Harnessing a Biocatalyst to Bioremediate the Purification of Alkylglycosides

As the world moves towards net‐zero carbon emissions, the development of sustainable chemical manufacturing processes is essential. Within manufacturing, purification by distillation is often used, however this process is energy intensive and methods that could obviate or reduce its use are desirable. Developed herein is an alternative, oxidative biocatalytic approach that enables purification of alkyl monoglucosides (essential bio‐based surfactant components). Implementing an immobilised engineered alcohol oxidase, a long‐chain alcohol by‐product derived from alkyl monoglucoside synthesis (normally removed by distillation) is selectively oxidised to an aldehyde, conjugated to an amine resin and then removed by simple filtration. This affords recovery of the purified alkyl monoglucoside. The approach lays a blueprint for further development of sustainable alkylglycoside purification using biocatalysis and, importantly, for refining other important chemical feedstocks that currently rely on distillation.


Introduction
Selective oxidation reactions are one of the most important transformations required by the chemical sector.Typically, industrial oxidation chemistry is carried out using stoichiometric quantities of often undesirable reagents, such as chromium compounds, although more attractive (and often catalytic) earth abundant metals are under development. [1]An alternative to catalytic chemical oxidation is the use of biocatalysts. [2]Enzymes are inherently sustainable, offering exquisitely chemoselective catalysis that operates under mild conditions. [3]As such, biocatalysis has transformed our capabilities and approach to synthetic chemistry, including for oxidation reactions. [4]Their mainstream use to date has been to effect key transformations for building molecules, [5][6][7][8] typically, via stepwise or combinatorial one-pot approaches.This delivers the desired transformation, and the enzyme is then removed/recycled.Herein we wanted to utilise biocatalysts in a slightly different manner, employing them instead to convert reaction by-products that are normally difficult to remove (e. g., using expensive, high energy and semi-destructive methods, such as distillation) into materials that can be easily separated from reaction product(s), effectively implementing a bioremediative purification rather than a biocatalytic synthesis (Figure 1a).
As an archetypal system and to encompass industrial relevance, we looked to manufacturing processes that traditionally use distillation to purify bio-based surfactants, a pertinent context as the world moves to develop sustainable manufacturing processes and replace traditional petroleum derived materials. [9][12] These materials have important application and market profiles, for example within the cosmetic and fermentation industries. [13,14]APGs are typically synthesised via Fischer glycosylation with the alkyl alcohol acceptor also used as the reaction solvent.Distillation processes are then employed to remove the excess alcohol, running under reduced pressure at elevated temperatures (e. g., over 120 °C at 100 mm Hg for > C4 alcohols). [15]This distillation process has been shown to consume significant energy, with 313 kWh per tonne of APG product quoted in a recent life cycle analysis of APG manufacturing. [16]Whilst there have been efforts to find enzymatic routes for glycosylation to replace the need for the chemical glycosylation step, these are plagued by inefficiency towards longer-chain alkyl donors, [12] underlining the need for improvements to other aspects of APG synthesis, including the purification process.
Accordingly, we sought an enzymatic approach to purify alkyl glycoside/long chain alcohol mixtures using biocatalysis (Figure 1b).Using an oxidase to selectively react with excess alkyl alcohol and thereby produce an aldehyde, we envisaged this material could be sequestered using an amino resin and enable a simple filtration, releasing pure alkyl glycoside and effectively demonstrating a purification mediated by enzymes.

