Stereoselective Bioreduction of α-diazo-β-keto Esters

Diazo compounds are versatile reagents in chemical synthesis and biology due to the tunable reactivity of the diazo functionality and its compatibility with living systems. Much effort has been made in recent years to explore their accessibility and synthetic potential; however, their preparation through stereoselective enzymatic asymmetric synthesis has been scarcely reported in the literature. Alcohol dehydrogenases (ADHs, also called ketoreductases, KREDs) are powerful redox enzymes able to reduce carbonyl compounds in a highly stereoselective manner. Herein, we have developed the synthesis and subsequent bioreduction of nine α-diazo-β-keto esters to give optically active α-diazo-β-hydroxy esters with potential applications as chiral building blocks in chemical synthesis. Therefore, the syntheses of prochiral α-diazo-β-keto esters bearing different substitution patterns at the adjacent position of the ketone group (N3CH2, ClCH2, BrCH2, CH3OCH2, NCSCH2, CH3, and Ph) and in the alkoxy portion of the ester functionality (Me, Et, and Bn), were carried out through the diazo transfer reaction to the corresponding β-keto esters in good to excellent yields (81–96%). After performing the chemical reduction of α-diazo-β-keto esters with sodium borohydride and developing robust analytical conditions to monitor the biotransformations, their bioreductions were exhaustively studied using in-house made Escherichia coli overexpressed and commercially available KREDs. Remarkably, the corresponding α-diazo-β-hydroxy esters were obtained in moderate to excellent conversions (60 to >99%) and high selectivities (85 to >99% ee) after 24 h at 30 °C. The best biotransformations in terms of conversion and enantiomeric excess were successfully scaled up to give the expected chiral alcohols with almost the same activity and selectivity values observed in the enzyme screening experiments.


Introduction
α-Diazo carbonyl compounds are widely recognized as versatile reagents in organic synthesis and chemical biology [1][2][3][4][5]. The chemistry of α-diazo-β-hydroxy carbonyl compounds attracts particular attention, as these functionalized compounds have been employed in the synthesis of amino acid analogues and heterocycles of biological relevance [6][7][8][9][10][11]. The conventional preparation of α-diazo-β-hydroxy carbonyl compounds relies on the aldol-type addition of a terminal diazo carbonyl compound with aldehydes or ketones mediated by a strong base such as the organometallics of lithium, magnesium, or zinc, DBU, and NaOH [12][13][14][15][16]. Asymmetric versions of the aldol reaction involving α-diazo esters and aldehydes have been reported by several authors [1,2,17,18], but the best yields and enantioselectivites were achieved by Trost and co-workers using a dinuclear magnesium complex [19,20]. Although each method possesses its own advantages, the use of basic medium generally requires Next, diazo keto esters 2a-i were chemically reduced by employing sodium borohydride (NaBH4) in tetrahydrofuran (THF) as the solvent at 0 °C for only 10 min to give the expected α-diazo-β-hydroxy esters 3a-i in moderate to high yields after column chromatography purification (62-85%, Table 1). With α-diazo-β-keto esters 2a-i and racemic α-diazo-β-hydroxy esters 3a-i in hand, robust analytical methods were developed for the determination of the conversion values in the corresponding bioreduction experiments and the measurements of the optical purity of 3a-i through the selection of HPLC techniques and a series of HPLC chiral columns (see Tables S1-S10 in the Supplementary Material).
With α-diazo-β-keto esters 2a-i and racemic α-diazo-β-hydroxy esters 3a-i in hand, robust analytical methods were developed for the determination of the conversion values in the corresponding bioreduction experiments and the measurements of the optical purity of 3a-i through the selection of HPLC techniques and a series of HPLC chiral columns (see Tables S1-S10 in the Supplementary Material).

