Nucleosomes in gene regulation: theoretical approaches

This work reviews current theoretical approaches of biophysics and bioinformatics for the description of nucleosome arrangements in chromatin and transcription factor binding to nucleosomal organized DNA. The role of nucleosomes in gene regulation is discussed from molecular-mechanistic and biological point of view. In addition to classical problems of this field, actual questions of epigenetic regulation are discussed. The authors selected for discussion what seem to be the most interesting concepts and hypotheses. Mathematical approaches are described in a simplified language to attract attention to the most important directions of this field.

💡 Research Summary

The paper provides a comprehensive review of the theoretical frameworks that have been developed to describe how nucleosomes—fundamental units of chromatin—affect gene regulation. It begins by summarizing the basic biophysical properties of nucleosomes: 147 bp of DNA wrapped around an octamer of histone proteins, the intrinsic DNA bending energy, and the role of histone post‑translational modifications (PTMs) in modulating nucleosome stability. From this foundation, the authors organize the existing literature into three principal modeling paradigms.

The first paradigm is the “nucleosome positioning energy model,” which treats the DNA‑histone interface as a sequence‑dependent energy landscape. By quantifying contributions from DNA curvature, base‑pair stacking, AT‑rich tracts, and the presence of CpG methylation, the model predicts the most favorable positions for nucleosome deposition along a given genome segment. The authors illustrate how this approach can be calibrated against high‑resolution nucleosome maps derived from MNase‑seq or chemical mapping, and they discuss the limitations of treating nucleosome placement as a purely equilibrium problem.

The second paradigm extends the analysis to the whole chromatin fiber using statistical‑mechanical lattice models. Here, each lattice site can be occupied by a nucleosome, a transcription factor (TF), or remain empty. The model incorporates steric exclusion (preventing nucleosome overlap), cooperative interactions between neighboring nucleosomes, and the energetic penalty or reward associated with TF binding to nucleosome‑occupied versus nucleosome‑free DNA. Monte‑Carlo sampling is employed to generate ensembles of nucleosome configurations under different biological conditions, such as the presence of a repressive complex or an activating enhancer. This framework reproduces experimentally observed phenomena, including nucleosome phasing downstream of transcription start sites and the formation of nucleosome‑depleted regions (NDRs) at active promoters.

The third paradigm introduces dynamics through kinetic Monte‑Carlo or Gillespie simulations that explicitly model ATP‑dependent chromatin remodelers (e.g., SWI/SNF, ISWI, CHD families). The remodelers are represented as stochastic agents that can slide, evict, or restructure nucleosomes with rates that depend on local histone PTMs, DNA sequence, and the occupancy of TFs. By coupling remodeler activity to TF binding, the authors demonstrate a feedback loop: TFs recruit remodelers, remodelers reposition nucleosomes to expose additional TF binding sites, and the newly exposed sites further stabilize TF occupancy. This dynamic view captures the transient nature of “open” chromatin states observed in single‑cell ATAC‑seq and live‑cell imaging experiments.

A substantial portion of the review is devoted to epigenetic regulation. The authors discuss how specific PTMs alter the parameters of each model. For instance, H3K27 acetylation reduces the DNA bending penalty, effectively lowering the nucleosome formation energy and facilitating nucleosome eviction at enhancers. Conversely, H3K9 trimethylation increases the histone‑DNA affinity term, stabilizing nucleosomes over heterochromatic regions and impeding TF access. DNA methylation is treated as a modifier of both the intrinsic DNA stiffness and the affinity for certain histone readers, thereby influencing nucleosome positioning indirectly.

The paper also critically evaluates current shortcomings. Most models are one‑dimensional, neglecting three‑dimensional chromatin looping and higher‑order folding that can bring distal regulatory elements into proximity. Remodeler complexes are often simplified to a single kinetic rate, ignoring the multi‑subunit coordination that determines substrate specificity. To address these gaps, the authors propose integrating machine‑learning approaches—such as deep neural networks trained on large‑scale nucleosome occupancy datasets—to infer more complex energy functions. They also suggest the incorporation of Cryo‑EM derived nucleosome structures and single‑molecule force spectroscopy data to refine the physical parameters governing DNA‑histone interactions.

In conclusion, the review argues that a synergistic combination of equilibrium energy models, statistical‑mechanical lattice frameworks, and kinetic simulations provides a powerful toolkit for dissecting the causal relationships between nucleosome arrangement and transcriptional output. By linking theoretical predictions with experimental datasets (MNase‑seq, ATAC‑seq, ChIP‑seq for histone marks, and live‑cell imaging), researchers can generate testable hypotheses about how epigenetic modifications, remodeler activity, and DNA sequence collectively shape the regulatory landscape. The authors anticipate that continued refinement of these models will be essential for interpreting disease‑associated chromatin alterations and for guiding the design of epigenome‑editing strategies.