Mutational Spectrum Induced by Chromium(III) in Shuttle Vectors Replicated in Human Cells: Relationship to Cr(III)−DNA Interactions

Mutational Spectrum Induced by Chromium(III) inShuttle Vectors Replicated in Human Cells:Relationship to Cr(III)-DNA Interactions

Tsui-Chun Tsou, Ren-Jye Lin, and Jia-Ling Yang*

Molecular Carcinogenesis Laboratory, Department of Life Sciences, National Tsing Hua University,Hsinchu 300, Taiwan, Republic of China

Received March 18, 1997X

Trivalent chromium (Cr(III)), the ultimate species of chromium(VI) intracellular reduction,can associate with DNA forming Cr(III) monoadducts and DNA-DNA cross-links. However,the mutational specificity of Cr(III) has not been determined partly because Cr(III) has difficultyentering cells. In this study, we have characterized the types of Cr(III)-induced DNA lesionsin two buffer systems and the mutational spectrum of Cr(III)-treated shuttle vectors replicatedin human 293 cells. Plasmids were treated with Cr(III) in buffers consisting of either 10 mMpotassium phosphate, pH 7.5 (designated as KP), or 0.2 mM Tris-HCl and 20 µM EDTA, pH7.4 (designated as TE/50). The amounts of Cr(III) bound to DNA increased as Cr(III)concentration increased in both buffers; these Cr(III)-DNA associations were stable in bothbuffers during a 24-h dialysis. The electrophoretic mobility of supercoiled DNA was markedlyretarded in samples treated with Cr(III) in TE/50 but not KP buffer, suggesting that Cr(III)-mediated DNA-DNA cross-links were generated in TE/50 but did not form in KP. Polymerase-stop assay showed that DNA polymerases were mostly blocked at the 3′ adjacent bases ofguanines on templates treated with Cr(III) in TE/50 but were not observed on those treated inKP. The signals of Cr(III)-mediated cross-links generated in TE/50 buffers were reduced whenthey were dialyzed against KP buffers. Similarly, Cr(III)-DNA monoadducts formed in KPwere converted to primer-template cross-links by dialysis against TE/50. The mutationfrequency of Cr(III) in the supF gene of pSP189 or pZ189 shuttle vectors replicated in humancells increased as Cr(III) concentration increased in both buffers. DNA sequencing analysisshowed that single-base substitutions (61-68%), two-base substitutions (3-5%), and deletions(21-34%) were induced in similar frequencies in plasmids treated with Cr(III) in either TE/50 or KP. The Cr(III)-induced base-substitution hot spots are different from those occurringspontaneously. Cr(III) enhances G‚C base substitutions, particularly G‚C f C‚G transversions,at 5′GA, 5′CG, and 5′AG sites. Base-substitution hot spots did not correlate with strongpolymerase-stop sites, suggesting that base substitutions are derived from Cr(III) monoadducts,not from DNA-DNA cross-links.

Introduction

Although Cr(III)1 is an essential trace element inglucose and lipid metabolism (1), Cr(VI) compoundsincrease risks of respiratory tract cancers (2, 3). Cr(VI)also induces chromosomal abnormalities (4, 5), celltransformations (6), apoptosis (7), signal transductions(8, 9), and gene mutations (10, 11) in cultured mam-malian cells. Cr(VI) compounds are more effective thanCr(III) compounds at inducing cytotoxicity and carcino-genicity because the cellular uptake capability of theformer is significantly higher than that of the latter (2,3, 12). Nevertheless, Cr(VI) does not directly interactwith DNA in vitro (13-15). Once Cr(VI) enters cells, itis believed to exert genotoxicity through metabolic acti-vation (16, 17). The ultimate kinetically stable Cr(III)and several reactive species, including short-lived chro-mium intermediates and ROS, are generated upon Cr-(VI) reduction (16-20). Exposing cultured cells to Cr(VI)

generates several types of DNA damage including single-strand breaks, radical-DNA adducts, and Cr-DNA ad-ducts, as well as Cr-mediated DNA-DNA and DNA-protein cross-links (7, 21-24).Intercellular chromium contents can rapidly reach

millimolar levels, hundreds of times greater than extra-cellular chromium concentrations (25, 26), and accumu-late in cell nuclei (25). Although Cr(III) is inefficient atentering cells (2, 3, 12), the major form of chromiuminside the cells has been estimated to be Cr(III) (23, 24),which can subsequently bind to DNA and proteins (13-15, 20). In the presence of H2O2, Cr(III) induces mark-edly more hydroxyl radicals than Cr(VI) does and sub-sequently leads to strand breakage and the formation of8-hydroxydeoxyguanosine in DNA (27, 28). Hexacoordi-nate Cr(III) aromatic bidentate amine complexes inducemutations in Salmonella typhimurium that may resultfrom DNA damage generated by an active redox center(29). Furthermore, Cr(III) can enhance mutations insingle-stranded phage DNA replicated in Escherichia coli(30, 31). At low concentrations, Cr(III) increases DNApolymerase processivity and decreases the fidelity of DNAreplication (30, 31), whereas DNA-DNA cross-linksinduced by high concentrations of Cr(III) stop DNAsynthesis (32). Cr(III) may thus be an important species

* Author to whom correspondence should be addressed. Fax: 886-3-5721746. E-mail: [email protected].

X Abstract published in Advance ACS Abstracts, August 1, 1997.1 Abbreviations: Cr(III), trivalent chromium; Cr(VI), hexavalent

chromium; TE/50, 0.2 mM Tris-HCl and 20 µM EDTA (pH 7.4); KP,10 mM potassium phosphate (pH 7.5); ROS, reactive oxygen species;ICP-MS, inductively coupled plasma-mass spectrometer.

962 Chem. Res. Toxicol. 1997, 10, 962-970

S0893-228x(97)00040-4 CCC: $14.00 © 1997 American Chemical Society

leading to Cr(VI) mutagenesis in cells.Shuttle vector systems have been applied to examine

the molecular mutagenesis of many environmental car-cinogens (33). The mutational spectra induced by aparticular carcinogen, e.g., benzo[a]pyrenediol epoxide,in shuttle vector systems (34) almost always reflect thatobserved in the endogenous genes of mammalian cellshaving similar DNA repair capabilities (35). In thisstudy, we adopted shuttle vector systems to examine themutational spectrum induced by Cr(III). Shuttle vectorscarrying the supF target gene were treated with Cr(III)in either phosphate-based (i.e., 10 mM potassium phos-phate, pH 7.5 (KP)) or Tris-EDTA-based (i.e., 0.2 mMTris-HCl and 20 µMEDTA, pH 7.4 (TE/50)) buffers. TheCr(III)-treated plasmids were transfected into the humancell line 293 and allowed 2 days for replication. Progenyplasmids were then rescued, purified, and introduced intoindicator bacterial strains. Mutant plasmids were am-plified and subsequently analyzed using DNA sequenc-ing. Moreover, the amounts of Cr(III) bound to plasmidsand the Cr(III)-induced DNA polymerase-stop sites gen-erated in the supF gene were determined. The relation-ship between the mutational specificity and types of DNAlesions generated in Cr(III)-treated plasmids in these twobuffers was compared to explicate the role of Cr(III) inmutagenesis during Cr(VI) intracellular reduction.

