Preferential binding benzo[a]pyrene diol epoxide to the linker DNA … · 2005. 4. 22. · carcinogenicity and DNAbinding has been observed for a series ofchemicals (2). Benzo[a]pyrene,

Proc. Natl. Acad. Sci. USAVol. 82, pp. 2769-2773, May 1985Cell Biology

Preferential binding of benzo[a]pyrene diol epoxide to the linkerDNA of human foreskin fibroblasts in S phase in the presenceof benzamide

(cell synchronization/specific adducts/S phase/nucleosome)

PONNAMMA KURIAN*, ALAN M. JEFFREYt, AND GEORGE E. MILO**The Ohio State University, 410 West 12th Avenue, Suite 302, Columbus, OH 43210; and tColumbia University, New York, NY 10032

Communicated by Leo A. Paquette, December 21, 1984

ABSTRACT Addition of benzamide (BZ) at the onset of Sphase inhibited expression of the neoplastic phenotype in hu-man foreskin fibroblasts treated in vitro with (±)-7a,8(3-dihy-droxy-9f3,10g3-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene(B[a]P diol epoxide) in early S phase. Analysis of the specificB[a]P diol epoxide-DNA adducts revealed that ca. 65% of thetotal adducts in BZ and non-BZ carcinogen-treated cells wasthe B[a]P diol epoxide-deoxyguanine adduct. Limited micro-coccal nuclease digestion of the early S phase nuclei from cellstreated with B[a]P diol epoxide indicated that the carcinogenbinds equally to linker and core DNA. However, when the cellswere predominantly in S phase, in the presence of BZ, therewas ca. three times more binding of B[a]P diol epoxide to thelinker DNA compared to the core region. The confluent cells inG1 cell arrest treated only with B[a]P diol epoxide also boundthe carcinogen preferentially to the linker region. These dataindicate that pretreatment of the cells with BZ at the onset of Sphase established a preferential binding pattern in the linkerDNA similar to that observed in the cells treated with B[a]Pdiol epoxide in G1 arrest.

The covalent interaction of a carcinogen with the target cellDNA is considered as one of the major events in neoplastictransformation (1-3). Proliferating cells in vivo and in vitroare found to be more sensitive to transformation by chemicalcarcinogens than nondividing cells (4, 5), and this sensitivityis increased when the cells are treated in early S phase (6, 7),presumably due to the higher binding of the carcinogen tothe newly synthesized DNA (8). A good correlation betweencarcinogenicity and DNA binding has been observed for aseries of chemicals (2).

Benzo[a]pyrene, a potent environmental carcinogen, re-quires metabolic activation by the microsomal enzymes tothe ultimate carcinogen metabolites (9, 10). The bay regionepoxide (±)-7a,8f3-dihydroxy-9f3,103-epoxy-7,8,9,10-tetra-hydrobenzo[a]pyrene (B[a]P diol epoxide) has been identi-fied as the ultimate metabolite that binds to the DNA. One ofthe major adducts formed results by the trans addition at the10-position of B[a]P diol epoxide to the exocyclic aminogroup of guanine (11). A correlation between the biologicalactivity of B[a]P diol epoxide and the presence of B[a]P diolepoxide-guanine adducts in the modified DNA has beenidentified (12). Furthermore, recent studies on the B[a]P diolepoxide-DNA adducts in the chromatin have revealed anonrandom distribution, the linker DNA being modified to agreater extent than the nucleosome core DNA (13, 14). Itseems likely that B[a]P diol epoxide is capable of formingsite-specific adducts under favorable conditions and thismight have a profound effect on gene expression.

Previous studies from this laboratory have shown that the

in vitro treatment of human foreskin fibroblasts, in early Sphase, with B[a]P diol epoxide induced a neoplastic transfor-mation in the low-passage responsive cell populations butnot in the high-passage refractory cell populations. More-over, we observed no significant difference in the level ofDNA modification or in the specific carcinogen-DNA ad-duct profiles of both of these treated cell types (15). Recent-ly, benzamide (BZ), a specific inhibitor of poly(ADP-ribose)polymerase (16), has been shown to inhibit the transforma-tion of the responsive treated cells initiated by B[a]P diolepoxide (17). Preliminary results, upon examination of spe-cific carcinogen-DNA adduct formation, indicated that BZdid not alter the DNA modification. The data presented hereexpand on the preliminary report (18) and attempt to explainhow BZ may play a role in the B[a]P diol epoxide-DNA ad-duct formation as well as distribution.

