Simultaneous and sequential enzymatic cascades for asymmetric synthesis of chiral beta-hydroxyamide derivatives promoted by alcohol dehydrogenases and nitrile hydratases

ABSTRACT

Asymmetric synthesis of optically pure (S)-3-hydroxy-3-phenylpropanamide derivatives from o-, m- and p-substituted benzoylacetonitrile was achieved by two enzymatic one-pot protocols, the simultaneous and the sequential two-step linear cascade promoted by alcohol dehydrogenase (ADH) and nitrile hydratase (NHase). A set of commercially available ADHs and NHases were individually screened, followed by the investigation of the impact of the order of the biocatalyst addition and the effect of the substrate’s substituent. While NHases were able to hydrate o-, m-, and p-substituted electron withdrawing and electron donating substituents, the ADH was selective for the p-substituted ones. To overcome the absence of ADH catalytic activity towards beta-ketoamides, a less active NHase was employed, and the desired p-substituted products were obtained in high conversions (>99%) and ee (>99%) in the simultaneous and sequential cascade modes. Both strategies lead to optically pure (S)-3-hydroxy-3-phenylpropanamide p-substituted derivatives without the isolation of intermediates, minimizing the environmental impact and offering a greener approach.

GRAPHICAL ABSTRACT

KEYWORDS:

• Cascade reactions
• alcohol dehydrogenase
• ketoreductase
• nitrile hydratase, biocatalysis, hydroxyamide

Introduction

Optically active β-hydroxyamides are essential building blocks for synthesizing various pharmaceuticals and fine chemicals (1–5). Therefore, due to their significance, the development of their efficient synthesis has been pursued in the last decades (1, 6, 7). The low-cost, readily available, and prochiral benzoylacetonitrile derivatives (1a-i) have been a versatile starting material in synthesizing several relevant intermediates, including β-oxoamides (2a-i), β-hydroxynitriles (3a-i), and β-hydroxyamides (4a-i) (Scheme 1) (8–12).

Scheme 1. Two-step bi-enzymatic linear cascade reaction approaches combining an enantioselective ketone reduction and a non-stereoselective nitrile hydration for the synthesis of (S)−3-hydroxy-3-phenylpropanamides 4a-i from benzoylacetonitrile derivatives 1a-i.

Nevertheless, typical procedures toward the synthesis of chiral alcohols using asymmetric reduction of ketones and hydration of nitriles suffer from some drawbacks, such as the use of high atmospheric pressure, scarce heavy metals as catalysts, the need for inert atmosphere conditions and purification steps (1–14). For instance, (S)-4a was obtained through enantioselective hydrogenation of 2a using iridium as a catalyst under 15 atm pressure, inert atmosphere, after 20 h with ee 98% (13). Wessjohann and coworkers synthesized 4a as a racemic compound via the Reformatsky reaction catalyzed by chromium under an inert atmosphere (15). González-Fernández and coworkers obtained 4a from 1a in a one-pot tandem reaction using milder reaction conditions, however, in its racemic form (1, 2). Kitanosono and coworkers synthesized both (R)-3a and (R)-4a using copper salts and enantioselective boron conjugate-addition reactions with ee 81% (16).

Therefore, the search for efficient, robust, and preferably sustainable and environmentally friendly protocols to overcome these issues should be prioritized. In doing so, biocatalytic routes have the proper credentials to perform reactions in a ‘greener’ fashion: usually, such reactions are carried out at room temperature, in aqueous medium or environmentally benign solvents, and under atmospheric pressure (17–22). In addition, biocatalysts are obtained from renewable sources, are nontoxic, and potentially chemo-, regio-, and stereoselective. An interesting methodology for obtaining (R)-3a (85% yield, 97% ee) is a combined lipase-ruthenium mediated dynamic kinetic resolution in a chemoenzymatic approach (23). However, this strategy has the limitation of producing only the (R)-enantiomer due to the lipases enantioselectivity and could not be used if the compound of interest is its counterpart, the (S)-enantiomer.

Overall, the biocatalytic processes become even more sustainable when several steps of synthesis are merged in one-pot synthesis, i.e. in a cascade mode, due to step economy which avoids time-consuming, yield-reducing, and waste production during isolation and purification steps, consequently reducing the process mass intensity (PMI, total mass of material per mass of product) and costs (24–29). Cascade reactions can be classified according to various parameters such as the type of catalysts involved (bio- and/or chemocatalysts), the number of catalysts, and if the reaction steps are performed simultaneously, sequentially, or in flow and with different approaches for process optimization (30–33). Many of the elegant enzymatic cascades using different combinations of enzymes (alcohol dehydrogenases, transaminases, nitrilases, and lipases) to produce β-hydroxyacids and β-aminoacids have been described.

