The Road to Structurally Defined β-Glucans

Abstract

β-glucans are polymers of glucose that have been isolated from a variety of organisms. Isolated β-glucans have been used for medical purposes for centuries; however, efforts to define the biological activities of β-glucans experimentally were initiated in the 1940’s. The diversity of structure associated with isolated β-glucans has impeded said investigations, and efforts to leverage the biological activity of β-glucans for clinical applications. In recognition of the need for defined β-glucans that retain the biological activity of isolated β-glucans, considerable investment has been made to facilitate the synthesis of structurally defined β-glucans. Here, we review the different approaches that have been applied to prepare β-glucans. In addition, we summarize the approaches that have been utilized to conjugate β-glucans to proteins.

Graphical Abstract

β-glucans are polysaccharides composed of repeating glucose units that have been shown to have intriguing biological activity. In this review, the numerous strategies that have been applied to gain access to structurally defined β-glucans are discussed.

1. Introduction

Homopolymers of glucose, present in bacteria, fungi, algae, and plants, are described as β-glucans.^[1] β-glucans derived from natural sources are heterogeneous, and exhibit unique and heterogenous branching patterns. These structural differences impact the three-dimensional conformations, physical characteristics, and biological activities of β-glucans.^[2] The biological activities of β-glucans arise from interactions with receptors associated with the innate immune system (e.g., Dectin-1, CR3, and Ficolin-2),^[3] which has led scientists to primarily investigate the immunostimulatory activities of β-glucans.^[4]

To facilitate the development of β-glucans as immunostimulators, scientists have sought to thoroughly characterize the molecular interactions underpinning the professed immunostimulatory activity. This has initiated efforts focused on characterizing interactions between isolated β-glucans and immune receptors;^[5] however, the heterogeneity of isolated β-glucans significantly complicate such efforts.^[6] Synthetically prepared structurally defined β-glucans have permitted in-depth biophysical and structural studies focused on characterizing the interactions between β-glucans and innate immune receptors.^[3b,7] In addition to these fundamental studies, the structurally defined β-glucans have been utilized to prepare conjugates that have been investigated in preclinal models.^[8]

To date, an array of unique synthetic strategies have been employed to gain access to structurally defined β-glucans. Here, we review these different synthetic strategies, and current approaches utilized to conjugate β-glucans to proteins.

2. Chemical Synthesis

Syntheses of β-(1,3)-glucans have most frequently been described using solution phase chemistry. Linear syntheses have been employed for the preparation of smaller β-(1,3)-glucans, while the preparation of larger β-(1,3)-glucans are typically prepared through convergent strategies. Common to all synthetic efforts is the need to form the β-(1,3)-linkage between glucose residues. As such, the chemical synthesis of β-(1,3)-glucans requires a protecting group strategy that aids regio- and diastereselective glycosyl bond formation. To this end, the protecting groups at the 2-, 4-, and 6-positions must remain intact until the final global deprotection, while the protecting group at the 3-position must be selectively removable to facilitate chain propagation. Additionally, the protecting groups about the ring have been utilized to impart the desired β-selectivity.

2.1. Kitamura Synthesis

The first chemical synthesis of β-glucans was described by the Kitamura group (Scheme 1).^[9] In this report, the linear synthesis of β-(1,3)-glucans using a common glucose monosaccharide building block (1) is described. Further elaboration of this building block via the installation of a 3-O-chloroacetyl group, and functionalization of the anomeric center provided glycosyl chloride intermediate 2. It is imperative to highlight that the chloroacetyl group is central to this synthetic strategy, as it can be selectively removed to permit chain propagation. Activation of the glycosyl chloride 2 using silver carbonate (Ag₂CO₃)/silver perchlorate (AgClO₄) in the presence of methanol provided a methyl glycoside. This methyl glycoside would serve as the reducing end of the β-(1,3)-glucan. The chloroacetyl group selectively removed to provide glycosyl acceptor 3 and a subsequent glycosylation reaction involving methyl glycoside acceptor 3 and glycosyl chloride donor 2, followed by removal of the chloroacetyl, gave rise to intermediate 5. The synthesis of donor 4, was achieved via a silver trifluoromethanesulfonate (AgOTf) mediated glycosylation involving donor 2 and acceptor 1, and subsequent manipulation of the reducing end anomeric center.

Scheme 1.

Kitamura synthesis of linear β-(1,3)-glucans.

