Mechanisms of the Carcinogenic Chromium(VI)-induced DNA-Protein Cross-linking and Their Characterization in Cultured IntactHuman Cells*

(Received for publication, May 6, 1996, and in revised form, October 4, 1996)

Subhendra N. Mattagajasingh and Hara P. Misra‡

From the Department of Biomedical Sciences and Pathobiology, Virginia-Maryland Regional College of VeterinaryMedicine, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061-0442

DNA-protein complexes (DPCs) were induced in hu-man leukemic T-lymphocyte MOLT4 cells by treatmentwith potassium chromate. DPCs were isolated by ultra-centrifugal sedimentation in the presence of 2% SDS and5 M urea. The complexes were analyzed by two-dimen-sional SDS-polyacrylamide gel electrophoresis. Threeacidic proteins of 74, 44, and 42 kDa and a basic proteinof 51 kDawere primarily complexed to DNA following 25mM chromate treatment. Higher concentrations of chro-mate cross-linked many other proteins to DNA. Aminoacid sequencing and immunoblotting studies indicatedthat the acidic 44-kDa protein could be nuclear b-actin.Lectin and aminoglycoside nucleotidyltransferase werealso found to cross-link with DNA by chromate treat-ment. The composition and stability of the DPCs werestudied using nucleases, proteinase K, and disruptivechemicals. Pretreatment of cells with antioxidants in-hibited the formation of DPCs, measured as K1-SDS pre-cipitable DPCs, indicating the involvement of oxidativemechanisms. Because chromate causes certain nuclearproteins to form complexes with DNA and the complexesare resistant to treatments such as 2% SDS and 5 M urea,but disruptable under gel electrophoretic conditions,chromium could be used as a cross-linking agent for theidentification of other proteins, such as transcriptionfactors, that transiently interact with DNA.

DNA-protein complexes have been shown to be induced by anumber of physical and chemical carcinogenic agents such asg-radiation and UV light (1, 2), alkylating agents (3), formal-dehyde (4), vinyl acetate (5), and metal compounds of chromate(6–8), nickel (9), cis- or trans-platinum (10) and vanidium(V)(11). Hexavalent chromium (Cr(VI)) compounds have been con-sidered as potent human carcinogens and have been shown tocause different types of DNA damage including DNA-proteincross-linking in various cells and tissues (see Ref. 12 for areview). Interestingly, Cr(VI) does not bind to DNA or proteinsin cell-free systems (13, 14). However, Cr(VI), which exists asan oxyanion at physiological pH, is readily transported into thecell through the cells’ sulfate anion transport system (15, 16).Inside the cell, Cr(VI) is believed to be reduced by the cells’redox system to its biologically most stable form, chromium(III)

(17, 18). Cr(III) binds to DNA as well as proteins in cell-freesystems (19) and has high affinity for many other biologicalligands (20). Cr(III), however, is poorly taken up into the celland is considered to be noncarcinogenic (21). During the intra-cellular reduction of Cr(VI) to Cr(III), reactive species such asintermediate valance states of chromium and active oxygenspecies are generated (17, 22–23), which may, in turn, initiatethe carcinogenic process by altering the structure of DNA (24).Hydroxyl radicals (zOH), which are generated during the cellu-lar reduction of chromate (25) are also capable of causing DNA-protein cross-linking (26, 27) and are considered as the “ulti-mate agents” in chromate carcinogenesis (28). Cr(III) and thereactive intermediate states of chromium may also be consid-ered as carcinogenic because zOH radicals are shown to begenerated by redox cycling of Cr(III) (29), and DNA damage hasbeen shown to be caused by intermediate valence states ofchromium, such as Cr(V) (30).Although chromate-induced DNA-protein complexes are im-

plicated in chromate carcinogenicity, the mechanisms of theirformation, composition, and biological significance are not wellunderstood. It has been postulated that cross-linking of pro-teins to DNA could disrupt chromatin structure and the normalregulation of gene expression (31). This, in turn, could play arole in carcinogenesis in that deletion of DNA bases may resultwhen portions of replicating DNA are buried under DNA-pro-tein complexes (32). Such deletions to “tumor suppressorgenes” (33) may give rise to loss or inactivation of the gene,leading to carcinogenesis. Furthermore, during normal regula-tion of gene expression, proteins, either alone or in cooperationwith other proteins, reversibly interact with specific DNA se-quences (34). Cross-linking of DNA with inappropriate proteinscould disrupt the normal regulation of DNA-protein interac-tions, causing serious genetic consequences, including disrup-tion in or alteration of gene expression. Therefore, it is neces-sary to know the identity of the proteins that participate inchromate-induced DNA-protein complexes and the nature oftheir interaction with DNA. Identification of proteins cross-linked to DNA may also assist in our understanding of chro-matin structure and protein interactions, including the three-dimensional orientation of proteins around DNA.In the present study, we have analyzed the proteins com-

plexed to DNA by chromate treatment of MOLT4 cells. We havepreviously explained the reasons for the choice of this cell line(23). Additionally, an increased yield of DNA-protein cross-links and a faster reaction kinetics have been reported inMOLT4 cells as compared with other cells (11). We have alsoattempted to identify some of the proteins that cross-link withDNA after chromate exposure, by isolation and partial N-ter-minal sequencing of these proteins, as well as using antibodiesto “candidate proteins,” because inappropriate complexing ofproteins of structural and/or functional importance to DNA

* The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby marked“advertisement” in accordance with 18 U.S.C. Section 1734 solely toindicate this fact.Part of this work was presented at the 10th International Conference

on Methods in Protein Structure Analysis, Snowbird, UT, September8–13, 1994.‡ To whom correspondence should be addressed. Tel.: 540-231-7174;

Fax: 540-231-7367; E-mail: misra@vt.edu.

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 271, No. 52, Issue of December 27, pp. 33550–33560, 1996© 1996 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

This paper is available on line at http://www-jbc.stanford.edu/jbc/33550

by guest on Novem

ber 20, 2020http://w

ww

.jbc.org/D

ownloaded from

rather than the DNA-protein complexes themselves may haveimportance in chromate carcinogenicity. Identification of suchproteins is essential for better understanding of the potentialconsequences of DNA-protein cross-links and the three-dimen-sional orientation of proteins around DNA. The compositionand stability of chromate-induced DNA-protein complexes andthe effect of antioxidants on the formation of such complexeshave also been reported here.

MATERIALS AND METHODS

Chemicals—Highest purity grade potassium chromate (K2CrO4) waspurchased from J. T. Baker (Phillipsburg, NJ). A protein determinationkit and all gel electrophoresis reagents were purchased from Bio-Rad.Polyvinylidene difluoride (PVDF)1 membranes, [3H]thymidine, [35S]me-thionine, 125I-protein A, [51Cr]potassium chromate, and Aquasure LSCmixture were purchased from DuPont NEN. DNase-free RNase andproteinase K were purchased from Boehringer Mannheim. All otherchemicals and enzymes were purchased from Sigma.Cell Culture and Treatment—Human leukemic T-lymphocyte

MOLT4 cells (ATCC CRL 1582) were purchased from American TypeCulture Collection (Bethesda, MD) and were maintained in suspensionat exponential growth phase in RPMI 1640 (HEPES-modified) mediumsupplemented with 10% heat-inactivated fetal bovine serum, 10 units ofpenicillin, and 10 mg/ml streptomycin solution as described before (22).Cellular DNA and proteins were radiolabeled with [3H]thymidine and[35S]methionine (0.02 mCi/ml each), respectively, for ;24 h, in methi-onine-free RPMI 1640 medium. Radiolabeled cells were collected bycentrifugation, washed three times in cold Saline A (5 mM NaHCO3, 6mM dextrose, 5 mM KCl, and 140 mM NaCl, pH 7.2), and resuspended insalts/glucose medium (SGM; 50 mM HEPES, 100 mM NaCl, 5 mM KCl,2 mM CaCl2, 5 mM dextrose, pH 7.2) at a concentration of 106 cells/ml.Potassium chloride (as control) or potassium chromate was added to thecell suspensions from freshly prepared stock solutions. Following treat-ment, cells were collected, and cytotoxicity was determined by exclusionof trypan blue as described previously (22).Isolation and Quantitation of DNA-Protein Complexes from Intact

