3206-3215 Nucleic Acids Research, 2000, Vol. 28, No. 17
`
`' 2000 Oxford University Press
`
`Purification and characterization of a mammalian
`homolog of Escherichia coil MutY mismatch repair
`protein from calf liver mitochondria
`Antony Parker, Yesong Gu and A-Lien Lu *
`
`Department of Biochemistry and Molecular Biology, University of Maryland at Baltimore, 108 N. Greene Street,
`Baltimore, MD 21201, USA
`
`Received June 15, 2000; Revised and Accepted July 18, 2000
`
`ABSTRACT
`A protein homologous to the Escherichia coil MutY
`glycosylase, referred to as mtMYH, has been purified
`from calf liver mitochondria. SDS-polyacrylamide
`gel electrophoresis, western blot analysis as well as
`gel filtration chromatography predicted the molecular
`mass of the purified calf mtMYH to be 35-40 kDa. Gel
`mobility shift analysis showed that the purified
`mtMYH formed specific binding complexes with N8-
`oxoG, G/8-oxoG and T/8-oxoG, weakly with C/8-oxoG,
`but not with NG and NC mismatches. The purified
`mtMYH exhibited DNA glycosylase activity removing
`adenine mispaired with G, C or 8-oxoG and weakly
`removing guanine mispaired with 8-oxoG. The
`mtMYH glycosylase activity was insensitive to high
`concentrations of NaCl and EDTA. The purified
`mtMYH cross-reacted with antibodies against both
`intact MutY and a peptide of human MutY homolog
`(hMYH). DNA glycosylase activity of mtMYH was
`inhibited by anti-MutY antibodies but not by anti-
`hMYH peptide antibodies. Together with the previously
`described mitochondrial MutT homolog (MTH1) and
`8-oxoG glycosylase (OGG1, a functional MutM
`homolog), mtMYH can protect mitochondrial DNA
`from the mutagenic effects of 8-oxoG.
`
`INTRODUCTION
`
`Base excision repair by Escherichia coli MutY glycosylase
`corrects the base-base mismatches A/G and A/C as well as
`adenine and guanine paired with 7,8-dihydro-8-oxo-deoxy-
`guanine (8-oxoG) that arise through DNA replication errors
`and DNA recombination (1-9). Together with MutM and
`MutT, the MutY protein helps to protect the bacteria from the
`mutagenic effects of 8-oxoG (10,11), the most stable product
`known caused by oxidative damage to DNA (12,13). The
`formation of 8-oxoG in DNA, if unrepaired, can lead to the
`mis incorporation of adenine opposite the 8-oxoG lesion
`resulting in a C:G(cid:151)A:T transversion (14-17). The MutT
`protein has nucleoside triphosphatase activity that eliminates
`8-oxo-dGTP from the nucleotide pool (18-20). The MutM
`
`protein (Fpg protein) provides a second level of defense by
`removing both mutagenic 8-oxoG adducts and ring-opened
`purine lesions (21,22). MutM efficiently removes 8-oxoG
`lesions opposite C but very poorly if opposite A. MutY glyco-
`sylase provides a third level of defense by removing the
`adenines or guanines misincorporated opposite 8-oxoG
`following DNA replication.
`Information regarding mammalian MutY proteins is
`emerging. Mammalian MutY homologous (MYH) activities
`have been detected in the nuclear fractions of calf thymus,
`Jurkat and HeLa cells (23-25). The mammalian MYH has
`adenine glycosylase and binding activities on A/8-oxoG and
`A/G mismatches and has recently been shown to possess
`glycosylase activity on 2-hydroxyadenine paired with A, G, T,
`C and 8-oxoG (24). eDNA encoding part of the mouse MutY
`homolog has been cloned (GenBank accession nos A10409068
`and AA409965), although expression and characterization of
`the gene product remains unpublished (26). The gene for a
`human MutY protein (hMYH) has been cloned (27) and the
`predicted size of this hMYH is 59 kDa similar to the size of a
`band detected in HeLa nuclear extracts with an anti-MutY anti-
`body (25). Recently the hMYH protein from the cloned eDNA
`in vitro transcription/translation
`has been expressed in an
`system (28) and in E.coli (26,29) and partially characterized.
`This expressed recombinant hMYH has adenine glycosylase
`activity on the A/8-oxoG mismatch but very weak activity on
`the A/G mismatch. Human cells have also been shown to
`possess MutT (hMTH1) and MutM homologs (hOGGI) (30-33).
`These three enzymes (hMYH, hMTHI and hOGGI) are
`proposed to function in the reduction of 8-oxoG in the human
`genome.
