RAD52-mediated single-strand DNA annealing: a single-molecule approach

Abstract

The integrity of the genetic information is a fundamental requirement for the maintenance of somatic cell lineages and crucially, the maintenance and transfer of genetic traits within a gene pool. However, there is a plethora of exogenous factors, such as UV-light, radiation, and various chemicals, and endogenous factors including reactive oxygen species, failed DNA replication, and meiosis that can affect genome integrity. The variety of factors and damaging agents result in a multitude of DNA damage types with different levels of cytotoxicity. Amongst all of them, DNA double-strand breaks are the most severe type of DNA damage. Since both strands are affected, it results in a more complex process of information restoration. Cells utilize two major pathways to repair DNA double-strand breaks: non-homologous end joining and homologous recombination. While the non-homologous end joining relies only on short homologous stretches of DNA, it leads to the accumulation of errors and point mutations within the genome. On the contrary, DNA repair via homologous recombination utilizes a sister chromatid to ensure faithful restoration of the genetic information. Conventional homologous recombination in mammalian cells is dependent on RAD51, which requires a variety of interaction partners to ensure its localization on DNA strands. In yeast, Rad52 has been identified as a mediator that displaces DNA end-protection proteins and facilitates RAD51 loading. In mammalian cells, this function is taken over by BRCA2. However, in the case of BRCA2-deficient cells, which is the hallmark of many cancers, human RAD52 can overcome this deficiency, secure RAD51 loading in vitro and promote the survival of cancer cells. Thus, RAD52 has become a very prominent target molecule in anti-cancer therapeutic studies. Nevertheless, the exact mechanism by which RAD52 performs DNA annealing remains unknown. The latest annealing model, utilizing the intrinsic propensity of the protein to self-oligomerize, has proposed that multiple ring structures would load onto single-stranded DNA and search for the homology while sliding along each other. This model has been established by studying only the N-terminal domain of RAD52 and by utilizing very high protein concentrations in the presence of large amounts of salt that could artificially favor the formation of ring structures. The inability of this model to explain the proof-reading mechanism together with other inconsistencies has raised a need for an alternative annealing model. By combining biochemical assays with single-molecule mass photometry, I investigated the dependency between the DNA binding and annealing activity of the full-length RAD52 and its N-terminal truncation and their oligomeric states under various experimental conditions. These experiments revealed a striking difference between the proteins. While RAD52(209) primarily formed undecameric rings, RAD52, which was known to form ring structures consisting of seven subunits, unexpectedly, exhibited the formation of a heterogeneous pool of rings with different numbers of subunits. Moreover, the mass photometry revealed monomers and lower-order oligomers that have been seldom observed before. Upon binding to single-stranded DNA, both proteins formed higher-order oligomers. Interestingly, the DNA excess caused the dissociation of these oligomers as well as single rings increasing the fraction of short oligomers. Overall, these experiments bring new information that will contribute to the revision of RAD52's annealing mechanism and help the design of more efficient therapeutic pathways in anti-cancer research.