Chemical tools to study bacterial glycans: a tale from discovery of glycoproteins to disruption of their function

Abstract

Bacteria coat themselves with a dense array of cell envelope glycans that enhance bacterial fitness and promote survival. Despite the importance of bacterial glycans, their systematic study and perturbation remains challenging. Chemical tools have made important inroads toward understanding and altering bacterial glycans. This review describes how pioneering discoveries from Prof. Carolyn Bertozzi's laboratory inspired our laboratory to develop sugar probes to facilitate the study of bacterial glycans. As described below, we used metabolic glycan labelling to install bioorthogonal reporters into bacterial glycans, ultimately permitting the discovery of a protein glycosylation system, the identification of glycosylation genes, and the development of metabolic glycan inhibitors. Our results have provided an approach to screen bacterial glycans and gain insight into their function, even in the absence of detailed structural information.

Keywords: bacteria; bioorthogonal chemistry; glycan; glycoprotein; metabolic labeling.

Figure 1.

Sampling of glycan structures found in (A) eukaryotic and (B) bacterial glycosylated proteins, lipopolysaccharides (LPS), and capsular polysaccharides (CPS). Common eukaryotic monosaccharides glucose (Glc), galactose (Gal), mannose (Man), N-acetylglucosamine (GlcNAc), N-acetylgalactosamine (GalNAc), fucose (Fuc), and sialic acid are shown with commonly modified N-acetyl positions highlighted in color. Exclusively bacterial monosaccharides, including bacillosamine (Bac), pseudaminic acid (Pse), legionaminic acid (Leg), N-acetylrhamnosamine (RhaNAc), N-acetylquinovosamine (QuiNAc), N-acetylfucosamine (FucNAc), and N-acetylpneumosamine (PneNAc), are shown uniformly colored to distinguish them from ubiquitous common monosaccharides found in both eukaryotes and prokaryotes. Asparagine-linked (N-linked) and serine/threonine-linked (O-linked) glycans from glycosylated proteins in eukaryotes and bacteria are shown.

Figure 2.

Metabolic glycan labelling (MGL) harnesses the promiscuity of carbohydrate biosynthesis enzymes to install unnatural monosaccharides into cellular glycans. A) This approach was inspired by Reutter and coworkers’ reports demonstrating that sialic acid biosynthesis enzymes are permissive of ManNAc homologs bearing additional methylene moieties, allowing installation of unnatural sialic acids on live cells. B) Bertozzi and coworkers built upon this precedent by incorporating unnatural monosaccharides bearing abiotic functional groups (e.g., azides) into cellular glycans and developing two bioorthogonal chemistries, Staudinger ligation and strain-promoted [3+2] azide-alkyne cycloaddition, that proceed with selectivity in living animals for the covalent delivery of probes to labelled glycans.

Figure 3.

Treatment of H. pylori with Ac₄GlcNAz followed by probing with Phos-FLAG led to the discovery of glycoproteins that had not been previously detected. Anti-FLAG western blot of lysates from treated cells reprinted with permission from Koenigs et al. Copyright 2009 Molecular BioSystems.

Figure 4.

Metabolic glycan labelling facilitated tagging, enrichment, and mass spectrometry-aided identification of 125 putative glycoproteins produced by H. pylori. Beta-elimination of O-linked glycans from enriched glycoproteins yielded a detectable HexNAz-Phos-FLAG-His₆ adduct.

Figure 5.

Metabolic glycan labeling formed the basis of a screen to identify genes involved in glycan biosynthesis in H. pylori. A) Wildtype (WT) H. pylori containing intact protein glycosylation pathways synthesize a robust glycoprotein biosynthesis fingerprint that can be detected by MGL with Ac₄GlcNAz and Phos-FLAG. B) Glycosylation mutants bearing insertionally inactivated glycosyltransferase genes (ΔGT) exhibit glycoprotein biosynthesis defects that are revealed by MGL. C) Western blot analysis with anti-FLAG revealed the full glycoprotein profile synthesized by WT H. pylori and strains containing mutations in GTs not involved in glycoprotein biosynthesis (Δ607, Δ1339); by contrast, common defective glycoprotein biosynthesis fingerprints were revealed in thirteen ΔGT mutants, implicating these mutated genes in the pathway. D) O-linked glycan removal by beta-elimination (+) yielded residual azide-dependent glycoprotein signal that was similar to the defective glycoprotein biosynthesis fingerprint observed in mutants. Western blots reprinted with permission from Moulton et al. Copyright 2020 American Chemical Society.

Figure 6.

Azide-containing rare bacterial monosaccharide analogs facilitated MGL in a range of bacterial pathogens, including C. jejuni, Vibrio vulnificus, and Pleisiomonas shigelloides. Synthesis and evaluation of A) d-sugar analogs and C) l-sugar analogs was described by Clark et al. and Luong et al., respectively. Representative data of metabolic incorporation in a bacteria-selective manner is shown in (B). Western blots reprinted with permission from Clark et al. Copyright 2016 American Chemical Society.

Figure 7.

Metabolic inhibitors interfere with cellular glycan biosynthesis. A) Fluorosugars act as chain terminators and benzyl glycosides act as substrate decoys that divert glycan biosynthesis. B) Fluorosugar and benzyl glycoside analogs of rare bacterial sugars inhibit glycoprotein biosynthesis in H. pylori. C) Lectin binding with ConA reveals altered cell surface glycan architecture in H. pylori upon treatment with metabolic glycan inhibitors. Carbo block represents the fluorescence intensity observed when untreated cells (WT) are probed with ConA pre-incubated with mannose. Flow cytometry plots reprinted with permission from Williams et al. Copyright 2020 Royal Society of Chemistry.