Chain propagation was initiated via a AgOTf mediated glycosylation reaction involving chloro-disaccharide 4 and disaccharide acceptor 5, which was carried out in the presence of 2,6-di-tert-butylpyridine (DTBMP). Subsequent removal of the chloroacteyl provides an intermediate that can be elaborated by repeating the glycosylation and deprotection sequence, to provide β-(1,3)-glucans with even number of glucose units. β-(1,3)-glucans with uneven numbers of glucose units were prepared in a similar manner; however, the synthesis was initiated with the glycosylation of methyl glycoside acceptor 3 with donor 4 to provide trisaccharide 7. Again, selective deprotection of the chloroacetyl, and a sequence of glycosylation and deprotection reactions, as outlined above, provide the desired polysaccharides. Finally, global deprotections were caried out via a two-step process, involving the acid mediated removal of the benzylidene groups, and a Zemplén deacylation. Purification of the resulting products was carried out using Bio-Gel P2 chromatography to provide the desired β-glucans.

This seminal work demonstrated the feasibility of preparing structurally defined β-(1,3)-glucans via chemical synthesis. Despite many advances in carbohydrate chemistry in the years preceding the dissemination of this work, the Kitamura groups approach remains among the most practical in terms of material throughput. The ability of the 4,6-O-benzilidene to impart the desired β-selectivity in glycosylation reactions was also highlighted in this approach and has been adopted by the majority of subsequent synthetic efforts.

2.2. Takahashi Synthesis

In order to investigate the interactions between β-(1,3)-glucans and Dectin-1, the Takahashi group prepared a series β-(1,3)-glucans of varying length, with and without 1,6-branching (Scheme 2).^[10] In this synthesis, the 4,6-O-benzylidene protected glycosyl acceptor 12 was regioselectively glycosylated with the levulinic ester bearing glycosyl donor 11 under Lewis acid promoted reaction conditions. The resulting intermediate was then protected using benzoyl chloride to provide intermediate 13. This disaccharide was then applied to gain access to both donor and acceptor species that would facilitate the synthesis of longer β-(1,3)-glucans. The acceptor 14 was prepared by removal of the levulinoyl group, while the donor 15 was prepared via desulfurization and conversion to the trichloroacetimidate. The disaccharide donor 15 and acceptor 14 were then submitted to a trimethylsilyl trifluoromethanesulfonate (TMSOTf) mediated glycosylation that provided a tetrasaccharide intermediate. This intermediate was acetylated to protect the free hydroxyl group, which gave rise to intermediate 16. This tetrasaccharide 16 was further elaborated at the reducing end via an N-iodosuccinimide (NIS)/triflic acid (TfOH) mediated glycosylation with the monotosyl-alkyl alcohol. Subsequent displacement of the tosyl group with sodium azide and cleavage of the levulinic esters provided acceptor 17. The glycosyl acceptor 17 and thioglycoside 16 where then subjected to an NIS/TfOH mediated glycosylation to yield an octasaccharide. Again, the free hydroxyl group was acylated and the levulinic esters were cleaved to provide 18. Sequential elaboration with the tetrasaccharide glycosyl donor 16, and analogous modifications provided the dodecasaccharide (19) and hexadeccasaccharide (20) intermediates.

Scheme 2.

Takahashi synthesis of linear and branched β-(1,3)-glucans.

The preparation of 1,6-branched β-(1,3)-glucans was initiated from intermediate 21, which was protected at the 6-position as the 2-(azidomethyl)benzoyl (AZMB) ester. The moiety pendent to the anomeric center was altered to a trichloroacetimidate, and was then applied to glycosylate acceptor 12. The free hydroxyl group of disaccharide 22 was protected as a benzoyl ester, the silyl ether was removed, and the intermediate was glycosylated with 14 to provide tetrasaccharide 23. The AZMB ester was then removed, and the resulting intermediate was elaborated with glycosyl donor 24. This pentasaccharide donor 25 was then applied to elaborate glycosyl acceptors 16, 18, and 19 to provide branched β-glucans 26a–c.

Global deprotections of the resulting β-(1,3)-glucans were caried out via a Birch reduction. From a technical standpoint, it is important to note that the Takahashi group used gel permeation chromatography, in addition to silica gel chromatography, for the purification and isolation of the polysaccharides of twelve or more glucose residues.

This work demonstrates the ability to efficiently prepare poly β-(1,3)-glucose chains that contain comparable numbers of glucose units to those observed in β-(1,3)-glucans isolated from natural sources. In addition, a feasible approach to incorporate branching via solution phase synthesis is described.

2.3. Vetvicka Synthesis

The synthesis of β-(1,3)-glucans reported by Vetvicka utilizes 1,2:5,6-di-O-isopropylidene-α-D-glucofuranose (Scheme 3).^[11] This starting material permits a facile route to selectively protect the hydroxyl group at C-3 as a naphthyl ether. Subsequent removal of the isopropylidene groups provided intermediate 29, which could be converted into thioglycoside 30 over 3 steps. The thioglycoside was further elaborated with a 4,6-O-benzylidene acetal and a benzoyl group to provide donor 31. Impressively, the synthesis of this donor was reported on 100 g scale with an average yield of 90% for each of the seven synthetic operations.

