Mechanism of metabolic activation of the potent carcinogen 7, 12

Mar 1, 1992 - James X. Chen, Annie Pao, Yi Zheng, Xiancang Ye, Alexander S. Kisleyou, Rebecca Morris, Thomas J. Slaga, Ronald G. Harvey, and ...
0 downloads 0 Views 1MB Size
Chem. Res. Toxicol. 1992,5, 220-226

220

Mechanism of Metabolic Activation of the Potent Carcinogen 7,12-Dimethylbenz[ a ]anthracene N. V. S. RamaKrishna,t P. D. Devanesan,t E. G . Rogan,t E. L. Cavalieri,*it H. Jeong,* R. Jankowiak,* and G. J. Small* Eppley Institute for Research in Cancer and Allied Diseases, University of Nebraska Medical Center, 600 South 42nd Street, Omaha, Nebraska 68198-6805,and Department of Chemistry and Ames Laboratory-USDOE, Iowa State University, Ames, Iowa 50011 Received August 27, 1991

The DNA adducts of 7,12-dimethylbenz[a]anthracene(DMBA) previously identified in vitro and in vivo are stable adducts formed by reaction of the bay-region diol epoxides of DMBA with dG and dA. In this paper we report identification of several new DMBA-DNA adducts formed by one-electron oxidation, including two adducts lost from DNA by depurination, DMBA bound at the 12-methyl to the N-7 of adenine (Ade) or guanine (Gua) [7-methylbenz[a]anthracene (MBA-12-CHz-N7Ade or 7-MBA-12-CH2-N7Gua, respectively]. The in vitro systems used to study DNA adduct formation were DMBA activated by horseradish peroxidase or 3-methylcholanthrene-induced rat liver microsomes. The biologically-formed depurination adducts were identified by high-pressure liquid chromatography and by fluorescence line narrowing spectroscopy. Stable DMBA-DNA adducts were analyzed by the 32P-postlabeling method. Quantitation of DMBA-DNA adducts formed by microsomes showed about 99% as depurination adducts: 7-MBA-12-CH2-N7Ade (82%) and 7-MBA-12-CHz-N7Gua (17%). Stable adducts (1.4% of total) included one adduct spot that may contain adduct(s) formed from the diol epoxide (0.2%) and unidentified adduds (1.2%). Activation of DMBA by horseradish peroxidase afforded 56% of stable unidentified adducts and 44% of depurination adducts, with 36% of 7-MBA12-CHz-N7Adeand 8% of 7-MBA-12-CH2-N7Gua. Adducts containing the bond to the DNA base at the 7-CH3 group of DMBA were not detected. These results show that with activation by cytochrome P-450 the DMBA-DNA adducts are predominantly formed by one-electron oxidation and lost from DNA by depurination, and the adducts formed by the diol epoxide pathway are very minor. The 12-CH3group is shown to be critical in metabolic activation of DMBA, a finding consistent with the results of carcinogenicity studies in rodents.

I ntroductlon Elucidation of the mechanisms of activation of polycyclic aromatic hydrocarbons (PAH)' is essential for understanding the process of tumor initiation. Studies thus far point to two major pathways of activation: one-electron oxidation to form intermediate radical cations and monooxygenation to produce bay-region diol epoxides (1-6). Some evidence supports the hypothesis that the bayregion diol epoxide of 7,12-dimethylbenz[a]anthracene (DMBA) is responsible for its carcinogenic activity. DMBA 3,4-dihydrodiol is the only dihydrodiol potent in tumor initiation in mouse skin and induction of lung adenomas in newborn mice (7,8).This proximate carcinogen is more potent than the parent DMBA. Analysis of DMBA-DNA adducts provides a second line of evidence. Three major adducts have previously been identified as arising from addition of the bay-region diol epoxide of DMBA to dA and dG (9,lO).These adducts and several minor ones have been separated by high-pressure liquid chromatography (HPLC) and 32P-postlabeling techniques (9-12);however, analysis by the 32P-postlabeling technique revealed several unidentified DMBADNA adducts formed in mouse skin (12). We propose that one-electron oxidation is the predominant mechanism of activation for the most potent PAH, including benzo[a]pyrene (BP), DMBA, and 3-methyl-

* To whom correspondence should be addressed. t University of Nebraska Medical Center.

* Iowa State University.

cholanthrene (MC). Mammalian peroxidases, including prostaglandin H synthase (13-16),and cytochrome P-450 (17-22) catalyze one-electron oxidation, and this mechanism is also involved in the binding of PAH to DNA (23-26). The DNA adducts formed when BP is activated by MC-induced rat liver microsomes have recently been shown to arise primarily by one-electron oxidation; almost all of these adducts are lost from DNA by depurination (27). A similar profile of BP-DNA adducts is observed in mouse skin treated with BP in vivo.2 Evidence for activation of DMBA by one-electron oxidation comes from observations concerning benz[a]anthracene (BA) and some of its derivatives, as well as the chemical properties of the DMBA radical cation. The parent compound BA is a borderline carcinogen (a), but substitution of a methyl group at C-7 leads to substantially increased carcinogenicity (29-31). Among the methylsubstituted BAS, the 7-CH3and 12-CH3are the most active, followed by the 6-CH3 and 8CH3derivatives, but the others are inactive (24-31). Substitution of two methyl groups at the meso-anthracenic positions of BA produces DMBA, one of the most potent carcinogenic PAH (28). 1,2,3,4-TetrahydroDMBA is a strong carcinogen, despite Abbreviations: BA, benz[a]anthracene;BP, benzo[a]pyrene;DMBA, 7,12-dimethylbenz[a]anthracene;DMSO, dimethyl sulfoxide; FLN, fluorescence line narrowed; FLNS, fluorescence line narrowing spectroscopy; HPLC, high-pressureliquid chromatography; HRP, horseradish peroxidase; MC, 3-methylcholanthrene; PAH, polycyclic aromatic hydrocarbonb);MBA, methylbenz[a]anthracene;PDA, photodiode array detector; ZDC, zero-drift controller. Unpublished results.

