Histone modifications influence the action of Snf2 family remodelling enzymes by different mechanisms

Abstract

Alteration of chromatin structure by chromatin modifying and remodelling activities is a key stage in the regulation of many nuclear processes. These activities are frequently interlinked, and many chromatin remodelling enzymes contain motifs that recognise modified histones. Here we adopt a peptide ligation strategy to generate specifically modified chromatin templates and used these to study the interaction of the Chd1, Isw2 and RSC remodelling complexes with differentially acetylated nucleosomes. Specific patterns of histone acetylation are found to alter the rate of chromatin remodelling in different ways. For example, histone H3 lysine 14 acetylation acts to increase recruitment of the RSC complex to nucleosomes. However, histone H4 tetra-acetylation alters the spectrum of remodelled products generated by increasing octamer transfer in trans. In contrast, histone H4 tetra-acetylation was also found to reduce the activity of the Chd1 and Isw2 remodelling enzymes by reducing catalytic turnover without affecting recruitment. These observations illustrate a range of different means by which modifications to histones can influence the action of remodelling enzymes.

Figure 1

Generating acetylated histones and nucleosomes. (a) Modified histones were generated by the ligation of chemically synthesised peptide bearing a C-terminal thioester group to an N-terminally truncated histone bearing an N-terminal cysteine. In this case the peptide consisted of amino acid residues 1–27 of the Xenopus laevis H3 tail in which lysine residues 9, 14, 18 and 23 were acetylated and K27 synthesised as a thioester derivative. The histone was histone H3 in which amino acid residues 1–27 had been truncated and the new N-terminal residue mutated from serine to cysteine. The ligation reaction proceeds via an irreversible intramolecular rearrangement to produce a covalent native peptide bond without the introduction of unnatural chemical moieties. (b) Full-length acetylated histones were purified away from the reactants by two rounds of ion exchange chromatography using a step gradient. Lane 1 shows a representative reaction in which ligation had proceeded to around 40%. The top two panels show the traces from the first and second chromatography runs used to purify the ligated product from the unreacted material. The bottom panel shows SDS–PAGE of fractions taken form the positions (i, ii, and iii) indicated on the traces. Fraction iii, consists of over 95% ligated full length histone H3. (c) Acetylated histone octamers formed nucleosomes similar to those formed by unmodified histone octamers. Interestingly, the mobility of tetra-acetylated H3 nucleosomes during native PAGE was slightly reduced compared to control nucleosomes of the wild-type sequence; compare lanes 1 and 2 or histones containing cysteine point mutations (not shown).

Figure 2

Effects of histone acetylation on intrinsic nucleosome mobility. (a) Outline of competitive remodelling assay used to accurately measure differences between nucleosomes. This setup has the advantage that the reaction times and conditions are exactly the same for two nucleosomes and avoids the possibility of observing effects due to different nucleosome reconstitution efficiencies. (b) Histone H4 tetra-acetylation does not alter the rate at which nucleosomes reposition thermally. Two pmol of H4 acetylated and H4 V21C unmodified nucleosomes assembled on 54A54 DNA were mixed and incubated at 47 °C for the specified amount of time. The images represent the Cy3 (H4 V21C) and Cy5 (H4 acetylated) scans of the same gel. The amount of remodelling is plotted to the right of the gels and a hyperbolic curve fitted to the data points. From this the initial rate of repositioning is calculated and the average of three independent repeats displayed in (e). (c) Tetra-acetylation of H3 results in a twofold increase in the rate of intrinsic nucleosome mobility. Reaction conditions are as for (b). (d) The increase in nucleosome mobility by H3 acetylation is not solely due to charge neutralisation. Two pmol of H3 wild-type and H3 in which the lysine residues at position 9, 14, 18 and 23 had been substituted to alanine were assembled on 54A18 DNA were mixed and incubated at 47 °C for the specified amounts of time. Substituting the lysine residues with alanine, which will reduce the basic charge of the tail, had no effect on nucleosome mobility. (e) Table indicating the initial rate of repositioning relative to control for the three constructs described above.

Figure 3

RSC preferentially repositions tetra-acetylated H3 nucleosomes. (a) 20 fmol of RSC were incubated with 1pmol unmodified S28C octamers assembled onto Cy3-labelled 54A18 DNA and 1pmol H3 tetra-acetylated nucleosomes on the same DNA labelled with Cy5 and incubated for the specified length of time at 30 °C in the presence of 1 mM ATP. RSC shows a dramatic preference for H3 tetra-acetylated nucleosomes: compare lanes 1–6 and 7–12; see also (d). On this fragment a proportion of nucleosomes are deposited at an alternative location, indicated by an asterisk (*), that has been characterised previously. Inclusion of nucleosomes deposited at this location had little effect on the calculated initial rate of sliding so they were excluded from quantitative analysis. (b) H4 tetra-acetylated nucleosomes, in contrast, are not repositioned faster by the RSC complex. (c) Nucleosomes that are both H3 and H4 tetra-acetylated are not repositioned any faster than H3 tetra-acetylated nucleosomes, confirming that H4 tetra-acetylation does not promote RSC catalysed repositioning. (d) Table indicating the average initial rate of repositioning relative to control and standard deviation from three independent experiments for the acetylated constructs described above.

