J Biol Chem, Vol. 275, Issue 14, 9924-9929, April 7, 2000

Role of the N-terminal Proline Residue in the Catalytic Activities of the Escherichia coli Fpg Protein^*

Olga M. Sidorkina and Jacques Laval

From the Groupe "Réparation des Lésions Radio- et Chimio- Induites," UMR 8532 Centre National de la Recherche Scientifique, Institut Gustave Roussy, 39, Rue Camille Desmoulins, 94805 Villejuif Cédex, France

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The Escherichia coli Fpg protein is a DNA glycosylase/AP lyase. It removes, in DNA, oxidized purine residues, including the highly mutagenic C8-oxo-guanine (8-oxoG). The catalytic mechanism is believed to involve the formation of a transient Schiff base intermediate formed between DNA containing an oxidized residue and the N-terminal proline of the Fpg protein. The importance and the role of this proline upon the various catalytic activities of the Fpg protein was examined by targeted mutagenesis, resulting in the construction of three mutant Fpg proteins: Pro-2  Gly (FpgP2G), Pro-2  Thr (FpgP2T), and Pro-2  Glu (FpgP2E). The formamidopyrimidine DNA glycosylase activities of FpgP2G and FpgP2T were comparable and accounted for 10% of the wild-type activity. FpgP2G and FpgP2T had barely detectable 8-oxoG-DNA glycosylase activity and produced minute Schiff base complex with 8-oxoG/C DNA. FpgP2G and FpgP2T mutants did not cleave a DNA containing preformed AP site but readily produced Schiff base complex with this substrate. FpgP2E was completely inactive in all the assays. The binding constants of the different mutants when challenged with a duplex DNA containing a tetrahydrofuran residue were comparable. The mutant Fpg proteins barely or did not complement in vivo the spontaneous transitions G/C  T/A in E. coli BH990 (fpg mutY) cells. These results show the mandatory role of N-terminal proline in the 8-oxoG-DNA glycosylase activity of the Fpg protein in vitro and in vivo as well as in its AP lyase activity upon preformed AP site but less in the 2,6-diamino-4-hydroxy-5-N-methylformamidopyrimidine-DNA glycosylase activity.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

7,8-Dihydro-8-oxoguanine (8-oxoG)¹ is the major mutagenic base lesion generated in DNA by active oxygen during normal metabolism (1). Because 8-oxoG pairs preferentially with adenine rather than cytosine, it generates transversion mutations after replication (2). To maintain the genetic integrity, this oxydized base is repaired, in Escherichia coli, by the Fpg protein (2, 3) coded for by the fpg gene (4). The physical and enzymatic properties of the protein have been established (5). It is a globular monomer of 30.2 kDa (269 amino acids), pI = 8.6, that contains one zinc atom/protein molecule present in a zinc finger motif (5). In vitro, the Fpg protein exhibits three catalytic activities: (i) it is a DNA glycosylase that liberates modified residues such as 2,6-diamino-4-hydroxy-5-N-methyl-formamido pyrimidine (Fapy) (6), 8-oxoG (3, 7), 8-oxoA, the imidazole ring-opened forms of adenine and guanine (7), and modified pyrimidines, such as 5-hydroxycytosine, 5-hydroxyuracil, -ureidoisobutiric acid, -R-hydroxy--ureidoisobutiric acid (8) (for review, see Ref. 9), (ii) it is an AP lyase that incises DNA at abasic sites by a --elimination mechanism (10, 11), and (iii) it is a deoxyribophosphodiesterase that removes 5'-terminal deoxyribose phosphate residues (12). Therefore the Fpg protein belongs to the class of DNA glycosylases endowed with a -lyase activity (9). It was hypothesized that the glycosylase and lyase functions of DNA glycosylases/AP lyases are mechanistically coupled (13-15). A mechanism involving the nucleophilic attack on C1' of the modified deoxynucleoside targeted for excision has been proposed (13-15) to explain the catalytic action of this class of enzymes. In this mechanism, the catalytic nucleophile attacks the glycosidic bond, displacing the aberrant base and forming a transient intermediate (Schiff base) with the C1'-deoxyribose moiety. The isomerization of this transient complex leads to cleavage of the AP site through a conjugate elimination mechanism. It has been proposed that most DNA glycosylases/AP lyases use the - NH₂ group of a lysine as the catalytic nucleophile (16). For the Nth, hOgg1, and hNth proteins, the lysine residues at positions 120, 249, and 212, respectively, are believed to be the catalytic nucleophiles (16-18). However, the NH₂ group of the N-terminal amino acid can be the catalytic amine, as shown in the case of the bacteriophage T4 endonuclease V (19). For the Fpg protein, the N-terminal proline residue was shown to be linked, in a Schiff base complex, with DNA containing 8-oxoG residues. It was suggested that this proline was the nucleophile that initiates the catalytic excision of 8-oxoG (20).

