Effect of Particle Wettability and Particle Concentration on the Enzymatic Dehydration of n‐Octanaloxime in Pickering Emulsions

Abstract Pickering emulsion systems have emerged as platforms for the synthesis of organic molecules in biphasic biocatalysis. Herein, the catalytic performance was evaluated for biotransformation using whole cells exemplified for the dehydration of n‐octanaloxime to n‐octanenitrile catalysed by an aldoxime dehydratase (OxdB) overexpressed in E. coli. This study was carried out in Pickering emulsions stabilised solely with silica particles of different hydrophobicity. We correlate, for the first time, the properties of the emulsions with the conversion of the reaction, thus gaining an insight into the impact of the particle wettability and particle concentration. When comparing two emulsions of different type with similar stability and droplet diameter, the oil‐in‐water (o/w) system displayed a higher conversion than the water‐in‐oil (w/o) system, despite the conversion in both cases being higher than that in a “classic” two‐phase system. Furthermore, an increase in particle concentration prior to emulsification resulted in an increase of the interfacial area and hence a higher conversion.


Materials
For the OxdB expression in E. Coli BL21-CodonPlus(DE3)-RIL cells (Fisher Scientific) were used. Luria broth (LB)-medium, terrific borth (TB) medium, Carbenicillin, Chloramphenicol, D-lactose and D-glucose were purchased from Carl Roth and agar-agar and the salts for the preparation of the potassium phosphate buffer (K2HPO4/KH2PO4) were purchased from VWR.
n-Octanal was purchased from Sigma Aldrich and was used without further purification.
n-Octanenitrile was used as a reference compound for gas chromatography (GC) analysis and was purchased from TCI Chemicals.
For the preparation of emulsions, n-dodecane (> 99%) was purchased from Alfa Aesar and was passed twice through a basic alumina column to remove polar impurities. Water was purified through a RiOs-DI 3 UV water purification system (Merck Millipore) comprising of an ion exchange cartridge and a UV lamp. Fumed silica particles were kindly supplied by Wacker-Chemie and were used as received. Particles of different SiOH content (100%, 79%, 65%, 51%, 25% and 15%) were used. The primary particles are approximately spherical of diameter between 5 and 30 nm although they can aggregate into larger units of about 200 nm in diameter.
Hydrophilic silica was silylated to various extents by reaction with dichlorodimethylsilane in the presence of water followed by drying at 300 °C for 2 h, leaving the particle surfaces containing silanol (SiOH) and dimethylsilane (SiOSi(CH3)2) groups. [1] The silanol content was determined by titration with aqueous NaOH, while the carbon content was measured using a carbon analyzer. From a previous investigation via the immersion test, [2] it is clear that the silanol content is proportional to particle wettability and we used the former to describe it.
Ethyl acetate (EtOAc) (p.a., Fisher Chemicals) was used as received to halt the reaction under study.

Oxd expression in E. coli
E. coli BL21-CodonPlus(DE3)-RIL cells harboring the plasmids with the OxdB-gene (Table   S1) were stored as glycerol stocks at -80 °C. To prepare the pre-culture, a sample from the glycerol stock was plated on 5 mL of Luria broth (LB)-agar containing 100 g mL -1 carbenicillin and 34 g mL -1 chloramphenicol (used as antibiotic) and it was incubated for 12 -18 h at 37 °C under stirring at 180 rpm. Main cultures for the expression of OxdB were performed using TB-autoinduction medium. Sterile 20 g L -1 lactose solution in Milli-Q water (160 mL) and sterile 50 g L -1 D-glucose solution in Milli-Q water (16 mL) were added to 1,424 mL of sterile TB-medium (Carl Roth) in a 2 L Erlenmeyer flask. 100 g mL -1 carbenicillin and 34 g mL -1 chloramphenicol were added to the medium. Main cultures were inoculated with 1% (16 mL) of the OxdB pre-culture and incubated for 2 h at 37 °C and 150 rpm. Afterwards, OxdB-cultures were cultivated at 30 °C for 72 h and 150 rpm.
Cell harvest was performed at 4,000 g for 15 min and 4 °C (Thermo Scientific Heraeus multifuge X3R). The supernatant was discarded and cells were washed three times with 50 mM K2HPO4/KH2PO4 buffer at pH 7.0. The biomass was determined (bio wet weight (bww)) and cells were re-suspended in 50 mM K2HPO4/KH2PO4 buffer (pH 7.0) to a final concentration of 333 mg mL -1 cells in buffer. Cell suspensions were stored in a fridge at 4 °C or on ice before use in biotransformations.

