Synthesis of Neisseria meningitidis X capsular polysaccharide fragments

Serotype X of Neisseria meningitidis bacterium (Men X) recently emerged as a substantial threat to public health. Since anti-meningococcal vaccines currently available or under investigation do not contain antigenic components of Men X capsular polysaccharide, there is the need to develop more comprehensive conjugate vaccines capable to offer higher protection. As a preliminary step towards this goal, the synthesis of three conjugatable Men X capsular polysaccharide fragments is described. The installation of the crucial α-glycosyl phosphodiester linkages is based on the hydrogenphosphonate methodology using pure α-glycosyl hydrogenphosphonates 10 and 12 obtained from hemiacetals 9 and 11 , respectively.


Introduction
Bacterial meningitis causes approximately 170,000 annual deaths upon more than 1,200,000 cases, with at least a 5-10% of case fatality in industrialized countries and a 20% in the developing world. 1 Streptococcus pneumoniae, Haemophilus influenzae type b (Hib) and Neisseria meningitidis are responsible for most of the cases of bacterial meningitis worldwide. 2n particular, thirteen different serogroups of N. meningitidis have so far been defined, but about 90% of the infections are due to serogroups A, B, C, Y and W135. 3 However, in the past 20 years sporadic cases or clusters of meningitis due to other N. meningitidis serogroups have emerged.N. meningitidis type X (Men X), first described in the 1960s, 4 has been found to cause a few cases of invasive disease across North America, Europe, Australia and Africa. 5In 2006, owing to an unprecedented incidence of meningitis caused by Men X in Niger, 6 the World Health Organization (WHO) recognized Men X as a substantial threat and epidemic potential.
Vaccination is considered by the WHO to be the most cost-effective strategy for controlling infectious disease, since it should confer long-term protective immunity in the population.In particular, carbohydrate-based vaccines have recently emerged as a powerful tool with enormous potential benefits for human health. 7They are designed to target pathogen-specific cell surface saccharide structures, such as the carbohydrate capsule (capsular polysaccharide, CPS), which represent a major virulence factor of encapsulated bacteria, including N. meningitidis.Saccharide vaccines currently present on the market, or under development, are based on carbohydrateprotein conjugates, 8 and they comprise also different formulations against H. influenzae type b, S. pneumoniae and several meningococcal serogroups.However, none of them include antigenic components of Men X, and therefore they do not offer protection against infections caused by this emerging serogroup.Although Men X currently causes only a small proportion of meningococcal disease, 5,6 it is possible that repeated vaccination against some serogroups (especially A and C) has the potential to select meningococci of other serogroups (for example, Men X) and might result in a changed profile of meningococcal disease.This possibility should be considered when conjugate vaccines carrying limited ranges of serogroups are introduced.5b The development of more comprehensive conjugate vaccines including Men X CPS fragments could therefore become an urgent issue in the near future.
Carbohydrate-based antigens needed for inclusion in a vaccine, however, are not readily available from natural sources.The isolation and purification of the polysaccharide from the bacterial source is indeed a challenging task, leading to the presence of biological contaminants and impurities, thus raising severe issues of quality assurance.On the other hand, synthetic carbohydrate-based vaccines have important advantages, including their well-defined chemical structures and a better safety profile.Consequently, a variety of pathogen-associated saccharide antigens have been synthesized in the last decade, coupled to carrier proteins and proved to elicit protective antibodies in animal models. 9In addition, synthetic oligosaccharides can help elucidate the structural moieties of the native bacterial polysaccharide that are essential to induce antibodies. 10This step is crucial for the design of a new generation of improved and safer vaccines obtained either from chemical synthesis or bacterial source.
On the basis of these considerations, we report herein the synthesis of the monomer, dimer and trimer of Men X CPS repeating unit (compounds 1-3, Figure 1).The synthetic fragments are provided with a phosphodiester-linked aminopropyl spacer at their reducing end, suitable for their eventual conjugation to a carrier protein.

