The list of proteins whose activity is controlled by transport into different subcellular compartments keeps growing. The latest additions are two regulators of gene expression.
Histone proteins serve double duty in the nucleus. First, they compress DNA and DNA-bound proteins into a compact structure called chromatin. Second, they have a say in which genes are expressed. Specifically, it is the way in which histones are labelled with different chemical groups that is important in controlling gene transcription. Acetyl (COCH3) groups are one type of label, and are added and removed by enzymes called histone acetylases and deacetylases1,2. These regulatory enzymes must themselves be kept under control — but how? By looking at how muscle cells differentiate to form muscle fibres in cell culture, McKinsey and colleagues ( page 106 of this issue3) provide some answers. Unusually, they find that two types of histone deacetylase are moved out of the nucleus during muscle differentiation, in a process that ultimately is controlled by calcium ions — ubiquitous chemical messengers.
The process of converting an unspecialized muscle cell to a skeletal muscle fibre (myogenesis) is a complicated business. The key point is that an array of muscle-specific genes (such as that encoding the myosin protein, a molecular motor) must be turned on. This requires the cooperation of several proteins4 including the transcription factor MyoD and myocyte enhancer factor-2 (MEF2).
Clearly, the activity of these factors must be controlled so they do not wantonly activate muscle-specific genes. This is where the histone deacetylases come in. Several groups have reported that MEF2 and histone deacetylases interact functionally and physically5,6,7,8,9, resulting in vitro in the repression of MEF2. In both T cells and early muscle cells in culture, MEF2 responds to Ca2+ signals by dissociating from histone deacetylases7,10. Moreover, histone deacetylase-4 can shuttle into and out of the nucleus5, and histone deacetylase-4 and -5 can be kept outside the nucleus by proteins of the 14-3-3 family11. McKinsey et al.'s triumph is to link these seemingly disparate observations, and to show their relevance to muscle differentiation and gene expression3. Their results also provide the first example of the 'spatial' regulation (by export from the nucleus) of a chromatin regulator.
The authors report that when early muscle cells are encouraged to differentiate by the removal of serum from their culture medium, histone deacetylase-5 exits the nucleus. MEF2 is left behind, where it promotes the transcription of genes essential for muscle differentiation. This response is thought to be mediated largely by Ca2+ signals generated at the cell surface. In light of this fact and earlier results7, McKinsey et al. set out to find a Ca2+-regulated enzyme that might induce the removal of histone deacetylase-5 from the nucleus. The answer turned out to be two Ca2+/calmodulin-dependent protein kinases (enzymes that add phosphate groups to other proteins). When added to muscle cells in culture, the constitutively active form of each of these enzymes (known colloquially as CaMKI and CaMKIV) induces the export of added histone deacetylase-5 (and -4) from the nucleus. The response is selective to histone deacetylase-4 and -5 — added CaMKI did not have the same effect on histone deacetylase-1 — and seems to require the phosphorylation of the histone deacetylases in question.
Previous work7 showed that only Ca2+ and a nuclear extract are required to induce the dissociation of histone deacetylase-4 from MEF2. So, putting all these results together, we can suggest how the pathway in early muscle cells might work (Fig. 1). Binding of an extracellular molecule to its receptor on the cell surface leads to an increase in cellular Ca2+. This then activates Ca2+/calmodulin-dependent kinases, which in turn phosphorylate histone deacetylase-4 and -5. These enzymes then detach themselves from MEF2, exposing their nuclear export signal (also discovered by McKinsey et al .). This signal consists of a sequence of amino acids that acts as an address label, allowing the proteins to be sent out of the nucleus. Alone in the nucleus, MEF2 works together with muscle-specific transcription factors to enhance the expression of muscle-specific genes, allowing differentiation of the cell into a muscle fibre. This all seems physiologically plausible, especially as the repression of MEF2 by a constitutively nuclear histone deacetylase-5, with mutations in its sites of phosphorylation by CaMKI, cannot be relieved by CaMKI (or, apparently, by Ca2+, although this was not shown directly).
This is all well and good, but there are still details of the signalling pathway to be resolved. The most unintuitive part of the pathway is that removal of serum induces a Ca2+-dependent process. Serum contents such as platelet-derived growth factor normally stimulate cells, so one might think that removing serum would terminate stimulation. One possible explanation is that a molecule that is activated by serum withdrawal is important — perhaps insulin-like growth factor-1, which induces muscle differentiation12, or phenylephrine, which invokes Ca2+ signalling and the dissociation of histone deacetylases from MEF2 (ref. 9). However, McKinsey et al.'s results3 do not establish that a Ca2+ stimulus alone is sufficient for histone deacetylase-4 and -5 to exit the nucleus. This question could be easily answered by using drugs, such as ionomycin, that cause selective Ca2+ influx at physiological levels and rates.
Another question is how relevant these mechanisms are to muscle differentiation in vivo. This may become clear only with the use of mice in which mutant histone deacetylase genes, encoding mutations in the phosphorylation sites, are inserted into the mice's own histone deacetylase gene. McKinsey et al.'s results predict that the histone deacetylase encoded by such mutant genes would not be able to respond to Ca2+ signalling. It might therefore block the differentiation of skeletal muscle (and, likewise, the programmed cell death of T cells, during which MEF2 also dissociates from histone deacetylases). But whatever the outcome of these in vivo studies, the new work3 is important not least because it adds to the growing body of evidence13 for the regulation of chromatin regulators by signals that begin at the cell membrane.
Brownell, J. E. et al. Cell 84, 843–851 (1996).
Taunton, J., Hassig, C. A. & Schreiber, S. L. Science 272, 408– 411 (1996).
McKinsey, T. A., Zhang, C.-L., Lu, J. & Olson, E. N. Nature 408, 106–111 (2000).
Davis, R. L., Weintraub, H. & Lassar, A. B. Cell 51, 987– 1000 (1987).
Miska, E. A. et al. EMBO J. 18, 5099–5107 (1999).
Wang, A. H. et al. Mol. Cell. Biol. 19, 7816– 7827 (1999).
Youn, H. D., Grozinger, C. M. & Liu, J. O. J. Biol. Chem. 275, 22563– 22567 (2000).
Lemercier, C. et al. J. Biol. Chem. 275, 15594– 15599 (2000).
Lu, J. et al. Proc. Natl Acad. Sci. USA 97, 4070– 4075 (2000).
Youn, H. D., Sun, L., Prywes, R. & Liu, J. O. Science 286, 790–793 (1999).
Grozinger, C. M. & Schreiber, S. L. Proc. Natl Acad. Sci. USA 97, 7835–7840 (2000).
Musarò, A. et al. Nature 400, 581–585 (1999).
Zhao, K. et al. Cell 95, 625–636 (1998).
About this article
The Interplay between the Glucocorticoid Receptor and Nuclear Factor-κB or Activator Protein-1: Molecular Mechanisms for Gene Repression
Endocrine Reviews (2003)
Induction and superinduction of growth arrest and DNA damage gene 45 (GADD45) α and β messenger RNAs by histone deacetylase inhibitors trichostatin A (TSA) and butyrate in SW620 human colon carcinoma cells
Cancer Letters (2002)
Journal of Biological Chemistry (2001)