Beate Pscheidt

Yeast cell factories for fine chemical and API production

This review gives an overview of different yeast strains and enzyme classes involved in yeast whole-cell biotransformations. A focus was put on the synthesis of compounds for fine chemical and API (= active pharmaceutical ingredient) production employing single or only few-step enzymatic reactions. Accounting for recent success stories in metabolic engineering, the construction and use of synthetic pathways was also highlighted. Examples from academia and industry and advances in the field of designed yeast strain construction demonstrate the broad significance of yeast whole-cell applications. In addition to Saccharomyces cerevisiae, alternative yeast whole-cell biocatalysts are discussed such as Candida sp., Cryptococcu s sp., Geotrichum sp., Issatchenkia sp., Kloeckera sp., Kluyveromyces sp., Pichia sp. (including Hansenula polymorpha = P. angusta), Rhodotorula sp., Rhodosporidium sp., alternative Saccharomyces sp., Schizosaccharomyces pombe, Torulopsis sp., Trichosporon sp., Trigonopsis variabilis, Yarrowia lipolytica and Zygosaccharomyces rouxii.

Laboratory evolved biocatalysts for stereoselective syntheses of substituted benzaldehyde cyanohydrins.

Hydroxynitrile lyases (HNLs) are biocatalysts employed industrially for the asymmetric addition of HCN to aldehydes or ketones. The resulting optically active cyanohydrins are important building blocks for pharmaceuticals and agrochemicals. Several genes of Rand S-selective HNLs have been cloned and expressed in suitable host systems, and can thus be ACHTUNGTRENNUNGengineered to improve their catalytic properties. The almond (Prunus amygdalus) (R)-HNL isoenzyme 5 (PaHNL5) is particularly suited as a starting point for engineering approaches. PaHNL5 can be heterologously expressed in Pichia pastoris and is highly stable even at pH values lower than 3.0. It is efficiently secreted into the culture supernatant, which can be directly employed for biocatalysis in aqueous or biphasic systems. Reduction of steric hindrance by structure-guided design has resulted in the simple and quick generation of new enzyme variants with significantly improved catalytic rates and enhanced stereoselectivity in cyanohydrin syntheses with employment of non-natural substrates. We wanted to explore whether the so far best, highly active, and stereoselective enzyme variant for the production of (R)-2-chloromandelonitrile (2a, Scheme 1), PaHNL5/L1Q/A111G, already industrially used, can reach an activity similar to that of the WT enzyme with its preferred substrate benzaldehyde. Compound 2a is a key intermediate for the production of a widely administered anticoagulant that reduces the risk of cardiovascular events in patients with acute coronary syndromes. As computational methods did not provide any indications for further improvement, we addressed the challenge of applying directed evolution. Bacteria, especially E. coli, are widely used for laboratory evolution, because of their simple genetic molecular manipulability, high transformation efficiency, and rapid growth rates. However, for eukaryotic proteins their use is often limited because of misfolding and the lack of typical eukaryotic posttranslational modifications. P. pastoris, an efficient system for heterologous protein production, can compensate for such disadvantages. Drawbacks, however, are a high input of linearized plasmid DNA and the concomitant relatively low integration rate into the genome. In addition, we have observed varying copy numbers or incomplete cassette integration, which seemed to impede the utilization of P. pastoris for highthroughput expression, screening, and laboratory evolution of proteins. Although expression of PaHNL in E. coli has been described, large amounts of highly active PaHNL5 could only be obtained with P. pastoris. Consequently, it was necessary to devise a new method for the efficient construction of reliable expression libraries in P. pastoris and the uniform expression of thousands of individual transformants, while circumventing time-consuming ligation and cloning steps in E. coli, which lead to a loss of diversity and library efficiency. Furthermore, there are no stable plasmids for Pichia transformation available. A concise procedure to generate PaHNL5/L1Q/A111G libraries by integration of linear expression cassettes produced by an overlap extension PCR (OE-PCR) strategy was thus designed, and so the randomly mutated Pahnl5 gene, a fragment consisting of a partial GAP (glyceraldehyde 3-phosphate dehydrogenase) promoter, a 5’ secretion signal sequence for the a-mating factor from bakers’ yeast and a 3’ zeocin resistance cassette were linked together (Figure 1). Both flanking arms were generated by proof-reading polymerases. To improve the efficiency of the OE-PCR, the overlapping region between the mutated gene and the selection marker (overlap 2) was redesigned. Then, without any prior ligation step, P. pastoris X33 was directly transformed with the linear PCR-based integration cassettes. In order to test this new strategy, the Pahnl5/L1Q gene was used to reassemble a linear integration cassette by overlap extension PCR. Since an L1Q mutation seemed to enhance exScheme 1. Synthesis of halogen-substituted (R)-mandelonitriles (TBME: tertbutyl methyl ether).

Engineering the Pichia pastoris methanol oxidation pathway for improved NADH regeneration during whole-cell biotransformation.

Industrial biocatalytic reduction processes require the efficient regeneration of reduced cofactors for the asymmetric reduction of prochiral compounds to chiral intermediates which are needed for the production of fine chemicals and drugs. Here, we present a new engineering strategy for improved NADH regeneration based on the Pichia pastoris methanol oxidation pathway. Studying the kinetic properties of alcohol oxidase (AOX), formaldehyde dehydrogenase (FLD) and formate dehydrogenase (FDH) and using the derived kinetic data for subsequent kinetic simulations of NADH formation rates led to the identification of FLD activity to constitute the main bottleneck for efficient NADH recycling via the methanol dissimilation pathway. The simulation results were confirmed constructing a recombinant P. pastoris strain overexpressing P. pastoris FLD and the highly active NADH-dependent butanediol dehydrogenase from S. cerevisiae. Employing the engineered strain, significantly improved butanediol production rates were achieved in whole-cell biotransformations.

Targeting Posttranslational Modifications – Perspectives for Biocatalyst Engineering

Engineering posttranslational modifications is of great importance in biomedical research. Recent advances in the field of biocatalyst engineering expanded the strategy of targeting posttranslational modifications to a novel research field, i.e. rendering the particular enzyme more efficient for the desired biocatalytic conversion. Thus enzymes could be trimmed to higher fitness for technical applications e.g. by improved activity and/or prolonged life-time. Furthermore, bottlenecks during enzyme expression caused by posttranslational processing in the cell can be overcome. Focusing especially on recent approaches in our laboratory, this article presents strategies for targeted manipulations of selected posttranslational modifications.