Introduction

Chromatin plays a central role in regulating access to genomic DNA during fundamental cellular processes such as transcription, replication, and DNA repair (1). The basic structural unit of chromatin is the nucleosome, which comprises approximately 147 base pairs of DNA wrapped around a histone octamer containing two copies each of histones H2A, H2B, H3, and H4 (2, 3). The positions in nucleosomal DNA are designated as superhelical locations (SHLs) (4). By defining the center of the nucleosomal DNA as SHL(0), which corresponds to the dyad axis of the nucleosome, DNA positions at roughly every 10 base pairs are assigned as SHL(±1) to SHL(±7) (2, 4). The dynamic nature of nucleosomes is tightly regulated by post-translational modifications (PTMs) of histones, which can modulate histone-DNA and histone-protein interactions to influence chromatin accessibility (5–7). Among these modifications, the acetylation of lysine residues on histone H3 has long been associated with transcriptional activation, by promoting an open chromatin conformation conducive to gene expression (8, 9).

While histone acetylation is generally thought to weaken histone-DNA contacts by neutralizing the positive charge of lysine side chains (10), how individual acetylation sites contribute to chromatin dynamics and transcriptional regulation remains poorly understood. In particular, the lysine residues within the histone H3 core domain, such as Lys56 (K56), Lys64 (K64), and Lys122 (K122), are positioned at key DNA-histone interfaces and have been implicated in modulating nucleosome stability (6, 11–13). H3K56 acetylation occurs at the DNA entry/exit region and has been linked to nucleosome assembly, disassembly, and replication-coupled chromatin remodeling (14–17). The acetylation of H3K122, located near the nucleosomal dyad, has been reported to directly destabilize histone-DNA contacts, promoting nucleosome sliding and transcriptional activation (18–22). Similarly, the acetylation of H3K64, positioned within the nucleosome core, facilitates nucleosome instability and promotes chromatin accessibility, correlating with transcriptionally active regions (13, 23). However, the mechanisms by which these modifications alter the nucleosome architecture to facilitate RNA polymerase II (RNAPII) transcription remain elusive.

In the present study, we found that the acetylation of histone H3 at Lys56 and Lys122 markedly enhances RNA polymerase II (RNAPII) transcription through nucleosomes, whereas the 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 containing site-specifically acetylated H3 at Lys56, Lys64, or Lys122. The structures revealed that H3K56ac and H3K122ac locally weaken histone-DNA interactions at the DNA entry/exit region and near the dyad, respectively, whereas H3K64ac causes no detectable changes. These distinct structural consequences of acetylation within the histone fold domain of H3 provide a mechanistic framework for understanding how specific lysine modifications fine-tune nucleosome dynamics to facilitate RNAPII transcription.

Results

Preparation of site-specifically acetylated nucleosomes for RNAPII transcription

To investigate how specific acetylation sites within the histone fold domain of histone H3 influence nucleosome structure and transcription, we prepared human histone H3 bearing the site-specific acetylation of Lys56, Lys64, or Lys122 (H3K56ac, H3K64ac, and H3K122ac) by chemical synthesis (Figure 1A and B, Supplementary Figure 1A). Each acetylated histone H3 was obtained through native chemical ligation with synthetic acetylated and non-acetylated peptides. In this method, we employed the H3.2 C110A mutant, in which the Cys110 residue of H3.2 was replaced by Ala, because the Cys residue is converted into the Ala residue during the desulfurization process after peptide ligation. We prepared histone octamers containing recombinant histones H2A, H2B, H4, and the synthesized H3.2 containing H3K56ac, H3K64ac, or H3K122ac. Nucleosomes containing H3K56ac, H3K64ac, and H3K122ac were then reconstituted using a 198 base-pair (bp) DNA fragment consisting of a modified Widom 601 sequence containing a 9-base mismatched region in a linker DNA (24) (Supplementary Figure 1B and C). RNAPII associates with the 9-base mismatched region and initiates transcription by the addition of nucleotide triphosphates and a primer RNA. We then performed in vitro nucleosome transcription assays using purified Komagataella phaffii RNAPII and its essential elongation factor, TFIIS (Figure 1C).

