Introduction

α-Amino acids (AAs), encompassing both natural and unnatural variants (UAAs), serve as fundamental building blocks for protein structure and function, underpinning biological systems and therapeutic design1,2,3,4,5,6,7,8,9. Recent advances in pharmaceutical chemistry and biotechnology have amplified the demand for enantiopure α-AAs, particularly in drug development, spurring innovations in asymmetric synthesis (Fig. 1A)10,11,12,13,14. Current approaches to access chiral α-AAs include isolation from protein hydrolysates, biocatalysis, and chemical synthesis15,16,17,18. While enzymatic methods, such as amino acid racemase-mediated interconversion of L- and D-AAs, offer step efficiency, their utility for α-UAAs is constrained by enzyme stability, catalytic activity, and narrow substrate scope (Fig. 1B)19,20,21,22. In contrast, chemical synthesis has evolved as a versatile and scalable platform for constructing chiral α-(U)AAs. State-of-the-art methodologies, including dynamic kinetic resolution (DKR) of azlactones or carboxylic acids23,24,25,26,27, N–H insertion reactions28,29,30,31,32, and asymmetric hydrogenation33,34, have significantly expanded the synthetic repertoire. These strategies further enable the preparation of peptides incorporating chiral α-UAAs, which show significant potential in the design of peptide-mimetic therapeutics35,36. However, most existing catalytic systems necessitate substrates with specific structural features or strategic pre-modification to align with their activation modes. Consequently, developing innovative methodologies to access structurally diverse α-(U)AAs remains a critical challenge. Herein, we present a catalytic strategy that employs readily available racemic (U)AAs or enantiopure α-(U)AAs (L/D-configuration) as starting materials, enabling stereoselective conversion to target (U)AAs with precise stereochemical control. This approach leverages N-heterocyclic carbene (NHC) catalysis to bypass traditional substrate limitations, offering a streamlined route to chiral α-(U)AAs with tailored functional and stereochemical diversity (Fig. 1C).

Fig. 1: Strategies for the transformation of α-(U)AAs.
figure 1

A Selected examples of drug molecules containing amino acid modules. B Enzyme-catalyzed chirality interchange of an (unnatural) amino acid stereocenter. C Chiral amino acid construction enabled by the oriented transformation of (racemic) stereocenters.

Full size image

Building on established principles of N-heterocyclic carbene (NHC) catalysis, we hypothesize that NHC-derived acyl azolium intermediates enable both the construction and stereochemical preservation of α-stereocenters in amino acids37,38,39,40,41,42. Upon activating carbonyl substrates—including anhydrides, imides, esters, amides, or aldehydes—the NHC catalyst engages the carbonyl carbon, facilitating chiral induction at the α-position via a proton shuttle mechanism38,39,43,44,45,46,47,48. This DKR process efficiently transfers stereochemical information from the NHC scaffold to the nascent α-stereocenter, aligning with our strategy for directed chiral α-AA synthesis (Fig. 2A). Herein, we introduce a methodology for chiral α-AA synthesis: sterically demanding chiral NHC catalysts drive enantioselective assembly of α-AA derivatives, while minimally hindered NHC variants preserve stereochemical fidelity during subsequent peptide elongation. This dual catalytic system resolves the competing demands of stereoselective synthesis and configurational stability, providing a unified platform for chiral α-AA construction and peptide stereocontrol (Fig. 2B).

Fig. 2: NHC catalyzed the construction of ɑ-(U)AAs.
figure 2

A NHC-mediated H-shuttling for constructing carbonyl α-stereocenters. B NHC-mediated approach enabling the synthesis of chiral (U)AAs and peptides (this study).

Full size image

Results

NHC-catalyzed stereoselective formation of α-amino esters

To test our hypothesis, we systematically explored the stereoselective protonation of α-amino acids (1) using diphenylmethanol (2a) as a sterically demanding nucleophile. Initial screening identified N,N’-diisopropylcarbodiimide (DIC) as the optimal activating agent, demonstrating superior efficiency among various coupling reagents. Further optimization revealed that combining 20 mol% NHC precatalyst 3a with 1 equivalent Cs2CO3 in dichloromethane (0.05 M) efficiently promoted the formation of target α-amino ester 4, achieving excellent yield and enantioselectivity (see Supplementary Tables S1–S3).

