Comparative Biochemistry and Physiology Part B 127 (2000) 399–410

Purification and characterization of a cold-adapteduracil-DNA glycosylase from Atlantic cod (Gadus morhua)

Olav Lanes a, Per Henrik Guddal b, Dag Rune Gjellesvik b,Nils Peder Willassen a,*

a Department of Biotechnology, Institute of Medical Biology, Faculty of Medicine, Uni6ersity of Tromsø, N-9037 Tromsø, Norwayb Biotec ASA, N-9008 Tromsø, Norway

Received 25 January 2000; received in revised form 12 July 2000; accepted 27 July 2000

Abstract

Uracil-DNA glycosylase (UDG; UNG) has been purified 17 000-fold from Atlantic cod liver (Gadus morhua). Theenzyme has an apparent molecular mass of 25 kDa, as determined by gel filtration, and an isoelectric point above 9.0.Atlantic cUNG is inhibited by the specific UNG inhibitor (Ugi) from the Bacillus subtilis bacteriophage (PBS2), and hasa 2-fold higher activity for single-stranded DNA than for double-stranded DNA. cUNG has an optimum activitybetween pH 7.0–9.0 and 25–50 mM NaCl, and a temperature optimum of 41°C. Cod UNG was compared with therecombinant human UNG (rhUNG), and was found to have slightly higher relative activity at low temperaturescompared with their respective optimum temperatures. Cod UNG is also more pH- and temperature labile than rhUNG.At pH 10.0, the recombinant human UNG had 66% residual activity compared with only 0.4% for the Atlantic cUNG.At 50°C, cUNG had a half-life of 0.5 min compared with 8 min for the rhUNG. These activity and stability experimentsreveal cold-adapted features in cUNG. © 2000 Elsevier Science Inc. All rights reserved.

Keywords: Uracil-DNA glycosylase; UNG; UDG; Catalytic efficiency; Stability; Processivity; Protein purification; Cold-adapted; Gadusmorhua

www.elsevier.com/locate/cbpb

1. Introduction

Uracil-DNA glycosylase (UDG; UNG) cataly-ses the hydrolysis of the N-glycosylic bond be-tween the deoxyribose sugar and the base in uracilcontaining DNA, and was first isolated and char-acterized from Escherichia coli (Lindahl, 1974).This is the first step in the base excision repair

(BER) pathway of removing uracil from DNA(Lindahl, 1979), and the apyrimidinic site gener-ated is thereafter repaired by an AP endonuclease,phosphodiesterase, DNA polymerase and DNAligase in the BER pathway (Kubota et al., 1996).

Several classes of UDGs have been described.The major cellular form of UDG in human cells isUNG encoded by the UNG gene (Slupphaug etal., 1995). Other classes consist of cyclin-like hu-man UDG2 (Muller and Caradonna, 1991), singlestrand selective monofunctional UDG (SMUG1)from human and Xenopus lae6is (Haushalter etal., 1999), G/T:U-specific mismatch DNA glyco-sylase (MUG) isolated from E. coli (Barrett et al.,1998), and Thermotoga maritima UDG(TMUDG; Sandigursky and Franklin, 1999).

Abbre6iations: cUNG, cod UNG; FPLC, fast protein liquidchromatography; hUNG, human UNG; PCR, polymerasechain reaction; rhUNG, recombinant human UNG; UNG,uracil-DNA glycosylase.

* Corresponding author. Tel.: +47-77-644651; fax: +47-77-645350.

E-mail address: nilspw@fagmed.uit.no (N.P. Willassen).

0305-0491/00/$ - see front matter © 2000 Elsevier Science Inc. All rights reserved.PII: S0305-0491(00)00271-6

O. Lanes et al. / Comparati6e Biochemistry and Physiology, Part B 127 (2000) 399–410400

Human UNG is a monomeric protein of about25–35 kDa in mass. It is not dependent on anycofactors and is highly conserved among differentspecies (Krokan et al., 1997). UDG is, however,affected by ionic strength. It has been shown toact both in a processive ‘sliding mechanism’,where it locates sequential uracil residues prior todissociation from the DNA, and a distributive‘random hit’ mechanism (Higley and Lloyd, 1993;Purmal et al., 1994; Bennett et al., 1995). UDGhas previously been isolated and characterizedfrom rat liver (Domena and Mosbaugh, 1985),human cells and tissues (Wittwer and Krokan,1985; Wittwer et al., 1989), calf-thymus (Talpaert-Borle et al., 1979), slime mold (Guyer et al.,1986), yeast (Crosby et al., 1981), wheat germ(Blaisdell and Warner, 1983), prokaryotes (Lin-dahl, 1974; Cone et al., 1977; Williams and Pol-lack, 1990; Sobek et al., 1996; Purnapatre andVarshney, 1998) and viruses (Winters andWilliams, 1993; Focher et al., 1993).

In human and rat, both mitochondrial- and anuclear UDG have been isolated (Domena andMosbaugh, 1985; Caradonna et al., 1996). In hu-man and mouse cells, these two are encoded bythe same gene, UNG (Nilsen et al., 1997). Byusing two different transcription start sites andalternative splicing, two forms are generated dif-fering only in the N-terminal signal sequence,which targets the enzyme to the nucleus (UNG2)and mitochondria (UNG1), respectively (Nilsen etal., 1997). Recently, several studies have beenperformed to further study the N-terminal signalsequences and the targeting of UNG to the nu-cleus and mitochondria (Bharati et al., 1998; Ot-terlei et al., 1998), and the nuclear UNG2 is foundto be phosphorylated (Muller-Weeks et al., 1998).The crystal structures of UDG from man, herpessimplex virus and E. coli have been solved (Mol etal., 1995; Savva et al., 1995; Ravishankar et al.,1998). The active site residues are conserved andthe mode of action in these enzymes seems to bethe same with a nucleotide flipping mechanism toremove uracil from DNA (Parikh et al., 1998).

