Insulators are DNA-protein complexes that function in chromatin either as barriers to oppose the inappropriate spread of silent heterochromatin or as enhancer blockers, preventing them from activating the wrong genes (
Wallace and Felsenfeld, 2007). The picture emerging from recent genomic profiling is that these functions are integrated into the over-all transcriptional pattern within cells through formation of long range interactions. The vertebrate insulator protein CTCF is an organizer of chromosomes and chromosome territories and forms a nuclear network of chromatin loops (
Botta et al., 2010;
Lieberman-Aiden et al., 2009). A global CTCF interactome study further suggests that CTCF configures the genome into domains with unique epigenetic and transcriptional properties (
Handoko et al., 2011). These findings raise the expectation that at least some insulator interactions will be cell type specific and dynamic as gene expression changes during development and differentiation. In this issue,
Wood et al. (2011) provide insights into how insulator activity can be established at specific
Drosophila genomic sites to achieve proper gene expression through recruitment of the insulator binding protein CP190.
In
Drosophila, the interaction of the protein Su(Hw), with the
gypsy insulator has provided the paradigm for insulator function. Su(Hw) recruits other components to insulator sites, including CP190, and these elements coalesce in nuclei to form insulator bodies, presumably based on loop formation among the elements. Recent 3C data support the idea that Su(Hw) insulators can form chromatin loops (
Comet et al., 2011;
Li et al., 2011). Genome profiling in two
Drosophila cell lines showed that among sites for insulator proteins Su(Hw), dCTCF and BEAF, a subset (5–37%) were cell type specific, similar to results in vertebrate cells for CTCF (
Bushey et al., 2009;
Cuddapah et al., 2009). Co-occupancy by CP190 was not universal and at some sites where CP190 occupancy differed between the cell lines, occupancy by BEAF, Su(Hw) or dCTCF was invariant. These data provided the basis to investigate cell type specificity and function of insulators in
Drosophila.
Drosophila cells experience dramatic changes in gene expression in response to heat-shock, which turns off transcription of most genes while inducing transcription of heat-shock genes to high levels. Using immunostaining of Kc cell polytene chromosomes, Wood et al observed that localization of insulator proteins Su(Hw), BEAF and dCTCF was not notably affected during heat shock, however, CP190 was released from most sites and Su(Hw) insulator bodies were severely disrupted. These results make it likely that CP190 recruitment to insulators is necessary for interactions that affect insulator body formation and that their disruption correlates with the global changes in gene expression after heat shock. Since heat shock primarily represses genes, the data suggest that CP190-mediated insulator interactions function broadly to maintain genes in the active state. The retention of DNA binding insulator proteins after heat shock might provide a bookmark for the return of CP190 after recovery.
To induce more modest transcriptional changes, Wood et al. treated Kc cells with the hormone ecdysone. The biologically active form of ecdysone, 20-hydroxyecdysone (20-HE), binds to the ecdysone nuclear receptor complex (ECR-C) and controls postembryonic development of Drosophila. 20-HE treatment over a time course of 48 hours resulted in up- or down-regulation of about 2000 genes, but insulator bodies were not obviously disrupted. Genome profiling by ChIP-Seq of insulator proteins revealed that BEAF and dCTCF generally occupied fewer sites over the course of hormone treatment, while Su(Hw) was reduced at some sites after 3 hours but increased or occupied new sites after 48 hours. CP190 binding was initially unchanged but by 48 hours of treatment new sites of occupancy appeared. CP190 consistently co-localized with at least one insulator protein before and after 20-HE treatment, although, over time, tended towards co-localizing with Su(Hw) or Su(Hw) together with dCTCF at the expense of co-localization with BEAF. The preferential recruitment of CP190 at sites already occupied by insulator proteins suggests that such sites may be poised insulators which become active upon recruitment of CP190 after 20-HE treatment.
To test this idea and to ask if CP190 is crucially involved in insulator looping, the authors chose ecdysone induced 75B (Eip75B), which is broadly occupied by the 20-HE receptor complex after induction and contains a CTCF site at which CP190 abundance dramatically increases upon 20-HE treatment. 3C was performed using this site as the anchor in cells with normal or depleted CP190 protein, with or without 20-HE treatment. Interactions were observed between the anchor site and two other downstream sites consistently occupied by CP190 and dCTCF. The interactions were initially fairly weak and became stronger after 20-HE treatment, but they were unchanged in the CP190-depleted background. The conclusion is that dCTCF occupancy at these sites suffices for a moderate frequency of interaction but that 20-HE induced CP190 recruitment to the anchor site stabilizes the loops. Interestingly, in the vertebrate
INFG and
TCRα loci, cohesin is recruited in a developmental stage specific fashion to CTCF sites and provides the basis for looping interactions that regulate gene expression (
Hadjur et al., 2009;
Seitan et al., 2011).
What is the functional significance of inducible CP190 loop stabilization at Eip75B? The four Eip75B alternative transcripts respond differently to 20-HE treatment. The RB transcript, uniquely, shows an initial increase and then levels off. Reduction of CP190 leads to aberrant up-regulation of Eip75B-RB and four neighboring genes. The picture that emerges is that CP190 recruitment to a pre-existing insulator site in response to a differentiation signal stabilizes an insulator loop necessary for protection of these genes from activation and that this protection is lost upon depletion of CP190 (). The promoters of the affected transcripts are separated from the receptor complex by the CP190-regulated insulator site; however, not all of them lie within the chromatin loop strengthened by CP190 recruitment. This result is consistent with the idea that the transcriptional effects of insulators may be imposed by alterations in chromatin structure on a more global scale than that represented by individual loop formation alone.
In principle, insulator loops might be regulated at the level of DNA binding by insulator proteins or through recruitment of additional components required for looping. Current information suggests that both mechanisms are used by vertebrate and dipteran cells (
Wallace and Felsenfeld, 2007). The new work of Wood et al shows that CP190 recruitment to an insulator site and participation in loop formation is a key regulatory point. It will be important to extend these observations to other differentially occupied CP190 sites. Still to be addressed is how insulator proteins and the partners they recruit find specific site of occupancy that will influence transcription among thousands of opportunities genome wide. We must also turn the question around and ask whether the nuclear organization that seems to be provided by insulators determines the transcription outcome or is determined by it.