`In the mitochondria, 8-oxoG is one of the most abundant
`lesions formed by exposure to reactive oxygen species (ROS),
`generated as by-products of cellular respiration (13). The accu-
`mulation of oxidative lesions and alterations in mitochondrial
`DNA (mtDNA) has been implicated in aging and several
`human diseases such as carcinogenesis, Parkinson’s disease
`and Alzheimer’s disease (34-36). Because the oxidative environ-
`ment of this organelle creates unfavorable conditions for DNA
`stability and, unlike nuclear DNA, the mitochondrial genome
`is not protected by histone proteins, it is reasonable to assume
`that the mitochondria possess some effective means of
`repairing DNA damage frequently generated in their genome.
`
`*To whom correspondence should he addressed. Tel: +i 410 706 4356; Fax: +1 410 706 1787; Email: aluchang@umaryland.edu
`
`
`
`GDX 1035
`
`(cid:9)
`

`

`Studies have indicated that the mitochondria contain base
`excision repair pathways responsible for the removal of oxida-
`tively damaged DNA lesions. It has been shown that DNA
`lesions caused by oxidative damage, in particular 8-oxoG,
`induced in Chinese hamster ovary cells are rapidly removed from
`the mitochondrial genome suggesting the presence of a 8-oxofl
`glycosylase/AP lyase (OGGI) (37). Croteau et al. partially
`purified a 25-30 kDa base excision endonuclease that prefer-
`entially cleaved C/8-oxoG mismatches but not A/G or A/8-
`oxoG (38). These OGG1 or MutM-like activities are consistent
`with several processed forms ofOGGI enzyme being localized
`to the mitochondrion from a single gene (39). In addition,
`hMTH1, which catalyzes the removal of 8-oxoGTP from the
`nucleotide pool, and endonuclease ITT-like (hNth) activities,
`which remove thymine glycol and fragmented pyrimidines,
`have been shown to be present in the mitochondria (39,40).
`Takao ci al. have shown that there are two types of human
`MYH protein: a mitochondrial form (Type 1) and a nuclear
`isoform (Type 2) (28). Type 1 hMYH, when transiently
`expressed, can be transported into the mitochondria and is a
`DNA glycosylase (28,39). Ohtsubo ci al. also showed that a
`57 kDa hMYH is localized to the mitochondria (24). However,
`Tsai-Wu ci al. showed that Type I hMYH is localized to the
`nucleus excluding the nucleoli (29). Recently, it was also
`shown that three different forms of hMYH exist in the nucleus
`with masses of 52, 53 and 55 kDa, respectively (24). These
`controversial results remained to be resolved. In this study we
`demonstrate biochemically, for the first time, that calf mito-
`chondria contains MYH-like DNA glycosylase activity. The
`purified native calf mitochondrial MYH has a molecular mass of
`35-40 kDa and cross-reacts with both anti-MutY and anti-
`hMYH antibodies. These findings suggest that this protein is a
`mammalian mitochondrial homolog of the E.coli MutY
`protein, which we have named mtMYH.
`
`MATERIALS AND METHODS
`
`Preparation of mitochondria
`Mitochondria were purified from calf liver using a combination of
`differential and Ficoll gradient centrifugation (41,42). All
`procedures were carried out at 4(cid:176)C. Briefly, fresh calf livers
`(6 kg) were chopped into pieces, rinsed in 154 mM NaCl and
`homogenized in a Warring blender with 2 1 of IM buffer
`(225 mM mannitol, 2 mM HEPES, pH 7.4, 75 mM sucrose,
`0.1 mM EDTA and 0.5 mM phenylmethylsulfonyl fluoride).
`After centrifugation at 750 g for 30 mm, supernatants were
`pooled, filtered through a gauze and centrifuged for a further
`30 min at 7500 g to pellet the mitochondria. The mitochondria
`were resuspended in 50 ml IM buffer, centrifuged and washed
`twice. The mitochondrial pellets were resuspended in 30 ml
`buffer A (20 mM potassium phosphate, pH 7.4, 2 mM dithio-
`threitol, 0.1 mM EDTA and 0.1 mM phenyimethylsulfonyl
`fluoride) containing 50 mM KCI with 5 ml aliquots loaded into
`ultracentrifuge tubes that contained 20 ml 30% (w/v) Ficoll in
`225 mM mannitol, 1 mM EDTA, 25 mM HEPES, pH 7.4,
`0.1% (w/v) bovine serum albumin and spun for 30 min at
`95 000 g in a Beckman L8-80M Ultracentrifuge using a
`Beckman 50.2Ti rotor. Mitochondria were collected from the
`bottom of the dense yellow/brown band, washed in IM buffer
`
`Nucleic Acids Research, 2000, Vol. 28, No. 17 3207
`
`by uitracentrifugation and resuspended in seven 50 ml aliquots
`in IM buffer and stored at (cid:151)80(cid:176)C until lysis.
