Chem. Res. Toxicol. 1991,4, 115-122
115
Characterization of Covalently Modified Deoxyribonucleosides Formed from Dibenr[ a ,]]anthracene in Primary Cultures of Mouse Keratinocytes Raghunathan V. Nair,t Rosalynn D. Gill,? Anusha N. Nettikumara,? Wanda Baer-Dubowska,t Cecilia Cortez,* Ronald G. Harvey,* and John DiGiovanni*tt Department of Carcinogenesis, Science Park-Research Division, University of Texas M. D. Anderson Cancer Center, P.O. Box 389, Smithville, Texas 78957, and The Ben May Institute, University of Chicago, 5841 Maryland Avenue, Chicago, Illinois 60637 Received August 13,1990
Identification of various deoxyribonucleoside adducts formed in primary cultures of mouse keratinocytes exposed t o dibenz[aj]anthracene (DB[aj]A) is presented. A preliminary analysis of the DNA adducts formed from 7-methyldibenz[aj]anthracene(7MeDB[aj]A) also is presented. Cultures of keratinocytes obtained from dorsal skins of female SENCAR mice were exposed to 0.5 hg of tritium-labeled hydrocarbons/mL of medium for 24 h. The total DNA binding was 2.23 f 0.54 and 5.28 f 0.97 pmol of hydrocarbon/mg of DNA for DB[aj]A and 7MeDB[aj]A, respectively. These binding values represented the radioactivity associated with the modified deoxyribonucleosides separated from the normal deoxyribonucleosides on Sephadex LH-20 columns following enzymatic digestion of isolated DNA. Treatment of keratinocytes with DB[a j ] A produced adduct peaks corresponding to marker adducts derived from trans addition of both deoxyguanosine as well as deoxyadenosine residues t o the (+) enantiomer of the anti-diol epoxide where the deoxyadenosine adducts were predominant. In addition, DNA adduct peaks corresponding t o markers of trans and cis addition, respectively, of deoxyguanosine and deoxyadenosine to the (+)-syn-diol epoxide were also noted in these chromatograms. A major DNA adduct in cells exposed to D B [ a j ] A was tentatively identified as resulting from the addition of deoxyadenosine to DB[aj]A-5,6-oxide. Several other later eluting DNA adduct peaks, not corresponding t o any of the marker adducts, were also present in these chromatograms. In comparison, when cells were exposed t o the more biologically potent 7-methyl analogue, a t least 12 DNA adduct peaks were consistently observed in HPLC chromatograms. One peak in these chromatograms coeluted with a marker adduct derived from trans addition of deoxyadenosine to the bay region (+)-anti-diol epoxide of 7MeDB[aj]A, while the additional DNA adducts remain unidentified a t the present time. The current results are discussed in terms of the mechanism(s) of tumor initiation by D B [ a j ] A and its derivatives.
Introduction The covalent modification of deoxyribonucleosides in cellular DNA by reactive metabolites of xenobiotics such as the polycyclic aromatic hydrocarbons (PAH)' is believed to be one of the critical events in the process of tumor initiation induced by these chemicals (1-5). Recent studies have attempted to identify similarities and differences in the levels of modifications of specific bases in nuclear DNA by various environmentally occuring PAH relative to their carcinogenic and/or tumor-initiating potencies (6). The formation of covalently modified DNA bases has been investigated in a wide variety of target tissues exposed to different PAH. Such studies utilizing the hydrocarbon benzo[a]pyrene (B[a]P) have consistently demonstrated deoxyguanosine (dGuo) as the major deoxyribonucleoside in cellular DNA modified by this hydrocarbon (6). Several methyl-substituted PAH are more potent carcinogens than the corresponding nonmethylated derivatives (7-10). Among the methyl-substituted PAH, comparison of the tumor-initiating ability of 7,12-dimethyl-
* To whom correspondence should be addressed. 'University of Texas M. D. Anderson Cancer Center. 1 University of Chicago.
benz[a]anthracene (DMBA) with that of B[a]P has revealed that, for an equivalent papilloma response, DMBA is nearly 70 times more potent than B[a]P in mouse skin (11). However, at these doses, the total covalent binding to mouse epidermal DNA was 2.2 pmol/mg of DNA for DMBA and 17.7 pmol/mg of DNA for B[a]P (11). Unlike B[a]P, DMBA has been found to bind substantially to deoxyadenosine (dAdo) in addition to dGuo residues in cellular DNA (12, 13). Thus, in contrast to the 2-39'0 binding to dAdo observed with B[a]P (14, 15), as much as 42-48% of the total DNA binding observed with DMBA was found in dAdo residues in the DNA (16, 17). This differential and selective modification of deoxyribonucleosides in cellular DNA by these hydrocarbons having substantially different tumor-initiating potencies may contribute to the lack of a linear correlation with total Abbreviations: PAH, polycyclic aromatic hydrocarbons; B[a]P, benzo[a]pyrene; dGuo, deoxyguanosine; dAdo, deoxyadenosine; dCyt, deoxycytidine; dN, deoxyribonucleoside; DMBA, 7,12-dimethylbenz[a]anthracene; B [c]Ph, benzo [c]phenanthrene; DB [ a i ]A, dibenz [ a i ]anthracene; 7MeDB[aJ]A, 7-methyldibenz[aj]anthracene;THF, tetrahydrofuran; MEM, minimum essential medium; FBS, fetal bovine serum; PBS, phosphate-buffered saline; GITC, guanidinium isothiocyanate; 7OHM-12-MBA, 7-(hydroxymethyl)-12-methylbenz[a]anthracene; TPA, 12-0-tetradecanoylphorbol13-acetate; ODS, octadecylsilane; DE, diol epoxide.
