Applications of fluorous tag methodology in carbohydrate synthesis
Published:15 Dec 2021
I. H. Mahmud and P. G. Goekjian, in Carbohydrate Chemistry: Chemical and Biological Approaches, Volume 45, ed. A. Pilar Rauter, T. K. Lindhorst, and Y. Queneau, The Royal Society of Chemistry, 2021, vol. 45, pp. 1-56.
Download citation file:
The use of fluorous tags to assist in solution-phase oligosaccharide synthesis is reviewed. A full panoply of light and heavy fluorous tagged protecting groups are now available, providing a choice of anomeric, terminal, late-stage, iterative, and activatable groups that allow fluorous tag methodology to be applied to most oligosaccharide synthetic strategies with very little adaptation. A variety of reaction and purification protocols have been developed that greatly accelerate the synthesis and purification steps in oligosaccharide synthesis. The outlook for fully automated oligosaccharide synthesis is highly promising, and at the same time, fluorous tag-assisted synthesis is an important asset for carbohydrate research groups, which can be adopted with relatively little effort and expense.
Fluorous-assisted synthesis is a methodology for enabling stoichiometric or catalytic chemical transformations based on the limited miscibility of partially or fully fluorinated compounds in nonfluorinated media.1 It allows for easy isolation of the fluorous compound from complex reaction mixtures and combines the advantages of solid- and liquid-phase synthesis.2–4 In practice, a number of parameters need to be adjusted for effective application of fluorous synthesis. These include the availability of fluorous tagging groups, the number of fluorines required to render organic molecules “fluorous”, and the overall efficacy with which fluorous substrates react in organic transformations. Over the years, a number of different incarnations of fluorous technology have appeared, each of which presents strong advantages in particular contexts.
In 1994, Horváth and Rábai reported a fluorous biphasic reaction system (FBS) consisting of a fluorous phase containing a dissolved reagent or catalyst and another phase, which could be any common organic or aqueous solvent with limited solubility in the fluorous phase.5 Under vigorous stirring and/or at higher temperatures, the fluorous and non-fluorous components become sufficiently miscible for the reaction to proceed efficiently, and upon cooling and sitting, there is complete separation of the fluorous and non-fluorous partners. In the case of a fluorous catalyst, the organic product can be decanted off and the catalyst can be efficiently recycled. A variety of fluorous solvents with different miscibilities, physical properties, and solvent properties are available. The ability to completely separate a catalyst or a reagent from the products under mild conditions could lead to industrial applications in homogeneous catalysis and to the development of more environmentally benign processes. A significant number of fluorous chains are needed to entice otherwise organic molecules or transition metal complexes to partition into a fluorous phase to the extent needed to avoid leeching of the catalyst, but the high molecular weights associated with heavy fluorous molecules are not detrimental, provided that the fluorous catalysts are efficient and can therefore be used in low molar ratios.6
In another incarnation, Dennis P. Curran and co-workers demonstrated in the late 1990's that the starting material of a multistep synthesis could be labeled with a fluorous “tag”, the reaction sequence could be run under standard conditions in organic solvents, and the final products could be extracted into a fluorous layer by liquid–liquid extraction (Fig. 1).7 The fluorous tag was then cleaved to provide the target organic molecules. The use of this methodology for conducting and purifying multicomponent reactions illustrated the power of the fluorous approach. Successfully combining the experimental conditions of traditional organic synthesis in solution with the ease of purification of solid-phase synthesis provided a broadly applicable method for the synthesis of small molecules.
Fluorous biphasic synthesis and fluorous liquid–liquid extraction require “heavy fluorous chain” tags leading to compounds with about 60% fluorine by molecular weight in order to ensure complete partitioning into the fluorous layer and efficient liquid extraction in organic/fluorous biphasic or aqueous/organic/fluorous triphasic systems. On the other hand, a number of methods have been developed based on “light fluorous tagged” protecting groups, with 40% fluorine or less by molecular weight, which rely on the use of a fluorous stationary phase rather than on liquid–liquid extraction.
Light fluorous-labelled compounds can be recovered with excellent purity by solid/liquid separation using fluorous-derivitized silica gel as an alternative to the liquid/liquid approach. Fluorous chromatography was first applied to F-tag-assisted sythesis in the form of fluorous solid-phase extraction (FSPE),8 which has since become the most widely used isolation technique for fluorous synthesis. Long popular in chemical analysis, solid-phase extraction is increasingly used in chemical synthesis due to its relative simplicity, its separation power, and its suitability for automation. A crude mixture of fluorous, organic and inorganic compounds is loaded onto fluorous silica gel and eluted in a first-pass with a “fluorophobic” solvent. The fluorous material adsorbs onto the column while the organic and inorganic materials elute with or near the solvent front. A second-pass elution with a “fluorophilic” solvent then elutes the fluorous compound in a single fraction or in several fractions. In addition to its simplicity, the SPE procedure is attractive because no fluorous solvent is needed for either the reaction or separation and because the fluorous silica is robust and reusable. SPEs feature high loading levels (sometimes 10% or more) and generally only two fractions, organic and fluorous, are collected. SPE of heavy fluorous compounds can also be attractive for its more robust separation.
At the other extreme of chromatographic resolution, fluorous-HPLC has been used to separate mixtures of fluorous-tagged compounds on a stationary phase bearing fluorocarbon chains in a high pressure/performance liquid chromatography mode. A number of such columns are available under trademarks such as Fluofix® and Fluoroflash™ among others, many of which show broad utility for the chemical analysis of organic or biological molecules beyond fluorous synthesis.3,9–14 Fluorous columns present a significantly superior resolving power for the purification of mixtures of fluorous compounds relative to standard reverse-phase columns, due to their considerably higher affinity for fluorous moieties. Alternatively, the more common pentafluorophenyl (PFP) columns can provide an intermediate compromise that combines fluorous and hydrophobic affinities.15 The significant retention affinity of fluorous motifs allows the efficient separation of multiple or differentially fluorous-tagged compounds. For example, Curran and co-workers reported the separation of differentially F-tagged libraries prepared by a “split and pool” approach.16 The fluorous-HPLC cleanly separated the different components based on fluorine content.
Bridging the gap between fluorous SPE and fluorous-HPLC is “fluorous flash chromatography”.6,17 Fluorous silica gel is prepared by silylation of standard silica gel with ClSi(Me2)CH2CH2C6F13 and imidazole in toluene at 100 °C.18 Fluorous chromatography can be conveniently performed using SPE cartridges with lower loading,6 or preparatively by traditional flash column chromatography with fluorous silica gel. The successful purification of fluorinated compounds by column chromatography on fluorous silica gel likely predates Dennis P. Curran's report,18,19 but is particularly useful in the context of fluorous tag-assisted synthesis.
These significant practical developments in fluorous-assisted synthesis have led to a considerable effort toward readily accessible and versatile building blocks containing perfluoroalkyl chains. In particular, fluorous protecting groups such as fluorous benzyloxycarbonyl (F-Cbz),20,21 F-tert-butyloxycarbonyl (F-Boc),22 F-methylsulfonylethoxycarbonyl (F-Msc),23 F-benzyl,24 F-silyl,25 F-alkoxy ethyl ether,26 F-tert-butyl,27,28 F-p-methoxybenzyl (F-PMB),28 F-trialkoxybenzhydryl (F-Rink-type),28,29 F-propionate (Bpf),30,31 among others,32 were reported in the literature to provide convenient F-Tags for fluorous synthesis. The fluorous chain is typically isolated from the reacting functional group by two or three methylene groups in order to maintain the reactivity of the corresponding non-fluorous protecting groups. In this chapter we will thus refer somewhat arbitrarily to a fluorous ponytail if the fluorous moiety is not intended to be removed from the molecule (for recycling, immobilization, etc.), a fluorous tag if it is intended to be cleaved to provide a non-fluorous product,6 and a fluorous protecting group in those cases where the fluorous group is designed as a direct analog of a common protecting group, and whose fluorous moiety is electronically isolated from the reacting functionality. In these cases, the selectivity during formation and cleavage is sufficiently similar that the fluorous protecting group can be used interchangibly with the non-fluorous group in the synthetic design. It is clear that a fluorous protecting group is per se a fluorous tag, and that a fluorous tag can be used as a fluorous ponytail, so that these terms are strictly contextual.