Entry
Substrate a Conversion and enantiomeric excess values measured by HPLC.
In-house made ADHs did not perform the reduction of the azido-substituted keto ester 2a satisfactorily (see Section 2.2.1), but LB-ADH was able to reduce both chloro analogues 2b and 2c to the corresponding alcohols 3b and 3c with reasonable conversion (71%) and high selectivity (99 and 98% ee, respectively, entries 8 and 13). At this point, the influence of the alkoxy group of the ester 2c Overall, the best results in terms of activity and selectivity were obtained with the commercially available KREDs. Remarkably, for the α-diazo-β-keto esters bearing different substitutions at the γ-position (Cl for 2b and 2c, Br for 2h, and SCN for 2i), it was found that three of the ketoreductases studied (KRED-P2-D11, KRED-P2-D12, and KRED-P2-G03) displayed excellent activity (92 to >99% conversion) and selectivity values (90 to >99% ee; entries 5-7, 10-12, 25-30) towards the formation of the corresponding (R)-alcohols.
In-house made ADHs did not perform the reduction of the azido-substituted keto ester 2a satisfactorily (see Section 2.2.1), but LB-ADH was able to reduce both chloro analogues 2b and 2c to the corresponding alcohols 3b and 3c with reasonable conversion (71%) and high selectivity (99 and 98% ee, respectively, entries 8 and 13). At this point, the influence of the alkoxy group of the ester functionality was also considered, and a simple comparison between ethyl 4-chloro-2-diazo-3-oxobutanoate (2b, entries 5-7) and its corresponding methyl ester 2c (entries 10-12) demonstrated that there were not significant differences among KRED-P2-D11, KRED-P2-D12, and KRED-P2-G03, with the reductions reaching quantitative conversion and excellent selectivity in all cases.

Absolute Configuration Assignment for the Optically Active Hydroxy Esters 3a-i Obtained through the Bioreduction Process
The best biotransformations in terms of conversion and enantiomeric excess were successfully scaled-up to a 0.23 mmol scale, moving the reactions from Eppendorf tubes to Erlenmeyer flasks. The preparative reductions were carried out with seven (2a-c,e,f,h,i) out of the nine α-diazo-β-keto esters 2a-i that gave conversions higher than 80% in the optimization studies (see Table 3). In all cases, almost identical results in terms of activity and selectivity were observed for the production of 3a-c,e,f,h,i. Additional trends in the behavior of the ADHs were observed after performing the bioreduction experiments. On the one hand, Sy-ADH and ADH-A, which are considered to act following the Prelog selectivity, displayed in all cases the opposite stereopreference to anti-Prelog enzymes such as LB-ADH and evo-1.1.200. On the other hand, the commercially available KREDs led to the same enantiomers as those observed for LB-ADH and evo-1.1.200. For these reasons, and taking into consideration the priority changes when assigning the absolute configurations, the synthesis of α-diazo-β-hydroxy esters (S)-3a,d,f,g and (R)-3b,c,e,h,i is here claimed (Scheme of Table 3).

General Methods
Alcohol dehydrogenases and glucose dehydrogenase (GDH-105) were purchased from Codexis Inc. (Redwood City, CA, USA), while evo-1.1.200 ADH was acquired from Evoxx technologies GmbH (Monheim am Rhein, Germany). In-house made ADHs were overexpressed in E. coli: Rhodococcus ruber (ADH-A), Thermoanaerobacter species (ADH-T), Lactobacillus brevis (LB-ADH), Ralstonia species (Ras-ADH), Sphingobium yanoikuyae (Sy-ADH), and Thermoanaerobacter ethanolicus (Tes-ADH) [50]. d-Glucose, NADPH, NADH, β-keto esters 1, and the other reagents for the development of chemoenzymatic transformations, were acquired from Sigma-Aldrich (Madrid, Spain) and used as received. β-Keto esters 1h and 1i were chemically synthesized by known methods, exhibiting physical and spectral data in agreement with those reported in the literature [51,52]. 1 H, 13 C, and DEPT NMR spectra were recorded on a Bruker AV300 MHz spectrometer (Bruker Co., Faellanden, Switzerland). All chemical shifts (δ) are given in parts per million (ppm) and referenced to the residual solvent signal as the internal standard (for some experiments the carbon atom directly linked to the diazo functionality was not detected). IR spectra were recorded on a Jasco FT/IR-4700 spectrophotometer (Jasco-Spain, Madrid, Spain), and ν max values are given in cm −1 for the main absorption bands. Melting points were measured in a Stuart apparatus SMP3 (Bibby Sterilin, Staffordshire, UK) by introducing the samples in open capillary tubes, and the measurements are uncorrected. High-resolution mass spectra (HRMS) experiments were carried out by electrospray ionization in positive mode (ESI + ) using a VG AutoSpec Q high-resolution mass spectrometer (Fision Instrument, Milford, Massachusetts, USA). Measurement of the optical rotation values was carried out at 590 nm on an Autopol IV Automatic polarimeter (Rudolph Research Analytical, Hackettstown, NJ, USA).