Scheme 3.

Vetvicka synthesis of β-(1,3)-glucans.

With this key donor in hand, they chose to cap the reducing end with benzyl alcohol, and subsequently removed the naphthyl ether using 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ), which provided acceptor 32. Chain propagation was carried out via series of NIS/triethylsilyl trifluoromethanesulfonate (TESOTf) mediated glycosylations and DDQ mediated naphthyl deprotections, which were successively repeated to provide intermediates 33a–b. A final NIS/Tin(II) trifluoromethanesulfonate (Sn(OTf)₂) mediated glycosylation provided the protected β-glucans. The authors note that significantly improved reaction yields were achieved by using Sn(OTf)₂ in place of TESOTf in this final glycosylation. A DDQ mediated deprotection was then carried out to cleave the naphthyl protecting group to provide 34a–c. A deprotection sequence involving the removal of the benzylidene groups, Zemplén deacylation, and lastly benzyl ether removal under reductive conditions afforded β-glucans 35a–c.

Vetvicka’s approach to synthesize β-(1,3)-glucans highlights the challenges associated with building polysaccharides in a linear manner using monosaccharide building blocks. Such an approach makes it challenging to obtain poly β-(1,3)-glucose chains comparable to those observed in isolated β-(1,3)-glucans. Nonetheless Vetvicka presents an efficient approach to access β-glucans building blocks that has been incorporated by subsequent syntheses.

2.4. Bundle Synthesis

The Bundle synthesis of β-(1,3)-glucans also started from 1,2:5,6-di-O-isopropylidene-α-D-glucofuranose (Scheme 4),^[12] which was protected with a benzyl ether prior to isopropylidene removal and global benzoylation. This intermediate (36) was then selectively deprotected at the anomeric center, and converted into key glycosyl donor 37. This donor was then applied to glycosylate 1,2:5,6-di-O-isopropylidene-α-D-glucofuranose giving rise to a disaccharide intermediate, which was subjected to hydrogenolysis to selectively remove the benzyl ether, thus providing intermediate 38. Further elaboration by repeating the glycosylation-deprotection sequence gave rise to trisaccharide 39. Further protecting group and reducing end manipulations provided donor 40. Unfortunately, attempts to apply 40 to prepare elongated structures via glycosylation reactions involving acceptors 38 or 39 led to consumption of donor 40 in the absence of desired product formation.

Scheme 4.

Bundle synthesis of β-(1,3)-glucans.

Disaccharide 38 was thus converted to 41 by removing the isopropylidene groups and subsequent benzoylation of the resulting free hydroxyl groups. The benzoyl groups decorating the C-4 and C-6 hydroxyl groups were then selectively removed, giving rise to an intermediate that could be reprotected with 4,6-O-benzildene acetals. Selective de-O-benzoylation was achieved using kinetic control, as the authors note that reactions byproducts increased with elongated reaction times. Disaccharide 42 was then treated with boron trifluoride diethyl etherate (BF₃·OEt₂) in the presence of ethanethiol to provide thioglycoside 43. Further functionalization of the reducing end using 5-hydroxy-pentanoic acid methyl ester, followed by the selective removal of the benzyl ether provided intermediate 46 in the absence of benzylidene acetal removal due to short reactions times. Further elaboration was achieved through sequential glycosylations and deprotections, which gave rise to the protected hexasaccharide 49. Finally, global deprotection provided the desired laminarahexose 50 carrying a functional handle that could be used to facilitate further studies.

The Bundle synthesis highlights the unpredictability associated with the synthesis of polysaccharide (i.e., glycosylation’s involving donor 40). However, reverting to incorporate the benzylidene acetal led to the efficient synthesis of a hexasaccharide. The use of kinetically controlled deprotections in this synthesis (i.e., de-O-benzoylations and de-O-benzylations) should be highlighted as they were carried out on gram scale and considerably enhanced route efficiency.

