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

J. Biol. Chem., Vol. 277, Issue 30, 26987-26993, July 26, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/30/26987    most recent M111100200v1 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 Saparbaev, M. Articles by Laval, J. Search for Related Content PubMed PubMed Citation Articles by Saparbaev, M. Articles by Laval, J. Social Bookmarking                      What's this?

1,N²-Ethenoguanine, a Mutagenic DNA Adduct, Is a Primary Substrate of Escherichia coli Mismatch-specific Uracil-DNA Glycosylase and Human Alkylpurine-DNA-N-Glycosylase^*

Murat Saparbaev^§, Sophie Langouët^¶, Cyril V. Privezentzev, F. Peter Guengerich^**, Hongliang Cai^**§§, Rhoderick H. Elder^¶¶, and Jacques Laval

From the  Groupe Réparation de l'ADN, Unité Mixte de Recherche 8532 CNRS, Laboratoire de Biotechnologies et Pharmacologie Génétique Appliquée-Ecole Normale Supérieure Cachan, Institut Gustave Roussy, 94805 Villejuif Cedex, France, ^¶ INSERM U456, Faculté des Sciences Pharmaceutiques et Biologiques, Université de Rennes I, 35043 Rennes Cedex, France, ^** Department of Biochemistry and Center in Molecular Toxicology, Vanderbilt University School of Medicine, Nashville, Tennessee 37232-0146, and ^¶¶ Cancer Research United Kingdom Carcinogenesis Group, Paterson Institute for Cancer Research, Christie Hospital National Health Service Trust, Wilmslow Road, Manchester M20 4BX, United Kingdom

Received for publication, November 20, 2001, and in revised form, April 4, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The promutagenic and genotoxic exocyclic DNA adduct 1,N²-ethenoguanine (1,N²-G) is a major product formed in DNA exposed to lipid peroxidation-derived aldehydes in vitro. Here, we report that two structurally unrelated proteins, the Escherichia coli mismatch-specific uracil-DNA glycosylase (MUG) and the human alkylpurine-DNA-N-glycosylase (ANPG), can release 1,N²-G from defined oligonucleotides containing a single modified base. A comparison of the kinetic constants of the reaction indicates that the MUG protein removes the 1,N²-G lesion more efficiently (k_cat/K_m = 0.95 × 10³ min¹ nM¹) than the ANPG protein (k_cat/K_m = 0.1 × 10³ min¹ nM¹). Additionally, while the nonconserved, N-terminal 73 amino acids of the ANPG protein are not required for activity on 1,N⁶-ethenoadenine, hypoxanthine, or N-methylpurines, we show that they are essential for 1,N²-G-DNA glycosylase activity. Both the MUG and ANPG proteins preferentially excise 1,N²-G when it is opposite dC; however, unlike MUG, ANPG is unable to excise 1,N²-G when it is opposite dG. Using cell-free extracts from genetically modified E. coli and murine embryonic fibroblasts lacking MUG and mANPG activity, respectively, we show that the incision of the 1,N²-G-containing duplex oligonucleotide has an absolute requirement for MUG or ANPG. Taken together these observations suggest a possible role for these proteins in counteracting the genotoxic effects of 1,N²-G residues in vivo.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The etheno ()¹ ring system is formed by the attack of reactive bifunctional epoxides or aldehydes at a nitrogen of the DNA base, followed by dehydration and ring closure (1-2). The -derivatives of purine and pyrimidine bases (e.g. 1,N⁶-ethenoadenine (A), N²,3-ethenoguanine (N²,3-G), 1,N²-ethenoguanine (1,N²-G), and 3,N⁴-ethenocytosine (C)) are generated in cellular DNA by reaction with epoxides that result from the metabolism of various industrial pollutants. The highly mutagenic and genotoxic properties of -adducts have been established in vitro by analyzing steady-state kinetics of primer extension assays and in vivo by site-specific mutagenesis in mammalian cells (3-7).

The increasing interest in exocyclic DNA adducts has been triggered by the discovery that they can be formed by endogenous processes through the interaction of lipid peroxidation-derived aldehydes and hydroxyalkenals with DNA (8-9). -adducts are ubiquitous and have been found in DNA isolated from tissues of untreated rodents and human controls (10). However, A and C levels were significantly increased by cancer risk factors contributing to oxidative stress/lipid peroxidation, such as dietary w-6 fatty acid intake, chronic infections, and inflammatory conditions (11). These findings strongly suggest that exocyclic DNA adducts, together with other forms of oxidative DNA damage, play a role in the multistage process of human carcinogenesis, in which persistent oxidative stress is increasingly recognized as a driving force toward malignancy.

Although 1,N²-G is yet to be detected in genomic DNA, possibly due to the lack of an appropriate detection method, it was found to be the major guanine adduct formed in reactions of lipid peroxidation products with DNA in vitro (Fig. 1) (12). While 1,N²-G is a moderate DNA polymerase-blocking lesion in vitro (13) widely differing misincorporation frequencies opposite 1,N²-G have been observed both with in vitro assays and in Escherichia coli (13-14). Furthermore, the presence of 1,N²-G in chromosomal DNA of Chinese hamster ovary cells induces deletions, rearrangements, double mutants, and base pair substitutions (7).

