An enhancer RNA recruits MLL1 to regulate transcription of Myb

SUMMARY The Myb proto-oncogene encodes the transcription factor c-MYB, which is critical for hematopoiesis. Distant enhancers of Myb form a hub of interactions with the Myb promoter. We identified a long non-coding RNA (Myrlin) originating from the −81 kb murine Myb enhancer. Myrlin and Myb are coordinately regulated during erythroid differentiation. Myrlin TSS deletion using CRISPR/Cas9 reduced Myrlin and Myb expression and LDB1 complex occupancy at the Myb enhancers, compromising enhancer contacts and reducing RNA Pol II occupancy in the locus. In contrast, CRISPRi silencing of Myrlin left LDB1 and the Myb enhancer hub unperturbed, although Myrlin and Myb expression were downregulated, decoupling transcription and chromatin looping. Myrlin interacts with the MLL1 complex. Myrlin CRISPRi compromised MLL1 occupancy in the Myb locus, decreasing CDK9 and RNA Pol II binding and resulting in Pol II pausing in the Myb first exon/intron. Thus, Myrlin directly participates in activating Myb transcription by recruiting MLL1.


INTRODUCTION
While the vast majority of the mammalian genome is transcribed, only a small fraction of these transcripts encodes proteins.The functional relevance of most non-coding transcription remains largely unknown.RNA polymerase II (Pol II) non-coding transcripts that are greater than 200 nt in length and lack coding potential are known as long non-coding RNAs (lncRNAs).LncRNAs, much like their protein coding counterparts, can be spliced and polyadenylated, but they are biased towards two-exon transcripts that remain localized to the nucleus [1][2][3] .Nuclear-localized lncRNAs can be involved in gene regulation via interactions with chromatin remodelers, histone modifying complexes and transcription factors [4][5][6][7][8] .LncRNAs derived from enhancer regulatory regions, known as enhancer RNAs or eRNAs, have the potential ability to function together with the enhancers from which they are derived to regulate expression of target genes [9][10][11] .
Enhancers increase the transcriptional output of target genes in a cell-type specific fashion, often bridging substantial genomic distances [12][13][14][15][16][17] .Both active enhancers and genes are marked by histone H3 lysine 4 (H3lys4) methylation.This modification is deposited by the MLL/COMPASS family of histone methyltransferases including SET1A/B and four mixed lineage leukemia complexes (MLL1-4) 18,19 .MLL1 forms a large macromolecular complex with WDR5, Menin, RbBP5, ASH2L and DPY-30 and interacts with the basic transcription machinery, including RNA Pol II 20 .MLL1 selectively binds non-methyl CpG DNA through its CXXC domain 21 .Additionally, MLL1 adaptor subunit WDR5 promotes the recruitment of MLL1 to genomic targets at a subset of genes 22 .The MLL1 complex is recruited by the lncRNA HOTTIP through direct binding with WDR5, establishing H3K4me3 deposition and driving HOXA gene transcription, while HoxBlinc recruits both Set1 and MLL complexes to hoxb genes through the SET domain. 23,24.
The Myb proto-oncogene encodes c-MYB (hereafter MYB), a critical hematopoietic regulator of cell proliferation and differentiation 25 .MYB is a repressor of human fetal hemoglobin production 26 .Given that elevation of fetal hemoglobin in adults moderates the symptoms of Sickle Cell Disease and β-Thalassemia, MYB is a potential target of therapeutic manipulation.
Myb is regulated by microRNAs and by a series of enhancers distributed over more than 100 kb between Myb and the adjacent upstream Hbs1l gene in mouse and human [27][28][29][30] .In the mouse, five enhancers, -36, -61, -68, -81 and -109 kb, with respect to the Myb transcription start site (TSS), establish proximity with the Myb promoter in an active chromatin hub 28 .Repression of Myb, which is required for terminal differentiation of erythroid cells, is accompanied by loss of these contacts.The enhancers are occupied by the LDB1 transcription factor complex, which mediates enhancer looping, and reduction of LDB1 using an shRNA compromises formation of the Myb enhancer hub 31,32 .However, how chromatin looping and transcription activation at the Myb locus are regulated remains unknown.
Long non-coding RNAs have been linked to erythropoiesis and the regulation of numerous erythroid genes, including the adult β-globin and fetal γ-globin genes [33][34][35][36][37][38] .We identified a novel lncRNA derived from the murine -81 kb Myb enhancer termed Myrlin for Myb enhancer long intergenic non-coding RNA.The Myrlin transcript was not required for formation of the Myb enhancer hub.However, Myrlin loss reduced MLL1/WDR5 recruitment in the Myb locus and decreased CDK9 and RNA Pol II occupancy.Furthermore, Myrlin loss resulted in pausing of RNA Pol II within the Myb first exon/intron.These results tie the Myb locus lncRNA Myrlin to the detailed mechanism of Myb transcription activation and suggest novel avenues that could become therapeutic targets for increasing HbF in β-globin hemoglobinopathies.