MATERIALS AND METHODS

Chemicals. [G-3H]B[a]P diol epoxide (specific activity,565 Ci/mol; 1 Ci = 37 GBq) was supplied by the MidwestResearch Institute (Kansas City, MO). [methyl-'4C]Thymi-dine (specific activity, 56 Ci/mol) and [methyl-3H]thymidine(specific activity, 80 Ci/mmol) were purchased from NewEngland Nuclear. Micrococcal nuclease, pancreatic RNase,proteinase K, and Sarkosyl were purchased from Sigma.

Cell Culture and Treatment. Human neonatal foreskin fi-broblast (HNF) cells were cultured and serially passaged byestablished procedures (19). These cells were maintained in75-cm2 flasks in complete growth medium (CM), consistingof Eagle's minimal essential medium (MEM): 25 mM Hepesbuffer (pH 7.2) supplemented with 0.1 mM nonessential ami-no acids, 1 mM sodium pyruvate, 2 mM glutamine, 50 ,g ofgentamicin per ml, 0.2% sodium bicarbonate, and 10% fetalbovine serum. Cells at PDL-5 (population doubling-5) wereblocked in the G1 phase by using a nutrient-deficient medi-um, released from the block, and treated in early S phasewith [G-3H]B[a]P diol epoxide (15). The final concentrationof [G-3H]B[a]P diol epoxide used in these experiments was0.34 ,g/ml of the treatment medium. After 3 hr of treatment,either the cells were harvested by mild trypsinization andpelleted by centrifugation at 1000 x g or the experimentalmedium was removed and the cultures were refed with CMand harvested 8 hr later.

In experiments in which the cells were treated with [G-3H]B[a]P diol epoxide in the presence of BZ, 1 mM BZ wasadded to the release medium at the onset of S phase of thecell cycle, 2 hr prior to the carcinogen treatment, and wasmaintained at this concentration throughout carcinogentreatment. The cytotoxic effects of BZ and B[a]P diol epox-

Abbreviations: B[a]P diol epoxide, (+)-7a,8,-dihydroxy-9/3,10f3-ep-oxy-7,8,9,10-tetrahydrobenzo[a]pyrene; BZ, benzamide; HNF, hu-man neonatal foreskin.

2769

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

4, 2

021

Proc. NatL Acad. Sci USA 82 (1985)

ide on the cells were determined in separate experiments bya colony-formation toxicity assay (20). The effect of 1 mMBZ on the S phase of the cell cycle during the treatment peri-od was measured by the incorporation of thymidine into thenuclear DNA by autoradiography (20) in the presence andabsence of BZ.

Cells used to isolate nuclei for limited micrococcal nucle-ase digestion were prelabeled with [methyl-14C]thymidine asdescribed. PDL-3 cells were serially seeded at 10,000 cellsper cm2 into 75-cm2 flasks and 16-18 hr later 0.2 nCi of[methyl-14C]thymidine per ml was added for 48 hr. At thistime the experimental radioactive medium was replaced withfresh CM and the cells were allowed to grow for 2-3 moredays to a confluent state. These prelabeled confluent cellswere subpassaged at 10,000 cells per cm into a nutrient-defi-cient blocking medium for 24 hr, released from the block,and treated with the carcinogen or BZ and carcinogen (de-scribed earlier). When confluent cell cultures in G1 arrestwere required, the [methyl-14C]thymnidine-prelabeled cellswere allowed to grow in the fresh CM for 5-6 days. Theywere then treated with 0.34 ,ug of [G-3H]B[a]P diol epoxideper ml of CM without fetal bovine serum for 3 hr and thenwere harvested.

Isolation of Nuclei. All isolation procedures were carriedout between 0C and 40C. The pelleted HNF fibroblast cellswere suspended in 0.25 M sucrose/10 mM Tris maleate, pH7.5, containing 1 mM dithiothreitol, 3 mM calcium acetate,and 2 mM magnesium acetate (pH 7.5) (buffer A) at 1D7 cellsper ml and were homogenized in a stainless steel mortar witha Teflon pestle with a 229-,um clearance connected to a Tri-Rhomogenizer. When 80% or more of the cells were broken,as evidenced by examining the cells under phase microscopyat x430 following staining by toluidine blue, the cells wereprepared for nuclei isolation in the following manner: 6.5 mlof 1.8 M sucrose in buffer A was added to 1 ml of homoge-nate, mixed by two up and down strokes, layered over 5 mlof 1.8 M sucrose in buffer A, and centrifuged at 40,000 x gunder vacuum at 4°C for 1 hr. The nuclei pellet was resus-pended in 1.0 M sucrose in buffer A by gentle homogeniza-tion and centrifuged at 10,000 x g for 10 min. The pellet atthe bottom was resuspended in the desired solution andquick-frozen at -70°C.

