Synthesis of Aminooxy Glycoside Derivatives of the Outer Core Domain of Pseudomonas aeruginosa Lipopolysaccharide

Pseudomonas aeruginosa is a highly prevalent gram-negative bacterium that is becoming more difficult to treat because of increasing antibiotic resistance. As chemotherapeutic treatment options diminish, there is an increased need for vaccines. However, the creation of an effective P. aeruginosa vaccine has been elusive despite intensive efforts. Thus, new paradigms for vaccine antigens should be explored to develop effective vaccines. In these studies, we have focused on the synthesis of two L-rhamnose–bearing epitopes common to glycoforms I and II of the outer core domain of Pseudomonas aeruginosa lipopolysaccharide, α-L-Rha-(1→6)-α-D-Glc-(1→4)-α-D-GalN-(Ala)-α-aminooxy (3) and α-L-Rha-(1→3)-β-D-Glc-(1→3)-α-D-GalN-(Ala)-α-aminooxy (4), respectively. The target trisaccharides were both prepared starting from a suitably protected galactosamine glycoside, followed by successive deprotection and glycosylation with suitably protected D-glucose and L-rhamnose thioglycosides. Global deprotection resulted in the formation of targets 3 and 4 in 22 and 35% yield each. Care was required to modify basic reaction conditions to avoid early deprotection of the N-oxysuccinamido group. In summary, trisaccharides related to the L-rhamnose–bearing epitopes common to glycoforms I and II of the outer core domain of Pseudomonas aeruginosa lipopolysaccharide have been prepared as their aminooxy glycosides. The latter are expected to be useful in chemoselective oxime-based bioconjugation reactions to form Pseudomonas aeruginosa vaccines.


INTRODUCTION
Pseudomonas aeruginosa is a widely distributed, encapsulated, gram-negative bacterium. In the early 1900s, it was recognized as a bacterial pathogen, and in the past 50 years, it has become one of the most concerning pathogens. It is reported that P. aeruginosa is the cause of 1 in 10 nosocomial infections associated with serious illness such as ventilator-associated pneumonia and various sepsis syndromes, and it has the highest mortality rate (37%) (Klevens et al., 2008;Lister et al., 2009). Most P. aeruginosa is resistant to at least one of the classes of antibiotics, and some P. aeruginosa is resistant to all the available antibiotics (Talbot et al., 2006;Pier, 2007). Therefore, vaccines are potential solutions to overcome the antimicrobial resistance (AMR) developed by P. aeruginosa. Lipopolysaccharide (LPS) is a complex glycolipid present on the outer layer of gram-negative bacteria. It plays a vital role as an essential virulence factor in the pathogenicity of P. aeruginosa strains and, hence, is a potential antigen target for a prophylactic vaccine (Pier, 2007). In a recent study, Liu et al. published that this bacterium could have one of twenty different O-polysaccharides as a part of its LPS (Liu et al., 1983;Liu and Wang, 1990). This is a proposed reason for inconsistent efficiency of some of the vaccines that were formulated based on isolated LPS as target antigens. Therefore, we have shifted our focus to synthetic outer core domains of the LPS that have been shown to react with protective monoclonal antibodies that promoted macrophage-mediated opsonophagocytosis against several serotypes and are also believed to target the rhamnose moiety and neighboring saccharides (Yokota et al., 1989;Terashima et al., 1991). It is noted that LPS is made up of O-polysaccharides linked to Lipid A via an intervening core oligosaccharide. Outer-membrane LPS consists of both smooth-type (S) LPS and rough-type (R) LPS based on the presence and absence of O-polysaccharides, respectively (Rocchetta et al., 1999;Raetz and Whitfield, 2002). Both the (S)-type laboratory strains and (R)-type clinical strains of P. aeruginosa share the presence of structurally similar outer-core glycoform I (1) and glycoform II (2) (Figure 1) consisting of one D-galactosamine (GalN) residue, three D-glucose residues, and one L-rhamnose residue. The GalN residue is acylated with an L-alanyl group in all LPSs (Sadovskaya et al., 1998;Sadovskaya et al., 2000;Knirel et al., 2001;Bystrova et al., 2002;Bystrova et al., 2003;Bystrova et al., 2004;Bystrova et al., 2006).
Glycan epitopes common to both S-type and R-type LPS are potentially ideal antigen targets as they are less serotype dependent. However, isolation of these minimal epitopes from the biological source is not possible. Therefore, it is essential to develop synthetic strategies to obtain these materials for study. In this context, synthesis of two common trisaccharide fragments found within the outer core domain of P. aeruginosa related to glycoform I and glycoform II, noted as α-L-Rha-(1→6)-α-D-Glc-(1→4)-α-D-GalN-(Ala)-α-aminooxy (3) and α-L-Rha-(1→3)-β-D-Glc-(1→3)-α-D-GalN-(Ala)-α-aminooxy (4) ( Figure 2), respectively, is reported herein. We have selected the succinimidyl group as the aminooxy precursor and introduced it at the preliminary stage of the synthetic strategies based on the previous reports (Marcaurelle et al., 2001;De Silva et al., 2009) for developing biologically stable oxime bonds for glycoconjugate vaccines. The use of N-hydroxysuccinimide (NHS) was essential for introducing the aminooxy (-ONH 2 ) functionality which has already proved very useful in several reports (Bourgault et al., 2014;Ghosh and Andreana, 2014;Ghosh et al., 2016;Ghosh et al., 2020;Kleski et al., 2020). The presence of an aminooxy group at the reducing terminal of the trisaccharides provides a readily available form for conjugation to an appropriate carbonylmodified carrier protein (Agten et al., 2016) without destroying the cyclic structure of the reducing end saccharide. The resulting glycoconjugates can then be evaluated in immunological experiments.

