Sequential Regiodivergent Polyol Sulfonylation and Functionalization Enable Precise Engineering of Carbohydrates

Abstract

Substituent functionalization of unprotected and partially protected carbohydrates with controlled regioselectivity remains challenging due to the difficulty in differentiating hydroxyl groups with similar reactivities. This study presents an efficient protocol for site-specific modification through a “two-stage,” regiodivergent polyol tagging and functionalization strategy. To achieve effective tagging, we developed two complementary Ag₂CO₃-ligand-based regimes that enable the regioselective sulfonylation of cis-diol and trans-diol in carbohydrates, controlled by simply toggling the presence of a [Pd] catalyst. Competition experiments and DFT simulations elucidated the underlying dual mechanisms accounting for the regioselectivity. [Pd] catalyst complexes to cis-diol as a bidentate ligand, enhancing the differentiated electrophilicities through stereoelectronic effects and preferentially activating the equatorial C3-OH groups. Conversely, without [Pd], the Ag­(I) complex switches the reaction position, directing sulfonylation to the axial hydroxyl within 1,2-cis-diol, a position that is typically kinetically inert under conventional conditions. And the Ag­(I) complex preferentially coordinates to cis-1,2-substituents on the sugar ring and selectively activates the C2-OH group. The sulfonylated products serve as versatile synthons for the following structural derivations and chemical glycosylations, facilitating efficient access to structurally unique rare sugars, deoxy- and aminosugar analogues, and complex oligosaccharides. This dual-catalytic approach provides a robust platform for precision carbohydrate engineering, advancing the synthesis of biologically relevant oligosaccharides and glycoconjugates.

Introduction

Precision synthetic strategies are increasingly important in modern drug discovery, enabling desired chemical transformations with both specificity and efficiency. These strategies are particularly crucial in carbohydrate chemistry, which bridges the fields of chemical glycobiology and glycopharmaceutics by providing essential molecular tools and drug candidates. However, achieving such precision remains a significant challenge, especially given the subtle differences between chemically similar and spatially proximate hydroxyl groups on native sugars (Scheme a). This complexity complicates the direct skeletal modification of sugar rings, often necessitating tedious and labor-intensive protecting group manipulations. Therefore, the development of universal and efficient protocols for direct precision modification of carbohydrates, bypassing the need for protecting groups, is highly desirable to streamline the synthesis of the advanced glycostructures.

1. Synthetic Challenges and Summary of Our Work .

^a (a) Challenge: differentiating an alcohol in carbohydrates. (b) Two-stage modification strategy. (c) Site-switchable functionalizations with electrophiles assisted by altering chiral catalysts or additives. (d) Dual-mode catalytic system enabling sequential regiodivergent sulfonylation and further adaptive modification.

The past decade has witnessed significant advancements in synthetic strategies aimed at achieving regioselective functionalization of (partially) unprotected carbohydrates, addressing the intricate reactivities of hydroxyl groups to enable straightforward carbohydrate engineering. − While photoredox-mediated radical reactions of native sugars have offered an alternative for sugar modifications, enabling epimerization, oxidation, and alkylation, the emerging “two-stage” functionalization (Scheme b) of native saccharides has provided an efficient, adaptive, and convenient strategy for versatile sugar modifications. − This approach allows for the facile preparation of valuable and uncommon sugar derivatives. In this context, Chi and co-workers developed an NHC-catalyzed regioselective acylation of sugars with 1,4-dihydropyridine (DHP)-based acid as the acylating agent, followed by its C–O bond activation using photoredox chemistry to realize site-specific functionalization. , Zhang and co-workers employed a photoactive 4-tetrafluoropyridinylthio group to achieve unique manipulation of sugar skeletons by the two-stage modification protocol. However, these radical-based transformations have potential limitations, including a narrow scope of transformations and inconsistency with structural characteristics commonly found in native rare sugars. Additionally, radical reactions on sugar may result in the loss of stereochemical information due to the planar-conformed nature of radicals, complicating product separation and target-oriented carbohydrate synthesis.

It is envisioned that the two-stage modification strategy is primarily hampered by challenges of regioselective tagging of carbohydrates. , Although pioneering works by Miller, Dong, Tang, Niu, and others have established catalyst-, ligand-, or additive-controlled site-switchable functionalization of carbohydrates, particularly the equatorial hydroxyls of cis-diols due to their increased nucleophilicity (Scheme c), these efforts remain case-sensitive. ,,− The careful tailoring of catalysts and ligands is essential to match the subtle reactivities among multiple hydroxyl groups. Broadly applicable approaches for carbohydrate tagging, such as sulfonylation, an important and reliable process for hydroxyl group conversion, are remarkably scarce. Addressing these limitations remains a key focus in this field to enable precision modification of the sugar ring.