2.5. Liu and Costantino Syntheses

A series of β-(1,3)-glucan syntheses have used allyl ethers as reducing end protecting groups. In all cases, 2-benzoyl-4,6-O-benzyilidene trichloroacetimidate glycosyl donors carrying either levulinic ester or acyl protecting groups at the three position have been utilized. The synthesis described by Liu utilized a 2+3 strategy to prepare pentasaccharide 58 (Scheme 5A).^[13] In this work, a TMSOTf mediated glycosylation reaction involving donor 51 and acceptor 52 proceeded to provide disaccharide intermediate 53. Manipulation of the reducing end via a PdCl₂ mediated allyl deprotection, and conversion of the resulting hemiacetal into a trichloroacetimidate provided donor 54. The acetyl group of disaccharide 53 could be removed via a HBF₄ mediated acyl deprotection to provide intermediate 55, which could be further elaborated with donor 51 to provide a trisaccharide intermediate. Subsequent acetyl removal provided trisaccharide acceptor 56, which underwent a glycosylation reaction with disaccharide 54 leading to the formation of the globally protected pentasaccharide 57. The allyl ether was then epoxidized with meta-chlororperoxynbenzoic acid (m-CPBA), and a global deprotection was carried out over two steps to provide epoxide functionalized β-(1,3)-glucan 58.

Scheme 5.

A) Liu and B) Costantino synthesis of β-(1,3)-glucans.

An analogous synthetic strategy was also employed by Costantino et. al. leading to the preparation of laminarahexose (Scheme 5B).^[14] Here, the levulinic ester functionalized glycosyl donor 59 was utilized to glycosylate acceptor 60, and the resulting intermediate was deprotected to provide acceptor 61. This disaccharide was glycosylated with donor 59 to provide trisaccharide 62. Removal of the levulinoyl group provided acceptor 63, while facile manipulation of the trisaccharides reducing end provided trichloroacetimidate donor 64. With these intermediates in hand, a 3+3 strategy was applied using donor 64 and acceptor 63, which provided protected hexasaccacharide 65. Global deprotection was completed in a two-step process that importantly maintained the allyl group at the reducing end which was needed subsequent functionalization. Photochemical mediated modification of the reducing end with cysteamine provided an adduct (67) that facilitated conjugation to the non-toxic diphtheria toxin mutant CRM₁₉₇.

The Liu and Costantino syntheses both provide β-glucans with functional handles amenable for protein conjugation. It is interesting to note that glycosylation reactions involving O-allyl protected glycosyl acceptors proceed with lower-than-expected yields. Nonetheless, the use of di- and trisaccharide building blocks provide access to elaborate structures rather efficiently.

2.6. Ensley Synthesis

In this report, Ensley and co-workers utilized the 4-acetoxy-2,2-dimethylbutanoate (ADMB) protecting group to provide the β-selectivity necessary for the synthesis of β-glucans.^[15] The desired selectively would arise due to the formation of an orthoester intermediate (70) that would dictate selectivity (Scheme 6A).^[16] In order to validate the synthetic utility of this protecting group, an ADMB enabled protecting group strategy to synthesize laminarahexose was developed (Scheme 6B). Their synthetic efforts began with a glycosylation reaction, involving donor 72 and acceptor 73, which provided disaccharide intermediate 74. Subsequent removal of the silyl ether gave rise to intermediate 75 that was amenable to further elaboration. The disaccharide donor 78 that would be utilized to facilitate said elaboration was prepared via a glycosylation reaction involving glycosyl donor 76 and ADMB functionalized glycosyl acceptor 77. With these disaccharide intermediates in hand, Ensley and co-workers began to assemble laminarahexose (83). Specifically, a glycosylation reaction involving disaccharides 75 and 78 gave rise to tetrasaccharide 79. Removal of the silyl group provided intermediate 80, which could be further elaborated via a glycosylation reaction with donor 78 to provide hexasaccharide 81. Finally, a global deprotection provided laminarahexose 83.

Scheme 6.

A) Mechanistic rationalization for ADMB β-selectivity. B) Ensley synthesis of β-(1,3)-glucans.

The Ensley synthesis of laminarahexose proceeds in a highly efficient manner. The ADMB group, developed in the Ensley lab, is indicated to facilitate the high selectivity and yields obtained in this synthesis. Interestingly, analogous syntheses carried out by Takahashi and Bundle proceed with similar efficiency and selectivity in the absence of the ADMB group, thus the necessity for the ADMB group in this synthesis is unclear.

2.7. Li Synthesis

The Li group reported an ionic liquid mediated synthesis of laminarahexose (Scheme 7).^[17] This synthesis was initiated via the functionalization of the trichloroacetimidate donor 59 with an imidazolium bearing benzyl alcohol. Subsequent cleavage of the levulinic ester provided glycosyl acceptor 84 that could be elaborated via successive glycosylations. Four sequential cycles of BF₃·OEt₂ mediated glycosylations followed by cleavage of the levulinic ester before a final glycosylation, resulted in the fully protected hexasaccharide 85. Deprotection was carried out over three steps to provide laminarahexose.

Scheme 7.

Li ionic liquid mediated synthesis of β-(1,3)-glucans.

