We have previously reported a general method for the deracemisation of racemic chiral amines (primary, [1] secondary [2] and tertiary [3] ) by using variants of the enzyme monoamine oxidase N (MAO-N) from Aspergillus niger. This deracemisation process employs a combination of an enantioselective enzymatic oxidation of the amine to afford the corresponding imine or iminium ion, together with a non-selective chemical reduction of the imine or iminium ion back to the racemic starting material (Scheme 1). The use of an (S)-selective MAO-N enzyme leads to accumulation of the (R)-amine in high enantiomeric excess and yield through several rounds of oxidation and reduction.Previously, we have applied this chemo-enzymatic approach to the deracemisation of the alkaloid (AE)-crispine A. [4] Crispine A (1) was first isolated from extracts of the plant Carduus crispus (welted thistle), along with the cytotoxic crispine B (2, Figure 1) and three other bicyclic isoquinoline alkaloids. [5] Although the deracemisation of (AE)-crispine A resulted in the generation of the R enantiomer in > 97 % ee, the reaction required 40 h to proceed to completion with a MAO-N-5 variant in an overall yield of 48 %. This previous study also highlighted that less-functionalised analogue 3 was more reactive with the same variant, taking only 6 h to reach completion (> 97 % ee), thus indicating that the two methoxy groups of racemate (AE)-1 resulted in lower activity. We reasoned that this drop in activity was due to steric interference between these two groups and residues within the active site of the enzyme. As a result, we employed a combination of molecular modelling and a rational re-design of the MAO-N-5 variant to identify and develop potential new MAO-N variants that could have enhanced activity towards enantiomer (S)-1. In addition, we sought to improve the efficiency of the overall synthesis of enantiomer (R)-1 by utilising the microwave synthesis of the racemic amine coupled with the enhanced deracemisation that is brought about by changes to the MAO-N enzyme.To identify MAO-N variants that have improved activity towards compound (AE)-1, we modelled (S)-crispine A into the active site of the MAO-N-5 enzyme (PDB code 2VVM). [6] Four residues (Phe210, Leu213, Met242 and Met246), which are located at the entrance to the active site channel, were identified as providing possible steric interactions with the methoxy groups of crispine A (1). To optimise these residues, two randomised libraries were created: The first library targeted amino acids Phe210 and Leu213 (library A) and the second library targeted Met242 and Met246 (library B). Both libraries were screened against compound (AE)-1 by using our previously reported solid-phase assay [7] (Figure 2); sixteen active "hits" were collected from each of the two libraries (A and B), which were then subjected to a second round of screening to eliminate any false positives. After the second round of screening, the two MAO-N variants from each library that had the highest activity, as judged by t...