Complexing of Amino Acids to DNA byChromate in Intact CellsVictoria Voitkun, Anatoly Zhitkovich, and Max CostaNelson Institute of Environmental Medicine, New York University Medical Center, New York, New York

Using o-pthaldialdehyde (OPT) fluorescence, the amino acids associated with DNA were studied following exposure of intact Chinese hamster ovarycells to chromate. Rigorous extraction with EDTA, acid, or base was required to release the amino acids cross-linked to the DNA isolated from con-trol or chromate-treated cells by standard procedures (i.e., proteinase K, phenol, etc.). Amino acids resisting extraction from DNA were not studiedsince analysis was limited to those that could be released by these procedures. There was a chromate dose-dependent increase in amino acidscomplexed with the DNA that could be released by EDTA, acid, and base, and these amino acids were separated by HPLC and identified.Substantial increases in cysteine, glutamine, glutamic acid, histidine, threonine, and tyrosine were found as a function of increasing concentrationsof chromate. There was also a time-dependent increase in complexing of these amino acids to the DNA by chromate. The amino acids found com-plexed to DNA in intact cells by chromate were thought to originate from reactions of free amino acids or small peptides with the DNA rather thanbeing proteolytic products derived from larger proteins that were cross-linked to the DNA. This was supported by a number of experiments: a) freeamino acids or bovine serum albumin (BSA) were cross-linked by chromium to DNA in vitro and the DNA was isolated by standard procedures. WithBSA, few amino acids were found cross-linked to the DNA, but with free amino acids, numerous amino acids were associated with DNA; b) whenradiolabelled threonine complexed to the DNA was examined in the absence and presence of a protein synthesis inhibitor, there was substantialstimulation of the threonine complex to DNA by chromate when protein synthesis was inhibited with cyclohexamide; c) there were substantialincreases in amino acids associated with DNA isolated without protease. Not only does the cross-linking of amino acids to DNA represent a newtype of lesion to study in intact cells but it may also be a useful biomarker of human exposure to cross-linking agents. - Environ Health Perspect102(Suppl 3):251-255 (1994).

Key words: cysteine, histidine, glutamine, HPLC, DNA extraction

Introduction

Even the most highly purified DNA hasresidual amino acids associated with it thatcan be detected by extraction with strong

acids, bases, or other rigorous extractionprocedures (1,2). These residual aminoacids are thought to have been part of theoriginal chromatin-scaffold proteins or

other tightly bound chromatin proteins(2). A number of proteins are covalentlycross-linked to the DNA and these pro-

teins are cross-linked through specificamino acid residues, which should bedetected when amino acid-DNA cross-

links are examined. An example of one

such protein is topoisomerase (3).Previous studies in our laboratory

have utilized DNA-protein cross-links as

an indicator of exposure to various chem-

This paper was presented at the SecondInternational Meeting on Molecular Mechanisms ofMetal Toxicity and Carcinogenicity held 10-17January 1993 in Madonna di Campiglio, Italy.

This work was supported by grants ES 04895, ES04715, ES 05512 and ES 00260 from the NationalInstitute of Environmental Health Sciences, and bygrant CA 13343, Kaplan Cancer Center.

Address correspondence to Dr. Max Costa,Nelson Institute of Environmental Medicine, NewYork University Medical Center, 550 First Avenue,New York, NY 10016. Telephone (914) 351-2369/(212)263-5280. Fax (914) 351-2118/(212) 263-5019.

ical agents, such as chromate, nickel,formaldehyde, UV light, etc. (4,5). Wehave developed assays for detection ofDNA-protein complexes (4,6) and pro-teins associated with DNA (5) have beenstudied. This type of lesion has beenused as a human biomarker of exposureto cross-linking agents (7). However, theability to extract DNA in the presence ofintact proteins is limited, and mostknown procedures are generally flawedwith background problems. DNA iso-lated by standard proteinaseK/phenol/chloroform extraction proce-dures, however, yielded highly purifiedDNA with a minimal of background pro-teins. This DNA also had a minimum ofassociated background amino acids (2,8).

