Binding the boundaries of chromatin domains
© Pirrotta; licensee BioMed Central Ltd. 2014
Published: 30 June 2014
Skip to main content
© Pirrotta; licensee BioMed Central Ltd. 2014
Published: 30 June 2014
A new study proposes an integrated framework to improve our understanding of the multiple functions of insulator elements, and their architectural role in the genome.
See related research; http://genomebiology.com/2014/15/6/R82
Since their discovery in Drosophila, insulator elements and the proteins that bind to them have fascinated researchers for their unusual topology-dependent enhancer-blocking activities, and for the diverse roles they play in the genome. In recent years, attention has focused on their ability to promote interactions between different binding sites, producing higher-order, long-distance folding of chromatin. This could be the underlying feature common to the different roles that have been ascribed to them. A new article from the laboratory of Victor Corces now proposes an integrative framework to understand their function through their ability to determine the folding and architecture of genomic chromatin .
In Drosophila, five different insulator DNA-binding proteins have been described: Su(Hw), dCTCF, Zw5, BEAF32 and, in some contexts, GAGA-binding factor (GAF). The first three of these, like mammalian CTCF, are multi-zinc finger proteins with a similar structure. Two other proteins, CP190 and isoforms of Mod(mdg4), are known to interact with insulator-binding factors and are recruited to many, but not all, of their binding sites. CP190 and Mod(mdg4) contain a BTB/POZ domain known to mediate strong protein-protein interactions. At sites known to act as enhancer blocking insulators, they are essential for insulator function and are thought be the ‘glue’ that tends to hold together two or more insulator sites.
How insulators act as enhancer blockers remains mechanistically obscure, although several models have been proposed. They do not act as enhancer decoys that compete with promoters; instead they interfere with physical contact between chromatin elements on one side, and elements on the other side of their binding site. The best formulation so far, based largely on Su(Hw), was offered by Capelson and Corces , who proposed that insulator-binding complexes interact with one another through the ‘glue’ proteins CP190 and Mod(mdg4), organizing genomic chromatin into loops radiating from clusters of interacting insulator elements. The topology of these structures somehow prevents interactions between different chromatin loops.
Surprisingly, most of the Drosophila insulator proteins are phylogenetically recently derived. The exception is CTCF, the only insulator protein so far known in mammals. Homologs of the ‘glue’ proteins are not conserved in mammals. Instead, many but not all CTCF binding sites also bind cohesin. Cohesin and condensin are multi-protein ring-shaped complexes whose key structural constituents are called structural maintenance of chromosome (SMC) proteins. Cohesin and condensin can use their ring structure to hold together two chromatin fibers, and they serve important roles in chromosome condensation and segregation during mitosis. They are also important in chromatin architecture and gene expression during interphase. It has been shown that cohesin binding at CTCF binding sites depends on CTCF and that cohesin is required for CTCF insulator function. Thus, in some way, cohesin, like the ‘glue’ proteins in Drosophila, may gather together multiple CTCF binding sites.
Genomic mapping studies have revealed hundreds of binding sites for each of the insulator proteins in Drosophila, and more than 10,000 for mammalian CTCF. However, many of these do not have insulator function; instead, in many cases they are involved in folding the chromatin fiber, thus bringing appropriate regulatory regions together to contact promoters. Most insulator protein binding sites are invariant in different tissues or developmental stages, suggesting that they play structural roles in shaping the architecture of the genome. In the last few years, studies using Chromatin Conformation Capture (3C) technologies for genomic analysis of chromatin-chromatin interactions have revealed the presence of complex higher order organization in mammalian and Drosophila genomes. The genome is organized in topologically associated domains (TADs) within which chromatin-chromatin interactions are extensive, while interactions across domain borders are suppressed [3–8]. The domain borders therefore have the properties expected for insulator elements, which then become understandable in the broader context of genomic architecture. Many TAD borders in fact correspond to binding sites for insulator proteins. However, many more binding sites are found within domains, indicating that most do not act as either domain borders or insulators.
In a new study, Van Bortle et al.  considered first of all that insulator proteins function as organizers of genomic architecture rather than as local insulators, and they proposed to rename them architectural proteins (APs) and their binding sites as APBSs. They used Drosophila as a model because of its variety of known APs and they mapped an additional kind of element that has AP-like properties: TFIIIC, a DNA-binding complex that recruits RNA polymerase III (RNA Pol III) to targets such as tRNA genes. Clusters of tRNA genes in budding yeast and in mammalian genomes have insulator properties dependent on TFIIIC. Binding sites for TFIIIC unconnected with RNA Pol III promoters are also found in yeast, Drosophila and mammals, and are known as extra TFIIIC (ETC.) sites. The condensin complex is enriched at TFIIIC binding sites and is responsible for their spatial clustering in the nucleus. Thus, TFIIIC might constitute yet another AP, linked to the others by the association with SMC-containing complexes such as cohesin or condensin. Van Bortle et al. find that both of these SMC complexes are enriched at ETC. sites, and that ETC. sites and SMC complexes often coincide with sites that bind multiple APs. Thus, contrary to earlier conclusions, in Drosophila, as well as in mammals, APs are often associated with SMC complexes. In fact, the association is correlated with the degree of AP occupancy of a site: the more APs bound the more likely the association with SMC complexes. Furthermore, knocking down APs such as dCTCF leads to loss of cohesin at these sites.
Mouse ES cell APs
CAP-H2 (condensin II)
CAP-H2, CAP-D3 (condensin II)
The discovery of the relationship between APBS clustering and the strength of TAD boundaries was made possible by the variety of different ABs known in Drosophila. This insight would not have been evident in mammals, where CTCF was the only insulator protein known. Going back to the extensive human and mouse Hi-C data and adding mapping data for TFIIIC, cohesin and condensin as well as CTCF and PRDM5 (another protein recently found to be frequently associated with the first four), Van Bortle et al. were able to confirm that the conclusions reached for Drosophila hold also for mammalian genomes.
The properties of TAD boundaries are those expected of a good old-fashioned enhancer-blocking insulator. Might clustering of APBSs also explain which APBSs have insulator properties and which do not? Going back to Drosophila, APBSs that have been directly found to have insulator activity, Van Bortle et al. show that they are well correlated with the degree of occupancy of APs. Sites with robust insulator activity bind at least seven APs. Sites whose insulator activity is context-dependent bind, on average, five APs, while sites lacking demonstrable insulator activity bind an average of three and a half. Furthermore, insulators correspond, as expected, to TAD boundaries and the degree of insulator activity corresponds to the strength of the boundary.
Architectural protein binding site
Extra TFIIIC (not associated with RNA Pol III promoters)
RNA polymerase III
Structural maintenance of chromosome
Topologically associated domain.
This article is published under license to BioMed Central Ltd. The licensee has exclusive rights to distribute this article, in any medium, for 12 months following its publication. After this time, the article is available under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.