The generation of DNA single-strand breaks during the reduction of chromate by ascorbic acid and/or glutathione in vitro.

The Generation of

DNA

Single-Strand Breaks

During

the

Reduction

of Chromate

by

Ascorbic

Acid and/or Glutathione

In Vitro

Andreas

Kortenkamp1

and Paul

O'Brien2

1Department

of Toxicology,

School

of Pharmacy, University

of

London;

2Department

of Chemistry,

Queen Mary and

Westfield College,

University

of

London, London, United Kingdom

The potential roleof iron and copper and the involvement of hydroxyl radicalsinthe DNA cleavage caused by chromate and glutathione(GSH)has been investigated. Wehave alsostudied theabilityof chromate, on reaction with ascorbateaswell asinmixedsolutions of ascorbate and GSH, to cause DNA strand breaks. Inbothfully demetalated and conventional (i.e., metal contaminated) systems, chromate and GSH induced similar num-bersof DNA strand breaks. This observation suggests thattracesof ironorcoppercontaminating the reaction mixtures donotplayamajorrolein the DNA cleavage causedby chromate and GSH. A series of hydroxyl radical scavengers exhibitedaprotective influenceonthe induction of DNA strand breaks. However, glucose and sucrose, both strong hydroxyl radical scavengers, showed no concentration-dependent inhibition of DNA cleavage. Competition kinetics studies yielded apparentrate constantsthatwere notconsistent with hydroxyl radicals being thespecies responsible forDNAstrandbreaks.Ascorbateincombination with chromatewasalsofoundtoinduce strand breaksinDNA; this damage could be attributedto reactiveintermediates generated during the reduction. When mixed systems of ascorbate and GSH in the presence of chromate were investigated, therewereclearly interactions between thetwo reductants.-Environ HealthPerspect102(Suppl3):237-241 (1994)

Key words:chromium(VI), DNAstrand breaks, ascorbate, glutathione, chromium(VI) reduction, competition kinetics

Introduction

Despite considerable research effort during the lasttenyears, the mechanismsby which

the known humancarcinogen chromatecan

causeDNAimpairment and mutations still

remain obscure. Chromate is rapidly taken

upby mammalian cells and readily reduced

by cellular constituents. It is nowgenerally

accepted that the intracellular reduction of the metal compound is crucial for the induction of DNA lesions(1). On the basis of kinetic studies in vitro, it has been

sug-gested that glutathione (GSH), ascorbic acid (AsA), and cysteine are the most

important nonenzymatic cellular

con-stituents for the reduction of chromate in the cytoplasm (2). However, the relative contribution of these various reductants to

the generation of DNA lesions is far from clear. Furthermore, it has as yet not been

possibleto identify the ultimate carcinogen

deriving from chromium(VI) upon

reduc-tion.

While the reduction of chromate by GSHhas beenextensively studiedinmodel

Thispaper waspresentedattheSecond International MeetingonMolecular Mechanismsof MetalToxicity andCarcinogenicity held 10-17January 1993 in Madonna diCampiglio, Italy.

Address correspondence to Dr. Andreas Kortenkamp, Department ofToxicology, School of

Pharmacy, University of London, 29-39, Brunswick,

Square, London WC1N 1AX, United Kingdom.

Telephone44 717535908.FAX 44 71 753 5811.

systems in vitro (3) thesignificance ofAsA as an intracellular reductant of chromate has, as yet, received comparatively little attention. Recently, Suzuki and Fukuda (4) reported that in the bronchoalveolar lining fluid and the lung parenchyma of ratsAsA is morereactive thanGSH in the reduction of chromate. In subsequent in vitrostudies Suzuki (5) observed a syner-gisticeffect ofAsAandGSH ontherateof reduction of chromate. Later, Standeven and Wetterhahn (6) provided evidence suggesting thatAsA may be the principal

nonenzymaticreductantof chromatein rat liver andkidney. Inview ofthese observa-tions it is now necessary to consider the

possibility that reduction by ascorbate mediates thegenerationofDNAlesions.

