Differential introduction of DNA damage and repair in mammalian genes transcribed by RNA polymerases I and II.

We have developed a general quantitative method for comparing the levels of drug-induced DNA crosslinking in specific mammalian genes. We observed a dramatic difference between the efficiency of the removal of both psoralen monoadducts and interstrand crosslinks from the rRNA genes and the efficiency of their removal from the dihydrofolate reductase (DHFR) gene in cultured human and hamster cells. While 90% of the interstand crosslinks were removed from the human DHFR gene in 48 h, less than 25% repair occurred in the rRNA genes. Similarly, in Chinese hamster ovary cells, 85% repair of interstrand crosslinks occurred within 8 h in the DHFR gene versus only 20% repair in the rRNA genes. The preferential repair of the DHFR gene relative to that of the rRNA genes was also observed for psoralen monoadducts in cells from both mammalian species. In human-mouse hybrid cells, the active mouse rRNA genes were five times more susceptible to psoralen modification than are the silent rRNA human genes, but adduct removal was similarly inefficient for both classes. We conclude that the repair of chemical damage such as psoralen photoadducts in an expressed mammalian gene may depend upon the class of transcription to which it belongs.

Differential repair and replication of damaged DNA in ribosomal RNA genes in different CHO cell lines.

We studied the repair of psoralen adducts in the pol I-transcribed ribosomal RNA (rRNA) genes of excision repair competent Chinese hamster ovary (CHO) cell lines, their UV sensitive mutant derivatives, and their UV resistant transformants, which express a human excision repair gene. In the parental cell line CHO-AA8, both monoadducts and interstrand crosslinks are removed efficiently from the rRNA genes, whereas neither adduct is removed in the UV sensitive derivative UV5; removal of both adducts is restored in the UV resistant transformant CHO-5T4 carrying the human excision repair gene ERCC-2. In contrast, removal of psoralen adducts from the rRNA genes is not detected in another parental CHO cell line CHO-9, neither in its UV sensitive derivative 43-3B, nor in its UV resistant transformant 83-G5 carrying the human excision repair gene ERCC-1. In contrast to such intergenomic heterogeneity of repair, persistence of psoralen monoadducts during replication of the rRNA genes occurs equally well in all CHO cell lines tested. From these data, we conclude that: 1) the repair efficiency of DNA damage in the rRNA genes varies between established parental CHO cell lines; 2) the repair pathways of intrastrand adducts and interstrand crosslinks in mammalian cells share, at least, one gene product, i.e., the excision repair gene ERCC-2; 3) replicational bypass of psoralen monoadducts at the CHO rRNA locus occurs similarly on both DNA strands.

DNA damage stimulates human cell transformation by integrative but not episomal Epstein-Barr virus-derived plasmid.

Previous work has demonstrated that ultraviolet (UV) irradiation of SV40-based plasmids can strikingly enhance the frequency of stable transformation of human cells. In this study we compared the effect of UV-induced DNA damage on transformation mediated by integrative versus autonomously replicating plasmids derived from human Epstein-Barr virus (EBV). We report that transfection of human fibroblasts with UV-irradiated integrative EBV-based plasmid results in enhanced transformation. However, transfection of UV-damaged episomal EBV-based constructs into the same human cell line does not enhance transformation; in fact, the extrachromosomal status of the plasmid is maintained irrespective of the UV dose to the plasmid. We conclude that enhanced transformation of human cells by damaged DNA requires its chromosomal integration.