Exploration of novel αβ-protein folds through de novo design

Most naturally occurring protein folds have likely been discovered. The question is whether natural evolution has exhaustively sampled almost all possible protein folds, or whether a large fraction of the possible folds remains unexplored. To address this question, we introduce a set of rules for β-sheet topology to predict novel folds, and carry out the systematic de novo protein design for the novel folds predicted by the rules. The rules predicted eight novel αβ-folds with a four-stranded β-sheet, including a knot-forming one. We designed proteins for all the predicted αβ-folds and found that all the designs are monomeric with high thermal stability and fold into the structures close to the design models, demonstrating the ability of the set of rules to predict novel αβ-folds. The rules also predicted about twelve thousand novel αβ-folds with five- to eight-stranded β-sheets ; the number is far exceeding the number of αβ-folds observed so far. This result suggests that the enormous number of αβ-folds are possible but have not emerged or become extinct due to evolutionary bias. The predicted novel folds should open up the possibility of designing functional proteins of our interests.