In the present article, we report some observations on the induction ofDNA strand breaks by chromate and AsA. We also studied the generation ofDNA cleav-age in the presence ofAsA, GSH, and chromate. Initial experiments in our labo-ratory revealed that AsA itself, in the absence ofchromate, can cause DNA strand breaks. Theeffectisduetothe gen-eration ofreactive oxygen species during the autoxidation ofAsA(i.e., the oxidation ofAsA in aqueous solution by dissolved

dioxygen) (7). Thereaction iscatalyzed by

tracesof metalions, namely copper(II) and iron(III). Ascorbate, in the absence of

cat-alytic metals, is stableeven in air-saturated

solutionsofneutralpH(8).

Inthelight ofa controversy concerning theability of chromate andGSH toinduce DNAstrand breaksinFe-andCu-free sys-tems (9) it has also become necessary to reassess the roleoftraces of catalytic metal ions in the chromate/GSH reaction.While we have observed the induction ofsingle strand breaksin PM2DNAwith chromate and GSH in both phosphate and HEPES

buffer(10),Aiyaretal.

(1])

usingpBR322 DNA, have detected very little, ifany, DNA cleavage in a Tris-buffered system. Theyremovedcontaminating iron from all solutions while we used all reagents with-outfurtherpurification.The resultsof sub-sequent studies carried out in our laboratory (12) inanattempt toclarifythe situationsuggestedthatFecould be impor-tant in these reactions. Moreover, we observed aprotective influenceofaseries ofhydroxyl radical scavengers on the gen-eration of strand breaks by chromate and GSH. Taken together, these observations led us to suggest hydroxyl radicals as the ultimate DNA-cleaving species. However, in an article primarily concerned with

controls 0.2 0.3 0.4 5 5 5

controls 0.2 0.3 0.4

5 5 5

Figure1. DNAstrandbreaks induced by chromateandGSHindemetalatedandconventionalsystems.II:nicked,c

circularform of PM2DNA,I:intactsupercoil.

distinct possibility, but moreexperimental

workis needed before firm conclusionscan

be drawn. Here, we communicate our recentobservations that might helpto

clar-ify

some ofthe earlier, apparently

contra-dictory,results.

Materials and

Methods

All reagents usedwere ofanalytical grade.

Glutathione, ascorbic acid, diethylendi-aminopentaacetic acid (DTPA) and ethid-ium bromide were from Sigma (Poole,

Dorset, UK); bromophenol blue and

agarose were from Biorad (Hemel

Hempstead, UK). Potassium chromate (K2CrO4) was purchased from

BDH-Merck (Poole, UK) and PM2 DNA was

obtained from Dr. IreneWitte, University ofOldenburg FRG, who isolated it as

describedbyEspejoetal.(14).

PM2 DNA(8pg/ml),inafinalvolume

of 20 pl, wastreated withvarying

concen-trations ofK2CrO4, GSH, and/orAsA in phosphate buffer, 0.1 M, pH 7.0, for at

least 3 hrat room temperature (close to

22°C). Where indicated ("demetalated

sys-tem"), traces ofFeand Cu were removed

from buffer solutions by stirring overnight with Chelex 100 resin. To remove metal traces fromDNA, itwas dialyzed (24 hr)

against 1 mMDTPA(dissolvedin demeta-latedphosphate buffer).

The successful removal ofcatalytic

met-als from buffer solutions was assessed by

using the ascorbate method of Buettner (8). The method involves monitoring the autooxidation of ascorbate at 265 nm in

the presence oftest solution. After tr mentwith Chelexresin, ascorbatewas

ble (i.e., loss of absorbance <0.5%) fc least30min,indicating that metalion(

taminationwasminimal. DNAwascon

ered free ofmetals when ascorbati concentrationsupto 1 mM(in theabst

ofchromate) did not induce any str

breaks.

Electrophoresis was carried ou

describedpreviously(10,12).

Results

and

Discussion

Chromate and GlutathioneCause DNAStrand Breaks in Iron(III)-an

Copper(II)-fiee

Systems

Inanattempttoprobethe involvemen

Fe inthegenerationofDNAstrand br by chromate and GSH, we studied

reaction in conventional (i.e., metal-co minated) and exhaustively demetala solutions. In both systems chromate

GSH induced similar numbersof sin strand breaks in isolated PM2 D (Figure 1). Thelevel of cleavage incre;

with risingconcentrations ofchromat

phosphate buffer,0.1 M, pH 7.0,cont

ing 5 mMGSH.

