Biocatalytic Oxidative Cascade for the Conversion of Fatty Acids into α‐Ketoacids via Internal H2O2 Recycling

Abstract The functionalization of bio‐based chemicals is essential to allow valorization of natural carbon sources. An atom‐efficient biocatalytic oxidative cascade was developed for the conversion of saturated fatty acids to α‐ketoacids. Employment of P450 monooxygenase in the peroxygenase mode for regioselective α‐hydroxylation of fatty acids combined with enantioselective oxidation by α‐hydroxyacid oxidase(s) resulted in internal recycling of the oxidant H2O2, thus minimizing degradation of ketoacid product and maximizing biocatalyst lifetime. The O2‐dependent cascade relies on catalytic amounts of H2O2 and releases water as sole by‐product. Octanoic acid was converted under mild conditions in aqueous buffer to 2‐oxooctanoic acid in a simultaneous one‐pot two‐step cascade in up to >99 % conversion without accumulation of hydroxyacid intermediate. Scale‐up allowed isolation of final product in 91 % yield and the cascade was applied to fatty acids of various chain lengths (C6:0 to C10:0).


Enzymes and chemicals
All chemicals, catalase (from bovine liver) and lysozyme (from chicken egg) were obtained from Sigma Aldrich (Steinheim, Germany) unless otherwise stated.
The plasmid for expression of GO-LOX enzyme was a kind gift from Prof. Ping Xu and Dr. Chao Gao (Shandong University, China). The pDB-HisGST plasmid used for cloning of P450 SPα was obtained from the DNASU plasmid repository. [1] 1.2. Biocatalysts

Cloning, expression and purification of P450 CLA (in-house number pEG306)
Cloning of P450 CLA from Clostridium acetobutylicum [2] was performed as reported elsewhere. [3] For heterologous expression, cells were grown overnight (140 rpm, 37 °C) in a 50 mL shake flask as preculture in 10 mL lysogeny broth (LB) medium supplemented with 50 µg/mL kanamycin. For preparing the main culture, 1 mL of grown pre-culture was transferred into 100 mL TB medium pre-filled into a 1 L shake flask supplemented with 50 µg/mL kanamycin and 100 µL of a sterile filtered trace elements solution [4] and incubated at 140 rpm and 37 °C. At an optical density (OD 600 ) of 0.8, the expression was induced by adding 100 µL of 100 mM IPTG solution (isopropyl β-D-1-thiogalactopyranoside) and 200 µL of 0.5 M δaminolevulinic acid (ALA) solution. After incubation for 20 h at 25 °C, the cells were centrifuged for 15 min at 3,000 x g and 4 °C. The supernatant was discarded and the pellet was frozen at -20 °C. Frozen cell pellets were resuspended in 20 mL purification buffer A (KPi, 100 mM, pH 7.0, 20% glycerol, 300 mM KCl and 50 mM imidazole). Lysozyme was added (1 mg/mL) followed by incubation at 37 °C for 2 h. Cells were finally disrupted by sonication (1 min at 30% amplitude; 2 sec on, 4 sec off, Digital Cell Disrupter, Branson, Emerson Electric). Cell debris was removed by ultracentrifugation (23,500 x g, 20 min, 4 °C). The cell free lysates were pressed through a sterile 0.45 µm Rotilabo® syringe filter (Roth, Karlsruhe, Germany) to eliminate residual particles.
Filtered cell free lysates were purified by a 5 mL His-Trap TM FF column (GE Healthcare Europe GmbH), which was washed with 50 mL water. For eluting the undesired enzymes, the column was washed with 50 mL (10 column volumes) of buffer A (without glycerol). After the washing step, P450 CLA was eluted by buffer B (KPi, 100 mM, pH 7.0, 300 mM KCl and 400 mM imidazole) and the fraction containing P450 CLA (~20 to 30 mL) was dialyzed for 36 h with imidazole-free buffer C (KPi, 100 mM, pH 7.0, 300 mM KCl) to remove residual imidazole using a dialysis tubing cellulose membrane (14 kDa cutoff, Sigma Aldrich, Steinheim, Germany). The purified protein was dialyzed three times against 300 mL buffer C (3 x 12 h) at 4 °C under continuous and slow stirring (50 rpm). Both activity and concentration of the enzyme were measured via analysis of reduced CO difference spectra as published by Omura and Sato. [5] Upon addition of 300 mM KCl, no visible precipitation occurred during the entire dialysis procedure and active P450 concentrations between 10 to 20 µM were regularly obtained without further concentration steps.