The attractive qualities of light tag fluorous synthesis were summarized by W. Zhang:33
1 Fluorous tags are inert to chemical reactions and have minimal effect on the attached molecules’ reactivity.
2 The solubility of the fluorous molecules in organic solvents may be equal or better than the untagged molecules because of the perfluoroalkyl chain's lipophilicity.
3 Homogeneous reactions in organic solvents are faster than polymer-bound heterogeneous reactions.
4 Literature reaction conditions can be applied with little or no modification.
5 A significant excess of fluorous reagent is not required in most cases.
6 The measure of the stoichiometry of fluorous compounds is easier and more accurate compared to immobilized compounds.
7 The reaction progress can be monitored by conventional analytical methods such as TLC, HPLC, MS, IR, and NMR.
8 Both intermediates and final products can be purified by fluorous separations as well as by regular and reverse-phase chromatography.
9 Scale-up and automation are possible.
2 Why use F-tags in carbohydrate synthesis?
The impressive progress in oligosaccharide synthesis over the last thirty years has made the synthesis of complex, biologically relevant oligosaccharides accessible to organic chemists by chemical and enzymatic routes,34,35 including the industrial syntheses of complex, chemically defined carbohydrate drugs that are currently on the market as anticoagulants36 and vaccines.37 The carbohydrate community has convincingly demonstrated that it can make oligosaccharides. Much of the progress to be made is thus in accelerating the process, so that new defined carbohydrates can be prepared in 6–12 months rather than 5–10 years. The work of Seeberger et al. and others in developing solid-phase oligosaccharide synthesis is a key step in that direction,38 but, despite the strong results, this methodology is not yet accessible to the average non-carbohydrate research group, unlike solid-phase peptide and oligonucleotide synthesis. There are several aspects of carbohydrate synthesis that make full automation particularly challenging:
1 Reaction conditions for the key glycosylation step are highly variable, are typically run at low temperatures in the presence of highly reactive Lewis acids, and are rather temperamental. Unlike peptide and nucleotide synthesis, where there are a few broadly applicable reaction conditions, there are quite a few glycosylation conditions, and more need yet to be developed. Functional group chemistry is a first approximation, and the idea that the activated acetal group of the glycosyl donor and the alcohol of the glycosyl acceptor will react in the same way regardless of the rest of the carbohydrate molecule is well… not always supported by experimental results.
2 There are complex issues of regiochemistry and stereochemistry that do not exist in peptide and nucleotide chemistry. The mantra of carbohydrate chemists is that there is one peptide sequence Ala-Ala-Ala; there is one nucleotide sequence AAA; there are 832 different trisaccharides Gal-Gal-Gal (and we'd like to be able to make each one). Finally, unlike phosphodiester and amide-forming reactions, the glycosylation reaction creates a new stereocenter, the control of which is of central importance in oligosaccharide synthesis.
3 Protecting groups play a highly strategic role in carbohydrate synthesis, far beyond the typical temporary capping of a functional group.39 In chemical oligosaccharide synthesis, the issue of regiochemistry is ultimately solved through protecting groups, and it is, therefore, necessary to control the regioselectivity of the protection steps as well as the order of deprotection of the different groups during oligosaccharide assembly. The structural variety of oligosaccharides thus requires placing terminal, late-stage, and iterative protecting groups at the appropriate positions of the monosaccharide precursors, as well as activatable protecting groups at the anomeric position. The nature of the protecting group, in particular at the C2 position, plays a key role in the stereoselectivity of the glycosylation reaction. The protection conditions, deprotection conditions, and regioselectivity in the protection steps all need to be taken into account in the choice of protecting groups and the synthetic design. There is, therefore, no universal set of protecting groups that are used for carbohydrate synthesis.
4 Purification is by far the most time-consuming part of oligosaccharide synthesis. The desired coupled product needs to be separated from the unreacted glycosyl acceptor, the hydrolyzed glycosyl donor (usually as a mixture of anomers), the by-product of the activating group (e.g., trichloroacetamide, disulfide compounds, etc.), degradation products, and other reaction components, with no large difference in polarity between them, unlike for example in peptide synthesis. The stereocontrol of the glycosylation step is not always perfect, in which case the product must also be separated from its anomer, which tends to be challenging. It is often necessary to perform multiple purifications to isolate the product in acceptable purity. In addition, there are typically a large number of regioselective protection steps to make the precursors, which adds considerably to the number of purifications.
Fluorous tag methodology offers a powerful tool in oligosaccharide synthesis by using fluorous tagged protecting groups, as this methodology can be adapted in a straightforward and direct manner to existing synthetic approaches. Fluorous analogs of many protecting groups used in carbohydrate synthesis are available for the synthesis of fluorous monosaccharide precursors, with relatively similar stabilities, protection and deprotection conditions, and regiochemical behavior to the traditional protecting groups. In short, while automated solid-phse and F-tag carbohydrate synthesis can bring oligosaccharide synthesis to non-carbohydrate groups, in a more immediate sense, F-Tag synthesis can bring accelerated synthesis to carbohydrate groups.
Oligosaccharide synthesis is therefore a particularly opportune application for fluorous tag-assisted synthesis. An overview of F-tagged carbohydrate synthesis is provided to show both the progress and the opportunities in this field. We will focus more on the aspects related to carbohydrate synthesis than on the preparation of the F-tags, which generally involve relatively acessible chemistry. The chapter is organized based on the most common role of the parent protecting group: a terminal protecting group will be removed at the very end of the synthesis, to provide the free alcohol or hemiacetal in the final oligosaccharide, and thus must be stable to the all reaction and deprotection conditions used during the synthesis; late-stage protecting groups are removed late in the synthesis, either to introduce branching points or modifications such as sulfate, phosphate or ester groups, and therefore must be stable to all conditions other than the removal of the terminal groups; iterative protecting groups are used in reducing-to-non-reducing iterative oligosaccharide synthesis, and must only be stable to the glycosidic coupling conditions. However, the principal practical consideration in the choice of protecting groups is the control of the regioselectivity in order to introduce each one in the appropriate position of the monosaccharide precursors. Furthermore, the protecting group at the C2-position of the glycosyl donor controls the stereochemistry of the glycoside linkage. It is therefore important for synthetic design to know how closely a fluorous protecting group mirrors the chemical reactivity of the parent protecting group, both in its regioselectivity during protection steps and its chemoselectivity during deprotection steps. The anomeric position may bear either an activated group (halogen, trichloroacetimidate, phosphate, etc.), a terminal protecting group, a permanent linker for immobilisation, conjugation, etc., or an activatable protecting group in non-reducing-to-reducing sense and bi-directional oligosaccharide synthesis. Fortunately, only one or two fluorous groups are needed for fluorous-assisted synthesis, so there is considerable flexibility in the protecting group strategy to be adopted.
3 Reducing sugar anomeric fluorous tags and ponytails
Placing a fluorous moiety at the anomeric position of the reducing sugar allows for reducing-to-non-reducing iterative oligosaccharide synthesis with a single fluorous tag or ponytail, in order to efficiently separate the fluorous oligosaccharide from the residual glycosyl donor and reaction by-products at each step. The reducing sugar anomeric tag has been the most widely used option, by analogy to solid-phase synthesis, as this position will not be involved in glycosylation reactions, by definition. Fluorous tag methodology will not generally separate the coupled product from the uncoupled glycosyl acceptor. Hence, as in solid-phase synthesis, it is necessary to push the coupling reaction to completion by recycling through the reaction conditions once or twice; however, in contrast to solid-phase synthesis, this is usually possible with only a few equivalents of the glycosyl donor. The anomeric tag can be removed at the end of the synthesis in which case it will be a fluorous tag, or may serve as a linker for immobilization and hence be considered a fluorous ponytail. The use of activatable fluorous anomeric thioglycosides for non-reducing-to-reducing sense oligosaccharide synthesis is treated later in the chapter.
