The emerging roles of eRNAs in transcriptional regulatory networks

Abstract

Following reports by ENCyclopedia Of DNA Elements (ENCODE; GENCODE) Consortium and others, it is now fairly evident that the majority (70–80%) of the mammalian genome has the potential to be transcribed into non-protein-coding RNAs (ncRNAs). Critical to our understanding of genetic processes is the mechanism by which ncRNAs exert their roles. Accordingly, ncRNAs are shown to regulate the expression of protein-coding loci (i.e., genes) at the transcriptional as well as post-transcriptional stages. We recently reported on a widespread transcription at the DNA enhancer elements in myogenic cells. In our study, we found certain enhancer RNAs (eRNAs) regulate chromatin accessibility of the transcriptional machinery at loci encoding master regulators of myogenesis (i.e., MyoD/MyoG), thus suggesting their significance and site-specific impact in cellular programming. Here, we examine recent discoveries pertinent to the proposed role(s) of eRNAs in regulating gene expression. We will highlight consistencies, discuss confounding observations, and consider a lack of critical information in a way to prioritize future objectives.

Introduction

With the advent of high-throughput sequencing, studies have taken a comprehensive approach in cataloguing regulatory genomic elements and the transcriptome in various cell types; as well as defining the relations between the global transcriptional activity and chromatin architecture.^1-5 These studies have unraveled a complex and surprising description of the human genome that has challenged the views on the non-protein-coding compartment (equivalent to ~98% of the entire genome). Of note is the observation that the genome is pervasively transcribed with a greater proportion of long non-coding RNAs (lncRNAs), rather than protein-coding transcripts, showing cell type-specific expression.^3,5,6 Subcategorized under lncRNAs are transcripts originating from regulatory enhancer elements (i.e., eRNAs). By knockdown approach, studies demonstrate the prominence of lncRNA as well as eRNAs in regulating gene expression and cellular programming.^7-12 Genome-wide techniques are used to annotate enhancers, their connectivity, and mature transcripts. According to these studies (as they will be discussed herein), eRNAs display perplexing features, distinct from the rest of the transcriptome. These observations suggest that despite the collected data, there is much to be discovered about eRNAs that would precisely depict their molecular mechanism, including detailed biochemical characterization, processing (i.e., splicing, editing), co-factor identification, and genome-wide distribution.

Enhancers

Although we will briefly highlight key features of DNA enhancer elements, readers are referred to extensive reviews on this topic published elsewhere.^13-17 Conventionally, transcriptional enhancers were shown to increase the expression of protein-coding genes in reporter-expression assays.^18,19 Since then, endogenous enhancers, estimated to be in tens-of-thousands in metazoans, are being discovered through distinct chromatin signature, ablation, and cloning strategies. These enhancers control the expression of genes over a relatively large genomic distance, occasionally reaching megabases.^13-17 Specifically, enhancers are binding sites for multiple transcription factors (TFs) that, to a large extent, are conserved, despite evolutionary divergence at regions flanking these binding sites.^2,20,21 Furthermore, depending on their activity status, enhancer sites are more sensitive to endonucleases and are modified at distinct nucleosome residues (lysine-4 methylation and lysine-27 acetylation on histone H3), thus distinguishing them from surrounding regions and other transcribed loci.^13,16,22-24 Active enhancers are also occupied by RNA polymerase II (PolII), though this observation was originally thought to be the result of indirect connectivity to proximal regulatory regions in chromatin immunoprecipitation (ChIP) studies. Nonetheless, enhancers are also sites of transcription,^25-30 thus raising questions on the relevance of eRNAs in cellular processes.

Enhancer sites are further classified in two configurations: (1) Typical enhancers (< 1 kb) associated with housekeeping genes and (2) super-enhancers (enhancer-clusters or chromatin regulatory beacons; ranging in number from 200–1000 in certain cell types), which are generally confined near key developmental genes.^31-33 Super-enhancers (~3–50 kb) are binding sites for unusually high levels of TFs, Mediator complex, and PolII. Segments of super-enhancers are found to be transcribed, whose eRNA levels correlate with the expression of nearby genes.^8,31 As compared with typical enhancers where simple one-to-one promoter-connectivity is shown to drive gene expression,^13,34 the significance of super-enhancer modules is not clear. Yet, given the high levels of Mediator complex at super-enhancers and the prominent role of Mediator complex in transmission of transcriptional instructions and 3D organization of the genome,^35-38 it could be proposed that super-enhancers serve as genomic connectivity centers where appropriate regulatory networks are structurally organized in factories for coordinated expression. In fact, recent PolII-assisted chromatin connectivity mapping has revealed that, depending on the cellular context, SOX2 and OLIG1 loci (including their super-enhancers) are directly connected to distinct developmental networks.^6,31,32 Overall, in addition to operating as information hubs,³⁹ it may well be that super-enhancers are structural centers that stabilize the transcriptional architecture and perform as major ports for integrating developmental networks.

