HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Masaoka, A. Articles by Ide, H. Search for Related Content PubMed PubMed Citation Articles by Masaoka, A. Articles by Ide, H. Social Bookmarking                      What's this?

J Biol Chem, Vol. 274, Issue 35, 25136-25143, August 27, 1999

Enzymatic Repair of 5-Formyluracil
I. EXCISION OF 5-FORMYLURACIL SITE-SPECIFICALLY INCORPORATED INTO OLIGONUCLEOTIDE SUBSTRATES BY AlkA PROTEIN (Escherichia coli 3-METHYLADENINE DNA GLYCOSYLASE II)^*

Aya Masaoka, Hiroaki Terato, Mutsumi Kobayashi, Akiko Honsho, Yoshihiko Ohyama, and Hiroshi Ide

From the Graduate Department of Gene Science, Faculty of Science, Hiroshima University, Kagamiyama, Higashi-Hiroshima 739-8526, Japan

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

5-Formyluracil (fU) is a major thymine lesion produced by reactive oxygen radicals and photosensitized oxidation. We have previously shown that fU is a potentially mutagenic lesion due to its elevated frequency to mispair with guanine. Therefore, fU can exist in DNA as a correctly paired fU:A form or an incorrectly paired fU:G form. In this work, fU was site-specifically incorporated opposite A in oligonucleotide substrates to delineate the cellular repair mechanism of fU paired with A. The repair activity for fU was induced in Escherichia coli upon exposure to N-methyl-N'-nitro-N-nitrosoguanidine, and the induction was dependent on the alkA gene, suggesting that AlkA (3-methyladenine DNA glycosylase II) was responsible for the observed activity. Activity assay and determination of kinetic parameters using purified AlkA and defined oligonucleotide substrates containing fU, 5-hydroxymethyluracil (hU), or 7-methylguanine (7mG) revealed that fU was recognized by AlkA with an efficiency comparable to that of 7mG, a good substrate for AlkA, whereas hU, another major thymine methyl oxidation products, was not a substrate. ¹H and ¹³C NMR chemical shifts of 5-formyl-2'-deoxyuridine indicated that the 5-formyl group caused base C-6 and sugar C-1' to be electron deficient, which was shown to result in destabilization of the N-glycosidic bond. These features are common in other good substrates for AlkA and are suggested to play key roles in the differential recognition of fU, hU, and intact thymine. Three mammalian repair enzymes for alkylated and oxidized bases cloned so far (MPG, Nth1, and OGG1) did not recognize fU, implying that the mammalian repair activity for fU resided on a yet unidentified protein. In the accompanying paper (Terato, H., Masaoka, A., Kobayashi, M., Fukushima, S., Ohyama, Y., Yoshida, M., and Ide, H., J. Biol. Chem. 274, 25144-25150), possible repair mechanisms for fU mispaired with G are reported.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Cellular DNA is constantly subjected to the threat of exogenous and endogenous DNA damaging agents (1, 2). The loss of critical genetic information stored in DNA due to the damage results in lethal and mutagenic events of cells. Furthermore, it has been shown that defective processing of DNA damage, i.e. impaired DNA repair and cellular responses to DNA damage, lead to tumorigenesis and carcinogenesis (1). Reactive oxygen species generated by cellular metabolism, exogenous redox active chemicals, and ionizing radiation produce a wide spectrum of oxidative DNA base damage, which constitutes major DNA lesions induced by exogenous and endogenous DNA damaging agents (3, 4). To cope with genotoxicity associated with oxidative base damage, cells have evolved a number of repair enzymes that specifically recognize unique structural features of altered bases and remove them from DNA (1, 2, 5). Among the repair enzymes that recognize oxidative base damage from four bases, those for thymine damage have been best identified and characterized, because the chemical nature of oxidative thymine damage has been established considerably well (3, 4). Oxidative thymine lesions are classified into four groups depending on the structure: group I, C-5,C-6 hydroxylation products, such as thymine glycol and thymine hydrate; group II, ring fragmentation products, such as urea and methyltartronylurea; group III, ring contraction products, such as hydantoin derivatives; and group IV, 5-methyl oxidation products, such as 5-formyluracil (fU)¹ and 5-hydroxymethyluracil (hU). In Escherichia coli, group I products are recognized by endonucleases III and VIII (6-10), encoded by the nth (11, 12) and nei (13, 14) genes, respectively. Recently, eukaryotic counterparts of the nth gene have been cloned, and the expressed gene products were shown to have activities similar to the E. coli enzyme (15-18). Group II lesions, such as urea residues, are recognized by a number of E. coli apurinic/apyrimidinic endonucleases without (exonuclease III and endonuclease IV) (19-21) and with associated N-glycosylase activities (endonucleases III and VIII, formamidopyrimidine DNA glycosylase) (7, 8, 10, 21, 22). Similar to E. coli endonuclease III, eukaryotic endonuclease III homologues appear to recognize urea type products, as well as group I products (15, 16, 18). However, somewhat surprisingly, bovine apurinic/apyrimidinic endonuclease 1 does not recognize urea residues (23), despite sharing considerable amino acid sequence homology with E. coli exonuclease III (24). Repair enzymes for group III and IV products have been much less clarified compared with those for group I and II products. Available data indicate that 5-hydroxy-5-methylhydantoin, a group III product, is a substrate of E. coli endonuclease III (8, 9, 25). Concerning the repair of the group IV products, mammalian cells contain an activity that releases hU from DNA (26, 27). However, hU itself is neither a lethal nor mutagenic lesion (28). Thus, further characterization of the activity and elucidation of the physiological role remain to be performed.