Experimental Procedures

Plasmids and Cells. Shuttle vectors pZ189 and pSP189 andthe E. coli strain MBM7070 were provided by Dr. M. M.Seidman (Otsuka Pharmaceuticals, Rockville, MD). Both pZ189and pSP189 vectors contained the mutagenic target supF, atyrosine amber-suppressor tRNA gene, flanked by the ampicillingene and the bacterial origin of replication (36, 37). Thesevectors also carried the replication origin and large-T antigengene from simian virus 40. The shuttle vector pSP189 alsocontained an 8-bp signature sequence with 2 × 48 possibleunique plasmids to permit unambiguous identification of inde-pendent mutant clones following mutagenesis (37). The num-bering used in this study, except where otherwise indicated, isthat of the gene in pZ189 where the unique EcoRI site GAATTCis denoted as position 1.The human embryonic kidney cell line, 293, which served as

the eukaryotic hosts for shuttle vectors and the E. coli SY204cells were obtained from Dr. V. M. Maher (Michigan StateUniversity, MI). The E. coli MBM7070 (36) and SY204 hostcells (38) carrying an lacZ amber mutation were used todistinguish plasmids containing supF mutations from thosecontaining wild-type supF sequences. The human 293 cells weregrown in Dulbecco’s modified Eagle media (Life Technologies,Grand Island, NY) supplemented with sodium bicarbonate(2.2%, w/v), L-glutamine (0.03%, w/v), penicillin (100 units/mL),streptomycin (100 µg/mL), and fetal calf serum (10%). Humancell cultures were maintained at 37 °C in a humidified incubatorcontaining 10% CO2 in air. The E. coli cells were grown inLuria-Bertani media at 37 °C.Determination of Cr(III) Binding to DNA. CrCl3‚6H2O

(99.995% pure) was purchased from Aldrich Chemical (Milwau-kee, WI). Plasmid pSP189 (10 µg, 3.08 × 10-8 mol of nucle-otides) was treated with CrCl3 (1-1000 µM, freshly preparedin H2O) at 37 °C for 30 min in 50 µL of either TE/50 or KP.Immediately after treatment, the reaction mixtures were passedthrough Sephadex G-50 columns (Boehringer Mannheim GmbH,Germany) at 2700 rpm for 4 min to remove the unbound Cr-(III) ions. One set of those Cr(III)-treated samples was placedon a membrance (VSWP 02500, Millipore, Bedford, MA),dialyzed against TE/50 or KP buffer at room temperature for24 h (drop dialysis), and then passed through Sephadex G-50columns. The amount of Cr(III) bound to the plasmid wasdetermined using an inductively coupled plasma-mass spec-

trometer (ICP-MS; SCIEX ELAN 5000, Perkin Elmer, Norwalk,CT). The ICP-MS conditions were set as follows: power, 5000W; plasma flow rate, 15 L/min; auxiliary flow rate, 0.8 L/min;and sample flow rate, 0.8 L/min. Cr(III) standard concentra-tions were prepared by series dilutions of 1000 mg/L chromium-(III) nitrate (Merck, Darmstadt, Germany) to 1-200 ppb ineither TE/50 or KP. The numbers of chromium adducts werecalculated and expressed as the numbers of chromiummoleculesbound per 1000 nucleotides.

Agarose Gel Electrophoresis and Southern Analysis.Immediately after Cr(III) treatment, DNA samples were ana-lyzed by electrophoresis in 0.8% agarose gels containing ethid-ium bromide (0.2 µg/mL). The gel was then placed on atransilluminator UV box and photographed from above withPolaroid type 665 positive/negative film. The band intensitiesof negatives were determined using a computing densitometerequipped with the ImagQuant analysis program (MolecularDynamics, Sunnyvale, CA). DNA samples in gels were thendenatured and Southern transferred to nylon membranes. Themembrane was washed, dried, and exposed to UV light (1200× 100 µJ/cm2) using a XL-1000 UV cross-linker (Spectronics Co.,Westbury, NY). The immobilized DNA samples in the mem-brane were hybridized with 32P-labeled probes which wereprepared from [R-32P]dCTP (specific activity, 800 Ci/mmol;Amersham, Little Chalfont, U.K.) and the NEBlot randomprimer kit (BioLabs, Beverly, MA). Hybridization was per-formed in a hybridization oven (Hybaid Ltd., Middlesex, U.K.).Next, the membrane was washed, dried, and plastic wrapped.Either phosphor screens or X-ray films were exposed to radio-activity on the membrane. The band intensities on the phosphorscreens were quantified using a PhosphorImager equipped withthe ImagQuant analysis program (Molecular Dynamics).

Polymerase-Stop Assay of Cr(III)-Treated Templates.The ability of chromium adducts to interfere with DNA syn-thesis was determined by the in vitro DNA polymerase-stopassay described previously, with modifications (32, 39). Briefly,EcoRI- and DpnI-linearized pZ189 were extracted with phenol/chloroform and further purified using a Centricon-30 concentra-tor (Amicon, Danvers, MA). The DNA sample (4 µg) was thentreated with CrCl3 in 15 µL of either TE/50 or KP as describedabove. The unbound Cr(III) was removed by drop dialysisagainst TE/50 or KP buffer at room temperature. The dialyzedDNA samples were used as templates for polymerase-stop assaythat was carried out in a manner similar to the DNA sequencingreaction, except that the dideoxyribonucleotides were omitted.To obtain Cr(III)-mediated polymerase-stop sites on the twostrands of the supF gene, [γ-32P]ATP (specific activity, 5000Ci/mmol; Amersham) end-labeled primer 1, 5′225ACGGGGTCT-GACG213, and primer 2, 5′-26GTATCACGAGGCCCT-12, wereannealed with the EcoRI- and DpnI-linearized templates,respectively. The Cr(III)-treated DNA templates (2 µg) andγ-32P-labeled primers (2.3 pmol) were mixed in a 10 µL buffercontaining 40 mM Tris-HCl (pH 7.5) and 10 mM MgCl2. Thetemplate-primer mixture was then heated at 95 °C for 3 minand rapidly cooled on ice. Next, 1-2 units of diluted Sequenase(US Biochemicals, Cleveland, OH) and four deoxyribonucleotides(final concentrations of 12.5 µM of each were used) were addedto 2.5 µL of the template-primer mixture. The polymerizationreaction was performed at 37 °C for 5 min. Dideoxy sequencingreaction was also performed on the untreated DNA template toprovide DNA size markers. The products of the polymerase-stop assays and dideoxy sequencing reactions were analyzed byelectrophoresis in a 6% polyacrylamide gel.