0893-228x19212705-0220~03.00/0 0 1992 American Chemical Society

Identification a n d Quantitation of DMBA-DNA Adducts

Chem. Res. Toxicol., Vol. 5, No. 2, 1992 221 Table I. Separation of Adducts by HPLC retention time, min CH,OH/ CH&N/HpO HZO gradientb adduct gradient4 7-MBA-12-CH2-N7Gua 73.05 55.10 12-MBA-7-CH2-N7Gua 74.56 56.15 7-MBA-12-CH2-N7Ade 68.00 57.33 12-MBA-7-CH2-N3Ade 70.12 58.55 ~

I

CH3

7-MBA-12-CH2-N7Gua

a

Cn2

7-MBA-12-CH,-N7Ade

(:c) N "

NH*

12-MBA-7-CH,-N7Gua 12-MBA-7-CH2-N3Ade Figure 1. Structures of depurination adducts analyzed.

being fully saturated in the angular ring and unable to be metabolicallyactivated to the bay-region diol epoxide (32, 33). Formation of the bay-region diol epoxide is also blocked by the fluoro substituent in 1-F- and 4-F-DMBA, but they are carcinogenic (32-35). One-electron oxidation of DMBA by Mn(OAc), (3) or iodine-pyridine (36) produces nucleophilic substitution at the two methyl groups, suggesting that they play an important role if one-electron oxidation is the mechanism of bioactivation. To assess the importance of one-electron oxidation in the metabolic activation of DMBA, stable and depurination DMBA-DNA adducts were analyzed and quantified. The adducts formed in the MC-induced rat liver microsome-catalyzed binding of DMBA to DNA were compared to the adducts formed when DMBA was activated by horseradish peroxidase (HRP) or DMBA 3,4-dihydrodiol, the proximate metabolite in the diol epoxide pathway, wm activated by microsomes. Deoxyribonucleoside adducts previously synthesized by electrochemical oxidation of DMBA (37) were used as standards in the identification of the depurination adducts by HPLC and fluorescence line narrowed spectrometry (FLNS). The stable adducts were quantitated by 32p-postlabelingand the depurination adducts by radioactivity during HPLC. We report here the discovery that two DMBA-DNA adducts formed by one-electron oxidation and loat from DNA by depurination are the major adducts (99%) when the binding of DMBA to DNA is catalyzed by MC-induced rat liver microsomes.

Materials and Methods Chemicals. [3H]DMBA (sp act. 1907 Ci/mol) was purchased from Chemsyn Science Laboratories (Lenexa, KS) and used a t a specific activity of 500 Ci/mol. DMBA trans-3,4-dihydrodiol was purchased from the National Cancer Institute Chemical Carcinogen Repository (Bethesda, MD). Authentic 7-methylbenz[a]anthracene(MBA)-12-CH2-N7Gua, 7-MBA-12-CH2N7Ade, 12-MBA-7-CH2-N7Gua, and 12-MBA-7-CH2-N3Ade (Figure 1)were synthesized by anodic oxidation of DMBA in the presence of deoxyguanosine or deoxyadenosine (37). DMBA and its dihydrodiol are hazardous chemicals; they were handled according to NIH guidelines (38). Binding of P A H to DNA. As previously described (39,40), [3H]DMBA (80 pM) and DMBA 3,4-dihydrodiol (15 pM) were bound to DNA in reactions catalyzed by MC-induced rat liver microsomes, and [3H]DMBA was also bound to DNA by HRP.

The ODs-AQ column was eluted with 30% CH30H in H20 for 5 min, followed by a linear gradient to 100% CH30H in 75 min at a flow rate of 1.0 mL/min. bThe ODs-AQ column waa eluted with 20% CH3CN in H20 for 5 min, followed by a linear gradient to 100% CH3CN in 80 min at a flow rate of 0.8 mL/min. The reactions with [3H]DMBAwere 15 mL in volume, whereas the reaction with DMBA 3,4-dihydrodiol was 1 mL; they were all incubated for 30 min at 37 "C. At the end of the reaction, a 1-mL aliquot of the 15-mL mixture was used to determine the level of PAH binding to DNA and the amount of stable adducts by the P1-nuclease version of the 32ppostlabeling method, as previously described (40).The DNA from the remaining 14-mL mixture was precipitated with 2 volumes of absolute ethanol, and the supernatant was used to identify and quantify the depurination adducts by HPLC and FLNS. Analysis of Depurination Adducts by HPLC. The supernatant from each binding reaction, which contained depurination adducts and metabolites, was evaporated to dryness under vacuum. The residue was dissolved in a minimal volume of dimethyl sulfoxide (DMSO), followed by addition of an equal volume of CH30H. After sonication to enhance solubilization, the undissolved residue was removed by centrifugation. A 75-pL aliquot was analyzed on a YMC 5-pm, ODs-AQ reverse-phase analytical column (6.0 X 250 mm) (YMC, Overland Park, KS) on a Waters 600E multisolvent delivery system with a Waters 700 satellite WISP autoinjector (Millipore Corp., Wood Dale, IL). The column was eluted for 5 min with 30% CH30H in H20, followed by a 75-min linear gradient to 100% CH30H, at a flow rate of 1.0 mL/min. The peaks were detected by UV absorbance (254 nm) using a Waters 990 photodiode array (PDA) detector and by radioactivity (Radiomatic Flc-one/Beta radiation monitor, A250 series, Radiomatic, Tampa, FL). The synthesized authentic adducts were used as reference markers. The radioactive adducts coeluting with the authentic adducts were collected in multiple runs (see Table I for retention times). The combined material was evaporated under argon, redissolved in DMSO/ CH30H (l:l), and further chromatographed on a CH3CN/H20 gradient. The column was eluted for 5 min with 20% CH&N in H20, followed by an 80-min linear gradient to 100% CH3CN a t a flow rate of 0.8 mL/min (see Table I for retention times). For each type of incubation, the depurination adducts were analyzed in four experiments. In a first experiment, the depurination adducts were identified by HPLC in both CH30H/H20 and CH3CN/H20 gradients in the presence of added authentic adducts. In a second experiment, the adducts were chromatographed in both solvent systems without added authentic adducts, and the material eluting at the respective retention times on the CH3CN/H20gradient was collected for identification by FLNS. In the third and fourth experiments, quantitative data were collected for calculation of the amounts of both the depurination and stable adducts. The quantity of each adduct varied in the two experiments by 5-15%, with the larger variations in the minor adducts. Analysis of Depurination Adducts by FLNS. The excitation source of the FLNS apparatus, shown in Figure 2, is a Lamda-Physik EMG 102 MSC excimer (XeC1) pumped tunable dye laser (Lamda-Physik FL-2002, Gottingen, Germany) system possessing a 10-ns pulse width and variable repetition rate up to 100 Hz. Average power densities used are in the 1-100 mW/cm2 range, sufficiently low to avoid nonphotochemical hole burning (41). Fluorescence is dispersed by 1-m McPherson 2061 monochromator (F7.0) (Acton, MA) equipped with a 136-mm X 116mm grating with 2400 grooves/". The linear dispersion is 0.416 nm/mm. Dispersed fluorescence is detected with a Princeton Instruments IRY-l024/G/R/B intensified blue enhanced gatable

