Direct Dehydrative Glycosylation of C1-Alcohols

Abstract

Due to the central role played by carbohydrates in a multitude of biological processes, there has been a sustained interest in developing effective glycosylation methods to enable more thorough investigation of their essential functions. Among the myriad technologies available for stereoselective glycoside bond formation, dehydrative glycosylation possesses a distinct advantage given the unique properties of C1-alcohols such as straightforward preparation, stability, and a general reactivity compatible with a diverse set of reaction conditions. In this Focus Review, a survey of direct dehydrative glycosylations of C1-alcohols is provided with an emphasis on recent achievements, pervading limitations, mechanistic insights, and applications in total synthesis.

Graphical Abstract

1. Introduction

Accumulating insights into the numerous roles of carbohydrates in cellular signaling processes and disease progression designate this class of molecules as an important source of inspiration for novel drug design.^[4] For example, carbohydrate-mediated signaling events are implicated in the angiogenesis,^[6] chronic inflammatory disease,^[6d,7] bacterial^[8] and viral infection,^[10] and neurodegenerative illnesses, such as Alzheimer’s disease.^[11] The practical utilization of carbohydrate-based therapeutics and glycomimetics, however, continues to be hampered by limited accessibility due to synthetic challenges imposed by their intrinsically complex frameworks. The preparation of carbohydrate scaffolds stands at a marked disadvantage relative to other classes of biomolecules, for example, peptides and nucleic acids,^[12] due to the lack of a template-driven bio-synthetic model from which we can adapt synthetic methodologies. Carbohydrates assemble into oligosaccharides and gly-cans in diverse ways resulting in many different linkage points as well as decorated, multifaceted saccharide cores which further enhances chemical complexity (Figure 1).^[6d,13] To address some of these problems, chemoenzymatic methods have been employed; however, the limited substrate scope for optimized enzymes, in addition to high expenses involved as you enhance specificity, prevent the practical utilization of enzyme-mediated glycosylation.^[14] Thus, organic synthesis remains the preferred method for accessing this privileged class of molecules.

Figure 1.

Selected examples of bioactive glycosides prepared under dehydrative conditions: Quillaja saponin variants (1),^[2] jadomycins (2),^[3] hygromycin A (3),^[5] and nucleoside antibiotic A201A (4).^[9]

Access to complex carbohydrate structures is enabled through chemical glycosylation and the principle challenge within preparative glycochemistry is developing general, stereoselective, and efficient glycosylation strategies. Despite its importance, the majority of current methodologies retain the same core synthetic obstacles confronted by researchers first attempting chemical glycosylation. Specifically, substrate-controlled stereoselectivities and cumbersome pre- and post-synthetic protecting group manipulations are still apparent in many methods. Most strategies for installing a new glycosidic bond are, thus, limited. A universal glycosylation method that has controllable and predictable anomeric selectivities, is applicable to a wide range of substrates, and is synthetically efficient is yet to be established, if ever attainable.

1.1. General aspects of chemical glycosylation

A significant hurdle in the development of a general protocol is inherent difficulties in the anomeric functionalization of the glycosyl donor. The glycosyl donor (5) in a coupling pair is typically the electrophile and needs to be activated in a way that facilitates the glycosylation reaction (Scheme 1a). Traditionally this is achieved by installing a more electronegative group at the C1 carbon thereby generating an activated donor with a latent anomeric leaving group (6). Such a manipulation serves to enhance the reactivity of the otherwise inactive donor prior to coupling with the nucleophilic acceptor to give the glycoside product (7).^[15] Over the last several decades, many glycosyl donors have been developed and their various mechanisms investigated (Scheme 1b).

Scheme 1.

Chemical glycosylation through selected isolable glycosyl donors.

In contrast to traditional glycosylation strategies, dehydrative glycosylation^[16] entails the direct utilization of the C1-hydroxy functionality in glycoside bond formation (Scheme 2). C1-hydroxy sugars represent some of the simplest carbohydrate donors and many classical approaches to glycosylation involve anomeric manipulations from this reactive center. For example, the Fischer glycosylation^[17] is a foundational transformation in preparative carbohydrate chemistry and involves the reaction of the C1 hydroxy of a carbohydrate donor with an alcohol acceptor in the presence of a catalytic amount of acid. The resulting O-glycosidic linkage is formed with the concomitant loss of water leading to the name dehydrative glycosylation. Since then, the term dehydrative glycosylation has developed to encompass the set of strategies wherein a C1-hemiacetal serves as the glycosyl donor and, in a one-pot procedure, is electrophilically activated to give reactive intermediate 8 in situ and then cross-coupled to a nucleophilic acceptor (Scheme 2b). The unique advantage in this glycosylation approach is the direct engagement of the C1-hydroxy which allows for the condensation of anomeric derivatization and activation/coupling into a single step without the need for isolation of intermediates (Scheme 2a). In addition to reducing the total number of synthetic steps thus enhancing the process’ overall efficiency, the direct use of C1-hemiacetals in glycosylation capitalizes on the unique properties of anomeric alcohols which readily react under mild activating conditions. Furthermore, these conditions usually tolerate wide range of other functional groups, specifically other glycosyl donors, making anomeric alcohols useful synthetic building blocks in orthogonal and iterative glycosylation techniques.^[18]

Scheme 2.