Results and Discussion
The CPS of N. meningitides X is a homopolymer of (14)-linked 2-acetamido-2-deoxy-α-Dglucopyranosyl phosphate residues (Figure 1). 11There is no doubt that the major challenge in the preparation of our target oligomers is the stereochemical control in the synthesis of the αphosphodiester linkages.Among other methodologies available in the literature for the formation of phosphodiester linkages, we relied on the well-established hydrogenphosphonate (Hphosphonate) protocol. 12The same strategy was employed by Shibaev and co-workers who, to the best of our knowledge, reported the first synthesis of Men X dimer. 13The Russian group synthesized an anomerically pure α-H-phosphonate (glycosyl phosphite) of a Nacetylglucosamine (GlcNAc) derivative which was then coupled with the 4-OH of the O-pnitrophenylglycoside of a second GlcNAc residue.On the other hand, our fragments are intended for the preparation of the corresponding protein conjugates, and each compound contains a phosphodiester-linked spacer available for this purpose.An additional advantage of our approach is the higher flexibility, since the protecting groups pattern of the synthetic precursors was designed to allow further elongation and synthesis of oligomers with variable length, as demonstrated by the preparation of trimer 3. Accordingly, suitably protected azide-containing monosaccharides 4 and 5 (Scheme 1) were employed as key building-blocks.In particular, compound 5 was selected as elongation block, since the 4-OH is easily available by Zemplén transesterification, whereas we envisaged intermediate 4 as a capping block to be introduced at the non-reducing terminus of each oligomer.Compound 4 -precursor of 5-was initially synthesized in four linear steps from D-glucosamine hydrochloride as described by Martin-Lomas et al. 14 This procedure resulted however unsuitable for large laboratory scale (over 20 g of starting material), and we applied the longer but more reliable reactions sequence outlined in Scheme 1, which provided intermediates 4 and 5 in good overall yield and much higher purity.Briefly, D-glucosamine was first converted into the 2-azidoglucose derivative 7 by diazotransfer reaction followed by standard O-acetylation (Ac2O, pyridine, N,N-dimethylaminopyridine, DMAP) in 77% yield.The crucial diazotransfer reaction was carried out using the recently described 15 imidazole-1-sulfonyl azide diazo donor as a cheap and safe alternative to the more popular trifluoromethanesulfonyl azide (TfN3). 16Besides the high cost of trifluoromethanesulfonic anhydride, used in its preparation, neat TfN3 has indeed explosive nature and its poor shelf life requires its preparation in solution just prior to use. 17On the contrary, the imidazole-1-sulfonyl azide can be easily prepared from relatively inexpensive starting materials and its hydrochloride salt (compound 6 in Scheme 1) is a crystalline solid that can be stored at 4°C for weeks without loss of efficiency.Scheme 1. Synthesis of key building-blocks 4 and 5.
Next stage of the synthesis of the target oligomers was the crucial installation of the anomeric phosphodiester linkages using the H-phosphonate protocol.We reasoned that the most reliable way to obtain the stereochemical control in the synthesis of anomerically pure α-glycosyl phosphodiesters is the condensation of a glycosyl α-H-phosphonate with the free hydroxyl of a sugar acceptor, followed by oxidation of the H-phosphonate diester intermediate.The preparation of anomerically pure glycosyl α-H-phosphonate is however a challenging task.Accordingly, no examples have been found in the literature describing the synthesis of exclusively α-anomeric H-phosphonate from 2-azido-2-deoxy-glucopyranosyl derivatives.
Literature data suggest however that the treatment of a sugar hemiacetal with a proper phosphitylating agent in the presence of phosphorous acid could lead to equilibration of the initially formed α,β mixture of the anomeric H-phosphonates into the pure, more thermodynamically stable α anomer. 18Compound 4 was therefore 1-O-desilylated using tetrabutylammonium fluoride and acetic acid in THF at -40°C to afford hemiacetal 9 in 80% yield (Scheme 2).The synthesis of the α-H-phosphonate from 9 was next thoroughly investigated testing different reaction conditions, varying the phosphitylating agent, the reaction solvent and temperature, and the stoichiometry of the reagents.Eventually, we found that the treatment of 9 with 2-chloro-4H-1,3,2-benzodioxaphosphinin-4-one (commonly named salicylchlorophosphite) and H3PO3 in 1:1.5:3 molar ratio in pyridine from room temperature to 40°C (see the Experimental Section) provided exclusively α-H-phosphonate 10, as evinced by 1 H-and 31 P-NMR spectra.The formation of the H-phosphonate was confirmed by the diagnostic doublet at 7.01 ppm in the 1 H-NMR spectrum with the characteristic value of 1 JH,P 640.4 Hz (δP 1.84), while the α-configuration was ascertained by the signal corresponding to H-1 (δH 5.70, J1,2 3.5 Hz, J1,P 8.8 Hz).Scheme 2. Synthesis of α-H-phosphonates 10 and 12.
The long reaction time (8-9 days are needed) represents a significant drawback of this procedure.The reaction progress could be however easily monitored by 1 H-and 31 P-NMR analysis.As shown in Figure 2 (part a), as the reaction proceeds one can observe the progressive disappearance of the signals corresponding to the β-glycosyl H-phosphonate with the concomitant formation of the hemiacetal 9 (see the doublet at δ 5.32).Figure 2b shows the same trend in 31 P-NMR spectra.According to the literature, 18 a reasonable explanation of the observed behaviour is that the more reactive β-glycosyl H-phosphonate is converted to either the α-anomer 10 (due to a SN2 displacement by H3PO3) or the hemiacetal 9 as a result of acid-catalyzed cleavage of the H-phosphonate group.In accordance with this mechanism, the α-H-phosphonate 10 was obtained in 41% yield along with 45% of the hemiacetal 9, which could be easily separated by chromatography and recycled.The same protocol was successfully applied to hemiacetal 11, obtained by 1-O-desilylation of compound 5, affording the α-glycosyl H-phosphonate 12 in 52% yield (along with 38% of recovered 11, Scheme 2).The formation of α-H-phosphonate 12 from 11 was slightly faster (6-7 days), and its structure was confirmed by 31 P-and 1 H-NMR data (δP 2.24, δH 7.03 for H-P with 1 JH,P 642.0 Hz, δH 5.73 for H-1, with J1,2 3.0 Hz and J1,P 8.7 Hz).NMR spectra of the reaction mixture during the conversion of 11 into 12 are illustrated in Figure 3. Now the stage was set for the assembly of the target oligomers.First, glycosyl Hphosphonate 10 was coupled with commercially available benzyl N-(3-hydroxypropyl)carbamate using pivaloyl chloride as condensing agent (Scheme 3). 12,19In situ oxidation of the Hphosphonate diester intermediate was carried out with iodine in aqueous pyridine, providing phosphodiester 13 in 62% yield as triethylammonium salt after quenching the reaction mixture with 1 M aq.triethylammonium hydrogen carbonate (TEAB) buffer solution (pH 7).The structure of compound 13 was confirmed by NMR and mass spectrometry.The 31 P-NMR spectrum exhibited a single signal with δP -0.84 characteristic for glycoside-linked phosphodiesters.
The subsequent azide reduction turned out much more difficult than expected.While the classical Staudinger reaction 20 failed, all other attempts (1,3-propanethiol, PMe3, Zn-Cu couple in THF-Ac2O-AcOH) gave disappointing results, with only traces of the desired product.Finally, the azide was converted into the acetamide using a combination of NiCl2 and NaBH4 followed by treatment with Ac2O, 21 affording compound 14 in 29% yield.In spite of this unsatisfactory result, we deemed this yield acceptable at this stage of the research, and the obtained acetamido intermediate 14 was submitted to final deprotection.The remaining protecting groups were removed via hydrogenolysis over Pd on carbon.Final purification was accomplished by eluting a water solution of deprotected compound over a column filled with Dowex 50W X8 resin (H + form), followed by a second ion exchange on the same resin in Na + form.Lyophilization of the eluates provided the spacer-linked monomer 1 as sodium salt in 93% yield (Scheme 3).

