Purification and characterization of a novel UV lesion-specific DNA glycosylase/AP lyase from...

Ž .Mutation Research 459 2000 307–316www.elsevier.comrlocaterdnarepair

Community address: www.elsevier.comrlocatermutres

Purification and characterization of a novel UV lesion-specificDNA glycosylaserAP lyase from Bacillus sphaericusq

Debra A. Vasquez a, Simon G. Nyaga b, R. Stephen Lloyd c,) ,1

a School of Medicine, The UniÕersity of Texas Medical Branch at GalÕeston, 301 UniÕersity BlÕd., GalÕeston, TX 77555, USAb Laboratory for Molecular Genetics, National Institute on Aging, National Institutes of Health, 5600 Nathan Shock DriÕe,

Baltimore, MD 21224, USAc Sealy Center for Molecular Science and The Department of Human Biological Chemistry and Genetics, The UniÕersity of Texas Medical

Branch at GalÕeston, Medical Research Building, Room 5.142, 301 UniÕersity BlÕd., GalÕeston, TX 77555-1071, USA

Received 3 December 1999; received in revised form 15 February 2000; accepted 16 February 2000

Abstract

The purification and characterization of a pyrimidine dimer-specific glycosylaserAP lyase from Bacillus sphaericusŽ .Bsp-pdg are reported. Bsp-pdg is highly specific for DNA containing the cis–syn cyclobutane pyrimidine dimer,

Ž .displaying no detectable activity on oligonucleotides with trans–syn I, trans–syn II, 6–4 , or Dewar photoproducts. Like` Xother glycosylaserAP lyases that sequentially cleave the N glycosyl bond of the 5 pyrimidine of a cyclobutane pyrimidine

dimer, and the phosphodiester backbone, this enzyme appears to utilize a primary amine as the attacking nucleophile. Theformation of a covalent enzyme–DNA imino intermediate is evidenced by the ability to trap this protein–DNA complex byreduction with sodium borohydride. Also consistent with its AP lyase activity, Bsp-pdg was shown to incise an APsite-containing oligonucleotide, yielding b- and d-elimination products. N-terminal amino acid sequence analysis of this 26kDa protein revealed little amino acid homology to any previously reported protein. This is the first report of aglycosylaserAP lyase enzyme from Bacillus sphaericus that is specific for cis–syn pyrimidine dimers. q 2000 ElsevierScience B.V. All rights reserved.

Keywords: Pyrimidine dimer; GlycosylaserAP lyase; Bacillus sphaericus

q Supported by ES04091, ES06676.) Corresponding author. Tel.: q1-409-772-2179; fax: q1-409-

772-1790.Ž .E-mail address: [email protected] R.S. Lloyd .

1 R.S.L. holds the Mary Gibbs Jones Distinguished Chair inEnvironmental Toxicology from the Houston Endowment.

1. Introduction

Ž .Ultraviolet light UV damages DNA, resulting inseveral types of potentially cytotoxic and mutagenic

w xDNA photoproducts 1,2 . The two major types ofŽ .lesions induced by UV-B 280–320 nm and UV-C

Ž .200–280 nm irradiation are cyclobutane pyrimidineŽ .dimers and pyrimidine–pyrimidone 6–4 photo-

0921-8777r00r$ - see front matter q 2000 Elsevier Science B.V. All rights reserved.Ž .PII: S0921-8777 00 00009-4

( )D.A. Vasquez et al.rMutation Research 459 2000 307–316308

Ž .products. Unlike the repair of 6–4 photoproducts,which can be initiated by the nucleotide excision

Ž .repair system or a 6–4 photolyase, cyclobutanepyrimidine dimer repair can be initiated by either thebase excision or nucleotide excision repair pathwaysas well as by a dimer photolyase.

The base excision repair pathway is initiated byDNA glycosylases, a class of lesion-specific en-zymes that removes the damaged base. These en-zymes may or may not have concomitant abasicŽ . w w xAP lyase activity reviewed in Ref. 3 . The mostwell characterized of the pyrimidine dimer glycosy-lases is T4 endonuclease V, now referred to as

Ž . wT4-pdg pyrimidine dimer-specific glycosylase re-w xviewed in Refs. 4,5 . T4-pdg was the first DNA

glycosylase to have its structure determined by X-raycrystallography and its active site residues identifiedw x6–12 . Upon recognizing and binding to DNA con-taining a cyclobutane pyrimidine dimer, T4-pdg

` Xcleaves the N1 C1 glycosyl bond of the dimer’s5X-pyrimidine and then cleaves the phosphodiesterbond between the two dimerized pyrimidines. This isaccomplished by formation of a covalent enzyme–DNA imino intermediate, which results in glycosylbond scission, followed by b-elimination, generating

w xphosphodiester bond cleavage 6,10,13 .Over the past several years, a number of other UV

lesion-specific glycosylases have been identified.Some of these enzymes are structurally homologousto T4-pdg, whereas others are related to the Es-cherichia coli endonuclease III family. One of thestructural homologs of T4-pdg has been identifiedfrom a virus, Paramecium bursaria chlorella virus-1Ž .PBCV-1 , which infects the green algae Chlorellaw x14 . The gene encoding the PBCV-1 pyrimidine