`
`Proteinase K treatment of mitochondria
`Purified intact mitochondria were treated with Proteinase K
`(10 mg/ml) for 20 min at 0(cid:176)C to remove any contaminating
`enzymes from the cytoplasm or nuclei. A control was run
`concurrently without Proteinase K. The reactions were stopped
`by the addition of 0.1 mM phenylmethylsulfonyl fluoride and
`2 Ig/ml Pefabloc. The mitochondria were then centrifuged at
`7500 g for 20 min at 4(cid:176)C, washed twice in TM buffer and lysed
`as described in the purification procedure.
`
`Fractionation of mitochondria
`Fractionation of calf liver mitochondria was performed as
`previously described (43). Briefly, 10 ml of mitochondria
`(-30 mg/ml protein) was diluted to 25 ml with buffer A, to
`which was added 25 ml of 1.2% digitonin and stilTed gently at
`0(cid:176)C. After 15 mm, 30 ml of IM buffer was added and the
`solution centrifuged at 10 000 g for 10 mm. The supernatant
`(S 1) was stored on ice and the pellet was resuspended in 20 ml
`TM buffer and centrifuged for 10 mm, The supernatant was
`added to the SI supernatant fraction to form the ’Outer
`Membrane fraction’. The pellet was resuspended in Lubrol
`WX (ICN, Costa Mesa, CA) and left to stand at 0(cid:176)C. After
`15 mm, the solution was diluted 2-fold with TM buffer and
`ultracentrifuged at 144 000 g for I h. The pellet was resus-
`pended in 20-30 ml IM buffer and represented the ’Inner
`Membrane fraction’ while the supernatant represented the
`’Soluble Matrix fraction’.
`
`Organelle marker enzyme assays
`Maiate dehydrogenase activity (a marker enzyme in the mito-
`chondrial matrix) was measured by monitoring the oxidation
`of NADH at 340 nm during the reduction of oxaloacetate to
`malate as previously described (40). Reactions (I ml)
`contained 89 mM potassium phosphate, pH 7.4, 250 IIM
`NADH, 20 ltM oxaloacetate and the protein sample.
`Monoamine oxidase activity (a marker enzyme in the outer
`mitochondrial membrane) was determined by measuring the
`increase in absorbance at 314 nm with the oxidation of
`kynuramine to 4-hydroxyquinoline (44). The reaction buffer
`contained 50 mM sodium phosphate, pH 7.2,0.2% Triton X-100,
`1 mM kynuramine and the protein sample in a I ml volume.
`Lactate dehydrogenase activity (a marker enzyme for the cyto-
`plasm) was determined by measuring the decrease at 340 nm
`due to the oxidation of NADH (45). The reaction buffer
`contained 50 mM HEPES, pH 7.5, 8 mM sodium pyruvate,
`0.2 mM NADH and the protein sample in a I ml volume.
`Sodium pyruvate was added last to minimize non-specific
`NADH oxidation.
`
`Purification of calf mtMYH from mitochondria
`Purified mitochondria (100 ml) were diluted 3-fold with
`0.5 mM
`buffer (50mM Tris(cid:151)HC1, pH 7.6, 0.1 mM EDTA,
`dithiothreitol, 0.1 mM phenylmethylsulfonyl fluoride and
`500 mM KC1) containing 2 j.tg/mi each of the following
`protease inhibitors: aprotinin, pepstatin A, chymostatin A and
`leupeptin. For lysis, 10% Triton X-100 was added to a final
`concentration of 1% and the resulting solution sonicated for
`10 s using a Fischer Scientific 550 Sonic Dismembrator. The
`
`GDX 1035
`
`

`

`3208 Nucleic Acids Research, 2000, Vol. 28, No. 17
`
`solution was clarified by centrifugation at 100 000 g for I h
`(Fraction I).