0 1991 American Chemical Society
116 Chem. Res. Toxicol., Vol. 4, No. I , 1991
DNA binding. In addition, such information has led to the hypothesis that binding to dAdo residues in DNA may be critical for the greater biologic activity of PAH such as DMBA (13). This hypothesis is further supported by recent studies showing specific A T transversion mutations in the Ha-ras gene following initiation with DMBA but not B[a]P (reviewed in ref 18). Preferential binding to dAdo residues in cellular DNA has also been noted in rodent embryo cell cultures exposed to benzo[c]phenanthrene (B[c]Ph) (19). The significance of such selective modification of dAdo or other residues in cellular DNA relative to the carcinogenic potency of B[c]Ph and other hydrocarbons remains to be determined. To further address this question, we have examined the relative levels of modification of deoxyribonucleosides in primary cultures of mouse keratinocytes exposed to two homologous hydrocarbons: dibenz[aj]anthracene (DB[aj]A) and 7-methyldibenz[aj]anthracene(7MeDB[aj]A). Previous work from our laboratory established that 7MeDB[aj]A is approximately 3 times more potent than DB[aj]A as a tumor initiator in mouse skin (10). In the present work, we have examined the binding of these xenobiotics to specific deoxyribonucleosides in mouse epidermal cell DNA.
-
Experimental Procedures Chemicals. D B [ a j ] A (20),7MeDB[aj]A (20),and racemic D B [ a j ] A anti-diol epoxide (21) were prepared according to published procedures. Additional quantities of DB[a,j]A were also acquired through the chemical carcinogen repository of the National Cancer Institute. T h e K-region (f)-DB[aj]A-5,6-oxide was synthesized from the parent hydrocarbon by the general method previously described (22). The DB[aj]A syn-diol epoxide and the 7MeDB[a,j]A anti-diol epoxide were synthesized by appropriate modifications of procedures previously utilized for the synthesis of the analogous diol epoxide derivatives of other polycyclic hydrocarbons (21,23);details of these syntheses will be published elsewhere. Tritiation of hydrocarbons was carried out by Chemsyn Science Laboratories, Lenexa, KS. Specific ativities were 1.75 Ci/mmol for [3H]DB[aj]A and 1.47 Ci/mmol for [3H]7MeDB[aj]A. Percoll, calf thymus DNA, DNase 1 (bovine pancreas, EC 3.1.21.1), snake venom phosphodiesterase (Crotalus atrox, EC 3.1.4.1), Escherichia coli alkaline phosphatase (type 111, EC 3.1.3.1) suspension in 3.5 M ammonium sulfate, and proteinase K (Tritirachium album type XI, EC 3.4.21.14) were purchased from Sigma Chemical Co., St. Louis, MO. RNase A (bovine pancreas, EC 3.1.4.22) was obtained from Worthington Biochemical Co., Freehold, NJ. Sephadex LH-20 was purchased from Pharmacia Fine Chemicals, Inc., Piscataway, NJ, and [ [N-[N'-[m-(dihydroxyboryl)phenyl]succinamyl]amino]ethyl]cellulose (dihydroboronyl cellulose) was a generous gift from Dr. William M. Baird, Purdue University, West Lafayette, IN. Eagle's minimum essential medium containing no ea2+was obtained from Whittaker M.A. Bioproducts, Walkersville, MD, and guanidine isothiocyanate was purchased from Bethesda Research Laboratories, Gaithersburg, MD. [3H]dGuo, [3H]dAdo, and [3H]dCyt having specific activities of 10-22 Ci/mmol were obtained from ICN Laboratories, Irvine, CA. T h e hydrocarbons and their epoxide derivatives should be considered carcinogens and should be handled with care. P r e p a r a t i o n of Deoxyribonucleoside-Hydrocarbon Add u c t M a r k e r s . The dGuo, dCyt, and dAdo adduct markers of DB[a,j]A and the dGuo and dAdo adduct markers of 7MeDB[aJA were prepared according to procedures described previously (24). Under similar reaction conditions, solutions of calf thymus DNA radioactively labeled by a nick translation procedure with [3H]dGuo, [3H]dAdo,or [3H]dCyt in 0.01 M Tris-HC1 buffer, p H 7, were individually reacted with the bay region racemic 7MeDB[aJ]A anti- or racemic DB[aj]A anti- or syn-diol epoxides, or the K-region (&:)-DB[aj]A-5,6-oxidedissolved in T H F to afford the respective adduct markers. Isolation of Deoxyribonucleoside-Hydrocarbon Adducts from Mouse Keratinocytes. Primary cultures of keratinocytes
N a i r et al. from the dorsal skins of adult female SENCAR mice (NCI, Frederick, MD) were prepared according to established procedures (25) in low ea2+ modified minimum essential medium (MEM) with growth factor supplements and 1% FBS. Cultures were switched to high ea2+ (1.4 mM) modified MEM lacking supplements 24 h prior to treatment. Cells were seeded at a density of 8 X lo6 per 100-mm2petri dish. Cell cultures were treated with [3H]DB[aj]Aor [3H]7MeDB[aj]Aa t a final concentration of 0.5 Fg/mL of medium for 24 h. The medium was removed, and cells were briefly washed with PBS and lysed by addition of 0.6 M guanidinium isothiocyanate (GITC). The collected cell lysate was dialyzed against 50 mM Tris, 10 mM EDTA, and 10 mM NaCl buffer, pH 8, for 36 h with three changes of the buffer. Experiments using isolated DNA samples failed to demonstrate any loss of DNA-associated radioactivity during the 36-h dialysis period. Following dialysis, the lysate was treated with proteinase K (100 wg/mL, 15 units/mg) for 3 h at 55 "C, extracted sequentially with water-saturated phenol (26)and then phenol chloroform/isoamyl alcohol (24:1), and digested with RNase A (2000 units/O.l mL) for 1 h a t 37 "C. T h e samples were further extracted with chloroform/isoamyl alcohol (24:1), and the DNA was precipitated by addition of 30% NaCl (l/lo volume) and ice-cold ethanol (2 volumes), washed sequentially with ethanol (3X) and ether (3X), and dried (N2). The isolated DNA was hydrolyzed to deoxyribonucleosidesby sequential treatments with DNase I, snake venom phophodiesterase, and alkaline phosphatase as described previously (27). Following digestion with DNase I and prior to addition of snake venom phosphodiesterase, the DNA content was estimated according to the method of Burton (28), and binding was expressed as pmol of hydrocarbon covalently bound/mg of DNA. DNA hydrolysates were subjected to Sephadex LH-20 chromatography to separate the hydrocarbondeoxyribonucleoside adducts from unmodified deoxyribonucleosides and unhydrolyzed DNA as described (29). Samples loaded on the Sephadex LH-20 columns were successively washed with HzO (10 mL) and 20% CH30H in H 2 0 (6 mL), and finally the hydrocarbon-deoxyribonucleoside adducts were eluted in CH30H (6 mL). A portion of the [3H]7MeDB[aJ]A-deoxyribonucleoside adducts were further separated on a column of dihydroboronyl cellulose (0.9 x 3 cm) according to the procedure of Sawicki et al. (30). Hydrocarbon-adducted deoxyribonucleoside samples both before and after chromatography on dihydroboronyl cellulose were analyzed by HPLC. H P L C Analyses. All HPLC analyses were carried out on an Altex Ultrasphere octadecylsilane (ODS) column (4.6 mm X 25 cm) using a Shimadzu LC-6A equipped with Shimadzu SCL-6A system controller, SPD-6A UV detector, RF-535 fluorescence detector, and CR501 or C-R4A data processors. The column flow rate was 1 mL/min. DB[aj]A-DNA adducts were separated by elution with 46% CH30H in H,O for 50 min followed by sequential linear gradients of 46-58% CH30H in H 2 0 for 50 min and 58-100% CH30H in H,O for 65 min. Likewise, the 7MeDB[aj]A-DNA adducts were separated by elution with 49% CH30H in H,O for 50 min followed by sequential linear gradients of 4 9 4 0 % CH30H in HzO for 50 min and 60-100% CH30H in HzO for 30 min. Individual 0.5-min fractions were collected in scintillation vials concurrent with injection of 100-pL samples into the column. Following addition of scintillation fluid (Ready Value, Beckman Instruments Inc., Fullerton, CA), radioactivity of fractions was determined by use of a Beckman LS 1800 liquid scintillation counter.
