Fig. 4

The representative structures and the results of sequence-dependent folding simulations of the designed single-chain three-helix bundles: GB-CCW9, GBB-CW11, and BAAB-CCW8. A The side-view and top-view of the designed structures with the α-helix shown as a cartoon and the hydrophobic side-chains represented as sticks. B The results of the folding and relax simulations. The vertical axis represents the Rosetta score, and the horizontal axis represents the root-mean-square deviation from the target structures. The black dots correspond to the final snapshots of the fragment-assembly folding simulations starting from extended conformations, and the red dots correspond to the final snapshots of the relax simulations starting from the native conformations. These designs are predicted to fold into the target topologies because the trajectories of the folding simulations can reach the near-native ensembles. C The lowest score models in folding simulations (orange) superimposed onto the design models (white). The predicted models and design models agrees well, which suggests the designed amino acid sequences fold well into the target conformations