Functional Analyses of an Evolutionarily Conserved Acidic Patch on the Nucleosome

Abstract

The eukaryotic canonical nucleosome has an acidic patch on each H2A/H2B dimer. This acidic patch is also detected in histone variants, such as the H2A.Z (yeast Htz1)/H2B dimer. Here, we screened a comprehensive histone point mutant library and identified 11 histone residues located in four distinct nucleosome domains (Homologous Recombination (HR) Domain I–IV (HRD-I–IV)) with a potential role in HR. H2A-L66, -E93, and -L94 residues in HRD-I are located in the acidic patch region. Equivalent residues (H2A-L66 and Htz1-L73) partly compensate the function of each dimer. A common residue H2B-L109, which is located underneath of the acidic patch in both dimers, also partly compensates the function of each dimer. Upon exposure to DNA double-strand break (DSB)-inducing agents, the fragmented chromosomes of H2A-L66A mutant cells exhibited slow and limited recovery into intact chromosomes, suggesting that the H2A-L66A mutant is partly deficient in DSB repair. Furthermore, strand invasion, one of critical steps of HR, could be less efficient in H2A-L66A cells. All 11 HRD residues, including H2A-L66, are highly conserved in extant eukaryotic cells; therefore, our screening reported in this study will provide a foundation for future studies about the mechanisms underlying eukaryotic HR based on chromatin.

INTRODUCTION

Unrepaired DNA double-strand breaks (DSBs) are lethal for cells.¹⁾ In eukaryotic cells, homologous recombination (HR) and non-homologous end-joining (NHEJ) are responsible for DSB repair.¹⁾ Mutant cells lacking proteins involved in HR, such as Rad51 and Rad52, show sensitivity to both the DNA damaging agent, methylmethane-sulfonate (MMS), and the DNA replication inhibitor, hydroxyurea (HU). Furthermore, such mutations are lethal in the absence of the RAD27 gene, which encodes an FEN1 nuclease involved in Okazaki fragment processing during DNA replication²⁾ (Fig. 1A).

Fig. 1. Eleven Histone Residues Produce Severe Growth Defects When Mutated in Combination with rad27 Deletion

(A) Schematic representation of representative genes that are synthetically lethal when mutated in combination with rad27 deletion. (B) Residues exhibiting MMS (left) or HU (right) sensitivity when mutated are mapped on the nucleosome surfaces (PDB ID: 1ID3). Blue and yellow represent residues that are sensitive to MMS and HU, respectively, when mutated. The Venn diagram indicates 48 residues that show sensitivity to both MMS and HU when mutated. (C) Schematic representation of the principal of screening for histone point mutants showing synthetic lethality with rad27 deletion by spot assay. (D, E) Results of synthetic lethal screening against the 48 histone mutants in Fig. 1B. Exposed residues are underlined. Eleven histone point mutants showing severe growth defects with rad27 deletion are marked in red. (F–J) Enlarged view of the position of the 11 histone residues in D and E in a canonical nucleosome (PDB ID: 1ID3). The 11 histone residues, named HRD residues, are located at four different regions within the nucleosome.

Although transacting proteins involved in HR have been extensively characterized,¹⁾ the role of the nucleosome core in repair of DSBs remains elusive. A nucleosome comprises both the wrapped 146 bp of DNA and a histone octamer composed of two histone H2A/H2B dimers and a (H3/H4)₂ tetramer.^3,4) To elucidate the function of each histone residue, comprehensive functional analyses of budding yeast histone point mutant cells have been performed^5,6) and have identified 90 MMS- and 50 HU-sensitive mutants (Fig. 1B). Of the histone mutants identified, 48 showed sensitivities to both MMS and HU, indicating their potential involvement in HR.

Here, we aimed to further select strong candidate histone residues involved in HR by performing synthetic lethal screening of these 48 histone mutants with the rad27 deletion (Fig. 1C). We identified 11 histone residues (Figs. 1D, E) localized at four distinct nucleosome domains, which we have named Homologous Recombination Domain I–IV (HRD-I–IV), with potential roles in HR (Figs. 1F–J). The molecular functions of HRD-I in DSB repair were characterized (Figs. 2–4).