+
Results Reactions of calf thymus DNA in vitro with the putative metabolites (&)-DB[aj]A-5,6-oxide,(*)-DB[aj]A syn-diol epoxide, and (f)-DB[aj]A anti-diol epoxide (Chart I) followed by isolation of modified deoxyribonucleosides (see Experimental Procedures) and subsequent HPLC analyses afforded the product profiles shown in Figure 1. Separate reactions utilizing DNA radioactively labeled with single tritiated deoxyribonucleotides ( [3H]dGuo, [3H]dAdo,or [3H]dCyt) were carried out with each of the reactive epoxide metabolites (Chart I) to identify the nature of the product peaks found in the HPLC chromatograms. Ra-
Characterization of DB[a,j]A-DNA A. 5,&0xide
Adducts
0 dGuo
I
W dAdo IJ
Unknown
1
105 110 115 120 125 130 1%
140
B(a]P-S,l&Dlol
f
Chem. Res. Toxicol., Vol. 4,No. 1, 1991 117 Table I. Distribution of DNA-Associated Radioactivity following Sephadex LH-20 Chromatographic Separation of DNA Hydrolysates" 90 dpm % dpm in 20% % dpm hydrocarbon in HzO MeOH in H,O in MeOH DMBA 28.1 1.2 70.7 7MeDB[aj]A 67.4 8.0 24.7 DB[aj]A 79.8 5.0 15.2
DNA hydrolysates obtained according to procedures described under Experimental Procedures were loaded on Sephadex LH-20 columns and washed successively with H 2 0 (10 mL), 20% CH30H in HzO (6 mL), and CH30H (6 mL). The hydrocarbon-bound deoxyribonucleosides eluted in CH30H. Values represent an average of two separate experiments giving nearly identical results. C. anti DE
$
2
Time (min)
Figure 1. HPLC elution profiles of deoxyribonucleoside adducts obtained from reactions of tritium-labeled calf thymus DNA with (A) (&)-DB[a i ]A-5,6-oxide, (B) (*) -DB[a j]A syn-diol epoxide, and (C) (*)-DB[a,j)A anti-diol epoxide. Column elution was monitored by fluorescence detection. The identity of various deoxyguanosine (dGuo) and deoxyadenosine (dAdo) adduct peaks was determined by radioactivity measurement of HPLC fractions obtained from separate reactions of respectively tritium-labeled DNA with the reactive epoxides. Stereochemical assignments shown in panel B are according to the relative retention times reported in ref 31. Assignments shown in panel C are according to those in ref 24. Chart I
Ani1 dlol epoxlde
Syn dlol epoxlde
DB(a,j)A-5,6-oxlde
R=H, DB(a,J)A
R=CH3,7MeDB(a,J)A
dioactivity measurement of the collected fractions during HPLC analyses identified the origin of hydrocarbon adducted deoxyribonucleosides formed in these reactions. Figure 2 represents the radioactive profiles obtained when DNA labeled with [3H]dCyt was reacted with the respective epoxide metabolites (Chart I). The stereochemical characterization of adducts formed from reaction of DNA with racemic anti-diol epoxides of DB[a,j]A (Figure 1, panel C) and 7MeDB[aj]A was reported recently (24). The stereochemical assignments for products of reaction of DNA with DB[a,j]A syn-diol epoxide (Figure 1, panel B) is based upon comparison of the relative retention times of individual adduct peaks with those reported by Chadha et al. (31). Both the 5,6-oxide as well as the syn-diol epoxide of DB[aj]A (Chart I) reacted with calf thymus DNA to afford four dGuo and four dAdo adducts, respectively. However, as reported previously (24), reaction of calf thymus DNA with the corresponding anti-diol epoxides afforded mainly three dGuo and four dAdo adducts, re-
Table 11. Comparison of Total DNA Binding in Mouse Keratinocytes and Tumor-Initiating Activities of Dibenz[a $]anthracene Derivatives with Benzo[a Ipyrene DNA binding, 7 ' 0 of pmol/mg of mice with papillomas/ hvdrocarbon DNA" uauillomas mouse* 100 7.6 BbIP 21.68 79 2.8 7MeDB[a,j]A 5.28 i 0.97 50 0.9 DB[a j]A 2.23 f 0.54
Binding levels represent radioactivity associated with adducts in the CH30H phase from Sephadex LH-20 chromatography of DNA hydrolysates from primary cultures of mouse keratinocytes treated with 0.5 wg of tritiated hydrocarbon/mL of medium for 24 h. Values represent an average of two (B[a]P) or at least three (7MeDB[aj]A, DB[aj]A) separate experiments i S D . *Average number of papillomas per mouse after 18 weeks of promotion with TPA following initiation at the dose of 400 nmol of hydrocarbon per mouse. Further details on tumor experiments can be found in ref 10.