The N-terminal formylmethionine of the cloned wild-type Fpg protein is post-translationally cleaved in E. coli (4). However, because in some mutant Fpg proteins, the formylmethionine could not be cleaved, the amino acid sequence of the Fpg protein was numbered Met-1, Pro-2, etc. (21, 22).

To further understand the role of the N-terminal proline, in the various catalytic activities of the Fpg protein, Pro-2 was mutated to three different amino acids: Gly, Thr, or Glu generating the FpgP2G, FpgP2T, and FpgP2E mutant proteins, respectively. These proteins were purified, and their catalytic properties toward several substrates as well as their binding constants were established. Finally, we have determined the in vivo properties of the mutant Fpg proteins. Our results show that Pro-2 is involved in the various enzymatic activities of the Fpg protein, although at different extent, and that one can observe the formation of a robust Schiff base intermediate with an oligonucleotide containing an abasic site despite the lack of a detectable AP lyase activity.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Bacterial Strains, Enzymes, and Chemicals-- The following strains, derived from the E. coli K12 strain, were used: CC 104 (ara (gpt-lac)₅, rpsL/F' (lacI₃₇₈,lacZ₄₆₁, proA⁺B⁺] (23), BH 990 (as CC 104 but fpg::kan^R,mutY::kan^R) (24), BH 20 (as AB 1157 but fpg::kan^R) (25), and DH5 (26). The plasmids containing the cloned fpg-gene pFPG60 (4), pFPG220 (27), and pUC19 were from laboratory stock.

Restriction endonucleases, DNA polymerase, DNA ligase, bovine serum albumin (molecular biology grade), and other DNA modifying enzymes were obtained from F. Hoffmann-La Roche (Switzerland), Taq and Pfu DNA polymerases were obtained from Amersham Pharmacia Biotech and Stratagene (La Jolla, CA), respectively, and used as recommended by the manufacturers.

[³H]Dimethylsulfate (3.9 Ci/mmol) was purchased from NEN Life Science Products. [³H]Thymidine (48 Ci/mmol), [-³²P]dATP (3000 Ci/mmol) were from Amersham Pharmacia Biotech. [-³²P]ATP (3000 Ci/mmol) and [ -³⁵S]dATP (1000 Ci/mmol) were from ICN (Paris, France). The substrates [³H]Fapy-poly(dG-dC) and [³H]thymine E. coli DNA containing abasic sites were prepared as described (4, 10). The preparation of plasmids, restriction enzyme digestion, agarose gel electrophoresis, DNA ligation, and bacterial transformation were performed using standard methods as described by Sambrook et al. (26) or by the manufacturers.

Site-directed Mutagenesis-- The plasmid pFPG60 was used for the construction of mutated fpg genes. The site-directed mutagenesis was achieved by polymerase chain reaction using Taq and Pfu DNA polymerases. The primers used to generate amino acid modifications of the Fpg protein are listed in Table I and were obtained from Genset (Paris, France). To obtain an unique restriction site HindIII close to the N terminus of the coding fpg sequence, the plasmid pFPG7 was constructed by producing a 156-base pair polymerase chain reaction product (primers 5'-FPG (HindIII site), reverse (Table I), that was restricted with BglII and HindIII, and then cloned into the BglII/HindIII-cut pFPG60. pFPG7 was used to produce the mutant Fpg proteins with amino acid changed Pro-2  Gly, Pro-2  Glu, and Pro-2  Thr. The corresponding plasmids are pFPG7P2G, pFPG7P2E, and pFPG7P2T. The DNA sequence was verified by the dideoxy chain termination method with T7 DNA polymerase (Sequenase). To further characterize the mutant proteins, the sequence of the six N-terminal amino acids of each protein was determined (data not shown).