3.1 Cleavable anomeric fluorous tags
Mizuno and co-workers developed several recyclable polyamide heavy fluorous tags for oligosaccharide and peptide synthesis using liquid–liquid fluorous extraction (Scheme 1). A glutarate-linked anomeric ester heavy tag Hfb 1 (hexakisfluorous chain-type butanoyl) was reported by Goto et al. for the efficient iterative oligosaccharide assembly and recycling of the F-Tag, as demonstrated by the synthesis of the trisaccharide 10 using the 6-O-TBDPS group as the iterative protecting group (Scheme 1A).40 The heavy tag carboxylic acid 1 was introduced onto the glycosyl acceptor 2 by esterification. Iterative deprotection of the silyl group and glycosylation of 3 with the trichloroacetimidate 4 and of 6 with 7 gave the heavy fluorous tagged trisaccharide 8 in 42% overall yield from the hemiacetal 2. Each synthetic intermediate could be monitored by TLC, NMR, and mass spectrometry. The use of the Hfb allows rapid purification of synthetic intermediates by partitioning between fluorous and organic solvents. Cleavage of the fluorous tag and the ester protecting groups under Zemplen conditions, followed by hydrogenolysis efficiently provided the trisaccharide 10, along with the recovered heavy fluorous tag 1, which was recovered in 81% yield after saponification of the methyl ester.
In a second example from the same group, Goto, Miura and Mizuno explored the concept of “tag and sacrificial linker” to develop a benzyl-type linker to efficiently introduce the hexakis(fluorous chain) alcohol 11 and perform reducing-to-non-reducing synthesis (Scheme 1B).41 Glycosylation of the benzyl alcohol tag 13 with a 2.0-fold excess of the phenyl tri-O-benzyl-6-O-Fmoc-thioglucoside donor 14, in the presence of N-iodosuccinimide (NIS) and triflic acid (TfOH) in dichloromethane – ethyl perfluorobutyl ether mixture, followed by deprotection of the Fmoc iterative protecting group gave the heavy F-tagged glycosyl acceptor 16. The reaction with a 2.0-fold excess of thioglycoside donor 14 under the same conditions afforded the fluorous disaccharide 17, thus setting the basis for iterative reducing-to-non-reducing oligosaccharide synthesis by a sequence Fmoc deprotection and glycosylation. The improved α-selectivity using the tribenzyl 6-O-Fmoc thioglycoside compared to the result above is noteworthy. Each of the fluorous intermediates was isolated in a straightforward manner by partitioning between FC72 and either MeOH or MeCN. The fluorous tag 11 was cleaved by treatment with NaOMe and the heavy flourous tag could be recovered in 91% yield with a FC72/MeOH partition system. The disaccharide 18 on the other hand was obtained in 56% overall yield from 13 after silica-gel column chromatographic purification of the crude MeOH extract and was deprotected by hydrogenolysis to the disaccharide 19.
In related work by Goto and Mizuno, a heavy fluorous tag 20 derived from meso-erythritol was attached by a dihydroquinone linker to the reducing sugar anomeric center for the preparation of disaccharide unit 31 (Scheme 2A).42 The dihyroquinone heavy fluorous tag 22 was introduced by O-glycosylation and was used to prepare both the glycosyl acceptor 25 and glycosyl donor 28, which were coupled to give the disaccharide 31. The fluorous synthetic intermediates were isolated by fluorous–organic solvent partition, with only three silica chromatography steps over the course of the synthesis. Unfortunately, the 1H NMR peaks from the meso-erythritol tag scaffold interfered with key signals of the sugar ring, complicating structure elucidation. The fluorous tag 30 was cleaved by acetolysis and recovered in 70% yield after 7 steps. In the same vein from the Mizuno group, an acid-resistant all-carbon-linked heavy fluorous tag 32 was introduced at the anomeric position as a bis-allylic ether by Fukuda et al. to provide the glycosyl acceptor 38 after routine deprotection-protection steps (Scheme 2B).43 Glycosylation with the trichloroacetimidate 39 gave the β-disaccharide 40. Fluorous liquid–liquid extraction-assisted deprotection, protection, and glycosylation reactions, conducted in high yields devoid of a chromatographic purification process, established a general iterative oligosaccharide synthetic methodology. The anomeric fluorous tag was removed by Pd(PPh3)4 mediated hydrolysis, and the heavy fluorous tag 32 was recovered efficiently (>80% yield over 7 steps)along with free anomeric sugar 41. A variety of additional fluorous tags were inspired by these precedents.44,45 These results collectively clearly establish the basis for efficient reducing-to-non-reducing iterative oligosaccharide synthesis by liquid–liquid extraction using recyclable heavy F-tags, which may find optimal applications on large scale in industry where the fluorous solvents can be recycled efficiently.46
The application of the acid- and base-stable light fluorous silyl ether tag developed by Rover and Wipf25 to fluorous oligosaccharide synthesis was reported by Manzoni et al.47,48 (Scheme 3). The silyl group was used as a reducing end anomeric tag to synthesize the disaccharide 50 and the Lewis a trisaccharide 59. Silylation of the anomeric hydroxyl group of tetraacetyl-d-glucosamine 43 with the fluorous silyl bromide 44 followed by routine O-deacetylation and benzylidene formation gave the glycosyl acceptor 47 (Scheme 3A). Coupling of the β-trichloroacetimidate of galactose tetraacetate 48 with 0.1 eq TMSOTf at −30 °C with the F-tagged glycosyl acceptor 47 gave the β-disaccharide 49 in 78% yield after two recycles. Convenient TLC and MALDI MS reaction monitoring allowed the protection and coupling reactions to be driven to completion. After a similar coupling strategy using the closely related Troc-protected fluorous glucosamine acceptor 54 and the trichloroacetimidate 48 to provide 55 (Scheme 3B), the regioselective reductive opening of the benzylidene group with triethylsilane, trifluoroacetic acid and trifluoroacetic acid anhydride to 56 and coupling with the 2-O-benzyl-3,4-O-acetylfucose trichloroacetimidate 57 under similar conditions gave the trisaccharide 58. The F-Tag was removed with TBAF to provide the protected Lewis a trisaccharide 59. The fluorous tag dramatically accelerated the synthesis of the trisaccharide by reducing the purification procedures to a simple fluorous solid-phase extraction, providing the trisaccharide 59 in seven steps and 23% overall yield (average 81% per step) in only 3 days.
Automated iterative reducing-to-non-reducing fluorous synthesis of polymannoside and polyrhamnoside oligosaccharides using a light fluorous allyl anomeric tag has been developed by Pohl and co-workers (Scheme 4).49–51 The synthesis of linear and branched α-mannose oligosaccharides 63 and 70, respectively, employing an anomeric fluorous tag along with reaction conditions and FSPE protocols amenable to automation, were reported by Jaipuri and Pohl52 (Scheme 4A). The fluorous allyl group 60 was introduced by glycosylation with 2-O-acetyl-tri-O-benzylmannose trichloroacetimidate 61 to provide 63a after deacetylation. Three additional cycles of glycosylation with the donor 61 catalyzed by TMSOTf in dichloromethane at 5 °C followed by methanolysis of the iterative C-2 acetyl group in the crude SPE methanol extract, provided the tetra-α-mannoside oligoaccharide 63d in 79% overall yield using only a total of 7 equivalents of the trichloroacetimidate donor. The fluorous tag permits fluorous SPE purification of all synthetic intermediates. The synthesis of the more complex branched oligomannoside 70 was achieved using an orthogonally protected 4-O-benzyl-6-O-tert-butyldiphenylsilyl-3-O-levulonoyl-2-O-pivaloyl glycosyl donor 64 (Scheme 4A). Glycosylation of the F-tagged glycosyl acceptor 66 with 64 gave the disaccharide 67, which after removal of both levulonoyl and silyl iterative groups was glycosylated with the trichloroacetimidate 61 to provide the branched pentasaccharide 70. The fluorous tag 60 can be removed under standard palladium-mediated deallylation conditions, and the linker design was of significant advantage in improving the solubility in the reaction solvent and aqueous-organic solvent for purification protocol as a simple handle to allow for scale-up.
