1 Introduction

Protecting groups are typically unavoidable in synthetic carbohydrate chemistry. However, their use inevitably involves additional synthetic steps, resulting in loss of materials, and the generation of a significant amount of waste; not only chemical waste, but also in terms of cost and time. Thus, there is an increasing need for the development of methods for the synthetic manipulation of carbohydrates that do not involve protecting groups, but yet still furnish the desired products in high yield and with the requisite selectivity.

2 Synthesis of Glycosyl Oxazolines in Water Using DMC

Glycosyl oxazolines have found wide application in carbohydrate chemistry, first in their protected forms as donors for oligosaccharide synthesis [1, 2], and subsequently, and perhaps more importantly as activated donor substrates for the enzymatic synthesis of glycopeptides and glycoproteins [3, 4].

In 2004, Kadokawa et al. [5] came close to developing a useful method for the direct synthesis of glycosyl oxazolines in water when they used a series of carbodiimides to activate the anomeric centre. However, these reactions only afforded the desired products in very low yield (<30%). Even so, the fact that this type of highly selective transformation was possible at all drew the attention of the Glycoscience community.

The dehydrative properties of the reagent, 2-chloro-1,3-dimethylimidazolium chloride 1 (DMC), were originally reported by Isobe and Ishikawa [6, 7]; however these reported transformations were all performed in organic solvents. In 2009, Shoda and co-workers [8] made a significant breakthrough in the field when they reported the use of DMC for the direct synthesis of glycosyl oxazolines from reducing sugars in D2O in high yield (Scheme 1).

Scheme 1
scheme 1

Synthesis of glycosyl oxazolines directly from reducing sugars in water by Shoda et al. [8]

Full size image

Prior to this report, the synthesis of glycosyl oxazolines typically required peracetylated 2-acetamido-2-deoxy sugars to be prepared. These could then be converted to the corresponding oxazolines using Lewis acids, such as ferric(III) chloride, tin(IV) chloride, boron trifluoride, or trimethylsilyl triflate, [9,10,11,12] before the final removal of acetate protecting groups under basic conditions. However, the use of strong Lewis acids can damage glycosidic linkages and, typically in the cases of larger oligosaccharides, may result in the formation complex reaction mixtures and low yields. This report, therefore, represented a significant advance in the production of N-glycan oxazolines, particularly as donor substrates for enzymatic synthesis.

The proposed mechanism [8] of oxazoline formation involves preferential attack of the hemiacetal hydroxyl on DMC; in the case of the β-anomer this yields reactive intermediate 2 (Scheme 2). Intramolecular attack of the 2-acetamido group at the anomeric centre, followed by abstraction of a proton by a suitable base, then affords the oxazoline. The α-imidazolinium intermediate 3 can also be formed by attack of the α-hemiacetal hydroxyl on DMC. However, since 3 cannot directly form an oxazoline it is then probably hydrolysed to regenerate the β-anomer of the free sugar; this β-anomer then follows the reaction pathway via intermediate 2 to the oxazoline product. Although this mechanistic pathway is plausible, Shoda has stated [8] that the intermediacy of a β-glycosyl chloride during the conversion of 3 into the oxazoline product cannot be ruled out.

Scheme 2
scheme 2

Proposed mechanism for sugar oxazoline formation using DMC [8]

Full size image

The use of DMC for the selective activation of reducing sugars in water has subsequently proved pivotal in facilitating the use of glycosyl oxazolines as donors for enzymatic glycosylation reactions catalysed by endo-β-N-acetylglucosaminidases (ENGases), such as those reported by the groups of Fairbanks [13,14,15,16], Wang [17,18,19], and Yamamoto [20,21,22].

Subsequently, Shoda and co-workers [23] reported another method for the direct synthesis of glycosyl oxazolines using the DMC analogue, 2-chloro-1,3-dimethyl-1H-benzimidazol-3-ium chloride 4, (CDMBI) as the dehydrative agent (Fig. 1).

