AI-guided design yields high-affinity nucleosome acidic-patch binders

Artificial-intelligence-assisted protein design has produced peptides with high affinity for the acidic patch of canonical nucleosome core particles, according to a brief research report in Frontiers in Bioengineering and Biotechnology . The study combined BindCraft-based sequence design with an adapted donor FRET-quenching assay to identify and quantitatively characterize the interactions. The reported result establishes a laboratory-validated peptide-binding workflow for the nucleosome acidic patch, an interaction interface used by numerous chromatin-associated proteins.

Chromatin is an assembly of DNA, RNA and proteins in the nuclei of eukaryotic cells, and the protein interactions governing its function are important regulators of genome activity. The researchers focused on the nucleosome, one of the primary structural units of chromatin, and specifically on its acidic patch. This region is a binding site for many chromatin-associated proteins, including proteins that engage it through intrinsically disordered regions.

Because engineering or modulating these interactions could provide a route to epigenetic therapeutics, the acidic patch offers a defined molecular target for designed binding peptides. The computational component used BindCraft, an AI protein-design approach, to generate candidate peptide sequences directed toward the nucleosome acidic patch. The authors then refined a donor FRET-quenching assay for this application so that it could provide sensitive and consistent measurements of the binding affinities of the designed sequences.

This combined design-and-validation framework meant that the reported binders were not selected solely from computational outputs: their interactions with canonical nucleosome core particles were evaluated experimentally through the adapted FRET assay. Among the results, the researchers successfully identified peptides with high affinity for the acidic patch. The work therefore demonstrates a route by which BindCraft-generated candidates can be selected through quantitative binding measurements and converted into experimentally verified nucleosome-binding peptides.

The demonstrated outcome is narrower: AI-assisted peptide design, coupled to a tailored FRET-based validation assay, yielded peptides that bound the intended nucleosome acidic-patch region with high affinity. For synthetic biology, the practical significance lies in the workflow itself. AI-assisted design can propose molecular sequences for a biologically important interface, while a dedicated quantitative assay tests those proposals and measures their interactions.

The authors describe the findings as groundwork for synthetic epigenetic modulators and for broader use of protein design in managing three-dimensional genome organization. At this stage, however, the evidence supports a biochemical binding result rather than a demonstrated epigenetic, cellular or therapeutic effect. The study’s central achievement is the experimental validation of BindCraft-designed peptides as high-affinity binders of the nucleosome acidic patch, together with a refined donor FRET-quenching assay developed to measure those interactions precisely.

Subject of Research: BindCraft-guided design and FRET validation of peptides binding the nucleosome acidic patch

Article Title: Design and validation of nucleosome binding peptides using BindCraft

Article References: Oleinikov, P. D., Armeev, G. A., Sidorkina, A. E., Rykunova, E. B., Komarova, G. A., Motorin, N. A., Gerasimova, N. S., Feofanov, A. V., Studitsky, V., & Shaytan, A. K. (2026). Design and validation of nucleosome binding peptides using BindCraft. Frontiers in Bioengineering and Biotechnology, 14, 1946220. https://doi.org/10.3389/fbioe.2026.1946220

Image Credits: AI Generated

DOI: 10.3389/fbioe.2026.1946220

Keywords: BindCraft, AI protein design, nucleosome acidic patch, nucleosome binding, peptide design, FRET assay, chromatin interactions, epigenetics, synthetic biology, Design, validation, nucleosome