RamaKrishna et al.

222 Chem. Res. Toxicol., Vol. 5, No. 2, 1992 DMBA with HRP

* o

Y

6

IZ DC

-i% 0,

COM

V

Figure 2. Block diagram of FLNS instrumentation: excimer laser

DMBA with Microsomes

(Exc.L), dye laser (Dye.L), photodiode (PD), zero-drift control (ZDC), cryostat (C), sample (S), photodiode array (PDA), monochromator (M), high-voltage pulse generator (FG-loo), optical multichannel analyzer (ST-120), and computer (COM). PDA interfaced with a Princeton Instruments SR-120 multichannel analyzer (Princeton, NJ). The monochromator-PDA provides a 9-nm spectral segment a t an optimum resolution of 0.08 nm. Gated detection is accomplished with a Lamda-Physik EMG-97 zero-drift controller (ZDC) that triggers a Princeton Instruments FG-100 high-voltage gate pulse generator which allows for adjustable delay and width of the detection window. An electronic comparator of the ZDC corrects for long-term temporal drift between the laser pulse (monitored by a photodiode with output fed to the ZDC) and the electronic synchronization pulse from the ZDC’s triggering circuit. The system provides a minimum gate delay of 5 ns. For the adducts studied, gate delays in the range of 25-75 ns with a 60-11s window are utilized. The same monochromator is used with a photomultiplier tube (Amperex-XP-2232) for the non-line-narrowed fluorescence measurements. The photomultiplier tube output and the output of a reference photodiode, which monitors a small portion of the laser pulse, are fed to a SRS Model SR280 boxcar averager (Stanford Research, Menlo Park, CA) with two Model SR250 processor modules (channels A and B). The ratio of the two signals compensates for pulse-to-pulse intensity jitter of the laser. An SR245 interface module is used for transfer of the normalized data to an IBM-PC computer. The glass forming solvent is a glycerol/water/ethanol mixture (0.5:0.40.1 v/v). Samples (ca. 5 pmol) are dissolved in ca. 10 pL of DMSO and mixed with ca. 30 pL of the glass forming solvent, providing a concentration about 5 X M. Approximately 20 pL of sample is contained in a 1cm long quartz tube (3-mm 0.d. X 2-mm i.d.) and rapidly cooled (ca. 2 min) to liquid He temperature (4.2 K) in a double-nested glass Dewar equipped with optical quartz windows. Calculation of Adduct Levels. The amount of stable adducts was calculated by the 32P-postlabeling method as previously described (40). For quantitation of the depurination adducts, each of the peaks eluting at the same time as an authentic adduct in HPLC using the CH30H/H20gradient was collected and counted. These peaks were then reinjected individually in a CH3CN/H20 gradient, and the percentage of the injected radioactivity eluting at the same time as the authentic adduct was measured and calculated using the radiation flow monitor. The total amount of each of the adducts was calculated from the specific activity of the PAH and normalized to the amount of DNA used in the reaction.

Results and Discussion Identification of DNA Adducts. To identify DMBA-DNA adducts formed by cytochrome P-450 in MC-induced rat liver microsomes, we compared the adducts to those formed by HRP-catalyzed one-electron oxidation of DMBA and those formed by microsomal activation of DMBA 3,ddihydrodiol. The stable adducts were quantitated and partially identified by the 32P-

DMBA 3,4-Dihydrodiol with Microsomes

6

*

Figure 3. (A) Autoradiogram of 32p-postlabeled DNA containing adducts formed after activation of DMBA by HRP. The film was exposed a t room temperature for 1.5 h. (B) Autoradiogram of 32P-postlabeledDNA containing adducts formed after activation of DMBA by MC-induced rat liver micromsomes. The film was exposed a t room temperature for 15 h. (C) Autoradiogram of 32P-postlabeled DNA containing adducts formed after activation of DMBA 3,4-dihydrodiol by MC-induced rat liver microsomes. The film was exposed for 0.5 h.

