Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2020 Sep 25;48(17):9550-9570.
doi: 10.1093/nar/gkaa671.

Enhancer RNAs predict enhancer-gene regulatory links and are critical for enhancer function in neuronal systems

Nancy V N Carullo  1 , Robert A Phillips Iii  1 , Rhiana C Simon  1 , Salomon A Roman Soto  1 , Jenna E Hinds  1 , Aaron J Salisbury  1 , Jasmin S Revanna  1 , Kendra D Bunner  1 , Lara Ianov  2 , Faraz A Sultan  1 , Katherine E Savell  1 , Charles A Gersbach  3 , Jeremy J Day  1   2
Affiliations

Enhancer RNAs predict enhancer-gene regulatory links and are critical for enhancer function in neuronal systems

Nancy V N Carullo et al. Nucleic Acids Res. .

Abstract

Genomic enhancer elements regulate gene expression programs important for neuronal fate and function and are implicated in brain disease states. Enhancers undergo bidirectional transcription to generate non-coding enhancer RNAs (eRNAs). However, eRNA function remains controversial. Here, we combined Assay for Transposase-Accessible Chromatin using Sequencing (ATAC-Seq) and RNA-Seq datasets from three distinct neuronal culture systems in two activity states, enabling genome-wide enhancer identification and prediction of putative enhancer-gene pairs based on correlation of transcriptional output. Notably, stimulus-dependent enhancer transcription preceded mRNA induction, and CRISPR-based activation of eRNA synthesis increased mRNA at paired genes, functionally validating enhancer-gene predictions. Focusing on enhancers surrounding the Fos gene, we report that targeted eRNA manipulation bidirectionally modulates Fos mRNA, and that Fos eRNAs directly interact with the histone acetyltransferase domain of the enhancer-linked transcriptional co-activator CREB-binding protein (CBP). Together, these results highlight the unique role of eRNAs in neuronal gene regulation and demonstrate that eRNAs can be used to identify putative target genes.

