Genotoxicity of coke-oven and urban air particulate matter in in vitro acellular assays coupled with...

Ž .Mutation Research 414 1998 77–94

Genotoxicity of coke-oven and urban air particulate matter in invitro acellular assays coupled with 32 P-postlabeling and HPLC

analysis of DNA adducts

Blanka Binkova a,), Jan Lenıcek b, Ivan Benes c, Paulına Vidova d, Ondrej Gajdos d,´ ´ˇ ˇ ´ ´ ˇe ˇ aMichael Fried , Radim J. Sram´

a Laboratory of Genetic Ecotoxicology, Regional Institute of Hygiene of Central Bohemia, cro Institute of Experimental Medicine,Academy of Sciences of the Czech Republic, Prague, Czech Republic

b ´Regional Institute of Hygiene of Northern Bohemia, Ustı nad Labem, Czech Republic´c District Institute of Hygiene, Teplice, Czech Republic

d Specialized Institute of Hygiene and Epidemiology, Kosice, SloÕak Republicˇe Institute of Hygiene, UniÕersity of Heidelberg, Heidelberg, Germany

Received 9 December 1997; revised 18 February 1998; accepted 20 February 1998

Abstract

This study is an in vitro part of the ongoing biomarker studies with population from a polluted region of NorthernBohemia and coke-oven workers from Czech and Slovak Republics. The aim of this study is to compare DNA adduct

Ž .forming ability of chemical compound classes from both the urban and coke-oven extractable organic mass EOM ofairborne particles. The crude extracts were fractionated into seven fractions by acid–base partitioning and silica gel column

Ž . Ž .chromatography. In in vitro acellular assays we used calf thymus DNA CT DNA with oxidative qS9 and reductiveŽ .activation mediated by xanthine oxidase qXO under anaerobic conditions. Both the butanol and nuclease P1 versions of

32P-postlabeling for detection of bulky aromatic andror hydrophobic adducts were used. The results showed that the spectraof major DNA adducts resulting from both the in vitro assays are within the fractions similar for both the urban andcoke-oven samples. The highest DNA adduct levels with S9-activation were detected for the neutral aromatic fraction,followed by slightly polar and acidic fractions for both samples. With XO-mediated metabolism, the highest DNA adductlevels were detected for both the acidic fractions. Assuming additivity of compound activities, then the acidic fraction, which

Ž .in the urban sample comprises a major portion of EOM mass 28% , may contain the greatest activity in both in vitro assays

w x w x w x w x w x w x w xAbbreviations: anti-BPDE, benzo a pyrene-r-7,t-8-dihydrodiol-t-9,10-epoxide " ; B a A, benz a anthracene; B b F, B j F, B k F,w x w x w x w x w x w x w x 2benzo b, j,k fluoranthenes; B a P, benzo a pyrene; 9-OH-B a P, 9-hydroxy-benzo a pyrene; B ghi P, benzo ghi perylene; BPDE-N -dG,

w 2 X x w x7R,8S,9S-trihydroxy-10S- N -deoxyguanosyl-3 phosphate -7,8,9,10-tetrahydrobenzo a pyrene; CHRY, chrysene; CT DNA, calf thymusw x w xDNA; DB ah A, dibenz ah anthracene; DCM, dichloromethane; DRZ, diagonal zone of radioactivity; EOM, extractable organic mass;

w x w xI cd P, indeno cd pyrene; 9 NA, 9-nitroanthracene; 3 NF, 3-nitrofluoranthene; 1 NP, 1-nitropyrene; PAH, polycyclic aromatic hydrocar-bons; PEI, polyethyleneiminine; PM10, particles -10 mm; POM, polycyclic organic matter; RRT, relative retention time; TSP, totalsuspended particles; XO, xanthine oxidase

) Corresponding author.

1383-5718r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved.Ž .PII S1383-5718 98 00040-0

( )B. BinkoÕa et al.rMutation Research 414 1998 77–94´78

Ž .39 and 69%, qS9 and qXO, respectively . In contrast, the aromatic fraction constituting only 8% of total urban EOMŽ .mass may account for comparable activity 34% with organic acids. The highest DNA adduct forming activity of the

Ž .coke-oven sample accounts for the aromatic fraction 82 and 63%, qS9 and qXO, respectively that also contains theŽ .greatest portion of the total EOM 48% . To characterize some of the specific DNA adducts formed, we coupled TLC on

20=20 cm plates with HPLC analysis of 32P-postlabeled adducts. In both S9-treated samples of the aromatic fraction, wew x Ž w x .tentatively identified DNA adducts presumably diolepoxide-derived from: 9-hydroxy-benzo a pyrene 9-OH-B a P ,

w x w x Ž . w x Ž w x w x w x .benzo a pyrene-r-7,t-8-dihydrodiol-t-9,10-epoxide " anti-BPDE , benzo b, j,k fluoranthenes B b F, B j F, B k F , chry-Ž . w x Ž w x . w x Ž w x .sene CHRY , benz a -anthracene B a A and indeno cd pyrene I cd P . These DNA adducts accounted for about 57% of

total DNA adducts detected in both S9-treated samples of the aromatic fraction. DNA adducts of XO-treated samples weresensitive to nuclease P1 and HPLC profiles of the major adducts were markedly different from the major adducts ofS9-treated samples. However, the combination of TLC and HPLC did not confirm the presence of DNA adducts derived

Ž . Ž . Ž .from 1-nitropyrene 1 NP , 9-nitroanthracene 9 NA and 3-nitrofluoranthene 3 NF that were detected by GC-MS in theslightly polar fraction. We concluded that the chemical fractionation procedure facilitates the assessing of DNA adductforming ability of different chemical compound classes. However, based on the results obtained with the whole extracts, itdoes not fulfil a task of the actual contribution of individual fractions within the activity of the whole extracts. Our resultsare the first in detecting of DNA adducts derived from urban air and coke-oven particulate matter. q 1998 Elsevier ScienceB.V. All rights reserved.

Keywords: Urban air particulate matter; Coke-oven particulate matter; Fractionation; Polycyclic aromatic hydrocarbon; PAH; Nitro-PAH;DNA adduct; 32 P-postlabeling; HPLC of 32 P-labeled DNA adduct

1. Introduction

The potential magnitude of health risks from ex-posure to complex mixtures of polycyclic organic

Ž .matter POM is uncertain at this time. However,there is a strong evidence that certain high-level

Ž .exposures to POM e.g., by coke-oven workers havew xcaused increased incidence of lung cancer 1,2 .

POM, generally associated with the air particles, areproducts of incomplete combustion of organic mat-ter, as well as products of chemical and photochemi-cal reactions in the environment. These complexmixtures consist of many diverse classes of com-

Žpounds, such as PAHs, substituted PAHs e.g., nitro-. Žand alkyl-PAHs and heterocyclics e.g., aza- and

.thio-arenes , that have both mutagenic and carcino-w xgenic activity 3–6 . Identification of biologically

active components in such complex mixtures, thatmight be minor components by mass, is a difficulttask. The bioassay-directed chemical fractionationprocedures have been mostly used in combinationwith the Ames Salmonella mutagenicity assays to

w xevaluate the mutagenic activity of fractions 7–11 orrecently to identify mutation spectra induced in

w xSalmonella revertants 12,13 . These studies haveshown that much of the mutagenic activity of variouscomplex mixtures is due to the compounds in one ora few chemical classes present in the mixture. Simi-

larly, it was found that the mutation spectrum of acomplex mixture reflects the dominance of particularclasses of chemical mutagens within the mixture.

The covalent modification of DNA by chemicalcarcinogens is believed to be one of the criticalevents in the initiation of mutations and the carcino-

w xgenic processes 14–16 . Therefore, the detection ofDNA adducts has become increasingly important inthe assessment of human exposure to carcinogensw x17–19 . Only a few studies were concerned with32 P-postlabeling analysis of DNA adducts derivedfrom complex mixture of organics associated withparticulate matter emitted to the environment by

w xdiesel and gasoline vehicles or coke-ovens 20–25 .These studies employed either an in vitro acellularsystem or an in vivo animal model to examine DNAadduct profiles, dose–response relationship and dis-tribution of DNA adducts arising from these com-plex mixtures in different organ tissues. Savela et al.w x26,27 have used similar model systems to charac-terize some of the DNA adducts derived from dieseland foundry particulate emissions by combining32 P-postlabeling, TLC and HPLC of 32 P-postlabeledDNA adducts. All these studies evaluated bulky-aromatic DNA adducts derived from crude organic

Ž .extracts extractable organic material, EOM withoutusing any fractionation procedures. In our pilot study,we used similar approaches to examine the DNA

( )B. BinkoÕa et al.rMutation Research 414 1998 77–94´ 79

adducts formed in vitro from an EOM extract ofurban air particles, fractionated according to Pyysalo

w xet al. 8 by silica gel column chromatography intow xthe fractions with increasing polarities 28 .