Enzymes from cold-adapted organisms, such asthe Atlantic cod (Gadus morhua), have to com-pensate for the reduction of chemical reactionrates at low temperatures to maintain sufficientmetabolic activities. This can be achieved byhigher transcription/translational levels, or im-proved catalytic efficiency (kcat/KM). Higher cata-lytic efficiencies can be reached by a more flexible

structure, compared with their warm-adaptedcounterparts, providing enhanced ability to un-dergo conformational changes during catalysis.The reduced stability to pH, temperature anddenaturing agents is regarded as a consequence ofa higher conformational flexibility (Feller andGerday, 1997).

In this paper, we describe the purification andcharacterization of a heat-labile UDG from acold-adapted organism, and represent the firstUNG purified and characterized from a teleostianspecies. The enzyme has similar characteristics, aspreviously described UDGs, with respect tomolecular weight, isoelectric point, pH and NaCloptimum. However, the cod enzyme is more pH-and heat-labile and has a higher relative activityat low temperatures, compared to the recombi-nant warm-adapted human UNG.

2. Materials and methods

2.1. Materials

Q-Sepharose FF, S-sepharose FF, Heparin Sep-harose HP (Hi-trap 5 ml), Poly-U-Sepharose 4B,superdex 75 HR10/30, Phast system and PhastIEF gels (3–9) and LMW gel filtration calibrationkit were obtained from Amersham PharmaciaBiotech (Uppsala, Sweden). Deoxy[5-3H]uridine5%-triphosphate (19.3 Ci/mmol) was purchasedfrom Amersham (UK). UNG inhibitor (Ugi) wasobtained from New England Biolabs (Beverly,MA), enzymes were purchased from Promega(Madison, WI). Protease inhibitors, calf-thymusDNA (D-1501), uracil, deoxyuridine and de-oxyuridine-monophosphate were purchased fromSigma (St Louis, MO). All other reagents andbuffers were purchased from Sigma or Merck(Darmstadt, Germany).

2.2. Purification of cUNG

Preparation of crude extract and all purificationsteps were performed at 4°C.

2.2.1. Preparation of crude extractTo 600 ml extract buffer (25 mM Tris/HCl, 100

mM NaCl, 1 mM ethylene diamine tetra aceti-cacid (EDTA), 1 mM DTT, 10% (v/v) glycerol,pH 8.0) 200 g of fresh cod liver was added andhomogenized in an Atomix homogenizer (MSE,

O. Lanes et al. / Comparati6e Biochemistry and Physiology, Part B 127 (2000) 399–410 401

UK). Before homogenization, the followingprotease inhibitor mix was added to the extractbuffer, 1 mM PMSF, 1 mM pepstatin, 1 mMleupeptin, 10 mM TPCK and 10 mM TLCK (finalconcentrations). The homogenate was centrifugedat 28 000×g for 15 min and the supernatant wasfiltered through glass wool. Finally, glycerol wasadded to 30% (v/v), and the cod liver crudeextract was frozen at −70°C.

2.2.2. Q-Sepharose Fast FlowOne liter crude extract was diluted with 1 l

buffer A (25 mM Tris/HCl, 10 mM NaCl, 1 mMEDTA, 1% (v/v) glycerol, pH 8.0 (fraction 1). Thesample was applied in two portions (1 l each) ona Q-Sepharose FF column (5.0/15), equilibratedwith buffer A, and then washed with 250 mlbuffer A, using a flow-rate of 10 ml/min. Proteinsbound to the column were eluted with 200 mlbuffer A+1.0 M NaCl, and the column wasre-equilibrated with buffer A before the next partof the sample was applied, as mentioned above.The UNG-containing eluate and wash fractionsfrom both two runs were pooled (fraction 2, 2340ml).

2.2.3. S-Sepharose Fast FlowFraction 2 was applied to an S-Sepharose FF

column (1.6/10) equilibrated in buffer A, flow rate10 ml/min. The column was washed with 300 mlbuffer A+60 mM NaCl, and eluted using a 200ml linear gradient of 0.06–0.4 M NaCl in bufferA, flow rate 5 ml/min. UNG-containing fractionswere pooled (55 ml) and dialyzed overnightagainst buffer A (fraction 3).

2.2.4. Heparin Sepharose HPFraction 3 was applied to a Heparin Sepharose

HP Hi-Trap column (1.6/2.5) equilibrated inbuffer A. The column was washed with 50 mlbuffer A+60 mM NaCl and was eluted in a 50ml linear gradient of 0.06–0.4 M NaCl in bufferA, flow rate 1 ml/min. UNG-containing fractionswere pooled (fraction 4, 20 ml).

2.2.5. Poly-U-Sepharose (4B)Fraction 4 was then diluted five times in buffer

A, and applied to a Poly-U-Sepharose column(1.6/10) equilibrated in buffer A. The column waswashed with 60 ml buffer A+60 mM NaCl andwas eluted in a 200 ml linear gradient of 0.06–0.4M NaCl in buffer A, flow rate 1 ml/min. UNG-

containing fractions were pooled (fraction 5, 70ml).

2.2.6. Superdex 75Fraction 5 was concentrated to 200 ml using

Ultrafree 15 and Ultrafree-MC ultracentrifuga-tion filters (Millipore), cutoff 5K, and applied ona Superdex 75 column (HR 1.0/30) equilibrated inbuffer A, with a flow rate of 0.5 ml/min. Frac-tions (350 ml) were collected, and those containingUNG-activity were pooled (fraction 6, 3 ml).

