Bidirectional transcription initiation marks accessible chromatin and is not specific to enhancers

Robert S Young,Yatendra Kumar,Wendy A Bickmore,Martin S Taylor

doi:10.1186/s13059-017-1379-8

Robert S Young, Yatendra Kumar + Show 2 more

Open Access

https://doi.org/10.1186/s13059-017-1379-8

Copy DOI

Abstract

BackgroundEnhancers are modular regulatory elements that are central to the spatial and temporal regulation of gene expression. Bidirectional transcription initiating at enhancers has been proposed to mark active enhancers and as such has been utilized to experimentally identify active enhancers de novo.ResultsHere, we show that bidirectional transcription initiation is a pervasive feature of accessible chromatin, including at enhancers, promoters, and other DNase hypersensitive regions not marked with canonical histone modification profiles. Transcription is less predictive for enhancer activity than epigenetic modifications such as H3K4me1 or the accessibility of DNA when measured both in enhancer assays and at endogenous loci. The stability of enhancer initiated transcripts does not influence measures of enhancer activity and we cannot detect evidence of purifying selection on the resulting enhancer RNAs within the human population.ConclusionsOur results indicate that bidirectional transcription initiation from accessible chromatin is not sufficient for, nor specific to, enhancer activity. Transcription initiating at enhancers may be a frequent by-product of promiscuous RNA polymerase initiation at accessible chromatin and is unlikely to generally play a functional role in enhancer activity.

Highlights

Enhancers are modular regulatory elements that are central to the spatial and temporal regulation of gene expression
Transcription initiation is a pervasive feature of accessible chromatin We systematically explored the relationship between chromatin state, transcription initiation, and DNA accessibility by focussing on four well studied cell lines (Gm12878, HepG2, Huvec, and K562), all with (i) detailed maps of chromatin modifications assimilated into chromatin state maps [9], (ii) cap analysis of gene expression (CAGE)-based measures of transcription initiation, and (iii) DNA accessibility as measured by DNase hypersensitivity
The same pattern of bidirectional transcription initiation was seen at DNase hypersensitivity site (DHS) located in each of the other evaluated chromatin environments and even at DHSs lacking sufficient chromatin modification to be assigned to any chromatin state (Fig. 1; Additional file 1: Figures S1–S4), illustrating that this is not a specific feature of enhancer-associated DHSs

Summary

Introduction

Enhancers are modular regulatory elements that are central to the spatial and temporal regulation of gene expression. Regulatory DNA elements that positively drive gene expression at a distance [1] They are thought to be central to controlling cellular differentiation and developmental gene expression profiles, and mutations disrupting them have been associated with several Mendelian disorders [2, 3]. The FANTOM5 consortium used cap analysis of gene expression (CAGE) transcriptome data to define active enhancers and validated 67–74% of their predictions [13]. These validation rates compare to the 75% obtained when enhancers are defined by their chromatin marks alone [14] and are lower than the 87% validation rate for enhancers defined by the binding of histone acetyltransferase p300 [15]. Direct comparison of validation rates between studies based on different discovery thresholds and validation systems is challenging, so it remains to be seen whether epigenetic marks or bidirectional transcription is more specific and accurate in identifying active enhancers

Methods

Results

Discussion

Conclusion