Enantiocomplementary Bioreduction of 1-(Arylsulfanyl)propan-2-ones

This study explored the enantiocomplementary bioreduction of substituted 1-(arylsulfanyl)propan-2-ones in batch mode using four wild-type yeast strains and two different recombinant alcohol dehydrogenases from Lactobacillus kefir and Rhodococcus aetherivorans. The selected yeast strains and recombinant alcohol dehydrogenases as whole-cell biocatalysts resulted in the corresponding 1-(arylsulfanyl)propan-2-ols with moderate to excellent conversions (60–99%) and high selectivities (ee > 95%). The best bioreductions—in terms of conversion (>90%) and enantiomeric excess (>99% ee)—at preparative scale resulted in the expected chiral alcohols with similar conversion and selectivity to the screening reactions.

Since KREDs are important elements of the toolbox of industrial biocatalysis [13][14][15][16], various forms of (S)-and (R)-selective KREDs (named also as alcohol dehydrogenases, ADHs) were explored for the production of a representative set of 1-[(substituted)arylsulfanyl] propan-2-ols.The best-performing KRED-catalyzed processes were scaled up to give the expected chiral alcohols at preparative scale.
The ketoreduction by whole-cell biocatalysts is an efficient technology because no external cofactor supplementation is needed, and the cofactor regeneration required for the biocatalysis happens within the cells.For an efficient (high-conversion) ketoreduction, an appropriate cofactor regeneration strategy is required.For the NAD(P)H regeneration in the KRED-catalyzed bioreductions, isopropanol, glucose, or lactose can serve as co-substrate providing the hydrogen to reduce the forming NAD(P) + .The nature of the co-substrate depends on the enzyme system applied for the cofactor regeneration.
The ketoreduction by whole-cell biocatalysts is an efficient technology because no external cofactor supplementation is needed, and the cofactor regeneration required for the biocatalysis happens within the cells.For an efficient (high-conversion) ketoreduction, an appropriate cofactor regeneration strategy is required.For the NAD(P)H regeneration in the KRED-catalyzed bioreductions, isopropanol, glucose, or lactose can serve as cosubstrate providing the hydrogen to reduce the forming NAD(P) + .The nature of the cosubstrate depends on the enzyme system applied for the cofactor regeneration.
Because the usefulness of various (S)-1-(arylsulfanyl)propan-2-ols was already indicated but no single-step bioreductions leading to (R)-1-(arylsulfanyl)propan-2-ols from the corresponding ketones were found, our goal was to develop high-efficiency methods for the preparation of both enantiomeric forms of 1-(arylsulfanyl)propan-2-ols using enantiocomplementary KREDs in their whole-cell forms.

Results and Discussion
In this study, four yeast strains from the Witaria culture collection The ketones with different substituents attached to the aromatic ring (2a-e) were prepared by alkylation of substituted thiophenols (3a-e) with chloroacetone (4) (Scheme 1).Racemic alcohols ((±)-1a-e) for setting up enantioselective GC were prepared by the sodium borohydride reduction of the corresponding ketones 2a-e.
The ketones with different substituents attached to the aromatic ring (2a-e) were prepared by alkylation of substituted thiophenols (3a-e) with chloroacetone (4) (Scheme 1).Racemic alcohols ((±)-1a-e) for setting up enantioselective GC were prepared by the sodium borohydride reduction of the corresponding ketones 2a-e.
The conversions of bioreductions were followed by GC analysis of samples taken from the reaction mixtures at 2, 4, 8, and 24 h reaction times (Figure 1).High enantiotopic selectivities were observed in all bioreductions, resulting in virtually enantiopure alcohols [ee > 99%; for (S)-1a-e with P. carsonii (WY1), L. elongisporus (WY2), C. norvegica (WY4), C. parapsilosis (WY12) and R. aetherivorans (S)-alcohol dehydrogenase (ReADH), and for (R)-1a-e with L. kefir (R)-alcohol dehydrogenase (LkADH)].The highest activities were observed with the WY1, WY2, and WY12 yeast strains in the cases of ketones 2a-c (80-99% conversions at 2 h reaction time).In the case of WY1, bioreduction of all ketones 2a-e resulted in products (S)-1a-e in good conversions with excellent selectivity (>99% ee) in 8 h reaction time (Figure 1a).The bioreduction of 2d was the slowest due to the highest steric hindrance of the aryl group bearing two chlorine The highest activities were observed with the WY1, WY2, and WY12 yeast strains in the cases of ketones 2a-c (80-99% conversions at 2 h reaction time).In the case of WY1, bioreduction of all ketones 2a-e resulted in products (S)-1a-e in good conversions with excellent selectivity (>99% ee) in 8 h reaction time (Figure 1a).The bioreduction of 2d was the slowest due to the highest steric hindrance of the aryl group bearing two chlorine substituents.In the case of WY2, the products formed with excellent selectivity (>99% ee) from 2a-d in good conversions within 4 h, while the conversion of 2e remained below 80% at 24 h (Figure 1b).WY4 was the least efficient but still quite selective biocatalyst among the investigated yeasts, producing (S)-1a-e (>99% ee) with the lowest conversions (Figure 1c).
The bioreductions of ketones 2a-e with enantiocomplementary recombinant alcohol dehydrogenases-the L. kefir (R)-alcohol dehydrogenase (LkADH) and R. aetherivorans (S)-alcohol dehydrogenase (RaADH)-were investigated as whole-cell biocatalysis.In the case of LkADH with (R)-(i.e., anti-Prelog) selectivity, excellent conversion and enantiotopic selectivity (c > 99%, ee > 99%) were obtained in the reduction of ketones 2a-c and 2e (Figure 1e).The steric hindrance by the dichloro-substituted phenyl moiety rendered the transformation of ketone 2d slower.The workup process of this biotransformation was also the most challenging due to the highest affinity of the hydrophobic alcohol (R)-1d to the cell constituents during extraction from the reaction mixture.Reactions with RaADH with the usual Prelog (S)-selectivity proceeded with lower conversions but also high enantioselectivities.Moreover, the non-complete conversions at longer reaction time possibly indicated product inhibition even at low substrate concentration (~10 mM)-except for the unsubstituted compound 2a with reasonable conversion in 4 h (Figure 1f).
Next, the best-performing bioreductions in terms of conversion of ketones 2a-e leading either to the (S)-or to the (R)-enantiomer of the alcohols (1a-e) (WY12 and LkADH, respectively) were performed at preparative scale to isolate and characterize the enantiopure products (Table 1).The most challenging aspect of the biotransformations was the emulsion formation during the extraction.Extraction was not feasible without previous centrifugation due to intense emulsion formation.Increasing the temperature of the centrifugation from 4 • C to 24 • C improved the efficiency of this step.Although extraction with the green solvent methyl tert-butyl ether would be desirable, the extraction by diethyl ether proved to be more efficient.The nature of forming alcohols (and the trace of residual ketones) had also a significant effect on the emulsion formations.The least intensive emulsion formation was observed during the working up of the reaction mixture of (S)-and (R)-1a with unsubstituted phenyl ring, leading to the highest preparative yields (entries 1 and 6, Table 1).In case of the other biotransformations, the handling of bioreactions with LkADH (entries 7-10) was less difficult than with the ones performed with the WY12 yeast whole cells (entries 2-5), resulting in higher isolated yields.All the isolated (S)-and (R)-products were virtually enantiopure (ee > 99%, determined by GC), having exactly opposite optical rotations.

