Characterization of the Recombinant MutY Homolog, an Adenine DNA Glycosylase, from Yeast Schizosaccharomyces pombe *

The mutY homolog (SpMYH) gene from a cDNA library of Schizosaccharomyces pombeencodes a protein of 461 amino acids that displays 28 and 31% identity to Escherichia coli MutY and human MutY homolog (MYH), respectively. Expressed SpMYH is able to complement an E. coli mutY mutant to reduce the mutation rate. Similar to E. coli MutY protein, purified recombinant SpMYH expressed inE. coli has adenine DNA glycosylase and apurinic/apyrimidinic lyase activities on A/G- and A/7,8-dihydro-8-oxoguanine (8-oxoG)-containing DNA. However, both enzymes have different salt requirements and slightly different substrate specificities. SpMYH has greater glycosylase activity on 2-aminopurine/G and A/2-aminopurine but weaker activity on A/C thanE. coli MutY. Both enzymes also have different substrate binding affinity and catalytic parameters. Although SpMYH has great affinity to A/8-oxoG-containing DNA as MutY, the binding affinity to A/G-containing DNA is substantially lower for SpMYH than MutY. SpMYH has similar reactivity to both A/G- and A/8-oxoG-containing DNA; however, MutY cleaves A/G-containing DNA about 3-fold more efficiently than it does A/8-oxoG-containing DNA. Thus, SpMYH is the functional eukaryotic MutY homolog responsible for reduction of 8-oxoG mutational effect.

Cellular and organism aging have been correlated with accumulated DNA damage (1,2). Oxygen is metabolized inside the cell by a series of one-electron reductions with the generation of reactive and potentially damaging intermediates called reactive oxygen species (3). The frequency of oxidative damage to DNA has been estimated at 10 4 lesions/cell/day in humans (4). 8-Oxo-7,8-dihydrodeoxyguanine (8-oxoG or GO 1 ) is one of the most stable products of oxidative DNA damage. The formation of GO in DNA, if not repaired, can lead to misincorporation of A opposite to the GO lesion and result in G:C to T:A transversions (5)(6)(7)(8). In Escherichia coli, a family of enzymes, MutY, MutM, and MutT, is involved in defending against the muta-genic effects of GO lesions (9 -11). The MutT protein has nucleotide triphosphatase activity, which eliminates 8-oxo-dGTP from the nucleotide pool (12). The MutM protein (Fpg protein) provides a second level of defense by removing both mutagenic GO adducts and ring-opened purine lesions (13)(14)(15)(16). The E. coli MutY is an adenine glycosylase that is responsible for the correction of A/GO as well as A/G and A/C mismatches (9,(17)(18)(19)(20)(21)(22). MutY removes misincorporated adenines paired with GO lesions and reduces the GO mutational effects. Recent results show that MutY and the N-terminal catalytic domain can be trapped in a stable covalent enzyme-DNA intermediate in the presence of sodium borohydride (23)(24)(25) and support the hypothesis that MutY contains both DNA glycosylase and AP lyase activities.
MutY homologous (MYH) activities have been identified in human HeLa (26) and calf thymus (27) extracts. Both human and calf MYH systems share similar features with the E. coli mutY-dependent pathway: mismatch specificities to A/G, A/C, and A/GO, and cleavage of the A but not G strand. Recently, a human cDNA of putative hMYH was cloned and its open reading frame predicts a 60-kDa protein (28), which is in good agreement with the size of a band detected in HeLa extracts with MutY antibodies (27). hMYH shares high homology with the E. coli MutY protein (28). However, no enzyme activity has been reported for the protein encoded by this open reading frame.
Here, we report the cloning and expression of the MYH gene from Schizosaccharomyces pombe. Expression of SpMYH suppresses the spontaneous mutation rate of E. coli mutY mutant strains. Like E. coli MutY, purified recombinant SpMYH has both adenine glycosylase and AP lyase activities on A/G and A/GO mismatches. Defined oligonucleotides containing various purines were used to examine the substrate specificity. SpMYH has slightly different substrate specificity from that of E. coli MutY protein. SpMYH has greater glycosylase activity on 2-aminopurine (2AP)/G and A/2AP but weaker activity on A/C than does E. coli MutY. Both enzymes also have different substrate binding affinity and catalytic parameters. These results suggest that SpMYH is a functional eukaryotic homolog of the bacterial MutY. The high homology of MutY homologs among different organisms suggests important roles in their cellular functions.