Fig. 2. Eleven HRD Residues Are Conserved in All Extant Eukaryotes

(A) Schematic representation of the principal of FALC strategy. (B) Function of fused H2B-H2A or fused H2B-Htz1 carrying a H2B L109A point mutation was evaluated using a spot assay against HU, MMS, benomyl, or 6AU. (C) Comparison of HRD residues and residues that are lethal when mutated, among a variety of eukaryotes. Colored closed circles represent conserved HRD residues; HRD-I (red), -II (green), -III (blue), and -IV (brown). Open circles represent conserved residues that are lethal when mutated. Several amino acid variations compared with S. cerevisiae H2A are indicated using the single letter amino acid code. Exposed residues are underlined.

Fig. 3. Cooperative Action of Acidic Patches of Both Canonical and Htz1 Nucleosomes

(A) Summarized sequence comparison among H2A histone variants. Conservation is represented as red closed, closed, and open circles (HRD-I, HU/MMS, and lethal residues when mutated, respectively). Exposed residues are underlined. (B) Schematic representation explaining the principal of the assay in C. (a) Wild-type H2A (orange nucleosome) and Htz1 (blue nucleosome) (b) H2A single point mutant containing red dot. (c) Htz1 single point mutant containing red dot. (d) Double point mutants in both H2A and Htz1 at the same position, where all nucleosomes have a red dot. (C) Spot assay against four different drugs as described in 2B. Double H2A and Htz1 mutants were compared with H2A or Htz1 single mutants. Exposed residues are underlined. (D) Scheme for analyses of PFGE showing a time course of cell sampling after 1 h Phl treatment. (E) PFGE analysis. In lanes a, f, and k (before Phl treatment), discreate bands represent budding yeast chromosomes. In lanes b, g, and l (just after Phl treatment), bands are smeared due the presence of a large number of DSBs.

Fig. 4. A H2A-L66A Mutant Exhibits Partial Deficiency in the Strand Invasion Step of HR

(A) Schematic representation of the ectopic inducible cut by HO endonuclease at the MAT locus. During homologous recombination repair of DSB induced by HO, resection occurs on both sides of the HO cut, Rad51 then interacts with ssDNA, and subsequently the Rad51-ssDNA filament invades template homologous DNA. (B) and (C) Resection steps were monitored by quantitative PCR on the right side of the HO cut. The principal of the resection assay is presented in Supplementary Fig. S1. Data points represent the mean of four independent experiments. Error bars represent the standard deviation. N.s. represents statistically not significant. (D) The strand invasion step was monitored by ChIP analyses of Rad51-3HA protein at both sides of the HO cut. We used different primer sets in the resection and Rad51 ChIP assays according to the different principals of these assays. Error bars represent the standard deviation (n = 4, ***, p < 0.005). Notably, the total amount of Rad51 during the HO cut experiment in H2A-L66A cells was almost equivalent to that of wild-type cells (Supplementary Fig. S8). (E) The graphical results presented in Figs. 4C and D were summarized.

MATERIALS AND METHODS

Yeast Strains

Yeast strains used in this study are listed in Supplementary Table S1.

Primer Sequences

DNA sequences of primer sets for the resection assays and chromatin immunoprecipitation (ChIP) analyses used in this study are listed in Supplementary Table S2.

Spot and Streaking Assays

Agar plates for the synthetic lethal assay were prepared with synthetic complete medium lacking histidine for YKH2AB, or leucine for YKH34, supplemented with 0.5 mg/mL 5-fluoroorotic acid (5-FOA). In the presence of 5-FOA, wild-type histone genes in the plasmid containing the URA3 marker are counter-selected and removed from cells, resulting in cells harboring only mutant histone genes. Agar plates for drug sensitivity assays were prepared with synthetic complete medium supplemented with 100 mM HU, 0.015% MMS, 10 µg/mL benomyl (diluted in dimethyl sulfoxide) (tubulin-depolymerizing agent), or 1 mg/mL 6-azaurasil (6AU) (transcription elongation inhibitor). Three-fold serial dilutions of the indicated strains were spotted onto agar plates with or without benomyl, starting at 1 × 10⁵ cells, and incubated for 3–5 d at 25 °C. Experiments were performed in duplicate and repeated several times. In the streaking assay, cells derived from a single colony of each genotype were streaked onto agar plates with or without 5-FOA, and incubated for 3–5 d at 25 °C.

