Protein design from in silico dynamic information: the emergence of the 'turn-dock-lock' motif

A protein design methodology based on ab initio folding simulations is described and illustrated. First, the time evolution of the chain topology is generated to identify a collapse-triggering nucleus. Then, a minimal spliced sequence of nuclear residues is created and systematically mutated in silico until it can sustain a stable conformation retaining the original nucleus topology. The mutations introduce a structural compensation for the deletions and eventually lead to the recovery of the native fold motif beyond topological identity. For ubiquitin, the systematically modified sequence is predicted to be a resilient folder, since it is 92% homologous to the hyperthermophile variant of B1-domain in streptococcal protein G. The methodology enabling us to identify the nucleus is independently validated vis-á-vis site-directed mutagenesis experiments on chymotrypsin inhibitor (CI2).