Synthesis, Antigenicity Against Human Sera and Structure-Activity Relationships of Carbohydrate Moieties from Toxocara larvae and Their Analogues

Stereocontrolled syntheses of biotin-labeled oligosaccharide portions containing the Galβ1-3GalNAc core of the TES-glycoprotein antigen obtained from larvae of the parasite Toxocara and their analogues have been accomplished. Trisaccharides Fuc2Meα1-2Gal4Meβ1-3GalNAcα1-OR (A), Fucα1-2Gal4Meβ1-3GalNAcα1-OR (B), Fuc2Meα1-2Galβ1-3GalNAcα1-OR (C), Fucα1-2Galβ1-3GalNAcα1-OR (D) and a disaccharide Fuc2Meα1-2Gal4Meβ1-OR (E) (R = biotinylated probe) were synthesized by block synthesis using 5-(methoxycarbonyl)pentyl-2,3,4,6-tetra-O-acetyl-β-D-galactopyranosyl-(1→3)-2-azide-4-O-benzyl-2-deoxy-α-D-galactopyranoside as a common glycosyl acceptor. We examined the antigenicity of these five oligosaccharides by enzyme linked immunosorbent assay (ELISA). Our results demonstrate that the O-methyl groups in these oligosaccharides are important for their antigenicity and the biotinylated oligosaccharides A, B, C and E have high serodiagnostic potential to detect infections caused by Toxocara larvae.


Introduction
In the course of our studies on unique glycoconjugates found in parasites, we have synthesized various glycosphingolipids and carbohydrate portions of glycoproteins with the aim of elucidating the mechanisms of host-parasite interactions [1][2][3][4][5][6][7]. In previous studies we have synthesized unusual carbohydrates from the parasites Echinococcus multilocularis [1,3,6], Schistosoma mansoni [2] and porcine roundworm (nematode) Ascaris suum [7]. In this paper we describe the synthesis of the carbohydrate portion of glycoproteins from Toxocara canis and T. cati. T. canis and T. cati are parasitic roundworms and are widely distributed in dogs and cats. Both nematodes cause severe infections in a human host affecting eyes, liver and the central nervous system [8,9]. Khoo et al., isolated Toxocara excretory-secretory (TES) antigen, a family of glycoproteins that are heavily O-glycosylated from the culture media of T. canis and T. cati larvae [10]. The TES antigen of T. canis is a mixture of mucin-type glycoproteins, containing a Fuc1-2Gal1-3GalNAc structure, and it has been found that the fucose part was O-methylated at the 2-position and approximately 50% of the galactose part was O-methylated at the 4-position, i.e., it contains the following two sequences  Figure 1) and studied their antigenicity [11,12]. The results showed that the sera from infected patients recognized the DiM more strongly than MoM α,β [11]. However, the both groups have not studied the effects of the di-O-methylated trisaccharide.

Chemical Synthesis
Syntheses of the target oligosaccharides A-E: In all cases we selected 5-(methoxycarbonyl)pentyl group as the protecting group of reducing end, because this group can be conveniently used for conjugation with biotin for the use in ELISA assay as previously shown by us [1]. The synthetic routes for target compounds A-E are outlined in Schemes 1-5.

