N-Heterocyclic Carbene (NHC) Catalysis in Enzymatic Organic Synthesis

1. Introduction

Enzymatic synthesis has played a vital role in modern synthetic chemistry and catalysis due to the generally high chemo-, regio- and stereo-selectivity observed in enzyme-catalyzed transformations [1–4]. In addition, the mild reaction conditions and the use of neutral aqueous media as solvents in enzymatic synthesis have made it highly attractive for sustainable chemistry and pharmaceutical industry [5–7]. In the past decade, the widely application of protein engineering and high throughput screening techniques have accelerated the pace of developing enzymes catalyzed non-natural transformations [8–11], as exemplified by many impressive reactions catalyzed by engineered P450 enzymes [12–19], non-heme iron enzymes [20–26] and flavin-dependent enzymes [27–30].

N-heterocyclic carbenes (NHCs) are cyclic compounds bearing a divalent carbene carbon atom that is embedded within the ring structure and connected to at least one nitrogen atom [31]. NHCs behave as good nucleophiles and Lewis bases [31], which make them one of the most ubiquitous ligands [32–41] in transition-metal catalysis as well as effective organocatalysts [42–46] (Scheme 1a). The connection between NHC organocatalysis and enzymatic catalysis originates from Nature. Thiamine diphosphate (ThDP, the biologically active form of vitamin B1) is a naturally occurring thiazolium salt that employs the same umpolung (inversion of polarity) strategy exploited by synthetic NHCs [47] (Scheme 1b). To elucidate the mechanism of this process, many leading scientists have dedicated significant efforts to this field [47]. A breakthrough came from observations of Ugai et al. that thiazolium salts could catalyze the benzoin reaction (Scheme 1c), which is reminiscent of acetoin formation in the enzymatic decarboxylation of pyruvate [48]. This finding was followed by the landmark studies of Breslow [49], in which the now widely accepted Breslow intermediate was proposed (Scheme 1d). These foundational discoveries, together with the seminal reports of isolable, stable NHCs by Arduengo [50] in the 1990s (Scheme 1e), built a solid foundation for the later broad applications of NHCs in organic synthesis [31–46, 51–53].

Scheme 1.

N-Heterocyclic carbene catalysis and thiamine diphosphate.

The emergence of NHC catalysis was inspired by ThDP-dependent enzymes, meanwhile, the flourishing of NHC catalysis has also triggered renewed interest to utilize these enzymes for synthetic applications beyond their natural roles, such as accepting non-natural substrates and catalyzing new reactions. During the past two decades, ThDP-dependent enzymes, including benzoylformate decarboxylase, benzaldehyde lyase, and transketolase, have been widely used in the benzoin-type reactions to create C-C bond [54–63]. Through protein engineering, these enzymes have been redesigned to catalyze synthetically challenging transformations, for example, asymmetric cross-benzoin condensations, Stetter reactions, and even asymmetric radical transformations (Table 1).

Table 1.

Selected examples of ThDP-dependent enzymes catalyzed C–C bond formations in organic synthesis (1999–2025)

Year	Enzyme	Author	Transformation	Ref.
1999	BFD	M.Müller	General synthesis of chiral benzoins	[64]
2002	BAL	M.Müller	First asymmetric cross-benzoin condensation	[65]
2010	YerE	M.Müller	Asymmetric aldehyde–ketone carboligation reaction	[66]
2010	PigD	M.Müller	Asymmetric intermolecular Stetter reaction	[67]
2015	BAL	D. Baker	First enzymatic Formose reaction	[68]
2017	BAL	P. Clapés	Asymmetric intramolecular benzoin reaction	[69]
2021	BFD	Y. Ma	C1 fixation via the Formose reaction to produce starch	[70]
2021	BAL	Q. Wu	Asymmetric intramolecular Stetter reaction	[71]
2024	BAL BFD	X. Huang Y. Yang	Visible light driven enantioselective radical acylation of redox-active esters	[72] [73]
2025	BAL	Y. Xie	Cross-benzoin reaction using highly enolizable 2-phosphonate aldehydes	[74]
2025	BAL	X. Huang	Asymmetric three-component radical reactions	[75]
2025	ALS	H.Takashi	Asymmetric radical acylation of α-bromo carbonyls	[76]
2025	BAL	R. Fasan X. Huang	Photobiocatalytic enantioselective benzylic C(sp3)–H acylation	[77] [78]
2025	HKAS	M.Ma	Cross-condensation of α-keto acids	[79]
2025	KdcA	X. Huang	Dynamic kinetic oxidation driven by electricity	[80]

BFD: benzoylformate decarboxylase

BAL: benzaldehyde lyase

YerE: cytidine diphosphate- yersiniose biosynthetic protein

PigD: an enzyme involved in the biosynthesis of cytotoxic agent prodigiosin

ALS: acetolactate synthase

HKAS: α-hydroxy-β-keto acid synthase

KdcA: branched-chain 2-keto acid decarboxylase

The major ThDP-dependent enzymes, including BAL, BFD, PDC, and TK, utilize a similar chemical mechanism, relying on the same cofactor (ThDP) to form the universal Breslow intermediate. Despite this common mechanistic foundation, the catalytic diversity of those ThDP-dependent enzymes arises not from their cofactor, but from their active site architectures. It is the different arrangement of amino acid variations surrounding this core and the specific binding mode of the substrate that direct this versatile intermediate toward a unique reaction and confer high selectivity. (Table 2).

Table 2.

The cofactor and amino acid variations of major ThDP-dependent enzymes (BAL/BFD/PDC/TK).

type	enzyme	cofactor	metal	amino acid variations	ref.
BAL	PfBAL	ThDP	Mg2+	WT	[65/69/71/81/89]
	PfBAL	ThDP	Mg2+	T481L/A480L	[72]
	PfBAL	ThDP	Mg2+	A28G	[74]
	PfBAL	ThDP	Mg2+	T481L/A480G/Q113H/N283F/H26E	[75]
	PfBAL	ThDP	Mg2+	A28S or A28S/A480W	[77]
	PfBAL	ThDP	Mg2+	T481L/A480G/Y397A/W163C	[78]
	PaBAL	ThDP	Mg2+	WT	[74]
	HeBAL	ThDP	Mg2+	A27I/V29I/G417S/A476L	[90]
	HeBAL	ThDP	Mg2+	A27I/V29I/ G417S/I552L/M553L	[91]
BFD	PpBFD	ThDP	Mg2+	WT	[64]
	PpBFD	ThDP	Mg2+	H281A	[65]
	PpBFD	ThDP	Mg2+	L476W or M365L/L461S	[98]
	PpBFD	ThDP	Mg2+	L461A or L461G	[99]
	PpBFD	ThDP	Mg2+	H281V	[73]
	PpBFD	ThDP	Mg2+	L476Q	[158]
PDC	ApPDC	ThDP	Mg2+	E469G	[89/101/159]
	ApPDC	ThDP	Mg2+	W388I	[101]
	ApPDC	ThDP	Mg2+	E469G/T384G/I468A	[102]
	ApPDC	ThDP	Mg2+	E469G, E469G/W543H	[103]
	ZmPDC	ThDP	Mg2+	WT	[104]
	SpPDC	ThDP	Mg2+	WT	[104]
TK	TKgst	ThDP	Mg2+	D469E or D469K	[107]
	TKgst	ThDP	Mg2+	H102T or H102L/H474S	[109]
	TKgst	ThDP	Mg2	L382F	[110]
	TKgst	ThDP	Mg2	L382N/D470S	[133]

PDC: pyruvate decarboxylase

TK: transketolase

In this review, we present an overview of important advancements achieved in enzymatic organic synthesis involving NHC catalysis from 1999 to 2025, focusing on reactions catalyzed by ThDP-dependent enzymes, and their diverse applications. The review is organized by the reaction type and further subdivided by the class of ThDP-dependent enzymes involved. The first section covers the widely investigated transformations, including Benzoin reaction, Stetter reaction and the recently exploited fascinating asymmetric radical transformations. Applications in the in situ carboligation processes, the formose reaction, and the preparation of chiral 1,2-amino alcohols and chiral vicinal diols are also discussed. We conclude with a discussion of the artificial metalloenzymes embedded with NHC-metal complexes as well as selected gold(I)-carbene catalysts containing thiamine analogs since they provide some valuable insights into the integration of metal-NHC complexes with enzymes, an appealing topic for both synthetic chemists and protein engineers. By discussing reactions promoted by ThDP-dependent enzymes and their applications, we hope to deepen understanding of enzymatic synthesis involving NHC catalysis and encourage more researchers to embrace this powerful and versatile approach within the field of both chemosynthesis and biosynthesis.

2. N-Heterocyclic Carbene (NHC) Catalysis with ThDP-Dependent Enzymes

2.1. Benzoin Reactions

2.1.1. Benzaldehyde Lyase

In 1989, González and Vicuña successfully identified a powerful ThDP-dependent enzyme, benzaldehyde lyase (PfBAL), from Pseudomonas fluorescens Biovar I [81]. They demonstrated that PfBAL could convert the α-hydroxy ketones benzoin and anisoin into benzaldehyde and anisaldehyde (reverse benzoin reaction). In 2001, Müller group showed that BAL could also catalyze the benzoin-type reaction [82]. Since then, many impressive examples have been reported[62]. The preparation of mixed benzoins is of significant interest due to their versatility as building blocks in organic synthesis. However, the chemoselectivity issues caused by undesired homocoupling reactions had made the asymmetric cross-benzoin reaction a difficult transformation [42]. In this context, Müller et al. [65] reported the first ThDP-dependent enzymes catalyzed asymmetric cross-benzoin reaction using BAL and a mutant of benzoylformate decarboxylase (BFD H281A7) (Scheme 2). Aldehydes containing ortho-substituents, such as 2-chlorobenzaldehyde, were chosen as acceptor substrates due to their inability to participate in the enzyme-catalyzed homocoupling. A broad range of aryl aldehydes efficiently participated in this transformation to give enantioenriched α-hydroxy ketones in good to excellent chemo- and enantioselectivity. This transformation serves as a notable example of how enzyme catalysis can address challenges that traditional small-molecule methods encounter.

Scheme 2.