Analysis of DNA Adducts. Nuclei prepared from cells har-vested immediately after the carcinogen treatment (3 hr) aswell as those refed with the CM for 8 hr after treatment weresuspended in 10 mM Tris HCl/1 mM Na2EDTA buffer, pH7.5, containing 0.1% NaDodSO4 (6-8 x 106 nuclei per ml ofbuffer) and quick-frozen at -700C until used for adduct anal-ysis. Nuclear DNA was purified and analyzed by the proce-dure described by Tejwani et al. (15).

Micrococcal Nuclease Digestion and Assay. The digestionand assay of the digested products were carried out bya modification of the combined procedures of Jack andBrookes (13) and Kaneko and Cerruti (21). Briefly, the nu-clei were suspended in 0.25 M sucrose/i mM Tris.HCl/0.5mM CaCl2, pH 7.4 (digestion buffer), at co. 5 x 106 nucleiper 100 ,ul and treated with micrococcal nuclease (2.5 ,ug/ml)at 37°C; 30-/l aliquots were removed at various times be-tween 0 and 60 min. The reaction was stopped by the addi-tion of an equal volume of 10 mM Tris HCI/10 mM EDTA,pH 7.4, followed by addition of Sarkosyl to a final concen-tration of 0.5%. The nuclear lysate was then incubated with20 ug of RNase (previously heat treated in the dark for 10min at 100°C) per ml at 37°C for 30 min followed by treatmentwith 100 Ag of proteinase K per ml for 60 min at 450C. Theincubation medium was mixed with an equal volume of re-distilled phenol made up as a phenol reagent [phenol/chloro-form/isoamyl alcohol, 24:24:1 (vol/vol)] by spinning in avortex for 2 min. After mixing the phases, the homogeneoussuspension was centrifuged at 5000 rpm for 5 min in a Beck-

man Microfuge, model 11. The top aqueous layer was re-moved and the phenol layer was extracted twice with half ofthe volume of 10 mM Tris HCl/10 mM EDTA, pH 7.4. Thecombined aqueous buffer phases were extracted once again(vol/vol), with the phenol reagent as above, and the finalaqueous phase was extracted twice with water-saturatedether to remove traces of phenol. Excess ether was removedby placing the tubes in a 60C water bath for 5 min. Twenty-five microliters of calf thymus DNA at 2 mg/ml was addedas a carrier and the DNA was precipitated with 2 vol of coldethanol at -20TC overnight. The precipitate was recoveredby centrifugation at 10,000 rpm (in the Microfuge) for 7 min.The precipitate was washed once carefully with cold 80%ethanol, and the pellet was dissolved in 50 ,l of 1 M HC1 byheating at 70TC for 1 hr. Following neutralization with 50 A.lof 1 M KOH (CO2 free), the solution was transferred quanti-tatively to a counting vial by using 150 ,4L of 50mM Tris HCl(pH 7.5). Ten milliliters of ScintiVerse II (Fisher) was usedfor assaying radioactivity in a Beckman model 9000 scintilla-tion counter. The phenol phase and the combined ethanolsupernatant also were assayed for the presence of radioac-tivity. The 3H dpm were corrected for spillover of the 14Cradiation. The ratio of 3H/14C dpm in the precipitate wasused as a measure of the concentration of B[aIP diol epox-ide-DNA adducts.

Determination of Adduct Concentration in Linker DNA.The concentration of the adducts on the linker DNA wascalculated by using the equation 200x = 146y + 54z (13), inwhich x = concentration of the adducts on undigested DNA,y = concentration of the adducts on core DNA (after 27%digestion), z = concentration of the adducts on the linkerDNA, 200 = base pairs in nucleosome, and 146 = base pairsin core DNA. To compute these concentrations of adducts in[methyl-'4C]thymidine prelabeled [G-3H]B[a]P diol epoxide-treated cells, the 3H/14C ratio of ethanol-precipitable nuclearmaterial at a definite time (t = 5-40 min) during limited mi-crococcal nuclease digestion was divided by the H/14C ratiomeasured at t = 0 min of digestion.