Ž .dimer-specific glycosylase cv-pdg predicts an en-zyme that is 41% identical to T4-pdg. Cv-pdg hasbeen shown to be specific for cis–syn and trans–synII cyclobutane pyrimidine dimers and functions as a

w xcombined glycosylaserAP lyase 15–17 .In addition, Micrococcus luteus has been shown

to contain two UV photoproduct-specific glycosy-w xlaserAP lyases 18,19 . One is an 18-kDa protein

w xwith a 27% amino acid identity with T4-pdg 19 .The other is a 31-kDa enzyme that exhibits a similarsubstrate specificity and catalytic mechanism as T4-

w xpdg 18 , but has structural homology to both MutYw x20,21 , a DNA mismatch repair glycosylase, and

w xendonuclease III 22 , a DNA glycosylaserAP lyasethat repairs oxidative damage in DNA. These differ-ences in structure, yet similarity in function, led tothe question of whether conserved structural motifsexist among glycosylases with AP lyase activity.

In order to explore the idea that conserved struc-tural motifs might exist within proteins that catalyzethe initiation of repair at pyrimidine dimers, a collab-

Ž .oration with New England Biolabs Beverly, MAwas begun in which a total of 108 microorganismswere screened for UV-dependent DNA glycosylaseor glycosylaserAP lyase activity. Cell-free extractsfrom various bacterial strains were screened using a

Ž .double-stranded deoxyoligonucleotide 49-mer con-taining a site-specific TpT cyclobutane pyrimidine

Ž .dimer or 6–4 photoproduct. Several strains showedpyrimidine dimer-specific nicking activity. These in-cluded: Bacillus sphaericus, Neisseria mucosa hei-delbergenesis, N. mucosa, N. sicca, Haemophilusgalinarum, H. parainfluenza and Moraxella boÕis.Interestingly, none of the 108 organisms demon-

Ž .strated 6–4 photoproduct-specific nicking activity,Ž .even though the 6–4 photoproduct is one of the

major lesions caused by UV exposure. In this study,we report the purification and biochemical character-ization of the pyrimidine dimer-specific glycosylaseenzyme from B. sphaericus.

2. Materials and methods

2.1. Materials

Cell-free lysates from 108 microorganisms wereŽthe very generous gift of Dr. Richard Roberts New

.England Biolabs . B. sphaericus cells were obtainedŽcourtesy of Dr. Peter Goldfarb New England Bio-

.labs .Deoxyoligonucleotides containing site-specific

Ž .cis–syn, trans–syn I, trans–syn II, 6–4 , and De-war photoproducts of TpT were kindly supplied by

ŽDr. John-Stephen Taylor Washington University, St.. w xLouis, MO 23 . Deoxyoligonucleotides containing

Ž .8-oxoguanine 8-oxoG were synthesized by theNIEHS Molecular Biology Core Facility at UTMB.

Phenyl sepharose CL-4B, heparin sepharose CL-6B, column chromatographic supplies and oligo-nucleotide sizing markers were purchased from Phar-

( )D.A. Vasquez et al.rMutation Research 459 2000 307–316 309

macia. Prestained molecular weight markers wereŽ .from BRL. X-ray films X-Omat were from Kodak.

w 32 x Ž .g- P dATP 6,000 Cirmmol was purchased fromDupont-NEN.

2.2. Purification of B. sphaericus pyrimidine dimer-( )specific glycosylase Bsp-pdg .

All purification procedures were performed at48C. A total of 57 g of frozen B. sphaericus cellswere thawed and suspended in 500 ml of 20 mM

Ž .Tris–HCl pH 7.5 , 1 mM EDTA, 50 mM KCl andŽ .10% ethylene glycol buffer A , disrupted with

Ž .French Press 1100 psi and the cellular debris pel-leted by centrifugation for 30 min at 10,000 rpm in aGSA rotor. The resulting supernatant was applied toa 400 ml single-stranded DNA–agarose column that

Žhad been equilibrated with 25 mM NaH PO pH2 4.6.8 , 1 mM EDTA, 100 mM KCl, 10% ethylene

Ž .glycol buffer B . After the sample had been loadedonto this column and thoroughly washed with bufferB, the bound proteins were eluted with a linear

Ž .gradient 700 ml each of buffer B to buffer BŽ .supplemented with 2.5 M KCl. Fractions 4 ml were

collected and monitored for thymine dimer-specificnicking activity. As in all subsequent purificationsteps and schemes, activity was assayed by reactionwith radiolabeled thymine dimer-containing oligo-

Žnucleotides followed by gel electrophoresis see be-.low . In addition, proteins were analyzed by SDS-

PAGE.For SDS-PAGE analysis, fractions from each of

the columns that showed pyrimidine dimer-specificnicking activity were precipitated with 30% tri-

Ž .chloroacetic acid TCA for 1 h at 48C. The precipi-tated proteins were pelleted by centrifugation for 15min at 48C, and the supernatant discarded. Sampleswere then washed with acetone and air-dried. Load-

Ž Ž .ing buffer 20 ml of 100 mM Tris–HCl pH 6.8 , 20Ž . Ž .mM DTT, 4% wrv SDS, 0.2% wrv bromophe-Ž . .nol blue, 20% vrv glycerol was added and sam-

ples were boiled for 10 min prior to separationthrough a 15% polyacrylamide SDS gel. Followingelectrophoresis, the gel was stained by silver salts forfurther analysis.

