Archaeal chromatin 'slinkies' are inherently dynamic complexes with deflected DNA wrapping pathways

Abstract

Eukaryotes and many archaea package their DNA with histones. While the four eukaryotic histones wrap ~147 DNA base pairs into nucleosomes, archaeal histones form 'nucleosome-like' complexes that continuously wind between 60 and 500 base pairs of DNA ('archaeasomes'), suggested by crystal contacts and analysis of cellular chromatin. Solution structures of large archaeasomes (>90 DNA base pairs) have never been directly observed. Here, we utilize molecular dynamics simulations, analytical ultracentrifugation, and cryoEM to structurally characterize the solution state of archaeasomes on longer DNA. Simulations reveal dynamics of increased accessibility without disruption of DNA-binding or tetramerization interfaces. Mg²⁺ concentration influences compaction, and cryoEM densities illustrate that DNA is wrapped in consecutive substates arranged 90^o out-of-plane with one another. Without ATP-dependent remodelers, archaea may leverage these inherent dynamics to balance chromatin packing and accessibility.

Keywords: AUC; archaea; chromosomes; cryo-EM; gene expression; histone; molecular biophysics; molecular dynamics; none; nucleosome; structural biology.

Figure 1.. Comparison of the eukaryotic nucleosome and Arc120, an archaeasome model system containing 120 base pairs of DNA and four histone dimers.

(A) The eukaryotic nucleosome (PDB 1AOI) containing two copies of H3, H4, H2A, and H2B (blue, green, gold, and red, respectively) arranged as H2A-H2B dimers flanking the (H3–H4)₂ tetramer and binding 147 base pairs of DNA (gray and light blue). (B) The model Arc120 system, derived from the Arc90 crystal structure (PDB 5T5K), with four HMfB homodimers participating in L1-L1 stacking interactions shown in pink. (C) Enhanced views of both the wild type and G17D mutant interfaces simulated in this study.

Figure 2.. Backbone RMSD for simulated systems compared to their initial conformations.

(A) Representative timeseries from each system demonstrating the largest change in structure from the initial states. The Arc90 trajectory (black) exhibits both a large peak value, as well as wide variances across the timeseries. Arc120 (gold) and Arc180 (teal) systems are significantly less dynamic, and the Arc120-G17D mutant (orange) shows an increased divergence from the initial state when compared to the wild-type Arc120 system. (B) Distribution of RMSD values sampled across all three independent trajectories of each system, post-equilibration (100 ns). Arc90 displays a sampling of two different states, while each of the larger systems are unimodal. (C) Distribution curves for center-of-mass separations between DNA ends and the neighboring superhelical gyre. Representative conformations of fully closed (~10 Å) and open (~30 Å) states identified in the Arc120-G17D simulations are shown. Arc90 simulations (black) show a bimodal distribution between these values, but Arc120 (gold) and Arc180 (teal) systems, containing one and three L1-L1 interactions each, sample unimodal values around the closed state. The Arc120-G17D mutant (orange) samples a unimodal distribution, but indicative of a more open archaeasome than the wild type.

Figure 2—figure supplement 1.. Root mean-squared fluctuation (RMSF) plots for measured for DNA bases in all four simulated systems.

Average values across the three independent systems are traced by solid lines, and error ranges are calculaged by the standard error of the mean across the three simulations and outlined with shaded regions. RMSF values of the 5’ (top) and 3’ (bottom) leading strands are shown separately. Periodicity in each plot coincides with DNA turns transitioning between solvent- and protein-facing residues.

Figure 2—figure supplement 2.. Representative timeseries for DNA end-to-neighbor separation distances from each system.

The Arc90 system (black) shows both the largest maximum separation, as well as the widest variance in values. The next largest separations and fluctuations are observed in the Arc120-G17D mutant (orange), with the Arc120 wild-type (gold) and Arc180 (teal) systems exhibiting the tightest DNA wrapping.

Figure 3.. Distribution of solvent-accessible surface areas observed per frame in the Arc120 and Arc120-G17D simulations.

