Community Research and Development Information Service - CORDIS

Final Report Summary - BIOOX (Developing a validated technology platform for the application of oxygen dependent enzymes in synthesis and transformation of alcohols)

Executive Summary:
Oxidations are amongst the most important reactions in industrial chemistry, used in the production of fine and bulk chemicals, many consumer products, and the treatment of waste streams. However, chemical oxidations are both hazardous and have high environmental impacts. The implementation of biocatalytic oxidations in European chemical manufacturing has the potential to transform industrial processes by improving costs and efficiency, whilst minimising hazards and environmental impacts.

The wealth of academic and industrial experience assembled in the consortium has enabled the development of cutting edge technologies addressing biocatalytic oxidations required by the chemical manufacturing sector. This includes the discovery and characterisation of novel biocatalysts, some representing previously unknown classes of enzymes, along with technologies to enable more rapid discovery and commercialisation of oxidative enzymes. In addition to established examples these form a broad enzymatic toolkit further developed, engineered, and optimised in BIOOX for application in industrial biotransformations. The research papers in preparation and already published will continue to enrich the field, providing new directions for future R&D efforts. Key examples of oxygen-dependent enzymes are now available for industrial screening and evaluation, in some cases achieving preliminary scale up and process development as part of BIOOX.

The four year research programme undertaken in BIOOX has delivered substantial progress beyond the state of the art, delivering technologies for applications in diverse markets, including chemicals and intermediates, biopolymers, consumer products, and flavours and fragrances. The technologies developed by the BIOOX consortium will achieve this whilst also delivering consumer benefits through quality, safety, and innovative products.

Project Context and Objectives:
BIOOX brings together a consortium of academic and industrial excellence to address the challenges of developing new technologies to fundamentally change chemical manufacturing in Europe and beyond. This collaborative project drew together 11 partners based in 6 European countries from academia, innovative SMEs, and major industrial partners supplying chemical markets worldwide. BIOOX supported technology development from enzyme discovery and optimisation, through process development and scale-up, to product evaluation. BIOOX aimed to develop and validate a technology platform providing industrial biocatalysts for transformation of industrially relevant compounds in viable process applications.

BIOOX would build on a wealth of European knowledge and skills to develop a set of tools that are validated and demonstrated by leading industrialists as an oxidation platform for applying Oxygen Dependent Enzymes (ODEs) in alcohol synthesis and oxidation. Biocatalytic oxidations are well studied in academic laboratories but few have been applied on an industrial scale. Due to inherent difficulties associated with large scale chemical oxidation methods, robust ODEs have great potential for application in industrial production of both bulk and high-value chemicals.

Oxidations are amongst the most important reactions in industrial chemistry, used in the production of fine and bulk chemicals, consumer products, and treatment of waste streams. However, chemical oxidations are both hazardous and have high environmental impacts. Industrial chemical oxidants are often transition metal-based systems used with chlorinated solvents and sometimes with significant thermal inputs or outputs. On chemical plants cooling can often be more expensive than heating whilst exothermic oxidants present significant hazards. The “greener” peroxide and chlorine based oxidants are hazardous to make, store, transport and use, which creates environmental impact by adding an extra reagent synthesis step that is both energy expensive and adds energy to the system that frequently has to be removed at next step. Also these reagents generate Cl+ or Br+ (peroxide generates these electrophiles in the presence of halide) and these can react as electrophiles with reactive aromatic rings (for instance, phenols and indoles) in the environment or in the organism to produce toxic, environmentally recalcitrant products. Even “benign” oxidants often involve precious or toxic metal catalysts

In contrast, biooxidations are characterised by relatively benign reaction conditions and exquisite selectivity, often achieving transformations of sustainable, cheap, and accessible precursors which are difficult or impossible via traditional chemistry. Europe has leading academic expertise in the area of industrial biotechnology (IB), and specifically in biocatalytic oxidations. In order to lead the future chemical industry in the direction of using IB to drive economic and sustainable production in global competition we need to continually lead the way in innovation. The aim of BIOOX was to build upon the strong knowledge base and foster strong and lasting research and development collaborations with industry, because although many different potential bioprocesses have been presented in academic literature, few have achieved commercial success. This represents unfulfilled potential where Europe has the opportunity to lead the world.

BIOOX was designed to deliver that underutilised potential, since industrial biocatalysis had previously remained primarily the last technology of choice for niche, high-value products with process methods for single products specifically implemented to address economics. Furthermore, historically the development of an optimized bioprocess took an unacceptably long time, while there also remained great uncertainties with respect to the cost of the final processes because of a lack of data on the different contributing factors.
In the fine chemical manufacturing industries the scope of previously applied biocatalytic methods were mainly limited to use of single-step reactions of hydrolase/lipase/amidase/esterase type enzymes. For high value products, typically pharmaceutical ingredients and intermediates, one-step redox reactions with cofactor recycling of a simpler sacrificial substrate have also become well established in industry. These are exemplified by use of enzymes such as ketoreductases for chiral secondary alcohol synthesis, for which IB has become the method of choice, and chiral amine synthesis using ω-transaminases. The latter of which as a method has limitations of both restricted amine donor choices, and in some cases unfavourable equilibrium. In both cases, these methods are limited to very few high value products. Although large economic advantages over traditional chemical routes have been obtained for these niche methods they are still relatively inefficient processes due to cofactor issues, and are essentially single step, single product processes developed at the later stages of the drug development program to reduce cost of goods.

The major objectives of BIOOX were:

1) Computational searching of “metagenomic” databases and screening of culture collections to identify the natural diversity of ODEs for biocatalysis with potential to carry out target transformations.
This aspect gained additional importance after the introduction of new legal frameworks for Access and Benefit Sharing (ABS) by the EU and member states governing the research and commercial use of “genetic resources”, i.e. the Nagoya Protocol on Access to Genetic Resources and the Fair and Equitable Sharing of Benefits Arising from their Utilization to the Convention on Biological Diversity and EU Regulation 511/2014. These frameworks aim to ensure fair and equitable sharing of benefits arising from the utilisation of genetic resources, thereby contributing to the conservation and sustainable use of genetic resources (such as the genes encoding enzymes). Development of metagenomes sampled from local environments rather than third party nations not only ensures legal compliance, but also provides the opportunity to guarantee freedom to operate by selecting sequences entirely unknown to public databases. Nature offers a huge diversity of sequences and the ability to affectively screen large numbers of potential enzymes targeted to the desired reactions was also therefore key.

2) Use of a range of enzymology techniques to develop the specificity, activity and robustness of the specific biocatalysts identified in these searches.
Although the enzyme activities targeted for application in BIOOX were already known, the need remained to improve specific aspects of performance so that they were fit for purpose in industrial application. In some cases, significant activity enhancements could be achieved through selection of alternative wild-type enzymes. Typically, this was as a precursor to additional protein engineering techniques to address specific parameters requiring improvement.

3) Use of multi-objective screening protocols to develop novel, improved ODEs to match process conditions to enzyme function.
A great deal of the body of academic literature on biocatalysis has focussed on improving the intrinsic catalytic properties, such as the turnover number or substrate binding constants, of the enzyme. However, industrial application requires that only is the biocatalyst but also that further properties are matched to process conditions, including stability, optimum operating temperature, along with catalytic efficiency.

4) Process modelling and design of experiments to optimize process performance and experimental protocols through optimal experimental design.
Importantly, process considerations should be incorporated into experimental design from the earliest stage, guiding decision making throughout the experimental process. This will ensure that fitness for purpose is addressed at the earliest opportunity, and also that challenges are identified and addressed at the appropriate stage of technology development.

5) Development of high throughput screening methods for all key steps in the process of identification of improvements in the full process development chain.
Previously, application of biocatalysis by industry has in part been limited by the time taken to develop novel bioprocesses versus development of competing tradition chemical routes. As such, the development of tools for rapid screening not only of novel and improved biocatalysts, but also of reaction and process conditions was critical as a general measure to reduce barriers to industrial uptake of biocatalysis.

6) Developing cost effective immobilization procedures for increased biocatalyst productivity and economic improvement.
One approach to economic application of ODEs is to prepare supported formulations of the biocatalysts that are matched to reactor configuration and enhanced practical downstream processing. Poor operational stability is a major limitation in many biocatalytic processes through enzyme inactivation mechanisms. Catalyst immobilization offers some advantages compared to the use of free catalysts, most notably good catalyst/product separation, easy operation, the possibility of reuse and enhanced stability of biological catalysts.

7) To investigate and analyse cost and effectiveness of reactor configurations for oxidations including flow reactors and optimum oxygen addition techniques.
Industrially, existing equipment has traditionally been used for IB processes due to infrastructure costs, and is often limited to packed bed reactors or fed-batch stirred tanks. These reactor configurations have been successful for virtually all IB products developed so far but in a significant number of companies using existing equipment leads to a limited uptake of IB, since many processes are unsuitable for such deployment in such configurations.

8) Development of cost effective fermentation and DSP protocols for enzyme fermentation and target hydroxylation reactions.
The cost of the whole process is often very dependent on the efficient production of the biocatalyst. An efficient production system is therefore key for the implementation of synthetic processes at industrial scale. Besides high titres it is especially important for industrial enzymes to have facile downstream-processes (DSP) and stable formulations to enable cost efficient biocatalysts.

9) To conduct economic value and environmental sustainability evaluations to quantify process improvements and maximize impacts of BIOOX.
While cost is the primary driver of chemical manufacturing processes it is also necessary to examine the environmental credentials of such processes given the original motivation for increased IB uptake.

10) Process validation on a larger scale of the technology developed in BIOOX.
The BIOOX project was designed with the target of developing an industry driven, validated technology platform, and therefore demonstration activities that bridge the gap between laboratory and industrial scale, and to prove the techno-economic viability of the targeted biotransformation were essential.

The project was split into 7 research Work Packages (WPs) in order to provide the necessary focus to address the challenges: WPs 1-4 focus on development of suitable biocatalysts; WPs 5-7 on the associated process technologies and metrics. WP8 and 9 were dedicated to dissemination to the wider civil and technical communities and to project management, respectively.

Work Package 1 focussed on determining target reactions; identifying new enzymes through bioinformatics analysis of databases and metagenomes; screening and characterizing ODEs for application in oxidation reactions; creating engineered strains to analyse the application of model enzymes for in vivo production.

Work Package 2 covers enzyme engineering for improvement and optimization of activity towards target substrates: 1) development of bioinformatics tools to identify mutation hotspots; 2) selection of target substrates for oxidations; 3) creation and screening of ODE variant libraries; 4) creation of engineered strains for substrate and enzyme production.

Work Package 3 is for development and delivery of whole cell biocatalysts for syntheses of hydroxylated fatty acids, alkenes and oxyfunctionalised terpenes: 1) development of robust and selective oxidising enzymes in whole cells; 2) fermentation strains suitable for application; 3) optimization of fermentation expression and production conditions.

Work Package 4 will develop immobilized enzymes and enzyme formulations with improved stability, applicability and recyclability for ODEs from WP1-3. To develop economical and environmentally-friendly scale up, WP4 will input to WP5 and 6.

Work Package 5 is focussed on evaluation of reactor configurations and oxygen feeding strategies, and how these influence stability and kinetics of ODEs. Evaluation will be based on isolated enzymes, immobilized enzymes, and whole-cell biocatalysts.