Scheme 3. Synthesis of the spacer-linked monomer 1.
The syntheses of dimer and trimer required the condensation of the glycosyl H-phosphonate 12 with benzyl N-(3-hydroxypropyl)carbamate in the presence of pivaloyl chloride, followed by in situ oxidation to phosphodiester 15 (64%) and careful Zemplén 4-O-deacetylation (96% yield, Scheme 4).The resulting alcohol 16 was coupled with the capping residue 10 to afford, after oxidation, phosphodisaccharide 17 in 45% yield.The azide reduction was accomplished using the NiCl2/NaBH4 protocol, followed by N-acetylation and hydrogenolytical removal of the protecting groups, furnishing after ion exchange fully deprotected dimer 2 as bis-sodium salt in 36% overall yield (from 17, Scheme 4).
On the other hand, the synthesis of trimer 3 was achieved by condensation of alcohol 16 with the elongation block 12 followed by oxidation, which gave phosphodisaccharide 18 in 40% yield (Scheme 5).Zemplén deacetylation of 18 and subsequent coupling with 10 provided phosphotrisaccharide 20 in 43% overall yield after oxidation.Reduction of the three azides and N-acetylation was carried out with NiCl2/NaBH4 followed by Ac2O addition.Eventually, hydrogenolysis of the remaining protecting groups and ion exchange performed as described above afforded trimer 3 as tris-sodium salt in 33% overall yield from 20 (Scheme 5).
Oligomers 2 and 3 were fully characterized by NMR spectroscopy and mass spectrometry. 1H-and 13 C-NMR spectra, using both mono-and bidimensional experiments, were in excellent agreement with the proposed structures (see the Experimental Section).In addition, 31 P-NMR spectra showed two peaks at δ 0.90 and 0.60 for dimer 2, and three peaks (δ -0.31, -0.65 and -0.71) for trimer 3.