With optimized reaction conditions established, we evaluated the substrate generality of this transformation (Fig. 3). The methodology demonstrated broad compatibility with both natural and unnatural amino acids. Racemic N-terminal-protected amino acids underwent convergent stereoselective conversion to (S)-configured derivatives (Phe, Trp, Ala, Met; 4a-4d, >90% ee average). α-Benzyl-substituted substrates, featuring electron-rich and electron-deficient aryl groups, proceeded smoothly with high enantioselectivity (products 4e-4j, 90–97% ee). A thiophene-containing phenylalanine analog was efficiently synthesized under these conditions, requiring only a single additional step to generate the racemic precursor from α-bromo carboxylic acid (4k, 87% yield, 94% ee; see SI). Notably, steric hindrance from α-aryl substituent did not impede the reaction, enabling the synthesis of chiral α-aryl analog 4l (65% yield, 91% ee). α-Alkyl-substituted substrates underwent DKR with high efficiency, yielding target chiral UAAs (4m-4p, 92–97% ee). Crucially, this strategy exhibited superior functional group tolerance compared to conventional transition metal- or enzyme-catalyzed approaches, preserving unsaturated side chains (4q-4r) without double/triple bond degradation—a critical advantage for post-synthetic modifications. The methodology further accommodated secondary α-cyclohexyl substituents (4s, 82% yield, 90% ee), though tertiary α-substituents completely suppressed reactivity (see Supplementary Fig. S5).

Fig. 3: The scope of chiral ɑ-amino acids.
figure 3

Reaction conditions: A mixture of ɑ-AAs 1 (0.05 mmol), diphenylmethanol 2a (0.075 mmol), NHC-3a (20 mol%), Cs2CO3 (0.05 mmol) and DIC (0.06 mmol) dissolved in DCM (0.05 M) was stirred at r.t. for 2–4 h. Isolated yields are presented. Chiral HPLC determined ee values. a NHC-3s was applied instead of NHC-3a.

Full size image

Using phenylalanine (Phe) as a model substrate, we systematically investigated the influence of N-protecting groups (PGs) on the reaction (Fig. 4). Substituted benzoyl derivatives proved highly effective (see Supplementary Fig. S3), while other acyl PGs—including acetyl, heterocyclic, cinnamyl, alkyl, and amide groups—were also well tolerated (4ab-4af). This diversity in compatible N-terminal PGs expands the range of feasible deprotection strategies for subsequent peptide modification, enhancing adaptability across different applications. We next examined nucleophiles for catalyst turnover, a critical component of the acyl-azolium catalytic cycle. Primary and secondary alcohols, including functionalized variants (e.g., hydroxyproline derivatives), performed exceptionally well (4ag-4ak). We extended our investigation to peptide substrates with this demonstrated functional group tolerance. Under standard conditions, racemic N-terminal peptide-linked amino acids underwent asymmetric proton shuttle via DKR, affording chiral peptide products with excellent stereocontrol (6a-6g). The protocol retained efficiency even with UAA moieties embedded in the peptide backbone, yielding chiral peptides with diverse side chains (6h-6j). The successful DKR of peptides into their chiral ester derivatives highlights the potential of this methodology for peptide chemistry and related applications.

Fig. 4: The scope of substrates with different N-PGs and peptides.
figure 4

Reaction conditions: A mixture of ɑ-AAs 1 or peptides 5 (0.05 mmol), alcohol (0.075 mmol), NHC-3a (20 mol%), Cs2CO3 (0.05 mmol) and DIC (0.06 mmol) dissolved in DCM (0.05 M) was stirred at r.t. for 4–10 h. Isolated yields are presented. Chiral HPLC determined ee or de values. a NHC-3c was applied instead of NHC-3a.