Pulsed-Field Gel Electrophoresis (PFGE)

PFGE experiments were performed as described previously.⁷⁾ PFGE can separate each yeast chromosome as a discrete band. Briefly, logarithmically growing cells were diluted to 8 × 10⁶ cells/mL, exposed to 100 µg/mL phleomycin (Phl) (DSB-inducing agent) for 1 h at 30 °C, washed to remove Phl, and then cultured at 30 °C in YPAD medium for the indicated periods of time. Agarose plugs containing chromosomal DNA were subsequently treated with the CHEF Yeast genomic DNA Plug kit (Bio-Rad, Hercules, CA, U.S.A.). After electrophoresis, gels were stained with 0.5 µg/mL ethidium bromide for 30 min, destained in deionized water for 20 min, and then imaged.

DSB End Resection Assay

HO (Homothallic switching) endonuclease, a type of restriction enzyme that recognizes a specific sequence at the mating-type (MAT) locus, can introduce only one DSB in the whole genome. Quantitative analysis of HO-induced DSB end resection using real-time PCR was performed as previously described.⁸⁾ Briefly, logarithmically growing cells were diluted in YPA + raffinose medium containing 15 µg/mL nocodazole (tubulin-depolymerizing agent) and cultured for 2.5 h to arrest cells at G2/M phase. Transcriptional expression of HO endonucleases was controlled by the GAL1 promoter; therefore, HO endonucleases were not expressed in medium containing raffinose. Cells were exposed to 2% galactose for 1 h to induce DSB at MATα locus mediated by HO endonuclease expression. Subsequently, cells were cultured in medium containing 2% glucose to repress HO endonuclease induction for the indicated periods of time. Genomic DNA from each cell was collected and treated with Sty I or Xba I (TaKaRa, Shiga, Japan). The principal of the resection assay is presented in Supplementary Fig. S1. After DNA purification, quantitative PCR data were obtained using the Thermal Cycler Dice Real-Time System (TaKaRa).

ChIP Assay

The ectopic HO endonuclease assay was performed the same as the DSB end resection assay, and ChIP analyses of yeast cells were performed as previously described.^7,9) Briefly, immunoprecipitation of hemagglutinin (HA)-tagged Rad51 was performed using Dynabeads Protein G (Invitrogen, Waltham, MA, U.S.A.) conjugated with anti-HA (3F10, Roche, Basel, Switzerland) antibody. Quantitative expression data were obtained using the Thermal Cycler Dice Real-Time System (TaKaRa).

Molecular Graphics

Molecular graphics were prepared using the PyMOL program (http://www.pymol.org).

RESULTS

Histone Residues Important for HR Are Either Exposed or Buried in the Nucleosome

Underlined histone residues are exposed, while non-underlined histone residues are buried in the nucleosome, as defined in a previous report,⁶⁾ and this nomenclature will be used henceforth. Synthetic lethal screening of 48 previously identified histone mutants in a rad27 deletion background resulted in the identification of 11 histone point mutants (H2A-L66A, -E93A, -L94A, H2B-D71A, H3-F54A, -I62A, -E97A, -H113A, -L130A, H4-R36A, and -Y98A) that showed severe growth defects under these conditions (Figs. 1D, E). We therefore hypothesized that these histone residues are likely to be important for HR. The 11 histone residues are located in four distinct domains of the nucleosome (Fig. 1F), I (H2A-L66, -E93, -L94) (Fig. 1G), II (H3-F54, -I62, -E97, H4-R36) (Fig. 1H), III (H3-H113, -L130) (Fig. 1I), and IV (H2B-D71, H4-Y98) (Fig. 1J), which we have called HRD-I-IV, respectively.