`An aliquot of 25% streptomycin sulfate (Sigma Chemical Co.,
`St Louis, MO), in buffer T was added slowly to the clarified
`mitochondria solution to a final concentration of 5% and then
`centrifuged for 30 min at 12 000 g. To the pooled supernatant
`(Fraction I) was added ammonium sulfate to 40% saturation
`and then centrifuged for 30 min at 12 000 g. Again the super-
`natant was pooled, ammonium sulfate was added to 50%
`saturation and then centrifuged for 30 min at 12 000 g. The
`pellet was resuspended in 65 ml of buffer A containing 50 mM
`KC1 and dialyzed extensively versus the same buffer. This was
`designated Fraction II (85 ml). Fraction II was diluted to
`250 ml with buffer A containing 50 mM KC1 and loaded onto
`a 40 ml phosphocellulose PC- Il (Whatman Laboratories,
`Kent, UK) column equilibrated with buffer A containing
`50 mM KC1. After the column was washed with three volumes
`of equilibration buffer, proteins were eluted using a 500 ml
`linear gradient of KCI (0.05-1.0 M) in buffer A. Fractions
`eluting at 250 mM KCI and exhibiting A!8-oxoG binding
`activity were pooled (Fraction III, 20 ml) and dialyzed against
`buffer B (10 mM potassium phosphate, pH 7.4, 50 mM KCI,
`2 mM dithiothreitol, 0.1 mM EDTA and 0.1 mM phenyl-
`methylsulfonyl fluoride). Fraction Ill was applied to 13 ml
`single-stranded DNA (ssDNA) cellulose (Sigma Chemical Co.)
`column equilibrated in buffer B. After the column was washed
`with three volumes of buffer B, bound proteins were eluted
`using a 40 ml linear gradient of KCI (0.05-1.0 M) in buffer B.
`Fractions eluting at 150 mM KCI and exhibiting A!8-oxofl
`binding activity were pooled (Fraction IV, 2.5 ml), concen-
`trated in a centricon-10 to 0.5 ml, washed five times with
`buffer P (10 mM HEPES, pH 7.4, 50 mM KC1, 2 mM dithio-
`threitol, 0.1 mM EDTA and 0.1 mM phenylmethylsulfonyl
`fluoride) in the centricon-lO and applied to a Q-sepharose column
`(0.5 ml) (Amersham Pharmacia Biotech, Piscataway, NJ)
`equilibrated in buffer P. The column was washed with three
`volumes of buffer P and bound proteins eluted with buffer P
`containing 1 M KCI. Fractions with A/8-oxoG binding activity
`eluting in buffer P with 1 M KC1 were pooled (Fraction V,
`0.5 ml), concentrated to 0.15 ml using a centricon-10, and
`applied to a 24.5 ml calibrated Superose 12 (Pharmacia
`Biotech) gel filtration column equilibrated with buffer A
`containing 200 mM KCI. Fractions containing A/8-oxoG
`binding activity were pooled, concentrated (Fraction VI,
`0.3 ml), using a centricon-10, divided into small aliquots and
`stored at -80(cid:176)C.
`
`Preparation of antibodies against MutY and hMYH
`Antibodies against full-length E.coli MutY and an E.coli MutY
`peptide (residues 192-211) were raised in rabbits and prepared
`essentially as described previously (23). Antibodies against the
`hMYH peptides, -344 (against residues 344-361, amino acid
`sequence FPRKASRKPPREESSATC) and (x-516 (against
`residues 516-534 with cysteine at the N-terminus, amino acid
`sequence CDNFFRSHISTDAHSLNSAA) were raised in
`rabbits, For purification of the peptide antibodies, CNBr
`sepharose matrices (Amersham Pharmacia Biotech) were acti-
`vated with 1 mM HCl, washed with 1 mM HCI and 0.125 M
`phosphate coupling buffer and then coupled with the synthetic
`peptide at 4(cid:176)C overnight. The coupled matrices were washed
`with coupling buffer and incubated with I M ethanolamine, pH
`
`8.0, blocking buffer for 5 h at 4(cid:176)C. The matrices were washed
`with phosphate buffered saline (PBS) and packed into a 10 ml
`column. An equal volume of PBS was added to 6 ml of
`antisera, filtered and loaded onto the column with a Waters 650
`FPLC system at 4(cid:176)C. After washing with 40 ml of PBS, the
`antibodies against the hMYH peptides were eluted with elution
`buffer containing 63 mM glycine, pH 2.3 and 0.2 M NaCl.
`Samples (0.75 ml) were collected into tubes containing 0.25 ml
`of neutralizing buffer (0.5 M potassium phosphate, pH 7.5).
`Antibody titer was performed by ELISA.
`
`Western blot (immunoblot) analysis
`Protein fractions were resolved on a 10% SDS-polyacrylamide
`gel (46) and transferred to a nitrocellulose membrane (47). The
`membrane was subjected to the Enhanced Chemiluminescence
`analysis system (Amersham Pharmacia Biotech) according to
`the manufacturer’s protocol, except that the blocking buffer
`contained 10% non-fat dry milk (Carnation) and the wash
`solution contained 1% non-fat dry milk.
`
`DNA substrates
`Annealed l9mer duplexes were prepared as previously
`described (23):
`5’-CCGAGGAATTZGCCTTCTG -3’
`3’- GCTCCTTAAXCGGAAGACG- 5’
`where Z = A, 2-aminopurine (2-AP), C, G or T and X = G, C or
`8-oxoG.
