During cell division, chromosomes compact into dense mitotic bodies where centromeres consistently localize at the chromatin surface to enable kinetochore-microtubule interactions. The mechanisms responsible for this surface positioning remain poorly understood. To address this, we combine cellular perturbations, biochemical reconstitution, and multiscale molecular dynamics simulations, with the goal of capturing interactions across molecular to mesoscale resolutions. In earlier work, we employed our multiscale chromatin model to investigate linker length dependence in chromatin arrays and to enhance the resolution of cryo-ET. These studies demonstrated the model’s ability to reproduce experimental results with high fidelity, establishing it as a reliable tool for connecting mesoscale chromatin behavior with molecular detail. Building on this foundation, we now deploy the model to interrogate biologically relevant questions of centromere organization. Simulations of centromeric condensed chromatin and associated proteins showed that surface localization emerges from electrostatic repulsion. Specifically, the highly negatively charged CENP-B protein exhibited unfavorable interactions with the compact chromatin phase, driving centromeric regions outward. Analysis of spatial distributions revealed the emergence of a biphasic centromere architecture: one domain enriched in chromatin-favoring interactions that embeds within the chromosome body, and another domain dominated by repulsive components that is excluded to the surface. Control simulations with synthetic, tethered negatively charged proteins reproduced the same behavior, demonstrating that electrostatic effects alone are sufficient to recapitulate centromere surface localization. These results were recapitulated in cells, in vitro and in silico: importantly, they demonstrate that centromere layering is not determined by predetermined folding motifs or condensin-driven loop extrusion, but rather arises from phase separation driven by charge asymmetry within the chromatin environment. Our work highlights how coarse-grained simulations can reveal emergent mesoscale behaviors inaccessible to either atomistic models or purely experimental approaches and establishes electrostatic polarity as a general, programmable physical principle for organizing chromatin.
Maristany et al. (Sun,) studied this question.