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Chemical structures of exocyclic adducts.

Therefore, as all -adducts are potentially mutagenic, mechanisms should exist for their removal from DNA. Indeed the repair of A and N²,3-G adducts in DNA by DNA glycosylases present in rat cell extracts was first described over fifteen years ago (15). More recent studies have shown that this activity is associated with mammalian alkylpurine-DNA-N-glycosylases (ANPG) (16). The purified AlkA protein from E. coli has also been shown to release A and N²,3-G when present in DNA (17, 18). For C, it has been shown that both the E. coli mismatch-specific uracil-DNA glycosylase (MUG) and its structural and functional homolog in human cells, thymine-DNA glycosylase (hTDG) can release this exocyclic pyrimidine from DNA (19, 20). Recently, it was also shown that the human mismatch-specific DNA N-glycosylase (MED1/MBD4) has a weak C-DNA glycosylase activity (21).

Human ANPG is a monofunctional N-glycosylase that releases a variety of structurally unrelated bases from DNA. In addition to the N-methylpurines, we have shown previously that ANPG efficiently excises A when present in DNA (K_m 24 nM and k_cat 0.05 min¹) (18, 22) and hypoxanthine (K_m 6 nM and k_cat 0.21 min¹) (23, 24). It should be stressed that the ANPG and AlkA proteins, in contrast to the MUG and hTDG proteins, are not able to excise C (19, 25). Similarly, the MUG and hTDG proteins do not release A from DNA (18, 26).² Two alternatively spliced transcripts of human ANPG have been isolated from human cells. The full-length cDNA first isolated by Samson et al. (ANPG70, 293 amino acids) (38) differs from the splice variant (ANPG60, 298 amino acids) (44) only in the sequence of the first eight and 13 N-terminal amino-acids respectively. Two truncated versions of human ANPG have also been described: ANPG40 lacks the first 63 amino acids and ANPG80 the first 73 amino acids from the N terminus when compared with APNG70 (33, 48).

Dosanjh and co-workers have shown that partially purified HeLa extracts can release A and 1,N²-G (16). However, because a cell extract was used, it was unclear whether the two -adducts were repaired by a single enzyme. In the present study using highly purified proteins, we have shown that MUG and ANPG can release 1,N²-G from a 1,N²-G-monosubstituted duplex oligonucleotide and that this activity was absent in extracts of E. coli mug mutants. Furthermore, as the murine ANPG protein (APNG/Aag) (27) shares significant sequence and functional homology with ANPG (28), we have used cell-free extracts from normal and APNG-deficient mice (29) to confirm that APNG is the only 1,N²-G N-glycosylase present in mouse cells under the experimental conditions used. These observations suggest a possible role for these proteins in vivo to counteract the genotoxic effects of 1,N²-G adducts.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Oligonucleotides-- Oligonucleotides containing single 1,N²-G residues at position X were prepared as described by Langouët et al. (13). The oligonucleotides were purified by ion exchange chromatography on a 7.5 × 50 mm Vydac Ion Exchange column (The Separations Group, Hesperia, CA) using a 0-300 mM NaCl gradient in 20 mM Tris-HCl buffer (pH 8.0) (flow rate 1 ml/min). The peak fractions were collected and desalted using 1.2 × 30 cm Sephadex G-10 columns (Amersham Biosciences) using H₂O as the solvent. The recovered oligonucleotides (monitored using A₂₆₀) were ~90% pure as judged by capillary electrophoresis (Beckman PACE 5000, Beckman, Fullerton, CA) (13). Further purification was achieved by preparative electrophoresis through a 16% (w/v) acrylamide gel containing 8 M urea in 50 mM sodium 3-(N-morphilino)propanesulfonate (13). The oligomers were eluted from the gel, concentrated by centrifugal evaporation and analyzed by matrix-assisted laser desorption ionization/time-of-flight mass spectrometry using a Perseptives Voyager instrument in the Vanderbilt facility. The matrix was an 8:1 (v/v) mixture composed of (i) 3-hydroxypicolinic acid (50 mg/ml) in CH₃CN-H₂O, 1:1, v/v and (ii) aqueous ammonium citrate (50 mg/ml). The measured M_r of the G30 oligonucleotide was 9302.0 ± 1.6 (mean of five measurements, calculated 9302), and the M_r of the G31 oligonucleotide was 9503.85 ± 4.0 (mean of six measurements, calculated 9504): G30 5'-d(TGACTGCATAXGCATGTAGACGATGTGCAT and G31 5'-d(TACATCGTCACCTGGGXCATGTTGCAGATCC).

