Tunable supramolecular polymerization from protein charge heterogeneity and architecture

Abstract

Multidomain proteins with flexible unstructured sequence regions are abundant in cellular signaling. This protein architecture enables self-assembly into supramolecular structures, but how structured interaction domains and overall protein architecture jointly regulate the assembly size, structure and kinetics remains unclear. Here we use the budding yeast protein Bem1 as a model multidomain system to show that supramolecular polymerization can be tuned by charge heterogeneity and protein architecture. We experimentally demonstrate that Bem1's isolated PB1 domain forms extended filaments, whereas full-length Bem1 forms substantially shorter assemblies, indicating that the PB1 domain drives assembly while the remaining protein architecture tunes filament length. To understand these observations, we develop minimal coarse-grained models approximating the PB1 as a polar 5-bead domain and the full-length Bem1 as a 6-bead model with an additional bead representing the remainder of Bem1. The weight distribution of supramolecular filaments assembled by the 5-bead model quantitatively follows reversible Flory-like polymerization theory, which is tunable within a narrow charge polarity regime. In contrast, the 6-bead model shifts chain-length distributions towards shorter polymers despite retaining the same driving domain. We show that this deviation arises from steric and geometric constraints imposed by the appended unstructured regions, where the rotational flexibility between the charge-polar structured domain and the unstructured region emerges as key physical parameter governing self-limited self-assembly. Together, our results establish charge polarity, protein architecture, and conformational flexibility as programmable control knobs for supramolecular polymerization and suggest a general framework for understanding how multidomain proteins assemble into tunable biomolecular structures.