Two recent studies show that some ATP-dependent chromatin remodeling complexes are subject to control by a surprising set of regulators — phosphoinositols. These studies extend earlier observations in mammalian cells and suggest that second messengers are regulators of chromatin remodeling.
Because the compaction of DNA into chromatin plays a role in gene transcription, there is a need to understand the regulation of chromatin structure to fully understand transcriptional regulation. Two broad classes of chromatin regulating activities have been described — enzymes that covalently modify histones, such as acetyltransferases and methylases, and ATP-dependent chromatin remodeling complexes, such as the Swi/Snf complex1. However, these complexes appear to have no intrinsic DNA sequence specificity. How, then, are these regulators targeted to chromatin? Recent evidence has shown that chromatin-remodeling enzymes are targeted to unique sites in the genome by interacting with sequence-specific DNA-binding proteins2. This model is neat and intuitive. However, with noted exceptions3, few studies have investigated other potential regulators of the chromatin remodeling factors. Two reports by Shen et al.4 and Steger et al.5 in a recent issue of Science now show that ATP-dependent chromatin remodeling activities are subject to regulation by a group of second messengers, inositol polyphosphates.
In vivo and in vitro ...
The role of inositol polyphosphates in regulating the activity of the Swi/Snf complex was identified through two fundamentally different experimental approaches (Fig. 1). Steger et al.5 conducted a genetic screen for regulators of PHO5 transcription in yeast and demonstrate that the ARG82/IPK2 inositol kinase, a critical enzyme in the production of the second messengers IP4/IP5 (inositol phosphate is abbreviated as IP; the subscripted numbers indicate the number of phosphates on the inositol moiety), is required for induction of the phosphate-responsive genes PHO5 and PHO84 in vivo5. Their data show that PHO5 transcription requires the Swi/Snf complex, as well as the Swi/Snf–like INO80 complex, and that these complexes, together with the ARG82/IPK2 inositol kinase, are required for the increased chromatin accessibility seen at the PHO5 locus upon transcriptional induction. To identify the step at which inositol phosphates function in the regulation of these complexes, Steger et al.5 demonstrate that the transcription factor PHO4, which is required for targeting chromatin remodeling complexes to the PHO5 and PHO84 promoters, continues to bind its cognate sequences in mutants where the ARG82 gene has been deleted. Thus, the defect is downstream of transcription factor activation and PHO4 promoter binding. However, INO80 and Swi/Snf complexes, which are recruited to the PHO84 promoter upon transcriptional activation in wild type yeast, do not associate with the promoters in the ARG82 deletion mutants. Thus, there is a defect in either the recruitment or the stable binding of INO80 and Swi/Snf complexes to the target promoter.
Shen et al.4, on the other hand, used an in vitro approach to study the regulation of Swi/Snf–like complexes. Intrigued by the observation that the promoter for the inositol metabolism gene, INO1, is regulated by several ATP-dependent chromatin remodeling complexes, they tested various inositol polyphosphates for their effects on chromatin remodeling activities in vitro. The results reveal that chromatin remodeling activity, as well as ATPase activity, is modulated by inositol phosphates in vitro. Interestingly, distinct complexes respond differently to these molecules. For example, INO80 is inhibited by IP6 and is unaffected by IP4, whereas Swi/Snf function is not affected by IP6 but is stimulated by IP4. Shen et al.4 also show that INO1 transcription requires the formation of inositol polyphosphates in vivo: INO1 transcription is repressed in mutants (including the ARG82 deletion strain) defective for formation of IP4 and IP5.
The complementary approaches taken by these two groups overlap in suggestive ways. For example, Steger et al.5 demonstrate that transcription of PHO5 requires either IP4 or IP5, because mutations blocking the pathways leading to IP6, PP-IP4 and PP-IP5 (pyrophosphate-inositol phosphates) have no effect on transcription. Shen et al.4 find that IP4 and IP5 have little effect on INO80 function in vitro, but both stimulate Swi/Snf function. These results suggest that a major effect of IP4 and IP5 at phosphate-responsive promoters (including, possibly, their effects on INO80 binding to the promoters) may be mediated through the Swi/Snf complex.
The two groups studied distinct functions of the complexes in question. While Shen et al.4 find that in vitro chromatin-remodeling activity and ATPase activity of Nurf and INO80 are modulated by inositol polyphosphates, Steger et al.5 show that stable association of Swi/Snf and INO80 at promoters depends on IP4/IP5 production. Taken together, these results suggest that chromatin remodeling may be required for continuous association of complexes with target promoters.
The finding that soluble inositol phosphates are involved in chromatin remodeling furthers previous work on nuclear inositides (for a review, see ref. 6). In particular, the two papers in Science4,5 confirm and extend a previous study showing that, in vitro, phosphoinositides regulate the function of the Swi/Snf-like BAF complex from T lymphocytes3 (Fig. 2). The T-cell receptor (TCR) mediates signals that direct the development of T lymphocytes. This process requires BAF complexes, which regulate the expression of genes essential for T-cell development7. TCR signaling modulates PIP2 levels in the cell and leads to nuclear retention of the BAF complexes3. Furthermore, PIP2 regulates the association of BAF complexes with crude chromatin fractions in vitro3. Thus, modulation of PIP2 levels by TCR signaling might regulate inositide levels in the nucleus, which in turn leads to chromatin association and targeting of Swi/Snf–like complexes to genes essential for T-cell development (Fig. 2). A mechanism for this inositide-dependent targeting was suggested by recent results showing that BAF complexes bind PIP2 directly through the Brg1 subunit8. In this case, PIP2 binding led to actin filament assembly in vitro8. (It should be noted that vertebrate BAF complexes and the Drosophila BAF complex contain actin; the yeast Swi/Snf complex does not, but INO80 complexes do contain actin.) Given that different inositol phosphates regulate chromatin remodeling complexes differently4, the results in vertebrate and yeast cells may herald a new appreciation for the roles of both soluble and membrane bound phosphoinositides in the control of chromatin remodeling.