Mutagenesis Assay. The mutagenicity of DNA containingCr(III) adducts was assayed using the shuttle vector systemdescribed previously (34). Briefly, 10 µg of untreated or Cr-(III)-treated plasmids (purified with Sephadex G-50 columns)were mixed with 1.5 mL of HEPES-buffered solution containingcalcium and phosphate (40). The DNA mixtures were trans-fected into human 293 cells (8 × 105) on 150-mm diameterdishes. After a 2-day replication period, plasmids were isolatedfrom the cells by alkaline lysis and treated withDpnI restrictionendonuclease to remove unreplicated plasmids. The replicated

DNA Damage and Mutational Spectrum of Cr(III) Chem. Res. Toxicol., Vol. 10, No. 9, 1997 963

plasmids were introduced into competent E. coli SY204 orMBM7070 cells, respectively, using calcium chloride (for pZ189)or electroporation (for pSP189) (40). The electroporation wasconducted by pulsing the competent bacteria-DNA mixtureonce using a Cell-Porator (Life Technologies) set at 1500 V, 330µF, and 4000 Ω. The transformed cells were then assayed forampicillin resistance and mutations in the supF gene. Bacteriareceiving plasmids with supF mutations formed white or lightblue colonies, whereas those receiving a functional supF geneformed blue colonies on agar plates containing ampicillin,5-bromo-4-chloro-3-indolyl â-D-galactopyranoside, and isopropylâ-D-thiogalactopyranoside. Mutant colonies were restreaked onthose agar plates to confirm the phenotypes. The plasmids withmutant supF genes were amplified, purified, and then analyzedusing electrophoresis in 0.8% agarose gels. Mutant plasmidswere sequenced according to the dideoxy sequencing procedureusing Sequenase and primer 1 (for mutants derived from pZ189)or primer 3, 5′GGCGACACGGAAATGTTGAA (65-46 basesupstream of EcoRI site; for mutants derived from pSP189).Either [R-35S]dATP (specific activity, 1000 Ci/mmol; Amersham)or [γ-32P]ATP end-labeled primer was used for autoradiographicdetection of the sequence. Mutant plasmids with supF deletionswere subjected for restriction mapping by EcoRI, BamHI, andHindIII.

Results

Characterization of Cr(III)-Treated Plasmids.Shuttle vector pSP189 was treated with various concen-trations of CrCl3 in either TE/50 or KP buffer. Followingremoval of unbound Cr(III) ions by Sephadex G-50 gelfiltration, the amounts of chromium bound to plasmidswere assayed by ICP-MS. Only a background level ofchromium was recovered when CrCl3 was applied to theG-50 column. As shown in Figure 1, the amounts ofchromium adducts per plasmid increased linearly whenCrCl3 concentrations in both TE/50 (open circles) and KP(open triangles) buffers were increased. The capabilityof Cr(III) bound to DNA in TE/50 was much higher thanthat in KP at the same dose ranges. The amounts ofchromium bound per 1000 nucleotides generated by 200µM CrCl3 in TE/50 and KP were ∼170 and ∼25, respec-tively. At 500 µMCrCl3, the amounts of chromium bound

per 1000 nucleotides increased to ∼150 in KP buffers.Conversely, DNA samples treated with 500 µM CrCl3 inTE/50 could not be recovered from the G-50 columns.One other set of the Cr(III)-DNA samples was drop

dialyzed for 24 h before passing through G-50 columnsto examine the stability of Cr(III)-DNA interactions.Figure 1 shows that the numbers of Cr(III) bound to DNAwere approximately the same in Cr(III)-treated DNAwith or without a 24-h dialysis against buffers used forCr(III) treatment. Furthermore, 92.0 ( 3.1% (mean (SD; 3 determinants) of chromium remain bound to DNAin KP buffer when those G-50-purified samples wereincubated for 24 h with EDTA at a 2-fold molar excessover the Cr(III) concentration and then drop dialyzed.Approximately 80% (78.2 ( 6.2%; average of 9 determi-nants) of chromium remain bound to DNA in TE/50buffer when those samples were chelated with EDTA ata 5-fold molar excess over the Cr(III) concentration.These results indicate that Cr(III) ions associate withDNA in both buffers. The above results also suggest thatdifferent types of Cr(III)-DNA associations may begenerated by these two buffer systems.Figure 2 shows agarose gel patterns and Southern blots

of supercoiled plasmids treated with Cr(III) in vitro. Theelectrophoretic mobility of supercoiled plasmids wasmarkedly reduced when DNA samples were treated withg100 µM Cr(III) in TE/50 buffer (Figure 2, lanes 3-6).On the other hand, when supercoiled plasmids werereacted with millimolar levels of Cr(III) in KP buffer,their electrophoretic mobilities remain the same as thatof untreated DNA (Figure 2, lanes 7-11). The phenom-enon that reaction buffers affect the electrophoreticmobility of Cr(III)-treated supercoiled plasmids has sug-

Figure 1. Determination of numbers of Cr(III) bound perplasmid. Supercoiled pSP189 was treated with CrCl3 at 37 °Cfor 30 min in either TE/50 buffer (circles) or KP buffer(triangles). The reaction mixtures were purified by SephadexG-50 gel filtration (open symbols) and analyzed by ICP-MS asdescribed in Experimental Procedures. One set of Cr(III)-treatedsamples was subjected for a 24-h dialysis before G-50 purifica-tion (filled symbols). Data were obtained by averaging 2-5experiments, and the bars represent SD.

Figure 2. Effect of Cr(III) on the electrophoretic mobility ofsupercoiled plasmids. Supercoiled plasmids (660 ng) werereacted with Cr(III) in either TE/50 buffer (lanes 1-6) or KPbuffer (lanes 7-11) as described in Experimental Procedures.The Cr(III) concentrations are indicated in panel a. Immediatelyafter Cr(III) treatment, DNA samples were analyzed by elec-trophoresis in an agarose gel containing ethidium bromide(panel a). The relative intensity of ethidium bromide fluores-cence within areas marked with brackets (i.e., the position ofsupercoiled DNA) was determined by a densitometer. DNAsamples in gels shown in panel a were denatured, Southerntransferred to a nylon membrane, and hybridized with a R-32P-labeled probe. The radioactivity of DNA was visualized byexposing to a phosphor screen (panel b). The relative amountof supercoiled DNA on the blots (area within brackets) wasdetermined using a PhosphorImager.

964 Chem. Res. Toxicol., Vol. 10, No. 9, 1997 Tsou et al.

gested that Cr(III)-mediated DNA-DNA cross-links weregenerated in TE/50 but did not form in KP buffer. Theagarose gel patterns also showed a reduced ethidiumbromide fluorescence intensity of supercoiled plasmidstreated with Cr(III) in both buffers. This quench effectwould suggest that Cr(III) may interfere with ethidiumbromide bound to DNA. Additionally, DNA aggregateswere observed in samples treated with 0.8 and 4 mM Cr-(III) in TE/50 and KP buffers, respectively.Polymerase-Stop Assay. Polymerase-stop assay was

