Preparative deracemization of unnatural amino acids.

Unnatural amino acids are a growing class of intermediates required for pharmaceuticals, agrochemicals and other industrial products. However, no single method has proven sufficiently versatile to prepare these compounds broadly at scale. To address this need, we have developed a general chemoenzymatic process to prepare enantiomerically pure L- and D-amino acids in high yield by deracemization of racemic starting materials. This method involves the concerted action of an enantioselective oxidase biocatalyst and a non-selective chemical reducing agent to effect the stereoinversion of one enantiomer and can result in an enantiomeric excess of >99% from the starting racemate, and product yields of over 90%. This approach compares very favourably with resolution processes, which have a maximum single-pass yield of 50%. We have developed efficient methods to adapt the process towards new target compounds and to optimize key factors that influence process efficiency and offer competitive economics at scale.

Methods for the synthesis of unnatural amino acids.

A wide variety of methods have been used to prepare unnatural amino acids. This review covers methods reported between 2000 and the early part of 2001. Some of the approaches discussed are applications of established methods, but emphasis has been given to new approaches that could be generally applicable to large-scale syntheses, including catalytic reactions such as the Strecker reaction and biological approaches.

The large scale synthesis of "unnatural" amino acids.

The introduction of a stereogenic centre to produce an "unnatural" amino acid can be accomplished in a variety of ways ranging from asymmetric hydrogenation to biotransformations based on transaminase enzymes. Our transaminase approach can be used to access a wide variety of L- and D-amino acids from an alpha-keto acid substrate. It is run as a whole cell biotransformation and uses coupled enzyme systems. In addition, formation of amino acids with small side chains, such as 2-aminobutyrate, can cause significant isolation problems due to the presence of small amounts of other amino acids, such as alanine. The improvements we have made to the approach are illustrated with 2-aminobutyrate as the example. Aspartic acid is used as the amino donor and gives rise to the formation of pyruvate, a substrate for the transaminase enzymes. We have now developed an alternative approach where lysine is used as the amino donor to allow formation of a cyclic by-product that is removed from the equilibrium.

Engineering biosynthetic pathways: new routes to chiral amino acids.

The engineering of microbial biosynthetic pathways is advancing rapidly because of new molecular genetic approaches, sophisticated analysis of metabolic flux and the rapid sequencing of diverse bacterial genomes. The classical methods of mutagenesis/selection, originally applied in the development of amino-acid-over-producing bacteria are now being complemented by an increasingly rational strategy.

Engineering of a novel biochemical pathway for the biosynthesis of L-2-aminobutyric acid in Escherichia coli K12.

L-2-Aminobutyric acid was synthesised in a transamination reaction from L-threonine and L-aspartic acid as substrates in a whole cell biotransformation using recombinant Escherichia coli K12. The cells contained the cloned genes tyrB, ilvA and alsS which respectively encode tyrosine aminotransferase of E. coli, threonine deaminase of E. coli and alpha-acetolactate synthase of B. subtilis 168. The 2-aminobutyric acid was produced by the action of the aminotransferase on 2-ketobutyrate and L-aspartate. The 2-ketobutyrate is generated in situ from L-threonine by the action of the deaminase, and the pyruvate by-product is eliminated by the acetolactate synthase. The concerted action of the three enzymes offers significant yield and purity advantages over the process using the transaminase alone with an eight to tenfold increase in the ratio of product to the major impurity.

Novel biosynthetic approaches to the production of unnatural amino acids using transaminases.

Transaminase enzymes are being increasingly applied to the large-scale synthesis of unnatural and nonproteinogenic amino acids. Typically displaying relaxed substrate specificity, rapid reaction rates and lacking the need for cofactor regeneration, they possess many characteristics that make them desirable as effective biocatalysts. By judiciously combining the transaminase reaction with additional enzymatic steps, this approach can be used very efficiently to prepare a broad range of D- and L-amino acids.

Characterization of the genes encoding D-amino acid transaminase and glutamate racemase, two D-glutamate biosynthetic enzymes of Bacillus sphaericus ATCC 10208.