Fig. 1
figure 1

DMC-analogue CDMBI 4, and the urea DMBI 5 formed by its hydrolysis [23]

Full size image

CDMBI was alleged to be superior to DMC, as it is less hygroscopic and easier to handle, although most applications still use the commercially available DMC for oxazoline formation. Shoda also stated that replacing the imidazolidine ring of DMC with the more electron-rich benzimidazole would reduce the reactivity of the chloroformamidinium group toward nucleophilic attack by water, and that additionally it made the reactive glycosyl-imidazolidinium intermediate more stable. It was also postulated that the introduction of the aromatic ring would also reduce the aqueous solubility of the hydrolysed product, namely the urea 5 (1,3-dimethylbenzimidazol-2-one, DMBI) (Fig. 1). Indeed, when the complex bi-antennary N-glycan decasaccharide 6 (derived from a sialyl glycopeptide isolated from egg yolks [24]) was converted to the corresponding oxazoline 7 using CDMBI 4, DMBI 5 precipitated from the mixture as the reaction proceeded, and was easily removed by filtration (Scheme 3). Moreover, the filtrate containing the oxazoline could then be directly used for a subsequent enzymatic glycosylation of the acceptor p-nitrophenyl 2-acetamido-2-deoxy-β-D-glucopyranoside 8 using the glycosynthase, Endo-M N175Q [21], without the need for further purification or isolation of the oxazoline intermediate.

Scheme 3
scheme 3

Enzymatic glycosylation of GlcNAc-β-pNP with the sialoglycan-oxazoline without isolation of the intermediate oxazoline [23]. Reaction conditions: i CDMBI 4, Na3PO4, H2O; ii 8, Endo-M N175Q

Full size image

3 Synthesis of 1,6-Anhydro Sugars in Water Using DMC

Subsequent to the first report on oxazoline formation, DMC and its derivatives have been employed for the synthesis of a variety of glycosides formed by intercepting the α-imidazolinium intermediate 3 (Scheme 2) with different nucleophiles. In the absence of any external nucleophile, the 6-hydroxyl group may attack the anomeric centre. Thus, Shoda et al. [25] reported the use of DMC for the formation of 1,6-anhydro sugars from unprotected glycopyranoses in water in almost quantitative yield. The DMC procedure mitigated the requirement for the harsh reaction conditions that had previously been used for the synthesis of 1,6-anhydro sugars, such as pyrolysis [26, 27] or thermal degradation [28, 29]. The proposed reaction mechanism is illustrated in Scheme 4, and is similar to that previously suggested [8] for the formation of glycosyl oxazolines. Herein the first step is a nucleophilic attack of the anomeric hydroxyl group of the predominant β-anomer of glucose 9 on DMC 1, giving rise to intermediate 10. Intramolecular attack of the 2-hydroxyl group of 10 at the anomeric carbon affords a 1,2-anhydro intermediate 11, which is subsequently converted to the 1,6-anhydro sugar 12 via intramolecular nucleophilic attack by the 6-hydroxyl group. The α-anomer of glucose 13 also reacts with DMC to give the corresponding α-intermediate 14. Although direct conversion of 14 into the 1,6-anhydro sugar 12 is possible by an attack of the 6-hydroxyl at the anomeric centre, Shoda suggests that in fact 14 is hydrolysed by an attack of water to regenerate β-glucose 9. β-Glucose 9 then follows the above pathway via 11 to give the 1,6-anhydro sugar 12. Evidence for this overall mechanistic pathway, and the requirement for a 1,2-anhydro intermediate such as 11 en route to the 1,6-anhydro sugar, was the fact that the corresponding 1,6-anhydro sugars were not formed when D-mannose, 2-deoxy-D-glucose, and 2-fluoro-2-deoxy-D-glucose were used as substrates.

Scheme 4
scheme 4

Proposed mechanism for 1,6-anhydro sugar formation [25]

Full size image

The published optimised method uses a large excess of triethylamine to drive 1,6-anhydro sugar formation. In cases where other nucleophiles are to be introduced at the anomeric centre, e.g., for glycosyl azide synthesis using DMC (see Sect. 5) [30], fewer equivalents of base are commonly used. However, in these instances, 1,6-anhydro sugar formation can still be an undesired competing side reaction, which reduces the yields of desired products and can be particularly problematic when oligosaccharides, synthesised using multi-step methods and obtained in small quantities, are used. In these cases, 1,6-anhydro sugar formation can typically be minimised by the use of a large excess of the external nucleophile. Alternatively, Shoda has also reported that acetonitrile may be used as a co-solvent to suppress 1,6-anhydro sugar formation [31].

4 Synthesis of Thioglycosides Using DMC

Thioglycosides have found widespread application throughout the carbohydrate field as glycosyl donors for chemical oligosaccharide synthesis [32,33,34,35]. There has also been interest in the synthesis of de-protected thioglycosides as stable analogues of O-glycosides and as potential enzyme inhibitors [36, 37]. The introduction of sulfur at the anomeric centre often involves the reaction of a peracetylated sugar with a thiol in the presence of a Lewis acid [38], or alternatively substitution of a glycosyl halide with a thiolate [39]. In general, these methods require multi-step reaction sequences, involving protection/de-protection strategies, and are usually performed in organic solvents. Although direct methods for the preparation of thioglycosides from hemiacetals have been reported using trifluoroacetic acid as an activator, these reactions gave poor anomeric selectivities and also resulted in the formation of dithioacetal by-products [40, 41].

4.1 Aryl Thioglycosides

In 2009, Tanaka et al. [31] reported a simple method for the direct synthesis of aryl thioglycosides from reducing sugars in an aqueous solvent system using DMC and excess triethylamine (Scheme 5).

Scheme 5
scheme 5

Synthesis of aryl thioglycosides from unprotected sugars [31]. Reaction conditions: i DMC 1, R-SH, Et3N, H2O/MeCN

Full size image

The use of benzenethiol, p-toluenethiol, and 4-methoxybenzenethiol as the thiol component afforded the corresponding aryl thioglycosides as mixtures of anomers, in high yield. However, when 4-nitrobenzenethiol was reacted with D-glucose and DMC, the corresponding thioglycoside product was formed exclusively as the β-anomer, in 90% yield. It was found that the procedure could also be applied to disaccharides, and the corresponding products were formed in quantitative yield, and as exclusively the β-anomer.

4.2 2-Pyridyl 1-Thio-Glycosides

Shoda et al. [31, 42] then demonstrated that it was also possible to glycosylate 2-mercaptopyridine 16 with oligosaccharides having an N-acetylglucosamine unit at the reducing end 15 using DMC to give the corresponding 2-pyridylthioglycosides 17 (Scheme 6).

Scheme 6
scheme 6

Glycosylation of mercaptopyridine 16 with unprotected sugars using DMC reported Shoda et al. [42] Reaction conditions: i DMC 1, Et3N, H2O/MeCN (4:1), 0 °∁, 1 h

Full size image

Interestingly, the glycosyl oxazoline 18 was obtained as a by-product from these reactions. However, it was found that treating oxazoline 18 with 1 M HCl led to the opening of the ring and an increased yield of the desired 2-pyridylthioglycoside 17. A plausible mechanism, which explains the formation of both anomers of the 2-pyridylthioglycoside products as well as the oxazoline, was suggested and is shown in Scheme 7. Nucleophilic attack of the β-hemiacetal of 19 on DMC affords 20, which may then cyclise to give the oxazolinium intermediate 21. Attack on intermediate 21 by mercaptopyridine (PySH) will yield the β-2-pyridylthioglycoside 22. However, if the proton on the nitrogen atom of 21 is abstracted by a base, the intermediate will be converted to the corresponding oxazoline 18, which is itself unreactive towards nucleophilic opening. The addition of acid to oxazoline 18 leads to the re-protonation of the nitrogen to regenerate 21, which can subsequently react with mercaptopyridine to give 22. Alternatively, imidazolidinium intermediate 20 can react directly with mercaptopyridine to give the α-2-pyridylthioglycoside 23.