Results and Discussion
Oxidases have received significant attention as sustainable biocatalysts, [4,17,18] typified here by an engineered choline oxidase, AcCO6 (EC 1.1.3.17),previously demonstrated to possess substrate promiscuity towards C8 through C11 primary aliphatic alcohols. [19]We thus initiated our efforts using AcCO6 for the oxidation (and ultimate removal) of octanol from mixtures with the corresponding alkyl monoglucoside.As stated, alkyl glycosides are produced using Fischer glycosylation of fatty alcohols under forcing conditions (temperature and pH). [20]Because of these conditions, the homogeneity of the product alkyl glycosides is problematic; the reaction regio-and stereodivergence is not controlled, leading to extensive product heterogeneity.In addition, APGs are often produced as mixtures, with alkyl alcohol mixtures used as reaction media, adding further to the heterogeneity of the product mixture.We viewed the excess alkyl alcohol mixtures, typically removed by reduced pressure distillation, as affording an attractive oppor-tunity to develop an alternative purification, and thus help to reduce the overall intensity of APG production.We therefore sought to establish a proof-of-concept purification using AcCO6 to facilitate alkyl alcohol removal from APG ingredients.

Development of mixture bioremediation using soluble AcCO6
Previously, only specific activity for AcCO6 was determined towards long-chain alcohols in a substrate scope panel. [19]herefore, our first steps were to establish activity of the AcCO6 mutant towards the substrates that are used in the production of AMGs.Using a colourimetric assay with horseradish peroxidase (see Supporting Information, S3), activity was established towards the panel of alcohols shown in Table 1.At this stage we also confirmed there was no AcCO6 activity towards a chemically synthesised panel of AMG standards (see Figure 1.a) Overview of component mixture purification using a biocatalyst and contrasted to traditional mixture purification approaches b) Utilising an oxidase to remove excess alkyl alcohol, followed by sequestration of the biocatalytic product aldehyde using a solid supported resin.
Table 1.Specific activities of AcCO6 towards a panel of C 8 -C 12 alcohols, where specific activity is the amount of product formed [μmol] per min per mg of enzyme (1 U is defined as 1 μmol min À 1 ).Assay conditions are described in detail in the Supporting Information, S3.

Entry
Substrate: alkyl alcohol (C n ) Specific activity (mU mg  1, entries 4-5), our focus turned to the optimisation of C 8 -C 10 chain lengths (Table 1, entries 1-3).Kinetic parameters were next determined for AcCO6 towards 1-octanol, 1-nonanol, 1-decanol, and 1-undecanol (Figure 2, Table 2).It is important to note that these kinetic parameters are apparent, due to the limited saturation of O 2 in aqueous buffer. [21]Overall, the K M values for the alcohols C8, C10, and C11 were similar, between 0.98 and 1.07 mM, with C9 being the outlier at 0.51 mM.However, the k cat values did not follow the same trend, with the k cat value for octanol being over 4 times that of nonanol and significantly higher than for the C10 and C11 alcohols (7 and 16 times, respectively).The same general trend observed in the k cat data was also observed in the k cat /K M data, which is expected based on the K M values.These results were in line with the general trend of the specific activity assay suggesting that octanol was the best substrate for AcCO6.Whilst the C12 alcohol was examined in the specific activity assay, it was not examined in the kinetics as its rate of oxidation was too slow following the same parameters as the other alcohols to generate accurate data.
Following attainment of activity and kinetic parameters, we next used GC-FID analysis on an analytical scale to screen a series of reaction conditions (Table 3).It was decided important to determine the ratio of product aldehyde to carboxylic acid, with alcohol oxidases known to be active towards aldehyde hydrates, completing over-oxidation. [22]Initially, 20 mM was chosen as the substrate concentration, and the enzyme was used as a cell free extract (CFE).Within a four-hour reaction time, a 74 % conversion of octanol to octanal was observed, and a more modest 51 % conversion observed for decanol (Table 3, entries 1&2).Notably, no over-oxidation to the acid was observed.Increasing the reaction time to 24 hours fully consumed the C 8 alcohol, in a 99 : 1 ratio of aldehyde to acid (Table 3, entry 3).This timeframe also increased C 10 alcohol conversion to 77 %, with a 10 : 1 ratio of aldehyde to acid observed (Table 3, entry 4).Increasing substrate concentration to 40 mM showed an 85 % conversion of octanol and 69 % of decanol over 24 hours (Table 3, entries 5&6).
The by-product of oxidase transformations is hydrogen peroxide; accordingly, addition of catalase can both degrade this and provide a higher concentration of soluble oxygen.Indeed, by adding catalase to the reactions at 40 mM the conversion for octanol was near quantitative, and > 75 % for decanol (Table 3, entries 7&8).Increasing substrate concentration to 60 mM still provided the C 8 and C 10 aldehydes, but conversions were lower at 66 % and 45 % respectively (Table 3, entries 9&10).Finally, we confirmed that our conditions were not optimal for C 9 or C 11 alcohol substrates (Table 3, entries 11&12).
With a set of optimal conditions, our attention next turned to testing AcCO6 against samples of alcohol mixtures, akin to those used in APG manufacture (C 8 /C 10 & C 9 /C 11 , Table 3).Pleasingly, we observed comparable conversion of C 8 and C 9 components from these mixtures, relative to their percentage composition within the mixture (Table 4, entries 1&2).There was an expected drop off in activity towards C 10 and C 11 chain lengths.The slightly lower overall activity towards C 8 and C 9 components compared with the single component testing could be derived from an inhibitory effect of the longer-chain substrates towards AcCO6.Importantly, this demonstrated that mixtures of the alkyl alcohols could be simultaneously oxidised biocatalytically, aligning to the mixtures which would be obtained from an industrial process.Assay conditions: 30 °C, 0.5 mg mL À 1 AcCO6, 0.5 mg mL À 1 HRP, 0.7 mg mL À 1 ABTS, air-saturated 100 mM KPi buffer (pH 8.0).The increase in absorbance at 420 nm was followed.Further details can be found in the supporting information.