These findings clearly indicate

traces of Fe or Cu contaminating

buffers or bound to DNA do not plh

majorrole in the DNAcleavagecauses

chromate and GSH. Our results ar

agreement with a recent report by

Wetterhahn group that strand breaks

formed in a demetalated system (13)

observation that contradicts earlier reports

by

thesamegroup

(11,15).

Thedemonstrationthat thereisno spe-cific need for Fe or Cu in this reaction makes itnecessary to reassess ourearlier

suggestion

(12)

ofa

thiyl

radical/superox-ide

anion/hydrogen

peroxide-mediated

pathway

leading

to the

generation

of

hydroxyl

radicals as the ultimate

DNA-cleaving

agent.This suggestion was based

on observations ofa

concentration-depen-dent

depression

of the level ofDNA cleav-age

by

benzoate andformate upon addition

to chromate and GSH.

Consequently,

we

extended these studies and explored the influence ofawider range ofhydroxyl radi-cal scavengers.

The results confirmourearlier

findings

(12).

Weobserveda

concentration-depen-dent inhibition of DNA

degradation

with sodium formate and sodium benzoate.

Dimethylsulfoxide(DMSO)and mannitol losed also showed a

concentration-dependent

inhibitoryeffect. To our surprise, glucose,

astrong

hydroxyl

radical scavenger, exhib-reat- ited noprotective influence on the

induc-sta- tion of strand breaksby chromate/GSH at

ir at concentrations of up to 10 mM. Sucrose con- showed effects very similartothose of

glu-isid- cose(datanotshown).Table 1 summarizes

e at typical resultsobservedconventional,

non-ence demetalated systems (for results with -and DMSO, see P. O'Brien and A.

Korten-kamp, in this issue). Essentially the same

t as results were obtained in systems with

exhaustivelydemetalated DNA and buffer solutions(datanot

shown).

Competition

Kinetics

Id

The somewhat

unexpected

results with

glu-coseandsucrosepromptedus toprobe fur-ther the involvement ofhydroxyl radicals

Itof in thechromate/GSH system byusing the eaks method of competition kinetics. This the method has beenwidelyused in the deter-Fnta- mination ofrateconstants ofradicals. It ated utilizes the fact that anyhydroxyl radicals and (OH) generated in the present reaction

gle- mixturesshouldreactwithDNA toforma NA strand break. Any other molecule S added ased that iscapableof reacting with -OHshould

e in

tain-that the

ay a Iby e in

the

;are , an

CrO42-(gmM)

L-AsA(mM) 01 1001 2001 5001

11

Figure 2. DNA strand breaks caused by chromate and ascorbate. Il:nicked, closed circular form of PM2 DNA,I:

intact supercoil.

Demetallated system

CrO42,

(mM) 0 GSH (mM) 5

Conventional system CrO42-(mM) 0 GSH (mM) 5

0.1 5

0.1

DNA CLEAVAGE INDUCED BYCHROMATE, ASCORBATE, AND GLUTATHIONE

Table 1. Inhibitoryeffect of formate, benzoate, mannitol, and glucose on DNA strand breaks generated by chromate (0.2 mM) and glutathione (5mM).

Number of strand breaks

Scavenger, mM Formate Benzoate Mannitol Glucose

0 0.70 0.66 0.67 0.83

1.0 0.53 0.20 0.60 0.67

2.5 0.36 0.16 0.58 0.85

5.0 0.33 0.11 0.48 0.60

7.5 0.22 0.13 0.42 0.84

10 0.11 0.04 0.40 0.26

compete with DNA for -OH, thereby reducingthe level ofDNAcleavage. The

extent of competition, and hence the

degreeof protection against strandscission will depend onthe rate constant for reac-tion ofS with -OH and its concentration

relative to DNA. Since the rate constants for reactionofall the above scavengers with ,OH areknown (16), analysis ofthe data

obtained (Table 1) should allow calcula-tion ofa consistentvalue for the rate of reaction ofDNAwith OH. Furthermore,

the rate constant calculated should agree with the literature value 0.8x109 M-1 s-I

(17) for the rate ofreaction of-OH with

DNA-P.