In Vitro Repair Assay-- The procedures used to purify the Fpg mutant proteins, assay of Fapy- or 8-oxoG-DNA glycosylase activities, nicking activity of DNA at AP sites, and formation of enzyme-DNA covalent complexes in the presence of NaBH₄, have been described (21). For DNA binding determination, the gel mobility schift assays, using radiolabeled duplex oligomer containing a single synthetic abasic site analog (tetrahydrofuran) (23F:23C; Table I) were performed as described (28).

In Vivo Repair Assay-- The mutator strain BH990 (fpg mutY) was transformed by pFPG7P2T, pFPG7P2G, or pFPG7 using electroporation. The transformation mixture was plated onto LB medium containing ampicillin (100 µg/ml). Individual colonies (4-6) were selected, grown overnight at 37 °C in 2 ml of LB broth medium containing ampicillin and then plated onto M9 minimal medium containing lactose (0.4%) and ampicillin (100 µg/ml) at a density of 200-500 colonies/plate. The spontaneous mutation rate in the cultures was monitored for the generation of Lac⁺ revertants (21, 24).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We have constructed various E. coli mutant Fpg proteins to ascertain the involvement of the N-terminal proline in the various enzymatic activities of this protein. Proline 2 was mutated to Gly, Thr, or Glu. Gly and Thr were chosen for substitution because they are amino acids with uncharged polar R-group and minimal differences (29, 30). Because in the wild-type Fpg protein, the primary structure of N-terminal side is PELPVET ... , we examined the effect of an additional negative charge when Pro-2 is changed to Glu and constructed FpgP2E.

The mutant proteins were expressed in E. coli BH20 (fpg) cells to avoid any contamination by the wild-type enzyme. The amount of overproduced mutant enzymes was comparable with that of wild type, as evaluated by the abundance of the 30.2-kDa protein in PAGE gels (data not shown). In crude extracts of BH20 E. coli cells expressing the mutant FpgP2G and FpgP2T, a Fapy-DNA glycosylase activity was easily detected, but to a lower level than in cells expressing the wild-type protein. No Fapy-DNA glycosylase activity could be detected in crude extracts of BH20 E. coli cells expressing the Fpg P2E mutant (Table II). In contrast, when apurinic [³H]thymine-labeled DNA was used as substrate, a barely detectable AP nicking activity was observed in all the extracts (Table II) in the same order of magnitude as the control and was presumably due to Xth, Nth, or Nei proteins present in the extracts.

Fapy- and 8-oxoG-DNA Glycosylase Activities of Wild-type and Mutant Purified Fpg Proteins-- After purification of the wild-type and mutant Fpg, the fractions obtained after MonoS chromatographic purification step were better than 95% of homogeneity (data not shown) and were used to determine the various enzymatic properties of the mutant Fpg proteins. To further characterize these enzymes, the N-terminal amino acids were microsequenced.

Because the efficacy of the post-translational cleavage of the formylmethionine within the cell is determined largely by the amino acid next to the formylmethionine in the protein (31), it was possible that such cleavage does not occur in cells expressing the mutant enzymes. In fact, the N-terminal methionine was absent, as shown by microsequencing, in the FpgP2G and FpgP2T purified proteins. However in the FpgP2E, the formylmethionine was present (data not shown). None of the various enzymatic properties of the Fpg protein measured below were detectable using the FpgP2E mutant protein (data not shown).

The ability of FpgP2G and P2T to excise Fapy residues was analyzed. The comparison of the activity of the Fpg wild-type and mutant proteins showed that FpgP2G and FpgP2T had a comparable but low (about 10% of wild type) Fapy-DNA glycosylase activity (Fig. 1A).

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Fapy-DNA glycosylase and 8-oxoG-DNA glycosylase activities of wild-type and mutant Fpg proteins. A, the Fapy-DNA glycosylase activity was measured by incubating increasing amounts of Fpg proteins: Fpg wild type (), FpgP2G (), and FpgP2T () with [3H]Fapy-[poly(dG-dC)] for 10 min at 37 °C. The [3H]Fapy residues released were quantified (for details see "Experimental Procedures"). B, the 8-oxoG-DNA glycosylase activity was measured by incubating 90 fmol of 8-oxoG containing oligonucleotide for 10 min at 37 °C with wild-type () or mutant FpgP2G () or FpgP2T () proteins. The reaction products were analyzed by 20% PAGE containing 7 M urea. The radioactivity of the bands corresponding to the native or incised oligonucleotide was measured. Results are the means ± S.D. of three independent experiments.