spectively. The dCyt residues in DNA were also found to react with all three reactive epoxides (Figure 2). The order of reactivity was anti-diol epoxide > syn-diol epoxide > 5,6-oxide as determined from the total amount of radioactivity recovered in methanol phases following Sephadex chromatography of product mixtures obtained from reactions with identical starting DNA concentrations. Of particular interest was the product profile obtained from reaction of dCyt with DB[aj]A syn-diol epoxide (Figure 2, panel B). In addition to five relatively major [3H]dCyt adduct peaks, several other minor adduct peaks were detected in this chromatogram. The chemical configurations of DB[aj]A-5,6-oxide-DNA adduct peaks found in Figure 1, panel A, or the dCyt adduct peaks found in Figure 2 were not further characterized. The tentative identification of adducts formed in keratinocytes exposed to [3H]DB[aj]A is based on coelution with UV markers prepared from the same reactive metabolites. The distribution of DNA-associated radioactivity following hydrolysis of DNA isolated from the keratinocyte cultures exposed to 0.5 pg/mL of tritium-labeled hydrocarbons and subsequent separation of hydrocarbon-adducted deoxyribonucleosides on Sephadex LH-20 columns is shown in Table I. A large proportion of the radioactivity (-80% for DB[a,j]A; -67% for 7MeDB[aj]A) eluted from the Sephadex columns was found in the aqueous phase. Correspondingly lower levels of radioactivity (-15% for DB[a,j]A; -25% for 7MeDB[a,j]A) were found to be associated with the modified deoxyribonucleosides eluted in the CH30H phase. The radioactivity associated with the adducted deoxyribonucleosides contained in the CH30H phase collected from the LH-20 columns was used to calculate the binding values shown in Table 11. Consistent with the tumor-initiation data (lo), 7MeDB[a,j]A showed higher levels of total DNA binding (5.28 f 0.97 pmol/mg of DNA) when compared with DB-
118 Chem. Res. Toxicol., Vol. 4, No. 1, 1991
1
Nair et al.
I.
A. 5,6-0xide
250
0
-
-0
50
100
150
xx)
250
300
350
FRACTION NUMBER
Figure 3. HPLC elution profile of deoxyribonucleoside adducts obtained from rimary cultures of mouse keratinocytes treated with 0.5 pg of [ H]DB[aj]A (1.75 Ci/mmol)/mL of medium for 24 h. Column elution was monitored by radioactive scintillation
P
counting. U represents unknown.
3503 3003 25M
m 15M 1003
5M 0
o
x)
IW
1x1
m
2x1
3m
3 s
Fraction Number
Figure 2. HPLC elution profiles of dCyt adducts obtained from reactions of [3H]dCyt-labeledcalf thymus DNA with (A) (&)DB[aj]A-B,B-oxide,(B) (*)-DB[aj]A syn-diol epoxide, and (C) (f)-DB[aj]Aanti-diol epoxide. Column elution was monitored by radioactive scintillation counting.
[aj]A (2.23 f 0.54 pmol/mg of DNA). As expected, these values were significantly lower than the value observed in DNA from keratinocytes exposed to 0.5 rg/mL of the more potent tumor initiator, B[a]P (Table 11). The high proportions of DNA-associated radioactivity obtained in the aqueous phases from Sephadex LH-20 chromatography in these experiments (Table I) were in contrast to results obtained with more potent carcinogens, such as DMBA, where more than 70% of the DNA-associated radioactivity is consistently found as adducted deoxyribonucleosides (Table I). However, these data are similar to the results previously reported with the weakly carcinogenic 7-methylbenz[a]anthracene(27) as well as previous studies in our laboratory with 7-(hydroxymethyl)-12-methylbenz[a] anthracene (7-OHM-12-MBA) (16). To further explore the basis for these observations, we employed an alternate digestion procedure as described by Eastman et al. (32). In addition, duplicate experiments using higher concentrations of snake venom phosphodiesterase and alkaline phosphatase during the digestion of DNA samples were performed. In neither case did we obtain a higher proportion of adducts eluting in the methanol phases from the LH-20 chromatography step. During the course of the present investigation, we observed a similar distribution of radioactivity associated with DNA isolated from mouse epidermis following topical application of either DB[aj]A or 7MeDB[aj]A. An in-depth analysis of the aqueous phases from these in vivo samples revealed that the majority of the radioactivity (>go%) was associated with normal deoxyribonucleosides and deoxyribonucleotides (W. Baer-Dubowska, R. V. Nair, R. D. Gill, A. N. Nettikumara, C. Cortez, R. G. Harvey, and J. DiGiov-
Table 111. Distribution of Radioactivity in DB[a ,j]A-DNA Adduct Peaks Found in Figure 3 HPLC fractions / Deaks % radioactivitp fractions 1-31 65.5 unknown 1 (trans-tetraol) 0.7 (+)-trans-anti-DE-dGuo 0.7 (+)-trans-syn-DE-dGuo 0.7 unknown 2 0.7 (+)-transanti-DE-dAdo 1.9 (+)-cis-syn-DE-dAdo 0.5 unknown 3 0.5 5,6-oxide-dAdo 4.1 unknown 4 5.5 unknown 5 unknown 6
1.5 0.8
Percentage of total radioactivity eluted from HPCL column. Total radioactivity injected into the column was approximately 109000 dpm.
anni, submitted for publication). A preliminary analysis of the LH-20 water phases obtained in the present study from both DB[aj]A and 7MeDB[aj]A samples revealed that the majority of the radioactivity in the water phases was tritium incorporated in normal deoxyribonucleosides and deoxyribonucleotides (data not shown). Thus, the distribution of radioactivity in the methanol and water phases generated after LH-20 was not the result of incomplete digestion of DNA samples. HPLC analyses of the DB[aj]A-DNA adducts obtained following Sephadex LH-20 chromatography afforded the profile shown in Figure 3. Comparison of retention times for various deoxyribonucleoside adduct peaks with known adduct markers revealed formation of adducts derived from (+) enantiomers of both DB[aj]A anti- and syn-diol epoxides as well as from DB[aj]A-5,6-oxide. Approximately 65% of total radioactivity eluted from the HPLC column was found in fractions associated with the solvent front (Table 111). Of the remaining radioactivity representing DNA adducts, one prominent deoxyribonucleoside adduct (-4.1% of total radioactivity eluted) in cells exposed to DB[aj]A was tentatively identified as formed from reaction of dAdo residues in the DNA with DB[a,j]A-5,6-oxide. In addition, both dGuo (-0.7% of total radioactivity eluted) and dAdo (-1.9% of total radioactivity eluted) trans addition products from (+)-DB[aj]A anti-diol epoxide as well as products of trans and cis addition, respectively, of dGuo (-0.7% of total radioactivity eluted) and dAdo (-0.5% of total radioactivity eluted) residues to the (+)-syn-diol epoxide were detectably
Characterization of DB[a,j]A-DNA
Chem. Res. Toxicol., Vol. 4, No. 1, 1991 119
Adducts
.,"" mn
'1En
. 400 300
-
200
-
100
-
0 0
BUFFER 2
BUFFER1
. .. ...- , 5
10
l T 1 l l . . . - ~ -
15
20
- .I . . . 25
- I
8
.
