A role for phase separation to prevent R-loops during transcription

The protein FUS (FUSed in sarcoma) binds nascent RNA transcripts and all three nuclear RNA polymerases in metazoans. FUS undergoes phase separation to bind RNA polymerase II, which affects phosphorylation of the polymerase, recruitment of RNA processing factors, and transcription elongation and termination. This study investigates an activity of FUS targeting transcription but does not involve binding to the polymerase. An in vitro transcription assay revealed FUS prevented R-loops formed between the RNA transcript and DNA template, which could enhance transcription productivity for a non-eukaryote polymerase that FUS does not bind. The ability to bind RNA and phase separate was required for FUS activity. DRIP-seq analysis of human cells found endogenous activity matched shown by in vitro assays. FUS prevented R-loops at genomic locations and with sequence compositions consistent with previously characterized targets of FUS in cells. The model interpreted from these findings is that the abilities for RNA-binding and phase separation combine in FUS to prevent nucleic acid structures that disrupt chromosome stability and gene transcription.


INTRODUCTION
A prominent member of the heterogeneous nuclear ribonucleoprotein (hnRNP) family is FUS (FUsed in Sarcoma), known for mutations to its gene that can cause amyotrophic lateral sclerosis or drive pediatric sarcomas (1)(2)(3). FUS is conserved throughout metazoan species and among the highest expressed proteins in human tissues (3). FUS activity has been found to affect transcription by all three nuclear polymerases in eukaryotes (4)(5)(6)(7). FUS is also an enigmatic protein. Its structure is mostly intrinsically disordered. Among its disordered domains, the arginine and glycine-rich (RGG/RG) and low complexity (LC) domains of FUS bind RNA Pol II. FUS binds RNA Pol I, RNA Pol III, and other transcription and RNA-processing factors (8)(9)(10). ChIP-seq analysis of FUS localization on chromatin shows enrichment at the transcription start sites (TSS) of most expressed protein-coding genes. By a mechanism not fully understood, a loss of FUS leads to increased RNA Pol II phosphorylation at Ser-2 (Ser2P) in the C-terminal domain (CTD) heptad repeat. FUS can block phosphorylation of CTD by kinases P-TEFb and CDK12 in in vitro assays (7,11). FUS is not observed to affect phosphorylation of the nearby Ser-5 position, which is crucial to transcription initiation (7).
FUS affinity for RNA is provided by its RNA-recognition motif (RRM) and three RGG/RG domains (12)(13)(14)(15)(16). FUS specificity for RNA-binding is degenerate but prefers G-rich sequences, RNA stem-loops, and complex RNA structure (13,16). The number of messenger and noncoding RNAs in a cell that bind FUS is large (3). RNA binding causes FUS to oligomerize (14,17). FUS self-assembly results in phase separation, yielding a large molecular scaffold or condensate. FUS condensates bind the CTD of RNA Pol II and those formed in cells have been found to contain >80% of the polymerase (4,8,18).
This nuclear body has been observed at sizes 10 times the diameter of the polymerase. Assembly of FUS and RNA Pol II into the granule is transcription-dependent (19) The role of protein phase separation during eukaryotic transcription has drawn a high level of interest due to its profound implications to current paradigms for regulation of gene expression (20)(21)(22).
We set out to use in vitro transcription assays to better understand how isolated FUS functions contribute to its activity during transcription. We anticipated the most influential FUS activities would involve direct binding to a eukaryotic RNA polymerase and RNA processing factors (6,7,9,12). Our study instead revealed an effect of FUS on nucleic acid structures that influence transcription. We now show that FUS can alter transcription without direct interactions to the polymerase or other transcription factors.

