|
Fig. 1. Identification of histone Kbz in yeast.a Detection of the Kbz signals from core histones extracted from BY4742 yeast cells. The specificity of pan anti-Kbz was confirmed because it did not recognize any band from the E.coli whole-cell lysate (E.coli WCL) or recombinant Xenopus laevis histone octamer (Xl octamer) purified from E.coli. The H3* represents clipping H3. Source data for panels a–c are provided in the Source Data file. b Sodium benzoate treatment significantly enhanced the Kbz and Kac levels on yeast histones in a dose-dependent manner. BY4742 cells were grown in YPD to log phase, and then sodium benzoate was added into the medium for 6 h, followed by extraction of histones and western blotting analysis. c The level of histone Kbz was regulated by the type of carbohydrate. BY4742 cells were grown in standard medium with 2% glucose to log phase and then transferred to H2O or other indicated mediums for 4 h, followed by extraction of histones and western blotting analysis. d Illustrations of all Kbz sites identified on core histones extracted from yeast cells treated with 10 mM sodium benzoate. The detected Kbz sites are shown in red. Nine Kbz sites detected in samples without sodium benzoate treatment are labeled by asterisks. Most Kbz sites overlap with the known Kac sites, except for H3K42, H3K115, and H4K44. e–h: MS/MS spectra of H3 and H4 peptides bearing Kbz modification. (e) H3K9; (f) H3K14; (g) H3K27; (h) H4K44. Predicted b- and y-type ions are listed above and below the peptide sequence, respectively. The circle symbol indicates the neutral loss of water. Matched ions are labeled in the spectra.
|
|
Fig. 2. Gcn5-containing complex is a histone benzoyltransferase.a Fluorescence quantitative western blot analysis comparing the Kbz and Kac levels in WT (BY4742) and HAT-deletion strains. Total H4 was used as the loading control. The fluorescence signals from mutant strains were normalized by setting the WT fluorescence signal to 1.0 and were labeled in the image. Source data for panels (a–c) are provided in the Source Data file. b The Ada2-Gcn5 subcomplex could catalyze acetylation of the histone octamer in vitro. The reaction products were detected by western blot analysis using a pan-Kac antibody. c Histone benzoylation can occur by enzymatic and non-enzymatic mechanisms. The reaction products in different combinations were detected by western blot analysis using a pan-Kbz antibody. d The list of LC-MS/MS-identified histone benzoylation sites deposited by the non-enzymatic mechanism and by Gcn5-Ada2. The unique sites catalyzed by Gcn5-Ada2 are labeled red. e Gcn5-Ada2 catalyzed histone benzoylation more efficiently than spontaneous reaction. The Y-axis represents the ratio of MS/MS peak areas under enzymatic and non-enzymatic conditions. This result is from a single experiment.
|
|
Fig. 3. Hst2 is a histone debenzoylase.a Fluorescence quantitative western blot analysis comparing the Kbz and Kac levels in WT (BY4742) and ten HDAC-deletion strains. The fluorescence signals from mutant strains were normalized by setting the WT fluorescence signal to 1.0 and were labeled in the image. The H3* represents clipping H3. Source data are provided in the Source Data file. b Debenzoylase activity of Hst2FL protein (2 μM) with H3K9bz peptide (10 μM) shown by the representative MALDI-TOF spectra at 0, 5, and 30 min. The peaks for H3K9bz (bz, m/z 2,337) and unmodified (un, m/z 2,233) products are labeled. c Deacetylase activity of Hst2FL protein (0.5 μM) with H3K9ac peptide (10 μM) shown by the representative MALDI-TOF spectra at 0, 5, and 30 min. The peaks for H3K9ac (ac, m/z 2,737) and unmodified (un, m/z 2695) products are labeled. d The Michaelis-Menten plot for Hst2 fixed at 0.1 μM with varied H3K9bz peptide concentrations. Data are presented as mean ± SD, n = 3 biological independent measurements. Source data for panels (d, e) are provided in the Source Data file. e The Michaelis–Menten plots for Hst2 fixed at 0.05 μM with varied H3K9ac peptide concentrations. Data are presented as mean ± SD, n = 3 biological independent measurements. f Hst2-mediated debenzoylation reaction monitored by time-resolved MALDI-TOF mass spectrometry. The Y-axis represents the percentage of Kbz peptide in total peptides calculated from the MALDI-TOF spectra. Data are presented as mean ± SD, n = 3 biological independent measurements. Source data are provided in the Source Data file. g ITC binding curves for Hst2 interaction with different Kbz peptides. h Sequence alignment of benzoylated H3 peptides used in the current study. The alignment is centered on the target lysine (yellow). Positively charged R/K residues are shown in blue. The dissociation constants (Kd) and enthalpy changes (ΔH) derived from the ITC assay (panel g) are shown.