- - I . *' ' I ,
30
35
40
FRACIWN NUMBER
b
sa
1w
l&
260
$0
Go
1m
Figure 5. Dihydroboronyl cellulose chromatographic elution profile of deoxyribonucleoside adducts obtained from primary cultures of mouse keratinocytes exposed to 0.5 pg of [3H]7MeDB[aj]A (1.47 Ci/mmol)/mL of medium for 24 h. Approximately 49% of total rdioactivity eluted from the column was found in fractions of 1 M morpholine buffer (buffer 1);the remainder was found to elute in the 1M morpholine-10% sorbitol buffer (buffer 2). However, taking into account the amount of
radioactivity in fractions associated with the solvent front in subsequent HPLC chromatograms, the actual ratio of adducts eluting in buffer 1 and buffer 2 was k1.6. Chart I1
YM'
'R
(+) trans anti diol epoxide-dN
(+) cis anti diol epoxide-dN
0
Ea
1w
150
2w
Eo
Fraction Number Figure 4. HPLC elution profile of deoxyribonucleoside adducts obtained from rimary cultures of mouse keratinocytes treated with 0.5 fig of [9H]7MeDB[aj]A (1.47 Ci/mmol)/mL of medium for 24 h. Panel A represents the profile prior to dihydroboronyl cellulose chromatography. Panel B represents adducts eluted in 1 M morpholine buffer, and panel C represents adducts eluted in 1 M morpholine-10% sorbitol buffer. Column elution was monitored by radioactive scintillation counting. U represents
unknown.
present. No adduct peaks corresponding in retention times to any of the dCyt adduct markers (Figure 2) were observed in these chromatograms. The chromatogram shown in Figure 3 also contained several other DNA adduct peaks, the identities of which remain to be determined. Among these peaks, unknown peak 1 showed retention times similar to that of the anti-diol epoxide hydrolysis product trans-tetraol (Chart 11). It should be stressed that the percentage values for specific DNA adducts reported here is based on the total radioactivity eluted from the column and reflects loss of a large proportion of eluted radioactivity in early fractions of the chromatogram (Figure 3, Table 111). In comparison, when cells were exposed to the more biologically potent 7MeDB[aj]A, at least 12 DNA adduct peaks were consistently found in the HPLC chromatograms (Figure 4, panel A). At the present time only marker adducts derived from the anti-diol epoxide of 7MeDB[aj]A are available in our laboratory (24). Coinjection of these marker adducts revealed one 7MeDB[a,j]A DNA adduct peak coeluting with a trans dAdo adduct from the (+)-anti-diol epoxide. In addition, one HPLC peak coe-
k
'R (+)cis syn diol epoxide-dN
(+)trans syn diol epoxide-dN
R,=OH, R 2 4 N trans 5,goxide-dN
or
Rl=dN, R2=OH cia 5, Ssxide-dN
'R
trans tetraol dN=dGuo or dAdo or dCyt R=H, CHI
luted with a marker of the trans-tetraol. In order to obtain additional information on the possible nature of the remaining DNA adduct peaks found in Figure 4 (panel A), a portion of this material was further analyzed by chromatography on dihydroboronyl cellulose. According to this procedure (30),products of syn-diol epoxide origin elute in the 1 M morpholine buffer (buffer 1,Figure 5) whereas those having the anti-diol epoxide origin elute with 10% sorbitol in 1 M morpholine (buffer 2, Figure 5). Fractions (1mL) collected during separation on the dihydroboronyl cellulose column were pooled, and the DNA adducts contained in the two buffer fractions were purified on Sephadex LH-20columns. Subsequent HPLC analysis of the purified DNA adducts showed (Figure 4, panels B and C)
120 Chem. Res. Toxicol., Vol. 4, No. 1, 1991 Table IV. Distribution of Radioactivity in 7MeDB[a ,j]A-DNA Adduct Peaks Found in Figure 4 HPLC fractions/peak % radioactivitp fractions 1-21 36.0 unknown 1 (buffer 2)b 5.5 unknown 2 (buffer 2) 8.2 unknown 3 (buffer 2) 3.1 unknown 4 (buffer 1) 1.1 unknowns 5 and 6 (buffers 2 and 1, 3.1 respectively) 7.3 unknown 7 (trans-tetraol) unknown 8 (buffer 1) 2.7 unknown 9 (buffer 1) 1.0 (+)-trans-anti-DE-dAdo 0.5 unknown 10 (buffer 2) 1.7 unknown 11 (buffer 2) 1.1
'Percentage of total radioactivity eluted from HPLC column. Total radioactivity injected into the column was approximately 126000 dpm. bBuffers 1 and 2 represent the buffer systems used for the dihydroboronyl cellulose chromatographic separation of the DNA adducts.
formation of a slightly higher proportion of DNA adducts eluted in 1 M morpholine containing 10% sorbitol. However, these adduct peaks found in Figure 4 (panel C) exhibited considerably different retentions on the ODS column than those for (*)-7MeDB[aj]A anti-diol epoxide-dGuo (or -dAdo) adduct markers except again for a peak which coeluted with the (+)-trans-anti-DE-dAdo ( 2 4 ) . Similar to the DB[a,j]A-DNA adduct samples (Figure 3), a high proportion of total radioactivity (-36%, Table IV) associated with the isolated 'IMeDB[aj]A-DNA adducts eluted in early fractions associated with the solvent front (Figure 4, panel A). In addition, these chromatograms (Figure 4, panel A) often exhibited, albeit inconsistently, several late eluting peaks of unknown identity. These peaks were not recoverable following chromatography on dihydroboronyl cellulose (Figure 4, panels B and C), and therefore, their nature remains unknown.