The authors demonstrate an efficient approach to prepare the protected hexasaccharide in 15 hours, with an average yield of 90% per glycosylation step. Additionally, this approach requires limited chromatographic purifications. However, the utility of this approach to prepare β-(1,3)-glucans that contain comparable numbers of glucose units to those observed in isolated β-(1,3)-glucans has yet to be validated.

2.8. Seeberger Synthesis

Seeberger and co-workers have reported syntheses of β-glucans using solid phase chemistry. In doing so, they are able to significantly reduce the number of purification steps.^[18] Initially, the Seeberger group used thioglycoside 87 (Scheme 8A),^[19] which was protected at the 3-position with an Fmoc group, as the building block for the synthesis of β-(1,3)-glucans. They elected to use an fluorenylmethyloxycarbonyl (Fmoc) protecting group as it is stable under acidic glycosylation conditions, can aid in glycosylation reaction monitoring, and can be selectively removed to facilitate chain elongation. Their initial proof of principle studies were carried out with an alkoxide labile linker, and they employed successive rounds of glycosylations, followed by cleavage of the protected β-(1,3)-glucans from the resin using sodium methoxide (NaOMe) in methanol. Their initial attempt to prepare a trisaccharide proved successful; however, attempts to prepare larger β-(1,3)-glucans via this protocol were unsuccessful. It is suspected that the 4,6-O-benzylidene acetal and pivalate ester protecting groups were responsible for the observed low reactivity.

Scheme 8.

Seeberger first generation solid phase synthesis of β-(1,3)-glucans. A) Proof of concept synthesis of β-(1,3)-glucans and B) synthesis of decasaccharide.

To overcome this challenge, the Seeberger group altered the glycosyl donor that was used to elongate the polysaccharide (Scheme 8B). In the second generation approach,^[19] they elected to use a glycosyl phosphate donor 91, exchanged the 4,6-O-benzylidene acetal for benzyl ethers, but retained the pivalate ester at the two position. In this synthesis they utilized a photocleavable linker to the Merrifield resin. With this altered synthetic strategy, the group was able to prepare the corresponding protected dodecasaccharide through twelve iterative glycosylations with a mean calculated glycosyltion yield of 89%. Subsequent cleavage from the resin and global deprotection, carried out over two steps, provided the dodecasaccharide β-(1,3)-glucan 95.

More recently, the Seebeger Group have reported a third-generation synthesis of β-(1,3)-glucans that permits the incorporation of 1,6-branching (Scheme 9).^[20] With the group’s previous knowledge, they employed a similar glycosyl donor (96), but substituted the pivalate ester for a benzoyl ester at the two position, and in order to facilitate branching, donor 97 which is functionalized at the 6-position with a levulinic ester, was applied. Through a series of synthetic operations, they successfully prepared an array of linear β-(1,3)-glucans, but importantly 1,6-branching was readily incorporated in a controlled manner.

Scheme 9.

Seeberger second generation solid phase synthesis of β-(1,3)-glucans.

Solid phase carbohydrate synthesis strategies are ideally positioned to facilitate the synthesis of polysaccharides. The β-glucans syntheses described by Seeberger and co-workers demonstrate the feasibility and flexibility of this approach. However, the ability of this approach to efficiently provide material in excess of 10 mg remains unclear due to the need for a large excess of the monosaccharide building blocks that are not readily accessible. Nonetheless, this approach provided sufficient material to enable subsequent biological investigations.

3. Enzymatic Syntheses

Enzymatic or chemoenzymatic syntheses of structurally complex carbohydrates provide an alternative approach to the traditional synthetic strategies outlined above. Ideally, the use of enzymes would circumvent the need for laborious protecting group manipulations, and would provide access to the desired β-(1,3)-glucans in a regio- and stereoselective manner. The promise of chemoenzymatic strategies to synthesize defined β-(1,3)-glucans is evident but has yet to be realized. Here we highlight synthetic strategies that leverage the utility of enzymes to prepare β-(1,3)-glucans.