Full size image

Control experiments were performed to elucidate the reaction mechanism (see Fig. 5 and Supplementary Fig. S13). The transformation required all key components (NHC catalyst, base, and DIC) for optimal performance. Notably, oxazolone intermediate Int-1a was identified as a crucial species formed without NHC catalyst. Screening of coupling reagents revealed EDCI provided a high yield (97%), albeit with reduced enantioselectivity (86% ee), while HATU and CDI proved ineffective, yielding only trace products. Isolation and subsequent treatment of Int-1a under optimized conditions (without DIC) afforded product 4a in high yield with excellent enantioselectivity, confirming oxazolone’s role as a key intermediate in the enantioselective transformation, as well as activated p-nitrophenyl ester has similar reactivity (see Fig. 5A and Supplementary Fig. S4). Further evaluation of N-PGs demonstrated that while Cbz, Fmoc, and Boc groups provided high yields of amino esters, they exhibited poor stereocontrol. This distinction arises from DIC’s ability to promote oxazolone formation with Bz and Ac groups but not with Cbz, Fmoc, or Boc as PGs (Fig. 5B). The methodology proved effective for α-chirality interconversion between (S)- and (R)-configurations (Fig. 5C). Starting from either (S)-, (R)-, or racemic 1ab, a single stereoisomer (R)-4ab was obtained. Using ent-NHC 3a as a catalyst inverted the selectivity to furnish (S)-4ab. This stereochemical control was further validated using (rac)-1aa as the substrate and N-fmoc-trans-4-hydroxy-L-proline methyl ester as the nucleophile, which enabled the synthesis of a single stereoisomer (S) or (R)-4aj. These results collectively support a mechanism involving initial substrate racemization followed by NHC-mediated α-chiral reconstruction.

Fig. 5: Mechanistic study about NHC-catalyzed esterification of AAs.
figure 5

A Standard conditionsa for control experiment: A mixture of AAs 1a (0.05 mmol), alcohol 2a (0.075 mmol), NHC-3a (20 mol%), Cs2CO3 (0.05 mmol) and DIC (0.06 mmol) dissolved in DCM (0.05 M) was stirred at r.t. for 2–10 h. Isolated yields are presented. Chiral HPLC determined ee and de values, see Supplementary Fig. S2. bOxazolone Int-1a as starting material without DIC in the std. cond. A, see Supplementary Fig. S1. B Na-protecting group impact. cDirect synthesis of oxazolone under DIC. C α-chirality interconversion between (S)- and (R)-configurations.

Full size image

NHC-catalyzed racemization-free peptide synthesis application

Building on these insights, we systematically evaluated the methodology’s capacity to stabilize α-stereocenters and suppress racemization during peptide synthesis (see Fig. 6A and Supplementary Table S4)49,50. Employing amino acid hydrochloride 2b as a nucleophile with minimally hindered NHC catalyst 7a could effectively minimize racemization. The reaction efficiently operated under base-free conditions, outperforming direct coupling approaches that exhibited moderate yield and excellent stereoselectivity (73% yield, 99% es vs. entries 1–3). Introducing exogenous bases significantly exacerbated racemization, underscoring the importance of base-free circumstances (entry 4, 34% yield, 29% es). Heatmap analysis of N-PGs revealed enhanced yields and diastereomeric excess in the presence of 7a, confirming its role in racemization suppression (products 8b–8f). We further designed a tandem sequence integrating chiral (U)AA synthesis and peptide elongation, enabling direct access to enantiopure peptides from racemic N-protected amino acids (product 8a, 62% yield, 94% de). Expanding this concept, we developed a Wang resin-based solid-phase peptide synthesis (SPPS) protocol that simultaneously establishes α-chirality and peptide bond, delivering target peptides with high stereo fidelity (product 8c, 20% yield, 97% de). Further scale-up reaction and representative examples proved highly effective (see Supplementary Figs. S7, S8).

Fig. 6: NHC-catalyzed racemization-free peptide synthesis.
figure 6

A Standard conditionsa for control experiment: a mixture of chiral AAs 1ab (0.05 mmol), amine 2b (0.055 mmol), NHC-7a (20 mol%) and DIC (0.06 mmol) dissolved in DCM (0.05 M) was stirred at r.t. for 3–6 h. Isolated yields are presented. Chiral HPLC determined de and es values, see Supplementary Fig. S6. bCompounds 1ab and 2b dissolved in DCM (0.05 M) were stirred at r.t. c,dFor concrete operations, see Supplementary Figs. S9 and S10. B Application in peptite synthesis. eStd. cond. C: step 1: MYTsA (1.2 eq.) in DCM; step 2: H-L-Phe-OMe.HCl (1.1 eq.) in DMF, see Supplementary Fig. S11. fFor concrete operations, see Supplementary Fig. S12.