In the present study, we haveattempted to examine whether residualamino acids associated with DNA iso-lated by the proteinase K method couldbe used to detect exposure to cross-link-ing agents. There are substantialincreases in residual amino acids associ-ated with DNA found in intact cellstreated with a cross-linking agent such aschromate. Direct examination of aminoacids by fluorescent derivatives andHPLC separation yielded an accurateassessment of the actual amino acidcrosslinked to the DNA.

Materials and Methods

Chinese hamster ovary cells (CHO) were

cultured in a-MEM at 37°C, supple-mented with 10% fetal bovine serum andantibiotics in an atmosphere containing

95% air and 5% CO2. Cells were culti-vated in 150 mm dishes to approximate30% confluency. Subsequently, they were

treated with various concentrations ofK2CrO4 ranging from 5 to 100 pM for 24hr in complete media. Cells were subse-quently rinsed with phosphate bufferedsaline, scraped, collected by centrifugationand lysed in 0.5% SDS solution containing100 mM sodium chloride and 10 mMHEPES pH 8.0. Cells were concentrated to

about 2 to 3 million per ml. Cells were

passed through a 25-gauge needle and 25pg/ml of RNase was added and thehomogenate was incubated for 30 min at

37°C. Subsequently, 300 pg/ml of pro-

teinase K was added and the cells were

incubated for 3.5 hr at 50°C. Thishomogenate was extracted three times withphenol/chloroform/isomyl alcohol at a

ratio of 25:24:1 using equal volumes. Theaqueous phase was subsequently extractedwith chloroform/isomyl alcohol at 24:1using equal volumes. The aqueous phase,which contained the DNA, was made 0.1M in NaCl two volumes of ethanol were

Environmental Health Perspectives 251

VOITKUN ETAL.

control 60.BJM oC _ I00Jum Cr rescent dye o-pthaldialdehyde (OPT) (9).Sixty milligrams ofOPT was dissolved in 1ml of ethanol, and 40 pl of 2-mercap-toethanol and 1 ml of borate buffer, pH9.5 were added. Fifty microliters of aminoacid solution was reacted with 5 pl of thisOPT reagent for exactly 1 min. The reac-tion was stopped with 5 1pl of 50 mM ofNaC2H302 pH 4.5. Total fluorescence waseither determined directly using a Perkin-

IOTA Acid MaOH Elmer fluorescent spectrophotometer

Effect of chromate on complexing of amino (Norwalk, CT) or DNA samples were pre-NA. CHO cells were treated with 50 pM, 100 cipitated with ethanol, the supernatant washout K CrO4 for 20 hr in complete a-minimal dried and the residue dissolved in 15 idrnedia fa-MEM). DNA was isolated by pro- was injected in a HPLC column. Amino

teinase K/phenol/chloroform extraction as outlined inMaterials and Methods. The DNA was treated with 20mM EDTA for 24 hr at 37°C, 0.5 N H2SO4 for 20 min at70°C or 0.2 N NaOH for 2 hr at 37°C. Samples of theseextracts were used directly to measure fluorescence witho-phthaldldehyde dye as outlined in Materials andMethods.

added and the aqueous phase was allowedto stand overnight at -20°C. DNA was col-lected by centrifugation for 20 min at12,000g and the pellet was rinsed in 70%ethanol. The final DNA pellet was dis-solved in 20 mM of HEPES buffer, pH8.0. DNA was subsequently extracted witheither EDTA, NaOH, or sulfuric acid asdescribed in the figure legends.

Amino acids associated with the DNAwere quantitated by reaction with the fluo-

1 23 5

acids were separated on AdsorbosphereOPA-US column (Altech, Dierfield, IL) ina NaC2H302 buffer, pH 5.7, containing3% 2-propanol against 100% methanolcontaining 1.5% 2-propanol for 40 min ata flow rate of 1.5 ml/min (grad: 0-80%methanol for 40 min).