We have recently shown that fU belonging to group IV products forms a mispair with guanine in addition to a correct base pair with adenine during DNA replication (29). Ionization of fU promoted by the electron-withdrawing formyl group is involved in the mispairing between fU and guanine. It has been also demonstrated that the nucleoside and nucleotide forms of fU undergo specific modification by cysteine derivatives, although the biological relevance of this modification remains to be elucidated (30). In addition, exogenous 5-formyl-2'-deoxyuridine added to the culture medium exhibits the mutagenic effect on Salmonella TA102 and human cells (31, 32). Thus, in light of such genotoxic and cytotoxic potentials of fU, this lesion should be subjected to repair in cells. Previously, Bjelland et al. (33) found that fU was released from aged and photo-oxidized DNA by E. coli AlkA protein (3-methyladenine DNA glycosylase II) during a survey of the substrate specificity of the enzyme. Subsequently, it was reported that the activity that release fU from damaged DNA was present in crude extracts of mammalian cells (32, 34).

In the present work and the accompanying report (68), we prepared oligonucleotide substrates containing fU as unique damage and studied the repair of fU to obtain further insight into the cellular repair mechanism. The use of defined oligonucleotide substrates, which was not achieved in the previous studies, is essential to compare the substrate specificity and kinetic parameters of repair enzymes. More importantly, fU could be present in two different base pairs in DNA, i.e. a correctly paired fU:A form and a mispaired fU:G form. We report here that fU paired with A is specifically removed by E. coli AlkA protein with an efficiency comparable to that of 7-methylguanine (this paper). Furthermore, an fU:G mismatch generated during DNA replication is likely to be subjected to repair by both AlkA protein and the MutHLS mismatch correction system (68).

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Chemicals-- 7-Methyl-2'-dGTP was purchased from Sigma. Four normal dNTPs were obtained from Amersham Pharmacia Biotech, and N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) was from Wako. Preparation of 5-formyl-2'-deoxyuridine 5'-triphosphate (fdUTP) was reported previously (29). 5-Hydroxymethyl-2'-deoxyuridine 5'-triphosphate (hdUTP) was synthesized by NaBH₄ reduction of fdUTP and extensively purified by HPLC as described for fdUTP (29). The structure and purity of hdUTP were confirmed by alkaline phosphatase digestion of the triphosphate to the corresponding nucleoside followed by reversed phase HPLC analysis (data not shown). 3',5'-Di-O-accetylthymidine and 3',5'-di-O-acetyl-5-formyl-2'-deoxyuridine were synthesized according to the reported methods (35).

Bacteria-- E. coli MV1161 (thr1, ara14, leuB6, DEL(gpt-proA)62, lacY1, tsx33, supE44, galK2, hisG4, rfbD1, mgl51, rpsL31, kdgK51, xyl5, mtl1, argE3, thi1, rfa550) and MV1571 (alkA51::Mudl(Amp^Rlac) in MV1161) were gifts from Humayun and Dunman (New Jersey Medical School) (36) and were originally from Volkert et al. (37). E. coli JM105 (endA1, supE, sbcB15, thi, rpsL, (lac-proAB)/F'(traD36, proAB⁺, lacI^q, lacZM15)) was purchased from Amersham Pharmacia Biotech.

Buffer-- The following buffers were used for enzyme reactions: Buffer A for Pol I Klenow fragment, 66 mM Tris-HCl (pH 7.5), 1.5 mM 2-mercaptoethanol, 6.6 mM MgCl₂; Buffer B for AlkA, 70 mM Hepes-KOH (pH 7.8), 1 mM EDTA, 5 mM 2-mercaptoethanol; Buffer C for hMPG, 35 mM Hepes-KOH (pH 7.4), 50 mM NaCl, 5 mM EDTA, 5 mM dithiothreitol; Buffer D for mNth1, 20 mM Hepes-KOH (pH 8.0), 50 mM KCl, 0.25 mM EDTA, 0.25 mM dithiothreitol, 0.1 mg/ml bovine serum albumin; Buffer E for endonuclease IV, 10 mM Tris-HCl (pH 7.5), 50 mM NaCl, 1 mM EDTA; and Buffer F for hOGG1, 50 mM Tris-HCl (pH 7.5), 50 mM KCl, 2 mM EDTA, 0.1 mg/ml bovine serum albumin. Gel loading buffer consisted of 0.05% xylene cyanol, 0.05% bromphenol blue, 20 mM EDTA, and 98% formamide.

Enzymes-- E. coli DNA polymerase I Klenow fragment and T4 polynucleotide kinase were purchased from Life Technologies, Inc. and New England BioLabs, respectively. E. coli endonuclease IV was a gift from Hatahet and Wallace (University of Vermont). Human methylpurine DNA glycosylase (hMPG/Anpg) and 7,8-dihydro-8-oxoguanine DNA glycosylase (hOGG1/hMMH, splicing isoform 1a) were obtained from Kubo (38) and Nishimura (39), respectively. The mouse endonuclease III homologue (mNth1) was purified from E. coli NKJ1004 (nth nei) harboring plasmid PEK/mNth1 (18).

