ABSTRACT
Proteins are malleable molecules and tolerate many mutations with marginal effects on structure and function. A limited number of three-dimensional folds are recycled repeatedly for different functions. This malleability results from surviving hundreds of millions of years of mutation and recombination. Natural proteins thus provide a gold mine of starting materials for those interested in modifying proteins to suit their needs. Unlike nature, however, protein engineers try to evolve molecules on a much shorter time scale. To that end, combinatorial gene synthesis and in vitro evolution methods (Stemmer 1994; Giver et al. 1998) have been developed to accelerate the mutation and selection process with impressive results. Rational protein engineering is another approach that applies our knowledge of intermolecular forces and protein structure toward the identication of mutations that confer the desired properties on a target protein (Marshall et al. 2003). Computation automates much of the design process, allowing larger, more complex targets to be rationally engineered.