Inspired by the elegant two-stage sugar engineering strategy and recognizing the need for universal site-specific carbohydrate functionalization approaches, we aim to develop an efficient carbohydrate modification protocol featuring the regiodivergent sulfonylation and ensuing transformations of carbohydrate polyols (Scheme d). By designing suitable metal catalysts, we can achieve specific coordination with the two oxygen atoms of the cis- and trans-substrate, thereby distinguishing stereoelectronic effects and activating a specific hydroxyl group. This coordination strategy enables the formation of stable metal-diol five-membered chelate rings, directing the sulfonylating reagent to preferentially attack the hydroxyl site with less steric hindrance or more favorable electronic properties. Through precise modulation of the metal center, ligands, and reaction conditions, high regioselectivity can be achieved for both cis- and trans-diols. The effectiveness of the modification stage was demonstrated by the versatile functionalization of halogenation, thiolation, deoxygenation, azidation, and uncommon glycosylation, showcasing the broad applicability of our strategy. Competitive experiments and density functional theory (DFT) calculations unveiled the reaction mechanisms and the origins of selectivity, shedding light on the key functions of the catalysts and ligands.

Results and Discussion

Development of the Site-Divergent Tagging Strategy

To initiate this project, we first investigated the key regiodivergent sulfonylation reaction of carbohydrate polyols. Considering single-electron reductive elimination is utilized as a method to develop enantioselective S–O coupling under a [Cu] catalysis with chiral multidentate ligands, we began with the methyl 4,6-O-benzylidene mannoside 1a containing free C-2 and C-3 hydroxyl groups model substrate to explore the optimal catalytic reaction conditions (Table ). Under the [Cu] system, no desired product was obtained, likely because the densely functionalized, stereochemically complex carbohydrate framework hinders effective catalyst–substrate recognition. The combination of achiral ligand bipyridine (L2) and [Cu] markedly improved the selectivity, yet affords an unsatisfactory 32% yield (see the Supporting Information Figure S1). We then proceeded to screen multiple metal catalysts with different ligands (M1–M7) in the presence of Ag₂CO₃ (Figure S2), and we found that the selectivity was significantly metal-dependent: the [Pd] catalysts yielded the 3-OTs product, whereas the [Ni] catalysts mainly produced the 2-OTs product.

1. Tagging Protocol Optimization with 1a ^,

entry	variation from the standard conditions	yields (%) (3a/4a)
1	none	98/0
2	No Ag2CO3	NR
3	Ag2CO3 (0.6 equiv)	93/0
4	K2CO3 instead of Ag2CO3	65/0
5	CuI instead of M1, Ag2CO3 (0.6 equiv)	22/0
6	PdCl2 instead of M1	61/31
7	M2 instead of M1, L2	87/0
8	M2 instead of M1, L1	98/0
9	M3 instead of M1, L1	88/0
10	M1 (0.2 equiv), no ligand	72/0
11	M2 (0.2 equiv) instead of M1, no ligand	83/0
12	No M1	12/60

^aUnless otherwise stated, reported yields are for isolated and purified products; Reaction condition A (entry 1): Substrate 1a (0.1 mmol), TsCl (1.2 equiv), M1 (0.1 equiv), L1 (0.1 equiv), Ag₂CO₃ (1.0 equiv), CHCl₃ (2.0 mL), r.t., 48 h.

^bReaction condition B (Entry 12): Substrate 1a (0.1 mmol), TsCl (1.2 equiv), L1 (0.1 equiv), Ag₂CO₃ (1.0 equiv), CHCl₃ (2.0 mL), r.t., 48 h. NR = No Reaction.

