Chemical Synthesis and Immunological Evaluation of the Inner-Core Oligosaccharide of Francisella tularensis

Abstract

Francisella tularensis, which is a gram negative bacterium that causes tularemia, has been classified by the Center for Disease Control and Prevention (CDC) as a category A bioweapon. The development of vaccines, immunotherapeutics and diagnostics for F. tularensis requires a detailed knowledge of the saccharide structures that can be recognized by protective antibodies. We have synthesized the inner core region of the lipopolysaccharide (LPS) of F. tularensis to probe antigenic responses elicited by a live and subunit vaccine. The successful preparation of the target compound relied on the use of a disaccharide which was modified by the orthogonal protecting groups diethylisopropylsilyl (DEIPS), 2-naphthylmethyl (Nap), allyl ether (All) and levulinoyl (Lev) ester. The ability to remove the protecting groups in different orders made it possible to establish the optimal glycosylations sequence to prepare a highly crowded 1,2,3-cis configured branching point. A variety of different methods were exploited to control anomeric selectivities of the glycosylations. A comparison of the ¹H NMR spectra of isolated material and the synthetic derivative confirmed the reported structural assignment of the inner core oligosaccharide of F. tularensis. The observation that immunizations with LPS lead to antibody responses to the inner core saccharides provides an impetus to further explore this compound as a vaccine candidate.

INTRODUCTION

Francisella tularensis is the etiologic agent of tularemia (rabbit fever) in humans and animals.¹ It is a gram-negative facultative intracellular pathogen that can survive and propagate within phagocytic cells. In nature, a disease cycle is maintained between wild animals such as rabbits, beavers, squirrels and water rats and biting vectors such as flies, ticks, mosquitoes and mites and the contaminated environment.²F. tularensis is highly virulent, requiring as few as 10-50 cells to cause human infection.³ It can survive for long periods of time under harsh environmental conditions. Tularemia may occur in different forms but the pneumonic form is associated with the highest mortality (30% without antibiotic treatment). F. tularensis has been classified by the CDC as a top-priority (Category A) bio-terrorism agent. Common to all Category-A select agents, F. tularensis transmits easily, has the capacity to inflict substantial morbidity and mortality on a large number of people and can induce widespread panic.⁴ Aerosol dispersal is considered the most hazardous mode of transmission, as it would affect the most people.

To prevent infections by F. tularensis, an attenuated live vaccine strain (LVS) was developed in the 1950's, but was not licensed for use as a human vaccine in the USA because the nature of its attenuation was not known and may not be stable. Considerable efforts are being expended to develop a subunit vaccine composed of a cell surface component such as a protein antigen or capsular and lipopolysaccharides (LPS).⁵ In particular, LPS-based vaccines are attractive and for example it has been shown that mice vaccinated with the O-antigen released by mild acid hydrolysis of LPS and conjugated to BSA can protect against an intradermal challenge with a highly virulent type B strain of F. tularensis, and partially protect against an aerosol challenge with the same strain.⁶ It has also been shown that mice intradermally inoculated with intact LPS from F. tularensis acquire varying degrees of resistance against systematic or aerogenic challenge with virulent strains of the pathogen.⁷ More recently, it was found that a detoxified LPS complex with an outer membrane protein of N. meningitidis group B can protect mice against a lethal respiratory challenge with the highly virulent F. tularensis SchuS4.⁸

The structure of LPS of F. tularensis has been determined and it contains a lipid A moiety, a core oligosaccharide and an O-chain polysaccharide. The O-antigen is composed of tetrasaccharide repeating units, which consist of two N-acetyl galactosamine uronamides and a N-acetyl quinovosamine and N-formyl-4-amino-quinovose moiety.⁹ Furthermore, structural studies have shown that the core region has a highly unusual composition.¹⁰ It is linked to the lipid A region by only one 3-deoxy-d-manno-2-octulosonic acid (KDO) moiety (A) instead of the usual two KDO residues (Figure 1). It does not contain heptosyl residues but contains two mannosyl moieties. One of the mannosides (C) is β-linked to another mannoside (B), and this disaccharide fragment is further substituted at C-2, C-2' and C-3' by a β-glucoside (E), an α-galactosamine (D) and an α-glucoside (F), respectively.

Figure 1.

Target hexasaccharide 1 and the monosaccharide building blocks required for its assembly.

The development of vaccines, immunotherapeutics and diagnostics for F. tularensis requires a detailed knowledge of the saccharide structures that can be recognized by protective antibodies. It also needs well-defined oligosaccharides conjugated to carrier proteins for immunizations to establish structural motifs that can provide protection. Although oligosaccharide fragments can be obtained by controlled hydrolysis of LPS,¹¹ chemical synthesis offers a much more attractive approach to obtain such compounds.¹² Obviously, isolation of oligosaccharides from a Class A bio-terrorism agent is undesirable. It is also difficult to conjugate short oligosaccharides to carrier proteins without destroying vital immunological domains. Synthetic chemistry can address these issues since it makes it possible to incorporate an artificial linker for controlled conjugation to proteins.¹² Furthermore, it can provide substructures for establishing structure-activity relationships or used to determine minimal epitope requirements to elicit protective immune responses.