eRNA Synthesis and Biochemical Properties

eRNAs are detected in most cell types examined, and as mentioned above, originate from enhancers with a distinct chromatin signature.^3,8,16 Aside from this distinction, molecular components governing the transcription of eRNAs thus far appear comparable to genes, including the contribution of transcriptional machinery (PolII), Mediator complex, nuclear receptors (estrogen receptor and Rev-Erbs), and transcription factors (Klf4; Egr1; FoxA1; MyoD, p53) in driving eRNAs transcription.^{8,9,11,12,40,41} To catalog nascent eRNAs and determine their rate of synthesis, scientists have used Global Run-On with Sequencing (GRO-Seq).^11,12,40 Accordingly, nascent eRNAs contains a 7-methylguanylated cap with a rate of synthesis and levels comparable to the nearest protein-coding transcripts. Moreover, enhancers have the potential to maintain bidirectional RNA synthesis, much similar to occurrences found around transcriptional start sites (TSS) of genes.^40,42 Still, the data regarding polyadenylation at 3′ end processing of eRNAs have been less clear. A report by Kim et al. (2010) suggests that rapidly inducible eRNAs are not polyadenylated (based on RNA-Seq data and circularization experiments), whereas studies examining steady-state and/or developmentally regulated eRNAs imply that eRNAs are subject to polyadenylation.^8,10,29 Although it is quite conceivable that eRNAs differ in their post-transcriptional processing, generalization on this matter requires closer inspection as it will unravel key characteristics relevant for future studies.

There are reasons to suspect that the processing of eRNAs differs from the rest of transcriptome. One observation is that despite matching PolII occupancy signals at enhancers and genes in ChIP-Seq data, mature eRNAs are barely detectable in total RNA-Seq data sets, whereas nascent eRNAs levels (in GRO-Seq) are as high as nearby mRNAs.^8,12,29,40 Another is the difficulty in cloning regulatory RNAs (including eRNAs) for characterization (our unpublished observations). These data suggest that eRNAs are either unstable for steady-state accumulation or not readily processed for sequencing/cloning with current protocols. Either way, with advances in molecular techniques, these questions should be addressed in the near future.

Proposed Molecular Mechanism(s)