Although fractions were examined by SDS-PAGEfollowed by silver staining, fractions were generallypooled by pyrimidine dimer-specific nicking activityrather than by their molecular mass profiles. Frac-

tions from the single-stranded DNA agarose columnŽ .that displayed activity were combined 150 ml and a

Ž .saturated solution of NH SO was added until the4 2 4Ž .final NH SO concentration reached 1 M. This4 2 4

solution was loaded onto a 40 ml phenyl-sepharosecolumn that had been previously equilibrated with 25

Ž .mM NaH PO pH 6.8 , 10 mM EDTA, 1.8 M2 4Ž . Ž .NH SO , 10% ethylene glycol buffer C . After4 2 4

the sample was loaded, the column was thoroughlywashed with 200 ml of buffer C. Proteins wereeluted by applying a 200 ml linear gradient of buffer

Ž . ŽC to buffer C minus the NH SO . Fractions 5.04 2 4.ml were collected and screened for pyrimidine

dimer-specific nicking activity. The majority of theŽ .activity eluted between 500 and 800 mM NH SO .4 2 4

Ž .The fractions with activity were combined 85 mland extensively dialyzed against 25 mM NaH PO2 4Ž .pH 6.8 , 10 mM EDTA, 50 mM KCl and 10%

Ž .ethylene glycol buffer D . This solution was loadedonto a 40 ml heparin sepharose column that had beenequilibrated with buffer D. The column was washedwith 200 ml of buffer D and proteins were eluted byapplying a 100-ml linear gradient of buffer D con-taining 100 mM KCl to buffer D containing 800 mMKCl. Pyrimidine dimer-specific nicking activityeluted between 350 and 500 mM KCl. Analysis ofthese fractions by SDS-PAGE followed by silverstaining revealed that the enzyme had been purifiedto apparent homogeneity. A summary of the purifica-

Ž .tion scheme Purification Scheme a1 can be seen inFig. 1.

Fig. 1. Purification schemes for Bsp-pdg.


A second independent purification was carried outin order to generate sufficient quantities of highlypurified enzyme for biochemical characterization andN-terminal amino acid sequence determination. Thefirst three chromatographic steps, single-strandedDNA agarose, phenyl sepharose and heparinsepharose column chromatography, were essentiallyidentical to the procedure described above. SDS-PAGE followed by silver staining revealed the pres-ence of two proteins in the heparin sepharose frac-tions exhibiting the most pyrimidine dimer-specificnicking activity. In an attempt to identify which ofthese two proteins possessed the observed activity, 4

Ž .ml of one of the active fractions a24 was loadedonto a reversed phase HPLC column that had been

Ž .equilibrated with 0.1% trifluoroacetic acid TFA inwater. Proteins were eluted by addition of increasingconcentrations of acetonitrile in 0.1% TFA. Fractionswere screened for pyrimidine dimer-specific nickingactivity, and activity was found to peak between

Ž .fractions 47 and 49 270 ml . The fractions weresubjected to SDS-PAGE followed by silver staining,which revealed that fractions 47 and 48 contained a

Žsingle band. Fractions 47 and 48 were pooled 175.ml and submitted for N-terminal amino acid se-

quencing.Protein concentrations were determined using the

ŽBCA Protein Assay Reagent System Pierce Chemi-.cal . For assays containing low protein concentra-

tions, the ‘‘enhanced protocol’’ was used accordingto the suppliers’ protocol in which assays were car-ried out at 608C for 30–60 min. Bovine serum

Ž .albumin BSA was used to develop standard curves.

2.3.Preparation of UV photoproduct-, mismatch- andoxidatiÕe damage-containing dsDNAs

Ž .The trans–syn I, trans–syn II, 6–4 , and Dewarphotoproduct-containing deoxyoligonucleotidesshared the following sequences in the damagedstrand, where the bold-faced TT is the modified site:5X-AGCTACCATGCCTGCACGAATTAAGCAAA-TTCGTAATCATGGTCATAGCT-3X. The cis–syndimer-containing strand had a slightly different se-quence in the damaged strand, where the bold-facedTT is the modified site: 5X-AGCTACCATGCCTG-CACGTA TTATGCAATTCGTAATCATGGTCAT-AGCT-3X. Prior to use, the photoproduct-containing

strands were 5X end labeled with 32 P and then madeinto duplex DNA by annealing with an excess ofunlabeled complementary 49-mer.

Ž .The deoxyoligonucleotides 30-mers that wereused to screen for incision at mismatches and sites ofoxidative damage had the following sequence: 5X-TACGAATTCGTTAATTCGTGCAGGCATGGT-3X,where the bold-faced A is the mismatched base thatwas placed opposite guanine or 8-oxoG. To createthe A:G and A:8-oxoG-containing duplexes, thestrand containing G or 8-oxoG was 32 P-labeled priorto annealing to the complementary oligonuceotide, asdescribed above.

