Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2015 Aug;22(8):603-10.
doi: 10.1038/nsmb.3053. Epub 2015 Jul 6.

Recognition of Microbial Glycans by Human intelectin-1

Affiliations
Free PMC article

Recognition of Microbial Glycans by Human intelectin-1

Darryl A Wesener et al. Nat Struct Mol Biol. .
Free PMC article

Abstract

The glycans displayed on mammalian cells can differ markedly from those on microbes. Such differences could, in principle, be 'read' by carbohydrate-binding proteins, or lectins. We used glycan microarrays to show that human intelectin-1 (hIntL-1) does not bind known human glycan epitopes but does interact with multiple glycan epitopes found exclusively on microbes: β-linked D-galactofuranose (β-Galf), D-phosphoglycerol-modified glycans, heptoses, D-glycero-D-talo-oct-2-ulosonic acid (KO) and 3-deoxy-D-manno-oct-2-ulosonic acid (KDO). The 1.6-Å-resolution crystal structure of hIntL-1 complexed with β-Galf revealed that hIntL-1 uses a bound calcium ion to coordinate terminal exocyclic 1,2-diols. N-acetylneuraminic acid (Neu5Ac), a sialic acid widespread in human glycans, has an exocyclic 1,2-diol but does not bind hIntL-1, probably owing to unfavorable steric and electronic effects. hIntL-1 marks only Streptococcus pneumoniae serotypes that display surface glycans with terminal 1,2-diol groups. This ligand selectivity suggests that hIntL-1 functions in microbial surveillance.

Figures

Figure 1
Figure 1
hIntL-1 selectivity for monosaccharides. (a) Structures of saccharides used for characterization of hIntL-1 by ELISA and SPR. (b) The specificity of hIntL-1 for β-Galf, β-ribofuranose (β-Ribf) and β-galactopyranose (β-Galp) evaluated by ELISA (See Supplementary Fig. 1b for schematic). Data are presented as the mean ± s.d. (n=3 technical replicates, data are representative of >3 independent experiments). Data were fit to a single site binding equation (solid lines) and therefore represent the apparent (App) affinity of trimeric hIntL-1. Values for hIntL-1 bound to immobilized β-Galf (Kd(App, Trimer) ± s.d.) are 85 ± 14 nM or 8.0 ± 1.3 µg/mL. (c) Representative real-time SPR sensorgrams of hIntL-1 binding to immobilized carbohydrates. Biotin served as a control. The SPR complete data set is available in Supplementary Fig. 1e.
Figure 2
Figure 2
Glycan selectivity of hIntL-1 assessed by glycan microarrays. (a) Recombinant hIntL-1 (50 µg/mL) binding to mammalian glycan microarray CFG v5.1 and a furanoside array. The concentrations given for the furanoside array represent those used in the carbohydrate immobilization reaction. Data are presented as the mean ± s.d. (n=4 technical replicates). The full data set can be found in Supplementary Tables 1 and 2. (b) Recombinant Strep-hIntL-1 (50 µg/mL) binding to microbial glycan array. For glycan array data organized by genus, see Supplementary Fig. 2a. Data are presented as the mean ± s.d. (n=4 technical replicates). The full data set can be found in Supplementary Table 3. (c) Structural representation of the putative key binding epitopes for hIntL-1 and the non-binding N-acetylneuraminic acid (α-Neu5Ac). A terminal vicinal diol (red) is a common feature of α-Neu5Ac and all of the ligands identified.
Figure 3
Figure 3
Structure of hIntL-1 bound to allyl-β-d-Galf. (a) Complex of hIntL-1 disulfide-linked trimer and allyl-β-d-Galf. Each monomer unit is depicted in green, wheat, or grey, the β-allyl Galf is shown in black, calcium ions in green, the inter-monomer disulfides in orange, and ordered water molecules in the binding site in red. The two orientations indicate the positioning of all three ligand-binding sites within the trimer. The trimeric structure is produced from Chain A in the asymmetric unit by a three-fold crystallographic operation. (b) Stereo image of the carbohydrate-binding site. Residues involved in calcium coordination and ligand binding are noted. Dashed lines are included to show the heptavalent coordination of the calcium ion and to highlight functional groups important for ligand and calcium ion binding. Difference density map (Fo-Fc, 3σ) of the allyl-β-d-Galf ligand is provided in Supplementary Fig. 4b.
Figure 4
Figure 4
Models for hIntL-1 interacting with relevant saccharide epitopes from humans (α-Neu5Ac) or microbes (α-KDO). (a) Docking of methyl-α-Neu5Ac into the hIntL-1 structure. The conformation shown is similar to that observed in other protein structures with a methyl-α-Neu5Ac ligand (PDB: 2BAT, 2P3I, 2P3J, 2P3K, 2I2S, 1KQR, 1HGE, 1HGH (refs. –60)). All models in this figure were generated from the allyl-β-d-Galf–bound structure by docking the relevant diol of each compound into the Galf diol electron density using Coot without further refinement. Calcium ions are shown in green and ordered water molecules are depicted in red. (b) Docking of methyl-α-KDO into the hIntL-1 structure. Comparison with methyl-α-Neu5Ac docked into the hIntL-1 structure reveal differences in the steric requirements for binding for each molecule.
Figure 5
Figure 5
Human IntL-1 binds to S. pneumoniae serotypes producing capsular polysaccharides with terminal vicinal diols. (a) Chemical structure of the capsular polysaccharides displayed on the S. pneumoniae serotypes (8, 20, 43, 70) tested. The Galf residues assumed to mediate hIntL-1 cell binding are shown in red and the phosphoglycerol moiety is shown in blue. (b) Fluorescence microscopy of hIntL-1 binding to S. pneumoniae serotype 20. Bacteria were treated with Strep-tagged hIntL-1 (15 µg/mL) and an anti-Strep-tag antibody conjugate (red). Cellular DNA was visualized with Hoechst (blue). Left: hIntL-1 marks the surface of serotype 20 bacteria; serotypes 43 and 70 gave similar results; no binding to serotype 8 was detected (Supplementary Fig. 6a) Right: EDTA addition abrogates hIntL-1 binding to the bacterial surface, supporting the role for Ca2+. Images are representative of >5 fields of view per sample. Scale bar, 2 µm. (c,d) Flow cytometry analysis of Strep-hIntL-1 binding to S. pneumoniae serotypes with an anti-Strep-tag antibody conjugate. In the anti-Strep control sample, recombinant hIntL-1 was omitted. Cells were labeled with propidium iodide. (c) Flow cytometry analysis of serotypes 8, 20, 43, and 70; data were collected consecutively with identical instrument settings. (d) The dependence of the hIntL-1–carbohydrate interaction on Ca2+ was tested by adding 10 mM EDTA and ligand selectivity was tested by adding 100 mM glycerol. Data are representative of two independent experiments. For a similar analysis of serotypes 20 and 70 (no binding to serotype 8 was detected), see Supplementary Fig. 6b.
Figure 6
Figure 6
Structures of the 20 most prevalent monosaccharides that are unique to bacterial glycans. The most common, l,d-α-heptose, is shown in the top left corner and number twenty, β-l-arabinose-4-N, is shown in the bottom right. This figure is derived from data in reference 32. Terminal acyclic 1,2-diol epitopes that could serve as ligands of hIntL-1 are highlighted with a red box.