3-Methyladenine DNA glycosylase II (AlkA protein) was purified as follows. Based on the published sequence data (40), the alkA gene containing upstream 2 and downstream 57 extra nucleotides from the start and stop codons, respectively, was amplified by PCR from E. coli K12 genomic DNA using primers PCR-FW and PCR-RV (Table I). The amplified gene (approximately 1 kilobase) was restricted by EcoRI and HindIII, and inserted into the EcoRI/HindIII site of pKK223-3 (Amersham Pharmacia Biotech). The plasmid containing the alkA gene was designated pAlkA. The whole nucleotide sequence of the inserted alkA gene was confirmed by DNA sequencing. E. coli JM105 transformed with pAlkA was grown in LB medium (5 l) supplemented with 50 µg/ml ampicillin at 37 °C in a jar fermenter, and the alkA gene was induced by the addition IPTG (0.4 mM) at an absorbance of 0.2. Incubation was continued for further 9 h at 37 °C with aeration. Cells were collected by centrifugation, washed with 40 mM Tris-HCl (pH 8.0), and disrupted by the lysozyme treatment and repeated freeze-thaw cycles. Crude AlkA protein was precipitated by ammonium sulfate (45% saturation) and purified by the successive chromatography using HiLoad Superdex 75 pg, Resource S, and Resource Q columns (all columns were from Amersham Pharmacia Biotech) following the reported procedure (33). Ten N-terminal amino acid residues of the purified protein were determined by a protein sequencer and were consistent with those expected from the DNA sequence. The AlkA protein concentration after the Resource Q column was determined by UV absorption assuming ₂₈₀ = 59,400,² and used for enzyme assays.

Oligonucleotides-- Oligonucleotides used in this study are listed in Table I. Oligonucleotides except 25T, 25FU, 25HU, 25MG, 25FP, 19TG, and M1 were synthesized by the standard phosphoramidite method and purified by reversed phase HPLC. 25FU and 25HU containing fU and hU, respectively, at the 16th position from the 5'-terminus were prepared by the enzymatic extension of primer P1. Typically, P1 was 5'-end labeled with [-³²P]ATP (110 TBq/mmol, Amersham Pharmacia Biotech) and T4 polynucleotide kinase and purified as described (41). P1 annealed to the template 30A (12.5 pmol as a template/primer) in Buffer A (total volume, 500 µl) was extended by E. coli DNA polymerase I Klenow fragment (90 units) with fdUTP (100 µM) or hdUTP (100 µM) at 37 °C for 5 min and then with additional dATP, dGTP, and dCTP (100 µM each) for 30 min (29). The reaction was terminated by the addition of EDTA (final concentration, 25 mM). The reaction mixture was extracted with phenol, and duplex 25FU/30A and 25HU/30A were precipitated by ethanol. They were resuspended in water (100 µl), purified by Sephadex G-25 chromatography (bed volume, 2 ml), and recovered by ethanol precipitation. Duplex 25T/30A was prepared by the same procedure except that dTTP was used instead of fdUTP in the DNA polymerase reaction. It is noted that fully replicated products 25FU and 25T were 25-mer in length. This design aided separation of these oligonucleotides from longer template 30A by polyacrylamide gel electrophoresis when single stranded 25FU and 25T were needed to construct heteroduplex substrates in MutS binding experiments described in the accompanying paper (68). Duplex 25MG/30C containing 7-methylguanine (7mG) was prepared and purified in a manner essentially similar to that described for 25FU/30A. 5'-Labeled P2 was annealed to the template 30C and elongated in the presence of dCTP, dTTP, 7-methyl-2'-dGTP (100 µM each). 7mG incorporated into 25MG was stable when stored at 4 °C or below so that no detectable sites sensitive to endonuclease IV (abasic sites) or formamidopyrimidine DNA glycosylase (2,6-diamino-4-hydroxy-5-N-methylformamidopyrimidine (Fapy)) were generated during the storage. Duplex 25FP/30C containing Fapy was prepared by alkali treatment of 25MG/30C. The details of the synthesis and characterization of the oligonucleotides containing 7mG and Fapy will be published elsewhere.³ 19TG containing a thymine glycol was prepared by KMnO₄ oxidation of the oligonucleotide containing a single thymine at the same position and was extensively purified by reversed phase HPLC (42). The marker oligonucleotide M1 was prepared by the treatment of 19TG with 10% piperidine at 90 °C for 30 min.

Enzyme Assay with Crude Cell Extracts-- To analyze induction of repair activities by MNNG, E. coli MV1161 (alkA⁺) and MV1571 (alkA) were treated with MNNG as reported previously (43). E. coli MV1161 and MV1571 were grown overnight in LB medium in the absence (MV1161) and presence (MV1571) of 50 µg/ml ampicillin. The overnight cultures were inoculated into the same fresh medium with 8-fold dilution. After 3 h of incubation at 37 °C with shaking, MNNG was added to the medium at a final concentration of 20 µM, and incubation was continued for 1.5 h at 37 °C. Cells were collected by centrifugation, and 0.2 g of wet cells was treated by 1 mg/ml lysozyme in 10 mM Tris-HCl (pH 8.0), 5 mM EDTA (total volume, 0.5 ml) and then subjected to repeated freeze-thaw cycles three times. Cell debris were removed by centrifugation, and the protein concentration of the supernatant was determined by the BCA protein assay reagent kit (Pierce). Repair activities for fU and 7mG in the crude cell extracts were assayed as follows. 5'-End labeled oligonucleotide substrates (25FU/30A and 25MG/30C, 20 nM) were incubated with 20 µg of crude cell extracts in Buffer B (50 µl) at 37 °C for 30 min. The reaction mixtures was extracted by phenol. Oligonucleotides were precipitated by ethanol and resuspended in Buffer E (10 µl). Endonuclease IV (4 ng) was added to the reaction mixture, and incubation was continued at 37 °C for 30 min. The reaction was terminated by adding gel loading buffer. Samples were subjected to polyacrylamide gel electrophoresis (PAGE) analysis.