2.4. Preparation of AP site-containing DNA

A single-stranded 49-mer deoxyoligonucleotidecontaining a single uracil at position 21, with thesequence 5X-AGCTACCATGCCTGCACGAA-UTAAGCAATTCGTAATCATGGTCATAG-3X, was

Ž .obtained from Midland Reagent. This 49-mer 35 ngX w 32 xwas radiolabeled on the 5 end using g- P dATP in

a 20 ml reaction and then annealed to a threefoldexcess of its complementary strand. Just prior to use,the resulting double-stranded DNA was incubated

Žwith 1 unit of uracil DNA glycosylase UDG; Epi-.centre Technologies at 378C for 30 min in a reaction

Ž .buffer consisting of 25 mM NaH PO pH 6.8 , 12 4

mM EDTA, 25 mM NaCl, and 100 mgrml BSA.

2.5. Assaying purified Bsp-pdg for UV photo-product-, AP site-, mismatch- and oxidatiÕedamage-specific nicking actiÕities

To monitor Bsp-pdg catalytic activity on the UVphotoproduct-, AP site-, mismatch- and oxidative

Ž .damage-containing substrates, aliquots 3.0 ml ofcolumn fractions were incubated at 378C for 30 min

Ž .in 20 ml of 25 mM NaH PO pH 6.8 , 1 mM2 4

EDTA, 100 mM KCl, 100 mgrml BSA with theŽappropriate duplex DNA 0.3 ng of cis–syn dimer-

containing DNA; 1.3 ng of trans–syn I, trans–syn IIor Dewar photoproduct-containing DNA; 1.9 ng ofŽ .6–4 photoproduct-containing DNA; or 0.5 ng of

.the mismatch- or 8-oxoG-containing DNA . Reac-tions were terminated by the addition of an equal

Ž Ž .volume of oligonucleotide loading buffer 95% vrv


Ž .formamide, 20 mM EDTA, 0.02% wrv xyleneŽ . .cyanole, 0.02% wrv bromophenol blue . All sam-

ples, except for those with AP site-containing DNA,were heated at 808C for 5 min prior to gel elec-trophoresis. Samples were separated on 15% denatur-ing polyacrylamide gels and exposed to X-ray filmbetween two DuPont Quanta III intensifying screens.For quantitative data analyses, the wet gels were

Žscanned on a PhosphorImager 450 Molecular Dy-.namics, Sunnyvalle, CA and data analyzed using

Ž .ImageQuant software Molecular Dynamics .

2.6. Sodium borohydride trapping of the coÕalentimino enzyme–DNA intermediate

A duplex 49-mer containing a single cis–synŽ .thymine dimer 32 fmolesrreaction was labeled on

X w 32 xthe 5 end using g- P -dATP and then incubated at378C with increasing concentrations of Bsp-pdg orT4-pdg in the presence of 100 mM NaBH . The4

Ž .reactions 20 ml were performed in siliconized tubesŽ .in 25 mM NaH PO pH 6.8 1 mM EDTA, 1002 4

mgrml BSA. Control reactions in the absence ofNaBH and with DNAs not containing lesions were4

also performed. After 1 h, reactions were heated to908C for 3 min and dried by vacuum centrifugationfor 30 min. The resulting pellets were dissolved in a95% formamide-containing loading buffer. The sam-ples were then separated by electrophoresis througha 15% polyacrylamide gel containing 8 M urea.Electrophoresis was carried out at a constant 20 Wfor 2 h. The wet gels were analyzed by autoradiog-raphy.

3. Results and discussion

3.1. Purification of Bsp-pdg

In an attempt to discover new UV photoproduct-specific DNA repair enzymes, and in collaborationwith New England Biolabs, soluble whole-cell ex-tracts from 108 microorganisms were screened forUV lesion-specific nicking activity using deoxy-oligonucleotides containing a single cis–syn cy-

Ž .clobutane thymine dimer or 6–4 photoproduct ofTpT as substrates. Of the 12 Bacillus species thatwere tested using these assays, only B. sphaericusshowed a dimer-specific nicking activity that was

Žconsistent with a glycosylaserAP lyase data not.shown . These results may suggest that this pyrimi-

dine dimer-specific nicking activity is not evolution-arily conserved among the Bacillus genus and thatthe gene encoding this DNA repair protein couldpotentially be associated with a mobile genetic ele-ment or integrated phage.

Since preliminary dimer-specific nicking assayssuggested that the activity in B. sphaericus may bethe result of a glycosylaserAP lyase, we rationalizedthat the purification scheme used for T4-pdg mightbe effective in the purification of the protein respon-sible for this activity. Thus we used the followingchromatographic steps: single-stranded DNA agarose,phenyl sepharose and heparin sepharose chromatog-raphy. An outline of the purification scheme is shownin Fig. 1, Scheme a1. During purification, fractionswere monitored routinely for their ability to incise a

Table 1Purification of a DNA glycosylaserAP lyase from B. sphaericusOne unit of Bsp-pdg is defined as the amount of enzyme required to incise 1.5 ng of duplex 49-mer oligonucleotide containing a singlethymine dimer in 30 min at 378C.