BAL- and BFD-catalyzed first asymmetric cross-benzoin condensation.

In 2003, Rosazza et al. employed benzaldehyde lyase (BAL) to catalyze the reaction between benzaldehyde derivatives and phenylacetaldehyde, affording mixtures of products 1 and 2 with excellent enantioselectivity, albeit in moderate yields. Additionally, this study represents the first use of biocatalytic method to synthesize 1-hydroxy-1,3-diphenylpropan-2-ones and α-(R)-hydroxydihydrochalcones, which are important intermediates in the chemoenzymatic preparation of flavonoids (Scheme 3) [83].

Scheme 3.

Condensation reaction between phenylacetaldehyde/benzaldehyde catalyzed by BAL.

In 2003, the first acyloin-type condensation reaction between methoxy or dimethoxy-substituted acetaldehyde and aromatic aldehydes catalyzed by BAL was reported by Müller et al.[84]. A variety of enantiomerically pure 2,3-dioxygenated aryl propanones were obtained with excellent chemo- and enantioselectivity. Due to the reversible nature of the Benzoin reaction (Scheme 1d), benzoin products can also be converted back to the corresponding aldehydes. This reversibility has been demonstrated in the kinetic resolution of racemic 2-hydroxy-1,2-diphenylethan-1-one through BAL-mediated C–C bond cleavage coupled with the simultaneous C–C bond construction. Notably, when racemic benzoin rac-3a was used as the substrate, only the R enantiomer reacted with 2-dimethoxyacetaldehyde and 2-methoxyacetaldehyde to form the corresponding products (R)-4a and (R)-5a with excellent enantioselectivity, while the S enantiomer (S)-3a remained unreacted (Scheme 4).

Scheme 4.

Enantioselective carboligation of methoxy-substituted acetaldehyde with aryl aldehydes.

In 2006, Pohl et al. broadened the product range of BAL-catalyzed benzoin-type reaction through systematic optimization of the reaction conditions. Highly substituted hydroxybutyrophenone was obtained from benzaldehyde and glyceraldehyde acetonide, while aliphatic acyloins could be synthesized from aliphatic aldehydes. The resulting enantiopure, substituted hydroxybutyrophenones represent valuable intermediates for the synthesis of chiral polyols [85]. In 2018, Müller group successfully employed this strategy to achieve the synthesis of chiral tetrols via a tandem reduction/deprotection of hydroxybutyrophenone 6, furnishing chiral tetrols 8 in high yields and with excellent stereoselectivity [86] (Scheme 5).

Scheme 5.

Synthesis of hydroxybutyrophenones and acyloins via BAL-catalyzed benzoin reaction.

In 2008, Muller group investigated the BAL-catalyzed C–C bond ligation reaction between aliphatic/aromatic aldehydes and a,β-unsaturated aldehydes with the latter serving as donor substrates in most cases[87]. Additionally, small a,β-unsaturated aliphatic aldehydes could also act as acceptor substrates with aromatic aldehydes as donor substrates (Scheme 6). It should be noted that only 1,2-addition products were observed in this study.

Scheme 6.

Carboligation reaction of aliphatic/aromatic aldehydes with α, β-unsaturated aldehydes.

BAL has traditionally been used to synthesize molecules with a single stereocenter. A breakthrough occurred in 2013 when Müller et al. demonstrated that BAL could catalyze the benzoin condensation between arylaldehydes and racemic branched aliphatic aldehydes, enabling the construction of two stereogenic centers with excellent stereoselectivity [88]. This approach combines enantioselective carboligation with kinetic resolution of racemic substrates, highlighting BAL’s capacity to create complex chiral architectures. Notably, when benzaldehydes and enantiopure (S)-2-methyl-butanal were used as substrates, a mixture of benzoin 9 and 10 was obtained in 70% and 30% yields, respectively, both with exceptional diastereoselectivity (>99%). These products could be easily separated using flash chromatography, providing access to potentially important chiral building blocks with two stereogenic centers for the preparation of natural products (Scheme 7).

Scheme 7.

Preparation of α-hydroxyl ketones with two stereogenic centers.

In 2016, Müller group reported a highly chemo- and enantioselective aliphatic–aromatic cross-benzoin reaction utilizing BAL as a catalyst, furnishing a wide variety of (R)-2-hydroxypropiophenone (HPP) derivatives 11 in moderate to high yields (up to 86%) with excellent enantioselectivity (up to >99% ee) [89]. Interestingly, the use of cyclopropanecarbaldehyde led to a reversal of selectivity, delivering (R)-PAC 12 as the major product with excellent stereoselectivity (Scheme 8). Notably, other ThDP-dependent enzymes exhibited distinct stereoselectivity; for instance, PpBFD-L461A led to the generation of (S)-HPP derivatives, while ApPDC-E469G (a variant of pyruvate decarboxylase) and LlKdcA (branched-chain 2-keto acid decarboxylase) predominantly produced (S)-PAC and (R)-PAC derivatives, respectively. This study demonstrated that choosing a suitable enzyme enables precise control over both the regio- and stereo-selectivity in the enzymatic cross-benzoin reaction.

Scheme 8.

BAL-catalyzed aromatic–aliphatic cross-benzoin reaction.

Logically, the intramolecular asymmetric benzoin condensation is easier to achieve than the intermolecular variant from the standpoint of organic synthesis. However, it was not until 2017 that Clapés group [69] reported the first example of ThDP-dependent enzymes catalyzed asymmetric intramolecular benzoin reactions (Scheme 9). PfBAL was identified as a suitable catalyst for the intramolecular carboligation reaction of two benzaldehyde derivatives linked by an alkyl chain, affording several enantioenriched cyclic benzoin-type products in good yields with moderate to excellent enantioselectivities (up to 99% ee). Notably, substrate bearing a pyridine moiety could also furnish the product with excellent enantioselectivity albeit with a low chemoselectivity.

Scheme 9.

BAL-catalyzed asymmetric intramolecular benzoin reactions.

BAL enzymes have also been engineered to promote the condensation reaction of diverse aldehydes with formaldehyde. In 2023, Ma group found that HeBAL A27I/V29I/G417S/A476L was an efficient catalyst for such transformation [90], enabling the construction of hydroxyl methylketone derivatives in excellent yields. Additionally, the reaction of formaldehyde with 3-hydroxypropanal afforded 1,4-dihydroxybutan-2-one, which could be easily reduced to 1,2,4-butanetriol (Scheme 10). This triol is an important intermediate for the preparation of 1,2,4-butanetriol trinitrate, a compound widely utilized both as a component of rocket propellants and a plasticizer in manufacturing of explosives.

Scheme 10.

HeBAL-catalyzed C–C bond formation of diverse aldehydes with formaldehyde.

Furthermore, the carboligation of arylaldehydes and formaldehyde using mutant HeBAL A27I/V29I/G417S has been reported [91], affording a broad range of aromatic α-hydroxymethyl ketones in high yields. A one-pot two-step biocatalytic method was also developed for synthesizing optically active 1,2-diols from simple aldehydes and formaldehyde with excellent stereoselectivity through sequential hydroxymethylation-reduction reactions (Scheme 11).

Scheme 11.

HeBAL A27I/V29I/G417S-catalyzed hydroxymethylation of arylaldehydes.

Due to the undesirable aldol-type side reactions and the deactivation of nucleophiles by acidic H at the α-position of aldehyde, the selective C-C coupling reaction of highly enolizable aldehydes with nucleophiles is more challenging than with simple alkyl or aryl aldehydes. In 2025, Xie and coworkers reported a convenient method for synthesis of a broad range of biologically active β-hydroxy phosphonates from arylaldehydes and highly enolizable 2-phosphonate aldehydes by using PfBAL and its variant PfBAL A28G as catalyst [74], achieving yields of up to 95% and enantioselectivities of up to 99% ee (Scheme 12). It should be noted that commonly used chiral NHC catalysts failed to give highly chemoselective cross-benzoin condensation products as a complex mixture of products were observed. Moreover, by employing benzaldehyde lyase from Polymorphobacter arshaanensis (PaBAL), the product configuration was successfully switched from R to S. This configuration reversal may be attributed to the different orientations of substrate entry into the enzyme’s active cavity.

Scheme 12.

PfBAL and PaBAL catalyzed benzoin-type reaction of arylaldehydes and highly enolizable 2-phosphonate aldehydes.

2.1.2. Benzoylformate decarboxylase

Benzoylformate decarboxylase (BFD) has been identified in several bacteria, such as Acinetobacter calcoaceticus, Pseudomonas aeruginosa and Pseudomonas putida [92, 93]. A primary function of BFD is the transformation of benzoylformate into benzaldehyde through a non-oxidative decarboxylation reaction as founded in the mandelate pathway[94]. In 1992, Wilcocks and co-workers first revealed the catalytic capability of BFD to facilitate C–C bond formation [95]. They found that crude extracts of Pseudomonas putida could promote the reaction between acetaldehyde and benzoylformate, generating (S)-2-hydroxy-1-phenylpropanone ((S)-2-HPP). In this process, BFD first mediates the nonoxidative decarboxylation of benzoylformate to form benzaldehyde. The benzaldehyde is then activated by the ThDP cofactor to form a Breslow intermediate, which subsequently condenses with acetaldehyde to furnish 2-hydroxy-1-phenylpropanone product.

In 1999, Müller and co-workers demonstrated the first general synthesis of (R)-benzoin via BFD-catalyzed aldehyde self-condensation [64]. By employing aromatic and heteroaromatic aldehydes as substrates, they established a reliable biocatalytic route to produce (R)-benzoin, offering a sustainable alternative to traditional chemical methods (Scheme 13). This study and the independent work by Patel et al. could be considered as a breakthrough in the ThDP-dependent enzyme-catalyzed formation of chiral α-hydroxy ketones from benzaldehydes [96].

Scheme 13.

First general synthesis of chiral benzoin catalyzed by BFD.

In 2000, Pohl, Müller and coworkers performed a systematic exploration of the substrate scope in BFD-catalyzed benzoin-condensation type reaction [97]. Employing acetaldehyde as the acceptor substrate, a diverse range of acyl donors could be efficiently transformed into the corresponding enantioenriched (S)-2-hydroxy ketones (Scheme 14).

Scheme 14.

