A DNA computing-based genetic program for in vitro protein evolution via constrained pseudomodule shuffling

An in vitro domainal shuffling strategy for protein evolution was proposed in (J. Kolkman and W. Stemmer, Nat. Biotech. 19 (423) 2001). Due to backhybridization, however this method appears unlikely to be an efficient means of iteratively generating massive libraries of combinatorially shuffled genes. Recombination at the domain level (30-300 residues) also appears too coarse to support the evolution of proteins with substantially new folds. In this work, the module (10-25 residues long) and pseudomodule are adopted as the fundamental units of protein structure. Each protein is modelled as an N to C-terminal tour of a digraph composed of pseudomodules. An in vitro method based on PNA-mediated Whiplash PCR (PWPCR), RNA-protein fusion, and restriction-based recombination, XWPCR is then presented for evolving proteins with a high affinity for a given motif, subject to the constraint that each corresponds to a walk on the pseudomodule digraph of interest. Simulations predict that PWPCR is an efficient method of producing massive, shuffled gene libraries encoding for proteins as long as roughly 600 residues.