Figure 1:

Site-specific acetylations of the histone fold domain of H3 differently modulate nucleosome transcription by RNAPII

When RNAPII transcribed through the unmodified nucleosome, an RNA transcript of approximately 40 nucleotides, which corresponds to the RNAPII pausing at the SHL(-5), was predominantly accumulated (Figure 1D, lane 3, Supplementary Figure 2). In addition, RNA products around 75-80 nucleotides, which correspond to the RNAPII pausing at the SHL(-1) and SHL(0) positions, were also detected (Figure 1D, lane 3). Notably, run-off transcripts were barely detectable in the unmodified nucleosome template (Figure 1D, lane 3). These results are consistent with the previous studies (24) and indicate that the nucleosome acts as a substantial barrier to RNAPII progression.

Strikingly, nucleosomes containing either H3K56ac or H3K122ac showed a pronounced increase in run-off transcript production by RNAPII (Figure 1D, lanes 4, 6). In contrast, H3K64ac had little or no effect on the transcription efficiency, which was comparable to the unmodified control (Figure 1D, lane 5). These results demonstrate that specific acetylations within the histone fold domain distinctly modulate the transcription efficiency of RNAPII through the nucleosome.

A quantitative analysis revealed that H3K56ac and H3K122ac increased the fractions of full-length transcripts by approximately three-and five-fold relative to unmodified nucleosomes, respectively (Figure 1E). These effects are comparable to the enhanced nucleosome transcription observed with the H3 N-terminal acetylations, including H3K4ac, H3K9ac, H3K14ac, H3K18ac, H3K23ac, and H3K27ac (Figure 1D, lane 7, and E). Consistent with a previous study (25), the H3 N-terminal acetylations drastically reduce RNAPII pausing at the SHL(-5) position, thereby increasing the pausing population at the SHL(-1) and SHL(0) positions (Figure 1F-H). In contrast, the RNAPII pausing at the SHL(-5) and SHL(-1) positions was not substantially altered in the H3K56ac and H3K122ac nucleosomes (Figure 1F and G). Notably, the RNAPII pausing at the SHL(0) position was significantly alleviated in the H3K56ac and H3K122ac nucleosomes (Figure 1F). Such alleviation of RNAPII pausing at the SHL(0) position was not observed in the nucleosome containing the H3 N-terminal acetylations (Figure 1D, lane 7, and H). Therefore, the acetylation within the histone fold domain may regulate RNAPII transcription differently from the acetylations on the H3 N-terminal tail.

Cryo-EM structures reveal local weakening of histone-DNA contacts by H3K56ac and H3K122ac, but not H3K64ac

To understand the structural basis for the distinct transcriptional outcomes, we determined the cryo-EM structures of the H3K56ac, H3K64ac, and H3K122ac nucleosomes and the unmodified nucleosome used for the nucleosome transcription assays at 2.96, 2.99, 3.07, and 3.07 Å resolutions, respectively (Figure 2A, Table 1, and Supplementary Figure 3-6). We avoided sample crosslinking, which stabilizes histone-DNA contacts in the nucleosome. Instead, we employed a single-chain variable fragment (scFv), PL2-6, which specifically binds to the acidic patch of nucleosomes and effectively isolates the nucleosome sample from the air-water interface, thereby stabilizing the cryo-EM specimen (26).

Table 1.

Figure 2:

Figure 3:

In the H3K56ac nucleosome, parts of the Lys56 and acetylated Lys56 side chains are resolved, with the visible density appearing to extend toward the SHL(±6) region (Figure 3A). In addition, the density corresponding to the DNA entry/exit region appeared substantially weaker than that of the unmodified nucleosome, suggesting the increased flexibility of the nucleosomal DNA ends (Figure 2A). Since lysine acetylation diminishes its positive charge and thereby weakens the local histone-DNA interactions, the flexibility of the nucleosomal DNA ends may be enhanced by the H3 Lys56 acetylation, which weakens the histone-DNA interactions near the SHL(±6) position despite the overall preservation of the side-chain orientation. Consistently, the DNA ends of the H3K56ac nucleosome were markedly more susceptible to micrococcal nuclease than those of the unmodified nucleosome (Figure 2B-D, Supplementary Figure 7). These results are consistent with previous fluorescence resonance energy transfer studies (19, 27–29). Notably, the leading edge of RNAPII sterically clashes with the distal DNA end of the nucleosome upon reaching the SHL(0) position, inducing RNAPII pausing (30). Therefore, the increased flexibility at the distal nucleosomal DNA end may alleviate the RNAPII pausing at the SHL(0) position (Figure 4).