(a) Dehydrative glycosylation overview. (b) Electrophilic activation of C1-hydroxy donor.

Despite these advantages, dehydrative glycosylation is much less developed compared to other glycosylation methodologies. This is largely due to challenges imposed by substrate-dependent reactivities and selectivities. Past developments in this methodology relied on various strategies generally involving one of two ways for donor-controlled, stereoselective glycosylation: (1) direct activation of the C1 hydroxy under mild (Brønsted) acidic conditions or (2) electrophilic activation generating a transient leaving group in situ.^[15b,16,19] Recently, methodologies for C1-hemiacetal glycosylation have further expanded on electrophilic activation methods with phosphoniumand pyridinium-derived salts^[20] as well as strategies which capitalize on the different reactivities of glycosyl pairs with various promoter systems,^[21] with a focus on developing more reagent-controlled techniques for dehydrative glycosylation.

Presented herein is an overview of direct dehydrative glycosylation strategies from C1-hydroxy donors. The aim of this Focus Review is to primarily discuss those methodologies in which the anomeric oxygen remains intact through the process of anomeric derivatization with an emphasis on developments in the last 10 years. Methodologies that proceed through glycosyl halides or other anomeric derivatives in which the anomeric alcohol is not incorporated into the active intermediate or result in stable, isolable intermediates are well-established and previously reviewed,^[15b,16] but are not the major focus of discussion in the following sections.

2. Mechanism of glycosylation and issues of stereoselectivity

Advancement in glycosylation methodology involves the competing aspects of reactivity or reaction efficiency and anomeric stereoselectivity. These two factors seem to be mutually exclusive with more reactive anomeric species resulting in high yields and a mixture of α/β anomers (stereochemical scrambling) and less reactive, more stable species giving lower overall yields but higher selectivities.^[15b] While both components are important when considering methodological advancement, access to the correct diastereomer in sufficient purity takes precedence as effective sugar mimics must adopt conformations similar to their natural counterparts in order to maximize their use as therapeutics or imaging tools.^[22]

2.1. Mechanistic considerations for anomeric selectivity

In short, the glycosylation reaction is straightforward: the donor glycoside is activated with an electrophilic reagent, and the glycoside acceptor reacts with the activated donor, establishing a glycosidic bond. Additionally, nucleophilic attack of the acceptor often occurs on an oxocarbenium intermediate 8c which is the more likely intermediate for strongly activated donors, or in the case of strongly activating electrophiles (Scheme 3). Glycosylation, however, loses the façade of simplicity upon considering the outcome of the stereoselectivity of the anomeric bond. The capricious nature of the oxocarbenium intermediate, which has only recently been spectroscopically confirmed,^[23] results in a complex system of interconverting intermediates (e.g., 8a–c) being part of any glycosylation reaction (Scheme 3a). Factors such as solvent polarity,^[24] acceptor nucleophilicity,^[25] donor nucleofugality,^[19] donor conformational preferences,^[26] donor electronics (armed and disarmed glycosides),^[15c] and the presence of semi-stable anomeric intermediates - namely anomeric triflates^[27] - all influence whether a given glycosylation is α- or β-selective. While the full consideration of the precise means of mechanism is beyond the scope of this review and covered thoroughly elsewhere,^[28] we will highlight important aspects relevant to dehydrative glycosylation.

Scheme 3.

(a) Potential glycosylation pathways depending on nature of intermediate (b) Diagram of anchimeric or neighboring group assistance to afford 1,2-trans-glycosides.

C1-hydroxy donor activation occurs in situ, so that the electrophilically-activated anomeric alcohol undertakes a nucleophilic attack from the acceptor, resulting in the loss of water and the formation of an anomeric bond. Because electrophiles must be highly reactive to remove water, intermediates formed during dehydrative glycosylation share this strong reactivity, often resulting in poor selectivity and competing degradation pathways, namely, self-condensation. Selectivity can be achieved through anchimeric assistance of, for example, a C2 acetoxy (9) or acetamido (10) participating groups (Scheme 3b).^[29] These groups generate cis heterocyclic intermediates (11), which, when attacked from the opposite face of the acceptor, yield 1,2-trans-glycosides 12. Although reliable, these methods depend highly on donor manipulations which inherently limit the reaction scope for glycosylation. More general means of obtaining high selectivities rely on tempering intermediate reactivity with the addition of a halide salt. Through stabilization, these intermediates ensure stereoselectivity since the β halide is much more reactive than the α halide, encouraging an S_N2-type attack from the acceptor, thus furnishing the α-glycoside.^[30] Stronger acceptors, however, often favor a direct attack on the more stable α halide, delivering the β glycoside.