PubMed Disclaimer

Figures

Figure 1.
Figure 1.
Genome-wide characterization of enhancers and eRNAs. (A) Analysis pipeline for localization and quantification of transcriptionally active putative enhancers (TAPEs). ATAC-seq datasets were generated using cultured cortical, hippocampal, and striatal rat neurons and used to identify regions of open chromatin (ROCs). ROCs were filtered to capture intergenic regions at least 1kb from annotated genes (iROCs). Total RNA-seq data from the same culture systems was used to identify 28 492 bidirectionally transcribed intergenic ROCs and termed TAPEs. TAPEs are characterized by an enrichment of ATAC-seq and bidirectional RNA-seq reads at TAPE centers. (B) Genome browser tracks showing ATAC-seq signal and total RNA expression at two example regions (Klf4 and Sik1), relative to tracks marking conserved DNA elements (phastConsElements20way), CTCF motifs, and enhancer-linked histone modifications. (C) TAPEs exhibit higher densities of H3K4me1, H3K4me3, H3K27ac, sequence conservation and CTCF motifs, and decreased H3K27me3 compared to surrounding regions. Chromatin immunoprecipitation datasets were obtained from the mouse forebrain at postnatal day zero (ENCODE project) and lifted over to the rat Rn6 genome assembly. (D) TAPE-gene pairs were determined by correlations of eRNA and mRNA levels at genes within a 1 Mb distance cutoff. (E) On average, TAPEs and closer genes show higher correlation values. (F) Only 20.9% of TAPEs have their maximal gene correlation at the closest gene, while the remaining 79.1% show higher correlations to genes at more distal positions. (G) For pairs with highest global TAPE–gene correlation, correlation strength does not decrease with gene distance from TAPE.
Figure 2.
Figure 2.
Identification of brain region-selective enhancers and eRNAs in primary neuronal culture systems. (A) Illustration of DESeq2-based identification of enhancers selective for cortex, hippocampus, and striatum. (B) Heatmap indicating transcription levels at region-selective TAPEs revealed 776, 390 and 898 TAPEs selective for hippocampus, cortex, and striatum, respectively. (C) Heatmap showing adjusted P-values of the six most enriched Gene Ontology Term groups for genes corresponding to region-selective TAPEs (left). HOMER analysis of transcription factor binding motifs shows an enrichment of characteristic transcription factors within region-selective TAPEs (representative examples on right, complete list in Supplementary Data Table S3). (D–F) Genome browser tracks showing ATAC-seq and total RNA-seq (reads mapped to + and – strand) signal at example loci separated by the three brain regions of interest (top, square brackets indicate y-axis max CPM value for all ATAC-seq and RNA-seq tracks at the respective loci). Region-selective TAPEs are represented by Kcnf1 for cortex, Prox1 for hippocampus, and Mn1 for striatum. Distance between TAPE and predicted target gene and global Pearson's correlation is indicated above each representative TAPE–gene pair, Rn6 chromosome coordinates are indicated below. Region-selective expression of the respective target genes is also evident in in situ hybridization images from the adult mouse brain (bottom, Image credit: Allen Brain Atlas, Allen Institute).
Figure 3.
Figure 3.
Activity dependence of enhancers and eRNAs. (A) Illustration of activity-dependent TAPE identification. (B) Volcano plots showing differentially expressed TAPEs in primary neuronal cultures generated from rat cortex, hippocampus, and striatum. (C) mRNA levels from TAPE-linked genes are predicted by the direction of eRNA changes at upregulated and downregulated TAPEs (Wilcoxon signed rank test with theoretical median = 0 for upregulated TAPEs, n = 77, P < 0.0001, and downregulated TAPEs n = 15, P < 0.0001). Chromatin accessibility decreases at gene promoters that correspond to upregulated TAPEs (Wilcoxon signed rank test with theoretical median = 0 for upregulated TAPEs, n = 78, P < 0.0001, and downregulated TAPEs, n = 15, P < 0.0730). ATAC-seq signal also decreases at both up- and downregulated TAPEs (Wilcoxon signed rank test