Mapping of the 5'-2-deoxyribose-5-phosphate lyase active site in DNA polymerase beta by mass spectrometry.

The mechanism of the 5'-2-deoxyribose-5-phosphate lyase reaction catalyzed by mammalian DNA beta-polymerase (beta-pol) was investigated using a cross-linking methodology in combination with mass spectrometric analyses. The approach included proteolysis of the covalently cross-linked protein-DNA complex with trypsin, followed by isolation, peptide mapping, and mass spectrometric and tandem mass spectrometric analyses. The 8-kDa domain of beta-pol was covalently cross-linked to a 5'-2-deoxyribose-5-phosphate-containing DNA substrate by sodium borohydride reduction. Using tandem mass spectrometry, the location of the DNA adduct on the 8-kDa domain was unequivocally determined to be at the Lys(72) residue. No additional amino acid residues were found as minor cross-linked species. These data allow assignment of Lys(72) as the sole Schiff base nucleophile in the 8-kDa domain of beta-pol. These results provide the first direct evidence in support of a catalytic mechanism involving nucleophilic attack by Lys(72) at the abasic site.

Mammalian DNA polymerase ␤ (␤-pol), 1 a constitutively expressed "housekeeping" enzyme, has been implicated in DNA base excision repair (BER) (1)(2)(3)(4). Base excision repair is thought to be the major repair pathway protecting cells against single base DNA damage. Recent evidence has indicated that BER in mammalian cells is mediated through at least two subpathways that are differentiated by the patch sizes and by the enzymes involved (5)(6)(7)(8). These subpathways are designated as "single nucleotide" BER and long patch BER (two to several nucleotide repair patches). The single nucleotide BER subpathway is a sequential multistep process, where ␤-pol is involved in two steps (9,10). The overall scheme for the single nucleotide BER subpathway can be outlined as follows: (i) recognition and cleavage of an altered/damaged base by a specific monofunctional DNA glycosylase (a glycosylase lacking intrinsic apurinic/apyrimidinic (AP) lyase activity) resulting in an AP site; (ii) class II AP endonuclease incision of the phosphodiester backbone 5Ј to the AP site, resulting in 3Ј-hydroxyl and 5Ј-2-deoxyribose-5-phosphate (dRP) containing termini; (iii) replacement of the missing base; (iv) removal of the 5Јsugar phosphate to provide a 5Ј-phosphate for DNA ligation by ␤-pol; and (v) DNA ligase sealing of the nick (3, 9 -13).
␤-Pol is a multifunctional enzyme consisting of an 8-kDa N-terminal domain with dRP lyase activity and a 31-kDa Cterminal domain with nucleotidyltransferase activity (9, 14 -16). These two domains appear to be packed together in ␤-pol in solution, as revealed by comparison of axial ratios of ␤-pol (5.0) and the isolated 8-and 31-kDa domains (2.3 and 5.5, respectively) (17). Circular dichroism analysis has revealed that the 8-kDa domain is essentially ␣-helical in nature (16), and NMR solution and crystal structures (18,19) have since confirmed the predominance of this secondary structure.
The 8-kDa domain of ␤-pol (residues 1-87), originally characterized as a single-stranded DNA binding domain, is formed from four ␣-helices. These helices are packed as two antiparallel pairs with 60°crossing between the pairs. Connecting segments between helices 1 and 2 and between helices 3 and 4 each contribute to DNA binding. Helix 3-turn-helix 4, which forms a "helix-hairpin-helix" motif is similar to the helix-hairpin-helix motif that has been found in a number of DNA repair proteins (18,20,21), including several DNA glycosylases and AP lyases (20,22). Residues of the helix-hairpin-helix motif have been proposed to contribute to recognition and excision of damaged nucleotides in DNA, as well as AP lyase chemistry (23,24). Alignment of the helix-hairpin-helix motifs from ␤-pol and endonuclease III suggested that Lys 68 in ␤-pol may be important in lyase chemistry, because mutation of the analogous lysine residue in endonuclease III, Lys 120 , resulted in a dramatic reduction in AP lyase activity, and Lys 120 has been proposed to be the Schiff base nucleophile in this enzyme (25). In addition, the crystal structure of ␤-pol bound to a gapped DNA molecule representing a product of the dRP lyase reaction suggested that lysine residues 35 and 68 coordinate the 5Јphosphate that exists in the gapped DNA/enzyme crystal structure (Ref. 26 and Fig. 1).
