Eukaryotic DNA is packaged into chromatin, consisting of repeating nucleosome units. Each nucleosome comprises ~146 base pairs (bp) of DNA wrapped around a histone octamer, connected by linker DNA, and sometimes associated with a linker histone. The arrangement of these nucleosomes into higher-order structures is crucial for gene regulation and genome function. While the structure of individual nucleosomes is well-understood, the organization of nucleosomes into higher-order structures remains less clear. In vitro studies often involve large, dynamic assemblies, while in vivo imaging techniques capable of resolving individual nucleosomes are relatively new. The prevailing model of chromatin structure was the folded, helical '30 nm' fiber, but recent in situ studies suggest a more irregular, interdigitated fiber structure with zigzag characteristics. This study aims to resolve this ambiguity using crystallography to determine the atomic-level structure of nucleosome fibers under compacting conditions and investigate the role of linker histone in this process. Understanding this structure is crucial for fully understanding genomic function and regulation.

For several decades, the dominant model for chromatin higher-order structure was the 30 nm fiber, proposed as either a solenoid or a two-start zigzag. These structures were observed in vitro under specific, low divalent cation concentrations. However, recent research using various imaging techniques (e.g., ChromEMT) on in situ chromatin reveals that compact chromatin often comprises irregularly interdigitated fibers with zigzag characteristics, rather than 30 nm structures. These findings challenge the traditional view of chromatin organization. While 30 nm fibers have been reported in contexts like terminal differentiation, the factors determining whether chromatin adopts interdigitated or 30 nm configurations remain unclear. Crystallographic studies have provided valuable insights into chromatin compaction, but the dynamic nature of nucleosomal systems makes it challenging to obtain well-ordered crystals of large assemblies. The existing crystal structures of larger nucleosome assemblies are based on blunt-ended DNA fragments, limiting resolution due to disorder and lack of double helix continuity. This study utilizes a different approach to address these challenges.

The researchers created two cohesive-ended dinucleosome constructs, 349 bp and 355 bp, differing in linker DNA length. These constructs were designed with near-maximal histone octamer affinity sequences flanked by histone octamer-refractory poly-A/T elements, terminating in single-stranded 3' TGCA overhangs. In divalent metal-containing buffers, these dinucleosomes self-assembled into lattices through Watson-Crick base pairing at the cohesive termini. This resulted in continuous DNA double helix continuity across the crystal lattice. The structures were solved using X-ray crystallography, reaching resolutions of 3.4 Å and 3.8 Å for the 349 and 355 constructs, respectively. Crystals were also grown with and without the linker histone H1.0 to assess its impact on fiber structure and packing. The recombinant human core histones (H2A, H2B, H3, H4) were expressed and purified from *E. coli*, while human H1.0 was similarly expressed, purified, and its molecular weight confirmed by mass spectrometry. Dinucleosomes were reconstituted using established protocols, and their ability to form fibers was confirmed via a dinucleosome ligation assay using T4 ligase. Crystallization was performed via hanging-droplet vapor diffusion, with crystals stabilized in cryoprotectant solutions for data collection. X-ray diffraction data were collected at beamline X06DA of the Swiss Light Source. Data processing involved using software such as iMosflm, XDS, autoPROC, SCALA and AIMLESS. Structure solutions were obtained using molecular replacement with existing nucleosome and H5 (avian H1.0) structures as search models. Atomic refinement and model building were done with REFMAC and COOT. 3DNA was used for double helix parameter analysis.

Both 349 bp and 355 bp dinucleosomes assembled into continuous nucleosome chains with open zigzag configurations. The fibers were composed of two parallel nucleosome columns, creating a staircase-like profile. The 355 bp fiber was highly extended (42.5 Å rise per nucleosome), with dinucleosome repeats perpendicular to the fiber axis. The 349 bp fiber, while also zigzag, was less extended (30.3 Å rise per nucleosome) and more folded. The difference in fiber structure was attributed to the linker DNA length, affecting the double helix twist and altering the relative orientation of nucleosomes. The fibers interdigitated with one another, creating a densely packed structure. Interdigitation involved nucleosomal face-to-edge contacts, supported by histone-DNA and divalent metal-mediated interactions. The presence or absence of linker histone did not significantly alter the overall fiber structure and packing. However, in the 349 bp fiber, two distinct linker histone binding modes were observed: an on-dyad mode (similar to previously observed structures), mediating internucleosomal contacts within a fiber, and a non-dyad mode, mediating interfiber contacts. Both binding modes utilized similar DNA binding motifs by H1.0, suggesting a flexible, adaptable role for linker histones in chromatin organization.

These findings demonstrate that nucleosome chains have a preference for interdigitating zigzag conformations, consistent with observations of heterochromatin and metaphase chromosome architecture. The open zigzag structure minimizes extreme linker DNA bending, and the linker DNA length profoundly influences fiber configuration and inter-fiber packing. The variable linker DNA lengths observed in vivo suggest substantial heterogeneity in local chromatin chain conformation. Various interactions stabilize the nucleosome-nucleosome interactions: core histones, linker histones, DNA, and divalent metal ions. Histone tails, while disordered, likely contribute to these interactions. The observed structures are consistent with proximity ligation studies showing enriched N to N+2 and N to N+3 contacts. The findings challenge the exclusive existence of 30 nm fibers in vivo and support the notion of a heterogeneous chromatin structure, influenced by factors like linker DNA length, histone variants, post-translational modifications, and other nuclear factors. The presence of linker histone, while not altering overall fiber structure, contributes to interfiber stability through varied binding modes.

This study provides near-atomic resolution structures of interdigitated nucleosome fibers, revealing open zigzag conformations as a common mode of chromatin organization. Linker DNA length is a critical determinant of fiber structure and packing, and linker histone binding stabilizes the compact state. The findings support a model of heterogeneous chromatin structure, shaped by an interplay of DNA sequence, histone modifications, and architectural factors. Future research could focus on investigating the influence of other nuclear factors on fiber structure and exploring the dynamic transitions between different chromatin states.

The study is based on in vitro crystal structures, which may not fully capture the complexity and dynamics of chromatin in vivo. The use of specific DNA sequences and purified components might limit the generalizability to diverse genomic regions. The resolution of the X-ray crystallography data is relatively low, preventing detailed modeling of all components such as the flexible histone tails.