Work Package 6 is for scalable demonstration of technologies developing in other WPs, including production of demonstration scale batches of formulated enzymes, and testing of BIOOX synthesis and transformation technologies at relevant scales.

Work Package 7 focussed on economic and environmental process evaluation, developing economic and environmental models for evaluation of oxidase bioprocesses.

BIOOX incorporated a fully integrated project structure in order to ensure that the process requirements are considered from the earliest stage of the project. This integration enabled efficient translation of research expertise into scalable process development, with outputs transferred directly to industrial partners for evaluation.

Project Results:
Work Package 1
BIOOX incorporated a fully integrated project structure in order to ensure that the process requirements are considered from the earliest stage of the project. This integration enabled efficient translation of research expertise into scalable process development, with outputs transferred directly to industrial partners for evaluation. A key early output from WP1 was a Vision for Success, defining performance criteria for enzymes, productivity, yields, and reactor types for the selected oxidations studied in BIOOX. Progress beyond the intial state of the art was the result of an iterative process of planning and evaluation of research throughout the project, guided by the Vision for Success.

Bio-Prodict BV (hereafter BPT) developed a P450 protein superfamily information system at the start of the project and granted access to the partners. Called 3DM, this system collects, combines, and integrates many forms of protein-related data in order to facilitate the exploration of structure-function relations. The use of 3DM in BIOOX helped to guide enzyme discovery and engineering efforts by the other partners in WP1-3, and feedback from partners also facilitated the development of new 3DM tools. All interested partners were trained in a three-day 3DM system course at BPT’s site. An update of the 3DM system for cytochrome P450 enzymes was made available in 2016 at the request of project partners, providing a rich resource for the design of discovery and engineering strategies for this important target enzyme family.

Furthermore, BPT has developed a prototype patent landscape analysis system that can scan and integrate patent data for a complete protein superfamily. A web-tool was made that can be used by the project partners to analyse the patent landscape of the P450 3DM system, allowing evaluation of patented modifications in BIOOX systems where appropriate, but also enabling 3DM users to select novel enzymes with freedom to operate (FTO).
New enzyme discovery by bioinformatics approaches and studies of in-house enzymes, have enabled allowed screening of hundreds of ODEs for activity towards substrates identified by the industrial partners, as well as the auxiliary enzymes for cofactor recycling. New metagenomic technologies have allowed identification and production of large panels of completely novel enzymes, including new classes of activity, with freedom to operate.

The University of Manchester (UNIMAN) used a 3DM database of alcohol oxidases to discover enzymes suitable for the oxidation of primary alcohol targets provided by Firmenich SA (FIR) with applications in the flavours and fragrances sector. One enzyme in particular showed activity towards many substrates and was chosen for further development. High-throughput screening of enzyme variant libraries targeted substitutions to the active site and in flexible loop regions. This work yielded a six-point variant of choline oxidase from Arthrobacter chlorophenolicus with improved activity (20 fold increase in kcat from the wild type) towards the oxidation of hexanol. 100mg of hexanol was oxidised to aldehyde with negligible production of the acid over-oxidation production, and was isolated with a yield of 72%. The enzyme was further improved in its tolerance for organic solvents and activity at higher temperatures. In a screen of 50 primary alcohols, this variant showed improved activity for 90% of the substrates compared to the wild type enzyme, thus yielding an enzyme that could be applied as a more general alcohol oxidase.

Prozomix Ltd (PROZO) used proprietary technologies to survey sequence space to achieve maximum diversity in the selection, cloning, and produce enzymes, rapidly creating panels of relevant target enzymes to supply the consortium. Panels were developed for three key enzyme activities required in BIOOX; NAD(P)H oxidases (NOXs), alcohol oxidases (KREDs) and cytochrome P450s (P450s). PROZO developed in-house metagenomic libraries as a rich genetic resource allowing discovery of enzymes entirely unknown in public database. This aspect gained additional importance after the introduction of new legal frameworks for Access and Benefit Sharing (ABS) by the EU and member states governing the research and commercial use of “genetic resources”, i.e. the Nagoya Protocol on Access to Genetic Resources and the Fair and Equitable Sharing of Benefits Arising from their Utilization to the Convention on Biological Diversity and EU Regulation 511/2014. These frameworks aim to ensure fair and equitable sharing of benefits arising from the utilisation of genetic resources, thereby contributing to the conservation and sustainable use of genetic resources (such as the genes encoding enzymes). Development of metagenomes sampled from local environments rather than third party nations not only ensured legal compliance, but is also consistent with the principles of responsible research and innovation. RRI is an increasing prominent aspect of research and development, and embodiment of those principles in the technology developed through the project serves both to raise awareness of the BIOOX early career scientists and distinguishes the technologies in the marketplace. The metagenomic approach to enzyme discovery also created the opportunity to ensure freedom to operate for all BIOOX enzymes by avoiding patented sequence space, with the additional commercial advantage to PROZO of avoiding known sequences commonly duplicated across the commercial enzyme panels of competitors.

A total of 698 putative KRED targets were identified from PROZO’s in-house strain collection. Over 180 of these were rapidly cloned by PROZO’s proprietary GRASP™ cloning protocol and of these, 145 yielded soluble protein. The first 96 enzymes were arrayed in kREDy-to-go™ format, a flagship colorimetric screening plate system which allows rapid and facile identification of enzymes with activity towards test substrates. Further enzymes were discovered to complete PROZO’s second such plate, i.e. the kREDy-to-go™ plate 2. UNIMAN screened kREDy-to-go™ plates 1 and 2 against a panel of 21 primary alcohols from FIR and identified hits for all targets. In particular, KRED(020) and KRED(150) were selected for further study.
A total of 29 highly diverse putative NOXs were targeted for development as part of WP1. These targets were distant homologues of known L. brevis (NADH specific) and L. sanfranciscensis (dual cofactor specific) enzymes. The former of these was the subject of a then current, but now lapsed, patent (US2005064570). However, it should be noted that the patent only claimed primary sequence space of >80 % identity to the L. brevis enzyme. Careful selection of the panel using bioinformatics tools meant that the closest expressed and active homologue developed (NOX 2) was at 62% identity, and all four novel NOXs of this type therefore had FTO in any case. Six constructs containing genes from a known nitroreductase family yielded five soluble novel NOX recombinant proteins, one of which - NOX 25 - exhibited significant NOX activity and all of which demonstrated strong preference for NADH as cofactor (Table 2). All enzymes were produced by PROZO and supplied to the rest of consortium on demand, with a number of enzymes explored as lead candidates in combination with KRED(020) and KRED(150) for the oxidation of hexanol by the Technical University of Denmark (DTU) and UNIMAN.

P450s available from PROZO’s in-house metagenomes cover a number of architectures, including Class I, VII and VIII, and were used to further develop panels of greater interest right to the end of the project. The number of targets that were available for harvesting was large and covered huge diversity. Interestingly, of those studied in greatest detail as part of BIOOX, the Class VII and Class VIII P450s form distinct clades, demonstrating that there is much more work to be done with this resource beyond the end of the project.

The Class VII enzymes were pursued with the greatest urgency, since the literature suggested activities of significant commercial interest both within and outside the consortium, including towards compounds such as diclofenac (a nonsteroidal anti-inflammatory drug) and 7-ethoxycoumarin (a fragrance molecule). Of 14 metagenomic Class VII members cloned, four produced soluble enzymes which were supplied to partner UNIMAN; CYP116B59, CYP116B60, CYP116B1 and CYP116B1_ortholog. Preliminary data also demonstrated that each soluble Class VII candidate had some degree of activity towards either diclofenac or 7-ethoxycoumarin. A panel of Class VIII enzymes were supplied to partners UNIMAN and the University of Stuttgart (USTUTT) and after screening, a number of lead enzymes were then utilised in further deliverables and transferred to FIR for further application development and evaluation in fermentation processes.
Class I P450 enzymes were further studied, collecting data for use in combination with non-native ferredoxin and ferredoxin reductase auxiliary partners. PROZO demonstrated that there is no necessity to find native electron transfer protein pairs to activate Class I P450 enzymes. Indeed, functional assays demonstrated no fewer than five non-native ferredoxin/ferredoxin reductase pairs to be more active than the most utilised system in the literature, that of P450cam; i.e. capable of transferring more electrons more efficiently. Further data was collected from an expanded panel of ferredoxin reductases which were assayed against a panel of approximately 25 ferredoxins. Table 3 outlines each of the reductases which are now part of the PROZO panel which can be screened for simple identification of auxiliary partners for use in P450 systems.

Ultimately, PROZO was to develop a novel panel of 48 catalytically self-sufficient Class VII and VIII P450s, each with known activity. Catalytically self-sufficient P450s are natural (or engineered) fusion proteins where the CYP and redox domains exist as a single protein. These are attractive systems for application because they do not require the auxiliary partner proteins essential for the activity of other P450 classes. Use of self-sufficient P450s simplifies development and application, and can also help to reduce the cost of the biocatalyst. Moreover, members of these classes have been identified with activity suitable for key BIOOX target biotransformations, thus demonstrating the biocatalytic utility of the panel. Subsequent to this successful evaluation by the consortium, the novel panel was produced in prototype 96-well commercial (freeze-dried) format and made available for screening by potential customers, thus significantly contributing to the PROZO Biocatalysis Enzyme Toolkit. Continued development beyond the project will see a final panel of 96 self-sufficient P450s fully commercialised in 2018, enabling new biocatalytic syntheses and drug metabolism studies, where P450s possess crucial utility. The new panel will be signposted in an imminent publication reporting the large networks of novel self-sufficient P450s discovered and annotated by PROZO as part of WP1.

USTUTT found approximately 20 putative monooxygenases by shotgun sequencing of two arene-degrading strains (Arthobacter sp. and Phenylobacterium immobile). While P. immobile did not grow on sesquiterpene substrates, Arthrobacter sp. was able to grow using the target molecules bisabolene and premnaspirodiene as sole carbon sources. This behaviour is typical of strains producing P450s with activity towards the target as part of the metabolitic pathway of the orgsanism. In addition to Arthrobacter sp., Beauveria sp. was investigated by proteomic analysis, since it was able to convert sesquiterpene substrates in initial investigations. Seven putative CYPs were identified along with one bifunctional reductase as promising candidates for further studies. The proteomic analysis data as well as the corresponding P450-sequences were shared with FIR and BiCT srl (BiCT) to support work towards functional enzyme production in Saccharomyces cerevisiae and Aspergillus oryzae respectively.

To further increase the panel of potential ODEs the organism Chondromyces apiculatus was investigated, which is closely related to Sorangium cellulosum a strain known for P450s active towards sesquiterpenes,. Putative P450 genes were cloned and the enzymes produced for activity screening. Hydroxylation activities were detected for a number of enzymes: a putative steroid hydroxylase CYP125E; a member of the CYP264 family; and a member of the CYP109 family. The redox partners from P. putida (Cam A and Cam B) were determined to be the most efficient electron transport system. The P450 from the CYP109C family demonstrated broad substrate scope, ranging from mono- and sesquiterpenes, to nonsteroidal anti-inflammatory drugs like diclofenac, and steroids.
In the course of this project, FIR has collaborated with USTUTT, PROZO and UNIMAN on the discovery and development of P450 monoxygenases capable of transforming selected terpene compounds to produce flavor and fragrance molecules. Candidate P450 monooxygenases were screened and the positive hits functional characterised by USTUTT and UNIMAN, prior to enzyme engineering to develop activities towards the prototypical terpene hydrocarbons produced by FIR using submerged fermentation of engineered bacterial cells. FIR produced and transferred to the academic partners approximately 10g of each of the sesquiterpene substrates to enable this initial P450 screening.