Figure 4:

The H3 Lys122 residue directly interacts with the DNA backbone in the unmodified nucleosome. The acetyl group neutralizes the positive charge of the lysine sidechain, thereby diminishing the electrostatic attraction between the histone and DNA, which likely facilitates the partial unwrapping of the nucleosomal DNA near the SHL(0) position. In fact, the acetylation of the H3 Lys122 residue reorients the sidechain away from the nucleosomal DNA (Figure 3B). This may be a major reason why the H3 Lys122 acetylation enhances RNAPII transcription, by alleviating its SHL(0) pausing (Figure 4).

In contrast, the H3K64ac nucleosome displayed nearly identical density and local resolution profiles to the unmodified nucleosome, indicating that this modification does not significantly alter the histone-DNA contacts within the nucleosome core particle (Figure 3C).

Discussion

Our results reveal that the acetylation of lysine within the histone fold domain of H3 has distinct and site-specific effects on RNAPII transcription through nucleosomes. Among the three tested sites, H3K56ac and H3K122ac markedly enhanced transcription, whereas H3K64ac had little influence. Structural analyses by cryo-EM demonstrated that H3K56ac and H3K122ac locally weaken the histone-DNA interactions at the DNA entry/exit region and near the dyad, respectively, providing a direct structural basis for their transcriptional effects. These findings establish a mechanistic link between the site-specific acetylation within the histone fold domain and the regulation of RNAPII transcription by histone acetylation.

The acetylations of the H3 Lys56 and H3 Lys122 residues likely promote transcription by distinct but complementary mechanisms (Figure 4). H3K56ac increases the flexibility of nucleosomal DNA ends, which may lower the energetic barrier for DNA unwrapping and alleviate the RNAPII pausing at the SHL(0) position by reducing steric clashes between RNAPII and the distal nucleosomal DNA end. In contrast, H3K122ac destabilizes the histone-DNA contacts around the dyad, directly weakening the central anchor point of the nucleosome and facilitating the passage of RNAPII through this region. Both effects converge to lower the nucleosomal barrier and enhance transcriptional output, consistent with the observed increases in full-length transcripts. Notably, these mechanisms differ from those mediated by H3 N-terminal acetylations, which mainly affect RNAPII pausing at upstream positions such as SHL(-5) (25). Thus, the acetylation within the histone fold domain modulates transcription through a structural mechanism distinct from that of histone tail acetylations. Previous biochemical studies have suggested that H3K56ac and H3K122ac contribute to chromatin remodeling, histone exchange, and transcriptional activation (18, 20, 21, 31, 32). Our structural data now provide direct evidence that these modifications alter the physical properties of the nucleosome to promote transcriptional elongation.

Although H3K64ac is enriched at transcriptionally active sites, our results indicate that acetylation of H3 Lys64 has minimal impact on RNAPII transcription through nucleosomes or on nucleosome structure. These findings suggest that H3K64ac alone does not directly modulate RNAPII transcription efficiency. Previous studies have reported that H3K64ac promotes histone eviction, raising the possibility that additional factors, such as histone chaperones, are required to facilitate transcription (13). Moreover, H3K64ac is known to co-localize with transcriptionally active histone marks, including H3K27ac, H3K4me1, and H3K122ac, suggesting that H3K64ac may regulate transcription through crosstalk with other PTMs (13, 23). In our H3K64ac nucleosome structure, acetylated Lys64 is shielded by the surrounding DNA, implying that this modification is inaccessible to reader proteins in an intact nucleosome. Therefore, recognition of H3K64ac likely requires nucleosome disruption. Crosstalk between H3K64ac and other PTMs may occur prior to chromatin assembly or during nucleosome-disruptive events.

More broadly, our findings highlight the functional diversity of histone acetylation beyond the N-terminal tails. Whereas the N-terminal tail acetylations predominantly recruit effector proteins through bromodomains and other reader modules, histone fold domain acetylations act directly on nucleosomal DNA dynamics(6, 8, 33). This dual mode of regulation, via effector-mediated signaling and direct modulation of nucleosome architecture, may enable cells to precisely coordinate chromatin remodeling and transcription in response to specific cues(34). Our work provides a foundation for future studies to investigate how distinct histone fold domain acetylations interact with other chromatin modifications, remodeling factors, and transcription machinery across different physiological and developmental contexts, thereby advancing our understanding of how the nucleosome structure encodes regulatory information.