Sequestration of product alkyl aldehydes
Removal of product alkyl aldehydes from solution was investigated next.We targeted using Schiff base formation between an amine functionalised resin (solid support) and the alkyl aldehyde product.Whilst such functionalised resins have been extensively used as a method for protein immobilisation, [23,24] there is minimal precedent for using them as a support to scavenge aldehydes. [23]We selected a commercial aminomethyl polystyrene resin and attempted extracting 40 mM octanal from an aqueous solution.The sequestration was monitored by IR spectroscopy, with a diagnostic frequency observed at 1667 cm À 1 , indicative of imine formation, alongside disappearance of the octanal carbonyl band at 1725 cm À 1 (see supporting information, S7).Analysis of the extracted aqueous phase by GC-FID also confirmed that no aldehyde remained in solution, validating the approach for further application to oxidation reaction products.The aldehyde sequestration was then applied within a batch biotransformation using AcCO6 (Scheme 1).First attempts proceeded with addition of the aminomethyl resin at the start of the reaction.However, this one-pot process caused protein precipitation and no imine was observable by IR.To ensure  inclusion of the resin did not deleteriously affect the pH of the solution resulting in precipitation of enzymes, 20 mmol of preswelled aminomethyl-functionalised polystyrene resin was added to 1 mL of 100 mM KPi buffer (pH 7).No change in pH was observed after 24 hours (see supporting information, Figure S3).In case of unwanted hydrophobic interactions between AcCO6 and the resin, the reaction was run to completion prior to addition of the resin.However, no sequestration of aldehyde was observed.As such, we filtered a completed biotransformation using a protein concentrator to afford a filtrate free of AcCO6.The resin was thus added to the aldehyde-containing filtrate and imine formation was observed, confirming sequestration of octanal.