In this study, we employed the procedure ofw xLewtas et al. 29 for fractionation of EOM extracts

of urban particulate emission and coke-oven particu-late emission into seven chemical compound classesby acid–base partitioning and silica gel column chro-

Žmatography. The urban air particles PM10 -10.mm were collected in the region where our human

w xbiomarker studies were also performed 30–33 . Theaim of this study was to compare DNA adductforming activity of the fractions and the whole ex-tract from both the environmental and coke-ovensamples evaluating total DNA adducts resulting fromthe in vitro acellular assays, and to compare themajor DNA adducts derived from the fraction withthe highest activity by confirmation of their identitiesusing TLC and HPLC separation procedures.

2. Materials and methods

2.1. Chemicals

Spleen phosphodiesterase was purchased fromŽ .Boehringer Mannheim; calf thymus DNA CT DNA ,

Ž .xanthine oxidase XO , micrococcal nuclease andnuclease P1 from Sigma; T4 polynucleotide kinase

32 Žfrom US Biochemical; g- P-ATP 3000 Cirmmol,.10 mCirml from Amersham; polyethylene-imine

Ž . Ž .PEI cellulose TLC plates 0.1 mm fromMacherey-Nagel; PAH compounds as individual

Ž . w xstandards 99% pure from Supelco; 9-OH-B a Pand anti-BPDE from Midwest Research InstituteŽ . ŽKansas City, MO ; rat liver S-9 Aroclor 1254

. Ž .induced from Organon Teknika West Chester, PA .All other chemicals and solvents were of HPLC oranalytical grade.

2.2. Air samples collection and extraction of EOM

Ž .Urban air particles PM10 particles -10 mmwere collected using HiVol air sampler Anderson

Ž .equipped with Pallflex filters 20=20 cm T60A20in Teplice City, a polluted area of the Northern

ŽBohemia, during the winter season 1993–1994 Oc-. 3tober–March . A total of 224 369 m of air was

sampled daily during 24-h or 12-h periods, using atotal of 205 filters. According to the data from thestationary monitoring station, during this winter pe-riod the daily concentrations of PM10 ranged from 5

3 w x Ž w x .to 180 mgrm , benzo a pyrene B a P from 0.3 to3 w x20 ngrm and carcinogenic PAHs including B a P,Ž . w x Ž w x .chrysene CHRY , benz a anthracene B a A ,

w x Ž w x . w xbenzo b fluoranthene B b F , benzo k fluorantheneŽ w x . w x Ž w x .B k F , dibenz ah anthracene DB ah A ,

w x Ž w x . w xindeno 1,2,3-cd pyrene I cd P and benzo ghi per-Ž w x . 3ylene B ghi P ranged from 2 to 113 ngrm . The

average concentrations for this winter period were3 w x 3following: PM10: 61 mgrm ; B a P: 5.2 ngrm ;

carcinogenic PAHs: 31 ngrm3. The same filterswere used in HiVol sampler Sierra to collect total

Ž .suspended particles TSP at the top-side of a coke-oven in an iron factory complex at Kosice during

Ž 3 .May 1995 1250 m , 20 filters . Particulate organicŽ .matter was extracted by dichloromethane DCM in

a Soxhlet apparatus for 24 h. The crude extracts wereconcentrated in a vacuum evaporator to reduce thevolume which was adjusted by DCM to a finalvolume of 200 ml for the following fractionationprocedure.

ŽTotal urban PM10 collected were 15.56 g aver-age winter concentration calculated from HiVol sam-

3.pler was 69 mgrm , total TSP collected at theŽcoke-oven were 32.94 g average particulate emis-

3.sion was 26 mgrm , total EOM extracted wereŽ . Ž3910 mg 25.1% of PM10 and 3310 mg 10.1% of

.TSP , for urban and coke-oven samples, respec-tively.

2.3. Chemical fractionation and analysis

ŽCrude extracts in gram quantities for the use in.different bioassays were fractionated according to

w xthe procedure of Lewtas et al. 29 by acid–basepartitioning into organic bases, acids and neutralfraction. The simplified overview is shown in Fig. 1.The neutral organic fraction was subsequently frac-tionated into five fractions according to increasingpolarity by silica gel column chromatography. Allfractions were concentrated by rotary evaporation.The aliquots of fractions were evaporated to drynessfor determination of mass distribution and redis-


Fig. 1. Scheme of fractionation procedure of air particulate crude extracts.

Žsolved in DMSO for in vitro assays stock solutions.2 mgrml DMSO .

Quantitative chemical analysis of eight carcino-genic PAHs in the crude extracts and aromatic frac-tions was performed by GC-MS at two independent

´Žlaboratories Ustı nad Labem, Czech Republic; Hei-´.delberg, Germany . The data are summarized in

Table 1. The nitro-PAHs detected by GC-MS in theŽurban slightly polar fraction were 1-nitropyrene 32

. Ž .ngrmg , 3-nitrofluoranthene 53 ngrmg and 9-Ž .nitroanthracene 11 ngrmg . In the coke-oven

slightly polar fraction were identified 1-nitropyreneŽ . Ž .27 ngrmg and 3-nitrophenanthrene 66 ngrmg . In

Ž .the coke-oven acids 9-nitroanthracene 78 ngrmgŽ .and 9-nitrophenanthrene 23 ngrmg were found.

The average urban winter concentrations of car-w xcinogenic PAHs and B a P calculated from HiVol

data were higher than those obtained from dailyŽ 3 3monitoring 49.5 ngrm and 7.4 ngrm vs. 31.0

3 3 w xngrm and 5.2 ngrm for PAHs and B a P, respec-.tively . The average concentrations of carcinogenic

w xPAHs and B a P at the coke-oven during period of

Table 1Carcinogenic PAHs in the crude extracts and aromatic fractions as quantified by GC-MS

Carcinogenic PAHs Urban Coke-oven

Crude extract Aromatic fraction Crude extract Aromatic fractiona b a bŽ . Ž . Ž . Ž .mgrmg mgrmg mgrmg mgrmg

w xBenzo a pyrene 0.42 4.26 6.36 12.75Chrysene 0.73 5.25 8.92 16.61

w xBenz a anthracene 0.35 5.49 8.85 16.24w xBenzo b fluoranthene 0.38 4.35 7.02 14.58w xBenzo k fluoranthene 0.19 2.83 6.92 11.53w xDibenz ah anthracene 0.07 1.48 0.84 1.69w xIndeno 1,2,3-cd pyrene 0.32 4.23 4.86 9.89w xBenzo ghi perylene 0.39 3.75 8.92 16.61

SUM of carcinogenic PAHs 2.84 31.64 47.95 91.93

aRefers to the mgrmg of total EOM.bRefers to the mgrmg of organic mass of aromatic fraction.


HiVol sampling were 127.0 and 16.8 mgrm3 forw xPAHs and B a P, respectively. These emission val-

ues are three order of magnitude higher as comparedto the urban imission values.

2.4. DNA adduct standards

2 Ž w 2BPDE-N -dG 7R,8S,9S-trihydroxy-10S- N -de-X xoxyguanosyl-3 phosphate -7,8,9,10-tetrahydrobenzo-

w x .a pyrene was purchased from Midwest ResearchŽ . w x w xInstitute MI, USA . B j F and I cd P-derived DNA

adduct standards were gifts from Dr. L. KingŽ .NHEERL, US EPA, RTP, NC . The anti-BPDE-DNA adduct standard was prepared by direct incuba-

w x w x w xtion with CT DNA. B b F-, B k F-, 9-OH-B a P-,w x w xB a P- and B a A-DNA adduct standards were pre-

pared by incubation of CT DNA with individualŽ .compounds final concentration 25 mgrml in the

presence of S-9 mix as described in Section 2.5. Theinternal UV standard, cis-9,10-dihydroxy-9,10-dihy-drophenanthrene, was a gift from Dr. D.H. PhillipsŽHaddow Laboratories, Institute for Cancer Re-

.search, Sutton, Surrey, UK .