performed to determine how templates containing Cr-(III) adducts interfered with DNA replication. PlasmidpZ189 was linearized with EcoRI or DpnI, purified, andtreated with CrCl3 (25-200 µM) in TE/50, and unboundCr(III) was removed using dialysis before the polymerase-stop assay. Complete polymerization was verified by theformation of a full-length product: i.e., 225 bases usingprimer 1 and EcoRI-linearized templates and 233 basesusing primer 2 and DpnI-linearized templates. Figure3 shows that EcoRI-linearized templates treated by CrCl3in TE/50 generated unique patterns in three regions ascompared with untreated templates. At Cr(III) concen-trations g 100 µM, polymerase was strongly blocked nearthe primer sites (region III). Several polymerase-stopsites (region II) were found in CrCl3-treated templates,and the highest signal intensities were observed intemplates treated with 50 µMCrCl3. These results showthat the DNA replication was mostly blocked at one baselocated at the 3′ side of a guanine on the template strand.Furthermore, full-length DNA products decreased withincreases in CrCl3 concentrations (Figure 3, top of regionII). At 50 µM, full-length DNA signals were 50% of thosefrom untreated templates; such signals were not observedin templates treated with g100 µMCrCl3 (Figure 3). Theradioactivity in the loading wells markedly increased asCrCl3 concentrations were increased; the relative intensi-ties were 3-, 11-, 33-, and 51-fold of those observed foruntreated templates when EcoRI-linearized DNA wastreated with 25, 50, 100, and 200 µM CrCl3, respectively(Figure 3, region I; average of 3-5 experiments). Theseenhanced radioactive signals in the loading well havebeen considered as primer-template cross-links (32).Similar results were obtained in DpnI-linearized pZ189treated with CrCl3.The polymerization was also carried out on templates

treated with CrCl3 in KP. Interestingly, apparentlyneither polymerase-stop sites nor radioactivity in theloading wells was enhanced above background levelswhen templates were treated with 50-1000 µM CrCl3in KP (data not shown). To examine the stability of Cr-(III)-mediated DNA cross-links generated in TE/50, Cr-(III)-treated DNA samples were drop dialyzed against oneof the two buffers before the polymerase-stop assay wasperformed. The intensities of the polymerase-stop sitescaused by Cr(III) in TE/50 were reduced after a 4-hdialysis against KP (data not shown); additionally, theradioactivity in the loading wells of the sequencing geldecreased (Figure 4, lanes 2 and 3). When the DNAsamples were treated with 50 µM CrCl3 in KP and thendialyzed against TE/50, the radioactivity in the loadingwells of the sequencing gel increased as dialysis periodsincreased (Figure 4, lanes 4-6), but still no significantpolymerase-stop sites were observed (data not shown).This result suggests that buffer environments affect thetypes of Cr(III)-DNA associations.Cr(III) Mutagenesis and Mutational Spectrum.

Immediately after treatment with CrCl3 in either TE/50

or KP, treated and untreated shuttle vectors were passedthrough G-50 columns and transfected into human 293cells for replication. The progeny plasmids were rescuedand introduced into indicator E. coli strains to determinethe supF mutant frequency. As shown in Table 1, thefrequency of supF mutants increased linearly as a

Figure 3. Polymerase-stop assay of Cr(III)-treated template.Plasmids were linearized with EcoRI before treatment with 25-200 µM CrCl3 in TE/50 buffers as described in ExperimentalProcedures. The Cr(III)-treated templates were replicated usingSequenase, four deoxyribonucleotides, and primer 1. The prod-ucts of these polymerase-stop reactions and dideoxy sequencingreactions of untreated plasmids (G, A, T, C) were analyzed in a6% acrylamide gel. The gel was cut into three regions, and thelanes were lined up from loading wells (region I) to the bottomof the gel (region III). The full-length DNA products were shownto be EcoRI at the top of region II. The dots shown in region IIon the right-hand sides of samples treated with 50 µM Cr(III)denote the bases located on the 3′ sides of template guanines.


function of CrCl3 concentration. DNA sequence analysiswas performed in 23 mutants derived from progeny ofpZ189 treated with Cr(III) in TE/50 (from six transfectionexperiments); 4 of the 23 mutants (17%) could have beensiblings and were not included in Table 2. Among the19 independent mutants sequenced, 13 contained single-base substitutions, 4 were deletions, and 2 were duplica-tions (Table 2, group A). The shuttle vector pSP189 thatwas designed for the identification of independent mu-tants (37) was chosen to obtain more detailed informationon Cr(III) mutagenesis. Of the 282 pSP189-derivedmutants analyzed in this study, 16 (5.7%) were found tobe actual sibling mutants. Table 2 shows a total of 76and 82 independent mutants derived in progeny ofpSP189 treated with Cr(III) in TE/50 buffer (group B)and in KP buffer (group C), respectively. Similar fre-quencies for each type of mutation derived from thesetwo populations were observed; i.e., 61-62% of thesemutants were single-base substitutions, 3-5% were two-base substitutions, and 32-34% were deletions. Tocompare Cr(III)-induced mutational specificity with thatoccurring spontaneously, DNA sequence analysis wasalso performed on 108 mutants obtained from untreatedpSP189 replication in human 293 cells (in several sepa-rate experiments). The results showed 63% of these

spontaneous mutants were single-base substitutions, 8%were two-base substitutions, 25% were deletions, and 4%were duplications (Table 2).Table 3 summarizes the specific kinds of base substitu-

tions generated by replication of Cr(III)-treated anduntreated plasmids in human 293 cells. All the basesubstitutions derived in group A were G‚C base pairtransversions: i.e., 10 G‚C f C‚G (77%) and 3 G‚C fT‚A (23%). All of those observed in group B occurred atG‚C-base pairs: i.e., 19 G‚C f A‚T (37%), 18 G‚C f C‚G(35%), and 14 G‚C f T‚A (28%). The major kinds of basesubstitutions observed in group C also occurred at G‚Cbase pairs (90%); only 6 of the 58 (10%) base substitutionsoccurred at A‚T base pairs. Similar frequencies of G‚Cf T‚A, G‚C f C‚G, and G‚C f A‚T were observed ingroups B and C. Among these base substitutions, thefrequencies of G‚C f C‚G observed in Cr(III)-treatedgroups were higher than those generated in untreatedplasmids (Table 3). These results indicate that thespecific kinds of mutations enhanced by Cr(III) treatmentwere similar in mutants derived from plasmids treatedin KP and TE/50; this is in spite of the different patternsof the Cr(III)-DNA interactions found in the two buffersdescribed above.The specific locations of base substitutions in the supF

t-RNA gene occurring spontaneously and induced by Cr-(III) are shown in Figures 5 and 6a, respectively. Single-base substitutions occurring at positions 104, 108, 123,

Figure 4. Effect of reaction environment on Cr(III)-inducedprimer-template cross-links. EcoRI-linearized plasmid pZ189 (3µg of DNA/15 µL) was treated with 50 µM Cr(III) in either TE/50 or KP buffer at 37 °C for 30 min. The unbound Cr(III) wasremoved by drop dialysis against either TE/50 or KP buffer forvarious times as indicated. The in vitro DNA polymerizationreaction was carried out and analyzed as described in Figure3. Only the radioactivity in the loading wells is shown. Lane 1was DNA sequencing reaction of untreated template usingdideoxycytidine triphosphate in chain termination.