In this contribution, we report the use of the low-cost benzoylacetonitrile (1a) as a model substrate for the synthesis of optically pure (S)-3-hydroxy-3-phenylpropanamide ((S)-4a) by coupling an alcohol dehydrogenase – ADH (EC 1.1.1.1), which catalyzes the asymmetric reduction of ketones producing the correspondent chiral secondary alcohol (34–38) and a nitrile hydratase – NHase (EC 4.2.1.84) which catalyzes the hydration of nitriles into the correspondent primary amides without the formation of undesired carboxylic acids (39, 40). Sequential and simultaneous linear bi-enzymatic cascades were investigated focusing on the reduction and hydration reaction rates. Additionally, the stereoelectronic effects of orto, meta, and para electron withdrawing and donating substituents in the aromatic ring were investigated to broaden the substrate scope.

Results and discussion

Prior to enzymatic investigations chiral chromatographic separations were done by GC-FID or HPLC. The compounds 2a-i were obtained from acid catalysis (H₂SO₄ in CH₂Cl₂) with around 40% yield in a mixture with the respective carboxylic acid (monitored by TLC and stained with bromocresol green) (41). As expected, the use of strong acids to hydrate nitriles is problematic due to the harsh conditions, low selectivity, and consequently low yields caused by the kinetically favored hydrolysis of the amides into carboxylic acids (2, 42). Also, during the neutralization step, there was extensive formation of salt, resulting in a significant environmental footprint. Moreover, dichloromethane is not an environmentally friendly solvent – it can be ozone-depleting, is carcinogenic, and should be avoided (43). Compounds rac-3a-i were synthesized from 1a-i using NaBH₄/MeOH reduction protocol in a range of 75–84% yield. The subsequent hydration of rac-3a, rac-3g-i to afford the correspondent β-hydroxyamide rac-4a, rac-4g-i was performed under basic catalysis (K₂CO₃) and hydrogen peroxide (34% aqueous solution) (15). After chromatographic purification, the product was recovered in 14% yield. As observed by Wessjohann et al., an inconvenience of this methodology is that during the work-up step considerable mass of the product is lost, due to its great affinity for the aqueous phase. The use of concentrated acid (H₂SO₄) to hydrate β-hydroxynitriles, such as rac-3a, rac-3g-i, is not a viable choice because of the competing elimination reaction, resulting in unsaturated byproduct formation and consequently lower chemical yields of the desired β-hydroxiamide.

As can be seen from these results, the hydration of nitriles to amides still presents a challenging task in organic synthetic chemistry and the search and development for more efficient and environmental methodologies is crucial and of practical use. In order to circumvent these issues, the enzymatic nitrile hydration method was explored.

We began our studies by screening a panel of 13 commercially available NHases from Prozomix^® on the hydration of the model substrates 1a and rac-3a (Tables 1 and 2). Having performed screening for both Fe and Co-type NHase in phosphate buffer solution (pH 7.0, 200 mM), 25°C, 24 h, it has been found that for compound 1a conversions >90% were achieved for all Co-type NHases, except for NHase 010 and 016, which gave low and moderate conversions (8% and 44%, respectively). On the other hand, both Fe-type NHases 012 and E0256 evaluated (Table 1, entries 12 and 13, respectively) afforded the lowest substrate conversions. These results agree with the literature data, where it is observed that Co-type NHase preferentially hydrates aromatic nitriles, whereas the Fe-type exhibits a high affinity for arylaliphatic nitriles due to the differences in their substrate binding pocket (44).

Table 1. NHase screening for the hydration of substrate 1a.

Entry	NHase	Cofactor	2a (%)^[a]
1	003	Co	>99
2	004	Co	94
3	008	Co	90
4	010	Co	8
5	011	Co	>99
6	014	Co	>99
7	015	Co	>99
8	016	Co	44
9	018	Co	93
10	E0257	Co	>99
11	E0259	Co	>99
12	012	Fe	6
13	E0256	Fe	0

Note: Reaction conditions: substrate 1a (1 mg; 7 μM), DMSO (100 μL), sodium phosphate buffer pH 7.0, 200 mM (1.0 mL), enzyme (5 μL), 24 h. [a] Conversions were determined by GC-FID analysis.