Figure 4

Determining the kinetic parameters of nucleosome remodelling by RSC with a fluorescent ATPase assay. Overview of ATPase assay. (a) Nucleosome remodelling generates the release of inorganic phosphate (Pi) as a result of ATP hydrolysis. (b) This level of Pi is detected by a fluorescently labelled phosphate binding protein (PBP–MDCC), whose fluorescence increases dramatically upon phosphate binding. (c) Chromatin remodelling is initiated by the addition of 0.3 nM RSC to different concentrations of 36W36 nucleosomes and the fluorescence intensity measured in real-time at 1 s intervals over approximately 10 min. (d) Kinetic parameters were calculated by non-linear fitting of the Michaelis–Menton equation to the plotted data. (e) K_m and K_cat of remodelling of different nucleosome substrates by RSC. H3 tetra-acetylated nucleosomes show lower K_m values without affecting K_cat. H4 tetra-acetylation does not affect either parameter, consistent with data from Figure 3. Mono-acetylation at K14 of H3 significantly affects the K_m of remodelling, largely mimicking H3 tetra-acetylation. Although the K_m and K_cat shown above were calculated for wild-type histones no difference was detected for H3S28C nucleosomes (Supplementary Data, Figure 4).

Figure 5

H4 tail regulates the catalytic activity of Isw2. (a) As Isw2 has been observed to move nucleosomes away from DNA ends, nucleosomes were assembled at a position close to a DNA end using the fragment 54A0. 3 fmol of Isw2 were incubated with 1pmol intact octamers assembled onto Cy3-labelled DNA and 1pmol octamers from which the first 19 amino acid residues of H4 had been deleted assembled on Cy5-labelled DNA for the times indicated at 30 °C in the presence of 1 mM ATP. gH4 nucleosomes were repositioned slower as shown in the graph. (b) In a similar comparison tetra-acetylated nucleosomes are repositioned slower than H4 V21C nucleosomes by Isw2. (c) Quantitative comparison of the effects of truncating and acetylating the H4 tail on Isw2 atpase activity. (d) Effects of H4 acetylation and truncation on the ATPase activity of Isw2. Reaction conditions are as for Figure 4 except using 0.2 nM Isw2 and different concentrations of 54W0 nucleosomes. Truncation of the H4 tail, and to a lesser extent histone acetylation, reduce the catalytic turnover of the enzyme without significantly affecting the K_m.

Figure 6

Chd1 requires the H4 tail for efficient nucleosome remodelling. (a) 40 fmol of Chd1 were incubated with 1pmol of wild- type octamers deposited on Cy3-labelled 54A0 DNA and 1pmol of octamers containing H4 with the first 19 amino acid residues deleted assembled on the same DNA labelled with Cy5 and incubated for the specified lengths of time at 30 °C in the presence of 1 mM ATP. Deletion of the H4 tail results in poor nucleosome remodelling by Chd.1 (b) Mutation of amino acid residues 16–19 to alanine significantly reduces the rate at which Chd1 repositions nucleosomes. (c) Tetra-acetylation of H4 causes nucleosomes to be repositioned at a slower rate by. (d) Quantification of the effect of alterations to histones on the initial rate of nucleosome repositioning by Chd1 shown in (a)–(c). (e) Effect of truncating the H4 tail on the ATPase activity of Chd1. The reaction conditions are as for Figure 4 except using 1 nM Chd1 and 54W0 nucleosomes initial rates of ATP hydrolysis were measured. A non-linear fit of these data to the Michaelis–Menton equation (with R2 confidence values above 0.98) allows K_m and K_cat to be determined for nucleosomes containing intact H4 and truncated H4. There is little effect on K_m, but K_cat is significantly reduced, suggesting that the H4 tail is an allosteric effector for Chd1.

Figure 7

H4 tetra-acetylation increases octamer transfer by RSC. (a) Octamer transfer assay: RSC is able to disrupt nucleosomes and transfer the histone octamer from unlabelled donor nucleosomes onto a separate DNA molecule, in this case a radiolabelled 147 bp 0W0 fragment derived from the 601 positioning sequence. This is measured by the shift in mobility of a radiolabelled DNA fragment to that of a nucleosome. (b) Efficiency of octamer transfer from different donor nucleosomes. H3 tetra-acetylated nucleosomes are transferred faster than unmodified nucleosomes consisting of wild-type H3 and H4: compare lanes 4–6 with lanes 1–3. Surprisingly, H4 tetra-acetylated nucleosomes are also transferred faster than control: compare lanes 7–9 with lanes 1–3. When both H3 and H4 are acetylated the effect is additive (compare lanes 10–12 with 1–3). Lane 13 shows an equivalent amount of free DNA in the absence of RSC or nucleosomes and lane 14 is a nucleosome reconstituted separately on the same DNA fragment as a mobility reference. (c) Table plotting the amount of octamer transfer from different donor nucleosomes as the average of three independent repeats. Error bars represent the standard deviation. Although histones of wild-type sequence are used as a control in the data shown, octamers bearing cysteine mutations behaved similarly.