Work Package 2
2,5-diformylfuran (DFF) is an attractive intermediate with potential applications in a variety of polymer formulations as a bio-based alternative to petroleum based plastics and other materials. The oxidation of hydroxymethylfurfural (HMF) to produce DFF was a high priority target identified by BASF during the development of the Vision for Success. Engineering galactose oxidase (GOase) for production of DFF from 5- HMF has been a primary target for UNIMAN within BIOOX. UNIMAN developed new GOase variants from two separate engineering experiments that showed improved conversion in biotransformations at high substrate loading 100mg/mL (~800mM) semi-crude HMF. One screen targeting active site residues was performed at reduced oxygen concentration to identify variants with greater oxygen reactivity. The other library targeted non-active site residues and was performed against low activity substrates to identify variants with generically improved alcohol oxidation activity. A generation of variants created through combining the unique mutations in these top hits has resulted in further productivity improvements at this high substrate loading, with the best yield to date of 76% conversion to DFF within 6 hours.

Continued optimisation of the reaction for oxidation of HMF by GOase G2 variants has identified industrially applicable solvents that are more suitable for application in biphasic reactions. Equally important, some of these new solvents lead to enhanced production of DFF in biotransformations scaled up to 100mg/ml of crude HMF substrate loading. Extensive GOase reaction optimisation and process development performed during BIOOX have resulted in joint patent applications by BASF SE (BASF) with UNIMAN and DTU currently under examination.

Further characterization by UNIMAN of the aldehyde oxidase activity of M3-5 towards a panel of substrates identified trends in the characteristics of substrates that can be oxidized by the enzyme. Notably, several of the nitrogen-containing heteroaromatic alcohols could be sequentially oxidized twice to produce carboxylic acids via the aldehydes in high conversions. This unique activity has potential applications in production of heteroaromatic carboxylic acids, which are often used in the synthesis of pharmaceuticals.

USTUTT developed enzymes identified for the hydroxylation of BIOOX target substrates including sesquiterpenes, norisoprenoids, and fatty acids. A number of examples from various enzyme classes, including Rieske non-heme iron dioxygenases (RO) and several P450 monooxygenases from the USTUTT in-house library, along with those discovered in WP1, were chosen as biocatalysts. Additionally, seventeen natural P450 fusion enzymes were provided as lyophilized Cell Free Extract (CFE) from PROZO for the screening of hydroxylation activity.

USTUTT pursued engineering of one enzyme provided PROZO, generating enzyme variant libraries with active site substitutions which were screened for improved activity toward BIOOX targets. The most active new variant showed 18-fold higher activity toward regioselective hydroxylation of β-ionone than the wild type enzyme, and others from the full panel of active enzymes provided access to new regioselectivities for hydroxylation of several sesquiterpene substrates. Similarly, UNIMAN has made and screened libraries of PROZO enzyme variants, resulting in an enzyme with 16-fold higher activity and greater product specificity in hydroxylation of β-bisabolene. Reaction optimisation through addition of an immiscible cosolvent further improved performance and reaction efficiency of this enzyme.

Through enzyme engineering, USTUTT generated a panel of variants of an RO, naphthalene dioxygenase, with expanded substrate scope for mono- and di-hydroxylation of arene substrates. A P450 variant with increased activity for C12 fatty acid ω-hydroxylation was also engineered. Omega hydroxylation of fatty acids is an important reaction both in the study of physiological processes, but also in industrial synthesis. Alongside this several variants of a novel P450 enzyme were created with significantly improved conversion and entirely new product profiles for production of hydroxylated sesquiterpenes and β-ionone. The P450 variants developed within BIOOX have enabled production of both target and unanticipated, novel flavour and fragrance compounds at greater titres than were previously accessible.

Critical to the industrial application of biocatalysis is the ability to produce enzymes in a choice of host organisms that provide sufficient activity with suitable handling characteristics of the production organism. Additionally, freedom to operate is key requirement in order to avoid costly licencing of production vectors and host strains. BiCT constructed a new class of modular expression plasmids for the FTO production of enzymes in industrial strains of Aspergillus fungus species. These Aspergillus sp./E. coli shuttle plasmids enable facile integrative transformation of different strains of Aspergillus. Transformation protocols were defined for a number of industrially relevant species, including A. oryzae, A. niger and A. niger var. awamorii. Aspergillus expression vectors were assembled employing modular design in order to allow the rapid replacement of one module when new sequences, either synthetic or cloned directly from the genome, become available. Moreover, BiCT have continued to expand their sequence database focusing on both constitutive and inducible promoter sequences, export signals, and the marker genes that can be used to rapidly test different combinations of both, in order to achieve efficient production of the proteins of interest.
The genetic constructs were optimised for production of bifunctional P450s CYP153/CPR from Marinobacter aquaeolei and CYP52A10 from Beauveria bassiana in Aspergillus sp. and cloned in a suitable vector. Additionally, two challenging ODEs were successfully produced in soluble and active forms using the BiCT FTO Aspergillus system: the unspecific peroxygenase UPO1 from Agrocybe aegerita and the laccase LCC1 from Coprinopsis cinereus. Unspecific peroxygenases are an exciting class of enzymes offering industrially important activities, but are extremely limited in application due to their notoriously difficult production. Development of production strains and protocols for UPO1 were therefore a key target for BiCT. Unspecific peroxygenase were effectively produced as concentrated, spray-dried material from Aspergillus, and these were made available to other BIOOX partners in multiple formulations. The galactose oxidase G2 variant was produced in Aspergillus oryzae and transferred to BIOOX partners for comparison with enzyme produced in the alternative expression systems available within the project (PROZO’s production in E. coli, and BASF’s production in Saccharomyces cerevisiae).

As part of WP1, a number of putative enzyme panels were developed by PROZO targeted towards BIOOX objectives and supplied to the consortium. As part of WP2, further effort was made to improve upon the panels, largely via homologue sampling. In the central panel, total of 27 new NOX homologues were identified, 12 of which were produced in soluble form and shown to be active.

Pre-commercial industrial prototyping of the KREDs discovered and produced in colorimetric screening format during BIOOX by both PROZO customers and consortium partners demonstrated the utility of large panels of this commonly employed class of biocatalysts, and thus showed the need to expand the panels during WP2. Despite the fact that KREDs were already industrially applied, primarily for chiral secondary alcohol synthesis, there was still demand for further expanded panels offering enhanced properties and broader substrate scopes to access a wider range of products. The results in BIOOX from both academic partners and large end-users were of great fundamental and strategic value to PROZO, as it meant a much larger panel was actually required to meet ultimate industrial appetite, where the identification of off-the-shelf, process-ready biocatalysts saves incredibly valuable process development time. WP2 efforts therefore focused on mining natural three-dimensional sequence space from in-house PROZO metagenomes, rather than mutating a small collection of earlier hits to create diversity as was previously proposed. This metagenome mining led to development of kREDy-to-go™ plates 2, 3 and 4. Furthermore, moving entirely to mining of UK derived metagenomic libraries protects against potentially huge future liabilities for PROZO under the Nagoya Protocol and provides significant commercial advantages. The utility of these panels was demonstrated in WP6 by providing screening materials to potential R&D partners and customers both within and external to the BIOOX project.

BPT has developed several 3DM protein information systems for the partners in the work package and trained these partners in the use of such 3DM systems. BPT hosted a 3-day course for the project partners to train them in using the generated systems in smart library design. This course started with training in visualizing protein structures and continued with an introduction to the 3DM systems technology. The participants were taught how to analyse data and how to answer biological questions using the 3DM system. Alongside maintenance of 3DM systems to enable enzyme discovery and engineering, BPT performed product development to launch the new 3DM patent analysis tool developed within BIOOX. The ability to mine patent literature represents an extremely valuable addition to the software package. Although it remains in early development, it is a novel tool with high potential to deliver value for the field, and has already attracted considerable attention through product validation activities performed in WP6. With further development planned, this tool will increase the value and appeal of the 3DM systems, giving clients a greater incentive to purchase a software license.

Over the course of BIOOX, FIR has collaborated with USTUTT and UNIMAN on the discovery and development of P450 monoxygenases active on selected terpene compounds. Candidate P450 monooxygenases have been screened, eventually engineered and functionally characterized by USTUTT and UNIMAN against prototypical terpene hydrocarbons produced by FIR using submerged fermentation of engineered bacterial cells. FIR has produced and delivered ~10g of each of the sesquiterpene substrates for this initial P450 screening (WP1).

The best candidate P450s identified by USTUTT, UNIMAN, and PROZO from approximately 20 different CYPs tested were transferred to FIR for evaluation in fermentative production strains. FIR functionally expressed these P450s in bacterial cells engineered to overproduce the monooxygenase terpene substrate from a simple, sustainable, and cost-effective carbon source. Performance of the engineered bacterial strains was initially assessed in small-scale functional assays, and the activity of these biocatalysts was optimised by systematically testing different expression constructs, strain genetic backgrounds and production conditions. In order to further evaluate the performance of these more productive E. coli strains at scale, two strains producing the P450 susbtrates humulene and bisabolene were moved to fed-batch, biphasic fermentations in laboratory-scale bioreactors. Under optimised conditions, the strains achieved peak titres of humulene- and bisabolene-epoxides of >300mg/L. A downstream process was then implemented to recover about 4g of product for analytical characterisation of the product molecules and organoleptic evaluation by an FIR Master Perfumer in WP 3 and 5.

Work Package 3
Biocatalysts offer many advantages over chemical routes, principally inherent selectivity and relatively benign reaction conditions, however they also often require expensive and unstable cofactors and can themselves become unstable when isolated from cells. Moreover, the cost of enzyme purification may be prohibitive in many cases where the overall economy of the process does not allow it. Although enzymes can often be deployed as cell free extracts which can be relatively cheaply produced requiring only that cells are disrupted to release a crude mixture of enzyme and other cellular debris, this may not result in the required enzyme or cofactor stability. In such cases, production of whole cell biocatalysts may be appropriate, where the enzyme is retained in either metabolically active growing cells or inactive resting cells. Additionally, whole cell biocatalysts can incorporate a number of enzymes acting in sequence to perform longer biosyntheses, deriving complex products from simple and sustainable feedstocks by exploiting combinations of natural and synthetic anabolic pathways to build molecules. The overall objective of WP3 within the BIOOX project was development and delivery of suitable whole cell biocatalysts for the syntheses of industrial relevant omega hydroxylated fatty acids, alkenes and oxy-functionalized terpenes. To achieve these objectives different Rieske non-heme dioxygenases (ROs) and catalytically self-sufficient P450 monooxygenases were selected in WP2 for initial studies (CYP116B3 from Rhodococcus ruber DSM 44319 and CYP153A from Marinobacter aquaeolei, fused to the non-natural redox partners CPR from Bacillus megaterium or Pfor from Rhodococcus ruber DSM 44319).