Recombinant protein expression and purification
FUS constructs used in this study are available upon request. All recombinant proteins were expressed and purified from E. coli BL21(DE3) cells. All FUS constructs were N-terminally fused to 6x His and maltose-binding protein (MBP) to improve purification and solubility. Protein expression was induced after cell growth reached an OD600 of 0.8 and then allowed to continue overnight at 17 ºC with shaking (200 rpm). Purification was made using 1 to 2 g of frozen pellets of induced E. coli, lysates incubated with Ni-NTA Sepharose beads (Cytiva Lifesciences, 17531802), and eluted in FUS-SEC buffer (1 M urea, 0.25 M KCl, 50 mM Tris-HCl pH 8.0) with 250 mM imidazole and either 1.5 mM b-mercaptoethanol or 1 mM DTT added (4,13,15).
Plasmids for protein expression of T7(P266L) polymerase fused with a 6xHis tag were provided by A. Berman (University of Pittsburgh) (23). Expression in E. coli BL21(DE3) cells and purification were done essentially as previously published (24). Protein expression was induced by 0.5 mM Isopropyl b-D-1-thiogalactopyranoside (IPTG) when transformed cells grew to an OD600 of 0.6 then continued for 3 hours at 37 ºC. Induced cells stored at -80 ºC were thawed, lysed in T7 Buffer (50 mM Tris-HCl pH 8.0, 100 mM NaCl, 5% v/v glycerol, and 5 mM b-mercaptoethanol added just before use) with 1 mM imidazole then incubated with Ni-NTA beads. T7 Buffer with increasing imidazole was used to wash beads 4 times (1 mM imidazole), 2 times (10 mM imidazole), and to elute protein (100 mM imidazole).
Size-exclusion chromatography (SEC) was performed with a 10/300 Superdex® 200 column (Cytiva Lifesciences, 17517501) for FUS using FUS-SEC buffer and for T7 Pol with T7-SEC buffer (20 mM Tris-HCl pH 7.5, 100 mM NaCl, 0.1 mM EDTA, and 1 mM DTT). FUS protein was stored at room temperature for up to 4 weeks or until activity appeared lost. T7 Pol activity was assessed by titrating protein into transcription assays as described below. The absence of nuclease activity was confirmed by comparing plasmid DNA and transcribed RNA incubated overnight with or without protein and electrophoresis using a Novex™ TBE-urea, 6% polyacrylamide gel (Invitrogen, EC68655BOX) and stained with SYBR-Gold (Invitrogen, S11494).
Additional nucleic acids were yeast tRNA (Invitrogen, 15401011) and a commercially synthesized ssDNA (Millipore-Sigma). Nucleic acid sequences can be found in Supplemental Table 1.

DRIP-seq assay and data analysis
DRIP-seq assays were performed essentially as a previously published protocol (27). Sample libraries were prepared and sequenced by Novogene Corporation Inc. using a NovoSeq6000 with 150 base paired-end reads (8 G raw data per sample). Reads were trimmed using TrimGalore (version 0.6.6) and aligned to GRCh38 using STAR (version_2.7.8a), essentially as previously described. Bam files were processed using the deepTools2.0 suite (version 3.5.1) to normalize to 1x depth (reads per genome coverage) with bamCoverage, merge replicates by bigwigCompare, and tags binned by computeMatrix, which was then used to compute scaling factors and generate heat maps and profiles. Merged bigwig files were converted to bedgraph and then peaks by Macs2 (version 2.1.1.20160309). HOMER (version v4.11.1) was used to calculate GCcontent and annotate peaks. Data was visualized, and figures were generated using Integrated Genomics Viewer (version 2.5.0). DRIP-seq data is available (GSE206740) from the Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/geo/).