Expression and purification of GO-LOX [(R)-α-HAO] (in-house number pEG359)
The plasmid for expression of GO-LOX enzyme from Gluconobacter oxydans 621H was a kind gift from Prof. Ping Xu and Dr. Chao Gao (Shandong University, China). [6] The plasmid was transformed into E. coli BL21 (DE3) and one colony was used to grow a new pre-culture in 10 mL LB medium containing 100 µg/mL ampicillin. For preparing the main culture, 1 mL of grown preculture was transferred into 100 mL LB medium pre-filled into a 1 L shake flask supplemented with 100 µg/mL ampicillin solution and incubated at 140 rpm and 37 °C. At an optical density (OD 600 ) of 0.4-0.6, the expression was induced by adding 1 mM IPTG solution and incubation was carried out at 16 °C for further

Cloning, expression and purification of (S)-α-HAO A95G from Aerococcus viridians (in-house number pEG358)
(S)-α-HAO [7] is a FMN-dependent and (S)-specific α-hydroxyacid oxidase from A. viridans. The gene coding for the protein bearing mutation A95G was ordered (Thermo Fisher Scientific) and cloned in pET28a(+) using restriction sites XhoI and NdeI such that the protein bears a His6-tag. Oxidase activity was measured at room temperature using a peroxidase-coupled assay containing 3,5dichloro-2-hydroxybenzenesulfonic acid (DCHBS) and 4-aminoantipyrine (AAP) as chromogenic substrates (HRP-AAP/DCHBS assay) in combination with lactic acid as substrate. [8] The oxidation of lactic acid by (S)α-HAO releases H 2 O 2 , which is monitored by the assay.
FMN was added to all reaction mixtures to allow full saturation of the enzyme following purification.

Cloning, expression and purification P450 SPα from Sphingomonas paucimobilis (in-house number pEG371)
The pDB-HisGST vector harboring a GST-tag (fusion protein to enhance soluble expression) followed by a N-terminal His 6 -tag was chosen as expression vector. [1] Additionally, this vector gives the ability to cleave off the GST-tag by a TEV protease. The gene coding for P450 SPα [9] was ordered from Thermo Fisher Scientific (Germany) and inserted into pDB-HisGST vector using NdeI and XhoI restriction sites. ONCs of the transformants were prepared in 50 mL tubes (10 mL LB Medium + 50 µg/mL kanamycin) and incubated at 37 °C and shaken at 120 rpm overnight. ONCs were used to prepare glycerol stocks (500 µL culture + 500 µL 30% glycerol stock) and stored until further use (at -20 °C or -80 °C).
For growing cells, shaking flasks (volume 250 mL or 1 L) were filled with TB medium (100 mL or 330 mL) and autoclaved. After cooling to room temperature, kanamycin (50 µg/mL) and 100 µL of a sterile filtered trace elements solution [4] were added. The prepared medium was inoculated with the ONC (1 mL) and shaken at 37 °C and 120 rpm. When OD 600 reached 0.6 -0.8, the cultures were cooled down to room temperature and -aminolevulinic acid was added (0.5 mM). Cells were then induced with IPTG (0.1 mM) and the culture was shaken overnight at 20 °C and 120 rpm. The next day the cells were harvested by centrifugation (12,040 x g, 20 min, 4 °C). The cell pellets were washed with buffer (KPi, 100 mM, pH 7.4, 10 mL buffer per g pellet used) and resuspended in the same buffer (if His-tag purification was performed, lysis buffer B was used). For cell disruption, the suspension was ultrasonicated on ice [(30% amplitude, 2 sec on, 4 sec off for 2 min) x 2]. Cell debris was removed by centrifugation (18,800 x g, for 20 min, at 4 °C) and the supernatant (soluble fraction) was used for enzyme purification. The pellet was resuspended in buffer B (KPi, 100 mM, pH 7.5, 100 mM NaCl, 0.8% w/v cholate, 1 mM PMSF and 15% v/v glycerol, 10 mL buffer per g pellet used) and a sample was taken (insoluble fraction). Expression level of the enzyme was analyzed by SDS-PAGE ( Figure S3). Both activity and concentration of the enzyme were measured via analysis of reduced CO difference spectra as published. [5]  with TEV protease. [1a] SDS-PAGE results during loading, after purification and after cleavage of the enzyme are shown in Figure S4. In both reactions, ethanol was used as co-solvent. It is a co-solvent of choice for fatty acid-type substrates [3] and is not accepted by the two hydroacid oxidases, which are highly specific for 2-hydroxy acid substrates.

General procedure for extraction and derivatization on GC-GC/MS
Analysis on GC-GC/MS was performed after extraction and derivatization. Dodecanoic acid (5 mM) was employed as internal standard. Silylation was found to be the most suitable derivatization method before achiral GC and GC-MS analysis (leading to quantitative yield to corresponding silylated substrate, hydroxy and oxo-products). Silylation of final oxo-product delivered a mixture of derivatized oxo-acid and

GO-LOX and (S)-α-HAO
The  Table S2 and Figure S6. Both enzyme preparations were tested in the cascade set-up B in combination with (S)--HAO ( Figure S7). Detailed methods are provided in section 3.  Only small differences between both systems could be observed due to the removal of the tag, confirming the functional activity of the protein in both preparations and indicating that the presence of the free tag in solution was not responsible for side activities.