Scheme 7
scheme 7

Proposed mechanism for 2-pyridylthioglycoside formation using PySH and DMC [42]

Full size image

The α-anomer of the starting material 24 can also react with DMC to give intermediate 25, which can be directly attacked by mercaptopyridine to give the β-2-pyridylthioglycoside 22.

4.3 4-Methyl-7-Thioumbelliferyl-Labelled Glycosides (MUS-Labelled Glycosides)

There is a growing need to develop quantitative ways to label sugars with chromophores. The most common current method of labelling is reductive animation, in which an oligosaccharide is reacted with an amine containing the chromophore, and the ensuing imine is then reduced by a source of hydride [43]. However, sugars labelled in this manner contain an open-chain 1-aminoalditol moiety, essentially rendering the reducing end inert for further useful transformations. This ring opening also reduces the structural relevance of that part of the oligosaccharide, for example with respect to binding or conformational studies. Furthermore, the conversion of 1-aminoalditols back into reducing sugars requires harsh conditions, such as the use of hydrogen peroxide [44, 45]. A method for introducing a detectable chromophore that could also be easily detached would, therefore, be highly advantageous.

In 2013, Shoda et al. [46] reported the direct introduction of 4-methyl-7-thioumbelliferone 26 (MUS) at the anomeric centre of monosaccharides and at the reducing terminus of oligosaccharides using DMC activation (Scheme 8). The fluorescence spectrum of the MUS-labelled glycosides produced indicated that the labelled derivatives showed a high sensitivity for fluorescence detection based on the maximum wavelengths for excitation and emission at 330 and 395 nm, respectively.

Scheme 8
scheme 8

Synthesis of MUS-labelled sugars using DMC by Shoda et al. [46]. Reaction conditions: i DMC 1, Et3N, H2O/MeCN (1:1), 0 °∁, 1.5 h

Full size image

The yields of the MUS-sugars made by this DMC labelling method were sufficiently high to suggest its application in quantitative studies. Indeed, analysis of a mixture of laminari-oligosaccharides of known composition, which were then modified by both MUS-labelling and pyridylamination, showed very similar quantitative values (Table 1). The MUS-glycosides were also easily de-protected by conventional halosuccinimide-mediated thioglycoside activation (e.g. N-bromo, N-chloro-, and N-iodosuccinimide) to re-form the corresponding reducing sugars, which could then be used for further functionalisation; a significant advantage compared to the pyridylamination labelling method.

Table 1 Quantitative analysis of MUS-labelled lamnari-oligosaccharide structures by Shoda et al. [46]
Full size table

4.4 Glycopolymerisation

Although the binding of carbohydrates to proteins is a fundamental process, with widespread importance throughout Biology, it is now appreciated that the majority of carbohydrate–protein interactions are of low affinity. Nature has compensated for this apparent paradox by the use of multivalent receptor-ligand presentation, which amplifies the affinity of single interactions. It is fair to say that all aspects of this so-called ‘glycocluster effect’ [47, 48] are not yet completely understood, although detailed discussions and theories have been presented [49, 50]. Despite shortcomings in our understanding, there have been many reports on the design and production of synthetic glycoclusters, such as glycopolymers [51], glycodendrimers [52], and glyconanoparticles [53]. Many of these synthetic constructs have demonstrated an amplification in carbohydrate-mediated binding, in a similar manner to that achieved by multivalent carbohydrate presentations of natural glycopeptides [54] and glycoproteins [55].

The preparation of glycopolymers often requires laborious, multi-step procedures involving protection and de-protection of a saccharide and which must also include the introduction of a polymerizable group, such as vinyl or norbornene, at the anomeric centre [56, 57]. Only after multiple synthetic steps can the resulting glycomonomers be used in a polymerization reaction. Clearly, more efficient synthetic methodologies for the preparation of glycomonomers would be advantageous.

In 2014, Tanaka et al. [58] reported a one-pot method for the production of glycomonomers directly from unprotected sugars, by the DMC-mediated synthesis of 4-aminophenyl 1-thio-glycosides in water in the presence of triethylamine, and their subsequent acrylamidation (Scheme 9).

Scheme 9
scheme 9

One-pot synthesis of glycomonomers [58]. Reaction conditions: i DMC 1, HS–C6H4–NH2, Et3N, H2O/MeCN; ii acryloyl chloride, Et3N, H2O/THF

Full size image

These acylamide containing glycomonomers were then subjected to reversible addition-fragmentation chain transfer (RAFT) living radical polymerisation to give glycopolymers, which were subsequently immobilised onto gold nanoparticles for investigations into glycocluster effects.

4.5 S-Linked Glycopeptides

S-Linked peptide and protein glycosylation, where the anomeric oxygen of O-linked glycosides has been replaced by sulfur, has attracted a significant amount of interest from the Glycoscience community [59,60,61]. S-Linked glycopeptides are well known for their functional surrogacy, in addition to their chemical and biological stability as compared to their O-linked counterparts, especially towards glycosidase-catalysed cleavage [62].

S-Linked glycopeptides can be accessed by several strategies. The most common involve either conjugate addition or nucleophilic substitution reactions, with a glycosyl thiol acting as the nucleophile [63,64,65,66]. However, more recent approaches, such as free radical thiol-ene ‘Click’ reactions [67], desulfurative rearrangements [68], and the opening of 1,6-anhydrosugars [69], are alternatives that may possess certain advantages. Nevertheless, all of these methods still require multi-step methodologies in order to prepare the required glycosyl thiol.

Winssinger et al. [70] recently reported the use of DMC for the solid phase synthesis of S-linked glycopeptides. By using a H2O/dioxane solvent mixture and cooling to −10 °C, they demonstrated that unprotected sugars could be ‘pre-activated’ by DMC, and found the glycosyl-imidazolinium intermediate that was formed was stable for up to 1 h. The addition of a polymer-bound peptide containing a thiol to the reaction mixture during that time period then led to the formation of the desired polymer-bound S-linked glycopeptide 27 (Scheme 10).

Scheme 10
scheme 10

Glycosylation of a thiol-containing resin with D-glucose using the DMC methodology [70]. Reaction conditions: i DMC 1, Et3N, H2O/dioxane (1:1), −10 °∁; ii RSH

Full size image

An attempt was also made to glycosylate a polymer-bound tripeptide containing cysteine (Fmoc-Cys-Phe-Phe-Rink) with various mono- and disaccharides. Interestingly, direct reaction with D-glucose with DMC in the presence of the tripeptide 28 (i.e. without the pre-activation process) led to the exclusive formation of the thioimidazolidinium by-product 29 (Scheme 11a). However, when the conditions previously identified for pre-activation were used, the desired glycopeptide 30 was obtained in >90% yield (Scheme 11b).

Scheme 11
scheme 11

Solid phase S-glycosylation of a resin-bound tripeptide 28 by Winssinger et al. [70]. Reaction conditions: 28, DMC 1, Et3N, H2O/dioxane (1:1), −10 °∁; ii DMC 1, Et3N, H2O/dioxane, −10 °∁, 15 min; iii 28

Full size image

The process was applied to other monosaccharides (D-galactose, D-mannose, L-fucose) and also to disaccharides (lactose, mellibiose, and Galα(1 → 4)Glc), though the yields were lower in the cases of the disaccharides. An elegant method, involving an iterative addition strategy, was found to increase efficiency of glycopeptide formation, and led to conversions to the desired product of >99%. The process was then exemplified by the solid phase synthesis of an analogue of the repeat unit of the cancer-associated MUC1 glycopeptide, in which the two natural O-linked carbohydrates sites were replaced with S-linked glycans.

5 Synthesis of Glycosyl Azides in Water Using DMC

Glycosyl azides are highly useful synthetic intermediates in sugar chemistry [71]. The traditional method for their formation involves at least three or four steps, invariably involving protecting group manipulations, to furnish the desired de-protected product. For example, a typical sequence involves the conversion of the sugar, e.g. N-acetyl-D-glucosamine 31, into a protected glycosyl halide 32, followed by nucleophilic displacement of the anomeric leaving group by azide to give the glycosyl azide 33 (Scheme 12). Finally, the ester protecting groups are removed, for example by Zemplén de-acetylation, to give the de-protected glycosyl azide 34.

Scheme 12
scheme 12

Synthesis of de-protected glycosyl azides. Reaction conditions: i AcCl; ii NaN3, sat. aq. NaHCO3, tBuNH4·HSO4, DCM; iii Na, MeOH

Full size image

In addition to the inherent inefficiency of these multi-step reaction sequences, this type of synthetic route can also occasionally become problematic when applied to higher oligosaccharides due to cleavage of inter-glycosidic bonds during the synthetic pathway.

Following on from earlier studies in the Shoda group, Tanaka et al. [30] reported that glycosyl azides of monosaccharides could be formed directly when unprotected sugars were treated with excess DMC in water in the presence of a large excess (typically 10 equivalents of each) of azide and triethylamine (Scheme 13).

Scheme 13
scheme 13

Direct synthesis of glycosyl azides from unprotected sugars using DMC [30]. Reaction conditions: i DMC 1, NaN3, base, H2O

Full size image

Interestingly, when the reaction was applied to disaccharides using triethylamine as the base, 1,6-anhydro sugars were formed as by-products. It was suggested that, for some reason, intramolecular nucleophilic attack of the 6-hydroxy group at the anomeric centre was increased in the presence of triethylamine [30]. A screen of various other bases was performed, and it was found that the use of either N,N-diisopropylethylamine (Hunig’s base, DIPEA) or 2,6-lutidine led to reduced formation of the 1,6-anhydro derivative, presumably an effect that was related to their greater steric bulk. The use of 2,6-lutidine as the base was also found to be essential in the case of 2-acetamido sugars to avoid significant oxazoline formation.

Tanaka et al. [30] then demonstrated that the reaction could be applied to larger oligosaccharides, including a sialic acid terminated complex biantennary decasaccharide, which was converted to the corresponding glycosyl azide in 87% yield, though this reaction did require the use of 20 equivalents of DMC and 40 equivalents of 2,6- lutidine. Glycosyl azides were generally produced stereoselectively as the 1,2-trans glycosides, except in the cases of 2-deoxy sugars, which were formed as anomeric mixtures. To date, a detailed mechanism has not been presented, though the results suggest that reaction of sugars containing a 2-hydroxy group probably proceeds via 1,2-anhydro sugar intermediates.

6 Synthesis of Glycosyl Triazoles in a One-Pot Reaction

The original Shoda method [30] for glycosyl azide synthesis, like all DMC activation procedures, potentially generates up to two equivalents of HCl per equivalent of DMC used; one equivalent is produced from de-protonation of the anomeric hydroxyl group, and a second one may additionally be produced from water in the case of a hydrolysis reaction. This production of acid during the course of the reaction necessitates the use of excessive amounts of triethylamine. Furthermore, the use of a large excess of azide is also required.

As an alternative approach which would avoid the requirement for this large excess of reagents, we envisaged replacement of the chlorine on the dimethylimidazolinium ring of DMC with azide; such a reagent should still be able to activate a reducing sugar, would only result in the formation of one equivalent of acid, and would additionally itself act as the source of azide. 2-Azido-1,3-dimethylimidazolinium hexafluorophosphate (ADMP), has been previously used as an efficient agent for diazo-transfers [72,73,74,75], migratory aminations [76, 77], and azide transfer [78] reactions. Although ADMP contains a significant amount of nitrogen, impact sensitivity and friction sensitivity tests have demonstrated that it is not explosive [73], making it in fact safer to handle than sodium azide. Additionally unlike DMC, which is hygroscopic [23, 73], ADMP is isolated as a stable crystalline solid and is easily handled. Although, unlike DMC, it is not currently commercially available, ADMP can be very easily synthesised, for example following procedures originally reported by Kitamura et al. [72, 79] (Scheme 14), simply by conversion of DMC 1 to the hexafluorophosphate salt 35, and then treatment with sodium azide to give ADMP 36.

Scheme 14
scheme 14

Synthesis of ADMP 36 [72, 79]. Reaction conditions: i NaPF6, MeCN, rt, 30 min, 99%; ii NaN3, MeCN, 0 °C, 3 h, quant

Full size image

In 2014, we [80] reported an alternative method for the synthesis of glycosyl azides in water. N-Acetylglucosamine was reacted with ADMP 36 and triethylamine in a D2O/MeCN (4:1) solvent mixture to give the glycosyl azide 33, as well as oxazoline 37 (Scheme 15). The use of D2O as solvent, as first reported by Shoda in the original oxazoline formation study [8], resulted in higher yields of products, presumably due to reduced rates of competitive solvent-mediated hydrolysis.

Scheme 15
scheme 15

Reaction of GlcNAc with ADMP to yield glycosyl azide 33 and oxazoline 37 [80]. Reaction conditions: i ADMP 36, Et3N, D2O/MeCN (4:1), 0 °C

Full size image

However, applying the idea reported by Shoda et al. [42] that the subsequent addition of acid would result in oxazoline ring opening, we found that the desired azide 33 was obtained as the sole reaction product when the crude reaction mixture was simply acidified with 1 M aqueous HCl.

Glycosyl azides are obvious substrates for further elaboration by the modified Huisgen cycloaddition [81, 82], the most well known ‘Click’ reaction. This high yielding transformation may be performed in water, leading to numerous applications, as reaction conditions are compatible with biological systems. In the carbohydrate field, ‘Click chemistry’ has already been applied to the synthesis of a wide range of sugar derivatives [83] having interesting biological properties [84], including inhibitory activity against glycosidases [85,86,87,88,89,90] and glycosyltransferases [91]. The development of a one-pot process for glycosyl azide and then triazole formation by Click reaction with an alkyne was, therefore, an obvious avenue for investigation.

We found that after the formation of the glycosyl azide was complete, the addition of propargyl alcohol, CuSO4·5H2O, and L-ascorbic acid, and then heating at 50 °C for 14 h, led to the formation of the glycosyl triazole 38 in excellent yield and with complete stereoselectivity; the 1,2-trans glycosyl triazole being formed in all cases (Scheme 16). This procedure was then applied to a variety of sugars ranging from in size from mono- to trisaccharides [80]. In all cases, the use of D2O as co-solvent led to a variable, but minor, the amount of deuterium incorporation into the triazole ring during the course of the reaction.

Scheme 16
scheme 16

Direct synthesis of glycosyl triazoles from reducing sugars in water by Fairbanks et al. [80]. Reaction conditions: i (1) ADMP 36, Et3N, D2O/MeCN (4:1), 0 °C, 3 h, then add propargyl alcohol, CuSO4·5H2O, L-ascorbic acid, 50 °C for 14 h

Full size image

The generality of the one-pot process with respect to the alkyne was explored. The reaction was found to be widely applicable, including with interesting and potentially biologically relevant substrates such as propargyl glycine. All alkynes investigated were successfully Clicked with monosaccharides, and products formed in high yield and with complete stereoselectivity. The only limitation appeared to be in terms of the solubility of the alkyne-coupling partner in the aqueous reaction medium. The one-pot Click reaction also allowed the direct conjugation of reducing sugars to a variety of other carbohydrates which themselves contained an alkyne functionality, giving rise to a variety of di-, tri- and pentasaccharide mimics. All reactions were high yielding and completely stereoselective [80].