Similar articles

See all similar articles

Cited by 35 articles

See all "Cited by" articles

References

    1. Stowell SR, et al. Innate immune lectins kill bacteria expressing blood group antigen. Nat. Med. 2010;16:295–301. - PMC - PubMed
    1. Vaishnava S, et al. The antibacterial lectin RegIIIgamma promotes the spatial segregation of microbiota and host in the intestine. Science. 2011;334:255–258. - PMC - PubMed
    1. Gallo RL, Hooper LV. Epithelial antimicrobial defence of the skin and intestine. Nat. Rev. Immunol. 2012;12:503–516. - PMC - PubMed
    1. Backhed F, Ley RE, Sonnenburg JL, Peterson DA, Gordon JI. Host-bacterial mutualism in the human intestine. Science. 2005;307:1915–1920. - PubMed
    1. Varki A. Essentials of glycobiology. Cold Spring Harbor, N.Y.: Cold Spring Harbor Laboratory Press; 2009. p. 784.

Methods references

    1. Blixt O, et al. Printed covalent glycan array for ligand profiling of diverse glycan binding proteins. Proc. Natl. Acad. Sci. U.S.A. 2004;101:17033–17038. - PMC - PubMed
    1. Stowell SR, et al. Microbial glycan microarrays define key features of host-microbial interactions. Nat. Chem. Biol. 2014;10:470–476. - PMC - PubMed
    1. Bentley SD, et al. Genetic analysis of the capsular biosynthetic locus from all 90 pneumococcal serotypes. PLoS Genet. 2006;2:262–269. - PMC - PubMed
    1. Song X, Lasanajak Y, Xia B, Smith DF, Cummings RD. Fluorescent glycosylamides produced by microscale derivatization of free glycans for natural glycan microarrays. ACS Chem. Biol. 2009;4:741–750. - PMC - PubMed
    1. Heimburg-Molinaro J, Song X, Smith DF, Cummings RD. Preparation and analysis of glycan microarrays. Curr. Protoc. Protein Sci. 2011;12(12):10. - PMC - PubMed

Publication types

MeSH terms

Feedback