`1 9mer duplexes were radiolabeled at the 3’-end of the upper
`strand with Klenow fragment of DNA polymerase I for 30 mm
`32P]dCTP (50 pCi at 3000 Ci/mmol)
`at 25(cid:176)C in the presence of [(x-
`and 20 pM of dGTP (48). The resulting blunt-end DNA
`duplexes were 20 bp long. The reaction mixture was passed
`through a G25 Quick-Spin column (Boehringer Mannheim)
`and the purified duplexes stored at -20(cid:176)C,
`
`DNA glycosylase assay
`The DNA cleavage activity (DNA glycosylase activity
`followed by heating) of E.coli MutY glycosylase was assayed
`as previously described (49). The glycosylase activity for
`mtMYH was assayed using similar conditions except that
`different buffer and incubation times were used. Protein
`samples were incubated with 1.8 fmol 3’-end labeled 20mer
`duplex DNA in a 10 p1 reaction mixture containing 10 mM
`Tris-HCI, pH 7.6,0.5 mM dithiothreitol, 0.5 mlvi EDTA and 15%
`(v/v) glycerol. After either a 30 or 60 min incubation at 37(cid:176)C,
`the reaction products were lyophilized and dissolved in 3 p1 of
`loading dye containing 90% (v/v) formamide, 10 mM EDTA,
`0.1% (wlv) xylene cyanol and 0.1% (wlv) bromophenol blue.
`The DNA samples were heated at 90(cid:176)C for 2 min and analyzed
`on 14% poly acrylamide-7 M urea sequencing gels (50). The
`gel was then autoradiographed. For some experiments, some
`samples were not heated at 90(cid:176)C for 2 min before loading the
`gel and some samples were supplemented with I M piperidine
`and heated at 90(cid:176)C for 30 min after mtMYH incubation.
`
`Gel mobility shift analyses
`The formation of protein-DNA complexes was analyzed by
`gel retardation assay on 8% polyacrylamide gels in 50 mM
`Tris-borate, pH 8.3 and 1 mM EDTA as previously described
`(49). Protein samples were incubated with 1.8 fmol 3’-end
`labeled 20mer duplex DNA in a 20 p1 reaction mixture
`
`GDX 1035
`
`

`

`containing the same buffer used in the glycosylase assay
`except that 20 ng of poly(dI-dC), 20 mM NaCl and 25 mM
`EDTA were added. Following either a 30 or 60 min incubation
`at 37(cid:176)C, 1 .5 f0 of 50% glycerol was added and the reaction
`products loaded directly onto the gel.
`
`Determination of DNA cleavage in the mtMYH-bound
`DNA fractions
`
`Calf mtMYH (fraction IV, 210 ftg) was incubated in the
`standard gel mobility assay for 60 min at 37(cid:176)C except that
`30 fmol 3’-end labeled 20mer duplex DNA and 200 ng of
`poly(dI-dC) were used. After fractionation on an 8% poly-
`acrylamide gel, the protein-bound and protein-free DNA were
`excised from the gel and eluted by electrophoresis into dialysis
`tubing containing 400 1.0 TBE buffer (50 mM Tris(cid:151)borate,
`pH 8.3, 1 mM EDTA) and 12 ftg/rnl tRNA. The DNA samples
`were twice extracted with phenol, precipitated with ethanol
`and fractionated on a 14% polyacrylamide-7 M urea
`sequencing gel. The gel was then autoradiographed.
`
`Trapping assay
`Covalent trapping of mtMYH with A/8-oxoG in the presence
`of sodium borohydride was performed using the method of Lu
`ci at. (5 1). Protein samples were incubated with 1.8 fmol 3’-end
`labeled 20mer duplex DNA in a 20 ftl reaction mixture
`containing the same buffer used in the glycosylase assay and
`different concentrations of NaBH 4 . A NaBH4 stock solution
`(1 M) was freshly prepared and added immediately after the
`enzyme was added. After incubation at 37(cid:176)C for 60 mm, the
`products were separated on a 12% polyacrylamide gel in the
`presence of SDS (SDS(cid:151)PAGE), and the gel was dried and
`autoradiographed.
`
`Other methods
`Protein concentrations were determined using the method of
`Bradford (52). Phosphoimager quantitative analyses of gel
`images were performed as mentioned previously (2).