The repair activity for fU in crude cell extract from E. coli JM105 harboring pAlkA or pKK223-3 (control) was assayed as follows. E. coli JM105 harboring pAlkA or pKK223-3 was grown in LB medium containing 50 µg/ml ampicillin at 37 °C overnight. The overnight culture was inoculated in fresh LB medium containing 50 µg/ml ampicillin with 16-fold dilution. After 3 h of incubation at 37 °C, IPTG was added to the medium at a final concentration 0.4 mM. Cells were collected at appropriate incubation times and stored frozen at 80 °C. Cell extracts were prepared, and the repair activity was assayed using substrates 25FU/30A and 25MG/30C, as described for the MNNG treatment.

Enzyme Assay with Purified Enzymes-- Oligonucleotide substrates, such as 25FU/30A and 25MG/30C (typically 20 nM), were incubated with AlkA protein (10 ng) in Buffer B (10 µl) at 37 °C for 10 min unless otherwise noted. After incubation, DNA was purified by phenol extraction and ethanol precipitation. The purified DNA was incubated with endonuclease IV (4 ng) at 37 °C for 30 min in Buffer E (10 µl), mixed with gel loading buffer, and finally subjected to gel electrophoresis. Kinetic parameters of AlkA for fU and 7mG were determined in an essentially similar manner using substrates 25FU/30A and 25MG/30C with a concentration range of 5-100 nM.

Reactions with mammalian enzymes were performed as follows. 19TG and 25OG were 5'-end labeled and purified as described above and annealed to the complementary strands 19A and 30C, respectively. 19TG/19A and 25FU/30A (20 nM) were incubated with mNth1 (1-10 ng) in Buffer D (10 µl) at 37 °C for 30 min. 25OG/30C, 25FP/30C, and 25FU/30A (20 nM) were also treated in a similar manner by hOGG1(60 ng) in Buffer F (10 µl). 25MG/30C and 25FU/30A (20 nM) were incubated with hMPG (5-50 ng) in Buffer C (10 µl) at 37 °C for 30 min. For hMPG, DNA was purified and further treated with endonuclease IV as described for AlkA.

Electrophoresis and Autoradiography-- After adjusting total radioactivity, samples in gel loading buffer were loaded onto a 16% denaturing polyacrylamide gel and electrophoresed at 2000 V for 2-3 h. The gel was autoradiographed at 80 °C. The radioactivity of separated bands was quantitated by Fuji BAS 2000.

NMR Spectra-- ¹H and ¹³C NMR spectra of 3',5'-diacetyl derivatives of thymidine and 5-formyl-2'-deoxyuridine were measured in CDCl₃ on a Varian Gemini 200 spectrometer at 200 (¹H) and 50.3 (¹³C) MHz, respectively, with tetramethylsilane (¹H) and the solvent signal (¹³C) as internal standards.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Parallel Induction of Repair Activities for fU and 7mG by MNNG Treatment of E. coli Cells-- E. coliMV1161 (alkA⁺) and MV1571 (alkA) were treated with a sublethal dose of MNNG, and induction of the repair activity for fU and 7mG in cells was examined using substrates containing these damage (25FU/30A and 25MG/30C). The substrates incubated with crude cell extracts were further treated with endonuclease IV to cleave abasic sites potentially formed in the substrates. Fig. 1 shows the results of product analysis by PAGE, where bands of original substrates (25FU and 25MG) and reaction products resulting from specific cleavage at fU or 7mG site are indicated by arrows. In MV1161 cells (alkA⁺), repair activity for fU was in a basal level and was not clearly seen over the background before the MNNG treatment (Fig. 1, lane 10), but it was noticeably induced after the treatment (lane 11). In contrast, such induction did not occur in the alkA mutant MV1571 (lanes 12 and 13). The induction pattern of the repair activity for fU was identical to that of 7mG so that a clear band indicating the repair of 7mG appeared over the background only in MNNG-treated wild type cells (MV1161, lane 5), but not in the alkA mutant (lane 7). These results suggest that the repair activity for fU are inducible by MNNG in E. coli and is probably associated with AlkA protein.

View larger version (25K): [in this window] [in a new window]   Fig. 1.   MNNG induction of the repair activities for fU and 7mG in E. coli cells. E. coli MV1161 (alkA+) and MV1571 (alkA) were treated with or without 20 µM MNNG at 37 °C for 1.5 h, and cell extracts were prepared by the freeze-thaw method. 25FU/30A containing fU and 25 MG/30C containing 7mG (both at 20 nM) were incubated with the cell extracts (20 µg as protein) in Buffer B (50 µl) at 37 °C for 10 min. After incubation, oligonucleotides were purified by phenol extraction followed by ethanol precipitation, and treated with endonuclease IV (4 ng) in Buffer E (10 µl) at 37 °C for 30 min. Reaction products were analyzed by 16% PAGE. Lane 1, marker oligonucleotide P1 (see Table I); lane 2, untreated 25MG/30C; lane 3, 25MG/30C incubated with endonuclease IV alone; lanes 4 and 5, 25MG/30C incubated with the extract from alkA+ cells treated without () or with (+) MNNG; lanes 6 and 7, 25MG/30C incubated with the extract from alkA cells treated without () or with (+) MNNG; lane 8, untreated 25FU/30A; lane 9, 25FU/30A incubated with endonuclease IV alone; lanes 10 and 11, 25FU/30A incubated with the extract from alkA+ cells treated without () and with (+) MNNG; lanes 12 and 13, 25FU/30A incubated with the extract from alkA cells treated without () and with (+) MNNG.