Table 1. Mutant Frequency Obtained by Transformationof E. coli with Progeny of Cr(III)-Treated and Untreated

Shuttle Vectors Replicated in 293 Cells

number of

CrCl3 (µM) transformants mutantsfrequency of supFmutant (×10-4)

pZ189 in TE/50 Buffera0 22 776 2 0.881 19 164 3 1.5710 19 092 13 6.8125 19 029 8 4.2050 14 343 13 9.06

pSP189 in TE/50 Bufferb0 52 355 1 0.1925 110 690 42 3.7950 85 450 44 5.15

pSP189 in KP Bufferb0 48 845 6 1.2350 41 780 13 3.11200 91 220 39 4.28500 83 900 39 4.65

a Data obtained from six independent transfection experiments.The transformation experiments were performed using the calciumchloride procedure, and the E. coli host strain was SY204. b Dataobtained from two independent transfection experiments. Thetransformation experiments were performed using electroporation,and the E. coli host strain was MBM7070.

Table 2. Types of Mutations Generated in the supF Geneof Cr(III)-Treated and Untreated Plasmids Replicated in

293 Cells

number of mutants observeda

Cr(III)-treatedb

supF mutations group A group B group C total untreated

single-basesubstitutions

13 (68) 47 (62) 50 (61) 110 (62) 68c (63)

two-basesubstitutions

0 2 (3) 4 (5) 6 (3) 9 (8)

deletionsd 3 (16) 11 (14) 19 (23) 33 (19) 13 (12)compounddeletionse

1 (5) 14 (18) 9 (11) 24 (14) 14 (13)

duplications 2 (11) 2 (3) 0 (0) 4 (2) 4 (4)total 19 76 82 177 108a Numbers in parentheses indicate percentages. b Groups A-C

represent mutants derived from pZ189 treated with Cr(III) in TE/50, pSP189 treated with Cr(III) in TE/50, and pSP189 treated withCr(III) in KP, respectively. c Two mutants having a single-basesubstitution plus a single-base insertion were categorized in thisgroup. d Simple deletions. e In addition to supF deletions, thesemutants had insertions and/or rearrangements.

Table 3. Kinds of Base Substitutions Generated in thesupF Gene of Cr(III)-Treated and Untreated Plasmids

Replicated in 293 Cells

number of mutations observeda

Cr(III)-treatedbkinds of basesubstitutions group A group B group C all untreated

transversionsG‚C f T‚A 3 (23) 14 (28) 15 (26) 32 (26) 27 (31)G‚C f C‚G 10 (77) 18 (35) 17 (29) 45 (37) 18 (21)A‚T f C‚G 0 0 2 (3) 2 (2) 1 (1)A‚T f T‚A 0 0 0 0 2 (2)

transitionsG‚C f A‚T 0 19 (37) 20 (35) 39 (32) 37 (43)A‚T f G‚C 0 0 4 (7) 4 (3) 1 (1)

total 13 51 58 122 86a Numbers in parentheses indicate percentages. b Groups A-C

are defined in Table 2.


129, 163, and 168 were frequently observed (46%) inuntreated plasmids (Figure 5). Three mutational hotspots occurring at positions 133, 139, and 155 wereenhanced by Cr(III) treatment in KP buffers; threemutational hot spots at positions 105, 133, and 164 wereenhanced by Cr(III) treatment in TE/50 buffers (Figure6a). The hot spot at position 133 was commonly en-hanced in both Cr(III)-treated groups. The Cr(III)-enhanced single-base substitutions occurring at positions133, 139, 155, and 164 were infrequently observed inuntreated plasmids (Figure 5).

The relative intensities of polymerase-stop sites ontemplates treated with 50 µM Cr(III) in TE/50 buffer(region II of Figure 3) were quantified using a Phosphor-Imager, and the results are shown in Figure 6b. Strongpolymerase-stop sites occurred at positions 116, 124, 126,164, 172, and 173 of the supF coding region. Onlyposition 164 was also a hot spot for base substitutionsinduced by Cr(III). Whereas positions 133, 139, and 155were mutational hot spots for Cr(III), these positionswere weak sites for stopping DNA synthesis. Thiscomparison indicates that most of the Cr(III)-mediated

Figure 5. Locations of spontaneously occurring base substitutions in the supF gene coding region. Underlines represent two-basesubstitutions; double underlines represent single-base substitutions plus single-base insertions. The 3′ downstream bases of thesupF gene coding region of pSP189 differ from those of pZ189; these positions are not numbered.

Figure 6. Locations of base substitutions (a) and polymerase-stop sites (b) induced by Cr(III) in the supF gene coding region. Themutations derived from plasmids treated with Cr(III) in TE/50 and KP buffers, respectively, are shown above and below the 5′ to 3′sequence of the nontranscribed strand. Stars shown below the supF sequences indicate spontaneously occurring hot sopts. The 5′upstream and 3′ downstream bases of the supF gene coding region of pSP189 differ from those of pZ189; these positions are notnumbered. Lower case letters indicate mutations derived from pZ189. Underlines represent two-base substitutions. One Cr(III)-induced G‚C base substitution derived from group A occurring in the 5′ promoter region is not shown. The polymerase-stop sitesobtained in templates treated with 50 µM Cr(III) were analyzed using a PhosphorImager (b). The relative signal intensities ofpolymerase-stop sites are shown by the varying lengths of bars.


polymerase-stop sites were cold Cr(III)-induced basesubstitution spots and vice versa.All the Cr(III)-induced G‚C base substitutions found

in the supF coding region were pooled; their flankingbases were analyzed and compared with those derivedspontaneously to determine the existence of sequencespecificity. Both spontaneous and Cr(III)-derived basesubstitutions were less frequently observed at 5′GC,5′GT, and 5′TG sites. The 5′GA and 5′GG sequenceswere the frequent sites observed for the G‚C basesubstitutions occurring spontaneously. Base substitu-tions at 5′GA, 5′CG, and 5′AG sites were enhanced byCr(III) treatment, but occurrences at 5′GG and 5′GG siteswere decreased as compared with those derived sponta-neously.DNA sequencing results showed that only 1 of the 13

spontaneous deletions (7.7%) had deleted sizes largerthan 100 bp, whereas 8 of the 33 Cr(III)-induced deletions(24.2%) had these large deletions. Previous analyses ofDNA sequences in deletion mutations showed that shortdirect and inverted repeats were observed frequently atthe deletion break points (41, 42). Analysis of supFdeletions with known sequences around break pointsshowed that 3 of the 11 spontaneous (27%) and 5 of the30 Cr(III)-induced (17%) supF deletions had such shortrepeats at the deletion break points. This analysisindicates that Cr(III) treatment can enhance deletions,and these deletions seemed to occur randomly in shuttlevectors.