Full size image

DFT calculation

Density functional theory (DFT) calculations were performed to elucidate the stereochemical control governing chiral reconstruction (see Fig. 7 and Supplementary Figs. S14–16). The computational studies delineate the stereoselective transformation pathway, beginning with NHC-catalyzed oxazolone ring-opening. Transition state analysis reveals a 1.2 kcal/mol energy preference for (S)-TS1 over (R)-TS1, driven by stabilizing phenyl π-π interactions between the substrate and NHC catalyst. Subsequent CsHCO₃-mediated protonation generates acyl azolium intermediate Int2, where steric constraints render the R-configured species 2.9 kcal/mol higher in energy than its S-counterpart. Two divergent pathways emerge from Int2: (1) direct CsHCO₃-assisted alcohol addition via TS3, or (2) enolate formation (Int-Z/Int-E) followed by alcohol addition through TS5. The former pathway dominates, with (S)-TS3 exhibiting a 2.5 kcal/mol energy advantage over (R)-TS3 due to favorable substrate-alcohol π-π interactions (see Supplementary Fig. S17). This step, characterized by an activation barrier of 17.5 kcal/mol, constitutes the rate-determining step. In contrast, the enolate pathway is kinetically disfavored, requiring significantly higher activation energy ((S)-TS5 = 32.7 kcal/mol). These findings collectively establish that stereoselectivity arises from NHC-mediated modulation of kinetic (transition state stabilization) and thermodynamic (intermediate stabilization) factors, with the catalyst dictating the energy landscape and stereochemical trajectory.

Fig. 7: Mechanistic investigation of NHC-catalyzed alcoholysis of oxazolone.
figure 7

The relative Gibbs free energies are reported from the SMD(DCM)-B3LYP-D3(BJ)/def2tzvp//B3LYP-D3(BJ)/def2svp level of theory, and the values are given in kcal/mol.

Full size image

Discussion

We present an NHC-catalyzed strategy for the stereocontrolled synthesis of chiral α-amino acids, enabling direct access to both natural and unnatural variants with precise stereochemical fidelity. This methodology unifies chiral α-(U)AA construction and peptide synthesis into a single catalytic framework, resolving persistent challenges in traditional approaches, including racemization and configurational instability. By leveraging NHC catalysts, the platform achieves dynamic stereochemical interconversion between S- and R-configured α-(U)AAs while maintaining chirality during peptide elongation—a dual capability critical for designing bioactive molecules. Operationally simple, base-free conditions and broad functional group tolerance accommodate diverse substrates, from sterically hindered α-aryl groups to unsaturated side chains, without compromising stereoselectivity. Mechanistic studies reveal that stereochemical outcomes are governed by NHC-mediated modulation of the reaction energy landscape, where π-π interactions and steric effects collaboratively dictate transition state preferences. This work bridges synthetic precision and biomolecular engineering, offering a versatile platform for biomolecular design and the development of peptide therapeutics.

Methods

General method for the NHC-catalyzed esterification of AAs and peptides

In an oven-dried 5 mL Schlenk tube equipped with a magnetic stir bar, α-amino acids 1 or peptides 5 (0.05 mmol), NHC precatalyst 3a (0.01 mmol, 20 mol%), alcohol 2 (0.075 mmol, 1.5 equiv), and Cs₂CO₃ (0.05 mmol, 1 equiv) were added, followed by 1 mL of dry dichloromethane (DCM). The resulting mixture was degassed and backfilled with argon (3 cycles), after which DIC (0.06 mmol) was added. The tube was sealed with a screw cap, and the reaction mixture was stirred vigorously at room temperature (25 °C) for 10 h. Upon completion (as monitored by TLC), the crude mixture was purified by flash column chromatography (petroleum ether/ethyl acetate = 10:1 to 4:1) to afford the corresponding ester products 4 or 6.

General method for the NHC-catalyzed racemization-free peptide synthesis

In a glove box under an argon atmosphere, the NHC precursor was pre-activated with NaH (1.2 equiv) for 1 h. After filtration, the resulting solution of free NHC-7a (0.01 mmol, 20 mol%) was added to a reaction mixture containing α-amino acid 1 (0.05 mmol) and amine 2b (0.055 mmol, 1.1 equiv) in an oven-dried 5 mL Schlenk tube equipped with a magnetic stir bar. Dry dichloromethane (DCM, 1 mL) and DIC (0.06 mmol) were subsequently added inside the glove box. The tube was sealed with a screw cap, and the mixture was stirred vigorously at room temperature (25 °C) for 6 h. Upon complete consumption of α-amino acid 1 (monitored by TLC), the crude reaction mixture was purified by flash column chromatography (petroleum ether/ethyl acetate = 2:1 to 1:1) to afford the desired peptide product 8.