BFD-catalyzed benzoin reaction to synthesize chiral 2-hyroxy ketones.

In 2003, Lingen et al. reported two mutations of BFD, L476W and M365L/L461S, exhibiting improved catalytic activity in carboligation reactions [98]. Notably, both variants could efficiently catalyze the condensation of acetaldehyde with steric hindered ortho-substituted benzaldehyde derivatives, delivering chiral 2-hydroxy ketones in excellent conversion yields (up to 100%) and high enantioselectivity (up to >99% ee) (Scheme 15).

Scheme 15.

BFD-catalyzed reaction of ortho-substituted benzaldehydes with acetaldehyde.

In 2008, Pohl and coworkers identified certain mutations of BFD (e.g., Leu461→Ala/Gly) that can accommodate bulkier aldehyde acceptors [99]. These engineered variants catalyzed the S-selective carboligation reaction of benzaldehydes and longer chain aliphatic aldehydes, achieving significantly higher enantioselectivity compared to the wtBFD. A variety of chiral 2-hydroxy ketones were obtained with excellent enantioselectivity. This study highlighted how enzyme engineering can expand the substrate scope of wild-type enzymes (Scheme 16).

Scheme 16.

BFD-catalyzed carboligation reaction of benzaldehydes and longer chain aliphatic acceptor aldehydes.

2.1.3. Pyruvate Decarboxylase

Pyruvate decarboxylase (PDC) has long been renowned for its application in decarboxylation of α-ketoacids to aldehydes. Interestingly, it can also promote the carboligation reaction of pyruvate with benzaldehyde, producing (R)-phenyl-acetylcarbinol (PAC) during fermentation [57]. The mechanism of this process is similar to that of BFD, PDC initially catalyzes the decarboxylation of pyruvate, generating acetaldehyde. The acetaldehyde is then activated by the ThDP cofactor to form a Breslow intermediate, which undergoes condensation with benzaldehyde to give PAC product.

In an earlier report, Crout et al. demonstrated that purified pyruvate decarboxylase from yeast could efficiently catalyze the reaction between benzaldehyde and pyruvate, affording a series of chiral phenyl acetyl carbinols with excellent enantioselectivity (Scheme 17) [100]. An advantage of this approach is the higher optical purity of the products compared to those obtained using whole-cell yeast. Additionally, the use of purified enzyme eliminates competing side reactions, such as the reduction of aldehydes to alcohols by dehydrogenases present in whole organisms.

Scheme 17.

PDC-catalyzed enantioselective synthesis of acyloins with pyruvate.

Typically, wild-type PDC prefers to catalyze the formation of R-enantiomer products. An impressive work by Pohl and coworkers [101] demonstrated that substituting the E469 residue with glycine in PDC from Acetobacter pasteurianus (ApPDC) enables the enzyme to synthesize (S)-2-hydroxy ketones. Interestingly, both ApPDC and its mutants preferentially utilize aliphatic aldehydes as donor substrates in the presence of aromatic aldehydes. However, product yields tend to decrease as the carbon chain length of the aliphatic aldehydes increases (Scheme 18). It should be noted that this variant expands the toolbox of ThDP-dependent enzymes, providing avenue to (R)-phenylacetylcarbinol derivatives.

Scheme 18.

PDC-catalyzed mixed carboligation between aliphatic and aromatic aldehydes.

The synthesis of (S)-benzoin has long presented a challenge in enzymatic catalysis. In this context, Pohl et al. addressed this issue by using ApPDC variants [102]. Substrate scope investigations revealed that both mutants can catalyze the homocoupling reactions of substituted benzaldehydes, enabling the preparation of enantiopure (S)-benzoins from benzaldehyde derivatives with excellent enantiomeric excess (up to >99%) and good yields. Interestingly, meta-substituted benzaldehydes exhibited higher yields and enantioselectivity compared to ortho- and para-substituted as well as unsubstituted benzaldehydes (Scheme 19).

Scheme 19.

ApPDC mutants-catalyzed homocoupling affording S-enantiomer products.

In 2017, Rother group [103] reported a highly efficient method for the enantioselective synthesis of (S)-PAC derivatives using ApPDC and its engineered variants. A variety of arylaldehydes underwent the carboligation reaction with pyruvate to give the corresponding (S)-PAC derivatives with high enantioselectivities. Under the optimized conditions, a product concentration of up to 320 mM could be obtained, highlighting the potential of this method for large-scale industrial applications (Scheme 20).

Scheme 20.

Synthesis of (S)-PAC derivatives by ApPDC and its variants.

In 2018, Zhu, Wu and Yao et al. [104] developed an inventive biocatalytic cascade reaction to synthesize chiral cyclic diols using PDC and alcohol dehydrogenases (ADH). The process commenced with a PDC-catalyzed intramolecular carboligation of aliphatic dialdehydes, yielding chiral α-hydroxy ketones, which were subsequently transformed into optically active vicinal diols via an asymmetric reduction mediated by ADH (Scheme 21). Various combinations of PDCs and ADHs from different sources were evaluated using glutaraldehyde or adipaldehyde as starting materials, resulting in the formation of the corresponding chiral cyclic diols with complementary stereoselectivity. This multi-enzyme system offers an efficient and highly versatile platform for the stereoselective conversion of adipaldehyde and glutaraldehyde into cyclohexane-1,2-diols and cyclopentane-1,2-diols, respectively.

Scheme 21.

Enzymatic cascade reaction to synthesize cyclic vicinal diols catalyzed by PDC and ADH.

2.1.4. Transketolase

Transketolase (TK) is a ThDP-dependent carbon–carbon bond forming enzyme, first identified in S. cerevisiae by Racker et al. [105]. Its role in Nature is to reversibly transfer a two carbon ketol fragment from D-xylulose-5-phosphate to D-ribose-5-phosphate and D-erythrose-4-phosphate [106]. TK is an efficient catalyst for the transformation of α-hydroxyaldehydes into α,α-dihydroxyketones. Although wild-type TKs can accept certain non-α-hydroxylated aliphatic aldehydes as substrates, these transformations typically proceed with a significantly lower catalytic activity compared to reactions involving α-hydroxyaldehydes [106]. In 2010, Hailes and co-workers [107] obtained several TK mutants through directed evolution that could utilize nonhydroxylated aldehydes, such as benzaldehydes and phenylacetaldehydes as acceptors, affording the corresponding α-dihydroxyketone products with high stereoselectivity (Scheme 22). Notably, the wild-type enzyme exhibited no activity towards benzaldehydes.

Scheme 22.

TK mutant-catalyzed synthesis of dihydroxyketones with nonhydroxylated aldehydes.

In 2012, Hanefeld and co-workers[108] identified several TK mutants, including R526Q, R526N, R526K/S525T and R526Q/S525T, which exhibit improved catalytic activity in the condensation reaction between nonphosphorylated lithium hydroxypyruvate (HPA) and D-glucose, glycolaldehyde (GO), or D-ribose (rib). The activity of mutant R526Q/S525T demonstrated up to a fourfold increase in activity over the wild-type TK for substrates LiHPA and glycolaldehyde, and a 2.6-fold improvement when using LiHPA and D-ribose.

In 2017, Fessner group [109] identified several TK variants that can accept pyruvate and longer-chain aliphatic analogues as donor substrates through the use of protein engineering (Scheme 23). Among them, the H102T single mutant exhibited optimal activity in reactions when 3-methyl-2-oxobutyrate was as the donor substrate. In contrast, variant H102L/H474S demonstrated good catalytic performance using pyruvate as donor.

Scheme 23.

TKgst variants-catalyzed asymmetric carboligation reactions between oxoacids and aldehydes.

In 2020, Charmantray, Hecquet, and co-workers [110] reported a structure-based approach to engineer TK from Geobacillus stearothermophilus (TKgst), obtaining several enzyme variants capable of stereoselectively synthesizing aliphatic acyloins with diverse carbon chain lengths (ranging from C5 to C10). Among them, the single-point mutant L382F efficiently transferred the ketol unit from hydroxypyruvate to various aliphatic aldehydes (ranging from C3 to C8), yielding a broad range of chiral 1,3-dihydroxy ketones. In addition, the triple mutant H102L/H474S/F435I was capable of transferring acyl group from 2-oxobutyrate (Scheme 24).

Scheme 24.

TKgst variants catalyzed formation of aliphatic acyloin.

2.1.5. Miscellaneous

YerE, a ThDP-dependent enzyme from Yersinia pseudotuberculosis (YpYerE), participates in the biosynthetic pathway of branched-chain sugar yersiniose A [111]. In 2010, Müller group[66] employed YpYerE to achieve the asymmetric aldehyde–ketone carboligation reaction (Scheme 25). Various aldehydes as well as open-chain and cyclic ketones underwent carboligation with pyruvate to afford chiral tertiary alcohols with moderate to excellent enantiomeric excess. Notably, this transformation represents the first example of ThDP-dependent enzymes catalyzed enantioselective aldehyde–ketone carboligation, offering a new route for creating chiral tertiary alcohols.

Scheme 25.

The first ThDP-dependent enzyme catalyzed-asymmetric aldehyde–ketone condensation.

In 2018, Dobritzsch, Müller and co-workers [112] demonstrated that YerE from Pseudomonas protegens (PpYerE), also exhibits comparable catalytic performance. Employing pyruvate as the acyl donor, various benzaldehyde derivatives and ketones were successfully transformed into optically active 2-hydroxy ketones products (Scheme 26). Moreover, the successful crystallization of PpYerE enabled detailed site-directed mutagenesis studies, providing valuable insights into the structure-function relationships of YerE.

Scheme 26.

PpYerE-catalyzed carboligation reactions of pyruvate with aldehydes or ketones.

Cyclo-hexane-1,2-dione hydrolase (CDH), a ThDP-dependent enzyme, is involved in an anaerobic degradation pathway of alicyclic alcohols [113]. CDH, originated from Azoarcus sp. strain 22Lin, is capable of cleaving carbon–carbon bonds and facilitating the asymmetric formation of non-physiological C–C bonds [114, 115]. In 2014, Müller group found that purified recombinant His-tagged CDH exhibited identical activities in C–C bond formation to those of the native enzyme [116]. A variety of arylaldehydes underwent carboligation with pyruvate to afford the corresponding (R)-PAC derivatives with excellent enantioselectivity. It should be noted that CDH can also accept hydroxybenzaldehydes as substrates, which have rarely been observed for other ThDP-dependent enzymes (Scheme 27).

