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.
The market for chiral raw materials and synthetic intermediates currently stands at $15 billion and is growing at 9.4% annually. The rapidly increasing industrial demand for these compounds is occurring principally to support the development of new, single-enantiomer pharmaceuticals and agrochemicals. In parallel, the use of biocatalysis for the industrial-scale manufacture of such chiral compounds is also increasing significantly . The high enantio- and regio-selectivity of biocatalysts is well suited to chiral organic synthesis and the continuing advances of methods in molecular biology, enzyme evolution, microbial genomics, strain engineering and bioinformatics are broadening the scope and availability of new biocatalysts . Nevertheless, the complexities of biocatalyst production and bioprocess optimization frequently pose significant hurdles for process implementation, particularly in the rapid time frames required by industry. It is therefore highly advantageous if a biocatalytic transformation or process can be applicable to a diverse range of compounds. Such a ‘platform’ approach not only reduces the development time for each new application but also mitigates against the loss of individual markets through the failure of specific products. Ingenza, an Edinburgh, U.K. based biocatalyst and bioprocess development company, is developing such a scalable platform bioprocess, which is adaptable to new targets within compound families, in order to develop a general route to unnatural amino acids and chiral amines at high optical purity. Unnatural amino acids now play an increasingly significant role in pharmaceutical development  and chiral amines are also of growing industrial importance and value in the fine chemical arena, in view of their application as resolving agents , chiral auxiliaries/bases  and catalysts for asymmetric synthesis . Ingenza's route to these compounds involves the deracemization of racemic mixtures by means of the stereoinversion of the undesired enantiomer through the concerted use of an oxidase biocatalyst and a chemical reducing catalyst, as described below. Recent efforts in the process development have been focused on amino acid deracemization due to the higher commercial priority of the specific target molecules. However, the methods now being optimized are equally applicable to the deracemization of chiral amines.
Background to the deracemization process
Since Ingenza was founded in 2002, the company has focused on the development of scalable industrial bioprocesses. The company has integrated expertise in molecular biology, high-throughput screening, fermentation, bioprocess development and synthetic chemistry to establish proprietary technology for the chemoenzymatic deracemization process. Currently, this process is applicable to over 100 amino acids or amines and can be customized towards specific targets of industrial interest using high-throughput screening, biocatalyst evolution and bioprocess optimization. The method employs the concerted use of a highly enantioselective oxidase biocatalyst and a non-enantioselective chemical reducing agent or catalyst (1). The imine, which is generated exclusively from one enantiomer of the target compound by the oxidase, can be converted in equal proportions into both enantiomers by the reductant. This reaction ultimately results in the near complete depletion of the oxidase substrate, with a concomitant increase in the opposite enantiomer. The process involves no substrate recycling and results in the stereoinversion of one enantiomer to yield the desired enantiomer, typically in >99% ee (enantiomeric excess) and with yields approaching 100%. The major advantages of the technology lie in the co-ordinated action of proven industrial oxidase biocatalysts and supported metal catalysts and efficient screening methods to adapt and optimize the approach towards new targets.
Deracemization of amino acids
Amino acid deracemization
Ingenza's deracemization process derives from earlier synthetic chemistry and biology in the laboratory of Professor Nicholas Turner at the School of Chemistry at Edinburgh University. Based on previous literature reports [7,8], the Turner group originally explored the deracemization of cyclic and acyclic amino acids using commercial DAAO (D-amino acid oxidase) from porcine kidney and the reducing agents sodium borohydride (NaBH4) and sodium cyanoborohydride (NaCNBH4) . The Turner group began a concerted programme of process improvement and achieved a 99% yield and 99% ee of L-proline from DL-proline  using porcine kidney DAAO and only three molar equivalents of NaCNBH4. Similar results were observed in the deracemization of DL-piperazine-2-carboxylic acid, which could be converted into the L-enantiomer in 86% yield and 99% ee using DAAO and NaCNBH4 . Further significant enhancements to the versatility and economic potential of the deracemization process followed in a collaboration between the Turner group and NSC Technologies, a unit of what was then Great Lakes Fine Chemicals, in 1999. In this collaboration, amine-boranes  and catalytic transfer hydrogenation using palladium/carbon  were introduced and proved more effective reducing agents in the deracemization of cyclic and acyclic amino acids. The use of supported metal catalysts in transfer hydrogenation to prepare amino acids by deracemization, which was introduced by Dr Scott Laneman at Great Lakes, has proven a particularly versatile and economical approach and has provided a basis for the current process development.
