Full‐Length FUS Protein Condensates Adopt a Domain‐Preferential Spatial Architecture

ABSTRACT Biomolecular condensates play crucial roles in cellular physiology and are implicated in neurodegenerative diseases and cancer. However, the mechanisms governing their formation and spatial organization remain poorly understood, largely due to technical challenges. Here, using FUS as a paradigmatic system, we reveal how single‐protein sequences determine condensate architecture that is intrinsically linked to biological function. We demonstrate a domain‐specific preferential distribution organization: the low‐complexity domain (LCD), which drives condensate formation, localizes to the inner layer, while the RNA recognition motif (RRM) preferentially occupies the interfacial layer. This spatial arrangement enhances RNA‐binding accessibility, suggesting a direct structure–function relationship. We further propose a sequence–structure–function paradigm for biomolecular condensates: through the cooperation emerging from multiple “stickers,” individual domains function as integrated units in shaping the structure and functionality of biomolecular condensates, which may represent a broader mode of protein–protein interaction (PPI) within condensates. Our findings elucidate the evolutionary logic of protein sequences in driving liquid–liquid phase separation (LLPS) and provide a foundation for designing therapeutics targeting aberrant condensates in disease.