The -81 kb murine Myb enhancer contains the TSS for a spliced, long intergenic noncoding RNA
The murine Myb-Hbs1l intergenic region contains previously characterized regulatory enhancers 28 .In MEL cells and in primary erythroid cells, the five enhancers, located -36, -61, -68, -81 and -109 kb upstream of the Myb promoter, are occupied by the LDB1 complex that includes DNA binding transcription factors GATA1 and TAL1, bridging protein LMO2 and looping protein LDB1 (Figure 1A) 28 .ChIP-seq and RNA-seq data for uninduced MEL cells indicates that RNA Pol II occupies several of the enhancer sites, but active RNA transcription is only observed at the -81 kb enhancer (Figure 1A) 28 .
Rapid amplification of cDNA ends (5' and 3' RACE) revealed an unannotated 2-exon transcript at the Myb -81 kb enhancer that exists as two spliced isoforms with a single intron spanning more than 20 kb (Figure 1B, C).The primary transcript is 809 nt with a minor 357 nt shorter form attributable to early termination in exon 2. Stranded RNA-seq from induced and uninduced MEL cells publicly sourced from ENCODE indicates that Myb and the -81 kb transcript are divergently transcribed from opposite DNA strands (Figure S1).The transcript has very low coding potential according to the Coding Potential Assessment Tool (CPAT) 39 (Figure 1D).Thus, the transcript qualifies as a long non-coding RNA and its low abundance (about 10-fold lower than Myb) is consistent with that of an enhancer RNA 40 .
We named the transcript Myrlin (Myb enhancer long intergenic non-coding RNA).In uninduced MEL cells Myrlin is primarily nuclear localized, consistent with a potential role in gene regulation (Figure 1E).Expression of Myb decreases upon erythroid cell maturation, which is mirrored by decreases in both Myb and Myrlin upon differentiation of MEL cells by DMSO (Figure 1F).Like Myb, Myrlin is expressed at very low levels in E8.5 yolk sac primitive erythroid cells and then more robustly in E14.5 fetal liver definitive erythroid cells 41 (Figure 1G).These findings show that expression of Myb and Myrlin is coordinately regulated in a developmental stage-specific fashion in murine erythroid cells and raises the possibility that Myrlin may play a role in regulating Myb expression.

Deletion of the Myrlin TSS reduces Myb expression
JASPAR motif analysis (http://jaspar.binf.ku.dk/) 42 identified a TATA box located -25 nucleotides upstream of the 5' end of Myrlin as determined by RACE, and a GATA1 binding motif, site of LDB1 complex occupancy, located -56 nucleotides upstream (Figure 2A, Figure S2).This organization is consistent with the finding that most mouse erythroid-expressed non-coding RNAs are transcribed from conventional promoters regulated by known transcription factors and are regulated by similar Pol II release mechanisms 35,43 .To investigate a role for Myrlin in Myb expression, we generated several MEL cell lines with small deletions that were designed to target the transcription start site (TSS) of Myrlin (Figure 2A, Figure S2).Three different mutations were obtained all of which reduced Myrlin and Myb transcription to varying degrees in uninduced MEL cells (Figure 2B).We chose for further study the 17 base pair deletion (Δ17, hereafter ΔTSS), which removes most of the sequence between the TATA box and the initiator element, and results in the strongest reduction of Myrlin and Myb.
To begin to characterize the impairment of Myb transcription ΔTSS cells, we performed ChIP to detect the occupancy of RNA Pol II.Compared to a control MEL cell line generated with a plasmid lacking an sgRNA, occupancy of Pol II in ΔTSS cells was reduced at the Myb promoter and at the -81 kb enhancer/Myrlin TSS as well as at the other enhancer sites (Figure 2C).TBP ChIP revealed that only the -81 kb enhancer was occupied by this member of the pre-initiation complex and in ΔTSS cells there was a marked reduction (Figure 2D).In accordance with results showing that fetal γ-globin transcription increases when Myb regulatory microRNAs are reduced in human cells 27,[44][45][46] , after differentiation of ΔTSS MEL cells, there was a several fold increase in murine embryonic βh1 globin transcription (Figure 2E), consistent with Myb reduction.These results support the idea that Myrlin, transcribed from the -81 kb Myb enhancer locus, is a positive regulator of Myb transcription.

Long-range Myb promoter and enhancer contacts are reduced after Myrlin TSS deletion
Myb expression is regulated by long-range interactions between the Myb promoter and LDB1bound enhancers within the Myb-Hbs1l intergenic region 28 .A CTCF site 30 kb downstream of the Myb promoter also participates in the Myb enhancer hub, likely through direct interaction between CTCF and LDB1 47 .Moreover, the disruption of this enhancer hub underlies Myb downregulation during erythroid differentiation 28 .We used chromatin conformation capture (3C) to determine whether ΔTSS influenced contacts between Myb and its intergenic enhancers.
Compared to control cells, ΔTSS cells displayed reduced interaction frequency between the Myb promoter and enhancers, which was particularly evident at the -36 kb and -81 kb enhancer sites and the -30kb CTCF site (Figure 3A).
The reduction of interaction frequency between Myb and its multiple enhancers by ΔTSS deletion closely resembles the reduced interactions observed upon LDB1 knock down using an shRNA in MEL cells 28 .To investigate further, we carried out ChIP experiments to determine the occupancy of the LDB1 complex after reduction of Myrlin in ΔTSS cells.We observed that diminished long-range interactions in ΔTSS cells correlated with reduced LDB1 and TAL1 across the enhancers but reduced GATA1 occupancy was only notable at the -81 and -36 enhancers (Figure 3B-D).The H3K27ac mark, which indicates active enhancers, was not significantly affected in ΔTSS cells (Figure 3E).Together, our results show that the decrease in Myb expression upon deletion of the Myrlin TSS is accompanied by reduced LDB1 complex occupancy across the Myb enhancers and Myb enhancer hub disruption, although the enhancers remain in a potentially active state, retaining the H3K27ac mark.

