The study of R-loops has recently gained increasing attention in the scientific community, with a growing number of studies devoted to these fascinating three-stranded nucleic acids, composed of DNA:RNA hybrids and displaced single-stranded DNA moieties. It is now generally accepted that co-transcriptional R-loops play a dual role in genome homeostasis. While they contribute to essential physiological functions, their unscheduled accumulation can have detrimental consequences for genetic integrity, leading to the prevailing view that these structures are inherently genotoxic. Consistently, inactivation of factors that limit R-loop formation has been linked to various pathological conditions, from cancer to neurological disorders.
#### **How it started – the screening challenge**
With this in mind, we set out to identify new factors that could prevent the accumulation of R-loops and their associated genotoxicity. Our approach? A (long) series of systematic screens designed to uncover budding yeast mutants that exhibit both DNA:RNA hybrid accumulation and increased genetic instability. What we did not expect, however, was to identify as a top hit Rad27, a well-characterized nuclease involved in the processing of DNA replication intermediates - the Okazaki fragments. Strikingly, the loss of Rad27 actually led to increased hybrid levels all over the transcribed regions of the yeast genome, as measured by genome-wide analyses. Similarly, interfering with the activity of its mammalian counterpart, Fen1, triggered hybrid accumulation in human cell lines. These unexpected findings raised a question: how was this nuclease connected to the regulation of hybrids?
#### **The twist: from _in vitro_ reconstitution to yeast genetics**
Initially, we wondered whether Rad27 could directly target the single-stranded part of R-loops, as other nucleases involved in hybrid regulation do. However, our _in vitro_ experiments revealed something quite different: Rad27 cleaved DNA flaps, its natural branched substrates in Okazaki fragment processing, much better than R-loops. This suggested that Rad27 does not actively remove hybrids, but rather prevents their formation by processing DNA flaps, and raised a new question: could DNA flaps themselves drive hybrid formation? To explore this alternative, we took advantage of the yeast model’s ability to genetically modulate flap accumulation. The outcome was striking: the more flaps accumulated, the more hybrids were detected! In fact, hybrid formation was not restricted to situations of flap accumulation: preventing Okazaki fragment ligation was similarly associated with high hybrid levels. Overall, our data pointed to DNA discontinuities – DNA flaps, single-stranded gaps - as the primary culprits in triggering hybrid formation.
#### The chicken and egg problem: do hybrids cause genetic instability?
These observations left us with a concern: if such natural DNA lesions precede the formation of DNA:RNA hybrids, do these structures contribute to the high levels of genetic instability detected in Rad27-deficient cells? To specifically address this question, we interfered with hybrid accumulation either by transcriptional repression, or by increasing the dosage of RNase H, a hybrid-specific ribonuclease. In both cases, genetic instability remained unchanged, suggesting that hybrids were not the cause, but rather the consequence of pre-existing DNA damage. Interestingly, this trend was not unique to Rad27 inactivation or replication defects: we made similar observations in several other yeast mutants with high genetic instability and hybrid accumulation, but also upon FEN1 inhibition in human cells, convincing us of the broad relevance of this discovery.
#### **Proposing a novel classification for DNA:RNA hybrids**
Reassuringly, we were also able to assess hybrid-accumulating situations where R-loops had previously been shown to be the cause of genetic instability - and confirm that in some cases, such as defective mRNA packaging or processing, hybrids do indeed precede the formation of genotoxic DNA lesions. This led us to propose a novel classification of DNA:RNA hybrids based on their stage of formation with respect to DNA damage: "_pre-lesion_" hybrids would accumulate in situations of defective RNA metabolism and subsequently trigger DNA damage and genetic instability, whereas “_post-lesion_” hybrids would form in the presence of DNA discontinuities resulting from defects in DNA metabolism, _e.g._ replication or repair. Whether such DNA lesions trigger hybrid formation by facilitating RNA invasion or by interfering with transcriptional dynamics remains however to be investigated.
#### **Why it matters**
Our study challenges the widely accepted notion that the accumulation of DNA:RNA hybrids is a systematic source of genetic instability. In the many pathological situations where hybrids or R-loops have been detected, these structures may thus be the byproducts rather than the drivers of genome instability. If so, DNA:RNA hybrids may hold promise as diagnostic tools or biomarkers rather than therapeutic targets, providing insight into disease progression.
Of course, it took a variety of expertise to complete this project - from systematic screens, _in vitro_ reconstitution, and yeast genetics to DNA:RNA hybrid imaging or molecular detection. While the screening effort involved the invaluable support of Rodney Rothstein at Columbia University Medical Center (New York, USA), _in vitro_ assays were initiated with Pierre-Henri Gaillard (Cancer Research Center, Marseille, France) and DNA:RNA hybrid imaging was performed with Sergio de Almeida at the Gulbenkian Institute for Molecular Medicine (Lisbon, Portugal) and Peter Stirling at the University of British Columbia (Vancouver, Canada). So the scientific journey has certainly been a human adventure.