EXPERIMENTAL PROCEDURES
Cloning of cDNA of S. pombe MYH-According to the published genomic sequence of S. pombe (accession no. Z69240), the putative mutY homolog (MYH) sequence contains two introns and codes for a 461-residue protein. To clone the S. pombe MYH gene, we synthesized two PCR primers, Chang 219 (5Ј-GGAGATATACATATGTCGGATTC-AAATCATTC-3Ј) and Chang 220 (5Ј-GCAGCCGGATCCTTAGCACTC-TGCTTTCGT-3Ј). Chang 219 and Chang 220 anneal at the first six and last six codons of the predicted coding sequence for SpMYH, respectively. DNA prepared from an S. pombe cDNA library in pGADGH (kindly provided by D. Beach, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY) was used as a template for the PCR reactions. PCR * This work was supported by Public Health Service Grant GM 35132 from NIGMS, National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF053340.
‡ To whom correspondence should be addressed. To obtain a mutation-free SpMYH clone, the 1.4-kb PCR product was labeled and used as a probe to screen the S. pombe cDNA library in pGADGH by colony hybridization. Out of 6 ϫ 10 6 colonies screened, two contained the SpMYH sequence. Subsequent restriction and PCR analysis indicated that the MYH cDNA was within a 1.6-kb BamHI fragment in one of the clones. This fragment was isolated and transferred to pUC19 to generate pSPMYH19. Based on the restriction map of pSP-MYH19, the 1.1-kb EcoRI fragment containing the C-terminal domain of SpMYH and 21 base pairs of pUC19 sequence was isolated and ligated with the 5.76-kb fragment of pSP11a-2.3 that contained the vector pET11a and the N-terminal portion of SpMYH to yield the plasmid pSPMYH11a-4. The SpMYH sequence of pSPMYH11a-4 (accession no. AF053340) was exactly the same as the predicted sequence for the S. pombe MYH cDNA. The recombinant expressed a 52-kDa protein in GBE943(DE3) (lacIp4000(LacI q )lacZp4008(Lac L8)srlC-300::Tn10 Ϫ IN(rrD-rrnE)1micA68::Tn10Kan) cells following induction with 0.4 mM isopropyl-1-thio-␤-D-galactopyranoside (IPTG). The expression host, GBE943 with DE3 lysogen, was constructed according to the procedures described by Invitrogen.
Measurement of Mutation Rate-Independent overnight cultures of each strain were grown to an A 590 of 0.7 in LB medium containing 50 mg/ml ampicillin when necessary. After 2.5 h of induction by the addition of 0.1 mM IPTG, 0.1 ml of cells from each culture was plated onto LB agar containing 0.1 mg/ml rifampicin. The cell titer of each culture was determined by plating a 10 Ϫ6 dilution onto LB agar. The ratio of Rif r cells to total cells was the mutation rate.