Recognition of fU by AlkA-- Because it was suggested by the MNNG induction experiments that AlkA had an repair capacity for fU as well as 7mG, the cloned alkA gene was overexpressed in E. coli, and the repair activity was examined using crude cell extracts and purified protein. As shown in Fig. 2, significant repair activities for both fU and 7mG were present in cell extracts prepared from IPTG-induced JM105 harboring pAlkA (lanes 5 and 9), but not pKK223-3, a control parental vector (lanes 4 and 8).

View larger version (28K): [in this window] [in a new window]   Fig. 2.   Overexpression of repair activities for fU and 7mG in E. coli harboring pAlkA. Cell extracts were prepared from JM105 harboring pKK223-3 (parental vector) or pAlkA containing the alkA gene after IPTG induction. 25MG/30C containing 7mG and 25FU/30A containing fU (both 20 nM) were incubated without or with the cell extracts and treated with endonuclease IV as described in Fig. 1. Lane 1, marker oligonucleotide P1 (Table I); lane 2, untreated 25MG/30C; lane 3, 25MG/30C incubated with endonuclease IV alone; lanes 4 and 5, 25MG/30C incubated with the extracts from JM105/pKK223-3 (indicated as pKK223) and JM105/pAlkA (indicated as pAlkA), respectively; lanes 6-9, the same as lanes 2-5 but the substrate was 25FU/30C.

AlkA protein was overproduced in JM105/pAlkA and purified by successive chromatography as described under "Experimental Procedures." Purified AlkA protein was physically homogenous and gave a single band with an apparent molecular mass of 30 kDa in SDS-PAGE (Fig. 3). Purified AlkA exhibited repair activity for both fU and 7mG (Fig. 4). The activity for the both damage increased in a parallel manner with the amount of AlkA protein added to the reaction, indicating that the activity for fU and 7mG resided on the same protein AlkA. Furthermore, comparison of the amounts of reaction products for fU and 7mG suggests that both damages are recognized by AlkA with a similar efficiency.

View larger version (87K): [in this window] [in a new window]   Fig. 3.   SDS-PAGE analysis of the purified fractions of AlkA. Protein was separated by 10% SDS-PAGE and visualized by silver staining. Lane 1, molecular weight markers; lane 2, cell extract from JM105/pAlkA (1 µg); lane 3, 45% ammonium sulfate precipitation (1 µg); lane 4, HiLoad Superdex 75 pg (0.2 µg); lane 5, Resource S (0.1 µg); lane 6, Resource Q (0.1 µg). The amount of protein loaded in each lane is indicated in the parentheses.

View larger version (31K): [in this window] [in a new window]   Fig. 4.   Repair activity of purified AlkA for fU and 7mG. 25FU/30A and 25MG/30C (both 20 nM) were incubated with varying amounts of purified AlkA (Resource Q fraction) in Buffer B (10 µl) at 37 °C for 10 min. After incubation, oligonucleotides were purified by phenol extraction followed by ethanol precipitation, and treated with endonuclease IV (4 ng) in Buffer E (10 µl) at 37 °C for 30 min. Reaction products were analyzed by 16% PAGE. Lane 1, untreated 25FU/30A; lanes 2-5, 25FU/30A treated with increasing amounts of AlkA (1, 4, 11, and 107 ng, respectively); lane 6, untreated 25 MG/30C; lanes 7-10, 25MG/30C treated with increasing amounts of AlkA (1, 4, 11, and 107 ng, respectively).

Kinetic Parameters for Recognition of fU and 7mG-- To compare the reaction efficiencies of AlkA for fU and 7mG, the initial reaction velocity was determined using 25FU/30A and 25MG/30C with a substrate concentration range of 5-100 nM. Fig. 5 shows the substrate-velocity plots for fU and 7mG. The kinetic parameters (K_m and V_max) were evaluated using a hyperbolic curve fitting program and are summarized in Table II. The data are averages of four independent experiments. Judging from K_m for fU (21 nM) and 7mG (30 nM), the apparent affinity of AlkA for fU was slightly higher than that for 7mG, whereas V_max values for fU and 7mG were essentially comparable. The ratio of the overall reaction efficiencies (V_max/K_m) for 7mG versus fU was 1:1.4, showing that fU and 7mG were equally good substrates for AlkA.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Substrate-velocity plots for the excision of fU and 7mG by AlkA. 25FU/30A and 25MG/30C were incubated with AlkA (10 ng) in Buffer B (10 µl) at 37 °C for 10 min and treated with endonuclease IV as described in Fig. 4. After gel electrophoretic separation, the radioactivity of substrates and nicked products in the gel were quantitated by BAS 2000. The initial velocity (V) was obtained with varying substrate concentrations (S) (5-100 nM). A, fU; B, 7mG. Mean values of four independent experiments are plotted; error bars show S.D.

Activity of AlkA to hU-- Reactive oxygen radicals generate two major stable methyl oxidation products of thymine, i.e. fU and hU (3, 31). To elucidate whether or not AlkA recognized hU, this damage was introduced into the same site as fU in the oligonucleotide 25HU. Duplex substrates 25HU/30A as well as 25T/30A and 25FU/30A were incubated with AlkA, and reaction products were analyzed by PAGE (Fig. 6). Although AlkA acted on fU very efficiently (lanes 7 and 9) under these conditions, hU introduced into the same site was not recognized (lanes 11 and 13). Similarly, intact thymine was not a substrate of AlkA either (lanes 3 and 5). Thus, AlkA differentially recognizes the effects of modifications in the 5-substituent group.