with theoretical median = 0 for upregulated TAPEs, n = 78, P < 0.0001, and downregulated TAPEs, n = 15, P < 0.0103). (D–F) Genomic locus of Fos, Fosb, and Nr4a1 genes and their surrounding enhancer regions. (G–I) Time course experiments following neuronal depolarization with 25 mM KCl revealed that eRNAs are induced prior (Fos eRNA1, eRNA3, and Nr4a1 eRNA) or at the same time (Fosb eRNA) as mRNAs are induced (two-way ANOVA for Fos eRNA1, F(6,153) = 41.81, P < 0.0001, Fos eRNA3 F(6,154) = 37.87, P < 0.0001, Fos mRNA, F(6,154) = 456, P < 0.0001, Nr4a1 eRNA, F(6,154) = 31.4, P < 0.0001, Nr4a1 mRNA, F(6,154) = 311.3, P < 0.0001, Fosb eRNA, F(6,154) = 8.341, P < 0.0001, Fosb mRNA, F(6,154) = 98.34, P < 0.0001, Sidak's post hoc test for multiple comparisons). Inverted triangles represent P < 0.05 as compared to vehicle treated controls. (J) Summary of KCl time course experiments plotted as percentage of maximal response. (K) RT-qPCR analysis of eRNA and mRNA expression in response to 1 hr treatment with KCl, AMPA, and NMDA reveals activity-dependent induction of Fos eRNA1 and eRNA3, while FSK and TTX treatment had no effects on Fos eRNA expression (Kurskal–Wallis test for eRNA1 KCl F (3,32) = 25.04, P < 0.0001, AMPA F(3,32) = 20.81, P = 0.0001, NMDA F(3,32) = 17.79, P = 0.0005, FSK F(3,32) = 1.967, P = 0.5793, eRNA3 KCl F(3,32) = 26.52, P < 0.0001, AMPA F(3,32) = 26.11, P < 0.0001, NMDA F(3,32) = 15.66, P = 0.0013, FSK F(3,32) = 5.961, P = 0.1135, and mRNA KCl F(3,32) = 28.26, P < 0.0001, AMPA F(3,32) = 29.79, P < 0.0001, NMDA F(3,32) = 29.79, P < 0.0005, FSK F(3,32) = 30.28, P < 0.0001, with Dunn's post hoc test for multiple comparisons, and unpaired t-test for eRNA1 TTX t(14) = 0.1740, P = 0.8644, eRNA3 TTX t(14) = 1.461, P = 0.166, and mRNA TTX t(14) = 2.346, P = 0.0342). Data expressed as mean ± s.e.m. Multiple comparisons, *P < 0.05, **P < 0.01, *** P <0.001, ****P < 0.0001.
Figure 4.
Figure 4.
Transcriptional activation at enhancers is sufficient to induce linked genes. (A) Illustration of CRISPR activation (CRISPRa) strategy for site-specific targeting of the transcriptional activator VPR. (B) Immunocytochemistry on DIV 11 cortical neurons. Top, no virus control. Bottom, neurons co-transduced with lentiviruses expressing dCas9-VPR (marked by FLAG) and a custom sgRNA (mCherry reporter). Scale bar = 5 mm. (C) CRISPRa targeting to distal enhancers near Fos, Fosb, and Nr4a1 loci activates eRNA and mRNA from linked genes. Gene expression differences were measured with RT-qPCR (Fos, n = 18; Fosb and Nr4a1 n = 9 per group; two-tailed Mann-Whitney test for all comparisons as compared to non-targeting lacZ sgRNA control). (D) CRISPRa at Fos, Fosb, and Nr4a1 promoters increases mRNA but does not consistently increase eRNA (Fos, n = 18; Fosb and Nr4a1 n = 9 per group; two-tailed Mann–Whitney test for all comparisons as compared to non-targeting lacZ sgRNA control). Data expressed as mean ± s.e.m. Multiple comparisons, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.
Figure 5.
Figure 5.
Fos eRNA1 is sufficient for Fos mRNA expression. (A) Illustration of Display plasmids (top), and CREB binding motifs, sgRNA target sites and eRNA regions chosen for CRISPR-Display constructs (acRNA1–3) (bottom). (B) RT-qPCR analysis reveals that while the lacZ-targeting 150 nt eRNA1 acRNA does not affect Fos mRNA (n = 9 per group, two-tailed Mann–Whitney test, U = 36, P = 0.7304), targeting eRNA1 to Fos enhancer-1 results in increased Fos mRNA expression (n = 21 per group, two-tailed Mann–Whitney test, U = 77, P = 0.0002, graph contains data from 12 replicates (from four experiments) shown in the right graph and none additional replicates (from 2 experiments)). Constructs with increasing acRNA lengths (150, 300 and 450 nt) did not result in stronger Fos mRNA induction (one-way ANOVA F(3,44) = 3.791, P = 0.0167, Tukey's post hoc test for multiple comparison). (C) Tethering 150 nt acRNAs based on eRNA1, eRNA3, and a control RNA revealed that only eRNA1 tethering to its own enhancer induced Fos mRNA expression (one-way ANOVA for control-targeting F(3,19) = 3.191, P = 0.0472, and Kurskal–Wallis test for E1-targeting F(3,44) = 23.27, P < 0.0001, E3 targeting F(3,40) = 5.183, P = 0.1589). Data expressed as mean ± s.e.m. Multiple comparisons, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.