`
`RESULTS
`
`A18-oxoG mismatch binding activity is located in the
`mitochondrial matrix
`To demonstrate that mammalian mitochondria contain a MYH
`DNA glycosylase, the binding activity of the A/8-oxoG
`mismatch was assayed with the mitochondrial extract. As
`shown in Figure 1 (lane I), the lysate from Ficoll-purified
`mitochondria contained an A/8-oxoG binding activity. To
`assess the purity of the purified mitochondria, the level of
`cytosolic contamination was estimated by measuring lactate
`dehydrogenase activity (a cytosolic enzyme). The lactate
`dehydrogenase activities were 11.7 and 0.25 ftmol/min/mg for
`crude liver homogenate and for the Ficoll-purified mito-
`chondrial lysate (Fraction I), respectively. Therefore, a 48-fold
`enrichment of mitochondria was achieved. To confirm that the
`binding activity of the A/8-oxoG oligonucleotide was in fact
`located within the mitochondria, Ficoll-purified intact mito-
`chondria were treated with proteinase K for 20 min at 0(cid:176)C
`before lysis. When compared with Ficoll-purified mitochondria
`that had not been exposed to proteinase K, >90% of A/8oxo-G
`
`Nucleic Acids Research, 2000, Vol. 28, No. 17
`
`3209
`
`1)
`2 (cid:9)
`
`K (cid:9)
`3 (cid:9)
`
`0 (cid:9)
`4 (cid:9)
`
`t (cid:9)
`5 (cid:9)
`
`N1
`6
`
`F
`
`Figure 1. Calf mtMYH is localized in the mitochondrial matrix. Lane 2
`contains the A/8-oxoG mismatch oligonucleotide only (D). One hundred
`micrograms of protein from the non-Proteinase K treated (lane 1, F) and
`Proteinase K treated Ficoll-purified mitochondria lysale (lane 3, K) were
`assayed for binding with A/8-oxoG-containing oligonucleotide as described in
`Materials and Methods. The mitochondrial lysate was fractionated using
`digitonin as described and each fraction was assayed for A/8-oxoG mismatch
`binding activity. Lanes 4-6 contain reactions which contained IOU pg of the
`outer membrane and inter-membrane protein fraction (lane 4, 0), 100 pg of
`the inner membrane fraction (lane 5,1) or 100 pg of the matrix protein fraction
`(lane 6, M). Arrows indicate the positions of free DNA (F) and enzyme-bound
`DNA complex (C). Quantitation of the gel is shown in Table I
`
`binding activity remained after proteinase K treatment (Fig. I,
`lane 3).
`To establish the intra-mitochondrial location of this A/8-oxoG
`binding activity, the mitochondria were incubated with digitonin
`as previously described (43) to produce sub-mitochondrial
`fractions: the matrix, inner membrane and outer membrane
`with inter-membrane space. Monoamine oxidase (44) and
`malate dehydrogenase (40) were assayed to ascertain the
`efficiency of this fractionation (Table 1). Monoamine oxidase
`activity was located in the outer membrane fraction whereas
`the majority of the malate dehydrogenase activity was detected
`in the matrix fraction (Table 1). A similar overflow of marker
`enzyme activities were found present in the sub-mitochondrial
`preparations of others (38) and may be due in part to the pres-
`ence of other enzymes utilizing the substrates. Nevertheless,
`the location of the marker enzymes showed that the fractiona-
`tion was effective. When the mitochondrial fractions were
`assayed for A18-oxoG binding, all the activity was detected in
`the mitochondrial matrix fraction (Table 1; Fig. 1, lane 6).
`
`Purification of calf mtMYH
`The A/8-oxoG binding activity was purified (cid:151)475-fold from
`the isolated mitochondria using a combination of ammonium
`sulfate fractionation and four column chromatography steps
`(Table 2). A narrow ammonium sulfate cut was employed in
`the purification in order to remove other non-specific DNA
`binding proteins. In the final purification step, by Superose 12
`gel filtration, a single protein of (cid:151)38 kDa (Fig. 2A) co-purified
`with the A/8-oxoO binding activity (Fig. 2B). There was good
`correlation between the intensity of the protein and the A/8-
`oxoG binding activity (fractions 54-56 contained the 38 kDa
`band and activity). The molecular mass of the native A18-oxoG
`binding activity protein was determined using a calibrated
`Superose 12 gel filtration column. A/8-oxoG binding activity
`protein eluted at a position of 35-40 kDa by comparison with
`
`GDX 1035
`
`

`

`3210 Nucleic Acids Research, 2000, Vol. 28, No. 17
`
`Table 1. Localization of total A/8-oxoG binding activity within calf liver mitochondria
`
`Fraction (cid:9)
`
`Marker enzymes (cid:9)
`
`Outer membrane and inter-membrane (cid:9)
`
`Inner membrane (cid:9)
`
`Matrix (cid:9)
`
`Malate dehydrogenase (units)’ (cid:9)
`174 – 45 (cid:9)
`
`428 – 9 (cid:9)
`
`1759 – 63 (cid:9)
`
`Monoamine oxidase A (units) 5
`
`131 – 8 (cid:9)
`
`0.9 – 0.1 (cid:9)
`
`03 – 0.1 (cid:9)
`
`mtMYH (units)’
`
`0
`
`0.2 – 0.1
`
`75 300 – 5400
`
`The total activities shown in this table were from 10 ml of Ficoll-purified mitochondria from 0.17 kg of calf thymus and each value was obtained from three
`experiments.