RESULTS

Earlier studies in this laboratory have revealed that 1 mMBZ inhibits the carcinogenic response of B[a]P diol epoxide-treated HNF cells (17). The relative cloning efficiency (20), ameasure of cytotoxicity, was determined over a range ofconcentrations ofBZ from 0.01 to 4 mM. No inhibition of therelative cloning efficiency compared to the untreated controlwas observed with 1 mM BZ, the concentration used in ourexperiments. At 4 mM, =15% inhibition was observed. Al-though B[a]P diol epoxide at 0.34 pg/ml used in the bindingstudies was toxic to the cells at the low cell density em-ployed in the cloning assay, at the high cell density condi-tions used in our experiment the cells grew well after a shortlag phase and transformants were observed. However, thelow specific activity and the low percentage of base modifi-cation ofDNA by tG-3H]B[a]P diol epoxide necessitated theuse of this concentration.To investigate the effect of BZ on the early S phase of

these cells, the time period at which they are very suscepti-ble to the carcinogen, [methyl-3H]thymidine incorporation inthe presence and absence of BZ was measured by autoradi-ography. BZ was added 10 hr after release from the G1 blockby the nutrient-deficient medium, at the onset of S phase.Radiolabeled nuclei containing 50 grains or more for a totalof 1000 nuclei were counted for each point on the graph. Thepercentages of radiolabeled nuclei were calculated, up to 21hr after release from synchrony. No difference was observedin the number of nuclei going through S phase in the pres-ence or absence of BZ.

2770 Cell Biology: Kurian et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

4, 2

021

Proc. NatL Acad. Sci. USA 82 (1985) 2771

DNA was isolated from the nuclei of the cells treated with[G-3H]B[a]P diol epoxide or with [G-3H]B[a]P diol epoxideand 1 mM BZ. The levels ofDNA modification were 6.7 and5.6 adducts per 106 bases, respectively, in the one experi-ment and 2.3 and 1.8 adducts per 106 bases, respectively, inthe second experiment. The reason for the variability in thelevel of binding of B[a]P diol epoxide to cells may reflect inpart the instability of the B[a]P diol epoxide and the prob-lems associated with adding it to the medium, in which it isvery unstable, and it being absorbed into the cells where itmay bind to DNA. Once inside cells, microenvironments ex-ist in which the diol epoxide is quite stable. However, therewas no significant variation in the number of DNA adductsin the two differently treated samples within the same ex-periment. The slight variations could not be correlated withthe extent of inhibition of the transformations (17). TheHPLC profiles of the modified deoxyribonucleosides (Fig. 1Upper and Lower) were qualitatively and quantitatively sim-ilar. Eight hours after concluding the B[a]P diol epoxidetreatment, when 50% of the radiolabeled carcinogen was re-moved from the cells, nuclei were isolated and the DNA wasanalyzed. The specific B[a]P diol epoxide-DNA modifiedadducts were 1.2 and 0.75 modified bases of 713B[a]P diolepoxide-deoxyguanine per 106 bases in the presence and ab-sence of BZ, respectively. Moreover, the ratios of specificmodified bases were also qualitatively and quantitativelysimilar in the BZ- and non-BZ-treated cells (Fig. 2 Upper and

200

*<- 7(3

150-

7a

100 Tetrols

50 ~~~~~~dA

0

I-7I

7a

Tetrols

I

*<- 7j

160

1207a

80 Tetrols

I~'40-

1'I,0 10 20 30 40

Time, min

FIG. 1. HPLC profile of [G-3H]B[a]P dio]ducts formed in the presence and absence of 1were blocked at G1, released, and treated 2 hrphase with [G-3H]B[a]P diol epoxide for 3 hr. Denzymatically digested. The modified deoxyriseparated by Sephadex LH-20 chromatographcochromatographed on HPLC with authenticB[a]P diol epoxide-deoxyguanine; 7,3: 7(-Edeoxyguanine. (Upper) DNA from non-BZ-trBZ at 1.0 mM was added to the incubationB[a]P diol epoxide treatment.

50Time, min

FIG. 2. HPLC profile of specific [G-3H]B[a]P diol epoxide-DNA adducts remaining in the cells at 8 hr after conclusion of 3-hrtreatment with [G-3H]B[a]P diol epoxide in the presence (Lower)and absence (Upper) of BZ. PDL-5 cells were incubated with B[a]Pdiol epoxide or with BZ and B[a]P diol epoxide as described in thelegend to Fig. 1. At the end of the 3-hr treatment period, the cellswere refed with CM for 8 hr. DNA was then isolated and analyzedby HPLC as described in the legend to Fig. 3. 7a: 7a-B[a]P diolepoxide-deoxyguanine; 7(3: 7,B-B[a]P diol epoxide-deoxyguanine;dA: B[a]P diol epoxide-deoxyadenine. (Upper) DNA from B[a]Pdiol epoxide-treated cells. (Lower) BZ at 1.0 mM was added to theincubation medium 2 hr prior to B[a]P diol epoxide treatment.