Figure 6.
Figure 6.
Fos eRNA1 is necessary for Fos mRNA expression in neurons. (A) Anti-sense oligonucleotide (ASO) targeting of Fos eRNA1 for 24 h decreased both eRNA1 and Fos mRNA (unpaired t-test t(10) = 20.69, P < 0.0001 and t(10) = 5.739, P = 0.0002), but did not alter eRNA levels from other Fos enhancers (unpaired t-test for Gapdh mRNA t(10) = 0.9696, P = 0.3551; eRNA2 t(10) = 0.8608, P = 0.4095; eRNA3 t(10) = 1.014, P = 0.3346). (B) Fos mRNA targeting ASOs decreased Fos mRNA (t(10) = 5.198, P = 0.0004) with no significant effect on eRNA levels (unpaired t-test for Gapdh mRNA t(10) = 3.744, P = 0.0038; eRNA1 t(10) = 2.056, P = 0.0668; eRNA2 t(10) = 0.6508, P = 0.5298; eRNA3 t(10) = 1.679, P = 0.124). (C) Fos eRNA1 ASO pretreatment for 24 h prior to 1 h Veh treatment or KCl stimulation reduced induction of eRNA1 (top) and mRNA (bottom) when compared to a scrambled ASO control (n = 9 per group, two-way ANOVA for eRNA1 F(1,32) = 154.4 P < 0.0001, for mRNA F(1,32) = 5.267, P = 0.0284). Data expressed as mean ± s.e.m. Multiple comparisons, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.
Figure 7.
Figure 7.
Fos eRNA is transcribed independently of but interacts with the histone acetyltransferase CBP. (A) shRNA-mediated knockdown of Crebbp mRNA resulted in deceased Fos mRNA but not eRNA expression (n = 18 per group, Mann–Whitney for Crebbp mRNA U = 0, P < 0.0001, Fos mRNA U = 55, P = 0.0004, eRNA1 U = 140, P = 0.5010, eRNA3 U = 132, P = 0.3550). (B) CREB inhibition (666–15; 1 μM) blunted the KCl response of Fos mRNA but not eRNA (n = 6 per group, two-way ANOVA for mRNA F(1,20) = 37.79, P < 0.0001, eRNA1 F(1,20) = 0.2461, P = 0.8769, eRNA3 F(1,20) = 0.1592, P = 0.6941, with Tukey's post hoc test for multiple comparisons). (C) CRISPR dCas-HAT targeting in C6 cells, in which dCas9 carrying a histone acetyltransferase domain is expressed with sgRNAs to target selected enhancers (left) induced Fos mRNA transcription (right, n = 8–9 per group, one-way ANOVA F(2,23) = 6.151, P = 0.0072). (D) Illustration of CREB-binding protein (CBP) domains (top), and recombinant glutathione-S-transferase (GST) tag-containing CBP-histone acetyltransferase (HAT) domain and CBP-bromodomain (Bromo) used in mobility shift assays. (E–G) RNA electrophoretic mobility shift assay (REMSA) with escalating concentrations of recombinant protein (0 = free probe (FP), 0.05–0.6 μM) reveal complete binding of synthetic Fos eRNA1 (151 bp, 20 nM), eRNA3 (153 bp, 20 nM), and control RNA (150 bp, 20 nM) to CBP-HAT. (H) Unlabeled eRNA1 competes for CBP-HAT binding with labeled eRNA1 (LP) in competition assay. (I) No binding of eRNA1 to CBP-Bromo domain was observed. For all REMSA experiments, n = 6 per group. (J) Model of enhancer function at promoters with associated eRNAs interacting with CBP. Data expressed as mean ± s.e.m. Multiple comparisons, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Heinz S., Romanoski C.E., Benner C., Glass C.K.. The selection and function of cell type-specific enhancers. Nature. 2015; 16:144–154. - PMC - PubMed
    1. Li W., Notani D., Rosenfeld M.G.. Enhancers as non-coding RNA transcription units: recent insights and future perspectives. Nat. Rev. Genet. 2016; 17:207–223. - PubMed
    1. Wang D., Garcia-Bassets I., Benner C., Li W., Su X., Zhou Y., Qiu J., Liu W., Kaikkonen M.U., Ohgi K.A. et al. .. Reprogramming transcription by distinct classes of enhancers functionally defined by eRNA. Nature. 2011; 474:390–394. - PMC - PubMed
    1. Furlong E.E.M., Levine M.. Developmental enhancers and chromosome topology. Science. 2018; 361:1341–1345. - PMC - PubMed
    1. Gray J.M., Kim T.-K., West A.E., Nord A.S., Markenscoff-Papadimitriou E., Lomvardas S.. Genomic views of transcriptional enhancers: essential determinants of cellular identity and activity-dependent responses in the CNS. J. Neurosci. 2015; 35:13819–13826. - PMC - PubMed

Publication types

MeSH terms