A special procedure was utilized todetermine cysteine content associated withthe DNA. DNA purified from cells wasused for alkylation of cysteine residues toconvert them to S-carboxymethyl cysteine(CMC). Dithiothreitol was added to theDNA to a final concentration of 2 mM.Nitrogen was blown over the surface of thesolution and the reaction mixture was incu-bated at 370C for 1 hr. Then 4 mMiodacetic acid and 0.2 N NaOH wereadded to the solution and nitrogen was

1 - Asp2- Cys3 - Glu4 - Asn5 - Ser

6- Gln7 - Thr8 - His9-Ala10 - Arg

11 -Tyr12 - Met13 - Val14 - Phe15- lleu16 - Leu17 - Lys

8

10

11 1316

17

again blown over the surface of the solu-tion. Release of amino acids from the DNAand alkylation of released cysteine residueswere allowed to proceed in the dark at37°C for 2 hr. DTT and iodoacetate werenot able to agree to release cysteine residuesfrom the DNA. Treatment of the DNAwith NaOH was required to render thiolgroups of cysteine accessible to the reagentsfor alkylation that allowed us to detect cys-teine residues bound to DNA followingCr-treatment ofCHO cells.

Experiments to determine complexingof proteins or amino acids with DNA invitro were performed using calf thymusDNA, standard amino acids, and bovineserum albumin. One hundred microgramsper milliliter DNA was incubated with 0.5mg/ml BSA or 50 ,uM each of amino acidsin the presence or absence of 100 gMCrCI3 for 24 hr at 37°C as indicated inFigure 6. DNA was purified with the pro-teinase K/phenol/chloroform procedures asdescribed for CHO cells, except that pro-teinase K was not applied to the samples ofDNA incubated with amino acids.

ResultsFigure 1 shows the total fluorescent aminoacids that could be extracted from theDNA in untreated or cells treated with 50or 100 pM chromate. Note the increase offluorescent derivatives of DNA associatedamino acids that was dose-dependent fol-lowing treatment of cells with increasingconcentrations of chromate. The aminoacids associated with DNA were separatedby HPLC to resolve the individual aminoacids contributing to the overall fluores-cence.

Figure 2 shows the separation of aminoacids present in control or chromate-treated DNA. A number of NaOH-extractable amino acids were increased intheir cross-linking compared with controlDNA. Particularly striking was the increasein carboxymethylcystiene, glutamic acid,glutamine, and histidine.

Figures 3 and 4 quantitate and comparea number of the amino acid peaks thatwere resolved by HPLC. Note that theconsistency of the method is illustrated bythe small standard-error bars and note thatNaOH was the most effective agent inreleasing amino acids. Figure 3 is a quanti-tative comparison from both control andchromate-treated cells, while Figure 4 illus-trates the cross-linked amino acids relativeto control. Glutamic acid and glutathionewere more strikingly associated with DNAafter chromate treatment. Other aminoacids were also increased in their associa-

Environmental Health Perspectives

4

c

0,02

1

-5

Figure 1.acids to DIpM, or witlessential r

Figure 2 HPLC profiles showing the separation of individual amino acids complexed to the DNA. DNA from control(dash lane) or CHO cells (solid lane) exposed to 100 pM, potassium chromate was treated with sodium hydroxide as out-lined in the legend of Figure 1. DNA was then precipitated with 2 volumes of ethanol and the supematant was used forthe reaction of cysteine alkylation and amino acid derivation as outlined in Materials and Methods.

=

L

252

PRELIMINARY CHARACTERIZATION OFAMINO ACID-DNA ADDUCTS

-o14

< UsCM 10

IEa<:8 4Ec!

zaEc

Ec

10

S

S

S1

< az0aE

2

e 2

MM uiucr- Ioop o

1i

- lUSpMor

0EDTA EDTA

#

! ISla

! IS

Asp Olu Am Se Oln His bw Ae Aug Tyr Mt Val Pes

Acid

IAs OluASAO Gin gHi Thr As AM Tyr NO Vol Ph*

NaOH

J1uAA A] .VII j -A am_-u-

Asp aim Ass Sw in HN Th ANe A Tr Mm Val Pi

Figure 3. Quantitation of amino acids associated withDNA. DNA was treated with EDTA, acid, or sodiumhydroxide as described in Figure 1. The amino acids wereseparated with HPLC and several amino acid standardswere used to identify and quantitate unknown peaks.

tion with DNA (Figures 3,4). There wasno particularly different pattern in theindividual amino acids extracted by eitherEDTA or acid or base.

