Introduction

In recent years, the combination of transition-metal catalysis and radical chemistry has emerged as a powerful synthetic method for the construction of a variety of C-C and C-heteroatom bonds under mild conditions1. In particular, the transition metal-catalyzed radical cross-coupling allows the stereoselective construction of C(sp3)-C(sp3) bonds, which is often difficult to realize through the traditional ionic-based cross-coupling reactions2. Although significant progress has been achieved, the stereoselective alkylation via asymmetric radical cross-coupling of non-acidic C(sp3)-H bonds is still challenging and highly desired. This could attribute to the unavailability of p-orbitals of C(sp3)-H bonds to interact with the transition metal catalyst3.

Alkylnitriles play important roles in many application fields due to their unique electronic and structural properties. A large number of pharmaceuticals, bioactive natural products and functional materials contain such structural motifs4,5,6,7,8,9. Meanwhile, the nitrile group is a versatile synthon as it can be readily transformed into a range of useful functional groups, such as carboxylic acids, amides, amines, aldehydes, ketones, etc.10,11,12,13,14. Ever since the pioneering study of Boivin et al. in the 1990s, the iminyl radical-triggered C-C bond cleavage of cycloalkanone oximes has been successfully applied to synthesize a series of functionalized alkylnitriles under cyanide-free conditions (Fig. 1, eq. 1)15,16,17,18,19,20,21,22,23,24. Despite significant progress, the catalytic enantioselective variants of these reactions remain largely unexplored, and the reported reactions were restricted to the Giese additions of alkenes or the couplings with alkynes25,26,27,28. In contrast, the direct stereoselective cross-couplings of the key alkylnitrile radicals with other alkyl radicals to form chiral C(sp3)-C(sp3) bonds were still underdeveloped, especially with the C(sp3)-H as alkyl radical precursors.

Fig. 1: Radical functionalization of glycine derivatives and peptides.
figure 1

(1) The generation of alkylnitrile radicals. (2) Photo-induced Cu-catalyzed asymmetric C(sp3)-H alkylation. (3) Asymmetric C(sp3)-H cyanoalkylation of glycine derivatives and peptides. PC photocatalyst, TM transition metals.

Unnatural amino acids have encountered widespread applications for the preparation of biologically active molecules and peptidomimetic drugs29,30. Introducing functional groups with characteristic spectroscopy and reactivity properties (e.g., -N3, -CN, and etc.) into amino acids and peptides is of special interest for potential imaging and therapeutic purposes31,32,33,34. Given the importance of cyano-containing unnatural amino acids in drug discovery and functional material development, the practical and stereoselective synthetic method for constructing this type of compounds is highly desired.

In recent years, copper-catalyzed asymmetric C(sp3)-H functionalization through radical processes has attracted extensive attention35,36,37,38,39. We recently developed a photo-induced and Cu-catalyzed asymmetric C(sp3)-H alkylation of glycine derivatives for the synthesis of unnatural amino acids40. The mechanistic studies revealed that the photo irradiation was essentially required to generate a high-valent organometallic copper (CuIII) intermediate, which was crucial to converting the glycine C(sp3)-H to the corresponding alkyl radical via single electron transfer (SET) reduction and deprotonation sequence (Fig. 1, eq. 2). Interestingly, Gong recently developed a Fe promoted glycine C(sp3)-H cyanoalkylation, in which FeIII served as the SET oxidant to generate the glycine N-radical41. Meanwhile, Zhou et al.42, Zuo et al.28 and Zhang et al.43 independently disclosed that the redox-active cyclobutanone oxime could forge the formation of CuIII species. Inspired by these reports and on the basis of our long-standing interest in Cu-catalyzed radical couplings40,44,45,46,47,48,49,50, we hypothesized that trapping of the alkylnitrile radical following ring-opening of cyclobutanone oxime by a Cu complex would in situ form a CuIII complex, which would possibly enable glycine C(sp3)-H to form an alkyl radical under basic conditions. Moreover, with the assistance of a chiral ligand, the asymmetric C(sp3)-H cyanoalkylation might be achieved as a practical complementary approach to the current photo-irradiation strategy40,50. Importantly, the achievement of this reaction would provide a useful tool for preparing cyanoalkylated unnatural α-amino acids and late-stage modification of peptides (Fig. 1, eq. 3).

Results

Investigation of reaction conditions

In this program, we first examined the capability of the chiral Cu complex for catalyzing the redox process and its ability to control the stereoselectivity of radical cyanoalkylation using ethyl (4-methoxyphenyl)glycinate 1a and cyclobutanone oxime ester 2a as model substrates. Asymmetric cyanoalkylation was tested employing different chiral ligands (L1L7 in Table 1) using DABCO as the base and acetone as the solvent, whereas no stereoselectivities were observed in all cases (Fig. 2, eq. 1). We assume that the intermolecular coordination between the chiral catalyst and alkyl radical is quite weak and random, and it is difficult to control the stereochemistry of competitive racemic background reaction. Thus, enhancing the interaction between the chiral catalyst and substrate is an effective strategy to realize the stereoselective cyanoalkylation of C(sp3)-H40.