The efficiency of Fpg mutant proteins to repair 8-oxoG was investigated using as substrate a double-stranded oligonucleotide containing a single 8-oxoG (Table I). As shown in Fig. 1B, the rate of incision of the duplex was dramatically reduced for the P2G or P2T mutants as compared with the wild-type Fpg protein (about 3%). It was verified that the excision of 8-oxoG residues was not underestimated, because, following the action of the enzyme, in the presence or absence of 10% piperidine (which nicks DNA at AP sites), the same level of incision of the substrate was observed (data not shown). These data indicate that the N-terminal proline is mandatory for 8-oxoG but less for Fapy residues excision.

Nicking Activity of Wild-type and Mutant Fpg Proteins at Abasic Site-- The ability of P2G and P2T mutant proteins to cleave abasic sites was evaluated using two different substrates. In the first assay, E. coli [³H]thymine-labeled DNA containing AP sites was used as substrate. In this assay, the Fpg protein nicks at AP sites and liberates acid-soluble [³H]thymine-labeled short oligonucleotides. Using the purified mutant Fpg proteins, no detectable AP nicking activity was observed for any of the mutant proteins (Fig. 2A). In the second assay, a 50-mer oligonucleotide with an unique preformed abasic site at a defined position was used as substrate (Table I). The P2G mutant had less than 1% of the wild-type activity (Fig. 2B), generating only -elimination pruducts (data not shown). All together, these experiments suggest that the N-terminal proline is mandatory for the AP lyase activity when measured upon preformed AP site.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   -Lyase activity of wild-type and mutant Fpg proteins. A, increasing amounts of the different Fpg proteins (wild type (), mutant FpgP2G (), and FpgP2T ()) were incubated with [3H]thymine. E. coli DNA containing AP sites and the released radioactivity of 3H-labeled short oligonucleotides was measured (for details see "Experimental Procedures"). B, a 50-mer oligonucleotide containing an abasic site was treated with the wild-type Fpg () or FpgP2G () proteins. The reaction products were analyzed by a 20% PAGE, 7 M urea gel. The bands corresponding to native and incised DNA were quantified.

Formation of Schiff Base Intermediate-- The Fpg protein forms a transient Schiff base intermediate with 8-oxoG/C containing DNA in a stoichiometry 1:1 (32). This intermediate, after reduction with sodium cyanoborohydride or sodium borohydride, is converted to a covalently linked Fpg protein-DNA complex (32, 33). We have tested the ability of the mutant FpgP2G and FpgP2T proteins to trap 8-oxoG/C DNA in the presence of NaBH₄. The trapping efficiency of FpgP2G and FpgP2T was dramatically reduced (Fig. 3A). This result is in agreement with the barely detectable 8-oxoG-DNA glycosylase activity (Fig. 1B).

View larger version (33K): [in this window] [in a new window]   Fig. 3.   A, NaBH4 trapping of 8-oxoG-containing DNA with wild-type and mutant Fpg proteins. 50 fmol of 8-oxoG oligonucleotide (34-mer) were incubated with increasing concentrations of wild-type Fpg protein ()) or mutant FpgP2G () and FpgP2T () in the presence of 100 mM NaBH4. The reaction products were analyzed by 12% SDS-PAGE. The results represent the means ± S.D. from three independent experiments. B, NaBH4 trapping of an oligonucleotide with an unique abasic site with Fpg wild type or mutant FpgP2G. 100 fmol of oligonucleotide (50-mer) containing a single dIMP residue was treated with AlkA protein to produce a single abasic site. Lanes 1-7, incubated with 100 mM NaBH4. Lane 1, no further treatment; lanes 2-4, with increasing concentrations of FpgP2G (2, 5, and 20 ng, respectively); lanes 5-7, with wild-type Fpg (2, 5, and 20 ng, respectively). The reaction products were analyzed by 12% SDS-PAGE. For details see "Experimental Procedures."