In Bacillus sphaericus and other Bacillus spp., D-amino acid transaminase has been considered solely responsible for biosynthesis of D-glutamate, an essential component of cell wall peptidoglycan, in contrast to the glutamate racemase employed by many other bacteria. We report here the cloning of the dat gene encoding D-amino acid transaminase and the glr gene encoding a glutamate racemase from B. sphaericus ATCC 10208. The glr gene encodes a 28. 8-kDa protein with 40 to 50% sequence identity to the glutamate racemases of Lactobacillus, Pediococcus, and Staphylococcus species. The dat gene encodes a 31.4-kDa peptide with 67% primary sequence homology to the D-amino acid transaminase of the thermophilic Bacillus sp. strain YM1.

Nucleotide sequence of the Bacillus licheniformis ATCC 10716 dat gene and comparison of the predicted amino acid sequence with those of other bacterial species.

The gene encoding the D-aminotransferase from Bacillus licheniformis was cloned and the complete DNA sequence was determined. The deduced D-aminotransferase protein sequence, consists of 283 amino acids and shows a high degree of homology with other Bacillus D-aminotransferases, branched chain aminotransferase of Escherichia coli and the 4-amino-benzoate-4-deoxychorismate lyase of Bacillus subtilis and Escherichia coli.

Significant improvement to the catalytic properties of aspartate aminotransferase: role of hydrophobic and charged residues in the substrate binding pocket.

The substrate specificity of tyrosine aminotransferase (eTAT) from Escherichia coli has been tested by transferring the critically different residues Leu39, Glu141, and Arg293 into equivalent positions of aspartate aminotransferase (eAAT). These residues are not directly involved in the catalytic process. The single mutant eAAT V39L possesses greater values of kcat/KM not only for tyrosine but also for aspartate and glutamate. In contrast, the double mutant eAAT P141E,A293R and also the triple mutant eAAT V39L,P141E,A293R exhibit smaller changes of kcat/KM. The converse mutants of tyrosine aminotransferase, in which critical residues of eAAT (Val39) and of mitochondrial AAT (Ala39, Val37) were transferred into equivalent positions of eTAT, exhibited generally decreased values of kcat/KM for both dicarboxylic and aromatic substrates. On the basis of the known structures of eAAT and eAAT V39L as well as of a refined model of eTAT, these results indicate that the different substrate specificities of eAAT and eTAT are due to multiple side chain differences and minor rearrangements of the backbone. The generally improved catalytic efficiency of the mutant eAAT V39L appears to be due to an indirect effect, namely, the facilitated closure of the active site upon substrate binding.

Novel mutations in the pheA gene of Escherichia coli K-12 which result in highly feedback inhibition-resistant variants of chorismate mutase/prephenate dehydratase.

The bifunctional enzyme chorismate mutase/prephenate dehydratase (EC 5.4.99.5/4.2.1.51), which is encoded by the pheA gene of Escherichia coli K-12, is subject to strong feedback inhibition by L-phenylalanine. Inhibition of the prephenate dehydratase activity is almost complete at concentrations of L-phenylalanine greater than 1 mM. The pheA gene was cloned, and the promoter region was modified to enable constitutive expression of the gene on plasmid pJN302. As a preliminary to sequence analysis, a small DNA insertion at codon 338 of the pheA gene unexpectedly resulted in a partial loss of prephenate dehydratase feedback inhibition. Four other mutations in the pheA gene were identified following nitrous acid treatment of pJN302 and selection of E. coli transformants that were resistant to the toxic phenylalanine analog beta-2-thienylalanine. Each of the four mutations was located within codons 304 to 310 of the pheA gene and generated either a substitution or an in-frame deletion. The mutations led to activation of both enzymatic activities at low phenylalanine concentrations, and three of the resulting enzyme variants displayed almost complete resistance to feedback inhibition of prephenate dehydratase by phenylalanine concentrations up to 200 mM. In all four cases the mutations mapped in a region of the enzyme that has not been implicated previously in feedback inhibition sensitivity of the enzyme.