Figure 5 examines the increase in anumber of amino acids (EDTA extracted)including histidine, methionine, and threo-nine exhibiting a dose-dependency in theirassociation with DNA following chromatetreatment of intact cells. The bottom por-tion of Figure 5 shows the amino acidsmost affected in this complexing withDNA. For cysteine, as well as glutamine,there was a 35- to 60-fold increase in cross-linking that was dose-dependent. Even 5pM chromate was found to substantiallyincrease cysteine, glutamine, and glutamicacid complexing with the DNA.

Figure 6 shows time-dependence in thecomplexing of amino acids with DNA.Relatively long incubation periods were

Asp as Am S_a. Nh T_ Al M VW se P.e

Acid

.

a02a

NaOHa.

115

01 w10 1J | X

Aspal Ami Seta.Hi Thu Mm Au W MM~Pisl

Figure 4. Amino acids associated with DNA followingtreatment with chromate. CHO cells were treated with 50or 100 ,uM potassium chromate for 20 hr in complete oc-MEM. Samples were prepared as described in Figure 1and amino acids were separated by HPLC as described inMaterials and Methods. The ratio of each amino acidamount extracted from the DNA after Cr treatment to theamount of amino acid released from control DNA wasdetermined.

required to produce substantial increases inamino acid DNA complexes (16-20 hr).

To evaluate whether the extraction pro-cedure measures amino acid residues fromprotein-DNA cross-links or cross-linkingof free amino acids with DNA, we reactedDNA with BSA in the absence or presenceof trivalent Cr(III) under conditions inwhich we have shown that BSA could becross-linked to the DNA Cr (10). We alsoreacted DNA with amino acids in theabsence and presence of trivalent Cr.Figure 7 illustrates that when BSA wascross-linked to the DNA by Cr, there waslittle increase of amino acids above back-ground. However, when free amino acidswere cross-linked with DNA by Cr(III),there was a substantial increase of aminoacids complexed with DNA above the

60 1 Cys

50 .-o

45 .-o

40.35 'o ,, gln

25 - v, ,,z20 -

1,,,

15 .,/,- glu1 0 . ,#5

.1

0 20 40 60

Cr ( uM )

so 100

40 60

Cr (pM )

Figure 5. Concentration dependence in the association ofamino acids with DNA. CHO cells were treated with 100pM potassium chromate for 4, 8, 16, and 20 hr in completemedia. Samples of amino acids were prepared asdescribed in Materials and Methods.

background. When protein synthesis wasinhibited by cyclohexamide, more radiola-beled threonine was associated with DNAfollowing chromate treatment than when itwas not inhibited (Figure 8). These resultssuggest that free amino acids react readilywith DNA in intact cells to produceDNA-amino acid complexes. Additionally,we used a recently developed nonproteasemethod to isolate cellular DNA followingchromate treatment (6). Protein-DNAcomplexes were separated by precipitationwith K+- SDS of DNA mechanically frag-mented to homogeneous sizes while DNAnot precipitated by the K+- SDS had little orno DNA-protein cross-links. Examinationof this soluble DNA revealed substantialincreases in the amount of amino acidDNA complexes in chromate-treated cellscompared to controls (data not shown).

