The H2A-H2B dimeric kinetic intermediate is stabilized by widespread hydrophobic burial with few fully native interactions

Abstract

The H2A-H2B histone heterodimer folds via monomeric and dimeric kinetic intermediates. Within ∼5 ms, the H2A and H2B polypeptides associate in a nearly diffusion limited reaction to form a dimeric ensemble, denoted I₂ and I₂*, the latter being a subpopulation characterized by a higher content of nonnative structure (NNS). The I₂ ensemble folds to the native heterodimer, N₂, through an observable, first-order kinetic phase. To determine the regions of structure in the I₂ ensemble, we characterized 26 Ala mutants of buried hydrophobic residues, spanning the three helices of the canonical histone folds of H2A and H2B and the H2B C-terminal helix. All but one targeted residue contributed significantly to the stability of I₂, the transition state and N₂; however, only residues in the hydrophobic core of the dimer interface perturbed the I₂* population. Destabilization of I₂* correlated with slower folding rates, implying that NNS is not a kinetic trap but rather accelerates folding. The pattern of Φ values indicated that residues forming intramolecular interactions in the peripheral helices contributed similar stability to I₂ and N₂, but residues involved in intermolecular interactions in the hydrophobic core are only partially folded in I₂. These findings suggest a dimerize-then-rearrange model. Residues throughout the histone fold contribute to the stability of I₂, but after the rapid dimerization reaction, the hydrophobic core of the dimer interface has few fully native interactions. In the transition state leading to N₂, more native-like interactions are developed and nonnative interactions are rearranged.

Figure 1

Comparison of urea and GdmCl induced equilibrium unfolding transitions. WT H2A-H2B (black) and H2A-L51A mutant heterodimer (grey) titrations in urea (circles) or GdmCl (squares). Solid lines represent global fits to a two-state dimeric unfolding model with native baselines shown as dashed lines. Inset: Representative fluorescence titration illustrating curvature of the native baseline below 0.5 M GdmCl. Conditions were 25 °C, 0.2 M KCl, 20 mM KPi (pH 7.2), 0.1 mM K₂EDTA.

Figure 2

Denaturant dependence of the folding and unfolding rates. Folding, circles; unfolding, squares. Data points represent semi-global fits of multiple SF-CD and SF-FL kinetic traces at a given GdmCl concentration. Lines represent the global fits to Eq. 8 for kinetic data at denaturant concentrations that exhibit a log-linear dependence (i.e. excluding the roll-over regions). (a) Comparison of WT H2A-H2B kinetics in GdmCl (black) compared to kinetics in urea (dashed gray lines).^; (b) Representative mutants H2A-L85A (red) and H2B-M59A (blue) kinetics. WT kinetic data are shown as grey dashed lines. Conditions were 10 µM monomer, 25 °C, 0.2 M KCl, 20 mM KPi (pH 7.2), 0.1 mM K₂EDTA.

Figure 2

Denaturant dependence of the folding and unfolding rates. Folding, circles; unfolding, squares. Data points represent semi-global fits of multiple SF-CD and SF-FL kinetic traces at a given GdmCl concentration. Lines represent the global fits to Eq. 8 for kinetic data at denaturant concentrations that exhibit a log-linear dependence (i.e. excluding the roll-over regions). (a) Comparison of WT H2A-H2B kinetics in GdmCl (black) compared to kinetics in urea (dashed gray lines).^; (b) Representative mutants H2A-L85A (red) and H2B-M59A (blue) kinetics. WT kinetic data are shown as grey dashed lines. Conditions were 10 µM monomer, 25 °C, 0.2 M KCl, 20 mM KPi (pH 7.2), 0.1 mM K₂EDTA.

Figure 3

Ribbon diagram of the H2A-H2B dimer highlighting the effects of Ala mutations on the ratio of k_f(H₂O)/k_{0 M}. The H2A and H2B monomers are shown in pale red and blue, respectively. The Cα atoms of residues whose mutation to Ala result in WT-like or greater ratios of k_f(H₂O)/k_{0 M} are shown as magenta spheres. The side chains of residues whose mutation to Ala result in 2-fold or greater reduction in the k_f(H₂O)/k_{0 M} ratio are shown in red and blue for H2A and H2B, respectively. The figure was generated with PyMol (Delano Scientific, LLC, San Carlos, CA) using the coordinates of H2A-H2B in the NCP X-ray crystal structure (1AOI.pdb).

Figure 4

Comparison of the ΔΔG values for the H2A and H2B mutants. ΔΔG_I2, white bars; ΔΔG_TS, grey bars (ΔΔG_TS = ΔΔG_Eq − ΔΔG^‡_unf); ΔΔG_Eq, black bars. The error bars depict the propagated error from the fitted data. (a) H2A ΔΔG values. (b) H2B ΔΔG values.

Figure 4

Comparison of the ΔΔG values for the H2A and H2B mutants. ΔΔG_I2, white bars; ΔΔG_TS, grey bars (ΔΔG_TS = ΔΔG_Eq − ΔΔG^‡_unf); ΔΔG_Eq, black bars. The error bars depict the propagated error from the fitted data. (a) H2A ΔΔG values. (b) H2B ΔΔG values.

Figure 5

Structure of the H2A-H2B dimer displaying the Φ_I2 values. The backbone of the H2A and H2B monomers are shown in pale red and blue, respectively. The poorly structured N- and C-terminal tails have been truncated for clarity. The Cα atoms of mutated residues are shown as spheres. The H2A and H2B residues with Φ_I2 values ≥ 0.75 are highlighted in red and blue, respectively. The H2A and H2B residues with Φ_I2 values between 0.4 and 0.7 are shown in magenta and cyan, respectively. The figure was generated as described in the legend of Figure 3.

Scheme 1

Working mechanism for the kinetic folding of H2A-H2B.