Complex determinants of R-loop formation at transposable elements and major DNA satellites

Aberrant activation of transposable elements (TEs) has been a well-documented source of genomic instability and disease, stemming from their insertion into genes and their imposition of epigenetic effects on nearby loci. However, the extent to which their disruptive effects involve concomitant or subsequent formation of DNA:RNA hybrids (R-loops) remains unknown. Here, we used DNA:RNA immunoprecipitation followed by high-throughput sequencing (DRIP-seq) to map the R-loop profiles of TEs and satellites in Drosophila melanogaster ovaries in control and rhino knockout flies, where dozens of TE families are derepressed. We observe that R-loops form primarily in LTR retrotransposons that carry A/T-rich sequence motifs, which are known to favor R-loop formation at genes in Drosophila and other species. We also report evidence of R-loop formation at 11 of 14 highly abundant D. melanogaster DNA satellites. R-loop formation is positively correlated with expression level for both TEs and satellites; however, neither sequence content nor expression fully explain which repeat families form R-loops, suggesting other factors are at play. Finally, by analyzing population frequencies of R-loop-forming TEs, we present evidence that TE copies with high R-loop signal may be under stronger negative selection, which suggests that R-loop formation by TEs may be deleterious to their host. Collectively, these results provide insight into the determinants of R-loop formation at repetitive elements.