Structural basis for site-specific histone H3 acetylation-dependent regulation of RNAPII transcription through nucleosomes

  1. Takumi Oishi
  2. Suguru Hatazawa
  3. Tomoya Kujirai
  4. Sae Suzuki
  5. Koki Nakatsu
  6. Gosuke Hayashi
  7. Junko Kato
  8. Yuki Kobayashi
  9. Mitsuo Ogasawara
  10. Haruhiko Ehara
  11. Shun-ichi Sekine
  12. Yoshimasa Takizawa
  13. Hitoshi Kurumizaka

  • Curated by eLife

    eLife logo

    eLife Assessment

    This manuscript describes a valuable study of the mechanism by which acetylation on the histone H3 core domain regulates RNA polymerase II transcription passing through nucleosomes. The authors provide convincing evidence that acetylation influences transcription in a context-specific fashion. Some questions relating to the static nucleosome structures and the polymerase passage remain, but this manuscript will be of considerable interest to researchers in the chromatin and transcription fields.

Reviewed by

Read the full article

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

Log in to save this article

Abstract

Acetylation of histone H3 regulates chromatin dynamics and transcription, but how specific acetylation sites affect RNA polymerase II (RNAPII) transcription through nucleosomes remains unclear. Here, we found that H3 acetylation at Lys56 and Lys122 markedly enhances RNAPII transcription through nucleosomes, whereas acetylation at Lys64 has little effect. To elucidate the structural basis for these functional differences, we determined cryo-electron microscopy (cryo-EM) structures of nucleosomes bearing site-specific acetylation at H3K56, H3K64, or H3K122. The cryo-EM structures revealed that H3K56ac and H3K122ac locally weaken histone-DNA interactions at the DNA entry/exit region and near the dyad, respectively, while H3K64ac induces no detectable structural changes. These structural differences correlate with the observed transcriptional outcomes, indicating that acetylation at H3K56 and H3K122, but not H3K64, alleviates the nucleosomal barrier to RNAPII progression. Our findings provide direct structural evidence that specific acetylations within the histone fold domain of H3 finetune nucleosome dynamics to facilitate RNAPII transcription.

Article activity feed

  1. eLife

    eLife Assessment

    This manuscript describes a valuable study of the mechanism by which acetylation on the histone H3 core domain regulates RNA polymerase II transcription passing through nucleosomes. The authors provide convincing evidence that acetylation influences transcription in a context-specific fashion. Some questions relating to the static nucleosome structures and the polymerase passage remain, but this manuscript will be of considerable interest to researchers in the chromatin and transcription fields.

  2. eLife

    Reviewer #1 (Public review):

    Summary:

    The authors investigate how site-specific acetylation within the histone H3 folded domain affects RNA polymerase II transcription through nucleosomes. They focus on H3K56ac, H3K64ac, and H3K122ac, prepare chemically defined nucleosomes carrying each modification, and compare their effects using an in vitro transcription assay, cryo-electron microscopy structures, and micrococcal nuclease sensitivity assays.

    The main finding is that H3K56ac and H3K122ac increase production of full-length run-off transcripts and reduce pausing near the nucleosomal dyad region, whereas H3K64ac has little detectable effect under the same reconstituted conditions. The structural analyses suggest that H3K56ac weakens or destabilizes DNA near the entry/exit region, while H3K122ac alters histone-DNA contacts near the dyad. These observations support a model in which different acetylation sites within the H3 folded domain influence nucleosomal transcription barriers through distinct local effects on histone-DNA interactions.

    This is a useful study because it examines histone core-domain acetylation using chemically defined nucleosomes and directly compares several modifications in the same experimental system. However, the broader cellular context of these modifications is not sufficiently developed, and some mechanistic conclusions rely on correlations between static nucleosome structures and endpoint transcription assays rather than direct observation of polymerase passage through modified nucleosomes.

    Strengths:

    (1) The study uses site-specifically acetylated H3 proteins and reconstituted nucleosomes, allowing direct comparison of H3K56ac, H3K64ac, and H3K122ac under controlled conditions.

    (2) The combination of transcription assays, cryo-electron microscopy, and nuclease sensitivity assays provides multiple lines of evidence, particularly for increased DNA end flexibility in H3K56ac nucleosomes.

