Structural and Quantitative Characterization of Mucin-Type O-Glycans and the Identification of O-Glycosylation Sites in Bovine Submaxillary Mucin.

Bovine submaxillary mucin (BSM) is a gel-forming glycoprotein polymer, and Ser/Thr-linked glycans (O-glycans) are important in regulating BSM’s viscoelasticity and polymerization. However, details of O-glycosylation have not been reported. This study investigates the structural and quantitative characteristics of O-glycans and identifies O-glycosylation sites in BSM using liquid chromatography–tandem mass spectrometry. The O-glycans (consisting of di- to octa-saccharides) and their quantities (%) relative to total O-glycans (100%; 1.1 pmol per 1 μg of BSM) were identified with 14 major (>1.0%), 12 minor (0.1%–1.0%), and eight trace (<0.1%) O-glycans, which were characterized based on their constituents (sialylation (14 O-glycans; 81.9%, sum of relative quantities of each glycan), non-sialylation (20; 18.1%), fucosylation (20; 17.5%), and terminal-galactosylation (6; 3.6%)) and six core structure types [Gal-GalNAc, Gal-(GlcNAc)GalNAc, GlcNAc-GalNAc, GlcNAc-(GlcNAc)GalNAc, and GalNAc-GalNAc]. O-glycosylation sites were identified using O-glycopeptides (bold underlined; 56SGETRTSVI, 259SHSSSGRSRTI, 272GSPSSVSSAEQI, 307RPSYGAL, 625QTLGPL, 728TMTTRTSVVV, and 1080RPEDNTAVA) obtained from proteolytic BSM; these sites are in the four domains of BSM. The gel-forming mucins share common domain structures and glycosylation patterns; these results could provide useful information on mucin-type O-glycans. This is the first study to characterize O-glycans and identify O-glycosylation sites in BSM.


Introduction
Mucins are macromolecular (molecular mass, 0.4-4 MDa) glycoprotein constituents of mucus [1], and are highly modified with Ser/Thr-linked glycans (O-glycans) [2]. Mucins play a role in the protection of the host infection [2,3] and are involved in various functions, including gel-formation, viscoelasticity, hydration, and lubrication [3], and O-glycosylation is important in the regulation of these properties.
The biosynthesis of mucin-type O-glycosylation is initiated by N-acetylgalactosamine (GalNAc) attached to the hydroxyl moiety of Ser or Thr, and it is usually extended to form one of several common core structures [4]. These mucin-type O-glycans consist of GalNAc, N-acetylglucosamine (GlcNAc),  Figure 2C and m/z = 441.2684 in Figure 2D) oxonium ions.  Figure 2C and m/z = 441.2684 in Figure 2D) oxonium ions. The informative fragmentation ions could not be clearly discerned in AB-labeled O-glycans, rendering their structure difficult to elucidate (Figure 2A These results indicate that the EIC of ProA-labeled O-glycans were more efficiently ionized than those of AB-labeled O-glycans, and this can be explained by the high proton affinity of the ProA basic tail [2-(diethylamino)ethyl group], which contributes to ESI-MS/MS sensitivity in positive mode [23]. All 34 O-glycan structures and corresponding peak numbers in the UPLC chromatogram and their mass data are summarized in Table 1. The informative fragmentation ions could not be clearly discerned in AB-labeled O-glycans, rendering their structure difficult to elucidate (Figure 2A These results indicate that the EIC of ProA-labeled O-glycans were more efficiently ionized than those of AB-labeled O-glycans, and this can be explained by the high proton affinity of the ProA basic tail [2-(diethylamino)ethyl group], which contributes to ESI-MS/MS sensitivity in positive mode [23]. All 34 O-glycan structures and corresponding peak numbers in the UPLC chromatogram and their mass data are summarized in Table 1.       a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1  a Peak numbers correspond to peak numbers in Figure 1

Quantitative Characterization of O-Glycans
The quantity (%) of each glycan relative to total O-glycans (as 100%) was determined from the EIC areas generated by LC-ESI-HCD-MS/MS (Table 1). All O-glycans showed a high mass accuracy of <10.0 ppm. The singly charged precursor ions were obtained from AB-labeled O-glycans, and ProA-labeled O-glycans produced singly or doubly charged precursor ions, enabling the detection of trace O-glycans.
Fluorescent labeling can directly quantify O-glycans based on their intensities due to the stoichiometric attachment of one fluorescent label per O-glycan [25]. The fluorescence intensity of the most distinct O-glycan peak, G6, was selected from the UPLC chromatogram ( Figure 1A) due to the multiple glycans that can exist in one peak, and calculated as 0.35 pmol using a linear calibration curve (r 2 = 0.99) generated from AB concentrations (0.09-1.56 pmol) (data not shown). The others were calculated using their relative quantities (Table 1)

O-Glycopeptide Structures
To identify sites at which O-glycans attach to BSM, glycopeptides obtained from proteinase K digestion of reduced and alkylated BSM were enriched and analyzed by nano-LC-HCD-MS/MS. The 1589 AA sequence of BSM was based on the previous report [12]. Figure 3 Figure 3A,B,D), m/z = 308.0976 ( Figure 3A), m/z = 512.1974 ( Figure 3A-C)) with a mass error of <5.0 ppm. The glycosylation sites included a peptide (Y 0 ) ion and a peptide with a HexNAc (Y 1 ) ion that was observed as the predominant peak at m/z 831.4570 ( Figure 3A) or m/z 1175.5433 ( Figure 3B) in the MS/MS spectra. Additionally, the spectra provided part of a peptide backbone sequence, yielding masses for b 1 (observed m/z = 129.0550) ( Figure 3A) and y 7 (observed m/z = 719.3585) ( Figure 3B).