H2A-E57, -Y58, -E62, -E65, -L66, -D91, -E93, and -L94, together with H2B-L109 (Supplementary Fig. S2A), interact with the arginine residues of a variety of histone interacting proteins.¹⁰⁾ Remarkably, H2A-Y58, -E62, -D91, and H2B-L109 are lethal when mutated,^5,6) suggesting that HRD-I residues (H2A-L66, -E93, and -L94) act together with these four essential residues.

An interaction matrix of histone residues in HRD-II (H3-F54, -I62, -E97, and H4-R36) (Supplementary Figs. S2B, C) suggests that they are structurally and functionally related to H3-L48, -I51, -Q55, H4-R39, and -R40 residues, all of which are lethal when mutated.^5,6) Of these nine residues, seven are buried in the nucleosome (three HRD-II and four lethal residues when mutated), which leads us to hypothesize that HRD-II acts as some form of internal structural switch during HR. The scenario is supported by the structural instability of the human H3.6 variant nucleosome because a variant nucleosome containing H3-V62, which corresponds to H3-I62 (HRD-II residue) of canonical H3, is structurally unstable in vitro¹¹⁾ (Supplementary Fig. S3). Moreover, in eukaryotic cells, where HR and NHEJ are responsible for repairing DSBs, double mutants in which the YKU70 gene, whose product is involved in NHEJ, was deleted and HRD-I was mutated (H2A-L66A) were viable (Supplementary Fig. S4A). By contrast, double mutant cells of HRD-II (H3-F54A, -I62A, or -E97A) with yku70 deletion showed severe growth defects (Supplementary Fig. S4B). This suggests that HRD-II acts as a critical switch for DSB repair choice.

An interaction matrix of histone residues for HRD-III (H3-H113 and -L130) (Supplementary Fig. S2D) suggests that HRD-III residues are structurally and functionally related to H3-R116, -T118, -D123, and H4-R45 residues, all of which are lethal when mutated.^5,6) Four of these six residues are buried in the nucleosome (one HRD-III and three lethal residues when mutated), suggesting that HRD-III might also act as an internal switch for (H3-H4)₂ tetramer–DNA interaction and/or H3–H3′ interaction at the dyad during HR.

Similarly, an interaction matrix of histone residues for HRD-IV (H2B-D71 and H4-Y98) (Supplementary Fig. S2E) suggests that HRD-IV might act as an internal switch for association/de-association between the H2A-H2B dimer and the (H3-H4)₂ tetramer during HR. Given that the critical HRD-II, -III, and -IV residues are predominantly buried in the nucleosome, it is likely that efficient nucleosome disassembly/assembly utilizing those HRD residues would be necessary for successful HR.

H2B-L109 Is Involved in Processes That Protect against HU or MMS Exposure

Nineteen of the residues that are HU/MMS sensitive (including HRD-I residues) and lethal when mutated (H2A-E57, -Y58 [lethal], -E62 [lethal], -E65, -L66 [HRD-I], -G68, -D73, -N74, -K76, -R78, -R82, -H83, -L86, -D91 [lethal], -E93 [HRD-I], -L94 [HRD-I], and H2B-L109 [lethal], -R119, -K123) are clustered at the nucleosome surface. Given that 15 of these residues are exposed to solvent, the surface resulting from these 19 residues would represent an excellent docking site for a variety of nucleosome interacting factors.¹⁰⁾ Although it is impossible to examine whether residues H2A-Y58, -E62, and -D91, all of which are lethal when mutated, are involved in protecting against HU or MMS exposure; this can be done for H2B-L109.