Thus, securing high stereoselectivity under dehydrative conditions depends on multiple interacting parameters many of which are dictated by the set of activation conditions employed for a given reaction. Dehydrative glycosylations discussed in the following sections focus on two modes of activation to enhance the nucleofugality of the anomeric alcohol: (1) direct activation using Fischer-type conditions or (2) direct activation with electrophilic reagents.^[19]

3. Direct activation of anomeric alcohols using acid-based promoter systems

3.1. Traditional acid-promoted C1-OH activation

The Fischer glycosylation and its variants continue to serve as reliable and straightforward methods in the preparation of simple alkyl and diglycosides 13 (Scheme 4). However, due to harsh reaction conditions like high temperatures and acidic media in addition to large excesses of glycosyl acceptor, classic Brønsted acid-promoted Fischer-type glycosylations fail to provide a practical platform from which we can access complex oligosaccharides and glycoconjugates.

Scheme 4.

Fischer glycosylation.

Advancements in acid-catalyzed activation of the C1-hydroxy focus on creating a set of conditions under which the mild and, preferably, catalytic activation of the anomeric hydroxy can occur. Such developments have included additives alongside the acid promoter to serve as desiccating reagents or acid scavengers as well as promote stereoselectivity. Early work by Koto and co-workers demonstrated effective glucosylation of simple alcohols and glycosyl acceptors in 65–85% yields and 0.8–1.4:1 α:β ratios using a promoter system of MeSO₃H (cat.) and CoBr₂.^[32] In a later study, they further elucidated mechanistic aspects of their methodology and provided ¹H NMR evidence for the intermediacy of glycosyl bromide in product formation as well as proposed a catalytic cycle (Scheme 5a,b). Improvements to the initial selectivity of the reaction were also pursued with enhanced α-selectivity observed from the addition of Et₄NClO₄ to the reagent mixture (Scheme 5c).^[31] Under the ternary promoter-system, α-glucosides were prepared with decent selectivity (α/β ratio ~3:1) in 63–88% yields.

Scheme 5.

Methanesulfonic acid-promoted glycosylation reported by Koto and co-workers.^[31]

In the same study, the extent of anomerization of β-isomer of 15 was investigated in the presence of TBABr and Et₄NClO₄ using CH₂Cl₂ as a solvent, as well as MeNO₂ without additives. The limited anomerization observed in the presence of TBABr further supported a glycosyl halide-mediated α-glucosylation under the proposed conditions; however, Et₄NClO₄ induced significant anomerization with an observed α/β ratio of 3.5:1 after two hours at room temperature suggesting hydrolysis of 14a and α-glucosylation through putative β-glycosyl perchlorate 14b (Scheme 5c). When anomerization was tested in MeNO₂ without additives, significant amounts of a-anomer 15a were detected which lead the authors to speculate the transient formation of the anomeric mesylate facilitated by an increase in solvent polarity (favoring S_N2) from the increased solubility of CoBr₂ in MeNO₂.

It is important to emphasize that, while these two studies constitute a very small segment of a much wider set of important methodological insights regarding direct dehydrative glycosylation, they helped established a foundation, pulling from the conceptual frameworks established by Lemieux,^[30] for studying the effects of acid-catalysis, halide salt additives, solvent, and reaction temperatures on C1-OH activation (Table 1).

Table 1.

Selected acid-catalyzed dehydrative glycosylation strategies.

Products	Reagent System	Solvent	Selectivity	Ref.
Alkyl- and diglycosides	MsOH/CoBr2	CH2Cl2	α-selective[b]	[31,32]
	MsOH/CoBr2/Et4NClO4	CH2Cl2
Deoxy and per benzylated a-pyranosides	H4SiW12O40[a]	MeCN	α-selective	[33]
α/β-olivosides	Montmorillonite K-10[a]	CH2Cl2	β-selective[c]	[34]
β-ribosides	[(MeCOCH2CO2H/Yb(OTf)3]	CH2Cl2	β-selective	[35]
O-, C-, N-glycosides	[Yb3+/Sn2+/La3+][OTf] CaSO4/HMSDO	MeNO2/CH2Cl2 (3:1)	β-selective	[36]
			α-selective[b]
α/β-nucleosides	TrB(C6F5)4	EtNO2	α-selective[b]	[37]
2-deoxy-glucosides	TrB(C6F5)4	PhH:PhMe	α-selective	[38]

^[a]Stoichiometric amounts required.

^[b]Achieved only with addition of LiClO₄.

^[c]Only observed with carbonyl-based C4-protecting group.

Despite the important insights gained from the work of Koto and co-workers, issues with narrow reaction scope as well as poor selectivities persisted. Due to the relatively acidic reaction media resulting from the use of strong, Brønsted acids (e.g., H₂SO₄, HBr), many sensitive and complex substrates were excluded as viable coupling partners. This poses no problem in the context of simple alkyl glycosides and disaccharides. However, for the preparation of complex oligosaccharides and glycoconjugates which often require late-stage coupling of multifunctional and expensive starting materials, this feature renders the approach unsuitable for accessing such targeted products. Thus, metal triflates and other Lewis acids emerged as milder alternatives.