Discussion In the present study, we have investigated the potential covalent modification of three deoxyribonucleosides, dGuo, dAdo, and dCyt, in primary cultures of mouse keratinocytes exposed to DB[a,j]A. A preliminary examination of the DNA adducts formed in identical experiments utilizing 7MeDB[a,j]A also has been presented. In order to facilitate identification, we determined the elution characteristics of hydrocarbon-bound deoxyribonucleoside markers prepared from reactions of radiolabeled calf thymus DNA with three possible metabolites of DB[aj]A, the bay ragion (h)-syn- and anti-diol epoxides and the K-region (f)-5,6-oxide, on reversed-phase ODS columns (Figures 1 and 2). Under identical chromatographic conditions, the deoxyribonucleoside adducts from the syn- and anti-diol epoxides exhibited close retention times on the ODS column, while the retention of K-region oxide adducts was markedly different. As expected, the DNA adducts derived from the 5,6-oxide, being less polar by nature, were retained longer on the ODS column (Figures 1 and 2). All three deoxyribonucleosides in the DNA reacted with the three reactive epoxides in vitro (Figures 1and 2). Although we have not estimated the efficiency of incorporation of tritium-labeled dGuo, dAdo, or dCyt into the DNA in the nick translation procedures, under similar reaction conditions, the amount of radioactivity recovered in the hydrocarbon-adducted deoxyribonucleosides was in the following order: anti-diol epoxide > syn-diol epoxide > 5,g-oxide. However, the relative proportions of specific adducts derived from an epoxide in the in vitro reactions
Nair et al. were not identical with those formed in cells exposed to these hydrocarbons. In particular, although a higher proportion of dGuo than dAdo residues in the DNA were modified by anti-diol epoxides of DB[a,j]A in vitro (24), the DNA isolated from cells exposed to DB[aj]A contained higher levels of modified dAdo residues (Figure 3). In addition, although detectable modification of dCyt residues in the DNA was observed with all three reactive epoxides (Chart 1) by using the nick-translated DNA procedure in vitro (Figure 2), HPLC chromatograms of DNA adducts obtained from keratinocytes treated with DB[aj]A (Figure 3) showed no significant peaks in the region of elution of the dCyt adduct markers. Thus, the extent of modification of this base in cellular DNA was considerably lower overall relative to dAdo and dGuo. On the other hand, we cannot rule out the possibility that dCyt adducts may be formed at levels below which we could detect. Small amounts of dCyt adducts have been previously detected in DNA reactions with (-)-DB[aj]A anti-diol epoxide (31), (i)-B[a]P anti-diol epoxide (33,34), and lO-azabenzo[a]pyrene-4,5oxide (35),as well as in hamster embryo cell cultures exposed to B[a]P (36). However, the significance of modification of dCyt residues in cellular DNA for the process of tumor initiation remains to be determined. Among the DB[aj]A-deoxyribonucleoside adduct peaks tentatively identified (Figure 3), a major peak was determined to be dAdo-DB[ajJA-5,6-oxide. Formation of deoxyribonucleoside adducts derived from K-region oxides have also been noted previously in mouse and rat skin treated with B[a]P (37) or (f)-B[a]P-4,5-oxide (38). However, earlier studies have shown that the K-region oxides of PAH were weaker carcinogens than the parent hydrocarbons (39-41 ). Thus, the significance of formation of deoxyribonucleoside adducts derived from K-region oxides of PAH remains speculative at the present time. In addition to the dAdo-DB[aj]A-5,6-oxide adduct peak, the HPLC chromatogram in Figure 3 also showed peaks due to modified dGuo and dAdo residues from the DB[aj]A (+)-anti- as well as (+)-syn-diol epoxides. Although formed in smaller amounts, the presence of these DNA adducts is consistent with the hypothesis that bay region diol epoxides of DB[a,j]A may be involved in the carcinogenicity of this hydrocarbon ( 4 2 ) . As reported previously, the bay region anti-diol epoxide of DB[aj]A is a more potent tumor initiator than the parent hydrocarbon, supporting this statement ( 4 2 ) . In keratinocytes exposed to 7MeDB[aj]A, the product of trans addition of dAdo to (+)-anti-diol epoxide was tentatively identified (Figure 4) by chromatographic comparison with marker adducts. The remaining peaks found in the HPLC chromatogram (Figure 4) did not correspond with any of the marker adducts prepared from (&)7MeDB[aj]A anti-diol epoxide. The syn-diol epoxide or the 5,6-oxide of 7MeDB[aj]A is not available for preparation of correspondingmarker adducts at the pesent time. Thus, although we were able to consistently separate the peaks in Figure 4A on dihydroboronyl cellulose columns (Figure 5) and subsequent HPLC as shown in Figure 4B,C, the assignment of these peaks as deoxyribonucleoside adducts with either syn- or anti-diol epoxides requires further investigation and verification. Unknown 7 (Figure 4) exhibited a retention time similar to that of trans7MeDB[aj]A-1,2,3,4-tetrahydrotetraol formed as a result of hydrolysis of the anti-diol epoxide; however, the nature of the other peaks found in these chromatograms (Figure 4) remains unknown at the present time. Nevertheless, these preliminary observations suggest possible differences in the metabolic activation of 7-MeDB[aj]A compared to
Characterization of DB[a,j]A-DNA
Adducts
DB[a,j]A since several major DNA adduct peaks eluted very early in these chromatograms (Figure 4, peaks 1-3). In DNA binding studies using mouse embryo cell cultures exposed to the structurally related hydrocarbon B[c]Ph (19) little or no adduct formation occurred with the (-)-anti-diol epoxide. Our present data indicated that DB[aj]A also formed little or no DNA adducts derived from its corresponding (-)-anti-diol epoxide (Figure 3). A further similarity between the present data with DB[aj]A and previous results with B[c]Ph in mouse embyro cells (19) was a higher proportion of binding observed with the dAdo residues in cellular DNA (dAdo to dGuo ratio -5, Table 111). Further work is required to assess the relevance of preferential binding of reactive metabolites to dAdo residues exhibited by these hydrocarbons relative to their tumor-initiating potencies. In addition, comparison of the levels of specific deoxyribonucleoside adducts formed from DB[aj]A and 7MeDB[a,j]A must await further characterization of the presently unknown DNA adducts formed from these hydrocarbons (Figures 3 and 4). However, the levels of total DNA binding exhibited by these structural analogues (Table 11) correlated well with the differences in their tumor-initiating potencies.
