Amphipathic β‑Sheet-Forming Octapeptide Self-Assembly Using the Martini Potential Family

Amphipathic peptides can self-assemble into β-sheet-rich fibrils and hydrogels. The potential applications of these nanomaterials in biomedicine, drug delivery, and tissue engineering have sparked significant research. Current computational screens for amphipathic peptides heavily rely on coarse-grained molecular dynamics (CGMD) simulations using the Martini force field to ascertain the self-assembly potential of novel peptide sequences. This approach, however, showed limited success in self-assembling amphipathic peptides into experimentally observed β-sheet-rich hydrogels. We have systematically investigated this phenomenon using FKFEFKFE, an amphipathic octapeptide with a solved high-resolution self-assembled three-dimensional (3D) structure, as our model system. We first performed all-atom molecular dynamics (AAMD) simulations, which support the essential role of neutral capping groups in reproducing the cryogenic electron microscopy (Cryo-EM) structure. Accordingly, CGMD simulations were conducted to compare the efficacy of the Martini versions 2.1, 2.2, 2.2P, and 3. Surprisingly, Martini 2.1 was the best at maintaining the self-assembled bilayer structure of the peptide observed in Cryo-EM. The self-assembly of 20 mM FKFEFKFE peptides from a random initial arrangement was then examined via 5 μs CGMD simulations, further confirming that Martini 2.1 can successfully simulate the formation of a β-sheet-rich bilayer with phenylalanine side chains embedded between the two layers. Moreover, modifications to the Martini 3 potential and the use of small water models, which were promising for shorter-peptide self-assembly, have limited success. We further showed that configurational entropy performs better at characterizing the assembly of ordered structures than the routinely used aggregation propensity score. Though the FKFEFKFE peptide was exclusively examined, the comprehensive benchmarking conducted here provides valuable insights into the factors influencing the CGMD of peptide self-assembly. Overall, the results support the development of robust, reproducible CGMD protocols and analysis tools for studying self-assembling peptides, enabling the discovery of novel supramolecular structures and biomaterials.