postlabeling technique. The depurination adducts were quantitated and identified by comparison with authentic adducts on HPLC (see Materials and Methods and Table I) and FLNS. In this study the stable DNA adducts were identified by 32P-postlabelingas arising via one-electron oxidation or the diol epoxide pathway. As can be seen from the autoradiograms in Figure 3, the pattern of adducts observed with microsomal activation shares a number of similarities with the pattern of adducts obtained with HRP activation. In contrast, only one adduct spot (1)obtained with microsomal activation of DMBA is also seen in the adducts obtained by microsomal activation of DMBA 3,4-dihydrdol to the bay-region diol epoxide. Under these

Chem. Res. Toxicol., Vol. 5, No. 2, 1992 223

Identification and Quantitation of DMBA-DNA Adducts

conditions, no adducts were observed in DNA from incubations without metabolic activation or without DMBA. These results indicate that the formation of the bay-region diol epoxide is a minor mechanism of microsome-catalyzed binding of DMBA to DNA and suggest that one-electron oxidation is the predominant mechanism of stable adduct formation. These results contrast sharply with previous studies of stable DMBA-DNA adducts conducted by using HPLC or 32P-postlabeling( s 1 2 ) . In those studies, the DMBADNA adducts formed in mouse skin or fetal mouse cell cultures were identified as arising primarily from reaction of the bay-region diol epoxides of DMBA with dA, and with dG to a minor extent. Although some of the unidentified adducts observed by others (12) may correspond to adducts apparently formed by one-electron oxidation catalyzed by HRP and microsomes (Figure 3A,B), the relative intensities are very different in our study compared to the previous ones. This apparent contradiction will remain unresolved until we have analyzed the DNA adducts formed in mouse skin treated with DMBA. Two depurination adducts were identified, 7-MBA-12CH2-N7Ade and 7-MBA-12-CH2-N7Gua (Figure 1). Preliminary identification of these adducts was made by mlution with authentic adducts on HPLC in two different solvent systems, namely, CH30H/H20and CH3CN/H20 gradients (Table I). Proof of structure was obtained by FLNS analysis. Adducts containing DMBA bound at the 7-CH3group to DNA, namely, 12-MBA-7-CH2-N7Guaand 12-MBA-7-CH2-N3Ade (Figure l),were not detected with either HRP or microsomal activation. In the CH30H/H20 gradient, 12-MBA-7-CH2-N3Ade coeluted with a major metabolite of DMBA, but with additional HPLC in the CH3CN/H20gradient, which separates the adduct from the metabolite, no adduct was detected. Discussion of the principles of FLNS and its application to the characterization of macromolecular DNA- and globin-carcinogen adducts as well as nucleoside-carcinogen adducts can be found in two recent reviews (42,43). More recently, we reported that FLNS, employed in the vibronic excitation mode, can differentiate between the following adducts of DMBA formed by the one-electron oxidation pathway: 7-MBA-12-CH2-C8Gua, 7-MBA-12-CH2N7Gua, 12-MBA-7-CH2-N7Gua,7-MBA-12-CH2-N7Ade, and 7-MBA-12-CH2-N3Ade (37). Distinction between these adducts is based on differences in vibronic intensity distributions and/or excited-state (S,) vibrational frequencies. It is not possible, however, to differentiate between 7-MBA-12-CH2-C8Gua and its dG analogue. This can be accomplished with fast atom bombardment tandem mass spectrometry (37). It also has been reported that DMBA and its nucleoside adducts have a tendency to self-aggregate in solution and that very broad oligomeric fluorescence is observed from the glass at 4.2 K, even for concentrations as low as lo+' M. Such fluorescence, which cannot be line-narrowed, is centered at ca. 405 nm, in comparison with the monomer fluorescence which has a maximum at ca. 400 nm. Interference from the broad oligomeric fluorescence to the fluorescence line-narrowed (FLN) spectra of the monomers can be reduced by selective excitation and gated detection (37). This interference is apparent in the spectra presented below. FLN spectra are shown in Figure 4 for the authentic 7-MBA-12-CH2-N7Gua (spectrum A) and the isolated adducts from the HRP- (spectrum B) and microsomecatalyzed (spedrum C) reactions. The spectra are obtained with vibronic excitation at A,, = 390 nm and under identical gating conditions (delay = 50 ns,width = 60 ns). The

I ex = 390 nm

I

396

398

4 '2

400

WAVELENGTH (nm)

Figure 4. Segment of vibronically excited FLN spectra of the synthesized 7-MBA-12-CH2-N7Gua standard (A), 7-MBA-12CH2-N7Guafrom HRP reaction (B), and microsomal reaction (C) obtained with A,, = 390 nm.

1

I

,

396

398

400

402

396

398

400

402

WAVELENGTH (nm)

Figure 5. Comparison of vibronically excited FLN spectra of the synthesized 7-MBA-12-CH2-N7Ade standard (a), 7-MBA12-CH2:N7Ade from HRP reaction (b), and microsomal reaction (c) obtaJned with & = 392.8 nm (A) and 390 nm (B), respectively.

prominent zero-phonon bands are labeled with their excited-state vibrational frequency in cm-' and appear superimposed on the high-energy tail of the fluorescence band due to aggregates. Spectra obtained with &, = 392.8 nm (not shown), which reveal excited-state vibrations at 220,273,311, and 355 cm-l(37), also show a close similarity between the FLN spectra of the standard adduct and the HRP and microsomal adducts. Therefore, we can confidently assign the structure of the two biologically-formed adducts as 7-MBA-12-CH2-N7Gua. The FLN spectra for the 7-MBA-12-CH2-N7Ade standard and adducts isolated from the HRP and microsomal reaction are compared in Figure 5 for A,, = 390.0 and 392.8 nm. For both excitation wavelengths the spectra are very similar. This suggests that the structure of both the HRP and microsomal adducts is 7-MBA-12-CH2N7Ade. It should be noted, though, that the A,, = 390.0 nm spectra of Figure 5 are similar to the A,, = 390.0 nm spectra of Figure 4. There are, however, significant differences. For example, the vibronic intensity distributions

224 Chem. Res. Toxicol., Vol. 5, No. 2, 1992

incubation system HRP DMBA MC-induced microsomes DMBA DMBA 3,4-dihydrodiol

RamaKrishna et al.

