Gene regulation

Local or global?

Article metrics

Modifications to histone proteins were thought to act only locally to control the expression of single genes. But the finding of such changes across the whole genome brings that view into question.

The US congressman Thomas 'Tip' O'Neill once famously said: “All politics is local”. In the same vein, environmental groups have adopted the slogan “Think globally, act locally”. In biology, a similar sentiment is thought to be relevant to gene expression. Here, it is held that changes in the structure of chromatin — a compact, compressed form of DNA — occur only 'locally', allowing the regulation of individual genes. But it seems that this picture is too simplistic, for, on page 495 of this issue1, Vogelauer and colleagues describe their discovery of genome-wide, 'global' chromatin modifications. The functions of these modifications remain something of a mystery.

The basic structural unit of chromatin is the nucleosome, which consists of a stretch of DNA wrapped tightly around a core of histone proteins. This compressed form makes it hard for transcription factors, which regulate gene expression, to gain access to target genes in the DNA. However, enzymes that add or remove phosphate, acetyl or methyl groups from histone 'tails'2,3 somehow affect chromatin structure in such a way that genes can be turned on or off. The best characterized of these enzymes, which in effect act as transcription factors themselves, are the histone acetyltransferases (HATs), which add acetyl groups to the histone tails, and their opposite numbers, the histone deacetylases (HDACs). In a commonly held view, these enzymes alter gene expression by acting locally, gene by gene, to modify histone proteins near the regulatory 'promoter' regions of the genes.

Three observations led to this model. The first nuclear HATs4 and HDACs5 to be identified were known previously as promoter- specific transcription factors. Then it was discovered that both HATs6,7 and HDACs8,9,10,11,12,13 are components of multiple-subunit complexes that are recruited to specific promoters by interaction with DNA-bound transcriptional activators and repressors. And early findings14,15,16 suggested that both HATs and HDACs are strictly limited in their method of action. They seem to work on histone proteins in only one or two nucleosomes, found next to a 'TATA box' — the site in a promoter region to which the main gene-transcribing enzyme, RNA polymerase II, binds14,15,16. A pleasing model emerged in which chromatin-modifying factors, such as HATs and HDACs, are recruited to specific promoters to regulate transcription directly (Fig. 1).

Figure 1: A commonly held view of gene regulation.

Local modification of histone proteins, around which DNA is wrapped, is thought to allow the regulation of gene expression on a gene-by-gene basis. A histone acetyltransferase (HAT) adds acetyl groups (Ac) to histone proteins near the TATA box — a DNA sequence near the point in a gene from which transcription starts. This somehow allows the target gene to be transcribed into messenger RNA. The HAT is part of a protein complex that is recruited by physical interactions with a DNA-bound transcriptional activator. RNA polymerase II and the TATA-binding protein (TBP) are two of the key protein complexes involved in transcription.

But Vogelauer et al.1 have discovered that this is not the whole story. They have found that the best characterized of the histone-modifying enzymes have a more global role, apparently allowing a rapid cycle of acetylation and deacetylation throughout the genome. The authors use a powerful technique, chromatin immunoprecipitation, that allows in vivo examination of proteins and their modifications at specific points in chromatin. They find that acetylation occurs ubiquitously throughout a large chunk of the genome of the baker's yeast Saccharomyces cerevisiae. The level of global acetylation is slightly higher than that seen at the ends of chromosomes (telomeres), which are transcriptionally silent. Moreover, at every point examined in the yeast genome, the amount of acetylation could be lowered by deleting a particular HAT, or raised by deleting a particular HDAC. The implication is that the state of acetylation of a genome is in constant flux. It is higher at promoter regions, but this seems to be superimposed over the global pattern.

A possibly related phenomenon was recently reported by Cohen et al.17, who used whole-genome microarray analysis in yeast to examine the relationship between the expression of 'linked' genes. They uncovered interesting correlations in the patterns of expression of adjacent genes that are normally regulated in different ways. This suggests that patterns of expression of large domains of genes underlie gene-specific patterns.