Material and methods

Preparation of histones

Human histones H2A, H2B, and H4 were expressed and purified as described previously(35). Briefly, H2A, H2B, and H4 were expressed in Escherichia coli cells. After denaturation of the insoluble fractions, the 6×histidine-tagged histones were purified by Ni-NTA agarose chromatography (Qiagen). The 6×histidine-tags were removed by thrombin protease treatment and the histones were further purified by MonoS cation-exchange chromatography.

Unmodified human histone H3.2 C110A was purified as previously described(36). In summary, the chemically synthesized H3 N-terminal peptide (residues 1-28) was ligated to the recombinant C-terminal regions of H3 (A29C-135, C110A) using the one-pot native chemical ligation method. After desulfurization, the H3 peptide was purified by high performance liquid chromatography. Because cysteine at position 29 was converted to alanine during desulfurization, the resulting peptide contained only the C110A mutation.

The human histone H3.2 C110A acetylated at Lys 56 (H3K56Ac) was previously synthesized(37) and used in this study. The other acetylated histones (H3K64Ac and H3K122Ac) were newly synthesized using a similar procedure. Briefly, the full-length H3.2 sequence was divided into N-terminal, internal, and C-terminal peptide segments. Each segment was prepared through Fmoc solid-phase peptide synthesis as an N-terminal free amine and a C-terminal thioester for the N-terminal segment, an N-terminal thiazolidine (Thz) and a C-terminal thioester for the internal segment, and an N-terminal Cys and a C-terminal carboxylic acid for the C-terminal segment. To assemble the peptide segments, sequential peptide ligation in the C-to-N direction was conducted according to the following procedure: 1) ligation between the internal and C-terminal segments and subsequent Thz-to-Cys conversion in a one-pot manner followed by HPLC purification and 2) ligation between the N-terminal and internal/C-terminal segments followed by HPLC purification. After desulfurization, full-length H3K64Ac and H3K122Ac were isolated as white powders after the final HPLC purification and lyophilization steps.

Preparation of DNA fragment

The 198 bp DNA fragment containing the 9-base mismatched region was purified as previously described(24). In brief, the plasmid containing the modified Widom 601 DNA fragment was amplified in E. coli. The Widom 601 DNA fragment was excised from the plasmid DNA by EcoRV (Takara), and purified by polyethylene glycol precipitation. After dephosphorylation with alkaline phosphatase (Takara), the DNA fragment was digested with BglI (Takara). The resulting fragment was purified by a Prep Cell apparatus (Bio-Rad). The purified 153 bp DNA fragment was ligated with the 45 bp DNA fragment containing the 9-base mismatched region by T4 DNA ligase (NIPPON GENE), and further purified by a Prep Cell apparatus (Bio-Rad).

Reconstitution of nucleosomes

The nucleosomes were reconstituted as previously reported(35). In brief, the histone octamers were assembled from H2A, H2B, H4, and either unmodified or acetylated H3 histones, and purified by size-exclusion chromatography on a HiLoad 16/60 Superdex 200 column (Cytiva). The resulting octamers were mixed with the 198 bp DNA fragment, and the nucleosomes were reconstituted using the salt-dialysis method. The nucleosomes were further purified by a Prep Cell apparatus (Bio-Rad).

Preparation of RNAPII and TFIIS

Komagataella phaffi RNAPII was purified as previously described(38–40). In summary, endogenous Rpb2 containing the tandem affinity purification (TAP) tag was purified from K. phaffi cells by FLAG-affinity column chromatography and anion-exchange column chromatography.

K. phaffi TFIIS was prepared as previously described(40). Briefly, the TFIIS protein was bacterially produced and purified by Ni-affinity column chromatography. After removing the His-tag by HRV-3C protease, the TFIIS protein was further purified by cation-exchange column chromatography.

Transcription assay

The transcription assay was conducted as previously described(24, 25). In brief, the nucleosome and RNAPII were incubated at 30°C for 30 minutes in reaction buffer, containing 0.1 μM nucleosome, 0.1 μM RNAPII, 0.1 μM TFIIS, 26 mM HEPES-KOH (pH 7.5), 50 mM potassium acetate, 0.2 μM zinc acetate, 20 μM Tris(2-carboxyethyl) phosphine, 1.5% glycerol, 0.1 mM dithiothreitol, 5 mM magnesium chloride, 400 μM each ATP, UTP, GTP, and CTP, and 0.4 μM fluorescently labeled RNA primer (5′-DY647-AUAAUUAGCUC-3′) (Dharmacon). The reaction was quenched by mixing 2 μl of the reaction mixture and 1 μl of deproteinization solution containing 200 mM Tris-HCl (pH 8.0), 80 mM EDTA, and 0.6 mg/ml proteinase K (Roche), followed by an incubation at room temperature for 10 minutes. Then, 12 μl of Hi-Di formamide was added, and the sample was heated at 95°C for 5 minutes. The sample was analyzed by 10% denaturing PAGE. The DY647 fluorescence of RNA products was detected by an Amersham Typhoon imager (Cytiva). The band intensities of the transcripts were quantified with the ImageQuant™ TL software (GE Healthcare) and normalized with the transcripts of the unmodified nucleosome template.