Scheme 27.

CDH-catalyzed condensation of pyruvate with aromatic aldehydes.

In 2014, Müller et al. [117] found that the dual reactivity of CDH could be selectively modulated through mutagenesis. The single mutant CDH-H28A exhibited diminished C–C bond formation activity while its native function of cleaving C–C bond remained unaffected. In contrast, mutant CDH-H28A/N484A displayed the reverse behavior, exhibiting good performance in promoting the coupling of pyruvate with cyclohexane-1,2-dione. Moreover, other 1, 2-ketones and monoketones could also be used as substrates, producing enantioenriched tertiary α-hydroxy ketones (Scheme 28). This study broadened the scope of ThDP-dependent enzymes catalyzed formation of chiral tertiary alcohols by aldehyde-ketone coupling, a class of compounds which are difficult to obtain through non-enzymatic organic synthesis.

Scheme 28.

CDH variant-catalyzed carboligation between pyruvate and aldehydes or ketones.

Branched-chain 2-keto acid decarboxylase (KdcA), identified from Lac-tococcus lactis sup. cremoris B1157, is a ThDP-dependent enzyme [118]. In 2007, Müller and colleagues [119] found that KdcA could efficiently transform diverse aromatic aldehydes and aliphatic aldehydes into 2-hydroxy ketones with excellent enantiomeric purity. Interestingly, in reactions involving aryl aldehydes combined with different aliphatic aldehydes, the selectivity is significantly affected by the selected substrate combination. Depending on the aliphatic aldehyde used, the reaction can shift from predominantly producing (R)-phenylacetyl carbinol (PAC) derivatives when larger aliphatic aldehydes were used, to generating an almost equal distribution of hydroxypropiophenone (HPP) and PAC derivatives using acetaldehyde. Notably, the reaction of 3,5-dichlorobenzaldehyde with acetaldehyde favored the formation of a HPP derivative (Scheme 29).

Scheme 29.

KdcA-catalyzed carboligation reactions.

The physiological role of ThDP-dependent enzyme Acetoin:dichlorophenolindo-phenol oxidoreductase (AoDCPIP OR) is to mediate the C-C bond cleavage of acetoin (reverse benzoin condensation) [120]. In 2015, Giovannini et al. [121] demonstrated that AoDCPIP OR from Bacillus licheniformis is a powerful biocatalyst for the enantioselective preparation of chiral tertiary alcohols through an enzymatic aldehyde–ketone condensation reaction (Scheme 30). Remarkably, this enzyme can utilize methylacetoin as an acyl anion surrogate, allowing for the selective formation of chiral tertiary α-hydroxy ketones and suppressing the undesired homocoupling side reactions typically observed with other donors. Notably, several chiral tertiary alcohols obtained with Ao:DCPIP OR display stereochemical outcomes opposite to those catalyzed by YerE, further highlighting its synthetic potential in the preparation of tertiary α-hydroxy ketones.

Scheme 30.

Ao:DCPIP OR-catalyzed cross-coupling reactions.

In 2016, Massi and Giovannini et al. [122] demonstrated that wild type Ao:DCPIP OR is capable of catalyzing condensation reaction of methylacetoin, a precursor of acetyl anion, with various arylaldehydes, leading to the efficient formation of chiral α-hydroxy ketones with enantiomeric excesses as high as 99%. Notably, substrates that are typically poor acceptors, such as hydroxybenzaldehydes and 4-(tert-butyl)benzaldehyde, were well tolerated and efficiently transformed into the desired products (Scheme 31).

Scheme 31.

Ao:DCPIP OR catalyzed enantioselective synthesis of (S)-PAC derivatives.

In addition to arylaldehydes and ketones, Ao:DCPIP OR can also exploit aliphatic aldehydes as acceptor substrates. In 2018, Giovannini et al. [123] reported a Ao:DCPIP OR catalyzed condensation reaction of methylacetoin or 4-hydroxy-4-methylhexan-3-one with aliphatic aldehydes, producing diverse chiral 1-alkyl-1-hydroxyketones in high yields (up to 99%). It is of note that most of these chiral aliphatic hydroxyketone products had not been previously accessed through either organocatalytic or enzymatic benzoin-type reactions. This study also represented the first example of employing 4-hydroxy-4-methylhexan-3-one as a propionyl anion surrogate in thiamine-catalyzed C–C bond formation (Scheme 32).

Scheme 32.

Synthesis of 1-alkyl-1-hydroxyketones catalyzed by Ao:DCPIP OR.

The ThDP-dependent enzyme MenD (2-succinyl-5-enolpyruvyl-6-hydroxy-3-cyclohexene-1-carboxylate synthase) participates in the biosynthetic route of menaquinone by catalyzing a unique Stetter-type conjugate addition of α-KG to isochorismate [124]. In 2009, Müller and co-workers [125] investigated the non-natural C–C bond-forming capabilities of MenD from Escherichia coli K12. It was demonstrated that MenD can use α-ketoglutarate as donor substrates and various aldehydes as acceptors to produce diverse chiral α-hydroxyketones. The use of halide-substituted aromatic aldehydes as acceptor substrates resulted in higher yields compared to aliphatic aldehydes and α,β-unsaturated aldehydes (Scheme 33). In addition to α-ketoglutarate, MenD can also accept oxaloacetate and pyruvate as donors. Interestingly, the reversal of regioselectivity occurred when α-ketoglutarate reacted with pyruvate (activated acetaldehyde), which was initially ascribed to a reversal in the roles of the substrates as acyl donor and acceptor.

Scheme 33.

MenD-catalysed condensation reactions of α-ketoglutarate with aldehydes.

In 2013, Müller et al.[126] employed NMR and CD spectroscopy to study the above transformations, which revealed that the opposite regioselectivity in the reaction between α-ketoglutarate and pyruvate may result from the sequential rearrangement/decarboxylation process of the α-hydroxy-β-keto acid intermediate instead of a change in the roles of donor and acceptor substrates (Scheme 34).

Scheme 34.

Generation of 4-hydroxy-5-oxohexanoic acid from α-ketoglutarate and pyruvate.

In 2013, the same group [127] conducted a systematic study on MenD-catalyzed asymmetric C–C bond formations using α-ketoglutarate as acyl donor and various aldehydes as electrophilic acceptors. A series of arylaldehydes bearing inductively electron-withdrawing groups (e.g., halides) or electron-donating groups (e.g., methoxy) both gave chiral δ-hydroxy-γ-keto acid products with high yields and enantioselectivity. Aliphatic aldehydes can also participate in this transformation to afford the corresponding products with high conversion, albeit with low to moderate enantioselectivity. In addition, they investigated enzymes SucA (ThDP-dependent 2-oxo acid decarboxylase) and Kgd (α-ketoglutarate decarboxylase), both renowned for their decarboxylase activity, in asymmetric C–C bonds formation reactions. In contrast to Mend, SucA and Kgd demonstrated superior performance with aliphatic aldehydes of varying chain lengths, particularly in terms of enantioselectivity (Scheme 35).

Scheme 35.

MenD/SucA/Kgd-catalyzed carboligation of α-KG with various aldehydes.

Acetohydroxyacid synthase (AHAS), a ThDP-dependent enzyme, participates in the biosynthetic route of branched-chain amino acids, including isoleucine, leucine and valine [128–130]. In a study by Gao and co-workers, a glutathione S-transferase-tag fused catalytic subunit of E. coli AHAS I was utilized as a biocatalyst to generate (R)-phenylacetyl carbinol in 80% yield with enantiomeric excess exceeding 98% (Scheme 36) [131].

Scheme 36.

CSU-GST of E. coli AHAS I catalyzed formation of (R)-PAC.

In 2025, Ma and co-workers [79] reported two novel α-hydroxy-β-keto acid synthases, CsmA and BbmA, which catalyze the C-C coupling reaction between two β-keto acids (Scheme 37). While both enzymes exhibit a broad substrate scope, they display distinct substrate preferences. CsmA efficiently catalyzes the carboligation of two aromatic α-keto acids, and favors aromatic α-keto acid donors with bulkier side chains, but it does not tolerate α-keto acids containing aromatic substitutions at the α-position. In contrast, BbmA primarily catalyzes the coupling reaction of aliphatic acyl donors with aromatic acyl acceptors, and has poor compatibility with aromatic α-keto acids as donors. CsmA and BbmA have also been tamed to catalyze the coupling reaction of aromatic acyl donors with aliphatic acyl acceptors. It should be noted that α-hydroxy-β-keto acids have been reduced using NaBH₄ to avoid the undesired spontaneous decarboxylation of α-hydroxy-β-keto acids.

Scheme 37.

CsmA and BbmA-catalysed coupling reactions between different α-keto acids and the subsequent NaBH₄ reduction.

Apart from aldehydes and ketones, nitroso compounds can also be used as acceptor substrates in ThDP-dependent enzyme-catalyzed condensation reactions for the construction of C–N bonds. In 2011, Ayhan and co-workers [132] reported a highly efficient method to produce N-arylhydroxamic acids from aromatic aldehydes and nitroso compounds using benzaldehyde lyase. BAL is also capable of kinetically resolving rac-benzoin with nitrosobenzene, leading to the unreacted (S)-benzoin with enantiomeric purity exceeding 95% (Scheme 38).

Scheme 38.

BAL-catalyzed formation of C–N bonds from aldehydes and nitroso compounds.

Furthermore, thermostable transketolase from Geobacillus stearothermophilus (TKgst) can take nitroso compounds as acceptor substrates in condensation reactions. In 2021, Fessner and co-workers [133] reported a TKgst (L382N/D470S) variant that catalyzed condensation of nitrosobenzene and hydroxypyruvate, affording a series of N-aryl hydroxamic acids in moderate yields (Scheme 39).

Scheme 39.

Preparation of N-aryl hydroxamic acids catalyzed by TK.

