Stereodivergent C‐(Hetero)Aryl Glycosylation by Nickel‐Catalyzed Redox‐Neutral Cross‐Coupling of Glycosyl Sulfonyl Hydrazides

ABSTRACT

Unprotected C‐aryl glycosides are ubiquitous in biologically active compounds and are widely used in chemical biology. The stereochemistry at the anomeric carbon of a C‐glycoside often dictates its function. Despite past advances in C‐glycosylation, methods that provide selective access to both α and β anomers are scarce due to the challenge of controlling stereoselectivity. Herein, we demonstrate that native sugars, when transformed into glycosyl sulfonyl hydrazide precursors, undergo efficient radical cross‐coupling with (hetero)aryl halides under redox‐neutral nickel catalysis. The method has broad scope and excellent functional group tolerance, enabling the stereodivergent synthesis of diverse C‐(hetero)aryl glycosides in either α or β anomeric forms through ligand control.

A nickel‐catalyzed C–C cross‐coupling protocol employing bench‐stable glycosyl sulfonyl hydrazides as practical glycosyl radical precursors is reported. A variety of (hetero)aryl iodides underwent cross‐coupling under redox‐neutral conditions to afford structurally diverse unprotected C‐(hetero)aryl glycosides. This stereodivergent strategy offers a powerful approach to access both anomers of C‐aryl glycosides.

1. Introduction

C‐aryl glycosides are a vital class of carbohydrate derivatives distinguished by an aryl substituent attached to the anomeric carbon, finding significant utility in pharmaceutical development and biochemical studies [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]. These compounds are often studied as O‐ or N‐glycoside mimetics in drug discovery by leveraging their exceptional resistance to enzymatic degradation and their ability to engage in productive interactions with target proteins to achieve potent and selective activity [12, 13, 14]. The stereoselective construction of these glycosides continues to pose a significant challenge, particularly in the context of unprotected C‐aryl glycosides bearing numerous reactive hydroxyl groups [15, 16, 17]. Representative examples of α‐anomeric C‐aryl glycosides include C‐mannosyl tryptophan, which functions as a post‐translational protein modification in key biological pathways, as well as antivirulence C‐mannosides, which exhibit antiviral activity [18, 19, 20]. On the other hand, C‐aryl glycosides featuring β‐anomeric motifs have been integrated into the core scaffolds of SGLT2 inhibitors against type II diabetes, as exemplified by dapagliflozin and empagliflozin (Scheme 1a) [21, 22, 23]. In the design of sugar‐based therapeutics, the stability of C‐aryl glycosidic bonds against hydrolytic enzymes in vivo has played a major factor in their successful deployment as robust and potent surrogates of native O‐glycosides [12].

SCHEME 1.

Minimally protected glycosyl sulfonyl hydrazides as precursors for stereodivergent radical C‐aryl glycosylation under redox‐neutral Ni catalysis.

Driven by their remarkable structural diversity and biological relevance, C‐aryl glycosides have become the focus of extensive methodology development, leading to a plethora of synthetic approaches [24, 25]. Traditional strategies for making C‐aryl glycosides typically relied on Fries rearrangements, Friedel‐Crafts reactions, or the addition of arylmetal species to glycosyl electrophiles via nucleophilic or electrophilic pathways [26]. In recent years, glycosyl radical coupling has gained widespread attention as a prominent strategy for synthesizing C‐aryl glycosides [27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37]. Notwithstanding the developments in radical C‐aryl glycosylation, most reported methods could only provide access to one stereoisomer of the product. A general reaction platform that could promote the selective formation of both α and β anomers, especially in unprotected form, from a single glycosyl donor remains largely elusive. Of note, our group previously reported a stereoselective C‐aryl glycosylation protocol in which bench‐stable heteroaryl glycosyl sulfones served as effective precursors for radical C–C cross‐coupling with nucleophilic arylzinc reagents or electrophilic aryl halides under redox‐neutral Fe or reductive Ni catalysis, respectively (Scheme 1b) [38]. However, this method only offers stereodivergent access to α and β isomers of C‐aryl glycosides for a narrow range of protected monosaccharides. Furthermore, the requirement of two distinct transition metal catalytic systems coupled with the laborious multi‐step preparation of glycosyl sulfone donors limited the method's practical utility in carbohydrate chemistry.

