Nucleosome assembly and disassembly pathways in vitro

Abstract

Structural fluctuations of nucleosomes modulate the access to internal DNA in eukaryotic cells; clearly characterisation of this fundamental process is crucial to understanding gene regulation. Here we apply PhAST (Photochemical Analysis of Structural Transitions) to monitor at a base pair level, structural alterations induced all along the DNA upon histone binding or release. By offering the first reliable, detailed comparison of nucleosome assembly and disassembly in vitro, we reveal similarities and differences between the two processes. We identify multiple, sequential intermediate states characterised by specific PhAST signals whose localisation and amplitude reflect asymmetries of DNA/histone interactions with respect to the nucleosome pseudo dyad. These asymmetries involve not only the DNA extremities but also regions close to the pseudo dyad. Localisations of asymmetries develop in a consistent manner during both assembly and disassembly processes; they primarily reflect the DNA sequence effect on the efficiency of DNA-histone binding. More unexpectedly, the amplitude component of PhAST signals not only evolves as a function of intermediate states but does so differently between assembly and disassembly pathways. Our observation of differences between assembly and disassembly opens up new avenues to define the role of the DNA sequence in processes underlying the regulation of gene expression. Overall, we provide new insights into how the intrinsic properties of DNA are integrated into a holistic mechanism that controls chromatin structure.

Fig 1. Outline of the PhAST experiments monitoring successive DNA structural changes induced during nucleosome assembly and disassembly by using stepwise decrease and increase of NaCl concentration.

Fig 2. Comparison of PhAST signals in free DNA at various ionic strengths.

The Phast signal is expressed in terms of intensity (I) that is the normalised peak height associated with each YpY step of the free 601 sequence. The data collected at 0.5, 1.0, 1.5 and 2 M NaCl were systematically compared to those at 0.18 M NaCl. Each point represents the value of I, averaged from 6 experiments related to assembly and disassembly studies; the vertical and horizontal bars are standard errors. The correlation coefficients (CC) are given in boxes in each panel.

Fig 3. DNA regions involved in the DNA/histone interface.

This schematic representation shows those DNA regions that interact with the different histone dimers in the fully formed NCP. To distinguish the two copies of the H2A/H2B dimer, they are labelled 5’ (light blue, chains G, H) and 3’ (dark blue, chains C, D). Similarly, the two pairs of H3/H4 histones are labelled 5’ (pink, chains E, F) and 3’ (red, chains A, B). The DNA sequence is labelled as SHL (Super Helical Location, defined in Material and Methods). The denominations A, B, … of the histone chains are those commonly used in X-ray structures.

Fig 4. Changes in the probability of YpY dimer formation in DNA during nucleosome assembly and disassembly experiments.

Changes in the probability of YpY dimer formation are presented in terms of absolute values of log2 of the intensity ratios (|log2(IR)|) along the 601 sequence expressed in SHLs; they are given for decreasing (top panel) or increasing (bottom panel) ionic strengths, as indicated by the green and black arrow respectively. The IR quantities are the ratios calculated between the normalised peak heights associated with each YpY step in the histone plus DNA mixtures and those of DNA alone. Red and blue bars correspond to DNA residues involved in the interface with H3/H4 and H2A/H2B dimers, respectively; black bars correspond to dinucleotides contacted by the two dimers. Minor-groove inward facing regions observed in the nucleosome structures are represented by grey boxes. Error bars are first-order estimations of standard errors calculated on at least 3 independent experiments (see Material and Methods for details).

Fig 5. Effect of histones on DNA structure as calculated by summing |log₂(IR)| values over the whole 601 sequence (excluding up to SHL ± 6).

The data associated with different stages of assembly (green triangles) and disassembly (black inverted-triangles) are compared at each ionic strength; values are divided by the total number of pyrimidine dinucleotides. Error bars are first-order estimations of standard errors; differences at 1.0 M and 1.5 M are statistically significant (p < 0.05, Student’s t-test). Sigmoidal fits to the data (black and green curves) are used to estimate the approximate concentration of NaCl at which 50% of the range of total change is attained (black horizontal line) for assembly (dotted green vertical line) and disassembly (dotted black vertical line).

Fig 6. Effect of histones on DNA regions interacting with histone dimers in the fully wrapped NCP.

A: Effect of histones as calculated by summing |log2(IR)| over the DNA regions interacting with the histone dimers in the complete nucleosome. Error bars are 1.5 times standard error, such that no overlap between the error bars indicates significant differences (with circa p < 0.05). B: schematic representation of the DNA regions that are associated with the histone dimers in the complete nucleosome. C: Statistical analysis of the traces in A interpreted as the behaviour of each region at each salt concentration, classified as either fully unwrapped, fully wrapped or intermediate (see Materials and Methods). The ordinate shows salt concentration and the abscissa is detailed in B. The top half of the figure (green) refers to assembly and the lower part (black) to disassembly.

Fig 7. A structural model of putative intermediates during nucleosome turnover in vitro.

The model shows a cartoon of proposed structures present at different stages of assembly (green arrow) and disassembly (black arrow). The 601 sequence is represented as a four-colour line that corresponds to DNA regions that are associated with the histone dimers in the complete nucleosome with regard to the pseudo dyad axis (see Fig 3 for details). The upper schematic represents the organisation of the SHL’s on the complete nucleosome as a left-handed helix going into the page from the 5’ end.