CATALYSTS

‣ condensation and therefore lead to the inhibitive effects . This identified inhibition is proposed to be the main cause for DERA inactivation under high aldehyde concentrations . Due to the catalytic importance of L167 , only the C47 residue can be engineered and C47M mutation was identified as the most tolerant variant . ⁷ DSM engineered and optimised DERA to catalyse a tandemaldol reaction for the industrial synthesis of ( 3R , 5S )-6-chloro-2,4,6- trideoxyhexapyranoside , which is the chiral precursor for different vastatin drugs such as atorvastatin and rosuvastatin . ^6-8 In this particular case , the beneficial mutations with high stability towards 2-chloroacetaldehyde found by random mutagenesis were combined and a tenfold improved variant was identified for the industrial application .

Chiral intermediates using PLP-dependent L-threonine aldolase

PLP-dependent aldolases are a further tool to catalyse stereoselective C-C bond formation generating chiral molecules . In another recent work , we engineered and evolved a low specificity L-threonine aldolase for the formation of a chiral intermediate ( Figure 1c ) in the synthetic route to chloramphenicol from achiral precursors . ⁹ Our enzymatic route is five steps ( 60 %) shorter than the traditional chemical route . However , the underlying wild-type aldolase from Pseudomonas putida shows low diastereoselectivity in the formation of ( 2S , 3R ) -2-amino-3- hydroxy-3- ( 4-nitrophenyl ) propanoic acid , with a diastereomeric excess of about 50 %. In addition , poor

Figure 4 – C-terminal domain of the aldolase & region that should be stabilised for improved binding of ( 2S , 3R ) -diastereomer

enzymatic activity leads to only 10 % conversion within eight hours . To transform the aldolase into an industrially applicable catalyst , we developed a method to determine the preferred binding mode of the desired ( 2S , 3R ) -diastereomer and evolved the respective enzyme specificity pocket to stabilise the transition state of the desired product ( Figure 3 ). In particular , we identified the correct binding orientation and engineered the aromatic binding pocket near the C-terminal G318 and W322 . Six sites in this region were targeted for stabilisation ( Figure 4 ), which results in about 33 million variants in the library with full diversity . Using a unique computational design within the BioNavigator toolbox , we were able to reduce the library to fewer than 1,000 variants . After nine rounds of directed evolution , the engineered aldolase achieved de values of

> 90 % and conversion rates of > 80 % within eight hours , while the possible substrate loading was highly improved from 40 g / L to 200 g / L .

Conclusion

The examples of biocatalytic C-C bond formation presented here represent only a small number of the diverse applications and opportunities that enzymes provide in organic and chiral synthesis . Computer-guided enzyme engineering can help leverage the full power of such enzymes and use these for sustainable chemical manufacturing at scale , where enzymatic processes can often be designed to outperform traditional chemistry in terms of selectivity , atom economy and a reduced number of synthetic steps . •

* BioEngine and BioNavigator are both trade marks of Enzymaster

42 SPECIALITY CHEMICALS MAGAZINE ESTABLISHED 1981