Lower). Adduct analysis at the 8-hr time point was done onlyin one experiment. It required a large number of plates and,based on the large variation in the extent of modification inthe above experiments, did not suggest that any dominanteffect existed that might suggest a limiting role for DNA re-pair.

Since no significant difference in specific DNA modifica-tion could be detected, the distribution of the adducts on thelinker versus core DNA regions was examined by the limitedmicrococcal nuclease digestion technique. The time courseof the excision of the adducts on the nuclear DNA was fol-lowed by limited digestion of the nuclei prelabeled with[methyl- C]thymidine and treated with [G-3H]B[a]P diol ep-

50 60 70 oxide or with BZ and [G-3H]B[a]P diol epoxide and purifica-tion of the undigested DNA. Confluent cells (G1), prelabeledwith [methyl-14C]thymidine and treated with [G-3H]B[a]P

I epoxide-DNA ad- diol epoxide for 3 hr, were also analyzed similarly for com-mM BZ. PDL-5 cells parison, as it has been shown previously that B[a]P diol ep-;after the onset of S oxide preferentially binds to the linker region in quiescent

ibonucleosides were cells in a confluent dense state (13). The extent of residualiy and subsequently radiolabel remaining in the ethanol precipitate is measuredstandards. 7a: 7a- as a function of time from 0 to 40 min. The values obtained,

3[a]P diol epoxide- following the limited micrococcal nuclease digestion, wereeated cells. (Lower) computed as the percentage of precipitable radioactivity be-iedium 2 hr prior to fore digestion. The 14C dpm remaining in the ethanol precipi-

tate following the limited digestion are a measure of the re-

120

100 _

80 _

601

40

20

E0.

Cell Biology: Kurian et al.

1.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

4, 2

021

Proc. Nati Acad Scd USA 82 (1985)

sidual undigested DNA and are an indirect measurement ofthe rate of digestion of the DNA in the linker and core re-gions. A measurement of residual 41 dpm in the ethanol pre-cipitate following limited micrococcal nuclease digestionrepresents the residual modified DNA. This is an indirectmeasurement of the rate of total DNA adduct removal. Theresults are shown in Table 1. There was no significant varia-tion in the rates of digestion of the nuclei (14C removal) fromthe three samples. The decrease in the rates of digestion withtime, as the digestion proceeded from the linker to the coreDNA, was also similar in all three samples, indicating thatmodification of the nuclear DNA by B[a]P diol epoxide didnot affect the micrococcal nuclease digestion. In contrast tothe [methyl-'4C]thymidine removal, there was a significantdifference in the rate of adduct excision (3H removal) fromthe DNA of the three samples. The rates of B[a]P diol epox-ide-DNA adduct excision as well as DNA digestion werealmost the same throughout the micrococcal nuclease diges-tion period for the B[a]P diol epoxide-treated S-phase nuclei.However, the rate of the adduct removal from the DNA ofBZ- and B[a]P diol epoxide-treated S-phase nuclei and ofB[a]P diol epoxide-treated confluent cell nuclei during thefirst 10-20 min of digestion, when the linker DNA was solu-bilized, was higher than the rate of DNA digestion.These differences in the rates became negligible during the

20-40 min of digestion. Further digestion of the DNA (datanot shown) appeared to proceed at a slightly higher rate thanthe rate of adduct removal in both of these samples. It can beseen from Table 1 that when =27% of the DNA (linker DNA)was digested, ca. 70% of the B[a]P diol epoxide adduct re-mained in the precipitable DNA (core DNA) of the B[a]Pdiol epoxide-treated S-phase nuclei, whereas only ca. 50% ofthe adducts were associated with the core DNA of the BZ-and B[a]P diol epoxide-treated nuclei and of the nuclei fromB[a]P diol epoxide-treated confluent cells. This indicatedthat B[a]P diol epoxide bound almost uniformly to the linkerand core DNA of the Bia]P diol epoxide-treated S-phase nu-clei, whereas B[a]P diol epoxide attached preferentially to

Table 1. Kinetics of micrococcal nuclease digestion of the nucleifrom HNF cells after treatment with [G-3H]B[aJP diol epoxideunder three different conditions