View larger version (32K): [in this window] [in a new window]   Fig. 6.   Differential activities of AlkA for T, fU, and hU. 25T/30A, 25FU/30A, and 25HU/30A (2 or 20 nM) containing T, fU, and hU, respectively, at the same position were incubated with AlkA (50 ng) in Buffer B (10 µl) at 37 °C for 20 min and then treated with endonuclease IV as described in Fig. 4. After adjusting radioactivity, reaction products were separated by 16% PAGE. Lane 1, marker oligonucleotide P1 (Table I). For lanes 2-13, substrates, their concentrations, and treatment without () or with (+) AlkA are indicated on the top of the lanes.

Activity of Mammalian Repair Enzymes to fU-- Mammalian cells have functional homologues of AlkA that exhibit a substrate specificity similar to AlkA (44-48). In addition, they have endonuclease III (16-18) and MutM (39, 49-52) homologues that recognize oxidized pyrimidine and purine bases, respectively. Three mammalian homologues, cloned human methylpurine DNA glycosylase (hMPG/Anpg) (46-48), mouse endonuclease III homologue (mNth1) (18), and human MutM homologue (hOGG1/hMMH) (39, 49-52), were tested for fU to ask whether fU was recognized by these enzymes. The enzymes were incubated with 25FU/30A and other control substrates (25MG/30C, 19TG/19A, 25OG/30C, and 25FP/30C), and products were analyzed by PAGE (Fig. 7). Although the tested enzymes recognized their canonical substrates (7mG (Fig. 7A, lane 14), thymine glycol (Fig. 7A, lane 16), 7,8-dihydro-8-oxoguanine (Fig. 7B, lane 3), and Fapy (Fig. 7B, lane 5), respectively), none of them recognized fU. Thus, no cleavage product appeared in the gel (Fig. 7A, lanes 6-11, and Fig. 7B, lane 7).

View larger version (31K): [in this window] [in a new window]   Fig. 7.   PAGE analysis of the repair activities of AlkA, mNth1, hMPG, and hOGG1 for fU and other lesions. A, 25FU/30A, 25MG/30C, and 19TG/19A (all 20 nM) were treated with AlkA, mNth1, or hMPG at 37 °C for 10 min (AlkA) or 30 min (mNth1 and hMPG). For AlkA and hMPG, reaction products were further treated with endonuclease IV as described in Fig. 4. Lane 1, marker oligonucleotide P1 (Table I); lane 2, untreated 25FU/30A; lanes 3-11, 25FU/30A treated with increasing amounts of AlkA (1, 5, 10 ng), mNth1 (1, 5, 10 ng), and hMPG (5, 25, 50 ng); lane 12, marker oligonucleotide P2 (Table I); lane 13, untreated 25MG/30C; lane 14, 25MG/30C treated with hMPG (50 ng); lane 15, untreated 19TG/19A; lane 16, 19TG/19A treated with mNth1 (10 ng); lane 17, marker oligonucleotide M1 prepared by hot piperidine treatment of 19TG (Table I). The -elimination product formed by mNth1 (lane 16) migrated more slowly than the -elimination product (M1) formed by the piperidine treatment (lane 17). B, 25OG/30C, 25FP/30C, and 25FU/30A (all 20 nM) were treated with hOGG1 (60 ng) at 37 °C for 30 min. Lane 1, marker oligonucleotide P2; lanes 2 and 3, 25OG/30C without () and with (+) hOGG1 treatment; lanes 4 and 5, 25FP/30C without and with hOGG1 treatment; lanes 6 and 7, 25FU/30C without and with hOGG1 treatment. The mobility of the reaction product from 25OG and 25FP (lanes 3 and 5) was slightly lower than that of P2 (lane 1) because hOGG1 generated strand breaks by the -elimination mechanism (38, 48-51).

NMR Spectra of 3',5'-di-O-acetyl-5-formyl-2'-deoxyuridine-- To obtain chemical insight into the substrate recognition and catalytic mechanisms of AlkA, ¹H and ¹³C NMR spectra of 5-formyl-2'-deoxyuridine and thymidine (as 3',5'-di-O-acetyl derivatives) were measured, and their chemical shifts were compared (Table III). Oxidation of the 5-methyl group in thymidine to the formyl group resulted in significant down-field shifts of H6 ( = 1.22 ppm) and C6 ( = 12.1 ppm) signals along with those of the formyl group itself. These shifts clearly indicate that the influence of the electron-withdrawing formyl group is extended to the pyrimidine ring via the  electron system, making H6 and C6 atoms electron-deficient relative to thymidine. The formyl group also caused deoxyribose C-1' to be somewhat electron-deficient, as judged from the noticeable down-field shift of the ¹³C signal of this atom ( = 6.1 ppm).