These sequence contexts were previously used to study the repair of a thymine fragmentation product (30-31) and hypoxanthine (23). Four complementary oligonucleotides containing dA, dG, dC, or T at the position opposite to 1,N²-G were purchased from Eurogentec (Angers, France). The resulting duplex oligonucleotides are referred to as G30·C(G,A,T) and G31·C(G,A,T), respectively. Oligonucleotides were 5'-end labeled by T4 polynucleotide kinase (New England Biolabs, Beverly, MA) in the presence of [-³²P]ATP (4500 Ci/mmol, ICN Pharmaceuticals, Inc., Costa Mesa, CA) or 3'-end labeled by terminal transferase (New England Biolabs) in the presence of the [-³²P]dCTP (3000 Ci/mmol, Amersham Biosciences) as recommended by the manufacturers. The oligonucleotides were then annealed to the appropriate complements as previously described (24).

Bacterial Strains-- The E. coli isogenic strains JM105 WT and MS1050 mug were from laboratory stock. Isolation of an C-DNA glycosylase (mug)-deficient mutant of E. coli K12, generated by insertion of a kanamycin cassette, will be published elsewhere. A Southern blot of genomic DNA from the mug mutant confirmed the insertion of the kanamycin cassette, whereas crude lysates exhibited no detectable C-DNA glycosylase activity (data not shown).

Enzymes-- Purification of the E. coli UNG, TagI (E. coli 3-methyladenine-DNA-glycosylase I), AlkA, MUG, Fpg (E. coli formamidopyrimidine-DNA glycosylase), Nth (E. coli endonuclease III), Nfo (E. coli endonuclease IV), and human TDG proteins was performed as described (19). The following recombinant proteins were overexpressed in E. coli BH290 (tag, alkA) cells and purified to apparent homogeneity: ANPG40 (26 kDa, 230 amino acids) (32), ANPG60 (32 kDa, 298 amino acids), ANPG70 (32 kDa, 293 amino acids) (33), and rat APDG (27.9 kDa, 253 amino acids) (34). The cDNAs encoding ANPG80, a truncated variant lacking the N-terminal 73 amino acids (24.4 kDa, 220 amino acids) (48) and hOGG1 (human 7,8-dihydro-8-oxoguanine-DNA glycosylase) were cloned into pET11a (Novagen, Madison, WI), and the proteins were overexpressed in E. coli BL21 CodonPlus cells (Stratagene, La Jolla, CA). Purification of ANPG80 was achieved using three chromatographic steps: AcA54 gel filtration (IBF, Villeneuve-la-Garenne, France), phenyl-Sepharose, and sulfopropyl-Sepharose cation exchange (Amersham Biosciences). The purification of hOGG1 and HAP-1 proteins was performed as described (35-36). Human Nth1 protein was generously provided by Dr. R. Roy (American Health Foundation, Valhalla, NY). The activity of the various proteins was tested using their well defined substrates and checked just prior to their use.

Enzyme Assays-- The release of 1,N²-G base adducts was measured by the cleavage of the oligonucleotide containing a single lesion at a defined position. The standard assay mixture for base excision activity (20 µl final volume) contained 0.2 pmol of the 5'-[³²P] or 3'-[³²P]dCMP-end labeled oligonucleotide duplex in 70 mM HEPES-KOH (pH 7.8), 100 mM KCl, 1 mM EDTA, 5 mM 2-mercaptoethanol, 100 µg of bovine serum albumin/ml, and limiting amounts of enzyme. For MUG, AlkA, and hTDG the incubation mixture was not supplemented with 100 mM KCl. For Nfo and HAP-1 the reaction buffer (20 µl) contained 20 mM HEPES-KOH (pH 7.5), 50 mM KCl, 0.1 mM EDTA (or 5 mM MgCl₂ for the HAP-1), 1 mM 2-mercaptoethanol, and 100 µg of bovine serum albumin/ml. Incubations were carried out at 37 °C for 60 min unless otherwise stated. Abasic sites were incised after glycosylase action by light piperidine treatment (10% (v/v) piperidine at 37 °C for 15 min). Reaction products were analyzed by electrophoresis through denaturing 20% (w/v) polyacrylamide gels (7 M urea, 0.5× TBE), visualized by PhosphorImager Storm 840 (Molecular Dynamics, Sunnyvale, CA), and quantified using ImageQuant software.

Electrophoretic Mobility Shift Assay (EMSA)-- The standard binding reaction mixture (20 µl) contained 70 mM HEPES-KOH (pH 7.8), 100 mM KCl, 1 mM EDTA, 5 mM 2-mercaptoethanol, 100 µg of bovine serum albumin/ml, 2 pmol G·C oligonucleotide duplex as nonspecific DNA, 0.2 pmol 5'-[³²P]-labeled 1,N²-G·C oligonucleotide duplex, and 2 pmol of a given repair protein unless otherwise stated. Following incubation on ice for 30 min, an aliquot of the reaction mixture was analyzed by electrophoresis through a 10% (w/v) nondenaturing polyacrylamide gel (29:1 acrylamide/bisacrylamide) containing 0.5× TBE at 100 V at 4 °C for 15 h. Gels were then exposed to a Storm 840 Phosphor Screen and the amount of radioactivity in the bands was quantified using ImageQuaNT^TM software.

