Modulation of ATP-Dependent Chromatin-Remodeling Complexes by Inositol Polyphosphates

See allHide authors and affiliations

Science  03 Jan 2003:
Vol. 299, Issue 5603, pp. 112-114
DOI: 10.1126/science.1078068


Eukaryotes use adenosine triphosphate (ATP)–dependent chromatin-remodeling complexes to regulate gene expression. Here, we show that inositol polyphosphates can modulate the activities of several chromatin-remodeling complexes in vitro. Inositol hexakisphosphate (IP6) inhibits nucleosome mobilization by NURF, ISW2, and INO80 complexes. In contrast, nucleosome mobilization by the yeast SWI/SNF complex is stimulated by inositol tetrakisphosphate (IP4) and inositol pentakisphosphate (IP5). We demonstrate that mutations in genes encoding inositol polyphosphate kinases that produce IP4, IP5, and IP6 impair transcription in vivo. These results provide a link between inositol polyphosphates, chromatin remodeling, and gene expression.

In eukaryotes, the SWI2/SNF2 family of ATP-dependent chromatin-remodeling complexes is widely used to regulate DNA accessibility for transcription. Four related classes of protein complexes (SWI2/SNF2, ISWI, Mi2, and INO80) use the energy of ATP hydrolysis to alter nucleosome architecture (1–3). Although there have been significant advances in understanding the mechanism and function of chromatin-remodeling complexes, the interaction of these complexes with cell signaling pathways has not been widely explored. One major mechanism for communicating environmental signals is the inositol signaling pathway. Activation of phosphatidylinositol-specific phospholipase C at the cell membrane leads to cleavage of phosphatidylinositol 4,5-bisphosphate (PIP2), generating secondary messengers inositol 1,4,5-trisphosphate (IP3), a regulator of calcium release and diacylglycerol (DAG), an activator of protein kinase C (4, 5). IP3 can undergo additional phosphorylation to IP4, IP5, or IP6, and di-phosphorylated derivatives (6). Recent advances have revealed multiple and varied functions for IP4, IP5, and IP6 in nucleic acid and viral metabolism (7–11).

The regulation of INO1, encoding inositol-1-phosphate synthase (12), by SNF2,ISW2, and INO80(13–16), encoding the core ATPases of three chromatin-remodeling complexes, prompted us to consider whether soluble inositol metabolites could influence ATP-dependent chromatin remodeling. We investigated this question by an in vitro nucleosome mobilization assay, which uses native gel electrophoresis to distinguish between nucleosomes at different locations on a DNA fragment. The Drosophila ISWI-containing complex NURF mobilizes reconstituted nucleosomes to favor one dominant position (N3) on hsp70 promoter DNA (17–19). We found that IP6 inhibits nucleosome mobilization by the NURF complex (Fig. 1A, see 40 μM and 100 μM), in the range of cellular IP6 levels (20–22). Control compounds (6) inositol hexasulfate (IS6), EDTA, EGTA, and IP3 showed little effect (Fig. 1A). We also tested IP4 and IP5 (500 μM) and found no effects for IP4 and some inhibition for IP5(23).

Figure 1

IP6 inhibits chromatin remodeling by NURF and ISW2. (A) Native polyacrylamide gel electrophoresis (PAGE) showing the effect of inositol polyphosphates on nucleosome mobilization. Positions of mononucleosomes (filled circles) are shown on the left. (N1+N2)/N3 ratios are given at the bottom; SD < 0.1. (B) Thin-layer chromatography analysis showing nucleosome (Nuc)-stimulated ATPase activity of rNURF. Percent ATP hydrolysis is shown at the bottom for all ATPase assays hereafter with standard deviation <1%. (C) Nucleosome mobilization by ISW2 complex. N4/N3 ratios are given at the bottom; SD < 0.1.

The nucleosome-stimulated ATPase activity of NURF was correspondingly inhibited by IP6 (19) (Fig. 1B), which suggests that inhibition by IP6 occurs, at least in part, through modulation of the ISWI ATPase. IP6 did not block the unrelated myosin ATPase or the ATPase activity of Fun30 (23), another SWI2/SNF2 family member (24). We tested the effect of IP6 on the yeast ISWI-containing chromatin-remodeling complex ISW2 (25). Yeast ISW2 mobilizes hsp70 nucleosomes from N4 to N3 positions. IP6 inhibited nucleosome mobilization (Fig. 1C) and nucleosome-stimulated ATPase activities of ISW2 (fig. S2A).

