Hyperacetylated Chromatin Domains Mark Cell Type-Specific Genes and Suggest Distinct Modes of Enhancer Function

Abstract

Stratification of enhancers by relative signal strength in ChIP-seq assays has resulted in the establishment of super-enhancers as a widespread and useful tool for identifying cell typespecific, highly expressed genes and associated pathways. We have examined a distinct method of stratification that focuses on peak breadth, termed "hyperacetylated chromatin domains" (HCDs), which classifies broad regions exhibiting histone modifications associated with gene activation. We find that this analysis serves to identify genes that are both more highly expressed and more closely aligned to cell identity than super-enhancer analysis does when applied to multiple datasets. Moreover, genetic manipulations of selected gene loci suggest that at least some enhancers located within HCDs work at least in part via a distinct mechanism involving the modulation of covalent histone modifications across domains, and that this activity can be imported into a heterologous gene locus. In addition, such genetic dissection reveals that the super-enhancer concept can obscure important functions of constituent elements. MainEnhancers are cis-regulatory DNA sequences that are bound by transcriptional activators which regulate genetically linked gene promoters 1-5 . They are also historically characterized by their ability to function over distances of anywhere from 100 bp to more than 1 Mb. Whole-genome methods of identifying enhancers rely on associated histone modifications, common transcriptional cofactors, and/or cell type-specific transcription factors (TFs). These methods suggest that mammalian genomes harbor a large number of enhancer sequences -perhaps 1-2 million -and that they represent the most numerous and significant cis-acting sequence determinants of cell type-specific gene expression.Fundamental issues regarding enhancer function, however, remain unclear. For example, there is no consensus for how enhancers communicate with their cognate gene promoters. The dominant model, termed "looping," involves direct interactions between factors bound to enhancers and factors bound near promoters. Evidence for such interactions, however, has provided little insight into how a distal enhancer finds a gene promoter, or how it distinguishes among potential promoters in gene-dense regions. Moreover, some evidence suggests that mechanisms of enhancer-promoter communication may be more varied 3,5 .