Acknowledgment. This research was supported by NIH Grants CA 36979 (J.D.) and CA 36097 (R.G.H.) from the Department of Health and Human Services and by American Cancer Society Grants FRA-375 (J.D.) and BC-132 (R.G.H.). We thank Joyce Mayhugh for her excellent secretarial skills in preparing the manuscript.
References (1) Heidelberger, C. (1975) Chemical carcinogenesis. Annu. Rev. Biochem. 44, 79-121. (2) Sims, P. (1980) The metabolic activation of chemical carcinogens. Br. Med. Bull. 36, 11-18.
(3) Phillips, D. H., and Sims, P. (1979) Polycyclic aromatic hydrocarbon metabolites: Their reactions with nucleic acids. In Chemical carcinogens and DNA (Grover, P. L., Ed.) Vol. 2, pp 29-57, CRC Press, Boca Raton, FL. (4) Miller, E. C. (1978) Some current perspectives on chemical carcinogenesis in humans and experimental animals: Presidential address. Cancer Res. 38, 1479-1496. (5) Dipple, A., Moschel, R. C., and Bigger, C. A. H. (1984) Polynuclear aromatic hydrocarbons. In Chemical Carcinogenesis (Searle, C. E., Ed.) ACS Monograph 182, Vol. 1, pp 41-163, American Chemical Society, Washington, DC. (6) Baird, W. M., and Pruess-Schwartz, D. (1988) Polycyclic aromatic hydrocarbon-DNA adducts and their analysis: A powerful technique for characterization of pathways of metabolic activation of hydrocarbons to ultimate carcinogenic metabolites. In Polycyclic Aromatic Hydrocarbon Carcinogens: Structure-Activity Relationships (Yang, S. K., and Silverman, B. D., Eds.) Vol. 2, pp 141-179, CRC Press, Boca Raton, FL. (7) Dipple, A. (1976) Polynuclear aromatic carcinogens. In Chemical Carcinogens (Searle, C. E., Ed.) ACS Monograph 173, pp 245-314, American Chemical Society, Washington, DC. (8) Hecht, S. S., Amin, S., Rivenson, A., and Hoffmann, D. (1979) Tumor initiating activity of 5,ll-dimethylchrysene and the structural requirements favoring carcinogenicity of methylated polynuclear aromatic hydrocarbons. Cancer Lett. 8, 65-70. (9) Slaga, T. J., Iyer, R. P., Lyga, W., Secrist, A., 111, Daub, G. H., and Harvey, R. G. (1980) Comparison of the skin tumor initiating activities of dihydrodiols, diol epoxides, and methylated derivatives of various polycyclic aromatic hydrocarbons. In Polynuclear Aromatic Hydrocarbons: Chemistry and Biological Effects (Bjorseth, A., and Dennis, A. J., Eds.) pp 753-769, Battelle Press, Columbus, OH. (10) DiGiovanni, J., Diamond, L., Harvey, R. G., and Slaga, T. J. (1983) Enhancement of the skin tumor-initiating activity of polycyclic aromatic hydrocarbons by methyl-substitution a t nonbenzo bay-region positions. Carcinogenesis 4, 403-407. (11) DiGiovanni, J., Sawyer, T. W., and Fisher, E. P. (1986) Correlation between formation of a specific hydrocarbon-deoxyribo-
Chem. Res. Toxicol., Vol. 4,No.I, 1991 121 nucleoside adduct and tumor-initiating activity of 7,lZ-dimethylbenz[a]anthracene and i b 9- and 10-monofluorderivatives in mice. Cancer Res. 46, 4336-4341. (12) Dipple, A., Pigott, M., Moschel, R. C., and Costantino, N. (1983) Evidence that binding of 7,12-dimethylbenz[a]anthraceneto DNA in mouse embryo cell cultures results in extensive subsitution of both adenine and guanine residues. Cancer Res. 43,4132-4135. (13) Bigger, C. A. H., Sawicki, J. T., Blake, D. M., Raymond, L. G., and Dipple, A. (1983) Products of binding of 7,12-dimethylbenz[alanthracene to DNA in mouse skin. Cancer Res. 43,5647-5651. (14) Ashurst, S. W., and Cohen, G. M. (1981) In vivo formation of benzo[a]pyrene diol epoxide-deoxyadenosine adducts in the skin of mice susceptible to benzo[a]pyrene-induced carcinogenesis. Int. J. Cancer 27, 357-364. (15) Ashurst, S. W., Cohen, G. W., Nesnow, S., DiGiovanni, J., and Slaga, T. J. (1983) Formation of benzo[a]pyrene-DNA adducts and their relationships to tumor initiation in mouse epidermis. Cancer Res. 43, 1024-1029. (16) DiGiovanni, J., Nebzydoski, A. P., and Decina, P. C. (1983) Formation of 7-hydroxymethyl-12-methylbenz[a]anthraceneDNA adducts from 7,12-dimethylbenz[a]anthracenein mouse epidermis. Cancer Res. 43,4221-4226. (17) DiGiovanni, J., Fisher, E. P., Aalfs, K. K., and Prichett, W. P. (1985) Covalent binding of 7,12-dimethylbenz[a]anthraceneand lO-fluor0-7,12-dimethylbenz[a]anthraceneto mouse epidermal DNA and its relationship to tumor-initiating activity. Cancer Res. 45, 591-597. (18) Beland, F. A., and Poirier, M. C. (1989) DNA adducts and carcinogenesis. In The Pathobiology of Neoplasia (Sirica, A. E., Ed.) pp 57-80, Plenum Press, New York. (19) Pruess-Schwartz, D., Baird, W. M., Yagi, H., Jerina, D. M., Pigott, M. A,, and Dipple, A. (1987) Stereochemical specificity in the metabolic activation of benzo[c]phenanthrene to metabolites that covalently bind to DNA in rodent embryo cell cultures. Cancer Res. 47,4032-4037. (20) Harvey, R. G., Cortez, C., and Jacobs, S.A. (1982) Synthesis of polycyclic aromatic