A fully automated solution-phase synthesis of the challenging β-(1 → 2), β-(1 → 3) and β-(1 → 6)-mannan oligomers 77–79, was achieved by a β-directing C-5 carboxylate manuronate strategy, as recently disclosed by Tang and Pohl53 (Scheme 4B). The automated synthesis began with conjugation of the F-tag allyl alcohol 60 and the trichloroacetimidate of methyl di-O-benzylmanuronate bearing a p-methoxybenzyl group at either the 2-O- or 3-O-position 71 and 72, respectively, catalysed by TMSOTf at −20 °C. After 30 min, the solvent was removed under reduced pressure, and a ceric ammonium nitrate (CAN) solution in MeCN/H2O (1/9) was added to the mixture and vortexed for 1h to remove the iterative p-methoxybenzyl group. Complete reaction was confirmed by TLC, and the mixture was robotically transferred to the SPE station for the FSPE purification. After the FSPE, the purified product passed additional glycosylation–deprotection–FSPE cycles. The ester groups were reduced with lithium triethylborohydride and the crude products were purified manually to afford the desired β(1 → 2) and β(1 → 3)-linked mannose oligosaccharides 77 and 78. The β(1 → 6)-oligomannoside 79 was obtained from the tribenzyl trichloroacetimidate 73 in a similar way by reduction of the maluronate ester group in place of CAN-mediated deprotection. The methodology significantly improved the glycosylation reactions and thus limited the number of glycosyl donor equivalents needed compared to solid-phse synthesis, thereby considerably reducing the cost of oligosaccharide synthesis. TLC monitoring allowed the reactions to be recycled until pushed to completion. Kohout and Pohl recently reported the automated synthesis of alpha(1 → 2),(1 → 3)-l-rhamnan and l-rhamnan sulfate fragments 80–82 by a similar methodology.51 As solution-phase fluorous tag-assisted chemistry evolves, automation for robust and rapid oligosaccharide synthesis may help meet the demand for biologically relevant oligosaccharides.
In a recent methodology paper, Kabotso and Pohl reported the use of a pentavalent bismuth complex for the activation of thioglycosides which they exemplified using a variety of anomeric fluorous tags e.g., 83, 84 and 60 (Scheme 5).54 For example, the sialic acid thiophenyl glycoside 85 was coupled to the dihydroquinone-linked fluorous tag 83 in the presence of Ph3Bi(OTf)2 and isopropylthiol at room temperature to provide the fluorous sialic acid derivated 86. Although the reaction was somewhat sluggish, the stereoselectivity was high. Similarly, coupling of adamantyl tetra-O-acetylthioglucoside 87 under the same conditions provided the anomeric fluorous glucoside 88; deacteylation, tritylation, benzylation and removal of the trityl group provided the fluorous glycosyl acceptor 89, which was coupled with the tri-O-benzoyl galacturonic acid thiopropyl glycoside donor 90 under the same conditions to give the disaccharide 91. All intermediates were efficiently purified by reverse phase chromatography or FSPE, which illustrates the dual use of fluorous tags as fluorous and hydrophobic groups. A similar preparation of the disaccharide 96 via the bis(allyl) linked fluorous tagged acceptor 94 bearing an iterative levulinoyl protecting group clearly shows the potential for iterative reducing-to-non-reducing oligosaccharide synthesis.
The Boons group has reported several efficient fluorous tag-assisted syntheses of complex oligosaccharides. Zong et al. established the modular, convergent synthesis of defined heparan sulfate fragments 103 and 104 (Scheme 6A) using an anomeric aminopentyl linker protected by a perfluorooctyl-tagged benzyloxycarbonyl fluorous protecting group (CBzF).55 Coupling of the F-tagged aminopentyl alcohol 98 to the trifluoro-N-phenylacetimidate of the l-iduronic acid/d-azidoglucose disaccharide 99 gave the F-tagged disaccharide 100 bearing late-stage levulinoyl groups for installation of the sulfate groups and an Fmoc iterative protecting group. Deprotection of the iterative Fmoc and coupling of 101 to the same disaccharide donor 99 in the presence of triflic acid in dichloromethane at −20 °C gives the suitably protected tetrasaccharide 102 in 72% yield after one recycle, with excellent α-stereoselectivity. Deprotection of the levulinoyl groups, O-sulfatation with SO3-pyridine, saponification of the ester and Fmoc groups, Staudinger reduction of the azido groups to the amines, and either N-sulfatation or N-acetylation provides the heparan sulfate tetrasaccharides 103 and 104, respectively, after debenzylation. The F-tag allowed the authors to drive reactions to completion by recycling and simplified the notoriously difficult purification of these compounds, particularly the late stages of installing the sulfate and carboxylate groups. The use of trifluoroethanol to avoid the formation of micelles is noteworthy. This approach greatly simplifies the synthesis due to improved selectivity, lower reagent use, and facile purification, and sets the basis for fluorous tag-assisted synthesis of defined heparan sulfate oligosaccharides.
Huang, Gao and Boons reported the synthesis of the highly complex, biologically relevant branched hexasaccharide moiety 117 of the GPI anchor of Trypanosoma brucei using a benzyl fluorous-tag at the anomeric position of the reducing sugar (Scheme 6B).56 The anomeric stereochemistry was controlled elegantly using complementary anchiomeric C2-participation, with C-2 pyrano[1,4]oxathiane glycosyl donors such as 109 to introduce 1,2-cis-glycosides and C-2 acetyl groups such as 105 and 115 for 1,2-trans-glycosides. The light fluorous benzyl alcohol tag 84 was glycosylated with the reducing mannose thioglycoside donor 105 bearing a C2-directing acetyl group and orthogonal naphthylmethyl ether and levulinoyl iterative protecting groups. DDQ-mediated deprotection of the naphthylmethyl group and subsequent glycosylation with the oxidized oxathiane donor 109 mediated by a stoichiometric amount of triflic anhydride, 1,3,5-trimethoxybenzene and 2,6-di-tert-butyl-4-methylpyridine in DCM at −40 °C, delivered exclusively the Gal-α(1 → 3)Man disaccharide 110 after acid-mediated removal of the C-2 auxiliary, thanks to the trans-anchiomeric participation of the C2 group. Glycosylation with 3,4,6-tri-O-acetyl-2-O-benzyl-d-galactosyl trifluoro-N-phenylacetimidate 111 with TfOH in DCM at −25 °C gave the linear trisaccharide 112. Deprotection of the galactose 6-O-Nap group and iterative α-galactosylation using the triacetylated pyrano[1,4]oxothiane derived from 114 gave the branched pentasaccharide in a two step procedure. Acetylation of the free C2 group on the non-reducing galactose and sequencial deprotection of the levulinoyl group on the reducing mannose allowed coupling with the tri-O-benzyl-2-O-acetylmannosyl trifluoro-N-phenylacetimidate donor 115 to provide the protected hexasaccharide 116. The synthesis of this compound was performed using a purification protocol based on fluorous solid-phase extraction at each step, and multiple glycosylation cycles to push the reactions to completion. Global deprotection by hydrogenation over Pd/C, followed by removal of the acetyl esters using sodium methoxide in methanol furnished the target GPI anchor hexasaccharide 117. This is perhaps the most general demonstration of fluorous tag-assisted oligosaccharide synthesis to date, and can be adapted to a wide variety of target structures.