CRISPRi for Myrlin affects Myb transcription but not enhancer interactions
We next employed CRISPRi as an alternative approach to reduce the Myrlin transcript without altering the sequence context at the Myb -81 kb enhancer.Catalytically dead CAS9 (dCAS9) retains the ability to target DNA and can be an adaptable block to transcription elongation when fused to a KRAB repression domain.Using two different gRNAs to target dCAS9-KRAB to Myrlin exon 1, we observed a 50% to 60% reduction in Myrlin transcription leading to a similar drop in Myb transcription as observed in the Myrlin ΔTSS deletion (Figure 4A).
ChIP localization of histone modifications at Myb enhancers revealed H3K9me3 at the -81kb enhancer after Myrlin CRISPRi, a signature heterochromatin mark of KRAB-mediated repression (Figure 4B).There was no change in H3K27ac at any enhancer after Myrlin CRISPRi, (Figure 4C), similar to what we observed after the Myrlin ΔTSS deletion.However, in contrast to broad reduction of LDB1 at enhancer sites after Myrlin ΔTSS deletion, LDB1 enhancer occupancy was not significantly reduced by Myrlin CRISPRi (Figure 4D).In addition, 3C experiments revealed relatively little change in interaction frequency between Myb and its enhancers after Myrlin CRISPRi compared to a control clone generated with a non-targeted dCAS9-KRAB vector (Figure 4E).Thus, downregulation of Myb after CRISPRi silencing of Myrlin does not involve loss of the Myb enhancer hub, separating looping and transcription activation at this locus.We conclude that Myb downregulation after Myrlin CRISPRi silencing does not require loss of the Myb enhancer hub, raising the possibility that Myrlin has a direct role in Myb transcription activation.

Myrlin maintains H3K4me3 in the Myb locus through MLL1-WDR5
To further explore the role of Myrlin in Myb transcription activation, we focused on the Myb promoter and first exon/intron, which contain a CpG island that is highly enriched for H3K4me3 when Myb is active (Figure S3).ChIP-qPCR revealed strong H3K4me3 enrichment across these sequences, which was reduced following Myrlin transcriptional repression by Myrlin CRISPRi (Figure 5A).The Mixed Lineage Leukemia 1 (MLL1) complex has been shown to deposit H3K4me3 modifications at the Myb locus and activate transcription 48 .Therefore, we carried out ChIP for MLL1 complex components MLL1 and WDR5.We found a strong decrease in both the MLL1 and WDR5 enrichment at the Myb locus after Myrlin CRISPRi (Figure 5B, C).
Interestingly, HOTTIP long non-coding RNA targets the MLL1/WDR5 complex to promoters of HOXA genes to facilitate gene expression but currently no known eRNAs have been associated with MLL/WDR5 complex in the context of gene regulation 23 .
The results so far suggest that Myrlin may directly participate in activation of Myb transcription through recruiting MLL1.To further investigate, we performed Myrlin RNA-ChIP and found that the Myrlin transcript interacts with the MLL1 component WDR5 (Figure 5D).RNA-ChIP also detected interaction that were not statistically significant between Myrlin and MLL1 and with Menin, another component of the MLL1 complex.Therefore, we performed RNA pull down using biotinylated Myrlin and blotted against these components of the MLL1 complex.The results confirmed interaction of Myrlin with WDR5 and further revealed that Myrlin could pulldown complex components MLL1 and Menin.Biotinylated Myrlin did not pull down SET1a or SET1b, which are members of different COMPASS-like complexes, nor was there any interaction with Tubulin (TUB), which served as a negative control (Figure 5E).These results strongly support that Myrlin plays a role in MLL1 complex recruitment to promote H3K4me3 deposition and subsequent transcription of Myb and that Myrlin loss compromises this series of events and Myb transcription activation.

RNA Pol II pausing and CDK9 occupancy are affected by CRISPRi of Myrlin
MLL1 complexes can maintain target gene expression through regulating both transcriptional initiation and elongation.MLL1 loss results in loss of the CDK9, the protein kinase subunit of pTEFb that confers phosphorylation on Pol II ser2 to drive transcription elongation in hematopoietic cells 49,50 .Previous studies have shown that inhibition of CDK9 strongly reduced transcription elongation through the Myb gene body and resulted in pausing of Pol II within the Myb first intron 28,51 .Using ChIP-qPCR, we found that CDK9 recruitment across the Myb We calculated the pausing index of Pol II Ser5P from the longer fragment data based on the Ser5P ratio at the TSS (exon1) and in the gene body and found about a 3-fold higher pausing index in Myb after Myrlin CRISPRi compared to controls (Figure 6G).Similar results were obtained for Pol II pausing in Myb using RNAseq peaks without separation by fragment size (Figure S4A, B).These results connect Myrlin to the mechanism by which transcription of Myb is regulated and suggest that Myrlin, through MLL1, may help to recruit or stabilize CDK9 at Myb, which is necessary for efficient RNA Pol II elongation through the gene.

