Background: Reading structural diversity in individual chromatin fibers

DNA wrapped around nucleosomes is not uniformly occluded. A study published in Nature used IDLI, a computational analysis method, to classify nucleosome footprints in mouse embryonic stem cells and inferred partial internal DNA accessibility in 86.6% of the analyzed set. Experiments that removed or altered transcription factors provided evidence that these proteins directly regulate nucleosome structure.

A nucleosome consists of about 147 base pairs of DNA wrapped around eight histone proteins. It is a basic unit of DNA packaging and transcriptional regulation, but it does not always maintain the same form. Evidence has accumulated over decades for noncanonical structures in which some histones are missing or DNA is partly unwrapped. The remaining challenge was to distinguish this diversity across individual chromatin fibers at genome scale. With high-resolution imaging, identifying DNA sequences is difficult, while analyses based on chromatin cleavage have trouble distinguishing whether short fragments result from histone loss or transient DNA unwrapping.

Key findings: Internal accessibility and transcription factor regulation

The researchers developed IDLI (Iteratively Defined Lengths of Inaccessibility) to iteratively analyze accessibility patterns in single-molecule long-read footprinting data. The underlying experimental method, SAMOSA, uses PacBio sequencing to read marks left by the methyltransferase EcoGII on accessible DNA. IDLI infers protein-occluded DNA segments, or footprints, and accessibility patterns within them. It does not directly image the structures.

The researchers classified 795,765 footprints from 145,065 sampled chromatin fibers in E14 mouse embryonic stem cells into 14 types and seven groups. Groups II–VII, accounting for 86.6% of the total, suggested some degree of accessibility within nucleosomal DNA. The researchers interpreted this as nucleosome “distortion.” The patterns ranged from types in which only specific points along the wrapped DNA path were accessible, to types with both ends unwrapped and types interpreted as having lost some histones. This proportion is specific to the sampling and classification results in these cells.

Structural types were associated with transcription factor binding sequences and gene expression levels. Separate perturbation experiments provided evidence for direct regulation. Acute depletion of the transcription factors CTCF and SOX2 changed the distribution of footprint types around their binding sites. After SOX2 depletion in particular, some types with low accessibility also decreased. This shows that transcription factors’ roles are not limited to opening DNA.

An analysis comparing human induced pluripotent stem cells before differentiation with cells on day 6 of in vitro endoderm differentiation also found changes in structural types at FOXA2 binding sites. In primary mouse hepatocytes carrying one copy of a FOXA2 variant allele lacking a nucleosome-binding region, certain types decreased. This provides additional evidence that FOXA2 directly contributes to nucleosome structure regulation in vivo.

Meaning and outlook: A map of regulated partial accessibility

Rather than dividing nucleosomes only into fully closed or open states, this study establishes a basis for analyzing diverse patterns of partial accessibility. The paper’s term “programmed” distortion means that structural patterns are regulated according to transcription factors and developmental context. It does not mean that the timing and degree of activation of specific genes are predetermined.

The association between structural types and expression levels alone cannot establish that distortion causes gene activation. The researchers described the work as a proof of concept and said that the significance of co-occurring types must be tested at higher resolution. More data from target binding sites are also needed for precise analysis of individual transcription factors. The study did not assess patient diagnosis or the therapeutic effects of gene editing.

The study’s findings and methods are available in the original Nature paper and the full text and methods provided by Europe PMC.