    (3) The authors analyze unmodified, H3K56ac, H3K64ac, and H3K122ac nucleosomes in parallel, with reported structural resolutions of approximately 3 Angstroms and accompanying validation materials.

    (4) The negative result for H3K64ac is informative, because it distinguishes the direct effect of this modification in a minimal reconstituted system from prior cellular associations with active chromatin and histone eviction.
    The comparison with H3 N-terminal acetylation highlights that acetylation within the folded domain may affect transcription at different positions or by different mechanisms than tail acetylation.

    Weaknesses:

    The rationale for focusing on H3K56ac, H3K64ac, and H3K122ac has not been developed sufficiently. The manuscript would benefit from a clearer summary of what is known about the abundance of these modifications in cells, the enzymes or histone metabolic pathways that may introduce or remove them, and whether they are thought to occur before histone deposition, on assembled nucleosomes, or during nucleosome remodeling.

    The central mechanistic model is based mainly on correlations between structures of free nucleosomes and endpoint transcription assays. The study does not directly observe RNA polymerase II paused at or passing through the relevant nucleosomal positions, so the proposed link between local structural changes and reduced pausing should be stated with appropriate caution.

    The H3K56ac interpretation is supported by both structural observations and nuclease sensitivity data, but the map comparison underlying the reduced entry/exit DNA density is still mostly qualitative. The manuscript should more clearly state the map comparison conditions, such as contouring and local map quality, so that non-specialist readers can judge how robust the local density differences are.

    The H3K122ac mechanism is plausible, but the evidence for dyad destabilization is more indirect. The main support comes from the orientation of the K122 side chain and its distance from DNA, while an independent biochemical test of dyad-region destabilization is not provided.

    The transcription assay appears to include statistical testing, but the figure legend and methods should more clearly state which tests were used, what comparisons were made, how n was defined, and whether multiple-comparison correction was applied.

    The relationship between the 198 bp transcription template, the linker DNA, the 9-base mismatched region, and the DNA regions modeled in the cryo-electron microscopy structures is somewhat difficult to follow. This does not necessarily require new experiments, but a clearer explanation would help readers connect the transcription assay design with the structural models.

    The use of H3.2 C110A for chemical ligation and the use of the PL2-6 single-chain antibody fragment for cryo-electron microscopy sample stabilization are reasonable technical choices, but their purposes and possible effects on interpretation should be explained more clearly for readers outside structural biology.

    Because the work uses a minimal in vitro system with human nucleosomes and Komagataella phaffii RNA polymerase II/TFIIS, the conclusions should be limited to direct physical effects on nucleosome transcription barriers unless cellular cofactors, remodelers, histone chaperones, additional modifications, and nucleosome positioning are addressed or discussed.

  3. eLife

    Reviewer #2 (Public review):

    Summary:

    Chromatin regulates a wide range of biological processes. The nucleosome, composed of 147 bp of DNA wrapped around a histone octamer containing histones H2A, H2B, H3, and H4, is the fundamental unit of chromatin. Post-translational modifications of histone proteins regulate the dynamic properties of nucleosomes and thereby influence chromatin accessibility and gene expression. Among these modifications, lysine acetylation on histone H3 is closely associated with transcriptional activation. While the epigenetic functions of acetylation on the histone H3 N-terminal tail have been extensively studied, the molecular mechanisms by which acetylation within the histone H3 core domain, particularly at Lys56, Lys64, and Lys122, modulates nucleosome architecture to facilitate RNA polymerase II (RNAPII) transcription remain unclear.

    In this study, Oishi et al. investigated the effects of histone H3 acetylation at K56, K64, and K122 on RNAPII transcription using in vitro transcription assays. Furthermore, the authors determined the three-dimensional structures of nucleosomes containing these acetylation marks by cryo-electron microscopy single-particle analysis, revealing distinct structural dynamics depending on the acetylation site. Overall, this study advances our understanding of the molecular mechanisms linking histone H3 core acetylation to transcriptional regulation.

    Strengths:

    (1) Site-specifically acetylated histone H3 proteins were chemically synthesized using a unique and rational peptide ligation strategy, representing a major technical strength of this study.

    (2) The in vitro transcription assays demonstrated that H3K56ac and H3K122ac increase the production of run-off transcripts, whereas H3K64ac has little effect on transcription efficiency. These findings highlight the distinct functional roles of individual acetylation sites within the histone H3 core domain.

