Broad-Spectrum Legionaminic Acid-Specific Antibodies in Pooled Human IgGs Revealed by Glycan Microarrays with Chemoenzymatically Synthesized Nonulosonosides

Abstract

The presence and the level of antibodies in human sera against bacterial glycans are indications of prior encounters with similar antigens and/or the bacteria that express them by the immune system. An increasing number of pathogenic bacteria that cause human diseases have been shown to express polysaccharides containing a bacterial nonulosonic acid called 5,7-di-N-acetyllegionaminic acid (Leg5,7Ac₂). To investigate the immune recognition of Leg5,7Ac₂, which is critical for the fight against bacterial infections, a highly effective chemoenzymatic synthon strategy was applied to construct a library of α2-3/6-linked Leg5,7Ac₂-glycans via their diazido-derivatives (Leg5,7diN₃-glycans) formed by efficient one-pot three-enzyme (OP3E) synthetic systems from a diazido-derivative of a six-carbon monosaccharide precursor. Glycan microarray studies using this synthetic library of a Leg5,7Ac₂-capped collection of diverse underlying glycan carriers and their matched sialoside counterparts revealed specific recognition of Leg5,7Ac₂ by human IgG antibodies pooled from thousands of healthy donors (IVIG), suggesting prior human encounters with Leg5,7Ac₂-expressing pathogenic bacteria at the population level. These biologically relevant Leg5,7Ac₂-glycans and their immune recognition assays are important tools to begin elucidating their biological roles, particularly in the context of infection and host-pathogen interactions.

Keywords: bacterial nonulosonic acid; carbohydrate; chemoenzymatic synthesis; human antibodies; legionaminic acid.

Figure 1

Structures of N-acetylneuraminic acid (Neu5Ac, 1); 5,7-di-N-acetyllegionaminic acid (Leg5,7Ac₂, 2); Leg5,7Ac₂ precursor 6deoxyMan2,4diNAc (3); and Leg5,7Ac₂ chemoenzymatic synthon 6deoxyMan2,4diN₃ (4).

Scheme 1

Chemoenzymatic synthesis of α2–6-linked Leg5,7Ac₂-glycosides (25–31) via OP3E produced α2–6-linked Leg5,7diN₃-glycosides (12–18).

Scheme 2

Chemoenzymatic synthesis of α2–3-linked Leg5,7Ac₂-glycosides (32–37) via OP3E produced α2–3-linked Leg5,7diN₃-glycosides (19–24).

Figure 2

Glycan microarray binding studies against Leg5,7Ac₂-glycans and sialoglycan counterparts. (a) Glycan microarrays fabricated with Leg5,7Ac₂-glycans and sialoglycans (Neu5Gc/Neu5Ac/Neu5Ac7NAc-glycans) were examined with IVIG (IVIG-1 Privigen or IVIG-2 OMRIgG-AR; 50 μg/sub-array), anti-Neu5Gc IgY (αNeu5Gc; 1:7000), SiaFindα2-6 (0.25 μg), and 11 biotinylated plant lectins (2 μg each: MAL-II, LEL, PNA, WGA, VVL, AAL, LcH, ConA, ECL, PSA, and GSL-II) at 100 μL/subarray in blocking buffer (PBS, 1% ovalbumin; the assay buffers for plant lectins also contained divalent cations, see ESI), then detected with Cy3-anti-human IgG (0.4 ng/μL) for IVIG, Cy3-anti-chicken IgY (0.375 ng/μL) for IgY, and Cy3-sterptavidin (1.2 ng/μL) for Sia-Find and Lectins, at each 200 μL/sub-array. Slides were scanned and analyzed for Relative Fluorescence Units (RFU), which were plotted as a heatmap. (b) Binding of IVIG (average IVIG-1 and IVIG-2) to Leg5,7Ac₂-glycans compared to Neu5Ac7NAc-glycans, non-Sia glycans, and glycans terminated with Neu5Ac, Neu5,9Ac2, Neu5Gc, and Neu5Gc9Ac (mixed effects analysis, with the Geisser-Greenhouse correction and individual variances computed for each comparison using uncorrected Fisher’s LSD; * p = 0.04, ** p = 0.001, *** p < 0.001).

Figure 3

Characterization of IVIG binding to Leg5,7Ac₂-glycans. (a) Binding of IVIG-2 (OMRIgG-AR) to glycan microarray at 50 μg/sub-array, after pre-treatment of the arrays with/without 10 mU sialidase (Arthrobacter ureafaciens sialidase, AUS; Clostridium perfringens neuraminidase; NCP) for 2 h at 37 °C, then detected with 0.4 ng/μL Cy3-anti-human IgG. Slides were scanned, analyzed, and RFU-plotted (mixed-effects analysis, with the Geisser-Greenhouse correction, and individual variances computed for each comparison using uncorrected Fisher’s LSD; ns = non-significant p > 0.05 (p = 0.6132, 0.3334, 0.7287, 0.7454 None vs AUS/NCP in Leg/Neu, respectively), * p = 0.02, ** p < 0.003). (b) Averaged binding of IVIG-1 and IVIG-2 to Leg5,7Ac₂-glycans on glycan microarrays in the presence of selected competing glycans. IVIG (IVIG-1 Privigen or IVIG-2 OMRIgG-AR, 50 μg/sub-array) were pre-incubated for 2 h on ice with various inhibitors on each sub-array, including 2-O-methyl-αNeu5Gc (Gc2Me; 4 mM), glycan ID-51 (Core1β, 2 mM), mouse serum glycopeptides (mGP, containing N/O-glycans [46], 0.45 mM), glycan ID-502 (Leg5,7Ac₂α3Core1β, 1 mM), and glycan ID-508 (Leg5,7Ac₂α6Core1β; 1 mM). Then, binding of the pre-complexed IVIG was examined and detected with 0.4 ng/μL Cy3-anti-human IgG. Slides were scanned, analyzed, and RFU results were plotted. Percent binding to each glycan was calculated as a ratio to binding of IVIG without inhibitor (none, 100% binding). (c) Statistical analysis of inhibition of IVIG binding to Leg5,7Ac₂-glycans of data presented in (b) (Friedman test with uncorrected Dunn’s test; ns p = 0.4631, * p = 0.0159, *** p = 0.0002, and **** p < 0.0001).