Matsumoto and Kim (9) suggested that ␤-pol catalyzes removal of dRP from the AP endonuclease-incised AP site via ␤-elimination, as opposed to hydrolysis, and that this dRP lyase activity resides in the N-terminal 8-kDa domain of ␤-pol. In the ␤-elimination, the dRP excision reaction would proceed via an imine or Schiff base intermediate. Fisher et al. (27) demonstrated in 1958 that an imine intermediate formed between a substrate and enzymatic amino group can be trapped by reduction with sodium borohydride (NaBH 4 ). Since then this chemical technique has been widely used to elucidate reaction mechanisms, e.g. acetate decarboxylase (28) and aldolase (29). In both cases, the ⑀-NH 2 group of a lysine was found to be involved in the formation of the imine intermediate. More recently, NaBH 4 trapping was used for the identification of imine intermediates in several DNA enzymes: bacteriophage T4 endonuclease V-DNA (30), Escherichia coli endonuclease III-DNA (31), E. coli Fpg-DNA (31,32), and Micrococcus luteus UV endonuclease-DNA covalent complexes (33). More direct evidence that the 8-kDa domain of ␤-pol catalyzes removal of the dRP group via ␤-elimination was also obtained by Piersen et al. (12). These investigators showed that a Schiff base intermediate is formed between the dRP-containing DNA substrate and the enzyme. The Schiff base nucleophile in the 8-kDa domain has been suggested to be Lys 72 by site-directed mutagenesis (34,35). This residue is in close proximity to the 5Ј-phosphate product group in the gapped DNA/enzyme crystal structure (26) and is part of the putative dRP lyase active site identified in NMR structures (23) of the 8-kDa domain (Fig. 1). Thus, based upon structural and site-directed mutagenesis data, it has been suggested that Lys 72 is the preferred but not necessarily the obligatory residue in the dRP lyase active center of the 8-kDa domain of ␤-pol (24,35,36). To understand the precise mechanism of the lyase reaction catalyzed by ␤-pol and the lysine residue(s) involved in Schiff base chemistry, we utilized the NaBH 4 trapping technique in combination with mass spectrometric (MS) analysis. The 8-kDa domain was first covalently cross-linked to a dRP-containing DNA substrate by NaBH 4 reduction. We next identified the covalently modified lysine residue by MS sequencing. The approach included proteolysis of the covalently cross-linked protein-DNA complex with tryp-sin, followed by isolation, peptide mapping, and finally MS and MS/MS analyses of the adducted peptides. Our results, which unambiguously show that Lys 72 in the 8-kDa domain of ␤-pol is the sole Schiff base nucleophile, are discussed in the context of structure of the dRP lyase active site.

EXPERIMENTAL PROCEDURES
Materials-The acetonitrile and ammonium bicarbonate (Fisher Scientific, Fair Lawn, NJ), trifluoroacetic acid (Pierce), formic acid (Aldrich), and porcine trypsin (Promega Corporation, Madison, WI) were used as delivered. All buffers were prepared using water with a conductivity of 18-M⍀ (Hydro Service and Supplies, Research Triangle Park, NC).
Synthetic oligodeoxyribonucleotides purified by high pressure liquid chromatography were obtained from Oligos Etc, Inc. (Wilsonville, OR).
UDG Treatment of DNA Substrate-The 32 P-labeled duplex oligonucleotide was treated with human UDG that resulted in the 32 P-labeled deoxyribose sugar phosphate at the nick. Typically, 50 nM DNA substrate was pretreated with 10 nM UDG in 50 mM Hepes, pH 7.4, 0.5 mM EDTA, and 0.2 mM dithiothreitol. The reaction mixture was incubated for 30 min at 37°C. Due to the labile nature of the UDG-treated DNA, the DNA substrate was prepared just before performing the NaBH 4 trapping experiment.