Cells—The method used to isolate DNA-protein complexes was modi-fied from our previously described method (35). Potassium chromate-treated and control cells were collected by centrifugation at 500 3 g for10 min and were washed three times in phosphate-buffered saline. Thecells were lysed in 30 ml of 10 mM Tris-HCl containing 2% SDS, 1 mM

PMSF (pH 7.5) by shaking on a platform shaker for 6 h at roomtemperature. The cell lysates were transferred into a tight-fitting ho-mogenizer and given ten strokes. The samples were sedimented at100,000 3 g for 16 h at 18 °C, using a Beckman SW 27 rotor (BeckmanInstruments, Fullerton, CA). The pellets were placed in 28 ml of 5 M

urea containing 1 mM PMSF and rocked on a platform shaker at 4 °C for6 h. The samples were again homogenized as above, and then SDS wasadded to a 2% final concentration. The DNA-protein complexes wereisolated by ultracentrifugation as above and were rinsed in 10 mM

Tris-HCl (pH 7.5) containing 1 mM PMSF and 2% SDS. The final pelletswere resuspended in 10 mM Tris-HCl containing 1 mM PMSF (pH 7.5) insiliconized Eppendorf tubes by gentle pipetting and overnight rockingon a Nutator shaker at 4 °C. The DNA-protein complexes were precip-itated in 70% acetone at 220 °C. The DNA-protein complexes werecollected by centrifugation at 12,500 3 g for 15 min at 4 °C using aBeckman microfuge, rinsed in 70% acetone, and resuspended in 1 ml of10 mM Tris-HCl containing 1 mM PMSF (pH 7.5) by gentle pipetting orby shaking on a Nutator shaker for about 16 h at 4 °C. The DNA contentand the purity of the samples were determined by measuring theabsorbance at 260 and 280 nm (36). Both 3H and 35S activities weredetermined by dissolving the samples in Aquasure Mixture (DuPontNEN) and counting in a Beckman LS 5800 Liquid Scintillation counter(Beckman Instruments, Inc., Irvine, CA). The protein content of celllysates was determined by using Bio-Rad dye and bovine g-globulin asstandards (37).Isolation and Quantitation of DNA-Protein Complexes from Purified

Nuclei—Cells were collected by centrifugation at 500 3 g for 10 min,after which cells were washed three times in phosphate-buffered salineand incubated for 15 min on ice in a cold hypotonic buffer (10 mM

Tris-HCl, pH 7.5, containing 10 mM NaCl, 1.5 mM MgCl2). Cells were

collected by centrifugation at 300 3 g for 5 min, resuspended in theabove solution supplemented with 0.5% Nonidet P-40 and 1 mM PMSF,and were given 8–10 strokes in a loose fitting glass homogenizer. Thenuclei were sedimented at 700 3 g for 5 min at 4 °C; resuspended in 10mM Tris-HCl containing 250 mM nuclease-free sucrose, 3 mM MgCl2,and 1 mM PMSF (pH 7.5); and layered over a similar solution butcontaining 880 mM sucrose. Nuclei were subsequently collected by cen-trifugation for 10 min at 1000 3 g at 4 °C using a swinging bucket rotorin an ICE Centra-7R refrigerated centrifuge. The nuclei were observedby phase contrast microscopy, Giemsa staining, and Acridine orangestaining, and were found to be free from cytoplasmic contaminations(data not shown). The purified nuclei were used for isolation of DNA-protein complexes using the method described above.Analysis of Proteins by Two-dimensional Gel Electrophoresis—DNA-

protein complexes were analyzed by nonequilibrium pH gradient elec-trophoresis as described by O’Farrell et al. (38) This procedure does notallow nucleic acids to enter into the first dimension focusing gel. DNA-protein complexes containing 150 mg of DNA were acetone-precipitatedor lyophilized (FTS Systems, Inc., Stone Ridge, NY) and solubilized in30 ml of solubilizing buffer (9 M urea, 4% Nonidet P-40, 2% b-mercap-toethanol, and 3% ampholines (Bio-Rad), pH range 3–10). Isoelectricfocusing and nonequilibrium focusing were carried out in 200-ml capil-lary tubes (1.5-mm diameter, Fisher) containing 4% polyacrylamideand 2% ampholines (pH range 3–10). Second dimensional separationwas carried out on 10 or 12% SDS-polyacrylamide gels by following themethod of Laemmli (39) except that 1% b-mercaptoethanol was used. Insome cases, the DNA-protein complexes were digested by DNase I orRNase A and concentrated by acetone precipitation or lyophilizationbefore analysis by two-dimensional gels. The gels were subjected topolychromatic silver staining by following the method of Sammons et al.(40).Subcellular Fractionation—Subcellular localization of the proteins

complexed to DNA upon chromate treatment was determined by ana-lyzing the cytoplasmic, nuclear, and nuclear matrix proteins of MOLT4cells by two-dimensional gel electrophoresis. The protein fractions con-taining the membrane and cytoplasmic fractions were prepared bysedimentation of nuclei after hypotonic lysis of cells in the presence ofNonidet P-40, as described above (Nonidet P-40-soluble cytoplasmicmaterial). The nuclear protein fraction was prepared from the SDS-soluble material of the isolated nuclei, or by sonication of the purifiednuclei. The nuclear matrix fractions were prepared by digestion of thepurified nuclei with DNase I (400 Kunitz units/ml) and RNase A (20Kunitz units/ml), followed by extraction with 1.6 M NaCl, as describedpreviously (41). The protein contents of different fractions were deter-mined by the dye binding method, as described above. Thirty mg ofprotein from each fraction was acetone-precipitated and solubilized in30 ml of the solubilizing buffer. Equilibrium or nonequilibrium focusing,separation in the second dimension on 10% polyacrylamide gels, andsilver staining of samples were carried out as described above.Electrophoretic Transfer of Proteins to PVDF Membrane and Amino

Acid Sequencing—DNA-protein complexes containing 250 mg of DNAwere analyzed by two-dimensional gel electrophoresis and were electro-blotted onto PVDF membrane as described previously (42). The seconddimensional gel was prerun in the presence of 1 mM sodium thioglyco-late to protect the proteins from N-terminal blocking. Proteins wereelectroblotted onto PVDFmembranes in a Bio-Rad Transblot apparatususing 10 mM CAPS and 10% high pressure liquid chromatographygrade methanol (pH 11) as the electroblotting buffer at 50 V for 1 h atroom temperature. Coomassie Brilliant Blue R-250 (Bio-Rad; 0.025% in40% methanol) was used to visualize the proteins. Acetic acid wasomitted from the staining and destaining solution, since it may causeN-terminal blocking. The protein band of interest was excised, andautomated Edman degradation was performed using an Applied Bio-systems 477A protein sequencer equipped with a 120 A analyzer (Ap-plied Biosystems, Inc., Foster City, CA).Electroblotting of Proteins to Nitrocellulose Membrane and Immuno-

detection of Actin Using an Anti-actin Antibody—Following one- ortwo-dimensional gel electrophoresis, the proteins were electrophoreti-cally transferred to nitrocellulose membrane by following a modifica-tion of the method of Burnette (43). Proteins were transferred to nitro-cellulose membrane in 25 mM Tris, 100 mM glycine, and 10% methanolovernight at 200 mA. The nitrocellulose membrane sheet was blocked in20 mM Tris-HCl containing 0.05% Tween 20, 1% NaCl, 0.05%NaN3, and4% nonfat dry milk (pH 7.5) for 1 h at room temperature. Then the blotwas incubated with a 1:1000 dilution of the antibodies in blocking bufferfor 3 h with shaking. The blot was washed in 20 mM Tris-HCl containing0.05% Tween 20, 1% NaCl, 0.05% NaN3 (pH 7.5) and was reacted with2 mCi of 125I-protein A for 3 h at room temperature. The unbound

1 The abbreviations used are: PVDF, polyvinylidene difluoride;CAPS, 3-(cyclohexylamino)propanesulfonic acid; DPC, DNA-proteincross-link; PMSF, phenylmethylsulfonyl fluoride; SGM, salts/glucosemedium.