2.3. Preparation of substrate by nick-translation

3H-dUMP DNA was prepared by nick-transla-tion and polymerase chain reaction (PCR). Thenick-translated substrate was made up in a totalvolume of 1 ml and contained 50 mM Tris/HCl,10 mM MgSO4, 1 mM DTT, pH 7.2, 250 mgcalf-thymus DNA (purified by phenol/chloroformextraction and ethanol precipitated, prior to use),0.1 mM dATP, dCTP, dGTP and dUTP, where2.6 mM of the dUTP was [3H]-dUTP (19.3 Ci/mmol). The reaction was initiated by adding 0.1ng(5.35×10−4 U) DNase I (bovine pancreas,Promega), and 30 s later 25 U of E. coli DNApolymerase. The nick-translation mix was incu-bated at 21°C for 24 h. The nick-translated DNAwas purified by phenol/chloroform extraction,and ethanol precipitation. DNA was resuspendedin 50 ml TE-buffer and purified using an NAP-5column (AP Biotech) equilibrated in TE-buffer(10 mM Tris/HCl, 1 mM EDTA, pH 8.0) toremove unincorporated nucleotides. Specific-activ-ity of the nick-substrate was 1.8×105 dpm/mgDNA (425 cpm/pmol uracil).

2.4. Preparation of substrate by PCR

The PCR-produced substrate was used for allcharacterization experiments and consisted of a761 bp fragment generated from cationic trypsi-nogen (sstrpIV) from Atlantic salmon (Salmosalar) (Male et al., 1995). The PCR was carriedout in a volume of 50 ml in a Perkin Elmer Cetusthermocycler. The PCR-mix contained 10 mMTris/HCl, pH 8.3, 50 mM KCl, 6 mM MgCl2,0.37 mM dATP, dCTP, dGTP and dUTP, where10.2 mM of dUTP was [3H]-dUTP (17.0 Ci/mmol,Amersham), 700 pg template DNA (sstrpIV in apgem7zt-vector), 2.5 mM of upstream and down-stream primers and 2 U Taq-polymerase (Roche,

O. Lanes et al. / Comparati6e Biochemistry and Physiology, Part B 127 (2000) 399–410402

Switzerland). The PCR-reaction was performedby 30 cycles of 94°C for 1 min, 45°C for 1 minand 72°C for 1 min. Then, an additional 2 U ofTaq-polymerase was added and the PCR-reactioncontinued with 30 new cycles, as described above.The PCR-substrate was purified with QIAquickPCR-purification kit (Qiagen), as described by themanufacturer, and eluted in 50X diluted TE-buffer, pH 8.0. Specific-activity of the PCR-sub-strate was 5.9×105 dpm/mg DNA (451 cpm/pmoluracil). All characterization experiments weredone using the PCR-substrate.

2.5. Detection of UNG acti6ity (standard assay)

UNG activity was measured in a final volumeof 20 ml, containing 70 mM Tris/HCl, 10 mMNaCl, 1 mM EDTA, 100 mg/ml BSA, pH 8.0 and230 ng nick-substrate or 71 ng PCR-substrate.The reaction mixture was incubated 10 min at37°C, and terminated with the addition of 20 ml ofice-cold single-stranded calf-thymus DNA (1 mg/ml) and 500 ml 10% (w/v) TCA. Samples wereincubated on ice for 15 min, and centrifuged at16 000×g for 10 min. Supernatants containingacid-soluble 3H-uracil were analyzed using a liq-uid scintillation counter. One unit of activity isdefined as the amount of enzyme required torelease 1 nmol of acid-soluble uracil per minute at37°C.

2.6. Analysis of assay products with thin layerchromatography

Reaction products after assays were mixed with20 nmol uracil, deoxyuridine and deoxyuridine-monophosphate. Thin layer chromatography wasperformed according to the method by Wang andWang (1967), using polyamide layer plates(BDH), and tetrachloromethane, acetic acid andacetone (4:1:4, by volume) as solvent. Spots de-tected under UV-light were cut out of the plateand radioactivity measured in a liquid scintillationcounter.

2.7. Molecular weight determination

The molecular weight was determined by gelfiltration, and was performed with a Superdex 75column (1.0/30) equilibrated in a buffer contain-ing 25 mM Tris/HCl, 1.0 M NaCl, 1 mM EDTA,1% (v/v) glycerol, pH 8.0. The flow rate was 0.5

ml/min, and activity was measured in the frac-tions collected (250 ml). Bovine serum albumin(BSA, 67 kDa), ovalbumin (43 kDa), chy-motrypsinogen A (25 kDa) and ribonuclease A(13.7 kDa) were used as standards. Blue dextranand sodium chloride were used to determine void-(V0) and intrinsic volume (Vi), respectively.

2.8. Protein determination

Protein concentrations were determined by themethod of Bradford (1976) with the microtiterplate protocol, using bovine serum albumin (BSA)as a standard.

2.9. Isoelectric point determination

Isoelectric point determination was done withthe Phast-system, isoelectric focusing gel 3–9 andsilver-stained according to methods described bythe manufacturer. Standards used were phyco-cyanin (pI 4.45, 4.65, 4.75), b-lactoglobulin B (pI5.10), bovine carbonic anhydrase (pI 6.00), hu-man carbonic anhydrase (pI 6.50), equine myo-globin (pI 7.00), human hemoglobin A (pI 7.10),human hemoglobin C (pI 7.50), lentil lectin (pI7.8, 8.0, 8.2), cytochrome c (pI 9.6, IEF standardspI range 4.45–9.6, Bio-Rad). After focusing, thegel was cut into 2 mm pieces and incubated in 250ml extraction buffer (50 mM Tris/HCl, 0.2 MNaCl, 1 mM DTT, 1 mM EDTA, 1% (v/v) glyc-erol, pH 8.0) overnight. Activity was measuredusing standard assay conditions.