Table 11. Quantitation of Biologically-Formed DMBA-DNA Adducts" stable adducts, (mol of adduct/mol of DNA-P) X depurination adducts, (mol of adduct/mol 106 of DNA-P) X lo6 total adducts, (mol/mol of stable depurination DNA-P) X adducts, % 7-MBA-127-MBA-12adducts, % of total CHz-N7Ade CHz-N7Gua of total 106 unidentified 24.2 8.8

13.6b 0.13d 6.4e

56 1.4

8.7 (36)'

2.0 (8)

44

7.2 (82)

1.5 (17)

99

ratio of depurination/stable

0.77 70

"Values are the average of determinations on two preparations. The amount of each adduct varied between 5% and 15% in the two DreDarations. Because DMBA 3.4-dihvdrodiol was not radiolabeled, the depurination adducts could not be analyzed. Total adducts repiesent 0.1% of the DMBA with HRP and 0.03% with microsomes. 13 adduct spots. Number in parentheses is percentage of to& adducts. 9 adduct spots. e 4 adduct spots.

in the 400-550-cm-' region are not the same, and the relatively intense 420-cm-' mode of 7-MBA-12-CH2N7Gua has a frequency of 426 cm-' for 7-MBA-12CH2N7Ade. In addition, the 360-cm-' mode of 7-MBA12-CH2-N7Ade(left frame of Figure 5) becomes 355 cm-' for 7-MBA-12-CH2-N7Gua (37). The distinction between the 7-MBA-12-CH2-N7Gua and 7-MBA-12-CH2-N7Ade adducts was previously discussed in greater detail (37). Thus, we can confidently assign the structure of both biologically-formed adducts in Figure 5 as 7-MBA-12CH2-N7Ade. In summary, we have identified two new adducts formed by one-electron oxidation and lost from DNA by depurination. In both of these adducts, 7-MBA-12-CH2-N7Ade and 7-MBA-12-CH2-N7Gua, DMBA is bound to the nucleic acid base at the 12-CH3group. No adduct containing DMBA bound at the 7-CH3 group was detected. The pattern of stable adducts seen in the 32P-postlabeling analysis of DMBA-DNA adducts formed by microsomes resembels that formed by HRP activation, but appears to contain only one adduct (spot 1in Figure 3) that may arise in part from the DMBA 3,Cdihydrodiol. Quantitation of DNA Adducts. Quantitation of stable adducts was accomplished by the 32P-postlabeling method, whereas quantitation of depurination adducts was conducted by HPLC with a radiation flow monitor. The total amount of stable DMBA adducts detected was greater with activation by HRP compared to microsomes (Table II). HRP was previously found to bind a relatively large amount of DMBA to DNA (23). A sharp contrast is noted between the ratio of detectable depurination and stable adducts when binding was catalyzed by HRP or microsomes. With HRP, more than half of the analyzed adducts were stable, whereas with microsomes depurination adducts were the overwhelmingly predominant ones detected. The relative amounts of individual stable adducts detected by 32P-postlabeling ranged from 0.3% to 88% (Table III). With activation by HRP, the most abundant adduct spot (spot 3 in Figure 3A) contained 88% of the total adducts detected. This intense spot probably contains several adducts, and it corresponds to adducts 3 and 3a in the DNA containing DMBA activated by microsomes (Figure 3B). This area contained 47% of the adducts observed with microsomal activation. Other adducts were relatively abundant with both HRP and microsomal activation; for example, adducts 1, 6, and 7. When DMBA 3,4-dihydrodiol was activated by microsomes, four adducts were observed (Figure 3C), with adduct 1 being the most abundant (60%). This adduct is the only one corresponding to an adduct of DMBA activated by microsomes (Figure 3B) and accounts for only

Table 111. Amounts of Stable DNA Adductso amount of adduct, (mol/mol of DNA-P) X IO6 (% of stable adducts) DMBA 3,4-diDMBA-mihydrodiol spot DMBA-HRP crosomes microsomes 1 0.38 (2.8) 0.015 (12) 3.83 (60) 2 0.04 (0.3) 3 12.00 (88) 0.009 (7) 0.050 (40) 3a 4 0.11 (0.8) 0.003 (2) 5 0.04 (0.3) 0.019 (15) 6 0.18 (1.3) 7 0.30 (2.2) 0.017 (14) 8 0.11 (0.8) 9 0.11 (0.8) 10 0.05 (0.4) 11 0.003 (2) 12 0.14 (1.0) 0.010 (8) 13 0.08 (0.6) 0.35 (5) 14 0.04 (0.3) 0.20 (3) 15 2.04 (32) 16 total 13.6 0.126 6.42 aValues were calculated from the 32Pin adduct spots, as previously described (38). Experiments were conducted in duplicate, and the amount of each adduct varied 5 1 5 % between the two preparations.

