Exploiting Designed Oxidase–Peroxygenase Mutual Benefit System for Asymmetric Cascade Reactions

A unique P450 monooxygenase–peroxygenase mutual benefit system was designed as the core element in the construction of a biocatalytic cascade reaction sequence leading from 3-phenyl propionic acid to (R)-phenyl glycol. In this system, P450 monooxygenase (P450-BM3) and P450 peroxygenase (OleTJE) not only function as catalysts for the crucial initial reactions, they also ensure an internal in situ H2O2 recycle mechanism that avoids its accumulation and thus prevents possible toxic effects. By directed evolution of P450-BM3 as the catalyst in the enantioselective epoxidation of the styrene-intermediate, formed from 3-phenyl propionic acid, and the epoxide hydrolase ANEH for final hydrolytic ring opening, (R)-phenyl glycol and 9 derivatives thereof were synthesized from the respective carboxylic acids in one-pot processes with high enantioselectivity.

E. coli C41 strain was used to express OleTJE, and BL21 (DE3) was used to express P450-BM3 and ANEH, while Rosetta (DE3) was used to express FDH. The KOD hot start DNA polymerase was obtained from Novagen. All chemicals were purchased from 9ding chemistry, Bide Pharmtech, Tokyo Chemical Industry, Macklin Biochemical, Thermo Fisher, Sangon, or Aladdin. Shimadzu LC-20AT high-performance liquid chromatography, Shimadzu UV-2450 UV-visible spectrophotometer were used for detection.

General procedure of directed evolution
To increase the enantioselectivity of P450-BM3 toward to the epoxidation of styrene, the same V library was screened as in past work 1 . Then we got three mutants which show enhanced (R)-selectivity (M1) and reversed (S)-selectivity (M2 and M3) compared to WT. Then these mutants were used as templates to do the second cycle of directed evolution; the same sites were selected to be scanned by phenylalanine, when M1 used as template; F4-R4, F5-R7 and F7-R5 were used to amplify the fragments 75-82-263-264, 263-264-328 and 328-437-438. To amplify the mega-primer, these fragments were mixed as templates, F4-R5were used as primers to amplify the fragments containing all the mutations, and the last step used the mega-primer to amplify the whole express plasmid. Same strategies have been applied to construct the F library when M2 and M3 were used as templates. For M2, F4-R6, F6-R7, F7-R8 were used to amplify the three fragments and F4-R8 were used to amplify the mega-primer. For M3, F4-R9, F8-R10, F9-R11 were used to amplify the three fragments, and F4-R11 were used to amplify the mega-primer. All libraries were transformed into BL21 (DE3), expression and library screening was performed as before; each library was screened considering 368 colonies. Two mutants from the library which used M2 as templates showed reversed (R)-selectivity (SO1 and SO2), and two mutants from the library which used M3 as template showed enhanced (S)-selectivity (SO3 and SO4), but better mutants were not achieved from the library which used M1 as template. To further increase the (R)-selectivity, SO2 was selected as template to do the next step direct evolution due to its high activity. Accordingly, 4 more sites were selected to do the V and F scan stimulatory, F10-R12 and F11-R13 were used to amplify 78-82-181 and 181-262 fragments, then F10-R13 were used to amplify the mega-primer. After screening 184 colonies, the best (R)-selective mutant SO5 was obtained. All genotypes of the mutants are list in Table S2.

Purification of OleTJE and SO5
After expression as described before, washing by the PBS buffer (100 mM, pH 8.0) was performed, followed by concentration to about 10-15 mL with 10% glycerin, and 300 mM NaCl (OleTJE). The lysis and ultracentrifugation procedure was the same as described before. After ultracentrifugation, the supernatant was mixed with 20 L hemin (10 mM) in ice-water for 30 minutes, then 5 mL Ni-NTA was added and the mixture was stirred for 30 minutes. The mixture was washed by buffer A (0.1 M KPi, 10% glycerin, 50 mM imidazole, pH 7.4) with more than 5 times the volume of Ni-NTA, untill the protein concentration (detected by nano-300 spectrophotometer UV-Visible Allsheng) of outflow liquid was less than 1M. Then the Ni-NTA with 6-Histag protein was combined and washed by buffer B (0.1M KPi, 10% glycerin, 500 mM imidazole pH 7.4). All the red liquid was collect in a corresponding ultrafiltration tube (OleTJE 30000 Da, SO5 50000 Da), and concentrated (5000 rpm 10 min) to no more than 2 mL. The liquid was then passed through PD-10 Columns (GE Healthcare), and eluted by buffer C (0.1 M KPi, 10% glycerol, pH 7.4) to remove the imidazole. The concentration was determined by Cheng's protocol 2 .

Lysate cascade reaction
Reactions were performed in 1.5 mL polypropylene centrifuge tubes with a final volume of 500 L. The reaction mixtures were prepared in KPi buffer (0.1 M, pH 8.0) containing 300 mM NaCl, 100 mM ammonium formate, 5 mM phenylpropionic acids 1a-j as substrates, 800 M NADP+, 2 g/L FDH, 10 g/L BM3-ANEH, 20 g/L OleT, 30 ℃, 800 rpm, 24 h on a mini shaker (MS-100, Allsheng). The reactions were quenched by adding 50 L concentrated hydrochloric acid. Substrates and products were extracted by 500 L ethyl acetate. Organic phases were blown to dryness with nitrogen, and dissolved with the same volume of isopropanol for HPLC detection. All products had the same characteristic properties as reported by Zhi Li et al 3 .

Catalase activation and inhibition experiments
Reactions were performed in 2 mL polypropylene centrifuge tubes with a final volume of 500 L.
The reaction mixtures were prepared in KPi buffer (0.1 M, pH 8.0) containing 100 mM ammonium formate, 5 mM phenylpropionic acid as substrate, 800 M NADP+, 2 g/L FDH, 5 M S7 SO5; 5 M OleTJE were involved when reaction B was performed. All mixtures were incubated in 30 ℃, 800 rpm for 12h. To evaluate the H2O2 influence, an excessive amount (>1200 U) of catalase was added to the system. Extraction and analysis of products were performed as described above.

Concentration of OleTJE and SO5-mutant detection
Freeze-dried lysate powder was dissolved in KPi buffer (0.1 M, pH 8.0) to 100 mg/mL respectively, then followed Cheng's protocol to determine the concentration of both P450s 2 .

Detecting uncoupling of SO5 experiments
Reactions were performed in 2 mL vials with a sealed lid. Freeze-dried lysate SO5 power was dissolved in 485 mL KPi buffer (0.1 M, pH 8.0) to 1 mM, and a blank was scanned by UV-vis (Allsheng, nano-300). Then 10 L NADPH solution (100 mM) was added, and gentle shaking was continued. Absorbance A1 was scanned at 340 nm. Then 5 mL of substrate styrene was added (1 M, dissolved by DMSO). Vials were sealed and shaken at 30 ℃ for 5 minutes. Absorbance A2 was scanned after shaking. Considering spontaneous degradation of NADPH, control groups were set without substrate. All reactions were performed in parallel three times. NADPH consumption was calculated by the reduced proportion of A2 to A1. Production of styrene oxide was measured by GC.

Detection methods
Conversion and enantioselectivity were determined by HPLC with an AS-H column 3 . When some of the graphics (products 2b, 2c, 2d, 2f, 2h, 2i, 2j ) failed to separate the (R)-and (S)-peaks clearly, we changed the proportion of mobile phase to 94:6 (n-hexane:IPA; v:v) for these reactions. All chiral diols have been described previously 3 in the literature, their analytical data matching ours. S8