Activity test with OxdB
Standard activity assays of Oxds in whole cells using n-octanaloxime were performed using whole cells (bww) (33 mg/mL or 0.44 mg/mL) in a total volume of 0.5 mL in 1.5 mL microreaction tubes with 100 or 10 mM oxime concentration. The reaction was conducted in 50 mM K2HPO4/KH2PO4 buffer (pH 7.0) containing 10% (v/v) ethanol as co-solvent. The activity assay was performed at 30 °C and 1,400 rpm in an Eppendorf ThermoMixer. The reaction was stopped by addition of 500 μL EtOAc and extraction of substrates and products into the organic phase. Phase separation was enhanced by pulse centrifugation to 14,000 g at room temperature.
The organic phase was analyzed by GC (Table S2).

Synthesis of n-octanaloxime
Sodium carbonate (0.75 eq) was dissolved in distilled water with 5% (v/v) of ethanol.
Afterwards hydroxylamine hydrochloride (1.5 eq) was added to the stirred solution. Octanal (1 eq) was then added dropwise. The reaction mixture was stirred at room temperature for 12 h, during which the oximes precipitated as colourless solids. The reaction progress was monitored by thin-layer chromatography (TLC). After completion of the reaction, the crude product was isolated as a colourless solid by filtration. The crude product was rinsed with distilled water (100 mL) before drying in vacuo. Pure n-octanaloxime was obtained after recrystallization from n-hexane.
NMR spectra were recorded on a Bruker Avance III 500 at a frequency of 500 MHz ( 1 H) or 125 MHz ( 13 C). The chemical shift δ is given in ppm and referenced to the corresponding solvent signal (CDCl3). They are given in the Appendix.
Accurate mass nano-electrospray ionization (ESI) measurements were performed using a Q-IMS-TOF mass spectrometer Synapt G2Si (Waters GmbH, Manchester, UK) in resolution mode, interfaced to a nano-ESI ion source. Nitrogen served both as the nebulizer gas and as the dry gas for nano-ESI. Nitrogen was generated by a nitrogen generator NGM 11. Helium  Emulsions of a 10 mM n-octanaloxime solution in n-dodecane and a E. coli cells solution in 50 mM K2HPO4/KH2PO4 buffer (pH = 7) were prepared as stated above. The concentration of E. coli cells which will be used in the emulsions prepared with silica particles will be selected from this experiment and will be that so E. coli cells are surface-inactive. Therefore, we will ensure that emulsion stability is solely given by the silica particles. The concentration range of E. coli cells studied is shown in Table S3. For clarity, concentrations are given both in the aqueous dispersion and in the emulsion. The concentration of cells in the primary solution is approximately 330 g L -1 . Therefore, a 33 g L -1 stock solution was used to prepare the above dispersions.

Preparation and characterisation of emulsions
Photos of the emulsions were taken at different times with a digital camera (Canon EOS 1100D). The emulsion type was inferred from the drop test. This consists in determining whether an emulsion drop disperses in the aqueous phase (o/w emulsion) or in the organic phase (w/o emulsion). Optical microscope images of stable emulsions were taken with an optical microscope (Olympus IX51) with a digital camera (Canon EOS 700D) one day after preparation. A drop of the emulsion cream was placed on a glass slide with a coverslip, unless stated otherwise. Emulsion stability was assessed one day and one month after preparation by measuring the volume of aqueous and organic phases released and the cream height. The fraction of aqueous phase (fw) or organic phase (fo) resolved is calculated as in Eq. (S1).
where ht is the height of organic or aqueous phase separated after some time and h0 is the height of organic or aqueous phase used to prepare the emulsion.

