A chromatin extension model for insulator function based on comparison of high-resolution chromatin conformation capture and polytene banding maps

Abstract

Insulator proteins bind to specific genomic loci and have been shown to play a role in partitioning genomes into independent domains of gene expression and chromatin structure. Despite decades of study, the mechanism by which insulators establish these domains remains elusive. Here, we use genome-wide chromatin conformation capture (Hi-C) to generate a highresolution map of spatial interactions of chromatin from Drosophila melanogaster embryos. We show that from the earliest stages of development the genome is divided into distinct topologically associated domains (TADs), that we can map the boundaries between TADs to subkilobase resolution, and that these boundaries correspond to 500-2000 bp insulator elements. Comparing this map with a detailed assessment of the banding pattern of a region of a polytene chromosome, we show that these insulator elements correspond to low density polytene interbands that divide compacted bands, which correspond to TADs. It has been previously shown that polytene interbands have low packing ratios allowing the conversion of small genomic distances (in base pairs) into a large physical distances. We therefore suggest a simple mechanism for insulator function whereby insulators increase the physical space between adjacent domains via the unpacking and extension of intervening chromatin. This model provides an intuitive explanation for known features of insulators, including the ability to block enhancerpromoter interactions, limit the spread of heterochromatin, and organize the structural features of interphase chromosomes. IntroductionBeginning in the late 19th century, cytological investigations of the polytene chromosomes of insect salivary glands implicated the physical structure of interphase chromosomes in their metabolic functions (1-5). Progressively more detailed optical and electron microscopic analysis of polytene chromosomes in Drosophila melanogaster have identified a stereotyped banding pattern with compacted, DNA-rich "bands" alternating with extended, DNApoor "interband" regions(6-10). While much is now known about the molecular properties of these two types of chromatin (11)(12)(13)(14)(15)(16)(17)(18)(19)(20), the precise molecular nature of the banding structure remains elusive.Several methods have been developed in the past decade for the isolation and highthroughput characterization of chromosomal regions that are colocalized within nuclei (21-24), yielding genome-wide maps of chromatin structure in a number of organisms and tissues. These experiments revealed that the interphase chromatin fiber is organized into topologicallyassociated domains (TADs), contiguous regions of the genome that exhibit enriched threedimensional contacts between loci within the TAD, and depleted contacts between loci in different TADs. Eagen et al. recently showed that TADs identified from genome-wide chromatin conformation capture (Hi-C) at approximately 15 kb resolution from polytene nuclei of Drosophila melanogaster largely correspond to the bands of polytene chromosomes, ...

Tethering of CHROMATOR and dCTCF proteins results in decompaction of condensed bands in the Drosophila melanogaster polytene chromosomes but does not affect their transcription and replication timing

“…Indeed, this proposal has been confirmed experimentally by comparing the molecularly mapped band/interband borders with the positions of TADs [ 18 ]. This correlation has been comprehensively analyzed by Eagen et al [ 20 ] who demonstrated that TAD borders matched the flanking regions of large bands found in polytene chromosomes and by Stadler and Eisen [ 21 ]. These studies have thus reinforced the hypothesis on the existence of a shared plan of interphase chromosome organization.…”