2.10. Determination of pH/NaCl-optimum

Assays were done in a volume of 20 ml, asdescribed in standard assay using the PCR gener-ated substrate and NaCl concentration from 0 to200 mM with 25 mM intervals, and pH rangingfrom 9.5 to 6.5 with 0.5 pH unit intervals. Allbuffers contained 100 mg/ml BSA and 1 mMEDTA. The buffers used were diethanolamine/HCl (pH 9.5–8.5), Tris/HCl (pH 8.5–7.5) andMOPS/NaOH (pH 7.5–6.5). All buffers were pH-adjusted at 37°C and used in 25 mM concentra-tion in the assay.

2.11. Determination of temperature optimum

Assays were performed in a volume of 20 ml, asdescribed in standard assay using the PCR-gener-

O. Lanes et al. / Comparati6e Biochemistry and Physiology, Part B 127 (2000) 399–410 403

ated substrate. The assay mixtures were as de-scribed in standard assay conditions and wereadjusted to pH 8.0 for all temperatures. Thetemperature range used was 5–60°C. The activitywas measured in a sequential manner with 15 minintervals between each temperature. The enzymesused were diluted in standard dilution buffer (5mM Tris/HCl, 10 mM NaCl, 1% (v/v) glycerol,pH 8.0) and placed on ice. Due to the instabilityof the enzyme sample on ice over a prolongedperiod, results were corrected with respect to thestability of the enzymes incubated in dilutionbuffer on ice, with the formula N(t)=cpm/e−0693(t/l), where half-life (l) of cUNG andrhUNG are 2.0 and 2.6 h, respectively.

2.12. Effect of pH and temperature on stability

2.12.1. pHUNG (0.01 U) was preincubated (in a total

volume of 75 ml) for 10 min at 37°C in bufferscontaining 10 mM buffer, 10 mM NaCl, 1 mMEDTA, 1% (v/v) glycerol, with pH ranging from10.0 to 5.5 with 0.5 pH unit intervals using piper-azine/HCl (pH 10.0–9.5), diethanolamine/HCl(pH 9.5–8.5), Tris/HCl (pH 8.5–7.5), MOPS/NaOH (pH 7.5–6.5) and MES/NaOH (pH 6.5–5.5) as buffer components. All buffers werepH-adjusted at 37°C. Aliquots of 5 ml were trans-ferred to the assay mixtures and residual activitywas measured using standard assay conditions.

2.12.2. TemperatureUNG (0.01 U) was preincubated (in a total

volume of 75 ml) in 10 mM Tris/HCl, 50 mMNaCl, 1mM EDTA, 1% (v/v) glycerol, pH 8.0(adjusted at each temperature). After differenttime-intervals, as indicated in figure legends, 5 mlaliquots were transferred to the assay mixturesand residual activity was measured using standardassay conditions.

2.13. Substrate specificity against ss/ds DNA

PCR and nick-translated substrate was incu-bated 3 min at 100°C and thereafter rapidlycooled on ice to generate ssDNA. Following de-naturation, the ssDNA-substrates were used instandard assay conditions with 6.65×10−4 Upurified cUNG, and compared with the dsDNAPCR- and nick-generated substrates.

2.14. Ugi and uracil product inhibition

Activity measurements using PCR-substratewere performed with 6.65×10−4 U of purifiedcUNG. Various concentrations of uracil (0, 1, 2and 5 mM) or Ugi (1.25×103–2.00×102 U) wereadded to the assay mixtures (on ice). Activity wasthen measured, as described under standard assayconditions.

3. Results

3.1. Purification of cUNG

cUNG was purified 17 000-fold with a recoveryof 2%, as shown in Table 1. Despite the highpurification factor, the enzyme was only partlypurified as determined by sodium dodecyl sul-phate-polyacrylamide electrophoresis (SDS-PAGE). The yield was low due to manychromatographic steps and the concentration stepof the dilute protein sample prior to the gel-filter-ation step.

3.2. Molecular mass and pI-determination

The molecular mass was determined by gelfiltration to be 25 kDa92 (S.D.) from threeseparate experiments. The isoelectric point deter-mination was done with an IEF Phast Gel withIEF standards, ranging from 4.45 to 9.6. Follow-ing IEF, cUNG activity was eluted from the gelfragments and activity measured as described inSection 2. The cUNG activity eluted from the gelwas in the same region as the focusing of thecytochrome c standard, which has an isoelectricpoint, 9.6 as shown in Fig. 1.

The cytochrome c and the cUNG activity werefound where the electrode contacted the gel;therefore, we can only conclude that the pI islarger than 9.0, which is the highest measurablevalue using this system.

3.3. Substrate specificity

cUNG activity was measured using both ss-DNA and dsDNA. A 1.8- and 1.9-fold higheractivity for ssDNA than dsDNA was found usingnick- and PCR substrates, respectively, as shownin Table 2. Assay products were analyzed by thinlayer chromatography, and the major part of the

O. Lanes et al. / Comparati6e Biochemistry and Physiology, Part B 127 (2000) 399–410404

Tab

le1

Pur

ifica

tion

ofA

tlan

tic

cUN

G

Pro

tein

conc

entr

atio

nbT

otal

acti

vity

Vol

ume

Spec

ific-

acti

vity

Yie

ldP

urifi

cati

onSt

epT

otal

prot

ein

Act

ivit

ya

(%)

(ml)

(fol

d)(m

g)(U

)(m

g/m

l)(U

/mg)

(U/m

l)

0.02

110

01

480

0017

20.

086

2000

Cru

deex

trac

t0.

044

992

Q-S

epha

rose

FF

2340

0.07

317

01.

6638

8411

.658

540

8.53

0.15

598

.8S-

Seph

aros

eF

F55

1.79

61.

6220

34.3

3215

972.

775

55.5

0.08

1H

epar

inSe

phar

ose

ND

Pol

y-U

Seph

aros

e18

70N

D0.

446

31.2

ND

cN

D37

92

1767

90.

009

0.00

31.