Early work by USTUTT focused on the optimisation of expression protocols for production of the P450 enzymes and ROs. For the RO, naphthalene dioxygenase (NDO) a fractional factorial screening design was employed using MODDE software. This design of experiment (DOE) approach allows efficient use of time and resources by enabling smaller numbers of individual experiments to be conducted in order to gather the data required to define specific operating parameters. Thereby, measures in four areas were considered for the optimisation of expression: expression temperature, expression time, aeration rate, and agitation speed required for efficient mixing. The results suggested the following conditions as optimal for expression: expression temperature of 18°C, expression duration of 20h, aeration rate of 1.6vvm, and agitation speed of 300 rpm. The calculated optimal conditions for expression were verified experimentally by performing biotransformations on two substrates, naphthalene and α-methylstyrene. In both cases, the highest conversion was achieved using the calculated optimised expression parameters. Complete conversion of naphthalene to the corresponding dihydridiol was observed, whereas the conversion of α-methylstyrene resulted in a product mixture containing both 2-phenyl-1,2-propanediol and the allylic monohydroxylated 3-hydroxy-2-phenylpropene.

Optimised expression protocols to produce the P450 enzymes were also established by USTUTT, and in particular the choice of expression vector was discovered to be critical to protein levels, rather than expression conditions. P450 enzymes are capable of catalyzing a broad range of reactions with excellent region- and stereo-selectivity, but their industrial use remains limited in no small part due to the genetic instability of constructs and low enzyme yields with often poor activity. Therefore, development of optimized expression strains and protocols represents significant progress beyond current applications in drug metabolism screening, towards wider use of P450s for biocatalysis. High level, reproducible expression of CYP153AM.aq(G307A)-PFOR L2 was achieved using the strictly regulated pBAD vector system. A whole cell biotransformation using this biocatalyst achieved conversion of palmitic acid to the ω-hydroxylated fatty acid, with fivefold improvement over the previous systems. However, scale-up applying the same conditions for aeration, feeding and induction reported in previous studies in order to produce the biocatalyst in 5L fermentations was unsuccessful, resulting in inclusion body formation.

USTUTT studied the stability of artificial P450 fusion constructs, noting the genetic instability of previously reported biocatalysts used for whole cell biotransformations. By using the arabinose-inducible pBAD vector system the integrity of the fusion construct, and therefore the enzyme activity retained over time, could be enhanced significantly compared to other vector systems. During whole cell biotransformations with palmitic acid decreased product formation was observed due to the depletion of the product by the host. To overcome this issue the β-oxidation pathway, which is part of the catabolic pathway through which molecules are broken down by bacteria, was inhibited by incorporation of the ∆fadD deletion into the arabinose-catabolism deficient E. coli BW25113 strain required for pBAD. Furthermore, the transporter system FadL was co-expressed to enhance the transport efficiency to achieve higher product yields. The E. coli host strain along with the plasmids pBAD33 CYP153AM.aq(G307A)-PFOR L2 and pBAD18 FadL were transferred to DTU for optimisation of the biocatalyst fermentation at 5L scale.

Biocatalyst improvements from WP2 were transferred directly into whole cells experiments performed by USTUTT in WP3. A possible anchor position for fatty acid substrates was identified in the crystal structure of CYP153AM.aq, which by applying semi-rational design strategies, led to discovery of an improved variant Q129R. Significant yields of ω-hydroxy dodecanoic acid were obtained using this biocatalyst, comparing favourably with the previous best variant identified G307A. Indeed, once normalized for P450 protein content, the new variant displayed twofold enhanced performance in whole cell biotransformations. The strain E. coli BW25113 ∆fadD and the plasmids encoding the variant CYP153AM.aq(Q129R)-PFOR L2 were transferred from USTUTT to BASF for further optimisation and evaluation in fermentation.

The RO libraries of naphthalene dioxygenase viariants generated in WP2 were screened in vivo. Nearly all variants could achieve 80% or higher conversion of the natural substrate naphthalene. The library was next screened for activity towards the unnatural substrates α-methylstyrene and R-limonene. Biotransformation of α-methylstyrene, an important drug precursor molecule, gave two detectable products in modest yield; the monohydroxylated alcohol 3-hydroxy-phenylpropene, and the dihydroxylated 2-phenylpropan-1,2-diol. The stereoselectivity was also determined and reached up to 88% enantiomeric excess for the S-Diol with the single variant V260I. In contrast, the fragrance molecule R-limonene, was converted to two monohydroxylated alcohols; carveol and menthe-1,8-dien-10-ol, once again in modest yields. Notably, variants were identified which displayed complementary product distributions; single variant H295A showed a high selectivity for production the monoterpene carveol, whereas the variant A206G was more active but showed a higher selectivity towards production of mentha-1,8-dien-10-ol.

USTUTT further focussed on the oxy-functionalisation of a set of terpenes for application by FIR. Activities in WP1, 2 and 3 encompassed a full work flow where ODEs were identified and improved, before implementation in whole cell processes for the fermentation of valuable flavour and fragrance products. In collaboration with PROZO a set of 17 natural P450 catalytically self-sufficient fusion enzymes was screened for hydroxylation activity towards the BIOOX target substrate set. One candidate in particular was identified with the required activity, and further enhancements were attained using the 3DM database of BPT to implement semi-ration design strategies for protein engineering. The in vivo conversion of β-ionone to the allylic 4-hydroxy-β-ionone was improved 17-fold in some of the variants, however two single variants of enzyme P450_014 F95A and F95G exhibited interesting changes in substrate specificity compared to that of the wild type. Hydroxylated products of the BIOOX target substrates humulene, premnaspirodiene and longifolene could be identified in experiments using the two variants in vivo. These best candidate enzymes were transferred to FIR for further characterization of the products in the terpene producing strain.
In parallel with the development of the fusion enzymes, novel enzymes from the Class I and III P450 systems for the oxy-functionalisation of terpene substrates were investigated by USTUTT. Whereas Class VII and VIII P450s incorporate the CYP and redox domains in a single protein, Class I and III systems consist of three separate protein components; the CYP and two auxiliary proteins required for electron transfer and therefore the overall the activity of the system. Although the use of fusion protein has the potential to simplify processes and applications, Class I and III systems remain of interest due to their important substrate scopes and activities.

Efficient P450 whole cell biocatalysts based on the pBAD18/pBAD33 system were developed in WP1 and 2. Following expression optimisation, the Arthrobacter sp. P450 system identified in WP1 was tested as whole cell biocatalysts in WP3. Biotransformations of the natural substrate homovanillic acid yielded high product titres of up to 1.77 g L-1 of demethylated product, representing good productivity for a system of this type. Moreover, significant product titres of 0.5g/L of the BIOOX target 4-hydroxy-β-ionone were achieved in biotransformations using the in vivo system. CYP109C from Chondromyces apiculatus was produced using a similar whole cell pBAD system, incorporating the non-physiological CamA/B redox partners from P. putida. Once again, titres of approximately 0.5 g L-1 of hydroxylated β-ionone products were observed. A double variant of this enzyme T229L, A280V was of interest due to good activity and selectivity towards norisoprenoids. Three hydroxylated products were identified, and notably both non-allylic positions could be addressed, a useful activity that is rarely observed in biocatalysis. Alongside the main product, allylic 4-hydroxy-β-ionone, the formation of 3-hydroxy-β-ionone and 2-hydroxy-β-ionone was also observed, with a product distribution of 74%, 7%, and 19% of the total respectively.

The successful creation of an efficient system for whole cell biocatalysis also enabled the investigation of further substrate biotransformations. The CYP109C wild type enzyme accepted all BIOOX targets, and variants with increased activity or selectivity could be generated in each case. However, process development was also required in order to ensure that the biocatalyst could be implemented in realistic applications. Taking advantage of cyclodextrin as co-solvent for in vivo enabled biotransformation and analysis of sesquiterpenoid products. Cyclodextrin was used to both increase the solubility of the substrates, thereby enhancing mass transfer in the reaction, and to prevent evaporation of volatile products. Valuable targets including solavetivol and longifolene-aldehyde were produced by in vivo biotransformation with the most active variant V80A, A280I. Longifolene is an important fragrance compound found naturally in plant sources such as pine resins and certain types of tea. The pBAD constructs and best candidate enzyme variants were transferred to FIR for validation though application evaluation in the terpene-producing E. coli strains in WP6.

USTUTT in collaboration with the UNIMAN attempted to scale up the selective hydroxylation by the natural fusion enzyme P450 RhF of diclofenac to 5-hydroxy-diclofenac. 5-hydroxy-diclofenac is a metabolite of this important nonsteroidal anti-inflammatory drug, and therefore P450 systems capable of producing the compound synthetically find utility in toxicology screening and pharmacokinetic assays for drug development and formulation. Additonally, the use of this well characterised system as a model for application of P450s at preparative scale allowed identification of critical parameters and process improvements to be implemented in the fermentation protocols assessed. The fermentations incorporate two stages; i) growth of the E. coli (biomass) and production of the enzyme, the expression phase and; ii) the biotransformation phase using either growing or resting cells. Based on an optimised expression procedure in M9 medium, a total concentration in the biomass of 33mg active P450 RhF enzyme per gram of cell dry weight could be achieved reproducibly, representing useful activity yields from the fermentation.

In the subsequent biotransformation phase, an average space-time yield was achieved of 0.54mg of hydroxylated diclofenac product per litre of fermentation volume, per hour, per milligram of enzyme, per gram (dry weight) of resting cells. This was much higher that the yields achieved with comparable growing cells, an observation that could not be easily rationalised since the enzyme activity for corresponding CFE biotransformations showed similar activities in vitro. The probable cause of this difference was that NADPH cofactor availability presented a bottleneck in the P450 reaction. The concentration of NADPH, a cofactor used in many cellular processes, is much higher in non-growing or glucose-limited cells compared to metabolically active growing cells. The use of resting cells was therefore favoured for production of 5-hydroxy-diclofenac using P450-RhF. Additionally, the non-enzymatically catalysed formation of an over-oxidation product (quinoneimine) was observed. Quinoneimine production was dependent on the salt concentration and temperature of the fermentation, and could therefore be controlled by optimising process parameters to avoid formation of this unwanted side product. Final fermentations performed by USTUTT using optimised reaction conditions, reached significantly improved product titres of 0.36g/L and an average productivities of 11.7mg L-1 h-1.

Work Package 4
Biocatalyst cost can limit the application of enzyme technologies in manufacturing, particularly where products have low market value. Even when applied for the production of high value products, use of enzymes can complicate downstream processing, requiring laborious sample isolation to achieve product specification. Thus, economically viable application of biocatalysts requires process options that address these challenges, and one such possibility is to employ immobilised enzymes which can be easily recovered from reactions either to be recycled or in order to simplify DSP. WP4 was dedicated to development of immobilisation technologies for simplified and economical application of the enzymes developed in WP1 and 2. In WP4, CLEA Technologies BV (CLEA) focussed on two enzymatic oxidations assigned high priority in the Vision for Success. The first route was the GOase catalysed oxidation of HMF to produce DFF, and the second route combined alcohol dehydrogenase (ADH) for primary alcohol oxidation with NAD(P)H-selective NOX for cofactor recycling as a platform technology.