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

E. coli has two DNA glycosylases for alkylated bases, Tag (3-methyladenine DNA glycosylase I) and AlkA (3-methyladenine DNA glycosylase II) (53). Both enzymes are monofunctional DNA glycosylases without any associated endonuclease or apurinic/apyrimidinic lyase activity. Tag is constitutively expressed in cells and acts only on 3-methyladenine (3mA), whereas AlkA is inducible upon exposure to alkylating agents (40, 54) and excises structurally diverse damaged bases, such as 3mA, 7mG, O²-methylcytosine, O²-methylthymine, etc. (55, 56). Functional homologues of AlkA have been cloned from several eukaryotic sources (44-48). In this work, we have unambiguously demonstrated that AlkA recognizes fU, a major oxidation product of thymine, using a defined oligonucleotide substrate. This conclusion is supported by several lines of evidence: (i) induction of the repair activity in response to the MNNG treatment, and its absence in the alkA mutant (Fig. 1), (ii) increased repair activity in the cell extract of E. coli harboring pAlkA (Fig. 2), (iii) activity of purified AlkA to fU (Fig. 4), and (iv) parallel observation of the above results (i-iii) for the substrates containing fU and 7mG. Similar to other known substrates, AlkA acted as a simple DNA glycosylase on fU leaving an abasic site because the nicked product of 25FU/30A was detected only after the treatment with endonuclease IV (data not shown).

The activity of AlkA for fU demonstrated in the present study is qualitatively consistent with the result reported by Bjelland et al. (33), who showed the release of ³H-labeled fU from aged or photo-oxidized DNA by AlkA. However, K_m for fU in their report varies dramatically depending on the substrates: K_m = 4.4 nM (aged DNA in a native state), 0.14 nM (aged DNA in a denatured state), and 116 nM (photo-oxidized DNA in a denatured state). Therefore, it is rather confusing that the K_m for native DNA, supposed to be a better substrate, is higher than that for denatured DNA, and K_m values differ by 10³-fold between aged and photo-oxidized DNA. These results could arise from the nature of the substrates. Both emission of  particles during the aging of ³H-labeled DNA and photo-oxidation of DNA generate multiple DNA lesions that may specifically or nonspecifically interact with AlkA. Similarly, kinetic parameters of AlkA and its functional homologues for methylated bases have been determined using DNA substrates prepared by the treatment with methylating agents (33, 43, 53, 57, 58), which generate N- and O-methylated purines and pyrimidines, methylphosphates, and abasic sites simultaneously (59). Accordingly, parameters obtained from these substrates tacitly contains the contributions from unintended lesions. In contrast, the parameters determined in this study (Table II) reflect intrinsic values and can be directly compared with a reasonable basis because the substrate contains a single fU or 7mG site as a unique lesion. The kinetic parameters for fU and 7mG indicate that fU is a comparable substrate with 7mG, known as a good substrate for AlkA. To our knowledge, this is the first report of kinetic parameters of AlkA for recognition of fU and 7mG site-specifically introduced into oligonucleotide substrates. AlkA excises 3mA more rapidly than 7mG despite the 3-8-fold abundance of 7mG over 3mA in alkylated DNA (59). In addition, hypoxanthine (Hx) and 1,N⁶-ethenoadenine (A) are extremely poor substrates (58). Combining the published and present data, the rank of damage as a substrate for AlkA is likely to be following: 3mA > fU ~ 7mG Hx, A > (hU). It is noteworthy that apparent affinity (K_m) of AlkA to these substrates also decreases in this order: 3mA (7 nM), fU (21 nM), 7mG (30 nM), Hx (420 nM), A (800 nM) (Refs. 42 and 57 and this work). K_m for hU estimated using SPO1 phage DNA containing ³H-labeled hU is approximately 4000 nM, indicating that hU is an extremely poor substrate (33). This is essentially consistent with the present result that hU incorporated in place for fU was not recognized by AlkA, because such a low activity, even if present, was well below the detection limit of the present experiment. The activity to hU observed in a previous work (33) might be due to the recently reported unique activity of AlkA that releases normal bases from intact DNA (60).

Mammalian cells contain activities that release fU from oxidized DNA (32, 34). Thus, we have tested three cloned mammalian repair enzymes, hMPG/Anpg, mNth1, and hOGG1/hMMH, for fU. hMPG/Anpg is a human N-alkylpurine glycosylase and has a certain overlapping substrate specificity with AlkA (55, 56), yet their genes show little sequence homology (46-48). mNth1 is a mouse homologue of endonuclease III, a major E. coli enzyme for oxidized pyrimidines (18). mNth1 and endonuclease III share common features, such as the helix-hairpin-helix motif, the 4Fe-4S cluster, and the amino acids (Lys-208 and Asp-227 in mNth1) essential for the catalytic activity. hOGG1/hMMH is a human functional homologue of E. coli MutM (Fpg) that is responsible for the repair of oxidized purines, such as 7,8-dihydro-8-oxoguanine and Fapy (39, 49-51). These three enzymes represent base excision repair enzymes for methylated and oxidized base damage thus far cloned from mammalian cells. However, MPG/Anpg, mNth1, and hOGG1/hMMH did not act on fU, although they recognized the canonical substrates under the same conditions (Fig. 7), indicating that the mammalian repair activity for fU so far reported is associated with a yet unidentified protein.

Expression of the hMPG gene in E. coli alkA mutants corrects the sensitivity to alkylating agents, suggesting that AlkA and hMPG have a common function in removing lethal alkylated damage, such as 3mA (46-48). However, precise modes of substrate recognition and cleavage of the N-glycosidic bond are likely different. For example, hMPG efficiently recognizes Hx and A, which bear no positive charge and are extremely poor substrates for AlkA (43, 58), whereas this enzyme does not remove fU, a good substrate for AlkA (this work). These differences might reflect the divergence of recognition and catalytic mechanisms probably resulting from the lack of amino acid homology between AlkA and hMPG.