2.2. Stetter Reactions

2.2.1. Benzaldehyde Lyase

The Stetter reaction is an organic transformation to form carbon-carbon bonds through a 1,4-addition reaction of aldehydes to Michael acceptors via umpolung mechanism utilizing a nucleophilic catalyst like thiazolium salts and cyanide anion, providing synthetically useful 1,4-dicarbonyl compounds [42]. The mechanism of thiazolium salts catalyzed Stetter reaction (Scheme 40) is similar to that of benzoin reaction (Scheme 1d), the key difference is the reaction pathway of the initially formed Breslow intermediate from thiazolium salts and aldehyde. In the Stetter reaction, the Breslow intermediate undergoes an irreversible 1,4-addition to Michael acceptors, while in the benzoin reaction, it undergoes 1,2-addition to another molecule of aldehyde in a completely reversible way.

Scheme 40.

Mechanism of Stetter reaction catalyzed by thiazolium salts.

The intermolecular asymmetric Stetter reaction presents significant challenges due to chemoselectivity issues arising from aldehyde homocoupling as well as difficulties in achieving high stereoselectivity. However, over the past two decades, several impressive examples have been reported in enzymatic stereoselective intermolecular Stetter reaction. In 2005, Salmond et al. [134] discovered a novel ThDP-dependent enzyme PigD from Serratia marcescens. This enzyme can catalyze decarboxylation of pyruvate and the resulting Breslow intermediate then reacted with 2-octenal to produce 3-acetyloctanal (Scheme 41).

Scheme 41.

Stetter reaction catalyzed by PigD.

In 2010, Müller and co-workers [67] investigated the reaction of pyruvate with aromatic and aliphatic α,β-unsaturated aldehydes catalyzed by PigD. However, only 1,2-addition product was observed even with 2-octenal as a substrate. Interestingly, 1,4-addition selectivity was obtained using α,β-unsaturated ketones. Various aliphatic, aromatic and heterocyclic unsaturated ketones were converted into the corresponding chiral 1,4-diketones with excellent enantioselectivity (Scheme 42). For example, ketone 17 containing two acetyl groups could be efficiently synthesized from either 15 or its isomeric substrate 16. Notably, this study represented the first example of enzymatic asymmetric Stetter reaction between α,β-unsaturated ketones and α-keto acids proceeding through an umpolung mechanism. Later, Müller and co-workers [135] identified two new ThDP-dependent enzymes SeAAS and HapD through protein BLAST analysis. These enzymes can catalyze Stetter reaction with high enantioselectivity as well as 1,2-addition, as observed in PigD.

Scheme 42.

The first asymmetric Stetter reaction catalyzed by PigD.

ThDP-dependent enzyme MenD is involved in the biosynthetic route of menaquinone, catalyzing a conjugate addition of ketoglutarate to isochorismate. However, the 1,4-addition catalyzed by MenD is highly specific and limited to isochorismate and its derivatives. In 2014, Müller and co-workers [136] found that MenD could also utilize short-chain α,β-unsaturated carboxylic acids and their derivatives as acceptors with α-ketoglutarate as the acyl donor in Stetter reaction. Notably, although acrylonitrile is a suitable substrate for MenD-catalyzed Stetter reaction, its structural analog methacrylonitrile failed to undergo this transformation. This difference reveals that the MenD-catalyzed Stetter reaction imposes strict limits on sterics and electronics of the acceptor substrates (Scheme 43).

Scheme 43.

MenD-catalyzed Stetter reaction.

In 2021, Wu et al. [71] reported the first enzymatic intramolecular enantioselective Setter reactions catalyzed by PfBAL. Molecular dynamic (MD) simulations showed that PfBAL features larger active-site pocket and shorter distance between the ThDP cofactor and the substrate’s carbonyl group compared to other ThDP-dependent enzymes. Thus, a wide variety of aromatic substrates featuring either electron-donating or electron-withdrawing groups were well tolerated in the PfBAL-catalyzed addition reaction, affording the corresponding 4-chromanone derivatives in excellent yields (up to 99%) and outstanding enantioselectivity (up to 98% ee). However, no product was observed when substrates bearing substituents at the 6-position were used, likely due to steric hindrance. This indicated that the efficient combination of ThDP cofactor and the carbonyl group plays a crucial role in the pfBAL-mediated intramolecular Stetter reaction (Scheme 44).

Scheme 44.

pfBAL-catalyzed intramolecular Stetter reaction.

2.3. Radical Reactions

2.3.1. Benzaldehyde Lyase

The control of chemo- and stereoselectivity in transformations involving free radical intermediates remains a major challenge in synthetic chemistry due to the high reactivity and short lifetime of radical species. In 2024, Huang and co-workers [72] reported an elegant example of a visible light driven enantioselective radical acylation reaction through a synergistic ThDP-dependent enzyme PfBAL/photoredox catalysis system. Through protein engineering, three highly active enzyme variants PfBAL T481L, PfBAL T481L A480L and PfBAL T481L A480G were identified for the radical cross-coupling of various aldehydes with redox-active esters. A series of α-chiral ketones (35 examples) have been synthesized using this strategy, achieving enantiomeric excesses up to 96% (Scheme 45). A possible mechanism was proposed as illustrated in Scheme 46, the Breslow intermediate Int. A generated from ThDP cofactor and arylaldehyde was oxidized to ThDP-derived ketyl radical Int. B by photoexcited eosin Y*. Int. B then undergoes enantioselective radical cross-coupling reaction with alkyl radical formed through the single-electron reduction of redox-active esters by Eosin Y^•−, leading to the regeneration of photocatalyst eosin Y and production of α-chiral ketones (Scheme 46). This impressive work not only significantly expanded the toolbox of biocatalytic radical transformations but also offered a powerful strategy for controlling radical intermediates using enzymes, complementing existing chemical strategies.

Scheme 45.

The engineered PfBAL-catalyzed enzymatic synthesis of chiral ketones.

Scheme 46.

Mechanism of light-driven enzymatic radical acylation catalyzed by PfBAL.

In 2025, Fasan group [77] reported a photobiocatalytic enantioselective C(sp³)–H acylation enabled by ThDP-dependent enzymes via intermolecular hydrogen atom transfer. In this system, the photocatalyst oxidizes Breslow intermediate to form an acyl radical, which subsequently captures a benzylic radical formed through HAT between an alkane and N-fluoroamide reagent, resulting in the formation of chiral ketone products. Various aromatic aldehydes bearing either electron-withdrawing or electron-donating groups can participate in this reaction to afford diverse chiral ketone products (Scheme 47). Although the yields and enantioselectivity were moderate, this study elegantly demonstrated that the combination of radical reactivity of ThDP-dependent enzymes with hydrogen atom transfer provides a promising strategy for the highly challenging asymmetric functionalization of C(sp³)–H bonds.

Scheme 47.

Photobiocatalytic enantioselective C(sp³)–H acylation enabled by engineered PfBAL.

In 2025, Huang, Long, Wang[78] and co-workers also reported a photobiocatalytic benzylic C–H acylation reaction enabled by engineered ThDP-dependent enzyme BAL using aldehydes as acyl donor. A variety of aldehydes and alkyl aromatic hydrocarbon substrates could participate in this reaction to afford the corresponding chiral ketone products in moderate to good yields with high enantioselectivity (up to 97% ee). Compared to Fasan’s work[77], this study used 1,3-dioxoisoindolin-2-yl acetate as the HAT regent and generally achieved higher enantioselectivity. A limitation of this and Fasan’s work is that aliphatic aldehydes and substrates without benzylic C–H bonds are not compatible. Based on the Density functional theory (DFT) calculations and wet experiments, a possible mechanism was proposed. Initially, aldehyde condenses with the ThDP cofactor to form a Breslow intermediate, which is then oxidized by an excited-state photocatalyst to generate a persistent ketyl radical and a reduced photocatalyst PC^•−. Then, the HAT reagent undergoes a single-electron reduction by PC^•− to form methyl radical, which could abstract a hydrogen-atom from the benzylic position of substrate to form a prochiral radical. Finally, within the enzyme’s active site, radical-radical cross-coupling between ThDP-derived ketyl radical and benzylic alkyl radical affords the chiral product (Scheme 48).

Scheme 48.

Asmmetric photobiocatalytic benzylic C–H acylation enabled by BAL.

Since an enzymatic active site is typically not suited to accommodate multiple substrates and the difficulty of achieving stereochemical control of radical intermediates, the realization of enzymatic asymmetric multicomponent radical reactions is a formidable challenge. In 2025, Huang group [75] reported an impressive example of enzymatic asymmetric three-component radical cross-coupling through the combination of 3CRE (three-component radical enzyme) engineered from PfBAL with photoredox catalysis. A broad range of readily available aldehydes, alkenes and α-bromo-carbonyls were efficiently converted to the corresponding chiral ketone products in good yields (up to 80%) with excellent enantioselectivity (⩾97% ee in 24 cases) (Scheme 49). This study, together with the well-established applications of radical NHC catalysis in organic synthesis, highlights the potential of chemomimetic ThDP-dependent biocatalysis as a powerful platform for developing new radical transformations that are both mechanistically intriguing and synthetically valuable.

Scheme 49.

PfBAL variants-catalyzed asymmetric three-component radical reactions.

2.3.2. Benzoylformate Decarboxylase

In 2024, Yang and co-workers [73] developed a related dual catalytic system for asymmetric radical alkylation by combining an organic dye fluorescein as a photoredox catalyst with two engineered ThDP-dependent enzymes BFD and BAL. Through single-site saturation mutagenesis (SSM), the PpBFD H281V mutant was identified as an optimal catalyst for this enantioselective photobiocatalytic decarboxylative alkylation of α-keto acids and benzaldehydes with redox-active esters, affording a series of chiral α-branched ketones with excellent enantioselectivity (up to 99% ee). Simultaneously, a comparison of the structures and sequences of PfBAL and PpBFD to analyze their differences in reactivity led to a SSM of PfBAL, which identified a variant PfBAL T481L A480L that could also give the same high level of enantioselectivity albeit with lower catalytic activity (Scheme 50).

Scheme 50.

Asymmetric alkylation of benzaldehydes and α-keto acids catalyzed by engineered PfBFD.