Purification of SpMYH-Eighteen liters of E. coli GBE943/De3 cells harboring overproduction plasmid pSPMYH11a-4 were cultured to an A 590 of 0.7 in LB broth containing 50 mg/ml ampicillin at 37°C. The cells were induced by adding IPTG to 0.4 mM and cultured overnight at 28°C and harvested by centrifugation. All column chromatography was conducted in a Waters 650 FPLC system at 4°C, and centrifugation was done at 16,5000 ϫ g for 30 min. Cells (54 g of cell paste) were resuspended in 200 ml of buffer A (20 mM potassium phosphate (pH 7.4), 1 mM dithiothreitol, 0.2 mM phenylmethylsulfonyl fluoride) containing 50 mM KCl and disrupted with a bead beater (Biospec Products, Bartlesville, OK) using 0.1-mm glass beads. The cell debris was removed by centrifugation, and the supernatant was treated with 5% streptomycin sulfate. After stirring for 30 min, the solution was centrifugated, and the supernatant was collected as fraction I (595 ml). Ammonium sulfate (134 g) was added to fraction I, and the protein was precipitated for 1 h. After centrifugation, 126 g of ammonium sulfate was added to the supernatant and the protein pellets collected by centrifugation were resuspended in 50 ml of buffer A containing 50 mM KCl and dialyzed against two changes of 3 liters of the same buffer for 4 h each. The dialyzed protein sample was diluted 2-fold with buffer A containing 50 mM KCl as fraction II (140 ml). Fraction II was loaded onto a 50-ml phosphocellulose column, which had been equilibrated with buffer A containing 0.05 M KCl. After washing with 100 ml of equilibration buffer, proteins were eluted with a 400-ml linear gradient of KCl (0.05-0.6 M) in buffer A. Fractions that eluted between 0.2 and 0.4 M KCl were pooled (fraction III, 50 ml). Fraction III was loaded onto a 30-ml hydroxylapatite column equilibrated with buffer B (0.01 M potassium phosphate (pH 7.4), 10 mM KCl, 0.5 mM dithiothreitol, 0.1 mM EDTA, 0.1 mM phenylmethylsulfonyl fluoride). After washing with 60 ml of equilibration buffer, the flow-through and early elution fractions were pooled and dialyzed against buffer A containing 0.05 M KCl and 10% (v/v) glycerol for 2 h (fraction IV, 53 ml). Fraction IV was loaded onto a 5-ml heparin-agarose column equilibrated with buffer A containing 0.05 M KCl and 10% glycerol. After washing with 10 ml of equilibration buffer, the column was developed with a 50-ml linear gradient of KCl (0.05-0.6 M) in buffer A with 10% glycerol. Fractions containing the MYH nicking activity, which eluted between 0.15 and 0.3 M KCl, were pooled and dialyzed against 2 liters of buffer A containing 0.05 M KCl and 10% glycerol to yield fraction V (78 ml). Fraction V was then applied to a 1-ml Mono S column that had been equilibrated in buffer A containing 0.05 M KCl and 10% glycerol. After washing with 20 ml of equilibration buffer, the column was eluted with a step gradient of 0.1, 0.3, and 0.5 M KCl in 20 ml of buffer A each with 10% glycerol. Fractions containing the MYH nicking activity, which eluted between 0.1 and 0.3 M KCl, were pooled (fraction VI, 8.5 ml), divided into small aliquots, and stored at Ϫ80°C. Cleavage of A/G-containing 44-mer DNA was assayed during the purification of the recombinant SpMYH enzyme. One unit of activity is defined as that resulting in cleavage of 0.018 fmol of labeled DNA in 30 min at 30°C. Protein concentration was determined by the Bradford method (29).
Oligonucleotide Substrates-Oligonucleotides of 19-mer and 40-mer containing base mismatches were labeled as described by Lu et al. (30).
SpMYH Nicking Assay-The nicking activity of SpMYH, which is the combined action of the glycosylase and AP lyase activities, was assayed similarly as described (31). The standard reaction mixture contained 10 mM Tris-HCl (pH 7.6), 0.5 mM dithiothreitol, 0.5 mM EDTA, 1.45% glycerol, 50 g/ml bovine serum albumin, and 1.8 fmol of labeled DNA in a total volume of 20 l. SpMYH 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 30°C for 30 min. The reaction products were analyzed on 8 or 14% polyacrylamide DNA sequencing gels. Kinetic analyses were performed using a concentration range of DNA substrates with 0.4 nM SpMYH. Following autoradiography, bands corresponding to cleavage products and intact DNA were excised from the gel and quantified by liquid scintillation counting. K m and V max values were obtained from results of three experiments by a computer-fitted curve generated by the Enzfitter program (32).