FUS can increase transcription without binding the polymerase
One model for FUS activity on RNA Pol II transcription would be to form protein assemblies at active genes that can concentrate polymerases, transcription machinery, or RNA processing factors that aid in transcription initiation and elongation. We have previously shown that RNA Pol II transcription of a naked DNA template is greatly enhanced by recombinant FUS protein added in concentrations near those of the cell, >2 µM (4). If protein-protein interactions with RNA Pol II were required for activity, we hypothesized that FUS would not act on a polymerase it cannot bind.
Multiple domains in FUS can bind RNA Pol II (8,14,28,29). FUS binding and activity has been observed for all three nuclear RNA polymerases (6,30). We turned to a DNA-dependent RNA polymerase from T7 phage (T7 Pol) to test FUS activity in the absence of binding to the polymerase (Supplemental Figure 1A). The P266L mutant of T7 Pol was chosen, which has higher stability during promoter clearance and produces fewer abortive transcripts (26). Run-off transcription of a linear DNA template was assessed by acrylamide gel electrophoresis (PAGE) and SYBR staining of the DNA and RNA present over time (Figure 1A, top). Contrary to expectation, transcription assays in the presence of freshly purified FUS (2 µM) yielded more RNA transcript ( Figure 1A). Titrating FUS concentration in T7 Pol transcription assays correspondingly increased RNA yield ( Figure 1B). Western blots of co-IP assays test FUS binding to T7 Pol. FUS protein (In) incubated with anti-FUS was pulled down, leaving no protein in the supernatant (S1). T7 Pol incubated with FUS-bound beads remained in supernatants (S2) and did not elute with FUS from beads (E). The presence of RNA ("+ RNA") (12 nM) did not cause FUS to bind T7 Pol. A lane with FUS and T7 Pol (+) served as a positive control for detection by western blot assay. See also Supplemental Figure 1C for non-specific IgG immunoprecipitation control. At FUS concentrations above 2 µM, a mobility shift was observed, suggesting an interference from protein aggregation on the detection of RNA by PAGE. Consistent with this interpretation, the well shift was observed when transcription was prevented by omitting NTP. The shift was also reduced by treating samples of run-off transcripts with proteinase K (Supplement Figure 1B). To eliminate this factor from transcript measurements, we also quantified run-off transcription assays using fluorescent dyes specific for RNA or DNA, which found that the increase to RNA yield could reach ≥ 3-fold with FUS added at concentrations ≥ 2 µM (Figure 1C).
We considered whether FUS had an unforeseen ability to bind T7 Pol. Co-immunoprecipitation (co-IP) assays revealed that a FUS-specific antibody could deplete FUS out of solution (S1) but not T7 Pol (S2) (Figure 1D). Likewise, only FUS was detected in eluates of beads bound to anti-FUS antibodies (E, Figure 1D). Co-IP assays that included an RNA known to cause FUS binding to RNA Pol II (RMRP) did not induce detectable binding to T7 Pol ( Figure 1D) (7,14). Linear plasmid DNA also had no effect (Supplemental Figure 1C). In conclusion, we found no evidence of direct binding between T7 Pol and FUS that might explain or contribute to the activity observed.

Activity on transcription is specific to FUS and not competed by RNA
We considered whether high protein concentrations might increase transcription by non-specific molecular crowding. T7 Pol activity has been reported to increase through molecular crowding by agents that include glycerol and PEG (31)(32)(33). We tested a well-folded soluble protein, bovine serum albumin (BSA), and detected no increase in RNA product (Figure 2A and 2B). We also saw FUS activity diminished and was lost for protein stored at -80 ºC. Activity was stable up to 4 weeks for FUS kept at room temperature (Supplemental Figure 2A). We noted the temperature effect on FUS matched what is for many proteins a result of structural instability during freezing.
Though RNA did not cause FUS to bind T7 Pol, we questioned whether RNA could increase or sequester activity on transcription. An RNA that is commonly used for a non-specific competitor to RNA-binding is tRNA. A second RNA, TET456, has tertiary structure that includes parallel stacked helices and is previously known to cause FUS binding to RNA Pol II. FUS activity was unchanged by the addition of tRNA at the same concentration by mass as the RNA produced by the assay ( Figure   2C, Supplemental Figure 2B). Titrating TET456 RNA to concentrations up to 12.8 µM also did not affect FUS activity (Supplemental Figure 2C). Lastly, single-stranded DNA had no effect on FUS activity on T7 Pol transcription (Supplemental Figure 2D).

FUS inhibits RNA:DNA hybrids formed during transcription
We considered whether the activity observed may involve nucleic acid structures. One structure is an RNA:DNA hybrid, formed by a nascent transcript displacing the non-template strand to bind its complementary template DNA strand (34,35). RNA:DNA hybrids internal to larger DNA molecules are called R-loops, which are found in cells at low but consistent levels (34,(36)(37)(38)(39). R-loops can affect transcription and chromosome stability, which may threaten many DNA functions, and multiple mechanisms exist in cells to prevent or remove R-loops (34,40).
We used a well-characterized antibody, S9.6, to detect RNA:DNA hybrids by dot blot assays (27,41). The results showed hybrids increase during run-off transcription ( Figure 3A). Digestion of RNA:DNA hybrids by RNaseH confirmed that the signals detected were specific to hybrids. Controls that lacked NTPs or T7 Pol showed that hybrids were a product of transcription, and the S9.6 antibody did not detect hybrids when transcription assays included the TET456 RNA, which lacks complementarity to the plasmid (Supplemental Figures 3A and 3B). While FUS (2 µM) did not affect the level of hybrids detected after 2 hours of transcription, hybrids were decreased between 20 and 40 minutes of transcription (Supplemental Figure 3C). As FUS concentration increased, RNA:DNA hybrids detected were reduced (Figure 3B). To test that transcripts made could still form hybrids, we performed dot blot assay with samples that had been heated and saw similar levels of hybrids across all samples (Figure 3B, anneal). Lastly, quantitating dot blot assays confirmed a 2-fold reduction in hybrids at the 20 minute timepoint for transcription ( Figure 3C).