KLF1 interacts with Myrlin transcripts within the Myb active chromatin hub
What could be the basis of Myrlin local function in Myb transcription?Myrlin interaction with MLL1 at the actively transcribed Myb promoter raised the question of whether Myrlin is retained within the enhancer hub by tethering to the -81 kb enhancer, adjacent to its transcription start site.We performed ChIRP experiments to determine Myrlin localization in the Myb locus.Two sets of Myrlin probes (odd and even probes) can specifically capture the Myrlin transcript (Figure 7A).We found prominent peaks for Myrlin at its two exons and at the -81 kb enhancer site but not at other sites across the locus (Figure 7B).Since the -81 kb enhancer is close to Myrlin exon 1, the signal at that site could have been contributed to by cross-linking of Myrlin bound to exon1.Therefore, we asked whether Myrlin interacted with -81 kb-bound transcription factors, specifically KLF1, which uniquely binds to the -81 site among the enhancers.KLF1 is required for full Myb transcription activation 28 .Moreover, transcription factors have recently been documented to commonly bind RNAs 53 .Indeed, RNA ChIP for KLF1 revealed interaction with Myrlin, which was confirmed by biotinylated RNA pull down (Figure 7C, D).ChIP-qPCR confirmed occupancy of KLF1 at the -81 kb Myb enhancer but not at -68 kb, as expected (Figure 7E).KLF1 occupancy was not affected by Myrlin CRISPRi.We conclude that KLF1 may contribute to the localization of Myrlin within the Myb enhancer hub.
In summary, structural components contributing to Myb-enhancer interactions within the hub, including the LDB1 complex and KLF1, are present and the hub forms normally when Myrlin is lost through CRISPRi (Figure 7F).However, in the absence of Myrlin, recruitment of MLL1/WDR5 is reduced, resulting in poor binding of CDK9 and RNA Pol II in the Myb locus.
Moreover, RNA Pol II pausing within the Myb first exon/intron and failure to complete Myb transcripts are observed upon Myrlin loss.We conclude that the Myrlin eRNA has important recruitment functions for MLL1 complex that are required for Myb transcription activation beyond formation of the Myb enhancer hub.

DISCUSSION
Hundreds of long non-coding RNAs are expressed during erythropoiesis but evidence for their function remains anecdotal [33][34][35]37 . We dentified Myrlin, an eRNA, transcribed from the -81 kb Enhancer interdependence has been observed at the IgK super enhancer cluster where deletion of one enhancer affects interactions among the others 56,57 .This model comports well with the idea that extrusion of chromatin loops through the action of cohesin complexes and CTCF will bring many points within an extended locus into close proximity where their interactions may be stabilized by specific transcription factors 58,59 .Such a hub of interactions and transcription factor density may facilitate accumulation of transcriptional components such as RNA Pol II, possibly involving liquid-liquid phase separation, favoring transcription 14 .
The precise role of Myrlin in Myb enhancer hub formation, if any, remains to be clarified, since the LDB1 complex was lost from the enhancer sites in the ΔTSS cells, which is well documented to compromise LDB1 enhancer looping in this locus 28 .Intriguingly, LDB1 binding, and the hub formation were not affected when Myrlin transcription was repressed by CAS9-KRAB, which did not affect LDB1 complex occupancy.It remains unclear why LDB1 is lost from the enhancer sites in ΔTSS cells but not after Myrlin CRISPRi.One possibility is that the 17 bp deletion of the TSS, near the LDB1 complex motif at the -81 kb enhancer, distorts the local chromatin architecture sufficiently to reduce LDB1 binding, which then destabilizes binding at the other hub enhancers, and looping.Regardless, losing Myrlin under the circumstances of KRAB repression did not have this effect on LDB1 binding yet still resulted in Myb downregulation, separating looping from transcription activation.Looping in the absence of transcription has been observed at the β-globin locus in erythroid cells and, more generally, in leukemia cells after drug treatment to inhibit BET proteins [60][61][62] .These data support the idea that enhancer looping mechanistically precedes transcription activation.
To explore the difference between looped and transcriptionally active Myb loci versus inactive looped loci, we localized RNA Pol II and found a decrease in Pol II recruitment to the Myb promoter after Myrlin CRISPRi and significant pausing in intron/exon 1. Pol II pausing is correlated with transcriptional repression during terminal erythroid repression 63 .Previous work had suggested a Pol II pause region in Myb intron 1 that was proposed to correspond to either a stem-loop forming sequence followed by a poly (dT) tract 1.7 Kb downstream of the Myb TSS or to a CTCF site about 2 Kb downstream 28,51 .These groups reported that transcription elongation through Myb was inhibited in erythroid cells or in breast cancer cells by CDK9 inhibitors.Our high-resolution CUT & TAG results for Pol II Ser5 and Pol II Ser3 showed localization of paused transcripts across Myb exon 1 and into intron 1, likely encompassing the sites previously suggested as pause sites.We were able to show that Myrlin functions to recruit or stabilize CDK9 in the Myb exon/intron1 Pol II pause region.It has been suggested that interactions of an enhancer with its target promoter can stimulate Pol II pause release (Ghavi-helm 2014, Rahl 2010) and that this property may be related to CDK9 activity [64][65][66] .Myrlin provides a link between an eRNA and CDK9 at a target gene.
We observed that Myrlin interacts with several subunits of the MLL1 complex, including MLL1, WDR5 and Menin, and that Myrlin is important for MLL1 occupancy in the Myb locus and for H3K4me3 modification at the Myb promoter CpG island.Promoter proximal H3K4me3 has recently been linked to RNA Pol II pause release 67 .Reduced MLL1 may underlie loss of CDK9 at the Myb locus, as CDK9 reduction was reported on a global scale after deletion of MLL1 in hematopoietic stem and progenitor cells 49 .In this scenario, the Myrlin eRNA becomes important for Myb transcription after the enhancer hub is formed.We propose that Myrlin interacts with MLL1 and the CDK9 component of pTEFb to recruit RNA Pol II to the locus and assure efficient elongation through the Myb exon/intron 1 pause region.After Myrlin CRISPR, MLL1 and CDK9 are reduced, and Pol II pausing reduces Myb transcripts.The above putative functions of Myrlin would require its localized presence in the Myb locus/enhancer hub.We explored the potential role of KLF1 in such localization since it binds uniquely to the -81 enhancer where the Myrlin eRNA is transcribed.Indeed, we documented interaction of Myrlin and KLF1 by biotinylated RNA pull down.
One of the most promising strategies for treating Sickle Cell Disease and β-Thalassemia is reactivation of fetal hemoglobin production in erythroid cells of adult patients.GWAS revealed an association between single nucleotide polymorphisms (SNPs) in the Myb ─ Hbs1l intergenic region, encompassing the several Myb enhancers, that reduce Myb gene expression and elevate fetal hemoglobin in human adult erythroid cells 26,29,30 .Two of these SNPs, i.e., rs66650371 and rs77755698, are located within the LDB1 complex GATA1/TAL1 binding peak at the human -84 kb Myb enhancer, which is the homologue of the murine -81 kb enhancer 29,68 .These SNPs reduce LDB1 binding to the -84 kb enhancer and decrease interaction frequency with the MYB promoter 29 .The effect of the SNPs on interactions of the other enhancers with MYB was not tested in this work but our results suggest that overall formation of the hub is likely affected.
Recently, a 1,283 bp non-coding RNA, HMI-LNCRNA, was reported to arise from the -84 kb human MYB enhancer 36 .Thousands of human lncRNAs have homologues in other species with similar expression patterns but low sequence conservation 35,69 .However, BLATN sequence comparison of HMI-LNCRNA to the mouse genome revealed a conserved 'patch' of 378 nt (85.8% homology) shared with the 5' end of Myrlin, something commonly observed for these poorly conserved lncRNAs 69 (Figure S5).Within this homology patch lie the GATA1/LDB1 complex binding site that mediates looping to the Myb promoter in human and mouse erythroid cells, the -81 kb KLF1 binding site and the Myrlin TATA box and first exon.Thus, it seems likely that the pause release function of Myrlin at the Myb gene may be conserved between species.While any Myb regulatory function ascribed to Myrlin in mouse cells remains to be established for the related transcript in human cells, the Myb transcription mechanisms participated in by Myrlin suggest the possibility of their utility as targets to increase HbF production to ameliorate the severity of hemoglobinopathies.