Two methods for the aerobic oxidation of alcohols were developed using enzymes previously developed in WP1 and 2 immobilised in the form of cross-linked enzyme aggregates (CLEAs). CLEAs are created using chemical treatments to form covalently cross linked aggregates of enzymes with good control over the size and composition of the final biocatalyst particles. Efficient application of CLEAs in biocatalysis requires that the reaction components can successfully penetrate the CLEA particles in order to fully access the enzymes therein. The first oxidation method developed by CLEA employed combi-CLEAs containing both GOase and catalase. This route was tested successfully for the oxidations of both the model substrate benzyl alcohol, and the priority target HMF, using a fed-batch hydrogen peroxide as the oxygen donor. The second method used a combi-CLEA containing ADH, NOX, and also catalase, and was successfully tested for oxidation of benzyl alcohol.

In order to produce a GOase CLEA with high recyclability, irreversible deactivation of the active site by hydrogen peroxide had to be prevented. For this reason it was decided to co-immobilise GOase with catalase, an enzyme capable of breaking down the peroxide, in one single CLEA. The activity of this combi-CLEA allowed lower global hydrogen peroxide concentrations within the CLEA and therefore increased GOase stability. Additionally, by using catalase in the GOase combi-CLEA, it was possible to generate high local oxygen concentrations, through hydrogen peroxide breakdown, only in proximity to the GOase where it was required. Catalase in combination with hydrogen peroxide therefore offered the potential to address the high KM¬ for oxygen of the GOase, one of the factors limiting the activity of GOase, due to the relatively low solubility of oxygen. In HMF oxidation reactions, activity recoveries and conversions observed for the combi-CLEA were substantially higher than those in a comparable system where soluble catalase was used in combination with a GOase CLEA.
It was found that at concentrations of HMF higher than 50mM the reaction could not reach completion due to product inhibition. Hydrogen peroxide addition rate was also identified as a critical parameter for the stability of the GOase/catalase combi CLEA.

BiCT focussed on developing carrier beads, procedures and protocols for the immobilisation of oxidase enzymes. In order to measure the activity of alcohol oxidase, a coupled colorimetric assay was optimised and tested with seven different commercial enzymes. Commercially available short chain (AOX2) and a long chain (AOX6) alcohol oxidases were used as model enzymes in BiCT’s high-throughput immobilisation platform, which enables simultaneous screening of multiple variables. The activity of the immobilised biocatalyst was measured using a standardised alcohol oxidase assay, with excellent results.
In order to identify optimal immobilisation conditions for the ODE target enzyme GOase M3-5, BiCT carried out full factorial screening of parameters that can influence the immobilisation. Although the immobilisation is a rather empirical process, the optimised immobilisation conditions were rapidly determined for the GOase M3-5 using BiCT’s high-throughput platform. Several carriers were tested in parallel, offering various different properties which can enhance performance: the functional groups that interact with the enzymes (the type and concentration of reactive groups); granulometry; bead composition (“medium”); pore size; surface hydrophobicity.

C-TECH Innovation Ltd (CTECH) tested supported oxidase formulations in a pressurised continuous flow reactor, the Coflore Agitated Cell Reactor (ACR). Continuous flow reactors such as the ACR offer many potential advantages over traditional batch reactors, including precise reaction control, easy scalability, and reduced catalyst usage. The ACR consists of ten cells interconnected by small channels. An agitator is placed in each cell, which mixes the contents of the cell when the reactor body is shaken by lateral movement. Agitators can take a number forms. Where catalysts are used in solution, springs can be employed for efficient agitation. If immobilised catalysts are used, then they can be introduced in mesh-walled containment units (“baskets”) which also perform agitation by moving within the reactor.

The rate of reaction for the oxidase catalysed process is primarily determined by the rate of oxygen uptake, which in turn is dependent on mixing efficiency and homogeneity of gas/liquid dispersion. Therefore, the use of a dynamically mixed continuous flow reactor with transverse mixing in preference over traditional rotational mixing had the potential for improved mass transfer under flow conditions, resulting in shorter reaction times and better enzyme consumption.
Following initial testing by CTECH of immobilised GOase G2 provided by BiCT, the basket containment units for the ACR had to be reengineered to provide adequate seals and prevent biocatalyst loss into the reactor. New stainless steel mesh inserts were manufactured to fit the existing baskets in order to contain the small 100-300 micron diameter beads. 100 micron mesh did not permit sufficient hydration of the beads so a 150 micron mesh was chosen, allowing sufficient hydration of the beads when under agitation. A small amount of the enzyme was lost into the system due to the compromise on mesh size, but this could be easily filtered from the system.

Tests of the immobilised enzyme supplied by BiCT were conducted by CTECH using two out of the ten baskets available in the ACR, in order to enable shorter testing times and more prudent enzyme usage during testing. During initial reactions, the baskets were completely filled with the immobilised enzyme. This resulted in poor conversion to DFF, with little or no improvement even over extended residence times. This was due to poor hydration of the beads by the reaction mixture. Therefore, in subsequent operation the baskets were only half filled, resulting in much better hydration of the beads. The improved bead hydration resulted in better conversions to DFF, and in a more linear fashion.

After successful conversions to DFF over short residence times, extended twenty four hour runs were conducted to determine the effective lifetime of the immobilised GOase enzyme. Consistent conversion was anticipated over the time course, however the activity of the immobilised enzyme quickly decrease first five hour period. CTECH and BiCT were able to determine that the DFF product was adsorbed onto the immobilised enzyme carrier, which in turn inhibits the enzyme. It is therefore unlikely that this immobilisation system could be applied for the oxidation of HMF, however work in WP4 provided excellent proof of principle for both BiCT’s enzyme immobilisation technologies and the reactor operation protocols developed by CLEA. The operational stability of the immobilised support was good, and more than 99% of the beads remained intact and retained within the baskets during twenty four hours of continuous operation. Alternative reactor configurations for HMF oxidation were identified and evaluated in WP5-7.

Work Package 5
This work package focused on the evaluation of reactor configurations and oxygen feeding strategies, and how these influence the stability and kinetics of oxygen dependent enzymes. The evaluation was conducted based on isolated enzymes, immobilized enzymes, cell free extracts, and whole-cell biocatalysts, representing each of the principal methods for preparation of biocatalysts. This allowed direct comparison between different biocatalyst options for a reaction, or of the optimised formulation where the biocatalyst was only available in a single form.

The work has focussed on identifying suitable reactor configuration options based on batch reactors and continuous flow reactors, and three strategies for oxygen supply. Two primary model reaction systems were considered; a fast reaction catalysed by GOase enzyme, and a relatively slow reaction with a P450 whole-cell catalyst. These reaction systems enabled study of the effects of different reactor configurations and oxygen supply on the biocatalyst operational stability and the overall reaction rate. These data provided inputs to WP7 to enable DTU to perform economic and environmental evaluations of the biooxidation processes.

Stability under process relevant conditions is crucial for the development of a biocatalytic oxidation process. Furthermore, the performance of BIOOX model systems was compared where different biocatlaysts were capable of performing similar chemical transformations. Therefore, enzyme stability was investigated by DTU for two biooxidation systems developed in WP1 and 2, using the conversion of benzyl alcohol to benzaldehyde as a model reaction system. The influence of the aeration, as well as substrates and products was investigated, and particularly the effects on each of the individual enzymatic components in both systems: i) GOase variants supplied by UNIMAN and the enzyme required for its reactivation, horseradish peroxidase (HRP); ii) KRED and its cofactor regeneration enzyme NOX from PROZO.

Supply of oxygen is critical in to the application of ODEs, since oxygen limitation can severely restrict productivity, and the choice of oxygen supply method can drastically effect the safety, economy, and overall viability of the process. Three different methods of oxygen supply in reactors were evaluated by DTU: i) bubble aeration (sparging); ii) membrane contactors and; iii) in situ oxygen supply by hydrogen peroxide degradation using catalase. A comparison of the performance of three different membrane contactors, where oxygen permeates a membrane in order to supply the reaction. In each case, relatively similar performance was observed for the different modules, but practical implementation at large scale was found to be complex and expensive. The feasibility of feeding hydrogen peroxide and catalase was evaluated, but was impractical due to the inefficiency of mixing in large scale reactors, the tendency of highly reactive peroxides to deactivate enzymes at even low concentrations, and the large increases in reaction volumes caused by feeding liquid hydrogen peroxide at even the maximum safe concentrations for operation. It was therefore shown that bubble aeration is the primary choice based on the comparatively higher oxygen transfer rate and simpler operation.

However, the final choice of oxygen supply method is ultimately determined by matching process conditions to the biocatalyst or vice versa. Further testing was performed by DTU to determine the impact of the oxygen supply method on the stability of two oxidases, and the bubbling of gas into the reactor was found to be a feasible alternative for GOase, whereas it clearly damaged the NOX.
Oxygen supply is often a limiting factor in bioxidation reactions, due to the relatively poor solubility of oxygen in unpressurised reactor systems. It is therefore critical to be able to experimentally determine the kinetics of oxygen dependent enzymatic reactions in order to enable rapid screening or reengineering of biocatalyst variants. Kinetic characterization of oxygen dependent enzymes is a time consuming and material intensive task due to the difficulties involved with setting and maintaining multiple different concentrations of poorly water-soluble oxygen. DTU developed a novel laboratory microscale reactor designed to obtain initial rates for oxygen dependent enzymes at a large range of different oxygen concentrations. A constant oxygen concentration was obtained by utilising a tube-in-tube micro reactor where oxygen is transferred from an oxygen-nitrogen mixture over a membrane with high oxygen permeability to the liquid reaction medium. The reactor can be pressurised to increase the solubility of oxygen in order to fully saturate even enzymes with low oxygen affinity. An online spectrophotometric detector then enables quantification of reaction substrates and products in the outlet of the reactor. The system is fully automated and can fully characterize an enzyme in less than 12 hours using a minimum amount of material. This novel system is an important innovation to enable rapid determination of enzyme kinetics, and will find further applications with other reaction systems and enzyme classes. The reactor is currently the subject of a patent application by DTU.

Galactose oxidase variants generated in WP1 and 2 were kinetically characterised by DTU using the tube-in-tube reactor. An enzyme variant discovered by UNIMAN using reduced oxygen screening was determined to have a 2.8-fold higher reactivity with molecular oxygen; an important validation of the reactor technology, and a critical step towards viable application of this biocatalyst. This makes the variant a much better choice for industrial application, because due to the higher reactivity at the low oxygen concentrations experienced in an industrial reactor, the amount of enzyme required can be reduced by more than 50%. This is an important cost saving for large scale commodity product manufacturing.

Biocatalytic oxidation reactions employing molecular oxygen are difficult to conduct in a continuous flow reactor because of the requirement for high oxygen transfer rates. DTU selected the oxidation of glucose to glucono-1,5-lactone by glucose oxidase as a model reaction to study the ACR reactor system. In this case, the enzyme was used in solution with spring agitators. Based on tracer experiments, a hydrodynamic model for the ACR was developed. The model consisted of ten tanks-in-series with backmixing, where the reacted stream mixes with unreacted components, occurring within and between each cell. The backmixing was a necessary addition to the model in order to explain the observed phenomenon that the ACR behaved as two continuous stirred tank reactors (CSTRs) at low flowrates, while it at high flow rates behaved as the expected ten CSTRs in series. The performance of the ACR was evaluated by comparing the steady state conversion at varying residence times with the conversion observed in a stirred batch reactor of comparable size. It was found that the ACR could more than double the overall reaction rate, which was solely due to an increased oxygen transfer rate in the ACR caused by the intense mixing as a result of the spring agitators.