138

3.41

Supe

rdex

753

aE

nzym

eac

tivi

tyw

asm

easu

red

asde

scri

bed

unde

rst

anda

rdas

say

inSe

ctio

n2

usin

gni

ck-s

ubst

rate

.b

Pro

tein

conc

entr

atio

nw

asde

term

ined

asde

scri

bed

inSe

ctio

n2.

cP

rote

inco

ncen

trat

ion

was

belo

wth

ede

tect

ion

limit

ofth

eB

radf

ord

tech

niqu

e.

O. Lanes et al. / Comparati6e Biochemistry and Physiology, Part B 127 (2000) 399–410 405

Fig. 1. pI determination of Atlantic cUNG. Following theisoelectric focusing using Phast Gel IEF 3–9, the gel was cutin pieces of 2 mm. Each piece was transferred to an eppendorftube and incubated overnight in extraction buffer at 4°C.Activity was then measured, as described in Section 2. Stan-dards are shown from the IEF-gel.

Fig. 2. (a) Product inhibition of cUNG with free uracil.Different concentrations of uracil was added to the assaymixture and the assay performed, as described in Section 2; (b)Inhibition of cUNG with the Bacillus subtilis bacteriophageUgi. 6.65×10−4 U of cUNG was incubated with 1.25×103–2.00×102 U of Ugi using standard assay conditions, as de-scribed in Section 2. One unit of Ugi inhibits one unit ofUNG-activity, where the UNG-activity is defined as releasing60 pmol of uracil per min at 37°C.

radioactivity was identified as uracil. However,some radioactivity was also co-localized with thedeoxyuridine marker, but this could be due to thepartial overlap of the two markers. In addition,the purified cUNG did not exhibit any significanthydrolysis of 3H-adenine-labeled DNA, thereforeexcluding nucleases as responsible for hydrolyzingthe DNA.

3.4. Inhibition studies

Product inhibition by free uracil was examinedand given more than 50% inhibition with 1 mMuracil in the assay mixture, as shown in Fig. 2a.Adding 5 mM free uracil to the assay mixture, a78% inhibition of the activity was observed. Theeffect of Ugi on cUNG was measured by addingUgi to the assay mixture. cUNG was clearlyinhibited by Ugi, as shown in Fig. 2b.

3.5. pH and NaCl optimum

The pH and sodium chloride optimum wasexamined by measuring the enzyme activity at

different pHs and NaCl concentration from 0 to200 mM, as shown in Fig. 3. The enzyme exhib-ited a broad pH-optimum, with maximal activitybetween pH 7.0–9.0, and 25–50 mM NaCl. Ashift in NaCl optimum was observed, where theoptimum NaCl concentration changed from lowconcentrations at high pH to higher concentra-tions at low pH. At pH 9.5, cUNG was inhibitedby NaCl.

3.6. Temperature optimum

The temperature optimum of cUNG was 41°C.To compare the activity of cUNG with the warm-adapted recombinant human UNG (rhUNG) atlow temperatures, enzyme activity was measuredfrom 5 to 60°C, and the activity at low tempera-

Table 2Substrate-specificity ssDNA vs. dsDNA

Activity (%)cpmSubstrate Fold increase

dsNick 1.0586 100183 1.81078ssNick100dsPCR 1158 1.0

1.91912214ssPCR

O. Lanes et al. / Comparati6e Biochemistry and Physiology, Part B 127 (2000) 399–410406

Fig. 3. pH- and NaCl-optimum of cUNG. Activity of cUNGwas measured with 0–200 mM sodium chloride in differentpH-series, as described in Section 2. The per cent UNGactivity is set relative to the highest value measured, at pH 7.5with 50 mM NaCl.

Fig. 5. pH stability of Atlantic cUNG and recombinantrhUNG. In the different pH-buffers, 0.01 U of cUNG (�) orrhUNG () were incubated for 10 min at 37°C. Then, 5 mlaliquots were transferred to the assay mixture, and residualactivity was measured using standard assay conditions asdescribed in Section 2. One hundred percent activity wasmeasured directly from a sample diluted at pH 8.0 without anyincubation step.

3.7. Stability

The stability of the two UNG enzymes wascompared by preincubating the enzymes at differ-ent pHs. Atlantic cUNG was shown to be moststable between pH 7.0 and 8.5. At pH 5.5 and10.0, it had less than 1% residual activity. rhUNGwas most stable between pH 7.0 and 9.5. At pH5.5, only 3% residual activity remained but at pH10.0, as much as 66% of the activity remained, asshown in Fig. 5. The temperature stability of thetwo UNG enzymes was compared at 4, 25, 37 and50°C (Fig. 6). At 50°C, the half-life was deter-mined to be 0.5 and 8 min for cUNG andrhUNG, respectively. At all temperatures exam-ined, rhUNG was more stable than cUNG. Half-lives determined were 20 (37°C) and 6 (25°C) minand 2 h (4°C) for cUNG and 30 (37°C) and 150(25°C) min and 2.6 h (4°C) for rhUNG.

4. Discussion

4.1. Purification and molecular weight

The UNG from Atlantic cod liver (G. morhua)was purified about 17 000-fold using several chro-matographic techniques. Still, the enzyme wasonly partly purified, as several other bands wereobserved on an SDS-PAGE gel. Human nuclear-

tures compared with their respective optimumtemperatures (Fig. 4). The activity profile of thesetwo enzymes showed little difference at 5–15°C.However, at temperatures from 20 to 40°C, ahigher relative activity was observed with cUNGthan rhUNG, whereas at high temperatures (50–60°C), the opposite was observed.

Fig. 4. Temperature profile of cUNG (�) and rhUNG ().Enzyme activity was measured as described in Section 2. Thepercent UNG activity is set relative to the highest value forcUNG (45°C) and rhUNG (50°C), respectively. Due to theprolonged incubation of the enzyme samples on ice during theexperiment, activity is corrected with respect to the stability ofthe enzymes, as described in Section 2.

