Enantioselective Radical Hydrocyanoalkylation of Alkenes via Photoenzymatic Catalysis

Abstract

Organic nitriles are significant in pharmaceuticals, agrochemicals, cosmetics, and materials. Although numerous cyanidation methods have been developed, more eco-friendly and green protocols for manufacturing alkyl nitriles are in high demand. Here, we report a photoenzymatic enantioselective intermolecular hydrocyanoalkylation of alkenes catalyzed by flavin-dependent “ene”-reductases. The discovery of stereocomplementary enzymes that provide access to both enantiomers of the high-value nitriles further showcases the synthetic applications of this method. Radical trapping, isotopic labeling, and spectroscopic experiments have elucidated the formation of a charge transfer complex at the protein active site. The single-electron reduction of the cyanoalkyl radical precursor by flavin hydroquinone yields a cyanoalkyl radical, which then undergoes intermolecular radical addition. This active site can stereoselectively control the radical-terminating hydrogen atom transfer, enabling the synthesis of enantioenriched γ-stereogenic nitriles. This work further expands the reactivity repertoire of biocatalytic transformations via non-natural radical mechanisms.

Introduction

Organic nitriles are a major class of compounds having wide significance in the pharmaceutical, materials science, agrochemical, and cosmetic industries. In particular, C­(sp²)- and C­(sp³)-CN-containing compounds are present in countless pharmaceutical compounds such as anastrozole, bicalutamide, vildagliptin, tofacitinib, etc. (Scheme A). In addition, nitriles are important precursors for constructing different functional groups, such as aldehydes, amines, amides, azoles, carboxylic acids, and their corresponding carboxyl derivatives, in synthetic chemistry, both on an industrial and laboratory scale. Therefore, numerous cyanidation methods have been developed, including the Sandmeyer reaction, the Rosenmund-von Braun reaction, and other transition-metal-catalyzed cyanation reactions of aryl, vinyl halides, and pseudohalides, which utilize metal cyanides as cyanide sources. Furthermore, the asymmetric cyanation reaction of aldehydes and ketones has undergone significant development over the past two decades. The other type of reaction is transition-metal-catalyzed hydrocyanation of unactivated alkenes; studies on these reactions are relatively rare yet (Scheme B). Therefore, although there are various methods to synthesize this class of compounds, considering the toxicity of the cyanide sources such as NaCN, KCN, and trimethylsilyl cyanide (TMSCN), or the requirement of harsh reaction conditions like high temperatures, more eco-friendly and green protocols for manufacturing alkyl nitriles are in high demand.

1. Current Research on Cyanation and Our Reaction Design .

^a (A) CN-containing pharmaceutical compounds. (B) Previous cyanidation methods. (C) Our designed photoenzymatic enantioselective hydrocyanoalkylation strategy.

Due to the increasing demand for nitrile-containing molecules in the fields of pharmaceutical synthesis and organic chemistry, the catalytic enantioselective hydrocyanoalkylation methods remain an active research area, offering excitement and opportunities for expanding their scope and applicability. Biocatalysis has emerged as a powerful approach and is widely utilized for the green and efficient synthesis of fine chemicals because of the desirable enantio-, regio-, and chemoselectivity. Recently, the integration of photocatalysis with biocatalysis, especially visible light-induced direct photoenzymatic catalysis via flavin-dependent “ene”-reductases (EREDs), has been considered an emerging synthetic strategy for radical reactions. Significant examples include the enantioselective radical hydroalkylation and other C–C bond-forming reactions. Although considerable progress has been made in this area, the reactivity repertoire of these transformations is still limited. Inspired by previous advancements, we hypothesized here that photoexcited flavin-dependent enzymes could be harnessed to generate alkyl nitrile radicals via inexpensive, nontoxic, and readily available iodinated alkyl nitriles (Scheme C). Then, an intermolecular radical addition occurs to form C–C bonds. Ultimately, a series of enantioenriched γ-stereogenic nitriles were synthesized through enzymatic stereoselective hydrogen atom transfer (HAT).