Chromium(VI)-induced DNA-Protein Cross-linking 33551

by guest on Novem

ber 20, 2020http://w

ww

.jbc.org/D

ownloaded from

antibodies were removed by washing the blot in 20 mM Tris-HCl con-taining 0.05% Tween 20, 1% NaCl, 0.05% NaN3 (pH 7.5) for five times(5 min each). The blot was blot-dried, and the antigens were detected byautoradiography of immunoblots using XRP-1 film (Eastman KodakCo.).DNA Sizing on Agarose Gels—DNA-protein complexes containing 10

mg of DNA from control and chromate-treated samples were fractionedon agarose gels (0.6% (w/v) in 40 mM Tris, 40 mM borate, 1 mM EDTA)(36) for ;16 h at 50 V on a model H4 horizontal gel electrophoresisapparatus (Life Technologies, Inc.). The gel was stained with ethidiumbromide (0.2 mg/ml in TBE), and DNA was visualized by ethidiumbromide fluorescence at 254 nm.Antioxidant Treatment of Cells and Its Effect on the Formation of

Chromate-induced DNA-Protein Complexes—The nontoxic levels of thetest compounds were determined by monitoring the effect of 0–100 mM

a-tocopherol succinate, 0–5 mM sodium ascorbate, 0–5 mM Tiron and0–100 mM mannitol, on the growth of MOLT4 cells up to 96 h. Cellswere treated with nontoxic levels (.98% viable) of a-tocopherol succi-nate (25 mM), sodium ascorbate (1 mM), Tiron (1 mM), or mannitol (10mM) for 16 h in complete RPMI at 37 °C before exposure to chromate inSGM for 3 h.DNA-protein cross-links were detected by modification of previously

published methods (44–46). Labeling of cells with [3H]thymidine andtreatment with chromate were carried out as described above. Follow-ing treatment, cells were washed three times in ice-cold phosphate-buffered saline and were frozen at 280 °C for 12–16 h. The cells werethawed and were lysed in 20 mM Tris-HCl containing 1 mM PMSF and2% SDS (pH 7.5). The cell lysates were briefly sonicated on ice using aHeat System sonicator/cell disrupter (model W 225 R, Ultrasonics, Inc.,Plainview, NY) by giving 10 pulses at 50% duty cycle. The samples wereincubated at 65 °C for 10 min, and KCl (in 20 mM Tris-HCl, pH 7.5) wasadded to 100 mM final concentration. The samples were chilled on ice for10 min, and the K1-SDS precipitates formed were collected by centrif-ugation at 3000 3 g for 10 min at 4 °C. The pellets were resuspended in20 mM Tris-HCl containing 100 mM KCl. The samples were incubated at65 °C for 10 min, cooled on ice, and collected by centrifugation as above.This shearing and washing step was repeated two more times. The finalpellets were resuspended in water and the protein bound [3H]DNA wasestimated by counting the samples in a Beckman LS 5800 LiquidScintillation counter (Beckman Instruments, Inc., Irvine, CA) usingAquasure LSC Mixture (DuPont NEN). Trichloroacetic acid-insolublematerial from the cell lysates was used to estimate the total [3H]DNA.The ratio of the percentage of K1-SDS-precipitable DNA in the treatedcells to that in the control cells was used to estimate the DNA-proteincross-link coefficient (DPC coefficient). Unlike the SDS-urea method,which requires 5–6 days, the K1-SDS method is more sensitive, andresults can be obtained in 1 day. Therefore, the K1-SDS method wasused to monitor the effects of antioxidants where isolation and charac-terization of proteins were not required.Estimation of Cellular Ascorbate and Vitamin E—The intracellular

ascorbate level of cells was estimated by using Folin phenol reagent asdescribed before (22). The vitamin E level of cells was determinedspectrophotometrically by using the method of Fabianek (47).Cellular Uptake of 51CrO4

22—Cells, in SGM, were incubated with 1mCi of K2

51CrO4 at various concentrations for 2 h at 37 °C. Cells werecollected by centrifugation and washed three times with ice-cold phos-phate-buffered saline, and the cell number was counted in a Coultercounter (model ZM, Coulter Electronics, Inc., Hialeah, FL). The cellularuptake of chromate was determined by measuring the 51Cr activity in aBeckman g counter (Beckman Gamma 5500 equipped with a DP 5500Data Reduction System, Beckman Instruments) and comparing with astandard curve generated by using 0–20 nM potassium 51chromate.Determination of Stability of DNA-Protein Complexes—DNA-protein

complexes containing 100 mg of DNA in 10 mM Tris-HCl, 1 mM PMSF(pH 7.5) were taken in siliconized microcentrifuge tubes. MgCl2 wasadded to a 5 mM final concentration in samples treated with DNase Iand RNase. DNase I (200 mg/ml), RNase A (40 mg/ml), proteinase K (2mg/ml), EDTA (10–50 mM), thiourea (100 mM), or b-mercaptoethanol(2%) was added and mixed, and the tubes were incubated at roomtemperature for 3 h. Either PMSF was omitted from the samplestreated with proteinase K, or the samples were incubated at 4 °C for24 h before treatment with proteinase K to allow inactivation of PMSF.SDS was then added to a final concentration of 0.5% to inhibit nonspe-cific cross-linking, and samples were centrifuged at 100,000 3 g for 16 hat 18 °C. The supernatants were carefully removed, and the pelletswere resuspended in 10 mM Tris-HCl (pH 7.5) by brief sonication. DNAand protein contents were determined from the 3H and 35S specificactivity, respectively, by liquid scintillation counting as described above.

Statistics—All experiments were performed at least three times.Paired, two-tailed Student’s t test was performed, and p values #0.05were considered significant.

RESULTS

Cytotoxicity—Exposure of MOLT4 cells to 0–2 mM potassiumchromate in SGM for 2 h was found to have little cytotoxiceffect, as assessed by trypan blue exclusion (viability waswithin 98 6 2% of the control). The viability of cells treatedwith 200 mM chromate was not affected within 4 h of treatment.The viability of cells treated with 200 mM chromate for 16 h inSGM was decreased to 72 6 3% of the control.Effect of Potassium Chromate on DNA-Protein Cross-linking

in MOLT4 Cells—Cell exposure to 0–2 mM potassium chromatefor 2 h resulted in a dose-dependent increase in the formationof DNA-protein complexes in MOLT4 cells. Cells treated with200 mM chromate for 2 h had about 175% more DNA-proteincomplexes as compared with the control (Fig. 1). Chromate-induced DNA-protein complex formation in MOLT4 cells wasalso found to be time-dependent. As shown in Fig. 2, the DNA-protein complex formation increased in a time-dependent man-ner in cells treated with 200 mM chromate for different timeperiods, and after 24 h a 10–12-fold increase in the formationof DNA-protein complexes was observed as compared with thecontrol cells.Analysis of Proteins Complexed to DNA by Two-dimensional

Gels—DNA-protein complexes isolated from both the potas-sium chloride (control) or potassium chromate-treated cellswere analyzed by nonequilibrium two-dimensional gel electro-phoresis. DNA-protein complexes were loaded on the acidic endof the gel in order to avoid the entry of nucleic acids into thefirst dimensional focusing gels. Silver staining of two-dimen-sional gels of DNA-protein complexes isolated from control cellsdid not show any protein in the gel, indicating that the SDS/urea method used for isolation of DNA-protein complexes ef-fectively dissociates the background DNA-protein complexes inthe control cells (Fig. 3A).The proteins that were complexed to DNA, in cells exposed to

25 mM chromate for 16 h, and that were resistant to SDS/ureaextraction are shown in Fig. 3B. Three acidic proteins (a, b, andc) and a basic protein, d, were primarily complexed to DNA

FIG. 1. Dose-dependent increase in the formation of DNA-pro-tein complexes following exposure of MOLT4 cells to chromatefor 2 h in salts/glucose medium. Chromate treatment of cells andisolation of DNA-protein complexes by the SDS/urea method was fol-lowed as described under “Materials and Methods.” *, significantlydifferent from control, p # 0.01 (n 5 5).