Herein, we report the synthesis of the complete hexasaccharide inner core domain of F. tularensis LPS and the preparation of biotin and protein conjugates thereof. Immune recognition of the hexasaccharide by antisera of mice immunized with a live-attenuated vaccine or LPS has been determined.

RESULTS AND DISCUSSION

The chemical synthesis of hexasaccharide 1 is challenging due to its highly branched nature, which complicates the installation of the various glycosidic linkages. Furthermore, the target compound contains a number of glycosides that are difficult to install in a stereoselective fashion and in particular the introduction of β-mannosides, α-glucosides and α-linked galactosamines often leads to the formation of a mixture of anomers, which may be difficult to separate and lower the yield of required products.¹³ Furthermore, hexasaccharide 1 has a free amine and carboxylic acid, which makes conjugation to protein carriers or biotin challenging (compounds 2 and 3). The latter type of conjugation is, however, required for immunological evaluations.

It was envisaged that disaccharide 4, which at C-1, C-2, C-2' and C-3' is modified by the orthogonal protecting groups allyl ether (All), levulinoyl (Lev) ester, diethylisopropylsilyl (DEIPS) and 2-methylnaphthyl (Nap), respectively, would provide a flexible intermediate to prepare the target compound.¹⁴ The orthogonal protecting groups made it possible to establish the optimal sequence of glycosylation to install the highly crowded branching points. It also minimized protecting group manipulations during oligosaccharide assembly and offers future opportunities to synthesize a library of structurally related oligosaccharides for immunological studies.

The α-linked 2-amino-2-deoxy-galactoside of 1 could be installed by using glycosyl donor 5 or 6 which are modified by a 4,6-O-di-tert-butylsilyl acetal, which sterically blocks the β-face thereby providing only an α-linked galactoside even in the presence of a C-2 participating group.¹⁵ Glycosyl donors 7 and 8 were prepared to explore the stereoselective introduction of the α-glucoside moiety of 1. In particular, compound 7 was deemed attractive because the C-2 (S)-(phenylthiomethyl)benzyl ether can perform neighboring group participation during glycosylation to give an intermediate anomeric β-sulfonium ion, which upon displacement by a sugar alcohol will selectively provide an α-glucoside.¹⁶ In the case of compound 8,¹⁷ which carries a non-participating benzyl ether at C-2, solvent effects will need to be exploited to control α-anomeric selectivity.¹⁸ Installation of the β-glucoside of 1 should be straightforward by employing trifluoro-N-phenylacetimidates 9¹⁹ or 10²⁰, which have an acetyl ester at C-2 that participates during the glycosylation to give selectively the required β-anomer. Finally, KDO building block 11 carries an aminopropyl linker at its anomeric center, which was expected to allow conjugation to a carrier protein or biotin moiety.

The preparation of KDO building block 11 commenced with known derivative 12,²¹ which was treated with diethylaminosulfur trifluoride (DAST) to afford glycosyl fluoride 13 in a moderate yield due to the formation of a 2,3-unsaturated by-product (Scheme 1). Glycosylation of 13 with benzyloxycarbonyl protected aminopropanol in the presence of BF₃^.Et₂O as the promoter led to the formation of 14 as a separable mixture of α/β anomers (α/β = 3/1). The isopropylidene acetals of 14 were hydrolyzed using a mixture of acetic acid and water and the exocyclic diol of the resulting compound was selectively reprotected as an isopropylidene acetal²² using 2-methoxy-propene and a catalytic amount of p-toluenesulfonic acid (p-TsOH) in DMF to give diol 15. The equatorial alcohol of the latter derivative was selectively benzylated by first forming an intermediate stannyl acetal which was treated with benzyl bromide in the presence of CsF to give the required acceptor 11.²²

Scheme 1.

Preparation of KDO building block 11. Reagents and conditions: a) DAST, CH₂Cl₂, -50°C, 15 min, 50% b) N-benzyloxycarbonyl-3-amino-propanol, BF₃^.Et₂O, CH₂Cl₂, 0°C, 1 h, 55% c) i. AcOH, H₂O, reflux, 1 h, 96% ii. 2-methoxy propene, p-TsOH, 1,4-dioxane, DMF, rt, 16 h, 85% d) Bu₂SnO, MeOH, reflux, 3 h, then BnBr, CsF, DMF, rt, 16 h, 84%.