Time ofnuclei % 14C in % 3fl in

digestion, ethanol ethanolCell cycle and treatment nin precipitate precipitateCells in early S phase

Treated with BMaIP diol 0 100 100epoxide 5 85.05 ± 3.6 81.8 ± 4.3

10 73.90 ± 3.9 68.9 ± 4.920 66.00 ± 2.9 61.2 ± 5.540 55.90 ± 6.6 51.5 ± 4.8

Treated with B[a]P diol 0 100 100epoxide and BZ 5 88.9 ± 2.3 74.8 ± 4.4

10 77.8 + 2.5 57.2 ± 4.620 70.4 ± 3.4 50.4 ± 5.540 59.6 ±5.9 40.9 3.2

Confluent cells treated 0 100 100with B[a]P diol epokide 5 90.5 ± 2.4 79.5 ± 2.0

10 81.7 ± 2.9 62.8 ± 3.420 72.0 ± 3.6 50.9 ± 3.740 62.2 ± 4.0 43.2 ± 3.4

Pt)L-5 cells were incubated with [G-3HJB[a]P diol epoxide 2 hrafter the onset of S phase for 3 hr with or without BZ, and the nucleiwere isolated and digested with micrococcal nuclease. Nuclei iso-lated from confluent cells treated with [G-3H]B[a]P diol epoxide atPDL-5 were also used for micrococcal nuclease digestion. Bachexperimental point is an average of five measurements. The notation± represents 1 oC standard deviation from the mean.

U~~~~~~~~~~

a 0.6 -<

'o 0.4090.6

a~0.2

0 10 20 30 40 50 60% digested DNA

FIG. 3. Relative adduct concentration in the residual DNA(3H/14C ratio of ethanol-precipitable material at various times of di-gestion divided by the 3H/14C ratio of ethanol-precipitable materialbefore digestion) during micrococcal nuclease digestion as a func-tion of the percentage of DNA digested in nuclei from B[a]P diolepoxide-treated cells. Values from three independent experimentsare used. o, Nuclei from S-phase cells treated with [G-3H]B[a]P diolepoxide; e, nuclei from S-phase cells treated with BZ and [G-3H]B[a]P diol epoxide; A, nuclei from confluent cells (G1 arrest)treated with [G- H]B[a]P diol epoxide.

the linker DNA compared to the core DNA of the other twodifferently treated nuclear preparations. Chromatography ofthe digests by slab gel electrophoresis on 6% polyacrylamidegels did not show any detectable distinct difference in theband patterns of the DNA of the three differently treatednuclear preparations. We observed a prominent band of ca.146 base pairs in all three nuclear preparations following a10- to 20-min digestion.A plot of the relative adduct concentration (3H/'4C ratio in

the ethanol-precipitable material at various times of diges-tion divided by the 3H/14C ratio in the ethanol-precipitablematerial before digestion) versus the percentage of DNA di-gested is shown in Fig. 3. There is a rapid decrease in theadduct concentration in the confluent cell nuclei as well as inthe BZ- and B[a]P diol epoxide-treated S-phase nuclei. Afterthe linker DNA was digested (27% digestion), the adductconcentrations became constant, indicating that Bta]P diolepoxide binds preferentially to the linker DNA. Nuclei iso-lated from carcinogen-only-treated S-phase cells exhibitedonly a slight decrease in adduct concentration during diges-tion.At 27% digestion, when all of the linker DNA was solubi-

lized, the relative adduct concentration in the core DNA ofthe confluent cells exposed to the B[a]P diol epoxide was70o of the undigested DNA (Fig. 3). This is equivalent to 2.6times more binding to the linker DNA compared to the nu-cleosome core DNA (see Materials and Methods for calcula-tion) and is in close agreement with the data of Jack andBrookes (13) for primary mouse embryo cells. S-phase cells,exposed to the carcinogen and 8IZ also, exhibited a similarbinding, the relative adduct concentration in the core DNAbeing 68% of the total DNA and hence 2.7-fold more bindingon the linker DNA. However, in the absence of BZ, thisbinding was considerably decreased, only 1.4 times morebinding to the linker region of the nucleosomes. Since only80% of the cells went through the S phase, the preferentialbinding observed in this case may be due to that fraction ofthe cells that did not undergo replication.