Purification of PL2-6 scFv

PL2-6 scFv containing (GGGS)4 linker peptides between the heavy and light chains was purified as previously described(25, 26, 41). In short, the scFv was bacterially produced and recovered as an insoluble fraction. After denaturation, the histidine-tagged scFv was purified by Ni-NTA agarose chromatography (Qiagen). The resulting scFv was refolded by dialysis with buffer exchange using a peristaltic pump, and further purified on a HiLoad 26/600 Superdex 75 pg column (Cytiva). Eluted fractions were analyzed by SDS-PAGE under both reducing conditions with an incubation at 95°C for 5 minutes and non-reducing conditions without heating. The binding activity of the scFv in each fraction was assessed by an electrophoretic mobility shift assay using nucleosomes. Fractions exhibiting nucleosome-binding activity were combined, concentrated, and stored at -80°C.

Sample preparation for cryo-EM analysis

Samples of unmodified H3, H3K56ac, H3K64ac, and H3K122ac nucleosomes for cryo-EM analysis were prepared as previously described(25). Nucleosomes (0.5 μM) and 1.5 μM of PL2-6 scFv were mixed in reaction buffer, containing 10 mM HEPES-KOH (pH 7.5), 0.5 mM dithiothreitol, 2.5% glycerol, 10 mM Tris-HCl (pH 7.5), 30 mM sodium chloride, and 0.2 mM EDTA. After an incubation at 30°C for 30 minutes, the mixture was applied to a Quantifoil R1.2/1.3 copper grid and plunge-frozen using a Vitrobot Mark IV (Thermo Fisher Scientific). Grids were glow discharged for 1 minute using a PIB-10 Bombarder (Vacuum Device Inc.) prior to sample application.

Cryo-EM data collection

The cryo-EM data were acquired using the EPU automation software on a Krios G4 electron microscope (Thermo Fisher Scientific). The microscope was operated at 300 kV, with a nominal magnification of 81,000× (corresponding to a pixel size of 1.06 Å), and a defocus range of -1.25 to -2.25 μm. Micrographs were recorded on a K3 BioQuantum direct electron detector (Gatan) with a 20 eV energy-filter, with a total of 40 frames per movie. The detailed cryo-EM data recording conditions are provided in Table 1.

Image processing

The overall workflow is shown in Supplementary Figure 3-6. The acquired movies were aligned and dose-weighted using MOTIONCOR2(42). CTFFIND4(43) was used to estimate the contrast transfer function (CTF). Relion 5.0(44) was used for image processing. Particles were picked by Laplacian-of-Gaussian based auto picking and subjected to two-dimensional classification, de novo initial model generation, three-dimensional (3D) classification, 3D auto-refinement, Bayesian polishing, and CTF refinement. Junk particles were removed during the classification steps. To focus on the nucleosome structure, 3D classification with a nucleosome mask was conducted with and without alignment. The resolution of the final maps was estimated from the gold standard Fourier shell correlation (FSC = 0.143) criterion. The final cryo-EM map of the H3K56ac nucleosome was flipped along the z-axis using ChimeraX(45).

Model building and refinement

The initial atomic models were built based on the coordinates of the modified Widom 601 DNA sequence from the RNAPII-nucleosome complex (Protein Data Bank (PDB) ID: 6A5O(24)) and the coordinates of the histone octamer from the human nucleosome (PDB ID: 7VZ4(46)). Acetyl-lysine residues were introduced using Coot(47). The atomic coordinates were fitted to the cryo-EM maps by rigid-body fitting using UCSF ChimeraX(45). The fitted coordinates were refined by phenix.real_space_refine(48), and manually adjusted by Coot(47) and ISOLDE(49). Refinement and validation statistics are summarized in Table 1. The distances between atoms were measured using USCF ChimeraX(45).

