Unnatural cyclic secondary amino acids such as substituted proline and pipecolic acid derivatives appear in a number of peptidomimetic drug candidates in development. Examples include the unnatural 3,4-cis-fused bicyclic L-proline analogs that are key constituents in the peptidomimetic Hepatitis C virus protease inhibitors boceprevir (Schering-Plough, Phase 2) and telaprevir (Vertex, Phase 3). These are candidates for a first-in-class Hepatitis C therapy to address an important unmet medical need.

Prior chemical routes to produce these proline analogs are "steppy" and involve wasteful late separations of stereoisomeric mixtures. Codexis envisioned shorter processes that avoid separating stereoisomers involving biocatalytic desymmetrization of the corresponding 3,4-cis-fused bicyclic pyrrolidines (the starting materials for prior routes) to the single imine stereoisomer that is precursor to the L-amino acid.

Wild-type monoamine oxidases were identified that are capable of desired stereoselective oxidation of the symmetric cyclic amines to the desired asymmetric cyclic imines, but their characteristics were otherwise wholly inadequate to enable practical processes. They were poorly produced recombinantly and their catalytic activities and in-process stability are extremely poor. Applying its advanced directed evolution technologies, Codexis evolved monoamine oxidases that are readily manufactured and fit to enable the development of practical, scaleable reactions. The subsequent steps provide the desired L-amino acids without any resort to separating stereoisomers. Process development also involved addressing some interesting reaction engineering issues, which will also be presented. These processes are now being scaled up to provide material for clinical trials.