DISCUSSIONSeveral chemical carcinogens, including B[a]P diol epoxide,can induce neoplastic transformation ofhuman diploid fibro-

2772 Cell Biology: Kurian et aL

Dow

nloa

ded

by g

uest

on

Aug

ust 2

4, 2

021

Proc. Natl. Acad Sci. USA 82 (1985) 2773

blasts (6, 15, 17). BZ, a competitive inhibitor of chromatin-associated poly(ADP-ribose) polymerase, has been shown toinhibit transformation by a variety of chemicals, such asB[a]P diol epoxide, N-methyl-N'-nitro-N-nitrosoguanidine,and methylazoxymethanol acetate (17, 22). The specificityof this inhibition appears to involve the covalent molecularbinding of the inhibitor to the poly(ADP-ribose) polymer-ase-poly(ADP-ribose)-DNA complex (unpublished data).

Earlier studies (17) have shown that BZ at 1.0 mM doesnot have any effect on methylazoxymethanol acetate-in-duced strand breaks or the extent of methylation of 06-gua-nine, though it inhibits poly(ADP-ribose) synthesis and alsothe expression of neoplastic transformation by methylazoxy-methanol acetate. A similar finding has been reported byBorek et al. (22) for the analogue 3-aminobenzamide. A 1.0mM concentration of 3-aminobenzamide significantly inhib-ited transformation by N-methyl-N'-nitro-N-nitrosoguani-dine as well as poly(ADP-ribose) synthesis, whereas it hadno significant effect on repair replication, DNA synthesis bysalvage pathways, and strand-breakage frequencies in x-ray-damaged cells. However, contrary to our results, they foundthat strand-breakage frequencies in carcinogen-damagedcells increased with increasing concentration of 3-aminoben-zamide. Moreover, in a more recent study, we have shownthat BZ had no effect on UV-induced thymidine dimer for-mation or the removal of these dimers (unpublished data).However, it did inhibit transformation. The above data sug-gest that factors other than DNA damage and repair may beinvolved in the inhibition of transformation by carcinogen inthe presence of BZ.The present study was designed for detecting any possible

qualitative or quantitative difference in the formation of thecarcinogen-DNA adduct in the presence of BZ. The cellswere treated at early S.phase of the cell cycle with the car-cinogen or with BZ and carcinogen. Determination of thepercentage of the radiolabeled nuclei by autoradiography, af-ter [methyl-3H]thymidine treatment, revealed that BZ didnot alter the extent of incorporation of [methyl-3H]thymidinein the nuclei involved in DNA replication during the timeperiod of the experimental treatment. Analysis of the specif-ic B[a]P diol epoxide-DNA adducts by HPLC, from the car-cinogen-modified DNA isolated at the conclusion of the ex-perimental treatment either with the carcinogen or with thecarcinogen and BZ treatment and also at 8 hr after the treat-ment, did not reveal any qualitative or quantitative differ-ence in the presence of specific carcinogen adducts formedat a transforming dose of the carcinogen.Recent studies on the distribution of the carcinogen modi-

fications in the DNA of the chromatin using the limited mi-crococcal nuclease technique have indicated preferentialbinding of carcinogens to the linker DNA of chromatin (14,23). This localized binding of the carcinogen in specific re-gions ofDNA may play an important role in the regulation ofgene expression. Jack and Brookes (13) have shown thatB[a]P diol epoxide binds three times more to the linker DNAcompared to the core DNA when confluent mouse embryocells were treated. Our studies with treated confluent HNFcells also revealed similar results. There was 2.6 times morebinding to the linker DNA versus core DNA. Moreover, notransformant was observed (unpublished data). HNF fibro-blasts, when treated in S phase with B[a]P diol epoxide inthe presence of BZ, behave similarly to the cells in G, arrest,with 2.7 times more binding to the linker DNA than to thecore DNA. But preferential binding of the B[a]P diol epoxideto the linker DNA was very low (1.4 times) in the absence ofBZ. It has been shown (24, 25) that nucleosomes are redis-tributed randomly along the nuclear DNA during replication.The mechanism proposed to account for this redistribution