AUTHOR CONTRIBUTION
AD and IK conceived the project; IK, JK and BL supervised experiments.LFD, JK and MJM performed experiments.LFD, JK and AD wrote the paper; all authors edited the paper.

Mice and ethics statement
Mouse protocols were approved by the NIDDK Animal Care and Use Committee in accordance with AALAC specifications.Yolk sacs from E8.5 and fetal livers from E14.5 were dissected and washed in phosphate-buffered saline.

CRISPR-Cas9 genome editing of MEL cells
CRISPR gRNAs were designed using GeneTargrter (http://genetargeter.mit.edu/)(see Table S1 for gRNA sequences).gRNAs targeting the Myrlin transcription start site were cloned into the CRISPR-Cas9 and gRNA expression vector pSpCas9(BB)-2A-GFP (PX458) (gift from Feng Zhang, Addgene plasmid #48138) as described 70 .MEL cells were transfected with Escort IV lipid transfection reagent (Sigma-Aldrich L3287) according to the manufacturer's instructions.Stable MEL cell clones expressing dCAS9-KRAB were generated using Lenti-dCAS9-KRABblast (Gift from Dr. Gary Hon, Addgene plasmid #89567).2 x 10 6 of MEL cells were suspended in 100ul of High-Performance Electroporation Solution (BTXpress, 45-0801) with 5-10 µg of plasmid DNA and electroporated with the Gene Pulser Xcell System (Bio-Rad) using 2 pulses at 200 V for 5 ms.Cells were diluted with 100 µL of pre-warmed media and transferred to 2 mL of media in a 12-well culture dish.Cells were selected in 10 µg/mL Blasticydin (Gibco, A1113903) for one week and plated at limiting dilution to isolate clones.Clonal lines were checked for production of S. pyogenes dCAS9 by RT-qPCR.dCAS9-KRAB MEL cells were electroporated as above with gRNAs targeting the Myrlin transcription start site cloned into LentiGuide-puro (gift from Feng Zhang, Addgene plasmid #52963).Cells were selected in 10 µg/mL Blasticydin and 1 µg/mL puromycin (Gibco, A1113803) for one week and plated at limiting dilution to isolate clones.Transfected cells were validated for expression of Myrlin by RT-qPCR.

5' and 3' rapid amplification of cDNA ends (RACE)
RACE was performed using the FirstChoice RLM-RACE kit (ThermoFisher Scientific, AM1700) following the manufacture's protocol.Total RNA from MEL cells was extracted and reverse transcribed using the 3' RACE adapter and the sequence of interest was amplified by nested PCR (3' RACE).Alternatively, a sample of the same RNA was treated with Calf Intestine Alkaline Phosphatase, then Tobacco Acid Pyrophosphatase and finally ligated to the 5' RACE adapter.De-capped adapter-ligated RNA was then reverse transcribed, and the sequence of interest was amplified by nested PCR (5' RACE).PCR products were separated on a 1% agarose gel, purified using the QIAquick gel extraction kit (Qiagen, 28704), and cloned into the pCR4-TOPO vector (Invitrogen, K457502) for sequencing.For RACE primers, see Table S1.

