Markov State Models Reveal How hLL-37Tetramers Relax from a Disordered State to a Cross-α Amyloid Geometry.

Cross-α amyloids are an emerging class of functional peptide assemblies that underlie biofilm formation and antibiotic resistance in bacteria. Beyond their biological significance, synthetic cross-α architectures exhibit remarkable rigidity and piezoelectric response, positioning them as promising templates for biomaterial design. A striking example is provided by the hLL-37peptide-an antimicrobial fragment of human LL-37-which forms supramolecular assemblies featuring a three-dimensional repetition of a four-helix bundle with the cross-α amyloid geometry. While the initial steps of monomer-to-tetramer association are known, the pathways by which these disordered aggregates organize into the ordered cross-α arrangement remain poorly understood. Here, we construct Markov state models (MSMs) from 50 μs of molecular dynamics simulations using detailed structural descriptors to elucidate the relaxation landscape of hLL-37tetramers. The model reveals a dominant metastable state that reproduces the crystallographic cross-α geometry while capturing the emergence of structural order from disordered peptide bundles. Remarkably, the cross-α arrangement emerges spontaneously in solution given that it was previously observed only in solid-state crystallographic studies. The relaxation process is driven by the consolidation of a hydrophobic core that maximizes interpeptide packing while preserving solvent exposure of hydrophilic residues. Notably, transient disruptions of this core are rapidly repaired, suggesting a self-healing mechanism that preserves aggregate stability. To our knowledge, this is the first detailed kinetic and structural characterization of the relaxation processes governing cross-α amyloid formation in hLL-37. These insights advance the molecular understanding of cross-α amyloids and provide design principles for stable, self-organizing peptide-based materials.