Chemical Synthesis of Ethyl 4-Azido-2-diazo-3-oxobutanoate (2a)
Azido diazo ester 2a was prepared according to the one-pot procedure from commercially available ethyl 4-chloroacetoacetate (1b), as described elsewhere [37]. The product 2a was obtained as a yellowish oil (86% yield; see Table 1) and its spectral data are in agreement with those reported in the literature [37,53]

Chemical Synthesis of α-Diazo-β-keto Esters 2b-i
The synthesis of α-diazo-β-keto esters 2b-i was performed following a similar protocol to the one described in the literature [37]. t-BuNH 2 (5.0 mmol, 525 µL) was added dropwise to a solution of the corresponding β-ketoester 1b-i (5.0 mmol) and 4-acetamidobenzenesulfonyl azide (5.0 mmol, 1.20 g) in dry THF (10 mL) under inert atmosphere at 25 • C. The mixture was stirred at room temperature and monitored by TLC analysis (4:1 hexane/EtOAc). After complete consumption of the starting material, the mixture was diluted with CH 2 Cl 2 (20 mL), washed with brine (15 mL), dried over Na 2 SO 4 , filtered, and concentrated under reduced pressure. After complete removal of the solvent, the residue was triturated with diethyl ether (3 × 10 mL) and the resulting mixture was again concentrated under reduced pressure. The final solid residue was repeatedly triturated with hexane (3 × 10 mL) to separate out the insoluble ABSNH 2 by decantation. The resulting mixture was filtered, concentrated under reduced pressure, and purified by column chromatography on silica gel (4:1 hexane/EtOAc), obtaining the corresponding α-diazo-β-ketoesters 2b-i as oils (81-96% yield; see Table 1).

Chemical Reduction of α-Diazo-β-keto Esters 2a-i using Sodium Borohydride
Sodium borohydride (1.0 mmol, 37.8 mg) was added at 0 • C to a solution of the corresponding α-diazo-β-keto ester 2a-i (1.0 mmol) in dry THF (2 mL) under an inert atmosphere. The mixture was stirred at room temperature for 30 min and monitored by TLC analysis (4:1 hexane/EtOAc). Next, the reaction was quenched with an aqueous saturated solution of NH 4 Cl (1 mL); the resulting mixture was extracted with EtOAc (3 × 5 mL), and the organic phases were combined and washed with brine (1 × 15 mL). The resulting organic phase was dried over Na 2 SO 4 , filtered, and concentrated under reduced pressure. Finally, the crude product was purified by column chromatography on silica gel (4:1 hexane/EtOAc) to give the corresponding α-diazo-β-hydroxy esters 3a-i as oils (62-85% yield).    The selected ADH (12 mg) was added to a 1.5 mL Eppendorf tube containing the corresponding α-diazo-β-ketoester 2a-i (0.015 mmol), DMSO (56 µL), i PrOH (30 µL), a 10 mM aqueous solution of NADPH (60 µL), and a 50 mM Tris/HCl pH 7.5 buffer (454 µL). Then, the Eppendorf tube was closed and kept under orbital shaking at 250 rpm at 30 • C for 24 h. After this time, the product was extracted with EtOAc (3 × 0.5 mL) and the combined organic layers were dried over anhydrous Na 2 SO 4 . Finally, the solvent was carefully evaporated by bubbling nitrogen gas through it, and an aliquot was taken for the measurement of conversion and enantiomeric excess values of the corresponding α-diazo-β-hydroxy ester 3a-i by HPLC.

