 8-OH-Gua. Among the bacterial and mammalian DNA glycosylases, which were investigated for their activity on lesions in oxidatively damaged DNA, only Escherichia coli formamidopyrimidine glycosylase (Fpg) exhibited a low activity on 8-OH-Ade in DNA containing multiple modified bases (35,36). However, this activity was insignificant when compared with the activity of Fpg on 8-OH-Gua, 2,6-diamino-4-hydroxy-5-formamidopyrimidine (FapyGua), and 4,6-diamino-5-formamidopyrimidine (FapyAde), which are the principal substrates of Fpg. The inactivity of Fpg was suggested to result from the absence of a C6-keto group in 8-OH-Ade (37). A recent study on the cellular repair of modified DNA bases showed that 8-OH-Ade is efficiently repaired in human cells with kinetics similar to the repair kinetics of pyrimidine-derived lesions rather than to that of purine-derived lesions (18). This suggests that human cells possess enzyme(s) to repair 8-OH-Ade. However, it is not known whether this repair activity involves BER or NER, or both. A yeast functional homolog of E. coli Fpg encoded by the OGG1 gene of Saccharomyces cerevisiae (yOgg1) was shown to excise 8-OH-Gua and FapyGua, but not FapyAde or 8-OH-Ade (38,39). Human homologues of yOgg1 were recently isolated (reviewed in Ref. 40). Two polymorphic forms of hOgg1 namely ␣-hOgg1-Ser 326 and ␣-hOgg1-Cys 326 are produced in human cells. Similar to yOgg1, both enzymes were reported to efficiently excise 8-OH-Gua and FapyGua, but not FapyAde or 8-OH-Ade from DNA with multiple lesions (41). Two enzymes from Drosophila melanogaster also exhibited similar activities that efficiently excised 8-OH-Gua and FapyGua from DNA, but not FapyAde and 8-OH-Ade (42,43). In the same context, it should be pointed out, that although 8-OH-Ade was repaired in human cells with kinetics similar to repair kinetics of pyrimidine-derived lesions, none of the known pyrimidine lesion-specific DNA glycosylases removed 8-OH-Ade from DNA either (reviewed in Ref. 19). These studies clearly show that there is a paucity of knowledge about enzymes that repair 8-OH-Ade in cells or in vitro. On the other hand, our results confirm results from the previous study on the cellular repair of 8-OH-Ade in terms of the ability of human cells to repair this lesion and in terms of its repair kinetics (18). By studying the biochemical contribution and expression regulation of CSB to cellular repair of 8-OH-Gua, we found that CSB itself participates in the catalytic process of 8-OH-Gua incision in vitro, and CSB facilitates the expression of hOgg1. 2 Further research is necessary to understand the mechanism by which CSB facilitates cellular repair of 8-OH-Ade and to identify other proteins that might contribute directly or indirectly to cellular repair of this lesion.
8-OH-Ade was identified as its nucleoside form 8-OH-dAdo, and its level and accumulation in genomic DNA of CS-B mutant cells was quantified by means of a recently developed methodology using LC/IDMS (44). The present study is the first application of this technique to the measurement of the formation and repair of 8-OH-dAdo in living cells. The specificity and sensitivity of LC/IDMS in the SIM mode permitted us to detect and quantify 8-OH-dAdo at a level of ϳ7 molecules/10 7 DNA nucleosides. Such a level of detection by LC/IDMS-SIM had previously been reported (44). For quantification, we isolated a stable isotope-labeled analog of 8-OH-dAdo in pure form for the first time and used it as the internal standard. It should be pointed out that the use of liquid chromatography/tandem mass spectrometry (LC/MS/MS) has also been reported for the identification of 8-OH-dAdo (45)(46)(47)(48). However, the isotope-di-lution technique has been applied for quantification in two instances only (45,48). The sensitivity level for 8-OH-dAdo of this technique in the multiple reaction-monitoring (MRM) mode is similar to that of LC/IDMS-SIM (17). On the other hand, the application of LC/MS/MS-MRM to the measurement of 8-OH-dAdo in living cells has not been reported.
In conclusion, the present study suggests that CSB plays a role in the cellular repair of 8-OH-Ade and that the putative helicase motif VI of CSB may be important in this process. In addition, 8-OH-Ade and other lesions such as 8-OH-Gua might accumulate in the DNA of CS patients and could potentially contribute to pathology associated with CS.