A new study shows that during transcription, the TH2 interleukin gene cluster is organized into several small chromatin loops connected at their base by the protein SATB1. This first detailed glimpse of chromatin folding provides a new perspective on the coordination of cell type–specific gene expression.
Chromatin loops represent short- or long-range physical juxtapositions between distal DNA sequences and are generally thought to involve interactions between chromatin proteins that bind to underlying regulatory sequences. This is exemplified by the TH2-specific locus control region (LCR), which is positioned within the Rad50 gene to coordinate transcription of the neighboring cluster of genes encoding interleukins 4, 5 and 13 (refs. 1,2). In naive T cells, the TH2 LCR of the interleukin gene cluster is physically juxtaposed with the promoter region of the interferon-γ gene, located on another chromosome3. This arrangement may organize a 'pre-poised' state before activation of the interferon-γ gene and the interleukin gene cluster. When naive T cells differentiate into TH1 or TH2 cells, this interchromosomal pattern of interaction is lost in favor of intrachromosomal interactions, and mutually exclusive expression of interferon-γ and interleukin genes is established3. The results reported by Shutao Cai and colleagues4 on page 1278 bear directly on this proposal. They show that a factor termed SATB1 folds the chromatin into many small loops in activated TH2 cells. This unexpected chromatin structure for an active gene cluster may coordinate gene expression in cis and facilitate transcription by restricting the physical volume of the gene cluster.
SATB1 is the anchor
The initial bet that this involved SATB1 paid off, as downregulation of SATB1 expression prevented both chromatin folding and the expression of the interleukin gene cluster. So what is SATB1? First identified by the Kohwi-Shigematsu laboratory in 1992 (ref. 5), SATB1 has turned out to have a major role in T cell development6. There are several additional features that, taken together, make SATB1 unique: it binds to regions of unpaired bases exposed to negative torsional stress (termed BURs), it interacts with chromatin remodeling factors and it forms a nuclear, cage-like structure7.
The initial strategy to monitor chromatin folding during transcriptional activation depended on the in vitro identification of SATB1 binding sequences from a 200-kb region covering the interleukin gene cluster and Rad50. Thus, equipped with a set of nine strong binding sites, Cai et al. carried out an extensive chromosomal conformation capture (3C) analysis. This method assays patterns of chromatin loops based on the close physical proximity between remote chromatin segments8. The authors found that some loops involving the BUR elements existed in resting cells, presumably contributing to the poised state of the interleukin gene cluster. This tentative conclusion found support in the observation that many of the preexisting interactions increased in frequency by an order of a magnitude in activated TH2 cells. Moreover, new interactions were established. In some instances, a single BUR element made contact with five different interactors within the interleukin gene cluster, although it remains to be determined if these interactions occurred simultaneously.
To examine the possibility that SATB1 was part of these chromatin loops, the authors used the ChIP-loop assay9. This method combines the virtues of the urea-ChIP method with those of the 3C technique. An intuitive objection to the ChIP-loop data is that the antibody-antigen complex might aggregate the cross-linked complexes before the ligation step. The expectation that this would generate a random pattern of interactions was not, however, borne out. Indeed, the ChIP-loop and 3C assays produced very similar data. We concur with the authors' conclusion that SATB1 is at the base of all the numerous, small chromatin loops in the interleukin gene cluster of activated TH2 cells. It remains to be seen if there are extensive additional chromatin loops from other regions within the interleukin gene cluster that do not interact with SATB1-binding sites. In addition, it cannot be formally ruled out that the interleukin gene cluster has interchromosomal interactors, although it is clear that this does not involve the homologous chromosome.
Poised versus active states
The presence of some TH2-specific factors, such as STAT6 and GATA3, on the chromatin fiber of the interleukin gene cluster in resting cells is in line with the observation that these factors are important to establish a 'poised' chromatin state2. The attainment of an active state required the SATB1-dependent recruitment of additional factors, including c-Maf, a regulator of interleukin 4 expression, and the chromatin-remodeling factor Brg1. In the authors' own words, the transition between the poised and active states requires “tethering of the cytokine gene locus onto the SATB1 architecture” to trigger the recruitment of these factors and RNA polymerase II. The SATB1-dependent induction of c-Maf expression might complicate this scenario. As c-Maf is downstream of SATB1, it might be of interest to examine the effects of downregulated c-Maf function on chromatin loop formation in activated TH2 cells. This might clarify whether the SATB1-dependent chromatin structure is the cause or the consequence of transcriptional activation. Surprisingly, STAT6, GATA3 and c-Maf were distributed along the chromatin fiber of the entire gene cluster rather than occupying distinct positions. This implies that these factors might participate in larger structures, such as transcription factories, that are exposed to the entire gene cluster.
Why small chromatin loops?
SATB1 is likely to coordinate transcription of this gene cluster by stabilizing or facilitating physical interactions between the enhancer within the Rad50 LCR and the interleukin promoters. The authors explain this in terms of SATB1 reducing the physical volume of the gene cluster to increase exposure to transcriptional regulatory factors. Alternatively, the SATB1-dependent compaction of the active chromatin into dense loops might increase anchorage of the interleukin gene cluster to the chromosomal scaffold (Fig. 1). This explanation bears on a recent proposal that differentiation processes are generally accompanied by restrictions in chromatin movement10. By extrapolation, the transition from a poised to an active state would antagonize chromatin mobility to fix patterns of expression of the interleukin gene cluster.
The activation of the interleukin gene cluster might represent only one of many examples of SATB1-dependent coordination of expression from gene clusters in TH2 cells. Indeed, the extensive overlaps in nuclear distribution of SATB1 and RNA polymerase II might reflect a direct physical interaction between these pivotal proteins, perhaps in the context of transcription factories. In conclusion, the new perspective emerging is that cell type–specific coordination of gene expression might involve regulated access to BUR elements and formation of small chromatin loops in combination with chromatin movements.
Wilson, C.B. & Merkenschlager, M. Curr. Opin. Immunol. 18, 143–151 (2006).
Lee, G.R., Kim, S.T., Spilianakis, C.G., Fields, P.E. & Flavell, R.A. Immunity 24, 369–379 (2006).
Spilianakis, C., Lalioti, M., Town, T., Lee, G. & Flavell, R.A. Nature 435, 637–645 (2005).
Cai, S., Lee, C.C. & Kohwi-Shigematsu, T. Nat. Genet. 38, 1278–1288 (2006).
Dickinson, L.A., Joh, T., Kohwi, Y. & Kohwi-Shigematsu, T. Cell 70, 631–645 (1992).
Alvarez, J.D. et al. Genes Dev. 14, 521–535 (2000).
Cai, S., Han, H.J. & Kohwi-Shigematsu, T. Nat. Genet. 34, 42–51 (2003).
Dekker, J. Nat. Methods 3, 17–21 (2006).
Horike, S., Cai, S., Miyano, M., Cheng, J.F. & Kohwi-Shigematsu, T. Nat. Genet. 37, 31–40 (2005).
Gasser, S.M. Science 296, 1412–1416 (2002).
About this article
Wiley Interdisciplinary Reviews: Systems Biology and Medicine (2016)
Seminars in Cancer Biology (2013)
Comparative Analysis of Genome Sequences of the Th2 Cytokine Region of Rabbit (Oryctolagus cuniculus) with those of Nine Different Species
Immunology and Immunogenetics Insights (2011)
Comparative analysis of cis-regulation following stroke and seizures in subspaces of conserved eigensystems
BMC Systems Biology (2010)