Table 2. NHase screening for the hydration of substrate rac-3a.

Entry	NHase	Cofactor	rac-4a (%)^[a]
1	003	Co	>99
2	004	Co	>99
3	008	Co	>99
4	010	Co	0
5	011	Co	>99
6	014	Co	>99
7	015	Co	>99
8	016	Co	97
9	018	Co	>99
10	E0257	Co	>99
11	E0259	Co	>99
12	012	Fe	0
13	E0256	Fe	0

Note: Reaction conditions: substrate rac-3a (1 mg; 6 μM), DMSO (100 μL), sodium phosphate buffer pH 7.0, 200 mM (1.0 mL), enzyme (5 μL), 24 h. [a] Conversions were determined by GC-FID analysis.

As is shown in Table 2, a similar profile was observed using this NHase toolbox for compound rac-3a, except for NHase 010 (Table 1, entry 4), and all the Co-type NHase tested produced the desired amide in conversions at >99%. However, for compound 1a where Fe-type NHases showed low substrate rate conversion, for compound rac-3a none of them were able to recognize this substrate and a quantitative amount of the starting material was recovered.

It is worth noting that for compound rac-3a no stereoselectivity was observed with all the evaluated NHase. It is proposed that NHase possesses a spacious ‘pocket’ able to accept a wide range of substrates but with zero or low enantioselectivity (45).

In parallel with NHase enzyme assays, three ADHs from Sigma-Aldrich (Merck) were screened for the asymmetric reduction of substrates 1a and 2a. The ADH enzyme from the yeast S. cerevisiae and horse liver are NADH-dependent whereas the recombinant one from E. coli is NADPH-dependent. Regeneration of cofactors plays an important role in economic viability of such processes. In this work isopropanol (IPA) was chosen as a recycling system for NAD(P)H (46). The regeneration of NADP(H) coupled with the oxidation of IPA to acetone makes the process more economical since a second enzyme, such as glucose dehydrogenase (GDH), is not required to regenerate the cofactor (47). Therefore, the great advantage of IPA over the glucose/GDH system for cofactor recycling is the cost. In addition, IPA is rated as a green solvent, it is not necessary to have the tight pH control required for enzymatic recycling systems and can be used both as auxiliary in the cofactor recycle and organic cosolvent facilitating substrate solubilization (43, 44, 48). As is shown in Table 3, both ADH from S. cerevisiae and horse liver did not show any activity towards substrate 1a. Fortunately, the recombinant ADH catalyzed the asymmetric reduction of 1a into (S)-3a with conversion >99% and ee >99%, after 24 h, and addition of 5% of IPA. Our procedure improves on previous methods. Bisogno and coworkers used the racemic octan-2-ol as cosolvent and auxiliary in the cofactor regeneration of NAD(P)H and obtained the product (S)-3a from 1a in 78% yield and ee >99%, after 24 h (49). Ankati and coworkers also obtained (S)-3a with ee >99% and 83% yield using glucose/GDH system after 48 h (50).

Table 3. ADH screening for the reduction of substrate 1a.

Entry	ADH	(S)-3a^[a]	ee (%)^[b]
1	From Saccharomyces cerevisiae	0	–
2	From Horse liver	0	–
3	Recombinant from E. coli	>99	>99

Note: Reaction Conditions: substrate 1a (1 mg; 6.9 μM), IPA (50 μL), sodium phosphate buffer (1.0 mL, containing 1.0 mM of Mg²⁺), enzyme (1 mg or 1 μL), NAD(P)H (0.5 mg; 0.67 μM). Reaction time was 24 h. [a] Conversions were determined by GC-FID. [b] Enantiomeric excess was determined by chiral GC-FID analysis and absolute configuration was assigned by comparison of elution order on HPLC with reported ones (8).

Regarding enzyme screening for substrate 2a, no conversions were observed with all three evaluated ADHs in the same reaction conditions used for substrate 1a.

With these results in hand, we started to investigate the cascade reaction’s modes. Our preliminary screening with ADH’s revealed that only the recombinant ADH from E. coli could recognize substrate 1a. On the other hand, for NHases several of them could be employed. The criterion used to choose NHase E0257 was its greater availability in our lab. Our first effort to obtain the pure enantiomer (S)-4a from 1a was through a simultaneous cascade reaction approach, since there were no buffer, salts, pH-values, and reaction temperature incompatibility between both enzymes (ADH and NHase).