Therefore, there is an urgent demand to develop transformations that enable stereodivergent C‐aryl glycosylation that works for a wide variety of sugar residues and aryl moieties with broad functional group compatibility. Of particular interest is the establishment of a versatile protocol using nonprecious base metal catalysis [39, 40] to convert readily accessible glycosyl precursors into C‐aryl glycoside products with precise control over the stereochemistry at the newly formed C–aryl bond. Recently, alkyl sulfonyl hydrazides have emerged as redox‐neutral alkyl radical precursors amenable to cross‐coupling under operationally simple Suzuki‐like conditions [41, 42, 43]. These reactions feature an inexpensive nickel catalyst and a mild base, circumventing the need for pyrophoric organometallics or exogenous redox mediators. On the other hand, glycosyl sulfonyl hydrazides were employed as bench‐stable intermediates for making glycosyl 1‐phosphates via a non‐radical polar mechanism more than a decade ago [44]. Based on these seminal discoveries, we envisioned that glycosyl sulfonyl hydrazides could serve as attractive redox‐active radical precursors to undergo radical cross‐coupling with appropriate aryl‐substituted electrophiles under redox‐neutral Ni catalysis. Herein, we describe the successful development of a single Ni‐catalyzed manifold that harnesses glycosyl sulfonyl hydrazides as radical precursors for C–C cross‐coupling with aryl halides [45, 46] to deliver both α and β isomers of C‐aryl glycosides (Scheme 1c).

Initial studies with fully unprotected glycosyl sulfonyl hydrazides were unsuccesful due to poor conversion. This led us to investigate the cross‐coupling of minimally protected glucosyl sulfonyl hydrazide 1a (isolated in pure β anomeric form) and commercially available methyl 4‐iodobenzoate 2a (Table 1). Upon completion of the reaction, Et₃N·HF was directly added, and the crude mixture was allowed to stir at room temperature overnight to promote desilylation [47]. After extensive evaluation of various reaction parameters (see Supporting Information (SI), Section 3 for details), we found that stereoselective α‐C‐aryl glycosylation was achieved in the presence of 20 mol% of NiCl₂ ^.DME/L1 and 1.5 equivalents (equiv.) of 2a, PMP as base and MeCN as solvent at 60°C, furnishing α‐C‐aryl glucoside 3a as the major product (α:β ratio = 6:1) in 56% NMR yield within 24 h (entry 1). Only trace amounts of product were detected in the absence of the catalyst, ligand or base (entry 2). Switching the catalyst to Ni complex A improved the reaction, affording 3a in 66% yield and 7:1 α:β selectivity (entry 3). Replacing NiCl₂ ^.DME/L1 with Ni complex B led to slightly diminished yields and selectivity (entry 4). Intriguingly, β‐C‐aryl glucoside 4a was observed as the major product in 55% yield and 8:1 β:α ratio when the sterically more hindered terpyridine ligand L2 was used (entry 5). Furthermore, when bipyridine ligand L1 was replaced with L3 using TMG as base and DMF as solvent, 4a was obtained in 60% yield with β:α selectivity of 10:1 (entry 6). Lowering the amount of base to 3 equiv. or the catalyst loading to 10 mol % did not improve results (entries 7 and 8). Performing the reaction at 40°C or 80°C gave lower yields of 3a without significantly compromising stereoselectivity (entries 9 and 10). Replacing PMP with DBU negatively impacted C‐glycosylation (entry 11). Likewise, 3a was generated in lower yield when the solvent was changed to DMF (entry 12). Contrary to the stereoretentive reactions of enantioenriched (non‐sugar) alkyl sulfonyl hydrazides [41, 42, 43], our studies revealed that stereodivergent cross‐coupling of glycosyl sulfonyl hydrazides to access either α or β anomer could be accomplished by harnessing ligand control.

TABLE 1.

Reaction optimization.

Entry	Conditions	Yield (%) a	α:β b
1	None	56	6:1
2	Without Ni(II) or L1 or PMP	<5	ND
3	Complex A	66	7:1
4	Complex B	45	5:1
5	L2 instead of L1	55	1:8
6 c	L3 instead of L1	60	1:10
7	3 equiv of PMP	56	7:1
8	10 mol% of Complex A	38	6:1
9	40°C	46	5:1
10	80°C	56	6:1
11	DBU instead of PMP	<5	ND
12	DMF as solvent	48	5:1

Abbreviations: bpy, bipyridine; DBU, 1,8‐diazabicyclo[5.4.0]undec‐7‐ene; DME, dimethoxyethane; DMF, N,N‐dimethylformamide; dtbbpy, 4,4′‐di‐tert‐butyl‐2,2′‐bipyridine; ND, not detected; PMP, 1,2,2,6,6‐pentamethylpiperidine; terpy, terpyridine; TMG, 1,1,3,3‐tetramethylguanidine.

