72 | Chapter 3 factors such as SP161, which is also reflected in our findings (Figure 5C), this motif is also CG-rich, which further emphasizes the preference of RAG for GC-rich motifs. Next to RSS enrichment, we found a strong enrichment for CA and GA simple repeats around the RAG-dependent NBS1 peaks. Repeat DNA has been associated with alternative, non-B DNA structures, which often assume an open chromatin configuration and are also prone to the acquisition of mutations62. Such DNA structures are thought to contribute to genomic instability, for instance, H-DNA forming regions have been identified near the breakpoints of the translocated c-myc in Burkitt’s lymphoma and t(12;15) BALB/c plasmacytomas63. Non-B DNA structures were also identified around the Bcl-2 major breakpoint region (Mbr), where the RAG complex was shown to bind and nick at duplex–single-strand transitions of non-B DNA structures, resulting in double-strand breaks in vitro, thus demonstrating the transposase activity of RAGs12. RAG1/2 has been shown to cleave non-B cell structure without the presence of RSS, thus also functioning as a structure-specific endonuclease50. In fact, CA repeats have previously been shown to be a type of cRSS (mimicking the nonamer and the heptamer) and target of RAG1/2 activity, thus inadvertently introducing DNA breaks at CA-rich sites14,66. A more recent study has even demonstrated that the Znc2 domain of RAG1 is required for binding to non-B DNA structures; mutations introduced in this domain hindered the nicking capacity of the endonuclease on DNA quadruplexes and single-stranded regions.67 Therefore, the presence of the CA repeat regions and the GC-rich regions around the RAG-dependent NBS1 peaks may reflect the off-target activity of RAGs at the non-B DNA structures. Taken together, RAG1/2 is able to introduce DNA breaks on non-Ig and repeat regions are one of the common denominators of the off-target lesions. Non-B DNA structures, previously also implicated in the off-target activity of AID in mature B cells65, are more prone to DNA damage, putting lymphocytes under additional DNA damage stress during their development and maturation. Acknowledgements We thank Linda Koster, Core Facility Genomics, Amsterdam UMC, for helping with sample handling and library preparations. We also thank Ing. Ivan Mackovic EUR ING for providing a server space for the sequencing data, and to Ing. Jakub Mackovic for developing and testing a Python script to analyze the overlap between different ChiP-Seq peak data sets.
RkJQdWJsaXNoZXIy MTk4NDMw