Comprehensive Modular Synthesis of Ganglioside Glycans and Evaluation of their Binding Affinities to Siglec-7 and Siglec-9

Abstract

In the present work, bacterial glycosyltransferases are utilized to construct ganglioside glycans in a convergent approach via a sugar‒nucleotide regeneration system and one-pot multienzyme reactions. Starting from β-lactoside enables the diversification of both the glycan moieties and the linkages in the lower α-arm and upper β-arm. Overall, a comprehensive panel of 24 natural a-series (GM3, GM2, GM1a, GD1a, GT1a, and fucosyl-GM1), b-series (GD3, GD2, GD1b, GT1b, and GQ1b), c-series (GT3, GT2, GT1c, GQ1c, and GP1c), α-series (GM1α, GD1aα, and GT1aα), and o-series (GA2, GA1, GM1b, GalNAc-GM1b, and GD1c) ganglioside glycans are prepared, which are suitable for biological studies and further applications. Moreover, a microarray is constructed with these synthesized ganglioside glycans to investigate their binding specificity with recombinant Fc-fused Siglec-7 and Siglec-9, which are immune checkpoint-like glycan recognition proteins on natural killer cells. The microarray binding results reveal that GD3 and GT1aα are specific ligands for Siglec-7 and Siglec-9, respectively, and this discovery can lead to the identification of appropriate ligands for investigating the roles of these Siglecs in immunomodulation.

Keywords: Siglec; carbohydrates; chemoenzymatic synthesis; gangliosides; glycan microarray.

Figure 1

The general chemical structure of a ganglioside glycan (top panel) and structures of major o‐, a‐, b‐, c‐, and α‐series ganglioside glycans (bottom panel) synthesized and evaluated for their binding specificity with Siglec‐7‐Fc and Siglec‐9‐Fc in this study.

Scheme 1

Synthesis of ganglioside glycans by the SNRS or SOPME systems coupled with glycosyltransferases. See supporting information for reaction system abbreviation and details conditions.

Scheme 2

Evaluation of bacterial ST‐catalyzed α2,6‐sialylation of ganglioside glycans for the synthesis of 23 (GM1α), 25 (GD1aα), and 27 (GT1aα). ^aIsolated yield from the mixture of GD1aα’ and GD1aα’’ (the isolated yield of GD1aα’ was 2%). ^bIsolated yield from the mixture of regioisomers (ratio of 8:1) (isolated yield of GT1aα was 7%).

Figure 2

Comparison of the ¹³C NMR data of: A) GM3, GM2, and GM1a; B) GD3, GD2, and GD1b; C) GA1, GM1a, GM1b, and GD1a; and D) the anomeric carbons in the sialic acid residues of GM1b, GT1b, GQ1b, GQ1c, and GP1c. A, B, C, D, S, and S’ represent Glc, Gal, GalNAc, Gal, α3Neu5Ac at lower arm, and α3Neu5Ac at upper arm (or α8Neu5Ac) moiety, respectively from reducing end. The number indicates the position of the carbon atom. (a) Additional sialic acid attached to upper α(2,3)sia moiety with α2,8‐sialyl linkage (from GT1b to GQ1b) would shift the anomeric carbon of upper α(2,3)sia into the downfield region. (b) Additional sialic acid attached to lower α(2,8)sia moiety with α2,8‐sialyl linkage (from GT1b to GQ1c) would not shift the anomeric carbon of upper α(2,3)sia into the downfield region. (c) ¹³C NMR spectra of GQ1c in comparison with that of GP1c, original anomeric carbon of upper α(2,3)sia was shifted to the downfield region, agreeing with the results observed from (a) and (b) and confirming the sialyl linkages of GP1c.

Figure 3

Binding of human Siglec‐7‐Fc to the ganglioside glycan microarray. Binding signals were quantified by fluorescence intensity, as shown in the bar graphs in the panels, which represent the means ± SDs. b‐Series glycans that contain the GD3 motif presented higher binding affinities than other glycans with similar numbers of Neu5Ac residues.

Figure 4

Binding of human Siglec‐9‐Fc to the ganglioside glycan microarray. Binding signals were quantified by fluorescence intensity, as shown in the bar graphs in the panels, which represent the means ± SDs.