We have developed an artificially linked H2B-H2A fusion, employing the FALC (functional analysis of linker mediated complex) strategy¹²⁾ (Fig. 2A), and used this to analyze the function of H2B-L109. H2B-L109 is functionally important in both the H2A/H2B dimer and the Htz1 (human H2A.Z homolog)/H2B) dimer. Although H2B-L109 is lethal when mutated,⁶⁾ cells harboring either a H2B(L109A)-H2A fusion or a H2B(L109A)-Htz1 fusion do not show lethality.¹²⁾ As shown in Fig. 2B, cells harboring the H2B(L109A)-H2A fusion, but not the H2B(L109A)-Htz1 fusion, showed sensitivity to both HU and MMS, as well as benomyl and 6AU, indicating that H2B-L109 in the H2A/H2B dimer is involved in protecting against HU or MMS exposure. This might suggest that the other three lethal when mutated residues, H2A-Y58, -E62, and -D91, are also involved in HU or MMS exposure protective processes.

Notably, 13 residues that are lethal when mutated (H2A-Y58, -E62, -D91; H2B-L109, H3-L48, -I51, -Q55, -R116, -T118, -D123, H4-R39, -R40, and -R45) are structurally and functionally related to HRD-I, -II, or -III (Fig. 2C, Supplementary Fig. S2). Additionally, we have previously identified 15 histone residues that are lethal when mutated,^5,6) and thus we further investigated a total of 26 residues (15 lethal when mutated residues as well as 11 HRD residues) in this study. Furthermore, although two essential residues, H4-Y72 and -L90, are not related to any HRD residues, we also included them in the following discussion.

Most HRD Residues and Histone Residues That Are Lethal When Mutated Are Evolutionarily Conserved

The 26 identified residues in budding yeast (15 lethal when mutated residues as well as 11 HRD residues) are almost identical to human ones, as well as those from other eukaryotes (Fig. 2C). Thus, the structural and functional roles of these 26 residues seems to have been established prior to the emergence of the last eukaryote common ancestor (LECA). This would suggest that common HR mechanisms are likely to operate in any extant eukaryote since the emergence of the LECA. Humans have a variety of H2A, H2B, and H3 variants. Structural analyses of nucleosome containing histone variants are summarized in Supplementary Fig. S3, and these indicate that HRD-I, -II, -III, and -IV residues as well as the residues that are lethal when mutated (Fig. 2C) are well conserved amongst any human nucleosome (Supplementary Fig. S3). Thus, in human cells, HR can operate at any region of chromatin in response to DNA damage.

The Function of the Acidic Patch Is Common between a Canonical Histone and Its Variants

Histone variants are derived from duplication of canonical histone genes.¹³⁾ Budding yeast has one H2A variant, Htz1. In humans, there are a variety of histone variants including H2A.Z, H2A.X, MacroH2A, and H2A.B.¹³⁾ Canonical eukaryotic histone residues located at the acidic patch are also well conserved in histone H2A variants (Fig. 3A, Supplementary Fig. S3). Thus, the structure and function of histone residues around the acidic patch seem to have been established before the emergence of the LECA and inherited into canonical H2A and its variants (Fig. 3A, Supplementary Fig. S3).

Although comprehensive histone point mutant screens have been performed in the canonical histone H2A^5,6) (Figs. 3B(a)(b)), as well as Htz1¹⁴⁾ (Fig. 3B(a)(c)), the effect of simultaneous histone mutation at corresponding residues on both H2A and Htz1 has never been tested (Fig. 3B(d)). To this end, we examined several combinations of H2A and Htz1 point mutants. As shown in our spot assay (Fig. 3C), when both H2A and Htz1 homologous residues are simultaneously mutated, the resultant phenotypes are more severe than those seen in cells with single mutations, suggesting that in both H2A and Htz1 these residues act in a compensatory manner in chromatin based cellular reactions.

H2A-L66 Is Involved in HR

We next focused on the function of H2A-L66 (Figs. 1D, 1G, 2C, 3A, 3C). Treatment of H2A-L66A cells with HU, MMS, or Phl did not affect phosphorylation of the checkpoint protein Rad53 (Supplementary Fig. S5), implying that the checkpoint is largely intact in H2A-L66A cells. Moreover, phosphorylation of H2A-S128, which is a hallmark of the DSB response, occurred at the same level as wild-type cells in H2A-L66A cells upon MMS or Phl exposure (Supplementary Fig. S5), suggesting that the signaling pathway from DSB to H2A-S128ph is similarly intact in H2A-L66A cells.