Preparation of Cell-free Extracts-- The E. coli cell-free extracts were prepared from cultures at OD₆₀₀ = 1.0 as described previously (37). Mouse embryonic fibroblasts (MEF) were derived from APNG+/+ and APNG/ mice as previously described (29). The MEFs were grown in Dulbecco's modified Eagle's medium-F12 medium containing 10% (v/v) fetal calf serum (Invitrogen) at 37 °C and 5% CO₂ until confluent. Cells were recovered by treatment with trypsin-EDTA, pelleted by centrifugation, snap-frozen in solid CO₂ and stored at 80 °C until used. For the preparation of cell-free extracts by sonication, the cell pellets were first washed with phosphate-buffered saline and, after centrifugation, resuspended in a buffer containing 50 mM HEPES-KOH (pH 7.5), 0.1 mM EDTA, 0.5 M KCl, 5 mM 2-mercaptoethanol, 0.1% Nonidet P-40 (v/v), 5% glycerol (v/v), and a protease inhibitor mixture (Roche Molecular Biochemicals). The sonicated extract was centrifuged, and enzyme activity in the supernatant was determined as described above.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Activity of Various E. coli and Human DNA Repair Proteins on Oligodeoxynucleotides Containing 1,N²-G-- To study the repair of 1,N²-G in DNA, we investigated whether this lesion was a substrate for previously characterized DNA repair enzymes. We challenged a 5'-[³²P]-labeled oligonucleotide duplex containing 1,N²-G with a variety of highly purified base excision repair enzymes. The G30·C and G31·C duplex oligonucleotides containing 1,N²-G in a different sequence context were employed as substrates. Because not all base excision repair glycosylases possess apurinic-nicking activity, the samples were treated with piperidine after the DNA glycosylase reaction to cleave DNA at the potential abasic sites generated by the reaction. When the various E. coli and human DNA glycosylases were tested on the G31·C (Fig. 2), only incubation with MUG, ANPG70, or APDG led to the cleavage of the labeled oligonucleotide at the position of the modified base. Interestingly, the excision of 1,N²-G by the bacterial enzyme was more efficient than that of the mammalian enzymes (Fig. 2, lanes 5 and 10-11). Moreover, the human enzyme was more efficient than rat ADPG (Fig. 2, lanes 10 and 11). Despite being present in excess amount, UNG, Tag, AlkA, Fpg, Nth, Nfo, hTDG, hNth1 (human endonuclease III), hOGG1, and HAP-1 proteins did not act on the G31·C (Fig. 2, lanes 2-4, 6-9, and 12-14). Similar results were obtained with the G30·C oligonucleotide, indicating that 1,N²-G-DNA glycosylase is not affected by the sequence context (data not shown). Furthermore, the MUG and ANPG70 proteins excise 1,N²-G in a dose-dependent manner, and the release of the adduct was linear as a function of time during the first 20 and 60 min of incubation, respectively (data not shown).

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Activity of various E. coli and human DNA repair proteins on the 31-mer G31·C duplex oligonucleotide. 5'-[32P]-labeled G31·C (5 nM) was incubated with each purified repair protein (100 nM) at 37 °C for 60 min unless otherwise stated. The products of reactions were subsequently subjected to light piperidine treatment to reveal abasic sites generated by DNA glycosylases devoid of -lyase activity. Lane 1, control G31·C-oligonucleotide; lane 2, as 1 but treated by Ung; lane 3, TagI; lane 4, AlkA; lane 5, MUG, 20 min; lane 6, Fpg; lane 7, Nth; lane 8, Nfo, 20 min; lane 9, hTDG, 30 °C; lane 10, ANPG70; lane 11, APDG; lane 12, hNth1; lane 13, hOGG1; and lane 14, HAP-1, 20 min. The products of the reaction were analyzed as described under "Experimental Procedures." a, 31-mer oligonucleotide; b, 16-mer.

Activity of Various Forms of ANPG on Oligodeoxynucleotides Containing 1,N²-- GBecause ANPG may exist in a number of alternative forms in human cells, resulting from differential splicing of the primary transcript, we studied the release of 1,N²-G by various forms of the ANPG protein. As shown in Fig. 3, MUG, ANPG-70, -60, and -40 (lanes 2-4, 6, 8-10, and 12) excise 1,N²-G when it is present in the G30·C and G31·C duplex oligonucleotides, generating 10- and 16-mer DNA fragments, respectively. Surprisingly, ANPG80, although used in excess amounts to detect even marginal activity, did not act on the G30·C and G31·C duplexes under our experimental conditions (Fig. 3, lanes 5 and 11).

