Recurrent Translocations in Topoisomerase Inhibitor-Related Leukemia Are Determined by the Features of DNA Breaks Rather Than by the Proximity of the Translocating Genes.

Topoisomerase inhibitors are widely used in cancer chemotherapy. However, one of the potential long-term adverse effects of such therapy is acute leukemia. A key feature of such therapy-induced acute myeloid leukemia (t-AML) is recurrent chromosomal translocations involving AML1 (RUNX1) or MLL (KMT2A) genes. The formation of chromosomal translocation depends on the spatial proximity of translocation partners and the mobility of the DNA ends. It is unclear which of these two factors might be decisive for recurrent t-AML translocations. Here, we used fluorescence in situ hybridization (FISH) and chromosome conformation capture followed by sequencing (4C-seq) to investigate double-strand DNA break formation and the mobility of broken ends upon etoposide treatment, as well as contacts between translocation partner genes. We detected the separation of the parts of the broken AML1 gene, as well as the increased mobility of these separated parts. 4C-seq analysis showed no evident contacts of AML1 and MLL with loci, implicated in recurrent t-AML translocations, either before or after etoposide treatment. We suggest that separation of the break ends and their increased non-targeted mobility-but not spatial predisposition of the rearrangement partners-plays a major role in the formation of these translocations.

The broken MLL gene is frequently located outside the inherent chromosome territory in human lymphoid cells treated with DNA topoisomerase II poison etoposide.

The mixed lineage leukaemia (MLL) gene is frequently rearranged in secondary leukaemias, in which it could fuse to a variety of different partners. Breakage in the MLL gene preferentially occurs within a ~8 kb region that possesses a strong DNA topoisomerase II cleavage site. It has been proposed that DNA topoisomerase II-mediated DNA cleavage within this and other regions triggers translocations that occur due to incorrect joining of broken DNA ends. To further clarify a possible mechanism for MLL rearrangements, we analysed the frequency of MLL cleavage in cells exposed to etoposide, a DNA topoisomerase II poison commonly used as an anticancer drug, and positioning of the broken 3'-end of the MLL gene in respect to inherent chromosomal territories. It was demonstrated that exposure of human Jurkat cells to etoposide resulted in frequent cleavage of MLL genes. Using MLL-specific break-apart probes we visualised cleaved MLL genes in ~17% of nuclei. Using confocal microscopy and 3D modelling, we demonstrated that in cells treated with etoposide and cultivated for 1 h under normal conditions, ~9% of the broken MLL alleles were present outside the chromosome 11 territory, whereas in both control cells and cells inspected immediately after etoposide treatment, virtually all MLL alleles were present within the chromosomal territory. The data are discussed in the framework of the "breakage first" model of juxtaposing translocation partners. We propose that in the course of repairing DNA topoisomerase II-mediated DNA lesions (removal of stalled DNA topoisomerase II complexes and non-homologous end joining), DNA ends acquire additional mobility, which allows the meeting and incorrect joining of translocation partners.

Identification of conserved features of LAGLIDADG homing endonucleases.

LAGLIDADG family of homing endonucleases are rare-cutting enzymes which recognize long target sequences and are of great interest in genome engineering. Despite advances in homing endonuclease engineering, effective methods of broadening the range of cleaved sequences are still lacking. Here, we present a study of conserved structural features of LAGLIDADG homing endonucleases that might aid further development of such methods. The protein-DNA interface of LAGLIDADG homing endonucleases differs considerably with the particular nuclease, and the analysis of conserved protein-DNA interactions could not identify any residues crucial for DNA binding and common to most nucleases of the family. For the homing endonuclease PI-SceI, a comparison of structural and experimental data derived from literature helped to identify 23 residues that are likely to be important for DNA binding. Analysis of the LAGLIDADG domain dimerization interface allowed the choosing of six positions that contribute to dimerization specificity most, while comparison of 446 sequences of LAGLIDADG endonucleases revealed groups of residues in these positions that appear to be most favorable for dimerization.