Yeast SWI/SNF mobilizes N1, N2, and N3 hsp70 nucleosomes to a novel position (N*) slightly above N2 (Fig. 2A, right). We found little effect of IP6 on yeast SWI/SNF (Fig. 2A, left). However, when SWI/SNF levels were reduced, we found that IP4 (1,4,5,6), the major IP4 isomer in wild-type yeast (26), and IP5 (500 μM) consistently stimulated nucleosome mobilization, whereas another isomer, IP4 (1,3,4,5) did not stimulate but was inhibitory (Fig. 2A, right). We also detected stimulation of SWI/SNF remodeling activity by IP4 (1,4,5,6) and IP5 (at 100 μM), whereas IP6 showed no stimulatory effects (23). The stimulatory concentrations are higher than steady-state estimates of IP4 and IP5 in yeast (26); this could be a limitation of the remodeling assay, or concentrations could vary locally or transiently in vivo. The ATPase activity of SWI/SNF was unaffected by IP4 or IP5 (Fig. 2B), which suggests that stimulation occurs through a different mechanism.

Figure 2

IP4 and IP5 stimulate chromatin remodeling by SWI/SNF. (A) Native PAGE showing nucleosome mobilization. N1/(N2+N*) ratios are given at the bottom; SD < 0.05. (B) ATPase activities of SWI/SNF.

We next analyzed the INO80 chromatin-remodeling complex, using mononucleosomes reconstituted on 359-base pair (bp) INO1promoter DNA (27). The INO80 complex mobilizes nucleosomes mainly from N3 to N1 and N2 positions in an ATP-dependent manner (Fig. 3A). We observed neither inhibition (Fig. 3A) nor stimulation (23) of nucleosome mobilizing activity by IP4 and IP5. However, IP6 inhibited INO80-induced nucleosome mobilization (Fig. 3A). The ATPase activity of INO80 was correspondingly inhibited by IP6 but not by other inositol polyphosphates (fig. S2B).

Figure 3

IP6 inhibits chromatin remodeling by INO80 complex. (A) Native PAGE showing nucleosome mobilization. Positions of mononucleosomes (filled circles) reconstituted on INO1 promoter DNA are shown on the left. N3/(N1+N2) ratios are given at the bottom; SD < 0.1. (B) Northern analyses of INO1 expression. Percent wild-type expression normalized using ACT1 is given at the bottom; SD <10%.

The integrity of the inositol signaling pathway is required for the expression of INO1 in vivo. INO1mRNA is reduced to 18% in the ipk2Δ mutant and is rescued by introduction of wild-type IPK2, but only partially by the mutant with an Asp131Ala substitution (D131A) , which impairs kinase activity (Fig. 3B). Hence, production of IP4 and IP5 is required for expression ofINO1. We observed a reduction to 41% of INO1mRNA in the ipk1Δ mutant defective for IP6production, which is rescued by introduction of wild-typeIPK1 (Fig. 3B) (19). Moreover, SNF2and IPK2 are synthetically lethal, and INO80 andIPK1 display synthetic phenotypes (fig. S3) (19).

Proper expression of INO1 likely involves the integration of contributions from INO80, SNF2, and ISW2, which act as positive or negative regulators of transcription (14, 28). Given thatINO80 and SNF2 regulate INO1positively (13, 15, 16) and that ISW2 regulates INO1 negatively (14), cellular levels of IP4, IP5, and IP6 could modulate the balance between synergistic and antagonistic chromatin-remodeling activities (fig. S4). The observed stimulation of SWF/SNF-induced nucleosome mobilization by IP4 and IP5 is consistent with findings of O'Shea and colleagues, who showed that transcription and chromatin remodeling of PHO5 in vivo, mediated by SNF2and INO80, is dependent on production of IP4 or IP5 (29).

The mechanism(s) by which inositol polyphosphates modulate the activities of ATP-dependent chromatin-remodeling complexes are unknown. Recombinant NURF and ISWI protein can bind to IP6(30), which suggests that inositol polyphosphates might alter their activities by effects on protein conformation (31). IP4 or IP5 might affect the interaction between SWI/SNF and chromatin, as has been seen for PIP2 (32). Knowledge of physiological conditions affecting intracellular levels of soluble inositol polyphosphates, as well as corresponding studies of chromatin remodeling and gene expression, will be essential to define the signaling pathway to chromatin.

Supporting Online Material

Materials and Methods

Figs. S1 to S4

Table S1

  • * Present address: Department of Carcinogenesis, University of Texas MD Anderson Cancer Center, Science Park Research Division, Smithville, TX 78957, USA.

  • To whom correspondence should be addressed. E-mail: carlwu{at}


View Abstract

Stay Connected to Science

Navigate This Article