High yields and good β-selectivities have been reported using promoter systems of catalytic Yb³⁺, Sn²⁺, La³⁺ salts as Lewis acid catalysts in the presence of HMDSO and anhydrous CaSO₄.^[36] When tested in a solvent mixture of MeNO₂/CH₂Cl₂, it was found that these catalytic promoters could be used in quantities as low as 10 mol%. In order to promote α-selectivities under these reagent conditions, the utility of perchlorate salts was again demonstrated. A combination of LiClO₄ and Sn(OTf)₂ successfully promoted the synthesis of α-C-nucleo-sides, presumably through the same mechanism illustrated in Scheme 5. Uchiro and Mukaiyama again exploited α-directing utility of LiClO₄ in combination with Lewis acidic trityl salt catalyst to deliver α-ribofuranoside in 73% and α:β ratio 96:4. The same trityl salt catalyst proved effective for α-glycosylations of 2-deoxy glucopyranose C1-hydroxy donors and furnished alkyl glycosides and disaccharides in excellent yields (82–93%).^[38]

Effective glycosylations of C1-hydroxy donors are also demonstrated via activation with Lewis acidic salts of Bi³⁺. Yamanoi and co-workers employed catalytic Bi(OTf)₃ (5 mol%) to promote the cross-coupling of various primary alcohols and 1-hydroxy sugars under reflux in CH₂Cl₂.^[39] When the reaction was run at room temperature using larger amounts (10 mol%) of catalyst, larger quantities of self-condensation products were observed.

Acid-catalysis is one of the most straightforward strategies for C1-OH activation and can deliver simple alkyl glycosides, glycoconjugates, and disaccharides in good yields. Stereocontrolled strategies typically exploit the unique properties of halide and perchlorate salts to direct a-selective cross-coupling as well as induce anomerization of the resultant glycoside. However, many strategies to reign greater control over highly reactive glycosyl intermediates and avoid stereochemical scrambling still heavily rely on donor-centered manipulations and acceptor nucleophilicity. Consequently, the scope of viable substrates under acid-catalysed conditions remains severely restricted. Techniques to overcome the narrow range of application and improve reaction compatibility with multifunctional and/or sensitive coupling partners are still needed and many of the methodological advancements in response focus more on direct activation using non-acidic electrophiles. However, noteworthy developments capitalizing on the unique properties of several acid-based, non-traditional reagent systems have been investigated.

3.2. Other strategies

The need for milder conditions of acid-promoted dehydrative glycosylations ushered in a series of alternative methods capitalizing on the unique properties of heterogeneous acid catalysis, surfactant-acid combined catalysis, and ionic liquids.

Heterogeneous acid reagents are commonly employed as environmentally-friendly alternatives to traditional acid catalysts.^[40] Compared with conventional protic acids, heterogeneous acid catalysts do not require the usual high reaction temperatures or large excesses needed to ensure maximum conversions. Expeditious glycosylations are possible at ambient temperatures and with reduced quantities of acid-promoter due to the high molar acidities of solid acids. Additionally, heterogeneous acids have the ability to be recycled after each use as they can be easily recovered via filtration from the reaction mixture. The preparation of various deoxyglycosides using solid acid catalysts such as montmorillonite K-10 and heteropoly acids such as H₄SiW₁₂O₄₀ is reported by Jyojima and co-workers^[34] and Matsumara and co-workers,^[33] respectively. In addition to being effective activating agents for the anomeric alcohol, these acid catalysts also functioned as dehydrating agents which removed water from the reaction mixture as it formed.

Surfactant-acid catalysts were investigated for the glycosylation of 1-hydroxy sugars and long-chain alcohols. Kobayashi and co-workers^[41] probed the effects of using dodecyl benzenesulfonic acid (DBSA) for the activation of C1-hemiacetals in O-benzyl protected pyranosyl and furanosyl C1-glycosyl donors. The same group later demonstrated this method in the stereoselective synthesis of C-nucleosides.^[42]

Kuroiwa and co-workers^[43] exploited the green properties of ionic liquids as non-volatile reaction media to enable glycosylation at high temperatures and under reduced pressure.^[15b,44] Monasson and co-workers^[45] investigated the glycosylation of various alcohols and amino acids under dehydrative conditions using 1-butyl-3-methylimidazolium triflate ([BMIM][OTf]) as the ionic liquid reaction medium. The application of this methodology to the glucosylation of serine derivatives positions this technique as expedient in the synthesis of glycopeptides; however, the reaction still suffers from lackluster anomeric stereoselectivities with α/β ratios of approximately 1:1.

While use of alternative, acid-based promoter systems have allowed for greater synthetic efficiency and milder conditions, a continued disadvantage of methods remains the loss of stereochemical integrity due to the high reactivity of key glycosyl intermediates. This results in undesirable diastereomeric mixtures in the product glycosides. While attempts to exert more control over highly reactive species and improve selectivity have been investigated through the use of additives,^[26] the stereoselectivities resulting from these approaches are still variable and largely dependent on the nature of the donor as well as the nucleophilicity of the acceptor and polarity of the solvent.