Chromium(VI)-induced DNA-Protein Cross-linking33552

by guest on Novem

ber 20, 2020http://w

ww

.jbc.org/D

ownloaded from

upon chromate treatment. Analysis of the molecular weightand pI of these proteins showed that the protein a has a pI of5.2–5.6 and a molecular mass of 74 kDa, the protein b has a pIof 5.2–5.4 and a molecular mass of 44 kDa, and the protein chas a pI of ;5.8 and molecular mass of 42 kDa, respectively.The protein d, on the other hand, is found to be a basic proteinwith a pI of 8.8–9.2 and a molecular mass of 51 kDa. Thenumber of proteins cross-linked to DNA upon chromate expo-sure was found to be dependent on the dose of chromate,because cells treated with 200 mM chromate for 16 h had manyother proteins, in addition to the above proteins, complexed toDNA (Fig. 3C). Since 25% of the cells were found to be killed bysuch treatments (200 mM chromate for 16 h), there was a reasonto believe that the additional proteins cross-linked to DNAcould be due to dead cells. That this was not the case wasdemonstrated by treating cells with 200 mM chromate for 4 h.This latter treatment did not affect cell viability but cross-linked the same proteins to DNA (Fig. 3D).There is of course, a possibility that the cytoplasmic proteins

might associate with DNA during the cell lysis. In order tolessen the likelihood of this subtle artifact, we analyzed DNA-protein complexes isolated from either whole cells or purifiedintact nuclei of cells treated with chromate. If cytoplasmicproteins become associated with DNA during cell lysis, addi-tional proteins should appear in two-dimensional gels of wholecells as compared with that obtained from the nuclear frac-tions. That this was not the case is illustrated by the fact thatidentical proteins were found to be complexed to DNA wheneither whole cells or nuclei of cells treated with chromate wereused as the starting material (Fig. 3, D and E). Hence, itappears likely that chromate induces the cross-linking of nu-clear proteins to DNA.Molecular mass and pI of the major proteins complexed to

DNA upon 200 mM chromate treatment are listed in Table I.These proteins were not seen in two-dimensional gels of DNA-protein complexes isolated from untreated control cells, evenwhen the initial DNA load was increased to visualize the back-ground proteins. When the control material was digested withDNase I before analysis by two-dimensional gel, only traceamounts of some of the above proteins were observed (data notshown), but some of these proteins were also present as con-

taminants in DNase I (data not shown). Therefore, we assumethat the traces of proteins observed in control cells are not fromDNA-protein complexes but from the contaminant proteinspresent in DNase I.Subcellular Localization of Major Proteins Complexed to

DNA upon Chromate Treatment—To determine the subcellularlocalization of the four proteins complexed to DNA, cytoplas-mic, nuclear, and nuclear matrix protein fractions were ana-lyzed by two-dimensional gel electrophoresis. The purified nu-clei were free from cytoplasmic contaminations (not shown).Proteins of similar molecular weight, isoelectric point, andcoloration after polychromatic silver staining were assumed tobe the same protein. Proteins b, and c were visualized and werefound to correspond to proteins in the cytoplasmic fraction (Fig.4A). Proteins a, and d were predominantly present in thenuclear fraction (Fig. 4B). Additional proteins found complexedto DNA (m, n, o, and p), upon treatment of cells with 200 mM

chromate, were also present in the nuclear fraction, althoughthey were predominantly present in the cytoplasmic fraction.Fig. 4C shows the two-dimensional resolution of nuclear matrixproteins of MOLT4 cells. As shown in this figure, all of theproteins cross-linked to DNA by 25 mM chromate treatment anda 63-kDa acidic protein (m) cross-linked to DNA by higherdoses of chromate were found in this fraction. These resultssuggest that nuclear matrix proteins are the target for chro-mate-induced DNA-protein cross-linking.Effect of Nucleases Digestion on the Resolution of Proteins

Dissociated from DNA-Protein Complexes—DNA-protein com-plexes were digested by DNase I or RNase A prior to analysisby two-dimensional gels to determine if nuclease digestionwould dissociate any other protein from the complex that is notresolved under the gel electrophoretic conditions. The proteinsresolved in two-dimensional gels without nuclease digestion ofchromate-induced DNA-protein complexes are shown in Fig.5A. Fig. 5, B and C, show the resolution of DNA-associatedproteins following treatment with DNase I and RNase A, re-spectively. There was no significant difference in the resolutionpattern of proteins dissociated from chromate-induced DNA-protein complexes with or without nuclease digestion. The dif-ference in protein resolution pattern in DNase I-treated sample(Fig. 5B) was due to the presence of proteins in DNase I (la-beled as D). Similarly, RNase A proteins are labeled as R inFig. 5C. These results suggest that nuclease digestion is notrequired for the resolution of chromate-induced DNA-proteincomplexes in two-dimensional gels. The resistance of chromate-induced DNA-protein complexes to treatments such as 2% SDSand 5 M urea, but their resolution in two-dimensional gelsindicates that these complexes are disruptable by the electro-focusing buffer, which contained 2% b-mercaptoethanol and 9M urea.Identification of the Proteins Complexed to DNA upon Chro-

mate Treatment of Cells by Partial Amino Acid Sequencing andImmunoblotting—Apart from the selection of candidate pro-teins based on similar molecular weights and isoelectric points,we have followed two different approaches to further charac-terize the proteins cross-linked to DNA by chromate. In oneapproach, proteins were partially sequenced from their N-ter-minal ends, and the amino acid sequence obtained was used tosearch for its homology in different protein data banks. So farwe have not been successful in identifying the proteins a, c, andd (Fig. 3C) by following the above approach. The few amino acidsequences obtained by N-terminal sequencing of these proteinshave not been found homologous to any proteins in the existingprotein data banks (data not shown). However, using thismethod, a 43-kDa protein (labeled as p in Fig. 3C, p43, pI6.0–6.5), which was predominantly detected in the cytoplasmic

FIG. 2. Time-dependent increase in the formation of DNA-pro-tein complexes following exposure of MOLT4 cells to 200 mMchromate in salts/glucose medium. DNA-protein complexes weredetected by the SDS/urea method as described under “Materials andMethods.” *, significantly different from control, p # 0.01 (n 5 5).

Chromium(VI)-induced DNA-Protein Cross-linking 33553

by guest on Novem

ber 20, 2020http://w

ww

.jbc.org/D

ownloaded from

FIG. 3. Nonequilibrium two-dimensional gel electrophoresis of DNA-protein complexes isolated from whole cells or nuclei ofcontrol (potassium chloride) or chromate-treated MOLT4 cells. Chromate treatment of cells and isolation of DNA-protein complexes by theSDS/urea method was followed as described under “Materials and Methods.” Each gel was loaded with DNA-protein complexes containing 150 mg

Chromium(VI)-induced DNA-Protein Cross-linking33554

by guest on Novem

ber 20, 2020http://w

ww

.jbc.org/D

ownloaded from

fraction but abundantly cross-linked to DNA, was identified aslectin. Thus, the N-terminal sequencing of p43 revealed sixconsecutive amino acids that have absolute homology withamino acid residues 24–29 of lectin Bra-3 (Fig. 6A). This se-quence is also partially homologous to many glycoproteins andthe human multidrug resistance protein 1. Based on the partialamino acid sequencing and the homology of the sequence, an-other 53-kDa protein (labeled as n in Fig. 3C) appears to beaminoglycoside nucleotidyltransferase. The seven amino acidssequenced from the N-terminal end (Fig. 6B) were found tohave absolute homology with the amino acids 23–29 of ami-noglycoside nucleotidyltransferase. The other approach used toidentify the proteins was immunoblotting, using commerciallyavailable antibody to candidate proteins. Using an anti-actinpolyclonal antibody (ICN Biochemicals, Inc.), the acidic 44-kDaprotein (labeled as b in Fig. 3C) has been identified as b-actin(Fig. 7). This protein is one of the most prevalent proteinscomplexed to DNA upon exposure of cells to chromate. Thehomology of the eight amino acids obtained by N-terminalsequencing of this protein with amino acids 19–26 of b-actin(Fig. 6C) and its molecular weight and pI similar to that ofb-actin suggest that this protein is nuclear b-actin.Effect of Antioxidants on Chromate-induced DNA-Protein

Complexes—During the biological reduction of Cr(VI), reactivespecies, such as chromium(V) and active oxygen species, aregenerated. Although oxygen radicals have been shown to cause

