We transfected a UV-irradiated shuttle vector carrying a tRNA gene into normal and xerodermal pigmentosum complementation group A (excision repair defective) cells, and measured mutations and UV-induced lesions at each base pair in the gene. Four mutation hotspots were found in XPA; two were photoreactivatable, implicating cyclobutane dimers and two were nonphotoreactivatable, implicating a nondimer lesion. The dimer mutation sites were not found in normal cells, implying that in normal cells dimers were not mutagenic because they were excised. A less-readily excised nondimmer lesion does generate mutations in normal cells. The lesion frequency was not correlated with mutation frequency at these hotspots; preferential excision and unidentified lesions were excluded as explanations. Mutation hotspots thus occur at particular spots in the sequence independent of lesion frequency. Such a spot would occur if lesions blocked DNA replication, reducing the number of mutants recovered, except at the mutation hotspots. To examine the origin of human skin cancer, we examined UV photoproducts in the cloned Ki-ras gene, which is reported to be activated in UV-induced mouse skin tumors. A DNA damage hotspot is predicted in codon 61 of Ki-ras and N-ras, but not Ha-ras nor at codon 12. The lack of hotspot in Ha-12, Ha-61, and Ki-12 has been found; Ki-61 is being examined. Concurrently, we are transfecting UV-irradiated cloned c-Ha-ras and the c-Ki-ras minigene into NIH 3T3 cells to locate sites of activation. We have also isolated genomic DNA from human basal and squamous cell carcinomas; these will be screened for activated ras oncogenes with oligonucleotide probes and by 3T3 transfection.