4. Direct C1-hydroxy activation using non-acid electrophilic reagents

In order for the direct utilization of anomeric alcohols to result in stereocontrolled glycosylation, conditions which favor the formation of a stable glycosyl intermediate must be considered in addition to those which confer enhanced reactivity. Thus, strategies to activate the C1-hydroxy using non-acid electrophilic reagents which generate active glycosyl intermediates are frequently investigated. The following section mainly focuses on developments in dehydrative glycosylation using pyridinium salts and carboxylic acid anhydrides (carbon-based electrophiles), triflic anhydrides and sulfoxonium triflates (sulfur-based electrophiles), and reagents generating glycosyl oxophosphonium intermediates (Appel, Mitsunobu, and Hendrickson derivatives). In all of these examples, the anomeric oxygen remains intact in the activated donor thus qualifying them as direct, dehydrative glycosylation techniques.

4.1. Carbon-based electrophilic activation

Mukaiyama and co-workers demonstrated successful dehydrative glycosylation of 1-hydroxy donors to yield N-glycosides (e.g., 18a,b) using pyridinium salts (e.g., 16 and 17) as carbon-based electrophilic promoters (Scheme 6a).^[46] These C-electrophiles proved to be effective in promoting N-glycoside bond formation through an S_N2-like manifold from the corresponding onium salt intermediate (21).

Scheme 6.

Dehydrative glycosylation using pyridinium salts. (a) Glycosylation reported by Mukaiyama, suitable with 1-hydroxy ribofuranose and pyranose donors to give N-glycosides. (b) Synthesis of 2-pyridyl-1-thioglycosides by Yoshida et al. promoted by DMC (20).

Further developments using pyridinium salt-promoted dehydrative glycosylations were reported by Shoda and co-workers, who implemented dehydrative glycosylation using 2-chloro-1,3-dimethylimidazolinium chloride (20) to furnish 2-pyridyl-1-thioglycosides 22 with 1,2-trans-selectivity (Scheme 6b).^[20b] Moreover, this method was compatible with free carbohydrates (19) and aqueous reaction media. The possible mechanistic pathways outlined in Scheme 7. When the C2-position contained a free amine or hydroxy, high β-selectivities were hypothesized to arise due to the likely intermediacy of 1,2-anhydro donor 23. Likewise, employing donors with a carbonyl-protecting group at C2 (i.e. NHAc, OAc) also provided similar results and, as anticipated, these selectivities were lost when the reaction involved 2-deoxy substrates (Scheme 7b).

Scheme 7.

Plausible mechanism for 1,2-trans-selectivity in DMC-mediated dehydrative glycosylation. (a) Formal attack on α-glycosyl onium salt. (b) Intramolecular attack from C2-carbonyl or free OH, NH₂ group to facilitate β-attack from nucleophilic S atom.

Carboxylic acid anhydrides can also promote glycosylations of C1-hydroxy donors. Jeon and co-workers established the use of unsubstituted and monosubstituted phthalic anhydrides 25 to facilitate 1,2-cis-glycosylations of benzylidene-protected mannose to give β-mannosides 27.^[47] Success of this method, though, was hampered by the competing self-condensation pathway leading to the formation of 28. To overcome this issue, 3-fluoro substituted 25b was found to sufficiently suppress the propensity toward self-condensation by stabilizing putative intermediate 26 (Scheme 8).

Scheme 8.

Phthalic anhydride promoted dehydrative glycosylation of benzylidene protected mannose.

Bennett and co-workers probed the glycosylation of deoxy-C1-hemiacetals promoted by cyclopropenium cations (31, 33, 34) (Scheme 9).^[48] They reported excellent selectivity for 2,6-dideoxy (29), 2-deoxy (30), as well as 2,3,6-trideoxy (32) donors-all of which are notorious for their uncooperative nature toward well-established stereoselective glycosylation methodologies. NMR studies^[21b] revealed the most likely mechanistic pathway involved the displacement of the cyclopropenium-activated hydroxy functionality by the halide counterion furnishing the product glycosides with good selectivity from the anomeric halide through an S_N2-like manifold.

Scheme 9.

(a) Cyclopropenium-assisted dehydrative glycosylation for α-selective glycosylations of deoxy and dideoxy C1 alcohols. (b) Extension of cyclopropenium-mediated glycosylation to 2,3,6-trideoxy donors.

Despite deviating from the framework of direct dehydrative glycosylation, the work by Bennett and co-workers represents an effective and interesting approach stereocontrolled glycosylation of C1-hemiacetals.

4.2. Sulfur electrophiles: triflic anhydride-, sulfonyl halide-, and sulfoxonium triflate-mediate dehydrative glycosylation

The strong electron-withdrawing properties^[49] of triflates are frequently exploited to promote dehydrative glycosylation. At low reaction temperatures and in the presence of a sterically hindered base, C1-hemiacetals react with triflic anhydride to furnish reactive anomeric triflates.^[50] However, the instability of the glycosyl triflate generated typically leads to a/b mixtures and low yields due to significant self-condensation. Decomposition due to self-condensation can be mitigated by addition of tetra n-butyl ammonium bromide or by adding excessive amounts of acceptor. A recent development in this methodology was reported by Shen and co-workers,^[51] who demonstrated that strained olefin 37 could be used as a non-nucleophilic cation scavenger to enable Tf₂O-mediated dehydrative glycosylations of benzylated C1 alcohols 36 to give glycosides 38 in modest yields and selectivities (Scheme 10).