In an innovative approach from the Chen group, one-pot multienzyme (OPME) glycan assembly using fluorous mixture synthesis for the efficient assembly of complex oligosaccharides was reported by Hwang et al.57 The light fluorous tag was extremely useful in overcoming the severe purification challenges and time constraints inherent in enzymatic synthesis. Light fluorous tagged glycosyltransferase glycosyl acceptors with different lengths of perfluoroalkanes and oligoethylene glycol linkers were synthesized and tested in order to find optimal fluorous tags that were well tolerated by the glycosyltransferases in OPME reactions, while allowing for facile purification of the products by FSPE. The results showed that lactosides bearing a triethyleneglycol (TEG) or hexaethyleneglycol (HEG) spacer, with a perfluorohexyl or perfluorooctyl tag provided sufficiently good yields for practical preparation of glycosylated products. Glycosylation of the TEG-C6F13 tagged lactoside with the sugar nucleotide synthase/glycosyl transferase system PmST1 E271F/R313Y/NmCSS, Pd2,6ST/NmCSS, or EcGalK/BLUSP/PmPpA/α1-3GalT gave the trisacchairdes 118–120 (Fig. 2). The products were isolated in excellent purity by centrifugation and FSPE, washing with water and eluting the tagged trisaccharides with methanol, in 86%, quantitative, and 89% yield, respectively. The TEG/HEG-C8F17/C6F13-tagged lactosides improved the substrates’ solubility and increased the OPME glycosylation yields without compromising the FSPE purification process. In related work, Hatanaka showed that lactose and N-acetylglucosamine bearing anomeric fluorous tags could be incorporated into cell membranes and be glycosylated by cellular enzymes.58
3.2 Non-cleavable anomeric fluorous ponytails
Beyond providing handles for purification and iterative oligosaccharide synthesis, fluorous tags can also be used to permit the direct immobilization of sugars into carbohydrate arrays for biological screening. Carbohydrates communicate with a wide diversity of proteins to mediate a plethora of biological processes that include inflammatory responses, pathogen invasion, cell differentiation, cell–cell communication, cell adhesion and development, and tumor cell metastasis, making glycosyl arrays a powerful tool for biological investigations.59 Carbohydrate chip technologies on glass slides, for instance, require minimal sample usage.60–63 However, these microarray methods rely on covalent attachment of a compound to the slide and therefore require unique functional handles.
Fluorous tagging strategy has found an appealing application in a carbohydrate microarray system that would permit a variety of bioassays. Pohl and others reported the concept of direct microarray formation based on noncovalent fluorous-based interactions. Ko, Jaipuri and Pohl demonstrated that the strength of these interactions was sufficient for the construction of carbohydrate microarrays for biological screening.64,65 For such applications, the fluorous tag must be installed at the reducing sugar's anomeric position, followed by reducing-to-non-reducing linear carbohydrate synthesis. Various trichloroacetimidate glycosyl donors were reacted with fluorous-tagged allyl alcohol 60 and were hydrogenated and deprotected to yield fluorous moieties 121–135a–d (Scheme 7A). These results provided an efficient methodology for the synthesis of carbohydrate microarray chips for biological screening.
In related work, Chen and Pohl demonstrated the application of a new hydroxylamine-modified fluorous ponytail for straightforward immobilisation of reducing sugars onto a noncovalent fluorous-based microarray platform. Simply mixing the secondary flourous N,O-substituted hydroxylamine 138 in acetonitrile with the unprotected free sugar in pH 2.5 phosphate buffer and incubating at 37 °C for 24 hours followed by purification by FSPE gave a libray of fluorous N-glycosides of glucose (139), glucosamine (140), galactose (141), galactosamine (142), mannose, lactose (143), maltose, maltotetraose, and maltohexaose (144a–c)66 (Scheme 7B). While most of the sugars existed in the closed pyranose form, the mannoside existed as a mixture of pyranose and furanose forms, which were separated after acetylation, and deprotected. Fluorous tag chemistry thus allows the direct formation of glycoarrays by noncovalent fluorous-fluorous interactions with minimal reaction steps, using either natural or synthetic oligosaccharides.
4 Terminal protecting group fluorous tags
Placing the fluorous tag on an alcohol as a protecting group rather than at the anomeric position of the reducing sugar allows one to adjust the number of fluorines by the number of protecting groups as well as the length of the fluorous chains on a given group. The strategy also allows for both reducing-to-non-reducing, non-reducing-to-reducing, and bidirectional oligosaccharide synthesis. The terminal protecting groups will be removed in the final steps of the synthesis in order to unmask the sugar's free hydroxyl groups. It must therefore be stable to all synthetic steps such as coupling, functionalization, and deprotection conditions. Benzyl and acetyl groups are the most popular terminal protecting groups in oligosaccharide synthesis, as they are converted to volatile by-products so that the oligosaccharide does not require further purification. In the fluorous case, the protecting group will be retained by fluorous liquid- or solid-phase extraction, so the main issue is any residues from the reagents. Although any fluorous protecting group can be used as a terminal protecting group, providing it is stable to previous coupling and deprotection steps, we will focus in this section on fluorous benzyl groups.
One of the earliest applications of fluorous tagged oligosaccharide synthesis was reported by the Curran group in 1998, using a heavy-fluorous-tagged glycal 149 to synthesize a disaccharide 151 bearing multiple benzyl fluorous tags. Curran, Ferritto and Hua attached three fluorous chains to a benzyl group by a silicon bridge, to afford a novel protecting group (Bnf) 147 for liquid–liquid extraction.24 d-Glucal (systematic name: 1,5-Anhydro-2-deoxy-d-arabino-hex-1-enitol) was protected with three tris(perfluoroalkane)-substituted benzyl groups and was coupled with diacetone galactose 150 in the presence of toluenesulfonic acid in trifluorotoluene to furnish fluorous disaccharide 151, which was isolated by three-phase extraction in 85% yield (Scheme 8A). The fluorous tags were removed by hydrogenolysis to provide the disaccharide 152 and the fluorous toluene intermediate 146. The disaccharide 151 was also obtained from 149 and 150 with NIS, followed by tin hydride reduction, with efficient purification by fluorous liquid–liquid extraction. Although limited in scope, this result demonstrates the possibility of placing the fluorous tag on the non-reducing sugar for non-reducing-to-reducing iterative oligosaccharide synthesis. Unfortunately, the need for multiple fluorocarbon tails for effective liquid–liquid extraction also hinders the solubility of the compounds in nonfluorocarbon reaction solvents, but this result set the basis for further development of liquid–liquid fluorous assisted carbohydrate synthesis.
Goto more recently used a heavy fluorous benzyl tag to synthesize the terminal disaccharide structure of class III mucin 16567 (Scheme 8B). The pentaerythritol-derived tag 158 was attached to the galactose glycosyl acceptor by a regioselective n-Bu2SnO-directed Williamson ether synthesis, followed by routine protection steps to give acceptor 161. Glycosylation of 161 with the trichloracetimidate of tri-O-acetylazidoglucose 162 activated by TMSOTf in DCM-diethyl ether provided the disaccharide 163. The heavy fluorous tag was removed by hydrogenolysis following Zemplen deprotection of the ester groups, to provide the disaccharide 165. The fluorous intermediates were isolated by liquid–liquid extraction without further purification, and the desired compound was obtained not only in high yield but in pure form. Due to the symmetrical nature of the fluorous moeity, the spectrum was sufficiently simple that the signals from the protecting group did not interfere with the analysis of the sugar ring protons. After hydrogenolysis, the fluorous alcohol 157 was recovered from the fluorous layer in 94% yield and was recycled.
A light fluorous benzyl protecting group was reported by Kojima, Nakamura and Takeuchi for fluorous tag-assisted carbohydrate synthesis (Scheme 9).68 A variety of fluorous benzylidene compounds 169 were prepared in the glucose and galactose series bearing different lengths of fluorous chains. These fluorous benzylidenes can be deprotected by hydrogenolysis or hydrolysis and can thus serve as iterative or late-stage fluorous protecting groups in their own right. Reductive cleavage of the acetylated 4,6-O-F17benzylidene acetal 172 with Et3SiH-TFA affords the corresponding 6-O-F17benzyl-4-O-hydroxyl derivative 174 in 98% yield. On the other hand, utilizing PhBCl2 as a Lewis acid on the benzyl-protected 173, provided the related 4-O-F17benzyl-6-O-hydroxyl compound 175 in 96% yield, making it possible to synthesize representative α(1 → 4), β(1 → 4), α(1 → 6), and β(1 → 6), disaccharides 177–180. The separation of the fluorous intermediates by FSPE was straightforward. The fluorine atom content was around 21% at the final stage, and the fluorous compounds could also be purified by standard silica gel column chromatography. The F17Bn-group deprotection was achieved by hydrogenation with 10% Pd–C in EtOAc and MeOH, and the fluorous toluene byproduct was easily recovered by FSPE.
5 Late-stage protecting group fluorous tags
Late stage protecting groups are deprotected in the penultimate steps of oligosaccharide synthesis, either to install multiple copies of a functionality (sugar, sulfate, phosphate, etc.), or simply before the deprotection of the terminal groups in the deprotection sequence.
5.1 Fluorous ester groups
Ester groups are often late-stage protecting groups of choice, provided they are not used as iterative groups and are compatible with the overall synthetic design. Miura and co-workers developed a novel ester-type heavy tag Bfp (bisfluorous chain propanoyl) group 181, prepared by alkylation and acylation of β-alanine30,69 (Scheme 10). Three Bfp groups were introduced onto the reducing sugar glycoside acceptor 184 for iterative reducing-to-non-reducing oligosaccharide synthesis, using N,N-dicyclohexylcarbodiimide (DCC) and 4-(dimethylamino)pyridine (DMAP), followed by removal of the trityl group. The synthesis of the β-1,6-linked gentiotetraose tetrasaccharide 188, by three cycles of glycosylation with tri-O-acetyl-6-O-tert-butyldiphenylsilyl-d-glucosyl trichloroacetimidate 185, catalyzed by TMSOTf in Et2O at 0 °C followed by desilylation, was achieved using minimal column chromatography purification of the target oligosaccharide. Each synthetic intermediate was purified by simple FC72-toluene extraction. Only three Bfp groups made it possible to extract the derivative of the tetrasaccharide synthesized with the FC-72-organic solvent extraction with monitoring by TLC, NMR, and MS.
The strategy was extended by the same group to the rapid synthesis of galabiose Gb2 195 and the Gb3 196 oligosaccharide derivatives using fluorous–organic extraction purification31,70 (Scheme 10). Miura et al. introduced two Bfp groups 181 onto the free hydroxyl functions of allyl 4,6-O-benzylidene galactoside 189 with DCC/DMAP. After benzylidene deprotection, 181 was again selectively introduced to the primary hydroxyl function to give the fluorous glycosyl acceptor 192, which was successfully glycosylated with an excess of the glycosyl donor 193 (6 eq.) promoted by TMSOTf in ether–EtOC4F9 to afford selectively the α-linked fluorous disaccharide 194 (No β-isomer could be detected). The trisaccharide 196 was synthesized efficiently using a galactosyl donor and acceptor each with two Bfp groups. Miura et al. also reported the novel heavy fluorous benzoyl protecting group TfBz 197 (tris-fluorous chain-type benzoyl, Scheme 10) and its application to the synthesis of oligosaccharides, in particular for cases were the Bfp group was insufficient to draw the oligosaccharide into the fluorous layer.71
In a recent application of a light fluorous ester tag approach, Maza, de Paz and Nieto synthesized a protected sulfated hexasaccharide 207 (Scheme 11) as a glycosaminoglycan mimetic and studied its interactions with different growth factors: midkine, basic fibroblast growth factor (FGF-2) and nerve growth factor (NGF).72 The fluorous tag was introduced on the reducing end sugar by selective acylation of the 4,6-diol 198 using heptadecafluoroundecanoyl chloride 199, triethylamine, DMAP in CH2Cl2 at 0 °C to form the glycosyl acceptor 200.73 Coupling of the fluorous tagged acceptor with trichloroacetimidate glycosyl donor 201 (2 eq.) catalysed by TBSOTf at 0 °C, followed by F-SPE purification provided disaccharide 202 without the need for silica gel chromatography. Treatment of the disaccharide with hydrazine monohydrate in pyridine/acetic acid buffer removed the iterative levulinoyl group to provide 203 which opens the way for chain elongation to the trisaccharide 205. Iterative reducing-to-non-reducing sense synthesis by cycles of deprotection of the levulinoyl group and glycosidation with either the trichoroacetimidate 201 or 204, with isolation of each synthetic intermediate by F-SPE provides the fully protected tetra-, penta- and ultimately the hexa-saccharide 206 in 39% overall yield over 9 steps without the need for further chromatography. Glycosylation steps were pushed to completion by recycling using 2.5–3 eq. of donor to reach the hexamer with an average of 90% yield per step. Selective removal of silylidine with (HF).py complex in THF at 0 °C resulted in hexaol, and sulfation with SO3·NMe3 at 100 °C by microwave heating delivered the target molecule 207. Although the fluorous ester could be removed by Zemplen conditions along with the other protecting groups under standard conditions, the deprotected sulfated carbohydrates were found to be less active than the protected ones, so in this case the fluorous group can be considered as a ponytail, if not a pharmacophore or a prodrug.
5.2 Fluorous O-carbamates
The first total synthesis of the unprotected natural product cucurbitoside A 212 was accomplished by Kojima et al. using a novel fluorous N-phenylcarbamoyl (FCar) protecting group (Scheme 12).74 The FCar group was introduced onto glucose diacetonide 208 in high yield by in situ Curtius rearrangement of the fluorous benzoic acid with diphenylphosphoryl azide, and after conversion to the triacetylated pyranose trichloroacetimidate donor, was glycosylated with the phenolic acceptor in the presence of BF3·Et2O in DCM at −20 °C. Deacetylation and formation of the benzylidene provided the glycosyl acceptor 210 in 8 steps, purified by FSPE with a single column chromatography. Coupling of the 6-O-benzoyl-d-apiosyl trichloroacetimidate donor 211 with TMSOTf in DCM at −20 °C gave the protected cucurbitoside A 212, and the FCar group was selectively removed with Bu4NNO2. Its stability to most reaction conditions and its orthogonal deprotection conditions make it an excellent late stage protecting group.
5.3 Fluorous silyl groups
Zhang et al. from the Liu group reported one of the few clearcut examples of iterative non-reducing-to-reducing oligosaccharide synthesis, using a single silyl fluorous tag on the trichloroacetimidate donor 223 (Scheme 13). The approach uses an excess of glycosyl acceptor to ensure a high yield with respect to the F-tagged donor, and the product formed after each glycosylation cycle is purified by simple fluorous solid-phase extraction (FSPE). The excess glycosyl acceptor is recovered and reused. A fluorous silyl group was attached to the primary hydroxyl group of the trichloroacetimidate glycosyl donor 223. Three equivalents of the toluyl thioglycoside acceptor 219 was coupled with the tagged donor in anhydrous DCM at −40 °C in the presence of a catalytic amount of TMSOTf. The expected disaccharide 224 was isolated in 92.5% yield by FSPE. 95% of the excess glycosyl acceptor was recovered in pure form and could be recycled. The disaccharide was subequently converted into the trichloroacetimidate 226 and purified by silica gel chromatography. After another cycle of glycosylation and F-SPE, the protected trisaccharide 228 was obtained in 94% yield (Scheme 13) along with the excess glycosyl acceptor in 98% recovery.
5.4 Fluorous cap and tag solid-phase oligosaccharide synthesis
The same fluorous silyl tag was used by the Seeberger group in a novel capping-and-tagging (cap-tag) strategy in order to facilitate the purification of oligosaccharides prepared by automated solid-phase synthesis. Cap-tags render deletion sequences silent in subsequent coupling reactions and, in a first incarnation, also serves as a handle to separate unwanted, capped and tagged products from the desired, untagged products. Palmacci et al. thus reported that the introduction of fluorous silyl ether caps (F-Cap-tags) during automated solid-phase synthesis of oligosaccharides greatly simplifies post-synthetic workup and purification, as demonstrated for the syntheses of several trisaccharides.76 The F-cap-tagged deletion sequences were easily separated from the desired untagged product by FSPE. In a second incarnation, Carrel and Seeberger used pivaloyl caps during the automated solid-phase synthesis and the silyl fluorous F-Tag to cap the final desired oligosaccharide, which could be easily purified by gradient fluorous solid-phase extraction (FSPE), as illustrated by the synthesis of the trisaccharide 231 (Scheme 14).77 In a very interesting work, Liu et al. carried the concept further by integrating solid-phase and fluorous tag-assisted oligosaccharide and peptide synthesis.78 Solid-phase peptide synthesis using the glycosylserine 232 gave the glycosylpeptide 233, which was then extended to the disaccharide 234 by solid-phase oligosaccharide synthesis (Scheme 14). The final solid-phase glycosylation was performed with a fluorous ester-tagged trichloroacetimidate glycoside donor 236, prepared by esterification of the corresponding tribenzoyl thioglycoside 219 with the fluorous acid chloride 235 and conversion to the trichloroacetimidate, leading to the complex glycopeptide fragment 237. The hydrolyzed glycosyl donor could be recovered and recycled from the solid-phase wash and the final glycopeptide 238 was purified efficiently after cleavage from the resin, both by SPE.