O. Lanes et al. / Comparati6e Biochemistry and Physiology, Part B 127 (2000) 399–410 407

and mitochondrial UDG are shown to be gener-ated by alternate splicing, and have an ORF of313 and 304 amino acids, respectively (Nilsen etal., 1997). The molecular mass of cUNG wasdetermined to 25 kDa. This is approximately thesame molecular mass as the UNG purified fromhuman placenta (29 kDa, Wittwer et al., 1989)and the rhUNG (UNGD84, 27 kDa), whichlacked 77 and 84 of the first N-terminal aminoacids, respectively, as predicted from the mito-chondrial ORF (Slupphaug et al., 1995). Thissuggests that the N-terminal signal sequence inthe purified cUNG is processed or artificiallycleaved during purification or that AtlanticcUNG lacks an N-terminal signal sequence. Dur-ing purification, we did not see any sign of twodifferent UNGs as previously described duringpurification of UNG from rat or human sources(Domena and Mosbaugh, 1985; Wittwer andKrokan, 1985). But, as a vertebrate, one should

expect that Atlantic cod possesses both a nuclearand a mitochondrial form of UNG, but so far noeffort has been made to reconcile this matter. TheUNG from Atlantic cod was similar to otheruracil-DNA glycosylases previously purified andcharacterized with respect to the high pI, and the2-fold preference to ssDNA than dsDNA (Dom-ena et al., 1988).

4.2. Inhibition by Ugi and uracil

The Bacillus subtilis bacteriophage PBS2 UDG-inhibitor (Ugi) inhibits UNG by forming a stablecomplex with UNG at physiological conditions(Bennett et al., 1993). Ugi binds to hUDG byinserting a b-strand into the conserved DNA-binding groove, and acts by mimicking DNA(Savva and Pearl, 1995). This indicates that thestructure of the substrate-binding site of cUNG issimilar to other UNGs inhibited by the Ugi.

Fig. 6. Temperature stability of Atlantic cUNG (�) and rhUNG (). Enzymes (0.01 U) were incubated at 50, 37, 25 and 4°C, and5 ml aliquots were transferred to the assay mixture after different time-intervals, and standard assays were performed as described inSection 2. Half-lives were determined to 0.5 (50°C), 20 (37°C) and 60 min (25°C) and 2 h (4°C) for cUNG and 8 (50°C), 30 (37°C) and150 min (25°C) and 2.6 h (4°C) for rhUNG.

O. Lanes et al. / Comparati6e Biochemistry and Physiology, Part B 127 (2000) 399–410408

Inhibition with free uracil was in agreement withvalues previously reported (Domena et al., 1988).

4.3. Optimum conditions

cUNG has a broad pH-optimum from 7.0 to9.0, and the activity was strongly affected byNaCl concentration. A broad pH-optimum haspreviously been reported for several other UDGs(Winters and Williams, 1993; Purnapatre andVarshney, 1998). Interestingly, the NaCl-optimumfor cUNG increases as pH decreases. It has previ-ously been demonstrated that UDG functions in aprocessive manner at low ionic strength, which, asNaCl concentration increases the enzyme,switches to a distributive mechanism (Higley andLloyd, 1993; Bennett et al., 1995). However, Pur-mal et al. (1994) reported that UDG acted in adistributive mechanism at low ionic strength. Theprocessive mechanism is a common featureamong several DNA-interactive proteins (poly-merases, repressors, restriction/modification en-zymes, DNA repair enzymes), and theinteractions are generally electrostatic in nature(Lohman, 1986; von Hippel and Berg, 1989). In aUV-endonuclease from Micrococcus luteus, theprocessive mechanism has also been shown to bepH-dependent (Hamilton and Lloyd, 1989).Therefore, we suggest that the shift in NaCl-opti-mum with increasing pH reflects the processive/distributive nature of cUNG and that thecontroversy in the previous reports may be due todifference in buffer components and pH, in addi-tion to the differences already discussed.

The temperature-optimum (41°C) for cUNGwas slightly lower than for the warm-adaptedrhUNG (45°C; Slupphaug et al., 1995). The rela-tive activity at temperature ranging from 5 to45°C was higher for cUNG than rhUNG. Attemperatures exceeding 45°C, rhUNG shows ahigher relative activity than cUNG, presumablyas a consequence of the low temperature stabilityof cUNG.

4.4. Temperature and pH stability

Enzymes characterized from cold-adapted spe-cies tend to be more temperature and pH-labile,likely due to their flexible structure in order tomaintain enzymatic activity at low temperatures(Outzen et al., 1996). cUNG was found to bemore both pH- and temperature labile than the

rhUNG, which are known features for other cold-adapted enzymes (Outzen et al., 1996; Berglund etal., 1998).

A psychrophilic UDG from a marine bacteriumhas previously been isolated, and was comparedwith an E. coli UDG with respect to temperaturestability (Sobek et al., 1996). This prokaryoticUNG had a half-life of 2 and 0.5 min at 40 and45°C, respectively, compared with 27 and 8 minfor the E. coli UDG. cUNG was compared withthe rhUNG with respect to both temperature andpH stability. At 50°C, rhUNG had a half-life of 8min, compared with 0.5 min for cUNG. At lowertemperature, the differences in half-life were less,although rhUNG was more stable than cUNG atall temperatures examined. Both enzymes wereshown to be labile at low pH, whereas at high pH,rhUNG was more stable than cUNG.

In conclusion, the UNG from Atlantic codanalyzed in this study was shown to be similar toother UNG previously purified and characterizedwith respect to molecular weight, high pI, a broadpH-optimum and a 2-fold preference to ssDNAthan dsDNA. However, the cUNG was moretemperature- and pH-labile and has a higher rela-tive activity at low temperatures than the warm-adapted rhUNG. Considering that UNG is aconserved enzyme among different species, cUNGcould be a suitable enzyme for further studies ofcold-adaptation at a molecular level.