Ž . Ž .Step Total units U Concentration Volume ml Amount of Specific activityŽ . Ž . Ž .mgrml protein mg Urmg

5 6 6Starting material 8.6=10 0.9 1.9=10 1.71=10 0.5035 5 4Single-stranded DNA 7.0=10 0.10 9.0=10 9.0=10 7.8

agarose5 5Phenyl sepharose 2.18=10 0.07 1.25=10 8750 2494 5Heparin sepharose 8.1=10 0.001 1.35=10 135 600


Fig. 2. Molecular weight analysis of Bsp-pdg. Purified Bsp-pdgwas obtained following heparin sepharose chromatography. Afraction exhibiting cis – syn pyrimidine dimer-specific nickingactivity was separated by electrophoresis through an SDS poly-

Ž .acrylamide gel and then silver stained Lane 4 . Lane 1: prestainedŽ .low molecular weight markers; Lane 2: lysozyme 14.3 kDa ;Ž .Lane 3: catalytically active domain of E. coli MutY 26 kDa .

49-mer deoxyoligonucleotide duplex containing asite-specific cis–syn cyclobutane thymine dimer. Theefficiency of the purification process was evaluatedby the determination of specific activity using this

Ž .substrate Table 1 .

To analyze the purification process further,aliquots of the heparin sepharose fractions were sep-arated by SDS-PAGE and then stained with silversalts. The silver-stained gel of proteins found infractions that demonstrated pyrimidine dimer-specificnicking activity revealed that Bsp-pdg had been

Žpurified to apparent physical homogeneity Fig. 2,.lane 4 . Bsp-pdg has a relative molecular weight of

26 kDa, as evidenced by its comigration with theŽ .catalytically active 26 kDa domain of MutY Fig. 2

w x21 . However, an insufficient amount of protein wasobtained for amino acid sequence determination.

A second round of purification included reversedphase HPLC after heparin sepharose column chro-

Ž .matography Fig. 1, Scheme a2 . SDS-PAGE andsilver staining of the HPLC fractions demonstratingpyrimidine dimer-specific nicking activity revealed asingle protein band with a relative molecular weight

Ž .of 26 kDa data not shown , consistent with thefindings of purification scheme a1. This band wassubmitted to the NIEHS Protein Chemistry Core atUTMB for N-terminal amino acid sequencing. Table2 shows the 30, N-terminal amino acid residues ofBsp-pdg. Database searches using this 30 amino acidsequence as the query did not detect any sequencesimilarity with T4-pdg or cv-pdg, but did reveallimited sequence similarities with a 30S ribosomal

Ž .protein 3S from B. stearothermophilus and a tran-scriptional regulator from Herpes simplex virus type

Ž . Ž .1 HSV-1 strain HFEM Table 2 . This limitedamino acid sequence comparison suggests that Bsp-pdg may not belong to the 4Fe4S cluster superfamilyof DNA glycosylaserAP lyase enzymes, which in-

Table 2Amino terminal sequence of the B. sphaericus protein


cludes endonuclease III, MutY, AlkA and M. luteus-pdg. However, such a tentative conclusion should betempered by the fact that significant sequence simi-larities between endonuclease III and M. luteus-pdgdo not begin until residue 18 of endonuclease III.Thus, it is still possible that the Bsp-pdg is a mem-ber of this superfamily. However, if further studiesreveal no additional similarities, these data may sug-gest that the Bsp-pdg structure constitutes a novelDNA glycosylase motif. Confirmation of this awaitsthe cloning of the gene and the expression andpurification of sufficient quantities of the enzyme forbiophysical characterization.

3.2. Biochemical characterization of Bsp-pdg

The highly purified Bsp-pdg protein describedabove was assayed for UV photoproduct-specificnicking activity by assaying on 49-mer oligonucleo-tides containing a single site-specific cis–syn,trans–syn I or trans–syn II thymine dimer. LikeT4-pdg and cv-pdg, Bsp-pdg was able to incise

Ž .DNA containing cis–syn dimers. Fig. 3 . Uponreaction with cis–syn dimer-containing DNA, T4-pdgand cv-pdg gave rise to both b-elimination andd-elimination products, which have 3X 4-hydroxy-2-

w xpentanal and phosphate ends, respectively 24 . TheseŽ .products are seen in Fig. 3 lanes 2 and 3 and are

recognized by their mobility in relation to markersX Ž .possessing 3 hydroxyl termini data not shown . In

contrast, Bsp-pdg gave rise only to the b-eliminationŽ .product Fig. 3, lane 4 . Since the accumulation of

the d-elimination product requires secondary encoun-w xters between T4-pdg and the DNA substrate 24 , the

lack of significant d-elimination product by Bsp-pdgcould be the result of the relatively low enzymeconcentration used in the reaction. Alternatively, thisenzyme may be very inefficient in catalyzing thed-elimination reaction. When each of the enzymeswere examined for activity against DNAs containinga trans–syn I cyclobutane dimer, very minor inci-