Next, attention was focused on the preparation of the protected β-d-Man-(1→4)-d-Man disaccharide 4. β-Mannosides, which are an important class of 1,2-cis glycosides, are difficult to introduce due to the axial C-2 substituent, which sterically blocks incoming nucleophiles from the β-face and the Δ-anomeric effect, which provides additional stabilization of the α-anomer.²³ Crich and co-workers have pioneered an attractive approach for the introduction of β-mannosides by in-situ formation of an intermediate α-anomeric triflate because of a strong endo-anomeric effect.²⁴ An S_N2 like-displacement of the α-triflate by a sugar hydroxyl will then result in the formation of a β-mannoside. A prerequisite of β-mannoside formation is that the donor is protected by a 4,6-O-benzylidene acetal. It has been proposed that this protecting group opposes oxacarbenium formation (S_N1 glycosylation) due to the torsional strain engendered by the half chair or boat conformation of this intermediate and a destabilizing electronic effect caused by placing the O-6 dipole antiparallel to the oxacarbenium ion.²⁵ Thus, low-temperature activation of 16 with p-nitrobenzenesulfenyl chloride²⁶ and silver trifluoromethanesulfonate (AgOTf) was complete within minutes and subsequent addition of glycosyl acceptor 17 led to the clean formation of β-mannoside 4 as mainly the β-anomer (β/α = >20/1). The use of trifluoromethanesulfonic anhydride (Tf₂O) and 1-benzenesulfinylpiperidine²⁷ (BSP) as the promoter system led to significant lower yields of product.

To explore the installation of the α-glucoside and α-galactosamine moieties, optimal conditions for the removal of the 2-naphthylmethyl and diethylisopropylsilyl ether needed to be established. The Nap ether of 4 could be readily removed by oxidation with DDQ in wet DCM to give compound 18 in a yield of 93%.^14,28 Treatment of 4 with TBAF to cleave the DEIPS ether²⁹ led to partial removal of the Lev ester, however, the use of TBAF buffered with acetic acid led to clean formation of alcohol 19 in 98% yield.¹⁴

Having glycosyl acceptors 18 and 19 at hand, attention was focused on the installation of the α-glucoside and α-galactosamine moieties. Pre-activated glycosyl donor 7 with TfOH to form an intermediate sulfonium ion followed by the addition of glycosyl acceptor 18 gave trisaccharide 20 in a good yield of 73% as only the α-anomer. Alternatively, a glycosylation of 19 with 5 in the presence of NIS³⁰ and triflic acid (TfOH) afforded trisaccharide 22 in a yield of 77% as only the α-anomer. The trisaccharides 20 and 22 were converted into glycosyl acceptors by removal of the DEIPS and Nap ether using the aforementioned conditions to give glycosyl acceptors 21 and 23, respectively.

Extension of trisaccharide acceptors 21 and 23 to give tetrasaccharide 24 proved to be challenging. Thus, a glycosylation of 21 with glycosyl donor 5 using the aforementioned conditions led to a low yield of tetrasaccharide 24 and extensive decomposition of the glycosyl donor and acceptor was observed (Scheme 2). Mass spectrometric analysis of the crude product showed cleavage of the (S)-(phenylthiomethyl)benzyl ether of 21 which probably arose from reaction with the thiophilic iodonium ion promoter. Therefore, the glycosylation was repeated using glycosyl bromide 6 which can be activated under mild conditions that were expected to be compatible with the (S)-(phenylthiomethyl)benzyl ether. Coupling of 6 with 21 in the presence of AgOTf and DTBMP gave fewer byproducts, however, tetrasaccharide 24 was still isolated in a low yield (Scheme 2).

Scheme 2.

Assembly of the trisubstituted β-mannoside core. Reagents and conditions: a) 16, p-NO₂C₆H₄SCl, AgOTf, DTBMP, 5 min, -78°C then 17, 3 h, -78°C→ -35°C, 73 %. b) DDQ, CH₂Cl₂, H₂O, 3 h, rt, (93%, 18), (72%, 23). c) TBAF, AcOH, THF, 16 h, rt, (98%, 19), (81%, 21) d) 7, TfOH, CH₂Cl₂, 30 min, -35°C→0°C then 18, DTBMP, 16 h, -35°C→rt, 73%. e) NIS, TfOH, CH₂Cl₂, 10 min, 0°C, 77%. f) 5 and 21, NIS, TfOH, CH₂Cl₂, 10 min, 0°C, ~10%. 6 and 21, AgOTf, DTBMP, CH₂Cl₂, 30 min, 0°C, ~10%. 7, TfOH, CH₂Cl₂, 30 min, -35°C→0°C then 23, DTBMP, 16 h, -35°C→rt, ~10%. g) 8, TfOH, Et₂O, 10 min, -35° C, 72%.

The distal mannoside of 24 is glycosylated at C-1, C-2 and C-3, which are oriented in a 1,2,3-cis configuration rendering the bisecting C-2 alcohol inaccessible when C-1 and C-3 are glycosylated. Therefore, the preparation of tetrasaccharide 24 was examined by glucosylation of the C-3 hydroxyl of 23. Thus, low temperature activation of glycosyl donor 7 with TfOH to form an intermediate sulfonium ion was complete within minutes, however, addition of trisaccharide acceptor 23 led only to the formation of a small amount of tetrasaccharide 24 (~10%) and mainly glycosyl acceptor 23 was recovered. The failure of the glycosylation may be due to the bulky nature of the intermediate sulfonium ion, which may not be able to react with a sterically hindered alcohol. To test this hypothesis, trisaccharide 23 was coupled with glycosyl donor 8 using a catalytic amount of TfOH in diethyl ether and fortunately these reaction conditions afforded tetrasaccharide 25 in yield of 72% as mainly the α-anomer (α/β = 20/1). Probably, glycosyl donor 8 reacts through a solvent stabilized oxa-carbenium ion, which is more reactive and less sterically demanding than the corresponding β-sulfonium ion of 7.