The changes in the number of

chro-mate/GSH-induced DNAstrand breaks in the presence of varying concentrations of

*OHscavengers can beexpressed by (NO/N)-l =kS[S]/kDNA[DNA]

in which

No

and N are the numbers of strand breaksin the absenceorpresence of

scavenger, respectively; [S] and [DNA] are theconcentrationofscavengerandDNA-P,

respectively; andks andkDNAarethe

bimol-ecular rate constants for the reactions of

OH with scavenger and DNA,

respec-tively. Thus,aplot of

(No/N)-l

versusthe

concentration ofscavengershould be linear withaninterceptof 1 and giveastheslope theratiokS/kDNA[DNA]. The rate constant

forthe reaction of-OH with PM2 DNA can then be calculated from known values CrO42- (jM) 50 50 50 L-AsA

(gM)

25 50 75

of ks. As shown inTable 2, consistent val-uesforkDNA are notobtained. Furthermore, far higher concentrations of scavengers are needed to inhibit the DNA cleavage

inducedby chromate and GSH than would be expected for areaction involving free

hydroxyl radicals. Consequently, the appar-entvaluesof

kDNA

exceeddiffusion limit.

Weconclude thatsimpleFenton chem-istrydoes not explain the abilityof chro-mate and GSH to induce single-strand breaks in DNA and that Fe or Cu is not necessaryforthis process. At this stage, we can only speculate which species are responsible for the DNA damage.

Hypervalent Cr species known tobe gener-atedduring reduction byGSHinclude sev-eral chromium(V) [Cr(V)] complexes. Both Lay and co-workers (18) and we (10) found evidence that Cr (V) com-plexes can cause DNA single-strand breaks. Thus, mechanisms involving the direct reaction ofa high oxidation state complex with DNA, as suggested byLay, areworth seriousconsideration. Furthermore, thereis now evidence for some accumulation of Cr(IV) during the reaction ofchromate withGSH (19), andchromium(IV) com-plexes are known to cause C-C bond scission(20).Afurtherpossibility includes

the reaction ofhypervalentCrwith

molec-ular oxygen to form some reactive DNA

cleavingspecies(P.O'BrienandA.

Korten-kamp, thisissue).

Another factor, which might help explain the conflicting results concerning controls 50 50 0

100 200 100

Figure3. The influence ofvaryingamounts

of ascorbateonthe DNAcleavageinduced in the presence ofconstant amountsof chromate. Il:nicked,closed circular

the ability of chromate and GSH to cause DNAcleavage, is the nature of the buffer-ing agent. Weshowed earlier (10) that the

buffering agent present in the incubation mixtures exhibits a strong influence on the

level of strand breaks induced by chromate

and GSH. When HEPES bufferwas used instead of phosphate buffer, higher chro-mate concentrations were required to pro-duce the same level of DNAstrand breaks. The useofTris seems tobe problematic; it

interferes with the reduction of chromate by AsA (21). We are exploring the effect of

bufferingagentsingreaterdetail.

Induction

of DNA

Stand Bks by

Chromate

in

Combinaon

with

Ascorbate

and

Glutathione

Chromate and AsA,inthe absenceoftraces ofcatalyticmetal ions, causedDNA single-strand breaks. Figure2showsthe resultsof

an experiment inphosphate-buffered

solu-tion containing 1 mM AsA. With rising concentrations ofchromate, the level of

strand breaks increased and reached a

plateau as the ratio chromate:AsA approached 1. In the presence of100 1M chromate, the numberof strand breakswas stillsignificantly higher thanintheabsence

of chromate. These results indicate that chromate, uponreduction byAsA, hasthe

potential to generate species that cleave DNAand lendfurthersupport tothe idea

that the reduction of chromateis acrucial step in the generation ofDNAlesions. It remains tobe seen whether this alsoapplies

to the induction ofDNA-protein cross-links.

Suzuki has reported that the rate of

reduction of chromate byAsA increases with rising concentrations ofAsA in vitro

(5). We conducted experiments using

chromate andAsA at concentrations simi-lar to those chosenbySuzukiandobserved adecrease in the level ofDNA cleavage

with increasing concentrations ofAsA (Figure 3). These results could be indica-tive ofan inverse relationship between the rate of reduction and the generation of

DNAdamage: the slower thereduction,

the greater thedamage. Alternatively, the

effect could be dueto anefficient scaveng-ingofthe cleavingspecies by an excess of AsA. The observation ofaprotectiveeffect

with increasing amounts ofAsAclearly indicates that the DNA damage is caused

byaspeciesgeneratedduring the reduction ofchromatebyAsAratherthanbythefinal product ofthereaction.Ifthe finalproduct

wereresponsible for theDNAimpairment, risingamounts ofAsAshould haveleftthe levelofstrand breaksunchanged.