As shown in
Biomolecules 2020, 10, 636 9 of 15 To identify sites at which O-glycans attach to BSM, glycopeptides obtained from proteinase K digestion of reduced and alkylated BSM were enriched and analyzed by nano-LC-HCD-MS/MS. The 1,589 AA sequence of BSM was based on the previous report [12].

TI G18
a O-glycosylation sites are presented with bold underlined S or T. b All glycopeptide masses wer calculated and observed as doubly charged forms. c Mass error was calculated as [(observed mass − theoretical mass)/theoretical mass] × 10 6 (a high mass accuracy of glycopeptide; <20.0 ppm). d GlcNAc; , GalNAc; , Gal; , Fuc; , Neu5Ac; , Neu5Gc. e Glycan numbers correspond to peak numbers in Figure 1 and Table 1.

O-Glycosylation Sites at Four BSM Domains
Nine types of O-glycopeptides (GP1-GP9) consisting of three peptide backbones and fi glycans were identified, and three peptide backbones (307RPSYGAL, 625QTLGPL   a O-glycosylation sites are presented with bold underlined S or T. b All glycopeptide masses were calculated and observed as doubly charged forms. c Mass error was calculated as [(observed mass − theoretical mass)/theoretical mass] × 10 6 (a high mass accuracy of glycopeptide; <20.0 ppm). d GlcNAc; , GalNAc; , Gal; , Fuc; , Neu5Ac; , Neu5Gc. e Glycan numbers correspond to peak numbers in Figure 1 and Table 1.

O-Glycosylation Sites at Four BSM Domains
Nine types of O-glycopeptides (GP1-GP9) consisting of three peptide backbones and fiv glycans were identified, and three peptide backbones (307RPSYGAL, 625QTLGPL  (for the quantification of O-glycans) and ProA-labeling (for the qualification of O-glycans) analyses were complementarily used in this study. Mucin-type O-glycans contain mono-to more than 20-saccharides [33], but the present study shows that all BSM O-glycans consist of di-to octa-saccharides. Almost all of the O-glycans were sialylated, with O-glycans containing Neu5Ac and Neu5Gc. Sialylation occurred on di-to hexa-saccharides (not hepta-and octa-saccharides). Additionally, mono-and di-fucosylation occurred on tetra-to hexa-saccharides and hexa-to octa-saccharides, respectively, and terminal-galactosylation occurred on tetra-to hexa-saccharides.
The detailed functional role of each O-glycan in mucin is unknown [24], but sialylated, fucosylated, and terminal-galactosylated O-glycans play various roles in the functions of mucin or other glycoproteins. For example, sialylated O-glycans provide an overall negative charge to the molecule, stabilize protein conformation, increase protein thermal stability, act as a protective barrier, alter protein solubility, entrap pathogens, and enhance the viscosity of mucins [34]. Fucosylated O-glycans mediate ligand adhesion, pathogen-host interactions, and cellular processes via signaling mechanisms [35], and terminal-galactosylated O-glycans may provide recognition epitopes for galactose-specific lectins [36].
Core 1 Gal-GalNAc and Core 2 GlcNAc-(Gal)-GalNAc structures were found in BSM, which are similar to most human mucins [34], but Core 3 GlcNAc-GalNAc, Core 4 GlcNAc-(GlcNAc)-GalNAc, and Core 5 GalNAc-GalNAc structures, which are less common in other mucins, were also found in BSM. This suggests that BSM has an additional protective function that other mucins may not because these core structures are important in forming mucin barriers that influence host-environment interactions and disease pathogenesis [37].
The present study also identifies O-glycosylation sites from O-glycopeptides using nano-LC-HCD-MS/MS. Sialylated O-glycans, Core 3 GlcNAc-GalNAc, and Core 5 GalNAc-GalNAc structures are extensively attached at BSM, but fucosylated and terminal-galactosylated O-glycans are partially attached.
Unfortunately, dense O-glycosylations of BSM are resistant to proteolytic digestion, resulting in limited hydrolysis, and the present results reveal one O-glycosylated Ser/Thr in each glycopeptide. Thus, further studies are necessary to analyze glycopeptides containing more than two O-glycosylated Ser/Thr in each glycopeptide. Additionally, LC-MS/MS with electron-transfer dissociation, which can be used to fragment the peptide backbone and identify AA sequences, is also useful for glycopeptide analysis when combined with the present HCD.

Conclusions
The gel-forming mucins, including BSM, share common properties, such as domain structures, glycosylation patterns, and biosynthetic pathways [38], and the present results provide useful information regarding the properties, multifunction, and expansion of the range of their applications. This is the first study to characterize and quantify mucin-type O-glycans and identify O-glycosylation sites within BSM.