Enzymes are highly evolved catalysts that make metabolism and life possible. They are proteins, often functionalized with cofactors, that catalyze a wide variety of reactions. Enzymes are so proficient that, frequently, it is the rate of diffusion of the substrate onto protein in water which becomes the slow step of the reaction. Because the expression of proteins in organisms such as Escherichia coli has become a relatively routine activity in molecular biology, the idea of creating protein catalysts for any desired reaction has attracted much attention. If methods to design proteins that could catalyze any desired reaction were perfected, then enormous capabilities for synthetic processes, therapeutics, defense against biological threats, the incorporation of new synthetic capabilities into organisms, and many other applications could be realized.

Whilst the factors that influence biological catalysis have been largely explored, not all aspects are yet understood in their entirety [1]. The design of new enzymes is built upon the premise that the process is well enough understood to create protein catalysts without the aid of evolution. In 1948, Pauling proposed that enzymes complement transition states similar to the way that antibodies complement and bind antigens [2]. It has been shown that only covalent catalysis, and a modification of the mechanism, could account for the up to 1023-fold acceleration that occurs with some of the most proficient enzyme catalysts [3]. It might be thought that the screening of proteins would lead to the identification of potential catalysts that could then be subjected to directed evolution to produce efficient catalysts [4, 5]. However, various estimates have been made and experimental studies conducted which have shown that the probability of finding an active catalyst for an arbitrary reaction is vanishingly small [6, 7]. Catalytic antibodies have been created for a variety of reactions,[8–14] and have given up to 106-fold rate accelerations,[15, 16] but the technology required is somewhat daunting for a typical lab. Previous studies towards