Sequence-specific detection of aristolochic acid-DNA adducts in the human p53 gene by terminal transferase-dependent PCR.

Abstract

The carcinogenic plant extract aristolochic acid (AA) is thought to be the major causative agent in the development of urothelial carcinomas found in patients with Chinese herb nephropathy (CHN). These carcinomas are associated with overexpression of p53, suggesting that the p53 gene is mutated in CHN-associated urothelial malignancy. To investigate the relation between AA-DNA adduct formation and possible p53 mutations, we mapped the distribution of DNA adducts formed by the two main components of AA, aristolochic acid I (AAI) and aristolochic acid II (AAII) at single nucleotide resolution in exons 5-8 of the human p53 gene in genomic DNA. To this end, an adduct-specific polymerase arrest assay combined with a terminal transferase-dependent PCR (TD-PCR) was used to amplify DNA fragments. AAI and AAII were reacted with human mammary carcinoma (MCF-7) DNA in vitro and the major DNA adducts formed were identified by the (32)P-postlabeling method. These adducted DNAs were used as templates for TD-PCR. Sites at which DNA polymerase progress along the template was blocked were assumed to be at the nucleotide 3' to the adduct. Polymerase arrest spectra thus obtained showed a preference for reaction with purine bases in the human p53 gene for both activated compounds. For both AAs, adduct distribution was not random; the strongest signals were seen at codons 156, 158-159 and 166-167 for exon 5, at codons 196, 198-199, 202, 209, 214-215 and 220 for exon 6, at codons 234-235, 236-237 and 248-249 for exon 7 and at codons 283-284 and 290-291 for exon 8. Overall guanines at CpG sites in the p53 gene that correspond to mutational hotspots observed in many human cancers seem not to be preferential targets for AAI or II. We compared the AA-DNA binding spectrum in the p53 gene with the p53 mutational spectrum of urothelial carcinomas found in the human mutation database. No particular pattern of polymerase arrest was found that predicts AA-specific mutational hotspots in urothelial tumors of the current p53 database. Thus, AA is not a likely cause of non-CHN-related urothelial tumors.