DiscussionRadiolabeled amino acids complexed withDNA following treatment of intact cellswith Ni and Cr compounds have beenstudied previously (11). Cysteine, histi-dine, tyrosine, threonine, and methioninewere increased in their association withDNA in a dose-dependent manner (11).The present study agrees with the previousones in finding these amino acids com-plexed with DNA; however, the results alsodemonstrate that glutamine and glutamicacid were increased substantially in theirassociation with DNA by chromate but

Volume 102, Supplement 3, September 1994

I.200'D0cop

16322

211

4D

011z

16'O

.2

253

I

-A A

VOITKUN ETAL.

gin

glu

arg

asn

time (h) C _ +.,L._Aa

met

thrhis

__~~~~~~~~~~~~~,- tyr,,, +

4 aG 12 18 20 *A,N-time (h)

Rgure 6. Time dependence in the association of aminoacids with DNA. CHO cells were treated with 5, 10, 25,50, or 100 pM, potassium chromate for 20 hr in completea-MEM. The amino acids were extracted with 0.2 NNaOH (2 hr at 37°C) and were separated with (Materialsand Methods). The data presented are the ratio of aminoacid amount released from the DNA of Cr-treated CHOcells compared to control.

these amino acids were not found to beincreased when radiolabeled amino acidswere used. The reason for this is presentlyunknown but may have to do with radiola-beling problems. Most of the cysteine asso-ciated with DNA using radiolabeled aminoacids was found to originate from glu-tathione cross-links induced by chromate(11). This was demonstrated by inhibitingprotein synthesis with cycloheximide,which actually stimulated the amount ofcysteine associated with DNA (11). Thus,the cysteine-DNA cross-links do not repre-sent residual proteins associated with theDNA, but most probably represent actualglutathione-DNA adducts. In fact, mostcysteine in proteins is generally not accessi-ble for cross-linking in intact cells.

The fact that glutamine and glutamicacid were found to be highly cross-linkedto the DNA in the present study when anonradioactive method was used may bedue to the high levels of glutamine thatwere added in the tissue-culture media. Infact, using radioactive glutamine, the spe-cific activity of glutamine would be so lowthat glutamine DNA cross-linking may notbe detected by these radioactive methods.These results suggest that, in fact, the resid-ual amino acids associated with DNA didnot, in large part, originate from proteincross-linked with DNA, but originated

Figure 7. HPLC separation of amino acids associated withDNA following in vitro cross-linking with bovine serumalbumin and amino acids. Calf thymus DNA was incubatedwith amino acids (D, E) or BSA (B, C) in the presence (C, E)or absence (A, B, C) of 100 ,uM CrCI3. DNA was purified asdescribed in Materials and Methods. Amino acids wereextracted with 0.2 N NaOH (2 hr at 37°C) and analyzed byHPLC (Materials and Methods). The peaks are designatednumbers according to the numbering of peaks in Figure 2.

from free amino acid pools or small pep-tides that were cross-linked to the DNA inintact cells. This was certainly the case withglutathione that was found complexed withthe DNA (11). Most of the cysteine thatwas found complexed with the DNA origi-nated from glutathione (11). Experimentsconducted here also suggest that whenalbumin was cross-linked to the DNA,there was little residual amino acid com-plexed with DNA; however, when freeamino acids were cross-linked to DNA andthe DNA was isolated by standard meth-ods, there was an abundance of amino acid.Further, when protein synthesis was inhib-ited in intact cells, there was more radiola-beled amino acid cross-linked to the DNA,suggesting that free amino acids do, in fact,react with the DNA. Although there maybe some contribution from residual pro-tein-DNA cross-links, a substantial per-centage of the amino acids associated withDNA probably originated from the reac-tion of free amino acids with DNA. In gen-eral, free amino acids are chemically morereactive than amino acids compexed aspeptides.

We previously studied the chemistry ofthe cross-linking of amino acids to theDNA and demonstrated in the case ofchromate that trivalent Cr was involved in

22002z

6050-

200

0

t 2 3

t Control C ex os uolAr

Figure 8. Effect of cycloheximide on the crosslinking ofthreonine to the DNA by chromate. 1) Cells were incu-bated with radiolabeled threonine (3H-Thr) 24 hr beforeexposure. 3H-Thr was removed when cells were treatedwith potassium chromate for 16 hr. 2) After 7 hr of Cr-exposure of non labeled CHO cells, radioactive threoninewas added and Cr treatment was continued for an addi-tional 9 hr. 3) After 6 hr of Cr exposure of nonlabeled CHOcells, 50 pMJ cycloheximide was added to the media;then in 1 hr, H-Thr was added and cells were incubatedfor an additonal 9 hr. DNA was isolated and amino acidswere extracted with sodium hydroxide as previouslydescribed. Radioactivity corresponded threonine HPLCpeak was collected and counted using a Beckman scintil-lation counter.

