| Summary: | 博士 === 國立清華大學 === 物理系 === 102 === Although it has been generally believed as the Anfinsen's dogma that protein structures are uniquely determined by their amino acid sequences, complex conformational transformations of native structures that occur in multiscales makes Anfinsen's dogma elusive. To illuminate mechanism behind conformational transformations in
multiscales, we start by revisiting the uniqueness problem of protein native structures by investigating the peptide 2DX4, comprising 18 amino acids: INYWLAHAKAGYIVHWTA. Based on an ab initio coarse-grained model that are comprehensively calibrated, we construct the complete free energy of 2DX4 and demonstrate that 2DX4 is an inherently multi-conformation peptide with two nearly degenerate native structures: one is a helix structure, while the other is a hairpin-like structure.
Although there exist pathways connecting two degenerate native structures, conformation switch between two native structures does not occur often due to the existence of an energy barrier of the order 10kcal/mol between them. These results provide important clues that the native structures for small proteins may not be unique so that conformation transformations may occur. Our investigation is then extended to simulate fibril aggregation of peptide A-beta(16-22). It is shown that A-beta(16-22) also possesses two minimum with alpha-helix and beta-hairpin structures. However, due to small energy barrier between alpha and beta structures, transformations
between these structures occur frequently. As a result, A-beta(16-22) does not exhibit a fixed native structure. This feature enables the aggregation and manifestation of multi-conformations for A-beta(16-22). The energy landscapes of A-beta(16-22) from dimer to hexamer is found to be completely uncorrelated with that of a single A-beta(16-22). Antiparallel beta sheets are found to be the ground state, while several intermediates are minor unstable minima. To simulate large scale aggregation, a box container with a seed in which other monomers start to interact with the seed is implemented. Using this new simulation scheme, we are able to simulate aggregations of A-beta(16-22) in different concentrations. It is shown that similar to the growth of nanowires, when the density of A-beta(16-22) is low. The aggregation is dominated by diffusion limited aggregation, while for high densities, the system tends to be amorphous with multiple ground states. Our analyses indicate that in the low density region, the aggregation can be characterized by classical nucleation theory, including the lag phase, critical nucleation, fibrilization and their pathway dependence on the concentration. These results shed light on the multi-formational nature of aggregated proteins and provide important clues to understand aggregations pathways for amyloid assemblies.
|
|---|