Discussion

Although Cr(III) directly binds to DNA, its inefficiencyin entering cells (2, 3, 12) hampered our understandingof the mutagenic potential of this metal. In this study,we adopted a shuttle vector system to illustrate the typesof mutations induced by replication of Cr(III)-treatedplasmids in human 293 cells. In comparison with themutations occurring spontaneously, Cr(III) induces G‚Cbase substitutions, particularly G‚C f C‚G transversions,located at highly specific sites (5′GA, 5′CG, and 5′AG).Cr(III) also enhances gene deletions and rearrangements.Moreover, we have determined the types of Cr(III)-DNAassociations using ICP-MS, agarose gel eletrophoreticmobility, and polymerase-stop assays. ICP-MS resultsshow that Cr(III) ions stably associate with DNA in bothTE/50 and KP buffers. DNA synthesis by Sequenase wasmarkedly blocked, primarily at one base located on the3′ sides of guanines in templates during in vitro replica-tion of templates treated with Cr(III) in TE/50. Theguanine-specific arrest of DNA replication we notedgenerally agrees with a recent report from Bridgewateret al. in which they treated an 89-bp synthetic DNAtemplate with Cr(III) in 10 mM Tris-HCl, pH 7.6 (32).However, we also found that such polymerase-stop sitesdo not form in templates treated with Cr(III) in KPbuffers, even in templates containing high amounts ofchromium adducts. These results suggest that bufferenvironments affect the types of Cr(III)-DNA associa-tions. Cr(III) is known to bind to both the phosphatebackbone and the nucleotide bases of DNA, particularlyguanines (13, 14). CrCl3‚6H2O in solutions at physiologi-cal pH values and temperatures is presented as the aquoform trans-[Cr(H2O)4Cl2]+ (15) which may attack thenegatively charged phosphate backbone and lead tostable Cr(III) monoadducts. These Cr(III) monoadductsmay subsequently react with the other DNA molecules

forming DNA cross-links between the two complementarystrands or interplasmid DNAs; however, high concentra-tions of phosphate ions present in KP buffers maycoordinate with Cr(III) monoadducts and thereby sup-press the ability of Cr(III) to induce DNA-DNA cross-links. Cr(III) monoadducts generated in KP could beconverted into primer-template cross-links during dialy-sis against TE/50 buffers. Results shown here suggestthat the major types of DNA lesions induced byCrCl3‚6H2O in KP buffers are Cr(III) monoadducts thatdo not block DNA replication, and those formed in TE/50 buffers are interstrand cross-links that terminatepolymerase processivity.Interestingly, despite the apparently different types of

Cr(III)-DNA adducts formed in shuttle vectors, similarmutational spectra were generated when they replicatedin human cells. A scenario was proposed for thisphenomena in which these different types of Cr(III)adducts generated in vitro are exchangeable due toaltered solution environments: e.g., the transfectionsolution and the cellular physiological conditions. In-deed, our data showed that the amounts of Cr(III)-mediated DNA-DNA cross-links are markedly alteredby substitution of buffer environments. Recently, Zhit-kovich et al. showed that Cr(III)-mediated DNA-proteincross-links were derived from the interaction between Cr-(III)-amino acids and DNA, whereas amino acids (cys-teine or histidine) do not cross-link to Cr(III)-DNA (20).This suggests that DNA-protein cross-links are notinvolved in the mutagenesis of Cr(III)-treated shuttlevectors.Comparison of the frequency of polymerase-stop signals

and mutational hot spots indicates that strong poly-merase-stop sites were cold spots for base substitutions.Similarly, base substitution hot spots were weak poly-merase-stop sites. These results suggest that the major-ity of Cr(III)-induced polymerase-stop sites would notinduce base substitutions. In human cells, DNA poly-merases encountering Cr(III)-mediated DNA cross-linksmay strongly block DNA replication and promote genedeletions and rearrangements by incomplete recombina-tion repair. In contrast, low numbers of chromiummonoadducts on the templates may enhance the proces-sivity of polymerase with a reduced fidelity (30). DNApolymerases may misincorporate deoxyribonucleotidesdue to a preference for dGMP across chromium monoad-ducts, thereby resulting in G‚C base substitutions, pri-marily G‚C f G‚C, in subsequent replication cycles.We have previously demonstrated that 8-hydroxydeox-

yguanosine and strand breakage in DNA are generatedby incubation of Cr(III)-DNA with H2O2 (27, 28). Thosestudies also showed that •OH radicals are possibly themajor ROS formed during the reaction of H2O2 with Cr-(III)-DNA and leading to DNA lesions. Hydrogen per-oxide is formed in aerobic cells as a result of normalcellular metabolism, and high levels of H2O2 are foundin several cancer cell lines (43, 44). Hydrogen peroxidefreely passes though cell membranes and can reach anycellular compartment (43, 44). Cr(III)-shuttle vectorsmay interact with H2O2 in cells, and subsequently gener-ate •OH radicals causing DNA lesions, e.g., 8-hydrox-ydeoxyguanosine. It has been shown that DNA poly-merases misincorporate dAMP across 8-hydroxydeoxy-guanosine in templates and subsequently result in G‚Cf T‚A transversions (45, 46). Moreover, Cr(III) alsoincreases polymerase bypass ROS-induced DNA damage


(31). Thus, ROS may also play a role in Cr(III) mutagen-esis.The molecular mutagenicity of hydrogen peroxide,

singlet oxygen, and peroxynitrite has been studied usingshuttle vector systems (42, 47, 48). Among the single-base substitutions enhanced in the supF gene in H2O2-treated monkey CV-1 cells, hot spots occurred at positions133, 159, and 168 (42); only position 133 was a hot spotfor Cr(III). Singlet oxygen induced mainly G‚C f T‚Aand G‚C f C‚G in the supF gene replicated in monkeyCOS7 cells (47). Nevertheless, Cr(III)-induced muta-tional hot spots are different from those derived fromsinglet oxygen treatment. The mutational hot spotsinduced by peroxynitrite-treated pSP189 replicated inhuman 293 cells were at positions 113, 124, 133, 156,and 164 predominantly with G‚C f T‚A involved (48);positions 133 and 164 are also Cr(III) hot spots, but themajor kinds of base substitutions were G‚C f C‚G. Atpresent, the mutagenicity of other ROS-induced basedamage leading to G‚C f C‚G transversions remainsunexplored.Snow and Xu have shown that replication of Cr(III)-

treated single-stranded phage DNA in vitro or in bacteriaenhanced the mutation frequency 2-3-fold above that ofuntreated controls (30). Cr(III)-enhanced reversion mu-tagenesis was significantly potentiated in mismatchrepair-deficient E. coli strains under non-SOS conditionsbut not after SOS-induction, suggesting that eithermismatch repair or SOS-associated processes decreasedCr(III) mutagenesis (31). Additionally, Cr(III) enhancedC f A transversions in that reversion mutagenesis assaysystem (31). Hexacoordinate Cr(III) aromatic bidentateamines can induce reversion mutations in Salmonellaassay with T‚A base pair mutant strains (29). Thosereversion assays, however, were incapable of detectingthe guanine-specific mutagenesis observed here. Nev-ertheless, whether Cr(III) can enhance DNA damage incytosine, adenine, or thymine that could subsequentlylead to mutations in other cells requires further inves-tigation.Recently, Juedes and Wogan reported a spontaneous

supF mutational spectrum derived from a pool of un-treated and KNO3-treated pSP189 replicated in human293 cells (48). The locations of spontaneous base sub-stitutions in the supF gene observed in this study werequite different from those reported by Juedes and Wogan(48). This may partly be due to different transfectionprotocols being performed, which could influence muta-tional hot spots (49).As discussed above, Cr(III)-mediated DNA-DNA cross-

links may enhance gene deletions and rearrangements.Strand breakage induced by ROS in Cr(III)-DNA mayalso be the etiology of gene deletions. However, theseDNA lesions are repaired much faster then Cr(III)-mediated cross-links (7, 50).Our previous study of the Cr(VI)-induced mutational

spectrum showed that T‚A base pair transversions arethe major kinds of mutations induced by three Cr(VI)compounds in the hypoxanthine (guanine) phosphoribo-syltransferase gene of Chinese hamster ovary cells (10).Cr(VI) compounds also induce revertants in Salmonellatyphimurium strains that are sensitive to oxidativedamages, leading to T‚A base pair substitutions (51).However, Cr(VI) induces more G‚C than T‚A base sub-stitutions in TK6 human lymphoma cells (11). Thesecontradictory results for Cr(VI) mutational specificitycould possibly be due to different antioxidant machinery,