Having established the proper order for the introduction of the α-glucoside and α-galactosamine moieites, attention was focused on the further addition of glycosyl residues to 22 (Scheme 3). Due to the flexibility of our approach, this elaboration can take place in a number of different ways but since the Lev ester is needed for the stereoselective introduction of the α-(1→5)-mannosyl linkage to KDO, this glycosylation was undertaken first. Thus, the anomeric allyl moiety of trisaccharide 22 was removed using Pd(PPh₃)₄ in a mixture of CH₂Cl₂ and AcOH and the resulting lactol was converted into the corresponding trichloroacetimidate 26 by treatment with trichloroacetonitrile and 1,8-diazazdicycloundec-7-ene (DBU) in DCM. A TfOH mediated glycosylation glycosyl donor 26 with KDO acceptor 11 in CH₂Cl₂ afford tetrasaccharide 27 in a yield of 61% and due to neighboring group participation of the Lev ester only the β-anomer was formed. The Lev ester of 27 could be selectively removed using hydrazinium acetate in a mixture of toluene and ethanol without affecting the other base sensitive functionalities to give acceptor 28 in a yield of 78%.³¹ A glucosylation of tetrasaccharide 28 with 9¹⁹ having a C-2 acetyl ester to control β-anomeric selectivity afforded the corresponding pentasaccharide 29 in good yield. Surprisingly, the use of similar donor 10²⁰ having benzyl ethers at C4 and C-6 instead of a benzylidene acetal provided a pentasaccharide in low yield. Oxidative cleavage of the Nap ether of 29 by DDQ in a mixture of CH₂Cl₂ and water afforded glycosyl acceptor 30, which was coupled with glucosyl donor 8 under aforementioned conditions to give the fully assembled hexasaccharide 31 (α/β = >20/1).

Scheme 3.

Synthesis of the inner-core of F. tularensis (compounds 1-3). Reagents and conditions: a) i. Pd(PPh₃)₃, AcOH, CH₂Cl₂, 3 h, rt then ii) TCA, DBU, CH₂Cl₂, 1 h, rt, 75%. b) 11, TfOH, CH₂Cl₂, 10 min, 0°C, 61%. c) N₂H ^.₄AcOH, EtOH, Toluene, 30 min, rt, 78%. d) 9, TfOH, CH₂Cl₂, 10 min, 0°C, 82%. e) DDQ, CH₂Cl₂, H₂O, 3 h, rt, 76%. f) 8, TfOH, Et₂O, 10 min, -35°C, 73%. g) i). TFA, CH₂Cl₂, H₂O, 1 h, rt, then ii) HF^.pyridine, THF, 30 min, rt, then iii) Zn, AcOH, CH₂Cl₂, 3 h, rt, 89%, then NaOMe, MeOH, H₂O, THF, 1 h, rt, 55%, then v) Pd(OH)₂, H₂, t-BuOH, H₂O, 16 h, rt, 85%. h) PBS pH 7.4, Biotin-OSu, 16 h, rt, 62%. i) SAMA-Opfp, DIPEA, DMF, 3 h, rt, 53%.

The deprotection of 31 started with the removal of the isopropylidene acetals using TFA in a mixture of CH₂Cl₂ and water. It was expected that the 4,6-O-di-tert-butylsilyl acetal would also be cleaved under these conditions, however, this functionality proved to be remarkably stable and therefore the resulting diol was treated with HF^.pyridine to remove the silyl acetal, which was complete in 30 min. The resulting derivative was treated Zn powder in a mixture of CH₂Cl₂ and AcOH to remove the Troc carbamate to afford a partially deprotected derivative in 89% yield over three steps after purification by LH-20 size exclusion chromatography. Next, the acetyl and benzoyl esters were removed using NaOMe in a mixture of MeOH/THF/H₂O and finally, hydrogenation using H₂ and Pd(OH)₂ in mixture of t-BuOH and water afforded hexasaccharide 1.

The ¹H NMR spectrum of compound 1 is in excellent agreement to the reported ¹H NMR spectrum of the isolated LPS fragment (see Figures 2A, 2B).¹⁰ In addition, the anomeric signals of 1 displayed the appropriate chemical shifts and homonuclear as well heteronuclear coupling constants consistent with the desired product (see Supporting Information). Together these findings unequivocally confirm the reported structural assignment of the isolated LPS fragment.

Figure 2.