Isolation and Purification of the 8-kDa Domain-DNA Complex-To prepare the covalently cross-linked 8-kDa DNA complex, the NaBH 4 trapping technique was used (12). Briefly, the reaction mixture (1 ml) contained 50 mM Hepes, pH 7.4, 0.5 EDTA, 0.2 mM dithiothreitol, 25 M 8-kDa domain of ␤-pol, 5 M duplex oligonucleotide substrate comprising 0.04 M 32 P-labeled DNA to monitor the cross-linked complex, and 25 mM NaBH 4 . The reaction mixture was incubated for 30 min on ice and then 30 min at room temperature. After incubation, the reaction mixture was precipitated with ice-cold 10% (v/v) trichloroacetic acid. Under these conditions, the free protein and protein-DNA complex precipitate leaving the majority of the free DNA in the supernatant. The protein was then pelleted by centrifugation, washed twice with ice-cold 100% acetone, and air-dried. The pellet was solubilized in 8 M urea and diluted with 50 mM Tris-HCl, pH 8.8, to give a final urea concentration of 1 M. Subsequently, the sample was loaded onto an FPLC Mono Q column (HR 5/5). The bound protein-DNA complex was eluted from the column using a NaCl gradient (0 -1.0 M), and all fractions were counted for radioactivity. Fractions with peak radioactivity were subjected to SDS-PAGE followed by autoradiography. At this stage, the fractions containing the protein-DNA complex were then pooled and digested with micrococcal nuclease (10 g/ml). The covalently linked protein-DNA complex was precipitated with ice-cold trichloroacetic acid (10%), washed twice with ice-cold acetone (100%), and air-dried. The pellet was solubilized in 8 M urea and diluted with 50 mM (NH 4 ) 2 CO 3 (pH 8.5) to a final urea concentration of 1 M. The protein-DNA complex was further purified using an FPLC Mono S column (HR 5/5). All fractions were counted for radioactivity and analyzed by SDS-PAGE and autoradiography. Fractions containing DNAadducted protein were pooled and concentrated by trichloroacetic acid precipitation. Finally, the protein pellet was dissolved in 8 M urea and diluted with 50 mM (NH 4  The 5Ј-phosphate binds near a lysine-rich pocket (blue) and the proposed lyase catalytic center of the 8-kDa domain. Structural and sitedirected mutagenesis data suggest that Lys 35 and Lys 68 interact with this phosphate, and Lys 72 has been proposed to be in the dRP lyase active site of the 8-kDa domain (34,35). Nearby lysine residues Lys 60 and Lys 68 are also shown. Helices ␣2, ␣3, and ␣4 are translucent and shown in white. This figure was generated using GRASP (44), Molscript (45), and Raster3D (46). diode array detector, and a Bio-Rad Model 2110 fraction collector (Bio-Rad). The column used for the purification was a Vydac C4 reverse phase column (25 cm ϫ 4.6 mm inner diameter). Separations were performed using a linear water:acetonitrile (0.1% trifluoroacetic acid) gradient of 10 -60% acetonitrile over 50 min at a flow rate of 1 ml/min. Fractions were collected at 1-min intervals and analyzed on a Beckman Model LS 6500 liquid scintillation counter (Beckman Coulter, Inc., Fullerton, CA) to detect the 32 P-containing fractions.
Tryptic Digestion Conditions-The HPLC-purified 8-kDa domain protein-DNA complex was reconstituted in 10 l of 90% acetonitrile and then diluted to 100 l with 50 mM ammonium bicarbonate buffer (pH 8). Porcine trypsin (1 g) (Promega Corporation, Madison, WI) was added to an aliquot of the complex and to a sample of native 8-kDa protein alone (0.3 g/l in 50 mM ammonium bicarbonate, pH 8.0). The reactions were carried out at 37°C for 3 h. All tryptic fragment nomenclature refers to predicted fragments of the native 8-kDa domain protein.