2.5. In Õitro assays

The in vitro assays were performed as previouslyw x Ž .described 22,23,28 . Briefly, CT DNA 1 mgrml

was incubated with EOM fractions under oxidativeŽ . Ž .aerobic and reductive anaerobic conditions using

Ž .the same doses of organic material 100 mgrml forall fractions. The metabolic activation systems used

Ž .were as follows: 1 an oxidative rat liver S-9 systemŽ . Ž .0.5 mg proteinrml and 2 a reductive XO-cata-

Ž .lyzed system 0.5 U of XOrml . The incubationmixture for S9-activation contained in a final volume8 mM MgCl , 33 mM KCl, 5 mM glucose 6-phos-2

phate, 4 mM NADP, 100 mM sodium phosphatebuffer pH 7.4, and for XO activation 16 mM potas-sium phosphate buffer pH 5.8 and 3 mM hypoxan-thine. The S9 incubation was performed under aero-bic conditions for 4 h at 378C. The XO activation

Žwas performed in the anaerobic chamber Oxoid,.UK with an atmosphere of nitrogen for 4 h at 258C.

DNA was isolated by phenolrchloroformriso-amylalcohol extraction and ethanol precipitation andthe samples were kept at y808C until 32 P-postlabel-ing analysis.

2.6. 32P-postlabeling

Ž .DNA samples 6 mg were digested by a mixtureof micrococcal endonuclease and spleen phosphodi-esterase at 378C for 4 h. Both the nuclease P1 andbutanol extraction procedures were used for adductenrichment as described by Reddy and Randerathw x w x34 and Gupta 35 . Adducted nucleotides were la-

32 Žbeled with 25 mCi of g- P-ATP specific activity.3000 Cirmmol in the presence of T4 polynu-

cleotide kinase. For HPLC analysis, 50 mg DNAwas used and the quantities of digestion enzymesand nuclease P1 were eight times increased and forbutanol extraction three times higher volumes wereapplied. The sample volume after nuclease P1 treat-ment was reduced to 10 ml before labeling for HPLCanalysis with 50 mCi of g-32 P-ATP. The labeledDNA adducts were resolved by two-directional thinlayer chromatography on 10=10 cm or 20=20 cmPEI-cellulose plates. Solvent systems used for TLCwere the following: D-1: 1 M sodium phosphate, pH6.8; D-2: 3.8 M lithium formate, 8.5 M urea, pH 3.5;D-3: 0.8 M lithium chloride, 0.5 M Tris, 8.5 M urea,

Ž .pH 8.0, D-4 reduction of background radioactivity :same as D-1 solvent. Autoradiography was carriedout at y808C for 1 to 24 h. The radioactivity of the

Ž .diagonal zone DRZ or distinct DNA adduct spotswas measured by liquid scintillation counting. Tocalculate DNA adduct levels, the aliquots of DNA

Ž .enzymatic digest 0.5 mg of DNA hydrolysate wereanalyzed for nucleotide content by reverse phaseHPLC with UV detection which simultaneously al-lowed a control of the purity of the DNA. Thebackground CPM of DRZ or spots of control sam-

Ž .ples incubated with DMSO only were subtracted.Total DNA adduct levels were expressed as

8 w xadductsr10 nucleotides. B a P-derived DNA adductstandard was run in triplicate in each postlabelingexperiment for controlling of interassay variabilityand for normalizing the calculated DNA adduct lev-els.

2.7. HPLC analysis of 32P-postlabeled DNA adducts

HPLC analysis was carried out on the WatersHPLC system equipped with Model 600 pump andcontroller, Model 486 absorbance detector, Model717 autosampler and Millenium manager. Radioac-


tivity was monitored with a Berthold detector LB506 C-1 equipped with Browin manager softwareusing 200 ml flow cell. The adduct spots separatedon 20=20 cm TLC plates were excised and mea-sured by Cerenkov in a scintillation counter. Theexcised spots were extracted with 1 ml 4 M pyri-dinium formate, pH 4.0 for 8 h. The extracts were

Ž . Ž .centrifuged 6000 rpm, 5 min and an aliquot 10 mlof supernatant was measured in scintillation counterfor controlling the efficiency of extraction whichvaried between 85–98% according to the type ofDNA adduct. The samples were evaporated to dry-ness, redissolved in 100 ml methanolr0.3 M sodium

Ž .phosphate buffer 9:1; pH 2.0 and centrifuged againŽ .6000 rpm, 5 min to discard possible particulate

Ž .contaminants. The supernatant 90 ml was spikedŽwith internal standard of the UV marker cis-9,10-di-

.hydroxy-9,10-dihydrophenanthrene and placed in thevials for HPLC analysis. The resolution of 32 P-labeled adducts was performed on the analytical

Ž .column ODS 18 250=4.6 mm, 5 mm, Ultraspherewith flow rate of 1 mlrmin. The solvents and gradi-

w xent were according to King et al. 36 as follows:solvent A: 0.5 M NaH PO buffer pH 2; Solvent B:2 4

90% methanol and 10% 0.3 M NaH PO pH 2.0.2 4

The gradient was: 0 to 12.5 min, 10 to 43% B; 12.5to 60 min, 43 to 47% B; 60 to 80 min, 47 to 90% B;80 to 85 min, 90 to 95% B; 85 to 105 min, 95 to

Ž .10% B. The relative retention times RRT werecalculated by dividing the retention times of thepeaks of interest by the retention time of the internalstandard. HPLC analysis was used only for qualita-tive measurement.

3. Results

3.1. Mass distribution by fractions

The distribution of the organic mass by fractionsand recoveries of the mass after acidrbaserneutralpartitioning, silica gel column chromatography andoverall recovery of EOM are summarized in Table 2.The mass recoveries, using the fractionation schemeshown in Fig. 1, were similar for the both urban andcoke-oven samples. The highest portion of organicmass of the coke-oven sample was present in the

Ž .aromatic fraction 48.1% , which accounted only for

8% of total mass in the urban sample. In contrast, thehighest portion of organic mass of the urban sample

Žwas found in organic acids 28.1% vs. 4.6% for.urban and coke-oven, respectively . The moderately

polar fractions were more pronounced for both sam-Žples as compared with the remaining fractions 18.6.and 16.5% for urban and coke-oven, respectively .

3.2. 32P-postlabeling and quantitation of the totalDNA adduct leÕels

DNA samples modified in in vitro assays byincubation with the same doses of organic mass ofthe fractions were analyzed with both the butanoland nuclease P1 version of 32 P-postlabeling for bulkyaromatic andror hydrophobic adducts. Chromatogra-phy using 10=10 cm TLC plates revealed mostlythe presence of diagonal radioactive zone from closeto the origin up to the opposite top of the plate as

Ž . Ž .shown in Fig. 2A qS9 and Fig. 2D qXO forcoke-oven crude extracts. The length and the width

Žof DRZ was dependent on the fraction applied not.shown . The better resolution of the overlapping

spots was achieved by 20=20 cm TLC plates thatwere used for HPLC analysis of 32 P-labeled DNA

Ž .adducts as shown in Fig. 2B,C qS9 and Fig. 2E,FŽ .qXO for crude extract of both the coke-oven and

Ž .urban samples for description, see Section 3.3 .However, the small plates were used to quantify thetotal DNA adduct levels derived from the fractionsdue to simplicity of the TLC procedure for analyzinga big set of samples in the same experiment.

The quantitative results obtained for all DNAsamples analyzed using the butanol version of 32 P-postlabeling for bulky aromatic andror hydrophobicadducts are summarized in Table 3. DNA adductlevels detected by both the butanol and nuclease P1procedures were similar for oxidative activationŽ .qS9 . The differences were observed for yS9 andqXO conditions. The butanol procedure resulted in2–3-fold higher adduct recovery in the DNA sam-ples with yS9 and qXO systems than the nucleaseP1-mediation.