A role for soluble phosphoinositides in the control of transcription had actually been suggested earlier9. In that study, IPK2, which is responsible for the phosphorylation of IP3 and production of IP4 and IP5, was found to be identical to ARG82, a transcriptional regulator involved in the transcription of arginine-responsive genes9. Although the precise functions of the transcriptional activity and the inositol kinase acitivity are not clear at the present time9,10, it will be interesting to see what genes require the kinase activity of ARG82/IPK2 and why two disparate functions appear to be linked in one protein.
Notably, ARG82/IPK2 is also required for efficient mRNA export11. However, mRNA export requires IP6. The role of inositol phosphates in general nuclear processes makes it tempting to speculate that the regulation of nuclear inositides may allow the cell to rapidly shift major patterns of cellular physiology in response to environmental stimuli.
Small molecule control of chromatin structure
The discovery of small-molecule regulators of chromatin-remodeling activity complicates our understanding of chromatin regulation but may help explain some puzzling results in the literature. For example, several whole-genome localization studies of chromatin-modifying proteins have concluded that these proteins are bound to the promoters of many genes at which they are not active. In one study, Grunstein and collaborators have shown that the histone deacetylase RPD3 is bound to highly expressed genes12, despite the fact that deletion of RPD3 has no effect on the transcription of many of these genes13. Thus, RPD3 may be associated with these promoters in an inactive form. Further observations from Struhl and collaborators14 have shown that the ATP-dependent chromatin remodeling complex RSC, which is generally recruited to promoters upon activation, is bound to a number of promoters prior to their transcriptional activation.
The scenario emerging from these observations is that chromatin-modifying enzymes may associate with rapidly inducible or repressible promoters prior to activation or repression, and may somehow be kept in an inactive state at these promoters. Upon transcriptional induction (or repression), the production of small signaling molecules such as IP4 and IP5 could provide a means of rapidly modulating the transcription of target genes. Interestingly, many of the genes to which RPD3 binds but remains inactive are rapidly repressed when yeast are exposed to environmental stresses15. Perhaps stress modulates the level of some signaling molecule that relieves RPD3 from its inactive state at the promoters of highly expressed genes. The identity of such a signaling molecule — if indeed it exists — is, of course, a mystery. However, rapidly regulatable promoters clearly display pre-localization of chromatin-modifying proteins. Coupled with evidence that small molecules can regulate chromatin-remodeling enzymes, the data suggest that the controlled synthesis of small molecule regulators of chromatin modifiers may be one of the means whereby cells could rapidly modulate gene expression in response to environmental stimuli.
Narlikar, G.J., Fan, H.Y. & Kingston, R.E. Cell 108, 475–487 (2002).
Cosma, M.P., Tanaka, T. & Nasmyth, K. Cell 97, 299–311 (1999).
Zhao, K. et al. Cell 95, 625–636 (1998).
Shen, X., Xiao, H., Ranallo, R., Wu, W.H. & Wu, C. Science. 299, 112–114 (2003).
Steger, D.J., Haswell, E.S., Miller, A.L., Wente, S.R. & O'Shea, E.K. Science. 299, 114–116 (2003).
Divecha, N. & Irvine, R.F. Cell 80, 269–278 (1995).
Chi, T.H. et al. Nature 418, 195–199 (2002).
Rando, O.J., Zhao, K., Janmey, P. & Crabtree, G.R. Proc. Natl. Acad. Sci. USA 99, 2824–2829 (2002).
Odom, A.R., Stahlberg, A., Wente, S.R. & York, J.D. Science 287, 2026–2029 (2000).
Dubois, E., Dewaste, V., Erneux, C. & Messenguy, F. FEBS Lett. 486, 300–304 (2000).
York, J.D., Odom, A.R., Murphy, R., Ives, E.B. & Wente, S.R. Science 285, 96–100 (1999).
Kurdistani, S.K., Robyr, D., Tavazoie, S. & Grunstein, M. Nat. Genet. 31, 248–54 (2002).
Bernstein, B.E., Tong, J.K. & Schreiber, S.L. Proc. Natl. Acad. Sci. USA 97, 13708–13713 (2000).
Ng, H.H., Robert, F., Young, R.A. & Struhl, K. Genes Dev. 16, 806–819 (2002).
Gasch, A.P. et al. Mol. Biol. Cell 11, 4241–4257 (2000).
Inokuchi, S. & Imboden, J.B. J. Biol. Chem. 265, 5983–5989 (1990).
Bultman, S. et al. Mol. Cell 6, 1287–1295 (2000).
Reyes, J.C. et al. EMBO J. 17, 6979–6991 (1998).
About this article
Cite this article
Rando, O., Chi, T. & Crabtree, G. Second messenger control of chromatin remodeling. Nat Struct Mol Biol 10, 81–83 (2003). https://doi.org/10.1038/nsb0203-81
Biochemical Pharmacology (2019)
Cancer and Metastasis Reviews (2016)
Repression of Transcription by WT1-BASP1 Requires the Myristoylation of BASP1 and the PIP2-Dependent Recruitment of Histone Deacetylase
Cell Reports (2012)
Trends in Cell Biology (2010)
Biochemical Society Transactions (2009)