In contrast, when the AP site containing DNA is used as substrate, the mutant P2G produces a Schiff base intermediate with an even better efficiency than that of the wild-type protein (Fig. 3B). However, FpgP2G has an almost indetectable AP lyase activity upon preformed AP site (Fig. 2). These results show that it is possible to produce a robust Schiff base intermediate using an enzyme having an impaired DNA repaired activity, i.e. no AP lyase activity.

DNA Binding of Mutant Fpg Proteins-- The purified mutant FpgP2G and P2E proteins were tested for their ability to bind a duplex DNA containing a single tetrahydrofuran residue, a noncatalytically competent substrate (33, 34). The apparent dissociation constants of mutant FpgP2G and P2E were 39 and 60 nM, respectively, comparable with that observed for purified wild-type Fpg protein (45 nM) (Table III). These results suggest that the substitution of the N-terminal proline for glycine or glutamic acid does not interfere with DNA binding.

G/C  T/A Spontaneous Mutagenesis in fpg, mutY-deficient E. coli Strain BH990 Expressing Mutant fpg Genes-- E. coli possesses two DNA glycosylase activities that prevent spontaneous mutagenesis by 8-oxoG: the Fpg protein, which excises 8-oxoG residues in DNA (3, 7), and the MutY protein, which excises the adenine residues incorporated opposite 8-oxoG (35-37). Inactivation of both fpg and mutY genes of E. coli (BH990 fpg mutY) results in a strong G/C  T/A spontaneous mutator phenotype (24, 36). To study the repair capacity of the partially active FpgP2G and FpgP2T proteins in vivo, we have measured the spontaneous mutagenesis in the BH990(fpg mutY) cells hosting plasmids coding for the wild-type or mutant Fpg proteins (Table IV). The strain BH990 (fpg mutY) hosting pUC19 shows a 800-fold increase in G/C  T/A transversion compared with the parental CC104 strain. The spontaneous mutator phenotype because of fpg and mutY mutations is completely suppressed by expression of the plasmid coding for the wild-type Fpg protein. In contrast, the plasmids coding for FpgP2G protein barely modify, and the FpgP2T protein does not modify the spontaneous mutagenesis of the E. coli double mutant (Table IV). This experiment points out the crucial role of N-terminal proline for the biological activity of this protein, presumably via the removal of the mutagenic 8-oxoG residues.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

It has been proposed that the N-terminal proline of the Fpg protein could be the nucleophile involved in the Schiff base formation with 8-oxoG residues (20). To investigate, in detail, the implications of this amino acid upon the various activities of the Fpg protein, we have constructed three mutated proteins at this position by site-directed mutagenesis: FpgP2G, FpgP2T, and FpgP2E. These mutated proteins were purified to near homogeneity and characterized for their various enzymatic activities.

The results reveal that the substitution of the N-terminal proline barely affects the DNA binding function of the mutated proteins as compared with the wild type. These data are in agreement with the results of Tchou and Grollman (32), showing that the presence of a maltose-binding protein at the N terminus of Fpg protein does not change DNA binding.

When using as substrate a DNA containing Fapy residues, the FpgP2G and FpgP2T proteins were partially active (about 10% of the activity of the wild-type protein). The evaluation of the 8-oxoG-DNA glycosylase activity of the mutated proteins shows that the activity of the FpgP2G and FpgP2T mutants is dramatically reduced and that these mutants barely produce Schiff base with oligonucleotides containing 8-oxoG residues. Taken together, these results suggest that the N-terminal proline is critical for the catalysis of the 8-oxoG residue excision.

It has been proposed that 8-oxoG is the primary damage recognized by the Fpg protein in vivo leading, if unrepaired, to G/C  T/A transversion (1, 2, 9, 21). The results show that the FpgP2G barely, if at all, prevents and that the FpgP2T mutant does not prevent the spontaneous G/C  T/A mutagenesis in vivo. These in vivo results are in good agreement with the minute 8-oxoG-DNA glycosylase activity of the respective proteins, measured in vitro. Therefore the recognition and excision of 8-oxoG residues are complex mechanisms because at least two other amino acids, Lys-57 and Lys-155, are also involved in these processes (21, 22).

