Ligand dependent gene regulation by transient ERα clustered enhancers

Abstract

Unliganded Estrogen receptor alpha (ERα) has been implicated in ligand-dependent gene regulation. Upon ligand exposure, ERα binds to several EREs relatively proximal to the premarked, unliganded ERα-bound sites and affects transient but robust gene expression. However, the underlying mechanisms are not fully understood. Here we demonstrate that upon ligand stimulation, persistent sites interact extensively, via chromatin looping, with the proximal transiently ERα-bound sites, forming Ligand Dependent ERα Enhancer Cluster in 3D (LDEC). The E2-target genes are regulated by these clustered enhancers but not by the H3K27Ac super-enhancers. Further, CRISPR-based deletion of TFF1 persistent site disrupts the formation of its LDEC resulting in the loss of E2-dependent expression of TFF1 and its neighboring genes within the same TAD. The LDEC overlap with nuclear ERα condensates that coalesce in a ligand and persistent site dependent manner. Furthermore, formation of clustered enhancers, as well as condensates, coincide with the active phase of signaling and their later disappearance results in the loss of gene expression even though persistent sites remain bound by ERα. Our results establish, at TFF1 and NRIP1 locus, a direct link between ERα condensates, ERα enhancer clusters, and transient, but robust, gene expression in a ligand-dependent fashion. Author summaryERα occupies its cognate binding sites on the genome, upon estrogen stimulation, to bring about the signaling mediated response. However, ERα has been implicated in gene regulation even in the absence of hormone and its underlying mechanisms are not known. Here we show that, unliganded ERα-bound genomic sites are necessary for the emergence of active enhancers upon estrogen stimulation. Several ERα sites emerge in proximity to these unliganded ERα sites which loop with each other in three-dimensional nuclear space forming the enhancer:ERα clusters that we call Ligand dependent enhancer PLOS Genetics | https://doi.

Transcriptional regulation and chromatin architecture maintenance are decoupled functions at the Sox2 locus

“…As such, the SCR is presumed to strongly activate target genes due to the combinatorial action of its composite transcription factor binding sites. However, the functional significance of “superenhancers” compared with nonclustered enhancers is markedly debated, with previous genetic dissections uncovering context-dependent transcriptional effects that may be functionally redundant or additive, but not synergistic ( Hay et al 2016 ; Moorthy et al 2017 ; Saravanan et al 2020 ; for discussion, see Blobel et al 2021 ). Although the 7.3-kb-long SCR contains four regions, each bound by more than six different transcription factors, we found that only two of these regions (SRR107 and SRR111) account for the vast majority of Sox2 transcription activation in cis.…”

Transcription decouples estrogen-dependent changes in enhancer-promoter contact frequencies and spatial proximity

“…1D and Supplemental Fig.1E). Residual GREB1 expression may result from multiple regulatory regions coordinating GREB1 expression in E2 treated cells (Saravanan et al, 2020). Deletion of the NRIP1 enhancer led to almost total loss of NRIP1 induction in response to E2 (Fig.…”

“…1E). Residual GREB1 expression may result from multiple regulatory regions coordinating GREB1 expression in E2 treated cells (Saravanan et al, 2020).…”

Transcriptional regulation and chromatin architecture maintenance are decoupled modular functions at the Sox2 locus

“…As such, the SCR is presumed to strongly activate target genes due to the combinatorial action of its composite transcription factor binding sites. However, the functional significance of “super-enhancers” compared to non-clustered enhancers is markedly debated, with previous genetic dissections uncovering context-dependent transcriptional effects that may be functionally redundant or additive, but not synergistic (Hay et al, 2016; Moorthy et al, 2017; Saravanan et al, 2020; discussed in Blobel et al, 2021). Although the 7.3 kb-long SCR contains four regions, each bound by more than six different transcription factors, we found that only two of these regions (SRR107 and SRR111) account for the vast majority of Sox2 transcription regulation in cis.…”