Cr(VI) metabolic pathways, and DNA repair capabilitiesexisting in those cell lines. At present, determiningwhich forms of chromium metabolites and DNA lesionsinduced during Cr(VI) reduction cause each specific kindof mutation remains a vital challenge. Nevertheless, thisstudy has clearly indicated that Cr(III) enhances G‚Cbase pair substitutions and gene deletions.

Acknowledgment. The authors would like to thankthe National Science Council, Republic of China, forfinancial support of the work presented in this manu-script under Contract No. NSC85-2621-B007-003Z.

References

(1) Anderson, R. A. (1993) Recent advances in the clinical andbiochemical effects of chromium deficiency. In Essential and ToxicTrace Elements in Human Health and Disease: An Update(Prasad, A. S., Ed.) Vol. 380, pp 221-234, Wiley-Liss Inc., NewYork.

(2) IARC. (1990) Chromium, nickel, and wedding. In IARC Mono-graphs on the evaluation of the carcinogenic risks of chemicals tohumans, Vol. 19, International Agency for Cancer Research, Lyon.

(3) Langard, S., and Norseth, T. (1986) Chromium. In Handbook onthe Toxicology of Metals, 2nd ed. (Friberg, L., Nordberg, G. F.,and Vouk, V., Eds.) pp 185-210, Elsevier Science, Amsterdam.

(4) Sen, P., and Costa, M. (1986) Incidence and localization of sisterchromatid exchanges induced by nickel and chromium com-pounds. Carcinogenesis 7, 1527-1533.

(5) Sen, P., Conway, K., and Costa, M. (1987) Comparison of thelocalization of chromosome damage induced by calcium chromateand nickel compounds. Cancer Res. 47, 2142-2147.

(6) Biedermann, K. A., and Landolph, J. R. (1990) Role of valencestate and solubility of chromium compounds on induction ofcytotoxicity, mutagenesis, and anchorage independence in diploidhuman fibroblasts. Cancer Res. 50, 7835-7842.

(7) Manning, F. C. R., Blankenship, L. J., Wise, J. P., Xu, J.,Bridgewater, L. C., and Patierno, S. R. (1994) Induction ofinternucleosomal DNA fragmentation by carcinogenic chromate:relationship to DNA damage, genotoxicity, and inhibition ofmacromolecular synthesis. Environ. Health Perspect. 102 (Suppl.3), 159-167.

(8) Ye, J., Zhang, X., Young, H. A., Mao, Y., and Shi, X. (1995)Chromium(VI)-induced nuclear factor-κB activation in intact cellsvia free radical reactions. Carcinogenesis 16, 2401-2405.

(9) Kim, G., and Yurkow, E. J. (1996) Chromium induces a persistentactivation of mitogen-activated protein kinases by a redox-sensitive mechanism in H4 rat hepatoma cells. Cancer Res. 56,2045-2051.

(10) Yang, J.-L., Hsieh, Y.-C., Wu, C.-W., and Lee, T.-C. (1992)Mutational specificity of chromium(VI) compounds in the hprtlocus of Chinese hamster ovary-K1 cells. Carcinogenesis 13,2053-2057.

(11) Chen, J., and Thilly, W. G. (1994) Mutational spectrum ofchromium(VI) in human cells. Mutat. Res. 323, 21-27.

(12) Lu, Y.-Y., and Yang, J.-L. (1995) Long-term exposure to chromi-um(VI) oxide leads to defects in sulfate transport system inChinese hamster ovary cells. J. Cell. Biochem. 57, 655-665.

(13) Tsapakos, M. J., and Wetterhahn, K. E. (1983) The interactionof chromium with nucleic acids. Chem.-Biol. Interact. 46, 265-277.

(14) Koster, A., and Beyersmann, D. (1985) Chromium binding by calfthymus nuclei and effects on chromatin. Toxicol. Environ. Chem.10, 307-313.

(15) Hneihen, A. S., Standeven, A. M., and Wetterhahn, K. E. (1993)Differential binding of chromium(VI) and chromium(III) com-plexes to salmon sperm nuclei and nuclear DNA and isolated calfthymus DNA. Carcinogenesis 14, 1795-1803.

(16) Wetterhahn, K. E., Hamilton, J. W., Aiyar, J., Borges, K. M., andFloyd, R. (1989) Mechanism of chromium(VI) carcinogenesis:reactive intermediates and effect on gene expression. Biol. TraceElement Res. 21, 405-411.

(17) Aiyar, J., Berkovits, H. J., Floyd, R. A., and Wetterhahn, K. E.(1991) Reaction of chromium(VI) with glutathione or with hydro-gen peroxide: identification of reactive intermediates and theirrole in chromium(VI)-induced DNA damage. Environ. HealthPerspect. 92, 53-62.

(18) Shi, X., Mao, Y., Knapton, A. D., Ding, M., Rojanasakul, Y.,Gannett, P. M., Dalal, N., and Liu, K. (1994) Reaction of Cr(VI)with ascorbate and hydrogen peroxide generates hydroxyl radicals


and causes DNA damage: role of a Cr(IV)-mediated Fenton-likereaction. Carcinogenesis 15, 2475-2478.

(19) Liu, K. J., Shi, X., Jiang, J. J., Goda, F., Dalal, N., and Swartz,H. M. (1995) Chromate-induced chromium(V) formation in livemice and its control by cellular antioxidants: an L-band electronparamagnetic resonance study. Arch. Biochem. Biophys. 323, 33-39.

(20) Zhitkovich, A., Voitkun, V., and Costa, M. (1996) Formation ofthe amino acid-DNA complexes by hexavalent and trivalentchromium in vitro: importance of trivalent chromium and thephosphate group. Biochemistry 35, 7275-7282.

(21) Hamilton, J. W., andWetterhahn., K. E. (1989) Differential effectsof chromium(VI) on constitutive and inducible gene expressionin chick embryo liver in vivo and correlation with chromium(VI)-induced DNA damage. Mol. Carcinogen. 2, 274-286.

(22) Liebross, R. H., and Wetterhahn, K. E. (1992) In vivo formationof chromium(V) in chick embryo liver and red blood cells.Carcinogenesis 13, 2113-2120.