5.5 Fluorous mixture synthesis of oligosaccharides
In another innovation, Tojino and Mizuno described the use of solution-phase chemistry with fluorous tags for carbohydrate mixture synthesis and deconvolution.79 Fluorous oligosaccharide libraries were formed by fluorous mixture synthesis (FMS) using 4-alkoxyphenyl fluorous labels of different lengths at the anomeric position of the glycosyl acceptors, and either a fluorous ester tagged or an untagged glycosyl donor (Scheme 15). The anomeric fluorous tags 239a–c were introduced into different sugar pentaacetates in the presence of boron trifluoride etherate, and could be cleaved oxidatively with cerium ammonium nitrate. Glycosylation of a mixture of three galactose acceptors bearing different length alkoxyphenyl tags C4F9-240, C6F13-241 and C8F17-242, obtained by routine protection deprotection sequences, with a mixture of a trichloroacetimidate donor bearing a C3F7-tag ester group 243 and the peracetyl 2-azido-2-deoxyglucosyl trichloroacetimidate 162 in the presence of TMSOTf in DCM provided a mixture of six disaccharides 244–249. Preparative fluorous silica gel HPLC resolved all six products, eluted in the order of total fluorine content of the tags. In most cases, it was also possible to resolve α,β mixtures, as well as unreacted starting materials and byproducts. This method has the advantage of a single reaction mixture and separation for the synthesis of six products and could be applied broadly to the synthesis of oligosaccharide libraries, given the wide range of lengths of fluorous tags available.
6 Iterative protecting group fluorous tags
6.1 Iterative double tag reducing-to-non-reducing oligosaccharide synthesis
Iterative reducing-to-non-reducing oligosaccharide synthesis requires a glycosyl donor with an “iterative” protecting group, which can be deprotected in the presence of all other functionalities in order for the free hydroxyl to react with the next glycosyl donor. Although any of the fluorous late stage protecting groups above can in principle be used as an iterative protecting group, only the Pohl group has explicitely invoked oligosaccharide synthesis with an iterative fluorous group, using a double fluorous label strategy for the synthesis of N-acetyl glucosamine oligomers. Park et al. thus selected an ester-fluorous protecting group for its simple and readily automated deprotection conditions (Scheme 16).80 The fluorous ester-tagged glucosamine trichloroacetamide donor 254, prepared from 250, was coupled to the anomeric fluorous tag 60 with TMSOTf in toluene at 0 °C to provide the doubly-tagged compound 255. Efficient separation was achieved by FSPE, eluting with increasing proportions of methanol in water; non-fluorous reagents were eluted in 80% methanol, followed in 90% methanol by the excess fluorous ester-tagged donor in its hydrolyzed form, along with traces of the unreacted alcohol 60; the desired doubly-tagged product was eluted last from the fluorous cartridge using 100% methanol as eluent. Deprotection of the fluorous ester tag and isolation of the glycosyl acceptor by simple extraction with toluene sets the stage for iterative reducing-to-non-reducing synthesis of oligomers of N-acetylglucosamine using the glycosyl donor 254. The ability to recover and recycle the excess glycosyl donor represents an important advantage. Efforts to adapt this methodology to automated iterative oligosaccharide synthesis are reportedly underway.
In a related double-tag approach, the first chemical synthesis of the conserved oligosaccharide unit corresponding to the dengue virus was achieved by the Liu group (Scheme 17). Zhang, Liu and Liu reported that coupling a light-fluorous-tagged glycosyl donor 259 to the untagged disaccharide acceptor 260 gave the tetrasaccharide 261.81 The fluorous benzoyl group was designed and synthesized in order to withstand the acidic coupling conditions and yet be removable with a catalytic amount of base. The di-fluorous tagged strategy allowed the authors to efficiently separate the desired tetrasaccharide 261 from the mono-fluorous-tagged donor 259 and the mono-adduct trisaccharides, as well as from the untagged glycosyl acceptor, by FSPE using a water–methanol-dichloromethane eluent. The convergent synthetic strategy offers high synthetic efficiency and significant savings in both time and solvents. In this case, activation of the thioglycoside would allow for fluorous-tag-assisted bidirectional oligosaccharide synthesis.
7 Activatable fluorous thioglycoside groups
Non-reducing-to-reducing and bidirectional oligosaccharide synthesis requires the presence of an activatable group, which is inert as the sugar acts as a glycosyl acceptor, and can then be activated into a glycosyl donor for the next iteration. In principle, any of the anomeric protecting groups described in Section 3 can be cleaved selectively providing the free sugar which can then be activated. However, thioglycosides are particularly popular activatable groups, as they are relatively inert, yet can be activated oxidatively. Goto, Nuermaimaiti and Mizuno reported the synthesis of a heavy fluorous tagged thioglycoside and showed its ability to act as both glycosyl donor 264 and glycosyl acceptor 269 (Scheme 18A).82 The fluorous thioglycoside 264 was successfully used as a glycosyl donor by the action of NIS/TMSOTf promotor system, giving the disaccharide 265 in 71% yield. The same thioglyocoside was converted to its 3-O-acetyl-6-O-benzyl derivative 269 in order to serve as a heavy fluorous acceptor. Glycosylation with 2-deoxy-2-phtalimidoglucosyl trichloroacetimidate donor 270 with Cu(OTf)2 in DCM at 0 °C gave the LacdiNPhth disaccharide 271, a cancer-specific carbohydrate marker. Activation of this thioglycoside disaccharide in the presence of Br(CH2)6OH, NIS and TMSOTf in DCM at 0 °C delivered the disaccharide 272 with an alkyl bromide chain for immobilisation in 39% yield over six steps by partitioning between HFE7100 : FC72 (2 : 1) and 95% aq. MeCN and a single column chromatography. These results clearly establish the use of this heavy fluorous tag for non-reducing-to-reducing oligosaccharide synthesis.
Jing and Huang developed a light fluorous anomeric thioglycoside activatable group 275, which after glycosylation and protecting group manipulation provided armed and disarmed fluorous glycosyl building blocks 278–280 (Scheme 18B).83 The thioglycoside was stable under esterification, etherification, and deacetylation conditions, but showed excellent reactivities in glycosylation reactions using different promotor systems such as N-iodosuccinimide (NIS)/AgOTf, NIS/TfOH, and p-TolSCl/AgOTf to give disaccharides 281–286. The fluorous chain allowed for facile purification of the thioglycosides by fluorous solid-phase extraction, and the fluorous thiol was recovered as the disulfide after the glycosylation reaction and could be recycled. Although the authors did not explicitely demonstrate its use as a glycosyl acceptor, these results nonetheless set the basis for non-reducing-to-reducing flourous tag-assisted iterative oligosaccharide synthesis.