RESULTS AND DISCUSSION
Earlier syntheses of pentasaccharide and trisaccharide fragments of outer core domains corresponding to glycoforms I and II of P. aeruginosa containing a methoxy group and tertbutyldiphenylsilyl (TBDPS)-protected hydroquinone (TPH) as a multifunctional reducing-end capping group have been reported (Komarova et al., 2006;Komarova et al., 2008Komarova et al., , 2012Vartak et al., 2018). In these studies, we sought a route to an aminooxy glycoside that could be used in an oxime-based conjugation which would leave the unique L-Ala-modified galactosamine in its native state.
The target trisaccharide fragments of glycoforms I and II were synthesized as their α-aminooxy glycosides from suitably functionalized monosaccharides using stereoselective sequential glycosylations and functional group manipulations. A set of suitably functionalized donor thioglycoside building blocks, A (Tam and Lowary, 2010), B (Rajput and Mukhopadhyay, 2008), C (He et al., 2019), and D (Mukhopadhyay et al., 2004) (Figure 3), were prepared from the naturally available reducing sugars applying a number of reaction conditions reported earlier. D-(+)-glucose-based thioglycoside donors A and C were obtained from commercially available D-(+)-glucose. L-rhamnose-based glycosyl donors B and D were synthesized starting from L-rhamnose according to reported literature (Mukhopadhyay et al., 2004;Rajput and Mukhopadhyay, 2008;Tam and Lowary, 2010;He et al., 2019).
To access the second trisaccharide common outer core fragment of P. aeruginosa LPS, we selected the chloroacetyl SCHEME 1 | Synthesis of trisaccharide fragment target 3.
With a reasonable amount of acceptor in hand, iodonium ion-mediated stereoselective 1,2-trans glycosylation was achieved with thioglucoside donor C in the presence of a combination of NIS and TMSOTf (as before) at −50°C to afford β (1 → 3) disaccharide 17 in 70% yield. Exclusive formation of compound 17 was confirmed from its spectral analysis: signals at δ 5.53 (d, J 3.7 Hz, 1H, H-1) and 4.93 ppm (d, J 7.7 Hz, 1H, H-1′) in the 1 H NMR and at 103.04 (C-1) and 102.83 ppm (C-1′) in the 13 C NMR spectra. To build the target trisaccharide, oxidative removal of the p-methoxybenzyl (PMB) (Oikawa et al., 1982) group from compound 17 was achieved by the treatment with 2,3-dicholoro-5,6-dicyano-1,4-benzoquinone (DDQ) to give disaccharide acceptor 18 in 57% yield. Acceptor 18 and rhamnoside donor D were then coupled using the NIS:TMSOTf (as before) promoter system at −20°C to furnish the product 19 in 65% yield. From spectral analysis it was determined that the product was exclusively the desired trisaccharide without producing unwanted orthoester. The stereochemistry at the glycosidic linkages in compound 19 was confirmed from its spectral analysis: signals at δ 5.52 (d, J 3.7 Hz, 1H, H-1), 5.06 (d, J 7.3 Hz, 1H, H-1′), and 4.83 ppm (brs, 1H, H-1″) in the 1 H NMR and at 102.96 (C-1), 102.27 (C-1′), and 97.78 ppm (C-1″) in the 13 C NMR spectra. The global deprotection was achieved by treatment of the compound 19 with 80% acetic acid at 80°C to remove the benzylidene acetal groups. This step was followed by acetylation with acetic anhydride and pyridine to afford trisaccharide 20 in 40% yield over two steps. The azido group was reduced to an amine by the treatment with Zn/AcOH and coupled with Boc-Ala-OH in the presence of T3P in one pot to obtain trisaccharide derivative 21 in 66% yield in two steps. Finally, trisaccharide was subjected to deprotection reactions including (a) removal of the Boc-group using trifluoroacetic acid (TFA) and (b) removal of the remaining acetyl groups and the replacement of the succinimidyl protecting group as the α-aminooxy group by the treatment with hydrazine monohydrate to afford the target trisaccharide 4 in 35% overall yield (Scheme 2). The formation of compound 4 was confirmed by spectroscopic analysis: signal δ 5.00 (brs, 1H, H-1″), 4.86 (d, J 4.0 Hz, 1H, H-1), and 4.43 ppm (d, J 8. 0 Hz,1H, in the 1 H NMR and at 103.50 (C-1′), 100.91 (C-1″), and 100.44 ppm (C-1) in the 13 C NMR spectra.
Difficult-to-remove side products were generated during the removal of the succinimide group due to the use of excess hydrazine monohydrate (Renaudet and Dumy, 2004;Ghosh and Andreana, 2014). Furthermore, it was challenging to purify both deprotected compounds using only a C18 silica gel column. The problem was solved by passing the compounds through size exclusion chromatography (Bio-gel P-2) using water as the eluent. Fractions containing aminooxy sugars 3 and 4 (identified by TLC staining) were collected, frozen, and lyophilized. The resulting white solid was characterized by NMR and HRMS.

