Gene transcription

Two worlds merged

Article metrics

Why would two distant genes — on separate chromosomes and from different nuclear locations — unite in response to signals for gene expression? They might be seeds for the formation of transcriptional hubs.

Gene transcription occurs largely at the submicroscopic scale. So although microscopic analysis of nuclear architecture has implicated various structures in this process1, it has lacked the power to unravel the role that higher-order organization of chromatin (complexes of DNA and histone proteins) has in the expression of individual genes. Consequently, it has been difficult to combine whole-cell approaches such as microscopy with the molecular and biochemical techniques2 that are primarily used to study gene expression. Fortunately, technological advances in both microscopy and molecular approaches are closing this gap. For instance, writing in Cell, Nunez et al.3 bring a refreshing concordance between these two types of method to show that gene expression associated with activation of the nuclear receptor ERα (oestrogen receptor-α) depends in part on a large-scale reorganization of the genome that involves interactions both within and between chromosomes.

Much of the recent progress in understanding how gene transcription factors interact with chromatin can be credited to 'ChIP-on-chip' technology4. For example, this method revealed that ERα interacts with a specific set of DNA sequences known as oestrogen-response elements5 (EREs), and that only some of these interactions enhance transcription. Other transcription-factor binding sites adjacent to EREs also seem to function cooperatively in facilitating transcription, indicating that various non-coding DNA sequences influence gene expression5.

For their analysis of chromatin reorganization in response to the activation of ERα transcription, Nunez et al. used many sophisticated approaches, one of which — chromatin-conformation capture — is also related to ChIP-on-chip. This technique allows the detection of long-range interactions between genes, even those on different chromosomes6. Specifically, the authors aimed to characterize the interaction between two genes regulated by the oestrogen hormone: TFF1 on chromosome 21 and GREB1 on chromosome 2. They find that, in response to hormone treatment, the chromosomal regions (loci) containing these genes physically reach out for each other.

Nunez and colleagues also find that genes at other oestrogen-regulated loci interact with each other, indicating that interchromatin pairing of DNA sequences is a common feature of ERα-mediated transcription. What's more, previous work using another approach — assessing the effect that restricting genes' movement within the nucleus has on their transcription — has shown that other functionally related genes, such as those with roles in immunity or cellular differentiation, also make interchromatin contacts3. So it seems that interchromatin interactions are not restricted to a particular cellular process.

Interchromatin contact between TFF1 and GREB1 depends on the binding of ERα to an ERE. Moreover, these interactions are also influenced by the activity of other proteins, including chromatin remodelling factors and proteins that bind to the cytoskeletal protein actin; this indicates that passive movement is not sufficient for chromosomal reorganization1. Nunez et al.3 add to the list of proteins known to mediate interchromatin interactions, showing that transcriptional coactivators such as SRC-3/AIB1, CBP, p300 and PBP are also involved in this process. Furthermore, they show that another co-regulatory protein — the ERα-associated demethylase enzyme LSD1, which does not influence interchromosomal gene interaction — is required for the subsequent colocalization of the gene pairs with nuclear speckles (structures composed of high concentrations of specific proteins involved in messenger-RNA splicing), thereby probably leading to mRNA maturation (Fig. 1).

Figure 1: Rambling genes.

a, Nunez et al.3 find that, in the absence of oestrogen, two genes (TFF1 and GREB1) that are activated by this hormone and found on different chromosomes, reside in different locations within the nucleus. b, On exposure to oestrogen, these genes make interchromatin contacts, facilitated by the nuclear receptor ERα, transcriptional coactivators (CoAs), actin and molecular motor proteins. c, Subsequently, the gene pair interacts with a nuclear speckle in an LSD1-dependent manner, creating a multiprotein complex that might act as a transcriptional hub for DNA transcription into mRNA.

Transcriptional coactivators are multiprotein complexes that, collectively, have diverse enzymatic activities7. So it is tempting to speculate that, together with LSD1, distinct enzymes of a coactivator complex mediate actin polymerization (which might facilitate gene movement within the nucleus), interchromatin interactions and chromatin interface with transcription 'factories' — structures that generate mRNAs at higher rates. Ultimately, it is likely that coactivators and other receptor-associated proteins work together to orchestrate the formation of an ERα-activated transcriptional supercomplex comprising ERα itself, coactivators, mRNA-splicing factors, transcription factories and two or more oestrogen-regulated genes (Fig. 1).

The possibility of ERα-regulated genes converging into distinct hubs gives rise to some interesting questions. How common is the above process among other genes regulated by nuclear receptors? What signals define how one gene finds its appropriate partner gene? Do these transcriptional hubs contain functionally related genes, or is their organization simply spatial? Could it be that sequences close to a gene, or possibly specific chromatin modifications at each ERα-regulated locus, provide proper pairing information?

Many EREs are located in stretches of DNA between genes, far away from known coding sequences. Do these sites also engage with interchromatin loci pairs, and, if so, to what extent do they influence the transcription of the partner locus? Do large-scale structural rearrangements induced by oestrogen, or other signals that activate oestrogen receptors, affect genome-wide expression of genes that are not directly regulated by ERα? These are some of the intriguing questions that remain. But with the availability of sophisticated techniques such as those used by Nunez et al.3, there is hope that the mysterious relationship between higher-order nuclear architecture, chromatin organization and gene expression will soon be revealed.


  1. 1

    Kumaran, R. I., Thakar, R. & Spector, D. L. Cell 132, 929–934 (2008).

  2. 2

    Speicher, M. R. & Carter, N. P. Nature Rev. Genet. 6, 782–792 (2005).

  3. 3

    Nunez, E. et al. Cell 132, 996–1010 (2008).

  4. 4

    Nègre, N., Lavrov, S., Hennetin, J., Bellis, M. & Cavalli, G. Methods Enzymol. 410, 316–341 (2006).

  5. 5

    Carroll, J. S. & Brown, M. Mol. Endocrinol. 20, 1707–1714 (2006).

  6. 6

    Simonis, M., Kooren, J. & de Laat, W. Nature Methods 4, 895–901 (2007).

  7. 7

    Lonard, D. M. & O'Malley, B. W. Mol. Cell 27, 691–700 (2007).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.