Electrostatic origin of salt-induced nucleosome array compaction

Abstract

The physical mechanism of the folding and unfolding of chromatin is fundamentally related to transcription but is incompletely characterized and not fully understood. We experimentally and theoretically studied chromatin compaction by investigating the salt-mediated folding of an array made of 12 positioning nucleosomes with 177 bp repeat length. Sedimentation velocity measurements were performed to monitor the folding provoked by addition of cations Na(+), K(+), Mg(2+), Ca(2+), spermidine(3+), Co(NH(3))(6)(3+), and spermine(4+). We found typical polyelectrolyte behavior, with the critical concentration of cation needed to bring about maximal folding covering a range of almost five orders of magnitude (from 2 μM for spermine(4+) to 100 mM for Na(+)). A coarse-grained model of the nucleosome array based on a continuum dielectric description and including the explicit presence of mobile ions and charged flexible histone tails was used in computer simulations to investigate the cation-mediated compaction. The results of the simulations with explicit ions are in general agreement with the experimental data, whereas simple Debye-Hückel models are intrinsically incapable of describing chromatin array folding by multivalent cations. We conclude that the theoretical description of the salt-induced chromatin folding must incorporate explicit mobile ions that include ion correlation and ion competition effects.

Figure 1

(A) van Holde-Weischet sedimentation velocity curves obtained for the 12-177-601 array with K⁺, Na⁺, Mg²⁺, Ca²⁺, Spd³⁺, CoHex³⁺, and Spm⁴⁺ (the concentration of the cation is indicated in the graphs; the curve obtained in TEK buffer is shown for comparison). There is a background presence of ions from the TEK buffer in all solutions studied (10 mM K⁺ and ∼8.0 mM Tris⁺ cations). (B) Validation by native 5% PAGE of the histone/DNA stoichiometry in the reconstituted 12-177-601 array with various amounts of histone octamer added to the DNA (octamer/DNA ratios of 12, 12.5, 13, and 13.5 were tested as indicated at the bottom of the figure). Lanes 2, 4, 6, 8, 10 and 3, 5, 7, 9, 11 are for the arrays correspondingly before and after ScaI digestion. Lane 1 is 12-177-601 DNA before reconstitution; lane 12 is the DNA marker (from 100 bp to 1500 bp, step 100 bp). A histone/DNA ratio of 13.0:1.0 was chosen as optimal. (C) Dependence of s_20,w-values on the concentration of added cation. Horizontal line indicates the s_20,w-value in the reference TEK buffer. (D). Maximal s_20,w-values obtained for the cations studied (numbers are shown at the top of the each bar). Cation concentration corresponding the maximal value of s_20,w is listed at the bottom of the graph.

Figure 2

Comparison of the results of MD simulations of the 12-177 array in the presence of K⁺, Mg²⁺, and CoHex³⁺. (A) Representative snapshots illustrate the typical degree of folding of the array. Variation of (B) sedimentation coefficient (s_20,w) and (C) radius of gyration (R_g) during the course of the MD simulations. (D) Core-core and (E) external tail-core RDFs. (In the core-core RDF (D) neighboring NCPs were excluded from statistics.) The three peaks clearly visible in the core-core RDF of the CoHex³⁺-array system are related to three possible types of NCP-NCP contacts and illustrated by sketches.

Figure 3

Summary of MD simulation results for the 12-177 array titration by Mg²⁺ (A–D), CoHex³⁺ (E–H), and Spd³⁺ and Spm⁴⁺ (I and K). (A and E) Representative snapshots illustrating folding of the array at different stages of titration by (A) Mg²⁺ and (E) CoHex³⁺; numbers under the snapshots indicate the concentration of the bulk (first line) and average (second line) Mg²⁺ or CoHex³⁺ concentration. (B, F, and I) Radius of gyration, R_g; (C, G, and J) sedimentation coefficient, s_20,w; (D, H, and K) intensity of the maximum in the external tail-core RDF. In graphs B–D and F–H, data for the array with charged tails are drawn as solid circles and lines; similar results for the tailless array are shown as hollow points with dashed lines. In graphs B–D and F–K, the x axis is the bulk concentration of Mg²⁺ (B–D, C_Mg), CoHex³⁺ (F–H, C_Co), and Spd³⁺ or Spm⁴⁺ (I and K, C_Spd/C_Spm). Numerical values (s_20,w, R_g) and their dependencies on the average concentration of the oligocation are given in Table S5, Fig. S8, and Fig. S9.

Figure 4

Simulations within the DH approximation as a function of ionic strength for different models of the array. (A) Sedimentation coefficient, s_20,w; (B) radius of gyration, R_g. Horizontal dashed lines indicate the values of s_20,w and R_g calculated for uncharged array model. Hollow ovals in A and B respectively show s_20,w and R_g values obtained in the MD simulations with explicit ions; cations are indicated near the points. Solid ovals in A show the maximal s_20,w-values observed in the AUC measurements for different cations (indicated in the graph). Indices Mg²⁺ and high Mg²⁺ mark the MD simulation data calculated for C_Mg = 3.89 and 15.83 mM, respectively (see text for details).