Confirmation of structural integrity by NMR. A) Reported ¹H NMR spectrum of the isolated hexasaccharide fragment.¹⁰ B) ¹H NMR spectrum of the synthetic hexasaccharide fragment 1. C) ¹H NMR spectrum of the synthetic biotin conjugate 2. D) HMBC spectrum of 2. The correlation between the amide carbonyl at 176.6 ppm and the CH₂ signals of the linker (3.21-3.35 ppm) and biotin (2.25 ppm and 1.62 ppm) moiety confirms that biotinylation occurred at the desired site.

To perform immunological experiments, it was imperative to selectively derivatize the aminopropanol linker of 1 with a biotin moiety. It was anticipated that the amine of the artificial spacer would be more reactive than the amine of the 2-amino-2-deoxy galactosyl moiety of 1. Indeed, reaction of 1 with N-hydroxysuccinimido biotin (1.0 eq.) in PBS buffer (pH 7.4) afforded a monobiotinylated product as the major reaction product as judged by MALDI-TOF mass spectrometry and TLC analysis. The compound was purified by reverse phase C-18 chromatography and separated from starting material 1 and a dibiotinylated derivative. ¹H NMR analysis of 2 revealed that the signals adjacent to the amine had moved downfield compared to 1, consistent with amide formation at this site (Figure 2C). Furthermore, the heteronuclear multiple bond coherence (HMBC) spectrum (Figure 2D) showed a coupling between the amide carbonyl and the CH₂ protons of the linker confirming the site of reaction. In addition to the biotin derivative 2, a Keyhole Limpet Hemocyanin (KLH) conjugate was prepared for future immunizations. The conjugation of 1 to KLH entailed a two-step procedure. First, hexasaccharide 1 was reacted with perfluorophenyl 2-(acetylthio)acetate and DIPEA in DMF to afford 3. The regioselectivity was again confirmed by the heteronuclear multiple bond coherence coupling between the amide carbonyl and the CH₂ protons on the linker (see Supporting Information). Next, the thioacetyl was cleaved using ammonia in DMF under an inert atmosphere to prevent disulfide formation. The resulting thiol was reacted with maleimide activated KLH to afford the corresponding KLH-conjugate. Analysis of the KLH-conjugate using high-pH anion-exchange chromatography (HPAEC) showed that 339 glycans were present per protein molecule.

Antigenic responses against the inner core region of F. tularensis LPS elicited by a live vaccine³² and a LPS preparation⁶ were investigated. Thus, streptavidin-coated microtiter plates were treated with biotin-modified compound 2 and serial dilutions of sera were added. Detection was accomplished with anti-mouse IgG antibodies labeled with alkaline phosphatase. No appreciable levels of IgG antibodies were observed in the serum samples of mice immunized with live vaccine strain. However, antibodies were detected in mice immunized with the subunit vaccine (Figure 3), highlighting that the inner-core is antigenic when presented in a proper context.

Figure 3.

Immunoreactivity of inner-core oligosaccharide (2) to antisera elicited by a live vaccine strain (LVS) and LPS of F. tularensis. Microtiter plates were coated with compound 2 and serial dilutions of mouse antisera and control serum (starting dilution 1:128) were applied to the coated microtiter plates. The optical density values are reported as the means of triplicate measurements (see the Supporting Information for the SD of the measurements).

CONCLUSION

The successful preparation of hexasaccharide 1, which is derived from the inner core of the LPS of F. tularensis, relied on the use of an orthogonal protected disaccharide that made it possible to establish the optimal glycosylations sequence to prepare a highly crowded 1,2,3-cis configured branching point. In particular, the approach employed a β-d-Man-(1→4)-d-Man disaccharide modified with the orthogonal protecting groups diethylisopropylsilyl (DEIPS), 2-naphthylmethyl (Nap), allyl ether (All) and levulinoyl (Lev) ester. Furthermore, a variety of methods were exploited to control anomeric selectivities of the glycosylations including steric, conformational and solvent effects and classical and auxiliary mediated neighboring group participation. These strategic considerations will be important for the preparation of other highly branched oligosaccharides. It also highlights that the branched nature of many biologically important oligosaccharides complicates the development of routine synthesis procedures based for example on automated polymer supported synthesis. The comparison of the ¹H NMR spectra of isolated material and the synthetic derivative confirmed the reported structural assignment of the inner core oligosaccharide of F. tularensis. The facts that no antigenic responses to the inner core oligosaccharide were observed in mice immunized with a whole bacterial vaccine (LVS) indicate that it may not be suitable for the development of a diagnostic tool. However, the observation that immunizations with isolated LPS lead to antibody responses to the inner core makes it a worthwhile candidate for further exploration as a vaccine candidate. Future studies will focus on immunizations with hexasaccharide 1 conjugated to KLH to establish potential protective properties of this compound.

