Rational design of peptides that can self-assemble into amyloid-like nanostructures with atomic-level precision is of interest for developing bio-functional materials. In previous work, we developed a computational and experimental workflow to design peptides that can self-assemble into amyloid structures. The workflow begins with the Hall group’s Peptide Assembly Design (PepAD) algorithm, which searches for peptides with high binding affinity. The self-assembling dynamics of the discovered peptides are evaluated via coarse-grained discontinuous molecular dynamics simulations coupled with the PRIME20 force field (DMD/PRIME20). This approach enabled the design of peptide fibrils and allowed us to determine whether the β-sheets were parallel or antiparallel. However, we were not able to experimentally confirm whether or not the designed peptides self-assembled into the predicted fibrillar structures. To experimentally determine the atomic-level organization of fibrils, ssNMR characterization is required. In this project, we enhanced the PepAD algorithm to design self-assembling peptides that satisfy NMR C-C distance constraints, making them suitable for ssNMR experiments. Two rounds of peptide design using PepAD and simulation-based validation were conducted. In the first round, we used Class-1 fibril backbones and 14-mer peptides without NMR constraints to design 14-mer NMR-constrained peptides and screened the resulting 12 candidates via DMD/PRIME20 simulations. In the second round, we used the peptides from the first round as initial structures, with and without terminal caps, as input to PepAD. An additional 24 candidates were screened via DMD/PRIME20 simulations. By calculating the shortest C-C distances between the NMR-constrained residues from atomistic MD simulation, we found that most of the designed peptides have at least three, and possibly all four, inter-sheet distances smaller than 6.3 Å. These peptides are ideal candidates for experimental validation, including isotopically labeling and characterization via ssNMR.