Analytical Methods
The NMR spectra were recorded in CDCl 3 and DMSO-d 6 on a DRX-300 spectrometer (Bruker; Karlsruhe, Germany) operating at 500 MHz for 1 H, and 125 MHz for 13 C, and signals are given in ppm on the δ scale.Infrared spectra were recorded on a ALPHA FT-IR spectrometer (Bruker; Karlsruhe, Germany), and wavenumbers of bands are listed in cm −1 .Optical rotation was measured on Model 241 polarimeter (Perkin-Elmer; Shelton, CT, USA) at the D-line of sodium.The conversion from 2a-e and the enantiomeric excess of the formed alcohol (S)-1a-e or (R)-1a-e in the bioreductions were determined by gas chromatography (GC) using an 4890 GC (Agilent; Santa Clara, CA, USA) equipped with an FID detector and a Hydrodex β-6TBDM column [25 m × 0.25 mm × 0.25 µm film with heptakis-(2,3-di-O-methyl-6-O-t-butyldimethylsilyl)-β-cyclodextrin; Macherey & Nagel (Düren, Germany); H 2 carrier gas (injector: 250 • C, detector: 250 • C, head pressure: 12 psi, split ratio: 50:1)].The temperature programs and retention times are included in the Supplementary Information as Table S1.

Sampling for Determination of Conversion in Bioreductions by GC
Samples taken from the reaction mixture (50 µL) were extracted with MTBE (700 µL).The organic layer was dried over sodium sulfate (70 mg, for ~5 min), and the decanted MTBE-extracts were analyzed directly.The molar response factors for FID detection are included in the Supplementary Information as Table S2.

Sampling for Determination of Conversion in Bioreductions by GC
For the detection of the enantiomeric excess, samples taken from the reaction mixtures (500 µL) were extracted with MTBE (700 µL).After drying the organic layer over sodium sulfate 70 mg, for ~5 min), trifluoroacetic anhydride (30 mg) was added to the decanted solution and the acylation was conducted at 30 • C for 12 h to give (R)-and (S)-5a-c,e.The excess of the acylating agent was quenched with water (25 mg, for ~30 min), the organic layer was dried over dry sodium carbonate (~100 mg), and the decanted MTBE extracts were analyzed directly.In the case of (R)-and (S)-1d, derivatization was not necessary to separate the enantiomers by GC.

Synthesis of 1-(Arylsulfanyl)propan-2-ones 2a-e
To a solution of the corresponding benzenethiol (3a-e, 10 mmol) in acetonitrile (40 mL) at room temperature was added anhydrous triethylamine (25 mmol, 2.5 eq.).After cooling to 0 • C, we added dropwise to the stirred arylthiol solution a solution of chloroacetone (4, 20 mmol, 2 eq., in 10 mL of acetonitrile; ~30 min).The resulting mixture was allowed to reach room temperature, and then it was stirred at 40 • C until completion of the reaction (~8 h).After evaporating off the solvent from the reaction mixture, water (40 mL) was added to the residue and extraction was performed with dichloromethane (3 × 40 mL).The combined organic phases were washed with water (20 mL) and brine (20 mL), dried over sodium sulfate, and concentrated in vacuum.The residue was purified by column chromatography over silica gel (eluent: hexane/ethyl acetate 20:1 to 10:1) to give the desired ketone 2a-e.

Table 1 .
Results of bioreduction of 2a-e with the best KREDs at preparative scale.