SpMYH Binding Assay-The binding of SpMYH to various oligonucleotides was assayed by gel retardation. 3Ј end-labeled 44-bp or 20-bp oligonucleotides (1.8 fmol) were incubated with various concentrations of SpMYH in a 20-l binding buffer containing 10 mM Tris-HCl (pH 7.6), 0.5 mM dithiothreitol, 40 mM NaCl, 5 mM EDTA, 1.15% glycerol, 50 g/ml bovine serum albumin, and 5 ng of poly(dI-dC) at 30°C for 30 min. Protein-DNA complexes were analyzed on 8% polyacrylamide gels in 50 mM Tris borate (pH 8.3) and 1 mM EDTA as described previously (19). The apparent dissociation constants (K d values) of SpMYH and DNA were determined using a range of protein concentrations. Following autoradiography, bands corresponding to enzyme-bound and free DNA were excised from the gel and quantified by liquid scintillation counting. K d values were obtained from results of three experiments by a computer-fitted curve generated by the Enzfitter program (32).
Formation of Enzyme-DNA Covalent Complex-Reactions were carried out as described in the SpMYH cleavage assay except that the reactions were performed in the presence of NaBH 4 . A NaBH 4 stock solution was freshly prepared immediately prior to use. After incubation at 30°C for 30 min, SDS dye was added to the samples, which were heated at 90°C for 2 min and separated on a 12% polyacrylamide gel in the presence of SDS according to Laemmli (33), and the gel was dried and autoradiographed.

RESULTS
Sequence Analysis of the SpMYH Gene-The SpMYH gene encodes a protein of 461 amino acid residues that displays 28 and 31% identity and 59 and 62% conservation to the E. coli MutY and human MYH, respectively, by comparison using an ALIGN program. The eukaryotic MYH sequences contain extra amino acid stretches at the N-and C-terminal regions as compared with the bacterial MutY sequences. The amino acid sequences of the N-terminal part of the SpMYH share significant homology to other DNA glycosylases including E. coli endonuclease III (endo III) (Fig. 1). Two regions of highest similarity shared among these enzymes can be identified in the known three-dimensional structure of E. coli endo III (34,35): the helix-hairpin-helix domain (residues 147-181 of SpMYH) and the iron-sulfur domain (residues 213-242 of SpMYH). There are also distinct residues (shaded in gray boxes in Fig. 1) in the MutY family that are different from that of endo III family. Interestingly, when these residues are placed on the structure of E. coli endo III, they are located at the edge of the cleft between the iron-sulfur domain and the six antiparallel ␣-helices.
Reduction of the Mutation Rate of the E. coli mutY Mutant by S. pombe MYH-To demonstrate that the open reading frame of putative SpMYH gene encodes a functional MYH protein, we have expressed this Sp-cDNA under the control of the T7 promoter in pET11a in E. coli. E. coli GBE943 (mutY) harboring the plasmid pSPMYH11a-4 were induced to express the Sp-MYH protein by addition of IPTG to the growth medium, and the mutation rate was measured. As shown in Table I, mutY mutant (GBE943/DE3) exhibited a 40-fold higher mutation rate than the wild-type E. coli. GBE943/DE3 cells expressing SpMYH had mutation rates almost as low as the wild-type cell, whereas the vector (pET11a) alone had no effect on the mutation rate ( Table I). Expression of SpMYH in wild-type E. coli cells caused a slightly lower mutation rate than cells with vector alone (Table I).
Purification of Recombinant SpMYH Protein-To further demonstrate that the putative SpMYH encodes a functional MYH protein, we purified the recombinant SpMYH from the overproducing E. coli GBE943(DE3) strain harboring the plasmid pSPMYH11a-4. The SpMYH protein was purified by ammonium sulfate precipitation and phosphocellulose, hydroxylapatite, heparin-agarose, and Mono S chromatographic separation. We recovered 19 mg of SpMYH protein from 54 g of cell paste with about an 18-fold increase in specific activity. The purity of the products at different stages of the procedure is illustrated in Fig. 2. As judged on a 10% SDS-polyacrylamide gel, the protein was purified to a very high degree. The mobility of SpMYH in the gel matched the predicted size (52 kDa). SpMYH appears as a monomer because it was eluted at a molecular mass of about 45 kDa through a gel-filtration column (Superose 12) (data not shown).
