The carcinogenicity of methoxyl derivatives of 4-aminoazobenzene: correlation between DNA adducts and genotoxicity.

To elucidate the cause of the difference in genotoxic activity between carcinogenic 3-methoxy-4-aminoazobenzene (3-MeO-AAB) and noncarcinogenic 2-methoxy-4-aminoazobenzene (2-MeO-AAB), we analyzed DNA adducts in the livers of rats exposed to either of these chemicals and studied the resulting biologic potential with the aid of in vitro modified M13 phage DNA. 32P-Postalbeling analysis revealed that the carcinogen 3-MeO-AAB produced 20-fold higher amounts of adducts than did 2-MeO-AAB. Five adducts were formed in the 3-MeO-AAB case whereas only one adduct was apparent in 2-MeO-AAB-treated rat. Studies of in vitro DNA replication using N-hydroxy (N-OH)-aminoazo dye-modified M13 phage DNA as a template demonstrated inhibition by 3-MeO-AAB adducts to be substantially greater than in the 2-MeO-AAB-adducts. The specificity of mutagenesis induced in M13mp9 phage DNA by these chemicals also was analyzed after transfection into SOS-induced Escherichia coli JM103, mutation frequencies being higher with N-OH-3-MeO-AAB- than N-OH-2-MeO-AAB-modified DNA. The mutation spectra differed in each case. Our data suggest that the difference in hepatocarcinogenic activity between the two chemicals depends not only on qualitative and quantitative variation in adduct formation but also on conformation changes in modified DNA.


Introduction
Over several years, we have focused our attention on elucidation of the cause of the difference in genotoxic activity between 3methoxy-4-aminoazobenzene (3-MeO-AAB) and 2-methoxy-4-aminoazobenzene (2-MeO-AAB) (1)(2)(3)(4). The carcinogenic potencies of these aminoazo dyes are known to be influenced by the position of methoxy substituents on the aminoazobenzene molecule. 3-MeO-AAB is a hepatocarcinogen in the rat and a mutagen in Escherichia coli and Salmonella typhimurium, whereas 2-MeO-AAB, differing only in the position of the methoxy substituted on the same benzene ring, is a noncarcinogen and nonmutagen (5)(6)(7). 32 This article concerns: a) P-postlabeling analysis of DNA adducts in the liver after intraperitoneal (ip) administration of 3-MeO-AAB or 2-MeO-AAB, b) effects of DNA adducts on in vitro DNA synthesis, and c) mutations induced by aminoazo dye adducts in M13 viral DNA.
Representative results are shown in   Maximal DNA binding of 3-MeO-AAB and 2-MeO-AAB was observed after approximately 24 and 48 hr, respectively, and thereafter decreased gradually (Table  1). After 7 days, 40% of the maximum adduct levels were still present in both cases. The maxima were 3.8 for 3-MeO-AAB and 0.16 for 2-MeO-AAB per 107 nucleotides. The carcinogen 3-MeO-AAB generated more than 20-fold higher amounts of adducts in the liver than did 2-MeO-AAB at all time points examined.

Effects of 3-MeO-AAB-and 2-MeO-AAB-Adducts on in Vitro DNA Synthesis
To determine the effects of DNA adducts caused by 3-MeO-AAB and by 2-MeO-AAB on the replication of DNA, we analyzed the reaction products of E. coli DNA polymerase I (pol I) action on aminoazo dye-modified M 13mp 10 DNA templates by DNA sequencing gel electrophoresis.
Single-stranded M13mplO DNA was modified with N-hydroxy (N-OH) derivatives of 3-MeO-AAB or 2-MeO-AAB in the presence of seryl-AMP, whereby adducts have been proposed to be formed through (9,10). The extents of aminoazo dye adducts were calculated from spectrophotometric analyses (2,11). Figure 2 shows representative sequencing gel-electrophoresis bands of primer elongation products. From the evidence of stronger arrested bands and shorter DNA aAdduct numbers correspond to those of spots in Figure 1. Radioactivity of each adduct and total nucleotide spot was determined by Cerenkov counting. bMean ± SE of data for two separate tissues. DNA adduct levels in each tissue were calculated from duplicate samples. products observed with the N-OH-3-MeO-AAB-modified template, a conclusion can be drawn in this case of more effective inhibition of DNA chain elongation. In clear contrast, the arrested bands with N-OH-2-MeO-AAB-modified template were much fainter and the DNA products generally were longer, suggesting that pol I was able to read through 2-MeO-AAB adducts. Elongation was arrested extensively at one base prior to every 3-MeO-AAB guanine adduct and at -GGGG-sequences but not always at 2-MeO-AAB-guanine adducts, but it was blocked at adenines in -GAG-sequences. The differences between 3-MeO-AAB and 2-MeO-AAB guanine adducts might be because of their chemical structures or surrounding altered structures of the DNA.