Recombinant bacterial strains producing cloned microbial DAAO or LAAO (L-amino oxidase) have also been introduced, facilitating the deracemization of DL-amino acids to yield either enantiomer and permitting molecular biology and directed evolution methodology to enhance and broaden the deracemization process towards natural and unnatural amino acids which are not favoured substrates for the native enzymes. The deracemization of DL-leucine to D-leucine was initially demonstrated using amine-boranes and cells expressing recombinant LAAO, with a 98% yield and 99% ee . Since 2003, when Ingenza began to commercialize amino acid deracemization, the process development has focused entirely on the use of catalytic transfer hydrogenation, because of the extensive knowledge base available for the industrial use, recovery and recycling of the metal catalysts and the highly favourable large-scale process economics offered by this approach. Using bioinformatics and screening approaches, Ingenza has also continued to clone and evolve novel amino acid oxidase genes to broaden the applicability of the process.
Production and laboratory evolution of oxidase biocatalysts
By coupling random gene mutation with a powerful and very high-throughput in vitro and in situ selection, highly process-suitable amino acid or amine oxidases can be evolved from the wild-type enzymes. The high enantioselectivity of the native enzymes is typically retained or improved following laboratory evolution. The screening procedure takes advantage of the fact that amino acid oxidase, like other members of the oxidase family, evolves hydrogen peroxide as a reaction by-product. The presence of peroxide (and therefore oxidase activity) can be detected colorimetrically by the addition of horseradish peroxidase and a substrate that yields a coloured product. This screen can be carried out directly upon individual bacterial colonies carrying randomly mutated oxidase variants. Ingenza has now optimized this approach to enable 105–106 isolates to be economically screened in a single experiment. As shown in Figure 1, the appearance of darker coloured colonies in an initial solid-phase screen indicates improved oxidase biocatalyst activity against the actual target substrate. Amino acid oxidase activity towards a range of targets is then characterized kinetically in a microtitre plate-based assay. Increases in activity, substrate range and biocatalyst stability can be achieved through multiple cycles of this process. In this way, oxidase biocatalysts are rapidly adapted towards commercial amino acid or amine targets of interest.
Engineering of process-suitable oxidase biocatalysts
Improvements in the overall yield of oxidase biocatalysts from fermentation of recombinant strains can also be readily detected using this screen. Such improvements are typically due to codon changes that improve the efficiency of transcription or translation of oxidase genes. Further rounds of directed evolution and the combination of individual beneficial mutations generate the industrially appropriate biocatalysts required for the process.
Deracemization process optimization
Successful industrial bioprocesses must meet aggressive cost targets and, for broad acceptance, should be sufficiently robust to be compatible with existing equipment and manufacturing practices. In order to establish deracemization as a competitive manufacturing route for unnatural amino acids, many process conditions required optimization in addition to the successful biocatalyst evolution described above. Ingenza therefore has concentrated on the optimization of the operating parameters for the process. Principally, these parameters include substrate and catalyst loading, reaction conditions such as temperature, pH, aeration and agitation, catalyst formulation (free or immobilized enzyme and reducing catalyst), catalyst recycling, process scale-up, and product recovery and purification. These process development activities are led by Dr Ian Archer, Head of Process Development. In one example, Ingenza is developing the deracemization process to prepare L-2-aminobutyric acid from inexpensive DL-2-aminobutyric acid starting material (2). Statistical design of the experimental procedures enabled a rapid screen of over 40 supported metal catalysts to be conducted with varying levels of ammonium formate and biocatalyst. These experiments established the ideal metal catalyst for the reaction as well as the appropriate balance of chemical and biocatalysts.