The formation of 2a was observed within the first hour in 92% conversion and after two hours no starting material remained. As expected, even after 24 h, 2a was the only product detected for this reaction condition, since recombinant ADH is not active towards it (data mentioned previously in the screening of ADH).

Knowing that the recombinant ADH converted substrate 1a into (S)-3a and, NHase E0257 is a non-selective enzyme and can react with substrates 1a and (S)-3a, our second effort was the sequential cascade reaction approach.

To find out how long it would take for 1a to be completely consumed in this first stage of the cascade, the reaction was monitored at each hour. After six hours, it was observed that 1a was 90% consumed, affording (S)-3a with quantitative conversion in eight hours. We also confirmed that the stereoselectivity was maintained during the conversion of compound (S)-3a into compound (S)-4a (ee >99%). In this way, the second step of this sequential cascade reaction was initiated after the first eight hours by adding NHase E0257 to the same Eppendorf tube. The compound (S)-3a was completely converted into (S)-4a after two hours. Therefore, within 10 hours in total, (S)-4a was obtained in a mild reaction condition with conversion >99% and ee >99% in both reaction steps (Scheme 2).

Scheme 2. Sequential biocatalytic cascade reaction of ketone reduction and nitrile hydration for the synthesis of (S)−4a from 1a. Enantiomeric excess was determined by chiral GC-FID analysis and absolute configuration was determined by comparison with optical rotation data in the literature (12).

Although the performance of the sequential cascade mode was quite good, we still aimed to improve it, turning the envisioned cascade into a simultaneous one. The design of this second attempt to perform a simultaneous cascade was based on the hydration reaction rate. As can be seen in Table 1 (entries 8 and 10), NHase 016 converts 1a into 2a more slowly when compared to NHase E0257. This means that if we combined the recombinant ADH and NHase 016 simultaneously at time zero, the reduction reaction rate would be higher than the hydration one, yielding preferentially (S)-3a, and the formation of 2a would be zero or minimum. As soon as compound (S)-3a has been formed, its conversion into (S)-4a could be catalyzed by NHase 016. These results prompted us to perform the reaction in a simultaneous cascade mode. In doing so, after 24 h it was observed the formation of compound (S)-4a (82.5% conversion and ee >99%), compound 2a (8.6%) and the remaining intermediate (S)-3a (8.9%) (Scheme 3).

Scheme 3. Simultaneous biocatalytic cascade reaction of ketone reduction and nitrile hydration for the synthesis of (S)−4a from 1a. Enantiomeric excess was determined by chiral GC-FID analysis and absolute configuration was determined by comparison with optical rotation data in the literature (12).

Based on these results and aiming to extend the application of the cascade reaction, the stereoelectronic effect of o-, m-, and p-electron withdrawing and electron donating substituents on the aromatic ring was investigated using the optimized experimental conditions for the model reaction with substrate 1a. As shown in Table 4, a screening with the two pre-selected NHase 016 and E0257 with substrates 1b-i was performed and conversion rates >99% were obtained for all meta substituents with both NHases. The results for the para substituents followed this trend, except for 2 h and NHase 016 which afforded a conversion rate of 36%. On the other hand, for orto substituents no generalization could be drawn. These results suggest that the sterically effects have a pronounced influence on the reactivity while the electronic effects did not.

Table 4. NHase screening for hydration of substrates 1b-i.

Entry	Product	R	NHase E0257 Conv. (%)^[a]	NHase 016 Conv. (%)^[a]
1	2b	o-CH3	82	71
2	2c	o-Br	67	0
3	2d	o-OCH3	>99	26
4	2e	m-CH3	>99	>99
5	2f	m-Br	>99	>99
6	2g	p-CH3	>99	>99
7	2h	p-Br	>99	36
8	2i	p-OCH3	>99	>99

Note: Reaction conditions: substrates 1b-i (1 mg), DMSO (100 μL), sodium phosphate buffer pH 7.0, 200 mM (1.0 mL), enzyme (5 μL), 24 h. [a] Conversions were determined by GC-FID analysis.

The screening of commercial ADH recombinant from E. coli (Sigma Aldrich/Merck) with substrates 1b-h revealed that, as expected, stereoelectronic effects are notable. Among the tested substrates, it was observed high conversions and ee only for the p-electron withdrawing and donating substituents 1g-i (Table 5). The evaluated ADH was inactive towards amides 2a-i (data not shown).

Table 5. ADH screening for the carbonyl reduction of substrates 1b-i.

