PerspectiveGene Regulation

NFAT Is Well Placed to Direct Both Enhancer Looping and Domain-Wide Models of Enhancer Function

Science Signaling  01 Apr 2008:
Vol. 1, Issue 13, pp. pe15
DOI: 10.1126/stke.113pe15


Nuclear factor of activated T cells (NFAT) plays a central role in activating gene expression at the level of chromatin structure. A study now reveals that NFAT may also help to organize chromatin domains and enable enhancer-promoter communication. In activated T cells, inducible intrachromosomal looping occurs between the tumor necrosis factor–α (TNF-α) gene promoter and two NFAT-dependent enhancers located at –9 kb and +3 kb. This topology places the TNF-α gene and the adjacent lymphotoxin (LT) genes in separate loops, thereby allowing independent regulation of the TNF-α gene within a multigene locus. These findings build on other studies that indicate that NFAT is intimately associated with activities that disrupt nucleosomes within enhancers and mobilize nucleosomes across extensive chromatin domains linking enhancers and promoters. Taken together, these studies highlight NFAT as a factor that creates a chromatin environment that is permissive for both the recruitment and the clustering of factors that control transcription at promoters and enhancers.

The fundamental mechanisms that allow transcriptional enhancers to activate gene expression from a distance have been a lively subject of debate ever since they were first described in the early 1980s. At the simplest level, the various models proposed include (i) direct communication between enhancers and promoters through "DNA looping," and (ii) recruitment of complexes at enhancers that then "track" along the chromosome until they encounter a promoter (15). More complex variations of these models include (iii) spreading or linking of complexes along the chromosome, and (iv) "facilitated tracking," whereby complexes remain in contact with the enhancer while tracking or hopping along the DNA and progressively looping out the intervening DNA (13, 5). Because of the recent surge of studies utilizing the chromosome conformation capture (3C) (6) assay, the looping model is the most widely accepted mechanism for enhancer function (7, 8). Furthermore, even regulatory elements on different chromosomes can colocalize (9, 10). A study by Tsytsykova et al. (11) now suggests that nuclear factor of activated T cells (NFAT) is one of the transcription factors that enables looping between enhancers and promoters. In this article, I will discuss both this and additional evidence that suggest that NFAT is a major player in promoting multiple mechanisms of enhancer function.

NFAT family proteins are Ca2+-inducible transcription factors that activate expression of a wide range of immune response genes, especially in activated T cells (12, 13). NFAT-dependent promoters and enhancers rapidly undergo extensive inducible chromatin remodeling to form deoxyribonuclease I–hypersensitive sites (HSSs) (1317). It is often assumed that NFAT is the driving force behind this remodeling, and it is known that NFAT sites are alone sufficient to recapitulate inducible HSSs in a chromatin context (16). Indeed, chromatin remodeling may well be the primary function of NFAT elements, because even high-affinity NFAT sites are comparatively poor transcriptional activators in the absence of the collaborating transcription factors with which they normally associate (16). The transcription factor–activating protein 1 (AP-1) is the most common partner directly recruited by NFAT (12, 13, 16, 18), and at the level of HSSs, composite NFAT–AP-1 elements are very efficient for evicting nucleosomes (16, 19). Aiding this process, NFAT–AP-1 complexes recruit both histone acetyltransferases (HATs) and the ATP–dependent SWI-SNF family of chromatin remodelers (2023), which together provide all of the ingredients needed both to modify and to rearrange nucleosomes.