DNA-protein cross-linking (26, 27) and several investigatorshave suggested zOH radicals as the “ultimate carcinogen” in thechromate-induced carcinogenic process (28), the role of oxygenradical species in chromate-induced DNA-protein cross-linkinghas not been reported. Therefore, the effect of antioxidants andfree radical scavengers on chromate-induced DNA-proteincross-linking was investigated. Because differential uptake ofchromate by cells would lead to alterations in chromate-in-duced DNA-protein cross-linking, the effect of antioxidants oncellular uptake of chromate was first investigated. As shown inTable II, the cellular uptake of chromate increased in a dose-dependent manner when MOLT4 cells were exposed to 5–15 nMof potassium 51chromate. Pretreatment of cells with a-tocoph-erol succinate (25 mM), Tiron (1 mM), mannitol (10 mM), orascorbate (1 mM) had no significant effect on the cellular up-take of Cr(VI). As shown in Fig. 8, pretreatment of cells with 25mM a-tocopherol succinate, an antioxidant, inhibited the chro-mate-induced DNA-protein cross-linking by about 50%. Pre-treatment of cells with a-tocopherol or a-tocopherol acetate alsohad similar effects (data not shown). Cells treated with 25 mM

a-tocopherol succinate for 16 h in complete RPMI medium andexhaustively washed had an approximately 4-fold increase incellular tocopherol level over the untreated controls (data notshown). This is consistent with the findings of Sugiyama et al.(48), who have shown a 10-fold increase in cellular a-tocopherollevel after a 24-h treatment with 25 mM vitamin E in Chinesehamster V 79 cells. Pretreatment of cells with 1 mM Tiron (avitamin E analog) or 10 mM mannitol (zOH scavenger) inhibitedthe chromate-induced DNA-protein cross-linking by 45 and20% of control, respectively. Pretreatment of cells with 1 mM

ascorbate, on the other hand, increased the chromate-inducedDNA-protein cross-linking by about 150% of the control (Fig. 8).Cells treated with such levels of ascorbate increased the intra-cellular level of this vitamin by about 2-fold (data not shown).Since ascorbate had little effect on cellular uptake of chromateand pretreatment of control cells with ascorbate had trivialeffects on the background level of DNA-protein complexes, theobserved augmentation of chromate-induced DNA-protein com-plexing following this antioxidant treatment may, most proba-bly, be due to the direct reduction of Cr(VI) by ascorbate, givingrise to an increased level of intracellular Cr(III).Stability of DNA-Protein Complexes—The stability of DNA-

protein complexes was tested by monitoring the recovery ofDNA and protein in the pellet following treatment of DNase I,RNase A, proteinase K, EDTA, b-mercaptoethanol, or thiourea.As determined by agarose gel electrophoresis, the average sizeof DNA was approximately 7500 base pairs (not shown). Thecontrol samples (without any treatment) had almost 100% re-covery of both DNA and protein in the pellet following ultra-centrifugation, as determined by 3H and 35S radioactivity, re-spectively. Treatment of DNA-protein complexes, isolated fromboth control and chromate-treated cells, with DNase I signifi-cantly reduced the recovery of 3H and 35S in the pellet (Fig. 9).RNase A treatment of DNA-protein complexes did not interferewith recovery of DNA or protein. Proteinase K treatment dis-

of DNA. A, two-dimensional gel of DNA-protein complexes isolated from the nuclei of control cells. No proteins were detected on the gel of controlDNA-protein cross-links, indicating that the SDS/urea extraction method dissociated most of the background DNA-protein complexes. B,two-dimensional gel of proteins dissociated from DNA-protein complexes (obtained from purified nuclei) generated by treatment of cells with 25mM chromate for 16 h. The proteins a, b, c, and d were primarily cross-linked to DNA upon 25 mM chromate treatment of cells. C, same as B, exceptthat cells were treated with 200 mM chromate for 16 h. The letters m, n, o, and p refer to proteins that were prevalent in the cytoplasmic proteinfraction but cross-linked to DNA after 200 mM chromate exposure of intact cells. D, two-dimensional gel of proteins dissociated from DNA-proteincomplexes isolated from the nuclei of cells treated with 200 mM chromate for 4 h. As seen in this figure, identical proteins were cross-linked to DNAafter 200 mM chromate treatment of cells for either 4 or 16 h. E, two-dimensional gel of proteins dissociated from DNA-protein complexes isolatedfrom intact cells treated with 200 mM chromate for 16 h. As shown in this figure, identical proteins were cross-linked to DNA when DNA-proteincross-links were isolated from either the intact cells or the nuclei of cells exposed to chromate. The electrofocusing gel in C contained biolytes ofpH range 3–10 and 8–10 in a ratio of 4:1. In all other gels biolytes of pH range 3–10 was used.

TABLE IMolecular mass and pI of the major proteins complexed to DNA upon

chromate treatment of intact MOLT4 cellsDNA-protein cross-links were isolated fromMOLT4 cells treated with

200 mM potassium chromate for 16 h as described under “Materials andMethods.” The DNA-protein cross-links containing 150–200 mg of DNAwere analyzed in two-dimensional gels, and the molecular mass and theisoelectric point (pI) of major proteins resolved on the gel were deter-mined. Data presented here are from results of this study and ourprevious study (Ref. 35).

Molecular mass pI

kDa

108 5.4–5.898 5.2–5.674 (a)a 5.2–5.663 (m)a 5.2–5.453 (n)a 5.251 (d)a 8.8–9.249 (o)a 5.4–5.844 (b)a 5.343 (p)a 6.0–6.542 (c)a 5.840 4.8–5.036 5.0–5.2

36–38 (CNP)b 5.5–7.229 6.8

25–28 (CNP) 7.0–8.519 6.4–6.816 (CNP) 5.6–6.8

a Proteins marked in Fig. 3C. The letters a, b, c, and d correspond tothe proteins that cross-linked to DNA upon 25 mM chromate treatmentof MOLT4 cells for 16 h.

b CNP, cluster of nuclear proteins.

Chromium(VI)-induced DNA-Protein Cross-linking 33555

by guest on Novem

ber 20, 2020http://w

ww

.jbc.org/D

ownloaded from

sociated most of the proteins from the DNA-protein complexeswithout affecting the recovery of DNA. These data indicate thatchromate treatment induces the cross-linking of proteins toDNA and does not cause sedimentable protein aggregates andare consistent with a previous report (8) for chromate-inducedDNA-protein complexes in cultured Chinese hamster ovarycells.In order to test whether chromium is directly participating in

the DNA-protein complexes, EDTA in excess was used as achelating agent to examine whether chelation of chromiumwould disrupt the complex. As shown in Fig. 9, EDTA (50 mM)treatment of DNA-protein complexes (isolated from cells la-beled with [35S]methionine and [3H]thymidine) decreased therecovery of 35S radioactivity without affecting the recovery of3H activity. These results indicate that the decrease in 35Sactivity was not due to fragmentation of DNA. Dissociation of35S or 51Cr increased with EDTA concentration up to 50 mM,and no further dissociation was observed beyond this concen-tration (data not shown). The maximum decrease in 35S recov-ery after EDTA (50 mM) treatment was found to be approxi-mately 18% of the control (Fig. 9). Since Cr(III) has highaffinity for EDTA, the dissociation of some of the complexes byEDTA may, in part, be due to a chelatable form of chromium,such as Cr(III). Treatment of the complexes with b-mercapto-ethanol (2%) or thiourea (100 mM) decreased the recovery ofproteins to about 50 and 55% of control, respectively, withoutaffecting the recovery of DNA (Fig. 9), indicating the partici-pation of sulfhydryl groups in chromate-induced DNA-proteincomplexes.

DISCUSSION

Although previous studies have shown that carcinogenicCr(VI) induces DNA-protein cross-links, little is known aboutthe characteristics of the chromate-induced DNA-protein cross-links. It has generally been believed that the reduced form ofthe carcinogenic Cr(VI), Cr(III), gives rise to DNA-proteincross-links by directly mediating the cross-linking between thecellular DNA and protein (49). This belief is mainly based onthe fact that chromate-induced DNA-protein cross-links couldbe disrupted by EDTA, a chelator of Cr(III) but not of Cr(VI),and that the proteins in the DNA-protein complexes could bevisualized on SDS-polyacrylamide gels without nuclease diges-tion (50). In these studies, chromate-induced DNA-protein com-plexes were only partially disrupted by EDTA. Again, the na-ture of the cross-link would not be explained by resolution of aprotein on SDS-polyacrylamide gel electrophoresis if the sameprotein is cross-linked to DNA by different mechanisms, be-cause a single protein could possess many different reactivegroups to react with different cross-linking agents. Most otherin vitro studies supporting Cr(III) as the cross-linking agentare based on the fact that reducing agents are required in thereaction mixtures for Cr(VI)-induced cross-linking of DNA andprotein to take place (13, 51), a condition that generates reac-tive species capable of causing DNA-protein cross-linking (52).Therefore, the nature of chromate-induced DNA-protein cross-links is not fully resolved. We have previously reported thatchromate induces an oxidative stress in the cells (22, 23) andthat pretreatment of cells with vitamin E, an antioxidant,inhibits chromate-induced DNA-protein cross-linking (53). Theresults of the present study, for the first time, indicate thatchromate-induced DNA-protein complexes may be formed bythe generation of active oxygen species during the intracellular

FIG. 4. Localization of major proteins cross-linking to DNAupon chromate treatment of cells, in the cytoplasmic, nuclear,and nuclear matrix protein fractions. Proteins from cytoplasmicfraction (A), nuclear fraction (B), and nuclear matrix fraction (C) con-

taining 30 mg of protein were analyzed by nonequilibrium two-dimen-sional gel electrophoresis and were stained with silver stain. The pro-teins a, b, c, and d refer to the proteins that primarily cross-linked toDNA with 25 mM chromate treatment of intact cells for 16 h.