In WP5, CTECH applied the ACR for the conversion HMF to DFF, measuring performance parameters as one of the inputs to WP7. Immobilized galactose oxidase was used successfully to partially convert HMF, however, significant stability issues with the biocatalyst were identified. Adsorption of the product onto the beads led to inactivation of the enzyme, and would additionally complicate downstream processing. These challenges would need to be addressed to enable industrial application of immobilized galactose oxidase for this type of reaction. Moreover, the ACR were used to convert HMF to DFF using free galactose oxidase in a biphasic reactor system using pure oxygen for aeration. In this setup 250 mM of HMF could be fully converted with a residence time of only three hours, significantly improved from reaction times taken to achieve similar conversions in small scale batch reactions. Alternative reactor configurations, including alternative and novel batch type reactions, were modelled in WP7 in order to evaluate the feasibility of different process options for HMF biooxidation.

Work Package 6
The goals of WP6 were aimed at demonstration activities for BIOOX technologies. Demonstration activities completed included a workshop on the use of flow reactors in ODE biocatalysis; presentation of 3DM tools for analysis of FTO; production of demonstration scale batches of enzymes/enzyme formulations, and testing of BIOOX materials and methodologies in industrially-relevant processes and at scale.

BPT demonstrated the patent analysis tool to five large Biotechnology companies, four large pharmaceutical companies and BPT has presented the patent analysis tool at two prestigious drug design conferences (Drug Discovery Chemistry, and Discovery on Target) in San Diego and Boston, respectively. This has already resulted in a follow up with two biotech companies and with several pharmaceutical companies. Continued development on the patent analysis tool will be required to accommodate the different use cases of each interested company. BPT also presented the patent analysis tool at the BIOTRANS 2017 symposium in Budapest. A presentation was given in the industrial session to illustrate the capabilities of the patent analysis tool for FTO analysis.
Offering large collections of diverse enzymes in colorimetric screening format is an ideal proposition for PROZO’s large pharmaceutical industry (“big pharma”) customers, as the probability of identifying a process-ready enzyme off the shelf is maximised, potentially saving a great deal of process and/or enzyme development time, which is critical to competitiveness. Traditionally, the ability to rapid develop new syntheses was limited by the time taken to discover or engineer suitable biocatalysts, and is a key reason for the historically poor uptake of biocatalytic technologies by industries adapted to the short timelines required for development of chemical routes.

The utility of the metagenomic plates was demonstrated as part of WP6 by offering the screening kits to the BIOOX consortium and within another EU grant consortium – the H2020 project CARBAZYMES – where multiple KREDs were found to be suitable in native form for key biotransformations of interest. Beyond the consortium, one big pharma customer that expressed initial scepticism about whether any useful hits would be obtained through this approach reported 24 good hits against a small collection of challenging compounds during prototype screening in-house. In this way, and given that all prototype KREDs were produced at the 1-2 g scale, 100mg quantities of any hit suggested by kREDy-to-go™ can be immediately provided to the customer for in-house prototype testing against their prospective application, thereby maximising the potential for new business for PROZO, and delivers strong impact for the work performed during BIOOX.

Significantly, after evaluation of the BIOOX prototype NOX enzyme panel, one of PROZO’s biggest customers disclosed to the community that they identified a NOX enzyme for use in a major new application. While NOX enzymes have been known for some years, an enzyme suitable for the demanding application in question was only found by mining diverse homologues from relevant three-dimensional sequence space using PROZO GRASPTM technology. That process has now been scaled-up to 6kg as part of product validation activities in WP6, enabling PROZO to demonstrate the challenging fermentation of the NOX at industrial scale. This was a crucial demonstration of the feasibility for application of this enzyme, because production by scale-out using fermenter-in-a-flask was not suitable, and the scale-up of production was unprecedented. Providing that demonstration has enabled the big pharma customer to negotiate production by PROZO beyond the BIOOX project.

Galactose oxidase variant G2 was selected as the candidate biocatalyst for evaluation by BASF in the oxidation of HMF. In WP2 GOase G2 was developed to produce extremely high activity towards HMF for oxidation to the commercially relevant intermediate chemical DFF. Optimisation of enzyme properties and reaction conditions for HMF oxidation were performed by UNIMAN to improve the suitability of the biocatalyst for application at a large scale. PROZO supplied GOase variants in hundreds of grams quantities to other BIOOX partners for experimental testing and evaluation in processes.

CTECH performed scale up in flow of HMF oxidations previously developed as shake flask or stirred vessel reactions at laboratory scale by UNIMAN and DTU in WP1/3. The HMF reaction was scaled up by CTECH to both 250ml and 1L, and the conditions used for a successful 3 hour run were then used to perform a continuous twenty-four hour run. This run enabled production of a larger batch, including downstream processing and product isolation, allowing calculation of the overall yield for the process and characterisation of the product. Overall conversion of 80% to DFF was achieved over the course of twenty-four hours of continuous operation and approximately 10% of the HMF remained unconverted. After isolation of the product, this resulted in an overall isolated product yield of 52% with a purity of 89% DFF. The remaining 11% was mainly HMF, with less than 0.1% FFCA/FDCA. DSP was not further optimised in this case, but significant potential exists for improvement of product isolation methodologies.

Biooxidation reactions were performed by BASF at scale using two different reaction systems: i) GOase to produce DFF, and; ii) using xanthine oxidase (XOD) for production of an agrochemical active ingredient. For demonstration purposes this material was sent to BASF internal customers for further assessment. Several field trials with sample material based on a biooxidation process implemented by BASF showed poor performance of the products, and further development work (to be undertaken outside of BIOOX) will be required to achieve the required product specification.

Although the quality of DFF from galactose-oxidase catalysed biooxidation was sufficient if not excellent, there remained severe issues concerning its further chemical conversion to downstream products. Further work to derivatise DFF (performed outside the scope of BIOOX) is not yet finished, and additionally there are ideas to replace these chemical syntheses with enzymatic alternatives. The biooxidation of HMF remains an important target for BASF and patent applications have been submitted for two process variants developed in BIOOX.

As part of WP6, CTECH developed a workshop on the application of plug flow reactors for gas fed ODE catalysed reactions. This workshop was held as a satellite event to the Organic Process Research & Development Conference held in Stockholm, Sweden in September 2017. Alongside the workshop presented by CTECH, other BIOOX partners delivered presentations on the selection use of ODE in processes developed using plug flow reactors. An overview of the project was provided by UNIMAN (BIOOX project summary and objectives), with additional presentations from DTU on novel tube-in-tube reactor systems (Continuous reactors for biocatalytic applications), and from PROZO on discovery and screening for selection of ODEs (Screening methodologies for the selection of oxygen dependent enzymes). A targeted audience was reached, including representatives from organisations such as Merck, Boehringer Ingelheim, AstraZeneca, University of Cambridge, GSK and Givaudan. Coupled to the workshop activities of CTECH, this created a high impact event delivering an overview of BIOOX technologies to potential end users.

Work Package 7
WP7 was focussed on economic and environmental process evaluation. The overall objectives were to develop economic and environmental models as well as green and sustainability metrics for the evaluation of oxidase bio-processes. This was a critical aspect of the project, representing the culmination of much of the biocatalysts optimisation and process development work undertaken in the other WP. Uptake of biocatalysis and other IB processes requires that they can compete on quality, performance, and economy, with alternative chemical technologies. Performance metrics could be compared against the targets established in the Vision for Success, allowing the project consortium to understand the gains made through development of technologies, and importantly to benchmark those technologies against the requirements of industrial processes where they find potential application.

Initial work by DTU was directed towards developing flowsheets for biocatalytic oxidations processes selected in WP5. Two important process aspects were prioritized. Firstly, different biocatalyst preparations, i.e. growing cells, resting cells or isolated enzyme, were compared in terms of the expected oxygen requirement. The work also indicated that at least some processes may be sufficiently aerated using membrane technology, which may prove beneficial for the stability of some biocatalysts developed in the project. Others may be feasibly operated using bubbled aeration, which is a well-established method in industry. Secondly, several different co-factor regeneration alternatives for the biooxidation of alcohols were selected for further investigation. During the project, flowsheets have been developed based on the reactor configurations evaluated in other work packages. This resulted in process models that allow direct comparison of cost and environmental impact of different process alternatives, and enables the identification of bottlenecks which is useful for further process development.

The design of a process flow-sheet is dependent on a number of factors, such as the reactor type, oxygen supply method, biocatalyst format, need for substrate feeding and/or product removal and the downstream processing for recovery of the product. The figures below show general examples of process flow-sheets, including a simple process with only one separation step and a more complex process in which several separations are required both for recycling of e.g. biocatalyst and a reactant and for the recovery of the product.

The flow-sheet illustrates a process using a whole cell catalyst in a resting state. In such a process the biocatalyst is first produced through a fermentation step and subsequently separated and used to catalyse the synthetic reaction.

For the work on flowsheet analysis conducted in WP7, the oxidation of HMF to DFF is considered as an example of what can be achieved. It was a carefully chosen system in order to reflect the challenges of oxidation for the synthesis of low-value products. At an early stage potential alternative process flow sheets for performing the GOase-catalysed oxidation HMF were considered. Based on a comparison of the physio-chemical properties of HMF and DFF it was concluded that an ex situ product removal strategy based on crystallization of DFF was the best alternative. The crude DFF crystal can be further purified via a series of recrystallization steps to obtain DFF crystals with sufficient purity to be used as starting material for further chemical modification or polymer synthesis. Nevertheless for practical operation in the demonstration activities, operation with the alternative two-phase system was assessed at BASF as part of WP6.
In addition using the flowsheet, the material and energy balance was also developed to give the basis for scaling and costing. Based on the current knowledge of the reaction conditions and a number of assumptions the energy and material consumption for a plant producing 100,000 tons of DFF per year was calculated. Based on the process simulation it was found that 1.14 kg of crude HMF (95% purity) and 3.3 g of enzyme were required per kg of pure DFF. The energy requirements without taking electricity for pumping into account was found to be 4.1kWh for heating and 4.3kWh for cooling per kg product. The current process is based on multiple assumptions that need to be investigated experimentally, but in general many assumptions are set conservatively.

The Vision for Success in BIOOX, has played a central role for establishing a process model, stating the process targets in terms of e.g. productivity (gL-1h-1) and product purity. With this information as a reference, the required performance in the preceding process steps could be assessed. This can then be used for evaluating performance realized in DTU model systems. In the final stage of the work an economic assessment of this exemplar system was carried out.
Oxygen supply to fermentations has been a widely studied field of research for many decades. With the recent developments in oxidative biocatalysis, the supply of oxygen to biocatalytic reactions has never been more important. Although biocatalytic oxidations and aerobic fermentations share many similarities, a number of differences are critical when evaluating oxygen supply methods, such as instability of certain enzymes in the presence of gas-liquid interfaces. Likewise biocatalytic reactions can potentially be much faster, than their fermentative counterparts. To circumvent the problems associated with the traditional method of supplying oxygen via bubble aeration of stirred tanks, a number of alternative oxygen supply methods have been proposed in the scientific literature. In this WP the most promising alternatives were evaluated by DTU based on their technical feasibility and cost, and compared to traditional methods of supplying oxygen via bubble aeration of stirred tanks. The methods investigated included enriched air aeration and pressurization of stirred tanks, membrane aeration with a submerged and an external membrane configuration, and in situ oxygen generation by hydrogen peroxide decomposition. In reality little difference in the price per kg based on the method of oxygen supply was found. This is the first investigation of its type and this is an important conclusion.