12% of the total DMBA adducts detected after microsomal activation (Table 111). A minor adduct is also detected in this area with activation by HRP, suggesting that two different adducts chromatograph similarly, one formed by one-electron oxidation and one by the diol epoxide pathway. Thus, with microsomes adduct 1may arise by either pathway of activation,or both. Assuming that all of adduct 1 is attributable to formation of the diol epoxide, this adduct represents only 0.2% (12% of 1.4% of total adducts) of the total adducts. Thus, from the data presented in Tables I1 and 111,99.8% of the DMBA-DNA adducts appear to be formed by one-electron oxidation, with 99% as the two depurination adducts and 1% as unidentified stable adducts. The level of stable adducts detected by the 32P-postlabeling method is 100-fold higher with activation by HRP compared to activation by microsomes (Table 111), although we have routinely found that the level of binding of DMBA to DNA catalyzed by HRP (23) is about 10-fold higher than that catalyzed by microsomes.2 Almost all of this difference resides in adduct 3, which is 88% of the stable adducts detected. The expected 10-fold difference is reflected in almost all of the other adducts formed by both enzymes (Table 111). The very high intensity of ad-

Identification and Quantitation of DMBA-DNA Adducts

duct 3 with HRP activation is not currently understood. These results suggest that some adducts are detected more efficiently by 32P-postlabelingthan others. This technique does not appear to measure the entire amount of some or all of the DNA adducts formed (40). Quantitation of depurination adducts by HPLC also suffers from losses of adducts during the sequential chromatography processes. In view of these technological limitations, the quantitation has some imprecision, but is the best currently achievable. Completion of these analyses would include identification of the major stable adducts detected by the 32Ppostlabeling technique. One of the expected stable adducts is DMBA bound at the 12-CH3group to the C-8 of dG (37). Analysis of stable and depurination adducts will be extended to mouse skin treated with DMBA.

Conclusions One-electron oxidation is the predominant mechanism in the covalent binding of DMBA to DNA catalyzed by MC-induced rat liver microsomes. The two major DMBA-DNA adducts formed by this mechanism are depurination adducts in which specifically the 12-CH3group is bound to the N-7 of Ade (82%) or Gua (17%). These two adducts constitute almost 99% of the total adducts. No adducts were detected in which the 7-CH3 group was bound to a nucleic acid base. The stable adducts constitute 1.4% of the total. Only 0.2% could correspond to adducts formed via the diol epoxide pathway of activation. Formation of 7-MBA-12-CH2-N7Gua and 7-MBA-12CH2-N7Ade by anodic oxidation of DMBA (37) and by cytochrome P-450- or HRP-catalyzed one-electron oxidation provides a link between electrochemical and enzymatic experiments, suggesting that a radical cation intermediate is involved in both processes. Activation of DMBA by one-electron oxidation and its specific reaction at the 12CH3 group with DNA nucleophiles is in agreement with the results of several relevant carcinogenicity experiments. Substitution of the two methyl groups of DMBA with ethyl groups renders the compound, 7,12-diethylbenz[a]anthracene, inactive (44). The inactivity of ethyl-substituted derivatives is a general phenomenon already observed for 7-C2H6BAvs 7-CH3BA and 6-C2H5BPvs 6CH3BP.2 This inactivity is consistent with the lack of nucleophilic substitution of the radical cations of C2H5 PAH (for example, 7-C2H5BA)at the benzylic methylene (36). Furthermore, 7-C2H5-12-CH3BAdisplays a carcinogenic activity similar to that of DMBA, whereas 7-CH312-C2HaAis a much weaker carcinogen (44).These data suggest that the 12-CH3group is involved in the critical carcinogenic activation of DMBA. Therefore, the tumorigenicity experiments reported above and the specific reactivity of the 12-CH3 group with DNA nucleophiles suggest that metabolic activation of DMBA occurs via one-electron oxidation. Acknowledgment. This research was supported primarily by USPHS Grant PO1 CA49210 awarded to both research groups by the National Cancer Institute. Additional support was from Grants R01 CA25176 and R01 CA44686. Core support at the Eppley Institute was from the National Cancer Institute (P30 CA36727). Ames Laboratory is operated for the U.S. Department of Energy at Iowa State University under Contract W-7405-Eng-82, and research at the Ames Laboratory was also supported by the Office of Health and Environmental Research. Registry No. DMBA, 57-97-6; 7-MBA-12-CH2-N7Gua, 138606-34-5; 7-MBA-12-CHz-N7Ade, 138606-33-4; 3,4-dihydro3,4-dihydroxy-7,12-dimethylbenz[a]anthracene, 68162-13-0; cy-

Chem. Res. Toxicol., Vol. 5, No. 2, 1992 225 tochrome P-450, 9035-51-2; peroxidase, 9003-99-0.