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Activity of various human ANPGs on the duplex RT/A oligonucleotides containing 1,N2-G. 5'-[32P]-labeled G31·C or G30·C (10 nM) was incubated with various pure human ANPG or MUG at 37 °C for 60 and 20 min, respectively. The reaction products were subjected to light piperidine treatment to reveal abasic sites generated by DNA glycosylases devoid of -lyase activity. Lane 1, control, G31·C duplex oligonucleotide; lane 2, as lane 1 but treated by ANPG70, 160 nM; lane 3, ANPG60, 160 nM; lane 4, ANPG40, 160 nM; lane 5, ANPG80, 320 nM; lane 6, MUG, 80 nM; lane 7, control G30·C duplex oligonucleotide; lane 8, as lane 7 but treated by ANPG70, 160 nM; lane 9, ANPG60, 160 nM; lane 10, ANPG40, 160 nM; lane 11, ANPG80, 320 nM; lane 12, MUG, 80 nM. The products of the reaction were analyzed as described under "Experimental Procedures." a and b indicate the 31- and 30-mer oligonucleotides; c, 16-mer; d, 10-mer.

EMSA of Various DNA Repair Proteins with a 1,N²-G-Containing Duplex Oligodeoxynucleotide-- We examined the interaction of the G31·C-containing duplex oligonucleotide with, MUG, hTDG, AlkA, ANPG-70, -60, -40, and -80, Nfo, Nth, hNth1, and hOGG1 using EMSA, which can be used to detect the difference in migration pattern between free and protein-bound DNA. To avoid nonspecific interactions the standard binding mixture contained a 17-fold excess of albumin and a 10-fold excess of unlabeled, non-modified 34-mer G·C duplex oligonucleotides. As shown in Fig. 4 the MUG protein interacted in a highly specific manner with G31·C-containing DNA, with more than 80% of the labeled DNA present as the bound form (lane 2). In contrast, Nfo and the various forms of ANPG interacted very weakly with G31·C, because less than 10% of labeled DNA was present in bound form (Fig. 4, lanes 5-9). It should be noted that the full-length ANPG70 and truncated ANPG80 proteins bound to the G31·C duplex with the same efficiency. This observation ruled out the possibility that the inability of the ANPG80 protein to catalyze glycosidic bond hydrolysis of the 1,N²-G adduct in DNA resulted from a failure of the enzyme to bind to this substrate. As expected the AlkA, hTDG, Nth, hNth1, and hOGG1 proteins did not form specific complexes with G31·C-containing DNA. Interestingly, although Nfo protein does not excise the 1,N²-G residue in DNA it binds to G31·C duplex oligonucleotide in our assay condition. To demonstrate binding specificity, increasing amounts of either unlabeled nonspecific G·C or specific G31·C duplex oligonucleotides were titrated into binding reactions of the MUG, ANPG70, ANPG80, and Nfo proteins with the ³²P-labeled 1,N²-G oligonucleotide. As shown in Fig. 5 (lane 12), a 10-fold molar excess of the cold G31·C duplex oligonucleotide efficiently inhibits complex formation by the MUG protein, whereas little inhibition was observed with the same amount of the cold G·C DNA. The binding of the ANPG70, ANPG80, and Nfo proteins to the 1,N²-G was inhibited equally well by the cold G·C and G31·C duplex oligonucleotides (data not shown), suggesting that the interactions observed in Fig. 4 are rather nonspecific. In conclusion, the data show that among proteins tested only MUG interacts in a highly specific manner with the 1,N²-G adduct in DNA.

View larger version (87K): [in this window] [in a new window]   Fig. 4.   Electrophoretic mobility shift assay of various DNA repair proteins with 1,N2-G-containing duplex oligonucleotide. 5'-[32P]-labeled G31·C duplex oligonucleotide (10 nM) was incubated in the presence of a pure repair protein (100 nM) at 4 °C for 30 min. Lane 1, control G31·C; lane 2, as lane 1 but in presence of MUG; lane 3, hTDG; lane 4, AlkA; lane 5, ANPG70; lane 6, ANPG60; lane 7, ANPG40; lane 8, ANPG80; lane 9, Nfo; lane 10, Nth; lane 11, hNth1; lane 12, hOGG1. The reaction products were analyzed as described under "Experimental Procedures." a, enzyme-bound oligonucleotide; b, free oligonucleotide.

View larger version (36K): [in this window] [in a new window]   Fig. 5.   Specificity of MUG protein binding to the 1,N2-G adduct in DNA. Unlabeled duplex oligonucleotide (with or without adduct) was titrated into binding reaction mixtures containing 10 nM of 5'-[32P]-labeled G31·C duplex oligonucleotide and 100 nM of MUG at 4 °C for 10 min. Lane 1, control G31·C; lane 2, as lane 1 but in presence of MUG; lanes 3-8, as lane 2 but in presence of 25, 50, 100, 200, 400, and 600 nM unlabeled G·C oligonucleotide; lane 9, as lane 1 but in presence of MUG; lanes 10-15, as lane 9 but in presence of 25 nM, 50 nM, 100 nM, 200 nM, 400 nM, and 600 nM unlabeled G31·C oligonucleotide. The reaction products were analyzed as described under "Experimental Procedures." a, enzyme-bound oligonucleotide; b, free oligonucleotide.

