Refine
Document Type
- Doctoral Thesis (3)
- Article (1)
Language
- English (4) (remove)
Has Fulltext
- yes (4)
Is part of the Bibliography
- no (4)
Keywords
- Transaminases (4) (remove)
Institute
Publisher
- Wiley (1)
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.
The synthesis of several bioactive compounds and active pharmaceutical ingredients relies on the development of general and efficient methods to prepare optically pure amines. Transaminases are industrially relevant enzymes and are useful for synthesizing a large number of compounds that contain a chiral amine functionality. Although the immense potential associated to the use of these biocatalysts, the equilibrium position is often unfavorable for amine synthesis. The use of an excess of amine donor, compared to the ketone substrate, combined with selective removal of the formed product, can help in overcoming this limitation. This work mainly focused on broadening the application of membrane-based in situ product recovery (ISPR) techniques for the transaminase-catalyzed synthesis of chiral amines. The
overall work was designed around the implementation of amine donors, possessing considerably larger molecular ‘size’ compared to commonly used amine donors. To clearly
distinguish these molecules from traditional donor amines, we designate them as High Molecular Weigh amine donors. With a molecular weight between 400 and 1500 g/mol, in contrast to traditional donor amines, HMW amine donors enable a size-based separation between amine donor and amine product molecules. HMW amines, provided in excess for thermodynamic equilibrium shifting can thus be simply retained by a size-exclusion mechanism by commercial membranes, while the smaller product amines are permeated. Therefore, a selective recovery of the desired chiral amine product is possible. The implementation of ISPR techniques using HMW amine donors can theoretically lead to (i) equilibrium shifting, (ii) alleviation of product inhibition, and (iii) a highly pure product stream.
The feasibility of using HMW amine donors in aqueous, organic solvent and solvent-free media for the transaminase-catalyzed synthesis of 1-methyl-3-phenylpropylamine (MPPA) was proven in this thesis. The latter two approaches were investigated with the aim to achieve higher product concentrations. Along with that, we demonstrated two membrane-assisted ISPR proof of concepts. Specifically, nanofiltration was coupled with the enzymatic reaction performed in aqueous media (Article I), while liquid-liquid (L-L) extraction in a contactor was applied for transamination in organic solvent media (Article II). As an alternative to membrane-based strategies we also designed a spinning reactor concept for the integrated chiral amine synthesis (in organic solvent) and recovery (Article III).
In 2010, the identification of 17 novel (R)-ATAs represented a breakthrough for the biocatalytic asymmetric synthesis of chiral amines, because only one (R)-ATA was described before. These novel ATAs were identified in a bioinformatic approach by studying the substrate acceptance of BCATs and DATAs to deduce the unknown substrate coordination of (R)-ATAs. Article I describes an alternative approach for the identification of (R)-ATA activity by reengineering the substrate- recognition site of α-AATs. While the engineering of the eBCAT led to the formation of an initial (R)-amine acceptance only, the (R)-ATA activity was successfully introduced in the DATA scaffold. These results demonstrate the transformation of an α-AAT in a moderately active (R)-ATA for the first time and highlight the evolutionary relationship between α-AATs and ATAs. Despite the availability of different ATAs nowadays, their substrate spectrum is limited due to the natural composition of their active sites. Several protein-engineering studies showed the widening of the substrate spectrum and the acceptance of bulky substrates by screening large mutant libraries to identify beneficial variants. In Article II, we developed an in silico engineering approach for amine transaminases to improve the conversion of bulky substrates and to reduce the number of variants to be tested in the laboratory. The resulting double-mutants of the (S)-ATA from C. violaceum displayed a >200-fold improved activity towards the bulky benchmark substrate. These variants expand the available biocatalytic toolbox for the synthesis of bulky amines, and the developed framework paves the way for rational protein-engineering protocols.
By studying unconventional transaminase substrates, we explored the potential of the available in- house transaminase toolbox in Articles III, IV, V, and VI. In Article III, we showed the transamination of a β-keto ester, leading to the synthesis of β-phenylalanine. The described cascade in Article IV enables the synthesis of amino carbohydrates. In addition, Article V describes an enzymatic cascade for the synthesis of amino fatty acids, which was extended in Article VI to obtain fatty amines.
The findings of this thesis clearly contribute to the understanding of the substrate scope and specificity of amine transaminases and expand the application of this versatile biocatalyst beyond classical ketone substrates.
Amine transaminases are versatile biocatalysts for the production of pharmaceutically and agrochemically relevant chiral amines. They represent an environmentally benign alternative to waste intensive transition metal catalysed synthesis strategies, especially because of their high stereoselectivity and robustness. Therefore, they have been frequently used in the (chemo)enzymatic synthesis of amines and/or became attractive targets for enzyme engineering especially in the last decade, mainly in order to enlarge their substrate scope. Certainly, one of the most notable examples of amine transaminase engineering is the
manufacturing of the anti-diabetic drug Sitagliptin in large scale after several rounds of protein engineering. Thereby, the target amine was produced in asymmetric synthesis mode which is the most convenient and favored route to a target chiral amine, starting from the corresponding ketone. The choice of the amine donor is highly relevant for reaction design in terms of economical and thermodynamic considerations. For instance, the use of alanine as the natural amine donor is one of the most common strategies for the amination of target ketones but needs the involvement of auxiliary enzymes to shift the reaction equilibrium towards product formation. In fact, isopropylamine is probably one of the most favored donor molecules since it is cheap and achiral but it is supposed to be accepted only by a limited number of amine transaminases.
This thesis focusses on the optimization and application of amine transaminases for asymmetric synthesis reactions en route to novel target chiral amines using isopropylamine as the preferred amine donor.