Abstract
While peptide aggregation and protein phase separation have been well-studied, the rules linking sequence patterning to aggregate morphology, particularly how microscopic hydrophobic/electrostatic arrangements dictate macroscopic structure, remain elusive. Here, we decode this link by systematically designing and analyzing 50, mainly phenylalanine-rich, tri- to heptapeptides with shuffled sequences excised from natural phase-separating proteins. Combining coarse-grained simulations with transmission electron microscopy, nuclear magnetic resonance, thioflavin assay and circular dichroism spectroscopy experiments, we demonstrate that peptide hydrophobicity and sequence patterning determine association propensity and show how hierarchical assembly is driven by clustering of aromatic (F) and charged residues. This patterning yields spheres, bilayers, web-like networks, and porous sponges. Notably, subtle sequence shuffling (e.g., KFF vs KDFF or KDFF vs KFFD) dramatically alters morphology, underscoring the precision of design. Our work establishes a framework for programming peptide self-assembly via sequence logic, bridging amino acid patterning to functional material properties.