Bioreduction of α-Diazo-β-keto Esters 2a-i using LB-ADH
LB-ADH (12 mg) was added to a 1.5 mL Eppendorf tube containing the corresponding α-diazo-β-ketoester 2a-i (0.015 mmol), DMSO (56 µL), i PrOH (30 µL), a 10 mM aqueous solution of MgCl 2 (60 µL), a 10 mM aqueous solution of NADPH (60 µL), and a 50 mM Tris/HCl pH 7.5 buffer (394 µL). Then, the Eppendorf tube was closed and kept under orbital shaking at 250 rpm at 30 • C for 24 h. After this time, the product was extracted with EtOAc (3 × 0.5 mL) and the combined organic layers were dried over anhydrous Na 2 SO 4 . Finally, the solvent was carefully evaporated by bubbling nitrogen gas through it, and an aliquot was taken for the measurement of conversion and enantiomeric excess values of the corresponding α-diazo-β-hydroxy ester 3a-i by HPLC.

Bioreduction of α-Diazo-β-keto Esters 2a-i using ADH-A
ADH-A (12 mg) was added to a 1.5 mL Eppendorf tube containing the corresponding α-diazo-β-ketoester 2a-i (0.015 mmol), DMSO (56 µL), i PrOH (30 µL), a 10 mM aqueous solution of NADH (60 µL), and a 50 mM Tris/HCl pH 7.5 buffer (454 µL). Then, the Eppendorf tube was closed and kept under orbital shaking at 250 rpm at 30 • C for 24 h. After this time, the product was extracted with EtOAc (3 × 0.5 mL) and the combined organic layers were dried over anhydrous Na 2 SO 4 . Finally, the solvent was carefully evaporated by bubbling nitrogen gas through it, and an aliquot was taken for the measurement of conversion and enantiomeric excess values of the corresponding α-diazo-β-hydroxy ester 3a-i by HPLC.