Requirements for RNA-binding and phase separation for FUS activity
We sought to uncover a mechanism for activity by investigating the role of individual FUS domains.
The LC domain contributes to protein binding, oligomerization, and phase separation (4,(42)(43)(44). The three RGG/RG domains are capable of both phase separation and RNA-binding (8,14,(45)(46)(47). We compared FUS activity to that of the LC domain alone (FUS-LC) or RNA-binding domains (del-LC) ( Figure 4A). Neither truncated protein had significant activity on transcript yields or hybrids ( Figure   4B and C). This suggested that both phase separation and RNA-binding were necessary for activity. Next, we tested FUS with reduced RNA-binding by substituting serine for arginine residues in one (RS1, RS2, RS3) or all (RS4) the RGG domains ( Figure 4A) (13). RS1 and RS2 were previously found to retain most their RNA-binding affinity measured by EMSA but lost activity in cells by 2-fold.
Addition of RS1 or RS2 could increase transcript yields by 2-fold ( Figure 4D) and did not reduce hybrids at 20 minutes ( Figure 4E). RS3 increased the transcript T7 Pol produced by >2-fold, while decreasing RNA:DNA hybrids a small but significant amount (p < 0.05, student t-test assuming equal variances, Figure 4D-E). RS4 was previously shown to have >30-fold lower affinity for RNA than FUS, with no RNA-binding observed in cells (13). RS4 activity on T7 Pol transcript levels equalled that of RS1 and RS2, and a small reduction of hybrids was observed (Figure 4D-E).
We next investigated transcription assay conditions for evidence of phase separation by FUS and RS proteins. To preserve conditions close to the transcription assays while aiding the detection of condensates by microscopy, transcription assays were incubated 24 hours at room temperature.
Visible condensates up to 0.5 µm in diameter were found for FUS, RS1, RS2, and RS3 proteins (2 µM), but no RS4 condensates were observed (Supplemental Figure 4A). Conditions also tested were incubation at 4 ºC, increased protein concentrations (20 µM), and with the addition of RNA.
While FUS formed many condensates of ≥1 µm in diameter, none were seen for RS4 (Supplemental Figure 4B and 4C). FUS at high concentrations (85 µM) formed condensates large enough to clearly show liquid characteristics, but phase separation by RS4 (100 µM) remained undetected (Supplemental Figure 4D). Since condensate sizes were invisible to microscopy at early timepoints for the transcription assay, RS4 may have reached this undetectable level in phase separation but continued no further. We therefore concluded that RS4 was less competent to phase separate than FUS and other RS proteins.

FUS prevents R-loops in human cells
We next asked whether the FUS activity found in vitro was also present in cells. If true, a measurable increase in R-loops may result from a knockdown of FUS protein. We used a next-generation sequencing approach to quantify R-loops enriched by the S9.6 antibody (DRIP-seq) (27,48,49). We used western blot assay to verify protein levels were reduced in HEK293T/17 cells by transfection of an siRNA targeted for FUS (siFUS) relative to an siRNA control of scrambled sequence (SCR) (Supplemental Figure 5A). Half of each lysate was treated with RNaseH as a negative control. Dot blot assays with the S9.6 antibody confirmed the presence of R-loops in eluted DRIP samples (Supplemental Figure 5B). Published real-time PCR controls also indicated R-loop enrichment by comparison of signals for R-loop positive gene promoters, RPL13A and TFPT, to those of RNaseHtreated negative controls and an R-loop negative promoter, EGR1 (Supplemental Figure 5C) (27).