Identifying the unique catalyst-enabled selective tagging mode, we screened for optimal reaction conditions by combining 2,2′-bipyridylpalladium dichloride with diverse additional ligands, including bathophenanthroline L1, bipyridine L2, and terpyridine (Figure S3). And we found that L1 afforded 3a in a high 93% yield with excellent site selectivity (Table , entry 3). Systematic variations of other reaction parameters, including the type and equivalence of bases, and solvents (Table S1), led to the identification of the optimal conditions: 0.1 equiv [Pd], 0.1 equiv L1, 1.0 equiv Ag₂CO₃, in chloroform at room temperature. Under optimal conditions, the desired product 3a was obtained in 98% yield with exclusive O-3 selectivity (Table , entry 1). Control experiments indicated that when Ag₂CO₃ was replaced with other carbonates, the yield decreased significantly (Table , entry 4). Notably, to ensure high site selectivity, [Pd] catalyst must be added into the reaction in the form of a metal ligand (Table , entry 6). Although [Pd] alone could achieve a favorable yield in the absence of the ligand, adding L1 was proven to be beneficial to the reaction yields in most cases (Table , entries 7–9). When the ligand was omitted, even 0.2 equiv of [Pd] could not promote the reaction yield to such an excellent level (Table , entries 10–11). Interestingly, when [Pd] was deleted from the reaction mixture, an inverse selectivity was observed. Under these conditions, we found that the tosylation occurred on more kinetically accessible but lower nucleophilic O-2, and 4a was obtained in 60% yield with 5:1 O-2/O-3 selectivity (Table , entry 12, simple screening; see Table S3). This serendipity provided operationally easy and tunable protocols to realize site-switchable carbohydrate tagging. Therefore, after extensive reaction optimization, we have identified two sets of reaction conditions for orthogonal functionalization of carbohydrate diol: condition A (Table , entry 1) contains palladium, ligand, and Ag₂CO₃ for cis-diol tosylation, while condition B (Table , entry 12) only uses the ligand and Ag₂CO₃ to realize the trans-diol functionalization.

Substrate Scope and Limitations of the Tagging Protocol

With the optimal reaction established, we next aimed to confirm the scope of the tagging approach. Diverse sulfonyl chlorides were first exposed to the reaction with 1a or 1b as the 1,2-cis/trans-diol model, respectively (as presented in Figure ), affording structurally diverse glycosyl sulfonates 3b–3m and 4b–4m in good to excellent yields with remarkable site selectivity regardless of the structures of the tagging agents. The results demonstrated a remarkable tolerance of a wide range of substituents (2b–2m) possessing distinct electronic or steric characteristics within the sulfonyl chlorides to these mild conditions. Notably, the compatibility of benzenesulfonyl chlorides (BsCl for 2b), p-nitrobenzenesulfonyl chlorides (p-NsCl for 2f), methanesulfonyl chloride (MsCl for 2g), and 2,4,6-triisopropylbenzenesulfonyl chloride (TIPBsCl for 2m), which are commonly employed in organic synthesis and hydroxyl functionalization, holds significant synthetic promise as it lays a solid foundation for a multitude of versatile follow-up chemical transformations.

1.

Sulfonyl chloride scope of the tagging approach. Unless otherwise stated, reported yields are for isolated and purified products; Reaction condition A: Substrate 1a (0.1 mmol), Sulfonyl chloride (1.2 equiv), M1 (0.1 equiv), L1 (0.1 equiv), Ag₂CO₃ (1.0 equiv), CHCl₃ (2.0 mL), r.t., 48 h; Reaction condition B: Substrate 1b (0.1 mmol), Sulfonyl chloride (1.2 equiv), L1 (0.1 equiv), Ag₂CO₃ (1.0 equiv), CHCl₃ (2.0 mL), r.t., 48 h.

Meanwhile, the sulfonyl groups are extensively used in drug molecule design, and they can be introduced into molecules as bioisosteres to enhance biological activity, regulate solubility, prolong bioavailability, and improve pharmacokinetic properties. − The ability of our approach to introduce diverse sulfonyl functionalities into sugars might provide structurally novel lead glycotherapeutics.

The substrate scope of carbohydrate polyols was further evaluated to showcase versatile tagging scenarios (Figure ). The catalytic systems were utilized in a wide range of monosaccharides containing primary or secondary hydroxy groups, using tosylation as the standard tagging demonstration. The α-d-thiomannopyranoside also showed excellent selectivity and a high yield for monotosylation at O-3 (5c). Beyond site selectivity, the functional group tolerance and sensitivity were also tested. Gratifyingly, base-labile groups such as acetyl (Ac, 5d, 5k) and benzyloxycarbonyl (Cbz, 5p) functionalities were well tolerated in the Ag₂CO₃-promoted reactions. Despite the acetyl group possessing an electron-withdrawing property, the 4,6-di-O-acetyl substrate can still achieve a high yield (5d, 92%) under condition A. However, the electron-withdrawing effect enhances the acidity of the C3-OH group, thereby promoting the formation of 2,3-disubstituted product 6d. Due to the steric hindrance effect of the tert-butyldimethylsilyl (TBS) group, which obstructs the coordination between the metal and the diol of 4,6-di-O-tert-butyldimethylsilyl-mannopyranoside, the yield of 5e drops sharply under condition A, while only O-2-tosylated product 6e was generated under condition B. These results collectively suggested that our developed tagging approach could not only achieve excellent site selectivity but also be seamlessly incorporated into common carbohydrate manipulations with other frequently used protecting groups, showcasing the mildness and compatibility of the approach.