hydrocarbons via a novel annelation method. J . Org. Chem. 47, 2120-2125. (21) Harvey, R. G., Cortez, C., Sawyer, T. W., and DiGiovanni, J. (1988) Synthesis of the tumorigenic 3,4-dihydrodiol metabolites of dibenz[a,j]anthracene and 7,14-dimethyldibenz[a,j]anthracene. J . Med. Chem. 31, 1308-1312. (22) Harvey, R. G., Goh, S.H., and Cortez, C. (1975) K-region oxides and related oxidized metabolites of carcinogenic aromatic hydrocarbons. J . Am. Chem. SOC. 97, 3468-3479. (23) Harvey, R. G., Pataki, J., and Lee, H. (1986) Synthesis of the dihydrodiol and diol epoxide metabolites of chrysene and 5methylchrysene. J . Org. Chem. 51, 1407-1412. (24) Nair, R. V., Gill, R. D., Cortez, C., Harvey, R. G., and DiGiovanni, J. (1989) Characterization of DNA adducts derived from (i)-trans-3,4-dihydroxy-anti-l,2-epoxy-l,2,3,4-tetrahydrodibenz[ajlanthracene and (f)-7-methyl-trans-3,4-dihydroxy-anti-l,2Chem. Res. TOXepoxy-1,2,3,4-tetrahydrodibenz[aj]anthracene. icol. 2, 341-348. (25) Sawyer, T. W., Gill, R. D., Smith-Oliver, T., Butterworth, B. E., and DiGiovanni, J. (1988) Measurement of unscheduled DNA synthesis in primary cultures of adult mouse epidermal keratinocytes. Carcinogenesis (London) 9, 1197-1202. (26) Water-saturated phenol was prepared as follows: Melted phenol (500 g) containing distilled water (55 mL) was mixed with 8-hydroxyquinoline (0.5 g), m-cresol (70 mL), and excess distilled water. Water-saturated phenol was kept frozen until use. (27) Baird, W. M., and Brookes, P. (1973) Isolation of the hydrocarbon-deoxyribonucleoside products from the DNA of mouse embryo cells treated in culture with 7-methylbenz[a]anthracene3H. Cancer Res. 33, 2378-2385. (28) Burton, K. (1956) A study of the conditions and mechanism of the diphenylamine reaction for the colorimetric estimation of DNA. Biochem. J . 62, 315-323. (29) DiGiovanni, J., Singer, J. M., and Diamond, L. (1984) Comparison of the metabolic activation of 7,12-dimethylbenz[a]anthracene by a human hepatoma cell line (HepGZ) and low passage hamster embryo cells. Cancer Res. 44, 2878-2884. (30) Sawicki, J. T., Moschel, R. C., and Dipple, A. (1983) Involvement of both syn- and anti-dihydrodiol-epoxides in the binding of 7,12-dimethylbenz[a]anthraceneto DNA in mouse embryo cell cultures. Cancer Res. 43, 3212-3218. (31) Chadha, A., Sayer, J. M., Yeh, H. J. C., Yagi, H., Cheh, A. M., Pannell, L. K., and Jerina, D. M. (1989). Structures of covalent nucleoside adducts formed from adenine, guanine, and cytosine
122 Chem. Res. Toxicol., Vol. 4 , No. 1, 1991 bases of DNA and the optically active bay-region 3,4-diol 1,2-epoxides of dibenz[aj]anthracene. J. Am. Chem. SOC. 111, 5456-5463. (32) Eastman, A., Mossman, B. T., and Bresnick, E. (1981) Formation and removal of benzo[a]pyrene adducts of DNA in hamster tracheal epithelial cells. Cancer Res. 41, 2605-2610. (33) Meehan, T., Straub, K., and Calvin, M. (1977) Benzo[a]pyrene diol epoxide covalently binds to deoxyguanosine and deoxyadenosine in DNA. Nature 269, 725-727. (34) Straub, K. M., Meehan, T., Burlingame, A. L., and Calvin, M. (1977) Identification of the major adducts formed by reaction of benzo[a]pyrene diol epoxide with DNA in vitro. R o c . Natl. Acad. SC~. U.S.A. 74, 5285-5289. (35) Okuda, H., and Shudo, K. (1983) Structure of nucleic acid bases in calf thymus DNA modified by the potent mutagen, 10-azabenzo[a]pyrene-4,5-oxide.Chem. Pharm. Bull. 31, 2924-2927. (36) Ivanovic, V., Geacintov, N. E., Yamasaki, H., and Weinstein, I. B. (1978) DNA and RNA adducts formed in hamster embryo cell cultures exposed to benzo[a]pyrene. Biochemistry 17, 1597-1603. (37) Baer-Dubowska, W., Frayssinet, C. H., and Alexandrov, K. (1981) Formation of covalent deoxyribonucleic acid benzo[a]pyr-
Nair et al. ene-4,5-epoxide adduct in mouse and rat skin. Cancer Lett. 14, 125-129. (38) Rojas, M., and Alexandrov, K. (1986) In vivo formation and persistence of DNA and protein adducts in mouse and rat skin exposed to (*)benzo[a]pyrene-4,5-oxide. Carcinogenesis 7 , 235-240. (39) Miller, E. C., and Miller, J. A. (1967) Low carcinogenicity of the K-region epoxides of 7-methylbenz[a]anthraceneand benz[a]anthracene in the mouse and rat. R o c . SOC.Exp. Biol. Med. 124, 915-919. (40) Sims, P. (1967) The carcinogenic activities in mice of compounds related to 3-methylcholanthrene. Znt. J. Cancer 2, 505-508. (41) VanDuuren, B. L., Langseth, L., Goldschmidt, B. M., and Orris, L. (1967) Carcinogenicity of epoxides, lactones, and peroxy compounds. VI. Structure and Carcinogenic - adivitv. J. Natl. Cancer inst. 39, 1217-1228. (42) Sawyer, T. W., Baer-Dubowska, W., Chang, K., Crysup, S. B., Harvey, R. G., and DiGiovanni, J. (1988) Tumor-initiating activitv of the-bay-region dihydrodiols and diol epoxides of dilkz[a,jjanthracene and cholanthrene on mouse skin. Carcinogenesis 9, 2203-2207.
