T3-Dependent Transcriptional Programming by TRβ in Thyroid Cells Requires SWI/SNF Chromatin Remodelers

Abstract

Abstract Transcriptional regulation in response to thyroid hormone (T3) is a dynamic and cell-type specific process that maintains cellular homeostasis and identity. A bimodal switch model, where T3 binding alters the co-regulator profile of a constitutively DNA-bound thyroid hormone receptor (TR) to affect downstream gene expression, is widely used to describe the interaction between TRβ and chromatin. To test this model on a genome-wide scale, we used an integrated genomics approach to profile and characterize the cistrome of TRβ by CUT&RUN, map changes in chromatin accessibility by ATAC-seq, and capture the transcriptomic changes in response to T3 by RNA-seq in the normal thyroid cell line, Nthy-ORI. Our CUT&RUN data demonstrated that T3 binding causes significant shifts in TRβ genomic occupancy; these shifts are associated with differential chromatin accessibility. Most of the T3-induced differentially expressed genes have a TRβ binding site associated with the proximal protomer region within one kilobase of the transcriptional start site, suggesting that these are direct TRβ regulatory target genes. Remarkably, the majority of TRβ binding sites were found in transgenic regions and distal regulatory elements. In order to identify the co-regulatory proteins that are required for execution of a T3-dependent transcriptional program in our thyroid cells, we used a TRβ-miniTurboID fusion construct to perform a proximity ligation assay followed by mass spectrometry. We identified 1,138 nuclear proteins that interact with TRβ. Of these proteins, 75 interact preferentially in the presence of T3 and 68 in the absence of T3. All of the core SWI/SNF complex subunits from both of the major subtypes (BAF and PBAF) were identified as TRβ. Interestingly, we found that the PBAF-specific subunit, PBRM1, was significantly enriched in the presence of T3. BAF complex-specific subunits, such as ARID1A, were also present in our dataset. To test whether TRβ could differentially recruit BAF and PBAF complexes to its binding sites, we performed CUT&RUN targeting BRG1, PBRM, and ARID1A to determine the degree of co-occupancy with TRβ. Based on our comprehensive genomic and proteomic analyses, we propose a new model for selective recruitment of BAF and PBAF SWI/SNF complexes to TRβ binding sites for differential functions in regulating chromatin accessibility.