Chromium(VI)-induced DNA-Protein Cross-linking33556

by guest on Novem

ber 20, 2020http://w

ww

.jbc.org/D

ownloaded from

reduction of chromate.In the present study, DNA-protein cross-linking increased in

a dose- and time-dependent manner when MOLT4 cells weretreated with chromate (Figs. 1 and 2) and did not attain aplateau under the present experimental conditions. The in-crease in the formation of DNA-protein complexes was not dueto cell death, because short term exposure to chromate, whichdid not affect cell mortality, substantially increased DNA-pro-tein cross-linking, and cross-linked similar proteins to DNA. Aspecific group of nuclear non-histone proteins seems to partic-ipate in chromate-induced DNA-protein complexes. These pro-teins must reside in close vicinity to DNA in relatively highconcentrations, because a protein must reside in close proxim-ity to DNA, and its reactive groups should be oriented such thatthey are able to interact with that of DNA in order for it to becross-linked to DNA by any form of cross-linking agent. Presentstudies show that only four proteins (a, b, c, and d) were foundto be primarily complexed to DNA, although several otherproteins were seen in the nuclear protein fraction (Fig. 3B).Since chromate was required for the cross-linking of theseproteins with DNA, it is apparent that their selective interac-tion with chromiumwas necessary for the cross-linking of theseproteins to DNA. Other investigators have reported the asso-ciation of a 45-kDa protein (similar in molecular weight and pIto protein b) to DNA by chromium (6, 50, 55) platinum (8), andionizing radiation (56). The identity of proteins a, c, and dremains to be determined. Although histones constitute a sub-stantial part of the chromatin, these basic proteins were notcomplexed to DNA upon chromate treatment. This is consistentwith the findings of Miller et al. (50). Because Cr(III) has highaffinity for sulfur-containing ligands and there is scarcity ofcysteine residues among histones, it appears plausible that

FIG. 5.Effect of nuclease digestion on the resolution pattern ofchromate-induced DNA-protein complexes. DNA-protein com-plexes (equivalent to 2–3 A260 units) isolated from cells treated with 200mM chromate for 16 h were incubated in the presence or absence ofnucleases for 2 h at room temperature, concentrated, and analyzed by

FIG. 6. N-terminal sequencing of proteins complexed to DNAupon chromate treatment of MOLT4 cells. Following two-dimen-sional gel electrophoresis, proteins were electroblotted to PVDF mem-brane. The protein band of interest was trimmed off the blot, andEdman degradation was performed in an Applied Biosystems 477 Aprotein sequencer. A, N-terminal sequencing of p43 (protein p). B,N-terminal sequencing of p49 (protein n). C, N-terminal sequencing ofp44 (protein b). See “Materials and Methods” for details. *, can be anyamino acid. Numbers represent the number of the amino acids in therespective proteins (single-letter amino acid codes are used).

two-dimensional gels. A, two-dimensional gel of undigested DNA-pro-tein complexes. B and C, two-dimensional gels of DNase I- (300 mg/ml)and RNase A- (200 mg/ml) digested DNA-protein complexes, respec-tively. The proteins a, b, c, and d refer to the proteins that primarilycross-linked to DNA with 25 mM chromate treatment of intact cells for16 h. The gel in A was electrofocused for 1800 V-h. The gels in B and Cwere electrofocused for 2300 V-h. The proteins labeled as D and R arethe proteins in DNase I and RNase A, respectively.

Chromium(VI)-induced DNA-Protein Cross-linking 33557

by guest on Novem

ber 20, 2020http://w

ww

.jbc.org/D

ownloaded from

histones may not complex to DNA by chromate due to unavail-ability of appropriate ligands.We have tentatively identified three different proteins that

cross-link with DNA when cells are exposed to chromate. Theiridentity and the following plausible pathophysiological conse-quences are considered. (i) The 44-kDa acidic protein (proteinb) could be nuclear b-actin based on its identical molecularweight, isoelectric point, and homology with amino acids 19–26of b-actin. Furthermore, it positively reacted with an anti-actinpolyclonal antibody. This is in accord with the finding thatactin does cross-link with DNA in chromate-treated CHO cells(57). Actin is present in nucleus, nucleolus as well as thenuclear matrix (58). Actin has been shown to be associated withDNA replication, DNA repair, and RNA transcription (59–61).Hence, chromate-induced actin-DNA cross-linking may, atleast in part, lead to altered gene expression, as has recentlybeen shown for inducible genes (62). (ii) The homology of p43(protein p) microsequence with amino acids 24–29 of lectinBra-3 (Fig. 6) indicates that it could be a human lectin. Lectinshave not been previously shown as DNA-binding proteins. Al-though lectin receptors have been found on the cytoplasmicsurface of intracellular membranes such as the nuclear enve-lope and mitochondrial outer membrane, recent evidence indi-cates that lectin binding takes place on the noncytoplasmicsurface of these organelles (63). Lectins are located in a widevariety of cells and cell membranes, and alteration in theirlevels has been reported upon malignant transformation (64).Lectins are shown to play important roles in the developmentalprocesses (65). Lectins have also been shown to function as

receptors (65) and mitogenic regulators (66). (iii) The sevenN-terminal amino acids of the 49-kDa protein (labeled as n inFig. 3C) were found to have absolute homology with aminogly-coside nucleotidyltransferase, a protein previously not knownto bind to DNA. Because chromate-induced DNA-protein com-plexes predominantly occur in transcriptionally active DNA(62), it remains to be seen if these proteins are involved in thetranscription process. Nonetheless, the cross-linking of theseproteins to DNA could lead to serious physiological and geneticconsequences.The nature of chromate-induced DNA-protein complexes was

analyzed by enzyme and chemical treatment of the complexes.Treatment of DNA-protein complexes isolated from control orchromate-treated cell nuclei with DNase I dissociated most ofthe proteins associated with DNA (Fig. 9), indicating that thesedimentable nature of the proteins is due to the association ofproteins with the genomic DNA and not due to protein aggre-gation following chromate treatment or altered solubility of themetal-bound proteins. The small amount of DNA-protein com-plexes sedimented after DNase I digestion appears to be mostly

FIG. 8. Effect of antioxidants on chromate-induced DNA-pro-tein complexes. Cells were pretreated with different antioxidants for16 h in complete RPMI medium, washed, and then treated with 200 mM

chromate for 3 h. Following chromate treatment, cells were washed andsubjected to the K1-SDS procedure immediately or frozen at 270 °Cand then subjected to the K1-SDS procedure for detection of DNA-protein complexes as described under “Materials and Methods.” *, sig-nificantly different from control, p # 0.01 (n 5 5). DPC coefficient,DNA-protein cross-link coefficient (the ratio of K1-SDS-precipitableDNA in treated cells to that in control cells).

TABLE IIEffect of antioxidants on the cellular uptake of 51CrO4

22

Cells, in logarithmic growth phase, were treated with different con-centrations of K2

51CrO4 in SGM for 3 h at 37 °C. Cells were washedthree times in ice-cold phosphate-buffered saline, and cellular 51CrO4

22

uptake was determined as detailed under “Materials and Methods.”Each value is a mean of at least three different experiments 6 S.D.Values not significantly different from control have p # 0.05 (n 5 3–5).