Recent discoveries support a role for eRNAs in promoting gene expression through chromatin accessibility, PolII recruitment, and enhancer-promoter contacts (EPCs). Common to these findings is the correlation between the levels of eRNAs and those of nearby mRNAs. Using RNA interference (RNAi), experiments show that targeted depletion of eRNAs results in significant reduction of nearby mRNAs in cultured cells and mice.^8-12 Furthermore, this reduction occurs at the transcriptional stage, where PolII occupancy is noticeably reduced at genes; and eRNA depletion culminates in the loss of EPCs.^8,9,12 In line with the latter observations, recent findings demonstrate that certain eRNAs directly associate with the Mediator complex to facilitate EPCs and augment mRNA transcription, whereas other experiments underscore the binding of specific eRNAs to Rad21+ Cohesin complex to stabilize chromatin looping and transcription.^9,12 These observations are consistent with co-recruitment of the Mediator and Cohesin complexes for steady-state mRNA synthesis.^37,43-45 Nonetheless, the model regarding the direct role(s) for eRNAs in shaping EPCs may not hold for several reasons. First, eRNA depletion by transcriptional inhibition indicates that PolII holoenzymes may reinforce EPCs.^12,40 Second, a simple all-or-none contact does not explain the dose-response impact of eRNAs on gene expression observed in RNAi experiments. Perhaps analogous to an internal combustion system, proximity of the gas tank (i.e., enhancer) to the engine (i.e., promoter/gene) could only dictate agility in signal transmission. A “throttle” (i.e., eRNAs), which unlocks the gate for flow of reactive elements (i.e., PolII), dictates the rate of acceleration. In this context, designated enhancers have previously been shown to open chromatin for transcriptional activity at distinct promoters.^46-49 In our recent published study, we asked whether the emanating eRNAs facilitate transcription by exposing proximal regulatory regions to PolII complex. Specifically, we focused on two eRNAs upstream of MYOD1, corresponding to a region classified as the abovementioned super-enhancer.^8,31 By RNAi, we reported that while an eRNA (i.e., ^CERNA) controls chromatin accessibility and transcription at MYOD1, another (i.e., ^DRRRNA) operates to expose regulatory regions at another region, MYOG.⁸ Certainly, similar observations regarding transcriptional activity have been reported at other enhancer/gene combinations.^9,10,12 Overall, these results demonstrate the specificity of enhancers/eRNAs and re-emphasize similarities to the “throttle” analogy, which satisfies the refractory nature of nucleosomes to transcription, role of enhancers in TSS specification, and dose-dependent enhancement of PolII occupancy and transcription.^50-52 Third, interactome data reveal that although EPCs are developmentally dynamic, the connections in transient networks (e.g., TNFα) are static and stably formed even before the induction of eRNAs.^12,40,53 Fourth, according to PolII-mediated chromatin connectivity mapping, a significant fraction of EPCs occur interchromosomally (i.e., between chromosomes), suggesting an interactome that is far more complex than a simple in cis contact afforded by eRNA-assisted looping model.⁶ Lastly, a comparison between evolutionarily divergent enhancers suggests an intricate regulatory function beyond the frequently observed higher-order chromatin configurations in metazoans.^20,54-59 Given the above considerations, enhancers/eRNAs regulate transcription by establishing chromatin accessibility and PolII recruitment; and may not be directly responsible for the formation of EPCs, while it is PolII and its associated macromolecular complexes that may define the transcriptional architecture.³⁸ These observations hint at a step-wise eRNA-mediated transcription activation events conceptualized in Figure 1.

Figure 1. Emerging roles of eRNAs in establishing chromatin accessibility and the subsequent formation of EPCs. (A) eRNA synthesis at an enhancer and its targeting to a defined regulatory region (i.e., Promoter). (B) eRNA-mediated chromatin accessibility and the subsequent recruitment of factors for transcription and the stabilization of EPCs.

Hierarchy Within Regulatory Networks

eRNA synthesis occurs prior to, and regulates the activation of genes in developmental regulatory networks. For example, chromatin marks associated with active enhancers are first observed at the core enhancer (CE) of MYOD1 before its transcriptional activation, and depletion of ^CERNA results in a significant reduction of MyoD transcript.^8,44 Similarly, an eRNA (ncRNA-a7) activates SNAI1, a gene belonging to a family of TFs with roles in mesodermal determination and epithelial–mesenchymal transition (EMT).¹⁰ These data invoke an interesting hypothesis that certain eRNAs are at the top of the hierarchy within certain transcriptional regulatory networks. If so, this notion suggests that failure to activate, or mutations within, critical eRNAs leaves limited capabilities for TFs and results in disorders.^33,60-64 Therefore, one investigative priority would be to resolve the targets of eRNAs genome-wide.

Conclusions and Perspectives

Latest data reaffirm the widespread transcription of the mammalian genomes, and that enhancers are among the transcribed regions. Thus far, evidence suggests that enhancers/eRNAs promote mRNA transcription by establishing chromatin accessibility at specified loci and over a large genomic space, resulting in PolII/Mediator/Cohesin complex recruitment and culminating in the formation of transcription networks. A question arising from these observations is “why would more complex biological systems evolve an enhancer-based regulatory system?” In the view of the positive relationship between the genome size and biological complexity, one can anticipate a superior governing and adaptive power of sequence-specific regulation rather than a more primitive TF-based system in multicellular and multilineage organisms.^21,56,65-67 This way, coordinated spatiotemporal regulation of distinct transcriptional networks is organized and fine-tuned while utilizing similar protein components to drive complexity.^{6,57-59,68-70} Notwithstanding remarkable advances made thus far, a great deal is yet to be discovered about eRNAs, including their regulatory networks, biochemical characteristics, structure–activity relationship, and precise molecular mechanisms that make them key transcriptional integrators.⁷¹

10.4161/rna.27950