Deracemization of DL-2-aminobutyric acid using Pd/C and DAAO
For the concerted reaction with the metal catalyst, the biocatalyst is prepared by high cell density fed-batch fermentation. In small-scale reactions, the oxidase is provided as a whole-cell biocatalyst, whereas at larger scale, it is immobilized on suitable supports. At substrate loading of 100 mM, the L-2-aminobutyric acid is recovered at 99.9% ee with 95% isolated yield. Both the metal catalyst and biocatalyst can be recovered and reused multiple times. Through catalyst reuse and efficient reprocessing, the metal catalyst now contributes less than €10/kg to the process cost. This cost compares very favourably with the earlier deracemization process employing amine-boranes that contributed several hundred euros per kg of product. The use of DAAO, a well-established large-scale industrial biocatalyst, also provides highly favourable economics for the biocatalyst usage. The process development group has recently established conditions for the operation of this process at 500 mM substrate loading. Additional processes, including the preparation of L-pipecolic acid and several D-amino acids, are following similar development strategies.
Ingenza continues to refine the enabling technologies required for deracemization including recombinant production of oxidases, high-density fermentation and biocatalyst immobilization. Through participating in the Sixth European Framework Programme: Datagenom, novel oxidase biocatalysts are being obtained from genomic sources, including extremophiles, to further expand the application and robustness of oxidase-based deracemization. In addition, a one-step oxidase purification/immobilization, using a novel silica-based support (BioTrap) co-developed by Ingenza, offers highly economical production and reuse of the biocatalyst. Ingenza has operated amino acid deracemization at up to 10 litres scale in-house and with substrate concentrations up to 20% and is currently scaling amine deracemization to the same level. In-house fermentations are conducted at up to 15 litres scale and the company works closely with customers and manufacturing partners for further process scale-up. The company has entered development and technology licensing agreements with commercial partners in Europe, Japan and the U.S. to deliver deracemization processes for target amino acids and amines and continues to identify specific compounds for in-house development.
Many of the enabling technologies used in the deracemization process, such as biocatalyst discovery, fermentation, directed evolution and enzyme immobilization, are applied in other bioprocesses. For example, strain construction and microbial pathway engineering are applied in contract biocatalysis and bioprocess services. Ingenza has established alliances with partners to develop new bioprocesses aimed at the production of glycochemicals, nutraceuticals and secondary metabolites, in addition to deracemization targets, and is actively seeking new alliances and opportunities to implement biocatalytic processes.
Biocatalysis offers an increasingly realistic option for the manufacture of new chiral molecules at large scale. Regulatory directives from the U.S. Food and Drug Administration are increasing the number of chiral molecules required by the pharmaceutical industry, where single-isomer pharmaceuticals offer lower dosage and reduced side effects. Similarly, demands for single-isomer agrochemicals are increasing, due to environmental pressures. In addition to these drivers, the timing for new biocatalytic applications for industry is highly appropriate from a technology standpoint. Advances in molecular biology, bioinformatics and microbial genomics have enabled access to genetic material encoding an enormous array of novel enzymes, as the raw material for new biocatalysts. Methods of enzyme evolution, including increasingly diverse screens and efficient gene mutation and shuffling, offer increasingly rapid improvement of native enzymes. However, a rapid development time frame is also critical for biocatalysis to compete with more established methods. Accordingly, the development of adaptable platform bioprocesses such as deracemization is important in order to enhance the acceptance and growth of biocatalysis in industrial applications.
Biocatalysis: Enzymes, Mechanisms and Bioprocesses: Biochemical Society Focused Meeting in association with Pro-Bio Faraday Annual Conference held at Manchester Conference Centre, Manchester, U.K., 21–22 November 2005. Organized and edited by N. Bruce and G. Grogan (York, U.K.).