2.3.3. Acetolactate Synthase

In 2025, Takashi et al. [76] reported an elegant example of ThDP/FAD-dependent enzyme acetolactate synthase derived from Thermobispora bispora (TbALS) catalyzed radical acylation reaction. This transformation utilizes the combination of cofactor thiamine and flavin within a single enzymatic site to simultaneously produce two distinct radicals, thus enabling highly selective radical–radical coupling. Notably, both wild-type TbALS and its variants TbALS^KYH demonstrated high catalytic efficiency in the acylation reactions of α-bromo carbonyls with α-ketoacids. It should be noted short-lifetime alkyl radicals generated from N-acyloxyphthalimides could also be coupled with acyl radical equivalents generated from pyruvic acid to produce dialkyl ketones using variant TbALS(V104Y/D283Y) as a catalyst under blue LED irradiation. It is worth noting that these reactions are sensitive to the steric hindrance as the yields decrease significantly with bulker substrates (Scheme 51). This study highlighted that ThDP/FAD-dependent enzymes can be utilized as a highly efficient biocatalysts for the fascinating asymmetric radical cross-coupling reactions, particularly those generating transient alkyl radical intermediates.

Scheme 51.

Asymmetric radical acylation of α-bromo carbonyls and N-acyloxyphthalimides catalyzed by engineered TbALS.

2.3.4. Branched-chain 2-keto acid decarboxylase (KdcA)

The development of novel enzyme reactivity driven by electricity is challenging, largely due to the compatibility problems and the difficulty of heterogeneous electron transfer. In 2025, Huang and co-workers[80] reported a dynamic kinetic oxidation of α-branched aldehydes through the integration of thiamine-dependent enzyme KdcA (BAL in few cases) with ferrocene-mediated electrocatalysis, affording a series of chiral α-branched acids with excellent enantioselectivity (up to 99% ee). It should be noted that this method is also applicable to whole-cells catalytic system, with the enzyme loading as low as 0.05 mol% and a turnover number (TON) as high as 1320. Based on the mechanistic studies, a possible mechanism was proposed. Ferrocene methanol (FcMeOH) is oxidized to its ferrocenium form at the anode. This species is capable of oxidizing the Breslow intermediate, generated from the ThDP cofactor and aldehyde, to the plausible radical intermediate. A subsequent single-electron oxidation of this radical intermediate by another molecule of ferrocenium species leads to the generation of an acyl azolium cation, which ultimately undergoes hydrolysis reaction to yield (S)-configured carboxylic acids (Scheme 52).

Scheme 52.

Preparation of chiral α-branched carboxylic acids from α-branched aldehydes through enzymatic dynamic kinetic oxidation driven by electricity.

2.4. Applications

2.4.1. In situ Carboligation Processes

The in situ generation of aldehydes, pyruvates and their derivatives coupled with ThDP-dependent lyase-catalyzed enantioselective formation of C–C bond in a one-pot manner offers a green and efficient method to produce valuable chiral α-hydroxyketones from readily available feedstocks. In 2013, de María and co-workers [137] reported a two-step oxidase-lyase-catalyzed reactions, where oxidase from Hansenula sp. first oxidized short unbranched aliphatic alcohols to aliphatic aldehydes, followed by a carboligation with benzaldehydes, affording the corresponding chiral α-hydroxyketones with excellent enantioselectivity (up to 99% ee) (Scheme 53).

Scheme 53.

Preparation of α-hydroxyketones from alcohols and benzaldehydes by oxidase and BAL.

In 2017, Hollmann, Schmidt and co-workers [138] developed a one-pot, two-step oxidase-lyase system to produce chiral α-hydroxyketones from benzylic alcohols by combining alcohol oxidase from Pichia pastoris (PpAOX) with PfBAL (Scheme 54). In this system, the low solubility of α-hydroxyketones in the reaction medium drives the equilibrium toward the product formation, thus improving both the reaction efficiency and facilitating product purification.

Scheme 54.

Preparation of enantiopure α-hydroxyketones via oxidase-lyase-catalyzed cascades using PpAOX and pfBAL.

In 2017, Gao group [139] reported a multi-enzyme cascade for synthesizing acetoin from lactate. In this method, racemic lactate isolated from corn steep water was first oxidized to pyruvate via L-lactate oxidase (L-LOX) and D-lactate oxidase (D-LOX) catalyzed dehydrogenation process. The resulting pyruvate was then decarboxylated to acetaldehyde by PDC, which was subsequently converted to acetoin. The use of Catalase (CAT) was to remove hydrogen peroxide generated from the dehydrogenation reaction (Scheme 55). Notably, this study represents the first report of converting lactate directly into acetoin, highlighting a promising strategy for valorizing biomass-derived lactate in biocatalytic processes.

Scheme 55.

Synthesis of acetoin from lactate via a multi-enzyme cascade.

Enzymatic catalysis typically exhibits remarkable chemo-, regio- and stereoselectivity, enabling efficient and direct synthesis of complex molecules under mild reaction conditions without the need for protecting group strategies[3]. This advantage renders biocatalysis a powerful tool in polyol and carbohydrate synthesis as sophisticated protective group manipulations are generally required using standard chemical methods[140]. In 2017, Hecquet, Charmantray and co-workers [141] reported a one-pot cascade system for synthesizing highly valuable L-erythro (3S,4S) ketoses. This approach initiated with the generation of HPA intermediate from L-serine catalyzed by a thermostable L-α-transaminase, followed by a TK-catalyzed C–C bond-forming reaction with hydroxylated aldehyde, resulting in the production of enantiopure L-erythro (3S,4S) ketoses with excellent stereoselectivity (> 95% de) (Scheme 56).

Scheme 56.

Enzymatic preparation of L-erythro (3S,4S) ketoses catalyzed by TA_tca and TK_gst.

Subsequently in 2018, Ward et al. [142] reported a one-pot, two-step system that converts L-serine and L-arabinose into L-gluco-heptoluse using transaminase and transketolase. Within this system, β-hydroxypyruvate intermediate was generated from L-serine by transaminase and subsequently participated in a stereoselective C–C bond-forming reaction with L-arabinose catalyzed by transketolases, yielding L-gluco- heptulose in high yield (Scheme 57).

Scheme 57.

Preparation of L-gluco-heptulose promoted by TAm and TK.

In 2019, Hecquet, Charmantray et al. [143] reported the preparation of (3S)-hydroxyketones via a one-pot cascade process, utilizing transketolase catalysis in combination with the D-amino acid oxidase catalyzed in situ generation of hydroxypyruvate from D-serine. By using transketolases (TKs) from diverse sources, a series of aldehydes ranging from aliphatic to hydroxylated derivatives were efficiently converted into the corresponding chiral hydroxyketones or ketoses in good yields and with excellent stereoselectivity (> 95% ee or de) (Scheme 58).

Scheme 58.

DAAO_Rg and TK-catalyzed enzymatic synthesis of (3S)-hydroxyketoses.

In 2020, Hecquet and co-workers [144] developed an elegant example of one-pot, three-step cascade approach for synthesizing ketoses. In this multi-enzyme catalytic system, transketolase efficiently catalyzes carboligation of hydroxypyruvate (HPA), generated in situ from S-serine by TA_tca-mediated transamination reaction, and a chiral hydroxylated aldehyde, formed in situ through D-fructose-6-phosphate aldolase (FSA) catalyzed condensation reaction of glycolaldehyde with formaldehyde. This strategy enabled the efficient preparation of valuable (3S,4S)-ketoses, such as L-psicose, L-ribulose and D-tagatose, with excellent diastereoselectivity (> 95% de). The irreversibility of the transketolase-catalyzed step serves as a thermodynamic driving force, effectively moving the equilibrium of the otherwise reversible transamination and achieving full conversion of all the starting materials. Notably, this study highlights the remarkable compatibility and efficiency of multi-enzyme cascades and provides a useful enzymatic alternative to the traditional chemical methods for producing chiral ketoses (Scheme 59).

Scheme 59.

TA_tca, FSA_eco and TK_gst-catalyzed enzymatic synthesis of (3S,4S)-ketoses.

2.4.2. Synthesis of Chiral 1, 2-Amino Alcohols

The combination of ThDP-dependent enzyme-catalyzed asymmetric benzoin or benzoin-type condensation with stereoselective reductive amination provides an effective strategy to produce optically pure 1,2-amino alcohols, compounds which represent crucial precursors in the development of chiral auxiliaries and ligands [145]. In 2013, Rother et al. [146] developed a two-step enzymatic cascade to synthesize Norephedrine (NE) and Norpseudoephedrine (NPE). In the first step, acetohydroxyacid synthase I (AHAS-I) promoted the carboligation reaction between benzaldehyde and pyruvic acid, delivering (R)-phenylacetyl-carbinol ((R)-PAC) with an optical purity of >99% ee. Subsequently, (R)-PAC was converted to (1R,2S)-NE and (1R,2R)-NPE (>98% de, >99% ee) by employing ω-transaminases exhibiting (S)- and (R)-selectivity, respectively (Scheme 60).

Scheme 60.

Synthesis of NE and NPE using AHAS-I and ω-transaminases.

In 2017, Rother and co-workers [147] reported both chemoenzymatic and purely enzymatic cascade approaches for synthesizing stereochemically complementary tetrahydroisoquinolines (THIQs) without purification of the reaction intermediates. In this enzymatic cascade reactions, EcAHAS-I first catalyzed the carboligation reaction of pyruvate with 3-hydroxybenzaldehyde, yielding (R)-1-hydroxy-1-(3-hydroxyphenyl)propan-2-one in 95% yield with excellent enantioselectivity (99% ee). Subsequent transamination gave the corresponding chiral amino alcohol in 91% yield with 97% ee. The chiral amino alcohol then underwent a Pictet-Spengler cyclization with phenylacetaldehyde, catalyzed by a Pictet-Spenglerase, to afford the final THIQ product. The chemoenzymatic strategy used a similar route except that the final Pictet-Spengler reaction was catalyzed by inorganic phosphate instead of an enzyme. Interestingly, this variation resulted in the formation of a THIQ bearing the opposite configuration at the C1 stereocenter compared to that produced in the fully enzymatic cascade (Scheme 61).

Scheme 61.

Enzymatic and chemoenzymatic three-step cascades to isoquinoline derivatives.

