a a b so
b antiepileptic activity . It is therefore commonly used for the treatment of epilepsy and neuralgia . Pregabalin was originally manufactured by Pfizer in the US . It was approved by the EU for the treatment of partial seizures in July 2004 and by the US FDA in 2005 .
The chemical synthesis of enantiopure Pregabalin via racemic intermediates and subsequent chemical or enzymatic resolution by lipase , amidase or nitrilase is fairly inefficient . 5 To reach a high atom economy , chemical re-racemisation of the undesired enantiomer is needed , but this often requires harsh conditions and is therefore not environmentally friendly .
An attractive alternative is the application of a hydantoinase for enantioselective ring opening of a symmetrical prochiral cyclic imide ( Figure 3 ). First , we assembled and screened a diverse panel of putative wild-type imidinases and hydantoinases , some of which showed only moderate activity
70 60
- 50 �40
>
20
Q) 30
10 0
� 20
10
0 0
20
···•··· 3nd Round
- -G - · 2nd Round c
18 16 14
� 12 C 0
C 0
10 8 6 4 2 0
40 60 80 100 120 140 160 180
2nd Round
Product Cone . ( g / L )
wild type AA64 AA67
···•··· 3rd Round 1 h pre-incubation - -0 -- 2nd Round 1 h pre-incubation
Stage of evolution
Amino acid redidue no .
■ Yield Selectivity
4th Round
Figure 4 – Molecular dynamics simulation of production inhibition ( a ), production inhibition studies ( b ) & decrease of product inhibition ( c )
a
Figure 5 - Identified d-hydantoinase hotspots in the active site ( yellow ) and dimer interface ( a ) & effect of amino acid exchanges at key residues on activity & selectivity ( b )
b
AA337
and enantioselectivity towards the desired Pregabalin intermediate . This was also demonstrated in a recent academic study . 6
The best wild-type enzyme identified in our panel was a d-hydantoinase . 7 It showed high (> 90 %) conversion and good (> 95 %) enantioselectivity but it suffered from severe substrate and product inhibition , and required high enzyme loading .
During four rounds of directed evolution , increased activity , stability under process conditions and reduced substrate / product inhibition were targeted . Product inhibition was analysed using molecular dynamics by placing the enzyme in a simulation box employing conditions of atmospheric pressure , 30 ° C and 200 g / L of 14 ( Figure 4a ). Using this method , less stable regions of the protein could be identified and targeted for increasing stability by introducing amino acid exchanges to evolve the enzyme . In product inhibition studies , evolved enzymes of evolution rounds
100
98 -
Q) Q)
96
-
94
+-'
>
Q)
Q)
92 V)
90
2 and 3 were incubated in immobilised form for 24 hours at 40 ° C with 10 g / L of substrate and increasing amounts of 14 ranging from 0-160 g / L ( Figure 4b ). Enzymes were either used directly or pre-incubated for one hour with the same amount of 14 used later in the inhibition study .
It can be clearly seen that the evolved enzyme of round 3 was already more stable than the one from round 2 . Furthermore , pre-incubation of the enzyme with product prior to the reaction destabilised the enzyme even more . The evolved enzyme from round 4 showed almost threefold improved conversion compared to the evolved enzyme from round 2 when pre-incubated with 70 g / L product and subsequently incubated with 150 g / L substrate at 40 ° C for 28 hours .
Besides stability , activity and selectivity were two further attributes of the d-hydantoinase that we looked into . A 3D structure of the enzyme generated by homology modelling based on 84.5 % identical dihydropyrimidinase from Pseudomonas aeruginosa hot spots located in the active centre ( yellow ) and dimer interface region were identified ( Figure 5a ).
In particular , the amino acid residues 64 , 67 and 337 were all positions that positively affected the activity when mutated under the conditions : 10 g / L substrate load , 50 g / L catalyst load , 30 ° C and 24 hours . Under the same conditions amino acid residues 64 and 337 were shown to affect the selectivity as well in a positive way while amino acid residue 67 was neutral and remained the same selectivity ( Figure 5b ).
Overall , the four rounds of evolution resulted in a d-hydantoinase variant with more than 100-fold activity improvement , corresponding to 99 % conversion within 12 hours and excellent ee of 99 % under the desired process conditions ( Figure 6 ). 8 Moreover , the catalyst load could be reduced by a factor of five , while the substrate load could be increased by a factor of 100 .
36 SPECIALITY CHEMICALS MAGAZINE ESTABLISHED 1981