EXPERIMENTAL SECTION

General procedures

¹H and ¹³C NMR spectra were recorded on a 300 MHz, 500 MHz or a 600 MHz spectrometer. Chemical shifts are reported in parts per million (ppm) relative to tetramethylsilane (TMS) as the internal standard. NMR data is presented as follows: Chemical shift, multiplicity (s = singlet, d = doublet, t = triplet, dd = doublet of doublet, m = multiplet and/or multiple resonances), coupling constant in Hertz (Hz), integration. All NMR signals were assigned on the basis of ¹H NMR, ¹³C NMR, COSY and HSQC experiments. Mass spectra were recorded on an MALDI-TOF mass spectrometer. The matrix used was 2,5-dihydroxybenzoicacid (DHB) and Ultamark 1621 as the internal standard. Column chromatography was performed on silica gel G60 (Silicycle, 60-200 μm, 60 Å). TLC-analysis was conducted on Silicagel 60 F₂₅₄ (EMD Chemicals inc.) with detection by UV-absorption (254nm) were applicable, and by spraying with 20% sulfuric acid in ethanol followed by charring at ~150°C or by spraying with a solution of (NH₄)₆Mo₇O₂₄^.H₂O (25 g/L) in 10% sulfuric acid in ethanol followed by charring at ~150°C. CH₂Cl₂ was freshly distilled from calcium hydride under nitrogen prior to use. Molecular sieves (4Å) were flame activated under vacuum prior to use. All reactions were carried out under an argon atmosphere unless it is stated otherwise. KLH was purchased from Thermo Fisher Scientific inc.

2-Deoxy-2-amino-α-d-galactosamine-(1→2)-3-O-[α-d-glucopyranosyl-(1→3)]-β-d-mannopyranosyl-(1→4)-2-O-[β-d-glucopyranosyl-(1→2)]-α-d-mannopyranosyl-(1→5)-3-amino-propyl-3-deoxy-α-D-manno-octulopyranosidonate (1)

Compound 31 (70 mg, 26 μmol) was dissolved in a mixture of CH₂Cl₂ (4 mL), H₂O (0.2 mL) and TFA (0.4 mL), and the resulting mixture was stirred for 2 h at rt. Toluene (4 mL) was added and the mixture was concentrated in vacuo. The residue was dissolved in THF (4 mL) and HF^.Pyridine (0.7 mL) was added and the resulting mixture was stirred for 2 h at rt. EtOAc (10 mL) and sat. aq. NaHCO₃ (4 mL) was added drop wise and the organic layer was separated, dried (MgSO₄), filtered and concentrated in vacuo. The residue was purified by LH-20 size exclusion column chromatography (CH₂Cl₂/MeOH, 1/1) and the appropriate fractions were concentrated in vacuo. The residue was dissolved in a mixture of AcOH (2 mL) and CH₂Cl₂ (1 mL) and Zn powder (30 mg) was added. The resulting suspension was stirred for 2 h at rt after which the mixture was filtered and the filtrate was concentrated in vacuo. The residue was purified by LH-20 size exclusion column chromatography (CH₂Cl₂/MeOH, 1/1) and the appropriate fractions were concentrated in vacuo. The residue (50 mg, 89 % for three steps) was dissolved in mixture of MeOH (1 mL), THF (1 mL) and water (0.3 mL), and 30% NaOMe in MeOH (0.05 mL) was added. The resulting mixture was stirred for 2 h at rt and AcOH (0.1 mL) was added. The mixure was concentrated in vacuo and the residue was purified by reverse-phase C-18 column chromatography (0 – 90% MeOH/H₂O). The appropriate fractions were collected and concentrated in vacuo. The residue (25 mg, 55%) was dissolved in a mixture of water (1 mL) and t-BuOH (1 mL) and Pd(OH)₂ (20 mg) was added. A hydrogen atmosphere was created and the mixture was stirred for 36 h at rt. The mixture was filtered and the filtrate was concentrated in vacuo to afford 1 (12 mg, 85 %); ¹H NMR (600 MHz, D₂O): δ 5.61 (d, 1H, J = 3.0 Hz ,H-1-D), 5.34 (d, 1H, J = 3.0 Hz ,H-1-F), 5.14 (s, 1H, H-1-B), 4.85 (d, 1H, J = 7.8 Hz ,H-1-C), 4.51 (d, 1H, J = 7.8 Hz ,H-1-E), 4.37-4,35 (m, 1H, H-2-C), 4.29-4.26 (m, 1H, H-2-B), 4.23 (t, 1H, J = 9.0 Hz, H-5-D), 4.17-4.13 (m, 1H, H-4-A), 4.10-3.61 (m, 29H, H-6a,b-B,C,D,E,F, H-3D, H-4-D, H-2-F, H-3-F, H-4-F, H-5-F, H-3-B, H-4-B, H-3-C, H-4-C, H-5-C, H-3-E, H-4-E, H-5-E, H-5-A, H-6-A, H-7-A, H-8-A), 3.52-3.33 (m, 6H, CH₂ Linker, H-2-E, H-2-D, H-3-E, H-5-B), 3.20-3.19 (m, 1H, CHH Linker), 3.08-3.04 (m, 1H, CHH Linker), 2.01 (dd, 1H, J = 4.2 Hz, J = 12.6 Hz, H-3a-A), 1.94-1.91 (m, 2H, CH₂ Linker), 1.75 (t, 1H, J = 12.0 Hz, H-3b-A); ¹³C NMR (125 MHz, D₂O) δ 175.4, 101.6, 100.2, 99.4, 99.4, 99.3, 85.2, 78.5, 76.9, 76.4, 76.0, 76.0, 76.0, 75.8 ,75.8, 75.1, 72.9, 72.6, 72.5, 72.5, 71.5, 71.0, 70.9, 70.9, 69.5, 69.4, 69.0, 68.3, 67.9, 66.9, 65.4, 62.6, 61.7, 60.7, 60.6, 60.5, 50.2, 59.6, 50.7, 38.3, 35.0, 26.0, 23.1; HR-MALDI-TOF/MS (m/z): [M+Na]⁺ calcd for [C₄₁H₇₂N₂O₃₂+Na]⁺ , 1127.3965; found, 1127.3918.