^a Yields (major anomer) were determined by ¹H NMR analysis using dibromomethane as an internal standard.

^b α:β ratios were determined by crude ¹H NMR analysis.

^c TMG (3.0 equiv.) was used as the base, and DMF was used as the solvent. Upon reaction completion, Et₃N^.HF was added, and the mixture was allowed to stir at room temperature overnight.

Next, the generality of the established α‐C‐aryl glycosylation protocol was assessed across a vast spectrum of native monosaccharides that could be reliably transformed into fully unprotected α‐C‐aryl glycosides via their minimally protected sulfonyl hydrazide precursors, which were either isolated or generated in situ and used (without purification) for cross‐coupling with aryl iodides (Table 2). Under the optimized conditions, glycosyl sulfonyl hydrazides derived from glucose (3a), galactose (3b), mannose (3c), xylose (3d) and rhamnose (3e) participated in arylation with 2a to furnish the desired α‐C‐aryl glycosides in good yields (45%–62% yield) and high stereoselectivity across the board. To showcase practicality, unprotected α‐C‐aryl glucoside (3a) and mannoside (3c) were obtained directly from their native sugars via in situ C1‐hydrazidation and C4,C6‐silylation followed by cross‐coupling/desilylation (single purification), significantly streamlining the synthetic process. Rare sugars such as allose (3g) and the non‐natural L‐glucose (3f) similarly underwent reaction under the standard conditions, affording the desired C‐aryl glycoside products. Importantly, disaccharide derivatives such as lactose could also be converted to the corresponding C‐aryl glycoside product (3h) under the standard reaction conditions. Substituents on the aromatic ring bearing different functionalities at the para position were also investigated. Halides (3i, 3j), electron‐withdrawing groups (3k, 3l) and electron‐donating groups (3m) were all tolerated in our Ni‐catalyzed cross‐coupling system, furnishing α‐C‐aryl glycosides 3i–3m in up to 85% yield. Polysubstituted groups such as 3,5‐bis(trifluoromethyl)phenyl (3n) as well as sterically hindered arenes including biphenyl (3o) and naphthyl (3p) could also be incorporated, delivering the corresponding products in good to excellent yields (76%–85% yield). To further test the functional group compatibility of the method, we subjected heteroaryl iodides to the standard conditions and successfully secured products 3q−3t in 42%−85% yield. These examples highlight the excellent chemoselectivity and functional group tolerance of the glycosylation transformation.

TABLE 2.

α‐Selective cross‐coupling to access α‐C‐(hetero)aryl glycosides.

Note: Yields are for α anomeric products. See the Supporting Information for details.

The scope of β‐C‐aryl‐glycosylation under the established conditions was evaluated by examining different functionalized (hetero)aryl iodides as well as sulfonyl hydrazide precursors prepared from various native sugar residues (Table 3). Similar to the cases in Table 2, a range of glycosyl sulfonyl hydrazides underwent efficient C–C cross‐coupling with 2a, providing the corresponding β‐C‐aryl glycosides 4a–4h in 31%–56% yield. Importantly, our protocol enabled access to β‑C‑aryl mannoside and rhamnoside (4c and 4e), which have been recognized as challenging targets in previous reports [29, 36]. Varying the electronic and/or steric effects of the aryl substituent did not have an appreciable impact on reaction efficiency and stereoselectivity, furnishing β‐C‐aryl glycosides 4i–4p in 64%–86% yield. Iodides containing heterocyclic rings, including thiophene, pyridine, dibenzothiophene, and 1,3‐benzodioxole, also took part in arylation, delivering 4q–4t in 30%–73% yield.

TABLE 3.

β‐Selective cross‐coupling to access β‐C‐(hetero)aryl glycosides ^a .

Note: See the Supporting Information for details. Yields are for β anomeric products.

^a NiCl₂·DME/L2 (20 mol%), PMP (5.0 equiv.), CH₃CN, 60°C, 24 h.

^b NiCl₂·DME/L3 (20 mol%), TMG (3.0 equiv.), DMF, 60°C, 24 h.

