Chromatin remodelers fine-tune H3K36me-directed deacetylation of neighbor nucleosomes by Rpd3S

Abstract

Chromatin remodelers have been implicated in the regulation of histone-modifying complexes. However, the underlying mechanism remains poorly understood. The Rpd3S histone deacetylase complex is recruited by elongating RNA polymerase II to remove histone acetylation at coding regions in a manner that is dependent on methylation of lysine 36 on histone 3 (H3K36me), and Rpd3S prefers dinucleosomes. Here, we show that the binding of Rpd3S to dinucleosomes and its catalytic activity are sensitive to the length of nucleosomal linker in a nonlinear fashion. Intriguingly, we found that H3K36me on one nucleosome stimulates Rpd3S to deacetylate the neighboring nucleosomes when those two nucleosomes are within an optimal distance. Finally, we demonstrate that chromatin remodelers enhance Rpd3S activity by altering nucleosomal spacing, suggesting that chromatin remodelers prime chromatin configuration to fine-tune subsequent histone modification reactions. This mechanism is important for accurate temporal control of chromatin dynamics during the transcription elongation cycle.

Figure 1. The binding of Rpd3S to dinucleosomes is sensitive to the length of linker DNA in a non-linear fashion

(A) A schematic illustration of dinucleosome templates. All nucleosomes are identical except for the length of linker DNA and their positions were confirmed in (B). (B) Nucleosome positioning of reconstituted dinucleosomal templates was confirmed by Bgl1 restriction digestion. As DNA probes were ³²P-labeled at the 5′ ends, only the left-side (601-L) nucleosomes and residual undigested dinucleosomes can still be visualized upon digestion. Their migration patterns indicated the correct translational positions of 601-L nucleosomes and the dinucleosomes. (C) The binding of Rpd3S to dinucleosomes with different linker DNA were measured by EMSA assays. (D) Quantification of the results from (C). Data are represented as mean +/− SEM.

Figure 2. Histone H3K36 methylation regulates Rpd3S HDAC activity toward neighboring nucleosomes within proper linker distance

(A) Preparation of hetero-dinucleosome templates. Gel-purified mononucleosomes (designated as Right or 601-R) containing either unmodified (white-filled) or K36me3 (black-filled) histone H3, were labeled with ³H acetyl-CoA using Ada2-TAP (H3 acetylation) and NuA4 (H4 acetylation). 601-R nucleosomes were ligated at a Bgl1 site to 601-L nucleosomes that carry various lengths of linker DNA. The resulting dinucleosomes were gel-purified and directly applied to HDAC assays in (C–E) or end-labeled using ³²P-γATP for EMSA assays in (F–H). (B) Confirmation of proper nucleosome positioning of the hetero-dinucleosomes. The restriction analyses were performed similarly to that described in Figure 1B, except that in this case both ends of the DNA probes were labeled. Therefore, both digested mononucleosomes can be seen. Lane 19–26 contain the mono-nucleosomes before they were ligated to form dinucleosomes, and serve as marks for the correct migration pattern of digested dinucleosomes. (C–E)Histone deacetylase assays showing that the HDAC activity of Rpd3S toward neighboring nucleosomes depends on the length of linker DNA. Data are represented as mean +/− SEM. (F–H) The binding affinity of Rpd3S to each hetero-dinucleosome was measured using gel mobility shift assays. Data are represented as mean +/− SEM.

Figure 3. Deacetylation preference of Rpd3S on dinucleosome substrates

(A) The strategy to generate hetero-dinucleosome templates in which each nucleosome contains an equal amount of acetyl-groups. During the first round of the acetylation reaction, ³H-labeled acetyl-CoA or cold acetyl-CoA were used for either 601-L nucleosomes or 601-R nucleosomes. Subsequently, additional acetylation was carried out for all mononucleosomes using cold acetyl-coA to reach saturated acetylation levels on all nucleosomes. After ligation and gel-purification, the resulting two types of dinucleosomes are identical in total acetylation status except for ³H-labeling depending on the reaction. (B) The acetylation levels of each mononucleosome after two rounds of acetylation were monitored by western blotting using anti acetylated H3 or acetylated H4 antibodies. An antibody against histone H4 was used to demonstrate the loading consistency. (C) The final hetero-dinucleosomes prepared above were resolved on a 3.5% acrylamide gel, and the dinucleosome concentrations were quantified based on the amount of DNA. (D) Quantification of HDAC results based on three independent experiments. Data are represented as mean +/− SEM. (E) A model for different mechanisms of action of Rpd3S on dinucleosomes containing different linker DNA. The thicknesses of the arrows reflect the strength of HDAC activity. 30bp (mid) is near optimal distance in that Rpd3S can efficiently deacetylate both methylated nucleosome and neighboring unmethylated nucleosome; 15bp (left) represents the situation where two nucleosomes are too close and Rpd3S can function on neighboring nucleosomes but with less efficiency; At 70bp (right) distance, Rpd3S can no longer deacetylate neighboring nucleosomes.

Figure 4. Chromatin remodeling factors fine-tune the Rpd3S HDAC activity on dinucleosomes by altering nucleosomal spacing

(A) The experimental design. (B) Rpd3S HDAC activity on dinucleosomes is enhanced by indicated chromatin remodeling complexesin an ATP-dependent manner. Data are represented as mean +/− SEM. (C) Chromatin remodeling per-se does not increase Rpd3S activity. Similar experiments were performed as in (B) except that mono-nucleosomes with indicated modification pattern were used as substrates. Data are represented as mean +/− SEM. (D–F) ISWI family of ATPases mediates nucleosome sliding on dinucleosome templates. (D) Chromatin remodeling assay using dinucleosome substrates containing a 70 bp linker DNA. The reactions were directly loaded on a 3.5% native PAGE gel to resolve dinucleosome populations that are positioned differently. (E) Predicted positioning changes of each nucleosome caused by Isw1-TAP and ACF-mediated dinucleosome sliding. Restriction sites Not1 and Bgl1, which are just outside each nucleosome, are used to monitor nucleosome movement. (F) Restriction site protection assays for dinucleosome sliding suggest that both nucleosomes moved toward the center. Following the remodeling assay described above, the reaction mixtures were subjected to restriction digestion with Not1 or Bgl1. The resulting digested and undigested nucleosomes were resolved on a 4% PAGE gel.