Segmented Flow Processes to Overcome Hurdles of Whole‐Cell Biocatalysis in the Presence of Organic Solvents

Abstract In modern process development, it is imperative to consider biocatalysis, and whole‐cell catalysts often represent a favored form of such catalysts. However, the application of whole‐cell catalysis in typical organic batch two‐phase synthesis often struggles due to mass transfer limitations, emulsion formation, tedious work‐up and, thus, low yields. Herein, we demonstrate that utilizing segmented flow tools enables the conduction of whole‐cell biocatalysis efficiently in biphasic media. Exemplified for three different biotransformations, the power of such segmented flow processes is shown. For example, a 3‐fold increase of conversion from 34 % to >99 % and a dramatic simplified work‐up leading to a 1.5‐fold higher yield from 44 % to 65 % compared to the analogous batch process was achieved in such a flow process.


Protein sequences, plasmids and expressions
3.1 Imine reductase

Preparation of E. coli BL21 DE(3) whole cells with IRED from Streptomyces viridochromogenes and GDH2 from Bacillus subtilis
The preparation of the E. coli whole-cell catalyst is following the experimental procedure of Zumbrägel et. al. 1 An E. coli strain BL21(DE3), which was used for expression, and pACYCDuet-1 vector were purchased from Novagen (Madison, USA). The whole-cell catalyst was constructed as a two-plasmid-system, harbouring the gene for the glucose dehydrogenase from Bacillus subtilis in a pACYCDuet-1 vector 2 and the gene for imine reductase from Streptomyces viridochromogenes in a pET-22b(+) vector 3 . A starting culture of E. coli BL21(DE3) carrying the two recombinant plasmids was cultivated over night at 37 °C in 10 mL LB-medium, containing 80 µg mL -1 of carbenicillin and 80 µg mL -1 of chloramphenicol. The main culture was incubated in 300 mL autoinduction medium (TB-medium with 2 g L -1 lactose and 0.5 g L -1 glucose) containing 100 µg mL -

Preparation AOS and AOC2 from Arabidopsis thaliana whole cells
The gene for the AOS from Arabidopsis thaliana (AtAOS) was cloned into a pET28a(+) vector using XhoI and NcoI restriction sites. The gene for the AOC2 from Arabidopsis thaliana (AtAOC) was cloned into a pQE30 vector using BamHI and SalI restriction sites. Afterwards, plasmid-DNA was added to chemical competent cells (50 μL) and incubated for 30 minutes on ice. The cells were heated at 42 °C for 90 seconds and incubated again for five minutes on ice. Afterwards 1 mL of LB media was added. The mixture was heated for three hours at 37°C and 800 rpm. Subsequently, the cells were cultured on LB agar plates with suitable antibiotic and incubated overnight at 37 °C. A colony was isolated and transferred into an autoclaved 100 mL flask with 20 mL LB-medium and 20 μL of antibiotica. The mixture was incubated over night at 37 °C. TB-medium (200 mL) with 200 μL of antibiotic were transferred in autoclavated flasks. The medium was inoculated with 1% overnight culture. The

SI-6
cultures were grown at 37 °C and 180 rpm. When the culture reached an OD of 0.6-0.8, cell cultures were induced with 200 μL of IPTG (1M). Afterwards cells were harvested (4000x g, 4 °C, 30 minutes) and stored at 4 °C.

Org. Solvent
Conversion to amine 2 / %  Table S2: Results of the IRED-catalysed reduction of 1-methyl-3,4-dihydroisoquinoline 1 in a two-phased (methyl cyclohexane/KP i -buffer) batch system with a whole-cell concentration of 2 mg dcm mL -1 , stirring at 30 °C and 1100 rpm (mixed phases). Results shown in percentage appearing in crude reaction mixture analysed by SFC-HPLC.     For phase separation, the solution was transferred into 50 mL falcons and were centrifuged at 20 000 x g for 20 min. The organic phase was separated, while the aq. phase was extracted two more times with ethyl acetate (25 mL) and were each time centrifuged (20 000 x g, 20 min) for phase separation. The combined organic phases were dried over MgSO 4 and concentrated under reduced pressure, yielding the product with 75% (88 mg, 0.6 mmol, >99% purity). Analysis of the product succeeded via 1 H NMR spectroscopy.

Synthesis of octane nitrile in biphasic systems (batch mode)
In a glass vial, whole cells containing OxdB (100 µL, 333 mg mL -1 wet cell mass) were suspended in KP i buffer (900 µL, 50 mM, pH 7). A solution of octanal oxime (17.3 mg, 0.1 mmol, c organic : 100 mM) in cyclohexane (to 1 mL) was added at 30 °C. The reaction mixture was stirred for 2 h with a magnetic stirring bar (400 rpm) at 30 °C. The reaction was monitored by GC analysis of diluted organic phase (20 µL) of the reaction mixture with ethyl acetate (100 µL).