Scheme 1. Synthesis of di-methylated trisaccharide A.
Non-methylated fucopyranosyl donor 13 was synthesized from known donor 12 [16] by two-step procedure (Scheme 2). Hydrolysis of the isopropylidene group in 12 with aqueous AcOH and subsequent benzoylation produced fucopyranosyl donor 13. Trisaccharide 14 was synthesized by a coupling of the fucopyranosyl donor 13 with the disaccharide acceptor 5. The presence of an α fucosidic linkage in 14 was indicated by a doublet at δ 5.71 ppm showing small homonuclear coupling constant of 3.3 Hz in the 1 H-NMR spectrum. Deprotection and biotinylation were performed as described for compound A to provide target trisaccharide B (Scheme 2).
The synthesis of the trisaccharide C is outlined in Scheme 3. As C does not contain the O-methyl substitution in the galactose residue, we chose phenyl 3,4,6-tri-O-benzyl-2-O-benzoyl-1-thio-β-Dgalactopyranoside (18) [17] as a galactosyl donor. Glycosylation of acceptor 3 with donor 18 in the presence of NIS/TfOH provided disaccharide 19 in 68% yield. Selective removal of the benzoyl group in 19 with NaOMe produced disaccharide acceptor 20. Fucosylation of 20 with the donor 7 afforded trisaccharide 21, whose benzylidene acetal was removed as described for 8 and 14 to give 22 in 88% yield. Reduction of the azide group together with removal of benzyl groups of 22 was initially performed as described for compounds 9 and 15 using Zn/Cu and Pd(OH) 2 . However, this resulted only in a 24% yield of 23. In contrast, significantly improved yield was obtained by catalytic hydrogenation of 22 with Pd-C followed by acetylation to produce the desired trisaccharide 23 in 60% yield. Deacylation and biotinylation were performed as described for compound A to give target trisaccharide C (Scheme 3).

Scheme 3. Synthesis of mono-methylated trisaccharide C.
Non-methylated trisaccharide D was synthesized from the glycosyl donor 13 and acceptor 20 as described for compound C in a good yield (Scheme 4).
Dimethylated disaccharide E, which does not contain a GalNAc residue was synthesized from galactosyl acceptor 30 with the fucosyl donor 7 as outlined in Scheme 5. Compound 30 was obtained by coupling of galactose-based thiophenyl glycoside 2 to methyl 6-hydroxyhexanoate and subsequent debenzoylation using standard conditions. Glycosylation of the acceptor 30 with 7 in the presence of MeOTf and TTBP afforded desired disaccharide 31 in 72% yield. Unfortunately, the anomeric ratio of the fucopyranosyl linkage was 10:1 (α:β, from 1 H-NMR). Although we cannot properly explain the reason of it, we presume the influence by the absence of a GalNAc derivative in monosaccharide acceptor 30. Deprotection of benzyl groups by catalytic hydrogenolysis using Pearlman's catalyst provided 32. It was possible to separate the major α-anomer form the minor β-glycoside at this stage. Deacylation and biotinylation were performed as described for compound A to provide target trisaccharide E in a good yield (Scheme 5).

Antigenicity of Oligosaccharides by ELISA
The reactivity of the five biotin-labeled oligosaccharides A (natural), B, C (natural), D and E to patients' sera was examined using streptavidin-coated microplates. In four of the five oligosaccharides, except for D, the antibody response of the T. canis-infected patient group (Tc) was significantly high compared with that of the normal healthy (N) group ( Figure 3). However, the non-methylated trisaccharide D did not show any antigenicity to the patients' sera. This is in contrast to previous findings by Kosma and Maizels, who have reported marked differences in the antigenicity between dimethylated disaccharide (DiM) and monomethylated trisaccharide (MoMα,β in the form of BSA conjugates [11]. Our study indicates that not only the natural monomethylated trisaccharide C containing a mono-Fuc2Me moiety but also non-natural trisaccharide B containing a mono-Gal4Me moiety have the antigenicity. Moreover, the antigenicity of di-methylated trisaccharide A was stronger than that of di-methylated disaccharide E indicating that the GalNAc portion contributes to the antigenicity. Significant differences of A-C and E were observed between Tc patient group and N group (p < 0.05, Student's t test).

Serum Samples
Serum samples examined by ELISA were obtained from 6 patients who were confirmed to have Toxocara canis-visceral larva migrans (Tc) and 4 normal healthy individuals (NH).

Conclusions
We have prepared oligosaccharide-biotin conjugates A-E in order to study the antigenicity of putative carbohydrate sequences at the parasite T. canis and their analogues. Antigenicity of these compounds was examined by ELISA. Mono-or di-O-methylated forms showed good serodiagnostic potential to detect infections caused by T. canis. These results demonstrate that biotin-labeled oligosaccharides may serve as a diagnostic tool to detect T. canis and T. cati infections in humans.