2.

Sugar scope of the tagging approach. ^a Unless otherwise stated, reported yields are for isolated and purified products; Reaction condition A: Substrate (0.1 mmol), TsCl (1.2 equiv), M1 (0.1 equiv), L1 (0.1 equiv), Ag₂CO₃ (1.0 equiv), CHCl₃ (2.0 mL), r.t.,48 h; Reaction condition B: Substrate (0.1 mmol), TsCl (1.2 equiv), L1 (0.1 equiv), Ag₂CO₃ (1.0 equiv), CHCl₃ (2.0 mL), r.t., 48 h. ^bMeCN (2.0 mL).

Various sugar substrates were subsequently exposed to the developed reaction conditions to further test the generality of this approach. Derivatives of d-galactose, l-rhamnose, d-galactosamine, and d-ribose, which contain 1,2-cis-diol, readily afforded equatorial substitution products 5f, 5g, 5h, and 5i in high yields (80–98%) under condition A, respectively. Moreover, under condition B, the corresponding products 6f, 6g, 6h, and 6i with axial substitutions were mainly obtained. The catalytic system A also exhibited remarkable catalytic activity against pyranose substrates containing 1,3-diol. Under condition B, the same product as in condition A is obtained, but the yield is significantly lower. Consequently, the 6-monosulfonylation products 5j–p were isolated in 91–98% yields starting from glycoside substrates whose O-4 and O-6 positions were unprotected. It is encouraging that we found that the thioglycoside functional group is tolerated by this catalyst system, enabling a dual functionalization pattern by either hydroxyl group functionalization or anomeric glycosylation. The reaction for ribofuranoside resulted in only moderate selectivity, perhaps due to the crowded conformation of furanoside. Selective monotosylation was then performed on the polyhydroxy carbohydrate derivatives containing cis-vicinal diol moieties, including the 6-O-triisopropylsilyl (TIPS)-protected pyranoside derivatives of galactoside (5r), mannoside (5s), and commercially available derivatives rhamnoside (5t), fucoside (5u), and arabinoside (5v, 5w). Consequently, the equatorial O-3 group underwent selective tosylation to afford the desired products in good to excellent yields ranging from 74 to 98%. For glycoside substrates methyl galactoside and mannoside containing both cis-vicinal diol and 1,3-diol, the selectivity between equatorial O-3 and the primary OH group was poorly discriminated by our reaction. We found that 34% of 3,6-OTs (5x1), 44% of 3-OTs (5x2) galactosides, as well as 38% of 3,6-OTs mannosides (5y), were formed, and the functionalizations on axial OH were not observed. With regard to free methyl glucosides and sialic acid, since no cis-diol competed with 1,3-diol in the sulfonylation reaction, sulfonylation products of primary alcohols 5z and 5za were generated with yields of 45% and 46%, respectively. The relatively low yields of 5z and 5za are likely attributable to the poor solubility of the substrates in MeCN.

Next, the sulfonylation reaction was carried out on the glycosyl derivatives containing trans-1,2-diol moieties under condition B. These conditions favor sulfonylation at the hydroxyl group adjacent to an axial alkoxy substituent in equatorial trans-diols. Applying this condition to the 2,3-O-unprotected α-d-glucopyranosides with different 4,6-O-protection (PhCH, NaphCH, tBu₂Si, [(iPr)₂Si]₂O) afforded the corresponding 2-O-sulfonylated products 6zb–6ze in good to excellent yields, assisted by the axial methoxy group. Similarly, the 2,3-O-unprotected β-d-galactopyranosides reacted selectively at the C3-OH position, providing products 6zf–6zh in excellent yields facilitated by the axial C4-O-substituent. The 2,3-O-unprotected α-d-galactopyranosides possessing two axial oxygen groups (axial methoxy group and axial C4-O-substituent) mainly afforded the tosylated O-3 product (6zi). The sulfonylation of 2,6-O-unprotected galactopyranoside revealed an unexpected regiochemical outcome. Contrary to the typical higher reactivity of primary alcohols, the O-2 was preferentially sulfonylated over the O-6, yielding 6zj. This selectivity is attributed to the significant influence of the proximal axial methoxy group.

In summary, while Conditions A and B proved to be broadly effective for a wide range of monosaccharides, several limitations were identified. Condition A exhibits high robustness for cis-diol motifs, including those bearing base-labile protecting groups (e.g., Ac, Cbz). Its efficiency, however, can be diminished by severe steric shielding near the diol (e.g., by a TBS group) or by competing 1,3-diol units within the same molecule, which may lead to mixed selectivity. Condition B reliably directs sulfonylation to the hydroxyl adjacent to an axial alkoxy group in trans-diol systems but shows negligible activity toward simple monoalcohols or symmetric cis-diols lacking this stereoelectronic guide. Substrate solubility in MeCN also posed a practical constraint in a few cases (e.g., 5z, 5za). Based on the collected data, Condition A is the method of choice for precise equatorial functionalization of cis-diols, whereas Condition B should be employed to functionalize axial hydroxyl groups and particularly equatorial hydroxyl activated by an adjacent axial oxygen atom. This delineation provides a clear guideline for selecting the appropriate catalytic regime.

Downstream Functionalizations of the Sugar Tosylates

With the core tagging strategy successfully established, we next explored the second functionalization stage, casting the innate reactivities of tosylates to expand the chemical space of carbohydrates. These modifications are anticipated to not only alter the physicochemical properties of carbohydrates but also apply to drug development, materials science, and glycobiology. Thus, the sugar tosylates were subjected to ensuing functionalization to acquire engineered uncommon sugars and rare sugars.

A series of post-tagging engineering were demonstrated using 5j as the model compound. The halogenation smoothly converted 5j into glycosyl iodides 7 or glycosyl bromides 8 with NaI or LiBr as the halogen source, respectively (Figure a). Reduction of 5j by lithium aluminum hydride (LiAlH₄) afforded 6-deoxy sugars 9. The 6-deoxy-6-S-acetyl-mannoside derivative was obtained in 76% yield through nucleophilic attack by the potassium thioacetate (KSAc)/DMF system from 5j; subsequent hydrolysis of the thioacetyl group under basic conditions formed 6-deoxy-6-mercapto-mannoside 10. A two-step conversion of tosyl glycoside into sugar-triazole glycoconjugate 11 was achieved by successive NaN₃ substitution and Cu-catalyzed click reaction, , which may provide applications in biorthogonal coupling and drug development. A structurally unique 6,6′-ether-bridged 13 disaccharide could be synthesized via Williamson ether synthesis under strongly alkaline conditions.

3.

Downstream tosylate derivatization. (a) Sulfonyl-directed diversification strategy. (b) Derivatization to access the relevant rare sugar scaffolds. (c) Deprotection of the p-toluenesulfonyl group. (d) Deprotection of the p-nitrobenzenesulfonyl group.

Uncommon amino sugars appear as rare sugars in nature, and they usually play essential roles in (i) constructing structural scaffolds for biomaterials, (ii) maintaining tissue hydration and lubrication, (iii) mediating cellular recognition, adhesion, and signaling, (iv) forming the structural basis of bacterial cell walls, capsular polysaccharides, and lipopolysaccharides (constituting critical targets for both bacterial survival and antibiotic action), and (v) serving as antibiotic components. These rare amino sugars have usually limited natural abundance, which confines their in-depth investigations. We envisaged that our two-stage modification could offer simple and convenient access to these compounds. Specifically, the protected 2-azido-2-deoxy-d-glucopyranoside 5m with the primary alcohol substituted by an O-6 Ts group was treated with NaN₃ to give the rare 6-azido sugar 14 in high yield, the skeleton of which is the key sugar unit of aminoglycoside antibiotics (Figure b). A base-mediated elimination of the Ts group of 6zb forms a three-membered 2,3-epoxide intermediate, and an O-2, O-3 double configuration inversion led to 3-azido-3-deoxy-d-altroside 15 when subjected to substitution by nucleophiles such as sodium azide, presumably due to a favorable chairlike transition state according to the Fürst–Plattner principle (trans-diaxial opening). On the other hand, 2,3-diazido-2,3-dideoxy-d-taloside 17 could be prepared from galactoside 6zi, which has a sulfonyl group at the O-3 position. Base-promoted oxirane formation, followed by NaN₃ treatment at a high temperature, affords 2-azido-2-deoxy-D-idoside 16. Triflation quantitatively gave the O-3 triflate derivative, and subsequent displacement of the resultant triflate with sodium azide produced the desired d-taloside 17 smoothly.

The current approach has been further extended to multiple functionalizations to give deoxy aminosugar 19. Reduction of O-6 tosylate by successive iodination and azobis­(isobutyronitrile) (AIBN)/tributyltin hydride (nBu₃SnH)-mediated dehalogenation delivered deoxysugar, which was subjected to triflation and then azide substitution, giving 2,4-diazido-2,4,6-trideoxy-d-glucoside derivative 19, accompanied by an inversion of the configuration at C-4. The two-step inversion strategy through triflation followed by azide substitution was similarly applied to l-fucoside 5v to generate the corresponding 2,4-diazido-2,4-dideoxy-l-rhamnoside 20 with inversion of configuration at both the C-2 and C-4 positions. The two azide groups were converted to acetamides by successive hydrogenation and selective acetylation, affording 21 in almost quantitative yield. The sulfonyl group was chemoselectively removed with a catalytic amount of acridine radical photoreductant, 9-mesityl-10-methylacridinium tetrafluoroborate (Mes-Acr-BF₄), under photoirradiation conditions at 390 nm, giving the desired 2,4-diacetamido-2,4-dideoxy-l-rhamnoside 22 in 89% yield.

Beyond acting as a handle, the sulfonates can also be used as an adaptive and removable protecting group, which can be seamlessly incorporated into the synthesis of complex oligosaccharides (Figure c,d). For example, the glycosylation of O-4 of tosylate 5l smoothly yielded β-1,4-linked disaccharide 24 with glycosyl trichloroacetimidate (TCAI) donor 23 under the catalysis of trimethylsilyl trifluoromethanesulfonate (TMSOTf) in DCM, and the tosylate remained untouched during the glycosylation stage. The sulfonyl group can be smoothly reduced by Mes-Acr-BF₄, with a 95% yield to obtain 6-OH compound 25 for the next cycle of glycosylation, which enabled the preparation of branch trisaccharide 26 in 85% yield with glycosyl TCAI donor 23. On the other hand, a 6-O-p-nitrobenzenesulfonyl (Ns)-thioglucosyl donor 27 was subjected to the glycosylation with acceptor 12 in the presence of N-iodosuccinimide (NIS)/TMSOTf in DCM at −40 °C, leading to the desired disaccharide 28 in 74% yield with an α/β ratio of 6:1. The significant α-selectivity could be attributed to the potential remote-directing effect by the O-6 sulfonyl group, further consolidating the synthetic utility of our strategy. The facile deprotection of the Ns group in 28 could be achieved by treating the disaccharide with p-toluenethiol and K₂CO₃ in MeCN, affording disaccharide 29 in satisfactory yield. Under the catalysis of TMSOTf, disaccharide acceptor 29 was glycosylated with donor 23, affording linear trisaccharide 30 in excellent yield. Therefore, by proceeding with the carbohydrate modifications, rare sugar synthesis, and oligosaccharide assembly using tagged sugar sulfonate linchpin, we have showcased the power of the two-stage functionalization strategy in carbohydrate modification.

Mechanistic Investigations

A series of control and competitive experiments were carried out to elucidate the reaction mechanism. We found that, under condition A, the sulfonylation of trans-diol (37), monohydroxy substrate (32), or 1,4-butanediol (34) failed to give any desired products (Figure ). In contrast, competitive sulfonylation experiments revealed distinct reactivity patterns depending on the substrate structures. For instance, the competitive sulfonylation of a mixture of 1-phenyl-1,2-ethanediol 31 and its corresponding monohydric alcohol (phenylethyl alcohol 32) only led to 38, yet 32 was completely inert under these conditions, indicating a strong preference for the diol over the monohydric alcohol as the substrate. Similarly, in a competitive sulfonylation reaction involving a mixture of ethylene glycol 33 and 1,4-butanediol 34, the formation of monofunctionalized 39 was exclusive, while when a mixture of 1,2-diol (33) and 1,3-diol (propan-1,3-diol, 35) was subjected to competitive sulfonylation, a nearly equimolar mixture of sulfonylated 1,2-diol and 1,3-diol was obtained, suggesting comparable reactivity between these diols under the given conditions. Notably, in terms of the competitive experiments involving cis-cyclopentanediol (36) and 1,4-butanediol (34), the monosulfonylated products of cyclopentanediol (41) were solely obtained. In contrast, no products were detected in the competitive experiments with trans-cyclopentanediol and 1,4-butanediol under the same conditions, underscoring the critical role of stereochemistry in the reaction. These results collectively support the hypothesis that the [Pd] catalyst facilitates the formation of five- and six-membered cyclic intermediates, which are more readily accessible from cis-1,2-diols and 1,3-diol. Also, these mechanistic experiments provided hints about the compatibility of the developed tagging approach with carbohydrates with multiple hydroxyl groups and complex chemical environments.

4.

Competitive experiments.

To further elucidate the mechanistic details leading to the controlled regioselectivity under different conditions for sugars with different stereoelectronic properties, we have proposed two catalytic cycles for the tosylation of sugars under conditions A and B (Figure ), and performed DFT simulations of the reaction routes along the reaction cycles at the MN15L/ma-def2-TZVPP/(SMD, chloroform) level of theory, employing model mannoside 1a (condition A) and glucoside 1b (condition B) with unfunctionalized O-2 and O-3 groups, sulfonyl chloride 2b, palladium complex M1 (condition A), ligand L1 (condition A), and Ag₂CO₃ (condition A and condition B). Gibbs free energy of the calculated species was reported in kcal/mol, with the starting materials taken as the reference (0.0 kcal/mol) and energy barriers reported as the Gibbs free energy difference between the transition state and the immediately reacting intermediates. ,

5.

Proposed catalytic cycles under conditions A and B.

A comparison of the thermodynamics of the putative products arising from tosylation of 1a and 1b at O-2 and O-3 positions indicates that for both, the 3-tosylated product is more stable compared to the 2-tosylated product (3b-O-2 tosylate from mannoside 1a, and 4b-O-2 tosylate from glucoside 1b; taking the energy of the O-2-tosylated products as the reference, ΔG _3b‑3OTs = −3.3 kcal/mol, ΔG _4b‑3OTs = −3.1 kcal/mol). Therefore, the distinct regioselectivity of sugars under different reaction conditions is probably kinetically controlled.

Under the specified condition A, we have simulated a proposed reaction cycle involving [Pd] coordination, regioselective hydroxyl deprotonation, and S_N2-attack of the sulfonyl chloride by the hydroxide nucleophile. M1 (PdL2Cl₂) and L1 undergoes ligand exchange to generate IM1 (PdL1Cl₂, ΔG = −1.7 kcal/mol). Then, in the presence of Ag₂CO₃, IM1 is converted to the active catalyst , IM3 ([Pd­(II)­L1]²⁺, ΔG = −10.4 kcal/mol), driven by the precipitation of AgCl solid from the solution. The [Pd] center of IM3 then coordinates with the cis-diol motifs in 1a to form IM4 (ΔG = −18.5 kcal/mol). Subsequently, Ag₂CO₃ acts as a mild base to preferentially deprotonate the equatorial 3-OH group (ΔG _{TS1_3O} ^‡ = +4.2 kcal/mol; ΔG _{TS1_2O} ^‡ = +5.5 kcal/mol), forming IM5_3deH (ΔG = −19.6 kcal/mol) with 3-oxide preferentially over IM5_2deH (ΔG = −18.0 kcal/mol) with 2-oxide. With excess and Lewis acidic Ag₂CO₃ noncovalently complexing to the chlorine atom in 1b, forming IM5′ (ΔG = −27.2 kcal/mol), activating the sulfonyl chloride, nucleophilic 3-oxide in IM5_3deH or 2-oxide IM5_2deH can undergo S_N2-attack of the sulfonyl chloride, passing through corresponding transition states (ΔG _{TS2_3O} ^‡ = +7.4 kcal/mol; ΔG _{TS2_2O} ^‡ = +8.6 kcal/mol, see Figure S4 for TS structures), forming the corresponding tosylated products IM6_3OTs (ΔG = −36.8 kcal/mol) at a faster rate compared to IM6_2OTs (ΔG = −33.7 kcal/mol). Dissociation of [Pd­(II)­L1]²⁺ from IM6 gives the corresponding products 3b_3OTs (ΔG = −34.4 kcal/mol) and 3b_2OTs (ΔG = −31.1 kcal/mol) and regenerates the active catalyst, completing the catalytic cycle. The regioselective formation of 3-tosylated product results from the thermokinetically favored 3-OH deprotonation in IM4 and the kinetically favored IM5_3deH S_N2-attack on [Ag]-activated TsCl. For IM4 arising from 1a, coordination of the [Pd­(II)­L1]²⁺ complex enhances the acidity of the equatorial 3-hydroxyl group more significantly compared to the axial 2-hydroxyl group, with stronger electron donation of the 3-position O atom to the metal center, predisposing the complex to selective 3-deprotonation. The stronger nucleophilicity of 3-oxide in IM5_3deH compared to 2-oxide in IM5_2deH can be explained by the stereoelectronic effect of the axial conformation of 2-OH in mannose 1a, which maximizes the donation of n­(O2) electron density to the antibonding σ*(C1–O1) MO, reducing the local electron density and nucleophilicity of the 2-oxide. The stereoelectronic effect is confirmed by orbital density analysis of IM5_3deH and IM5_2deH (see SI) and calculation of the nucleophilic Fukui function of 3-oxide in IM5_3deH and 2-oxide in IM5_2deH (Table S4).

