Conventional and pioneer modes of glucocorticoid receptor interaction with enhancer chromatin in vivo. Mus musculus

Glucocorticoid hormone plays a major role in metabolism and many related diseases. The hormone-bound glucocorticoid receptor (GR) binds to a specific set of enhancers in different cell types, resulting in unique patterns of gene expression. GR-responsive enhancers have an accessible chromatin structure prior to hormone treatment (“pre-programmed”), whereas unresponsive enhancers specific to other cell types are inaccessible and inactive. Here we have addressed the role of chromatin structure in cell-specific GR-enhancer programming by precise mapping of nucleosome positions in mouse adenocarcinoma cells. We show that, before hormone treatment, some pre-programmed GR-enhancers are nucleosome-depleted, associated with the Brg1 chromatin remodeler and flanked by nucleosomes exchanging histones H2A.Z and H2A. However, most pre-programmed GR-enhancers are assembled into a nucleosome that exchanges H2A.Z and H2A, with little or no Brg1. After hormone treatment, nucleosomes at both types of pre-programmed GR-enhancer shift apart, coinciding with increased levels of Brg1, and continue to exchange H2A.Z and H2A. Hormone removal rapidly reverses these nucleosome shifts. In contrast, inactive GR-enhancers are nucleosomal, lack Brg1, do not exchange H2A.Z or H2A and do not respond to hormone. Thus, pre-programmed GR-enhancers are marked by a dynamic, mutable chromatin structure characterized by high levels of H2A.Z exchange. We propose that nucleosome-depleted GR-enhancers result from the binding of other transcription factors together with Brg1 to assist loading of GR after hormone treatment. At nucleosomal enhancers, GR binding may be directly facilitated by increased transient exposure of enhancer DNA associated with H2A.Z exchange. Overall design: Performed micrococcal nuclease digests of nuclei prepared from 3134 cells and sequenced mono-nucleosomal DNA using paired-end sequencing to map nucleosome positions genome wide. These data are correlated with DNase-seq and ChIP-seq data.