Entry	Product	Conv. (%)^[a]	ee (%) (S)
1	2b	0	–
2	2c	0	–
3	2d	0	–
4	2e	0	–
5	2f	8	>99
6	2g	>99	>99
7	2h	>99	36
8	2i	>95	>99

Note: Reaction Conditions: substrates 1b-i (1 mg), IPA (50 μL), sodium phosphate buffer (1.0 mL, containing 1.0 mM of Mg²⁺), enzyme (1 mg or 1 μL), NADPH (0.5 mg; 0.67 μM). Reaction time was 24 hours. [a] Conversions were determined by GC-FID. [b] Enantiomeric excess was determined by chiral HPLC or GC-FID or analysis. The absolute configuration were determined by comparison with literature optical rotation data or order of elution on HPLC.

Based on the results from individual enzymatic screening of NHases and ADHs, the experiments of simultaneous and sequential cascades reactions were performed with the p-substituted substrates 1g-i and the enzymes NHase 016, NHase E0257, and ADH, as shown in Table 6.

Table 6. Simultaneous and sequential biocatalytic cascade reaction of ketone reduction and nitrile hydration for the synthesis of (S)−4a, 4g-i from 1a, 1g-i.

R	Code	NHase	Cascade mode	2a,g-i Conv (%)	3a,g-i Conv (%)	4a,g-i Conv (%)
p-H	A	E0257	Sequential	0	0	>99
		016	Sequential	0	3	97
		016	Simultaneous	8.6	8.9	82.5
p-CH3	g	E0257	Sequential	0	0	>99
		016	Sequential	0	13	87
		016	Simultaneous	4	0	96
p-Br	h	E0257	Sequential	0	0	>99
		016	Sequential	0	0	>99
		016	Simultaneous	0	0	>99
p-OCH3	si	E0257	Sequential	0	5	95
		016	Sequential	0	0	>99
		016	Simultaneous	0	0	>99

The best cascade mode for the substrate 1 g was the sequential one using ADH-NHase E0257 which afforded the desired hydroxyamide (S)-4 g in >99% conversion and 99% ee in 24 h. When the system ADH-NHase 016 was used as sequential cascade, 13% of the intermediate (S)-3 g remained in solution in 24 h and in the simultaneous cascade 4% of the intermediate 2 g remained unreacted. For the electron-withdrawing substituent 1 h all cascade systems were equally efficient affording the product (S)-4 h in >99% conversion and 99% ee. Finally, for the electron-donating substituent 1i both simultaneous and sequential cascade modes using ADH-NHase 016 afforded the product (S)-4i in >99% conversion and 99% ee. Unfortunately, as the whole amino acid sequence for the evaluated commercial enzymes are undisclosed it was not possible to model docking experiments to rationalize the observed results. Nevertheless, it was possible to establish efficient cascade reaction modes for all the evaluated p-substituted derivatives demonstrating the potential of this synthetic strategy. This could be extended to the enantiocomplementary bioreduction product using a (R)-selective ADH.

Conclusions

In this study, we established both a one-pot simultaneous and one-pot sequential linear bi-enzymatic cascades with (S)-selective ADH and NHase enzymes to synthesize various substituted chiral 3-hydroxy-3-phenylpropanamide derivatives with high conversions (>99%) and ee (>99%). Overall, the stereoelectronic effects of substituents played an important role in ADH catalyzed carbonyl reduction but were less significant in NHase hydration reactions. In addition, the choice between the simultaneous or sequential cascade approaches depended on the reaction rate of nitrile hydration, since the ADH used herein was unreactive toward β-oxo- or β-hydroxyamides, while it readily converted both β-oxo- and hydroxynitriles. Both cascade modes completely remove intermediates workups and isolations therefore, resulting in reductions in PMI, solvent usage thus providing a more sustainable process compared with the ‘standard’ ones.

Experimental section

Chemicals and commercial enzymes

Substrates 1a-i and all three ADH enzymes were purchased from Sigma-Aldrich Co. (St. Louis, USA). The ADH recombinant from E. coli (catalogue number 49641) showed a specific activity of ≥500 U/mL. The kit of NHase enzymes was purchased from Prozomix Limited (Northumberland, UK). Solvents were used without further purification.