8 N-protecting groups
Carbohydrates with an amino functionality are the second most abundant group in the realm of sugar biology. The polarity, basicity, or reactivity associated with the free amino, N-acetyl or N-sulfate functionality of naturally occurring derivatives renders them incompatible with many, if not most synthetic reaction conditions. Hence, the necessity for protection. Amino sugars are commonly protected using allyloxycarbonyl chloride (AllocCl), phthalic anhydride (Phth), 2,2,2-trichloroethyl chloroformate (TrocCl), and trichloroacetyl chloride (TCACl) to form carbamates or amides. Alternatively the amino group can be carried as an azide (N3). Most importantly, C2-participation by the protecting group of 2-aminosugars play a key role in controlling the stereoselectivity of the glycosylation reactions, but the possibility of forming a relatively stable oxazoline precludes the use of the acetyl group itself, as well as many of the more common amino protecting groups.84–86 In most cases, deprotection of these groups in the target oligosaccharide affords the free amine, which is derivitized to the final N-acetyl or N-sulfate moiety. The trichloroacetyl amide, on the other hand, is transformed directly into an N-acetyl group in one step by reductive dehalogenation using AIBN and Bu3SnH.87 Unfortunately, inseparable mixtures may result from the incomplete conversion when applied to higher oligosaccharides (e.g., hexamers), and removing the tin by-products can be a burden. Phthalimido groups require rather stringent deprotection conditions, which may lead to base-catalyzed side reactions, in particular in the case of uronic acids.88 The Alloc group can be removed under mild conditions, but does not tolerate N-iodosuccinimide and related halonium ion-mediated activation strategies of thioglycosides. To date, a small number of fluorous variants of nitrogen protecting groups that are amenable to fluorous separation and stereoselective glycosylation protocols of amines have been proposed. Of significant note is the development by Manzoni and Castelli of a new fluorous protecting group, Froc, analogous to the trichloethoxycarbonyl (Troc) group frequently used in carbohydrate synthesis (Scheme 19A).89 d-Glucosamine hydrochloride 287 was reacted with the FrocCl tagging agent 288 and NaHCO3/H2O to give the corresponding Froc-protected glucosamine 289 in excellent yield. Subsequent protection steps afforded the fluorous glycosyl acceptor 294, which was coupled with tetra-O-acetylgalactose trichloroacetimidate 48, mediated by TMSOTf in CH2Cl2, to furnish the desired fluorous tagged Gal-β1,3-GlcNAc disaccharide 295 in 85% yield. Purification of the products was generally performed by fluorous solid-phase extraction techniques (F-SPE), but standard chromatographic purifications are also possible. The Froc group was cleanly removed by reductive elimination with Zn and acetic anhydride to the N-acetyl glycosamine disaccharide 296. A similar base-labile fluorous N-protecting group (FOC) was also reported by Roychoudhury and Pohl (Scheme 19B).90 Their chemical precursors could be made from nontoxic reagents amenable to scale up. The FOC N-methyl imidazolium reagent 297 was added to tetra-O-acetyl glucosamine 298, leading to the FOC protected monosaccharide 302 after several common protection/deprotection steps that the urethane was found to survive, including basic and acidic conditions. Given the stability of FOC under Zemplén conditions, a strongly basic medium was used for deprotection, as the free amine was obtained using refluxing 6 M aqueous sodium hydroxide and subsequently acetylated to provide the N-acetylated moiety 303. The FOC group was found to be an effective C2-participating group under BF3-etherate mediated glycosylation (Scheme 20).
Two new fluorous photolabile protecting groups (FNBC and FNB) were also reported by Roychoudhury and Pohl (Scheme 20).90 A linker 304 bearing an amine protected with a FNB (fluorous o-nitrobenzyl) group was coupled with the FOC-protected tetra-O-acetylglucosamine 299 above to provide the doubly tagged molecule 305 with C2-participation of the FOC group. The FNB could be cleaved in high yield in the presence of FOC by light sources of wavelengths ranging from 300 to 365 nm, to provide N-FOC glucosamine 306. A related fluorous variant of the Cbz group, named FNBC (fluorous o-nitrobenzyl carbamate) group 307, was also prepared to facilitate stereoselective glycosylation through anchimeric assistance (Scheme 20). The FNBC group was introduced onto glucosamine tetraacetate 298 under mild conditions to give 308 , which was converted to 311, showing the group's tolerance to various protection/deprotection conditions and its ability to act as a C2-participating group for the introduction of p-methoxybenzyl alcohol. Unfortunately, the photochemical deprotection was relatively low yielding, but it should be possible to cleave the group under non-photochemical conditions.
9 Phosphate protecting groups
A fluorous protecting group to facilitate purification in the synthesis of sugar phosphates was reported by Liu and Pohl (Scheme 21).91,92 Its application to the synthesis of the disaccharide 319 from Leishmania demonstrated its stability towards glycosylation protocols and its orthogonality to other protecting groups. The fluorous tag was introduced using the phosphoramide 313 followed by oxidation. The phosphorous group provides an additional handle for monitoring by 31P NMR. Glycosylation of 316 with tetra-O-benzyl-d-glucosyl trichloroacetimidate gave the disaccharide 318 in 94% yield. Deprotection was successful under mild reducing conditions using Zn/NH4HCOO in CH3CN/THF within 2 h. The fluorous phosphate group provides for easy purification of the intermediates and an alternative to late-stage installation of the phosphate group.
10 Conclusions and future developments
The growing demand for efficient and rapid construction of oligosaccharides in a diversity of contexts has incited emerging progress in exploring novel reagents, protecting groups, solvent systems, and purification methods, as well as various possibilities for automation. Fluorous-tag-assisted synthesis has contributed significantly to this effort. Highly fluorinated heavy fluorous tags for liquid–liquid extraction may find important applications in industrial contexts, and have undoubtly set the blueprint for later developments in the field. Solid-phase extraction with fluorous stationary phases have rendered the methodology more practical and amenable to automation. Both light of heavy fluorous analogs of most of the protecting groups used in carbohydrate chemistry, such as benzyl, acetate, p-methoxybenzyl, and silyl, have been developed. We have focused here on the apects related to oligosaccharide synthesis rather than on the preparation of the labels, which has been discussed elsewhere.39 The fluorous moiety is typically electronically isolated from the reacting center, so that the reactivity is sufficiently analogous to the non-fluorous group to be able to adapt fluorous protecting groups to existing synthetic schemes with little or no modification, while providing very significant advantages for the purification at each step, which is usually the most time-consuming part of carbohydrate synthesis. Anomeric fluorous tags, activatable fluorous thioglycosides, and fluorous protecting groups for amino and phosphate groups complete the arsenal available to the carbohydrate chemist. Iterative oligosaccharide synthesis in both reducing-to-non-reducing and non-reducing-to-reducing sense have been demonstrated, which fully opens the way for bidirectional synthesis. The fluorous groups are compatible with most routine transformations and with robust solution glycosylation methods under existing reaction conditions, thus paving the way for rapid, high yielding, cost effective, environmentally compatible oligosaccharide synthesis. The products can be isolated by fluorous solid-phase extraction or fluorous high pressure/performance liquid chromatography at low cost with strong resolving affinity. Unfortunately supply issues for fluorous protecting groups and fluorous stationary phases have recently hindered somewhat the broader application of this methodology, although lab-scale preparation of fluorous silica gel is remarkably straightforward, and should not be overlooked. Fluorous chemistry has progressively been reengineered over the last decades to suit carbohydrate synthesis and the continued effort will hopefully fulfill its full potential for the efficient chemical synthesis of defined complex oligosaccharides of biological importance. In the short term, there remain opportunities in several areas: in applying new purification methods and strategies to fluorous tag-assisted oligosaccharide synthesis; in developing ever simpler fluorous tags; in taking advantage of the electronic effects of the fluorous group in controlling carbohydrate reactivity; and, more broadly, in fully integrating fluorous tag methodology into the full range of oligosaccharide synthetic strategies. Reducing oligosaccharide synthesis to a mature, “plug and play” technology accessible to all is an important goal that will require, and therefore will drive, a much deeper understanding of the structural basis for reactivity, stereoselectivity, and regioselectivity in carbohydrate chemistry. In the meantime, its compatibility with existing synthetic strategies, methods, and purification protocols make fluorous tag-assisted synthesis an attractive tool that should be adopted by all carbohydrate research groups.
The Laboratoire Chimie Organique 2 was founded by Prof. Gérard Descotes in 1968.