Two possible functions for the 'base' level of acetylation discovered by Vogelauer et al.1 come to mind. The first is that it helps to turn transcription on or off, priming or dampening transcription (Fig. 2). This would mean that chromatin, like an idling truck, is always ready to go. Just a bit more acetylation at a particular gene's promoter is needed to turn that gene on. Then only a little deacetylation is needed to return it quickly to the idling state. A role in priming would help to answer the thorny question of how transcriptional activators penetrate the chromatin barrier to reach target genes. However, Vogelauer et al. propose, and provide some evidence in support of, a dampening model only — that the global acetylation pattern allows chromatin to return quickly to the idling state after activation. But of course both may be true, and could be working together.

Figure 2: Processes that might be regulated by the base level of histone acetylation discovered by Vogelauer et al.1 throughout the Saccharomyces cerevisiae genome.

Such global acetylation, brought about by the histone acetyltransferases and histone deacetylases (not shown), might somehow loosen the entire genome, allowing access of: a, the DNA-replicating enzyme DNA polymerase II; or b, DNA-repair enzymes. c, Vogelauer et al.1 offer evidence in support of another model. They propose that the base level of acetylation amounts to a transcriptional 'ground state' (the pale green shading denotes low RNA polymerase II activity), allowing target genes to be turned on and off rapidly. Ac, acetylation.

A second hypothesis is that these global modifications have a totally different role, perhaps not related to transcription at all. For example, when DNA is replicated as part of the cell-division cycle, it may be necessary to 'loosen' all the chromatin to allow the replication machinery access to the entire genome (Fig. 2). In this case, one would predict that the global acetylation pattern changes during the cell cycle. Perhaps higher levels of global acetylation are seen during the DNA-replication phase, with lower levels occurring during cell division itself, when DNA is more condensed. Another role for global acetylation might be to allow the DNA-repair machinery access to the genome, to 'shop' for mutations that need fixing ( Fig. 2). Other functions can be envisaged.

An unanswered question in this field is how changes to histone proteins translate into changes in the level of gene expression. Two possibilities are in favour. First, changes to histone proteins may lead directly to physical changes in nucleosome or chromatin structure, making it easier for other transcription factors to gain access to target genes. Alternatively, the modifications may create a special pattern that provides a surface to which relevant factors can bind. Investigation of the global modifications discovered by Vogelauer et al.1 may offer further insight into this problem.

During the past 40 or so years of molecular biology, researchers have generally used single genes to study genomic regulation, in part because individual genes are easier to access. The whole genome is much more complex, and much harder to study. No doubt much research into gene regulation will continue to look at the effects of local changes. But, with the advent of genomics and proteomics, as well as techniques such as that used by Vogelauer et al., our view of gene regulation is becoming less restricted.


  1. 1

    Vogelauer, M., Wu, J., Suka, N. & Grunstein, M. Nature 408, 495–498 (2000).

  2. 2

    Van Holde, K. E. in Chromatin: Springer Series in Molecular Biology (ed. Rich, A.) 111–148 (Springer, New York, 1988).

  3. 3

    Bradbury, E. M. Bioessays 14, 9–16 ( 1992).

  4. 4

    Brownell, J. E. et al. Cell 84, 843–851 (1996).

  5. 5

    Taunton, J. et al. Science 272, 408–411 (1996).

  6. 6

    Grant, P. A. et al. Genes Dev. 11, 1640– 1650 (1997).

  7. 7

    Utley, R. T. et al. Nature 394, 498–502 (1998).

  8. 8

    Laherty, C. D. et al. Cell 89, 349–356 (1997).

  9. 9

    Zhang, Y. et al. Cell 89, 357–364 (1997).

  10. 10

    Kadosh, D. & Struhl, K. Cell 89, 365 –371 (1997).

  11. 11

    Nagy, L. et al. Cell 89, 373–380 (1997).

  12. 12

    Heinzel, T. et al. Nature 387, 43–48 (1997).

  13. 13

    Alland, L. et al. Nature 387, 49–55 (1997).

  14. 14

    Kuo, M.-H et al. Genes Dev. 12, 627–639 (1998).

  15. 15

    Rundlett, S. E. et al. Nature 392, 831–835 (1998).

  16. 16

    Kadosh, D. & Struhl, K. Mol. Cell. Biol. 18, 5121–5127 (1998).

  17. 17

    Cohen, B. A. et al. Nature Genet. 26, 183– 186 (2000).

Download references

Author information

Correspondence to Shelley L. Berger.

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.