R-loop processing is impaired by the disassembly of the RPA/RNase H1 complex in oncogene-induced senescence

Genome stability is important for organisms and its maintenance depends significantly on the successful activation of the DNA damage response (DDR) and the role of non-canonical chromatin structures (NCSs), which regulate genome replication, transcription and DNA repair. In senescent cells, defective DDR often results in the accumulation of unrepaired DNA damage. Among NCSs, R-loops are three-stranded RNA-DNA hybrids, consisting of a nascent RNA strand that attaches to the DNA template, forming a hybrid that displaces the non-template single-strand DNA (ssDNA). R-loops influence crucial cellular processes and many factors, such as RNase H enzymes, regulate their processing. Additionally, the ssDNA-binding protein RPA enhances RNase H activity to repress R-loops and prevent genomic instability. Dysregulation of proper R-loop processing - excessive accumulation or excessive processing - can lead to increased genomic instability and DNA damage. In this study, we aimed to understand the mechanisms that ensure proper control of R-loops, leading to efficient DNA repair of exogenously induced DSBs, in particular in a senescent context. Efficient DSB repair is ensured by proper kinetics of R-loop processing, which enables precise non-homologous end joining (NHEJ) and homologous recombination (HR). Our experiments revealed that late processing or early degradation of R-loops at DSBs induced by endonucleases alters accurate DNA repair. Therefore, the precise timing of R-loop formation and resolution is crucial for the HR and NHEJ pathway selection and for DSB repair. Using three different models (U2OS EJ5-GFP and U2OS TRI-DR-GFP cells, a deadCas9-based delivery system and LacO-TetO system), we found that forced recruitment of both catalytically active and inactive RNase H1 to DSBs delays DNA repair. This suggests that both premature and delayed removal of R-loops affects DNA repair and the recruitment of repair protein at DSBs. By investigating R-loops’ role in genome stability, our findings showed impaired R-loops control and processing at DSBs in BJ fibroblasts engineered to express the HRASG12V oncogene, a model for oncogene-induced senescence (OIS), compared to proliferating cells and BJ fibroblasts escaping OIS via HDAC4 overexpression. Cells undergoing OIS accumulate more R-loops than proliferating fibroblasts and only a small fraction of them supported BRCA1 loading. DRIP-seq and R-ChIP experiments revealed that unsuccessful recruitment of RNase H1 by the RPA complex to R-loops impairs DDR protein loading at DSBs in OIS cells compared to proliferating and OIS-escaped cells. The trimeric RPA complex, consisting of RPA70, RPA32 and RPA14 subunits, binds to ssDNA and regulates RNase H1 recruitment to R-loops. Our research identified RPA32 phosphorylation (pRPA32) as the signal controlling RNase H1 activity at R-loops in correspondence to DSBs. Hyperphosphorylation of the RPA complex observed in pre-senescent cells alters this mechanism, leading to disassembly of the RPA/RNase H1 complex, accumulation of long-lived R-loops and irreparable DNA damage. Mass spectrometry, surface plasmon resonance (SPR) and in vitro experiments confirmed the decreased processing kinetics of RNase H1 in the presence of pRPA32. Since the molecular details of proper R-loop resolution are not yet fully understood in the literature, modulating R-loop processing kinetics could offer a new therapeutic approach to prevent the development of senescence and replication stress, thereby enhancing DSB repair efficiency.