(23) Sugiyama, M., Tsuzuki, K., and Ogura, R. (1991) Effect of ascorbicacid on DNA damage, cytotoxicity, glutathione reductase, andformation of paramagnetic chromium in Chinese hamster V-79cells treated with sodium chromate(VI). J. Biol. Chem. 266, 3383-3386.

(24) Sugiyama, M. (1994) Role of paramagnetic chromium in chromi-um(VI)-induced damage in cultured mammalian cells. Environ.Health Perspect. 102 (Suppl. 3), 31-33.

(25) Sechlmeyer, U., Hechtenberg, S., Klyszcz, H., and Beyersmann,D. (1990) Accumulation of chromium in Chinese hamster V-79cells and nuclei. Arch. Toxicol. 64, 506-508.

(26) Wise, S. J. P., Stearns, D., Wetterhahn, K. E., and Patierno, S.R. (1994) Cell-enhanced dissolution of carcinogenic lead chromateparticles: the role of individual dissolution products in clastoge-nesis. Carcinogenesis 15, 2249-2254.

(27) Tsou, T.-C., Chen, C.-L., Liu, T.-Y., and Yang, J.-L. (1996)Induction of 8-hydroxydeoxyguanosine in DNA by chromium(III)plus hydrogen peroxide and its prevention by scavengers. Car-cinogenesis 17, 103-108.

(28) Tsou, T.-C., and Yang, J.-L. (1996) Formation of reactive oxygenspecies and DNA strand breakage during interaction of chro-mium(III) and hydrogen peroxide in vitro: evidence for a chro-mium(III)-mediated Fenton-like reaction. Chem.-Biol. Interact.102, 133-153.

(29) Sugden, K. D., Geer, R. D., and Rogers, S. J. (1992) Oxygenradical-mediated DNA damage by redox-active Cr(III) complexes.Biochemistry 31, 11626-11631.

(30) Snow, E. T., and Xu, L. S. (1991) Chromium(III) bound to DNAtemplates promotes increased polymerase processivity and de-creased fidelity during replication in vitro. Biochemistry 30,11238-11245.

(31) Snow, E. T. (1994) Effects of chromium on DNA replication inVitro. Environ. Health Perspect. 102 (Suppl. 3), 41-44.

(32) Bridgewater, L. C., Manning, F. C. R., Woo, E. S., and Patierno,S. R. (1994) DNA polymerase arrest by adducted trivalentchromium. Mol. Carcinogen. 9, 122-133.

(33) Kraemer, K. H., and Seidman, M. M. (1989) Use of supF, anEscherichia coli tyrosine suppressor tRNA gene, as a mutagenictarget in shuttle-vector plasmids. Mutat. Res. 220, 61-72.

(34) Yang, J.-L., Maher, V. M., and McCormick, J. J. (1987) Kinds ofmutations formed when a shuttle vector containing adducts of(()-7â,8R-dihydroxy-9R,10R-epoxy-7,8,9,10-tetrahydrobenzo[a]py-rene replicates in human cells. Proc. Natl. Acad. Sci. U.S.A. 84,3787-3791.

(35) Yang, J.-L., Chen, R.-H., Maher, V. M., and McCormick, J. J.(1991) Kinds and location of mutations induced by (()-7â,8R-dihydroxy-9R,10R-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene in thecoding region of hypoxanthine (guanine) phosphoribosyltrans-ferase gene in diploid human fibroblasts. Carcinogenesis 12, 71-75.

(36) Seidman, M. M., Dixon, K., Razzaque, A., Zagursky, R. J., andBerman, M. L. (1985) A shuttle vector plasmid for studyingcarcinogen-induced point mutations in mammalian cells. Gene 38,233-237.

(37) Parris, C. N., and Seidman, M. M. (1992) A signature elementdistinguishes sibling and independent mutations in a shuttlevector plasmid. Gene 117, 1-5.

(38) Sarkar, S., Dasgupta, U. B., and Summers, W. C. (1984) Error-prone mutagenesis detected in mammalian cells by a shuttlevector containing the supF gene of Escherichia coli. Mol. Cell.Biol. 4, 2227-2230.

(39) Yang, J.-L., Maher, V. M., and McCormick, J. J. (1988) Kindsand spectrum of mutations induced by 1-nitrosopyrene adductsduring plasmid replication in human cells. Mol. Cell. Biol. 8,3364-3372.

(40) Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) MolecularCloning: A Laboratory Manual, 2nd ed., Cold Spring HarborLaboratory Press, Cold Spring Harbor, NY.

(41) Thacker, J., Chalk, J., Ganesh, A., and North, P. (1992) Amechanism for deletion formation in DNA by human cell ex-tracts: the involvement of short sequence repeats. Nucleic AcidsRes. 20, 6183-6188.

(42) Moraes, E. C., Keyse, S. M., and Tyrrell, R. M. (1990) Mutagenesisby hydrogen peroxide treatment of mammalian cells: a molecularanalysis. Carcinogenesis 11, 283-293.

(43) Toyokuni, S., Okamoto, K., Yodoi, J., and Hiai, H. (1995)Persistent oxidative stress in cancer. FEBS Lett. 358, 1-3.

(44) Wiseman, H., and Halliwell, B. (1996) Damage to DNA by reactiveoxygen and nitrogen species: role in inflammatory disease andprogression to cancer. Biochem. J. 313, 17-29.

(45) Moriya, M. (1993) Single-stranded shuttle phagemid for mutagen-esis studies in mammalian cells: 8-oxoguanine in DNA inducestargeted G‚C f T‚A transversions in simian kidney cells. Proc.Natl. Acad. Sci. U.S.A. 90, 1122-1126.

(46) Grollman, A. P., and Moriya, M. (1993) Mutagenesis by8-oxoguanine: an enemy within. Trends Genet. 9, 246-249.

(47) de Oliveira, R. C., Ribeiro, D. T., Nigro, R. G., Mascio, P. D., andMenck, C. F. M. (1992) Singlet oxygen induced mutation spectrumin mammalian cells. Nucleic Acids Res. 16, 4319-4323.

(48) Juedes, M. J., and Wogan, G. N (1996) Peroxynitrite-inducedmutation spectra of pSP189 following replication in bacteria andin human cells. Mutat. Res. 349, 51-61.

(49) Seetharam, S., and Seidman, M. M. (1992) Modulation of ultra-violet light mutational hotspots by cellular stress. J. Mol. Biol.228, 1031-1036.

(50) Xu, J., Manning, F. C. R., and Patierno, S. R. (1994) Preferentialformation and repair of chromium-induced DNA adducts andDNA-protein crosslinks in nuclear matrix DNA. Carcinogenesis15, 1443-1450.

(51) Sugden, K. D., Burris, R. B., and Rogers, S. J. (1990) An oxygendependence in chromiummutagenesis.Mutat. Res. 244, 239-244.

TX970040P


Documents

Mutational Spectrum Induced by Chromium(III) in Shuttle Vectors Replicated in Human Cells: Relationship to Cr(III)−DNA Interactions