Chromatin conformation capture assay (3C)
Cells were cross-linked with 2% formaldehyde in PBS at room temperature for 5 minutes, Glycine weas added to a final concentration of 0.125 M and sample were incubated at room temperature for 5 minutes.Cells were washed twice with cold PBS and resuspended in lysis buffer (10 mM Tris-HCl pH 8.0, 10 mM NaCl, 0.2% NP40, proteinase inhibitor).After lysis for 30 minutes on ice, nuclei were collected by centrifugation at 800g for 10 minutes at 4°C and resuspended in 1.2 X of NEB buffer 2.1 (New England Biolabs, B7202). 1 x 10 7 nuclei were then solubilized with 0.3% SDS for one hour at 37°C followed by adding 1.8% of TritonX-100 and incubating one hour at 37°C.Chromatin was digested with 1000 U of HindIII (New England Biolabs, R0104M) overnight at 37°C.Digested genomic DNA control were taken and stored at -20°C.Restriction enzyme was inactivated using 1.6% of SDS for 25 minutes at 65°C and chromatin was ligated by adding 4000 U of T4 ligase (New England Biolabs, M0202M) in 1× T4 DNA ligase buffer (New England Biolabs, B0202S) containing 1% TritonX-100 and for 4 hours at 16°C followed by additional incubation for 30 minutes at room temperature.Chromatin was de-cross-linked with a final concentration of 100 µg/mL of proteinase K at 65°C for 6-10 hours.Then 0.5 µg/mL of RNase A was treated for one hour at 37°C.Samples were purified via phenol-chloroform extraction and ethanol precipitation.Relative cross-linking between the Myb promoter and fragments of interest was analyzed by real-time qPCR with published TaqMan probes and primers 28 .Ligation products of HindIII digested BAC DNA containing the mouse Myb-Hsbl1 intergenic region were used to determine primer efficiency.Additional primers are listed in Table S1.

Myrlin RNA cellular localization
MEL cells were washed with ice cold PBS and lysed in hypotonic buffer (25 mM HEPES, 2 mM EDTA, 0.5% Tween-20).Cytoplasmic and nuclear fractions were obtained by centrifugation at 800 g for 10 minutes at 4°C.RNA from the supernatant cytoplasmic material and the nuclear pellet were purified and RNA was reverse transcribed, and cDNA was measured by real-time qPCR.For RT-qPCR primers see Table S1.

Chromatin Isolation by RNA purification (ChIRP)
The Myrlin probes were designed using the Stellaris Probe Designer version 4.2 and synthesized with 3' Bio-TEG modification by IDT.LacZ probes were used as a negative control (Millipore, CS216572).ChIRP-qPCR was performed as described (Chu et al., 2012). 2 X 10 6 of MEL cells per IP were fixed with 1% glutaraldehyde (Sigma-Aldrich, G5882) for 10 minutes at room temperature.Glycine was added to a final concentration of 0.125 M and sample were incubated at room temperature for 5 minutes.Cells were washed three times with cold PBS by centrifugation.Cell pellets were snap frozen and thawed on ice on the day of starting the ChIRP experiment.Cell pellets were resuspended in lysis buffer (10 X the mass of pellet, 50 mM Tris pH 7.0, 1% SDS, 10 mM EDTA) with proteinase inhibitors, 0.1 U/µL Superase-in (Invitrogen, AM2694) and 100 mM PMSF, then subjected to sonication with Bioruptor (Diagenode) for 50 minutes at highest setting with 30 seconds ON, 45 seconds OFF pulse intervals.Samples were spun at 16,000 g for 10 minutes at 4°C and supernatant was transferred to a new tube.Chromatin was snap frozen and thawed on ice on the day of starting the ChIRP experiment.Chromatin is diluted in two times volume of hybridization buffer (750 mM NaCl, 1% SDS, 50 mM Tris pH 7.0, 1 mM EDTA, 15% Formamide) with proteinase inhibitors, 0.1 U/µL Superase-in and 100 mM PMSF.Pre-designed probes were separated into two groups which are even probe set (probe 2, 4, 6 and 8) and odd probe set (probe 1, 3, 5 and 7) and each probe set (100 pmol /1 mL chromatin) were added to diluted chromatin, which was mixed by end-to-end rotation for 4 hours at 37°C.Dynabeads MyOne Streptavidin C1 (Invitrogen, 65001) were washed three times in lysis buffer and resuspended in its original volume.100 µL of streptavidin beads were added per 100 pmol of probes, and samples were mixed for 30 minutes at 37°C.Chromatin bound beads were washed five times with 1 mL of pre-warmed wash buffer (2× SSC, 0.5% SDS, 100 mM PMSF) for 5 minutes at 37°C.At last wash, 100 µL of resuspended beads aliquoted for RNA isolation and 900 µL for DNA fraction.RNA was extracted with the miRNAeasy Mini Kit (Qiagen, 217004), and Superscript III First-Strand Synthesis System was used to reverse transcribe RNA to cDNA.DNA fraction was isolated by phenol-chloroform extraction and ethanol precipitation.qPCR was performed using iTaq Universal SYBR Green Supermix with the ABI 7900HT.Percent of input was calculated against input.For probes see Table S1.

RNA pulldown
Full-length Myrlin RNA was generated using the MEGAscript T7 Kit (Invitrogen, AM1333) according to the manufacturer's protocol.RNA pulldown was performed using the Pierce Magnetic RNA-Protein Pull-Down Kit (ThermoFisher Scientific, 20164) according to the manufacturer's protocol.Briefly, 1 μg of biotinylated Myrlin RNA was incubated with 1 mg of precleared protein extracted from MEL cells for 4 hours at 4°C.Following this, Streptavidin magnetic beads was added and incubated for an additional 2 hours.Finally, proteins were eluted and subjected to western blotting.