`’One unit of activity is defined as the production of I imol NAD’ per minute.
`bO ne unit of activity is defined as the production of I fLinol 4-hydroxyquinolone per minute.
`’One unit of binding activity is defined as that resulting in binding of 1% (0.018 fmol) of 3’-labeled DNA 20mer with an A/8-oxoG mismatch at 37(cid:176)C for 30 mm.
`’All values are presented as mean values – the standard deviation.
`
`Table 2. Purification of mtMYH from calf liver mitochondria
`
`Fraction (cid:9)
`
`Volume (ml) (cid:9)
`
`Protein (mg)’ (cid:9)
`
`Specific activity (cid:9)
`
`Fold purification
`
`Nicking (units/mg) 5 (cid:9)
`
`Binding (units/mg)’ (cid:9)
`
`(A18-oxoG) binding
`
`(1) Mitochondrial lysate
`
`(II) Ammonium sulphate
`
`(III) Phosphocellulose
`
`(IV) ssDNA cellulose
`
`(V) Q-Sepharose
`
`(VI) Superose-12
`
`400
`
`85
`
`20
`
`2.5
`
`0.5
`
`0.3
`
`2700
`
`1040
`
`40
`
`3.1
`
`0.4
`
`01
`
`10
`
`55
`
`1100
`
`2450
`
`16250
`
`45000
`
`200
`
`900
`
`2800
`
`4800
`
`18400
`
`92700
`
`1
`
`5
`
`15
`
`25
`
`94
`
`475
`
`’Protein concentration was measured by Bradford assay.
`5One unit of nicking activity is defined as that resulting in cleavage of 1% (0.018 fmol) of 3’-labeled DNA 20mer with an A/8-oxoG mismatch at 37(cid:176)C for 30 mm.
`’One Unit of binding activity is defined as that resulting in binding of 1% (0.018 fmol) of 3’-labeled DNA 20mer with an A/3-oxoG mismatch at 37(cid:176)C for 30 mm.
`
`molecular mass markers (Fig. 2C) consistent with the single
`band of (cid:151)38 kDa shown by SDS(cid:151)PAGE (Fig. 2A). Thus, this
`protein appears to be a monomer. To verify that the purified A/8-
`oxoG binding activity protein is similar to E.coli MutY, the
`adenine glycosylase activity on A18-oxoG(cid:151)containing DNA
`was determined in the fractions from the Superose 12 column.
`As shown in Figure 2D, the adenine glycosylase activity co-
`purified with the A/8-oxoG binding activity (Fig. 2B) and the
`38 kDa protein (Fig. 2A). The adenine glycosylase activity
`could also be detected in the earlier fractions of the purification
`although it was much lower in fractions I and II (Table 2).
`Thus, we termed the 38 kDa protein with A/8-oxoG binding
`and glycosylase activities ’mtMYH’.
`
`Specificities of DNA binding and glycosylase activities of
`calf mtMYH
`The binding and glycosylase activities of calf mtMYH and
`E.coli MutY with several mismatches were compared. As
`shown in Figure 3A, in gel retardation assays, mtMYH bound
`efficiently to A18-oxoG-, G/8-oxoG- and T/8-oxoG-containing
`DNA (lanes 1, 7 and 8) and much more weakly to C/8-oxoG-
`containing DNA (lane 6). However, unlike MutY, A/U, A/C
`and 2AP/G mismatches were not bound by mtMYH. Because
`mtMYH bound all four 8-oxoG mismatches at a fixed concen-
`tration we sought to determine its relative affinity for these
`
`substrates. Figure 3B showed that substrate binding increased
`with increased mtMYH amount. As seen with MutY (2), the
`maximum binding of mtMYH with 8-oxoG-containing
`substrates did not reach 100% saturation. This may be due to
`an inherent problem with the band shift assay or the nature of
`the DNA substrates. At lower concentrations of mtMYH, the
`affinity was the best with A18-oxoG mismatch while at higher
`concentrations of mtMYH, the affinity with A/8-oxoG and UI
`8-oxoG was similar.