Instrumentation

Melting points were taken using a MQAPF – 302 apparatus. Precoated TLC plates of silica gel 60 UV₂₅₄ in aluminum support from Macherey-Nagel were used while for preparative thin layer chromatography (20 × 20 cm; 0.75 mm thinner), Kiesegel DF silica gel from Riedel-de Haën was applied. Compounds were visualized under ultraviolet irradiation in 254 nm or stained with bromocresol green. For purification by dry-column flash chromatography, silica 230–400 mesh from Macherey-Nagel was used. For purification by preparative thin-layer chromatography, Kieselgel DF silica from Riedel-de Haën at 0.75 mm, on glass plates of 20 × 20 cm, was used. Enzymatic and non-enzymatic reactions were monitored by GC-FID analysis on a Shimadzu – GC2010 Plus gas chromatograph (GC) equipped with automatic injector AOC – 20i coupled to flame ionization detector (FID). Hydrogen carrier gas was at 1.17–1.22 mL/min with constant flow mode in a Restek® RTX-5 (30 m × 0.25 mm × 0.25 μm, 5% phenyl, and 95% dimethylpolysiloxane) capillary column (Column A). The enantiomeric excess (% ee) of biocatalytic reduction products were determined by gas chromatograph (GC) equipped with automatic injector AOC – 20i coupled to flame ionization detector (FID) GC-FID and high-performance liquid chromatography (HPLC). For chiral GC-FID analyses Hydrodex^®β−3P (25 m × 0.25 mm × 0.25 μm) capillary column from Macherey-Nagel (Colum B) and Lipodex^® E (25 m × 0.25 mm × 0.25 μm) capillary column from Macherey-Nagel (Colum C) were used. Injector and detector were set at 260^°C and 280°C, respectively. Injection was in split mode (100:1) and injection sample volume of 1 μL on a concentration 0.5–1.0 mg/mL. Oven temperature settings are described in the Supporting information. The chiral HPLC analyses were performed on Agilent^® 1260 Infinity Quaternary LC, equipped with a photodiode array detector (PAD) G4212B and JASCO^®, LC-NET II/ADC equipped with a photodiode array detector (PAD) MD-2018PLUS using columns packed with chiral stationary phases as follows: Chiralcel^® OJ-3 (15 × 0.46 cm × 3 μm) (Column D), Chiralpak^® IB (25 × 0.46 cm × 3 μm) (Column E), Chiralpak^® IC (25 × 0.46cm × 3 μm) (Column F) equipped with dedicated pre-columns (4 mm × 10 mm, 5 μm). The respective mixtures of n-hexane/2-PrOH were used as mobile phases in the appropriate ratios; the HPLC analyses were executed in an isocratic, and isothermal (25^°C or 30^°C) mode; flow is given in mL/min; the method conditions for the resolution of the appropriate racemates are shown in Supplementary information. GC-MS analyses were recorded on an Agilent 7890B gas chromatograph (GC) coupled to a 5977A mass spectrometer (MS). Helium carrier gas was set at 1 mL/min with constant flow mode in a HP-5 MS capillary column (30 m × 0.25 mm×0.25 μm; 5% phenyl and 95% dimethylpolysiloxane), injector at 260^°C set in split mode (100:1), injection in the volume of 1 μL and sample concentration in the range of 0.5–1.0 mg/mL. The GC oven temperature programming is described in Supplementary information. Mass spectrometer transfer line into the quadrupole was set at 280^°C, ionization voltage of 70 eV and mass spectra were obtained in full scan mode (40–400 m/z). High resolution mass spectrometry (HRMS) analyses were performed on a Xevo G2-XS QTof mass spectrometer (Waters, Milford, MA, USA) equipped with the electrospray ionization source (ESI) in positive and negative ionization modes. Attenuated total reflectance-Infrared (ATR-IR) spectra were recorded on a Bruker Vertex 70 spectrometer with FT spectral range from 400 to 4000 cm⁻¹. ¹H and ¹³C NMR spectra were carried out in CDCl₃ (with TMS as an internal standard), DMSO-d6 or in CD₃OD on a Bruker 300 Fourier (7.1 T) spectrometer. Chemical shifts (δ) are recorded in ppm and spin–spin coupling constant (J) in Hz. Multiplicities are reported by the following abbreviations: s = singlet, d = doublet, t = triplet, m = multiplet, q = quartet, and br s = broad singlet.

Synthesis

General procedure for the synthesis of 3-oxo-3-phenylpropanamides, 2a-i

Sulfuric acid 98% (1.7 mL) was added slowly to 1a-i (1.0 mmol) and dissolved in anhydrous dichloromethane (4.0 mL), under magnetic stirring at room temperature. The reaction was monitored by GC-FID. After 6 h, the reaction mixture was cooled to 0^°C and quenched with NaOH (10 mol/L) up to pH 12.0 followed by repeated (3 times) extractions with equal volumes of ethyl acetate (10.0 mL). The combined organic extracts were washed with brine and water and dried over anhydrous MgSO₄ followed by solvent evaporation under reduced pressure. The crude products were resuspended in MeOH or EtOH and analyzed by GC-FID.

