Molecular analysis of Fanconi anaemia

Disruption of mouse SNM1 causes increased sensitivity to the DNA interstrand cross-linking agent mitomycin C

DNA interstrand cross-links (ICLs) represent lethal DNA damage, because they block transcription, replication, and segregation of DNA. Because of their genotoxicity, agents inducing ICLs are often used in antitumor therapy. The repair of ICLs is complex and involves proteins belonging to nucleotide excision, recombination, and translesion DNA repair pathways in Escherichia coli, Saccharomyces cerevisiae, and mammals. We cloned and analyzed mammalian homologs of the S. cerevisiae gene SNM1 (PSO2), which is specifically involved in ICL repair. Human Snm1, a nuclear protein, was ubiquitously expressed at a very low level. We generated mouse SNM1(-/-) embryonic stem cells and showed that these cells were sensitive to mitomycin C. In contrast to S. cerevisiae snm1 mutants, they were not significantly sensitive to other ICL agents, probably due to redundancy in mammalian ICL repair and the existence of other SNM1 homologs. The sensitivity to mitomycin C was complemented by transfection of the human SNM1 cDNA and by targeting of a genomic cDNA-murine SNM1 fusion construct to the disrupted locus. We also generated mice deficient for murine SNM1. They were viable and fertile and showed no major abnormalities. However, they were sensitive to mitomycin C. The ICL sensitivity of the mammalian SNM1 mutant suggests that SNM1 function and, by implication, ICL repair are at least partially conserved between S. cerevisiae and mammals.

Fanconi anemia, complementation group A, cells are defective in ability to produce incisions at sites of psoralen interstrand cross-links

The hypersensitivity of Fanconi anemia, complementation group A, (FA-A) cells to agents which produce DNA interstrand cross-links correlates with a defect in their ability to repair this type of damage. In order to more clearly elucidate this repair defect, chromatin-associated protein extracts from FA-A cells were examined for ability to endonucleolytically produce incisions in DNA at sites of interstrand cross-links. A defined 140 bp DNA substrate was constructed with a single site-specific monoadduct or interstrand cross-link produced by 4,5',8-trimethylpsoralen (TMP) plus long wavelength (UVA) light. Our results show that FA-A cells are defective in ability to produce dual incisions in DNA at sites of interstrand cross-links. Specifically, there is defective incision on the 3'- and 5'-sides of both the furan and pyrone sides of the cross-link. This defect is corrected in FA-A cells transduced with a retroviral vector expressing FANCA cDNA. At the site of a TMP monoadduct, FA-A cells can introduce incisions on both the 3'- and 5'-sides of the furan side monoadduct, but are defective in ability to produce these incisions on the pyrone side monoadduct. These studies also indicate that XPF is involved in production of the 5' incision by the normal extracts on these substrates. These results correlate with our previous work, which showed that FA-A cells are mainly defective in ability to repair psoralen interstrand cross-links with a lesser defect in ability to repair psoralen monoadducts. This defect in endonucleolytic incision at sites of TMP interstrand cross-links could be related to reduced levels of non-erythroid alpha spectrin (alphaSpIISigma*) in the extracts from FA-A cells. alphaSpIISigma* could act as a scaffold to align proteins involved in cross-link repair and enhance their interactions; a deficiency in alphaSpIISigma* could thus lead to reduced efficiency of repair and the decreased levels of incisions we observe at sites of interstrand cross-links in FA-A cells.

Genetic disorders associated with cancer predisposition and genomic instability

Genomic instability in its broadest sense is a feature of virtually all neoplastic cells. In addition to the mutations and/or gene amplifications that appear to be a prerequisite for the acquisition of a neoplastic phenotype, human cancers exhibit other "markers" of genomic instability--in particular, a high degree of aneuploidy. Indeed, many studies have shown that aneuploidy is an almost invariant feature of cancer cells, and it has been argued by some that the emergence of aneuploid cells is a necessary step during tumorigenesis. The functional link between genomic instability and cancer is strengthened by the existence of several "genetic instability" disorders of humans that are associated with a moderate to severe increase in the incidence of cancers. These disorders include ataxia telangiectasia, Bloom's syndrome, Fanconi anemia, xeroderma pigmentosum, and Nijmegen breakage syndrome, all of which are very rare and are inherited in a recessive manner. Analysis of the cells from such cancer-prone individuals is clearly a potentially fruitful approach for delineating the genetic basis for instability in the genome. It is assumed that by identifying the underlying cause of genetic instability in these disorders, one can derive valuable information not only about the basis of particular genetic diseases, but also about the underlying causes of genomic instability in sporadic cancers in the general population. In this article, we review the clinical and cellular properties of genetic instability disorders associated with cancer predisposition. In particular, we focus on the rapid advances made in our understanding of these disorders that have derived from the cloning of the genes mutated in each case. Because in many instances the affected genes have analogs in lower eukaryotic species, we shall discuss how studies in yeasts in particular have proved valuable in our understanding of human diseases and predisposition to cancer.

Interstrand cross-links induce DNA synthesis in damaged and undamaged plasmids in mammalian cell extracts

Mammalian cell extracts have been shown to carry out damage-specific DNA repair synthesis induced by a variety of lesions, including those created by UV and cisplatin. Here, we show that a single psoralen interstrand cross-link induces DNA synthesis in both the damaged plasmid and a second homologous unmodified plasmid coincubated in the extract. The presence of the second plasmid strongly stimulates repair synthesis in the cross-linked plasmid. Heterologous DNAs also stimulate repair synthesis to variable extents. Psoralen monoadducts and double-strand breaks do not induce repair synthesis in the unmodified plasmid, indicating that such incorporation is specific to interstrand cross-links. This induced repair synthesis is consistent with previous evidence indicating a recombinational mode of repair for interstrand cross-links. DNA synthesis is compromised in extracts from mutants (deficient in ERCC1, XPF, XRCC2, and XRCC3) which are all sensitive to DNA cross-linking agents but is normal in extracts from mutants (XP-A, XP-C, and XP-G) which are much less sensitive. Extracts from Fanconi anemia cells exhibit an intermediate to wild-type level of activity dependent upon the complementation group. The DNA synthesis deficit in ERCC1- and XPF-deficient extracts is restored by addition of purified ERCC1-XPF heterodimer. This system provides a biochemical assay for investigating mechanisms of interstrand cross-link repair and should also facilitate the identification and functional characterization of cellular proteins involved in repair of these lesions.

Initiation of DNA interstrand cross-link repair in humans: the nucleotide excision repair system makes dual incisions 5' to the cross-linked base and removes a 22- to 28-nucleotide-long damage-free strand

Most DNA repair mechanisms rely on the redundant information inherent to the duplex to remove damaged nucleotides and replace them with normal ones, using the complementary strand as a template. Interstrand cross-links pose a unique challenge to the DNA repair machinery because both strands are damaged. To study the repair of interstrand cross-links by mammalian cells, we tested the activities of cell extracts of wild-type or excision repair-defective rodent cell lines and of purified human excision nuclease on a duplex with a site-specific cross-link. We found that in contrast to monoadducts, which are removed by dual incisions bracketing the lesion, the cross-link causes dual incisions, both 5' to the cross-link in one of the two strands. The net result is the generation of a 22- to 28-nucleotide-long gap immediately 5' to the cross-link. This gap may act as a recombinogenic signal to initiate cross-link removal.