Effects of Salt and EDTA on SpMYH Activities-When Sp-MYH was initially assayed in the MutY buffer (20 mM Tris-HCl, pH 7.6, 1 mM dithiothreitol, 1 mM EDTA, 80 mM NaCl, and 2.9% glycerol), there was little cleavage activity (Fig. 3, lane 4). This prompted us to find the optimal conditions for SpMYH. The SpMYH glycosylase activity was reduced by adding 40 mM NaCl to the reaction buffer (Fig. 3, lane 8) and was abolished at NaCl concentration higher than 80 mM (Fig. 3, lanes 9 and 10). SpMYH cleavage activity was not inhibited by 8 mM EDTA (Fig. 3, lane 13) but was abolished by 32 mM EDTA (Fig. 3, lane  15). The effects of salt and EDTA on the SpMYH binding to  20-mer DNA containing an A/G mismatch were similar to that on the SpMYH cleavage except 40 mM NaCl had little effect on binding (data not shown). Therefore, the low SpMYH activity in MutY buffer is due to the presence of 80 mM NaCl.
SpMYH Contains Both Adenine DNA Glycosylase and AP Lyase Activities-As shown in Fig. 4, SpMYH can cleave a 20-mer oligonucleotide containing an A/G mismatch in a dosedependent manner. Heating the samples at 90°C for 2 min before loading to the sequencing gel enhanced the cleavage activity (Fig. 4, compare A and B). However, further treatment of the products with piperidine at 90°C (a condition promotes ␤-elimination) did not significantly increase the extent of cleavage (Fig. 4, compare B and C). Thus, SpMYH does contain intrinsic AP lyase activity, although it is not strictly coupled to the glycosylase activity. These properties of SpMYH are similar to the E. coli MutY protein, which contains both DNA glycosylase and AP lyase activities. If SpMYH has an AP lyase activity and uses the similar mechanism like MutY (23)(24)(25), an imino intermediate should be reduced by NaBH 4 to form a stable covalent protein-DNA complex. Thus, DNA containing an A/G mismatch was incubated with SpMYH in the presence of different concentrations of NaBH 4 . As shown in Fig. 5, SpMYH can be trapped in two covalently linked protein-DNA complexes as a doublet in the presence of NaBH 4 although not as efficiently as MutY. The reason for the formation of two complexes is not quite clear. The optimal trapping concentration of NaBH 4 is about 20 -30 mM (Fig. 5A) (less trapping is observed at 10 mM NaBH 4 with data not shown). The covalent complexes of SpMYH and DNA migrated slower than the MutY-DNA complex due to the larger size of SpMYH protein (the molecular mass of SpMYH is 52 kDa and that of MutY is 39 kDa). At 100 mM NaBH 4 , E. coli MutY can be trapped efficiently but the trapping ability of SpMYH is minimal (data not shown). The effect of NaBH 4 is consistent with the inhibition of SpMYH cleavage activity by 80 mM NaCl that is present in the MutY reaction (Fig. 3). In the presence of 30 mM NaBH 4 , DNA containing an A/G mismatch was tested in the trapping assay with increasing amount of SpMYH protein (Fig. 5B). At enzyme:DNA molar ratios ranging from 10 (Fig. 5B, lane 3) to 320 (Fig. 5B, lane 8), covalent complexes were detected. Because Schiff's base formation is an important criteria for the class of DNA glycosylase with AP lyase activity, our results strongly suggest that SpMYH protein possesses AP lyase activity.
SpMYH and MutY Have Different Substrate Specificities-We have shown that E. coli MutY can cleave several mismatches with different efficiencies (30). To compare the SpMYH with MutY, we tested the glycosylase activity on different mismatches. As shown in Fig. 6, both enzymes cleave A/G, N/G, and A/GO in a similar order and have very weak or no cleavage on I/G, A/I, C/GO, and C/G. However, SpMYH has greater activity on 2AP/G and A/2AP but weaker activity on A/C than E. coli MutY.