2-Deoxy-2-amino-α-d-galactosamine-(1→2)-3-O-[α-d-glucopyranosyl-(1→3)]-β-d-mannopyranosyl-(1→4)-2-O-[β-d-glucopyranosyl-(1→2)]-α-d-mannopyranosyl-(1→5)--N-biotinyl-3-amino-propyl-3-deoxy-α-D-manno-octulopyranosidonate (2)

Compound 1 (5.0 mg, 4.5 μmol) was dissolved in PBS buffer pH 7.4 (0.5 mL) and BiotinOSu (1.8 mg, 4.5 μmol) in PBS buffer pH 7.4 (0.2 mL) was added. The resulting mixture was stirred for 3 h at rt. The mixture was directly transferred to a reverse phase C-18 column and purified by eluting with 0 -10% MeOH/H₂O. The appropriate fractions were collected and concentrated in vacuo to afford 2 (3.7 mg, 62%) as a white solid. ¹H NMR (600 MHz, CDCl₃): δ 5.51 (d, 1H, J = 3.6 Hz ,H-1-D), 5.31 (d, 1H, J = 3.6 Hz, H-1-F), 5.15 (s, 1H, H-1-B), 4.84 (s, 1H, H-1-C), 4.61 (dd, 1H, J = 4.5 Hz, J = 7.8 Hz, CH Biotin), 4.49 (d, 1H, J = 7.8 Hz, H-1-E), 4.61 (dd, 1H, J = 4.2 Hz, J = 8.4 Hz, CH Biotin), 4.33 (m, 1H, H-2-C), 4.28 (m, 1H, H-2-B), 4.22-4.04 (m, 3H, H-4-A, H-3-B, H-3-C), 3.95-3.22 (m, 29H, H-6a,b-B,C,D,E,F, H-2-E, H-3-E, H-4-E, H-5-E, H-2-F, H-3-F, H-4-F, H-5-F, H-3-D, H-4-D, H-5-D, H-4-C, H-5-C, H-4-B, H-5-B, H-5-A, H-6-A, H-7-A, H-8-A), 3.11 (dd, 1H, J = 10.8 Hz, J = 4.2 Hz, H-2-D), 3.00 (dd, 1H, J = 12.6 Hz, J = 4.8 Hz, CHH Biotin), 2.78 (d, 1H, J = 13.2 Hz, CHH Biotin), 2.24 (t, 1H, J = 7.2 Hz, CH₂ Biotin), 2.05 (dd, 1H, J = 12.6 Hz, J = 4.2 Hz, H-3a-A), 1.79-1.56 (m, 8H, CH₂ Linker, 2 × CH₂ Biotin, H-3b-A), 1.42-1.38 (m, 2H, CH₂Biotin); ¹³C NMR (125 MHz, CDCl₃) δ 176.5, 175.0, 165.2, 101.5, 100.1, 99.8, 88.7, 99.2, 97.2, 79.0, 76.7, 76.5, 76.5, 76.2, 75.9, 75.8, 75.1, 72.9, 72.7, 72.6, 72.5, 71.4, 71.2, 71.0, 70.8, 69.6, 69.4, 69.3, 68.3, 68.1, 66.9, 65.5, 63.0, 61.8, 60.9, 60.7, 60.6, 60.4, 60.1, 59.5, 55.1, 53.7, 50.6, 39.5, 36.5, 35.4, 35.0, 28.1, 27.7, 27.5, 25.0; HR-MALDI-TOF/MS (m/z): [M+Na]⁺ calcd for [C₅₁H₈₆N₄O₃₄S+Na]⁺, 1353.4741; found, 1353.4726.