Electrospray Mass Spectrometry-A Micromass Q-Tof (Altrincham, UK) hybrid tandem mass spectrometer was used for the acquisition of the electrospray ionization (ESI) mass spectra and tandem mass spectra (39). The instrument is equipped with a nanoflow electrospray source and consists of a quadrupole mass filter and an orthogonal acceleration time-of-flight mass spectrometer. The needle voltage was ϳ2800 V, the cone voltage was 25 eV, and the collision energy was 4.0 eV for the MS analyses. For the MS/MS experiments, a parent ion is selected with the first mass analyzer and transmitted into a collision cell where fragmentation is induced by collision with argon atoms. The collision energy used for these experiments was 30 eV. The resulting fragment ions were detected with the second mass analyzer. In this type of experiment, only ions resulting from fragmentation of the selected parent ion are observed. Data analysis was accomplished with a Mass-Lynx data system and MaxEnt software supplied by the manufacturer.
Samples for flow injection analyses were infused into the mass spectrometer at ϳ200 nl/min using a pressure injection vessel (40). For the LC/MS and LC/MS/MS analyses, gradients were formed and delivered using a Gilson Gradient HPLC system and controller (Gilson, Inc., Middleton, WI). The HPLC system consisted of Gilson model 305 and 306 pumps, a model 811C manometric module, and a model 805 mixing chamber. Injections of 2.5 l were made using a FAMOS automatic injector (LC Packings, San Francisco, CA). A 5-95% acetonitrile (0.1% formic acid) linear gradient over 30 min was used for the chromatographic separations. A flow rate of 0.35 ml/min was delivered from the Gilson HPLC system to an Acurate Splitter (LC Packings, San Francisco, CA), which reduced the flow rate through the column to approximately 200 -300 nl/min. The column used was a 15 cm ϫ 75 m inner diameter Hypersil C18 ("pepmap") column (LC Packings, San Francisco, CA).
Matrix-assisted Laser Desorption Ionization Mass Spectrometry-The MALDI analyses were performed using a Voyager RP (PerSeptive Biosystems, Framingham, MA) time-of-flight dual-stage reflector mass spectrometer as has been previously described (41). A saturated MALDI matrix solution (0.5 l) of ␣-cyano-4-hydroxycinnamic acid in 45:45:10 ethanol:water:formic acid and 0.5 l of the sample solution were spotted onto the MALDI target. Co-crystallization of the sample and matrix was allowed to proceed at room temperature. 16 and a nick between positions 15 and 16 was chosen as a matter of convenience; the uracil-containing oligonucleotide was 5Ј-end-labeled prior to annealing. Next, the duplex DNA was treated with uracil-DNA glycosylase to create a dRP-containing single nucleotide-gapped DNA substrate. The DNA substrate, thus prepared (12), contained a 32 P-labeled dRP flap at the nick ( Fig. 2A).

Optimization of Covalent 8-kDa DNA Complex Formation by
For preparation of large amounts of the covalent 8-kDa DNA complex, conditions for trapping the imino complex by NaBH 4 reduction were evaluated. Results depicted in Fig. 2B (lanes  1-3) show that when the 8-kDa and 32 P-labeled DNA were used in varying ratios and incubated with NaBH 4 for 30 min at room temperature, little difference was detected in the amount of protein-DNA complex formed. Time course analysis of complex formation showed that a 30-min incubation of the 8-kDa protein and DNA with NaBH 4 gave results similar to longer incubations (Fig. 2B, lanes 4 -6). For subsequent large scale preparations, isolation, and purification of the complex, 8-kDa FIG. 4. MALDI mass spectrum of the micrococcal nuclease-digested 8-kDa protein-DNA complex following purification by FPLC using a Mono S column. MALDI spectra were acquired prior to HPLC purification (A) and after HPLC purification (B). Ions corresponding to the singly and doubly charged protein-DNA complex were observed at approximately 10,000 and 5,000 Da, respectively. The mass spectrum was acquired as described under "Experimental Procedures." Ions labeled with an asterisk (*) correspond to background and/or micrococcal nuclease ions. The additional ion of m/z 9783 in B is most likely a decomposition ion due to loss of guanine. protein and DNA were used in a 5:1 ratio (protein:DNA) and incubated for 30 min at room temperature.
