1. Biocatalysis to life science API’s
Enoate reductases
Enoate reductases (ERs) selectively reduce carbon-carbon double bonds in a/b-unsaturated carbonyl compounds and thus can be employed to prepare enantiomerically pure aldehydes, ketones, and esters (see Schem below). We cloned the genes and expressed three non-related ERs, two of them novel, in E. coli: XenA from Pseudomonas putida, KYE1 from Kluyveromyces lactis, and Yers-ER from Yersinia bercovieri. All three proteins showed broad ER specificity and broad temperature and pH optima but different specificity patterns. All three proteins prefer NADPH as cofactor over NADH and are stable up to 40oC.
By coupling Yers-ER with glucose dehydrogenase (GDH) to recycle NADP(H), conversion of >99% within one hour was obtained for the reduction of 2-cyclohexenone (see Scheme 1).
Scheme 1. Enzyme-coupled cofactor regeneration system with enoate reductase from Yersinia bercovieri and glucose dehydrogenase
Structure-guided consensus concept
Instability under non-native processing conditions, especially at elevated temperatures, is a major factor preventing widespread adoption of biocatalysts for industrial synthesis. Here we test the structure-guided consensus for the generation of a thermostable glucose dehydrogenase (GDH) from Bacillis subtilis (Vazquez-Figueroa, accepted). The consensus sequence in combination with additional knowledge-based criteria was used to select amino acids for substitutions. Using this approach we generated 24 variants, 11 of which showed higher thermal stability than the wild-type GDH, a success rate of 46%. Of the 24 variants, seven where located at the subunit interface—known to influence GDH stability—and six where more stable (86% success). The best variants feature a half-life of ~ 4 days at 65oC in contrast to ~ 20 minutes at 25oC for the wild type, enhancing stability 106-fold. In addition, the three most stabilizing single mutations were transferred to two GDH homologs from Bacillus thuringiensis and Bacillus licheniformis. The thermal stability as measured by half-life and CD222nm of the GDH variants was increased as expected. The resulting stability changes provides further support that these residues are critical for stability of GDHs and reinforces the success of the consensus approach for identifying stabilizing mutations.
The results described here show that it is possible to generate thermostable enzymes with minimal effort.
Finding relevant residues with Boolean learning/support vector machine techniques
A method for predicting the positions in the amino acid sequence that are critical for the fitness of a protein using Support Vector Machines (SVM) is introduced and analyzed. SVM are supported by an efficient learning algorithm and are flexible enough to take into account some prior knowledge about the structure of the problem in hand. A new kernel was formulated for this problem based on a defined feature space. The fitness of a protein is defined as its catalytic activity towards the substrate. The amino acid sequences of the variants of a protein, created by Directed Evolution protocols, along with their fitness are required as input data by this formulation to predict its critical positions. To investigate the performance of this algorithm, variants of the β-lactamase enzyme were created in silico using a mutagenesis and a recombination protocol. Results from literature on β-lactamase were used to test the accuracy of this method and the results were also compared with a simple search algorithm. The algorithm was shown to be able to predict critical positions that can tolerate up to two amino acids.