Synthesis and isolation of octane nitrile in biphasic systems (batch mode)
In a 50 mL pear-shaped flask, whole cells containing OxdB* (2 mL of a 333 mg mL -1 suspension) were suspended in KP i buffer (8 µL, 50 mM, pH 7) and Tween 20 (17.4 mg) was added. A solution of octanal oxime (145.1 mg, 1.01 mmol, c organic : 101 mM) in cyclohexane (to 10 mL) was added at 30 °C. The reaction mixture was stirred for 3:15 h with a magnetic stirring bar (600 rpm) at 30 °C. The poorly separated suspension was centrifuged (10 000 x g, 1 min) and the phases were separated. The solvent from the organic phase was evaporated to give the desired product (49.4 mg, 39% yield) as colourless liquid. While the aqueous phase was two times extracted with cyclohexane (each 10 mL) and centrifuged (20 000 x g, 20 min). From the combined organic phases, the solvent was removed under reduced pressure and the product 4 obtained as colourless liquid (31.8 mg, 25%). The products were analysed via GC and 1 H NMR analysis. The combined product (81.2 mg) was obtained in 64% yield (with >95% purity determined by 1 H NMR) and the conversion was determined to be >99%. *This wet-cell mass loading is higher compared to the other reactions since the activity for the freshly prepared whole-cell catalyst was found to be lower. Activities have been compared to a test reaction (see Table S8 Entry) and accordingly adjusted.

Synthesis of octane nitrile in liquid/liquid segmented flow
A syringe pump was connected to a Y-mixer (0.5 mm ID), which was connected to a tubular reactor (PFE, 0.8 mm ID, 1 mL). Reaction temperature (30 °C) was controlled by a water bath. For the organic solution, octanal oxime (0.5 mmol, 71.5 mg, c organic : 100 mM) was dissolved in cyclohexane (to 5 mL). For the aqueous solution, wet cells containing OxdB (500 µL of a 333 mg·mL -1 suspension) were suspended in KP i buffer (pH 7, 50 mM, to 5 mL). The reaction solution was collected in glass vials containing HCl solution (2 M aq. soln., 0.5 mL) for quenching of the reaction. To start the reaction, the syringe pump was set up (1 mL·h -1 , residence time: 30 min). For GC analysis, reaction mixture (20 µL) was diluted with ethyl acetate (100 µL) and analysed via GC.

Synthesis and isolation of octane nitrile in liquid/liquid segmented flow
A syringe pump was connected to a Y-mixer (0.5 mm ID), which was connected to a tubular reactor (PFE, 0.8 mm ID, 1 mL). Reaction temperature (30 °C) was controlled by a water bath. For the organic solution, octanal oxime (1.01 mmol, 144.87 mg, c organic : 101 mM) was dissolved in cyclohexane (to 10 mL). For the aqueous solution, wet cells containing OxdB* (2 mL of a 333 mg·mL -1 suspension) were suspended in KP i buffer (pH 7, 50 mM, to 10 mL). Both solutions were transferred into glass syringes and mounted on a syringe pump. To start the reaction, the syringe pump was set up (1 mL·h -1 , residence time: 30 min). Two fractions were collected. The first fraction (F1) from 1:09 -4:39 h run time, and the second (F2) from 4:40 -8:52 h. After phase separation, the solvent from both volumes was removed to give the product as colourless liquid (F1: 37.2 mg, 89% conversion; F2: 43.1 mg, 94% conversion (determined by GC), 82% crude yield (94% purity determined by 1 H NMR)). The product was analysed via GC and 1 H NMR. *This wet-cell mass loading is higher compared to the other reactions since the activity for the freshly prepared whole-cell catalyst was found to be lower. Activities have been compared to a test reaction (see Table S8 Entry) and accordingly adjusted.

Synthesis of 13-HPOT
α-Linolenic acid (5, 300 mg, 1.08 mmol) was diluted in ethanol (1 mL). The solution was dissolved in 300 mL ammoniumchloride buffer (100 mM, pH 9) and 14 mL ethanol. Lipoxygenase from glycine max (9.12 mg) was dissolved in 1 mL ammoniumchloride buffer (100 mM, pH 9). The reaction was carried out under constant oxygen stream at room temperature. Reaction was controlled via TLC. Afterwards the reaction mixture was acidified to pH 2 with hydrochlorid acid (2 M) and extracted twice with MTBE (1:1, v/v). Magnesium sulfate was added to remove residual water. The solvent was removed via rotary evaporator. 13-HPOT (280 mg, 0.90 mmol, 83%) was isolated as a yellow oil. The successful synthesis could be proven by literature.

Synthesis of 12-OPDA in batch (analytical scale)
Buffer (100 mM, pH 8) was saturated with oxygen. E.coli BL21CodonPlus(DE3)-RIL containing AtAOS and AtAOC2 (10/20/30 mg) was diluted in 1 mL saturated in buffer. 13-HPOT (7.50 mg, 0.02 mmol) was dissolved in 1 mL solvent (with or without the addition of 1 vol% Tween® 20 ). The reaction was done at room temperature, at 400 rpm and for 30 minutes. Afterwards the reaction was quenched with 2 M HCl (500 µL). The supernatant has been removed and the aqueous was extracted with dichloromethane (1 mL). The solvent was removed in vacuo and analyzed by the means of 1 H NMR-spectroscopy.  The successful synthesis has been confirmed by comparison of the analytical data with those from literature.

Synthesis of 12-OPDA in batch (preparative scale)
NaPi-buffer (100 mM, pH 8) was saturated with oxygen. E.coli BL21CodonPlus(DE3)-RIL containing AtAOS and AtAOC2 (200 mg) was diluted in 1 mL saturated in buffer. 13-HPOT (53.0 mg, 0.17 mmol) was dissolved in 5 mL isooctane with addition of 1 vol% Tween® 20. The reaction was done at room temperature, at 400 rpm and for 30 minutes. Afterwards the reaction was quenched with 2 M HCl (500 µL) and centrifuged (10.000x g, 5 min). The supernatant has been removed and the aqueous was extracted with dichloromethane (5 mL). The solvent was removed in vacuo and analyzed by the means of 1 H NMRspectroscopy. The crude product 8 was isolated with a yield of 36% (19 mg, 0.07 mmol).