Conclusions
In conclusion, as a preliminary step towards the development of more comprehensive antimeningococcal conjugate vaccines including new emerging serotypes, we described the synthesis of three conjugatable fragments (monomer 1, dimer 2 and trimer 3) with structures corresponding to the natural CPS of N. meningitidis type X.The challenging synthesis of anomerically pure α-hydrogenphosphonate, a crucial aspect for our strategy, has been accomplished for the first time on 2-azido-2-deoxy derivatives using a combination of H3PO3 and salicylchlorophosphite.Our approach is featured by high flexibility and, in principle, it allows the synthesis of even longer oligomers by iteration of the 4-O-deacetylationhydrogenphosphonate condensation sequence either using the capping residue 10 or the elongation block 12. On the other hand, a major drawback of our approach are the disappointing results obtained in the reduction of the azido functions, which caused a significant drop of the chemical yields.The use of the azido group was due to its non-participating and electronwithdrawing properties, that were expected to enhance the stability of the anomeric phosphodiester linkages.The difficulties encountered in the final deprotection of the oligomers however induced us to take into account a different approach, based on GlcNAc instead of azido glucose building blocks, which is currently under investigation.
The chemical conjugation of oligomers 1-3 to immunogenic carrier proteins, and the biological evaluation of the resulting neo-glycoconjugates will be reported in due course.

Experimental Section
General.All commercially available reagents including dry solvents were used as received.The only exception is H3PO3, which was coevaporated three times with dry toluene, and dried by high vacuum pump prior to use.Reactions were monitored by thin-layer chromatography on precoated Merck silica gel 60 F254 plates and visualized by staining with a solution of cerium sulfate (1 g) and ammonium heptamolybdate tetrahydrate (27 g) in water (469 mL) and concentrated sulfuric acid (31 mL).Chromatographic purifications have been performed either using the flash purification apparatus Biotage® SP1™ by gradient elution, or by standard column chromatography on Fluka silica gel 60.NMR spectra were recorded at 300K on spectrometer operating at 400 MHz.Proton chemical shifts are reported in ppm (δ) with the solvent reference relative to tetramethylsilane (TMS) employed as the internal standard (CDCl3 δ 7.26 ppm).J values are given in Hz.Carbon chemical shifts are reported in ppm (δ) relative to TMS with the respective solvent resonance as the internal standard (CDCl3, δ 77.0 ppm).In 13 C-NMR spectra, signals corresponding to aromatic carbons are omitted and, apart quaternary carbons, signals attribution was derived by HSQC experiment.Signals attribution in NMR spectra are designated using indexes a, b and c, and referred to ring a (reducing end, directly linked to the spacer), ring b (either the non-reducing end or the inner ring), and ring c (the nonreducing terminus).Optical rotations values are given in 10 -1 deg cm 2 g -1 and were measured at 25 °C on a polarimeter at 589 nm using a 5 mL cell with a length of 1 dm.High resolution mass spectra (HRMS) were performed at CIGA (Centro Interdipartimentale Grandi Apparecchiature), with mass spectrometer APEX II & Xmass software (Bruker Daltonics).ESI-MS spectra were recorded on a JEOL AX-505 spectrometer.

