Evidence for Prenucleated Fibrilogenesis of Acid-Mediated Self-Assembling Oligopeptides via Molecular Simulation and...

An important step in controlling biomimetic amyloid systems is understanding the self-assembly reaction kinetics. We are interested in a family of such materials characterized by symmetric sequences of amino acids flanking a π-conjugated functional core. Many of these materials rapidly self-assemble into long fibers upon protonation in an acidic environment. Despite extensive investigation of these materials' properties, little is yet understood regarding their reaction kinetics. Based on previous studies, we have chosen DFAG-4T-GAFD as a representative system and conducted molecular dynamics simulations to show that although large-scale assembly is induced by lowering pH, some degree of assembly is thermodynamically favorable in high-pH nonprotonating environments. These results are consistent with findings for other systems such as DFAG-OPV-GAFD. The nonprotonated aggregation also appears to be concentration dependent, occurring at concentrations of 100 nM and above. Single molecule measurements using fluorescence correlation spectroscopy provide experimental support for these computational predictions. We find evidence of spontaneous aggregation in aqueous solutions of peptides with concentrations as low as 100 nM; however, 10 nM solutions appear to be largely homogeneous solutions of unassembled monomer. These results indicate that the simplest explanations for kinetics of acid-mediated assembly-protonation-induced nucleation by monomeric addition followed by subsequent stages of aggregation and elongation-are inappropriate in this system. In fact, the system only exists as pure monomer in very low concentrations, nucleation actually occurs in the absence of protonating elements at concentrations typically used for experiments, and pH triggered assembly proceeds from these preassembled aggregates. Accordingly, triggered assembly must be considered to operate outside the domain of nucleation-dependent models.