Figure 5. Knockdown of FUS increases R-loop abundance in human cells. DRIP-seq was performed on HEK293T/17 cells with knockdown of FUS protein, siFUS, or negative control, SCR. (A)
Normalized DRIP-seq signal profiles at human genes were averaged and plotted for siFUS (red) or SCR (blue) samples. Gene lengths were scaled to align transcription start (TSS) and end (TES) sites. The largest R-loop signals were found at rRNA genes in siFUS treated samples.
We immunoprecipitated DNA fragments and generated libraries for sequencing. DRIP-seq analysis revealed R-loop enrichment at transcription start sites (TSS) for expressed genes ( Figure   5A). For samples with FUS knockdown, R-loop signals measured at promoters were increased by >2fold in siFUS samples for 29% (N=5260) of protein coding genes (N=18188) (Figure 5B). For the representative and well-expressed GAPDH, increased R-loop signals overlapped previous reports of FUS enrichment determined by ChIP-seq ( Figure 5C) (7). The largest increase in R-loop signals was found in rRNA genes ( Figure 5D). Since rRNA genes are transcribed by RNA Pol I, this result was consistent with the prediction for FUS activity to extend to any RNA polymerase. More peaks were detected in R-loop signals from siFUS (n=16184) relative to SCR (n=5046) samples. FUS knockdown also preferentially increased R-loop signals at sites in DNA with high GC content ( Figure 6A). The greatest number of R-loop peaks shifted from intergenic and satellite DNA in SCR samples to protein-coding genes in siFUS samples (Figure 6B). A representative simple DNA repeat affected by FUS knockdown was seen in the CAG-repeat of the ATXN7 gene, which when expanded can cause neurodegenerative disease (Figure 6C) (50). FUS knockdown also caused Rloops to increase at retrotransposons (n=2036). The most affected were SINE retrotransposons (n=1237), with nearly half (48%, n=987) being the SINE-VNTR-Alu (SVA) sub-type ( Figure 6D) (51).
We noted that SINE elements are expected to be transcribed by RNA Pol III. Inspection of other RNA Pol III genes found R-loop signals increased in an 5S rRNA gene cluster on chromosome 1 (Supplemental Figure 6B). R-loops did not appear increased at RNA Pol III genes RMRP, RN7SK, or tRNA gene clusters found in chromosomes 1 and 6 (Supplemental Figure 6C).
In summary, DRIP-seq analysis revealed FUS activity did suppress R-loops in HEK293T/17. Enrichment in GC content was notable due to the previous reported preference of FUS to bind GCrich RNA sequences. We concluded that the characteristics of FUS activity on R-loops in human cells matched expectations based on that seen in T7 Pol run-off transcription assays.