Micrococcal nuclease assay

Nucleosomes (0.125 μM) were incubated in a reaction mixture, containing 10 mM HEPES-KOH (pH 7.5), 2.5% glycerol, 1.9 mM dithiothreitol, 40 mM Tris-HCl (pH 8.0), 25 mM sodium chloride, 2.5 mM calcium chloride, and 10 mU/μl micrococcal nuclease (Takara) at 37°C for 10 minutes. A 4 μl portion of the reaction mixture was mixed with 2 μl of deproteinization solution, containing 20 mM Tris-HCl (pH 8.0), 20 mM EDTA, 0.1% SDS, and 0.5 mg/ml proteinase K (Roche), and incubated at room temperature for 10 minutes. The samples were analyzed by 8% non-denaturing PAGE and stained with ethidium bromide. Gel images were acquired by an Amersham Imager 680 (Cytiva). The band intensities of the DNA were quantified using the ImageQuant™ TL software (GE Healthcare) and normalized with those of the undigested DNA.

The use of Large Language Models

ChatGPT, a Large Language Model developed by OpenAI, was used in a strictly auxiliary role to support the writing process and graph generation. This included assistance with phrasing, grammar correction, and the generation of code examples in Python. The authors remain fully responsible for the content and conclusions of the manuscript.

Data availability

The cryo-EM maps of the unmodified, H3K56ac, H3K64ac, and H3K122ac nucleosomes were deposited to the Electron Microscopy Data Bank (EMDB) and the Protein Data Bank (PDB), respectively. The accession codes are as follows: unmodified nucleosome (EMD-66767 and PDB ID 9XDM), H3K56ac nucleosome (EMD-66768 and PDB ID 9XDN), H3K64ac nucleosome (EMD-66769 and PDB ID 9XDO), and H3K122ac nucleosome (EMD-66770 and PDB ID 9XDP).

Acknowledgements

We thank Y. Takeda and M. Dacher (The University of Tokyo) for their assistance. We also thank M. Goto and M. Henmi (RIKEN) for their assistance with protein purification.

Additional information

Author contributions

Conceptualization: TO, TK, GH, YT, SSe, HK

Methodology: TO, SH, TK, SSu, KN, GH, SSe, YT, HK

Resources: TO, SSu, KN, GH, JK, YK, HE, HK

Investigation: TO, SH, SSu, KN, GH, YT, HK

Visualization: TO, SH, MO, YT, HK

Funding acquisition: TO, SH, TK, GH, SSe, YT, HK

Project administration: TK, GH, SSe, YT, HK

Supervision: TK, GH, SSe, YT, HK

Writing – original draft: TO, HK

Writing – review & editing: TO, SH, TK, GH, HE, SSe, YT, HK

Funding

MEXT | Japan Society for the Promotion of Science (JSPS) (JP25KJ1033)

• Takumi Oishi

MEXT | Japan Society for the Promotion of Science (JSPS) (JP25K18410)

• Suguru Hatazawa

MEXT | Japan Society for the Promotion of Science (JSPS) (JP23K17392)

• Tomoya Kujirai

MEXT | Japan Society for the Promotion of Science (JSPS) (JP24H00062)

• Shun-ichi Sekine

• Tomoya Kujirai

MEXT | Japan Society for the Promotion of Science (JSPS) (JP25K24594)

• Hitoshi Kurumizaka

MEXT | Japan Society for the Promotion of Science (JSPS) (JP24H02319)

• Hitoshi Kurumizaka

MEXT | Japan Society for the Promotion of Science (JSPS) (JP24H02328)

• Hitoshi Kurumizaka

MEXT | Japan Science and Technology Agency (JST) (JPMJSP2108)

• Takumi Oishi

MEXT | Japan Science and Technology Agency (JST) (JPMJPR25N9)

• Tomoya Kujirai

MEXT | Japan Science and Technology Agency (JST)

https://doi.org/10.52926/jpmjer1901

• Hitoshi Kurumizaka

MEXT | Japan Science and Technology Agency (JST)

https://doi.org/10.52926/jpmjcr24t3

• Hitoshi Kurumizaka

Japan Agency for Medical Research and Development (AMED) (JP25ama121002)

• Yoshimasa Takizawa

Japan Agency for Medical Research and Development (AMED) (JP25ama121009)

• Hitoshi Kurumizaka

• Gosuke Hayashi

Additional files

Supplementary figures