General procedure B. Synthesis of α-hydrogenphosphonate.
A 2 M solution of H3PO3 in pyridine (3:1 molar ratio as to the hemiacetal) was added dropwise to a solution of the hemiacetal in dry pyridine (4 ml/mmol), and thereafter the solution was cooled to 0°C and salicylchlorophosphite (1.5:1 molar ratio as to the hemiacetal) was slowly added.The reaction was stirred at rt and, after disappearance of the starting material (TLC), the reaction temperature was raised up to +40°C.The reaction mixture was stirred under argon until complete disappearance of the β anomer (evidenced by 1 H-and 31 P-NMR spectra).A 1 M solution of TEAB (4 ml/mmol) was added to the reaction mixture at rt, then it was diluted with CH2Cl2, washed three times with cold TEAB solution (0.5 M), dried (Na2SO4), filtered and concentrated.The crude product was purified by flash chromatography (CH2Cl2/MeOH 8/2 + 1% Et3N).The obtained compound was further washed with cold TEAB solution (0.25M) to afford the α-Hphosphonate as a triethylammonium salt.General procedure C. Hydrogenphosphonate coupling and oxidation.The alcohol and the H-phosphonate (10 or 12) were co-evaporated with dry pyridine for three times under high vacuum.The residue was then dissolved in dry pyridine(10 ml/mmol), and PivCl (2.5:1 molar ratio as to the H-phosphonate) was added.The mixture was stirred under nitrogen atmosphere for 45 min-1 h (TLC in CH2Cl2/MeOH mixtures).After cooling to -40°C, a freshly prepared 0.5 M solution of I2 (2.5:1 molar ratio as to the H-phosphonate) in a 19:1 mixture of pyridine-H2O was added.The oxidation was completed at 0°C and quenched by dropwise addition of a 0.5 M solution of Na2S2O3•5H2O (10% w/v).The reaction mixture was diluted with CHCl3, and the organic layer was washed two times with a 0.5 M solution of Na2S2O3•5H2O (10% w/v), then 0.5 M TEAB, dried (Na2SO4), filtered, and concentrated.The crude residue was purified by flash chromatography (CH2Cl2/MeOH + 1% of Et3N), providing the phosphodiester derivative.The obtained compound was further washed with cold TEAB solution (0.25 M) to afford the corresponding triethylammonium salt.

General procedure D. Azide reduction and N-acetylation.
To a solution of the azidecontaining compound in MeOH (10 ml/mmol), first NiCl2•6H2O (3:1 molar ratio as to the number of azido groups) was added under stirring and N2 atmosphere, then NaBH4 (8:1 molar ratio as to the number of azido groups) portionwise (1 hour) at 0 °C.A black precipitate indicated the formation of a Ni-B species.The mixture was stirred at 0 °C and, after consumption of the starting material (amine formation is monitored by ninhydrin-detection), Ac2O (20:1 molar ratio as to the azide-containing compound) was added.The mixture was concentrated under reduced pressure, diluted with CH2Cl2 and washed three times with water.The organic phase was dried (Na2SO4), filtered and concentrated, and purified by flash chromatography to give the acetamide.The obtained compound was further washed with cold solution (0.25 M) to afford the corresponding triethylammonium salt.General procedure E: Hydrogenolysis and ion-exchange.The protected acetamide was hydrogenolysed over Pd/C (10%) in a 5:1 mixture of MeOH and H2O at room temperature for 24-48 h.The mixture was filtered over a Celite pad and the filtrate was concentrated.Then the residue was dissolved in H2O and first eluted through a column filled with Dowex 50W-X8 resin (H + form), and then through a column filled with the same resin in Na + form.The eluate was concentrated and lyophilized to afford the target compound as sodium salt.

Figure 1 .
Figure 1.Structures of repeating unit of Men X CPS and target oligomers 1-3.

Figure 2 .
Figure 2. 1 H-NMR (part a) and 31 P-NMR spectra (part b) showing the reaction progress from 9 to 10.

Figure 3 .
Figure 3. 1 H-NMR (part a) and 31 P-NMR spectra (part b) showing the reaction progress from 11 to 12.