Bioreduction of α-Diazo-β-keto Esters 2a-i using Ras-ADH
Ras-ADH (12 mg) was added to a 1.5 mL Eppendorf tube containing the corresponding α-diazo-β-ketoester 2a-i (0.015 mmol), DMSO (75 µL), an aqueous solution of GDH (60 µL, 10 U), a 50 mM aqueous solution of d-glucose (60 µL), a 10 mM aqueous solution of NADPH (60 µL), and a 50 mM Tris/HCl pH 7.5 buffer (345 µL). Then, the Eppendorf tube was closed and kept under orbital shaking at 250 rpm at 30 • C for 24 h. After this time, the product was extracted with EtOAc (3 × 0.5 mL) and the combined organic layers were dried over anhydrous Na 2 SO 4 . Finally, the solvent was carefully evaporated by bubbling nitrogen gas through it, and an aliquot was taken for the measurement of conversion and enantiomeric excess values of the corresponding α-diazo-β-hydroxy ester 3a-i by HPLC. . Then, the Eppendorf tube was closed and kept under orbital shaking at 250 rpm at 30 • C for 24 h. After this time, the product was extracted with EtOAc (3 × 0.5 mL) and the combined organic layers were dried over anhydrous Na 2 SO 4 . Finally, the solvent was carefully evaporated by bubbling nitrogen gas through it, and an aliquot was taken for the measurement of conversion and enantiomeric excess values of the corresponding α-diazo-β-hydroxy ester 3a-i by HPLC.
The selected commercially available Codexis KRED (1 mg) was added to a 1.5 mL Eppendorf tube containing the corresponding α-diazo-β-ketoester 2a-i (0.013 mmol), i PrOH (95 µL), and an aqueous solution with Na 3 PO 4 (128 mM), MgSO 4 (1.7 mM) and NADP + (1.1 mM) resulting in pH 7.0 (450 µL). Then, the Eppendorf tube was closed and kept under orbital shaking at 250 rpm at 30 • C for 24 h. After this time, the product was extracted with EtOAc (3 × 0.5 mL) and the combined organic layers were dried over anhydrous Na 2 SO 4 . Finally, the solvent was carefully evaporated by bubbling nitrogen gas through it, and an aliquot was taken for the measurement of conversion and enantiomeric excess values of the corresponding α-diazo-β-hydroxy ester 3a-i by HPLC.
The selected commercially available Codexis KRED (12 mg) was added to a 25 mL Erlenmeyer flask containing the corresponding α-diazo-β-ketoester 2a-c,e,f,h,i (0.23 mmol), i PrOH (920 µL), and an aqueous solution with Na 3 PO 4 (128 mM), MgSO 4 (1.7 mM), and NADP + (1.1 mM), resulting in pH 7.0 (8.28 mL). Then, the Erlenmeyer flask was closed and kept under orbital shaking at 250 rpm at 30 • C for 24 h. After this time, the product was extracted with EtOAc (3 × 10 mL), and the combined organic layers were dried over anhydrous Na 2 SO 4 . Finally, the solvent was carefully evaporated by bubbling nitrogen gas through it and an aliquot was taken for the measurement of conversion and enantiomeric excess values of the corresponding α-diazo-β-hydroxy ester 3a-c,e,f,h,i by HPLC. At this point, the resulting reaction crude was purified by column chromatography on silica gel (4:1 hexane:EtOAc) to give the pure corresponding α-diazo-β-hydroxy esters 3a-c,e,f,h,i as yellowish oils.

Conclusions
A family of α-diazo-β-keto esters 2a-i were chemically synthesized in good to excellent yields (81-96%), which were later successfully employed as substrates for bioreduction using a variety of E. coli overexpressed and commercially available alcohol dehydrogenases to give the corresponding optically active α-diazo-β-hydroxy esters 3a-i. Different structural patterns were considered, including the substitution at the adjacent position of the ketone group and in the alkoxy portion of the ester functionality. For most substrates, in particular, those bearing heteroatoms substituted at the 4-position (Cl, Br or SCN), the commercial enzymes KRED-P2-D11, KRED-P2-D12, and/or KRED-P2-G03 led to complete conversions to the corresponding alcohols 3 with excellent selectivities (97-99% ee). Alternatively, KRED-P1-C01 and KRED-P2-B02 were shown to be effective for the less reactive substrates, while the conversions decreased significantly when considering methyl ketones (up to 60% conversion). On the other hand, the type of the alkyl ester studied was not as relevant for the bioreduction, with methyl, ethyl, or benzyl esters being appropriate substrates to deliver the corresponding alcohols with high selectivity (85 to >99% ee) depending on the choice of the enzyme. Thus, straightforward access to the corresponding optically active α-diazo-β-hydroxy esters 3a-i has been provided. They were readily obtained in a preparative scale with high conversions and excellent stereoselectivities, depending on the enzyme selection.
Supplementary Materials: A pdf file containing the structures of α-diazo-β-keto esters 2a-i and hydroxy esters 3a-i, the development of analytical methods in HPLC for the measurement of conversion and enantiomeric excess values of the bioreduction processes, extensive enzyme-catalysed screenings, and full characterization of novel compounds by NMR spectra ( 1 H and 13 C-NMR experiments) is available on-line. Figure S1. Structures of α-diazo-β-keto esters 2a-i and the corresponding hydroxy esters 3a-i described in this contribution; Table S1. Retention times of α-diazo-β-keto esters 2a-i and their corresponding alcohols 3a-i in HPLC analyses; Tables S2-S10. Calibrate curves in HPLC for the bioreduction of compounds 2a-i; Tables S11-S18. Extensive enzyme screenings for the bioreduction of compounds 2b-i.