3 RAG1/2 introduces double-stranded breaks at non-Ig loci | 71 Table 3A and 3B) overlapped with the publicly available RAG1 Chip-seq peaks published by another group13. The cells used for the previously reported RAG1 ChIP-seq were primary mouse pre-cells, whose DNA damage-response pathways are unimpaired, as opposed to the Artemis-/- cell line used in our study, where the absence of Artemis would cause accumulation of the DNA breaks. In primary WT pre-B cells, the DSBs (including those caused by RAG1/2) would be efficiently repaired, which would perhaps result in a RAG1 binding pattern than in Artemis-/- cells. Also, the binding of RAG1 to a certain place in the genome of WT bone marrow pre-B cells might be transient and represent a “snap-shot” at a particular time point, which could explain the low overlap degree. It should also be noted that our analyses represent a collective portrayal of the cell population and it remains uncertain to what degree the individual lymphocytes mirror the population averages as shown by the ChIP-seq. The association of genomic lesions in B-ALL to RAG1/2 activity is based either on the proximity of conserved RSS near the breakpoint, the presence of N-nucleotides which are reminiscent of V(D)J joints at the breakpoints, or the proximity of the breakpoints to cRSS10,12. Though sequences directly under the RAG-dependent NBS1 peaks did not show enrichment for RSSs, we found a clear enrichment especially for nonamers, and in a lesser extent also for heptamers 500bp and 1000bp around the NBS1 peaks on Igk locus. In the rest of the genome a moderate enrichment for the nonamer, but not the heptamer, was observed. This seems to be in line with the observation that nonamers, but not heptamers, are enriched in the proximity of RAG1 binding regions in mouse v-Abl cell lines and mouse thymocytes13. Interestingly, the nonamer was found mostly enriched in the sequences with 1000bp offset, while the heptamer was mostly enriched in the sequences with 500bp offset. Though MRN complex, including NBS1, binds the DNA at the broken ends58, it is possible that the association of NBS1 with the DNA around the DSBs takes place a few hundred bps away from the actual break, as also suggested by the NBS1-chromatin binding studies59. The motifs identified around RAG-dependent NBS1 peaks were associated with various zinc-finger binding domains, the most common structural motif in eukaryotic cells. Most of the known motifs were shared by several transcription factors or DNA-binding proteins. Interestingly, several of the identified motifs were known binding sites of DNA repair proteins, such as PRDM9, EGR1, MAZ, or RREB141–44, or transcription factors involved in p53 signaling and apoptosis such as PATZ1or SP145,46, on one hand suggesting the susceptibility of these regions to RAG-mediated damage, on the other hand underscoring the potential threat that RAG1/2 off-target activity can inflict. Another identified motif was associated with MYC-associated zinc finger protein (MAZ), a transcription factor regulating the MYC oncogene expression and affecting large-scale genome organization through interaction with cohesin60. Part of the MAZ binding motif is partially shared by other transcription
RkJQdWJsaXNoZXIy MTk4NDMw