|
|
Fig. 4. The structural basis for H3K9bz recognition by Hst2.a The overall structure of the Hst28-294–H36-14K9bz complex. The large domain and small domain in Hst2 are shown in green and cyan, respectively. H3K9bz peptide is presented as a stick model (yellow). Four glycine residues located at the junctions between the large and small domains are labeled. b Superimposition of Hst2-H3K9bz and Hst2-2′-O-benzoyl-ADP-ribose complexes indicates the relative 17° rotation of the small domain and the rearrangement of the α2-α3 loop. The large domain and small domain in the Hst2-H3K9bz complex are shown in green and cyan, respectively. Hst2 in the Hst2-2′-O-benzoyl-ADP-ribose complex is shown in gray. The cofactor product 2′-O-benzoyl-ADP-ribose is shown in slate. c The interface between Hst2 and H3 peptide. Hst2 is shown as a surface model colored according to its electrostatic potential (positive potential: blue; negative potential, red). H3 residues from T6 to K14 are shown in yellow stick models. An acidic ring formed by a series of acidic residues is labeled. d The H3K9bz binding pocket of Hst2. The side chain of K9bz (shown as a ball model) is inserted into a relatively hydrophobic tunnel of Hst2. e ITC measurements reveal that Hst2 mutations and H3K9bzR8A weaken the Hst2–H3K9bz interaction. The dissociation constant (Kd) and their fitting errors are shown. f MALDI-TOF-based debenzoylase assays show that Hst2 mutations and H3K9bzR8A decrease the debenzoylase activity of Hst2. Data are presented as mean ± SD, n = 3 biological independent measurements. Source data are provided in the Source Data file. g Comparison of the almost identical Kbz and Kac-binding pockets in Hst2-H3K9bz and Hst2-H4K16ac (PDB: 1Q1A) structures. H3K9bz is shown in yellow, and H4K16ac is shown in red. Hydrogen bonds are shown as magenta dashed lines. h Comparison of the acyl-group binding pockets by superimposition of Hst2-H3K9bz, Hst2-H4K16ac, and SIRT3-H3K4cr structures. The benzoyl, acetyl, and crotonyl groups are surrounded by four conserved hydrophobic residues. i ITC results reveal that Hst2 I117F and I117V mutants decrease and increase binding with H3K9bz peptides, respectively. The dissociation constant (Kd) and their fitting errors are shown. j MALDI-TOF-based debenzoylase assays show that Hst2 I117F and I117V mutants decrease and increase debenzoylase activities on H3K9bz peptides, respectively. Data are presented as mean ± SD, n = 3 biological independent measurements. Source data are provided in the Source Data file.
|
|
Fig. 5. YEATS domains from Taf14 and Sas5 are Kbz readers.a ITC analyses show that Taf14YEATS has different binding affinities with Kbz peptides. b ITC analyses show that Sas5YEATS could bind Kbz peptides with different binding affinities. c ITC analyses show that Yaf9YEATS does not bind any Kbz peptides tested or that the binding affinity is beyond the detection limit. d Superimposition of Taf14YEATS-H3K9bz and Sas5YEATS-H3K27bz structures. The Taf14YEATS and Sas5YEATS are shown in yellow and green; H3K9bz and H3K27bz peptides are shown in blue and magenta. e The benzoyl groups in H3K9bz and H3K27bz peptides are inserted into a conserved binding pocket sharing nearly identical interacting networks. The residues involved in hydrophobic contacts and hydrogen bonds (orange dashed lines) are localized in two arms, embracing the benzoyl group. The red dot is the water molecule. f ITC results show the effects of Sas5YEATS mutations on the interaction between Sas5YEATS and H3K27bz peptide. g ITC results show the effects of Taf14YEATS mutations on the interaction between Taf14YEATS and H3K9bz peptide. h Superimposition of four YEATS-Kbz complex structures, including AF9-H3K9bz (gray-red), YEATS2-H3K27bz (orange-navy blue), Taf14-H3K9bz (yellow-cyan), and Sas5-H3K27bz (green-magenta). The residues at the tip regions of the benzoyl-binding pocket are different and indicated by arrows. i The acyl-binding pocket of Taf14YEATS. Four Taf14YEATS structures are superimposed, including Taf14-H3K9bz, Taf14-H3K9ac (PDB: 5D7E), Taf14-H3K9cr (PDB: 5IOK), and Taf14-H3K9bu (PDB: 6MIQ). The benzoyl, acetyl, crotonyl, and butyryl groups are shown in cyan, magenta, orange, and green, respectively. The tip residues of V29 and A63 are involved in the fine-tuning of pocket shapes to accommodate different acylation marks. j ITC results reveal that the mutations of tip residues on Taf14YEAST and Sas5YEATS have negligible effects on binding with acetylated peptides.
|
|
Fig. 6. Sth1 bromodomain has Kbz binding activity.a ITC analyses of four yeast bromodomains binding with H3K14bz peptides. Sth1BD exhibits a detectable interaction with the H3K14bz peptide (Kd = 89 μM). b Superimposition of Sth1BD-H3K14bz (Sth1:orange; H3, cyan) and Sth1BD-H3K14ac (Sth1: yellow; H3, purple) complexes shows almost identical conformations of Sth1BD and H3 peptide, except for the C-terminal 310-helix of H3. c Detailed interaction networks between acylated lysine and Sth1BD. Benzoyl and acetyl groups are shown in cyan and purple, respectively. Hydrophobic contacts and hydrogen-bonding networks are shown. d ITC assays showing the effects of Sth1 mutations on the interaction between Sth1BD and H3K14bz peptide.
|
|
Fig. 7. Proteome-wide screening of Kbz-modified proteins in yeast cells.a The workflow used for HPLC–MS/MS-based non-histone Kbz sites identification in S. cerevisiae. b The MS/MS spectrum of an Eno1 peptide (LAK+104.0268LNQLLR) harboring one benzoylated site. Predicted y- and b-type ions are listed above and below the peptide sequence, respectively. Matched ions are labeled in the spectra. c Distribution of non-histone Kbz sites based on the site number per protein. d Venn diagram showing the subcellular compartment distribution of non-histone Kbz proteins. e Gene ontology analysis associated with significant regulated genes (P < 0.05) of non-histone Kbz proteins. P values are derived from one-sided Fisher’s exact test. f Overview of glycolysis pathway. The enzymes with identified benzoylation sites are shown in red. g Eight benzoylation sites on Eno1 are shown in the structural model of Eno1 (PDB: 1ONE). The Eno1 substrate, 2-phosphoglycerate, is shown in cyan. K346, K397, and K409 are adjacent to the substrate-binding pocket.
|