    (3) The cryo-EM structures of nucleosomes containing either H3K56ac or H3K122ac revealed that H3 acetylation weakens histone-DNA interactions, providing a structural basis for the observed effects on transcription.

    Weaknesses:

    (1) Although the biochemical and structural data are convincing and sufficiently support the authors' conclusions, complementary cellular experiments would further strengthen the physiological relevance of the in vitro findings. While such experiments are not essential for supporting the main claims of the study, they would enhance the overall impact and biological significance of the work.

    (2) Although the authors demonstrate the structural consequences of individual H3 core acetylation events, the study does not investigate potential synergistic effects among multiple acetylated lysine residues within the H3 core domain. Consequently, the relationship between combinatorial acetylation patterns and their collective impact on RNA polymerase II-mediated transcription remains unclear.

  4. eLife

    Reviewer #3 (Public review):

    This is a short and punchy manuscript that nicely summarises the 4 structures that are determined and provides a basis for the differences seen for acetylation sites shown for RNAPII activity.

    The authors build on previous biochemical work that determined the functional outcomes of H3 core acetylation, adapting an assay they have previously used extensively to investigate RNAPII transcription on nucleosomes and, indeed, even H3 N-terminal tail acetylation. This assay is as such well set up and has a wealth of confirmatory previous studies from this lab and the authors are careful not to overanalyse their results, leading to robust and well-considered results. The structures are determined to a high resolution, allowing the interpretation put forward about side chain orientations, with clear densities shown for the regions of interest.

    Further discussion or experiments would strengthen the conclusions further:

    (1) The conclusion on the role of H3K56Acetylation could be strengthened, especially as the results are somewhat counterintuitive. It is conceptually surprising that acetylation near the entry/exit DNA that destabilises this region also leads to a reduced stall propensity at the dyad but has a limited effect at SHL5? While it can be explained by the clash at the dyad pause being reduced, the more direct effect of DNA breathing amplification would be expected to have a larger effect at SHL 5. Indeed, the density for DNA at SHL5 appears to be weaker in Figure 2A, suggesting the entry/exit DNA flexibility is amplified past this region.

    Perhaps another assay that looks more directly at the flexibility of the entry/exit DNA would be useful, either through restriction enzyme-mediated cleavage or FRET (DNA ends and H2AK119 labels), providing stronger evidence of this effect. MNase is rather indirect and similar to the RNAPII assay itself.

    Similarly, were the authors surprised by the modest effect (less than 2-fold) in transcriptional pause at SHL 0 for the K122Ac? Presumably, based on the model in Figure 4, this would be expected to be the area with the largest effect? The results of K56Ac and K122Ac almost seem swapped to what would be expected in Figure 1H. Further discussion of this observation would be useful.

    (2) Could the local weakening of DNA, especially at the dyad, be observed in the cryo-EM structures? Perhaps comparison of local resolution estimation differences in this region compared to unmodified would be useful.

    (3) Caution should be taken, and discussion should include that the structural data presented is after extensive processing. Many nucleosome averaging classes were discarded in the 3D classification steps (nicely summarised in Table 1 as "particles for 3d classification" and "particles in final map"). Indeed, it is likely that higher DNA flexibility particles would be thrown away during this processing step. This can be observed for K56Ac DNA ordering, for example, in Supplementary Figure S4, yellow and cyan classes from the round of 3D classification look to be high resolution and have a higher order of DNA, so there has been some selection here. How was this done? While this is not fully quantifiable, it gives an idea of the extent of wrapping. We would suggest discussing the methodological limitations and showing the models after the first auto refinement to see if the features discussed on end flexibility and dan ordering are retained.

    (4) Di Cerbo et al. (reference 13) showed acetylation at K64 alters salt stability and affects transcription. Why do the authors think there is a discrepancy, albeit with different assays? Direct reference and discussion of this in the text should be included.

    (5) Why was H3.2 used, while this is relatively abundant in mouse cells, human protein was used, and this appears to be less common than H3.1 and H3.3. We are sure that the effect is not likely to be substantive on structure (as shown by the Kurumizaka lab previously), but should be addressed in the text

  5. Version published to 10.7554/elife.111802.1 on eLife
  6. Version published to 10.7554/elife.111802 on eLife
  7. Version published to 10.64898/2026.05.02.722397 on bioRxiv