Summary
Overall, the aims of BIOOX have been achieved, and the technologies developed during the lifetime will find applications in industrial context. Key early successes beyond the project are likely to be through the commercialisation of tools and methods for ODE discovery and production. Equally, the body of research literature and laboratory technologies has been greatly enhanced over the course of the project, with new experimental approaches and capabilities available, alongside tools for decision making which will inform future research and development in the area. Further development will be required for the application of BIOOX processes in commercial chemical manufacturing, but planning for those efforts is already underway, and the desire to continue that development is reflected in patent applications from within the project, and further process development and evaluation activities being undertaken by the industrial partners beyond BIOOX.

(See attached PDF)

Potential Impact:
The aerobic biocatalytic oxidation reaction is one which would currently have significant impact on the future uptake of industrial biotechnology (IB) in Europe because chemical oxidation is both hazardous and has high environmental impacts. Many biocatalytic reactions have been identified in academic laboratories but few have been applied on an industrial scale. The project gathered European manufacturers and research organizations with complementary interdisciplinary skills across Europe in order to form a strong transnational collaboration to generate new knowledge and new technologies that will be implemented in new products.

Oxidations are amongst the most important reactions in chemical manufacturing, and as such established technologies are well embedded in industrial processes. These traditional chemical routes offer rapid development timelines for the implementation of robust processes which can be easily addressed using existing infrastructure. However, they also often rely on hazardous and dirty reactions, using fossil feedstocks and transition metal-based catalysis, with concomitantly high costs to the environment and human health. Efforts to improve the sustainability of chemical methods have concentrated on the use of molecular oxygen as the least polluting stoichiometric oxidant, commonly employing metal catalysts in organic solvents. The use of solvents other than water represents a two-fold challenge. Firstly, highly polluting chlorinated solvents often offer the best reactivity. Secondly, the use of oxygen in combination with volatile organic solvents presents a potential fire or explosion hazard. Thus, the most sustainable chemical catalysis methodologies require a difficult balance between reactivity and environmental and safety profile, limiting the reaction scope. Moreover, chemical catalysis is often limited by inherent lack of regio- and stereo-selectivity, often requiring complex synthetic strategies to ensure product identity. These limitations stand in direct contrast with the general benefits of biocatalytic alternatives, which are generally highly selective and require relatively benign reaction conditions. Indeed, enzymatic catalysis offers exquisite selectivity, meaning that transformations can be performed without protection-deprotection chemistries, requiring additional reaction steps, needed by their traditional counterparts. Even where complex, multistep processes are implemented, enzymes are often capable of operation under the same reaction conditions, in aqueous environments, requiring physiological temperatures and pH. The inherent substrate selectivity of enzymes also facilitates progressive and sequential reaction of substrates without the need to isolate intermediates. Therefore multistep enzyme syntheses can often be performed as one-pot processes.

In many cases, biocatalytic systems are capable of transformations that are inaccessible via chemistry, enabling production of novel molecules previously unavailable to the market. Biocatalysis thus has the potential to deliver not only improved industrial processes, but also better products with enhanced or advanced properties. These benefits are already starting to be realised by the pharmaceuticals sector, where chiral synthesis of APIs and privileged scaffolds enables enantiopure production, avoiding chiral resolutions where previously 50% of a racemic reaction product would be discarded during tedious product isolation steps. However, even in those cases, biocatalysis remains the technology of last resort, employed only where the economy of a process demands. Moreover, the use of biocatalysis is generally restricted to a small number of very specific processes, such as the use of ketoreductases in synthesis of chiral alcohols, or transaminases for chiral amine production. In most cases, these reactions are limited by substrate scope and the requirement that they address high value product molecules.

BIOOX therefore aimed to address the barriers to commercial implementation of biocatalysts, specifically by developing a validated aerobic biooxidation platform for synthesis and modification of alcohols. Alcohols are a critical class of molecules as intermediates and as products in their own right. Thus, both biocatalytic synthetic routes to alcohols, and biocatalytic modification of alcohols are important technology targets. The general classes and activities of the oxygen dependent enzymes required for the biooxidations targeted were known during the design of the project, however a wealth of genetic data exists in public databases containing sequences of many hundreds of thousands of putative and candidate ODE sequences. Bioinformatics tools enabling rational searches and filtering of that data to identify specific candidates and panels of enzymes as starting points for development in BIOOX were therefore essential.

The bioinformatics software developed through BIOOX allow interrogation of sequence data based on structure-function relationships. This allowed the laboratory researchers to identify candidates for activity screening, and also to design protein engineering strategies to improve the activity and other properties for application. Feedback from those experiments allowed further refinement of the bioinformatics platform, including the design and implementation of a tool to mine patent texts. The patent landscape analysis system allows identification of patented technology which can be evaluated to improve new technologies during development, but more importantly can be used in future projects or by industry to assess freedom to operate (FTO). The ability to identify other relevant IP at the earliest possible opportunity avoids wasting time and resources on the development of technology which could never be realised as commercial products due to lack of FTO. This software was further validated within the project and by demonstration to potential external customers, and has attracted significant commercial interest beyond the project.

New screening panels of KREDs, NOX, P450s and reductases have been developed. This is an important offering to the chemical manufacturing industries, since ready to screen panels of diverse enzymes can help to reduce the development time for biocatalytic process to timescales similar to those of traditional chemistry. This addresses one of the key barriers to uptake by industry, meaning that evaluation of biocatalytic alternatives can be undertaken at an early stage of product development, rather than being considered only as a last resort (often very late in process development, necessitating that the biocatalyst tolerates restrictions imposed by other chemistry). These panels will be of interest to various sectors, but are likely to initially attract interest from pharmaceuticals companies, where previous use of enzymatic and biocatalytic processes has already reduced barriers to use. From a societal perspective, novel methodologies for simplified production of medicines will enable cheaper manufacturing and thus increase their availability to treat more patients and in less wealthy countries. In particular, a hit will enable a large pharmaceutical company bioprocess to become economically viable for the first time, as no other enzyme was found until the BIOOX NOX was screened. The company plans to patent the new process for manufacturing a drug targeting a major disease, and has planned commercial manufacturing to commence in 2018.

BIOOX has provided new tools to find and exploit novel enzyme catalysts to produce nature-identical fragrances and flavours. Traditional chemistries employed within the F&F industry have employed harsh reaction conditions and heavy metals catalysts, in most cases achieving only a mixture of differently hydroxylated products, or resulting in over-oxidation of alcohols to carboxylic acids. The limitations in selectivity and the environmentally unfriendly reaction conditions can be easily overcome by using biocatalysts and the selection of more environmental-friendly reagents. In general, biotechnologically produced fragrances and flavours are more expensive than the well-established chemical procedures, but increasing waste disposal costs associated with the toxic substances generated by traditional chemistry makes room for novel and green processes.

Enzyme engineering was undertaken to optimise their properties for application, encompassing technologies working both parallel to and complementary with discovery of new wild type enzymes. In some cases, enzymes were already well-characterised in work prior to BIOOX, where in other cases completely novel enzymes were engineered as required. Once again, engineering strategies were supported by the bioinformatics tools and scalable enzyme production supplied by BIOOX SMEs, further validating the workflow for development of biocatalysts, and those specific technology platforms to support it. Thus, the project contributed to the sustainable growth of EU SMEs, but also provides a transferable approach to R&D in this area. This is an important output of the project, since it provides a future model for collaborative research, and establishes strategies which can be transferred to industrial R&D. The validation of the overall approach thus helps to mitigate risks in future research programmes, increasing the confidence of the sector to undertake development of biocatalysis.

P450 enzymes are capable of catalysing diverse and industrially important transformations with applications in many sectors: in pharmaceutical manufacturing, including for challenging syntheses of complex antibiotic molecules, steroids, and other drugs, and for deriving metabolic intermediates for toxicology studies; in the synthesis of monomers from sustainable feedstocks for use in bioplastics; in production of biosynthetic flavour and fragrance compounds. Some of the key limitations of P450 systems were genetic and protein instability, and relatively low product titres in biotransformations, rendering the application of such systems economically and technically inviable. Variants isolated from screening libraries of P450 enzymes were engineered for increased activity towards BIOOX substrates, including F&F targets. This was achieved through active site engineering based on the structure function relationships elucidated using the improved bioinformatics tools. The best candidates were characterised as purified enzymes before being developed as whole cell biocatalysts. A series of high impact publications were prepared describing the discovery and development of P450s in BIOOX, with enhancements achieving modified substrate scope and product spectra, along with improved protein levels and stability. Together with better genetic control of protein production, along with modification of the E. coli to improve substrate influx and avoid product metabolism by the host, these improvements have achieved more robust biocatalyst with higher product titres. Taken alongside the P450 enzyme and reductase panels now available as a result of the project, these technologies represent significant additions to the body of P450 research, and progress towards broader application of this family in industrial processes.

Terpenes are one of the most important chemicals to the Flavour and Fragrance (F&F) industry. Due to supply shortages and price inflation, biosynthetic routes to some of these structurally complex compounds are currently being developed as an alternative to the conventional plant extraction processes. Most of the terpenes are natural compounds found in essential oils of plants and they are extracted from these plants. Fragrance and flavour industries use terpenes as starting materials for chemical synthesis to produce terpenoids, which are highly sought-after molecules. The discovery of novel compounds with potential anti-microbial, anti-proliferative effects or a new smell/taste is a highly interesting research field. One of the key but also most challenging biosynthetic steps is the oxyfunctionalisation of terpene hydrocarbons, which is essential for conferring organoleptic properties to this class of compounds. The work accomplished in BIOOX has contributed to the discovery and optimization of P450s for the selective hydroxylation of F&F terpenes in high density fermentations. This work provides a basis for the development of alternative routes to F&F terpenes that are both sustainable and cost effective. In addition, the results made available from BIOOX will push the state of the art field of selective terpene hydroxylation and will further influence research programmes undertaken in Horizon 2020, the successor of the 7th Framework Programme for Research and Development.

Oxidation of hydroxymethylfurfural (HMF) to diformylfuran (DFF) was a high priority target for the project. DFF is an attractive intermediate with potential applications in a variety of polymer formulations as a bio-based alternative to petroleum based plastics and other materials. The starting material HMF can be derived from sugars from sustainable biomass, and industrial production is already possible. However, HMF production results in an impure product, and extensive downstream is required to purify it. An ideal HMF biooxidation therefore requires a biocatalyst with not only excellent selectivity without the need to highly purify the substrate, but also tolerance of the impurities present in the mixture. This was achieved through the identification of a galactose oxidase variant displaying remarkable activity towards HMF, greater even than that of the wild type galactose towards its natural substrate. This variant was further engineered to increase its reactivity towards molecular oxygen, a key requirement for implementation at scale, where dissolved oxygen concentrations are a potential bottleneck.