The highest DNA adduct levels with S9-activationresulted from the aromatic fraction for both the

Žurban and coke-oven samples 386 and 385adductsr108 nucleotides as detected by butanol en-

.hancement procedure , followed by slightly polar


Table 2The mass distribution by fractions and recoveries of the organic mass after fractionation procedures for both urban and coke-oven samples

Ž .Fraction Mass distribution %

Urban Coke-oven

Bases 1.4 3.7Acids 28.1 4.6Aliphatics 8.5 2.1Aromatics 8.2 48.1Slightly polar 9.6 4.3Moderately polar 18.6 16.5Highly polar 5.9 1.5Recovery of EOM after acidrbaserneutral partitioning 92.3 99.4Recovery of neutral organic mass after column chromatography 81.1 79.5

Ž .Total recovery of EOM % 80.4 80.7

Ž 8 .151 and 196 adductsr10 nucleotides and acidicŽ 8 .fractions 132 and 217 adductsr10 nucleotides .

The total DNA adduct levels were significantly higher

for coke-oven fractions with the exception of PAH-aromatic fractions that were comparable. DNA adduct

Ž .levels detected without presence of S9 mix yS9

Fig. 2. Autoradiographs of TLC maps of 32 P-postlabeled adducts of CT DNA modified in vitro in the presence of S9 oxidative system byŽ . Ž . Žincubation with coke-oven A,B and urban C crude extracts and under anaerobic conditions in the presence of XO D,E: coke-oven and F:

. Ž . Ž .urban crude extracts . A,D A total of 6 mg of DNA analyzed, TLC plates 10=10 cm, autoradiography 6 h, at y808C. B,C,E,F A totalŽ . Ž . Ž . Ž .of 50 mg of DNA analyzed, TLC plates 20=20 cm, autoradiography 2 h B,E and 4 h C,F , at y808C. A–C nuclease P1; D–F

butanol extraction version of 32 P-postlabeling.


Table 3Total DNA adduct levels derived from urban and coke-oven fractions

aDNA adduct levels

Fractions Incubation

yS9 qS9 qXO

Urban Coke-oven Urban Coke-oven Urban Coke-ovenb )Bases 25.9"5.2 ND 65.5"5.5 96.4 "9.4 30.5"5.2 18.5"6.1

) ) ) ) )Acids 33.2"5.6 18.7 "0.9 131.5"3.0 217.1 "13.0 95.6"12.4 299.9 "38.0Aliphatics ND ND ND ND ND ND

) )Aromatics 20.9"4.3 39.6 "0.9 386.2"25.7 384.6"29.0 56.4"2.6 86.6 "1.8) )Slightly polar 22.7"1.1 12.1 "2.4 151.4"15.0 196.0 "17.4 38.1"1.2 37.1"3.1

) ) )Moderately polar 14.3"5.2 13.0"0.6 55.6"12.1 103.4 "2.6 18.1"0.5 45.1 "5.0) ) ) )Highly polar 6.8"1.7 14.8 "1.5 6.4"2.0 109.4 "21.1 1.6"0.3 42.8 "0.4

) )Crude extract 52.4"1.9 51.9"7.2 137.7"5.5 257.7 "21.9 87.0"6.7 95.0"13.4

CT DNA was incubated with the fractions using the doses of 100 mgrml of organic mass. The data are for butanol extraction version of the32 P-postlabeling.a DNA adducts expressed as adductsr108 nucleotides.b Mean"SD from two separate incubations of each sample analyzed in three separate postlabeling experiments.ND: not detectable DRZ or spots on autoradiograms.) ) ) Ž .p-0.05; p-0.01, comparison of urban and coke-oven samples t-test .

were, for both the urban and coke-oven fractions,Ž .significantly lower up to one order of magnitude as

compared with S9-activation.In contrast, with XO-mediated metabolism, the

highest DNA adduct levels resulted from the organicacids. However, DNA adduct levels were 3-foldhigher for coke-oven acids as compared to the urban

Ž 8acids 300 vs. 96 adductsr10 nucleotides, butanol.extraction .

The highest differences between the urban andcoke-oven samples in both in vitro assays werefound for the highly polar fraction that exhibitedvery low activity by the urban sample. No DNAadducts were detected by both the aliphatic fractions

Table 4ŽDNA adduct forming activity of the fractions according to their mass partition in the crude extract for urban and coke-oven samples butanol

32 .extraction version of the P-postlabelingaŽ .DNA adduct forming activity %

Fractions Incubation

yS9 qS9 qXO

Urban Coke-oven Urban Coke-oven Urban Coke-oven

Bases 2.1 0 1.0 1.5 1.1 1.0Acids 56.0 3.8 38.9 4.4 69.0 21.0Aliphatics 0 0 0 0 0 0Aromatics 10.3 83.6 33.5 82.0 11.9 63.3Slightly polar 13.1 2.3 15.3 3.7 9.4 2.4Moderately polar 16.0 9.4 10.9 7.6 8.7 11.3Highly polar 2.4 0.9 0.3 0.7 0 0.9

bŽ .Recovery of DNA adduct forming activity % 31.8 43.9 68.9 87.5 44.9 69.2

a wŽ . Ž . x ŽCalculated as: total DNA adduct level ) f rS total DNA adduct level ) f )100; f smass partition factor of i-fraction according toi i i i i i.mass distribution in Table 1 .

b w Ž . x xCalculated as: S total DNA adduct level ) f rtotal DNA adduct level of crude extract )100.i i i

()

B.B

inkoÕa

etal.r

Mutation

Research

4141998

77–

94´

85

Ž .Fig. 3. Autoradiographs of TLC maps of CT DNA modified in vitro in the presence of S9 oxidative system by incubation with PAH-aromatic fraction of coke-oven A andŽ . w x Ž . w x Ž . w x Ž . w x Ž . Ž . w x Ž .urban E samples and DNA adduct standards derived from 9-OH-B a P B , B a P C , B b F D , B k F F , CHRY G and B a A H that were run in the same postlabeling

Ž . Ž . Ž . 32and TLC experiment. A total of 50 mg of DNA was analyzed in the case of A,E and 20 mg DNA in the case of B–D and F–H using nuclease P1 procedure. P-postlabeledŽ .DNA adducts were resolved on 20=20 cm plates with filter wicks 5=20 cm in both the D2 and D3 directions. Autoradiography was at y808C for 1 h in the case of A , 2 h in

Ž . Ž . Ž .the case of E and 15 to 30 min in the case of B–D and F–H .


in both the in vitro assays under 32 P-postlabelingconditions used.

Comparing the total DNA adduct levels derivedfrom the crude extracts with S9-activation then theratio of coke-oven vs. urban sample was 1.9 and 2.1as determined by the butanol and nuclease P1, re-spectively. When the concentrations of eight carcino-

Ž .genic PAHs Table 1 were totaled, the PAH-ratio incrude extracts of coke-oven vs. urban sample was16.9. In the aromatic fractions, the PAH-ratio was2.9, however, the total DNA adduct levels derivedfrom these fractions were the same as determined byboth the butanol and nuclease P1 versions. Thenature of organic compounds, their combinationandror differences in relative concentrations mayaccount for the different metabolic rate of PAHs inthese complex mixtures, and consequently, for thedifferences in ratio of total DNA adduct levels asdetected by S9-mediated metabolism.

3.3. The eÕaluation of weighted DNA adduct formingactiÕity

The weighted DNA adduct forming activity ofŽeach fraction, as defined in Table 4 see table note

.a , was used to calculate recovery and contributionof each fraction to the total activity. Assuming theadditivity of compound activities without major syn-ergistic or antagonistic effects then the sum ofweighted activities is equal to recovered activity. Therecoveries of DNA adduct forming activity with S9metabolic activation were near to the mass recover-

Ž .ies 80% assuming only a weak synergistic effect ofŽcompounds present in the urban sample 69% recov-

.ery and a weak antagonistic effect in the coke-ovenŽ .sample 88% recovery . The recoveries of DNA

adduct forming activity without the presence of S9-Ž .mix direct DNA reactive compounds and under

Ž .reductive conditions qXO were lower indicating a

32 Ž .Fig. 4. HPLC analysis of P-postlabeled DNA adducts resulting in vitro from aromatic fractions qS9 mix and PAH-DNA adductŽ . Ž . Ž . Ž . Ž . Ž . Žstandards. Profiles shown are those of the adduct 1 A , adduct 3 B , adduct 4 C , adduct 2 D , adduct 5 E adduct 6 F see Fig. 3A and

. w x Ž . Ž . w x Ž . w x Ž . Ž . w x Ž .E and DNA adduct standards of 9-OH-B a P G , CHRY H , B a A I , B k F J , anti-BPDE K and B b F L . Retention time of UVŽ .marker cis-9,10-dihydroxy-9,10-dihydrophenanthrene , used as an internal standard, was 37.2–37.4 min during these HPLC runs. Relative

Ž . Ž .retention time is calculated from the retention time min of the peak of interest divided by the retention time min of the UV marker. Forchromatographic conditions, see Section 2.