General procedure for synthesis of (±)-3-hydroxy-3-phenylpropanenitriles, (±)-3a-i

Sodium borohydride (62.4 mg, 1.65 mmol) was added to 1a-i (0.50 mmol) dissolved in ethanol (5.0 mL), under slow magnetic stirring at 0^°C with gradual warming. The reaction was monitored by TLC (CH₂Cl₂/n-hexane/EtOAc 4:1:1). After 45 min, the reaction mixture was evaporated under reduced pressure and the crude product resuspended in brine followed by repeated (3 times) extractions with equal volumes of ethyl acetate (15 mL). The combined organic extracts were dried over anhydrous MgSO₄ followed by solvent evaporation under reduced pressure affording, (±)-3a-i. The crude products were resuspended in ethyl acetate, MeOH, or EtOH and analyzed by GC-FID.

General procedure for synthesis of (±)-3-hydroxy-3-phenylpropanamides, (±)-4a, 4g-i

Hydrogen peroxide 34% (190.0 μL) and K₂CO₃ (0.172 mmol) were added to a (±)-3a, 3g-i (0.58 mmol) dissolved in DMSO (580.0 μL), under magnetic stirring at room temperature. The reaction was monitored by GC-FID. After 6 h, brine was added (1.0 mL) and followed by repeated (3 times) extractions with equal volumes of ethyl acetate (2.0 mL). The combined organic extracts were dried over anhydrous MgSO₄ followed by solvent evaporation under reduced pressure. The crude products were resuspended in ethyl acetate, MeOH, or EtOH and analyzed by GC-FID.

General screening procedure for ADHs

Reactions were performed in 2.0 mL Eppendorf tubes containing 1a-i or 2a-i (∼ 1.0 mg; 6–7 μM), IPA (50.0 μL) and 500.0 μL of phosphate buffer solution (pH 7.0; 100 mM, enriched with MgCl₂ (1.0 mM) and NAD(P)H 0.5 mg/mL). This solution was mixed in a thermomixer at 25^°C; 1000 rpm for 2 min. In a second Eppendorf tube, ADH (1.0 μL; Enzyme activities: ADH from S. cerevisiae 300 U/mg; ADH from horse liver 1.8 U/mg; ADH recombinant from E. coli 10 U/mL. One unit corresponds to the amount of enzyme which reduces 1.0 μM of nonanone per min at pH 7.0 and 25^°C) was solubilized in phosphate buffer solution (500.0 μL) and stirred at 25^°C; 1000 rpm for 2 min. After that, the enzymatic solution was added to substrate solution Eppendorf tube and the reaction was allowed to proceed for 24 h. The reaction mixture was centrifuged (12.000 rpm/4 min) and the supernatant was saturated with NaCl and extracted with ethyl acetate (3 × 500.0 μL). The combined organic extracts were dried under anhydrous MgSO₄ and solvent was evaporated. The crude product was resuspended in ethyl acetate, MeOH, or EtOH and analyzed by GC-FID. The ee values were determined by chiral GC or HPLC analysis. The absolute configuration was determined by comparing the sign of optical rotation or the elution order on HPLC with literature data.

General screening procedure for NHases

Reactions were performed in 2.0 mL Eppendorf tubes containing sodium phosphate buffer (1.0 mL; pH 7.0; 200 mM) enriched with CoCl₂ˑ6H₂O (1.0 μL, stock solution 2.0 mg/500 μL) or FeCl₃ˑ6H₂O (1.0 μL, stock solution 2.0 mg/500.0 mL), depending on the NHase specificity and NHase (5.0 μL). Solutions were thoroughly mixed with a pipette and placed in an ice bath for 1.5 h under natural light irradiation (for Fe-type photoreactive NHase) followed by stirring and warming at 1000 rpm, 25^°C in a thermomixer for 2 min. A DMSO (100.0 μL) solution of 1a-1i or (±)-3a-i (∼1.0 mg; 6–7 μM) was added to the enzymatic solution. After 24 h the reaction mixture was extracted with ethyl acetate (3 × 500.0 μL). The organic combined extract was dried over anhydrous MgSO₄ and evaporated. The crude product was resuspended in ethyl acetate, MeOH, or EtOH and analyzed by GC-FID. The ee values were determined by chiral GC or HPLC analysis. The absolute configuration was determined by comparing the sign of optical rotation or the elution order on HPLC with literature data.