51CrO422

5 nMK2

51CrO4

10 nMK2

51CrO4

15 nMK2

51CrO4

pmol/106 cells

Control 1.97 6 0.94 5.18 6 0.81 6.83 6 0.73a-Tocopherol succinate(25 mM)

2.26 6 0.26 4.86 6 0.84 7.31 6 0.89

Tiron (1 mM) 2.15 6 0.65 5.37 6 0.79 7.17 6 0.82Ascorbate (1 mM) 1.81 6 0.50 5.29 6 0.92 6.97 6 0.67Mannitol (10 mM) 2.07 6 0.63 4.90 6 0.68 7.07 6 0.88

FIG. 7. Two-dimensional gel analysis and immunoblotting of chromate-induced DNA-protein complexes for identification of actin.A, silver-stained two-dimensional gel of DNA-protein complexes containing 100 mg DNA. B, an autoradiogram of a duplicate gel of A electroblottedto nitrocellulose, reacted with anti-actin antibody and 125I-protein A. The proteins a, b, c, and d in A refer to the proteins that primarily cross-linkedto DNA with 25 mM potassium chromate treatment of intact cells for 16 h. The protein recognized by the anti-actin antibodies corresponds to the44-kDa acidic protein b.

Chromium(VI)-induced DNA-Protein Cross-linking33558

by guest on Novem

ber 20, 2020http://w

ww

.jbc.org/D

ownloaded from

in the form of stable chromium-nucleoprotein complexes. Thisis consistent with the findings of other investigators who havedemonstrated the resistance of chromium-bound nucleoli tonuclease digestion (67) and have shown the cross-linking ofnuclear matrix proteins to DNA by heavy metals and UV irra-diation (55, 68). The resistance of the DNA-protein complexesto RNase A digestion indicates that chromate treatment doesnot induce the formation of RNA-protein complexes. The sta-bility of the DNA-protein complexes was further assessed bymonitoring the resistance of the complexes to EDTA treatment.Treatment with EDTA caused dissociation of only 18% of 35Sactivity from the DNA-protein complexes. Because EDTA ef-fectively chelates Cr(III) but poorly binds with the oxyanion ofCr(VI), EDTA-dissociable proteins from DNA-protein com-plexes could have been mediated by a chelatable form of chro-mium such as Cr(III). However, the majority of the chromate-induced DNA-protein complexes were resistant to EDTAtreatment. These data suggest that the predominant form ofDNA-protein cross-links caused by chromate may not involvethe metal. Such cross-links could be generated by direct inter-action between DNA and protein via formation of proteinand/or DNA radicals produced during intracellular reduction ofchromate. It is also possible that some of the Cr(III) in thecross-link is not accessible to EDTA chelation. The dissociationof some of the chromate-induced DNA-protein complexes byb-mercaptoethanol or thiourea suggests that some of the cross-links involve sulfhydryl groups. Although Cr(III) is known toform complexes with sulfur-containing ligands, the -SH groupscould directly be involved in the complex via thiyl radicals ordisulfides that are generated during chromate-induced oxida-tive stress. The later contention is further supported by the factthat pretreatment of cells with antioxidants or free radicalscavengers such as a-tocopherol succinate, Tiron, and mannitoldid inhibit the chromate-induced DNA-protein cross-linking.The observed increase in chromate-induced DNA-protein cross-linking following ascorbate treatment may, in part, be due toan increase in the reduction of Cr(VI), leading to the increasedaccumulation of intracellular Cr(III), which may, in turn, giverise to DNA-protein complexes. Intracellular Cr(III) is predom-

inately generated by the reduction of Cr(VI), a process shown togenerate oxygen free radicals (52, 69). Free radical generatingsystems such as ionizing radiation as well as Fenton typereactions have been shown to cause DNA-protein cross-linking(56, 69). Collectively, these results suggest that free radicalsmay, at least in part, be involved in chromate-induced DNA-protein cross-linking.Free radical independent mechanisms may also play a role in

some chromate-induced DNA-protein cross-linking, becausethe electrophoretic conditions would not disrupt the radical-induced covalent DNA-protein cross-links, and we were able tovisualize the proteins cross-linked to DNA in two-dimensionalgels without digesting DNA. Furthermore, nuclease digestiondid not cause the appearance of additional proteins on two-dimensional gels. Visualization of proteins in two-dimensionalgels without nuclease digestion of the complexes may, at leastin part, be due to the reduction of sulfhydryl groups of proteinsby b-mercaptoethanol and the presence of a high concentrationof urea in the sample buffer, leading to disruption of the com-plex. Such mechanisms have been shown to be the leadingcause for dissociation of chromate- and cisplatin-induced DNA-protein complexes (55). Another possibility is that the sameproteins are complexed to DNA via both the Cr(III) and oxida-tive mechanisms due to interaction of different amino acidresidues with DNA. In that case, two-dimensional gel patternsof DNA-protein complexes may be the same with or withoutnuclease digestion. That this was the case was shown by diges-tion of the complexes with DNase I and RNase A (Fig. 6).In view of the difficulties and limitations of the methods, it is

not possible to propose a definite mechanism(s) of action ofchromate in inducing DNA-protein complexes at this time.However, it is likely that chromate-induced DNA-protein com-plexes are formed via more than one mechanism and that atleast some of the complexes are noncovalent in nature. There-fore, use of the term “cross-link” to indicate the association ofproteins with DNA may not be appropriate, although we haveused this term to express the association of DNA and protein,as has been used in the literature.The results presented in this study indicate that chromate

treatment of cells complexes a selected group of non-histoneproteins to DNA. Actin, lectin, and aminoglycoside nucleotidyl-transferase are among other proteins that appear to participatein chromate-induced DNA-protein complexes. The exact natureof the interaction between the DNA and protein remains to bedetermined. However, our results suggest both the participa-tion of a chelatable form of chromium such as Cr(III) and theinvolvement of oxidative mechanisms in the process of chro-mate-induced DNA-protein cross-linking. Our results also sug-gest the involvement of sulfhydryl groups in chromate-inducedDNA-protein cross-links. Although chromate-induced DNA-protein complexes are found to be resistant to treatments suchas 2% SDS and 5 M urea, their reversibility in the gel electro-phoretic conditions indicates that their association is in theform of noncovalent interactions. These characteristics of chro-mate-induced DNA-protein complexes suggest that it is possi-ble to use chromium in studies involving chromatin structureas well as identification of proteins participating in DNA-pro-tein interactions, specifically those that undergo transient in-teraction with DNA, such as transcription factors.

REFERENCES

1. Banjar, Z. M., Hnilica, L. S., Briggs, R. C., Stein, J., and Stein, G. (1983)Biochem. Biophys. Res. Commun. 114, 767–773

2. Oleinick, N. L., Chiu, S., Ramakrishnan, N., and Xue, L. (1987) Br. J. Cancer55, Suppl. 8, 135–140

3. Grunicke, H., Bock, K. W., Becker, H., Gang, V., Schnierda, J., and Puschen-dorf, B. (1973) Cancer Res. 33, 1048–1053

4. Cosma, G. N., Jamasbi, R., and Marchok, A. (1988) Mutat. Res. 201, 161–1685. Kuykendall, J. R., Taylor, M. A., and Bogdanffy, M. W. (1993) Toxicol. Appl.

FIG. 9. Analysis of composition and stability of chromate-in-duced DNA-protein complexes. DNA-protein complexes obtainedfrom chromate-treated cells (200 mM, 16 h) by following the SDS/ureaextraction method were treated with different chemicals and enzymes,as shown in the figure, and were ultracentrifuged to determine theeffect of these agents on the recovery of DNA and protein in the pellet.See “Materials and Methods” for details. Each value is a mean of fiveindependent experiments 6 S.D. *, significantly different from controlp # 0.01. 2-ME, 2-mercaptoethanol.