(c) Emulsions with silica particles
Emulsions with ~1 wt.% silica particles of different hydrophilicities (15%, 25%, 51%, 65%, 79% and 100% SiOH) were prepared under the standard procedure. The two phases were: 10 mM n-octanaloxime in n-dodecane and a 0.44 g L -1 E. coli cells (containing OxdB) solution in 50 mM K2HPO4/KH2PO4 buffer (pH = 7). Therefore, the concentration of the substrate and the cells in the emulsion was 0.082 wt.% and 0.025 wt.%, respectively. Two sets of emulsions were prepared by mixing equal volumes of the two phases: one for the study of the long-term emulsion stability and a second one to follow the conversion of the reaction at room temperature at different times. Silica particles of different hydrophilicities were dispersed by hand-shaking in either the organic (15, 25 and 51% SiOH) or the aqueous (65, 79 and 100% SiOH) phase prior to emulsification.
Emulsions with different concentrations of 65% SiOH silica particles and fixed concentration of E. coli cells (containing OxdB) (0.025 wt.%) and substrate (0.082 wt.%) were prepared to determine the influence of particle concentration on both emulsion stability and the conversion of the reaction at room temperature. The particle concentrations studied were: 0.1, 0.2, 0.5, 0.7 and 0.9 wt.%. (e) Control two-phase systems at planar interface In order to compare the biotransformation in Pickering emulsions with that in the two-phase system, the two phases containing the substrate and the E. coli cells (containing OxdB) at the same concentrations as the emulsions prepared previously and at room temperature were put together in a 14 mL screw-cap glass vial. No silica particles were added. The aqueous phase containing E. coli cells (containing OxdB) and the organic phase containing n-octanaloxime were either added as they are or they were previously homogenised with the Ultra-turrax for 2 min at 13,000 rpm. The two phases were either left to stand or stirred with a magnetic stirrer to enhance the diffusion of the substrate to the interface. The stirring was performed with a 3 x 8 mm magnetic bar and it was kept at 1,000 rpm (IKA RCT classic). Therefore, six twophase systems were studied containing equal volumes of organic and aqueous phases.

Determination of the conversion of the reaction (a) Sample preparation for GC
For the Pickering emulsions, 100 L of the emulsion were added to 400 L EtOAc. After mixing in a vortexer (VWR analog vortex mini 945304) phase separation was induced by pulse centrifugation to 14,000 g at room temperature (VWR microstar 17). 200 L of the organic phase was filtered through a small pad of silica (packed in a Pasteur pipette) into a glass vial and analysed by GC (Shimadzu GC 2010 Plus). The conversion of the reaction is defined as the area of n-octanenitrile peak over the total area (n-octanenitrile + n-octanaloxime). For the determination of conversion after work-up, the Pickering emulsion was broken by centrifugation (5 min at 5,000 rpm) (Thermo Scientific Heraeus multifuge X3R). Both phases were analysed separately. 100 L of each phase were added to 400 L EtOAc and after mixing (vortex) phase separation was induced by pulse centrifugation to 14,000 g (VWR microstar 17) at room temperature in the case of the aqueous sample. 200 L of the organic phase was filtered through a small pad of silica (packed in a Pasteur pipette) into a glass vial and analysed by GC. For the monophasic system, the whole reaction in the aqueous phase was extracted with 5 mL of EtOAc. 200 L of the organic phase was transferred to a glass vial for GC analysis.
For the two-phase system (no emulsion), both phases were analysed separately. 100 L of each phase was added to 400 L EtOAc, and mixed (vortex) by pulse centrifugation to 14,000 g at room temperature. 200 L of the organic phase was transferred to a glass vial for GC analysis.  Table S4). Measurements were conducted on a chiral SGE Analytik An initial rate was also calculated as shown below by considering that the conversion at t = 0 is 0%.

Calculation of the total interfacial area in emulsions
If emulsion droplets are considered spherical and monodisperse, then the volume of an emulsion droplet is that of a sphere 4/3πr 3 . As the amount of organic phase used to prepare the emulsion was 5 cm 3 , by dividing this volume by the volume of one sphere, the number of droplets in each emulsion could be calculated. The total interfacial area in the emulsion is equal to the surface area of one sphere (4πr 2 ) times the total number of droplets in the emulsion.

