Exploring and expanding the synthetic potential of a biocatalytic dynamic kinetic resolution

Abstract

The market demand for enantiopure non-canonical amino acids (NCAA) has been increasing in recent years, given the importance of these compounds as chiral building blocks employed in the synthesis of many pharmaceuticals and agrochemicals. An example is the use of L-phosphinothricin (L-PT), a bioactive amino acid with potent herbicidal properties produced as a secondary metabolite by the Gram positive bacterium Streptomyces viridochromogenes. Recently, the use of NCAAs has been expanded to include applications in protein engineering, thanks to the advances achieved in the field of genetic code expansion. Over the last decade various biocatalytic strategies have been developed for the enantioselective production of these compounds, using a variety of enzymes, such as aminotransferases and ammonia lyases. A well-established biocatalytic protocol allows for the efficient preparation of NCAAs from their N-acylated derivatives via a dynamic kinetic resolution (DKR). This is achieved by coupling an engineered Amycolatopsis N-acyl amino acid racemase (NAAAR) with a compatible stereoselective deacetylase (amide hydrolase). Two recombinant hydrolytic enzymes: E. coli ArgE and S. viridochromogenes Tü494 Dea have been investigated in this work. Purified ArgE is a zinc-dependant, L-selective N-acetyl-L-ornithine deacetylase and a thorough kinetic study revealed that it displays a wide substrate scope but poor activity for acidic and aromatic substrates. Various tests were carried out to optimise biotransformation conditions, which were employed in the coupling of ArgE with the engineered NAAAR double mutant (NAAAR DM: G291D F323Y) for the DKR of N-acetyl NCAAs derivatives. Directed evolution was employed to increase deacetylase activity towards acidic N-acetylated amino acids; unfortunately, thus far no improved variant has been isolated. Prism-shaped protein crystals were obtained, but with poor resolution, which prevents the determination of the E. coli ArgE X-ray structure. Despite this, homology modelling and site-directed mutagenesis were successfully employed to probe the enzyme active and substrate binding sites. The deacetylase S. viridochromogenes Dea is a novel member of the hormone-sensitive lipase (HSL) family. It was proposed that this enzyme catalyses the hydrolysis of N-Ac-L-bialaphos, the precursor of the tripeptide bialaphos (PTT), a natural herbicide that contains the L-PT building block. Similar to other members of the HSL family SvDea catalyses the deacetylation reaction by employing a catalytic triad of highly conserved serine, histidine and glutamate residues in the active site, but the basis of its substrate specificity is unknown. To study its synthetic utility recombinant SvDea was isolated from E. coli and activity screening using various assays (including a novel 1H NMR assay) revealed a narrow substrate scope, limited to small acetylated esters and N-acetyl-L-PTT. Initial crystal trials failed to deliver positive hits, so structure and sequence analysis generated a homology model to obtain further insights into the substrate binding and direct future engineering efforts. In the last section of this thesis the synthetic scope of a novel NAAAR quadruple mutant (NAAAR QM: Q26I M50I G291D F323Y) was explored. This engineered racemase was found to display enhanced activity for bulky N-acetylated amino acids and could be coupled with either L- and D-selective acylases to prepare enantiopure target L- or D- amino acids. Following small scale screening, eight DKRs were set up at 1 g scale to highlight the biocatalytic potential of this versatile enzyme for the preparation of valuable, optically pure phenylalanine derivatives from racemic starting materials.