Binding Affinity of SpMYH and MutY for Different Mismatches-Because SpMYH and MutY have different glycosylase activities on different substrates, we then determined the apparent dissociation constants (K d ) of SpMYH from different mismatches. The apparent dissociation constants (K d values ) of SpMYH and DNAs were determined using a range of protein concentrations with a fixed DNA concentration (90 pM). Representative autoradiograms of the binding assays and the corresponding binding curves for SpMYH to A/G-and A/GO-containing 44-mer DNA are shown in Fig. 7. When the concentration of SpMYH protein was higher than 50 nM, an extra slower migrating complex was observed in binding assays with A/G-44 DNA (Fig. 7A). SpMYH binding is saturated below 70% of A/GO-44 DNA at the highest MutY concentration tested (Fig. 7D).
As shown in Table II, SpMYH has great affinity to A/GOcontaining DNA as does MutY, but the binding affinity to A/G-containing DNA is substantially lower for SpMYH than MutY. The difference in the binding affinity of SpMYH with

A/G-44 and A/GO-44 is 155-fold and with A/G-20 and A/GO-20
is 700-fold. The difference in the binding affinity of MutY with A/G-44 and A/GO-44 is 13-fold and with A/G-20 and A/GO-20 is 80-fold. SpMYH has higher nonspecific binding to 44-mer homoduplex than MutY. Therefore, SpMYH has only a 4 -7-fold higher binding affinity to A/G mismatch than C/G pair, whereas MutY has a 70 -175-fold higher binding affinity to A/G mismatch than C/G pair. As shown above, SpMYH has greater cleavage activity on 2AP/G and A/2AP than MutY; however, the binding affinities of SpMYH to these two mismatches were not greater than that of MutY.
Kinetic Parameters of SpMYH-The cleavage efficiencies of 20-mer oligonucleotide containing A/G or A/GO by SpMYH and MutY were compared (Table III). As measured at 0.4 nM protein concentration, the K m values for SpMYH on both substrates were slightly higher than that of MutY. The turnover number (K cat ) for SpMYH on an A/G 20-mer is 2 times lower than that of the MutY. SpMYH has similar reactivity (K cat /K m ) to both A/G-and A/GO-containing DNA; however, MutY cleaves A/G-containing DNA about 3-fold more efficiently than it does A/GO-containing DNA. When 44-mer DNA substrates were tested for cleavage, SpMYH also displays similar specificity constants (K cat /K m ) with both A/G and A/GO mismatches.
The specificity constants of SpMYH with 44-mer is 3-4 times higher than with 20-mer DNA substrates. DISCUSSION As part of the genome project, the genomic sequence of mutY homolog of S. pombe was obtained (SPAC26A3.02 in entry Z69240). To confirm that this sequence codes for a functional MutY-like protein, we expressed the SpMYH cDNA in E. coli and assayed the functions of the recombinant protein. SpMYH is shown to be the functional MutY homolog by its ability to complement the mutator phenotype of an E. coli mutY mutant containing expressed SpMYH (Table I). This proves the conservation of the repair function in these organisms. The SpMYH protein displays similar extent of identity and conservation to both E. coli MutY and human MYH. SpMYH belongs to a superfamily of base-excision DNA repair proteins that recognize diverse lesions such as oxidized purines, fragmented and oxidized pyrimidines, UV-cross-linked bases and base-base mismatches (36). In addition to the MutY and endo III families, this superfamily also includes the OGG1 (8-oxoG glycosylase) family (37)(38)(39)(40) and AlkA family (41,42). As shown in Fig. 1, MutY and endo III families share significant homology. Two regions of highest similarity can be identified in the known FIG. 4. SpMYH exhibits DNA glycosylase and AP lyase activities on A/G substrates. Purified recombinant Sp-MYH was used in a cleavage assay with 1.8 fmol of A/G substrate at the indicated enzyme to DNA ratios. Lane 1 represents 1.8 fmol of untreated DNA. Reactions were carried out at 30°C for 1 h, and DNA samples were fractionated on 14% sequencing gels. A, samples were resuspended in sequencing dye and loaded to the gel. B, samples were resuspended in sequencing dye, heated at 90°C for 2 min, and then loaded to the sequencing gel. C, after SpMYH reaction, piperidine was added to the samples to a final concentration of 1 M and then heated at 90°C for 30 min. Samples were treated similar to B. The positions of intact oligonucleotide (I) and nicking product (N) are indicated by arrows. three-dimensional structure of E. coli endo III (34,35): the helix-hairpin-helix (HhH) domain (residues 147-181 of Sp-MYH) and the iron-sulfur domain (residues 213-242 of Sp-MYH). The HhH motif with a ␤-hairpin structure was first identified as the binding site for thymine glycol in crystals of endo III (34). Due to the diverse substrate specificities, binding to HhH presumably reflects interactions that are common to all substrates. Residues in the vicinity of the iron-sulfur cluster have also been proposed to be involved in DNA binding, and the conservation of residue spacing for the cysteines (with the exception of endo III of S. pombe; see Fig. 1) that ligate the iron-sulfur cluster suggests that this structure is very similar in the endo III and MutY protein families (35).