Kinetic Parameters of Release of 1,N²-G from Duplex DNA by MUG and Anpg-70, -60, and -40-- Based upon the above qualitative observations, we further investigated the efficiencies of MUG and the various active forms of ANPG to recognize 1,N²-G in DNA. The K_m and k_cat values and the k_cat/K_m ratios for the MUG and ANPG-70, -60, and -40 proteins acting on 1,N²-G are shown in Table I. For comparison, the respective kinetic constants for these proteins acting on their other well characterized substrates (C and A) are also presented. Interestingly, the apparent K_m for MUG, calculated from the Lineweaver-Burk plot (11 nM), suggests that MUG has stronger affinity than does the ANPG proteins (Table I). The comparison of the kinetic constants for the excision of 1,N²-G shows a dramatic difference among the four enzymes used, the bacterial one being by far the most efficient, whereas the efficiency of repair by ANPG proteins is much lower. Moreover, the analysis of the kinetic constants of the various forms of human enzymes suggests that the biologically active protein for the excision of 1,N²-G in vivo is most probably the full-length and splicing variant proteins. However, it is also worth noting that the kinetic constants of the MUG and ANPG proteins for excision of the C and A, respectively, are more efficient.

View this table: [in this window] [in a new window]   Table I Kinetic constants of the E. coli MUG and various human ANPG proteins for the excision of -bases

Base-Pair Specificity of the MUG and ANPG70 Proteins-- We investigated the specificity of MUG and ANPG70 when acting on various 1,N²-G oligonucleotides containing each of the four naturally occurring deoxynucleotides opposite 1,N²-G. Fig. 6 presents the initial velocity of the two enzymes when acting upon the four different base pairs. Both MUG and ANPG70 preferentially excise the 1,N²-G adduct when it is opposite a dC. All of the duplexes containing G31·dN base pairs are substrates for MUG, whereas ANPG70 does not excise 1,N²-G when it is opposite dG. The relative order of -adduct excision for the MUG protein was dC > dG  dA > T, and for ANPG70 it was dC > T  dA dG.

View larger version (13K): [in this window] [in a new window]   Fig. 6.   Cleavage of duplex oligonucleotides containing different bases opposite 1,N2-G by MUG and ANPG70. The oligonucleotide-containing 1,N2-G was annealed to various complements to generate the following mismatches: G31·dC, G31·dG, G31·dA, and G31·T. These were used as substrates for MUG (A) and ANPG70 (B). The excision of 1,N2-G was measured by determining the amount of oligomer migrating to the position of the 16-mer using 10 nM oligonucleotide and limiting amounts of enzyme. For details see "Experimental Procedures."

Lack of 1,N²-G-Oligonucleotide Nicking by Extracts of E. coli mug Mutants and APNG/ MEFs-- To confirm that MUG and APNG are the major 1,N²-G-releasing activities in E. coli and mouse cells, we prepared cell-free extracts from a mug mutant of E. coli and MEFs derived from APNG+/+ and APNG/ mice (Fig. 7). To exclude the possibility that the observed reaction products resulted from the activity of an unknown double-stranded 3'-5'-exonuclease that is blocked by 1,N²-G, the G31 oligonucleotide was radioactively labeled on the 3' side with [-³²P]dCTP using terminal transferase, and this was used as the substrate in addition to the 5'-labeled G31 oligonucleotide (not shown). As shown in Fig. 7, lanes 2 and 4, the oligonucleotide is not degraded, showing that an enzyme(s) in cells excise(s) the 1,N²-G residues by an N-glycosylase/AP-endonuclease mechanism (lane 6). The oligonucleotide nicking by the bacterial extract (lane 2) was again more efficient than that of the mammalian extract (lane 4). As expected, we did not observe cleavage of the G31·C oligonucleotide in cell extracts from E. coli mug and APNG/ MEFs (lanes 3 and 5). These results show that the MUG and APNG proteins are indeed the enzymes that excise the 1,N²-G in E. coli and mouse cells.

View larger version (20K): [in this window] [in a new window]   Fig. 7.   Cleavage of a duplex oligonucleotide containing 1,N2-G by cell-free extracts of E. coli and MEFs. 3'-[32P]dCMP-labeled G31·C duplex oligonucleotide 10 nM was incubated with cell-free extracts from E. coli and MEFs at 37 °C for 30 and 120 min, respectively. Lane 1, control G31·C; lane 2, as lane 1 but incubated in the presence of 10 µg of extract from E. coli JM105 (mug+); lane 3, 10 µg of E. coli MS1050 (mug::KnR); lane 4, 100 µg of MEF extract (APNG+/+); lane 5, 100 µg of MEF extract (APNG/); lane 6, 10 ng of each purified MUG and Nfo proteins. The products of the reaction were analyzed as described under "Experimental Procedures." a, 32-mer oligonucleotide; b, 15-mer.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The results presented herein show that the E. coli MUG and ANPG proteins excise 1,N²-G when present in duplex oligonucleotides. The human monofunctional ANPG is the only glycosylase that excises N-methylpurines (32, 38, 39), A (16, 18), and hypoxanthine (23) in human cells. Similarly, experiments on mouse cells lacking APNG have demonstrated that it has a similar substrate specificity (26, 29, 40). The role of the MUG in the repair of C (19) and mismatched uracil and thymine bases (19, 41, 42) is well documented.