Table 5Contribution of the major DNA adducts of in vitro modified CT DNA by aromatic fractions with S9-activation to the total DNA adductlevel

aAromatic fraction % Of total radioactivity

Adduct no. Adduct derived from Urban Coke-oven

w x1 9-OH-B a P 20.5 17.9w x w x w x2 anti-BPDE, B b F, B k F, B j F 19.8 19.8

3 CHRY 7.4 10.2w x4 B a A, unknown 3.8 6.5

w x5 I cd P 6.2 2.16 Unknown 8.3 4.97 Unknown 9.8 10.3Sum 1–5 57.7 56.5Sum 1–7 75.8 71.7

a Total radioactivity of all DNA adducts detected as measured by Cerenkov counting.

possible synergistic effect of organic compoundspresent in these mixtures.

As shown in Table 4, the acidic fraction, which inthe urban sample comprises a major portion of EOM

Ž .mass 28% , may contain the greatest DNA adductŽforming activity in the both in vitro assays 39 and

.69%, qS9 and qXO, respectively . In contrast, thearomatic PAH fraction which constitutes only 8% of

Fig. 5. Autoradiographs of TLC maps of CT DNA modified in vitro in the presence of S9 oxidative system by incubation with coke-ovenŽ . Ž . Ž . Ž .A and urban C organic acids and coke-oven B and urban D slightly polar fractions. A total of 50 mg of DNA was analyzed bynuclease P1 procedure; TLC plates 20=20 cm; chromatographic conditions same as in the legend in Fig. 3. Autoradiography was at

Ž . Ž .y808C for 2 h in the case of A and 4 h in the case of B–D .


total EOM mass may account comparable activityŽ .with organic acids 34% . The highest DNA adduct

forming activity of the coke-oven sample accountsŽfor the aromatic fraction 82 and 63%, qS9 and

.qXO, respectively that also contains the greatestŽ .portion of organic mass 48% . The contribution of

each fraction to activity without the presence of S9mix is rather speculative due to low recoveries ofcrude extract activity and low DNA adduct levelsdetected.

3.4. Separation of the major DNA adducts formed byurban and coke-oÕen samples using 20=20 cm TLCplates coupled with HPLC analysis of 32P-post-labeled adducts

DNA adducts formed in in vitro assays by crudeextracts and the fractions that accounted for themajor DNA adduct forming activity were resolvedusing 20=20 cm TLC plates. Fig. 2 shows DNAadduct maps obtained from crude extracts of coke-

oven and urban samples using both S9- and XO-Ž .mediated Fig. 2B,CFig. 2E,F, respectively

metabolism. DNA adduct profiles obtained by S9-activation of the aromatic fractions are shown in Fig.

Ž . Ž .3A coke-oven and Fig. 3E urban together withTLC maps of DNA adduct standards derived from

w x Ž . w x Ž . w x Ž9-OH-B a P Fig. 3B , B a P Fig. 3C , B b F Fig.. w x Ž . Ž . w x3D , B k F Fig. 3F , CHRY Fig. 3G and B a A

Ž .Fig. 3H . The major adducts obtained from both thearomatic fractions and PAH-DNA adduct standardswere excised, Cerenkov counted, extracted fromPEI-cellulose and analyzed by HPLC. DNA adductstandards exhibiting the same chromatographic mo-bilities on TLC plates as the corresponding majoradducts derived from aromatic fractions were runsubsequently. The results of HPLC analysis areshown in Fig. 4. The adducts 1 and 3 that exhibited

Ž .markedly different TLC mobilities Fig. 3A,E elutedŽ .from the column similarly Fig. 4G,H as the corre-

sponding DNA adduct standards derived from 9-OH-w x Ž . ŽB a P Fig. 3BFig. 4G and CHRY Fig. 3GFig.. Ž .4H . HPLC profile of the adduct 4 Fig. 3A,E , with

Ž . Ž .Fig. 6. HPLC analysis of major adducts of CT DNA modified in vitro under oxidative qS9 and reductive qXO conditions withŽ . Ž . Ž . Ž .coke-oven crude extract see Fig. 2A,E . Profiles shown for S9-treated sample are those of adduct 1 A , adduct 2 B , adduct 3 C and

Ž . Ž .adduct 5 D . Retention time of UV marker was 41.6–42.0 min. Profiles shown for XO-treated sample are those of adduct 1 E , adduct 2Ž . Ž . Ž .F adduct 3 G and adduct 4 H . Retention time of UV marker was 40.5–40.6 min. Relative retention time of the peak of interest iscalculated as in the legend in Fig. 4.


w xthe same TLC mobility as major adduct of B a A-Ž .DNA adduct standard Fig. 3H , revealed two chro-

Ž .matographic peaks Fig. 4C , one of them with thew x Ž .same RRT as B a A-DNA adduct standard Fig. 4I ,

the second peak of unknown origin with RRTs0.46.Ž .The adduct 2 Fig. 3A,E exhibited the similar TLC

w x w xmobility as anti-BPDE-, B b F- and B k F-DNAŽ .adduct standards Fig. 3C,D,F . HPLC analysis of

Ž .the adduct 2 Fig. 4D showed multiple peaks withRRTs 0.92, 0.94, 1.01 that corresponded to RRTs of

w x Ž . Ž . w xB k F- Fig. 4J , anti-BPDE- Fig. 4K and B b F-Ž .DNA Fig. 4L adduct standards. Comparing the

peak of RRTs0.84 with RRTs of DNA-adductstandards, we could suggest that it is derived from

w xone of the major adducts of B j F-DNA adductstandard that exhibited on the small TLC plate thesame mobility as the anti-BPDE-DNA adduct stan-

Ž .dard. Similarly, we could compare RRT Fig. 4E ofŽ .the adduct 5 Fig. 3A,E with RRT of the major

w xadduct of I cd P-DNA adduct standard and its mobil-

ity on the small TLC plate. HPLC analysis of theadduct 6, more pronounced in the urban aromatic

Ž .fraction Fig. 3A,E , demonstrated a single peak withŽ .RRTs0.98 Fig. 4F of unknown origin. HPLC

Ž .analysis of the adduct 7 Fig. 3A,E resulted inmultiple peaks just under the detection limit withRRTs of 0.71, 1.44 and 1.78. The identity of theadducts 1, 2, 3 and 4 was also confirmed by spikingof the extracted adduct spots with the corresponding

ŽDNA adduct standards before HPLC analysis data.not shown .

The contribution of the major DNA adducts, thatwere observed by S9-treated aromatic fractions, tothe all adducts detected along the diagonal zone ofradioactivity on 20=20 cm TLC plates as measuredby Cerenkov counting, is summarized in Table 5.DNA adducts tentatively identified using the combi-nation of TLC and HPLC accounted for about 57%of the total adducts derived from both the urban andcoke-oven samples.

Ž . Ž .Fig. 7. Autoradiographs of TLC maps of CT DNA modified in vitro in the presence of XO by incubation with coke-oven A and urban DŽ . Ž . Ž . Ž .organic acids, with coke-oven B and urban E aromatic fractions and with coke-oven C and urban F slightly polar fractions. A total of

50 mg of DNA was analyzed using butanol extraction; TLC plates 20=20 cm; chromatographic conditions same as in the legend in Fig. 3.Ž . Ž .Autoradiography was at y808C for 1 h in the case of A and 4 h in the case of B–F .