Acknowledgements

Purified recombinant human UNG (UNGD84)was kindly provided by Dr Hans E. Krokan atthe Institute for Cancer Research and MolecularBiology, Norwegian University of Science andTechnology. We thank Chris Fenton for help withthe manuscript. This work was supported by theResearch Council of Norway and Biotec ASA,Tromsø, Norway.

References

Barrett, T.E., Savva, R., Panayotou, G., Barlow, T.,Brown, T., Jiricny, J., Pearl, L.H., 1998. Crystalstructure of a G:T/U mismatch-specific DNA glyco-sylase: mismatch recognition by complementary-strand interactions. Cell 92, 117–129.

O. Lanes et al. / Comparati6e Biochemistry and Physiology, Part B 127 (2000) 399–410 409

Bennett, S.E., Sanderson, R.J., Mosbaugh, D.W., 1995.Processivity of Escherichia coli and rat liver mito-chondrial uracil-DNA glycosylase is affected byNaCl concentration. Biochemistry 34, 6109–6119.

Bennett, S.E., Schimerlik, M.I., Mosbaugh, D.W.,1993. Kinetics of the uracil-DNA glycosylase/in-hibitor protein association. UNG interaction withUgi, nucleic acids, and uracil compounds. J. Biol.Chem. 268, 26879–26885.

Berglund, G.I., Smalas, A.O., Outzen, H., Willassen,N.P., 1998. Purification and characterization ofpancreatic elastase from North Atlantic salmon(Salmo salar). Mol. Mar. Biol. Biotechnol. 7, 105–114.

Bharati, S., Krokan, H.E., Kristiansen, L., Otterlei, M.,Slupphaug, G., 1998. Human mitochondrial uracil-DNA glycosylase preform (UNG1) is processed totwo forms one of which is resistant to inhibition byAP sites. Nucleic Acids Res. 26, 4953–4959.

Blaisdell, P., Warner, H., 1983. Partial purification andcharacterization of a uracil-DNA glycosylase fromwheat germ. J. Biol. Chem. 258, 1603–1609.

Bradford, M.M., 1976. A rapid and sensitive methodfor the quantitation of microgram quantities ofprotein utilizing the principle of protein–dye bind-ing. Anal. Biochem. 72, 248–254.

Caradonna, S., Ladner, R., Hansbury, M., Kosciuk,M., Lynch, F., Muller, S., 1996. Affinity purifica-tion and comparative analysis of two distinct hu-man uracil-DNA glycosylases. Exp. Cell Res. 222,345–359.

Cone, R., Duncan, J., Hamilton, L., Friedberg, E.C.,1977. Partial purification and characterization of auracil DNA N-glycosidase from Bacillus subtilis.Biochemistry 16, 3194–3201.

Crosby, B., Prakash, L., Davis, H., Hinkle, D.C., 1981.Purification and characterization of a uracil-DNAglycosylase from the yeast, Saccharomyces cere-6isiae. Nucleic Acids Res. 9, 5797–5809.

Domena, J.D., Mosbaugh, D.W., 1985. Purification ofnuclear and mitochondrial uracil-DNA glycosylasefrom rat liver. Identification of two distinct subcel-lular forms. Biochemistry 24, 7320–7328.

Domena, J.D., Timmer, R.T., Dicharry, S.A., Mos-baugh, D.W., 1988. Purification and properties ofmitochondrial uracil-DNA glycosylase from ratliver. Biochemistry 27, 6742–6751.

Feller, G., Gerday, C., 1997. Psychrophilic enzymes:molecular basis of cold-adaptation. Cell Mol. LifeSci. 53, 830–841.

Focher, F., Verri, A., Spadari, S., Manservigi, R.,Gambino, J., Wright, G.E., 1993. Herpes simplexvirus type 1 uracil-DNA glycosylase: isolation andselective inhibition by novel uracil derivatives.Biochem. J. 292, 883–889.

Guyer, R.B., Nonnemaker, J.M., Deering, R.A., 1986.Uracil-DNA glycosylase activity from Dictyosteliumdiscoideum. Biochim. Biophys. Acta 868, 262–264.

Hamilton, R.W., Lloyd, R.S., 1989. Modulation of theDNA scanning activity of the Micrococcus luteusUV endonuclease. J. Biol. Chem. 264, 17422–17427.

Haushalter, K.A., Todd Stukenberg, M.W., Kirschner,M.W., Verdine, G.L., 1999. Identification of a newuracil-DNA glycosylase family by expressioncloning using synthetic inhibitors. Curr. Biol. 9,174–185.

Higley, M., Lloyd, R.S., 1993. Processivity of uracilDNA glycosylase. Mutat. Res. 294, 109–116.

Krokan, H.E., Standal, R., Slupphaug, G., 1997. DNAglycosylases in the base excision repair of DNA.Biochem. J. 325, 1–16.

Kubota, Y., Nash, R.A., Klungland, A., Schar, P.,Barnes, D.E., Lindahl, T., 1996. Reconstitution ofDNA base excision-repair with purified humanproteins: interaction between DNA polymerase betaand the XRCC1 protein. EMBO J. 15, 6662–6670.

Lindahl, T., 1974. An N-glycosidase from Escherichiacoli that releases free uracil from DNA containingde-aminated cytosine residues. Proc. Natl. Acad.Sci. USA 71, 3649–3653.

Lindahl, T., 1979. DNA glycosylases, endonucleases forapurinic/apyrimidinic sites, and base excision-re-pair. Prog. Nucleic Acid Res. Mol. Biol. 22, 135–192.

Lohman, T.M., 1986. Kinetics of protein–nucleic acidinteractions: use of salt effects to probe mechanismsof interaction. CRC Crit. Rev. Biochem. 19, 191–245.