The recombinant SpMYH protein has been shown to incise A/G-, N/G-, A/GO-, 2AP/G-, and A/2AP-containing DNA. Cleavage is achieved by both DNA glycosylase and AP lyase activities. Two lines of evidence strongly suggest that SpMYH possesses AP lyase activity. 1) The detection of the nicked product in a sequencing gel without heating the samples at 90°C (Fig.  4A) argues that the cleavage activity observed in SpMYH reaction is not caused by heating at high pH, which may catalyze a ␤-elimination at the AP site. 2) SpMYH can be trapped in a stable covalent enzyme-DNA intermediate with A/G-containing DNA in the presence of sodium borohydride. Imino enzyme-DNA intermediates are characteristic of a group of glycosylase/AP lyases including T4 endonuclease V, Micrococcus luteus UV endonuclease, E. coli MutY (23)(24)(25), E. coli endonuclease III, and E. coli FPG (MutM) (43)(44)(45). These enzymes use an amino group as the nucleophile, resulting in an imino enzyme-DNA intermediate (43). In addition, SpMYH cleavage activity is not inhibited by 8 mM EDTA. Thus, SpMYH belongs to that class of DNA glycosylases that possess concomitant AP lyase activity.
The glycosylase/AP lyase activity of SpMYH on A/G-20 and AGO-20 is comparable with that of MutY enzyme on A/GO-20 but is about 3-fold lower than that of MutY on A/G-20 (Table  III). Thus, the catalytic activity of SpMYH is similar to MutY, although the imino intermediate of SpMYH and A/G-containing DNA is less easily trapped than MutY and the same substrate. The weak trapping activity of SpMYH with A/Gcontaining DNA is partly attributable to its weakened binding affinity and catalytic activity in the presence of salt. In contrast to MutY, addition of NaCl higher than 80 mM abol-  ished both SpMYH glycosylase and binding activities. We expect sodium borohydride will have the same effect on Sp-MYH activity.
The MutY and endonuclease III families are highly homologous ( Fig. 1) (31,46,47). Lysine 120 of E. coli endonuclease III at the HhH motif is conserved in the endo III family and has been suggested to be necessary for the formation of the enzymesubstrate intermediate (35). Lys-249 of human OGG1 located at the same position as Lys-120 of endo III has been shown to be the active amine for Schiff's base formation (48). However, MutY and SpMYH have a serine and tyrosine residues at this position, respectively. In fact, MutY was grouped initially as one of the monofunctional glycosylases based on its lack of this conserved lysine (43,49). It will be interesting to see which amino acid of SpMYH is involved in the nucleophilic attack upon the C-1Ј carbon of the sugar of adenine.