Conversion of octanoic acid (1) to 2-oxooctanoic acid (3) with P450 CLA /(S)--HAO
A test cascade was performed on octanoic acid (1)   Despite some variations in the recovery (attributed to analytical error), the amount of product 3 obtained from 3.2 mM H 2 O 2 corresponds to the expected theoretical value (measured 7.3 mM, with 6.8 mM corresponding to complete H 2 O 2 -recycling) with maximum allowed formation of (R)-2 (3.3 mM). In the absence of (S)--HAO, P450 CLA could efficiently and regioselectively hydroxylate octanoic acid and the amount of 2 formed equaled the maximum theoretical concentration that can be reached under these conditions (i.e. same as concentration of oxidant used, entry 8, Table S3). In the cascade set-up and independent on the amount of oxidant used, the amount of recovered intermediate 2 was found to correlate with the amount of oxidant used (entries 1-4, Table S3). This, together with the rather steady ratio 2/3 (Table   S3), confirms that both steps have reached maximum product levels after 12 h (ratio 2/3 close to theoretical 32:68). This was confirmed by calculating ee values for 2 after 15 h reaction time: starting from 10 mM 1 and 3.2 mM oxidant, (R)-2 was obtained in 88% ee (83% starting from 0.5 mM oxidant), indicating only minor amount of residual (S)-2 in the cascade reaction mixture (only low amount of (S)--HAO used), thus corroborating that the oxidation step of (S)-2 was close to completion.  (Table S4).

Testing different concentrations of GO-LOX
The effect of GO-LOX concentration was tested in cascade A using following reaction conditions: 1 mL  (Table S5). Cascade reactions were performed in closed glass-vials at 4 ºC and 170 rpm shaking for 24 h and 48 h (Table   S6). Results reveal that higher conversions were achieved at room temperature compared to 4 ºC and, as expected by using higher H 2 O 2 concentration, higher conversions were achieved at 0.5 mM.
In addition to room temperature and 4 ºC, the cascade was performed at 30 ºC as well. Due to enzyme inactivation at this temperature, no conversion was observed in all tested conditions.  (Table S7).

GC, GC-MS and NMR
Silylation was found to be the most suitable derivatization method (quantitative yield to corresponding silylated substrate, hydroxy-and oxo-products). Silylation of final oxo-product delivered a mixture of derivatized oxo-acid and derivatized enol-form of the oxo-acid ( Figure S8). final product is therefore possible (no individual standard available for the enol form and likely different response factors for both forms). Figure S8. Silylated compounds obtained from derivatization procedure 1 H and 13 C-NMR spectra were recorded using a Bruker AVANCE III 300 MHz spectrometer using a 5mm BBO probe at 300 K. Chemical shifts () are given in parts per million (ppm) relative to TMS (δ = 0 ppm) or to the residual solvent signal, and coupling constants (J) are reported in Hertz (Hz).

Determination of ee value of 2-hydroxyoctanoic acid on chiral GC
Derivatization using ethyl chloroformate/methanol was employed using the following procedure: Reaction  Figure S9. Achiral GC chromatogram of derivatized octanoic acid 1 (internal standard dodecanoic acid with t ret 11.4 min) Figure S10. Achiral GC chromatogram of derivatized 2-hydroxyoctanoic acid 2 (internal standard dodecanoic acid with t ret 11.4 min) Figure S11. Achiral GC chromatogram of derivatized 2-oxo octanoic acid 3 (internal standard dodecanoic acid with t ret 11.4 min) Figure S12. Achiral GC chromatogram obtained from cascade A using 3.0 mM H 2 O 2 as described in Table 1 of main text (internal standard dodecanoic acid with t ret 11.4 min)     Figure S22. Achiral GC chromatogram of derivatized hexanoic acid with dodecanoic acid (t ret 11.3 min) as internal standard Figure S23. Achiral GC chromatogram obtained from cascade A using 3.0 mM H 2 O 2 and hexanoic acid as substrate (as described in Table S8, entry 2) with dodecanoic acid (t ret 11.3 min) as internal standard Figure S24. Achiral GC chromatogram obtained from cascade B using 3.0 mM H 2 O 2 and hexanoic acid as substrate (as described in Table S8, entry 2) with dodecanoic acid (t ret 11.3 min) as internal standard   Table S8, entry 4) Figure S31. Achiral GC chromatogram obtained from hydroxylation step of the cascade A using 3.0 mM H 2 O 2 and heptanoic acid as substrate Figure S32. GC-MS chromatogram obtained from cascade A using 3.0 mM H 2 O 2 and decanoic acid as substrate (as described in Table S8, entry 6) with dodecanoic acid (t ret 10.3 min) as internal standard    Figure S36. Achiral GC chromatogram obtained from cascade A using 3.0 mM H 2 O 2 and decanoic acid as substrate (as described in Table S8, entry 6) with dodecanoic acid (t ret 11.3 min) as internal standard Figure S37. Achiral GC chromatogram obtained from cascade B using 1.0+1.0 mM H 2 O 2 and decanoic acid as substrate (as described in Table S8, entry 5) with dodecanoic acid (t ret 11.3 min) as internal standard