The study by Tsytsykova et al. (11) has revealed new insights into NFAT-dependent mechanisms that govern induction of the tumor necrosis factor–α (TNF-α) gene in activated T cells. In their search for distal regulatory elements, the Goldfeld group also had the opportunity to take a new look at an old question; they investigated how the TNF-α gene is regulated independently within a densely packed multigene locus. The mouse TNF-α gene sits snugly between the lymphotoxin (LT)-α and LT-β genes. These three genes occupy just 12 kb embedded within the gene-rich major histocompatability complex locus (Fig. 1A). Using conventional approaches, the authors identified numerous HSSs spanning the TNF-LT locus and found that two of these elements, located 9 kb upstream and 3 kb downstream of the TNF-α gene, both consist of arrays of conserved NFAT-binding sites. It is noteworthy that these two HSSs are present before activation but, nevertheless, function as inducible enhancers in activated T cells. Furthermore, the HSS+3 and HSS–9 enhancers and the TNF-α promoter all associate with each other in an inducible manner in activated T cells when analyzed by the 3C assay. This implies that these elements are all recruited to a single active chromatin hub (7) or transcription factory (4).

Fig. 1.

(A) Organization of the TNF-α–LT locus in T cells before stimulation, showing the locations of HSSs (arrows) and conserved enhancers located at the –9 and +3 kb HSSs. (B) Architecture of the TNF-α–LT locus after activation of TCR signaling pathways by phorbol myristate acetate (PMA) and the calcium ionophore ionomycin. The HSS–9 and HSS+3 enhancers, which encompass conserved NFAT sites, are both recruited to an active chromatin hub together with the TNF-α promoter. The LT gene promoters are not recruited to this active site. The chromatin hub is most likely a transcription factory that contains many chromatin-remodeling factors in addition to transcription factors and RNA polymerase.

The concept of an active chromatin hub that is created at sites of transcription is now widely accepted, but this was not always so. The idea that active genes are anchored within discrete transcription factories was originally pioneered by Peter Cook (24, 25). It was only revealed in 2002 that active promoters and their enhancers colocalize within the nucleus. The confirmation of the looping model was made possible with the advent of the very elegant RNA TRAP (tagging and recovery of associated proteins) method devised by Peter Fraser (26), at the same time as the introduction of the 3C chromatin ligation method (6) used by Wouter de Laat (27). The RNA TRAP method is an RNA in situ hybridization procedure that tags distal regulatory elements that lie in close proximity to a transcribed gene. The 3C method involves formaldehyde cross-linking of chromatin segments encompassing regulatory elements that lie in close proximity. In the 3C procedure, cross-linked chromatin is digested with restriction enzymes and then religated to reveal DNA ends that have been brought close enough together to enable religation. Richard Flavell has also used the 3C assay to show that regulatory elements and HSSs spanning 120 kb of the T helper cell type 2 (TH2) cytokine gene locus all colocalize in T cells (8). The Flavell studies found that although most of these interactions exist in a wide range of cell types before activation, some interactions at the locus control region (LCR) are either calcium-inducible or only occur in activated T cells. In contrast, the TNF-α study (11) finds that all of the interactions between elements within the TNF-α–LT locus are dependent on activation of T cell receptor (TCR) signaling pathways. Furthermore, these interactions create a nuclear architecture within the TNF-α–LT locus that joins together all of the elements that control the TNF-α gene, but leaves the LT genes and other HSSs looped out in different chromatin domains, as depicted in Fig. 1. This very neatly provides a mechanism that allows the TNF-α gene to be regulated independently of the LT genes. In addition, this structure also places the 5′ and 3′ ends of the TNF-α gene in close proximity to each other in a circular conformation that is believed to promote recycling of RNA polymerase (2830). This may represent the first example of a mammalian gene that is thus circularized in an inducible manner.

For the most part, little is known regarding the specific factors that mediate direct enhancer-promoter communication within the genome. The Flavell studies of the TH2 locus revealed that the transcriptional factor GATA-3 and the signal transducer and activator of transcription 6 (STAT6) (8) are important for establishing enhancer-promoter communication in T cells. The Tsytsykova study suggests, but does not yet prove, that NFAT also plays a role in enabling long-range enhancer-promoter communication by a DNA-looping mechanism. The Flavell group had already provided evidence for this concept by showing that LCR-promoter interactions are induced in 3T3 cells by calcium signaling pathways in the presence of GATA-3 (8), which implies that NFAT is required for this interaction. At the TNF-α locus, NFAT is clearly present at the sites of interaction, but that does not necessarily mean that NFAT itself does the engaging, and its function, alternatively, could be to create a permissive environment. It would be oversimplistic to assume that the functions of the HSS+3 and HSS–9 enhancers are driven solely by NFAT, and it is likely that additional factors also participate in enhancer function and communication.