References (1) Cavalieri, E. L., and Rogan, E. G. (1984) One-electron and two-

electron oxidation in aromatic hydrocarbon carcinogenesis. In Free Radicals in Biology (Pryor, W. A., Ed.) Vol. VI pp 323-369, Academic Press, New York. (2) Cavelieri, E., and Rogan, E. (1985) One-electron oxidation in aromatic hydrocarbon carcinogenesis. In Polycyclic Hydrocarbons and Carcinogenesis (Harvey, R. G., Ed.) ACS Sympceium Series 283 pp 289-305, American Chemical Society, Washington, DC. (3) Cavalieri, E., and Rogan, E. (1985) Role of radical cations in aromatic hydrocarbon carcinogenesis. Enuiron. Health Perspect. 64,6944. (4) Cavalieri, E. L., and Rogan, E. G. (1990) Radical cations in aromatic hydrocarbon carcinogenesis. Free Radical Res. Commun. 11, 77-87. (5) Sims, P., and Grover, P. L. (1981) Involvement of dihydrodiols and diol epoxides in the metabolic activation of polycyclic hydrocarbons other than benzo[a]pyrene. In Polycyclic Hydrocarbons and Cancer (Gelboin, H. V., and Ts'o, P. 0. P., Eds.) pp 117-181, Academic Press, New York. (6) Conney, A. H. (1982) Induction of microsomal enzymes by foreign chemicals and carcinogenesis by polycyclic aromatic hydrocarbons: G. H. A. Clowes Memorial Lecture. Cancer Res. 42, 4875-4917. (7) Slaga, T. J., Gleason, G. L., DiGiovanni, J., Sukumaran, K. B., and Harvey, R. G. (1979) Potent tumor-initiating activity of the 3,4-dihydrodiol of 7,12-dimethylbenz(a)anthracenein mouse skin. Cancer Res. 39, 1934-1936. (8) Wislocki, P. G., Gadek, K. M., Juliana, M. M., MacDonald, J. S., Chou, M. W., Yang, S. K., and Lu, A. Y. H. (1981) Tumorigenicity and mutagenicity of 7,12-dimethylbenz(a)anthraceneand selected metabolites including the 3,4-dihydrodiols. In Polynuclear Aromatic Hydrocarbons: Chemical Analysis and Biological Fate (Cooke, M., and Dennis, A. J., Eds.) pp 675-685, Battelle Press, Columbus, OH. (9) Cheng, S. C., Prakash, A. S., Pigott, M. A,, Hilton, B. B., Lee, H., Harvey, R. G., and Dipple, A. (1988) A metabolite of the carcinogen 7,12-dimethylbenz[a]anthracenethat reacts predominantly with adenine residues in DNA. Carcinogenesis 9, 1721-1723. (10) Cheng, S. C., Prakash, A. S., Pigott, M. A., Hilton, B. D., Roman, J. M., Lee, H., Harvey, R. G., and Dipple, A. (1988) Characterization of 7,12-dimethylbenz[a]anthracene-adeninenucleoside adducts. Chem. Res. Toxicol. l, 216-221. (11) Schmeiser, H., Dipple, A., Schurdak, M. E., Randerath, E., and Randerath, K. (1988) Comparison of 32P-postlabeling and high pressure liquid chromatographic analyses for 7,12-dimethylbenz[alanthracene-DNA adducts. Carcinogenesis 9,633-638. (12) Yang, P.-F., and Randerath, K. (1990) High-resolution TLC mapping and characterization of 32P-postlabeledmonophosphate 7,12-dimethylbenz[a]anthracene-DNAadducts. Carcinogenesis 11,159-164. (13) Boyd, J. A.,and Eling, T. E. (1984) Evidence for a one-electron mechanism of 2-aminofluorene oxidation by prostaglandin H synthase and horseradish peroxidase. J. B i d . Chem. 22, 13885-13896. (14) Josephy, P. D., Eling, T. E., and Mason, R. P. (1983) Cooxidation of benzidine by prostaglandin synthase and comparison with the action of horseradish peroxidase. J. Biol. Chem. 258, 5561-5569. (15) Degen, G. H., Eling, T. E., and McLachlan, J. A. (1982) Oxidative metabolism of diethylstilbestrol by prostaglandin synthase. Cancer Res. 42, 919-923. (16) Cavalieri, E., Devanesan, P., and Rogan, E. (1988) Radical cations in the horseradish peroxidase and prostaglandin H synthase mediated metabolism and binding of benzo[a]pyrene to DNA. Biochem. Pharmacol. 37,2183-2188. (17) Cavalieri, E., Rogan, E., Cremonesi, P., and Devanesan, P. (1988) Radical cations as precursors in the metabolic formation of quinones from benzo[a]pyrene and 6-fluorobenzo[a]ppene: Fluoro substitution as a probe for one-electron oxidation in aromatic substrates. Biochem. Pharmacol. 37, 2173-2182. (18) Hanzlik, R. P., and Tullman, R. H. (1982) Suicidal inactivation of cytochrome P-450 by cyclopropylamines. Evidence for cationradical intermediates. J. Am. Chem. SOC.104, 2048-2050. (19) MacDonald, T. L., Zirvi, K., Burka, L. T., Peyman, P., and Guengerich, F. P. (1982) Mechanism of cytochrome P-450 inhib-