DNA adduct maps resulted from organic acidsand slightly polar fractions with S9-activation areshown in Fig. 5A–D. DNA adduct patterns derivedfrom organic acids of both coke-oven and urbansamples were similar and only small differencescould be seen for slightly polar fractions. TLC mo-bilities some of the major DNA adducts were closeto those major adducts identified in the DNA sam-ples modified by aromatic fractions with S9-activa-tion. HPLC analysis of the major adducts apparenton DNA adduct maps in Fig. 5A–D did not confirmtheir identities with the major DNA adducts derivedfrom aromatic fractions but we were not able toidentify them. These DNA adducts could be over-lapped on the DNA adduct maps derived from S9-treated crude extracts. Therefore, we can suggest thatby HPLC analysis of the major DNA adduct spots

Ž .1–7 Fig. 2B,C derived from crude extracts weidentified only their dominant part. HPLC profiles inFig. 6A–D illustrate that the adduct spots 1, 2, 3 and5 eluted with the same RRTs as adduct 1,2, 3 and 5derived from aromatic fractions under the same con-

Ž .ditions Fig. 4A,D,B,E .DNA adduct maps which resulted from organic

acids, aromatic and slightly polar fractions withXO-mediated metabolism are shown in Fig. 7. Thediffuse DNA adducts derived from organic acids andslightly polar fractions showed by HPLC analysismostly have multiple broad peaks or they were be-low the detection limit. HPLC analysis of the distinctDNA spots visible on the autoradiograms of DNAsample modified by aromatic fractions in the pres-

Ž .ence of XO Fig. 7B,E confirmed their dissimilari-ties from the DNA adducts observed in S9-treated

Ž .samples RRTs of major spots were over 1.4 .Fig. 6E–H demonstrates the differences in HPLC

profiles of the major DNA adducts 1, 2, 3, and 4derived from coke-oven crude extract through XO-mediated metabolism as compared with S9-treated

Ž .samples Fig. 6A–D . All these major DNA adductseluted with markedly different retention times thanthe major adducts of S9-treated samples. With theHPLC gradient used, in this region are eluted DNAadducts with more hydrophobic characteristics such

w xas nitro-PAH-DNA adducts 36 . We have expectedthe presence of these adducts in the slightly polarfractions according to GC-MS analysis. However,neither on TLC map of the slightly polar fractions

nor on the crude extracts did we observe any spotcorresponding to the major spots of DNA adductstandards derived from 1-NP, 3-NF and 9-NA pre-

Žpared under the same condition XO-catalyzed ni-.troreduction .

4. Discussion

32 P-postlabeling is widely used in human biomon-w xitoring studies 17–19 for detection of bulky aro-

matic DNA adducts in human blood cells. The lowDNA adduct levels and a small amount of DNAaccessible from human blood restrict the identifica-tion of DNA adducts arising from exposure to com-plex environmental or occupational mixtures. Theuse of an in vitro system of CT DNA under specific

w xactivation conditions 23,24,28 allows the prepara-tion of a sufficient amount of DNA modified by

w xthese complex mixtures. Adams et al. 37 showedthat this acellular system in combination with 32 P-postlabeling for the direct detection of DNA damagemay provide a valuable mechanistic insight in ge-

w xnetic toxicology testing. Reddy et al. 38 alsodemonstrated a good correlation between the DNAadduct forming activity as evaluated using the invitro acellular system and Salmonella mutagenicityfor various complex mixtures. Several studies showedthe possibilities of the combination of 32 P-postlabel-ing method, TLC and HPLC analysis of 32 P-post-labeled DNA adducts in the identification of the

w xPAH-adducts formed in vitro by coal tar 39 , dieselw x w xexhaust 26 and foundry air EOM extracts 27 .

In this study, we fractionated EOM extracts of theurban air particulate imission and the coke-ovenparticulate emission into seven chemical compoundclasses by acid–base partitioning and silica gel col-umn chromatography. We utilized the in vitro acellu-

Ž .lar assays with oxidative qS9 and nitroreductiveŽ .qXO activation to investigate DNA adduct form-ing ability of these fractions using 32 P-postlabelingfor the detection of bulky aromatic andror hy-drophobic adducts.

The similar DNA adduct pattern of the majoradducts were observed for the most active fractionsof both the urban and coke-oven samples with S9-activation. The highest DNA adduct levels resulted


from the aromatic fractions and were comparable,even if the content of carcinogenic PAHs that areknown to form DNA adducts was 2.6-fold higher inthe coke-oven aromatic fraction as compared to theurban fraction. These results may indicate the in-volvement of components other than these PAHswhich may not actually bind to DNA themselves, butmay, nevertheless, affect the total DNA adduct levelsdetected. This observation is consistent with resultsfound in other in vitro and in in vivo studies of

w xcomplex mixtures 23,39,40 .The current use of 10=10 cm multidirectional

TLC alone was not sufficient for the separation ofDNA adducts derived from these complex mixtures.We examined the use of ammonium hydroxide andisopropanol solvents, as it was recommended by

w xSpencer et al. 41 , but we did not achieve betterresults. Therefore, for an improved separation ofDNA adducts, we used 20=20 cm TLC plates withour chromatographic solvents. To identify some ofthe specific DNA adducts formed, we coupled TLCwith HPLC analysis of 32 P-postlabeled adducts.Comparing TLC mobilities and HPLC profiles of themajor DNA adducts with corresponding PAH-DNAstandards, we tentatively identified in both DNAsamples of S9-treated aromatic fractions the DNAadducts presumably diolepoxide-derived from: 9-

w x w x w x w xOH-B a P, anti-BPDE, B b F, B j F, B k F, CHRY,w x w xB a A, and I cd P. Their identities were also con-

firmed by spiking of the extracted adduct spots ofinterest with corresponding standards before HPLCanalysis. These DNA adducts accounted for about57% of total DNA adducts detected by PAH-aromatic

Ž .fraction modified DNA samples qS9 . HPLC anal-ysis showed that the DNA adduct spot usually evalu-ated from TLC maps of DNA adducts derived fromcomplex mixtures as the major anti-BPDE-adductcould be a mixture of adducts possessing the sameTLC mobilities. In the current stage of our proce-dures for HPLC analysis of 32 P-postlabeled DNAadducts we were not able to identify the major DNAadducts derived from the another active fractionsŽ .organic acids and slightly polar .

The major DNA adduct spots apparent on TLCplates and analyzed by HPLC in the DNA samplesmodified by aromatic fractions were also apparent inthe samples of S9-treated crude extracts. However,these adduct spots could be partially overlapped by

DNA adduct spots derived from organic acid andslightly polar fractions possessing the similar TLCmobilities. As was further shown by HPLC analysis,the major DNA adducts resulting from the crudeextracts were identical to those derived from aro-matic fractions but they were detected by HPLC withlower sensitivity. The contributing adduct spots de-rived from the another fractions were probably underthe detection limit. Based on the results obtainedwith oxidative activation of crude extracts, we cansuggest that the contribution of the acidic fractionthat consist a major portion of total organic massŽ .28% in the urban sample is less than that calculated

Žaccording to the mass partition 39% of total DNA.adducts forming activity assuming additivity of

compound activities. This suggestion supports alsothe lower recovery of the total DNA adduct forming

Ž .activity with S9-activation 69% than the recoveryŽ .of total EOM mass 80% after fractionation. The

aromatic fraction seems to be more pronounced inthe activity of the total urban extract as could beexpected according to its mass partition. The DNAadduct forming ability of coke-oven crude extractwith oxidative activation is predominantly deter-mined by the aromatic fraction that also accounts for

Ž .the major portion of organic mass 48% . The resultsobtained in this in vitro study are consistent with the

w xin vivo studies 42,43 that evaluated the carcino-genic potency of particular fractions of diesel engineexhaust condensate and emissions from coal-firedresidential furnaces and concluded that PAHs withfour to six aromatic rings accounted for the majorityof carcinogenic activity.

