Glycan:glycan interactions: High affinity biomolecular interactions that can mediate binding of pathogenic bacteria to host cells

Abstract

Cells from all domains of life express glycan structures attached to lipids and proteins on their surface, called glycoconjugates. Cell-to-cell contact mediated by glycan:glycan interactions have been considered to be low-affinity interactions that precede high-affinity protein-glycan or protein-protein interactions. In several pathogenic bacteria, truncation of surface glycans, lipooligosaccharide (LOS), or lipopolysaccharide (LPS) have been reported to significantly reduce bacterial adherence to host cells. Here, we show that the saccharide component of LOS/LPS have direct, high-affinity interactions with host glycans. Glycan microarrays reveal that LOS/LPS of four distinct bacterial pathogens bind to numerous host glycan structures. Surface plasmon resonance was used to determine the affinity of these interactions and revealed 66 high-affinity host-glycan:bacterial-glycan pairs with equilibrium dissociation constants (K(D)) ranging between 100 nM and 50 µM. These glycan:glycan affinity values are similar to those reported for lectins or antibodies with glycans. Cell assays demonstrated that glycan:glycan interaction-mediated bacterial adherence could be competitively inhibited by either host cell or bacterial glycans. This is the first report to our knowledge of high affinity glycan:glycan interactions between bacterial pathogens and the host. The discovery of large numbers of glycan:glycan interactions between a diverse range of structures suggests that these interactions may be important in all biological systems.

Keywords: adherence; glycoconjugates; lipooligosaccharide; lipopolysccharide.

Fig. 1.

Structures of LOS/LPS used in this study. LOS/LPS structures of the organisms used in this study. *, C. jejuni 11168-O is 90% sialylated at 37 °C and less than 50% at 42 °C (59). The underlined section is the terminal structure used in the ITC experiment. #, H. influenzae 2019 LOS is phase variable only one product is shown (60).

Fig. 2.

Analysis of LPS/LOS glycan interactions with glycan array and SPR. Heat map of glycan binding from glycan array and SPR experiments. Comparison between whole bacteria and isolated LOS/LPS molecules. Binding to glycans on the array have been grouped into common terminating structures; for a full list of binding to individual structures see Dataset S1. Interactions are noted in red for binding or white if no binding was observed by glycan array analysis.

Fig. 3.

Analysis of affinities of interactions between bacterial LOS/LPS structures and free host glycans by SPR. Heat map of affinities of binding between bacterially derived LOS/LPS structures with various host glycans. Measurements were done with whole LOS/LPS and amine-modified O-deacylated LOS/LPS glycans. Increased affinity is indicating by a darker color (see key). A white box with an “N” indicates that no concentration-dependent interaction could be identified. Full K_D measurements can be found in Tables S1–S4.

Fig. 4.

(A) Analysis of interactions between synthetic glycans of structures matching C. jejuni terminal LOS structure [asialoG_M1]; (B) and blood group B tetra saccharide (C) by isothermal calorimetry. Sigmoidal fit of the interaction with self interactions for aG_M1 and blood group B and heat of injection subtracted.

Fig. S1.

Representative thermodynamic plot of glycan-glycan interactions using ITC. (A) Representative response curve for interaction between blood group B trisaccharide and asialo-G_M1 tetrasaccharide. y axis is in µJ/s and x axis is time (s). (B) Sigmodal curve of blank injection (PBS into aG_M1) and blood group B only (blood group B into PBS) injections.

Fig. 5.

Adherence assays in the presence of free host or bacterial glycans. (A) C. jejuni 11168-O wild-type adherence to Caco-2 epithelial cells in the presence or absence of free blood group antigens (A-trisaccharide 0.8 µM, 8 µM, 80 µM; B-trisaccharide 0.014 µM, 0.14 µM, 1.4 µM; or H-disaccharide 0.14 µM, 1.4 µM, 14 µM); C. jejuni 11168-O 42 °C LOS oligosaccharide (Cj OS) at 0.33 and 3.3 µM. (B) S. flexneri adherence to T-84 epithelial cells in the presence or absence of free blood group antigens (A-trisaccharide 0.08 µM, 0.81 µM, 8.1 µM; B-trisaccharide 84.9 µM, 849 µM or H-disaccharide 803 µM). (C) S. flexneri adherence to T-84 epithelial cells in the presence or absence of free bacterial oligo/polysaccharides Sf S-PS (0.32 or 4.3 µM) or Sf S-LPS (4.3 µM) or Sf R-LPS (7.4 or 4.3 µM). (D) C. jejuni adherence to Caco-2 epithelial cells in the presence or absence of antibodies to cell and bacterial surface glycans (A and O blood group antigens and G_M1 ganglioside) at 0.2 and 2 µg/mL. The O blood group antibody will target the O-blood group antigen on the surface of Caco-2 cells with high affinity, the A blood group antibody will target the O blood group and nonfucosylated α-GalNAc structures on the Caco-2 cells with low affinity and the anti-G_M1 antibody will target structures on the Caco-2 cells and C. jejuni 11168-O (G_M1 mimic) with high affinity. (E) S. typhimurium 180 and Caco-2 cells in the presence or absence of free blood group antigens (A-trisaccharide 26 µM; B-trisaccharide 21 µM; or H-disaccharide 19 µM). (F) Cell association between H. influenzae LOS mutants and 16HBE14 bronchial epithelial cells expressed as a percentage of the original inoculums at 2 h. H. influenzae WT (2019); H. influenzae ΔsiaP (siaP). *, significant difference compared with untreated/wild-type control.

Fig. 6.

Labeling of human small intestines shows colocalization of S. flexneri type 2a Alexa 488-polysaccharide and the M-cell marker anti-GP2. Human ileum section viewed at 4x (1) and 40x objectives with phase contrast optics (2 and 3). At 40x, cells were also viewed under fluorescence for Alexa 488-polysaccharide (4; green) and GP2 (5; red). Overlay images (3 and 6) show areas of colocalization in yellow. Arrows point to cells enlarged for better clarity (bottom corner of each frame).