In 2019, Rother et al. [148] reported a highly efficient method for synthesizing four stereoisomers of methoxamine using pyruvate and 2,5-dimethoxy-benzaldehyde. This approach utilizes a combination of a triple variant of pyruvate decarboxylase or EcAHAS-I and a transaminase from Bacillus megaterium, affording all stereoisomers in high overall yields (59–80%) and with excellent stereoselectivity (94–99% ee) (Scheme 62).

Scheme 62.

Divergent synthesis of four stereoisomers of methoxamine.

In 2021, the Lin group reported a one-pot sequential benzoin-type condensation/transamination reaction for the preparation of chiral phenylserinol derivatives from β-hydroxypyruvate and arylaldehydes by utilizing a combination of engineered TK (from E. coli) and transaminase ATA117 (from Arthrobacter sp) [149]. Through structure-based directed-evolution, they successfully converted the enantioselectivity of TK from S to R (93% ee to 95% ee), reversed the enantiopreference of ATA117 (from E(S) = 9 to E(R) = 12), and altered ATA117’s selectivity towards ketone/aldehyde substrates. Using the engineered TK and TA, a variety of (1R,2R)-Phenylserinol derivatives were obtained in high yields (up to 80%) with excellent enantioselectivity (up to 97% de and > 99% ee), offering a convenient method to produce chiral intermediate of Florfenico (Scheme 63).

Scheme 63.

Preparation of (1R,2R)-Phenylserinol derivatives from β-hydroxypyruvate and arylaldehydes.

In 2022, Zhu, Ma and co-workers [150] developed a highly efficient one-pot, two step system to synthesize chiral N-substituted 1,2-amino alcohols from simple aldehydes and amines (Scheme 64). In the first step, the benzaldehyde lyase PaBAL efficiently catalyzed the hydroxymethylation reaction between aldehydes and formaldehyde to generate α-hydroxymethyl ketones. This was followed by an asymmetric reductive amination catalyzed by an imine reductase from Mesorhizobium sp. 1M-11, yielding the corresponding enantiopure 1,2-amino alcohols with excellent enantioselectivity (up to >99% ee).

Scheme 64.

Preparation of N-substituted 1,2-amino alcohols from aldehydes and amines by PaBAL and IREDs.

In 2024, Ferrandi group [151] reported a biocatalytic two-step process for the stereoselective preparation of (1S,2S)-NPE and (1S,2R)-NE from benzaldehyde derivatives and methylacetoin. This approach first utilized the acetoin: dichlorophenolindophenol oxidoreductase (Ao:DCPIP OR) catalyzed carboligation reaction, followed by an asymmetric transamination reaction, affording the final chiral amino alcohols in high overall yields (46–72%) and with excellent stereoselectivity (95–98% ee) (Scheme 65). It is noteworthy that this study established a green and efficient biocatalytic platform to synthesize sterically demanding NE and NPE derivatives from methylacetoin.

Scheme 65.

Stereoselective synthesis of NE and NPE from benzaldehydes and methylacetoin.

2.4.3. Synthesis of Chiral Vicinal Diols

The ThDP-dependent enzyme-catalyzed enantioselective benzoin-type condensation combined with the asymmetric reduction of α-hydroxy ketone intermediates have been extensively used for the production of optically pure 1,2-diols. In 2002, Liese et al. [152] developed an enzymatic cascade approach to access four stereoisomers of 1-phenylpropane-1,2-diol (PPD). This approach first utilized either BAL or BFD promoted carboligation reaction of acetaldehyde and benzaldehyde, followed by stereoselective reduction using alcohol dehydrogenase (ADH). All four chiral vicinal diols products were obtained in high yields with perfect enantioselectivities (Scheme 66).

Scheme 66.

Preparation of enantiopure 1-phenylpropane-1,2-diol via an enzymatic tandem strategy.

In 2013, Giovannini et al. [153] reported a biocatalytic approach for producing chiral vicinal diol (2S,3R)-2,3-dihydroxy-3-methylnonan-4-one, a green tea flavor. This approach initiated with the homo-coupling of 2,3-octanedione catalyzed by acetylacetoin synthase (AAS), producing (3R)-3-hydroxy-3-methylnonane-2,4-dione with an enantiomeric excess of 44%. Then, this intermediate underwent acetylacetoin reductase (AAR)-catalyzed regio- and diastereoselective reduction (de >95%) to give the final chiral diol product in enantiomerically pure form (Scheme 67).

Scheme 67.

Synthesis of (2S,3R)-2,3-dihydroxy-3-methylnonan-4-one by AAS and AAR.

In 2014, Rother et al. [154] utilized a combination of BAL from Pseudomonas fluorescens and ADH from Ralstonia sp. to synthesize (1R,2R)-PPD in a micro-aqueous medium with an excellent yield up to 327 gL⁻¹d⁻¹. This approach enabled the use of high substrate concentration, reducing the reaction time and highlighting environmental benefits of enzymatic catalysis (Scheme 68). In the same year, the authors reported that encapsulation of the lyophilised whole-cell catalysts within teabags for (1R,2R)-PPD production could be used to simplify the isolation of product and catalyst recovery from the reaction mixture [155].

Scheme 68.

Synthesis of (1R,2R)-PPD stereoisomers from benzaldehyde and acetaldehyde by BAL and ADH.

In 2018, Rother group [156] also reported a highly stereoselective method to prepare all four enantiopure 4-methoxypheny-l, 2-hydroxy-propanone isomers using a combination of PfBAL or PpBFD L461A with ADH (Scheme 69). Notably, 4-methoxybenzyl alcohol could serve as co-substrate in the second step, where it was oxidized to 4-methoxybenzaldehyde and simultaneously regenerated NADPH, thereby enabling a self-sufficient cascade process that improved both atom economy and overall process efficiency. Later, in 2019, Krauss group also reported a self-sufficient cascade for the preparation of (1R,2R)-PPD using benzyl alcohol as co-substrate [157].

Scheme 69.

Synthesis of all stereoisomers of 4-methoxypheny-l,2-hydroxy-propanone by tandem enzymatic cascade.

In the same year, Pohl et al. [158] reported a highly efficient method to produce (1S,2S)-PPD from benzoylformate and acetaldehyde using HaloTag^™ fusion enzymes BFD L476Q and ADH in a continuous processes. Under the optimized conditions, this process efficiently delivered (1S,2S)-1-phenylpropane-1,2-diol with excellent enantioselectivity (96% ee) and a high space–time yield (1850 g L⁻¹ d⁻¹) (Scheme 70).

Scheme 70.

Preparation of (1S,2S)-PPD from benzoylformate and acetaldehyde by BFD L467Q and ADH.

Furthermore, in 2022, Rother group [159] reported a two-step, one-pot cascade to synthesize aliphatic (4S,5S)-octanediol from butanal catalyzed by a combination of a PDC variant (ApPDC E469G) and a butanediol dehydrogenase from Bacillus licheniformis (BlBDH). In this approach, ApPDC E469G efficiently catalyzed enantioselective carboligation of two butanal molecules to form (S)-butyroin with high (S)-selectivity. Subsequently, BlBDH reduced (S)-butyroin to afford aliphatic (4S,5S)-octanediol in good yield and with perfect stereoselectivity (Scheme 71). It should be noted this study represents the first enzymatic synthesis of (4S,5S)-octanediol, an important aliphatic diol that previously was accessible only through osmium-catalyzed asymmetric dihydroxylation [160].

Scheme 71.

Synthesis of (4S,5S)-octanediol from butanal by ApPDC E469G and BlBDH.

2.4.4. Formose Reaction

The conversion of formaldehyde (FALD) to dihydroxyacetone (DHA), commonly known as formose reaction, is well established in organic chemistry[161]. However, enzymatic catalysis of this process remains challenging due to formaldehyde’s toxicity to biological cells and its potential to disrupt cellular structures. Inspired by earlier findings that thiazolium salts could catalyze the formose reaction [162], Baker group [68] reported a computationally-designed enzyme capable of converting three FALD molecules into DHA (Scheme 72). This enzyme, named as Formolase (FLS), was obtained through the design and modification of BAL derived from Pseudomonas fluorescens biovar I, where its activity increased by 100-fold to catalyze two consecutive carboligation reactions, thus effectively converting FALD to DHA. Notably, this study reported the first enzyme that directly catalyzes the conversion of FALD into DHA.

Scheme 72.

The first enzymatic formose reaction catalyzed by a BAL variant.

In the same year, Siegel group [163] conducted a mechanistic study of FLS, observing that glycolaldehyde (GALD) is produced as a by-product in the formose reaction, which contrasted with the previous findings that GALD is not observed in the non-enzymatic synthesis using thiazolium salts in organic solvents. Moreover, the product distribution was found to be dependent on the concentration of FALD with high concentrations favoring the formation of dihydroxyacetone (DHA), while lower concentrations led to the predominant formation of GALD. This study highlighted potential directions for future research in precise engineering the formose pathway to improve multi-carbon assimilation pathway.

Efficient utilization of C1 carbon sources for the synthesis of highly-valued molecules has attracted intense interest from both academic and industrial communities owing to the profound significance of this process for sustainable development, environmental protection and energy-related applications. Recently, several impressive examples that employed FLS for C1 fixation in the preparation of erythrulose, L-lactic acid, furanose 6-phosphate, L-alanine and starch have been reported. In 2021, Sieber et al. [164] successfully obtained two high-production variants FLS_A3 (H29D/Q113C) and FLS_B2 (H29I/Q113S) from ‘wild type’ FLS [68]. Using these variants, the yield of erythrulose (ERY) from glycolaldehyde reached up to 98%. A maximum conversion efficiency of 74% was obtained when employing a combinatorial system of FLS_A3 and FLS_B2, with an ERY selectivity of 33%, attesting the suitability of this process for industrial applications. Moreover, an elegant artificial starch anabolic pathway (ASAP) was designed by Ma and co-workers for starch synthesis by utilizing methanol condensation in a cell-free system (Scheme 73) [70]. In the C3a module of ASAP, the formaldehyde-utilizing enzyme FLS-M3 (FLS, formolase from Pseudomonas putida) was obtained through directed evolution strategy, enhancing the enzyme’s catalytic activity with the K_cat value reaching 0.2375 s⁻¹. FLS-M3 was also introduced to other C1 utilization pathways (e.g., methanol to alanine pathway (MAP) [165], artificial CO₂-to-sugars pathways (ACSP) [166], and methanol to lactic acid pathways (MLPs) [167]) to synthesize L-alanine with 90% yield, fructose-6-phosphate (F6P) with diverse stereoisomers, and lactate with excellent stereoselectivity (>99% ee) (Scheme 65).

