In today’s loopy world, the wise words “Those who cannot remember the past are condemned to repeat it” have a particular resonance. Thankfully, our epigenomes are dynamic historians that get the message and make use of past transcripts by forming R-loops.
R-Loops and the Epigenome
R-loops are RNA-DNA hybrids: a 3 stranded nucleic acid structure. The interaction of RNA with genomic DNA results in the displacement of one strand to form a loop. R-loops occur in regions of GC skew, which is where C and G are not evenly distributed across the plus and minus strands.
The lab of Frédéric Chédin at the University of California Davis has previously shown that R-loops occur in unmethylated CpG island promoters and the terminals of human genes. Furthermore, R-loops appear to have a function in the brain as they are involved in genomic imprinting and Prader–Willi syndrome.
DRIPc-seq Interrogates R-Loops
Now, the Chédin lab is at it again and with an upgrade to their DNA:RNA immunoprecipitation sequencing (DRIP-seq) technique known as DRIPc-seq. DRIPc-seq involves following immunoprecipitation with DNase I treatment and then a reverse-transcription of the RNA to cDNA that allows for a strand-specific RNA-seq with higher resolution than DRIP-seq. Putting DRIPc-seq to the test, the team gained unprecedented insight into the nature of R-loops in human embryonic carcinoma (Ntera2) cells.
R-loops are “Prevalent, Dynamic, and Conserved”
Here’s what they found:
- R-loops are widespread and occupy 5% of the human genome
- They are also regulated in terms of frequency and residence time
- The majority of R-loops come from RNA polymerase II-dependent genes
- Over 90% of R-loops come from the sense strand of their respective genes and form co-transcriptionally
- By comparing to mice, they found that R-loops are conserved across species at gene promoters and terminals
The team then dug deeper into promoter and terminal R-loops by comparing them to ENCODE data. This revealed that R-loops are associated with open chromatin and that there are distinct types of R-loops.
Promoter R-Loops:
- Occur in 8,112 genes
- Are typically found in CpG islands with a GC skew
- Have a distinct epigenetic signature:
- Enriched for H3K4me2, H3K4me3, H3K9ac, and H3K27ac
- DNA hypomethylation
- Distinct enrichments and depletions for chromatin binding factors representative of open chromatin
Terminal R-Loops:
- Observed in 9,320 genes
- Genes are dependent on polyadenylation
- No association with GC skew
- Associated with regions of open chromatin but with a different signature: H3K4me1 and p300
- Overlap with enhancers
- Contain increased levels of CTCF and related proteins that suggest insulator function
- Involved in transcriptional termination by releasing RNA polymerase
Together, these findings suggest that R-loops are a widespread and dynamic feature of the genome that associate with epigenetic regulation. Given that R-loops form co-transcriptionally, it appears that they could be a mechanism for past transcriptional events to leave their mark.
Staying in the Loop
In order to gain more insight into R-loops, the EpiGenie team reached out to Professor Chédin. Chédin summarizes, “R-loops correspond to a novel path for nascent mRNAs: instead of being released normally and processed, the nascent RNA strand hybridizes to the DNA template strand, creating a metastable RNA:DNA hybrid that can reach up to 1 kilobase in length. The non-template DNA strand, no longer having a partner to pair with, exists in a looped-out single-stranded state. This explains the name for these non-B DNA structures: RNA-loops or R-loops.”
R-Loop Function
In terms of functionality, Chedin shares, “We are just beginning to learn about the interactions between R-loops and the epigenome. Obviously, R-loops, owing to their length and rigidity, will exert an impact on the local chromatin. We find that R-loop forming regions tend to be hyper-accessible, suggesting that R-loops prevent nucleosome formation. R-loops are also known to impede transcription in vitro and our study shows that R-loops associate with a high density of RNA Polymerase II most likely corresponding to transiently stalled complexes. This opens the door to R-loops regulating histone modifications that are deposited co-transcriptionally via interaction of chromatin modifying enzymes with the RNA Polymerase. Our data supports this possibility. Finally, it is entirely possible that R-loops are directly recognized by chromatin modifying complexes. Recent work by the Fazzio lab substantiates this exciting possibility.”
However, the regulatory potential doesn’t end there, as Chédin opines, “It is in principle possible that lncRNAs may target distant loci via R-loop formation in trans, in effect utilizing the formation of complementary RNA:DNA hybrids as a targeting mechanism. In fact a recent study in yeast offers evidence for this mechanism. It is too soon to know if a similar mechanism applies in mammalian systems but this is undoubtedly a topic of intense interest.”
The Future of R-Loops
Chédin also mentioned his vision for the future of R-loop research. “The R-loop field has been rapidly expanding over recent years, in part due to technological advances in detecting endogenous structures. On the technology side, we would love to measure R-loops on single alleles at high throughput and are optimizing a method to do just that. This will be very exciting. Conceptually, much remains to be learned about the protein complexes involved in making, sensing, resolving, stabilizing or preventing these structures in cells. Gathering this information is not only critical to our understanding of basic mechanisms governing R-loop metabolism and function but also to how R-loop dysfunction is tied to human diseases such as Amyotrophic Lateral Sclerosis (ALS) and other neurodegenerative disorders.”
Chédin concludes, “Our recent study and several others have profoundly shifted our views of R-loops from improbable, rare structures formed accidentally in pathological situations associated with mutations or disease to prevalent structures that form dynamically over conserved regions. In fact, R-loops are now arguably, the best documented, most abundant, non-B DNA structure in mammalian genomes.”
Go learn how to decipher R-loops over at Molecular Cell, July 2016.