Purification of the 8-kDa DNA Complex-The 8-kDa domain of ␤-pol contains 15 lysine and 4 arginine residues and has an isoelectric point of 10.3. Hence at pH 8.8, below the isoelectric point, the protein should have a net positive charge, whereas, after cross-linking to the DNA substrate, it should have a net negative charge. This charge difference was used to separate the covalently linked protein from the free 8-kDa protein on an FPLC Mono Q column. Under the conditions used (pH 8.8), free protein does not bind to the Mono Q column and emerges in the flow-through, whereas the cross-linked protein-DNA complex remains bound to the column. The bound protein-DNA complex was eluted from the column using a NaCl gradient. A small portion of each fraction was analyzed by SDS-PAGE. Fractions containing radiolabeled protein-DNA complex were pooled (Fig. 3, lane 1) and subjected to micrococcal nuclease digestion. Micrococcal nuclease digested the entire length of the DNA except for the covalently cross-linked nucleotide. Because most of the DNA was digested, the remaining protein-DNA complex should have a net positive charge and, hence, should bind to the Mono S column. FPLC Mono S column chromatography, therefore, was used to purify the protein-DNA complex after micrococcal nuclease digestion. Based upon SDS-PAGE analysis, the fractions that contained the protein-DNA complex were pooled (Fig. 3, lane 2) and used for subsequent MS and peptide mapping analyses.
Mass Spectrometric Analyses of the 8-kDa DNA Complex-Following micrococcal nuclease digestion and purification of the 8-kDa protein-DNA complex by FPLC using the Mono S column, the positive ion MALDI mass spectrum of the complex was acquired and is shown in Fig. 4A. The mass spectrum reveals ions corresponding to the protonated molecule of the protein-DNA complex at approximately 10,000 Da as well as ions that correspond in mass to protonated molecules of the micrococcal nuclease (ions labeled with an asterisk). Because the protein-DNA complex and the nuclease coelute from the Mono S column, the protein-DNA complex was further purified by reverse-phase HPLC using UV detection (Fig. 5A). As the DNA contains a 32 P label, the fractions containing the 8-kDa protein-DNA complex were easily identified (Fig. 5B). The native 8-kDa domain protein, as well as the HPLC-purified protein-DNA complex, were then analyzed by both flow injection ESI/MS (Fig. 6) and MALDI/MS (Fig. 4B). The molecular mass of the 8-kDa protein-DNA covalent complex as determined by ESI/MS was 10,010 Da (Fig. 6B) in comparison to 9467 Da for the native 8-kDa protein (Fig. 6A). The mass accuracy of this instrument with external calibration is 0.01%, therefore, for molecular masses at 10,000 Da, the accuracy is Ϯ 1 Da. These data suggest a molecular mass for the protein-DNA complex of 543 Da (Ϯ1 Da) greater than that of the molecular mass of the protein (or less if the protein had become oxidized). Similar results were obtained from the MALD/MS analysis (Fig. 4B).
LC/MS Analyses of the Tryptic Digests-To determine which amino acid(s) in ␤-pol interacts with the DNA 5Ј-phosphate, the 8-kDa domain alone and the 8-kDa domain-DNA complex were subjected to tryptic digestion and analysis by LC/ESI/MS. The potential tryptic cleavage sites in the amino acid sequence of the 8-kDa domain of ␤-pol and the resulting tryptic fragment numbers are shown in Fig. 7A. The mass chromatograms of the major tryptic fragments of the native 8-kDa protein were generated and compared with the mass chromatograms of these same tryptic fragments from the digest mixture of the protein-DNA complex. A notable difference observed between the mass chromatograms was that the relative abundance of tryptic fragment T15 (amino acids 73-81) was greatly reduced in the complex mixture in comparison to the native 8-kDa digestion mixture, indicating that the cross-linked DNA may be contained within these amino acid residues (data not shown). Note that the sample used for the tryptic digest of the 8-kDa protein-DNA also contained some residual native 8-kDa domain (Fig.  6B). It is, therefore, expected that tryptic fragments corresponding in mass to the native 8-kDa protein would be observed in the analysis of the digestion mixture of the 8-kDa protein-DNA complex. The most abundant new ion observed upon comparison of the LC/MS analysis of the digestion mixture of the complex (Fig. 7C) with that of the 8-kDa protein alone (Fig. 7B) was an ion [(M ϩ 2H) 2ϩ ϭ 981.5] that eluted at 77.8 min and corresponds in mass to amino acid residues 69 -81 (T14 -15) plus 527 Da, in the protein-DNA complex digestion mixture, which was absent in the control digest mixture. The mass of this ion may correspond to the formation of the Schiff base intermediate followed by reduction. These data also indicate that 16 Da of the mass difference between the 8-kDa protein and the 8-kDa protein-DNA adduct are probably because of oxidation of one of the amino acid residues (i.e. mass increase of protein-DNA complex, 543 Da, less mass increase of adducted peptide, 527 Da, ϭ oxygen, 16 Da). No signal was observed for the addition of 527 Da for any other predicted tryptic fragments above the baseline noise level (Table I). These data suggest that the DNA is cross-linked to one of the lysine residues in T14 -15 (residues 69 -81).