The G17D mutation at the key L1-L1 interaction between stacking dimers significantly increases the surface area by ~1000 Ų.

Figure 4.. Biophysical characterization of archaeasomes in vitro.

(A) van Holde-Weischet plots of Arc207 (circles) and Nuc147 (squares) samples. While Arc207 has a larger mass than the eukaryotic Nuc147, it sediments more slowly, indicative of increased drag caused by extended particle configurations. (B) Representative de-noised cryoEM micrograph of Arc207 sample. Particles are observed with DNA wrapped in a nucleosome-like pattern, but a significant separation between neighboring DNA turns is observed. (C) Two-dimensional classifications of particles extracted from these micrographs. No tightly wrapped DNA is identified, in agreement with our SV-AUC inferences of an extended Arc207 particle shape in comparison to compact Nuc147.

Figure 4—figure supplement 1.. Titration of HTkA histone on to 207 bp DNA strand, observed by native PAGE gel.

M: Marker; 1: Widom 207 DNA fragment (100 ng); Lanes 2–7: 3, 5, 7, 9, 11, and 13 molar equivalents of HTkA histone dimers. Full saturation is observed at the stoichiometric ratio of 7 histone dimers per DNA fragment (lane 4).

Figure 5.. Effects of Mg²⁺ concentration on sedimentation behavior of Arc207 and Nuc147 samples.

(A) van Holde-Weischet plots of (top) Arc207 and (bottom) Nuc147 samples. Increases from 0 to 5 mM MgCl₂ results in an increase in Arc207 sedimentation rate, but little change in Nuc147 sedimentation. (B) Effects of MgCl₂ concentration on (top) frictional ratio and (bottom) absorbance at 260 nm for Arc207 (circles) and Nuc147 (squares). Frictional ratios of Arc207 samples follow the same profile as sedimentation rate, where samples compact from 0 to 5 mM MgCl₂ but with little change from 5 to 10 mM MgCl₂. Similarly, Nuc147 samples showed very modest changes in compaction. OD_260nm measurements show Nuc147 particles aggregating as a result of increased MgCl₂ concentration, but Arc207 samples compact without losses to aggregation.

Figure 6.. Single particle cryoEM analysis of the Arc207 complex in the presence of 5 mM MgCl₂.

(A) Representative denoised micrograph. Particles are observed with both ‘nucleosome-like’ wrappings of the 207 base pair DNA fragment (blue boxes, Class I) as well as particles showing out-of-plane ‘lid’ extensions of DNA from similarly wrapped ‘cores’ (gold boxes, Class II). Also shown are two-dimensional classes (B, Class I) and (C, Class II). Three-dimensional densities from these classes are shown in (D) and (E), respectively. Class I particles were fit by rigid-body docking of Arc150 coordinates extracted from Arc180 simulations, and Class II particles were fit by rigid-body docking separate Arc90 and Arc120 components and bridging the connecting DNA through energy minimization.

Figure 6—figure supplement 1.. DNA parameters of the zero-length linker segment calculated from a 10 ns implicit solvent simulation of the Arc207 system restrained to the ‘Class II’ configuration.

The linker segment is highlighted in blue in the accompanying structure, and blue shaded regions in the plot outline the standard deviation of values across the trajectory, whereas the average values are shown by a dark blue line. Horizontal dotted lines define the 2σ range of values observed in the eukaryotic nucleosome (PDB 1aoi).

Figure 6—figure supplement 2.. Low fidelity density of Arc150 volume subclassification, showing a potential Arc60 ‘lid’ oriented out-of-plane with Arc150 ‘base’ reconstructed from the ‘Class I’ particles.

(top) The density shown with no docked model. (bottom) The same density, but with the addition of Arc150 and Arc60 model structures docked within the density. The Arc60 conformation was selected to mimic the complete Arc207 density solved previously (Figure 6E), and this conformation is centered well in the Arc60 extension despite having no a priori information regarding the observed conformation in this classification.