I-1-1-:.!::-CrO42-

(M) 0 100 100 100 200 200 200 L-AsA (mM) 1 1 1 0 1 0 1

GSH (mM) 0 0 1 1 0 1 1

II

CrO42-

(xM)

0 500 500 500 L-AsA (mM) 1 1 0 1 GSH (mM)

~~~~~~~~~~~~~~~~~...

1 0 1 1

Figure4. DNAstrand breaksgeneratedbychromate and mixed solutions ofascorbate andglutathione.11:nicked,closed circular form of PM2DNA,1: intactsupercoil.

Westartedexploring the DNA damag-ingeffects of chromate in mixed solutions of AsA and GSH. Glutathione, when

administered separately, appeared to be more effective than AsA in generating DNA strandbreaks in the presenceof iden-tical amounts of chromate. As shown in Figure 4, samples containing chromate (100, 200, 500 pM), but with GSH (1 mM) instead of AsA, yielded slightly higher levels ofstrand breaks. The effect of

Table 2. Values for theapparent rate constant

k,..

for the reaction of hydroxyl radical with DNA calculated from the results shown in Table1.

Scavenger

k,NA

ks,

M_'s-l M_'s-' Formate 4.9x1Ol6 3.2 x109 Benzoate 3.2x10'5 5.9x109 Mannitol 4.8 x1015 1.0x10O Values fork0,, shown compareto aliterature value of 0.8x109M-'s-1 ( 17).The otherconstants aretakenfrom Dorfman and Adams(16).

mixed systems on the induction ofDNA strand breaks varied with theconcentration of chromateinthesamples.Uponaddition

of AsA (1 mM) to solutions containing GSH(1 mM) and relatively small amounts of chromate (100 pM, 200

pM)

the num-bersof strand breaks returned to the levels observed with chromate and AsA alone (Figure4). Withrelatively high amounts of chromate (500

PM)

however, the combina-tion ofAsA and GSH exhibited a slightly protective influence on the generationof strand breaks; the cleavage observed was less pronounced than with AsA and chro-matealone.

Conclusions

Ourefforts to definewhich DNA

damag-ing species aregenerated upon reductionof

chromateby GSH now lead us to conclude that OH are not the cause for the strand scission observed in PM2 DNA. The processdoes notdepend on the presence of Fe or Cu,and simple Fenton chemistry can not explain the experimental results. It seems to usthat the direct interaction of some hypervalent Crspecies with DNA, probably involving the participation of

molecular oxygen, isworth serious consid-eration.

We have demonstrated that the reduc-tionof chromatebyAsAleads to the gener-ation of species that are able to cleave

DNA. The finalproduct of the reaction is not involved in the induction of these effects. The ultimately active species gener-ated in such reaction mixtures are not yet known; however it is likely that they are

different from the species present in solu-tions containingchromate and GSH.

The study of systems in mixed solu-tions ofAsAand GSH has yielded results that areindicative ofaninverserelationship between rate ofreduction and generation

ofDNAdamage.Another important factor is theinfluence ofscavenging propertiesof AsA or GSH when present in excess. More work is needed to probe this suggestion

further.

Although the levels of AsA and GSH in the human lung are not precisely known, available data for rats suggest that the cyto-plasmic levels of both reductants are

equimolar and fall in the millimolar range

(4). Considerableamountsof Cr can accu-mulate inside cells after exposure to chro-mate. Sehlmeyer et al. (22) have shown that after the exposure ofV79-cells to 10

jiM

chromate the total intracellular, cytosolic, Cr concentrations can be in the millimolar range. Under similar conditions Crlevels innucleican beevenhigherthan in thecytosol. The concentrationsof chro-mate, AsA, and GSH used in our experi-ments are similar to levels that might be attained within cells; our observations are

