Mutagenic specificity of syn-benzo[g]chrysene 11,12-dihydrodiol 13,14-epoxide in the dihydrofolate reductase gene of Chinese hamster ovary cells.

Polycyclic aromatic hydrocarbons (PAHs) with sterically hindered fjord region diol epoxides are interesting with respect to their potency as carcinogens, interactions with DNA and mutagenic specificities. Unlike the bay region PAH derivative, trans-7,8-dihydroxy-anti-9,10-epoxy-7,8,9, 10-tetrahydroxybenzo[a]pyrene (BPDE), reactive metabolites of two fjord region PAH, trans-3,4-dihydroxy-anti-1, 2-epoxy-1,2,3,4-tetrahydrobenzo[c]-phenanthrene [(+/-)-anti-BcPHDE] and trans-11,12-dihydroxy-syn-BgCDE], react with DNA to yield high levels of adenine adducts. We previously found that forward mutations induced by (+/-)-anti-BcPHDE in the dihydrofolate reductase (dhfr) gene of Chinese hamster ovary (CHO) cells preferentially targeted mRNA splice acceptor sites. (+/-)-anti-BcPHDE and (+/-)-syn-BgCDE are structurally similar; they differ only by the presence of an additional benzene ring. Thus we used (+/-)-syn-BgCDE to learn if the mutational target bias reflects aspects of the mutagen structure or its capacity to efficiently modify deoxyadenosine (dA) in vivo. dhfr(-) mutants were induced after treatment of hemizygous UA21 cells with a 0.75 microM dose of (+/-)-syn-BgCDE. Cell survival after carcinogen exposure was 40%. The induced mutation frequency was 9 x 10(-6), nearly 10-fold higher than the spontaneous one, but approximately 19-fold lower than formerly observed using (+/-)-anti-BcPHDE. In the 26 confirmed null dhfr(-) mutants 27 mutations were identified by DNA sequencing. The types of (+/-)-syn-BgCDE-induced mutations were very similar to those formerly induced by (+/-)-anti-BcPHDE. Consistent with the binding specificity, both chemicals induced transversion base substitution at purines (R-->T). The most prevalent type of mutation was A-->T, which represented 59% of the induced changes, compared with 42% for (+/-)-anti-BcPHDE. (+/-)-syn-BgCDE mutated mostly novel targets in the dhfr gene, sites not found mutated with any of the several other mutagens we have used in former studies. Whereas the 25 kg dhfr gene contains six coding exons, the majority (16/27) of (+/-)-syn-BgCDE-induced mutations were located in a single one (exon 4). A random distribution of mutations affecting splice acceptor sites (22%) was induced by (+/-)-syn-BgCDE. Hence, preferential mutation of these sites by (+/-)-anti-BcPHDE may reflect aspects of chromatin structure in vivo which make these sequences better targets for modification. Alternatively, the sequence context of these sites may dictate an adduct conformation which is poorer for damage recognition and/or efficient repair.