Scheme 10.

Triflic anhydride-promoted dehydrative glycosylation enabled by strained olefin.

Alternative methods derived from sulfur-based electrophilic activation focus on installing transient glycosyl sulfonates to facilitate glycosylation. Leroux and Poulin^[52] found that simple O-alkyl glycosides were attainable via mesyl chloride-promoted dehydrative glycosylation; Szeja similarly found tosyl chloride effectively mediated the coupling of anomeric alcohols and alkyl acceptors in aqueous solution under phase transfer conditions.^[53] Koto and co-workers investigated activation of the anomeric alcohol using this promoter system consisting of p-nitrobenzenesulfonyl chloride, Et₃N (acid scavenger), and AgOTf as a Cl⁻ scavenger (Scheme 11).^[54] Using this ternary system, they were able to secure disaccharides 40 in good yield and β-selectivities; however, selectivities were primarily found to correlate with substrate-mutable properties such as those conferred by the appropriate C2-participating groups. Self-condensation became a viable decomposition pathway when the scope was tested on glycosyl acceptors with sterically hindered alcohols. Their method was demonstrated in the syntheses of several branched-chain oligosaccharides including common structural motifs found in the antibiotic saponins^[55] as well as glucobiose derivatives.^[56] The highly substrate-dependent reactivities could be subdued via introduction of N,N-dimethylacetamide to the mixture to furnish α-selectivities in good yields.

Scheme 11.

Dehydrative glycosylation through glycosyl sulfonates (Ns=nosyl group).

Gin and co-workers pioneered the use of the highly reactive oxosulfonium reagent 45-generated in situ from the corresponding sulfide or sulfoxide reagent and triflic anhydride-as effective promoters of dehydrative glycosylation (Scheme 12).^[57] They found that stable glycosyl oxosulfonium intermediates 44 can be accessed by reactions of glycosyl alcohols with Ph₂SO/Tf₂O promoter-system at low temperature.^[58] The activated oxosulfonium donor 42 (verified using ¹⁸O-labelling techniques and ¹H NMR)^[58b] reacts with the nucleophile furnishing glycosides 43 in good yields and selectivities. Despite the promising regeneration of diphenyl sulfoxide following nucleophilic attack for potential extension of this promoter into catalytic activation, NMR experiments revealed that competing self-condensation pathway warranted excess quantities of Ph₂SO to ensure maximum yields at low temperature. Later work expanded on the foundational studies with Ph₂SO/Tf₂O and ventured into covalent catalysis mediated by alkyl sulfide reagents.

Scheme 12.

Dehydrative glycosylation using Ph₂SO/Tf₂O promoter system.

Gin’s work represents an important perspective into C1-hydroxy activation and dehydrative glycosylation as well as provided critical mechanistic insights into the activation pathways prevalent in anomeric alcohol activation.^[57c] However, the anomeric selectivities do not pose noteworthy advantages over previous methods. Thus, in order to secure adequate anomeric selectivities, protecting group manipulations around the saccharide core were necessary to favor formation of specific anomers. For example, 1,2-trans-β-selectivities were achieved through a C2-acyl substituent (ester or amide) as directing group or through a benzylidene protecting group on C4 and C6 which are known to furnish 1,2-cis-mannosylations.^[59]

The promoter system developed by Gin and co-workers has been applied to total synthesis of natural products such as the saponin-derived vaccine adjuvants (Scheme 13).^[1,60] In the first stages of the synthesis, dehydrative glycosylation was utilized to combine rhamnospyranoside 46 and xylopyranoside 47 to furnish 48 in 84% yield with high α selectivity. Additionally, rhamnose building block 49 was combined with 50 under similar conditions to furnish the β-disaccharide 51 (75%).

Scheme 13.

Gin’s synthesis of two units in vaccine adjuvant QS-7-Api.^[1]

4.3. Activation by phosphorus: applications of Appel, Mitsunobu, and Hendrickson-type reagents

Nifant’ev and co-workers^[61] reported the use of Appel reagents (Scheme 14) for the construction of fucosyl bromides, specifically noting that subsequent α-glycosylations with mercuric halide salts were undertaken without further treatment or purification of these privileged intermediates. Kobayashi later condensed this methodology into one-pot procedure wherein C1-alcohols (53) were treated with triphenyl phosphine/CBr₄ followed by addition of the acceptor, Et₄NBr, and catalytic N,N-tetramethylurea (TMU) to promote α-glycosylation via halide-ion and nucleophilic catalysis (Scheme 15).^[62] TMU was included, alongside the extra Br⁻ equivalent, to mediate anomerization reactions amid likely intermediates 52 and 54, thus maximizing the yield of the desired α-glycoside, 55. Other investigations using these conditions reported successful α-glycosylations of C2-benzyl-protected 1-hydroxy donors in DMF without halide salt or nucleophilic base additives.^[63] An additional study by Nishida and co-workers revealed that running the reaction in DMF induced the formation of α-glycosyl imidate halide salts, replacing glycosyl halide 54 as the active intermediate.^[64] In all cases, the preliminary intermediacy of the glycosyl oxophosphonium 52 is surmised and, given its relatively high reactivity, rapidly converts into 54 or imidate salt. These attempts showcase the dehydrative abilities of phosphonium species generated in situ and provide evidence for expedited C1-OH activation under mild conditions capitalizing on the thermodynamic driving force of generating phosphine oxide byproduct.