ofrelevance to the in vivo situation. However, it is too early forfirm conclu-sions because the relative contribution of the extracellular reduction ofchromate in lung tissueand theinfluence of

intracellu-lar compartmentalization is not known. Suzuki andFukuda (4) havereported that noothereffectivereductant apartfromAsA exists in the alveolar-lining fluid ofrats.

Thus, in the earlierperiods ofexposure to

chromate, extracellular reduction may be the overalldetermining step. After

deple-tion ofextracellular AsA, intracellular

reductions may become more important. However, more information is needed to delineate between thesetworoutesof chro-matereduction.

REFERENCES

1. Bianchi V, Levis AG. Recent advances inchromium genotoxicity. Toxicol Environ Chem 15:1-24 (1987).

2. Connett P, WetterhahnKE. Metabolism of the carcinogen chro-mate by cellular constituents. Struct Bonding54:93-124 (1983). 3. De Flora S, Wetterhahn KE. Mechanisms ofchromium

metabo-lismand genotoxicity.Life Chem Rep 7:169-244 (1989).

4. SuzukiY,FukudaK.Reduction of hexavalent chromiumby ascor-bic acid and glutathione with special reference to the rat lung. Arch Toxicol 64:169(1990).

DNACLEAVAGE INDUCEDBYCHROMATE, ASCORBATE,ANDGLUTATHIONE

of chromium(VI) in rat liver and kidney ultrafiltrates. Carcinogenesis 12:1733-1737(1991).

7. DaviesMB. Reactions of L-ascorbic acid with transition metal complexes.Polyhedron11:285-321 (1992).

8. Buettner GR.In the absenceof catalytic metals ascorbate doesnot

autoxidiseatpH 7ascorbateas a testforcatalytic metals.JBiochem BiophysMeth 16:27-40 (1988).

9. Standeven A, Wetterhahn KE. Is there arole for reactive oxygen species inthe mechanism of chromium(VI) carcinogenesis? Chem Res Toxicol 4:616-625(1991).

10. KortenKamp A,Ozolins Z Beyersmann D, O'Brien P. Generation ofPM2 DNAbreaksinthecourseof reduction of chromium(VI) by glutathione. Mutat Res 216:19-26 (1989).

11. Aiyar J, Borges KM, Floyd RA. Wetterhahn KE. Role of chromium(V), glutathionyl and hydroxyl radical intermediates in

chromium(VI)-induced DNAdamage. Toxicol Environ Chem 22:135-148(1989).

12. Kortenkamp A, Oetken G, Beyersmann D. The DNA cleavage induced by a chromium(V) complex and by chromate and glu-tathione is mediated by activated oxygen species. Mutat Res 232:155-161 (1990).

13. Borges KM,WetterhahnKE.Chromium bound to DNA alters cle-vage by restriction endonuclease. ChemResToxicol 4:638-641 (1991).

14. Espejo RF,Canelo ES,Sinsheimer RL. DNA of the bacteriophage

PM2: aclosed circular double-stranded molecule Proc NatlAcad Sci USA63:1164-1168(1969).

15. Borges KM, Boswell JS, LiebrossWetterhahn KE. Activation of chromium(VI) by thiols results in chromium(V) formation, chromium binding to DNAand altered DNA conformation. Carcinogenesis 12:551-561 (1991)

16. Dorfman LM, Adams GE. Reactivity of the hydroxyl radical in aqueous solution. Washington:US Government Printing Office,1972.

17. Scholes G.The radiation chemistry of pyrimidines, purins and related substances. Photochem Photobiol Nucleic Acids 1:521-527 (1976).

18. Farrell RP, Judd RJ, Lay PA, Dixon NE, Baker RSU, Bonnin AM. Chromium(V) induced cleavageof DNA: arechromium(V)

com-plexes the active carcinogens inchromium(VI)-inducedcancers?

ChemResToxicol2:227-230(1989).

19. Bose RN,Moghaddas S, Gelerinter E. Long-lived chromium(IV) and chromium(V) metabolites in the chromium(VI)-glutathione reaction: NMR, ESR, HPLC, and kinetic characterization. Inorg Chem31:1987-1994(1992).

20. Srinivasan VS. Chromium(IV)-induced carbon-carbon bond

scis-sion.InorgChem 21:4328-4330 (1982).

21. Goodgame DML, JoyAM. EPRstudy of the Cr(V)and radical species producedinthe reduction ofCr(VI) by ascorbate. Inorg ChimActa135:115-118 (1987).