Control experiments supported the intermediacy of radicals in the Ni‐catalyzed transformations (Scheme 2). When 1,1‐diphenylethylene was employed as the radical acceptor under the standard α‐C‐aryl glycosylation conditions, two competitive reaction pathways were observed (Scheme 2a). The desired Giese adduct 3x was obtained in 41% yield, together with the elimination byproduct 3w in 50% yield (both as α anomers). Meanwhile, the yield of the cross‐coupling product 3u decreased to 28%, accompanied by the glycal byproduct 3v in 30% yield. Further evidence for the involvement of radical intermediates was obtained by performing the reaction in the presence of 1.0 equiv. of the radical scavenger 2,2,6,6‐tetramethylpiperidin‐1‐oxyl (TEMPO). Upon addition of TEMPO (1.5 equiv.) under standard conditions, the cross‐coupling reaction was completely inhibited. Instead, the glycosyl TEMPO adduct 3y arising from the reaction between the in situ‐generated glycosyl radical intermediate and TEMPO, was isolated in 74% yield. Notably, when 1,1‐diphenylethylene was used as the radical acceptor under the standard β‐C‐aryl glycosylation conditions, no stereoretentive Giese adduct was formed (Scheme 2b). Instead, the α‐anomeric 3x was generated in only 26% yield, along with 11% of 3w. These findings suggest that the cross‐coupling process is not stereoretentive under the established conditions, which differs from results reported in previous literature [41]. Under the standard β‐C‐aryl glycosylation conditions, addition of 1.0 equiv. of TEMPO completely inhibited the cross‐coupling reaction, and 3y was detected as expected in 38% yield.

SCHEME 2.

Mechanistic studies.

Even though a full mechanistic analysis of the origin of stereochemical control is beyond the scope of this work, we reasoned that a combination of stereoelectronic and steric effects likely play a role in controlling the α:β stereoselectivity observed in our reactions (see Supporting Information, Section 7 for details). In the α‐C‐aryl glycosylation pathway under redox‐neutral Ni‐catalyzed conditions, we postulated that an in situ‐generated aryl–Ni(bipyridine) species reacts faster from the α‐face of the putative glycosyl radical intermediate (kinetic anomeric effect) due to stabilizing orbital interactions between the endocyclic oxygen and the incipient C1–Ni bond in the transition state [48, 49, 50], which likely outweigh any steric clash with the neighboring C2‐hydroxyl group and the axial C1–Ni(bipyridine) bond. This mechanistic rationale is fully consistent with related investigations on metal‐catalyzed radical α‐C‐aryl glycosylation [28]. The ensuing stereoretentive reductive elimination affords the product in good α selectivity. On the other hand, when the bulky tridentate terpyridine ligand is used in the β C‐aryl glycosylation pathway, the relatively short C1–Ni bond [51] induces severe steric repulsions with the C2‐hydroxyl unit, forcing the aryl–Ni(terpyridine) intermediate to react preferentially from the sterically more accessible β‐face of the glycosyl radical to form the β C‐aryl glycoside product, in congruence with earlier observations [29, 38]. However, for mannose and rhamnose substrates, the reason for formation of the β anomer is not clear at this stage. Detailed studies are ongoing to fully investigate these findings. Overall, the stereochemical outcomes observed in Tables 2 and 3 could be rationalized by the above analysis. In‐depth studies are underway to unravel more insights into the complex steroselectivity profiles of different sugar residues.

The synthetic utility of this protocol is further highlighted by the concise syntheses of marine natural product neopetrosin C 3z [52] and antidiabetic agent empagliflozin 4v (Scheme 3). Specifically, starting from commercially available mannose, the in situ‐generated mannosyl sulfonyl hydrazide 1c was directly coupled with methyl 2‐iodo‐1H‐indole‐3‐carboxylate under the optimized α‐selective C‐aryl glycosylation conditions, delivering 3z in 49% yield after in situ desilylation. Attempts to obtain the β anomer of 3z under standard β‐selective C‐aryl glycosylation conditions only led to trace product. On the other hand, glucose was activated to form glucosyl sulfonyl hydrazide 1a and subjected to reaction with the appropriate aryl iodide under β‐selective and α‐selective C‐aryl glycosylation conditions followed by deprotection, furnishing 4v and its α anomer 3za in 48% and 75% yields, respectively.

SCHEME 3.

Synthetic applications.

In summary, we have shown that bench‐stable glycosyl sulfonyl hydrazides serve as efficient glycosyl radical precursors for C–C cross‐coupling with (hetero)aryl iodides under redox‐neutral Ni‐catalyzed conditions. This stereodivergent method enables access to a wide variety of unprotected C‐(hetero)aryl glycosides through in situ silylation and post‐coupling deprotection. Isolable sulfonyl hydrazides derived from a broad range of sugar residues are amenable to the transformation, affording both α and β anomers of C‐aryl glycoside products with good stereochemical control. We expect this new catalytic manifold to find extensive utility in carbohydrate research for the stereoselective assembly of complex glycan derivatives.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: anie73356‐sup‐0001‐SuppMat.pdf.

Data Availability Statement

The data that supports the findings of this study are available in the supplementary material of this article.