Determination of the Amino Acid in ␤-Pol Cross-linked to DNA-To determine whether the (M ϩ 2H) 2ϩ ion at m/z 981.5 (T14 -15 plus 527 Da) contains the cross-linked DNA, an LC/ MS/MS spectrum was acquired. The resulting MS/MS spectrum after transformation of all ions to the single charge state is shown in Fig. 8 and Table II. The major fragment ions are observed at m/z 1810.85, 1712.86, 1632.76, and 1534.79 (labeled as -G, -dG, -pdG, and -pdG -p) and correspond to the loss of guanine, deoxyguanosine, deoxyguanosine plus a phos- phate, and deoxyguanosine plus two phosphate groups, respectively, from the peptide. These observed masses are within 0.05 Da of the calculated masses. These data confirm that the crosslinked DNA was contained within the T14 -15 peptide. In addition, a series of both y and b ions (42,43) are observed, which correspond to cleavages along the peptide backbone. The y series ions result from C-terminal peptide backbone cleavages and the b series ions result from N-terminal backbone cleavages. The y 1 through y 9 series ions correspond in mass to sequential loss of amino acids beginning at the C-terminal Lys 81 and ending at Ile 73 . The b series ions (b 4 -dG through b 11 -dG) correspond in mass to cleavages of amino acids from the N terminus of the peptide backbone minus deoxyguanosine. The observation of the y 10 -dG ion in addition to the observation of the y 9 ion provides the necessary data to definitively assign the location of the DNA adduct. The observed mass difference between these two ions plus the mass of deoxyguanosine equals the mass of a lysine residue plus the mass of the DNA adduct. These structurally informative fragment ions allow the unequivocal assignment of the cross-linked DNA to the Lys 72 residue of the 8-kDa domain of ␤-pol. No additional amino acid residues were found as minor cross-linked species (Table I). DISCUSSION In the present study, the mechanism of the dRP lyase reaction of human DNA ␤-polymerase has been investigated. To study the mechanism, the 8-kDa domain was covalently crosslinked to dRP-containing DNA by NaBH 4 reduction, and the trapped intermediate was purified and then subjected to peptide mapping and MS sequencing analyses. Identification of the precise location of the DNA adduct on the 8-kDa protein has provided the first direct evidence in support of a catalytic mechanism involving nucleophilic attack by Lys 72 at the abasic site. Based upon earlier site-directed mutagenesis studies (24, 34 -36), Lys 72 was shown to be involved at the dRP lyase active center of the 8-kDa domain of ␤-pol, but the precise role of Lys 72 could not be assigned. Other residues that potentially could be involved in Schiff base formation include Lys 35 , Lys 60 , Lys 68 , and Lys 84 (Fig. 1). To determine precisely the amino acid residue(s) covalently cross-linked to the DNA, the 8-kDa protein-DNA complex was subjected to tryptic digestion followed by mass spectrometric sequencing. Because of the complexity of the digestion mixture, on-line LC was used in conjunction with the MS analyses. The results of the LC/MS analysis of the tryptic digest of the 8-kDa protein-DNA complex were com- LC/MS analysis of the tryptic digest of the 8-kDa protein-DNA complex showed the same tryptic digest ions that were observed in the LC/MS analysis of the native 8-kDa protein (data not shown). This observation is not surprising given the fact that some residual native 8-kDa protein was present in purified 8-kDa protein-DNA complex (Fig. 6B). In addition to the observation of the methionine-containing tryptic fragment ions T3-4 and T4, the oxidized form of these ions were also observed (data not shown). Based on ion counts, over 80% of tryptic fragment T4 appears to be oxidized. This indicates that the protein-DNA complex had become oxidized during the isolation and purification procedures. This confirms that the mass difference attributable to the DNA adduct is most likely 527 not 543 Da.
Mass chromatograms of ions corresponding to the addition of 527 Da to the tryptic fragments containing the suspected amino acid residues involved in dRP lyase activity based on mutagenesis studies were generated from the LC/MS data of the digests of both the protein-DNA complex and the native protein (Table I). The mass chromatogram corresponding to the doubly charged ion of tryptic fragment T14 -15 (amino acid residues 69 -81) plus 527 Da (m/z 981.5) was observed in the LC/MS analysis of the protein-DNA complex digestion mixture (Fig. 7C), and this ion was not observed in the LC/MS analysis of the digestion mixture of the native protein used as a control (Fig. 7B). The additional mass of 527 Da apparently results from the covalently bound abasic site and the attached 3Ј-  deoxyguanylic acid. Because no signal was observed for the addition of 527 Da to any of the other potential amino acid residues, these data suggested that the DNA adduct is covalently cross-linked to one of the amino acid residues of 69 -81.
To verify these results, the LC/MS/MS spectrum of the doubly charged ion of the T14 -15 peptide was acquired. The tandem mass spectrum of the (M ϩ 2H) 2ϩ ion of m/z 981.5 in the tryptic digest mixture of the 8-kDa protein-DNA complex showed structurally informative fragment ions indicating the location of the DNA adduct on the tryptic peptide. After transformation of all ions to the single charge state, the resulting MS/MS spectrum showed abundant fragment ions corresponding to cleavages of the DNA that was adducted to the tryptic peptide ( Fig. 8 and Table II). These ions included the loss of guanine (-G), loss of deoxyguanosine (-dG), loss of deoxyguanosine plus a phosphate (-pdG), and loss of deoxyguanosine plus two phosphates (-pdG -p). These data are consistent with the formation of a DNA adduct, which would contain an abasic site. In addition, a nearly complete series of C-terminal ions (i.e. y 1 to y 9 ) were observed, which correspond to the amino acid sequence IDEFLATGK. N-terminal ions minus the deoxyguanosine were also observed (i.e. b 4 -dG to b 11 -dG). The observation of these structurally informative fragment ions, most importantly y 9 and y 10 -dG, allows for definitive identification of Lys 72 as the amino acid in the 8-kDa domain, which has been modified by covalent cross-linking to the DNA. Based upon the LC/MS and LC/MS/MS results, a proposed structure for the intermediate involved in dRP lyase activity is shown in Table II. The sole Schiff base intermediate in this enzyme is formed between the abasic site and the Lys 72 residue of the 8-kDa domain.
In summary, using various purification and MS sequencing methodologies, the Schiff base intermediate trapped by reduction was identified. Tandem mass spectrometry provided structural information as to the location of the DNA adducted to the 8-kDa domain of ␤-pol. The amino acid residue located at the center of the lyase activity was unequivocally determined to be   Table II. The most abundant fragment ions are labeled -G, -dG, -pdG, and -pdG -p and correspond to loss of guanine, deoxyguanosine, deoxyguanosine plus a phosphate, and deoxyguanosine plus two phosphate groups, respectively. The y and b series ions correspond to C-terminal and N-terminal cleavages, respectively, along the peptide backbone.
Lys 72 of the 8-kDa domain of ␤-pol. Thus, peptide mapping in combination with mass spectrometry is an extremely powerful technique for investigating the structure of covalent intermediates in protein-DNA interactions.