Crystallization and preliminary X-ray studies of an aspartate aminotransferase mutant from Escherichia coli.

Mutant aspartate aminotransferase V39L (Val39 replaced by Leu) from Escherichia coli has been crystallized into a monoclinic cell from a polyethylene glycol solution (pH 7.5) by vapor diffusion. The space group and the unit cell dimensions have been determined using a precession camera, a CAD4 diffractometer and a Nicolet Xentronics area detector to be P2(1) with a = 86.8 A, b = 79.9 A, c = 89.4 A, beta = 118.74 degrees. The crystals diffract to better than 2.3 A and are suitable for X-ray structure analysis.

The cloning and sequence analysis of the aspC and tyrB genes from Escherichia coli K12. Comparison of the primary structures of the aspartate aminotransferase and aromatic aminotransferase of E. coli with those of the pig aspartate aminotransferase isoenzymes.

In this paper we describe the cloning and sequence analysis of the tyrB and aspC genes from Escherichia coli K12, which encode the aromatic aminotransferase and aspartate aminotransferase respectively. The tyrB gene was isolated from a cosmid carrying the nearby dnaB gene, identified by its ability to complement a dnaB lesion. Deletion and linker insertion analysis located the tyrB gene to a 1.7-kilobase NruI-HindIII-digest fragment. Sequence analysis revealed a gene encoding a 43 000 Da polypeptide. The gene starts with a GTG codon and is closely followed by a structure resembling a rho independent terminator. The aspC gene was cloned by screening gene banks, prepared from a prototrophic E. coli K12 strain, for plasmids able to complement the aspC tyrB lesions in the aminotransferase-deficient strain HW225. Sub-cloning and deletion analysis located the aspC gene on a 1.8-kilobase HincII-StuI-digest fragment. Sequence analysis revealed the presence of a gene encoding a 43 000 Da protein, the sequence of which is identical with that previously obtained for the aspartate aminotransferase from E. coli B. Considerable overproduction of the two enzymes was demonstrated. We compared the deduced protein sequences with those of the pig mitochondrial and cytoplasmic aspartate aminotransferases. From the extensive homology observed we are able to propose that the two E. coli enzymes possess subunit structures, subunit interactions and coenzyme-binding and substrate-binding sites that are very similar both to each other and to those of the mammalian enzymes and therefore must also have very similar catalytic mechanisms. Comparison of the aspC and tyrB gene sequences reveals that they appear to have diverged as much as is possible within the constraints of functionality and codon usage.

Expression in Bacillus subtilis of the gene for human urogastrone using synthetic ribosome binding sites.

A chemically synthesised gene coding for human urogastrone which was earlier cloned in E. coli (Smith et al. 1982) has now been cloned into expression vectors for Bacillus subtilis. Two types of constructs have been made, one giving production of methionyl-urogastrone and the other giving rise to a methionyl-urogastrone-beta galactosidase fusion polypeptide facilitating quantification of expression levels. The ribosome binding sites used in the expression plasmids are synthetically made oligonucleotides residing on short restriction fragments to allow easy replacement by other ribosome binding sites. Using "shuttle" vectors and constitutive promoters from Bacillus phages phi 105 and SPP1, we were able to detect levels of expression amounting to a few thousand molecules per cell during logarithmic growth in both E. coli and B. subtilis.

Chemical synthesis and cloning of a gene for human beta-urogastrone.

A DNA duplex coding for the 53 amino acids of human beta-urogastrone has been synthesised. Computer assisted design of the gene included restriction endonuclease sites for plasmid insertion, a termination codon and two triplets coding for lysine at the 5'-end of the structural gene. The synthesis involved preparation of 23 oligodeoxyribonucleotides by phosphotriester procedures coupled to rapid HPLC techniques. The gene was constructed in two halves by enzymatic ligation of the oligonucleotides and cloned into a specially constructed chimeric plasmid vector. Escherichia coli K12 MRC8 was transformed by the plasmid and clones containing the full gene sequence were isolated and characterised.