the cross-link, since DNA-bound cysteinecould be extracted with EDTA (I11).Trivalent Cr is the only form of chromatethat can be chelated by EDTA, and we arenot studying amino acid cross-links exceptthose involving trivalent Cr with EDTA.However, we have also observed in previ-ous studies that amino acids can be cross-linked to the DNA by oxidativemechanisms that do not actually involvetrivalent Cr (11).

The biologic significance of the aminoacids cross-linked to the DNA is not cur-rently understood; however, these werethought to be amino acid residues arisingfrom chromatin-scaffold proteins (1-3).We show here that peptides and freeamino acid cross-linking in intact cellsmust also be considered as a possible majororigin of these amino acid cross-links. Thefact that one can demonstrate such a largeincrease in amino acids associated withDNA following treatment with cross-link-ing agents indicates that these aminoacid-DNA complexes may be a substantialimpediment to DNA replication thatcould possibly lead to mutations and maybe responsible for carcinogenic effects ofchromate. Complexes of DNA and aminoacids probably enhance the stability of theCr bound to the DNA. An additional uti-lization of this lesion as a potential bio-marker ofhuman exposure to cross-linkingagents is possible since amino acids associ-ated with DNA can be quantitated withhigh precision.

Environmental Health Perspectives

15

m

1316

.90112

a22

.9011102

tz

0

0

ML + am

I5 DU -. AGIM

13 16

254

PREUMINARY CHARACTERIZATION OFAMINOACID-DNA ADDUCTS

REFERENCES

1. Gianfranceschi GL, Barra D, Bossa F, Coderoni S, Paparelli M,Venanzi F, Cicconi F, Amici D. Small peptides controlling tran-scription in vitro are bound to chromatin DNA. Biochem BiophysActa 69:138-148 (1982).

2. Juodka B, Pfuitz M, Werner D. Chemical and enzymatic analysis ofcovalent bonds between peptides and chromosomal DNA. NucleicAcid Res 19:6391-6398 (1991).

3. Sutcliffe JA, Gootz TD, Barrett JF. Biochemical characteristics andphysiological significance of major DNA topoisomerases.Antimicrob Agents Chemother 33:2027-2033 (1989).

4. Miller CA, Costa M. Analysis of proteins cross-linked to DNA aftertreatment of cells with formaldehyde, chromate, and cis-diamminechloroplatinum(II). Mol Toxicol 2:11-26 (1989).

5. Costa M. Analysis of DNA-protein complexes induced by chemicalcarcinogens. J Cell Biochem 44:127-135 (1990).

6. Zhitkovich A, Costa M. A simple, sensitive assay to detect DNA-protein cross-links in intact cells and in vivo. Carcinogenesis

13:485-1489 (1992).7. Costa M, Zhitkovich A, Toniolo P. DNA-protein cross-links in

welders: molecular implications. Cancer Res 53:460-463 (1993).8. Neuer B, Plagens U, Werner D. Phosphodiester bonds between

polypeptides and chromosomal DNA. J Mol Biol 104:213-235(1983).

9. Lindroth P, Mopper K. High performance liquid chromatographicdetermination of subpicomole amounts of amino acids by precol-umn fluorescence derivatization with o-phtaldialdehyde. AnalChem 51:1667-1674 (1979).

10. Salnikow K, Zhitkovich A, Costa M. Analysis of the binding sitesof chromium to DNA and protein in vitro and in intact cells.Carcinogenesis 13:2341-2346 (1992).

11. Lin X, Zhuang Z, Costa M. Analysis of residual amino acid-DNAcross-links induced in intact cells by nickel and chromium com-pounds. Carcinogenesis 13:1763-1768 (1992).

Volume 102, Supplement 3, September 1994 255