Next, upon 1 h exposure to the DSB-inducing agent, Phl (Fig. 3D), all chromosomes were fragmented in wild-type, H2A-L66A, and rad52 cells, as measured by PFGE (Fig. 3E, lanes b, g, l) compared with untreated cells (lanes a, f, k). In wild-type cells, chromosomes were recovered during an 8 h incubation, but not in HR-deficient mutant rad52 cells. Under the same conditions, H2A-L66A cells showed slow and limited recovery of chromosomes after Phl exposure (Fig. 3E, lanes h–j), suggesting that repair of DSBs is partly deficient in H2A-L66A cells.

Next, we examined the repair of a single DSB introduced by induction of HO endonuclease in cells (Figs. 4A, B). To identify which step of HR-mediated DSB repair is defective in H2A-L66A cells, we examined two steps: 1) resection of the right arm of a HO endonuclease induced single DSB, and 2) strand invasion to homologous DNA mediated by the Rad51-DNA filament. When resection was examined using genomic PCR, the resection rate of H2A-L66A cells was comparable with that of wild-type cells (Fig. 4C). Since strand invasion requires Rad51 binding to single-stranded DNA (ssDNA) regions (Fig. 4A), ChIP analyses were used to measure Rad51 binding to ssDNA. These analyses demonstrated that the amount of DNA bound to Rad51 was lower at both ends in H2A-L66A cells than in wild-type cells (Fig. 4D). The results presented in Figs. 4C and D are graphically summarized in Fig. 4E and predict that strand invasion was less efficient in H2A-L66A cells.

DISCUSSION

We identified 11 HRD residues, which were classified into HRD-I-IV (Fig. 1). Most of these 11 HRD residues are interconnected with other HRD residues and/or residues that are lethal when mutated (Fig. 2, Supplementary Fig. S2). Here, we mainly characterized HRD-I residues (Figs. 2–4). HRD-I seems to have existed before the emergence of the LECA (Fig. 2C), and been inherited by canonical H2A and its variants (Fig. 3A, Supplementary Fig. S3). Furthermore, our findings suggest that H2A-L66 (HRD-I) is involved in HR repair (Fig. 4), especially in efficient strand invasion steps. Comprehensive proteomic analyses demonstrate that half of nucleosome-binding proteins interact with the acidic patch,¹⁰⁾ suggesting that nucleosome-binding proteins involved in HR facilitate H2A-L66-mediated HR repair. A variety of chromatin remodelers interact with the acidic patch of the nucleosome (Supplementary Fig. S6). Furthermore, the Rad54 chromatin remodeler is known to be involved in efficient strand invasion steps during HR repair,¹⁾ suggesting that Rad54 is a strong candidate for interaction with the H2A-L66 residue (Fig. 4E).

Apart from the nucleosome, a variety of mechanical switches that operate inside proteins and/or multi-subunit complexes such as Myoglobin, Hemoglobin, Rhodopsin, Aquaporin, Potassium channel, RNA polymerase II, Ribosome, and G-protein coupled receptor, have been structurally solved. Despite numerous structural biology studies of nucleosome and histone subunits,¹⁵⁾ it is still difficult to identify mechanical switches inside the nucleosome because unlike hemoglobin, the nucleosome is disassembled into two H2A-H2B dimers, a (H3-H4)₂ tetramer, or two H3-H4 dimers, and DNA. Structural biology can provide insights into only two states: either the intact nucleosome or disassembled histone subunits. Consequently, the action of a mechanical switch inside the nucleosome cannot be easily resolved using structural biology analyses alone. Thus, incorporation of information from the histone-GLibrary^5,6,9,14) (Supplementary Fig. S7) into cellular, biochemical, and structural analyses of nucleosomes is essential to understand highly conserved (but uncharacterized) mechanical switches hidden inside the nucleosome.