3.1. Linear β-(1,3)-glucans

Seminal studies carried out by Driguez and co-workers took advantage of the reversible reaction that is catalyzed by 1,3/1,4-β-glucanase from Bacillus licheniformis to prepare linear β-(1,3)-glucans.^[21] In this synthesis, they employed the glycosyl fluoride disaccharide donor 101. The development of this reaction was hindered by the instability of the β-glucanase enzyme. Subsequent investigations, focused on studying the 3D conformation of β-(1,3)-glucans, led to the preparation of linear β-(1,3)-glucans composed of 30–34 glucose residues, using a mutant of the Barley β-(1,3)-D-glucan endohyrdrolase (Scheme 10).^[22] To prepare these polysaccharides, they utilized the glycosyl fluoride disaccharide 101, which could be polymerized, using the mutated endohydrolase, to provide linear β-(1,3)-glucans. Interestingly, they also prepared the 3-thio-α-laminaribiosyl fluoride 102, which was also submitted to the enzyme mediated polymerization reaction, giving rise to the corresponding thioether β-(1,3)-glucans derivatives. Reactions using 102 proceed at a slower rate relative to those carried out with 101, thus, the linear thioether containing β-(1,3)-glucans that were formed contained fewer monosaccharide units. This approach gave rise to β-(1,3)-glucans that contain comparable numbers of glucose units to those observed in β-(1,3)-glucans isolated from natural sources, and it is reasonable to conceive that this approach could be optimized to prepare structurally defined β-(1,3)-glucans.

Scheme 10.

Enzymatic synthesis of linear β-(1,3)-glucans.

3.2. Branched β-(1,3)-glucans

Many naturally occurring β-(1,3)-glucans are modified by β-(1,6) branching that alters the conformation of the β-glucans, thus, impacting the biological activity of these polysaccharides. An enzymatic strategy to prepare β-(1,3)-glucans that have β-(1,6)-branching was developed by Driguez and co-workers (Scheme 11).^[23] Specifically, they applied a (1,3)-β-D-glucan endohyrolase mutant (Glu231Gly) to facilitate the formation of the β-(1,3)-glucan backbone linkage using glycosyl fluoride disaccharide 101 and a glycosyl acceptors 105 and 107. Their efforts highlight that branched glycosyl acceptors can be utilized by this enzyme, and while the preparation of larger branched β-glucans have yet to be reported, this strategy represents an alternative route to branched β-glucans that is devoid of arduous protecting group manipulations.

Scheme 11.

Enzymatic synthesis of branched β-(1,3)-glucans.

4. β-Glucan Mimetics

The application of synthesized β-(1,3)-glucans to functional studies promptly led to the realization that shorter structures (e.g., laminarahexose) are less effective inducers of immunostimulatory signals relative to isolated β-glucans which are typically composed of more than ten glucose residues. Efforts to overcome this shortcoming, in the absence of synthesizing longer β-(1,3)-glucans, has led to the preparation of β-glucan mimetics. Here we will highlight two effective strategies that have been applied to prepare β-glucan mimetics.

4.1. Bertozzi – Glycopolymer based Mimetics

To facilitate their studies, focused on investgating the interactions between β-(1,3)-glucans and Dectin-1, the Bertozzi lab used smaller di- and tri-saccharide units, and an N-carboxyanhydride mediated polymerization to prepare glycopolymers.^[24] These multivalent molecules exhibit increased avidity, and provide access to biological phenomena that would be challenging to study using the di- and trisaccharide polymer building blocks.^[25]

The preparation of this polymer was initiated by functionalizing glycosyl bromides 109a–c via a glycosylation reaction involving N-carboxylbenzyl-L-serine benzyl ester (Scheme 12). The resulting intermediates 110a–c were then submitted to hydrogenolysis, prior to N-carboxyanhydride (NCA) formation via treatment with phosgene. The resulting glycosyl-NCA’s 111a–c were thoroughly purified prior to subsequent polymerization reactions. Initial polymerization reactions, carried out using the glycosyl-NCA building blocks, provided fully glycosylated homopolymers; however, concerned that steric crowding would inhibit receptor binding the authors set out to prepare copolymers which would integrate alanine and glutamic acid spacing units between glycosylated units. The copolymers were prepared via a polymerization reaction involving the glycosyl-NCA’s, L-Ala-NCA, and t-Bu L-Glu-NCA that was catalyzed by Ni^II catalyst 112. Upon completion of the polymerization, the acetyl groups were removed by saponification, while the t-butyl esters protecting the glutamic acid side chains were removed using trifluoracetic acid. The azide terminated polymers were purified by dialysis prior to conjugation to polystyrene beads. Using these glycopolymer coated beads, Bertozzi and co-workers were able to demonstrate the ability of the synthetic glycopolymers to activate Dectin-1 signaling to levels comparable to those observed with the Dectin-1 agonist curdlan.

Scheme 12.

Bertozzi’s glycopolymer based β-(1,3)-glucan mimetics.

4.2. Crich – Carbocycle based Mimetics

Due to the biological activity of β-(1,3)-glucans, Crich and co-workers have reported syntheses of β-(1,3)-glucan mimetics in a series of manuscripts. Insights garnered from structural studies investigating the interactions between β-(1,3)-glucans and Dectin-1 suggest that the hydrophobic α-face of β-(1,3)-glucans interact with binding site aromatic residues. Based on these observations, Crich and co-workers sought to prepare β-(1,3)-glucan mimetics that would readily interact with the binding sites aromatic residues, and they hypothesized that increasing these interactions would circumvent the need for larger β-(1,3)-glucans.