The N-terminal proline is also critical for the -lyase activity of the Fpg protein at preformed AP site: two different substrates containing preformed AP site were used, none of the mutant proteins had detectable activity upon these substrates. The mechanism of action of the N-glycosylase/AP-lyase has been proposed to occur via a covalent Schiff base intermediate that can be trapped by treatment of the enzyme-substrate complex with a reducing agent (13-15). In fact, an efficient borohydride-dependent trapping of a number of DNA N-glycosylases possessing AP-lyase activity with their substrates has been observed (Refs. 15, 20-22, 32, 33, 38, and 39 and this paper). The results presented above using 8-oxoG containing oligonucleotide confirm this trapping by the wild-type Fpg protein. It efficiently excises 8-oxoG residues and readily produces Schiff base intermediate, whereas the FpgP2G or FpgP2T proteins having a poor 8-oxoG-DNA glycosylase activity barely produce Schiff base with 8-oxoG/C DNA. At variance, when using the oligonucleotide containing a preformed abasic site and FpgP2G, a mutant protein devoid of AP-lyase activity (see above), the formation and the amount of the covalent complex in the presence of NaBH₄ was comparable with the amount of complex obtained with the wild-type Fpg protein. This unexpected result shows that it is possible to dissociate the formation of the complex from the incision of the DNA at abasic sites. The mutagenesis of the N-terminal proline residue reveals an interesting example of a N-glycosylase readily effective to trap its substrate without its excision: the Fpg P2G mutant efficiently produced the covalent Schiff base intermediate with preformed AP site but did not excise AP site.

The E. coli Mut Y protein is a DNA glycosylase excising adenine residues in a 8-oxodG/A mismatch (40, 41). Mut Y, at higher than 1:1 protein-substrate molar ratio, forms a covalent intermediate with its substrate involving Lys-142 (41) but has little or no AP lyase activity (40, 41). It has been recently shown that the single amino acid substitution Lys-142 to Ala generates a Mut Y protein retaining its DNA glycosylase activity but incising DNA at abasic site by a - elimination mechanism without the formation of a Schiff base intermediate (42). This mutant protein has just the opposite properties of the Fpg P2G. These two observations point out that it is possible to dissociate the formation of the Schiff base complex from the incision of the DNA at abasic sites.

The results point out the different role of the N-terminal proline of the Fpg protein in the processing of the different substrates. However, it is not excluded that the amino acid substitution could modify the environment and/or the role of some specific amino acid(s) in the active site of the protein. For example, we have shown that the mutation P2G in the FpgP2G mutant did not change the overall structure of this protein as determined by measuring the acrylamide quenching and the lifetime of tryptophane fluorescence (43). However, subtle changes were observed in the structure dynamics of the P2G mutant at 20 °C, indicating that the tryptophane residues are differently surrounded as compared with the wild-type protein. The FpgP2E protein is devoid of any enzymatic activity but has a binding constant similar to the FpgP2G mutant. There are neither changes in the structure dynamics of FpgP2E at 20 °C nor at 4 °C, which means that its conformation at 20 °C is the same as at 4 °C, whereas wild-type Fpg, FpgP2G, and FpgK57G exhibit changes under such conditions. The lack of the structure dynamics at variant temperatures and/or the presence of the first Met of the protein could explain the total lack of any enzymatic activity of the Fpg P2E protein.

	ACKNOWLEDGEMENTS

We thank Drs. J. Lhomme and J. F. Constant (UMR 5616 CNRS, Grenoble) for the gift of oligonucleotide containing the tetrahydrofuran residue and for helpful discussion and Dr. J. Derancourt (UPR 1086 CNRS, Montpellier) for protein sequencing.

	FOOTNOTES

^* This work was supported by European Community Grant ENV4-CT97-0505 and by funds from Fondation pour la Recherche Medicale (fellowship to O. S.) and Association pour la Recherche sur le Cancer.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.: 33-1-42114824; Fax: 33-1-42114454; E-mail: jlaval@igr.fr.

	ABBREVIATIONS

The abbreviations used are: 8-oxoG, 7,8-dihydro-8-oxoguanine; Fapy, 2,6-diamino-4-hydroxy-5-N-methylformamidopyrimidine; AP, apurinic/apyrimidinic, abasic; PAGE, polyacrylamide gel electrophoresis.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES