Recent Advances in Chemical Synthesis of Amino Sugars

Amino sugars are a kind of carbohydrates with one or more hydroxyl groups replaced by an amino group. They play crucial roles in a broad range of biological activities. Over the past few decades, there have been continuing efforts on the stereoselective glycosylation of amino sugars. However, the introduction of glycoside bearing basic nitrogen is challenging using conventional Lewis acid-promoted pathways owing to competitive coordination of the amine to the Lewis acid promoter. Additionally, diastereomeric mixtures of O-glycoside are often produced if aminoglycoside lack a C2 substituent. This review focuses on the updated overview of the way to stereoselective synthesis of 1,2-cis-aminoglycoside. The scope, mechanism, and the applications in the synthesis of complex glycoconjugates for the representative methodologies were also included.


Introduction
A particular kind of carbohydrates known as amino sugars have one or more hydroxyl groups replaced by amino groups. Amino sugars exist ubiquitously in nature and are present in many biologically important and naturally occurring oligosaccharides and natural products. These molecules are closely related to people's lives and play a vital role in the human battle against diseases, and increase the quality and the yield of grain. For instance, erythromycin and other aminoglycoside and macrolide antibiotics, including amikacin and ribostamycin, which all incorporate the amino sugar sub-unit and are used as the first-line therapeutic medications for the treatment of bacterial infections, have prevented millions of deaths ( Figure 1A). Amino sugars are also found in a wide range of natural products, such as Blasticidin S, Ningnanmycin, and Xinaomycin, and have been utilized successfully as agricultural antibiotics [1][2][3] (Figure 1B). Additionally, sialic acids, which are essential for cell mutual recognition, differentiation, adhesion, and malignancy, and 2-amino sugars, which are widely prevalent in various sugar complexes like polysaccharides, glycopeptides, glycolipids, and glycoproteins are both unique amino sugars. ( Figure 1C) [4]. At the present stage, the availability of such compounds relies heavily on microbial fermentation, however, it is hard to quickly construct compound libraries containing multiple structural analogues in efficient ways. As a result, they are not suitable for a broad range of bioactivity screening and structure-activity relationship studies, which, ultimately, restricts the development of their activity and exploration of their functionality.
On the other hand, chemical synthesis of aminoglycosides is always associated with great challenge because: (i) the inherent regioselectivity and stereoselectivity issues in chemical synthesis of carbohydrates resulted in the lack of a general protocol to construct glycosides in an efficient manner [5,6]; (ii) competitive coordination of the basic amine to the Lewis acid promoter renders the traditional glycosylation conditions inefficient [7][8][9][10][11]; and (iii) many amino monosaccharides are not commercially available or available at a high price and often require the introduction of amino groups into the scaffolds by chemical Despite these challenges, in view of the structural and pharmacological diversity of aminoglycosides, many elegant chemical methods were thus developed for the synthesis of aminoglycosides. In this review, we wish to give the reader an overview of the recent achievements in stereoselective synthesis of aminoglycosides by using amino sugars as glycosyl donors. Since 6-amino-6-deoxy sugars could be readily obtained from corresponding 6-OH analogues, and there are several graceful articles and reviews on the synthesis of 1-amino-1-deoxy sugars (N-glycosides) [13,14], this review does not include the synthesis of 6-amino-6-deoxy sugars and 1-amino-1-deoxy sugars, but focuses on the chemical synthesis of 2-amino-2-deoxy, 3-amino-3-deoxy, and 4-amino-4-deoxy glycosides. This mini-review is divided into the following parts: (1) glycosylation of 2-amino-2-deoxysugars with different glycosyl donors; and (2) glycosylation of 3-amino-3deoxysugars and 4-amino-4-deoxysugars. The formation of 1,2-trans-aminoglycoside which exists in nature such as chitin, chitosan, as well as many O-linked GlcNAc targets and N-linked glycans, can be reliably performed with the assistance of a neighboring participating group at C(2)-amide to provide the β-linked product 10 (Scheme 1a) [20][21][22]. There are several elegant reviews on the stereoselective formation of 1,2-trans-2-aminoglycoside [16,[23][24][25]. Here, we only deal with two of the very recent representative examples. As for a non-participatory group at C(2)-amide, Li's group developed an effective strategy for β-selective glycosylation of 2,4dinitrobenzenesulfonyl (DNs)-protected glucosaminosyl and galactosaminosyl donors using Yu's glycosyl o-hexynylbenzoates as the donor (Scheme 2) [26]. The methodology is compatible with a wide range of substrates, including O-, N-, and C-nucleophiles. A plausible mechanism suggested that the reaction was achieved through a SN2-like pathway. It was supposed that, compared to solvent-separated ion pair (SSIP) intermediate C, the equilibrium heavily shifts toward the intermediates glycosyloxyisochromenylium A and contact ion pair (CIP) B due to the instability of C caused by the strong electron-withdrawing effect of C(2)-DNsNH. Moreover, nucleophiles preferentially a ack intermediates A and B to stereocontrollably result in β-glucosaminosides in an SN2-type process ( Figure 3). Very recently, Liu and co-workers developed a donor-acceptor cyclopropane (DAC)/Sc(OTf)3 system to activate C(2)-DNsNH-protected selenoglycoside donors (Scheme 3) [27]. The mechanism suggested the DNs group would stabilize the glycosyl zwi erion intermediate (III) to engage in the SN2-like displacement to form the 1,2-trans glycosaminyl bonds in a nonparticipating manner.   1,2-cis-Aminoglycoside, the key constituents in many naturally occurring bioactive glycoconjugates, are found to have a wide range of biological activities [15,16], therefore, it is meaningful to study the corresponding synthetic methodologies. The preparation of 1,2-cis-aminoglycoside 6 is challenging as the most readily amide protecting group for amine would produce the undesired 1,2-trans-product 5 due to the participation of the neighboring-group (Scheme 1a), while a C(2)-non-participatory group would unpredictably lead to the anomeric mixtures at the newly formed glycosidic bond, even though the 1,2-cis-anomer 6 would be produced as the major product due to thermodynamic stability (Scheme 1b). Herein, we will discuss the methodologies for the construction of 1,2-cisaminoglycoside, according to the employed glycosyl donors. The associated mechanism, the influence of additive to the stereoselectivity, and the application of the methodology to the synthesis of some biologically important oligosaccharides and glycoconjugates will also be included.

Glycosylation with C(2)-Azido Donors
The C(2)-azido functionality serves as an excellent latent amine-precursor that is stable in both acidic and basic circumstances, and can, subsequently, be selectively converted into the amino group, giving the most versatile route for the formation of 1,2-cis-aminoglycoside. Due to the chemical compatibility of C(2)-azido with many different kinds of glycosyl donors, including glycosyl halides, trichloroacetimidates, and thioglycosides, it has been applied in total syntheses of a broad range of biologically relevant oligosaccharides and glycoconjugates [28]. The needed C(2)-azide glycosyl donors could be easily prepared from the corresponding 2-amino sugars or by azidation of 2-O-triflate. The radical process from glycal was also developed by Tingoli and co-workers by using PhI(OAc)2 as radical initial and NaN3 as N3 source [29]. All of these protocols supported the application of C(2)-azido glycosyl donor for glycosylation.

Glycosylation with C(2)-Azido Donors
The C(2)-azido functionality serves as an excellent latent amine-precursor that is stable in both acidic and basic circumstances, and can, subsequently, be selectively converted into the amino group, giving the most versatile route for the formation of 1,2-cis-aminoglycoside. Due to the chemical compatibility of C(2)-azido with many different kinds of glycosyl donors, including glycosyl halides, trichloroacetimidates, and thioglycosides, it has been applied in total syntheses of a broad range of biologically relevant oligosaccharides and glycoconjugates [28]. The needed C(2)-azide glycosyl donors could be easily prepared from the corresponding 2-amino sugars or by azidation of 2-O-triflate. The radical process from glycal was also developed by Tingoli and co-workers by using PhI(OAc) 2 as radical initial and NaN 3 as N 3 source [29]. All of these protocols supported the application of C(2)-azido glycosyl donor for glycosylation.
Glycosylation by in situ activation of 1-OH sugar derivatives is convenient because it avoids preparation of reactive glycosyl donors. Koto et al. reported the synthesis of α-/(1,2-cis)-glycosylation of 2-azido-3,4,6-tri-O-benzyl-2-deoxy-D-glucose and 2-azido-3,4,6tri-O-benzyl-2-deoxy-D-galactose, which were prepared from corresponding glycals via azido-nitration followed by hydrolyzing 1-O-nitrated product on silica-gel column. By employing p-nitrobenzenesulfonyl chloride (NsCl)-silver triflate-triethylamine and N,Ndimethylacetamide (DMA) as the activating system, the anomeric free hydroxyl group could be activated in situ to produce desired products in moderate-to-good yields. The addition of DMA is crucial to control the stereoselectivity as the similar reaction in the absence of DMA lead to the formation of 1,2-trans-isomers (Scheme 4) [30]. tri-O-benzyl-2-deoxy-D-galactose, which were prepared from corresponding glycals via azido-nitration followed by hydrolyzing 1-O-nitrated product on silica-gel column. By employing p-nitrobenzenesulfonyl chloride (NsCl)-silver triflate-triethylamine and N,Ndimethylacetamide (DMA) as the activating system, the anomeric free hydroxyl group could be activated in situ to produce desired products in moderate-to-good yields. The addition of DMA is crucial to control the stereoselectivity as the similar reaction in the absence of DMA lead to the formation of 1,2-trans-isomers (Scheme 4) [30]. Glycosyl nitrates are important synthetic intermediates which could be applied in subsequent functional group transformations. Demchenko's group reported the utility of glycosyl nitrates as donors for O-glycosidation of 2-amino-2-deoxysugars in the presence of ytterbium(III) trifluoromethanesulfonate (Yb(OTf)3) (Schemes 5 and 6) [31]. Coupling the benzoylated donor 48 with acceptors 50-52 generates the desired disaccharides in 26-65% yields with good to excellent stereoselectivity (Scheme 5). In contrast to the Bz-protecting group, however, Bn-protected glycosyl nitrates delivered the corresponding products in shorter reaction time, higher yield, but worse stereoselectivity. The author concluded that this is because of the inverse correlation between the reactivity and stereoselectivity. Besides, conformation-restrainin sugar such as 2-azido-4,6-O-benzylidene-2-deoxy-3-O-triisopropylsilyl-β-D-glucosyl nitrate donor 59 , was subjected to the standard reaction condition to couple with glycosyl acceptors 50-52. Despite the fact that the conformation-restraining 4,6-O-benzylidene-protected glucosyl donor produces the glycosylated product with α-selectivity [32][33][34][35], this reaction merely produced the products with lower yield and inferior selectivity (Scheme 6). Glycosyl nitrates are important synthetic intermediates which could be applied in subsequent functional group transformations. Demchenko's group reported the utility of glycosyl nitrates as donors for O-glycosidation of 2-amino-2-deoxysugars in the presence of ytterbium(III) trifluoromethanesulfonate (Yb(OTf) 3 ) (Schemes 5 and 6) [31]. Coupling the benzoylated donor 48 with acceptors 50-52 generates the desired disaccharides in 26-65% yields with good to excellent stereoselectivity (Scheme 5). In contrast to the Bz-protecting group, however, Bn-protected glycosyl nitrates delivered the corresponding products in shorter reaction time, higher yield, but worse stereoselectivity. The author concluded that this is because of the inverse correlation between the reactivity and stereoselectivity. Besides, conformation-restrainin sugar such as 2-azido-4,6-O-benzylidene-2-deoxy-3-Otriisopropylsilyl-β-D-glucosyl nitrate donor 59, was subjected to the standard reaction condition to couple with glycosyl acceptors 50-52. Despite the fact that the conformationrestraining 4,6-O-benzylidene-protected glucosyl donor produces the glycosylated product with α-selectivity [32][33][34][35], this reaction merely produced the products with lower yield and inferior selectivity (Scheme 6).
During synthesis of the core structures of glycosyl phosphatidylinositol (GPI), Guo's group used the 2-azido-2-deoxy-D-glucopyranose donor 63 to couple with inositol acceptor 64, generating the desired product 65 and its β-isomer 65' in 70% yield with 4:3 α-/β-selectivity (Scheme 7) [36]. The two isomers can be separated by silica gel column, facilitating its further application. Even though no detailed experimental procedure was disclosed, they used dichlorohafnocene (Cp 2 HfCl 2 ) and AgOTf as promotors, which may limit its application in large-scale reactions.
During synthesis of the core structures of glycosyl phosphatidylinositol (GPI), Guo's group used the 2-azido-2-deoxy-D-glucopyranose donor 63 to couple with inositol acceptor 64, generating the desired product 65 and its β-isomer 65' in 70% yield with 4:3 α-/βselectivity (Scheme 7) [36]. The two isomers can be separated by silica gel column, facilitating its further application. Even though no detailed experimental procedure was disclosed, they used dichlorohafnocene (Cp2HfCl2) and AgOTf as promotors, which may limit its application in large-scale reactions. In 2011, Sewald's group reported a one-pot procedure to prepare 2-azido glycosyl chlorides via FeCl 3 /H 2 O 2 mediate azidochlorination of glycals. A series of mono-and disaccharide 2-azido glycosyl chlorides were synthesized from corresponding glycals in 37-78% yields with high α-selectivity. The reaction could also be carried out at 15 g scale which produced the product in similar yield and stereoselectivity. The produced 2-azidated galactosyl donor 66 was then used to prepare glycosylated threonine building block 69 under Koenigs-Knorr glycosylation conditions. The glycosylated amino acid derivative 69 could be used as the substrate for the synthesis of T N -antigen (α-GalNAc-Ser/Thr) (Scheme 8) [37]. The yields and diastereoselectivities of the threonine glycosylation are also comparable to the corresponding bromide 67, which was obtained from an azidonitration reaction followed by bromination [38,39].
In 2011, Sewald's group reported a one-pot procedure to prepare 2-azido glycosyl chlorides via FeCl3/H2O2 mediate azidochlorination of glycals. A series of mono-and disaccharide 2-azido glycosyl chlorides were synthesized from corresponding glycals in 37-78% yields with high α-selectivity. The reaction could also be carried out at 15 g scale which produced the product in similar yield and stereoselectivity. The produced 2-azidated galactosyl donor 66 was then used to prepare glycosylated threonine building block 69 under Koenigs-Knorr glycosylation conditions. The glycosylated amino acid derivative 69 could be used as the substrate for the synthesis of TN-antigen (α-GalNAc-Ser/Thr) (Scheme 8) [37]. The yields and diastereoselectivities of the threonine glycosylation are also comparable to the corresponding bromide 67, which was obtained from an azidonitration reaction followed by bromination [38,39]. Heparin, a kind of glycosaminoglycan, is crucial to a number of important biological functions [40]. Highly selective installation of the 2-amino-α-glycosidic linkage in the heparin backbone is challenging [41]. In 2002, Seeberger's group reported a new method for the stereoselective installation of α-glucosamine linkages by locking the conformation of the monosaccharide acceptors (Scheme 9) [42]. Coupling the donor 70 with conformationlocked glucuronic acid acceptor 71 and iduronic acid acceptor 73 provides the desired disaccharides 72 and 74 in 92% and 90%, respectively, with completely α-selectivity. Heparin, a kind of glycosaminoglycan, is crucial to a number of important biological functions [40]. Highly selective installation of the 2-amino-α-glycosidic linkage in the heparin backbone is challenging [41]. In 2002, Seeberger's group reported a new method for the stereoselective installation of α-glucosamine linkages by locking the conformation of the monosaccharide acceptors (Scheme 9) [42]. Coupling the donor 70 with conformationlocked glucuronic acid acceptor 71 and iduronic acid acceptor 73 provides the desired disaccharides 72 and 74 in 92% and 90%, respectively, with completely α-selectivity. In 2007, Geert-Jan Boons and co-workers reported the glycosylation of 2-azido-2-deoxy-glycosyl trichloroacetimidates 75 and 76 with a variety of glycosyl acceptors in the presence of PhSEt or thiophene to provide disaccharides in good-to-high yields with excellent α-selectivity (Scheme 10) [43]. The enhanced α-selectivity by addition thioether was proved to be derived from successive configurational inversion via SN2-type reactions, that is, the thioether reacted with α-triflate which was confirmed by chemical shift and a Scheme 9. The construction of a 2-amino-α-glycosidic bond by using conformation-locked acceptors.
In 2007, Geert-Jan Boons and co-workers reported the glycosylation of 2-azido-2deoxy-glycosyl trichloroacetimidates 75 and 76 with a variety of glycosyl acceptors in the presence of PhSEt or thiophene to provide disaccharides in good-to-high yields with Molecules 2023, 28, 4724 9 of 49 excellent α-selectivity (Scheme 10) [43]. The enhanced α-selectivity by addition thioether was proved to be derived from successive configurational inversion via S N 2-type reactions, that is, the thioether reacted with α-triflate which was confirmed by chemical shift and a smaller coupling constant, producing the β-substituted anomeric sulfonium ion then glycosyl acceptors attacked the anomeric sulfonium ion to generate the product with high α-selectivity. It was also found that in most cases: (i) elevating the reaction temperature could dramatically improve the stereoselectivity; and (ii) thiophene produced the products with better α-selectivity compared with ethyl phenyl sulfide (PhSEt). However, a similar reaction with acceptor 52 produced the product 79 in lower yield with only α-selectivity regardless of an increase in the reaction temperature or in the presence of thioether; the lower yield was because of the formation of an anomeric trichloroacetamide by-product. They also synthesized the 3,4,6-tri-O-benzyl-glucopyranosyl donor 76, and carried out glycosylation reaction under the same conditions with or without thiophene. Similar to the benzyl-protected glycosyl nitrate donors, benzyl-protected glycosyl trichloroacetimidate 76 also delivered the corresponding glycoside in lower stereoselectivity. The author speculated that the glycosylation may take place directly between acceptors with the oxacarbenium ion intermediate produced from the high reactivity of the benzyl-protected glycosyl donor. In 2008, Field's group disclosed synthesis of mucin-related glyco-aminoacids and glycopeptides by using phenylseleno 3,4,6-tri-O-acetyl-2-azido-2-deoxy-α-D-galactopyranoside 81 as glycosyl donor under the promotion of I2 and DDQ. In this case, selenogalactoside 81, which could be readily prepared from galactal via azidophenylselenation, was reacted with serine 82 and threonine 83 derivatives in toluene/dioxane to achieve exclusive α-stereoselectivity (Scheme 11) [44]. It was found that the use of toluene/dioxane mixture solvent and DDQ are critical for improving the yield and stereoselectivity. Even though the exact mechanism for DDQ-promoted glycosylation was not clear, they believed that high α-stereoselectivity was perhaps due to the formation of the O-β-glycosyl oxonium ion derived from dioxane. In 2008, Field's group disclosed synthesis of mucin-related glyco-aminoacids and glycopeptides by using phenylseleno 3,4,6-tri-O-acetyl-2-azido-2-deoxy-α-D-galactopyranoside 81 as glycosyl donor under the promotion of I 2 and DDQ. In this case, selenogalactoside 81, which could be readily prepared from galactal via azidophenylselenation, was reacted with serine 82 and threonine 83 derivatives in toluene/dioxane to achieve exclusive α-stereoselectivity (Scheme 11) [44]. It was found that the use of toluene/dioxane mixture solvent and DDQ are critical for improving the yield and stereoselectivity. Even though the exact mechanism for DDQ-promoted glycosylation was not clear, they believed that high α-stereoselectivity was perhaps due to the formation of the O-β-glycosyl oxonium ion derived from dioxane.
In 2008, Field's group disclosed synthesis of mucin-related glyco-aminoacids and glycopeptides by using phenylseleno 3,4,6-tri-O-acetyl-2-azido-2-deoxy-α-D-galactopyranoside 81 as glycosyl donor under the promotion of I2 and DDQ. In this case, selenogalactoside 81, which could be readily prepared from galactal via azidophenylselenation, was reacted with serine 82 and threonine 83 derivatives in toluene/dioxane to achieve exclusive α-stereoselectivity (Scheme 11) [44]. It was found that the use of toluene/dioxane mixture solvent and DDQ are critical for improving the yield and stereoselectivity. Even though the exact mechanism for DDQ-promoted glycosylation was not clear, they believed that high α-stereoselectivity was perhaps due to the formation of the O-β-glycosyl oxonium ion derived from dioxane. Scheme 11. The formation of TN-antigen building blocks. Scheme 11. The formation of T N -antigen building blocks.
Very recently, during the synthesis of deoxyfluorinated T N antigen analogues, Karban and co-workers prepared a series of C3 and C4 deoxyfluorinated galactosazide thiodonors to react with a variety of glycosyl acceptors and threonine derivatives in order to evaluate their stereoselectivity. It was found that for the threonine derivatives, the O-benzylated C4 fluoro-donor gave the coupling products in 63% to 69% yield with α:β selectivity ranged from 2.5:1 to 3.1:1, regardless of solvent (ether solvents such as diethyl ether and dioxane gave similar stereoselectivity as DCM alone). For C3 fluoro-donors, even though the donors with acyl and silyl protection at the 4-and 6-positions did not improve the α-selectivity, the reaction between 4,6-di-tert-butylsilylene-protected 3-fluoro galactosazide thiodonor 88 and threonine benzyl ester 83 produced the glycoside product 89 in excellent 1,2-cis-stereoselectivity (Scheme 12) [45]. Very recently, during the synthesis of deoxyfluorinated TN antigen analogues, Karban and co-workers prepared a series of C3 and C4 deoxyfluorinated galactosazide thiodonors to react with a variety of glycosyl acceptors and threonine derivatives in order to evaluate their stereoselectivity. It was found that for the threonine derivatives, the Obenzylated C4 fluoro-donor gave the coupling products in 63% to 69% yield with α:β selectivity ranged from 2.5:1 to 3.1:1, regardless of solvent (ether solvents such as diethyl ether and dioxane gave similar stereoselectivity as DCM alone). For C3 fluoro-donors, even though the donors with acyl and silyl protection at the 4-and 6-positions did not improve the α-selectivity, the reaction between 4,6-di-tert-butylsilylene-protected 3-fluoro galactosazide thiodonor 88 and threonine benzyl ester 83 produced the glycoside product 89 in excellent 1,2-cis-stereoselectivity (Scheme 12) [45]. Due to the protecting groups' significant impact on the stereoselectivity of glycosyl donors, Li's group was committed to examining the stereochemistry effect of the protecting groups [46][47][48]. They synthesized a panel of GalN3 thioglycoside donors with different protecting groups at 3-, 4-, and 6-positions, and compared the stereoselectivity of the glycosylation with primary alcohol 14. They discovered that the protecting group at 3-position had no noticeable effect on the stereoselectivity because both benzyl-and acetyl-protected thioglycoside donors delivered the products in comparable selectivity, whereas the acyl-protecting group at 4-position could increase the stereoselectivity with 4-benzoylprotected thioglycoside producing the product in 96% yield with 1:6 of α:β. Furthermore, 2-azido-3,4,6-tri-O-acetyl-1-thio-glycoside produced the glycosylation product in a 3:1 α:β Due to the protecting groups' significant impact on the stereoselectivity of glycosyl donors, Li's group was committed to examining the stereochemistry effect of the protecting groups [46][47][48]. They synthesized a panel of GalN 3 thioglycoside donors with different protecting groups at 3-, 4-, and 6-positions, and compared the stereoselectivity of the glycosylation with primary alcohol 14. They discovered that the protecting group at 3-position had no noticeable effect on the stereoselectivity because both benzyl-and acetyl-protected thioglycoside donors delivered the products in comparable selectivity, whereas the acyl-protecting group at 4-position could increase the stereoselectivity with 4-benzoylprotected thioglycoside producing the product in 96% yield with 1:6 of α:β. Furthermore, 2-azido-3,4,6-tri-O-acetyl-1-thio-glycoside produced the glycosylation product in a 3:1 α:β selectivity, while coupling the 6-TBDPS-protected similar donor 100 with the acceptor 14 gave the desired disaccharide product 102 in 99% yield with α:β > 20:1 [49] (Scheme 13). They also applied the optimized acyl-thioglycoside with the TBDPS protecting group at 6-position to the synthesis of precursors of T N antigen. The strategy of reagent-controlled glycosylation was also employed in the synthesis of oligosaccharides. Codée's group reported methyl(phenyl)formamide (MPF) could be employed as an additive for the glycosylation of 2-azido-2-deoxyglucose building blocks [50]. By using the strategy of MPF mediated glycosylation, an all-1,2-cis-linked tetraglucosamine was successfully synthetized. The repetitive coupling step of glycosylation reactions mediated by MPF between donor 103 and acceptor under the above identified reaction conditions provided the desired disaccharide in good yield with excellent α-selectivity (Scheme 14). Furthermore, using the same method, a fragment of Pel hexasaccharide 116 that contained both GalN and GlcN residues was also effectively synthesized in 86% yield with α:β = 10:1 (Scheme 15). The strategy of reagent-controlled glycosylation was also employed in the synthesis of oligosaccharides. Codée's group reported methyl(phenyl)formamide (MPF) could be employed as an additive for the glycosylation of 2-azido-2-deoxyglucose building blocks [50]. By using the strategy of MPF mediated glycosylation, an all-1,2-cis-linked tetraglucosamine was successfully synthetized. The repetitive coupling step of glycosylation reactions mediated by MPF between donor 103 and acceptor under the above identified reaction conditions provided the desired disaccharide in good yield with excellent α-selectivity (Scheme 14). Furthermore, using the same method, a fragment of Pel hexasaccharide 116 that contained both GalN and GlcN residues was also effectively synthesized in 86% yield with α:β = 10:1 (Scheme 15).

Glycosylation with Ring-Fused 2,3-Oxazolidinone Thioglycosides
In 2001, Kerns and coworkers reported a protocol using 2,3-trans-oxazolidinone as an effective protective group with high α-selectivity toward glycosylation. A notable feature of this class of α-selective D-glucosaminyl donors is that the oxazolidinone group affords simultaneous protection of the 2-N and 3-O positions [77].

Glycosylation with Ring-Fused 2,3-Oxazolidinone Thioglycosides
In 2001, Kerns and coworkers reported a protocol using 2,3-trans-oxazolidinone as an effective protective group with high α-selectivity toward glycosylation. A notable feature of this class of α-selective D-glucosaminyl donors is that the oxazolidinone group affords simultaneous protection of the 2-N and 3-O positions [77]. Later, the same group expanded this methodology to the formation of α-linked serine and threonine derivatives [78]. No matter what the protective groups on the sugar were, the desired products could be obtained in high yield (78-85%) with exclusive α-selectivity (Scheme 25), however, when the similar reaction was performed with fused-ring 2,3-oxa-Scheme 24. Ring-fused oxazolidinone for the stereoselective formation of α-linked glycoside.
Later, the same group expanded this methodology to the formation of α-linked serine and threonine derivatives [78]. No matter what the protective groups on the sugar were, the desired products could be obtained in high yield (78-85%) with exclusive αselectivity (Scheme 25), however, when the similar reaction was performed with fused-ring 2,3-oxazolidinone galactosamine donor, poor stereoselectivity was observed. Even the glycosylated production was isolated in high yield. The same poor selectivity was also observed with other 2-azido-2-deoxy-D-galactosyl donors (sulfoxide, thiophenyl, and selenoglycoside donors) in the previous reports [79][80][81].
Later, the same group expanded this methodology to the formation of α-linked serine and threonine derivatives [78]. No matter what the protective groups on the sugar were, the desired products could be obtained in high yield (78-85%) with exclusive α-selectivity (Scheme 25), however, when the similar reaction was performed with fused-ring 2,3-oxazolidinone galactosamine donor, poor stereoselectivity was observed. Even the glycosylated production was isolated in high yield. The same poor selectivity was also observed with other 2-azido-2-deoxy-D-galactosyl donors (sulfoxide, thiophenyl, and selenoglycoside donors) in the previous reports [79][80][81]. In 2004, Kerns's group modified the oxazolidinone group to its N-acetyl analogues, leading to switch the reaction from α-selectivity to β-selectivity [82]. In Kerns's case, stereoselective formation of α-linked or β-linked glycoside was supposed to be dependent on the reactivity and steric hindrance of acceptors. Due to the steric hindrance from the acetyl group on N-atom, and the equilibrium between α-triflate (a major product but with Scheme 25. The formation of α-linked serine and threonine derivatives. In 2004, Kerns's group modified the oxazolidinone group to its N-acetyl analogues, leading to switch the reaction from α-selectivity to β-selectivity [82]. In Kerns's case, stereoselective formation of α-linked or β-linked glycoside was supposed to be dependent on the reactivity and steric hindrance of acceptors. Due to the steric hindrance from the acetyl group on N-atom, and the equilibrium between α-triflate (a major product but with less reactivity) and β-triflate (a minor product but with high reactivity), the more-reactive and less-steric-hindered alcohol reacted rapidly with α-triflate to produce β-linked glycoside; conversely, less-reactive alcohols with steric hindrance could react with minority but more-reactive β-triflate to obtain α-linked glycoside with extended reaction time (Scheme 26).
In 2005, Stefan Oscarson and co-workers found that when N-acetylated 2N,3O-oxazolidinone thioglycoside was activated with NIS and AgOTf in the presence of glycosyl acceptors, the corresponding glycosylation products could be formed in excellent yield [83]. Interestingly, the stereochemistry of the anomeric carbon could be switched by controlling the amount AgOTf. When the catalytic amount of AgOTf was used, the reaction delivered the products in exclusive β-selectivity (Scheme 27, condition A), while 0.4 equiv. of AgOTf resulted in forming the products in total α-selectivity (Scheme 25, condition B). A mechanism study has shown that β-coupling products could take place in situ anomerization to form α-glycosylation products through Ag cation-promoted endocyclic C-O bond cleavage intramolecular cyclization [84]. less reactivity) and β-triflate (a minor product but with high reactivity), the more-reactive and less-steric-hindered alcohol reacted rapidly with α-triflate to produce β-linked glycoside; conversely, less-reactive alcohols with steric hindrance could react with minority but more-reactive β-triflate to obtain α-linked glycoside with extended reaction time (Scheme 26). In 2005, Stefan Oscarson and co-workers found that when N-acetylated 2N,3O-oxazolidinone thioglycoside was activated with NIS and AgOTf in the presence of glycosyl acceptors, the corresponding glycosylation products could be formed in excellent yield [83]. Interestingly, the stereochemistry of the anomeric carbon could be switched by controlling the amount AgOTf. When the catalytic amount of AgOTf was used, the reaction delivered the products in exclusive β-selectivity (Scheme 27, condition A), while 0.4 equiv. of AgOTf resulted in forming the products in total α-selectivity (Scheme 25, condition B). A mechanism study has shown that β-coupling products could take place in situ anomerization to form α-glycosylation products through Ag cation-promoted endocyclic C-O bond cleavage intramolecular cyclization [84]. In 2008, Ito and co-workers reported a stereoselective α-glycosylation procedure using N-benzyl-2N,3O-oxazolidinone thioglycoside as a highly α-directing glycosyl donor [85]. It was found that the glycosylation reaction between thioglycoside donors 223-224 and primary alcohols produced the products in lower selectivity, however, the stereoselectivity could be improved by using toluene and 1,4-dioxane as mixture solvent, meanwhile, a similar reaction with less-reactive secondary alcohol could produce the products In 2008, Ito and co-workers reported a stereoselective α-glycosylation procedure using N-benzyl-2N,3O-oxazolidinone thioglycoside as a highly α-directing glycosyl donor [85].
It was found that the glycosylation reaction between thioglycoside donors 223-224 and primary alcohols produced the products in lower selectivity, however, the stereoselectivity could be improved by using toluene and 1,4-dioxane as mixture solvent, meanwhile, a similar reaction with less-reactive secondary alcohol could produce the products in high selectivity regardless of solvents (Scheme 28). The key features of this reaction included: (i) the reaction was performed at room temperature and could be run on large scale; and (ii) the procedure could be used for the preparation of trisaccharide with two 1,2-cis-glycosidic bonds by one-pot two-step sequential glycosylation reaction (Scheme 29), which rendered it possible to use in polymer-supported or automated oligosaccharide synthesis. Although great advances in this field have been achieved by use of 2,3-trans-oxazolidinone as a non-participating group for glucosamine donors, several drawbacks, such as Although great advances in this field have been achieved by use of 2,3-trans-oxazolidinone as a non-participating group for glucosamine donors, several drawbacks, such as the undesired glycosylation and sulfenylation of the nitrogen atom [77,78], reduction of α-selectivity [82], low yields [85], and limited scope of acceptors [84,85], still existed. To tackle these problems, a new efficient strategy based on pre-activation protocol for both α-and β-stereoselective glycosylation of glucosamine donors was developed by Ye's group [86]. In this case, a glycosyl donor such as 4,6-di-O-acetyl-N-acetyl-oxazolidinone protected donor 237, was completely activated and consumed first, then a glycosyl acceptor was added. By using this strategy, either excellent αstereoselective or β-stereoselective glycosylation could be achieved in the absence of any base or by the addition of a hindered base, such as 2,4,6-tri-tert-butylpyrimidine (TTBP) [87] (Schemes 30 and 31). Compared with the traditional glycosylation protocols, controllable glycosylation could be realized by simply adding a steric-hindered base or without addition of any base under pre-activation procedure, which facilitated the construction of complex oligosaccharides with important biological functions. The same group also discovered that the protected group in the oxazolidinone donors could dramatically influence the stereoselectivities of the pre-activation procedure [88]. Although great advances in this field have been achieved by use of 2,3-trans-oxazolidinone as a non-participating group for glucosamine donors, several drawbacks, such as the undesired glycosylation and sulfenylation of the nitrogen atom [77,78], reduction of α-selectivity [82], low yields [85], and limited scope of acceptors [84,85], still existed. To tackle these problems, a new efficient strategy based on pre-activation protocol for both α-and β-stereoselective glycosylation of glucosamine donors was developed by Ye's group [86]. In this case, a glycosyl donor such as 4,6-di-O-acetyl-N-acetyl-oxazolidinone protected donor 237, was completely activated and consumed first, then a glycosyl acceptor was added. By using this strategy, either excellent α-stereoselective or β-stereoselective glycosylation could be achieved in the absence of any base or by the addition of a hindered base, such as 2,4,6-tri-tert-butylpyrimidine (TTBP) [87] (Schemes 30 and 31). Compared with the traditional glycosylation protocols, controllable glycosylation could be realized by simply adding a steric-hindered base or without addition of any base under pre-activation procedure, which facilitated the construction of complex oligosaccharides with important biological functions. The same group also discovered that the protected group in the oxazolidinone donors could dramatically influence the stereoselectivities of the preactivation procedure [88]. The Nguyen's group focused on the development of novel, scalable approaches to avoid the issues associated with the formation of 1,2-cis-2-aminoglycoside by using cationic nickel (II) catalysis [89,90]. Trihaloacetimidate (TCA) is viewed as a good "leaving group" in carbohydrate chemistry whose nitrogen can be coordinated with metals to play a role as a "directing group" in transition-metal-catalyzed reactions. Additionally, the benzylidene at C(2)-amino group of sugars serves not only as a "protecting group" but also as a "directing group" in transition-metal-catalyzed reactions [91]. In 2009, The Nguyen's group reported a nickel-catalyzed stereoselective glycosylation with C(2)-Npara-(methoxy)benzylideneamino donor 254 for the formation of 1,2-cis-2-aminoglycoside (Scheme 32) [89]. The thorough adjustment of the reaction conditions revealed that the catalyst, electronic properties of the ligands, and the substituted group on the benzylidenamino protecting-group all played crucial roles in regulating the stereoselectivity and speeding up the reaction.

The Development of Nickel-Catalyzed Stereoselective Glycosylation
The Nguyen's group focused on the development of novel, scalable approaches to avoid the issues associated with the formation of 1,2-cis-2-aminoglycoside by using cationic nickel (II) catalysis [89,90]. Trihaloacetimidate (TCA) is viewed as a good "leaving group" in carbohydrate chemistry whose nitrogen can be coordinated with metals to play a role as a "directing group" in transition-metal-catalyzed reactions. Additionally, the benzylidene at C(2)-amino group of sugars serves not only as a "protecting group" but also as a "directing group" in transition-metal-catalyzed reactions [91]. In 2009, The Nguyen's group reported a nickel-catalyzed stereoselective glycosylation with C(2)-N-para-(methoxy)benzylideneamino donor 254 for the formation of 1,2-cis-2-aminoglycoside (Scheme 32) [89]. The thorough adjustment of the reaction conditions revealed that the catalyst, electronic properties of the ligands, and the substituted group on the benzylidenamino protecting-group all played crucial roles in regulating the stereoselectivity and speeding up the reaction. In order to prove the superiority of the catalytic system in the glycosylation reaction, comparisons against conventional trichloroacetimidate donor-activation protocol and the developed nickel-catalyzed formation of 1,2-cis-aminosugar was performed. Classical Lewis acid activator of TCA donor, such as TMSOTf and BF3·OEt2, either resulted in the decomposition of the glycosyl donor or led to the formation of a trace amount of the desired disaccharide 261 [92] (Scheme 33). Compared to these established glycosylation methods [88,92], the coupling of 260 with 14 proceeded smoothly to provide disaccharide 261 in 77% yield with excellent α-selectivity. In order to prove the superiority of the catalytic system in the glycosylation reaction, comparisons against conventional trichloroacetimidate donor-activation protocol and the developed nickel-catalyzed formation of 1,2-cis-aminosugar was performed. Classical Lewis acid activator of TCA donor, such as TMSOTf and BF 3 ·OEt 2 , either resulted in the decomposition of the glycosyl donor or led to the formation of a trace amount of the desired disaccharide 261 [92] (Scheme 33). Compared to these established glycosylation methods [88,92], the coupling of 260 with 14 proceeded smoothly to provide disaccharide 261 in 77% yield with excellent α-selectivity. Scheme 32. Nickel-catalyzed stereoselective formation of 1,2-cis-aminoglycoside.
In order to prove the superiority of the catalytic system in the glycosylation reaction, comparisons against conventional trichloroacetimidate donor-activation protocol and the developed nickel-catalyzed formation of 1,2-cis-aminosugar was performed. Classical Lewis acid activator of TCA donor, such as TMSOTf and BF3·OEt2, either resulted in the decomposition of the glycosyl donor or led to the formation of a trace amount of the desired disaccharide 261 [92] (Scheme 33). Compared to these established glycosylation methods [88,92], the coupling of 260 with 14 proceeded smoothly to provide disaccharide 261 in 77% yield with excellent α-selectivity.

Scheme 33. Comparison between Lewis acids and nickel catalyst.
Glucosamine donors 260, 262-264 were then reacted with a broad range of glycosyl acceptors, such as primary, secondary, and tertiary alcohols (Scheme 34). All the glycosylations with cationic nickel(II) catalyst produced the desired products 261, 270-282 in high yield with excellent α-selectivity. The donor with steric hinderance, such as tertiary adamantol, also produced the desired product 282 in 96% yield and with α-selectivity (α:β = Scheme 33. Comparison between Lewis acids and nickel catalyst. Glucosamine donors 260, 262-264 were then reacted with a broad range of glycosyl acceptors, such as primary, secondary, and tertiary alcohols (Scheme 34). All the glycosylations with cationic nickel(II) catalyst produced the desired products 261, 270-282 in high yield with excellent α-selectivity. The donor with steric hinderance, such as tertiary adamantol, also produced the desired product 282 in 96% yield and with α-selectivity (α:β = 17:1). Dihydrocholesterol 266, a secondary alcohol acceptor, was employed to couple with the donor 260 to produce glycoconjugate 275 in 85% yield with α-selectivity (α:β = 11:1). Increased yield and α-selectivity are observed when glucosamine donors are combined with acceptor 260 when the para-substituent on the benzylidene group is simply changed from p-OMe-benzylidene in 277 (α:β = 6:1) into p-CF 3 -benzylidene (α only). Meanwhile, it was found that electron-withdrawing C(2)-N-para-fluoro-and trifluoromethyl-benzylidenamino donors 263 and 264 are more stable compared to the methoxy donor 260 for purification [62], which would facilitate it for preparation on a large scale.
Additionally, galactosamine trichloroacetimidate 283 was also employed as a donor to couple with both primary and secondary alcohol acceptors (Scheme 35). Once again, these results showed the cationic nickel (II) catalyst in the glycosylation reaction could produce 1,2-cis-2-aminoglycosides in high yield with outstanding α-selectivity regardless of the kind of the acceptors.

Scheme 35. α-Selective coupling with D-galactosamine trichloroacetimidate.
Due to the lower reactivity, using D-glucuronic acid derivatives as acceptors to couple with donors in a glycosylation reaction generally gave the products in low yield and poor selectivity. However, the nickel-catalyzed strategy might be able to partially resolve this issue [93,94] (Scheme 37). For instance, coupling acceptors 303-305 with donors 260, 262-264 proceeded smoothly to provide the desired disaccharides in moderate-to-good yields (55-87%) with good α-selectivity (α:β = 9:1 to α only). Among different benzylidenamino protected donors (260, 262-264), the one with an electron-withdrawing group, such as para-(trifluoromethyl)benzyl (264), was found to be the most efficient. Glycosylation of the methyl ester of D-glucuronic acid acceptor 305 with donor 264 produced the desired disaccharide 311 in 87% yield with exclusive α-selectivity.
Glycosylphosphatidylinositol (GPI) anchors are a large family of glycolipids that play significant roles in various biological processes, including signal transduction, prion disease pathogenesis, and immune response [95][96][97][98][99]. Total synthesis of GPI anchors is challenging due to structural complexity. The available methods of construction of the pseudodisaccharide unit of D-glucosamine and inositol components had several limitations, such as poor α-selectivity and long reaction time [100][101][102][103][104]. Due to the lower reactivity, using D-glucuronic acid derivatives as acceptors to couple with donors in a glycosylation reaction generally gave the products in low yield and poor selectivity. However, the nickel-catalyzed strategy might be able to partially resolve this issue [93,94] (Scheme 37). For instance, coupling acceptors 303-305 with donors 260, 262-264 proceeded smoothly to provide the desired disaccharides in moderate-to-good yields (55-87%) with good α-selectivity (α:β = 9:1 to α only). Among different benzylidenamino protected donors (260, 262-264), the one with an electron-withdrawing group, such as para-(trifluoromethyl)benzyl (264), was found to be the most efficient. Glycosylation of the methyl ester of D-glucuronic acid acceptor 305 with donor 264 produced the desired disaccharide 311 in 87% yield with exclusive α-selectivity.

Scheme 37. α-Selective glycosylation of glucuronic acid acceptors.
Glycosylphosphatidylinositol (GPI) anchors are a large family of glycolipids that play significant roles in various biological processes, including signal transduction, prion disease pathogenesis, and immune response [95][96][97][98][99]. Total synthesis of GPI anchors is challenging due to structural complexity. The available methods of construction of the pseudodisaccharide unit of D-glucosamine and inositol components had several limitations, such as poor α-selectivity and long reaction time [100][101][102][103][104]. Coupling inositol nucleophile 312 as the acceptor with C(2)-N-substituted benzylidene D-glucosamine TCA donors 260, 262-264 produced desired products 313-315 in excellent α-selectivity (Scheme 38), addressing the issue of poor α-selectivity in the literature. A mechanism for explaining α-selectivity of the glycosylation with cationic nickel (II) catalyst was shown in Figure 4. In pathway A, the formation of seven-member ring intermediate 318 via reversible coordination of C(2)-benzylidene nitrogen and the nitrogen of the C(1)-TCA donor 317 by nickel catalyst facilitated the 1,2-cis-glycoside. Alternatively, the cationic nickel catalyst could act as a mild Lewis acid (pathway B) to activate C(1)-TCA donor 317, promoting ionization to generate the oxocarbenium intermediate 321.

Scheme 38. α-Selective glycosylation of inositol acceptor.
A mechanism for explaining α-selectivity of the glycosylation with cationic nickel (II) catalyst was shown in Figure 4. In pathway A, the formation of seven-member ring intermediate 318 via reversible coordination of C(2)-benzylidene nitrogen and the nitrogen of the C(1)-TCA donor 317 by nickel catalyst facilitated the 1,2-cis-glycoside. Alternatively, the cationic nickel catalyst could act as a mild Lewis acid (pathway B) to activate C(1)-TCA donor 317, promoting ionization to generate the oxocarbenium intermediate 321.

Scheme 38. α-Selective glycosylation of inositol acceptor.
A mechanism for explaining α-selectivity of the glycosylation with cationic nickel (II) catalyst was shown in Figure 4. In pathway A, the formation of seven-member ring intermediate 318 via reversible coordination of C(2)-benzylidene nitrogen and the nitrogen of the C(1)-TCA donor 317 by nickel catalyst facilitated the 1,2-cis-glycoside. Alternatively, the cationic nickel catalyst could act as a mild Lewis acid (pathway B) to activate C(1)-TCA donor 317, promoting ionization to generate the oxocarbenium intermediate 321.
Subsequently, the formation of five-membered cyclic intermediate 323 was followed by ligand exchange 322, providing the 1,2-cis-glycoside 325. The potential of the nickel-catalyzed glycosylation strategy was further demonstrated by extending it to couple with thioglycoside acceptors. A key feature for synthesis The potential of the nickel-catalyzed glycosylation strategy was further demonstrated by extending it to couple with thioglycoside acceptors. A key feature for synthesis of thioglycoside was that it was easily synthesized and generally stable under a variety of glycosylation conditions, making it compatible with other glycosylation strategies, which, in turn, facilitated it for the multistep synthesis of oligosaccharides and glycoconjugates. However, using thioglycoside as glycosyl acceptor under nickel catalyst is much more of a challenge because: (i) intermolecular aglycon transfer of the anomeric sulfide to the trichloroacetimidate donor could result in undesired thioglycoside [105][106][107][108][109][110][111][112][113][114][115][116][117][118][119][120]; and (ii) sulfur could coordinate to the cationic nickel catalyst leading to its inactivation. However, after a series of trials, the desired disaccharide products were obtained in good yields with excellent α-selectivity, and only a minor mount of undesired thioglycoside was detected. The key for the success of this transformation was using N-phenyl trifluoroacetimidate (PTFA) as donors due to their attenuated activity compared to trichloroacetimidates [121][122][123]. The scope of substrates was then expanded to the glycosylation of a variety of thioglycosides with both armed and disarmed donors [124] (Scheme 39). Three L-rhamnose acceptors equipped with different anomeric sulfides (329-331) were coupled with corresponding donors (324 and 325). Coupling of armed donor 324 with phenyl thioglycoside 141 generates disaccharide 332 in 70% yield with no undesired thioglycoside by-product was detected, albeit in lower α-selectivity (α:β = 6:1). Similarly, coupling of phenyl D-galactosyl thioglycoside 327 with both donors 325 and 326 (entries 2 and 3) provids disaccharides 333 and 334 in exclusive 55% and 61% yield with exclusive α-selectivity but a large quantity of sulfide transfer products 341 and 342 were detected (8 vs. 24%), respectively. The donor 325 then reacted with D-galactosyl thioglycoside 328. No trace amount of transfer product 341 was observed, and exclusive α-disaccharide 335 was isolated in 71% yield. The desired disaccharides 336 and 337 were obtained in 81% and 72% yield with exclusive α-selectivity, also aglycon transfer products 341 and 342 were detected in 15% and 9% yield, respectively. Compared to the acceptor 330, the coupling of bulky 2,6-dimethylphenyl (DMP) [125] acceptor 331 with donor 325 produced the desired product 340 in a slight lower yield (98% vs. 74%) with the α-only selectivity maintained but no undesired thioglycoside by-product was detected. Additionally, 2-trifluoromethylphenyl thioglycoside acceptor 330 and galactosamine acceptors 327 and 328 were also investigated, providing α-disaccharides, and aglycon transfer products were not observed. Galactosamine donors were also employed to couple with corresponding acceptors, but the desired products were obtained in lower yield and minor amounts of undesired thioglycosides were observed in some reactions. To compare, the author also performed nickel-catalyzed glycosylation using C(2)azido of N-phenyltrifluoroacetimidate as the glycosyl donor. It was found that the yield of disaccharide products decreased dramatically but α-only selectivity was maintained (Scheme 40). In the same case, such as with 141 as glycosyl acceptor, the major product of the reaction was thioglycoside by-product derived from the intermolecular aglycon transfer of the anomeric sulfide to N-phenyltrifluoroacetimidate. To compare, the author also performed nickel-catalyzed glycosylation using C(2)azido of N-phenyltrifluoroacetimidate as the glycosyl donor. It was found that the yield of disaccharide products decreased dramatically but α-only selectivity was maintained (Scheme 40). In the same case, such as with 141 as glycosyl acceptor, the major product of the reaction was thioglycoside by-product derived from the intermolecular aglycon transfer of the anomeric sulfide to N-phenyltrifluoroacetimidate. To further explore the utility of this method, the sequential formation of trisaccharides was performed (Scheme 41). Coupling disaccharide thioglycoside 338 obtained from nickelcatalyzed glycosylation with acceptor 132 under NIS-AgOTf activation condition at 35 • C produced trisaccharide adduct 349 in 84% yield with the a slight α-selectivity (α:β = 3:2). Similarly, trisaccharide 351 was obtained by coupling disaccharide thioglycoside 350 with acceptor 132 under the same conditions. Subsequent deprotection of imine by treatment of 351 with HCl, acetone/DCM, followed by acetylation in the presence of acetic anhydride and pyridine provided corresponding product 352 in quantitative yield.

Synthesis of Biologically Active Glycans via Nickel-Catalyzed Glycosylation
Nguyen's group further extended nickel-catalyzed glycosylation to synthesis of several biologically relevant glycans and glycoconjugates including mycothiol, precursors of GPI anchor and TN antigen.

Synthesis of Biologically Active Glycans via Nickel-Catalyzed Glycosylation
Nguyen's group further extended nickel-catalyzed glycosylation to synthesis of several biologically relevant glycans and glycoconjugates including mycothiol, precursors of GPI anchor and T N antigen.

Synthesis of GPI Anchor
The GPI anchor, a glycophospholipid, is post-translationally tethered to a protein for attachment to a cell membrane [135]. These anchored proteins play a pivotal role in biochemical processes, such as signal transduction, immune response, and cancer metastasis [136,137]. All GPI structures share the same basic core, including a phosphoethanolamine Scheme 43. The formation of mycothiol.

Synthesis of GPI Anchor
The GPI anchor, a glycophospholipid, is post-translationally tethered to a protein for attachment to a cell membrane [136]. These anchored proteins play a pivotal role in biochemical processes, such as signal transduction, immune response, and cancer metastasis [137,138]. All GPI structures share the same basic core, including a phosphoethanolamine linker, a tetrasaccharide attached to the C(6)-position of myo-inositol, and phospholipid tail [139]. Despite the syntheses of GPI anchors have been reported extensively, there are several challenges to stereoselectively form the 1,2-cis-glycosidic bond between the glucosamine unit and the myo-inositol component [140][141][142]. The cationic nickel (II) catalysis developed by Nguyen's group is thought to be suitable for the GPI anchor core due to its similarity with mycothiol 353 (Scheme 44) [139]. Coupling C(4)-TES-C(2)-N-benzylidene donor 369 with C(6)-hydroxyl myo-inositol 370 under the standard condition generated the desired pseudodisaccharide 371 in 61% yield with α:β = 11:1. Subsequent removal of the TES group with TBAF generated the disaccharide 372. Finally, glycosylation of 372 with tetrabenzylated D-mannose trichloroacetimidate 373 afforded the desired GPI core pseudotrisaccharide 374 in 56% yield.

Synthesis of GPI Anchor
The GPI anchor, a glycophospholipid, is post-translationally tethered to a protein for attachment to a cell membrane [135]. These anchored proteins play a pivotal role in biochemical processes, such as signal transduction, immune response, and cancer metastasis [136,137]. All GPI structures share the same basic core, including a phosphoethanolamine linker, a tetrasaccharide attached to the C(6)-position of myo-inositol, and phospholipid tail [138]. Despite the syntheses of GPI anchors have been reported extensively, there are several challenges to stereoselectively form the 1,2-cis-glycosidic bond between the glucosamine unit and the myo-inositol component [139][140][141]. The cationic nickel (II) catalysis developed by Nguyen's group is thought to be suitable for the GPI anchor core due to its similarity with mycothiol 353 (Scheme 44) [138]. Coupling C(4)-TES-C(2)-N-benzylidene donor 369 with C(6)-hydroxyl myo-inositol 370 under the standard condition generated the desired pseudodisaccharide 371 in 61% yield with α:β = 11:1. Subsequent removal of the TES group with TBAF generated the disaccharide 372. Finally, glycosylation of 372 with tetrabenzylated D-mannose trichloroacetimidate 373 afforded the desired GPI core pseudotrisaccharide 374 in 56% yield. Scheme 44. The formation of the GPI anchor core. Scheme 44. The formation of the GPI anchor core.

Synthesis of T N Antigen
T N antigen is a core structure in O-linked glycoproteins, which consists of the linkage of an N-acetyl-galactosamine (GalNAc) and the hydroxyl group of serine or threonine via α-glycosidic bond [143]. T N antigen 375 and some other analogs, such as T F antigen 376 and ST N antigen 377 (Figure 6), are widely distributed on cell-surface mucin glycoproteins, which participate in cell-adhesion events associated with cancer metastasis, having potential as a biomarker or as a vaccine against cancer [144][145][146].
Early work employed the C(2)-oxazolidinone and the C(2)-azido as donors to couple with the acceptor, affording a mixture of anomers with the ratio of α:β ranged from 1:1 to 4:1 [78,147]. Large-scale synthesis of T N antigen core in good yield with α-selectivity could be achieved under the condition of cationic nickel (II) catalysis (Scheme 45). Two small-scale (Scheme 45a,b) and scale-up reactions (Scheme 45c) employing donor 355 were performed by using 10 mol% of Ni(4-F-C 6 H 4 -CN) 4 (OTf) 2 as the catalyst [148]. These scale-up reactions without loss in yield and selectivity showed that the cationic nickel (II) catalyst might be suitable for formation of 1,2-cis-aminoglycoside in large-scale industry production.

Synthesis of TN Antigen
TN antigen is a core structure in O-linked glycoproteins, which consists of the linkage of an N-acetyl-galactosamine (GalNAc) and the hydroxyl group of serine or threonine via α-glycosidic bond [142]. TN antigen 375 and some other analogs, such as TF antigen 376 and STN antigen 377 (Figure 6), are widely distributed on cell-surface mucin glycoproteins, which participate in cell-adhesion events associated with cancer metastasis, having potential as a biomarker or as a vaccine against cancer [143][144][145]. Early work employed the C(2)-oxazolidinone and the C(2)-azido as donors to couple with the acceptor, affording a mixture of anomers with the ratio of α:β ranged from 1:1 to 4:1 [78,146]. Large-scale synthesis of TN antigen core in good yield with α-selectivity could be achieved under the condition of cationic nickel (II) catalysis (Scheme 45). Two smallscale (Scheme 45a,b) and scale-up reactions (Scheme 45c) employing donor 355 were performed by using 10 mol% of Ni(4-F-C6H4-CN)4(OTf)2 as the catalyst [147]. These scale-up reactions without loss in yield and selectivity showed that the cationic nickel (II) catalyst might be suitable for formation of 1,2-cis-aminoglycoside in large-scale industry production. Finally, gram-scale synthesis of TN antigen from 379 was performed (Scheme 46). Treatment of 379 with HCl(g) (in situ formed from the reaction between AcCl and MeOH) followed by the addition of Ac2O-generated N-acetylated 380. Full deprotection of 378 with 0.2N NaOH in methanol produced 0.4 g TN antigen 375 in 82% yield. Additionally, the synthesis of TN/TF antigens was also completed by Nguyen's group via the nickel methodology in 2016 [148] and 2019 [149].   Early work employed the C(2)-oxazolidinone and the C(2)-azido as donors to couple with the acceptor, affording a mixture of anomers with the ratio of α:β ranged from 1:1 to 4:1 [78,146]. Large-scale synthesis of TN antigen core in good yield with α-selectivity could be achieved under the condition of cationic nickel (II) catalysis (Scheme 45). Two smallscale (Scheme 45a,b) and scale-up reactions (Scheme 45c) employing donor 355 were performed by using 10 mol% of Ni(4-F-C6H4-CN)4(OTf)2 as the catalyst [147]. These scale-up reactions without loss in yield and selectivity showed that the cationic nickel (II) catalyst might be suitable for formation of 1,2-cis-aminoglycoside in large-scale industry production. Finally, gram-scale synthesis of TN antigen from 379 was performed (Scheme 46). Treatment of 379 with HCl(g) (in situ formed from the reaction between AcCl and MeOH) followed by the addition of Ac2O-generated N-acetylated 380. Full deprotection of 378 with 0.2N NaOH in methanol produced 0.4 g TN antigen 375 in 82% yield. Additionally, the synthesis of TN/TF antigens was also completed by Nguyen's group via the nickel methodology in 2016 [148] and 2019 [149]. Finally, gram-scale synthesis of T N antigen from 379 was performed (Scheme 46). Treatment of 379 with HCl(g) (in situ formed from the reaction between AcCl and MeOH) followed by the addition of Ac 2 O-generated N-acetylated 380. Full deprotection of 378 with 0.2N NaOH in methanol produced 0.4 g T N antigen 375 in 82% yield. Additionally, the synthesis of T N /T F antigens was also completed by Nguyen's group via the nickel methodology in 2016 [149] and 2019 [150]. Early work employed the C(2)-oxazolidinone and the C(2)-azido as donors to couple with the acceptor, affording a mixture of anomers with the ratio of α:β ranged from 1:1 to 4:1 [78,146]. Large-scale synthesis of TN antigen core in good yield with α-selectivity could be achieved under the condition of cationic nickel (II) catalysis (Scheme 45). Two smallscale (Scheme 45a,b) and scale-up reactions (Scheme 45c) employing donor 355 were performed by using 10 mol% of Ni(4-F-C6H4-CN)4(OTf)2 as the catalyst [147]. These scale-up reactions without loss in yield and selectivity showed that the cationic nickel (II) catalyst might be suitable for formation of 1,2-cis-aminoglycoside in large-scale industry production. Finally, gram-scale synthesis of TN antigen from 379 was performed (Scheme 46). Treatment of 379 with HCl(g) (in situ formed from the reaction between AcCl and MeOH) followed by the addition of Ac2O-generated N-acetylated 380. Full deprotection of 378 with 0.2N NaOH in methanol produced 0.4 g TN antigen 375 in 82% yield. Additionally, the synthesis of TN/TF antigens was also completed by Nguyen's group via the nickel methodology in 2016 [148] and 2019 [149].

Glycosylation of 3-Amino-3-deoxysugars and 4-Amino-4-deoxysugars
Over the past few years, Wan's group has been working on the development of novel and effective methodologies for stereoselective formation of 3-aminoglycoside and applying the methods developed by them for the synthesis of natural products containing 3-amino-3-deoxysugars [151,152]. Inspired by their earlier report where an intramolecular hydrogen bond (HB) between a C3-nosyl group and an α-hydroxyl group on the anomeric carbon was observed in X-ray crystallography [153], they reported an HB-assisted protocol for β-selective glycosylation of rare 3-amino sugars (Scheme 47) [154]. By employment of 381 as the glycosyl donor and phosphine oxide as an exogenous nucleophile, an αoxyphosphonium species 382 was generated, in which there was an intramolecular hydrogen bond between the C3-axial sulfonamide group and the oxygen of the α-aligned phosphine oxide. Oxyphosphonium species 382 blocked α-face leading to nucleophiles attacked to 382 from β-face, which, in turn, resulted in β-selective glycosylation. carbon was observed in X-ray crystallography [152], they reported an HB-assisted protocol for β-selective glycosylation of rare 3-amino sugars (Scheme 47) [153]. By employment of 381 as the glycosyl donor and phosphine oxide as an exogenous nucleophile, an α-oxyphosphonium species 382 was generated, in which there was an intramolecular hydrogen bond between the C3-axial sulfonamide group and the oxygen of the α-aligned phosphine oxide. Oxyphosphonium species 382 blocked α-face leading to nucleophiles attacked to 382 from β-face, which, in turn, resulted in β-selective glycosylation.

Scheme 47. H-bond-stabilized oxyphosphonium species for β-glycosylation.
This H-bond-assisted glycosylation reaction was found to have broad application in the construction of β-glycosidic bonds (Scheme 48). For example, coupling of donor 363ad with bulky primary alcohol offered glycosylation products in better selectivity than that with the less sterically-hindered and electron-rich primary alcohols (366a and 366c) due to the latter one being a good H-bond acceptor to compete with the phosphine oxide additive. Secondary alcohols generally gave the products with β-selectivity. Additionally, this strategy was also extended to modification of natural products and drugs in good yield with satisfactory-to-excellent β-selectivity (391g-391i). This strategy was also suitable for the one-pot formation of trisaccharide such as 394 with Ph3PAuBAr4 F as the only catalyst. The glycosylation reaction between donor 392 and acceptor 393 occurred selectively on the 6-OH of 392 with α-selectivity in the presence of Ph3PAuBAr4 F catalyst. Then, 384a and 390 were added to the resulting mixture, furnishing the 394 with exclusive βselectivity in 52% yield over two steps; impressively, the whole process used Ph3PAuBAr4 F as the only catalyst for the two glycosylation reactions.

Scheme 47. H-bond-stabilized oxyphosphonium species for β-glycosylation.
This H-bond-assisted glycosylation reaction was found to have broad application in the construction of β-glycosidic bonds (Scheme 48). For example, coupling of donor 363a-d with bulky primary alcohol offered glycosylation products in better selectivity than that with the less sterically-hindered and electron-rich primary alcohols (366a and 366c) due to the latter one being a good H-bond acceptor to compete with the phosphine oxide additive. Secondary alcohols generally gave the products with β-selectivity. Additionally, this strategy was also extended to modification of natural products and drugs in good yield with satisfactory-to-excellent β-selectivity (391g-391i). This strategy was also suitable for the one-pot formation of trisaccharide such as 394 with Ph 3 PAuBAr 4 F as the only catalyst. The glycosylation reaction between donor 392 and acceptor 393 occurred selectively on the 6-OH of 392 with α-selectivity in the presence of Ph 3 PAuBAr 4 F catalyst. Then, 384a and 390 were added to the resulting mixture, furnishing the 394 with exclusive β-selectivity in 52% yield over two steps; impressively, the whole process used Ph 3 PAuBAr 4 F as the only catalyst for the two glycosylation reactions.
The methodology was further applied in the synthesis of trisaccharide fragment 399 of avidinorubicin. Avidinorubicin is an antibiotic isolated from the cultured broth of Streptomyces avidinii in 1991, it exhibited platelet-aggregation-inhibitory activity [155]. The synthesis of avidinorubicin is challenging due to its complex structure. Coupling 3-ADS 384d with 394 produced the desired disaccharide 396 in 66% yield with exclusive β-selectivity. Subsequent removal of the acetyl group generated 397 in 93% yield. Finally, glycosylation between 2,6-dideoxy sugar 398 and disaccharide 397 under Yu's glycosylation conditions proceeded smoothly, generating the protected trisaccharide unit of avidinorubicin 399 in 86% yield with exclusive α-selectivity (Scheme 49).

Scheme 52. The synthesis of β-aminoglycoside.
This methodology was also suitable for glycosylation of 4-amino-4-deoxysugars [116]. Coupling of acosamine derivatives 418 and 419 with MTHP electrophile 405 and 406 under Condition A generated desired disaccharides 420-423 in 60−93% yield with α-selectivity (Scheme 53). While disaccharide 424 was obtained in 64% yield with β-selectivity (β:α > 50:1) under Condition B (see Scheme 53 for detail). Under Condition B, disaccharide 426 was obtained in 52% yield with β-selectivity (β:α > 50:1) following addition of secondary electrophile 408, and derivative 419 was transformed into disaccharide 424 in 56% yield with β-selectivity (β:α > 2.6:1). Sialic acid (Neu5Ac) is an acidic nine carbon saccharide, which usually plays an essential role in terminating the carbohydrate portions of glycoproteins and glycolipids of mammals in the form of an α-glycoside. Silalylation via chemical methods with high α-stereoselectivity is challenging. For instance, the anomeric selectivity at the newly formed glycosidic bond is unpredictable due to the absence of a C3 participating group. On the other hand, the presence of Scheme 53. The synthesis of 4-aminoglycoside.
Sialic acid (Neu5Ac) is an acidic nine carbon saccharide, which usually plays an essential role in terminating the carbohydrate portions of glycoproteins and glycolipids of mammals in the form of an α-glycoside. Silalylation via chemical methods with high α-stereoselectivity is challenging. For instance, the anomeric selectivity at the newly formed glycosidic bond is unpredictable due to the absence of a C3 participating group. On the other hand, the presence of the C1 electron-withdrawing carboxylic group could reduce the reactivity of sialic acid donor, resulting in a poor yield and formation of a significant amount of undesirable 2,3-glycal side-product. In view of the above challenges, Yang's group developed an efficient α-sialylation methodology for various primary, secondary, and tertiary alcohol acceptors by using C4-Pico or 4-nitropicoloyl as directing group (Scheme 54) [166]. Sialic acid (Neu5Ac) is an acidic nine carbon saccharide, which usually plays an essential role in terminating the carbohydrate portions of glycoproteins and glycolipids of mammals in the form of an α-glycoside. Silalylation via chemical methods with high α-stereoselectivity is challenging. For instance, the anomeric selectivity at the newly formed glycosidic bond is unpredictable due to the absence of a C3 participating group. On the other hand, the presence of the C1 electron-withdrawing carboxylic group could reduce the reactivity of sialic acid donor, resulting in a poor yield and formation of a significant amount of undesirable 2,3-glycal sideproduct. In view of the above challenges, Yang's group developed an efficient α-sialylation methodology for various primary, secondary, and tertiary alcohol acceptors by using C4-Pico or 4-nitropicoloyl as directing group (Scheme 54) [165]. As for the scope of substrates, it was found that the combination of C4 Pico or 4nitropicoloyl groups with the DCM/CH 3 CN [167,168] mixture solvent system is more effective and selective for α-glycosylation ( Figure 7). Additionally, glycosylation of donor 432 with various acceptors all generated the desired products with satisfactory α-selectivity. As for the scope of substrates, it was found that the combination of C4 Pico or 4nitropicoloyl groups with the DCM/CH3CN [166,167] mixture solvent system is more effective and selective for α-glycosylation ( Figure 7). Additionally, glycosylation of donor 432 with various acceptors all generated the desired products with satisfactory α-selectivity. In order to further explore this methodology, the synthesis of sialylated trisaccharide 446 was examined (Scheme 55). Coupling donor 432 with acceptor 441 provided sialylated disaccharide 442 in 52% yield with exclusive α-selectivity. Then, removal of the two 2naphthylmethyl (Nap) protecting groups in 442 in the presence of 2,3-dichloro-5,6-dicyanobenzo-1,4-quinone (DDQ) gave the desired disaccharide triol 443 in 91% yield, which In order to further explore this methodology, the synthesis of sialylated trisaccharide 446 was examined (Scheme 55). Coupling donor 432 with acceptor 441 provided sialylated disaccharide 442 in 52% yield with exclusive α-selectivity. Then, removal of the two 2-naphthylmethyl (Nap) protecting groups in 442 in the presence of 2,3-dichloro-5,6dicyanobenzo-1,4-quinone (DDQ) gave the desired disaccharide triol 443 in 91% yield, which was followed by acylation with benzoyl chloride (BzCl) to produce 444 in 85% yield. Subsequently, deprotection of 1-O-p-methoxyphenyl group in 444 with ceric ammonium nitrate (CAN) followed by imidation with trichloroacetonitrile (CCl 3 CN) produced trichloroacetimidate 445 in 51% over two steps as inseparable α:β (1:1) isomers. Finally, coupling trichloroacetimidate 445 with monosaccharide alcohol 446 [169] under the activation of trimethylsilyl trifluoromethanesulfonate (TMSOTf) in DCM at −78 • C, generated the target trisaccharide 447 in 51% yield with exclusive α-selectivity. Additionally, similar strategies of glycosylation with various picoloylated sialyl donors were reported by Cristina De Meo's group [169,170] and other strategies are also reported [171−174]. There are also several elegant reviews about glycosylation of sialic acid [175−177]; we will not include these herein.

Conclusions
Amino sugars exist widely in nature and have a range of biological activities. The fine distinctions in their glycan structures may lead to the difference of the activities and, thus, can be exploited for the development of drugs and vaccines. The new methodologies and strategies for stereoselective aminoglycoside construction and convergent oligosaccharide assembly are critical for the development of carbohydrate chemistry. Many new leaving groups and catalyst systems have been developed with the goal to develop a versatile method for stereoselective glycosylation of amino sugars.
In this review, we summarized the advances of the stereoselective glycosylation protocols of amino sugars. A number of glycosylation strategies under mild conditions with potential applications mediated or catalyzed by transition-metal and (or) organic catalysts have been developed for stereoselective glycosylation of amino sugars. These methodologies render the glycosylation reaction more effective and extend the scope of donors and acceptors with a broad range of protecting groups, being no longer only dependent on neighboring group participation or anomeric effect. However, the results of these glycosylation methods are also affected by other factors, such as solvent effect, remote participation, steric hindrance, and conformational constraints, as well. After generous traditional approaches for the development of the synthesis of aminoglycosides, glycosidation of amino sugars via glycosyl radical intermediates has been a promising method and powerful tool, owing to its inherent features, including mild reaction conditions, broad functional groups tolerance, and being free of side reactions such as elimination and/or epimerization. Notably, the stereoselective formation of 1,2-cis-aminoglycosides via radical activation of glycosyl precursors is much less developed compared to conventional ionic activation strategies. In this context, it remains highly desirable to exploit general and user-friendly photocatalyzed or electrocatalyzed glycosylation methods to construct the Additionally, similar strategies of glycosylation with various picoloylated sialyl donors were reported by Cristina De Meo's group [170,171] and other strategies are also reported [172][173][174][175]. There are also several elegant reviews about glycosylation of sialic acid [176][177][178]; we will not include these herein.

Conclusions
Amino sugars exist widely in nature and have a range of biological activities. The fine distinctions in their glycan structures may lead to the difference of the activities and, thus, can be exploited for the development of drugs and vaccines. The new methodologies and strategies for stereoselective aminoglycoside construction and convergent oligosaccharide assembly are critical for the development of carbohydrate chemistry. Many new leaving groups and catalyst systems have been developed with the goal to develop a versatile method for stereoselective glycosylation of amino sugars.
In this review, we summarized the advances of the stereoselective glycosylation protocols of amino sugars. A number of glycosylation strategies under mild conditions with potential applications mediated or catalyzed by transition-metal and (or) organic catalysts have been developed for stereoselective glycosylation of amino sugars. These methodologies render the glycosylation reaction more effective and extend the scope of donors and acceptors with a broad range of protecting groups, being no longer only dependent on neighboring group participation or anomeric effect. However, the results of these glycosylation methods are also affected by other factors, such as solvent effect, remote participation, steric hindrance, and conformational constraints, as well. After generous traditional approaches for the development of the synthesis of aminoglycosides, glycosidation of amino sugars via glycosyl radical intermediates has been a promising method and powerful tool, owing to its inherent features, including mild reaction conditions, broad functional groups tolerance, and being free of side reactions such as elimination and/or epimerization. Notably, the stereoselective formation of 1,2-cis-aminoglycosides via radical activation of glycosyl precursors is much less developed compared to conventional ionic activation strategies. In this context, it remains highly desirable to exploit general and user-friendly photocatalyzed or electrocatalyzed glycosylation methods to construct the challenging 1,2-cis-aminoglycosidic bonds that avoid sensitive reagents, strong (Lewis) acids, precious (metal) catalysts, or labile substrates.

Conflicts of Interest:
The authors declare no conflict of interest.