Synthesis and binding affinity analysis of α1-2- and α1-6-O/S-linked dimannosides for the elucidation of sulfur in glycosidic bonds using quartz crystal microbalance sensors
Graphical abstract
Introduction
With the increasing awareness of the role of glycans as key recognition entities in many fundamental cellular processes, more carbohydrate structures are targeted for therapeutic development, with several drugs already on the market [1], [2], [3]. Many glycans and glycoconjugates have so far been associated with cancer, viruses, bacteria and parasites, and have thus been targeted as vaccines. However, since carbohydrates are much less immunogenic than peptides, glycans conjugated with adjuvants have recently received much attention [1], [4], [5]. However, enzymatic degradation of the glycan-antigen remains a problem, thereby reducing the immune response, and more stable glycan analogs are needed. In this context, several types of glycans have been envisaged in which the glycosidic oxygen linkage has been exchanged for a carbon, sulfur or nitrogen atom [6], [7], [8], [9], [10]. Although these analogs are generally more stable, the structural modification may affect the resulting recognition effects. Not only may the interactions with the glycosidic bond change, but the overall conformation of the glycan is an important factor to consider [11].
Studies have shown that various S-linked disaccharides, such as thiolactose [10], thiomaltose [12], thiocellobiose [13], and glycosyldisulfides [14], display increased conformational flexibility compared to their natural analogs, and varying degrees of inhibition towards targeted enzymes. Contrary to these findings, Zhong et al. found that a series of α1-3- and α1-6-linked thiooligomannosides, derived from the natural substrate, resulted in complete loss of binding when tested against the enzyme Golgi α-mannosidase II [15]. Glycosidases have proven unable to hydrolyze thioglycosides [16], [17], [18], [19], [20], hypothetically due to the poor hydrogen bonding ability of the sulfur atom, thus hampering the acidic cleavage of S,O-acetals [21], [22], [23]. This effect is however not universal, and O- and S-linked N-acetylglucosamine (GlcNAc) residues have been shown to undergo cleavage with comparable efficiencies in the presence of a human O-GlcNAcase [24]. Although these effects are difficult to predict, thioglycosides are progressively being identified as new therapeutics [25]. For instance, recent studies explored S-linked glycan analogs as hydrolytically stable immunogens, used either by themselves or coupled to tetanus toxoid or bovine serum albumin for increased immunogenicity. The results showed that the glycans efficiently induced immune responses, which detected not only the S-linked glycan analog but also the naturally occurring O-linked glycan target [26], [27], [28]. Another example is morphine- and codeine-6-glucuronide (M6G and C6G), two abundant analgesic metabolites of morphine which are both more potent than morphine and display less side effects making them more attractive as analgesics. However, the bioavailability is low due to hydrolysis of the glycosidic bond, an issue which was addressed by MacDougall et al. by introducing an S-linkage [29]. The results showed that the thioanalogs were full agonists with M6G and that the S-linkage is a viable strategy for improving the pharmacological properties. Glycan thiols further present potential advantages in the in vitro synthesis of glycoconjugates/-proteins where site-specific glycosylation is problematic due to the many competing functional groups in natural glycans. Sulfhydryl groups present increased nucleophilicity compared to hydroxyl groups, can be oxidized to the corresponding disulfides, as well as react selectively in many other types of reactions [30], [31], [32], [33], [34], [35], [36], [37].
Herein, the role of sulfur in specific glycosidic bonds, with respect to protein binding effects, has been evaluated using quartz crystal microbalance (QCM) instrumentation where the chips were functionalized with α-D-mannopyranoside monosaccharides [38], [39], [40], [41]. We present the synthesis of a series of O- and S-linked dimannosides, together with subsequent affinity evaluation against lectins with known specificity towards oligomannoside glycans. The binding studies were performed as competition assays where the binding of the tested lectins towards mannose-functionalized surfaces was detected in a QCM flow-through instrumentation, which also enabled comparison of the technique with isothermal titration calorimetry (ITC). Recurring injections of solutions of lectin with the dimannoside inhibitors of varying concentration generated inhibitory response curves, which fitted well to the binding data. The resulting EC50-values correlated well Kd-values estimated by ITC for the O-linked dimannosides.
Section snippets
Results and discussion
Two sets of disaccharides were chosen for the study (Fig. 1), based on previous binding affinity studies towards the target lectin Con A [42]. α1-6-Linked dimannoside 1 has been shown to bind with relatively low affinity (similar to the affinity of methyl α-D-mannopyranoside) whereas the corresponding α1-2-linked dimannoside 3 displayed a 17-fold increase in affinity. Consequently, binding affinity analysis of the target disaccharides would illustrate the effect of sulfur in glycosidic
Conclusions
A series of O- and S-linked carbohydrates have been designed and synthesized for the evaluation of S-glycosides in lectin binding. Glycosyl acceptors possessing a single hydroxyl/sulfhydryl group were synthesized in few steps in generally high yields. The O-linked disaccharides were obtained with thioethyl-linked glycosyl donors under NIS/TfOH conditions. The same methodology however failed for the S-linked disaccharides, likely because of low selectivity between the sulfur entities. The
Experimental section
General Methods: All commercially available starting materials were of reagent grade and used as received. 1H and 13C Nuclear Magnetic Resonance (NMR) data were recorded on a Bruker Avance 400 instrument at 400 MHz (1H) or a Bruker DMX 500 instrument at 500 MHz (1H) or 125 MHz (13C). Chemical shifts are reported as δ values (ppm) with either CHCl3/CDCl3 (1H: δ = 7.26, 13C = 77.16) or HDO (1H: δ = 4.79) as internal standards. 1H peak assignments were made by first order analysis of the spectra
Acknowledgments
This study was supported in part by the Swedish Research Council, the National Institutes of Health (R01GM080295 and 2R15GM066279, to MY), the National Nature Science Foundation of China (Nos. 21272083), the Royal Institute of Technology and Attana AB. NT thanks the Wenner-Gren Foundation for a postdoctoral fellowship grant.
References (60)
- et al.
Eur. J. Org. Chem.
(2016) - et al.
Carbohydr. Res.
(1994) - et al.
FEBS Lett.
(1998) - et al.
Bioorg. Med. Chem.
(2006) - et al.
J. Biol. Chem.
(1996) - et al.
J. Biol. Chem.
(1980) - et al.
Tetrahedron
(2008) - et al.
Biosens. Bioelectron.
(2005) - et al.
Tetrahedron Lett.
(1990) Adv. Carbohydr. Chem. Biochem.
(1997)