Little is known about the biological consequences of 1,N²-G adduct persistence in DNA. However, previously reported properties of the lesion (7, 14) and the demonstration that it is repaired by the MUG and ANPG/APDG/APNG proteins in vitro with significant efficiency (present study) strongly suggest that such a modified base cannot be tolerated in vivo. The present data add a new substrate to the previously reported ethenobases, suggesting that an additional function of MUG, ANPG, and APDG is in the repair of 1,N²-G (Fig. 2). Also, this is a first example of a shared substrate specificity for these enzymes and suggests that some structural similarity in the conformation of the active site could exist between the MUG/TDG family and the mammalian ANPGs as well as similarity in the mechanism of damaged base recognition.

Although the core portions of the ANPG70 and APDG proteins display 85% sequence identity (32), it seems that the human enzyme repairs 1,N²-G somewhat more efficiently than its rat counterpart (Fig. 2). In fact, a similar difference between human and mouse ANPG in the recognition of 3-methylguanine and 7-methylguanine residues has been reported (43). Interestingly, the human full-length and splice variants of ANPG have a direct repeat in the N-terminal amino acid sequence, which is absent in the rat and murine homologues. Therefore, the results presented above suggest that the nonconserved N-terminal part of mammalian ANPG might play a significant role in substrate specificity.

Cloning of the ANPG cDNA revealed that the protein could be present as several alternatively spliced forms, including a truncated version (32, 38-39, 44), and it has been concluded that the nonconserved, N-terminal part apparently contributes little to their damage recognition and N-glycosylase activity (33, 45). Surprisingly, therefore, we have shown that ANPG80, which lacks 73 amino acid residues at the N terminus, is unable to excise 1,N²-G from the duplex oligonucleotide (Fig. 3). Using EMSA, we have demonstrated that the inability of the ANPG80 protein to catalyze glycosidic bond hydrolysis of the 1,N²-G residue in DNA is not due to a failure of the protein to bind to this substrate (Fig. 4). Instead, a change in active site conformation or lack of amino acid residues involved in catalysis must be involved. Interestingly, ANPG40, which lacks the first 70 amino acid residues was found to be less active compared with the full-length and splice variant proteins (ANPG60 and ANPG70, Table I). Taken together these observations indicate that the nonconserved, N-terminal part of ANPG is essential for 1,N²-G-glycosylase activity but is dispensable for the release of A, hypoxanthine, and N-methylpurines.

The k_cat/K_m ratio of MUG for 1,N²-G was calculated to be 0.95 × 10³ min¹ nM¹, whereas ratios for the full-length and active truncated ANPG proteins were 10- and 50-fold lower, respectively (Table I). Interestingly, in contrast to the MUG protein the interaction of ANPG and Nfo proteins with the 1,N²-G adduct in DNA is very weak and rather nonspecific (Fig. 5 and data not shown). These indicate that the bacterial enzyme is the more efficient and that the efficiency of repair by the ANPG proteins is low. The comparison of the kinetic data for the excision of 1,N²-G by various forms of ANPG proteins suggests that the adduct could be repaired by the ANPG-70 and -60 proteins in vivo.

The preferential recognition of a modified base paired with one of the four naturally occurring bases in duplex DNA is an important characteristic of DNA glycosylases. As shown previously for MUG and ANPG, C and A are released irrespective of the opposing base (18, 23, 46-47). This suggests that these DNA repair proteins recognize ethenobases per se via the flip-out mechanism. Here we have shown that MUG preferentially excises 1,N²-G when it is opposite dC, followed by dG, dA, and T (dC > dG  dA > T) (Fig. 6A). Similarly, ANPG70 preferentially excises 1,N²-G when it is opposite dC, but followed by T and dA (dC > T  dA dG) (Fig. 6B). Unexpectedly, we found that the ANPG70 protein is unable to excise 1,N²-G when it is opposite dG (Fig. 6B). This differs from the excision pattern of the ANPG70 protein reported previously for hypoxanthine and A (18, 23, 46). However, recent studies of base pair specificity of the ANPG protein has demonstrated that the structure of the base pair rather than simply the damaged DNA base plays a role in the mechanism of recognition and excision by the enzyme (47). Therefore, we propose that the guanine opposite 1,N²-G might affect excision by influencing the alignment of the flipped base in the active site of the enzyme. Thus, improper alignment results in the inability of the ANPG protein to form a transition state intermediate structure upon the 1,N²-dG nucleotide, in which a substantial positive charge accumulates on the deoxyribose sugar and leads to glycosylic bond cleavage.

The crystal structures of ANPG80 complexed to DNA duplex containing pyrrolidine and A have been published (48-49). ANPG80 is a single domain protein of mixed / structures that forms a distinct structural group that does not resemble any other base excision repair protein (50). The enzyme bends the DNA by about 20 ° and intercalates into the minor groove of DNA causing the abasic pyrrolidine and A to flip into the enzyme active site. The structure of the ANPG80 nucleotide-binding pocket changes little upon flipping in the A base.