The role of functional groups in catalysis by SpMYH and MutY proteins were elucidated and compared by cleavage assay using defined oligonucleotides containing various purine derivatives. Both enzymes cleave A/G, N/G, and A/GO in similar order and have very weak or no cleavage on I/G, A/I, C/GO, and C/G. The C6-amino group of a mismatched A is not critical for cleavage since duplex DNA containing N/G is cleaved efficiently by SpMYH. Because the C6-amino group of adenine is involved in hydrogen bonding with guanine in all three A/G conformations (50,51), N/G pairing would be expected to be less stable than the A/G base pairing. Thus, the reactivity of SpMYH protein, like MutY, cannot be rationalized on the basis of the stability of the mismatched pair alone. The presence of a C6-keto group as in the I/G pair blocks catalysis of both enzymes. The C2-amino group of mismatched G, presented in the major groove as in G(syn)-A(anti), is critical for SpMYH recognition since A/I mismatch is poorly cleaved by the enzyme. The major difference in substrate specificity between SpMYH and MutY is their activities on 2AP/G, A/2AP, and A/C. The introduction of a C2-amino group, presented in the minor groove as in G(syn)-A(anti), slightly reduces SpMYH catalysis (compare 2AP/G with N/G); in contrast, it abolishes MutY glycosylase activity. Moreover, the C6-keto group of mismatched G, presented in the major groove as in G(syn)-A(anti), is less critical for SpMYH reactivity than MutY by comparing the activities of both enzymes with A/G and A/2AP mismatches. The N-1 of adenine (located in the minor groove) is protonated in the A/C mispair (52). The difference of cleavage on A/C mismatch by SpMYH and MutY may be related to the A/C structure. Thus, SpMYH and MutY may interact differently with DNA in both major and minor grooves.
SpMYH and MutY also have different substrate binding affinities. SpMYH has higher nonspecific binding to 44-mer homoduplex than MutY. The exact nature of this difference is not clear. The large size and extra amino acid sequences of SpMYH may contribute to the substrate affinity. The binding affinity of SpMYH for A/G-containing DNA is only 4 -7-fold higher than that for homoduplexes, suggesting that this protein recognizes A/G-mismatched sites mainly through nonspecific binding. The binding of SpMYH to A/GO-containing 20mer DNA is as great as that of MutY. The affinity of SpMYH protein for A/GO-containing DNA is 710-and 3,000-fold higher than that for homoduplex 20-mer and 44-mer DNA, respectively, yet the cleavage activity of SpMYH on A/GO mismatches is about the same as compared with the A/G mispairs. The presence of a C8-keto group in A/GO pair changes the nature of hydrogen bonding between N7 of GO and N1 of A. The N1 of A is protonated in the G(syn)-A(anti) pair (53) but is not protonated in GO(syn)-A(anti) pair (54). These structural differences between A/GO and A/G are located in the minor groove (54). The relative binding efficiency of SpMYH for the A/GO-containing DNA substrate may have biological significance. This may prevent 8-oxoG glycosylase (OGG) from acting upon the GO lesions paired with adenines, so information on both DNA strands would not get lost.
In E. coli, MutY, MutM, and MutT are involved in defending against the mutagenic effects of GO lesions (9 -11), and similar mechanisms to protect cells from the deleterious effects of GO are present in human cells. hMYH is active in removing adenines from A/G-and A/GO-containing DNA (26,27). An activity similar to MutM has been detected in human and rodent cells (55,56), and a human gene similar to yeast OGG1 which encodes 8-oxoG glycosylase has been identified recently by several groups (37)(38)(39)(40). Human cells have an enzyme similar to E. coli MutT, which hydrolyzes the nucleotide precursor 8-oxo-dGTP to 8-oxo-dGMP (57). The human 8-oxo-dGTPase has some sequence homology to the E. coli MutT protein (58). Therefore, the DNA repair pathways that protect the cells from the mutagenic effects of 8-oxoG are apparently highly conserved among diverse organisms. Although OGG1 and OGG2 with 8-oxoG glycosylase (MutM-like) activity have been identified in Saccharomyces cerevisiae (36), we have yet to identify the corresponding MutY-like protein sequences through a search of the entire genome of S. cerevisiae (one endo III homolog is found in accession no. P31378). There are two open reading frames (accession nos. P53164 and Q01976) in the genome of S. cerevisiae and one open reading frame (SPAC19A8.12) in the genome of S. pombe that contain a core MutT domain but their functions have not been confirmed. Hence, with the MutY-like activity of SpMYH described here, it would be interesting to see if S. pombe possesses a MutT-and MutM-like activity and how the GO lesions are repaired in this organism. The mutation rates and mutation specificities in the mutants defective in these gene functions should give important answers in this regard.