Ž .sion activity could be detected Fig. 3, lanes 6–8 .This residual activity is most likely the incision of asmall contaminant of cis–syn dimer DNA that arose

w xin the synthesis and purification of this DNA 15,16 .Additionally, Bsp-pdg appears to be unable to

efficiently incise DNA containing the trans–syn IIdimer, whereas this is an excellent substrate for

Fig. 3. Incision of cis – syn, trans – syn I and trans – syn IIthymine dimer-containing oligonucleotides. Double-stranded

Ž . Ž .oligonucleotides 49-mers containing a cis – syn lanes 1–4 ,Ž . Ž .trans – syn I lanes 5–8 , or trans – syn II lanes 9–12 thymine

dimer, radiolabeled on the 5X end of the dimer-containing strand,Ž .were incubated at 378C for 30 min with Bsp-pdg lanes 4, 8, 12 ,

Ž . Ž .T4-pdg lanes 2, 6, 10 or cv-pdg lanes 3, 7, 11 . Lanes 1,5, and9 represent controls in the absence of enzyme. After reaction, thesamples were heated at 908C for 5 min and then separated byelectrophoresis through a 15% polyacrylamide gel containing 8 Murea. The positions of the b- and d-elimination products areindicated.

Ž .T4-pdg and cv-pdg Fig. 3, lanes 10–12 . The inabil-ity to cleave the trans–syn II dimer was somewhatsurprising, since the 5X pyrimidine of this lesion isessentially in the same conformation as the 5X pyrim-idine in the more common cis–syn dimer.

In addition to examining its specificity for cy-clobutane pyrimidine dimer-containing DNA, Bsp-pdg was tested as to its ability to cleave DNAcontaining other common UV photoproducts, namely

Ž .the 6–4 photoproduct and its Dewar isomer. Tothis end, Bsp-pdg was reacted with site-specifically

Ž .modified oligonucleotides containing either a 6–4or Dewar photoproduct of TpT. Upon gel elec-trophoresis of the reaction products, no evidence for

Žincision of these photoproducts was found data not.shown . This result is not surprising, since neither

T4-pdg nor cv-pdg are able to incise these substrates.Overall, the specificity toward UV-induced lesionsare similar to what has been described for the M.

w xluteus pyrimidine dimer glycosylase 18 . However,in the M. luteus study, the trans–syn II dimers werenot tested.


Since Bsp-pdg appeared to catalyze a combinedglycosylaserAP lyase reaction on cis–syn dimer-containing DNA, as evidenced by the formation of acleavage product consistent with a b-eliminationmechanism, it was predicted that Bsp-pdg would beable to cleave AP site-containing DNA by b-elimination. AP site-containing DNA was preparedby reacting a radiolabeled 49-mer oligonucleotidecontaining a site-specific uracil with uracil DNAglycosylase. Immediately after preparation, this DNAwas incubated with aliquots of pooled Bsp-pdg fromheparin sepharose chromatography. Bsp-pdg wasfound to incise AP sites as predicted, resulting in

Ž .both b- and d-elimination products Fig. 4, lane 3 .The enzyme was found to have no activity on an

Fig. 4. Incision of an AP site-containing oligonucleotide byBsp-pdg. A double-stranded 49-mer oligonucleotide containing asingle AP site, labeled on the 5X end of the AP site-containingstrand, was incubated at 378C for 30 min with heparin sepharose-

Ž .purified Bsp-pdg lane 3 . Samples were treated as described inFig. 3, except without heating to 908C. Lane 4 shows a 20-mermarker with a 3X OH; the sequence of the 20-mer is identical tothe 5X end of the labeled AP site-containing DNA. Lane 1:deoxyoligonucleotide standards; lane 2: AP site-containing DNAin the absence of enzyme.

unmodified, control duplex 49-mer oligonucleotideŽ .data not shown . To positively identify the b- andd-elimination products, a 20-mer oligonucleotidemarker, having the same sequence as the putativereaction products but with a 3X hydroxyl terminus,was electrophoresed with the reaction products. This

Ž .marker migrated to a position Fig. 4, lane 4 thatwas distinct from the b- and d-elimination productsresulting from reaction of the substrate with Bsp-pdg.The mobilities of the Bsp-pdg reaction products areconsistent with the expected b- and d-eliminationproducts containing 3X 4-hydroxy-2-pentenal and 3X

phosphate termini, respectively.In data not shown, it was also determined that

Bsp-pdg was unable to incise DNA containing eitherA:G or A:8-oxoG mismatches, the known substrates

w xof MutY 21 . Bsp-pdg was also unable to cleaveduplex DNA containing 8-oxoG:C, a major substratefor the Fpg glycosylaserAP lyase. Based on theresults presented above and the enzyme’s inability tocleave mismatched base pairs or oxidative damage,Bsp-pdg appears to have a narrow substrate speci-ficity. These data indicate that of the substratestested in this study, Bsp-pdg most efficiently incisesDNA containing UV-induced cis–syn cyclobutanepyrimidine dimers and that it shares a similar cat-alytic mechanism to that of T4-pdg, a structurallydistinct glycosylaserAP lyase.