Results and Discussion

We initiated our investigation by exploring the radical hydrocyanoalkylation of 2-isopropenylpyridine 1a using iodoacetonitrile as a radical precursor with a series of EREDs previously identified as catalysts for photoenzymatic reactions. While many EREDs were able to facilitate the desired reaction, the stereoselectivity is unsatisfactory (Table S1). Among them, OYE2 from Saccharomyces cerevisiae formed product 2a in 87% yield with 14:86 e.r., favoring the S-enantiomer. Alternatively, we found that NCR from Zymomonas mobilis provided product 2a in 91% yield with 78:22 e.r., favoring the R-enantiomer (Table , entries 1–2). Control experiments confirmed the essential features of this reaction. The light, enzyme, and NADPH regeneration system were required for the hydrocyanoalkylation reaction (Table , entries 3–5). It is worth noting that when free flavin mononucleotide (FMN) was used instead of OYE2, only trace amounts of product were detected, implying the enzyme’s critical role in the initiation step of the reaction. As expected, owing to the radical-trapping ability of O₂, this reaction was significantly suppressed under an ambient atmosphere (<5% yield). A test of different reaction factors finally obtained the optimal reaction conditions (Tables S2–S11).

1. Effect of Reaction Parameters .

entry	conditions	P/yield	e.r.
1	as shown	87%	14:86
2	NCR-WT instead of OYE2-WT	91%	78:22
3	w/o light	N.D.
4	w/o enzyme	trace
5	w/o NADPH regeneration system	N.D.
6	FMN instead of OYE2	trace
7	air instead of Ar	<5%

^aReaction conditions: 2-isopropenylpyridine 1a (3 μmol, 1.0 equiv), ICH₂CN (9 μmol, 3.0 equiv), NADPH regeneration system (GDH (0.3 mg) + NADP⁺ (15 mol %) + glucose (6.0 equiv)), “ene”-reductase (0.2 mol %), 1 mL of buffer (50 mM PBS buffer, pH 8.0), hν (40 W, 495 nm cyan LEDs), 15 °C, 16 h. The yield was determined via LC–MS using a standard curve, and the enantioselectivity was assessed by chiral HPLC.

Next, we sought to improve the efficiency and selectivity of this transformation using iterative saturation mutagenesis (ISM) and two promising enzymes, OYE2 and NCR. We began by solving the cocrystal structure of wild-type OYE2 with 2-isopropenylpyridine 1a at 3.4 Å resolution (PDB ID: 9KUH, Figure S4 and Table S12). As shown in Figure (left), the residues that line the active site are considered the most significant potential contributors to influencing reactivity and selectivity, based on crystallographic information. After saturation mutagenesis applied to 10 residues, the best mutant in this study proved to be OYE2-F297N (OYE2-S), which facilitates the formation of (S)-2a in 91% yield with 90% ee (Tables 13 and 14). To improve the enantioselectivity of the R-enantiomer, we first conducted a structure alignment of NCR-WT based on the obtained OYE2 cocrystal structure (Figure S5). Furthermore, we selected 7 residues around the substrate pocket for saturation mutagenesis and finally found that NCR-Y343T (NCR-R) can obtain (R)-2a in improved yield (93%) with excellent stereoselectivity (92% ee) (Tables S15 and S16).

1.

Stereoselective hydrocyanoalkylation catalyzed by engineered “ene”-reductases.

Having demonstrated that we can optimize the product yield and enantiopurity using the OYE2-S or NCR-R, the substrate scope of the photoenzymatic hydrocyanoalkylation reaction was investigated. As shown in Scheme , both enantiomers of a variety of 2-isopropenylpyridine derivatives with ortho, meta, and para substituents on the aromatic ring were efficiently prepared in high yields and with good levels of enantioselectivity (2b–2i). Analogues bearing an electron-withdrawing group (6-Br, 6-F, 5-CF₃ or 6-CF₃) and electron-donating group (4-OMe, 6-OMe) of the pyridyl moiety were well-tolerated, affording the desired products in 78–96% yields with up to 1:99 e.r. for (S)-enantiomer (2n) and with up to 98:2 e.r. for (R)-enantiomer (2m). Substituting the pyridyl ring with cyano groups was viable in the transformation, and dicyano-substituted products (2p, 2q) could be obtained with satisfactory yields and stereoselectivity. Additionally, bulky methyl formate-substituted substrate (2r) was accepted by the enzyme’s binding pocket, leading to a moderate yield and stereoselectivity.