Their first series of β-(1,3)-glucan mimetics utilize a hydroxylamine moiety to link the glycomimetic units (Scheme 13A).^[26] These mimetics were synthesized from epoxide 114, which could be altered over 4 steps to prepare intermediates 115 and 116. Boc protection of the hydroxyamine moieties provided intermediates 117 and 118. Preparation of intermediate 120, which corresponds to the reducing end of the β-(1,3)-glucan, was prepared via the oxidative cleavage of 117 with osmium tetroxide and sodium metaperiodate, followed by a double ring closing reductive amination with O-Allylhydroxylamine hydrochloride. Removal of the tert-butyloxycarbonyl (Boc) group from the resulting intermediate provided 120. Intermediates 122 and 123 could be prepared via the oxidative cleavage of intermediates 117 and 121, followed by a ring closing reductive amination with hydroxyamine 120. Removal of the Boc group from intermediate 123, provided intermediate 124 that could be applied to prepare trisaccharide mimetic 125, via a reductive amination with the dialdehyde derived from 121. Global deprotection, using BCl₃, provided the desired β-(1,3)-glucan mimetics. With the β-(1,3)-glucan mimetics in hand, Crich and co-workers demonstrated the ability of the mimetics to bind to both Dectin-1 and CR3, and the ability of the mimetics to stimulate macrophages phagocytosis. Interestingly, trimer 126b stimulated phagocytosis much more efficiently than dimer 126a.

Scheme 13.

Carbocycle based β-(1,3)-glucans mimetic. A) hydroxyamine derivatives and B) thioether mimetics of β-(1,3)-glucans.

Encouraged by these results, Crich and co-workers developed a second-generation of β-(1,3)-glucan mimetics composed of 2,4-dideoxy-thioester-linked carbocycles (Scheme 13B).^[27] These mimetics were prepared starting from chiral 3-cyclohexene-1-carboxylic acid 127, which was transformed into key intermediate 128 over a 5 step sequence. This intermediate was then treated with sodium ethanethiolate (NaSEt) to provide diastereomers 129 and 130, which could be readily separated. Diastereomer 130 was then reduced with L-selectride, and the resulting hydroxyl group was activated using methanesulfonyl chloride (MsCl), which was subsequently displaced with cesium thioacetate. Removal of acetyl group provided a thiol intermediate that could undergo a Michael addition with intermediate 127 in the presence of 1,8-diazabicyclo(5.4.0)undec-7-ene (DBU) to provide a disaccharide mimetic. This 5-step process could be repeated to provide extended β-(1,3)-glucan mimetics. Finally, the elongated structures were reduced using sodium borohydride, and the methoxymethyl (MOM) ethers were removed under acidic conditions to provide the desired β-(1,3)-glucan mimetics. Biological characterization of these thioether mimetics revealed that the tetrameric mimetic exhibits significant binding to both CR3 and Dectin-1. Further functional studies, facilitated by the conjugation of the thioether mimetics to polymeric particles, revealed the ability of the mimetics to stimulate macrophage mediated phagocytosis in a receptor dependent manner. These thioether mimetics compared equally to both the hydroxyamine derivatives and synthetic β-(1,3)-glucans of equivalent length.

5. β-Glucan-Protein Conjugates

The biological application of β-(1,3)-glucans to date have sought to leverage the immunostimulatory activity of β-(1,3)-glucans for vaccine development. Towards this goal, a variety of strategies to conjugate β-(1,3)-glucans to antigenic carrier proteins have been developed. In this section, we will describe how said strategies facilitate the conjugation of synthetic and isolated β-(1,3)-glucans to proteins.

5.1. Lysine Conjugation

A common strategy to prepare β-glucan-protein conjugates involves the modification of lysine residues with β-glucans functionalized with an activated ester. This strategy has been employed to conjugate β-glucans to a variety of different proteins. One example of this strategy involves the conjugation of synthetic β-(1,3)-glucans to CRM₁₉₇ (Scheme 14).^[28] Specifically, an amine bearing synthetic β-(1,3)-glucan was first functionalized with di-N-hydroxysuccinamide to provide the activated ester 133, which was subsequently conjugated to CRM₁₉₇ in sodium phosphate buffer at pH 7. A number of β-(1,3)-glucan-CRM₁₉₇ constructs were prepared by varying the degree of β-(1,3)-glucan conjugation, and these constructs were evaluated for their ability to elicit the production of anti-CRM₁₉₇ antibodies. It was observed that the conjugation of β-(1,3)-glucans to CRM increased the production of anti-CRM₁₉₇ antibodies relative to co-administration of the individual conjugate components or CRM₁₉₇ alone.

