Non-natural α-amino acids (α-AAs) are indispensable tools in modern chemistry, serving as protease-resistant pharmacophores and chiral ligands.^[1] However, existing methods for their synthesis often lack modularity, particularly for N-heteroaryl variants, which are underrepresented despite their pharmaceutical relevance.^[2-3] An illustrative instance of this approach involves incorporating a side chain carbonyl group, a synthetically significant functional moiety that can effectively enhance molecular complexity through diverse transformations. The presence of a carbonyl group in the side chain of amino acids, which is naturally produced through metabolic pathways, holds crucial significance in biological processes.^[4-7] Various methodologies for synthesizing γ-carbonyl amino acids have been documented in the literature.^[8-13] Nitrogen-containing heteroaromatic compounds are prevalent in pharmaceuticals due to their direct influence on solubility, metabolic stability, and binding affinity of the molecules.^[14-15] Incorporation of nitrogen-containing heteroaromatic rings like pyridine and quinine significantly alters the properties of amino acids and peptides. For instance, nicotinylalnine, a neuroprotective agent, acts as a potent inhibitor of two enzymes, kynurenine hydroxylase and kynurenine enzyme, whereas 2-amino-4-oxo-4-(pyridin-2- yl)butanoic acid, a pyridine derivative of aspartate, shows potential as an antagonist of NMDA receptors (Scheme 1a).^[16-18] However, limited research has been conducted on the synthesis of analogous compounds.^[16-17,19] Aryl formaldehyde is a commonly used class of reagents in chemical reactions. Recent advancements have been achieved in the activation of aldehyde C-H bonds using photocatalysts,^[203420-34] leading to the synthesis of unnatural α-amino acids (Scheme 1b).^[38-39] Despite this progress, the photoinduced activation of aryl formaldehyde containing nitrogen-containing heterocyclic rings such as pyridine and quinoline has not been documented. In this study, we present a novel photocatalytic method for the synthesis of nitrogenous heterocyclic aromatic α-amino acids from nitrogen-containing heterocyclic aromatic formaldehydes (Scheme 1c). In this protocol, heteroaryl aldehydes bearing diverse functionalities can be incorporated. The reaction of peptides and the late-stage reduction of non-natural α-AAs demonstrate the practical application potential of this approach.

Our study commenced with an examination of the two model substrates involving nicotinaldehyde (1a) and methyl-2-(bis[(tert-butoxy)carbonyl]amino)-prop-2-enoate (2) (Table 1). When a mixture of 1a (0.2 mmol), 2a (1.5 equiv) and NaDT (3 mol%) in 1.0 mL of CH₃CN was irradiated by 390 nm LED for 24 h at room temperature, nonnatural α-AAs 3a was yielded in 96.5% yield (Table 1, Entry 1). Careful screening of solvents proved the irre- placeability of CH₃CN, all other solvents led to diminished yields (from trace to 58.3%, Table 1, Entry 2). Reducing the ratio of the feeding amounts of 1a:2a would not benefit the reaction, as well as increasing the ratio did (63.0%~ 80.1%, Table 1, Entry 3). Replacing NaDT with other photocatalysts also resulted in unsatisfactory yields (Table 1, Entry 4). The best NaDT dosages was confirmed to be 3% (Table 1, Entries 5~6). Changing the reaction times to other durations was found to be unsuitable for the transformation (41.9%~63.0%, Table 1, Entry 6).

With optimal reaction conditions, a study in the substrate scope of aryl aldehydes was commenced (Table 2). Aryl aldehydes bearing various alkyl or alkoxyl groups at different sites are tolerated by this method, delivering 3b~3l in a moderate to excellent yield of 41.5%~92.7%. When the halogen atoms of aryl aldehydes were introduced at different sites at the pyridine rings, the reactions still worked smoothly, affording unnatural α-AAs 3 in moderate to excellent yields of 42.6%~76.0% (3m~3x). The yields for ortho-halogen-substituted aldehydes are lower than those of para-halogen-substituted aryl aldehydes, likely due to electron density effects and steric effects of the halogen atoms, hindering the capture of H atoms on aldehyde groups by NaDT. Similarly, stronger electron- withdrawing para-halogen groups increase the electron density at the para-aryl carbon atoms and stabilized the aldehyde radicals, thus leading to higher yields in this reaction (3n, 3q and 3u). Electron-withdrawing substituents like trifluoromethyl and ester groups were also compatible in this approach, yielding in lower yield (28.5%~47.0%, 3y~3z), this disparity may be attributed to the electronic effect of the electron-withdrawing groups, rendering the substrate challenging to activate. Additionally, a substantial amount of unreacted raw material residue was observed after reactions were finished. Altering the pyridine group in 1a with various N-containing heteroaryl groups, for example, pyrimidinyl (3aa~3ab), quinolinyl (3ac~3ae) and isoquinolinyl (3ag~3ah), did not impede the reaction, yielding products 3 with 27.5%~85.1% efficiency. When exploring the practical potential of this protocol, it was found that peptides 4a and 4b were also contained in the appropriate substrates, yielding the target products 5a and 5b successfully (Scheme 2a), presenting a fresh concept for peptides modification. The late-stage reduction of the unnatural α-AAs 3a was also conducted, converting the carbonyl group to valuable hydroxyl group in 6a, which proved the more possibilities for derivatization of this method. To gain further insight of the reaction mechanism, a series of mechanistic experiments were conducted. Firstly, when 3.0 equiv. TEMPO (2,2,6,6- tetramethylpiperidin-1-oxyl) radical scavenger was added to the model reaction, the formation of 5a was completely suppressed (Scheme 3a). This indicates a radical reaction pathway of this transformation. Secondly, under 390 nm illumination (5 min), various photocatalysts underwent C—H activation through a reaction with 1a.^[37] The results depicted in Scheme 2b indicate that only NaDT was able to activate 1a into acarbonyl radical under illumination, which was abducted with DMPO and detected by electron paramagnetic resonance spectrum (EPR) with g=2.0042, AN=1.43 mT, AH=2.06 mT. In contrast, photocatalysts like FeCl₃, CuCl₂, anthraquinone, and others did not exhibit the signal of the respective carbon centered free radical under the same illumination conditions. Based on the above mechanistic experiments, the following reaction mechanism is proposed (Scheme 4). Under the light irradiation, NaDT is excited to NaDT*, which then reacts with 1a to abstract a hydrogen atom from the carbonyl group, generating acyl radical 1a-r and NaDT-H. Subsequently, 1a-r is captured by the alkene 2 to form a Giese-type carbon radical 3a-r. Finally, 3a-r undergoes a reverse hydrogen abstraction with NaDT-H to afford 3a and regenerate NaDT.

In conclusion, we merged photocatalysis with radical chemistry to establish a versatile platform for α-AAs synthesis, overcoming long-standing challenges in N-hete- roaryl aldehyde activation. This method leverages the generation of acyl radicals from heteroaryl aldehydes, combined with the Giese addition of alkenes. The reaction exhibits fabulous functional group tolerance and offers an access for the precise installation of a wide range of aryl carbonyl group to α-AAs precursors. Besides these, the resulted unnatural α-AAs bearing various functional groups, such as methoxy, halogen atom and ester groups, can be further derivatized to valuable molecule based on α-AAs skeletons. Together with the compatible reaction with peptides, all these experiments illustrated the practical potential of this method. Mechanistic studies included EPR study revealed the proposed pathway to the synthesis of unnatural α-AAs. In short, this approach provides a powerful and facile pathway for constructing unnatural α-AAs with carbonyl and N-heteroaryl groups.