N-Glycan structure annotation of glycopeptides using a linearized glycan structure database (GlyDB)

Abstract

While glycoproteins are abundant in nature, and changes in glycosylation occur in cancer and other diseases, glycoprotein characterization remains a challenge due to the structural complexity of the biopolymers. This paper presents a general strategy, termed GlyDB, for glycan structure annotation of N-linked glycopeptides from tandem mass spectra in the LC-MS analysis of proteolytic digests of glycoproteins. The GlyDB approach takes advantage of low-energy collision-induced dissociation of N-linked glycopeptides that preferentially cleaves the glycosidic bonds while the peptide backbone remains intact. A theoretical glycan structure database derived from biosynthetic rules for N-linked glycans was constructed employing a novel representation of branched glycan structures consisting of multiple linear sequences. The commonly used peptide identification program, Sequest, could then be utilized to assign experimental tandem mass spectra to individual glycoforms. Analysis of synthetic glycopeptides and well-characterized glycoproteins demonstrate that the GlyDB approach can be a useful tool for annotation of glycan structures and for selection of a limited number of potential glycan structure candidates for targeted validation.

Figure 1

GlyDB linear representation of a bianntenary glycan with a single sialic acid. (A) Conversion of standard glycan symbols to letters. Numbering of glycosidic bonds corresponds to the fragments listed in Table 2. (B) Examples of two potential linearizations of the structure in (A), dashed lines illustrate the process of linearization. Three potential fragmentation schemes and (C) the corresponding B- and Y- ions represented in linearized form. (D) Representation of glycopeptide in a fasta format suitable for database searching using conventional peptide database search programs. All four linear sequences representing the same glycan structure were concatenated into a single sequence. Symbols: □ N – HexNAc (GlcNAc, GalNac), ● H – Hex (Gal, Glc, Man), ▲ F – deoxyHex (fuc), ◇ S - Sialic acid (Neu5Ac), and K - peptide. Note: Slashes in the linear sequences were added to visualize individual branches; however, these symbols are not used in the actual database.

Figure 1

GlyDB linear representation of a bianntenary glycan with a single sialic acid. (A) Conversion of standard glycan symbols to letters. Numbering of glycosidic bonds corresponds to the fragments listed in Table 2. (B) Examples of two potential linearizations of the structure in (A), dashed lines illustrate the process of linearization. Three potential fragmentation schemes and (C) the corresponding B- and Y- ions represented in linearized form. (D) Representation of glycopeptide in a fasta format suitable for database searching using conventional peptide database search programs. All four linear sequences representing the same glycan structure were concatenated into a single sequence. Symbols: □ N – HexNAc (GlcNAc, GalNac), ● H – Hex (Gal, Glc, Man), ▲ F – deoxyHex (fuc), ◇ S - Sialic acid (Neu5Ac), and K - peptide. Note: Slashes in the linear sequences were added to visualize individual branches; however, these symbols are not used in the actual database.

Figure 2

Analysis of lectin from E. cristagalli. Automatically annotated MS/MS spectrum of a glycopeptide with a known glycan structure. ★ - xylose, other symbols as in Fig. 1.

Figure 3

Analysis of a synthetic glycopeptide with known glycan structure. (A) Correct glycan structure and hypothetical structures with the same glycan composition. See text for more details and Table 4 for the database search results. (B) Example of an annotated MS/MS spectrum acquired using direct infusion of a synthetic glycopeptide. Circled structures show confirmation for bisecting HexNAc. The blue letters (a – d) indicate what linear structure can be used to generate a particular fragment ion. Numbers with the plus sign show the charge state. Glycan symbols as in Fig. 1.

Figure 3

Analysis of a synthetic glycopeptide with known glycan structure. (A) Correct glycan structure and hypothetical structures with the same glycan composition. See text for more details and Table 4 for the database search results. (B) Example of an annotated MS/MS spectrum acquired using direct infusion of a synthetic glycopeptide. Circled structures show confirmation for bisecting HexNAc. The blue letters (a – d) indicate what linear structure can be used to generate a particular fragment ion. Numbers with the plus sign show the charge state. Glycan symbols as in Fig. 1.

Figure 4

Analysis of bovine fetuin. (A) Glycan structures identified in LC-MS analysis of a fetuin tryptic digest. The star indicates structures reported in ref . (B) Example of an MS/MS spectrum of one glycoform of bovine fetuin automatically identified by GlyDB approach with annotation of individual fragments. Symbols as in Fig. 1.

Figure 4

Analysis of bovine fetuin. (A) Glycan structures identified in LC-MS analysis of a fetuin tryptic digest. The star indicates structures reported in ref . (B) Example of an MS/MS spectrum of one glycoform of bovine fetuin automatically identified by GlyDB approach with annotation of individual fragments. Symbols as in Fig. 1.