RNA and DNA end structure dependent 5’ end processing during in vitro NHEJ

Abstract

Double-strand breaks (DSBs) are the most hazardous lesions sustained by DNA, with potential for the loss and rearrangement of genetic information, if not repaired correctly. Non-homologous end-joining (NHEJ) and Homologous Recombination (HR) are the 2 major pathways in cells for repairing DSBs. NHEJ processes the broken ends if required, and ligates them to each other. On the other hand, HR requires the presence of a homologous sister chromatid to serve as a template for synthesis, and is thus restricted to the S and G2 phases of the cell cycle.

NHEJ begins with binding of the Ku heterodimer to DNA ends. The binding of DNA-PKcs and the subsequent activation of the DNA-PK holoenzyme through phosphorylation brings the ends together for ligation by the XRCC4-LigIV complex. NHEJ is an iterative process where, if necessary, the ends are processed by a hierarchy of enzymes until the ends are fit for ligation. Besides the repair of DSBs generated by exogenous sources like radiation, NHEJ is crucial for cellular processes like V(D)J recombination during lymphocyte development.

Using human cell extracts that perform in vitro NHEJ on linearized plasmid substrates, I describe a novel 5’ end processing activity that is specific for DNA end structures with a protruding 5’ end. Treatment of the extract with RNase A or blocking DNA-PK activation inhibited the described processing. The processing occurs at ends that are already compatible for direct joining, implying that it is independent of the typical known NHEJ processing and could act as a mechanism to generate genomic diversity.

In order to identify the processing factor, biotinylated substrates were used to pulldown proteins binding to defined ends from RNase-treated and untreated extracts. Mass spectrometry showed a clear and significant change in the end-binding proteome after RNase treatment. Multiple proteins and complexes were found to not associate with the DNA after RNase treatment of the extract. Further work is required to validate these hits and identify the processing factor.

Parallel to studying RNA-dependent end processing, I also investigated the effect of nucleotide sequence on NHEJ. I generated a library of substrates with random nucleotides at or proximal to the end. Analysis of NHEJ junctions showed a preference towards thymidines in highly joined blunt ended breaks. For substrates with defined ends/overhang, positions far away from the break (>10 bp) exhibited mild bias for certain nucleotides. This altogether suggests that sequence indeed affects repair by NHEJ.