Mutations Induced by 3-MeO-AAB and 2-MeO-AAB Adducts in M13
Viral DNA The ability of 3-MeO-AAB and 2-MeO-AAB adducts to induce mutations was analyzed using M13mp9 viral DNA modified by these chemicals as described above (2). Modified DNA containing 20 to 40 adducts/molecule was transfected into E. coli host cell JM103 and mutants were screened for expression of the marker enzyme, ,-galactosidase.
Mutation frequencies were increased up to 7-fold (21 x 10 ) for N-OH-3-MeO-AAB-modified DNA and 4-fold (8 x   for N-OH-2-MeO-AAB-modified DNA in SOS-induced host cells, as compared to the uninduced cells. The mutagenesis induced by aminoazo dye adducts therefore seems to be largely dependent on SOS functions. DNA sequence analysis was performed by Sanger's methods (12) within the lacZ' region base pairs 6200 to 6400, corresponding to the N-terminal sequence of the ,B-galactosidase (Figure 3). The mutational hotspots were G at position 6300 in both modified DNAs. The most frequent mutation being G-*A transition, but G-4T and G-*C transversions also occurred with 3-MeO-AAB adducts.
in agreement with our previous finding that these aminoazo dyes specifically form adducts with guanine bases (11).

Discussion
The aim of this study was to elucidate differences in genotoxic activity between 3-MeO-AAB and 2-MeO-AAB. 32P-Postlabeling analysis of DNA adducts from the liver of rats exposed to these chemicals revealed that the carcinogen 3-MeO-AAB produced 20-fold higher amounts of adduct than the noncarcinogen 2-MeO-AAB. Moreover, five spots were detected on TLC sheets from 3-MeO-AAB-treated rats, whereas only one adduct was found in 2-MeO-AAB-treated rats. Essentially similar results were obtained with DNA of E. coli uvrA strains treated with N-OH-3-MeO-AAB or N-OH-2-MeO-AAB (3).
Although the adducts formed by 3-MeO-AAB and 2-MeO-AAB have not yet been fully characterized, C8-substituted deoxyguanosine (spot no. 1) may well be a major product. Minor spots (no. 2-5) in 3-MeO-AAB-modified DNA may be deoxyguanosin N -yl-substituted and deoxyadenosin N -yl-substituted 4-aminoazobenzene derivatives (13,14). In this context, it is of particular importance that the kind of DNA damage-the nature of the adduct itself or the conformational changes in DNA-which is responsible for the biologic activity of 3-MeO-AAB be clarified.
We therefore compared the biologic potential of M13 phage DNA in vitro modified by these chemicals. As a result, DNA replication by pol I was found to be blocked at or one base prior to guanine bases in the templates, 3-MeO-AAB adducts having a significantly greater inhibitory effect on DNA synthesis than 2-MeO-AAB adducts. The arrested pattern also differed in each case.
While 3-MeO-AAB adducts induced mutations more frequently than did 2-MeO-AAB adducts, they were both mainly derived from guanine base. However, 3-MeO-AAB adducts induced a higher percentage of frameshift mutations as well as base substitutions, suggesting a possibly greater alteration in DNA conformation. In contrast, 2-MeO-AAB adducts induced mostly base substitutions.
Our data thus suggest that the difference in hepatocarcinogenic activity between 3-MeO-AAB and 2-MeO-AAB depends not only on qualitative and quantitative variations in adduct formation, but also on conformation changes in DNA modified by these chemicals.