`Given that MutY can cleave several mismatches with
`different efficiency (4), we compared mtMYH and MutY glyco-
`sylase activities on different DNA mismatches. As shown in
`Figure 4A, similar to MutY, mtMYH cleaved A18-oxoG and Al
`U efficiently but cleaved 2AP/G weakly (lanes 1, 2, 4, 9, 10
`and 12). However, when compared to MutY, A/C mismatch
`was a better substrate and G/8-oxoG was a poorer substrate for
`mtMYH (Fig. 4A, lanes 3, 7, Il and 15). The minor band
`above the major cleaved product was also a product of the
`glycosylase activity and was derived from an impurity of DNA
`substrates that appeared above the intact DNA (Fig. 4A).
`Time course studies to determine the extent of glycosylase
`activity of mtMYH on different DNA substrates over the
`60 min period were shown in Figure 4B. No cleavage was observed
`with C/8-oxoG and T/8-oxoG mismatches even with long time
`incubation. The reactivity of mtMYH on G/8-oxoG-containing
`
`GDX 1035
`
`

`

`Nucleic Acids Research, 2000, Vol. 28, No. 17 3211
`
`52535355 5657 (cid:9)
`
`(B) (cid:9)
`
`52 53 54 55 56 57
`
`0)kl)a (cid:9)
`175 -*
`83 -*
`
`37 .
`
`Figure 2. Purification of calf mtMYH by Superose 12 column, Calf mtMYH (fraction V) was purified as described in Materials and Methods and loaded Onto S
`24.5 ml calibrated Superose 12 gel filtration column and 275 (LI fractions were collected, (A) SDS(cid:151)polyacrylamide analysis. Superose 12 column fractions were
`concentrated by TCA precipitation, electrophoresed on a 10% SDS.-PAGE and stained with silver. The positions of the molecular weight standards (New England
`Biolabs prestained markers) are marked. (B) A/8-oxoG binding activity of the fractions from Superose 12 column. The positions of free DNA (F) and binding
`complex (C) are marked. (C) The native molecular mass of mtMYH was estimated to be 35-40 kDa. The Superose 12 column was calibrated with blue dextran
`(2000 kDa). BSA (66 kDa), ovalbumin (45 kDa), carbonic anhydrase (29 kDa) and lysozyme (14 kDa). Molecular weight markers (marked by filled circles) were
`monitored by DV absorption and mIMYH (marked by a filled rectangle) was analyzed by A/8oxo-G binding activity. (D) A18-oxoG glycosylase activity of the
`fractions from Superose 12 column. The samples in the loading dye were heated at 90(cid:176)C for 2 min and electrophoresed in a 14% sequencing gel. The positions of
`intact DNA substrate (1) and cleaved products (N) are marked. The numbers above (A), (B) and (D) represent the fraction number eluted from the Superose 12
`column,
`
`DNA is very weak. At short reaction times the rate of cleavage
`of calf mtMYH was much faster for A18-oxoG mismatch than
`other substrates. However, at reaction times longer than 3 mm,
`the reactivity on AIG is greater than on the A18-oxoG
`mismatch.
`To examine whether the binding activity and glycosylase
`activity of mtMYH work simultaneously we excised the
`rntMYH-bound and mtMYH-free DNA bands shown in
`Figure 3A and analyzed them on a sequencing gel (Fig. 5). On
`A18-oxoG and G18-oxoG mismatches, mtMYH did not disso-
`ciate from the DNA substrate after the glycosylase action as
`the cleaved product was only present in the enzyme-bound
`fraction but was not observed in the enzyme-free fraction
`(Fig. 5, lanes 2-5). No cleaved product was present in enzyme-
`bound and enzyme-free fractions with the T18-oxoG substrate
`(Fig. 5, lanes 6 and 7).
`
`Calf mtMYH cross-reacts with antibodies raised against
`E.coli MutY and hMYH but only the E.coli MutY
`antibodies inhibit the DNA glycosylase activity of hMYH
`To test for homology of calf mtMYH with E.coli MutY and
`hMYH, we performed western blotting with polyclonal anti-
`bodies raised against full-length E.coli MutY, an E.coli MutY
`peptide (residues 192-211) and two hMYH peptides (residues
`
`344-361, termed cs-344, and residues 516-534, termed a-5 16).
`Calf mtMYH cross-reacted with both the anti-full-length
`MutY and the anti-MutY peptide antibodies (Fig. 6A, lanes 2
`and 3). Thus, calf rntMYH showed considerable homology to
`E.coli MutY. Calf mtMYH also showed some homology with
`hMYH because it cross-reacted with the (x-516 hMYH peptide
`antibodies (Fig. 5, lane 4) but not cx-344 hMYH peptide anti-
`bodies (data not shown). Both (x-516 and a-344 antibodies
`were raised against peptides representing the C-terminus of
`hMYH.
`Since calf mtMYH cross-reacted with antibodies raised
`against both MutY and hMYH in western blot analysi