In this study, we also utilized the in vitro assayŽwith nitroreductive activation catalyzed by XO a

.mammalian nitroreductase to enhance the selectiveformation of DNA adducts derived from nitroarenes.Both the butanol and nuclease P1 versions of the32 P-postlabeling assay were used to distinguish the

w xnature of DNA adducts 23,44–46 . The sensitivityof DNA adducts formed by XO-mediated metabolismto nuclease P1 treatment suggests the involvement ofnitroarenes. DNA adduct pattern of XO-treated crudeextracts and organic acids, the most active fractionfound, revealed more diffuse array of DNA adductsas compared with S9-treated samples. HPLC profilesof the major adducts resulting from XO-mediatedmetabolism were markedly different from HPLC


profiles of the major adducts of S9-treated samples.We have expected to detect at least 1 NP-, 3 NF- and9 NA-derived DNA adducts in the samples modifiedby the slightly polar fraction where these compoundswere detected by GC-MS. However, we did notobserve any adduct spots corresponding to the majorspots of DNA adduct standards derived from 1 NPand 9 NA that have markedly different TLC mobili-ties as the spots inside DRZ. The major DNA adductobserved in the samples of slightly polar fractionneither matched on TLC nor by HPLC with 3 NF-DNA adduct standard. The concentrations of nitro-PAHsrmilligram of organic mass of the fractionswas about three orders of magnitude lower than theconcentrations of PAHs in the aromatic fractions.The current in vitro assay used for enhanced forma-tion of nitroarene-DNA adducts did not facilitatetheir detection.

The highest DNA adduct levels with XO-media-ted nitroreduction were found for both the organicacid fractions. Only the presence of 9 NA and 9-nitrophenanthrene in the coke-oven acids was provedby GC-MS. The complete chemical analysis of or-ganic acids is out of the feasibility of the participat-ing laboratories. To date, we can only suggest thepresence of other compounds such as hydroxylatednitroaromatic and nitro-polyaromatic compounds ornitrophenols, that were detected in organic acids ofan ambient particulate extract by Nishioka and

w xcoworkers 47,48 . It is also not known to whatextent do oxidative or reductive metabolic pathwayscontribute to the bioactivation of these chemicals invivo. Further studies are in progress to show howthese mixtures behave in the human cell cultures.

Our results are the first in detecting DNA adductsderived from urban air and coke-oven particulatematter. It was shown that the spectra of major DNAadducts resulting from both the in vitro assays aresimilar within the fractions, for both the urban andcoke-oven samples. The fractionation procedure fa-cilitated the study of DNA binding ability of differ-ent chemical compound classes. However, it doesnot fulfil the task of the actual contribution of indi-vidual fractions within the activity of the wholeextracts. Based on the results obtained, we suggestthat the assumption of the additivity of compoundsactivities according to mass partition in the complexmixtures is uncertain considering the interaction of

compounds in the whole extract. Further effort shouldbe given to improve the strategy of sample prepara-tion and enhance the sensitivity of HPLC analysis of32 P-labeled DNA adducts derived from the originalcomplex mixture.

Acknowledgements

The authors would like to thank Dr. P.B. Farmer,University of Leicester, UK, for his critical com-ments in the preparation of this manuscript. Dr. B.Binkova acknowledges the collaboration with Dr. L.´King, NHEERL, US EPA, RTP, NC who facilitatedthe establishment of HPLC analysis of 32 P-post-labeled DNA adducts in her laboratory in Prague.This study was supported by Czech Ministry of

Ž .Environment Teplice Program , Czech Ministry ofŽ .Health grant a3159-3 and by EC grant CIPA-CT-

94-0113.

References

w x1 Polynuclear aromatic compounds: Part 3, IARC Mononogr.Eval. Carcinog. Risks Chem. Hum., Vol. 34, InternationalAgency for Research on Cancer, Lyon, France, 1984.

w x2 Overall evaluations of carcinogenicity: an updating of IARCMonographs, IARC Monogr. Eval. Carcinog. Risks Hum.,Vols. 1–42, Suppl. 7, International Agency for Research onCancer, Lyon, France, 1987.

w x3 R.R. Watts, R.J. Drago, R.G. Merril, R.W. Williams, E.Perry, J. Lewtas, Wood smoke impacted air: mutagenicityand chemical analysis of ambient air in residential area of

Ž .Juneau, Alaska, JAPCA 38 1988 652–660.w x4 J. Lewtas, C. Lewis, R. Zweidinger, R. Stevens, L. Cupitt,

Source of genotoxicity and cancer risk in ambient air, Phar-Ž .macogenetics 2 1992 288–296.

w x5 J. Lewtas, Complex mixtures of air pollutants: characterizingthe cancer risk of polycyclic organic matter, Environ. Health

Ž .Perspect. 100 1993 211–218.w x6 H. Iwado, M. Koyano, S. Goto, S. Kira, H. Hayatsu, Ubiqui-

tous presence of mutagenic and antimutagenic components inairborne particulates of two Japanese cities, Mutat. Res. 322Ž .1994 329–339.

w x7 D. Schuetzle, J. Lewtas, Bioassay-directed chemical analysisŽ .in environmental research, Anal. Chem. 58 1986 1060A–

1075A.w x8 H. Pyysalo, J. Tuominen, K. Wickstrom, E. Skytta, L. Tikka-

Ž .nen, Polycyclic organic material POM in urban air: frac-tionation, chemical analysis and genotoxicity of particulate


and vapour phases in an industrial town in Finland, Atmos.Ž .Environ. 21 1987 1167–1180.

w x9 J. Lewtas, Genotoxicity of complex mixtures: strategies forthe identification and comparative assessment of airbornemutagens and carcinogens from combustion sources, Fun-

Ž .dam. Appl. Toxicol. 10 1988 571–589.w x10 A.G. Miguel, J.M. Daisey, J.A. Sousa, Comparative study of

the mutagenic and genotoxic activity associated with inhal-able particulate matter in Rio de Janeiro air, Environ. Mol.

Ž .Mutagen. 15 1990 36–43.w x11 M. Casellas, P. Fernandez, J.M. Bayona, A.M. Solanas,

Bioassay-directed chemical analysis of genotoxic compo-nents in urban airborne particulate matter from BarcelonaŽ . Ž .Spain , Chemosphere 30 1995 725–740.

w x12 D.M. DeMarini, M.L. Shelton, D.A. Bell, Mutation spectrain Salmonella of complex mixtures: comparison of urban air

Ž . Ž .to benzo a pyrene, Environ. Mol. Mutagen. 24 1994 262–275.

w x13 D.M. DeMarini, M.L. Shelton, D.A. Bell, Mutation spectraof chemical fractions of a complex mixture: role of ni-troarenes in the mutagenic specificity of municipal waste

Ž .incinerator emissions, Mutation Res. 349 1996 1–20.w x14 J.M. Essigmann, M.L. Wood, The relationship between the

chemical structures and mutagenic specifities of the DNAlesions formed by chemical and physical mutagens, Toxicol.

Ž .Lett. 67 1993 29–39.w x15 A.P. Butler, D.G. Johnson, A.P. Kumar, S. Narayan, S.H.

Wilson, M.C. MacLeod, Disruption of transcription in vitroand gene expression in vivo by DNA adducts derived from a

Ž .benzo a pyrene diol epoxide located in heterologous se-Ž .quences, Carcinogenesis 18 1997 239–244.

w x16 D.A. Scichitano, I. Mellon, Transcription and DNA damage:Ž .a link to a kink, Environ. Health Perspect. 105 1997

145–153, Suppl. 1.w x17 P.B. Farmer, Monitoring of human exposure to carcinogens

through DNA and protein adduct determination, Toxicol.Ž .Lett. 82r83 1995 757–762.

w x18 D.K. La, J.A. Swenberg, DNA adducts: biological markers ofexposure and potential applications to risk assessment, Muta-

Ž .tion Res. 365 1996 129–146.w x19 M.C. Poirier, A. Weston, Human DNA adduct measure-

Ž .ments: state of the art, Environ. Health Perspect. 104 1996883–893, Suppl. 5.

w x20 J.A. Bond, R. Wolff, J.R. Harkema, J. Mauderly, R.F. Hen-derson, W.C. Griffith, R.O. McClellan, Distribution of DNAadducts in the respiratory tract of rats exposed to diesel

Ž .exhaust, Toxicol. Appl. Pharmacol. 96 1988 336–346.w x 3221 B. Schoket, A. Hewer, P.L. Grover, D.H. Phillips, P-post-

labeling analysis of DNA adducts in the skin of mice treatedwith petrol and diesel engine lubricating oils and exhaust

Ž .condensates, Carcinogenesis 10 1989 1485–1490.w x22 J.E. Gallagher, M.A. Jackson, M.H. George, J. Lewtas,

Dose-related differences in DNA adduct levels in rodenttissues following skin application of complex mixtures from

Ž .air pollution sources, Carcinogenesis 11 1990 63–68.w x23 J.E. Gallagher, M.J. Kohan, M.H. George, J. Lewtas, Im-

provement in the diagnostic potential of 32 P-postlabeling

analysis demonstrated by selective formation and compara-tive analysis of nitrated-PAH-derived adducts arising from

Ž .diesel particle extracts, Carcinogenesis 12 1991 1685–1691.w x24 J. Gallagher, M. George, M. Kohan, C.C. Thompson, T.