2. Substrate Scope of Photoenzymatic Asymmetric Hydrocyanoalkylation .

^a Reaction conditions: olefin 1 (30 μmol, 1.0 equiv), ICH₂CN (90 μmol, 3.0 equiv), EREDs (0.2 mol %), NADP⁺ (15 mol %), glucose (6.0 equiv), GDH (1.0 mg, about 9 U), 3 mL of buffer (50 mM PBS buffer, pH 8.0), hν (40 W, 500 nm cyan LEDs), 15 °C, 16 h under an Ar atmosphere. See the Supporting Information for the detailed conditions in each reaction. The yields were determined by LC and reported as the average of duplicates based on a 30 μmol reaction scale. The e.r. was determined by HPLC analysis on a chiral stationary phase. Absolute configurations were assigned based on the known reported synthetic procedure of (R)-2a.

Next, the scope of the heteroaromatic group was examined. In addition to 2-isopropenylpyridine, 3-isopropenyl and 4-isopropenyl were effective in this reaction, providing products 3a and 3b, respectively, in good yields with promising levels of enantioselectivity. The privileged heterocyclic structures in drug discovery, such as pyrimidine, pyrazine, and quinoline, were also accommodated under the optimized conditions. Replacing the six-membered aromatic group with a five-membered thiazole (3f) and benzofuran (3g) also proved viable, generating the enantioenriched products in satisfactory yields (77%–83%) with an 8:92 e.r. for (S)-3f and a 91:9 e.r. for (R)-3g. Surprisingly, this reaction could also accommodate alkyl substituents at the α-position. The ethyl group and larger groups, such as i-ropyl, were well accepted and provided the respective products (3h-3i) in 85–91% yields and high enantioselectivity (up to 3:97 e.r. for (S)-enantiomer). As for the bulky phenyl substituent at the α-position, this reaction could generate both enantiomers (3j) in good yields with satisfactory enantioselectivity. Notably, α-methylstyrene derivatives were also accessible, affording 4a and 4b in 89–94% yield with high enantioselectivities for both enantiomers. Beyond iodoacetonitrile as a radical precursor, we found that this transformation also accommodates 3-iodopropanenitrile (I–CH₂CH₂CN) and 2-iodo-2-methylpropionitrile (I–C­(CH₃)₂CN) with similar redox potentials, affording δ-stereogenic nitrile 5a and bulky 5b in yields of 48–67% with satisfactory enantioselectivity. To evaluate the synthetic potential of this transformation, a scaled-up experiment was performed using the model reaction (1a, 0.40 mmol, 0.1 mol % NCR-Y343T), affording the desired product (R)-2a in 79% yield (50.6 mg) with 90% ee (Figure S3). In addition, other substrates are also accommodated with this reaction (S1–S10, Figure S17). Unfortunately, this catalytic system is not compatible with nonaromatic substrates and cannot control the stereoselectivity through C–C bond formation (S11–S13, Figure S17). This reaction showcases the promiscuity and potential of enzymes for the asymmetric synthesis of nitriles, which is challenging for chemocatalysis.

To further demonstrate the synthetic utility of this biocatalytic reaction, we explored (R)-2a, which could be diverted to afford many other useful compounds (Scheme ). Through direct diversification of CN itself, the desired amines (6), ketones (7), esters (8), and diaminotriazines (9) were smoothly accessed in moderate to good yields with excellent stereoretention.

3. Representative Transformations of Product (R)-2a .

^a See the Supporting Information for the detailed experimental procedure.