Whereas the current focus of most enhancer studies remains firmly fixed on the 3C assay, other approaches have been somewhat neglected. Earlier studies of enhancers suggested that they create local chromatin accessibility, which is then propagated over long distances (2). Furthermore, there is now evidence that DNA looping is not the only way that NFAT contributes to the long-range reorganization of chromatin. It was recently demonstrated that an NFAT-dependent enhancer directs inducible chromatin remodeling across several kilobases of DNA, which includes the whole region between the enhancer and promoter at the granulocyte-macrophage colony-stimulating factor (GM-CSF) locus (19). This effectively means that some aspects of the "spreading" model of enhancer function are also at work, at least for enhancers located within a few kilobases of their promoters. Before activation, the GM-CSF locus exists as a highly regular array of sequence-specific positioned nucleosomes, typical of condensed chromatin. GM-CSF enhancer activation is followed by two types of chromatin reorganization: (i) nucleosomes within the enhancer are evicted and replaced by enhanceosome-like complexes encompassing NFAT sites; and (ii) nucleosomes are mobilized over at least 5 kb of flanking sequences, which creates a randomized nucleosome array that represents an open chromatin structure (19) (Fig. 2). From all of the 3C studies performed to date, it would be reasonable to assume that direct interactions are also taking place between the GM-CSF promoter and enhancer, looping out 3 kb of intervening DNA, as suggested in the speculative model presented in Fig. 2. Hence, the various models of enhancer function may be best seen as concurrent models rather than alternative models. It is also anticipated that direct interactions between promoters and enhancers will only occur when the stage has been set and a permissible environment exists. It must also be considered that the colocalization of enhancers and promoters at transcription factories does not necessarily depend on direct contacts between them.

Fig. 2.

Model of the nucleosomal organization of the human GM-CSF promoter and enhancer before and after activation of TCR signaling pathways. Before activation, the locus exists as a regular array of positioned nucleosomes. Upon activation, two enhanceosome-like complexes containing NFAT and AP-1 assemble on the enhancer, displacing two nucleosomes. This is followed by extensive reorganization of nucleosomal architecture across the entire 3-kb region that separates the enhancer from the promoter. It is likely that HATs and additional remodeling complexes such as SWI-SNF are also recruited by NFAT–AP-1 complexes and are responsible for the long-range chromatin remodeling. The GM-CSF promoter also exists as a nucleosome-free region in activated T cells. On the basis of the precedent of the many gene loci now studied by the 3C assay, it is highly likely that the promoter and enhancer are both recruited to a single active chromatin hub. Ets1, Runx1, and Sp1 are transcription factors.

It still remains unclear how promoters and enhancers find each other, but in the dynamic three-dimensional space in which they exist, they may never be far away from each other. In at least some cases, this process of coming together may be directly facilitated. In the example of the hepatocyte nuclear factor–4α locus, complexes recruited at the enhancer may track along DNA, which could lead to the final encounter between the enhancer and promoter (31). In the case of the β-globin locus, histone-methylating complexes are recruited to the LCR and then spread across the entire locus, which leads to histone H3 lysine 4 methylation within the βmaj-globin gene located 38 kb downstream (32). This may, in turn, help to create a dynamic state that enables LCR-promoter communication of the type detected by the 3C and RNA TRAP assays (26, 27). The specific role that NFAT plays in the complex process of locus activation remains unclear, but if nothing else, it is a great facilitator that initiates the first essential step of creating an accessible chromatin environment.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
View Abstract

Navigate This Article