2-Deoxy-2-amino-α-d-galactosamine-(1→2)-3-O-[α-d-glucopyranosyl-(1→3)]-β-d-mannopyranosyl-(1→4)-2-O-[β-d-glucopyranosyl-(1→2)]-α-d-mannopyranosyl-(1→5)-N-thioacetylacetyl-3-amino-propyl-3-deoxy-α-D-manno-octulopyranosidonate (3)

Compound 1 (5.0 mg, 4.5 μmol) was dissolved in DMF (0.5 mL) and SAMAOpfp (1.4 mg, 4.5 μmol) in DMF (0.2 mL) and DIPEA (0.2 uL, 9.0 μmol) were added. The resulting mixture was stirred for 3 h at rt. The mixture was directly transferred to a reverse phase C-18 column and purified by eluting with 0 -10% MeOH/H₂O. The appropriate fractions were collected and concentrated in vacuo to afford 3 (2.9 mg, 53%) as a white solid. ¹H NMR (600 MHz, D₂O): δ 5.34 (s, 1H, H-1-D), 5.09 (d, 1H, J = 3.6 Hz ,H-1-F), 4.90 (s, 1H, H-1-B), 4.60 (s, 1H, H-1-C), 4.27 (d, 1H, J = 7.8 Hz ,H-1-E), 4.12-4.14 (m, 1H, H-2-C), 4.05-4.00 (m, 1H, H-2-B), 3.97 (t, 1H, J = 9.0 Hz, H-5-D), 3.88-3.86 (m, 1H, H-4-A), 4.10-3.00 (m, H, H-6a,b-B,C,D,E,F, H-3D, H-4-D, H-2-F, H-3-F, H-4-F, H-5-F, H-3-B, H-4-B, H-3-C, H-4-C, H-5-C, H-3-E, H-4-E, H-5-E, CH₂ Linker, H-2-E, H-2-D, H-3-E, H-5-B, CH₂ Linker), 1.92-1.87 (m, 1H, H-3a-A), 1.56-1.46 (m, 3H, CH₂ Linker, H-3b-A). HR-MALDI-TOF/MS (m/z): [M+Na]⁺ calcd for [C₄₅H₇₆N₂O₄₃S+Na]⁺,1234.3897; found, 1243.3847.

Conjugation of 3 to keyhole limpet hemocyanin (KLH)

Compound 3 (1.5 mg, 1.3 μmol) was dissolved in DMF (0.5 mL) and 5% ammonia in DMF (50 μL) was added. After 2 h, MALDITOF showed complete removal of the S-acetyl and the mixture was concentrated in vacuo. The residue was dissolved in PBS buffer pH 7.2 (1.0 mL) and a solution of maleimide activated mcKLH (4 mg in 0.5 mL water) was added. The resulting mixture was stirred for 2 h at rt. The mixture was purified by spin filtration. Analysis of the KLH conjugate using high performance anion exchange chromatography (HPAEC) showed that on average, 339 hexasaccharides were conjugated per KLH molecule (see Supporting Information). Since one KLH molecule has 522 maleimide molecules this corresponds to a conversion of 65%.

Dose and immunization schedule

Specific-pathogen-free, female BALB/c mice were purchased from Charles Rivers Laboratories (St. Constant, Que.). Mice were maintained and used in accordance with the recommendations of the Canadian Council on Animal Care Guide to the Care and Use of Experimental Animals and entered the experiments between 7 and 10 weeks of age. For LVS immunization, mice were immunized p.o. on day 0 and 14 with 2 × 10⁸ CFU F. tularensis LVS (actual confirmed inocula: 1.6 × 10⁸/mouse for the 1^st immunization and 2 × 10⁸/mouse for the 2^nd immunization) or PBS as control as described previously.³² The mice were killed on day 35 for serum collection. The serum samples from the mice immunized with a vaccine consisting of the O-polysaccharide of the F. tularensis chemically lipopolysaccharide (LPS) conjugated to bovine serum albumin (BSA-O-PS conjugate) were kindly provided by Dr. Wayne Conlan (National Research Council Canada, Ottawa, Canada). The preparation of the glycoconjugate vaccine was described previously in details.⁶ Mice were immunized subcutaneously at 0, 28 and 56 days with 20 μg of the glycoconjugate emulsified in a 1:3 ratio with incomplete Freud's adjuvant in a total volume of 0.1 mL. Mice were killed on day 70 for serum collection.

Serologic assay

IgG antibody titers against the inner core of F. tularensis LPS were determined by enzyme-linked immunosorbent assay (ELISA). Reacti-bind NeutrAvidin coated and pre-blocked plates (Thermo Scientific) were incubated with compound 2 (a stock solution in DMSO (2 mM) was diluted to 5 μM; 100 μL/well) for 2 h. Next, serial dilutions of the sera were allowed to bind to immobilized compound for 2 h. Detection was accomplished by the addition of alkaline phosphatase-conjugated anti-mouse IgG (Jackson ImmunoResearch Laboratories Inc.). After addition of p-nitrophenyl phosphate (Sigma), the absorbance was measured at 405 nm with wavelength correction set at 490 nm using a microplate reader (BMG Labtech). The antibody titer was determined by linear regression analysis, plotting dilution vs. absorbance. Titers were defined as the highest dilution yielding an optical density of 0.1 or greater over that of control mouse sera. Experiments were performed in triplicate.