Western blot
Proteins were separated on NuPAGE 4%-12% Bis-Tris Gel (Thermo Fisher Scientific, NP0321) and transferred to a PVDF membrane using Trans-Blot Turbo Transfer Pack (Bio-RAD, 1704156).Membrane was blocked in blocking buffer (1 X TBS, 0.1% Tween-20, 5% w/v Nonfat dry milk) for 1 hour at room temperature.Primary antibodies in blocking buffer were treated overnight at 4°C.Membranes were washed with TBST three times with shaking for 10 minutes, incubated with HRP-conjugated secondary antibody in blocking buffer for 1 hour at room temperature, and washed three times in TBST.Blots were exposed to SuperSignal West Dura Extended Duration Substrate (Thermo Fisher Scientific, 34075) and scanned by Syngene PXi (Syngene).Antibodies are listed in Table S1.

RNA Polymerase II (Pol2) pausing index at Myb locus
For Pol2 abundance over specific regions and Pol2 pausing analysis, bam files were filtered for properly paired reads only with samtools view, and then converted to bedpe format using bedtools bamtobed.The subset of fragments that were shorter than 1000 bp and mapped to chromosomes chr 1-19, X, Y was extracted to generate a bed file of fragments.Subsets of bed files were generated based on the desired size ranges (total, <120 bp, 120-270 bp, >270 bp).
Fragments over genomic regions of interest were counted with bedtools intersect.The Pol2 pausing index was calculated as the ratio of the fragments per million (FPM) in the first exon of Myb over the rest of the gene body.

Statistical analysis
As indicated in the figure legends, data values reported in the figures are the mean and standard error of the mean (SEM).GraphPad Prism 8.0 (GraphPad Software) and Excel (Microsoft) were used to perform the statistical analyses.Unpaired Student's t-test was used for significance test.

FIGURE TITLES AND LEGENDS
enhancers and in Myb exon 1 and intron 1 was severely diminished after Myrlin CRISPRi (Figure6A).Subsequent ChIP-qPCR analysis of RNA Pol II localization in the Myb locus revealed decreased occupancy at several Myb enhancers but no significant difference at the Myb promoter (Figure6B).Interestingly, after Myrlin CRISPRi, RNA Pol II accumulated within the Myb gene body, particularly at Myb exon1/intron 1, which is consistent with reduction of Myb transcripts.To further investigate this result, we performed CUT & TAG for the Ser5P and Ser2P phosphorylated forms of RNA Pol II, which represent the initiating and elongating forms of the enzyme, respectively.The genome browser view in Figure6Cillustrates the reduction of both phosphorylated forms of RNA Pol II across the Myb locus after Myrlin CRISPRi.Quantitation of these data shows about a 30% reduction of Pol II Ser5P and Ser2P within the Myb gene body and reduced occupancy at each of the enhancers (Figure6D, E).To obtain increased resolution of this result, we analyzed the signals of the shorter (<120 bp) and longer (>270 bp ) Pol II Ser5P CUT & TAG fragments, which distinguishes Pol II in the pre-initiation state at the TSS and paused Pol II at peaks up-and downstream of the TSS, respectively52 (Figure6F).Myrlin CRISPRi reduced the signal for the shorter fragments, suggesting some decrease in Pol II recruitment upon Myrlin loss.At the same time, the signal associated with longer fragments, indicative of paused Pol II, strongly increased across exon/intron 1.

murine
Myb enhancer as a positive regulator of Myb.A rigorous paradigm to deduce the function of lncRNA and eRNA loci involves dissection of the functional roles of the RNA transcript versus the act of enhancer transcription, both of which can influence target gene expression and/or chromatin organization 54 .To this end, we analyzed the role of the RNA molecule itself by using Myrlin RNA ChIP and biotinylated Myrlin pull down assays to document the importance of the transcript, per se.We found that Myrlin interacts with MLL1/WDR5 to promote Myb transcription.In the absence of Myrlin, MLL1 recruitment to Myb and Myb promoter H3K4me3 modification is reduced.In addition, CDK9 is reduced at the Myb promoter and RNA Pol II accumulates excessively within the Myb first exon/intron.Thus, Myrlin is required for Pol II pause release to promote Myb transcription.In our initial experiments to reduce Myrlin transcription, we deleted 17 bp of the Myrlin TSS.This resulted in loss of the LDB1 looping complex from all the Myb enhancer sites, disruption of the Myb enhancer hub and reduction of Myb transcription.The results suggest interdependence among multiple Myb enhancers for formation of the hub.Hubs are understood from single allele interaction experiments to represent close association of all hub sequences within which multiple enhancers show preferential aggregation with each other and with the genes they regulate 55 .

Fluorescent cells were sorted 48
hours later and plated at limiting dilution to isolate clones.Clonal lines were genotyped by PCR using EmeraldAmp GT PCR Master Mix (Takara, RR310A) and target specific primers flanking the Myrlin transcription start site.Deletions were validated by sequencing.
and Stadhouders   et al., 2012 28 .Chromatin immunoprecipitation (ChIP)10 6  of MEL cells per IP were cross-linked with 1% formaldehyde in PBS at room temperature for 10 minutes.Glycine was added to a final concentration of 0.125 M and sample were incubated at room temperature for 5 minutes.Cells were washed three times with cold PBS by centrifugation.For LDB1 ChIPmentation cells were double cross-linked.MEL cells were washed three times with cold PBS with 1 mM MgCl2.Disuccinimidyl glutarate (ThermoFisher Scientific, 20593) in DMSO were added to a final concentration of 2 mM and samples were rocked at room temperature for 45 minutes.Cells were washed three times with PBS by centrifugation and cross-linked with 1% formaldehyde in PBS at room temperature for 10 minutes, Glycine weas added to a final concentration of 0.125M and sample were incubated at room temperature for 5 minutes.Cells were washed three times with cold PBS by centrifugation.