Molecular basis of the substrate specificity of AlkA accepting diverse base damage has not been fully established, but the three-dimensional structure recently solved by x-ray crystallography brought about suggestive insight into the substrate recognition and catalytic mechanism of this enzyme (61, 62). Judging from available data (Refs. 43, 58, and 63 and this work), damage recognized by AlkA could be roughly divided into two groups that are excised efficiently (3mA, 7mG, fU, and O²-methylcytosine) and very poorly or not at all (Hx, A, hU, O²-methylthymine, and O⁴-methylthymine). The excision efficiency of the second group appears to be at least 2 or 3 orders of magnitude lower than that of the first group. The positive charge in the base moiety and concomitant destabilization of the N-glycosidic bond were originally suggested as key features for the recognition by AlkA (64). The half-lives of the N-glycosidic bond of alkylated bases, fU and T (pH 7 and 37 °C, in free deoxyribonucleosides) are 3mA (0.8 h) < 7mG (4.6 h) < O²ethylC (26 h) < fU (17 days) < O²ethylT (220 days) < O⁴ethylT ~ T (30, 65, 66), where bases bearing an apparent positive charge at pH 7 are underlined (see also Fig. 8A), and the last two bases (O⁴ethylT and T) are not substrates for AlkA (63). In general, the damaged bases with a positive charge or a labile N-glycosidic bond (3mA, 7mG, O²ethylC, and fU) are good to fair substrates, although the order of the stability of the N-glycosidic bond does not exactly correlate with the excision efficiency by AlkA. For example, as shown in the present work, 7mG and fU are recognized with comparable efficiencies, yet their half-lives of the N-glycosidic bond are significantly different. Therefore, these data imply that additional factors, either steric or electronic, may also influence the base excision efficiency of AlkA. It has been implicated by the structural and site-directed mutagenesis data that the electron deficient bases flip out the DNA duplex to form strong -donor/acceptor interactions with electron-rich aromatic amino acids present in the active site of AlkA (62, 67). We suggest that a similar interaction is principally operating in the recognition and/or excision of fU by AlkA. As shown by the NMR data (Table III), the electron-withdrawing 5-formyl group in fU causes the pyrimidine ring and deoxyribose C-1' electron deficient relative to parent thymidine. The resonance structures of 5-formyl-2'-deoxyuridine are illustrated in Fig. 8B. As presumed for the substrates bearing the apparent positive charge (3mA, 7mG, and O²ethylC), the delocalized positive charge through the  electron system would increase the affinity of fU to the active site composed of electron-rich aromatic rings and/or accelerate the rupture of the N-glycosidic bond in the transition state. In the latter case, the induced positive charge in deoxyribose C-1' will promote the attack of a hydroxide ion on this atom, hence effecting the N-glycosylase reaction (62). In contrast, the 5-methyl and 5-hydroxymethyl groups in thymine and hU, respectively, are intrinsically electron-donating groups and are therefore unable to induce positive charges in the pyrimidine ring and deoxyribose C-1'. Probably, this is the primary reason that hU and thymine are recognized poorly or not at all.

View larger version (22K): [in this window] [in a new window]   Fig. 8.   Structures and charges of base damage differentially recognized by AlkA. A, structures and charges at pH 7. B, resonance structures of 5-formyl-2'-deoxyuridine.

Although fU resulting from oxidation of thymine mostly pairs with adenine, forming a correct base pair during DNA replication, it can also form an fU:G mismatch with a certain frequency (29, 68). Therefore, it is likely that cells need to deal with fU existing as two different base pairs, i.e. fU:A and fU:G. The present study has shown that fU in the fU:A pair is efficiently removed by AlkA, an E. coli repair enzyme for alkylated bases. In the accompanying paper (68), we report that fU present as an fU:G mispair is recognized by two E. coli repair proteins, AlkA and MutS.

	ACKNOWLEDGEMENTS

We thank Drs. Kihei Kubo, Shuji Seki, Susumu Nishimura, Zafer Hatahet, and Susan S. Wallace for generous gifts of DNA repair enzymes and Drs. Zafri Humayun and Paul M. Dunman for E. coli MV1161 and 1571. We are also grateful to Hironobu Nakano and Takao Yamada for technical assistance and preparation of oligonucleotides containing thymine glycol and Fapy.

	FOOTNOTES

^* This work was supported by grants-in-aid from Ministry of Education, Science, Sports and Culture of Japan (to H. I.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed. Tel. & Fax: +81-824-24-7457; E-mail: ideh@ipc.hiroshima-u.ac.jp.

² Y. Yamagata, personal communication.

³ H. Terato, T. Yamada, K. Asagoshi, Y. Ohyama, and H. Ide, manuscript in preparation.

	ABBREVIATIONS

The abbreviations used are: fU, 5-formyluracil; hU, 5-hydroxymethyluracil; 7mG, 7-methylguanine; 3mA, 3-methyladenine; Hx, hypoxanthine; A, 1,N⁶-ethenoadenine; fdUTP, 5-formyl-2'-deoxyuridine 5'-triphosphate; hdUTP, 5-hydroxymethyl-2'-deoxyuridine 5'-triphosphate; Fapy, 2,6-diamino-4-hydroxy-5-N-methylformamidopyrimidine; MNNG, N-methyl-N'-nitro-N-nitrosoguanidine; AlkA, E. coli 3-methyladenine DNA glycosylase II; hMPG/Anpg, human methylpurine DNA glycosylase; mNth1, mouse endonuclease III homologue; hOGG1/hMMH, human 7,8-dihydro-8-oxoguanine DNA glycosylase/human MutM homologue; PCR, polymerase chain reaction; PAGE, polyacrylamide gel electrophoresis; IPTG, isopropyl-1-thio--D-galactopyranoside; HPLC, high pressure liquid chromatography.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES