Refine
Document Type
- Article (3)
- Doctoral Thesis (1)
Language
- English (4) (remove)
Has Fulltext
- yes (4)
Is part of the Bibliography
- no (4)
Keywords
- Biocatalysis (1)
- Catalytic Activity (1)
- Machine Learning (1)
- Stereoselectivity (1)
- Transaminase (1)
- Transaminases (1)
- amine transaminases (1)
- asymmetric synthesis (1)
- pH optimum (1)
- protein engineering (1)
Institute
Publisher
- MDPI (1)
- Nature Publishing Group (1)
- Wiley (1)
Fast screening of enzyme variants is crucial for tailoring biocatalysts for the asymmetric synthesis of non-natural chiral chemicals, such as amines. However, most existing screening methods either are limited by the throughput or require specialized equipment. Herein, we report a simple, high-throughput, low-equipment dependent, and generally applicable growth selection system for engineering amine-forming or converting enzymes and apply it to improve biocatalysts belonging to three different enzyme classes. This results in (i) an amine transaminase variant with 110-fold increased specific activity for the asymmetric synthesis of the chiral amine intermediate of Linagliptin; (ii) a 270-fold improved monoamine oxidase to prepare the chiral amine intermediate of Cinacalcet by deracemization; and (iii) an ammonia lyase variant with a 26-fold increased activity in the asymmetric synthesis of a non-natural amino acid. Our growth selection system is adaptable to different enzyme classes, varying levels of enzyme activities, and thus a flexible tool for various stages of an engineering campaign.
Amine transaminases (ATAs) are powerful biocatalysts for the stereoselective synthesis of chiral amines. However, wild-type ATAs usually show pH optima at slightly alkaline values and exhibit low catalytic activity under physiological conditions. For efficient asymmetric synthesis ATAs are commonly used in combination with lactate dehydrogenase (LDH, optimal pH: 7.5) and glucose dehydrogenase (GDH, optimal pH: 7.75) to shift the equilibrium towards the synthesis of the target chiral amine and hence their pH optima should fit to each other. Based on a protein structure alignment, variants of (R)-selective transaminases were rationally designed, produced in E. coli, purified and subjected to biochemical characterization. This resulted in the discovery of the variant E49Q of the ATA from Aspergillus fumigatus, for which the pH optimum was successfully shifted from pH 8.5 to 7.5 and this variant furthermore had a two times higher specific activity than the wild-type protein at pH 7.5. A possible mechanism for this shift of the optimal pH is proposed. Asymmetric synthesis of (R)-1-phenylethylamine from acetophenone in combination with LDH and GDH confirmed that the variant E49Q shows superior performance at pH 7.5 compared to the wild-type enzyme.
With the aim to discover and create suitable biocatalysts for the synthesis of chiral amines in a faster and more efficient way, this thesis includes protein engineering studies (Article I), explores transaminase substrate specificities (Articles II and IV), and an ultrahigh-throughput growth system-based for the directed evolution of amine-forming enzymes (Article III).
The protein engineering studies described in Article I deal with the creation of a (R)-amine transaminase activity in the α-amino acid transaminase scaffold to expand our knowledge of the evolutionary relationship between amine transaminase and α-amino acid transaminase. Article II describes the broadening of the limited substrate scope of transaminases to enable the conversion of bulky substrates. In Article III, a growth selection system is described for an ultra-high throughput screening strategy to accelerate the identification of desired mutants, which can be widely applied to the directed evolution of amine-forming enzymes.
Amine transaminases (ATAs) are powerful biocatalysts for the stereoselective synthesis of chiral amines. Machine learning provides a promising approach for protein engineering, but activity prediction models for ATAs remain elusive due to the difficulty of obtaining high-quality training data. Thus, we first created variants of the ATA from Ruegeria sp. (3FCR) with improved catalytic activity (up to 2000-fold) as well as reversed stereoselectivity by a structure-dependent rational design and collected a high-quality dataset in this process. Subsequently, we designed a modified one-hot code to describe steric and electronic effects of substrates and residues within ATAs. Finally, we built a gradient boosting regression tree predictor for catalytic activity and stereoselectivity, and applied this for the data-driven design of optimized variants which then showed improved activity (up to 3-fold compared to the best variants previously identified). We also demonstrated that the model can predict the catalytic activity for ATA variants of another origin by retraining with a small set of additional data.