Chromium(VI)-induced DNA-Protein Cross-linking 33559

by guest on Novem

ber 20, 2020http://w

ww

.jbc.org/D

ownloaded from

Pharmacol. 123, 283–2926. Wedrychowski, A., Ward, W. S., Schmidt, W. N., and Hnilica, L. S. (1985)

J. Biol. Chem. 260, 7150–71557. Sugiyama, M., Patierno, S. R., Catoni, O., and Costa, M. (1986) Mol. Pharma-

col. 29, 606–6138. Miller, C. A., III, and Costa, M. (1989) Mol. Toxicol. 2, 11–269. Patierno, S. R., and Costa, M. (1985) Chem. Biol. Interact. 55, 75–9110. Banjar, Z. M., Hnilica, L. S., Briggs, R. C., Stein, J., and Stein, G. (1984)

Biochemistry 23, 1921–192611. Cohen, M. D., Klein, C. B., and Costa, M. (1992) Mutat. Res. 269, 141–14812. Cohen, M., Latta, D., Coogan, T., and Costa, M. (1990) in The Biological Effects

of Heavy Metals (Foulkes, E. C., ed) Vol. 2, pp. 19–76, CRC Press, Inc., BocaRaton, FL

13. Fornace, A. J., Jr., Seres, D. S., Lechner, J. F., and Harris, C. C. (1981) Chem.Biol. Interact. 36, 345–354

14. Koster, A., and Beyersmann, D. (1985) Toxicol. Environ. Chem. 10, 307–31315. Jennette, K. W. (1981) Environ. Health Perspect. 40, 223–25216. Arslan, P., Beltrame, M., and Tomasi, A. (1987) Biochem. Biophys. Acta 931,

10–1517. Connett, P. H., and Wetterhahn, K. E. (1983) Struct. Bond 54, 93–12418. Nieboer, E., and Jusys, A. (1988) in Chromium in Natural and Human Envi-

ronment: Advances in Science and Technology (Nieboer, E, and Nriagu, J.O., eds) Vol. 20, pp. 21–79, John Wiley & Sons, Inc., New York

19. Tsapakos, M. J., and Wetterhahn, K. E. (1983) Chem. Biol. Interact. 46,265–277

20. Earley, J. E., and Cannon, R. D. (1965) in Transition Metal Chemistry (Carlin,R. L., ed) Vol. 1, p. 33, Marcer Dekker, New York

21. DeFlora, S., and Wetterhahn, K. E. (1989) Life Chem. Rep. 7, 169–27722. Mattagajasingh, S. N., and Misra, H. P (1995) Mol. Cell. Biochem. 142, 61–7023. Mattagajasingh, S. N., and Misra, H. P. (1997) Toxic Sub. Mech. J., 16, 63–7924. Kawanishi, S., Inoue, S., and Sano, S. (1986) J. Biol. Chem. 261, 5952–598925. Shi, X., and Dalal, N. S. (1990) Arch. Biochem. Biophys. 281, 90–9526. Margolis, S. A., Coxon, B., Gajewski, E., and Dizdaroglu, M. (1988) Biochem-

istry 27, 6353–635927. Gajewski, E., and Dizdaroglu, M. (1990) Biochemistry 29, 977–98028. Shi, X., and Dalal, N. S. (1990) Arch. Biochem. Biophys. 277, 342–35029. Sugden, K. D., Geer, R. D., and Rogers, S. J. (1992) Biochemistry 31,

11626–1163130. Kortenkamp, A., Ozolins, Z., Beyersmann, D., and O’Brien, P. (1989) Mutat.

Res. 216, 19–2631. Bedinger, P., Hochstrasser, M., Jongeneel, C. V., and Alberts, B. M. Cell 34,

115–12332. Briggs, J. A., and Briggs, R. C. (1988) Cancer Res. 48, 6484–649033. Bouck, N. P., and Benjamin, B. K. (1989) Chem. Res. Toxicol. 2, 1–1134. Stein, G. S., and Kleinsmith, L. J. (eds) (1979) Chromosomal Proteins and

Their Role in Regulation of Gene Expression, Academic Press, Inc., NewYork

35. Mattagajasingh, S. N., and Misra, H. P. (1995) in Methods in Proteins Struc-ture Analysis (Atassi, M. Z., and Appella, E., eds) pp. 295–307, PlenumPress, New York

36. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning: ALaboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

37. Bradford, M. M. (1976) Anal. Biochem. 72, 248–25438. O’Farrell, P. Z., Goodman, H. M., and O’Farrell, P. H. (1977) Cell 12,

1133–1142

39. Laemmli, U. K. (1970) Nature 227, 680–68540. Sammons, D. W., Adams, L. D., and Nishiziwa, E. E. Electrophoresis 2,

135–14141. Kaufmann, S. H., Coffey, D. C., and Shaper, J. H. (1981) Exp. Cell Res. 132,

105–12342. Mattagajasingh, S. N., and Misra H. P. (1994) J. Protein Chem. 13, 449–45043. Burnette, W. N. (1981) Anal. Biochem. 112, 195–20344. Liu, L. F., Rowe, T. C., Yang, L., Tewey, K. M., and Chen, G. L. (1983) J. Biol.

Chem. 258, 15365–1537045. Trask, D. K., DiDonato, J. A., and Muller, M. T. (1984) EMBO J. 3, 671–67646. Zhitkovich, A., and Costa, M. (1992) Carcinogenesis 13, 1485–148947. Fabianek, J., DeFilippi, J., Rickards, T., and Herp, A. (1968) Clin. Chem. 14,

456–46248. Sugiyama, M., Tsuzuki, K., Matsumoto, K., and Ogura, R (1992) Photochem.

Photobiol. 56, 31–3449. Standeven, A. M., and Wetterhahn, K. (1991) Chem. Res. Toxicol. 4, 616–62550. Miller, C. A., III, and Costa, M (1988) Mol. Carcinogen. 1, 125–13351. Kortenkamp, A., O’Brien, P., and Beyersmann, D. (1991) Carcinogenesis 12,

1143–114452. Jones, P., Kortenkamp, A., O’Brien, P., Wang, G., and Yang, G. (1991) Arch.

Biochem. Biophys. 286, 652–65553. Mattagajasingh, S. N., and Misra, H. P. (1993) FASEB J. 7, 46954. Lesko, S. A., Drocourt, J.-L., and Yang, S.-U. (1982) Biochemistry 21,

5010–501555. Wedrychowski, A., Schmidt, W. N., and Hnilica, L. S. (1986) J. Biol. Chem.

261, 3370–337656. Chiu, S., Friedman, L., Sokany, N., Xue, L., and Oleinick, N. (1986) Radiat.

Res. 107, 24–3857. Miller, C. A., III, Cohen, M. D., and Costa, M. D. (1991) Carcinogenesis 12,

269–27658. Peters, K. E., Okada, T. A., and Comings, D. E. (1982) Eur. J. Biochem. 129,

221–23259. Scheer, U., Hinssen, H., Franke, W. W., and Jockusch, B. M. (1984) Cell 39,

111–12260. Eagly, J. M., Miyamoto, N. G., Moncollin, V., and Chambon, P. (1984) EMBO

J. 3, 2363–237161. Rungger, D., Rungger-Brandle, E., Chaponnier, C., and Gabbiani, G. (1979)

Nature 282, 320–32162. Hamilton, J. W., and Wetterhahn, K. E. (1989) Mol. Carcinogen. 2, 274–28663. Lis, H., and Nathan, S. (1986) in The Lectins: Properties, Function, and

Applications in Biology and Medicine (Liener, I., Sharon, N., and Goldstein,I. J., eds.) pp. 293–370, Academic Press, Inc., New York

64. Gabius, H., Engelhardt, R., Graupner, G., and Cramer, F. (1986) in Lectins:Biology, Biochemistry, Clinical Biochemistry (Bog-Hansen, T. C., and vanDriessche, E., eds) Vol. 5, pp. 237–242, Walter de Gruyter & Co., Berlin

65. Kolb-Bachofen, V. (1986) in Lectins: Biology, Biochemistry, Clinical Biochem-istry (Bog-Hansen, T. C., and van Driessche, E., eds) pp. 197–206, Walter deGruyter & Co., Berlin

66. Lis, H, and Nathan, S. (1986) in The Lectins: Properties, Function, and Appli-cations in Biology and Medicine (Liener, I., Sharon, N., and Goldstein, I. J.,eds) pp. 265–291, Academic Press, Inc., New York

67. Ono, H., Wada, O., and Ono, T. (1981) J. Toxicol. Environ. Health 8, 947–95768. Bouliakas, T. (1986) Cell Biol. 64, 474–48469. Shi, X., and Dalal, N. S. (1989) Biochem. Biophys. Res. Commun. 163, 627–634

Chromium(VI)-induced DNA-Protein Cross-linking33560

by guest on Novem

ber 20, 2020http://w

ww

.jbc.org/D

ownloaded from

Subhendra N. Mattagajasingh and Hara P. Misraand Their Characterization in Cultured Intact Human Cells

Mechanisms of the Carcinogenic Chromium(VI)-induced DNA-Protein Cross-linking

doi: 10.1074/jbc.271.52.335501996, 271:33550-33560.J. Biol. Chem. 

  http://www.jbc.org/content/271/52/33550Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/271/52/33550.full.html#ref-list-1

This article cites 61 references, 8 of which can be accessed free at

by guest on Novem

ber 20, 2020http://w

ww

.jbc.org/D

ownloaded from