DISCUSSION
Here we explore a mechanism for a FUS activity on nucleic acid structures and transcription that did not involve binding to the polymerase. This activity did require binding to RNA and was correlated with competency for phase separation. Our working model is that FUS binding prevents hybridization of nascent RNA transcripts to their complementary DNA template. DRIP-seq analysis showed in human cells that FUS prevents R-loops at expressed genes and other loci throughout the genome. This study highlights the role R-loops can have for biology and disease. Its results also offer new insights that unite common themes to FUS activities seen in and out of cells.
FUS has been found to influence the activity of all three nuclear RNA polymerases (3,6). We add to this list the RNA polymerase of T7 phage virus. This surprising result appears to be allowed because the activity does not involve binding to the polymerase (Figure 1D). The in vitro activity was seen at micromolar protein concentrations for FUS, which is within the range of FUS concentrations (4 to 8 µM) measured in the nucleus of cells (4,52,53). The largest effect of FUS found in cells was at rRNA genes, which may be explained by the high activity of RNA Pol I relative to other polymerases that can produce and sustain high levels of R-loops detected by DRIP-seq (54,55). Evidence for FUS activity on RNA Pol III was suggested by R-loops detected near retrotransposons and 5S rRNA genes.
Interestingly, the FUS activity on R-loops at RNA Pol II promoters matched well with previous reports of FUS localization on chromatin measured by ChIP-seq (7). This may indicate that FUS activity is most pronounced early during transcription at regions that span promoters and pause sites (7,11).
However, if R-loops found at the simple DNA repeats in gene bodies were formed by the nascent transcript originating from the distant mRNA promoter, it could be argued that FUS activity extends throughout the gene. The FUS knockdown also produced R-loops at sites not previously reported to have detectable FUS enrichment, such as retrotransposons.
Like other activities of FUS, RNA-binding was needed to prevent R-loops (14). FUS is implicated in RNA processing factor recruitment to RNA Pol II and nascent transcripts (9,12,56). Failure to recruit snRNPs, SR proteins, and hnRNPs has been reported to give rise to R-loops (48,(57)(58)(59)(60)(61). These effects could also be seen during T7 Pol transcription, which further supports the model interpreted from this study (60). This study has highlighted a feature that competing RNA or DNA could not interfere with activity ( Figure 2C, Supplemental Figure 2B-C). At least two considerations tie this observation to cell biology. First, high concentration of RNA is expected in the nucleus. Nascent transcripts make more secondary structure available for binding that resemble substrates preferred by FUS determined by in vitro studies (62)(63)(64). Once transcription is complete, many secondary structures become inaccessible as RNA forms tertiary structure, which is also true of RNAs in our competition assays with tertiary structures that are known from x-ray crystallography (25,62). Second, FUS concentrations measured in cells are >100-fold above the binding KD of FUS measured for a wide range in RNA substrates (13,15). While amount of FUS produced by cells may appear excessive for binding to RNA, it provides ideal conditions for phase separation (4). Therefore, the phase separation component to FUS activity has likely imposed a helpful condition to counter the diluting effect of RNA competing for FUS in the cell.
An ability to phase separate correlated to activity on transcription and R-loops. FUS phase separation has been hypothesized to recruit RNA Pol II to gene promoters, but recruitment is irrelevant to activity on T7 Pol, which FUS does not bind (3,14,17). Instead, phase separation likely increases FUS affinity for sites adjacent along the RNA, providing FUS an advantage to compete RNA from its nearby DNA template. Another aspect of phase separation considered by this study was the striking reduction of this activity in RS4. Two models may reconcile the observations of high phase separation activity for the LC domain, while FUS activity is limited without RNA (14). In one model, FUS C-terminal domains bind and sequester the LC domain whenever RNA is not bound (65). In a second model, the bulk of full-length FUS adds a steric hindrance to phase separation that must be overcome by the concentrating effect of binding to RNA (66)(67)(68). A challenge to the second model is that fusion to a large protein, MBP, affects LC domain activity less than might be expected, and phase separation activity seen in the RGG/RG domains would be predicted to add to activity, not subtract (14,69). If the 25 arginine residues replaced in RS4 formed intramolecular interactions with the LC domain, phase separation activity would be predicted to increase. With so little evidence for activity in RS4, future studies of how FUS activity is controlled might find reason to prioritize the surface residues of the RRM and ZnF domains, as well as the few RGG/RG domain residues not changed in RS4 (13,47,70).
R-loops play important roles in cell biology and disease (34,40,71). A long-standing paradigm has been that R-loops negatively impact transcription, but this assumption has been challenged by evidence of R-loops providing necessary effects in cells (34,37,72,73). If the latter proves correct, cells must have some mechanism to turn away FUS activity. Our study did not find how FUS activity is sustain past 20 minutes of transcription, but RNA Pol II transcription in cells has been found to occur as bursts (74)(75)(76). This provides periodic rests in transcription of 1 to 6 minutes, which may reset FUS activity and avoid the decreases observed in our assays. Finally, if evidence continues to tie DNA instability in ALS or Ewing sarcoma to R-loops, the significance of this FUS activity should rise above others that are not defective in portions of disease without FUS pathology (34,48).
To summarize, these results can improve future research by rephrasing the questions studied to be more specific. For example, if the activity described here extends to any RNA polymerase, can endogenous FUS act on viral or mitochondrial polymerases? The loss of FUS function produced a large increase in R-loops at rRNA genes. Can this lead to a depletion of rRNA and harm translation?
Why do R-loops appear at retrotransposons? Is it a backup mechanism to preserve silencing or merely reflect that protection against transposon activity was lost? Concluding with a question about RNA Pol II, do the R-loops prevented by FUS alter splicing or transcription termination, and what relationship do these R-loops have to RNA Pol II phosphorylation? Some of the questions rephrased are longstanding because they are difficult to address through unambiguous experiments (3,7,77,78).
If understanding the FUS activity studied here can lessen challenges to investigate other FUS activities, these future studies should provide important new insights on the mechanisms that regulate gene expression, drive neurodegenerative disease, and prevent childhood cancers.

SUPPLEMENTARY DATA
Supplementary Data are available at NAR online.      The largest R-loop signals were found at rRNA genes in siFUS treated samples.