Table 1 Optimization of the reaction conditions
Fig. 2: Copper-catalyzed asymmetric C(sp3)-H cyanoalkylation of glycine derivatives and peptides.
figure 2

(1) Asymmetric C(sp3)-H cyanoalkylation of ethyl (4-methoxyphenyl)glycinate. (2) Asymmetric C(sp3)-H cyanoalkylation of tert-butyl (5-methoxyquinolin-8-yl)glycinate. PMP 4 -methoxyphenyl.

Therefore, we embed the easily removed N donor unit into the substrate, by which the chiral catalyst and substrate would jointly constitute a new chiral catalytic system. The results show that electron transfer (or charge transfer) and stereocontrol can be defined within one molecule, and the asymmetric cyanoalkylation of glycine derivatives was achieved through the stereoselective reductive elimination of CuIII species (Fig. 2, eq. 2). Using tert-butyl (5-methoxyquinolin-8-yl)glycinate 1b and cyclobutanone oxime ester 2a as model substrates, to our delight when Cu(MeCN)4PF6 (10 mol%) and (S)-Phanephos (L1, 11 mol%) were employed, the ee of C(sp3)-H cyanoalkylation product was dramatically increased to 80%.

After exploring various chiral ligands, basic additives, Cu catalysts, solvents, and other parameters (see Supplementary Information for details), the optimized reaction conditions were determined (Table 1). The desired product 3ba was obtained in 80% yield and 93% ee at 0 °C for 18 h in the presence of Cu(MeCN)4PF6 (10 mol%), (S)-An-Phanephos (15 mol%), DABCO (2.0 equiv), and acetone (0.2 M) (Table 1, entry 1). Control experiments indicated that copper salt and basic additive are indispensable for the reaction (entries 2–3). Furthermore, Cu(MeCN)4PF6 also can independently catalyze the C(sp3)-H cyanoalkylation of glycine derivative in the absence of (S)-An-Phanephos, leading to the formation of the racemic product in 28% yield (entry 4). Notably, the efficiency of cyanoalkylation was not improved under LED irradiation (entry 5). The coupling reaction was sensitive to oxygen, and the yield dropped sharply to 36% under air conditions (entry 6).

Substrate scopes

With the optimized reaction conditions in hand, we explored the substrate scope of oxime esters. As summarized in Fig. 3, the catalytic system exhibited a broad substrate scope and good functional group tolerance. It could accommodate cyclobutanone, 3-oxetanone, Cbz and Boc protected azetidinone derived oxime esters (2a-d), delivering products 3ba-bd in satisfactory yields and enantiomeric excesses. X-ray crystallographic analysis of product 3ba confirmed the (S)-configuration of the newly formed stereochemistry. Meanwhile, (R)-3ba was obtained in 70% yield and −90% ee when (R)-An-Phanephos was used instead of (S)-An-Phanephos. The reaction of disubstituted oxime esters 2e and 2f also proceeded smoothly to afford the desired products 3be and 3bf in good yields and ees, respectively. In addition, the monosubstituted oxime esters 2g-i with -NHBoc, ester and ether groups at the 3-position were also well tolerated, with the corresponding products 3bg-bi isolated in 70–74% yields and 83–94% ees. As a limitation of this protocol, the secondary and tertiary cyanoalkyl radicals derived from 2l-2o were less efficiently captured by the copper catalysts compared to primary cyanoalkyl radicals and gave negative results.

Fig. 3: Substrate scope with respect to oxime esters.
figure 3

Standard conditions: 1b (0.1 mmol), 2 (2.0 equiv), Cu(MeCN)4PF6 (10 mol%), L2 (15 mol%), DABCO (2.0 equiv), acetone (0.5 mL), 0 °C, 18 h, argon atmosphere. Isolated yields based on 1b after chromatographic purification. Ee was determined by chiral HPLC. D.r. (diastereo ratio) was determined by 1H NMR analysis.

Further expansion of the substrate scope was focused on glycine derivatives (Fig. 4). Glycine amide (1c), glycine ester (1d) and glycine derivative bearing different N-aryl groups (1e) were well tolerated (3ca-ea, 65–77% yields). Furthermore, a variety of dipeptides (Gly-Val, Gly-Pro, Gly-Glu, Gly-Phe, Gly-Met, Gly-Thr and Gly-Lys) substrates were prepared to test the functional group tolerance of this reaction with various amino acid residues incorporated. Gratifyingly, the corresponding cyanoalkylated peptides 3fa-la afforded good yields and excellent stereoselectivities with other amino acid residues untouched (Fig. 4A).