Scheme 73.

Application of FLS in the fixation of C1 carbon sources to produce starch, L-Alanine, F6P and L-lactic acid.

In addition to enzymes engineered from BAL, variants of BFD have also been extensively investigated in the formose reaction. In 2019, Jiang et al. designed a synthetic metabolic route termed the SACA (Synthetic Acetyl-CoA) pathway [168], which utilized formaldehyde to synthesize acetyl-CoA, a key metabolite in central metabolic pathways. In this study, they obtained a glycolaldehyde synthase (GALS), engineered from ThDP-dependent enzyme PpBFD, which could achieve a glycolaldehyde yield of 1.8 g/L within two hours when starting from 2 g L⁻¹ formaldehyde. Subsequently, acetyl-phosphate synthase (ACPS) catalyzed the transformation of glycolaldehyde into acetyl-phosphate (AcP). The AcP intermediate was then converted to acetyl-CoA catalyzed by phosphate acetyltransferase (PTA) (Scheme 74).

Scheme 74.

Carboligation of formaldehyde by BFD to synthesize acetyl-CoA.

In 2019, Ma group [169] reported a two-step chemoenzymatic approach for converting formaldehyde into lactic acid. In this strategy, BFD M6, engineered from BFD M3 [168], first catalyzed the conversion of formaldehyde to DHA via the formose reaction, followed by a NaOH-promoted rearrangement that produced lactate with an overall yield of 82.9%. Compared to BFD M3, BFD M6 exhibited enhanced catalytic activity and improved robustness (Scheme 75).

Scheme 75.

BFD-catalyzed carboligation of formaldehyde to produce lactate.

In 2022, Ma group further reported a streamlined two-step chemoenzymatic process for producing glycolic acid from formaldehyde. Initially, BFD M4V2 catalyzed the formose reaction to generate a mixture of GALD and DHA (92.88% conversion), which were subsequently oxidized to glycolic acid (89.87% yield) (Scheme 76) [170].

Scheme 76.

Carboligation of formaldehyde by BFD to produce glycolic acid.

In a recent study, Ma, Zhu and co-workers analyzed the structure of the BFD M3 and found that the modification of α21 and α17 significantly improved its catalytic activity at low formaldehyde concentration [171]. To improve the catalytic activity of FLS and its preference for DHA production, Wang group employed scanning saturation mutagenesis to explore key residues around the BFD active site [172], leading to the identification of a new variant FLS S26F with higher DHA production and enhanced resistance to formaldehyde.

Inspired by the ability of 2-hydroxyacyl-CoA lyase (HACL) to catalyze the conversion of 2-hydroxyacyl-CoA into aldehyde and formyl-CoA, Gonzalez group utilized this enzyme to condense formyl-CoA and formaldehyde for producing glycolate, developing a new method of C1 bioconversion (Scheme 77A) [173]. Later, the same group [174] also confirmed the feasibility and orthogonality of this reaction through thermodynamic and stoichiometric analysis, providing a platform for the subsequent production of multi-carbon compounds from formaldehyde. In 2019, Erb, Burgener and co-workers [175] found that oxalyl-CoA decarboxylase (OXC) from Methylorubrum extorquens (OXC_Me) can efficiently catalyze condensation between different aldehydes and formyl-CoA for C1-extension. Based on this finding, a quadruple variant MeOXC4 named as glycolyl-CoA synthase (GCS) was obtained through iterative saturation mutagenesis of OXC_Me (Scheme 77A) [176]. MeOXC4 could efficiently catalyze the synthesis of glycolyl-CoA from formaldehyde and formyl-CoA, achieving a 200-fold activity increase compared to the native OXC enzyme. In 2022, Park group reported that a variant of glyoxylate carboligase from Escherichia coli (EcGCL R484M N283Q L478M) could be used to produce GALD with a rate of 4.3 mM h⁻¹ (Scheme 77B) [177], and the resulting GALD was then reduced to ethylene glycol using a lactaldehyde reductase (FucO) with a 66% conversion. Later, the same group [178] also reported that EcGCL could condense two formaldehydes for the production of GALD with a conversion of 83%, while an engineered variant of the E1 subunit from Vibrio vulnificus α-ketoglutarate dehydrogenase complex (VvSucA) could transform two GALD into C4-erythrulose (Scheme 77C).

Scheme 77.

Application of HACL, GCS and GCL in bioconversion of formaldehyde.

3. Metalloenzymes

3.1. Artificial Metalloenzymes for Hydroamination Reactions

While enzymatic organic synthesis involving NHC catalysis has achieved considerable progress in asymmetric synthesis for constructing C–C bonds, the integration of NHC–metal complexes into protein scaffolds offers a promising strategy to expand the scope of abiotic bond forming reactions. As exemplified by the impressive work of Ward et al. [179] in 2021, a biotin-tagged NHC–Au(I) complexes embedded in hybrid streptavidin scaffold capped with a SOD lid (artificial metalloenzymes) facilitates the regioselective hydroamination of terminal alkynes (Scheme 78). Two mutants with distinct regioselectivity were identified through three rounds of directed evolution. The Sav-SOD mutant S112N T114S T115N N118G K121G S122G favored the formation of anti-Markovnikov indole products via gold π-activation of the alkyne, while the S112T T115A N118S K121F S122Gb mutant led to the production of Markovnikov quinazolinone products through a dual σ-/π-activation mechanism, where both reactions achieved excellent regioselectivity. This study illustrated how NHC–metal complexes could be deployed to combine with evolvable protein scaffolds to achieve enzyme-like control over abiotic transformations, thus expanding the toolbox for regioselective C–N bond formation in aqueous environments.

Scheme 78.

Highly regioselective hydroamination catalyzed by artificial metalloenzymes.

3.2. Gold(I)-Carbene Catalysis with ThDP Analogs

ThDP-dependent enzymes have the distinct capacity to generate a carbene intermediate within their active site, offering the potential to form metal–carbene complexes upon coordination to a transition metal. Due to challenges associated with direct metalation of ThDP, thus far researchers have turned to synthesizing gold–carbene catalysts based on thiamine analogs. In 2017, Hilvert and coworkers [180] developed a single-step, aqueous-based method to synthesize gold(I)–carbene catalysts with ThDP analogs. The resulting gold(I)–carbene 18 could efficiently catalyze 5-endo-dig cyclization of alkenes and hydroalkoxylation of allenes, affording the corresponding cyclopentene derivatives (63% yield) and 2-(2-methylpropen-1-yl)tetrahydrofuran (90% yield). While gold(I)–carbene catalyst 18 exhibited no catalytic activity in aqueous buffer solution, its analog, pyrophosphate gold(I)-carbene 19, showed excellent performance in water, enabling the hydroalkoxylation of allenes in 99% yield (Scheme 79). Given that the crucial role of the pyrophosphate moiety in anchoring thiamine diphosphate to thiamine-dependent enzymes through supramolecular interactions [181–183] and the high catalytic activity of pyrophosphate gold(I)–carbene observed in this study, the assembly of pyrophosphate gold(I)–carbenes with protein scaffolds provides a promising strategy for the development of gold-enabled biocatalysis.

Scheme 79.

Gold(I)-carbenes with thiamin analogs and their application in carbocyclization and hydroalkoxylation reactions.

Furthermore, in 2023, Ward group [184] designed PTdP, a ThDP-mimetic with a pyridine ring replacing the pyrimidine ring, and complexed it with gold(I) to form AuPTdP (Scheme 80). To test the catalytic activity of AuPTdP, the hydroamination of 2-ethynylaniline was selected as a model reaction. Experiments conducted in the presence of SeBFD and TaTK or their alanine variants revealed that both proteins partially inhibited the activity of AuPTdP. Native mass spectrometry (MS), inhibition assays, and size-exclusion chromatography (SEC) further provided compelling evidence that AuPTdP and likely other gold-based complexes bind to SeBFD in a nonspecific yet reversible manner, resulting in reduced catalytic activity. These findings reveal that ThDP-dependent enzymes intrinsically exhibit a high, nonspecific affinity toward gold(I) complexes, thereby posing an intriguing question about the active-site-specific binding.

Scheme 80.

Synthesis of Au(I)–NHC complex containing a ThDP mimic.

4. Conclusions

The examples discussed in this review demonstrate that NHC catalysis originally inspired by the biochemical mechanisms of ThDP-dependent enzymes has in turn promoted the development of enzymatic synthesis involving NHC catalysis. During the past two decades, a wide range of ThDP-dependent enzymes have been successfully employed to catalyze benzoin-like condensations and Stetter-type reactions via Breslow intermediate, proving a robust and practical tool for the preparation of diversified chiral α-hydroxy ketones, 1,4-diketones and derivatives in excellent yields and often with perfect enantioselectivity. ThDP-dependent enzymes have also been engineered to catalyze synthetically challenging reactions, including asymmetric cross-benzoin condensations, Stetter-type reactions, and asymmetric radical transformations. Furthermore, the development of artificial metalloenzymes incorporating NHC–Au(I) complexes has enabled highly regioselective hydroamination of alkynes, opening new avenues for abiotic C–N bond-forming reactions within enzymatic scaffolds.

Despite considerable progress has been made in this important area, a key limitation remains. In contrast to the vast range of transformations accessible through NHC catalysis in small-molecule organic synthesis, the types of reactions catalyzed by ThDP-dependent enzymes are still limited, primarily focusing on benzoin-type condensation reactions. Inspired by the achievements made in dual NHC/transition-metal catalysis and NHC/photoredox catalysis in organic synthesis, we envision that, with the aid of protein engineering and high throughput screening techniques, integrating ThDP-dependent enzymes with transition-metal catalysts and/or photocatalysts will unlock new-to-nature reactivities and greatly expand the synthetic potential of this powerful biocatalytic platform.