Click chemistry has previously been applied to access different glycopeptides, [92, 93] and also protein scaffolds decorated with oligosaccharides [94]. However, reported procedures require multiple protecting group manipulations. In terms of a biologically interesting application, glycosylated versions of the tandem repeat domain of the cancer-associated mucin MUC1 [95, 96] have shown potential as components of synthetic anti-cancer vaccines [97, 98]. A further exemplification of the one-pot azide/Click reaction, we, therefore, conjugated a variety of sugars to two synthetic MUC1 peptides that incorporated propargyl glycine (Pra) residues. Conversion of GalNAc to its azide using ADMP and then direct reaction with peptides 39 and 40, comprising one and two propargyl glycines, respectively, in the presence of CuSO4·5H2O, gave the corresponding glycopeptides 41 and 42 in good yield (Scheme 17).

Scheme 17
scheme 17

Conjugation of GalNAc with 10-mer 39 and 20-mer 40 by Fairbanks et al. [80]; Reaction conditions: i GalNAc, ADMP, Et3N, D2O, MeCN, 0 °C, 3 h, then add alkyne, CuSO4·5H2O, L-ascorbic acid, and heat to 50 °C for 14 h [80]

Full size image

Similarly tetrasaccharide 43 [99], which corresponds to a core region of N-glycans, was also directly reacted with peptide 39 and gave glycopeptide 44 in 42% yield (Scheme 18).

Scheme 18
scheme 18

Reaction of tetrasaccharide 43 with 10-mer 39 by Fairbanks et al. [80]. Reaction conditions: i ADMP, Et3N, D2O, MeCN, 0 °C, 3 h, then add 10-mer 39, CuSO4·5H2O, L-ascorbic acid, and heat to 50 °C for 14 h

Full size image

As a final example of the utility of the method, the complex biantennary N-glycan 6 decasaccharide [21, 100] was converted to the glycosyl azide using ADMP, and then directly reacted with peptide 39 in the presence of CuSO4·5H2O to furnish glycopeptide 45 in 42% yield (Scheme 19).

Scheme 19
scheme 19

Conjugation of decasaccharide 6 to 10-mer 45 by Fairbanks et al. [80]. Reaction conditions: i 6, ADMP, Et3N, D2O, MeCN, 0 °C, 3 h, then add 10-mer 39, CuSO4·5H2O, L-ascorbic acid, and heat to 50 °C for 14 h

Full size image

7 Conclusion

Traditional methods of accessing glycoconjugates typically involve multi-step syntheses and require complex and protracted protecting group strategies. These methods are generally technically demanding, inefficient, expensive, and the production of significant amounts of the target material is usually logistically difficult to achieve.

Since the initial report by Shoda and co-workers on the use of 1,3-dimethylimidazolinium chloride (DMC) for the selective conversion of unprotected 2-acetamido sugars to glycosyl oxazolines in water, various synthetic targets have been accessed using this, or related, highly selective activating agent. In particular, selective activation of the anomeric centre under aqueous conditions and direct reaction with a diverse range of nucleophiles allows ready access to a range of glycoconjugates without recourse to any protecting group manipulations. Additionally, the products of nucleophilic substitution at the anomeric centre may be further derivatised in the same reaction vessel, for example by Click chemistry.

As a wide variety of structurally complex reducing oligosaccharides are available from natural sources, protecting group free methods based on DMC and similar activating agents show significant promise for rapid access to a broad range of biologically interesting glycoconjugates.