during DNA synthesis is that of proteins sliding along theDNA and dissociation and reassociation of the nuclear pro-teins. If so, we can easily envision how preferential bindingof B[a]P diol epoxide to the linker DNA of the chromatinobserved in the confluent cells may be absent when the cellsat S phase are treated with the carcinogen. The re-establish-ment of the preferential binding of the carcinogen to the S-phase chromatin in the presence of BZ, a potent poly(ADP-ribose) polymerase inhibitor and an anticarcinogen, suggeststhat poly(ADP-ribosyl)ation of the nuclear protein may beinvolved in an as yet unknown manner in retaining the integ-rity of the linker and core regions of the DNA during replica-tion in early S phase. This, in turn, might mask the modifica-tion of certain critical sites by the carcinogen, thereby inhib-iting the neoplastic transformation. Another possibleexplanation for the observed preferential binding of B[a]Pdiol epoxide to the linker DNA of B[a]P diol epoxide- andBZ-treated cells is that DNA repair occurs preferentially inthe linker regions of the chromatin and this repair is inhibitedby BZ. If so, the B[a]P diol epoxide- and BZ-treated cellswould be expected to have higher levels of binding to linkerDNA than the control S-phase cells. However, our datashow that the levels of modification in the total DNA werethe same in the presence and absence of BZ. If the abovehypothesis is true, BZ, in addition to inhibiting repair in thelinker region, would also have to cause an increased excisionof adducts from the core regions to explain our results. Al-though BZ may have different and opposite effects on thelinker and core regions, we feel it more plausible that theinhibitor may act through its effect of poly(ADP-ribose)polymerase.We thank Dr. Ernest Kun for his expert advice during the course

of this work, Louise Bartels for technical help, and Kathy Bartunekfor typing the manuscript. This work was supported in part by Envi-ronmental Protection Agency Grant R810407-01-0 (G.E.M.).

1. Heidelberger, C. & Davenport, G. R. (1961) Acta Unio Int. ContraCancrum 17, 55-63.

2. Brookes, P. & Lawley, P. D. (1964) Nature (London) 202, 781-784.3. Miller, J. A. (1970) Cancer Res. 30, 559-576.4. Craddock, V. M. & Freii, J. V. (1974) Br. J. Cancer 30, 503-511.5. Kakunaga, T. (1978) Cancer Res. 35, 1637-1642.6. Milo, G. E. & DiPaolo, J. A. (1978) Nature (London) 275, 130-132.7. Kaufmann, W. K., Kaufman, D. G., Ricek, J. M. & Wenk, M. L.

(1981) Cancer Res. 41, 4653-4660.8. Cordeiro-Stone, M., Topal, M. D. & Kaufman, D. G. (1982) Carci-

nogenesis 3, 1119-1127.9. Sims, P., Grover, P. L., Swaisland, A., Pal, K. & Hewer, A. (1974)

Nature (London) 252, 326-328.10. Philips, D. H. (1983) Nature (London) 303, 468-475.11. Jeffrey, A. M., Weinstein, A. B., Jennette, K. W.,

Grzeskowiak, K., Nakanishi, K., Harvey, R. G., Autrup, H. &Harris, C. (1977) Nature (London) 269, 348-350.

12. Yang, L. L., Maher, V. M. & McCormick, J. J. (1980) Proc. Nat!.Acad. Sci. USA 77, 5933-5937.

13. Jack, P. L. & Brookes, P. (1981) N. A. Res. 9, 5533-5552.14. Jack, P. L. & Brookes, P. (1982) Carcinogenesis 3, 341-344.15. Tejwani, R., Jeffrey, A. M. & Milo, G. E. (1982) Carcinogenesis 3,

727-732.16. Purnell, M. R. & Whish, W. J. D. (1980) Biochem. J. 185, 775-777.17. Kun, E., Kirsten, E., Milo, G. E., Kurian, P. & Kumari, L. (1983)

Proc. Nat!. Acad. Sci. USA 80, 7219-7223.18. Kurian, P., Jeffrey, A. M., Kun, E. & Milo, G. E. (1984) Fed.

Proc. Fed. Am. Soc. Exp. Biol. 43, 1887.19. Milo, G. E. & DiPaolo, J. A. (1980) Int. J. Cancer 26, 805-812.20. Oldham, J. W., Allred, L. E., Milo, G. E., Kindig, 0. & Capen,

C. C. (1980) Toxicol. App!. Pharmacol. 52, 159-168.21. Kaneko, M. & Cerruti, P. A. (1980) Cancer Res. 40, 4313-4319.22. Borek, C., Morgan, W. F., Ong, A. & Cleaver, J. E. (1984) Proc.

Nat!. Acad. Sci. USA 81, 243-247.23. Kaneko, M. & Cerutti, P. A. (1982) Chem. Biol. Interact. 38, 261-

274.24. Seale, R. L. (1976) Proc. Nat!. Acad. Sci. USA 73, 2270-2274.25. Hancock, R. (1978) Proc. Nat!. Acad. Sci. USA 75, 2130-2134.

Cell Biology: Kurian et aL

Dow

nloa

ded

by g

uest

on

Aug

ust 2

4, 2

021