For d-glucoside 1b with anomeric α-OMe, the equatorial 2,3-trans-diol configuration renders 2,3-coordination of the sugar to the metal catalyst complexes, either [Pd] (condition A) or [Ag] (condition B), relatively unfavorable, while the 1,2-cis-configuration of 1-OMe and 2-OH renders coordination of the [Ag] complex feasible. In the catalytic cycle of tosylation of 1b under condition B, Ag₂CO₃ is first converted to the active catalyst [Ag­(I)­L1]⁺ (ΔG = −9.4 kcal/mol), which then coordinates with 1-axial OMe and 2-equatorial OH in 1b to form IM8 (ΔG = −17.2 kcal/mol). The coordination of 2-OH to the [Ag] metal center reduces the electron density of 2-OH and enhances its acidity, and carbonate anions in the solution selectively deprotonate 2-OH as a base (ΔG _TS1 ^‡ = +4.0 kcal/mol) to form IM9 (ΔG = −22.0 kcal/mol) with 2-alkoxide. Then, with the chlorine atom in 1b activated by noncovalent coordination to the excess Lewis acidic Ag₂CO₃, nucleophilic 2-oxide in IM9 undergoes S_N2-attack of the sulfonyl chloride, passing through corresponding transition states (ΔG _TS2 ^‡ = +5.6 kcal/mol, see Figure S5 for TS structures), forming the corresponding O-2-tosylated products IM10 (ΔG = −32.3 kcal/mol). Dissociation of [Ag­(I)­L1]⁺ from IM10 gives the corresponding product 4b_2OTs (ΔG = −34.4 kcal/mol) and regenerates the active catalyst to complete the catalytic cycle. Notably, deprotonation of 2-OH in IM8 results in dissociation of the [Ag] center from 1-OMe in IM9, and the 1-OMe coordination is recovered in IM10 after tosylation (Figure S5). The regioselectivity arises from the preferential 1,2-coordination of the [Ag] complex to 1b, which results in selective deprotonation of coordinated 2-OH and produces 4b_2OTs as the exclusive product under condition B.

Collectively, our dual-mode catalytic system operates on complementary design principles, enabling the predictable and divergent site-selective modification of hydroxyl groups on sugar rings. The core mechanistic foundations are as follows. For cis-diol units, the [Pd]/L1 complex engages in bidentate chelation, preferentially activating the equatorial hydroxyl through stereoelectronic effects that enhance its electrophilicity and acidity, thereby directing functionalization to the equatorial position. In the absence of [Pd], the Ag­(I)/L1 system switches the reaction pathway, guiding sulfonylation to the axial hydroxyl (a site that is typically kinetically inert under conventional conditions). For substrates containing axial oxygen substituents (commonly found in trans-diol configurations), the Ag­(I)/L1 complex recognizes the 1,2-cis pair formed by the axial alkoxy group and its adjacent equatorial hydroxyl, selectively activating and guiding sulfonylation to that equatorial position, illustrating the directing role of the axial oxygen. Together, these mechanisms establish a programmable derivatization platform in which selectivity is dictated primarily by the choice of catalytic regime ([Pd] present or absent) rather than by inherent substrate distinctions, thereby providing a versatile and robust methodology for the precise modification of complex carbohydrate molecules.

Conclusions

In summary, a dual-mechanism sulfonylation tagging and ensuing functionalization strategy has been developed to establish a powerful molecular modification platform for (partially) unprotected carbohydrates. The presence or absence of [Pd] dictates the site selectivity. The [Pd] system activates the equatorial hydroxyl in cis-diols, while the Ag­(I) system (in the absence of [Pd]) targets the kinetically inert axial hydroxyl of cis-diols, as well as trans-diol substrates. Extensive substrate scope studies confirm the versatility of this approach, accommodating structurally diverse monosaccharides with varying protecting groups and stereochemical configurations. The divergent utility of this approach has been demonstrated by plenty of post-tagging transformations in skeletal modification, rare sugar synthesis, and oligosaccharide assembly, enabling expeditious and facile preparation of sugar entities of medicinal and biological interest. Mechanistic proposal, supported by competitive experiments and DFT calculations, underscores the critical role of cyclic intermediates between palladium and cis-diol and the axial oxygen effects in directing regioselectivity. Overall, this methodology provides a robust toolkit for direct, programmable engineering of carbohydrate architectures, significantly advancing the utilization of functional sugars in various fields of medicinal chemistry, chemical biology, and materials science.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacsau.5c01530.

• Full details of the experimental procedures for the synthesis of all new substances, together with characterization data, including computational details (PDF)

∇.

S.Z., C.H., A.G., and H.D. contributed equally to this work. H.C., F.D., X.L., and S.Z. conceived the project. S.Z., C.H., A.G., J.L., and L.Y. performed the experiments. S.Z., C.H., A.G., H.D., and G.B. cowrote the paper and the Supporting Information

The authors declare no competing financial interest.