Recyclability study
An emulsion with 1 wt.% fumed silica particles (65% SiOH) was prepared in a 15 mL Falcon tube between an organic phase containing n-octanaloxime and an aqueous phase containing E.
coli cells (with OxdB). The concentration of the substrate and the cells in each phase was the same as in the previous experiments, 10 mM and 0.44 g L -1 , respectively. After homogenisation with an Ultra-turrax homogeniser at 13,000 rpm for 2 min, the emulsion was left to stand for 1 hour without stirring. Afterwards, the emulsion was centrifuged for 5 min at 20,000 g and the organic phase was separated with a Pasteur pipette and analysed by GC. The aqueous phase separated containing the cells and silica particles was re-suspended, a fresh volume of ndodecane containing n-octanaloxime (10 mM) was added and the two phases were rehomogenised. After 1 h of reaction time, the emulsion was broken by centrifugation and the organic phase analysed. Three cycles were performed in total. The recycling study was performed twice and the average conversion was determined. The average droplet diameter after 1 h was measured for the emulsion prepared in each cycle.

Controls to check the surface-activity of the substrate and the product
The surface-activity at the oil-water interface of the substrate (n-octanaloxime) and the product (n-octanenitrile) of the reaction was evaluated. Dispersions containing either n-octanaloxime or n-octanenitrile in n-dodecane at various concentrations were homogenised with a K2HPO4/KH2PO4 buffer solution with an Ultra-turrax homogeniser as explained above. After emulsification, photos of the vials were taken at different times to assess the emulsion stability. Figure S1, n-octanaloxime is not surface-active at the oil-water interface as complete phase separation is attained several min after preparation at all concentrations. The same occurs for emulsions prepared with n-octanenitrile ( Figure S2). Therefore, these two species are not active at the n-dodecane-water interface, as expected from them being small molecules possessing no amphiphilic structure.

Effect of the concentration of E. coli cells on emulsion stability
E. coli is a bacteria, whose outer membrane consists of lipopolysaccharide, phospholipid, protein and Enterobacterial common antigen. [3] Therefore, due to the presence of both hydrophilic and hydrophobic moieties within its structure, it can stabilise emulsions. Here, however, we want silica particles to be the sole emulsifier. Therefore, the concentration below which E. coli cells are not capable to stabilise an emulsion has to be identified. In order to do so, a series of emulsions were prepared by increasing the [E.  Figure S4 and Figure S5(a)). The stability of these emulsions did not change substantially after 1 month. The fraction of aqueous and organic phase released was measured one month after preparation ( Figure S5 in the emulsion were 0.082 wt.% and 0.025 wt.%, respectively. As shown in Figure S6, complete phase separation was achieved several seconds after homogenisation, showing the same behaviour as the emulsion prepared with empty E. coli cells.

Limited coalescence model of particle-stabilised emulsions
In particle-stabilised emulsions in which the initial emulsifier concentration is varied, two main régimes can be distinguished with respect to emulsion formation. [4] At low stabilizer concentration (emulsifier-poor régime), droplet interfaces are partially covered by the particulate emulsifier. As a result, droplets coalesce to a limited extent once homogenisation is halted. By increasing the particle concentration, the degree of interfacial coverage increases and the average droplet diameter decreases. This leads to an increase of the total interfacial area between oil and water and prevents further coalescence events. [5] However, at high emulsifier concentrations (emulsifier-rich régime) the interfaces are sufficiently covered by particles and the average droplet diameter does not decrease further by increasing the particle concentration. Therefore, the amount of particles available determines the final droplet surface area, as well as their packing at interfaces. Assuming all particles are spherical and become adsorbed at the interface, the surface coverage can be defined as the ratio of the interfacial area that can be covered by particles of diameter and the total interfacial area, , equal to 6 / where is the volume of disperse phase and is the drop diameter. [6] For a hexagonal close-packed monolayer of monodisperse particles, should be equal to around 0.9.
If is, where is the particle density (g cm -3 ). Then, By plotting the inverse of the droplet diameter versus the mass of the particles, a straight line is expected to be obtained. From the slope, and considering that is 0.18 g cm -3 and is 20 nm (primary particle diameter), [2] can be determined. The surface coverage calculated is 22. As it is > 0.9, this suggest that particles are closely packed at the interface and there is more than one layer of particles and some excess particles in the continuous aqueous phase.
For this calculation, the primary particle diameter has been accounted as the particle diameter.
However, particle aggregates instead of discrete particles are more likely to be present at the interface, so a lower value of surface coverage is more reasonable.

Appendix Oxd Sequences
Oxd  Table S5. Average droplet diameter of the emulsions measured 1 h after preparation and conversion of the reaction and relative activity measured from the organic phase separated after three cycles for two independent runs. The initial emulsion (1)