Scheme 14.

Conjugation of β-(1,3)-glucans to proteins via lysine residues

5.2. Tyrosine Conjugation

An alternative strategy to prepare β-glucan-protein conjugates involves the modification of tyrosine residues. Unlike lysine residues, tyrosine residues are not abundant on protein surfaces, thus, modification of tyrosine residues results in fewer conjugation sites and the formation of less heterogenous products. The conjugation of β-glucans to tyrosine residues has been carried out in a two-step process (Scheme 15),^[29] which involves the modification of the protein’s tyrosine residues with an alkyne functionalized 1,2,4-triazole-3,5-dione 136, followed by the conjugation of a synthetic azide bearing β-(1,3)-glucan via a copper catalyzed 1,3-dipolar cycloaddition. This strategy has been applied to conjugate synthetic β-(1,3)-glucans to CRM₁₉₇. Interestingly, studies to assess the ability of the resulting constructs to elicit anti-β-(1,3)-glucan antibodies demonstrate that the tyrosine-based conjugates performed comparatively to laminarin-CRM₁₉₇ based conjugates that were prepared via lysine conjugation, even though there was a lower degree of conjugation and the attached glycans were considerably shorter in length.

Scheme 15.

Protein conjugation of Protein conjugation of β-(1,3)-glucans through lysine residues via tyrosine residues.

5.3. Chemical Modification of Natural β-(1,3)-glucans

Naturally occurring β-glucans are a valuable source of material for biological investigation; however, in many cases functionalization of the β-glucans is required for downstream applications. The hemiacetal at the reducing end of β-glucans provides an ideal handle for functionalization, and this strategy has been applied to conjugate laminarin to vaccine carrier proteins. In one such example, involving the conjugation of laminarin to tetanus toxoid,^[30] the reducing end of laminarin was modified via a reductive amination with propargylamine in the presence of sodium cyanoborohydride (Scheme 16A). The functionalized laminarin could subsequently be conjugated to azidinated tetanus toxoid (144) via a copper catalyzed 1,3-dipolar cycloaddition. While this approach is productive, it is important to note that a significant portion of isolated laminarin cannot be functionalized via this procedure due to the presence of a 1-O-D-mannitol terminating group at the reducing end of the polysaccharide.

Scheme 16.

Protein conjugation of naturally isolated laminarin via A) oxidative cleavage and B) reductive amination.

An alternative approach involves the oxidation of 1,2-diols to provide functional handles for further derivatization. This strategy has been broadly applied by a variety of different groups resulting alkyne to conjugate β-glucans to proteins (Scheme 16B).^[31] In one example, this strategy was applied to conjugate laminarin to recombinant calreticulin. Specifically, laminarin was oxidized using sodium periodate (NaIO₄) and the resulting isolated material was then conjugated to the protein via a sodium cyanoborohydride (NaBH₃CN) mediated reductive amination.

6. Concluding Remarks

β-glucans are a class of polysaccharides that have been shown to have immunostimulatory activity. In this review, we surveyed the numerous approaches that have been applied to prepare structurally defined β-glucans. Said approaches demonstrate that structurally defined β-glucans can be accessed synthetically, and that the synthetic molecules can be applied to investigate β-glucan-protein interactions; however, it is also clear that larger β-glucans, which remain challenging to access synthetically, are needed for functional studies. Efforts to prepare β-glucans mimetics, such as those described by Bertozzi and Crich, may circumvent this need, but further studies will be required to validate the general applicability of β-glucan mimetics. In addition, it is evident from our review of the methods utilized to conjugate β-glucans to proteins that careful selection of both the carrier protein and the conjugation method are necessary in order to prepare β-glucan-protein conjugates with the desired immunostimulatory activity.

Biographies

Mark P. Farrell received his B.Sc. (2010) and Ph.D. (2014) from the National University of Ireland, Galway. His Ph.D. research, carried out under the tutelage of Paul V. Murphy, focused on studying glycosyl bond anomerizations. Subsequently, he joined the lab Amos B. Smith, III at the University of Pennsylvania where he developed new reaction methods and HIV-1 entry inhibitors. In 2017, he joined the Department of Medicinal Chemistry at the University of Kansas, where he studies carbohydrate-protein interactions, and develops molecular strategies to leverage the potential of the immune system.

Patrick Ross was born in West Chester, Pennsylvania. In 2016 he obtained his B.Sc. in chemistry from Immaculata University. He then moved to the University of Kansas where he is a Ph.D. candidate in the Department of Medicinal Chemistry studying immune system modulating carbohydrate-protein interactions.