Fig. 4: Substrate scope with respect to glycine derivatives and peptides.
figure 4

Standard conditions: 1 (0.1 mmol), 2 (2.0 equiv), Cu(MeCN)4PF6 (10 mol%), L2 (15 mol%), DABCO (2.0 equiv), acetone (0.5 mL), 0 °C, 18 h, under Ar. Isolated yields based on 1 after chromatographic purification. Ee was determined by chiral HPLC. D.r. was determined by 1H NMR analysis. aDMF (0.5 mL) as solvent, 36 h. A Substrate scope with respect to glycine derivatives and dipeptides. B Substrate scope with respect to peptides.

Having established proof-of-concept with the above results, we became interested if our reactions could be applied in C(sp3)-H cyanoalkylation of polypeptides (Fig. 4B). Gratifyingly, the late-stage C(sp3)-H cyanoalkylation of pentapeptide (Gly-Leu-Phe-Ser-Lys) and hexapeptide (Gly-Leu-Tyr-Ser-Phe-Ala) derived substrates reacted smoothly and gave the corresponding cyanoalkylation products in good yields (3ma, 63%; 3na, 60%) and high diastereoselectivities (>20:1). It was worth noting that the reactions between hexapeptide (Gly-Leu-Tyr-Ser-Phe-Ala and Gly-Leu-Phe-Gly-D-Thr-Tyr) substrate and different oxime esters (2a, 2b, 2e, and 2f) were achieved in uniformly good yields and high diastereoselectivities, which further highlighted the generality of this method in modification of complex molecules.

Synthetic applications

To further illustrate the application potential of the product, several transformations were conducted (Fig. 5). Firstly, the deprotection of the cyanoalkylation product 3ba proceeded smoothly under simple procedures in high yield and did not erode the ee (Fig. 5, eq. 1). Secondly, treatment of the 3ba with Raney-Ni and H2 in pyridine/EtOH/H2O afforded the corresponding amine 5 in 70% yield and 90% ee (Fig. 5, eq. 2). Furthermore, the cyano group of 3ba was easily converted to amide 6 in the presence of Pd catalyst (Fig. 5, eq. 3).

Fig. 5: Synthetic applications.
figure 5

(1) Deprotection of 3ba. (2) Conversion of cyanide to amine. (3) Conversion of cyanide to amide. CAN Ce(NH4)2(NO3)6, Py pyridine.

Mechanistic studies

In order to gain some insight into the mechanism, several radical trapping experiments were carried out (Fig. 6). The radical trapping experiments with TEMPO (2,2,6,6-tetramethylpiperidin-1-oxyl, completely suppressed) or BHT (2,6-ditert-butyl-4-methylphenol, partially suppressed) indicated that a radical pathway was involved in this transformation. Meanwhile, radical trapping product 7 was isolated in 30% yield in the presence of TEMPO, which suggested the formation of cyanoalkyl radical in the reaction system (see Supplementary Information for details). Furthermore, the HRMS analysis of the original reaction mixture detected the formation of glycinate homo-coupling product 8 when the dosage of 2a was 1.0 equiv (Fig. 6, eq. 1). Moreover, when the reactions of 1b and 2a were carried out under air, oxidized glycine derivative 9 was obtained in 40% yield (Fig. 6, eq. 2). These results indicated that glycinate radical was generated under standard reaction conditions. N-CH3-substituted glycinate 1p did not give any cyanoalkylation product with quantitive recycling of the starting material, indicating that a single free hydrogen atom on N atom is crucial for this reaction (Fig. 6, eq. 3). Notably, imine 1q failed to give the cyanoalkylation product 3da, revealing the Cu-mediated SET pathway likely involved rather than the two-electron oxidation process (Fig. 6, eq. 4).

Fig. 6: Mechanistic studies.
figure 6

(1) The formation of glycinate homo-coupling product. (2) The α-oxidation of glycinate. (3) The reaction with N-CH3-substituted glycinate. (4) The reaction with imine. (5) The reaction with ethyl-(4-methoxyphenyl) glycinate. (6) The reaction with copper complex A. (7) The reaction with copper complex B.

