Programmable Artificial Hydrolase Based on Minimalistic Cross-α/Cross-β Amyloid Assemblies

Abstract Enzymes, which catalyze nearly all biochemical processes, are delicately organized special substances composed of large proteins, specific secondary structures, distinct amino acid residues with precise positioning, and often metal cofactors. Reconstruction of enzymatic catalytic domains within fully artificial systems has become an interesting but challenging subject, largely due to limited understanding of their evolutionary processes. A long-standing opportunity in the generation of artificial or miniature enzymes has been to design building units capable of performing similar chemistry without the structural complexity of full proteins. Herein, by using a reductionist approach, we explore the sequence–structure–activity landscapes of enzyme evolution by investigating minimal amyloid-forming analogous peptides. Our results reveal that amyloid peptides without encoding canonical catalytic amino acids (like Lys, Ser, His, Glu/Asp) or metal cofactors can nevertheless catalyze ester hydrolysis through fibril-dependent substrate binding. Minor sequence variations alter supramolecular organization and reorganize solvent-exposed terminal groups, creating distinct catalytic microenvironments and activities. Combining experiments with molecular simulations, we show that these fibril surfaces present topologically different binding grooves that modulate the association of the model substrate. This sequence–structure–activity paradigm of de novo biocatalysts holds promise for predictable and robust construction of peptide-based new functional materials applicable in chemical and biochemical synthesis.