The crystal structures of MUG and MUG complexed with an oligonucleotide containing a non-hydrolyzable deoxyuridine analog mismatched with guanine have also been published (51-52). The MUG structure contains 5-stranded  sheets at the center flanked by  helices, thus forming a -- topology. The structure of DNA-protein complex reveals an essentially nonspecific pyrimidine-binding pocket that allows MUG/TDG enzymes to excise the  base C.

The structure of the base-binding pocket of MUG and ANPG reveals the lack of specific interactions for methylated bases. This probably contributes to their remarkable multifunctionality, providing deamination, methylation, and exocyclic base damage glycosylase activity within the same enzyme. Moreover, the specific interaction of both the MUG and ANPG proteins with 1,N²-G when present in DNA suggests that these enzymes are similar with respect to the conformation of their active site pockets. This hypothesis is supported by the computer modeling of C in the active site of ANPG, which reveals a good fit of the -base into the active site of the enzyme despite its lack of activity (53). Surprisingly, neither hTDG (the human structural homologue of MUG) nor AlkA (the E. coli functional homologue of ANPG) excises or specifically binds 1,N²-G when present in DNA (Figs. 2 and 4). Co-crystallization of MUG or ANPG with 1,N²-G-containing DNA might lead to a better understanding of the structural requirements for recognition and catalytic mechanisms.

The murine APNG protein is a structural homolog of the ANPG protein (31), and both proteins have a similar substrate specificity (28, 43). Therefore, to further substantiate the role of MUG and APNG in the release of 1,N²-G, we tested the activity in cell-free extracts from an E. coli mug mutant and ANPG/ MEFs (Fig. 7). The results clearly demonstrate that the incision of the 1,N²-G-containing duplex oligonucleotide in cells extracts from E. coli and MEFs has an absolute requirement for the mug and aag genes, respectively. These observations indicate that MUG and APNG are the only detectable DNA glycosylases excising 1,N²-G adducts in bacteria and mammalian cells, respectively. Again it appears that the bacterial enzyme is more efficient than its mammalian counterpart.

Pyrimido[1,2-]purin-10(3H)-one (M₁G) (Fig. 1), a structural six-membered ring homologue of 1,N²-G, has been detected in liver, white blood cells, pancreas, and breast from healthy human beings at levels ranging from 1-120 per 10⁸ nucleotides (54). Because MUG and ANPG can excise various -adducts they may also play a role in the repair of other exocyclic adducts such as M₁G.

In conclusion, the putative role for the MUG, ANPG, APDG, and APNG proteins as 1,N²-G-DNA glycosylases based on biochemical and genetic studies has been determined. Although 1,N²-G has not yet been identified as a biologically relevant lesion, the potential biological relevance of this adduct is supported by the fact that the excision of 1,N²-G by MUG and ANPG is effected at rates comparable with the repair of C and A lesions.

	ACKNOWLEDGEMENT

We thank Dr. R. Roy for human Nth1 protein.

	FOOTNOTES

^* This research was supported by European Community Grant QLK4-2000-00286 (to M. S. and R. H. E.), the Association pour la Recherche sur le Cancer and Fondation Franco-Norvegienne pour la Recherche Scientifique et Technique et le Développement Industriel (to J. L.), and by United States Public Health Service Grants R35 CA44353, P30 ES00267, and R01 ES10375 (to F. P. G.).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: 33-1-42115404; Fax: 33-1-42115276; E-mail: smurat@igr.fr.

Supported by a postdoctoral fellowship from the Fondation pour la Recherche Médicale. Present address: Dept. of Biochemistry and Molecular Biology, University College, London, UK.

To whom questions concerning DNA synthesis and characterization should be addressed. Tel.: 615-322-2261; Fax: 615-322-3141; E-mail: guengerich@toxicology.mc.vanderbilt.edu.

^§§ Present address: Pharmacokinetics, Dynamics, and Metabolism Pfizer Global R & D, Ann Arbor, MI 48105.

Published, JBC Papers in Press, May 16, 2002, DOI 10.1074/jbc.M111100200

² M. Saparbaev, S. Langouët, C. V. Privezentzev, F. P. Guengerich, H. Cai, R. H. Elder, and J. Laval, unpublished observations.

	ABBREVIATIONS

The abbreviations used are: , etheno; C, 3,N⁴-ethenocytosine; A, 1,N⁶-ethenoadenine; N², 3-G, N²,3-ethenoguanine; 1, N²-G, 1,N²-ethenoguanine; ANPG, human alkylpurine-DNA N-glycosylase; MUG, mismatch-specific uracil-DNA glycosylase; hTDG, human thymine-DNA glycosylase; AlkA, E. coli 3-methyladenine-DNA-glycosylase II; EMSA, electrophoretic mobility shift assay; MEF, mouse embryonic fibroblasts; APNG, murine ANPG; APDG, rat ANPG.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES