DNA determinants and substrate specificities of Escherichia coli MutY.

Potential DNA contacts involved in the specific interaction between the Escherichia coli MutY protein and a 40-mer oligonucleotide containing an A/G mismatch have been examined by alkylation interference techniques. Ethylation interference patterns suggest that more than five phosphates are involved in electrostatic interactions between MutY and DNA. Interestingly, MutY has more contacts on the G-strand than on the A-strand. Methylation at both the N-7 position of the mismatched G and the N-3 position of the mispaired A interfere with MutY binding. In addition to these mismatched bases, MutY also contacts purines on both sides of the mismatch. Binding and endonuclease activities of MutY were assayed with 20-mer oligonucleotides containing A/G, A/C, A/7,8-dihydro-8-oxo-guanine (A/GO), A/inosine (A/I), A/2-aminopurine (A/2AP), nebularine/G (N/G), inosine/G (I/G), 2AP/G, and 7-deaza-adenosine/G (Z/G) mispairs. The C-8 keto group of GO in A/GO contributes to a much tighter binding but weaker endonuclease activity than is seen with A/G. Because A/I is not specifically well recognized by MutY, the 2-amino group of G in A/G is essential for recognition. The C-6 keto group present in A/G but absent in A/2AP is also important for recognition. The 6-amino group of adenine appears not to be required for either binding or endonuclease activity because N/G is as good a substrate as A/G. The 2AP/G mispair is bound and cleaved weaker than is the A/G mispair. Binding and endonuclease activities are abolished when the N-7 group of A is replaced by C-7 as in the Z/G mispair. When a C-6 keto group is present as in the I/G pair, its binding by MutY is as good as for A/G, but no endonuclease activity is observed. Taken together, our data suggest that DNA sequences proximal to and specific functional groups of mismatched bases are necessary for recognition and catalysis by MutY protein.

Multiple mismatch repair pathways with different mispair specificities and different size repair tracts are utilized by Escherichia coli to reduce replicative errors and to protect its DNA from various types of damage (1). One of the short-patch repair pathways requires mutY gene function and is independent of dam-methylation (2)(3)(4). E. coli mutY (or micA) mutants have higher mutation rates for C⅐G to A⅐T transversions (3,5). The MutY pathway specifically repairs A/G mismatches to C/G base pairs (2)(3)(4)6) and repairs A/C to G/C at a much lower rate (2,3,7). MutY can also act on adenines mispaired with 7,8dihydro-8-oxo-guanine (GO) 1 or 7,8-dihydro-8-oxo-adenine (AO) (8,9). The GO lesion is one of the most stable products of oxidative damage to DNA known. A role for the MutY pathway in E. coli is to remove misincorporated adenines opposite G or GO following DNA replication (8,10). Adenines are frequently incorporated opposite GO bases during DNA replication in vitro (11) and in vivo (12). A second round of replication through this mismatch subsequently leads to a C⅐G to A⅐T transversion (12)(13)(14)(15). E. coli uses MutY, MutM, and MutT to defend against the mutagenic effects of GO lesions (10,16). The MutT protein eliminates 8-oxo-dGTP from the nucleotide pool by its nucleoside triphosphatase activity (17)(18)(19). The MutM protein (FPG protein) provides a second level of defense by removing both ring-opened purine lesions and mutagenic GO adducts (20,21). MutM removes GO lesions efficiently from C/GO but poorly from A/GO (21). MutY works at a third level by correcting replicative errors that result from misincorporation of A opposite GO (8,9).
The 39-kDa MutY protein is an iron-sulfur protein, which has homology with E. coli endonuclease III (7,22,23). The MutY protein was shown by Tsai-Wu et al. (7) to have both DNA glycosylase and apurinic/apyrimidinic (AP) endonuclease activities, although the AP endonuclease activity could not be detected by Au et al. (24) in their MutY preparation. The DNA glycosylase activity removes the adenine bases from the A/G, A/C, A/GO, and A/AO mismatches (7,9,24), and the AP endonuclease activity cleaves the first phosphodiester bond 3Ј to the AP site (7,25).
The activity of MutY on its mismatched DNA substrates is influenced by the neighboring sequence composition (2,3,25). Structural studies have demonstrated that A/G mispairs can adopt three possible configurations (A(anti)-G(anti), A(anti)-G(syn), and A(syn)-G(anti)), depending on their neighboring sequences (26 -30). It is unclear which configuration of A/G mispair is recognized by MutY. Based on the common features of MutY substrates (A/G, A/GO, A/C, and A/AO), it has been suggested that the N-1 of adenine may be protonated and/or that the guanine is in the syn configuration (7,9). In the work described here, we have explored the potential purine and phosphate contacts involved in MutY recognition. Alkylation interference experiments have demonstrated that MutY specifically interacts with mismatched A and G as well as neighboring sequences. This may explain the sequence effect on the repair efficiency of MutY. Defined oligonucleotides containing various purines were used in this study to examine the substrate specificity of MutY protein and to establish the role of mispair functional groups in MutY recognition and catalysis.
Our results add nebularine/G (N/G), inosine/G (I/G), and 2-amino-purine (2AP/G) to the growing list of recognized substrates of MutY, although I/G and 2AP/G are not cleaved well by MutY.

EXPERIMENTAL PROCEDURES
MutY Protein-The purification of homogenous MutY protein from an overproducer strain of E. coli has been described previously (7). The enzyme used in our studies had a specific activity of about 13 ϫ 10 6 units/mg. 1 unit of endonuclease activity is defined as that resulting in cleavage of 0.018 fmol of labeled A/G-DNA in 30 min at 37°C. Iron analysis (31) of purified MutY showed that it contained 3.8 iron atoms/ monomer.
Oligonucleotide Substrates-The sequences of 19-mer oligonucleotides are given in Table I. The 40-mer DNA substrates, with fournucleotide overhangs at both 5Ј-ends, were identical to those used by Tsai-Wu et al. (23). The synthesis of the oligonucleotide containing a single GO has been described previously (32). Other modified phosphoramidites were purchased from Glen Research. The oligonucleotides were synthesized on an Applied Biosystems 381A automated synthesizer using standard procedures by Macromolecular Resources (Colorado State University). The deprotected oligonucleotides were electrophoretically purified on 20% sequencing gels (33). One strand of the 40-mer was labeled at its 5Ј-end with T4 polynucleotide kinase and [␥-32 P]ATP (34) prior to annealing with the other strand. Two complementary oligonucleotides were annealed in 7 mM Tris-HCl (pH 7.6), 7 mM MgCl 2 , and 50 mM NaCl at 90°C for 2 min and then cooled gradually to room temperature over 30 min to form heteroduplexes. The annealed 40-mer or 19-mer duplexes were radiolabeled at the 3Ј-end on the upper strand with Klenow fragment of DNA polymerase I for 30 min at 25°C in the presence of [␣-32 P]dCTP (50 Ci at 3,000 Ci/mmol), 20 M dATP, 20 M dTTP, and 20 M dGTP (34). The resulting blunt-ended duplex DNAs were 44 or 20 base pairs in length, respectively. The reaction mixture was passed through a Quick-Spin column (G-50 for the 40-mer and 44-mer and G-25 for the 20-mer) (Boehringer Mannheim).
MutY Endonuclease Assay-The endonuclease activity of MutY, which is the combined action of the glycosylase and AP endonuclease activities, was assayed as described previously (25). The standard reaction mixture contained 20 mM Tris-HCl (pH 7.6), 80 mM NaCl, 1 mM dithiothreitol, 1 mM EDTA, 2.9% glycerol, and 1.8 fmol labeled DNA in a total volume of 10 l. MutY protein, diluted in a buffer containing 20 mM potassium phosphate (pH 7.4), 1.5 mM dithiothreitol, 0.1 mM EDTA, 50 mM KCl, 200 g/ml bovine serum albumin, and 50% glycerol, was added to the reaction mixture and incubated at 37°C for 30 min. The reaction products were analyzed on 8 or 14% polyacrylamide DNA sequencing gels.
MutY Binding Assay-The binding of MutY to various oligonucleotides was assayed by gel retardation (35,36). 3Ј-end-labeled 44-or 20-base pair oligonucleotides (1.8 fmol) were incubated with various concentrations of MutY as in the endonuclease assay, except that 20 ng of poly(dI-dC) was added to each reaction. Protein-DNA complexes were analyzed on 4 or 8% polyacrylamide gels in 50 mM Tris borate (pH 8.3) and 1 mM EDTA as described previously (25).
Ethylation Interference Assay-5Ј-end-labeled 40-mer heteroduplex DNA containing an A/G mismatch was ethylated according to Siebenlist and Gilbert (37) with some modifications. About 2 pmol of DNA in 100 l of 50 mM sodium cacodylate (pH 7.0) was mixed with an equal volume of freshly prepared ethylnitrosourea and incubated at 45°C for 1 h. The reaction was stopped by adding 20 l of 3 M sodium acetate (pH 7.0), 5 g of tRNA, and 150 l of cold ethanol. The ethylated DNA was precipitated with ethanol several times and resuspended in water. Purified MutY (7.2 pmol) was incubated with the ethylated DNA (0.72 pmol) in a 15-l reaction containing 20 mM Tris-HCl (pH 7.6), 5 mM dithiothreitol, 0.1 mM EDTA, 0.01% Nonidet P-40, 50 mM methoxyamine, and 1 g of poly(dI-dC) per ml. After incubating at 30°C for 30 min, the reaction mixture was fractionated on a 4% polyacrylamide gel. Protein-bound and free DNA fractions were electroeluted and ethanol precipitated. Strand cleavage at phosphodiester bonds was performed as described (37), and the samples were analyzed on 10% sequencing gels. A control experiment was performed in the same way except that MutY diluent was used in the binding reaction.
Methylation Interference Assay-The 40-mer DNA substrate (2 pmol, 5Ј-end labeled) was partially premethylated by treatment with 0.05% dimethyl sulfate for 10 min at 23°C (37). The reaction (200 l) was terminated by the addition of 50 l of 1.5 M sodium acetate, 1 M 2-mercaptoethanol, and 250 g/ml tRNA. DNA was precipitated twice with ethanol and dissolved in water. Binding reactions were performed as in the ethylation interference experiments except that a 5-fold molar excess of MutY protein was added to each reaction. After incubating at 30°C for 30 min, the reaction mixture was fractionated on a 4% polyacrylamide gel. Protein-bound and free DNA fractions were electroeluted, ethanol precipitated, and subjected to the Maxam-Gilbert A Ͼ G cleavage reaction (33). The DNA samples were denatured in formamide and electrophoresed through a 10% polyacrylamide sequencing gel. A control experiment was performed in the same way except that MutY diluent was used in the binding reaction.

RESULTS
Ethylation Interference-Ethylnitrosourea ethylates the phosphate backbone and provides a tool to study the contact between proteins and DNA phosphates. The presence of cleavage activity in MutY AP endonuclease creates one problem in the determination of the ethylation effect on the first phosphodiester bond 3Ј to the mispaired A. Therefore, the binding reactions were performed in the presence of Nonidet P-40 and methoxyamine. The addition of Nonidet P-40 enhances binding while methoxyamine inhibits the AP endonuclease activity. Under this condition, more DNA-protein complexes were formed, and the majority of protein-bound DNA remained intact. Alkaline hydrolysis at ethylated phosphates yields products terminating in a 3Ј-hydroxyl or an ethylphosphate (appearing as doublet bands at each nucleotide position) (37). The 3Ј-hydroxyl product of the first phosphodiester bond 3Ј to the mispaired A runs at the same position as the MutY cleavage product in sequencing gel. In this case, the band with 3Јethylphosphate was used to determine the interference effect. Ethylation interference patterns ( Fig. 1) on 40-mer DNA with a Position X represents A, C, I, 2AP, or Z, and position Y represents C, G, GO, I, or 2AP.
an A/G mismatch at position 22 (see Table I) suggest that more than five phosphates (with a free/control ratio greater than 3) are involved in interactions between MutY and its DNA substrate. Two of these phosphates are five to seven phosphodiester bonds away from the mismatched G. Interestingly, although MutY has more contacts on the G-strand than on the A-strand, it is the A-strand that is cleaved by the MutY glycosylase and AP endonuclease. Methylation Interference-Methylation interference (37) was utilized to deduce potential purine base contacts involved in specific complex formation between MutY and A/G-containing DNA. A 5Ј-end-labeled 40-mer duplex oligonucleotide containing an A/G mismatch at position 22 (see Table I) was used in this study. The MutY cleavage product at the first phosphodiester bond ran differently from the chemical cleavage products. The methylation of either the N-7 position of mismatched G or the N-3 position of mispaired A interfered with MutY binding (Fig. 2). Besides these mismatched bases, MutY also contacted purines on both sides of the mismatch. Substantial interference was observed when G-23 on the A-strand and A-20 and G-24 on the G-strand were methylated (Fig. 2). These residues are key determinants of MutY binding to A/G-containing DNA. Thus, two nucleotides on both sides of the mismatched base have a strong influence on MutY reactivity. Methylation of several other bases farther from the mismatch also showed some interference in MutY binding. Fig. 3 summarizes the distribution of potential purine and phosphate contacts of MutY as deduced from the alkylation interference experiments. MutY covers about 12 base pairs of the A/G-containing DNA and has more contacts on the G-strand.
Inosine Substitution Effects-The role of G-23 on the Astrand and G-24 on the G-strand of 40-mer DNA containing an A/G mismatch was examined by substitution with inosine, a guanine base analog lacking a C-2 amino group (Fig. 4). Based on ethylation interference data, 20-mer oligonucleotides were designed to investigate base substitution effects (Table I). The 20-mer oligonucleotides containing an A/G mismatch were bound and cleaved very well by MutY (Fig. 5, lane 1). DNA containing inosine in place of guanine at position 23 or 24 The DNA was partially methylated with dimethylsulfate, purified, and bound to MutY. After binding to MutY, free and bound DNA were separated by 4% native polyacrylamide gel electrophoresis, eluted, and chemically cleaved by the A Ͼ G reaction (33). Control samples are 40-mer heteroduplex DNA treated by the same procedures without adding MutY protein. Cleavage products were then analyzed on 10% DNA sequencing gels and exposed to x-ray films. Panel B shows a representative autoradiogram of methylation interference on the Gstrand. Interference was quantitated and shown in panel A as described in Fig. 1. The data shown here are the average of at least five experiments. Error bars are Ϯ1 S.D. resulted in a reduction of MutY endonuclease activity by 46 and 22%, respectively (Fig. 5, lanes 2 and 6).
Uracil Substitution Effects-Individual thymines within three nucleotides on both sides of the mismatched base were replaced with deoxyuracil. Uracil substitution for T at the first and second nucleotide 5Ј to mismatched A had no effect on MutY endonuclease activity (Fig. 5, lanes 3 and 4). When the third nucleotide 3Ј to the mismatched G was substituted by uracil, the endonuclease activity was slightly enhanced (Fig. 5,  lane 7). Thus, the 5-methyl group of thymine at these positions was not essential for the interaction of MutY protein with its substrates.
Endonuclease Activity of MutY-Four mismatches (A/G, A/C, A/GO, and A/AO) have been reported to be substrates for MutY (7)(8)(9)24). To further delineate the base specificities of MutY and to determine the functional groups in the mispair required for MutY recognition, the A/G mispair was substituted with different purines without changing the surrounding sequence. These purines included nebularine (N), inosine (I), 2-aminopurine (2AP), and 7-deaza-adenine (Z) (Fig. 4). As shown in Fig.  6, MutY nicked different mismatches with different efficiencies. Duplexes containing A/C (Fig. 6, lane 2) and A/GO (Fig. 6,  lane 3) were cleaved about one-third and one-half the extent, respectively, of DNA containing an A/G mismatch (Fig. 6, lane 1). The duplex containing N/G (Fig. 6, lane 4) was cleaved as efficiently as the A/G mismatch. Duplexes containing I/G and Z/G (Fig. 6, lanes 5 and 7) and homoduplexes (Fig. 6, lanes 10  and 11) were not cleaved by MutY, and the duplexes containing A/2AP (lane 9), A/I (Fig. 5, lane 5, and Fig. 6, lane 8), and 2AP/G (Fig. 6, lane 6) were very weak substrates for MutY. Further treatment with piperidine after MutY reaction did not increase the amount of cleavage products, suggesting these base substitutions affect the glycosylase activity of MutY (data not shown). The endonuclease activity of MutY was of the following order: Binding Affinity of MutY for Different Mismatches-In the gel mobility shift assay, MutY protein formed complexes with 20-mer oligonucleotides containing different mismatches with four slightly different mobility shifts, although the free DNA duplexes had the same mobility ( Fig. 7 and  The apparent dissociation constants (K d ) of MutY from the different mismatches were determined. Representative autora-diograms of the binding assay and the corresponding binding curve for MutY to A/G-and A/I-containing 20-mer DNA are shown in Fig. 8 (panels A-D). In A/G binding assays, there was a band that migrated faster than the free DNA and represented about 5 and 15% of input DNA at 3.4 and 53.7 nM MutY, respectively, indicating that some A/G-containing DNA was cleaved by and dissociated from MutY (Fig. 8A). Cleavage products were not used in the K d calculation because their incusion would negligibly and improperly increase the apparent values. The cleaved free DNA band in native gel was not observed in the binding of MutY at 53.7 nM to A/GO-containing DNA (data not shown). Some slower mobility complexes (either polymers or aggregated forms) were observed in binding assays with low affinity DNA substrates (A/2AP, Z/G, A/I, A/T, and C/G) when the concentration of MutY protein was higher than 0.34 M (Fig. 8B). The results of these experiments, performed in triplicate, are summarized in Table II.
The apparent K d values for 20-mer and 44-mer DNA containing an A/G mismatch were 5.3 and 1.8 nM, respectively. MutY bound strongly to the 20-mer duplex with an A/GO mismatch (apparent K d ϭ 66 pM). This binding is 80-fold greater than that to A/G-containing DNA. The binding affinities of MutY for 20-mer oligonucleotides containing N/G and I/G were comparable to that for A/G. MutY bound weakly to duplex DNA containing A/C and 2-AP/G, much weaker to DNA with A/2AP, Z/G, and A/I, and very weakly to duplex DNA containing a matched A/T or C/G base pair. Reproducibly, MutY bound slightly better to A/T-than to C/G-containing DNA (Fig. 7,  lanes 10 and 11). The binding affinity of MutY was of the following order:

DNA containing A/G, A/C, A/GO, or A/AO mismatch has been
shown to be the substrate of MutY protein (7,9,24,25). The MutY protein removes the mispaired adenines from the mismatches by glycosylase activity (7,24). Tsai-Wu et al. (7) showed that MutY also has AP endonuclease activity. However, the AP endonuclease activity could not be detected by Au et al. (24) in their MutY preparation. The reason for this discrepancy is not clear. The detection of the dissociated nicked product in a native gel (Fig. 8A) argues that the endonuclease activity observed in MutY reaction is not caused by heating at high pH, which may catalyze a ␤-elimination at the AP site. Therefore, the AP endonuclease (or lyase) activity of MutY appears intrinsic. DNA phosphate ethylation experiments suggest that MutY interacts with phosphate residues spanning about 12 nucleotide pairs encompassing the A/G mismatch (Fig. 3). MutY binds DNA mainly via five major ionic bonds. Two phosphates with significant effects upon ethylation are located 5 and 7 phosphodiester bonds 3Ј and 5Ј to the mismatched G, respectively. This defines a large component of electrostatic binding energy involved in the binding of MutY to phosphates outside of the A/G mismatch. Interestingly, ethylation at the first phosphodiester bond 3Ј to the mispaired A that is attacked by the MutY AP endonuclease did not have a substantial interference effect on MutY binding (Fig. 1).
Methylation interference experiments with 40-mer DNA containing an A/G mismatch (Table I) reveal that MutY interacts with purines including the mismatched A and G and two bases on either side of the mismatch (Fig. 3). The N-3 group of A-20 on the G-strand is located in the minor groove while the N-7 groups of G-23 on the A-strand and G-24 on the G-strand are located in the major groove. Substitution of these two guanines by inosines indicates that the 2-amino groups of these guanines, located in the minor groove, are also important in MutY recognition. If substantial helix perturbation is not associated with MutY binding, these findings suggest that MutY binds DNA in both the major and minor grooves in the vicinity of the A/G mismatch. The involvement of purines flanking the mismatch in MutY binding may reflect the effect of neighboring sequences on repair and cleavage efficiencies (2, 25, 38). Al-though flanking sequence effects on MutY reactivity have not been explored systematically, our findings provide a rationale for earlier observations. The locations of the N-7 group of mismatched G and the N-3 group of mismatched A depend on the structure of the A/G mismatch. Structural analyses have shown that A/G can form three possible conformations: A(anti)-G(anti), A(anti)-G(syn), and A(syn)-G(anti), depending on the neighboring environment (26 -30). Because A/GO is a substrate for MutY protein and it can form a very stable base pair when A is in the anti and GO is in the syn configuration (39), it is suggested that the A(anti)-G(syn) conformation may be the favored substrate for MutY (9). The protonated N-1 form of adenine is also suggested for the A/C structure (40). Although these DNA conformations with bound MutY have yet to be confirmed, we will assume in the following discussion that the A/G-containing substrate for MutY is in the A(anti)-G(syn) conformation, in which both the N-3 group of mismatched A and the N-7 group of mismatched G are located in the minor groove.
The role of functional groups in mismatch recognition and catalysis by MutY protein were further elucidated by binding and endonuclease assays using defined oligonucleotides containing various purine derivatives. The C-6 amino group of mismatched A is not critical for DNA binding or cleavage since duplex DNA containing N/G is bound and cleaved efficiently by MutY. This is a surprising finding because the C-6 amino group of adenine is involved in hydrogen bonding with guanine in all  Table I for sequence).  Table I for sequence). Endonuclease activity is defined as that resulting in cleavage of 0.018 fmol (1%) of labeled DNA in 30 min at 37°C.  Fig. 6. The protein-bound and free DNA are indicated by arrows. Under this condition, some cleaved product of A/G-and N/G-containing DNA were dissociated from MutY and migrated off the gel (see Fig. 8A). Note that little free DNA was present in lanes 1-6 at this MutY concentration and complexes migrated at slightly different rates. three A/G conformations (26,30). The loss of one hydrogen bond would be expected to destabilize the A/G base pairing. Thus, the reactivity of MutY protein cannot be rationalized on the basis of the stability of the mismatched pair alone. However, the presence of a C-6 keto group as in the I/G pair blocks catalysis but not binding by MutY. The C-6 keto group may prevent the use of a protonated N-1 or N-3 as an "electron sink" in glycosidic bond cleavage. The replacement of the N-7 group of adenine by carbon as in the Z/G pair results in the loss of both binding and endonuclease activities by MutY. The introduction of a C-2 amino group decreases MutY binding and catalysis (compare 2AP/G with N/G). The C-2 amino and the C-6 keto groups of mismatched G, presented in the major groove as in G(syn)-A(anti), are critical for MutY recognition since A/I and A/2AP mismatches are poorly bound by the enzyme.
When the C-8 keto group of GO is present as in GO(syn)-A(anti), the apparent K d for MutY decreases by 2 orders of magnitude, yet the endonuclease activity is reduced 2-fold as compared to the A/G mispair. The presence of a C-8 keto group in A/GO pair changes the nature of hydrogen bonding between N-7 of GO and N-1 of A. The N-1 of A is protonated in the G(syn)-A(anti) pair (27) but is not protonated in the GO(syn)-A(anti) pair (39). These structural differences between A/GO and A/G are located in the minor groove (39). The relative binding and catalytic efficiencies of MutY for A/G-and A/GOcontaining DNA substrates may have biological significance. This may prevent MutM from acting upon the GO lesions so information on both DNA strands would not get lost. According to Tchou et al. (41), the apparent K d values of MutM (FPG) protein for A/GO-and C/GO-containing DNA were 340 and 8.9 nM, respectively. The affinity of MutY to A/GO-containing DNA is about 5,000-fold higher than that of MutM.
When the potential DNA contact sites of MutY elucidated in this study are mapped onto a B-DNA model with an A(anti)-G(syn) mispair, several interesting features are revealed. MutY appears to interact with its DNA substrate on both sides of the double helix as well as in both the major and minor grooves. Most base determinants are accessible from one side of the helix. Contacts on the other side of the helix are mediated were excised and quantified by liquid scintillation counting. The % of bound A/G-containing DNA was determined, plotted versus MutY concentration, and the value for K d was determined by Enzfitter (42). The apparent K d for A/I-containing DNA was determined as the concentration of MutY that results in 50% binding of input DNA. In this case, MutY binding to A/I-20 is not saturated at the highest MutY concentration tested and is not suitable for Enzfitter analysis. mainly through three phosphates (between T-15 and C-16, between T-20 and T-21, and between A-26 and A-27). It is interesting to note that these three phosphates are approximately five base pairs apart and form a line nearly parallel to the helical axis. The information obtained in this study provides a basis for understanding the interaction of MutY with DNA and the molecular mechanisms involved in the recognition of damaged or mismatched bases. Identification of I/Gcontaining DNA as a recognized but not a catalyzed substrate will facilitate the formation of protein-DNA co-crystal.