We prepared ((S)-An-Phanephos)Cu(CH3CN)4PF6 complex A and copper complex B (combined complex A and 1b) according to the literature procedure51. The ground state redox potentials of 2a (Eredox = −2.11 V vs. SCE in CH3CN) and copper complex B (Eredox = −2.38 V vs. SCE in CH3CN) demonstrated that the SET oxidation of 2a by copper complex B was feasible (see Supplementary Information for details). Some control experiments were also carried out, and the results are shown in Fig. 6 (eqs. 5–7). Ethyl (4-methoxyphenyl) glycinate 1a could provide a racemic product under standard conditions (3aa, 30% yield, 0% ee, Fig. 6, eq. 5). The result revealed the importance of the coordination with 5-methoxyquinoline group for the chirality induction. When the isolated Cu complex A was used as the catalyst, the corresponding cyanoalkylation product was obtained in 78% yield and 93% ee (Fig. 6, eq. 6). Furthermore, the isolated Cu complex B (1.0 equiv) was directly used as the substrate without any other catalyst in the system, the coupling product was obtained in 68% yield and 91% ee (Fig. 6, eq. 7). These results indicated that complex B likely formed in the reaction and act as the active catalytic species.

Proposed mechanism

Based on the above mechanistic studies and previous reports52,53,54,55, a plausible mechanism was proposed in Fig. 7. The 5-methoxyquinolinyl-8-glycinate ester 1b coordinates to the CuIL* and in situ forms a chiral Cu(I) intermediate A. Then, a SET reduction of oxime ester 2a by intermediate A occurs, followed by fragmentation to afford cyclic iminyl radical 2a-A and oxidized Cu(II) intermediate B. Next, cyclic iminyl radical 2a-A undergoes regioselective ring-opening C-C bond cleavage to form cyanoalkyl radical 2a-B. At this point, cyanoalkyl radical 2a-B was captured by Cu(II) intermediate B to produce the high-valent Cu(III) intermediate C. Then, the copper-mediated intramolecular oxidation of the N atom produces Cu(II) intermediate D, and DABCO-promoted deprotonation gives radical intermediate E, which subsequently attacks the copper center to form a chiral Cu(III) intermediate F. Finally, the stereoselectively reductive elimination delivered the desired product 3ba.

Fig. 7: Proposed mechanism.
figure 7

Copper-catalyzed asymmetric C(sp3)-H cyanoalkylation of glycine derivatives.

Discussion

In conclusion, we report a copper-catalyzed asymmetric C(sp3)-H cyanoalkylation of glycine derivatives and peptides. The reactions feature mild conditions, excellent enantioselectivity and broad substrate scope. Given the significance of introducing cyanoalkylation into amino acids and peptides for potential imaging and therapeutic purposes, we predict that our asymmetric C(sp3)-H cyanoalkylation would provide new approaches to the synthesis of unnatural α-amino acids and late-stage functionalization of bioactive compounds, and would be useful for modern peptide synthesis and drug discovery.

Methods

General procedure A (3ba-bi, 3ca-la)

To an oven-dried 10-mL quartz test tube with a stirring bar, derivatives of glycine (0.1 mmol) were added, followed by the addition of Cu(MeCN)4PF6 (0.01 mmol, 3.7 mg) and (S)-An-Phanephos or (R)-An-Phanephos (0.015 mmol, 10.5 mg). Then, the air was withdrawn and backfilled with Ar (three times). Acetone (0.25 mL) was added, and the mixture was stirred at room temperature for 40 min. Subsequently, oxime esters (0.2 mmol) and DABCO (0.2 mmol, 22.4 mg) dissolved in acetone (0.25 mL) were added to the abovementioned mixed solution by syringe. Thereafter, the test tube was transferred to a low-temperature device, where it was reacted for 18 h at 0 °C. Then, the reaction was quenched with water (1 mL), extracted with ethyl acetate (3 × 1.5 ml), dried over anhydrous sodium sulfate, concentrated in vacuo, and purified by column chromatography (hexane/ethyl acetate) to give the product.

General procedure B (3ma-oa, 3ob, 3oe, and 3of)

To an oven-dried 10-mL quartz test tube with a stirring bar, peptide substrates (0.1 mmol) were added, followed by the addition of Cu(MeCN)4PF6 (0.01 mmol, 3.7 mg) and (S)-An-Phanephos (0.015 mmol, 10.5 mg). Then, the air was withdrawn and backfilled with Ar (three times). DMF (0.25 mL) was added, and the mixture was stirred at room temperature for 40 min. Subsequently, oxime esters (0.2 mmol) and DABCO (0.2 mmol, 22.4 mg) dissolved in DMF (0.25 mL) were added to the abovementioned mixed solution by syringe. Thereafter, the test tube was transferred to a low-temperature device, where it was reacted for 36 h at 0 °C. Then, the reaction was quenched with water (1 mL), extracted with ethyl acetate (3 × 1.5 ml), dried over anhydrous sodium sulfate, concentrated in vacuo, and purified by column chromatography (hexane/ethyl acetate or dichloromethane/methanol) to give the product.