CONCLUSION
In conclusion, oligosaccharide fragments corresponding to the outer core domain of P. aeruginosa LPS were synthesized in good yield using a sequential glycosylation approach. During the synthesis of the target molecules, similar reaction conditions were used in each of the intermediate glycosylation reaction steps to obtain excellent stereochemical outcomes. This report again presents the importance of the NHS group for the synthesis of aminooxy glycosides where the base-sensitive NHS group remains stable through multiple glycosylations, protecting group modifications, and deprotections. However, care was required to avoid basic conditions to maintain NHS stability. For example, the modification of TBAF with 2 equivalents of acetic acid was essential for NHS stability. Furthermore, the route to compound 3 was tolerant of carefully controlled hydrogenolysis to avoid reduction of the N-O bond. Formation of the aminooxy linkage at the reducing end is expected to afford a convenient handle for highly chemoselective oxime conjugation with an appropriately modified carrier protein.

EXPERIMENTAL General Methods
All chemicals and solvents were purchased from Fisher Scientific, Acros Organics, Alfa Aesar, or Sigma-Aldrich. Reactions are carried out under an atmosphere of nitrogen using a nitrogen balloon. Solvents were dried using a solvent purification system by passing them through activated alumina and copper catalyst columns. Reactions were monitored by TLC (silica gel, f 254 ) under UV light or by charring (5% H 2 SO 4 -MeOH), and the purification was performed by column chromatography on silica gel (230-400 mesh), C-18, and P-2 biogel using the solvent system specified; solvents were used without purification for chromatography. 1 H NMR was recorded on a Bruker Avance III 600 MHz spectrometer using CDCl 3 and D 2 O as an internal reference. 13 C were recorded on a Bruker Avance III 600 MHz spectrometer using CDCl 3 and D 2 O as the internal reference. Highresolution mass spectrometry was recorded on a Thermo LTQ XL Orbitrap instrument from the Ohio State University Mass Spectrometry Center.