Glycosyl bond scission at a DNA lesion, such as apyrimidine dimer, is believed to be initiated by anucleophilic attack on the 5X sugar linking the lesionto the phosphate backbone. The nucleophile caneither be a primary amino group of the glycosylase

wenzyme or an activated water molecule reviewed inw xRef. 13 . If the attacking nucleophile is a primary

amino group, a Schiff base enzyme–DNA intermedi-ate results and cleavage of the DNA backbone isaccomplished by a b-elimination or b,d-eliminationmechanism. This imino intermediate can be reducedby NaBH , resulting in a stable covalent linkage4

between the enzyme and the DNA that can be de-tected as a shifted band by electrophoresis through a

w w xdenaturing polyacrylamide gel reviewed in Ref. 3 .Conversely, if the attacking nucleophile is an acti-vated water molecule, no covalent enzyme–DNAcomplex would be formed and no phosphodiester

Ž .bond cleavage lyase activity would occur. Further-more, no shifted band would be observed upon gel


Fig. 5. NaBH trapping of a covalent enzyme–DNA complex. A4

5X end-labeled 49-mer duplex containing a cis – syn cyclobutanepyrimidine dimer was incubated at 378C for 30 min with increas-

Žing concentrations of either Bsp-pdg 0.1, 0.2, 1.0 and 2.0 ng,. Žlanes 1–4, respectively or T4-pdg 0.1, 1.0 and 2.0 ng, lanes 5–7,

.respectively in the presence of 100 mM NaBH . Lane 8 repre-4

sents the control, untreated oligonucleotide. The covalently trappedcomplexes were separated by electrophoresis through a 15%polyacrylamide–SDS gel containing 8 M urea. Wet gels wereanalyzed by autoradiography.

electrophoresis of reaction products formed in thepresence of NaBH .4

In order to better understand the catalytic mecha-nism of Bsp-pdg, covalent trapping experiments us-ing NaBH were performed. As described above, the4

electrophoretic mobilities of the Bsp-pdg reactionproducts were consistent with the expected mobilityof cleavage products resulting from a b-eliminationmechanism. For this reason, we hypothesized thatthe enzyme would be trapped on cis–syn thyminedimer-containing DNA when the reaction was car-ried out in the presence of NaBH . As expected, a4

covalent enzyme–DNA intermediate was revealedupon gel electrophoresis of such reaction productsŽ .Fig. 5, lanes 1–4 , and the amount of this covalentenzyme–DNA complex increased with increasingenzyme concentration. A similar experiment per-formed using T4-pdg, which is known to initiateN-glycosyl bond cleavage via attack of a primaryamino group, also revealed a dose-dependent cova-

Ž .lent enzyme–DNA complex Fig. 5, lanes 5–7 . Noevidence of a covalent Bsp-pdg–DNA complex was

Ž .detected in the absence of NaBH data not shown .4

3.3. Proposed reaction mechanism of Bsp-pdg

Based on these results, a proposed reaction mech-anism consistent with that of a glycosylaserAP lyasehas been postulated for Bsp-pdg. This mechanismpredicts that Bsp-pdg has two associated activities:N-glycosylase activity, which cleaves the glycosylicbond connecting the 5X pyrimidine of the cis–syndimer to the deoxyribose, and an AP lyase activity,which cleaves the phosphodiester bond 3X to themodified sugar by a b-elimination mechanism. Thesecleavage reactions are believed to result in productswith 3X

a ,b-unsaturated aldehyde and 5X phosphatetermini. Bsp-pdg is further proposed to utilize anamino group to initiate the attack on the C1X of thesugar attached to the 5X damaged base. This attackresults in a covalent enzyme–DNA imino intermedi-ate that can be trapped by reduction with sodiumborohydride. Thus, this work has revealed a novelDNA glycosylaserAP lyase in B. sphaericus.

Acknowledgements

We are indebted to Drs. Richard Roberts andŽ .Peter Goldfarb New England Biolabs for providing

the cell-free extracts and cell pellets of B. sphaeri-cus.

This work would not have been possible withoutthe generous gifts of site specifically modified oligo-nucleotides from John-Stephen Taylor, WashingtonUniversity. Special thanks are also extended to Dr.Kathy Latham and Mary Lou Augustine for perform-ing the initial screenings of the cell-free extracts andto Dr. Latham for expert critiquing of this study.Thanks also are given to Ms. Rosemary Martinezand Mrs. Lisa Pipper Stephenson for their help in thepreparation of this manuscript.