Shank, T.J. Lewtas, Detection and comparison of DNAadducts after in vitro and in vivo diesel emission exposures,

Ž .Environ. Health Perspect. 99 1993 225–228.ˇ 32w x25 B. Binkova, L. Dobias, T. Wolff, R.J. Sram, P-postlabeling´ ´ˇ ´

analysis of DNA adducts in tissues of rats exposed to coke-Ž .oven emissions, Mutation Res. 307 1994 355–363.

w x 3226 K. Savela, L. King, J. Gallagher, J. Lewtas, P-postlabelingand HPLC separation of DNA adducts formed by dieselexhaust extracts in vitro and in mouse skin and lung after

Ž .topical treatment, Carcinogenesis 16 1995 2083–2089.w x27 K. Savela, M.J. Kohan, D. Walsh, F.P. Perera, K. Hemminki,

J. Lewtas, In vitro characterization of DNA adducts formedby foundry air particulate matter, Environ. Health Perspect.

Ž .104 1996 687–690, Suppl. 3.w x28 B. Binkova, L. King, J. Lewtas, DNA adducts formed in´

vitro from fractionated extract of urban air particles, Poly-Ž .cyclic Aromatic Compounds 10 1996 243–250.

w x29 J. Lewtas, J. Chuang, M. Nishioka, B. Peterson, Bioassay-di-rected fractionation of the organic extract of SRM 1649urban air particulate matter, Int. J. Environ. Anal. Chem. 39Ž .1990 245–256.

ˇw x30 B. Binkova, J. Lewtas, I. Mıskova, J. Lenıcek, R. Sram,´ ´ˇ ´ ´ˇ ´DNA adducts and personal air monitoring of carcinogenicpolycyclic aromatic hydrocarbons in an environmentally ex-

Ž .posed population, Carcinogenesis 16 1995 1037–1046.ˇw x31 B. Binkova, J. Lewtas, I. Mıskova, P. Rossner, M. Cerna, K.´ ´ˇ ´ ¨ ´

ˇPeterkova, G. Mrackova, J. Mumford, S. Myers, R. Sram,´ ˇ ´ ´Biomarker studies in the Northern Bohemia, Environ. Health

Ž .Perspect. 104 1996 591–597, Suppl. 3.ˇw x32 R.J. Sram, I. Benes, B. Binkova, J. Dejmek, D. Horstman, F.´ ˇ ´

Kotesovec, D. Otto, S.D. Perreault, J. Rubes, S.H. Selevan, I.ˇˇ ˇSkalık, R.K. Stevens, J. Lewtas, Teplice Program—the im-´pact of air pollution on human health, Environ. Health

Ž .Perspect. 104 1996 699–714, Suppl. 4.w x33 J. Topinka, B. Binkova, G. Mrackova, Z. Stavkova, V.´ ˇ ´ ´ ´

ˇPeterka, I. Benes, J. Dejmek, J. Lenıcek, T. Pilcık, R.J. Sram,ˇ ´ˇ ˇ́ ´Influence of GSTM1 and NAT2 genotypes on placentalDNA adducts in an environmentally exposed population,

Ž .Environ. Mol. Mutagen. 30 1997 184–195.w x34 M.V. Reddy, K. Randerath, Nuclease P1-mediated enhance-

ment of sensitivity of 32 P-postlabeling analysis test for struc-Ž .turally diverse DNA adducts, Carcinogenesis 7 1986 1543–

1551.w x 3235 R.C. Gupta, P-postlabeling analysis of bulky aromatic

adducts, Postlabeling Methods for Detection of DNAAdducts, Vol. 124, IARC Sci. Publ., 1993, pp. 11–23.

w x36 L.C. King, M. George, J.E. Gallagher, J. Lewtas, Separationof 32 P-postlabeled DNA adducts of polycyclic aromatic hy-drocarbons and nitrated polycyclic aromatic hydrocarbons by

Ž .HPLC, Chem. Res. Toxicol. 7 1994 503–510.w x37 S.P. Adams, G.M. Laws, R.D. Storer, J.G. DeLuca, W.W.

Nichols, Detection of DNA damage induced by human car-cinogens in acellular assays: potential application for deter-


Ž .mining genotoxic mechanism, Mutation Res. 368 1996235–248.

w x38 M.V. Reddy, G.R. Blackburn, C.A. Schreiner, C.R. Mack-erer, Correlation of mutagenic potencies of various petroleumoils and oil coal tar mixtures with DNA adduct levels in

Ž .vitro, Mutation Res. 378 1997 89–95.w x39 N.C. Hughes, W. Pfau, A. Hewer, J. Jacob, G. Grimmer,

D.H. Phillips, Covalent binding of polyaromatic hydrocarboncomponents of coal tar to DNA in mouse skin, Carcinogene-

Ž .sis 14 1993 135–141.w x 3240 B. Schocket, A. Hewer, P.L. Grover, D.H. Phillips, P-post-

labeling analysis of DNA adducts in the skin of mice treatedwith petrol and diesel engine lubricating oils and exhaust

Ž .condensates, Carcinogenesis 10 1989 1485–1490.w x41 G.G. Spencer, A.C. Beach, R.C. Gupta, Improved thin-layer

chromatographic separation of 32 P-postlabeled DNA adducts,Ž .J. Chromatogr. 612 1993 295–301.

w x42 G. Grimmer, H. Brune, R. Deutsch-Wenzel, G. Dettbarn, J.Jacob, K.W. Naujack, U. Mohr, H. Ernst, Contribution ofpolycyclic aromatic hydrocarbons and nitro-derivatives to thecarcinogenic impact of diesel engine exhaust condensateevaluated by implantation into the lungs of rats, Cancer Lett.

Ž .37 1987 173–180.w x43 G. Grimmer, H. Brune, R. Deutsch-Wenzel, G. Dettbarn, J.

Misfeld, Contribution of polycyclic aromatic compounds tothe carcinogenic impact of flue gas condensate from coal-firedresidential furnaces evaluated by implantation into the rat

Ž .lung, J. Natl. Cancer Inst. 78 1987 935–942.w x44 J.E. Gallagher, M.A. Jackson, M.H. George, J. Lewtas,

I.G.C. Robertson, Differences in detection of DNA adductsin the 32 P-postlabeling assay after either 1-butanol extraction

Ž .or nuclease P1 treatment, Cancer Lett. 45 1989 7–12.w x 3245 R.C. Gupta, K. Earley, P-adduct assay: comparative recov-

eries of structurally diverse DNA adducts in the variousŽ .enhancement procedure, Carcinogenesis 9 1988 1687–1693.

w x46 W.Z. Whong, J.D. Stewart, T. Ong, Comparison of DNAadduct detection between two enhancement methods of the32 P-postlabeling assay in rat lung cells, Mutation Res. 283Ž .1992 1–6.

w x47 M.G. Nishioka, C.C. Howard, D.A. Contos, L.M. Ball, J.Lewtas, Detection of hydroxylated nitro aromatic and hy-droxylated nitro polycyclic aromatic compounds in an ambi-ent air particulate extract using bioassay-directed fractiona-

Ž .tion, Environ. Sci. Technol. 22 1988 908–915.w x48 M. Nishioka, J. Lewtas, Quantification of nitro- and hydrox-

ylated nitro-aromaticrpolycyclic aromatic hydrocarbons inselected ambient air daytime winter samples, Atmos. Envi-

Ž .ron. 26A 1992 2077–2087.

Genotoxicity of coke-oven and urban air particulate matter in in vitro acellular assays coupled with...

Documents

Transcript of Genotoxicity of coke-oven and urban air particulate matter in in vitro acellular assays coupled with...