General procedure for the sequential cascade reactions

Reactions were performed in 2.0 mL Eppendorf tubes containing 1a-i (1.0 mg; 7.0 μM), IPA (50.0 μL) and 500.0 μL of phosphate buffer solution (pH 7.0; 100 mM, enriched with Mg²⁺ (1.0 mM), NAD(P)H 0.5 mg/mL and CoCl₂ˑ6H₂O (1.0 μL, stock solution 2.0 mg/500.0 μL)). This solution was mixed in a thermomixer at 25^°C; 1000 rpm for 2 min. In a second Eppendorf tube, ADH (1.0 μL) was solubilized in phosphate buffer solution (400.0 μL) and stirred at 25^°C; 1000 rpm for 2 min. After that, the ADH enzymatic solution was added to substrate solution Eppendorf tube and the reaction was allowed to proceed for 8 h. Then, NHase E0257 or NHase 016 (5.0 μL) solubilized in phosphate buffer (100.0 μL, pH 7.0; 100.0 mM) was added to the reaction mixture. After 2 h, the reaction mixture was centrifuged (12.000 rpm/4 min) and the supernatant was saturated with NaCl and extracted with ethyl acetate (3 × 500.0 μL). The combined organic extracts were dried under anhydrous MgSO₄ and solvent was evaporated. The crude product was resuspended in ethyl acetate, MeOH, or EtOH and analyzed by GC-FID. The ee values were determined by chiral GC or HPLC analysis. The absolute configuration was determined by comparing the sign of optical rotation or the elution order on HPLC with literature data.

General procedure for the simultaneous cascade reactions

Reactions were performed in 2.0 mL Eppendorf tubes in the same experimental conditions used for sequential cascade reactions, except that NHase E0257 was replaced by NHase 016 and both enzymes (NHase 016 and recombinant ADH) were added simultaneously, at the beginning of the reaction. The ee values were determined by chiral GC or HPLC analysis. The absolute configuration was determined by comparing the sign of optical rotation or the elution order on HPLC with literature data.

Preparative transformation for the sequential cascade reaction for compound 1g

Reactions were performed in 100 mL round bottom flask containing 1 g (50.0 mg), IPA (2.5 mL), and 25.0 mL of phosphate buffer solution (pH 7.0; 100 mM, enriched with MgCl₂ (1.0 mM), NAD(P)H 0.5 mg/mL and CoCl₂ˑ6H₂O (1.0 μL, stock solution 2.0 mg/500.0 μL)). This solution was magnetically stirred at 25^°C; for 2 min. In an Eppendorf tube, ADH (50.0 μL) was solubilized in phosphate buffer solution (400.0 μL) and stirred in a Thermomixer at 25^°C; 1000 rpm for 2 min. After that, the ADH enzymatic solution was added to substrate solution and the reaction was allowed to proceed for 8 h and monitored by GC-FID. After complete substrate conversion into 3 g, the NHase E0257 (25.0 μL) solubilized in phosphate buffer (100.0 μL, pH 7.0; 100.0 mM) was added to the reaction mixture. The reaction was monitored by GC-FID and after 2 h, the reaction mixture was centrifuged (12.000 rpm/4 min) and the supernatant was saturated with NaCl and extracted with ethyl acetate (3 × 15 mL). The combined organic extracts were dried under anhydrous MgSO₄ and solvent was evaporated. The crude product was resuspended in ethyl acetate and analyzed by GC-FID. The ee values were determined by chiral HPLC analysis.

Acknowledgements

Fellowships from the Coordination for the Improvement of Higher Education Personnel (CAPES) (Finance Code 001) and the National Council for Scientific and Technological Development (Proc. Number 134686/2016-0) are gratefully acknowledged. The authors also thank CAPES for maintaining the CAPES Periodical Portal.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Chiral HPLC-DAD and GC-FID chromatograms, NMR, IR, and MS spectra are available in the supplementary information.

Additional information

Funding

This work was supported by the Sao Paulo Research Foundation – FAPESP under grants [#2014/50249-8 and #2019/15230-8]. Conselho Nacional de Desenvolvimento Científico e Tecnológico [134686/2016-0]; Coordenação de Aperfeiçoamento de Pessoal de Nível Superior [88887.341669/2019-00]; Fundação de Amparo à Pesquisa do Estado de São Paulo [2014/50249-8].