226 Chem. Res. Toxicol., Vol. 5, No. 2, 1992 ition by cyclopropylamines. J . Am. Chem. SOC. 104, 2050-2052. (20) Augusto, O., Beilan, H. S., and Ortiz de Montellano, P. R. (1982) The catalytic mechanism of cytochrome P-450. Spintrapping evidence for one-electron substrate oxidation. J. Biol. Chem. 257,11288-11295, (21) Burka, L. T., Guengerich, F. P., Willard, R. J., and MacDonald, T. L. (1985) Mechanism of cytochrome P-450 catalysis. Mechanism of N-dealkylation and amine oxide deoxygenation. J . Am. Chem. SOC. 107, 2549-2551. (22) Stearns, R. A., and Ortiz de Montellano, P. R. (1985) Cytochrome P-450-catalyzed oxidation of quadricyclane: Evidence for a radical cation intermediate. J. Am. Chem. SOC. 107,4081-4082. (23) Cavalieri, E. L., Rogan, E. G., Roth, R. W., Saugier, R. K., and Hakam, A. (1983) The relationship between ionization potential and horseradish peroxidase/ hydrogen peroxide-catalyzed binding of aromatic hydrocarbons to DNA. Chem.-Biol. Interact. 47, 87-109. (24) Devanesan, P., Rogan, E., and Cavalieri, E. (1987) The relationship between ionization potential and prostaglandin H synthase-catalyzed binding of aromatic hydrocarbons to DNA. Chem.-Biol. Interact. 61, 89-95. (25) Rogan, E., Cavalieri, E., Tibbels, S., Cremonesi, P., Warner, C., Nagel, D., Tomer, K., Cerny, R., and Gross, M. (1988) Synthesis and identification of benzo[a]pyrene-guanine nucleoside adducts formed by electrochemical oxidation and horseradish peroxidase-catalyzed reaction of benzo[a]pyrene with DNA. J. Am. 110,4023-4029. Chem. SOC. (26) Cavalieri, E. L., Rogan, E. G., Devanesan, P. P., Cremonesi, P., Cerny, R. L., Gross, M. L., and Bodell, W. J. (1990) Binding of benzo[a]pyrene to DNA by cytochrome P-450-catalyzed oneelectron oxidation in rat liver microsomes and nuclei. Biochemistry 29, 4820-4827. (27) Devanesan, P. D., RamaKrishna, N. V. S., Todorovic, R., Rogan, E. G., Cavalieri, E. L., Jeong, H., Jankowiak, R., and Small, G. J. (1992) Identification and quantitation of benzo[a]pyreneDNA adducts formed by rat liver microsomes in vitro. Chem. Res. Toxicol. (in press). (28) Dipple, A. (1976) Polycyclic aromatic carcinogens. In Chemical Carcinogens (Searle, C. E., Ed.) pp 245-314, American Chemical Society, Washington, DC. (29) Dunning, W. F., and Curtis, M. P. (1960) Relative carcinogenic activity of monomethyl derivatives of benzo[a]anthracene in Fisher line 344 rats. J. Natl. Cancer Znst. 25, 387-391. (30) Stevenson, J. L., and Van Ha", E. (1965) Carcinogenicity of benzo[a]anthracene and benzo[c]phenanthrene derivatives. Am. Inst. Hyg. Assoc. J. 26, 475-478. (31) Wislocki, P. G., Fiorentin, K. M., Fu, P. P., Yang, S. K., and Lu, A. Y. H. (1982) Tumor-initiating ability of the twelve monomethylbenz[a]anthracenes. Carcinogenesis 3, 215-217. (32) DiGiovanni, J., Diamond, L., Singer, J. M., Daniel, F. B., Witiak, D. T., and Slaga, T. J. (1982) Tumor-initiating activity of

RamaKrishna et al. 4-fluoro-7,12-dimethylbenz(a)anth1aceneand 1,2,3,4-tetrahydro7,12-dimethylbenz(a)anthracenein female SENCAR mice. Carcinogenesis 3, 651-655. (33) Cavalieri, E. L., Rogan, E. G., Higginbotham, S., Cremonesi, P., and Salmasi, S. (1990) Tumorigenicity of 7,12-dimethylbenz[a]anthracene, some of its fluorinated derivatives, and 1,2,3,4-tetrahydro-7,12-dimethylbenz[a]anthracenein mouse skin and rat mammary gland. Polycyclic Aromat. Compd. 1, 59-70. (34) Harvey, R. G., and Dunne, F. B. (1978) Multiple regions of metabolic activation of carcinogenic hydrocarbons. Nature (London) 273, 566-568. (35) DiGiovanni, J., Decina, P. C., and Diamond, L. (1983) Tumor initiating activity of 9- and lO-fluoro-7,12-dimethylbenz[a]anthracene (DMBA) and the effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin on tumor initiation by monofluoro derivatives of DMBA in SENCAR mice. Carcinogenesis 4,1045-1049. (36) Cavalieri, E., and Roth, R. (1976) Reaction of methylbenzanthracenes and pyridine by one-electron oxidation: A model for metabolic activation and binding of carcinogenic aromatic hydrocarbons. J. Org. Chem. 41, 2679-2684. (37) RamaKrishna, N. V. S., Cavalieri, E. L., Rogan, E. G., Dolnikowski, G., Cerny, R. L., Gross, M. L., Jeong, H., Jankowiak, R., and Small, G. J. (1991) Synthesis and structure determination of the adducts of the potent carcinogen 7,12-dimethylbenz[a]anthracene and deoxyribonucleosides formed by electrochemical oxidation: Models for metabolic activation by one-electron oxi(in press). dation. J. Am. Chem. SOC. (38) NIH Guidelines for the Laboratory Use of Chemical Carcinogens (1981) NIH Publication No. 81-2385, US. Government Printing Office, Washington, DC. (39) Wong, A. K. L., Cavalieri, E., and Rogan E. (1986) Dependence of benzo[a]pyrene metabolic profile on the concentration of cumene hydroperoxide with uninduced and induced rat liver microsomes. Biochem. Pharmacal. 35, 1583-1588. (40) Bodell, W. J., Devanesan, P. D., Rogan, E. G., and Cavalieri, E. L. (1989) 32P-Postlabelinganalysis of benzo[a]pyrene-DNA adducts formed in vitro and in vivo. Chem. Res. Toxicol. 2, 312-315. (41) Jankowiak, R., Cooper, R. S., Zamzow, D., Small, G. J., Doskocil, G., and Jeffrey, A. M. (1988) Fluorescence line narrowingnonphotochemical hole burning spectrometry: Femtomole detection and high selectivity for intact DNA-PAH adducts. Chem. Res. Toxicol. 1, 60-68. (42) Jankowiak, R., and Small, G. J. (1989) Fluorescence line-narrowing spectroscopy in the study of chemical carcinogenesis. Anal. Chem. 60, 1023A-1032A. (43) Jankowiak, R., and Small, G. J. (1991) Fluorescence line narrowing: A high-resolution window on DNA and protein damage from chemical carcinogens. Chem. Res. Toxicol. 4, 256-269. (44) Pataki, J., and Balick, R. (1972) Relative carcinogenicity of some diethylbenz[a]anthracenes. J . Med. Chem. 15, 905-909.