Figure 1 .
Figure 1.The Myb -81 kb enhancer is the transcription start site for the noncoding Myrlin RNA.(A) ENCODE ChIP-Seq of LDB1 complex components (GATA1, LDB1 and TAL1), CTCF, and PolII in the Myb-Hbs1l locus in MEL cells.PolyA RNA-seq is shown in MEL cells.Intergenic LDB1 complex binding sites are highlighted (grey vertical bars).(B) Nested PCR products from 5' and 3' RACE in MEL cells.(C) Poly-A RNA-seq and Ldb1 ChIP-seq tracks from ENCODE for MEL at the -81 kb enhancer and downstream Myrlin exon 2. (D) Prediction of coding potential for Myrlin and select other transcripts as determined by CPAT (see text).(E) Relative expression of Myrlin in nuclear and cytoplasmic fractions of MEL cells determined by RT-PCR.MALAT1 lncRNA and ActB provided nuclear and cytoplasmic controls, respectively.(F) Total RNA of uninduced and induced MEL cells was used to determine relative expression of Myrlin and Myb by RT-PCR.Expression was normalized to ActB.(G) Total RNA of E8.5 yolk sac cells (YS) and E14.5 fetal liver cells (FL) was used to determine relative expression of Myrlin, Myb and Hbb-b1 by RT-PCR.Expression was normalized to ActB.Error bars indicate SEM of 3 independent biological experiments.(*) P < 0.05, (**) P<0.01, (***) P<0.001 by Student's t-test.See also Figures S1.

Figure
Figure 3. Chromatin organization of the Myb locus and transcription factor occupancy is

3 .
Figure 3. Chromatin organization of the Myb locus and transcription factor occupancy is affected by reduction of Myrlin in ΔTSS cells.(A) Chromatin conformation capture (3C) interaction frequency between the enhancers found within the Myb-Hbs1l intergenic region using the Myb promoter as the anchor (black bar) observed for ΔTSS and control cell uninduced MEL cells.(B) GATA1 occupancy at Myb enhancers in control and ΔTSS uninduced MEL cell lines determined by ChIP-qPCR.(C) TAL1 occupancy at Myb enhancers in control and ΔTSS uninduced MEL cell lines determined by ChIP-qPCR.(D) LDB1 occupancy at Myb enhancers in control and ΔTSS uninduced MEL cell lines determined by ChIP-qPCR.(E) H3K27ac normalized to H3 occupancy at Myb enhancers in control and ΔTSS uninduced MEL cell lines determined by ChIP-qPCR.Error bars indicate SEM of 3 independent biological experiments.(*) P < 0.05 and (**) P<0.01 by Student's t-test.

Figure 4 .
Figure 4. CRISPRi targeting of Myrlin compromises Myb transcription but not hub formation.(A) Expression of Myb and Myrlin monitored by RT-qPCR in CRISPRi uninduced MEL cells targeted with dCas9 sgRNA1, sgRNA3 or without an sgRNA (control).(B) ChIP-qPCR for H3K9me3 across the Myb locus before and after Myrlin CRISPi in uninduced MEL cells.(C) ChIP-qPCR for H3K27ac across the Myb locus before and after Myrlin CRISPi.(D) ChIP-qPCR for LDB1 across the Myb locus before and after Myrlin CRISPi.(E) Chromatin conformation capture (3C) interaction frequency between the Myb enhancers using the Myb promoter as the anchor (black bar) observed after Myrlin CRISPRi in uninduced MEL cells targeted with dCas9 sgRNA or without an sgRNA (control).Error bars indicate SEM of 3 independent biological experiments.(*) P < 0.05 and (**) P<0.01 by Student's t-test.

Figure
Figure 6.Pol II and CDK9 occupancy within the Myb enhancer hub is affected by CRISPRi

6 .
Figure 6.Pol II and CDK9 occupancy within the Myb enhancer hub is affected by CRISPRi targeting of Myrlin.(A) ChIP-qPCR for CDK9 in Myrlin CRISPRi uninduced MEL cells targeted with dCas9/KRAB or without an sgRNA (control).(B) RNA Pol II occupancy in the Myb promoter/exon 1 region after Myrlin CRISPRi and in control cells.(C) CUT&Tag for RNA Pol II Ser5 and Ser2 phosphorylated forms.(D, E) Quantitation of CUT&TAG data showing Pol II Ser5P and Ser2P occupancy in the Myb gene body and at each of the enhancers.(F) Separate analysis of the shorter (<120 bp) and longer (>270 bp) Pol II Ser5P CUT & TAG fragments displaying Pol II Ser5 occupancy 52 .(G) Pausing index calculated for Pol II Ser5 across Myb.Error bars indicate SEM of 2 independent biological experiments.(*) P < 0.05 and (**) P<0.01 by Student's t-test.See also Figures S4.

Figure
Figure 7. KLF1 interaction with Myrlin contributes to localization within the Myb enhancer

7 .
Figure 7. KLF1 interaction with Myrlin contributes to localization within the Myb enhancer hub.(A) Myrlin RNA pull down was conducted to determine efficiency of probes for ChIRP.(B) ChIRP DNA pull down by Myrlin across Myb and the Myb enhancers.(C) RNA ChIP for KLF1.(D) Biotinylated Myrlin pull-down and blotting with KLF1 antibodies.(F) Model of the Myb locus enhancer hub.Transcription of Myb is depicted with and without the -81 kb enhancer Myrlin eRNA after Myrlin CRISPRi.Large, shaded circle represents Pol II and LDB1 transcription factor density within the Myb enhancer hub which is diminished when Myrlin transcription is reduced.Error bars indicate SEM of 3 independent biological experiments.(*) P < 0.05 and (**) P<0.01 by Student's t-test.