Acknowledgments

We thank Dr. M. Horikoshi (Jobu University) for our initial analyses of histone-GLibrary.^6,9,12) The practical studies here have been based on research projects conducted by several past graduate students (Y. Okada, H. Yoshida, and G. Ueno in Tohoku University) and past undergraduate students (H. Shitara and C. Matsukawa in Tohoku Medical and Pharmaceutical University); we thus thank them all. This study was financially supported by Tohoku Medical and Pharmaceutical University.

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

This article contains supplementary materials.

REFERENCES

• 1) Jasin M, Rothstein R. Repair of strand breaks by homologous recombination. Cold Spring Harb. Perspect. Biol., 5, a012740 (2013).
• 2) Debrauwère H, Loeillet S, Lin W, Lopes J, Nicolas A. Links between replication and recombination in Saccharomyces cerevisiae: a hypersensitive requirement for homologous recombination in the absence of Rad27 activity. Proc. Natl. Acad. Sci. U.S.A., 98, 8263–8269 (2001).
• 3) Luger K, Mader AW, Richmond RK, Sargent DF, Richmond TJ. Crystal structure of the nucleosome core particle at 2.8 Å resolution. Nature, 389, 251–260 (1997).
• 4) White CL, Suto RK, Luger K. Structure of the yeast nucleosome core particle reveals fundamental changes in internucleosome interactions. EMBO J., 20, 5207–5218 (2001).
• 5) Matsubara K, Sano N, Umehara T, Horikoshi M. Global analysis of functional surface of core histones with comprehensive point mutants. Genes Cells, 12, 13–33 (2007).
• 6) Sakamoto M, Noguchi S, Kawashima S, Okada Y, Enomoto T, Seki M, Horikoshi M. Global analysis of mutual interaction surface of nucleosomes with comprehensive point mutants. Genes Cells, 14, 1271–1330 (2009).
• 7) Ogiwara H, Ohuchi T, Ui A, Tada S, Enomoto T, Seki M. Ctf18 is required for homologous recombination-mediated double-strand break repair. Nucleic Acids Res., 35, 4989–5000 (2007).
• 8) Chen H, Lisby M, Symington LS. RPA coordinates DNA end resection and prevents formation of DNA hairpins. Mol. Cell, 50, 589–600 (2013).
• 9) Kawashima S, Nakabayashi Y, Matsubara K, Sano N, Enomoto T, Tanaka K, Seki M, Horikoshi M. Global analysis of core histones reveals nucleosome surfaces required for chromosome bi-orientation. EMBO J., 30, 3353–3367 (2011).
• 10) Skrajna A, Goldfarb D, Kedziora KM, Cousins EM, Grant GD, Spangler CJ, Barbour EH, Yan X, Hathaway NA, Brown NG, Cook JG, Major MB, McGinty RK. Comprehensive nucleosome interactome screen establishes fundamental principles of nucleosome binding. Nucleic Acids Res., 48, 9415–9432 (2020).
• 11) Taguchi H, Xie Y, Horikoshi N, Maehara K, Harada A, Nogami J, Sato K, Arimura Y, Osakabe A, Kujirai T, Iwasaki T, Semba Y, Tachibana T, Kimura H, Ohkawa Y, Kurumizaka H. Crystal structure and characterization of novel human histone H3 variants, H3.6, H3.7, and H3.8. Biochemistry, 56, 2184–2196 (2017).
• 12) Nakabayashi Y, Kawashima S, Enomoto T, Seki M, Horikoshi M. Roles of common subunits within distinct multisubunit complexes. Proc. Natl. Acad. Sci. U.S.A., 111, 699–704 (2014).
• 13) Talbert PB, Henikoff S. Histone variants at a glance. J. Cell Sci., 134, jcs244749 (2021).
• 14) Kawano A, Hayashi Y, Noguchi S, Handa H, Horikoshi M, Yamaguchi Y. Global analysis for functional residues of histone variant Htz1 using the comprehensive point mutant library. Genes Cells, 16, 590–607 (2011).
• 15) Zhou K, Gaullier G, Luger K. Nucleosome structure and dynamics are coming of age. Nat. Struct. Mol. Biol., 26, 3–13 (2019).