A library of chemically defined human N-glycans synthesized from microbial oligosaccharide precursors

Abstract

Synthesis of homogenous glycans in quantitative yields represents a major bottleneck to the production of molecular tools for glycoscience, such as glycan microarrays, affinity resins, and reference standards. Here, we describe a combined biological/enzymatic synthesis that is capable of efficiently converting microbially-derived precursor oligosaccharides into structurally uniform human-type N-glycans. Unlike starting material obtained by chemical synthesis or direct isolation from natural sources, which can be time consuming and costly to generate, our approach involves precursors derived from renewable sources including wild-type Saccharomyces cerevisiae glycoproteins and lipid-linked oligosaccharides from glycoengineered Escherichia coli. Following deglycosylation of these biosynthetic precursors, the resulting microbial oligosaccharides are subjected to a greatly simplified purification scheme followed by structural remodeling using commercially available and recombinantly produced glycosyltransferases including key N-acetylglucosaminyltransferases (e.g., GnTI, GnTII, and GnTIV) involved in early remodeling of glycans in the mammalian glycosylation pathway. Using this approach, preparative quantities of hybrid and complex-type N-glycans including asymmetric multi-antennary structures were generated and subsequently used to develop a glycan microarray for high-throughput, fluorescence-based screening of glycan-binding proteins. Taken together, these results confirm our combined synthesis strategy as a new, user-friendly route for supplying chemically defined human glycans simply by combining biosynthetically-derived precursors with enzymatic remodeling.

Figure 1

Glycan synthesis strategies. Precursor glycans (1–3) were derived from (a) yeast invertase or (b) lipid-linked oligosaccharides (LLOs) from glycoengineered E. coli cells carrying plasmid pYCG, and subsequently used to synthesize glycans 4–21. Enzymatic steps: (i) PNGase F (product shown is representative of high mannose yeast glycans); (ii) α1,2- and α1,6-mannosidase; (iii) jack bean α-mannosidase and α1,6-mannosidase; (iv) GnTI; (v) β1,4-galactosyltransferase; (vi) GnTIV; (vii) non-enzymatic hydrolysis of extracted LLOs; (viii) GnTII; (ix) GnTV; and (x) GnTIII.

Figure 2

Biosynthesis of precursor oligosaccharides for enzymatic remodeling. MALDI-TOF MS analysis (top panels) and 600-MHz ¹H NMR characterization (bottom panels) corresponding to: (a) Man₅GlcNAc₂ glycan synthesized by enzymatic deglycosylation of S. cerevisiae oligosaccharides; and (b) the Man₃GlcNAc₂ glycan synthesized by glycoengineered E. coli cells.

Figure 3

Generation of hybrid glycans using the yeast-derived precursor. MALDI-TOF MS analysis (top panels) and 600-MHz ¹H NMR characterization (bottom panels) for the following: (a) glycan 6; (b) glycan 7; (c) glycan 8; and (d) glycan 9.

Figure 4

Generation of hybrid and complex-type glycans using the bacteria-derived precursor. MALDI-TOF MS analysis of the following products: (a) glycan 11; (b) glycan 14; (c) glycan 12; (d) glycan 17; (e) glycan 13; (f) glycan 20; (g) glycan 18; (h) glycan 21; (i) glycan 10; (j) glycan 19; (k) glycan 15; (l) glycan 16.

Figure 5

Binding of lectin ConA to microarray of synthesized human N-glycans. (a) Structures of reference compounds that were derivatized with AEAB and immobilized onto NHS-activated glass slides, alongside glycans 2–21. (b) Probing of immobilized glycans with biotinylated ConA (10 μg/ml). The amount of bound ConA was determined by streptavidin-Cy5 (5 μg/ml) fluorescence. Background subtracted mean fluorescence values are shown. Error bars represent the standard deviation of the mean. Representative structures for glycans 3, 12, 15 and 19 are shown. All glycans were printed at 100 μM in PBS in replicates of four. PBS and streptavidin-Cy5 spots served as controls. Biotin-LC-hydrazide spots printed on each subarray serve as an alignment feature to localize each subarray on the slide.