A Methodological Approach to Relate Pentapeptide Sequence, Atomistic Self-Assembly, and Gelation Behavior

The development of biomaterials that mimic the extracellular matrix of the native tissue represents an exciting frontier for tissue engineering and regenerative medicine. Injectable hydrogels made of short, self-assembling peptides offer a promising platform for the delivery and directed differentiation of therapeutic stem cells. However, the rational design of peptide hydrogels remains a significant challenge in tissue engineering due to our lack of understanding of the molecular mechanisms that underlie self-assembly. Although these materials hold great promise, most computational design efforts have focused on studying how peptide sequence impacts aggregation propensity. While useful as an initial indicator of self-assembly, aggregation propensity can be misleading as it is nearly synonymous with hydrophobic precipitation, thus highlighting the need for more robust screening and design strategies. To address this limitation, we introduce a systematic approach to study self-assembly beyond aggregation via molecular dynamics for designing peptide hydrogels. Our approach introduces several new atomistic descriptors-end-to-end distance, π-π stacking interactions, and residue-specific contacts-derived from molecular dynamics simulations to capture the nuances of the sequence-dependent assembly. We additionally uncovered key interactions among hydrophobic, aromatic, and charged residues that reliably predict gel formation, enabling a more rational approach to hydrogel design. We apply these molecular features to successfully predict a previously undiscovered, yet robust self-assembling sequence, KYYYL. An analysis of variance (ANOVA) confirms that our parameters provide significant differences among sequences, whereas aggregation propensity failed to reject the null hypothesis. Finally, we establish the sensitivity of simulation parameters to ensure methodological rigor and enable future study expansion in peptide sequence space. Our findings reveal that amino acid selection and position influence self-assembly. Furthermore, we demonstrate the key interactions among varying residues that reliably predict gel formation, enabling a more rational approach to supramolecular hydrogel design.