Scheme 14.

Phosphorus electrophiles employed in C1-hemiacetal activation.

Scheme 15.

Dehydrative glycosylation mediated by Appel reagents.

Glycosylation under Mitsunobu conditions was initially reported by Jones and co-workers who applied it to the synthesis of C6 N-alkylphthalimides on glucopyranose systems.^[65] Given the compatibility of these conditions with acidic acceptors, Mitsunobu glycosylation was found to be highly effective for the synthesis of N-nucleoside derivatives.^[66] In addition to establishing the viability of the method with anomeric alcohols, Szarek established precedent for the anomeric oxophosphonium as a promising intermediate and could, possibly in combination with halide salt additives, provide a new pathway toward stereocontrolled dehydrative glycosylation. The Mitsunobu glycosylation from C1-OH has been demonstrated in the synthesis of N-, O-, and C-glycosides and reviews on its many applications are available in the literature (Scheme 16).^[15b,c,67]

Scheme 16.

Mitsunobu dehydrative glycosylation (a) reaction overview and (b) selected examples.

Similar to Gin, dehydrative glycosylation finds its way into total synthesis with this coupling strategy. Recently, Mitsunobu conditions were used in the total synthesis of nucleoside antibiotic A201A (Scheme 17). The total synthetic strategy entailed an elaborate, multistep sequence using both linear and modular methods. Thus, dehydrative tactics were logically implemented as the reduction in preparative steps, high yields, and good selectivities ultimately translated into greater synthetic efficiency which often eludes larger syntheses of complex natural products. More importantly, these conditions enabled the formation of a 1,2-cis-furanosidic linkage, a challenging construction to accomplish stereoselectively, to generate the central unit 61. Under conditions optimized by Donohoe et al, who applied it in the total synthesis of hygromycin A (3),^[5] building blocks 59 and 60 were effectively coupled to give 61 in 79% yield and good β-selectivity (α/β 1:10).

Scheme 17.

Key cross-coupling step between building blocks 60 and 61 in the total synthesis of antibiotic A201A.

Glycosyl oxophosphonium intermediates offer a way into stereocontrolled glycosylations from the direct activation of the C1-hydroxyl functionality. Such stereocontrol is conferred by stable glycosyl intermediates formed using diphosphonium electrophiles (e.g., 63) and subsequent glycosylation under basic conditions. Reaction of phosphine oxide and triflic anhydride generated the highly reactive triphenyl phosphine ditriflates-acyclic phosphonium anhydride/bis(phosphonium) salts-similar to Hendrickson’s “POP” reagent (66).^[68] Originally employed as a general activating agent for any oxygen-bearing functional group, the capacity of diphosphonium salts to promote dehydrative glycosylation reactions was established by Mukaiyama and Suda in their synthesis of 1,2-cis-ribofurano-sides 64 (Scheme 18).^[69] In their work, ³¹P and ¹³C-NMR techniques provided evidence of an anomeric oxophosphonium intermediate which undergoes nucleophilic attack to furnish the product glycoside as well as an equivalent of phosphine oxide. Noteworthy observations regarding this approach are the short reaction times, high yields, and good α-selectivities using non-participating groups around the saccharide core.

Scheme 18.

Glycosylation developed by Mukaiyama et al. Bis(phosphonium) salts effectively promoted dehydrative glycosylation of protected furanoses.

The utility of Hendrickson’s reagent as an activator of 1-hydroxy pyranoses 66 was later demonstrated by Mossotti and Panza (Scheme 19a).^[20a] Per-benzylated C1-hydroxy pyranosyl donors 65a,b readily coupled with various acceptors which included sterically hindered alcohols as well as monosaccharides. Despite effectively establishing a precedent for the extension of phosphonium anhydride application to pyranose ring systems, the methodology proposed by Panza and co-worker were subject to several shortcomings. For instance, the oxophosphonium intermediate generated using Hendrickson’s reagent was relatively unstable leading to high degrees of self-condensation as well as significant scrambling of stereochemistry despite efforts to reign greater control over reactivity (e.g., reaction times, temperature, order of reagent addition). A later study by Filippov and co-workers, who attempted the stereoselective ribosylation of adenosine under the conditions developed by Mukaiyama without success, revealed methodological restraints regarding the reaction’s compatibility with sterically hindered alcohol acceptors.^[70]

Scheme 19.

Dehydrative glycosylation with phosphonium salt anhydrides. (a) Hendrickson’s ‘POP’ reagent used in glycosylation of C1-alcohols. (b) Cyclic phosphonium anhydride-mediated dehydrative glycosylation. (c) Selected glycosides prepared under phosphonium-salt promoted dehydrative conditions.

