Fig 1.
Structures of eukaryotic and archaeal histones and a summary table of studied systems.
(A) The scaffold of eukaryotic histone dimer H2A/H2B (blue/orange, PDB: 1AOI) consists of three α helices in each protein, plus two terminal helices and long histone tails. (B) The structure of homotypic archaeal histone dimer (HmfB)2 (green, PDB: 1A7W) has three α helices each with no histone tails. (C) Polymer scaling fitting of the Rg and sequence length suggests that histone dimers act more as “monomeric” proteins. 7 histone(-like) dimer structures from PDB and their 11 monomers are included in this plot. The dashed line is the empirical relation of Rg and N from a survey study of 403 globular monomeric proteins [25]. (D) This table summarizes the 13 types of heterotypic and homotypic histone(-like) dimers that are studied in this work by different methodologies. A blank cell in Tails/PDB column means “non-existent” while in other columns it means “not conducted”.
Fig 2.
Histone heterodimers fold in MD simulations and NMR experiments, not monomers.
(A) Q values as a function of the annealing temperature are plotted for simulated histone monomers, with the mean and standard deviation displayed as circles and error bars. Aligned final snapshots of H2A by α2 helix (orange) show that no stable tetiary structure formed. (B-C) Q analysis shows that histone monomers (blue, green) cooperatively fold and bind into a dimer (magenta). Two example systems are shown, human H2A/H2B (B) and archaeal histones (HMfA)2 (C). The final conformations (magenta) are well aligned with their native states (gray). (D-E) NMR and CD studies of H2A and H2B upon their complex formation. (D) 1H-15N NMR spectra of 15N-labeled H2A alone (i) and in the presence of unlabeled H2B at a 1:1 molar ratio (iii). Heteronuclear steady-state 15N{1H} NOE spectra with amide proton presaturation recorded for 15N-labeled H2A alone (ii) and in the presence of unlabeled H2B at an equimolar ratio (iv). In these spectra, contours with positive intensities are colored black while negative intensities are blue. (E) CD spectra of H2A (blue) and H2B (red) alone and in an equimolar mixture (black). The total concentrations of the proteins are the same in all three cases. Also shown (dashed magenta) is the expected CD spectrum of the equimolar mixture if no structural changes in either protein. The table shows the secondary structure composition of the proteins estimated from these experimental CD data using K2D algorithm.
Fig 3.
Inverted non-native conformation is found to be a stable formation of histone dimer.
(A) The native and non-native conformation of H2A/H2B found in AWSEM simulations are shown. Their major difference is measured by the angle between the α2-α2 helices (green) with each vector arrow pointing from N- to C-terminal. (B-C) Qdimer analysis shows that both native and non-native conformations are observed in the simulated annealing runs of eukaryotic H2AH2B (B) and the archaeal dimer (HMfA)2 (C). (D) Free energy profiles of H2A/H2B are projected on Qmonomer and the angle of α2-α2 helices (left), and on Qinterface and the α2-α2 helices angle (right). Two energy minima are found, corresponding to the native-like and inverted non-native conformation respectively. (E) All-atom MD simulations show comparable stability of inverted non-native conformation to that of native structure of histone dimer.
Fig 4.
Sequence symmetry of hydrophobic residues explains the predictability of inverted histone-fold structure.
(A) Hydrophobic interactions (orange) dominate the formation of binding interface of H2A/H2B (colored in green/purple). (B) Protein sequence alignments of histones H2A, H2B and their reversed sequences highlight the symmetrical distribution of hydrophobic residues. Hydrophobic residues are particularly marked. (C) Cartoon schemes illustrate previously proposed histone evolution hypotheses: i) histones may originate from one single helix-strand-helix structural motif HSH; ii) two HSH patterns form one monomer through duplication, differentiation and fusion; iii) domain swapping between two monomers (colored in green and blue) forms a histone-fold structure. (D) The inverted arrangement based on hydrophobic interactions could be an alternative formation of histone-fold structure.
Fig 5.
Predicted structures of eukaryotic histone homodimer.
(A) AWSEM and AlphaFold2 predicted homodimer structures of H2A/H2A align well with the native heterodimer structure of H2A/H2B (colored in blue, orange, and green). (B) RMSF analysis of all-atom simulations demonstrates comparable stabilities of AWSEM- and AlphaFold2-predicted homo-complex structures (in blue and orange). RMSF of two chains are plotted separately and helix regions are animated by cartoon in grey.
Fig 6.
Histone tails disfavor the formation of histone homodimer, yet stabilize the histone fold once formed.
(A) AWSEM-predicted trancated and full-sequence homodimer H2A/H2A are assessed by the maximum Qdimer of the histone-fold region (colored in blue and orange) and compared with predictions of trancated and full-sequence heterodimer H2A/H2B as a control (green and red). (B) AWSEM-predicted structures of full-sequence H2A/H2B (green/cyan) show native-like and inverted non-native orientations with displaced tail regions. (C) AWSEM-prediction of full-sequence H2A/H2A shows that the histone tails inhibit the formation of the hand-shake motif. α1 helices are colored in grey to help illustrate their native or non-native arrangements. (E) All-atom simulations of AlphaFold2-predicted truncated and full-sequence H2A/H2A homodimer (colored in blue and orange) are analyzed through RMSD and compared with that of truncated and full-sequence H2A/H2B (green and red). (F) The histone-fold region RMSF of the truncated and full-sequence H2A/H2A homodimer (colored in orange and blue). The two chains of H2A/H2A are plotted separately and their helix regions are animated by a cartoon diagram.