Few, if any, studies to date have focused on the understanding of oxygen supply to biocatalytic oxidations. It is clear that the developments in BIOOX are of great importance in this regard, not least to compare supply methods and test in full the agitated cell reactor (ACR) equipment now on several demonstration systems, including with galactose oxidase. This provided the basis for reactor selection, linked with a deeper knowledge about when to use the various methods of supply and when not. Several papers emphasize these methods. Of even greater importance has been the development of the tube-in-tube reactor, and the data subsequently generated using this novel equipment. This interesting method of collecting oxidation kinetics overcomes the standard limitation of having to sample high pressure apparatus. Nevertheless the need for such pressures was unexpected and has proven interesting in its own right. Work on the “low oxygen” variants of galactose oxidase in BIOOX has now shown that the development of such a system can also lead to the protein engineering to derive improved variants for the first time. This is a hugely important result, and the tube-in-tube reactor also has implications for other types of reaction and will have generic and wide applicability. Commercialisation of this reactor will be undertaken beyond the project, and to this end, a patent has been submitted and received favourable feedback from the examiners. BIOOX has thus delivered a set of technology tools of great value to the kinetic analysis and operation of oxygen dependent enzymes. Once again, this reduces a barrier to uptake by wider industry through the provision of practical laboratory scale tools and methodologies that can be easily adopted to accelerate research during new ODE and process development. The work has been well documented and disseminated via a series of publications, many conference presentations and posters as well as a webinar.

In certain bioprocesses, catalyst costs can be reduced and downstream processing can be simplified by the use of immobilised biocatalysts, permitting easy separation of the catalyst from reaction streams and its subsequent recycling. The use of immobilised biocatalysts is particularly advantageous in flow reactors, where it can reduce biocatalyst loading and enable better control of the order of operations by containing the biocatalyst within a section of reactor. The ability to immobilise enzymes with good residual activity is crucial to the economic performance of a process, since protein loading is a key cost driver and also influences the productivity and footprint of processes. Manual optimisation of this process is laborious, requiring screens of carrier mediums, porosity, immobilisation conditions, cross linking chemistries, and a constellation of other factors. The capability to rapidly screen these options using automation reduces the time taken to produce the biocatalyst formulations, allowing faster progress to evaluating their performance in reactions. In BIOOX, an automated high throughput platform for screening and optimisation of biocatalyst immobilisation was further developed using the galactose oxidase enzyme as model system. Immobilisation of the enzyme was achieved with very high residual activity, however evaluation in the ACR revealed that the DFF product was adsorbed onto the immobilised enzyme carrier, which in turn inhibits the enzyme. It is therefore unlikely that this immobilisation system could be applied for the oxidation of HMF, however the operational stability of the immobilised support was very good, even in this rigorous reaction environment. These results therefore represent excellent proof of principle for both the high throughput immobilisation technologies and the reactor operation protocols for application in other biocatalytic systems. By understanding the benefits and limitations of these technologies, they can be more easily incorporated into other research and development projects.

Next, reactor alternatives were assessed for application of galactose oxidase on scales suitable for manufacturing. Process optimisation was performed and the galactose oxidase variants were assessed in various scalable reactor configurations, in both free and immobilised forms. Laboratory scale reactions were performed to generate process metrics used to develop detailed flow sheets and allow economic evaluation of the various options. This is of great importance, primarily in order to evaluate when biocatalytic oxidation is beneficial and when not. It is very clear from this work that there are significant environmental benefits to biooxidations versus traditional alternatives. Economic evaluation was likewise based on assessment of the biocatalytic oxidation of HMF to DFF, which is a low-value product and therefore represents one of the most challenging scenarios. Through this evaluation, it was identified where improvements can take place and what developments are possible to drive the process further. This has resulted in applications for initial patents on the DFF process, which will be further developed beyond the project.

The greatest impact of biocatalysis on sustainable chemical manufacturing would be realised through implementation of bioprocesses for sustainable production of commodities, such as polymers. The sheer volume and intensity of production means that these processes have high energy costs and environmental footprints. Not only are the processes themselves polluting, they also tend to rely on fossil feedstocks and the products are often environmentally recalcitrant, meaning that there is are high environmental and resource costs at both ends of the supply chain. Global production of plastics is in the tens of megatonnes, and therefore significant disruption of this market requires robust processes that can compete with existing low costs production routes. The economic evaluation for HMF oxidation demonstrated that this target is within reach. The quality of the DFF produced in the biooxidation was found to be excellent, even where relatively crude HMF was used as substrate, and this is a very important result. The galactose oxidase systems generated in BIOOX clearly represent good progress towards practical application of HMF biooxidation, and the requirements for future biocatalyst and process development in order to achieve economic and technical viability are clearly understood. There remains a need to develop further processes for derivatisation of DFF to give final products, and initial work to perform such transformations chemically has been without success. However, biocatalytic routes to derivatised products are being considered, meaning that a one-pot or telescoped bioprocess may be possible. This is an important step towards biocatalytic production of sustainable low-cost, bulk polymers.

There is widespread interest in biocatalytic oxidations today. There are many small scale chemical studies of alternative biocatalysts and new products. The aim here has been to place that in context and understand what is required from the perspective of industrialisation and implementation of a new process. In this regard the key limitations have been found and applying a methodology developed in other projects to the strong collaboration in BIOOX has resulted in a unique and valuable output. BIOOX, through the potential impact of the new knowledge has already attracted young scientists to join the biotechnology industry. Furthermore, it will in turn be a source of highly trained, Europe-based personnel thanks to the increased demand for qualified and skilled personnel in industry and academia. Increasing industrial awareness of biocatalysis will, in part, be driven by dynamic and motivated young scientists entering the chemical manufacturing industries, and this will further increase acceptance of the technology as an option from the earliest stages of product and process design.

BIOOX will contribute to sustained European research and boost collaboration with universities, thus allowing European research organizations and industries to provide global leadership in this area. Efficient technology transfer from academic laboratories to European industrials is also crucial to ensure the competitive development of the chemical industry in the EU. High impact publications and presentations to major conferences have been a key output of the project, ensuring timely dissemination of the technical results to the scientific community. This was complemented by the presentation of the entire project at an OPRD Conference workshop, extending the reach beyond the established biocatalysis community to potential end users with lower awareness of the state of the art in biooxidation technologies. Additionally, BIOOX partners are involved in complementary research programmes in H2020, including for instance ROBOX (project ID 635734), CARBAZYMES (635595), and METAFLUIDICS (685474) where the technologies and collaborative strategies from BIOOX provide precedent for further research and development in similar systems.

A prosperous ecosystem of companies and research organisations is critical to the growth of high technology sectors, and SMEs often possess an agility and technology focus that drives innovation for that entire network, helping to translate academic discoveries into large industrial processes. The support provided to SMEs in BIOOX has contributed to their sustainable growth through the collaborative research and development efforts, allowing product prototyping and validation by project partners and external customers, thereby expanding their understanding of the requirements of end users and increasing their visibility to industry. BIOOX has already enabled an SME partner to confidently recruit new personnel on the basis of anticipated orders directly arising from participation in the project.

Collaborative projects are important to allow large industrial partners to undertake R&D which would otherwise be deemed too high risk. In the case of BIOOX, collaboration with academic organisations and SMEs has enabled development and testing of biocatalytic processes for the production of oxyfunctionalised terpenes and the platform intermediate DFF by major chemical manufacturers within the consortium, and evaluation of the NOX enzyme in a drug manufacturing process by an external large manufacturer of pharmaceuticals. In both of the former cases, projects progressed from fundamental enzyme discovery and engineering, to process development, and finally resulted in the isolation of multigram yields, sufficient for internal product evaluation purposes. The knowhow and IP generated over the course of the project therefore not only provides general insights into the development and application of ODEs, but has also allowed progress towards practical application for the production of commercially important products. It is this sort of considerable progress which will ultimately increase the adoption of biocatalysis as a primary option for R&D in the area of oxidations for chemicals manufacturing, leading to new processes and products.

Dissemination of basic concepts and the potential of industrial biotechnology beyond the scientific community has also formed an important role for BIOOX researchers, helping to raise awareness of biocatalysis and IB, whilst communicating the value of research programmes publically funded by the EU. The optimistic messages of environmental benefits and sustainability will also help to attract young people to study STEM subjects, and to consider careers in the area. For this reason, outreach to students approaching or in secondary level education was a priority. More general scientific outreach activities were designed to inform family audiences with no prior education in the relevant sciences, and these serve not only to promote IB technologies, but also to demonstrate the responsible approach to research being employed, and to address questions and concerns around the use, safety, and benefits of the technology in relatively relaxed social context. By communicating the scope and benefits of IB to civil society, the project helps to raise public awareness of these technologies. This is important, because although industry pull is increasing and helping to drive translation of biocatalysis research, additional consumer demand for the better sustainability and improved functionalities offered by these processes will also drive industrial uptake.

Overall, the aims of BIOOX have been achieved, and the technologies developed during the lifetime will find applications in industrial context. Key early successes beyond the project are likely to be through the commercialisation of tools and methods for ODE discovery and production. Equally, the body of research literature and laboratory technologies has been greatly enhanced over the course of the project, with new experimental approaches and capabilities available, alongside tools for decision making which will inform future research and development in the area. Further development will be required for the application of BIOOX processes in commercial chemical manufacturing, but planning for those efforts is already underway, and the desire to continue that development is reflected in patent applications from within the project, and further process development and evaluation activities being undertaken by the industrial partners beyond BIOOX.

The development of metagenomic resources was a critical aspect of responsible research and innovation that gained additional importance after the introduction of new legal frameworks for Access and Benefit Sharing (ABS) by the EU and member states governing the research and commercial use of “genetic resources”, i.e. the Nagoya Protocol on Access to Genetic Resources and the Fair and Equitable Sharing of Benefits Arising from their Utilization to the Convention on Biological Diversity and EU Regulation 511/2014. These frameworks aim to ensure fair and equitable sharing of benefits arising from the utilisation of genetic resources, thereby contributing to the conservation and sustainable use of genetic resources (such as the genes encoding enzymes). Development of metagenomes sampled from local environments rather than third party nations not only ensures legal compliance, but also provides the opportunity to guarantee freedom to operate by selecting sequences entirely unknown in public databases. Thus, responsible stewardship of natural resources also confers commercial advantages upon the technologies developed in BIOOX.

In 2015, the Paris Agreement on combating climate change was reached, providing ambitious targets to limit global climate change to below 2 degrees Celsius above pre-industrial levels, and preferably to no more than 1.5 degrees Celsius. The goals of the agreement will only be reached through actions such as development and implementation of low greenhouse gas emissions technologies, and reduced reliance on fossil resources by European (and global) economies. The application of industrial biotechnology processes, including biocatalysis, are absolutely key to this, since the demand for drugs, materials, and other chemical products is unlikely to significantly decrease. Biocatalytic manufacturing is inherently lower energy and can exploit lower quality but sustainable feedstocks, including waste streams from other industrial processes. The technologies developed in BIOOX will thus contribute directly to the sustainability and safety of the EU chemical manufacturing industries in the coming decades, and also help Europe to provide leadership in the global efforts to combat climate change.

List of Websites:
Website address: bioox.eu
Primary contact email address: Mark.Corbett@manchester.ac.uk
Coordinator email address: Nicholas.Turner@manchester.ac.uk

Related information

Reported by

THE UNIVERSITY OF MANCHESTER
United Kingdom
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top