Having established the unique photoenzymatic synthesis of enantioenriched nitriles by engineered EREDs, we focused our investigation on elucidating the mechanism of this transformation. Initially, we hypothesized that electron transfer occurred via photoexcitation of a charge transfer (CT) complex. To investigate this possibility, we analyzed our reaction mixture via UV–vis. The cofactor flavin mononucleotide (FMN) in OYE2 was completely reduced to flavin hydroquinone (FMN_hq) with sodium dithionite, which showed negligible absorption around 500 nm (Figure A, purple). In addition to either 1a or iodoacetonitrile alone, no significant changes were observed (Figures A and S13). The addition of 1a to the complex of OYE2-FMN_hq, and iodoacetonitrile produced a new broad absorption band from 440 to 530 nm, providing strong evidence for the formation of a quaternary CT complex between OYE2-FMN_hq, iodoacetonitrile, and 1a (Figure A, green). Subsequent absorption studies of the above CT complex of the same type and concentration of OYE2-FMN_ox and a series of fluorescence quenching experiments further supported the conclusion of the above CT complex (Figures S12, S14, and S15).

2.

Mechanistic investigations. (A) Ultraviolet–visible absorption experiments. (B) Radical trapping experiments. (C) Deuterium incorporation experiments.

To validate the proposed intermediates in this transformation, we carried out a series of radical-trapping experiments. When this model reaction was performed in the presence of radical scavengers such as 5,5-dimethyl-1-pyrroline N-oxide (DMPO), 2,2,6,6-tetramethylpiperidinyloxy (TEMPO), and 2,6-di-tert-butyl-4-methylphenol (BHT), respectively (Table S17), the reactivity was inhibited significantly. In addition, the TEMPO adducts Int.1 and Int.2 were obtained in 55% and 38% isolated yields, respectively, revealing the involvement of cyanomethyl radical and benzylic radical (Figure B). Collectively, these results supported the radical-mediated process (Figure S7).

Deuterium incorporation experiments were then implemented to determine the source of the hydrogen atom. As shown in Figure C, when the standard reaction is carried out in a deuterated buffer, only 4% deuterium incorporation is observed (Figure S10). Alternatively, when flavin was labeled in situ using d ₁-glucose and glucose dehydrogenase (GDH), a 93% D-incorporation product was observed (Figure S11), supporting that the γ-stereocenter was set via HAT from flavin as the primary mechanism of radical termination.

Based on these studies, we proposed a mechanism in which a cyanoalkyl radical was generated via photoinduced electron transfer between FMN_hq and iodoacetonitrile (Figure ). Subsequently, the cyanomethyl radical undergoes an intermolecular radical addition to the vinylpyridine to deliver the prochiral benzylic radical Int. A. Stereoselective HAT is readily carried out by FMN_sq to form the desired products and FMN_ox. Finally, product–substrate exchange and the regeneration of FMN_hq with the NADPH regeneration system complete the catalytic cycle. Furthermore, we also investigated the stereoselectivity by computational simulation (Figure S16). Due to the different orientations of Int. A in the catalytic pockets of the two enzymes, the stereoselectivity of the HAT process was differentiated, achieving stereoselective complementation.

3.

Proposed reaction mechanism.

Conclusion

In conclusion, we have developed a biocatalytic radical hydrocyanoalkylation of alkenes to afford a variety of γ/δ-stereogenic nitriles with high yields and excellent enantioselectivity. The discovery of stereocomplementary enzymes that provide access to both enantiomers of the product further enhances the synthetic applications of this method. Mechanistic studies elucidated the formation of a CT complex in the protein active site and clarified the transformation of radical intermediates in this reaction. Moreover, our work explores the potential of enzymes for the asymmetric synthesis of high-value nitriles, utilizing readily available iodoalkylnitriles as a source of cyano groups. This study further expands the biocatalyst toolbox through photobiocatalysis, offering a new approach to addressing the selectivity challenges of radical-mediated functionalization reactions.

The data and methods supporting the findings of this work are presented in the paper and its Supporting Information section.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacsau.5c00633.

• Source data (PDF)

†.

These authors contributed equally: D.W. and Z.S.

The authors declare no competing financial interest.