Male, R., Lorens, J.B., Smalas, A.O., Torrissen, K.R.,1995. Molecular cloning and characterization ofanionic and cationic variants of trypsin from At-lantic salmon. Eur. J. Biochem. 232, 677–685.

Mol, C.D., Arvai, A.S., Slupphaug, G., Kavli, B.,Alseth, I., Krokan, H.E., Tainer, J.A., 1995. Crystalstructure and mutational analysis of human uracil-DNA glycosylase: structural basis for specificity andcatalysis. Cell 80, 869–878.

Muller, S.J., Caradonna, S., 1991. Isolation and charac-terization of a human cDNA encoding uracil-DNAglycosylase. Biochim. Biophys. Acta 1088, 197–207.

Muller-Weeks, S., Mastran, B., Caradonna, S., 1998.The nuclear isoform of the highly conserved humanuracil-DNA glycosylase is an Mr 36 000 phospho-protein. J. Biol. Chem. 273, 21909–21917.

Nilsen, H., Otterlei, M., Haug, T., Solum, K., Nagel-hus, T.A., Skorpen, F., Krokan, H.E., 1997. Nu-clear and mitochondrial uracil-DNA glycosylasesare generated by alternative splicing and transcrip-tion from different positions in the UNG gene.Nucleic Acids Res. 25, 750–755.

O. Lanes et al. / Comparati6e Biochemistry and Physiology, Part B 127 (2000) 399–410410

Otterlei, M., Haug, T., Nagelhus, T.A., Slupphaug,G., Lindmo, T., Krokan, H.E., 1998. Nuclear andmitochondrial splice forms of human uracil-DNAglycosylase contain a complex nuclear localisationsignal and a strong classical mitochondrial locali-sation signal, respectively. Nucleic Acids Res. 26,4611–4617.

Outzen, H., Berglund, G.I., Smalas, A.O., Willassen,N.P., 1996. Temperature- and pH sensitivity oftrypsins from Atlantic salmon (Salmo salar) incomparison with bovine and porcine trypsin.Comp. Biochem. Physiol. 115B, 33–45.

Parikh, S.S., Mol, C.D., Slupphaug, G., Bharati, S.,Krokan, H.E., Tainer, J.A., 1998. Base excision-repair initiation revealed by crystal structures andbinding kinetics of human uracil-DNA glycosylasewith DNA. EMBO J. 17, 5214–5226.

Purmal, A.A., Lampman, G.W., Pourmal, E.I.,Melamede, R.J., Wallace, S.S., Kow, Y.W., 1994.Uracil DNA N-glycosylase distributively interactswith duplex polynucleotides containing repeatingunits of either TGGCCAAGCU or TGGC-CAAGCTTGGCCAAGCU. J. Biol. Chem. 269,22046–22053.

Purnapatre, K., Varshney, U., 1998. Uracil-DNA gly-cosylase from Mycobacterium smegmatis and itsdistinct biochemical properties. Eur. J. Biochem.256, 580–588.

Ravishankar, R., Bidya Sagar, M., Roy, S., Purnapa-tre, K., Handa, P., Varshney, U., Vijayan, M.,1998. X-ray analysis of a complex of Escherichiacoli uracil DNA glycosylase (EcUDG) with aproteinaceous inhibitor. The structure elucidationof a prokaryotic UDG. Nucleic Acids Res. 26,4880–4887.

Sandigursky, M., Franklin, W.A., 1999. Thermostableuracil-DNA glycosylase from Thermotoga mar-itima, a member of a novel class of DNA repairenzymes. Curr. Biol. 9, 531–534.

Savva, R., McAuley Hecht, K., Brown, T., Pearl, L.,1995. The structural basis of specific base-excision

repair by uracil-DNA glycosylase. Nature 373,487–493.

Savva, R., Pearl, L.H., 1995. Nucleotide mimicry inthe crystal structure of the uracil-DNA glycosy-lase–uracil glycosylase inhibitor protein complex.Nat. Struct. Biol. 2, 752–757.

Slupphaug, G., Eftedal, I., Kavli, B., Bharati, S.,Helle, N.M., Haug, T., Levine, D.W., Krokan,H.E., 1995. Properties of a recombinant humanuracil-DNA glycosylase from the UNG gene andevidence that UNG encodes the major uracil-DNA glycosylase. Biochemistry 34, 128–138.

Sobek, H., Schmidt, M., Frey, B., Kaluza, K., 1996.Heat-labile uracil-DNA glycosylase: purificationand characterization. FEBS Lett. 388, 1–4.

Talpaert-Borle, M., Clerici, L., Campagnari, F., 1979.Isolation and characterization of a uracil-DNAglycosylase from calf-thymus. J. Biol. Chem. 254,6387–6391.

von Hippel, P.H., Berg, O.G., 1989. Facilitated targetlocation in biological systems. J. Biol. Chem. 264,675–678.

Wang, K.T., Wang, I.S., 1967. Polyamide-layer chro-matography of nucleic acid bases and nucleosides.Biochim. Biophys. Acta 142, 280–281.

Williams, M.V., Pollack, J.D., 1990. A mollicute (my-coplasma) DNA repair enzyme: purification andcharacterization of uracil-DNA glycosylase. J.Bacteriol. 172, 2979–2985.

Winters, T.A., Williams, M.V., 1993. Purification andcharacterization of the herpes simplex virus type-2-encoded uracil-DNA glycosylase. Virology 195,315–326.

Wittwer, C.U., Bauw, G., Krokan, H.E., 1989. Purifi-cation and determination of the NH2-terminalaminoacid sequence of uracil-DNA glycosylasefrom human placenta. Biochemistry 28, 780–784.

Wittwer, C.U., Krokan, H., 1985. Uracil-DNA glyco-sylase in HeLa S3 cells: interconvertibility of 50and 20 kDa forms and similarity of the nuclearand mitochondrial form of the enzyme. Biochim.Biophys. Acta 832, 308–318.

.