Application of aldehyde sequestration to AMG/alkyl alcohol mixtures
As full proof of concept, we sought to demonstrate octanal removal from an AMG/alkyl alcohol mixture.Our prior confirmation that AcCO6 was inactive towards a panel of AMG synthetic standards meant we were confident that it would be selective toward a long-chain alcohol in the same solution.Pleasingly, oxidation of octanol was observable in the presence of a 90 % excess of octyl monoglucoside.Next, a full process was attempted with the aim of purifying a 50 mg mixture of octyl monoglucoside and octanol in a 9 : 1 ratio.This preparative scale reaction required three days of oxidation using AcCO6 CFE, whereafter the AMG was confirmed as unchanged by 1 H NMR, and GC-FID confirmed full oxidation of octanol to octanal (see supporting information, S6).(Scheme 2) Next, a filtrate containing octyl monoglucoside and octanal was treated with an excess of aminomethyl polystyrene resin.Following a final filtration of the sequestered aldehyde, 1 H NMR analysis of the crude sample indicated a purity of 80 % (see supporting information, S6).The by-product remained uncharacterised, but removal of the protein proved laborious within this process and would not be amenable to scale.As such, we sought to simplify the procedure through immobilisation of the enzyme.

Enzyme immobilisation and bioprocess development
Enzyme immobilisation is frequently used within industrial bioprocesses. [25]Using an immobilised enzyme effectively con-verts the protein into a heterogeneous catalyst, thus simplifying removal once it has been used.The immobilisation of AcCO6 has been demonstrated previously by us and others, [26,27] so we followed general procedures to immobilise AcCO6 onto a range of resins (see supporting information) including our own previously reported procedures for AcCO6 and other enzymes. [6,26]The resins examined included EziG and Lifetech ECR, with ECR8309F shown to be the most effective in this process.Both CFE and purified AcCO6 were immobilised seperately, with the purified form proving to be the most effective and immobilised at 1 wt%, with respect to the protein.
Validation showed that the immobilised preparation could still effectively oxidise octanol, so we next turned to oxidation in the presence of the AMG.
The process proceeded using the same conditions employed in the soluble batch approach, but using the AcCO6: ECR8039F preparation (Scheme 3).The final immobilised preparation was used at a concentration of 200 mg mL À 1 , equivalent to an enzyme concentration of ~2 mg mL À 1 .The reaction was run with a 9 : 1 ratio of AMG:octanol for 48 h, and a > 95 % conversion of alcohol was observed (by GC-FID analysis).Following this, the biocatalyst was removed from the reaction mixture and the aminomethyl resin added to sequester the aldehyde from solution.The purified solution was then lyophilised and final analysis by 1 H and 13 C NMR showed clean AMG, with a recovery yield of 86 % (average over three attempts, see supporting information for full experimental details).

Conclusions
We have successfully developed a biooxidative, reactive removal process for the bioremediation of an industrially relevant alkylmonoglucoside.Using an immobilised choline oxidase coupled with an innovative aldehyde by-product resin sequestration, we demonstrate capability to purify AMG sample mixtures.We first show that AcCO6 is capable of selectively oxidising C 8 and C 10 alcohols to their respective aldehydes in the presence of the AMG substrate, delivering 97 % conversion for C8 (octanol to octanal) and 75 % conversion for C 10 (decanol to decanal).To effectively sequester the aldehyde products following biocatalytic oxidation, heterogenous sequestration using amine-functionalised resins was used.By switching to an immobilised AcCO6, we demonstrate a proof-of-concept, twostep product removal process, enabling a clean, biocatalysis-Scheme 2. Reagents and Conditions: a) AcCO6 lyophilised CFE (110 mg.mL À 1 ), catalase CFE (0.04 mg.mL À 1 ), KPi buffer (100 mM, pH 7), 72 h, 37 °C, 200 rpm b) Filtration through filter paper 0.22 μm and protein concentrator 10k MWCO Amicon Ultra-15 c) Aminomethyl resin, orbital rotation at 18 rpm, 4.5 h, rt.based purification of AMG, with an average isolated yield of 86 %.30][31]

Figure 2 .
Figure 2. Plot of average velocity versus substrate concentration for oxidation of C8 to C11 alcohols with AcCO6.Data fitted using Hill model (OriginPro 9.6).