References

w x1 E. Sage, Yearly review — distribution and repair of photole-sions in DNA: genetic consequences and the role of sequence

Ž .context, Photochem. Photobiol. 57 1993 163–174.w x2 K.H. Kraemer, Commentary — sunlight and skin cancer:

another link revealed, Proc. Natl. Acad. Sci. U. S. A. 94Ž .1997 11–14.


w x3 A.K. McCullough, M.L. Dodson, R.S. Lloyd, Initiation ofbase excision repair: glycosylase mechanisms and structures,

Ž .Annu. Rev. Biochem. 68 1999 255–285.w x4 R.S. Lloyd, Base excision repair of cyclobutane pyrimidine

Ž .dimers, Mutat. Res. — DNA Repair 408 1998 159–170.w x5 R.S. Lloyd, The initiation of DNA base excision repair of

Ž .dipyrimidine photoproducts, Progr. Nucl. Acid Res. 62 1998155–175.

w x6 R.D. Schrock, R.S. Lloyd, Reductive methylation of theamino terminus of endonuclease V eradicates catalytic activi-ties. Evidence for an essential role of the amino terminus inthe chemical mechanisms of catalysis, J. Biol. Chem. 266Ž .1991 17631–17639.

w x7 K. Morikawa et al., X-ray structure of T4 endonuclease V: anexcision repair enzyme specific for a pyrimidine dimer,

Ž .Science 256 1992 523–526.w x8 T. Doi et al., Role of the basic amino acid cluster and Glu-23

in pyrimidine dimer glycosylase activity of T4 endonucleaseŽ .V, Proc. Natl. Acad. Sci. U. S. A. 89 1992 9420–9424.

w x9 N. Hori et al., Participation of glutamic acid 23 of T4endonuclease V in the beta-elimination reaction of an abasic

Ž .site in a synthetic duplex DNA, Nucl. Acids Res. 20 19924761–4764.

w x10 M.L. Dodson, R.D.D. Schrock, R.S. Lloyd, Evidence for animino intermediate in the T4 endonuclease V reaction, Bio-

Ž .chemistry 32 1993 8284–8290.w x11 R.D. Schrock, R.S. Lloyd, Site-directed mutagenesis of the

NH2 terminus of T4 endonuclease V. The position of thealpha NH2 moiety affects catalytic activity, J. Biol. Chem.

Ž .268 1993 880–886.w x12 R.C. Manuel et al., Involvement of glutamic acid 23 in the

catalytic mechanism of T4 endonuclease V, J. Biol. Chem.Ž .270 1995 2652–2661.

w x13 M.L. Dodson, M.L. Michaels, R.S. Lloyd, Unified catalyticŽ .mechanism for DNA glycosylases, J. Biol. Chem. 269 1994

32709–32712.w x14 M. Furuta et al., Chlorella virus PBCV-1 encodes a homolog

of the bacteriophage T4 UV damage repair gene denV, Appl.Ž .Environ. Microbiol. 63 1997 1551–1556.

w x15 A.K. McCullough et al., Characterization of a novel cis–synand trans–syn-II pyrimidine dimer glycosylaserAP lyasefrom a eukaryotic algal virus, Paramecium bursaria chlorella

Ž .virus-1, J. Biol. Chem. 273 1998 13136–13142.w x16 J.F. Garvish, R.S. Lloyd, The catalytic mechanism of a

Ž .pyrimidine dimer-specific glycosylase pdg rabasic lyase,Ž .Chlorella virus-pdg, J. Biol. Chem. 274 1999 9786–9794.

w x17 J.F. Garvish, R.S. Lloyd, Active-site determination of aŽ .pyrimidine dimer glycosylase, J. Mol. Biol. 2000 in press.

w x18 C.E. Piersen et al., Purification and cloning of Micrococcusluteus ultraviolet endonuclease, an N-glycosylaserabasiclyase that proceeds via an imino enzyme-DNA intermediate,

Ž .J. Biol. Chem. 270 1995 23475–23484.w x19 S. Shiota, H. Nakayama, UV endonuclease of Micrococcus

luteus, a cyclobutane pyrimidine dimer–DNA glycosylaserabasic lyase: cloning and characterization of the gene, Proc.

Ž .Natl. Acad. Sci. U. S. A. 94 1997 593–598.w x20 M.L. Michaels et al., A repair system for 8-oxo-7,8-dihydro-

Ž .deoxyguanine, Biochemistry 31 1992 10964–10968.w x21 R.C. Manuel, E.W. Czerwinski, R.S. Lloyd, Identification of

the structural and functional domains of MutY, an Es-cherichia coli DNA mismatch repair enzyme, J. Biol. Chem.

Ž .271 1996 16218–16226.w x22 C.D. Mol et al., Crystal structure and mutational analysis of

human uracil–DNA glycosylase: structural basis for speci-w x Ž .ficity and catalysis see comments , Cell 80 1995 869–878.

w x23 C.A. Smith, J.S. Taylor, Preparation and characterization of aset of deoxyoligonucleotide 49-mers containing site-specific

Ž .cis–syn, trans–syn-I, 6–4 , and Dewar photoproducts ofŽ X X . Ž .thymidylyl 3™5 -thymidine, J. Biol. Chem. 268 1993

11143–11151.w x24 K.A. Latham, R.S. Lloyd, Delta-elimination by T4 endonu-

clease V at a thymine dimer site requires a secondary bindingŽ .event and amino acid Glu-23, Biochemistry 34 1995 8796–

8803.

Purification and characterization of a novel UV lesion-specific DNA glycosylase/AP lyase from...

Documents

Transcript of Purification and characterization of a novel UV lesion-specific DNA glycosylase/AP lyase from...