Inspired by the work of Mukaiyama and seeking to increase the viability of phosphonium salt-assisted dehydrative glycosylation, Walczak and co-workers investigated the efficacy of cyclic phosphonium anhydride salts as C1-OH activators (Scheme 19b).^[20c] Because of their broad functional group tolerance, convenient methods of formation, and facile modulation of structural and electronic properties, this class of reagents offers advantages over other dehydrative reagents. Furthermore, activation of the anomeric alcohol by formation of a mixed anhydride would lead to an intermediate with a cationic nature, thus diminishing its propensity to undergo unwanted dimerization because exchange at the phosphonium center with a glycosyl acceptor would be disfavored (Scheme 20).^[71] Thus, after probing the reaction conditions using various cyclic phosphonium anhydrides 71, it was determined that the cyclized derivative of 1,4-bis (diphenylphosphino)butane bis-oxide (74, DPPBO₂) effectively promoted dehydrative glycosylation, yielding O-, N-, S-, and C-glycosides (67–72). Beyond the wide scope of acceptors tolerated by these conditions, Walczak and co-workers demonstrated the method’s utility in the synthesis of axinelloside A building blocks^[72] as well as in orthogonal glycosylation strategies employing anomeric nucleophiles as a new type of glycosyl acceptor.^[73]

Scheme 20.

Mechanistic pathways of both cyclic and acyclic phosphonium anhydride promoted glycosylation.

Despite the high yields afforded under the optimized conditions, anomeric selectivities were suboptimal with initial results showing α:β diastereoselectivities of 1:1.2. As highlighted at various points in this review, as well as in the seminal work published by Lemieux,^[30] poor anomeric selectivities can often times be ameliorated by halide-ion assisted anomerization of the glycosyl intermediate and can result in more reactive β anomeric halides. Walczak and co-workers found that the putative anomeric phosphonium could undergo further activation from the addition of TBAI which furnished exclusive α-selectivity, however, at the expense of a high yield. The use of acyclic and cyclic diphosphonium salts show promise as a unique class of dehydrating agents. Further data provided suggests the mediation of a semi-stable anomeric phosphonium species capable of modulating a-selective glycosylation; however, more evidence is necessary to substantiate such claims.

5. Summary and Outlook

Over time, moving away from simply activating the C1 hemiacetal for glycosylation to employing dehydrative strategies allowed more mild reaction conditions and greater selectivities than ever before. The achievements of the strategic use of Brønsted and strong Lewis acidotic conditions cannot be understated for driving dehydrative glycosylation forward. These methods not only provided early insights into the stabilization of donor intermediates, but continued investigation provides milder and even catalytic means of activating the donor for glycosylation. Beyond this, many have capitalized on the strong thermodynamic driving force of the generation of ureas and phosphine oxides to provide even milder ways of dehydrating anomeric hydroxyl groups to provide a more significant scope to dehydrative glycosylation. Granted, self-condensation quickly reared as a threat to this methodology, but stabilization of highly reactive intermediates using glycosyl oxosulfonium and oxophosphonium species provided means to suppress this detrimental pathway.

While issues with variable yields, narrow substrate scopes, and modest anomeric selectivities persist, developments in the direct dehydrative glycosylations of C1-hydroxy donors by acid activation show significant improvement from original methods. Noteworthy successes involve the strategic employment of additives to promote stereoselectivity as well as influence other parameters controlling reaction outcomes. The combined implementation of acid activators with additives to confer stereospecific results considerably widened the utility of Fischer-type glycosylations in the controlled synthesis of complex oligosaccharides. There is still, however, a substantial need for further investigations into different strategies to improve dehydrative glycosylations. The challenges of stereochemical control at the anomeric position and reaction generality require more experimentation in order to provide a better understanding of the extent of this methodology and chemical glycosylation in general.

Biographies

Jacob Rodriguez received his B.A. in chemistry, biochemistry, and cell biology from the University of Colorado at Boulder. He has conducted research under the supervision of Prof. M. Walczak primarily focusing on carbohydrate synthesis, completing an honors thesis describing novel methods to synthesize carbohydrate-based immunostimulants. He will be pursuing a doctoral degree in chemistry at the Massachusetts Institute of Technology starting in the fall 2018.

Sloane O’Neill received her BA in biochemistry from the University of Colorado Boulder in May 2018. She started in the Walczak group in 2016 as an undergraduate researcher and has worked on several projects focusing on methodological development of novel strategies for installing non-traditional functional groups in carbohydrates and peptides. She currently works as a research assistant in the Walczak group and will be relocating to Boston in August 2018.

Maciej Walczak is an Assistant Professor in the Department of Chemistry and Biochemistry at the University of Colorado at Boulder. Prof. Walczak received his Ph.D. at the University of Pittsburgh and continued his education as a postdoctoral associate in Memorial Sloan Kettering Cancer Center. In 2013 he started his independent career focused on the synthesis and chemical biology of carbohydrates, peptides, and bioactive small molecules. Prof. Walczak leverages his expertise in organic chemistry to address present-day problems in chemistry, biology, and medicine.