Pregnancy-Associated Changes of IgG and Serum N-Glycosylation in

Jul 18, 2016 - The dromedary camel (Camelus dromedarius) is an agriculturally important species of high economic value but of low reproductive efficie...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/jpr

Pregnancy-Associated Changes of IgG and Serum N‑Glycosylation in Camel (Camelus dromedarius) Barbara Adamczyk,*,†,‡,# Simone Albrecht,†,# Henning Stöckmann,† Ibrahim M. Ghoneim,§ Marzook Al-Eknah,§ Khalid A. S. Al-Busadah,∥ Niclas G. Karlsson,‡ Stephen D. Carrington,⊥ and Pauline M. Rudd† †

GlycoScience Group, NIBRT−The National Institute for Bioprocessing Research and Training, Fosters Avenue, Mount Merrion, Blackrock, Co. Dublin, Ireland ‡ Department of Medical Biochemistry and Cell Biology, Institute of Biomedicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg, 40530, Sweden § Department of Clinical Studies, Faculty of Veterinary Medicine, and ∥Department of Physiology and Pharmacology, Faculty of Veterinary Medicine, King Faisal University, Al-Ihssa, 31982, Saudi Arabia ⊥ School of Agriculture and Food Science, University College Dublin, Belfield, Dublin 4, Ireland ABSTRACT: The dromedary camel (Camelus dromedarius) is an agriculturally important species of high economic value but of low reproductive efficiency. Serum and IgG N-glycosylation are affected by physiological and pathogenic changes and might therefore be a useful diagnostic tool in camel livestock management. This study presents the first comprehensive annotation of the N-glycome from dromedary camel serum as well as their single-domain and conventional antibodies and its subsequent application for camel pregnancy diagnostics. N-glycans were released by PNGaseF, labeled with 2-aminobenzamide (2-AB), and analyzed by hydrophilic interaction liquid chromatography with fluorescent detection (HILIC−UPLC-FLD), enzymatic sequencing and mass spectrometry (MS). The use of a high-throughput robotic platform for sample preparation allowed the rapid generation of glycomics data from pregnant (n = 8) and nonpregnant (n = 8) camels of the Majaheem and Wadha breed. IgG N-glycans dominate the glycan profile of camel serum and present a mixture of core-fucosylated and noncore-fucosylated N-glycans which can contain N-glycolylneuraminic and N-acetylneuraminic acid. Significant pregnancyassociated but breed-independent increases in galactosylation, core-fucosylation, sialylation, and decreases in serum O-acetylation were observed. The monitoring of IgG and serum N-glycosylation presents an attractive complementary test for camel pregnancy diagnostics and presents an interesting tool for biomarker discovery in camel health and breeding. KEYWORDS: dromedary camel, pregnancy, N-glycome, serum, IgG, HILIC−UPLC-FLD−MS

■

INTRODUCTION Serum is an attractive source for biomarker discovery as its collection is straightforward, cost-effective, and minimally invasive.1,2 IgG is the most abundant glycoprotein in serum which contains a single N-glycosylation site in the constant region of each of its heavy chains. Many studies have demonstrated that alterations in serum glycoprotein and IgG N-glycosylation result from disease-related and environmental factors such as immunodeficiency diseases, cancer, age, gender, smoking, and body mass index3,4 but can also be attributed to cytokine and hormonal levels.5−7 Detailed structural libraries of IgG and serum N-glycosylation were therefore established for many species including humans and domestic animals for studying physiological and pathogenic glycan changes.8−13 However, no such library has yet been established for the dromedary camel (Camelus dromedarius), which is of high economic value and has been used for milk and © 2016 American Chemical Society

meat production as well as riding and leisure use in desert regions for many centuries. A main issue in camel breeding is their poor reproductive efficiency. This is mainly due to the fact that camels are induced ovulators with a short breeding season, a long gestation period (13 months), a delayed onset of puberty and a high incidence of early embryonic mortality.14 The diagnosis of pregnancy at an early stage is thus highly desirable for the efficient management of camel reproductive performance. Methods for camel pregnancy diagnosis include ultrasonography, visual inspection (i.e., cocking of the tail), recto-genital palpation as well as plasma- and urinary-based tests, such as progesterone and estrogen levels or chemical tests.15−17 Pregnancy-associated glycosylation changes of human IgG have been observed, such as increased galactosylation and Received: May 13, 2016 Published: July 18, 2016 3255

DOI: 10.1021/acs.jproteome.6b00439 J. Proteome Res. 2016, 15, 3255−3265

Article

Journal of Proteome Research sialylation and a decrease in IgG core-fucosylation.18−21 As, to date, methods for the detection of camel pregnancy lack accuracy; the discovery of pregnancy-associated glycan changes in serum and IgG could be a valuable asset for camel pregnancy diagnosis. Antibodies as present in the sera of humans and domestic animals are composed of two heavy and two light chains. Camelids are exceptional in this respect as they additionally produce very specific and unusual single-domain antibodies (VHHs).22 VHHs are of special biomedical interest due to their beneficial biochemical and economic properties. They were first described by Hamers et al. in 1993, who isolated three immunoglobulin G (IgG) subclasses from the serum of dromedary camel.23 For only one of these subclasses (IgG1), heavy (MWVH = 50 kDa) and light chains (MWVL = 30 kDa) were obtained after reduction of the glycoprotein, whereas for the other two subclasses, which constitute 75% of the total IgG pool, only heavy chains (IgG2, MWVH = 46 kDa; IgG3, MWVH = 43 kDa) were obtained.23 Since the discovery of single-domain antibodies, the field of recombinant antibody technology has rapidly progressed and multiple biotechnological applications of VHHs have emerged.24,25 Their high stability and solubility, increased target binding affinity, and substrate accessibility and the ability to express them in bacterial systems make them well suited as a research tool.25,26 The unique properties of nanobodies can facilitate their further development into a new generation of antibody-based therapeutics.27,28 Understanding protein glycosylation of recombinant therapeutics is crucial to ensure their safety and efficacy.29,30 Studies investigating protein glycosylation of camel antibodies can play an important role in defining the potential and utility of nanbodies as therapeutic agents. In this study we therefore performed a comprehensive annotation of the N-glycome from the different camel IgG subclasses and serum, and evaluated its usefulness in camel pregnancy diagnostics. Sera from nonpregnant as well as pregnant animals of the Majaheem and Wadha breed were studied. This allowed us to assess pregnancy-related glycan changes and additionally address interspecies variability. Methods for the characterization of N-glycan structures remain challenging due to the vast number of structural possibilities that present considerable overall complexity. Our approach is based on a well-established and robust chromatography-based workflow for the analysis of 2-AB labeled glycans,10 including HILIC-UPLC-FLD along with exoglycosidase digestions, DMB sialic acid speciation and structural confirmation by HILIC-UPLC−MS. In addition, the use of an automated sample preparation workflow implemented on a liquid handling platform31,32 enabled the rapid generation of glycomics data from our animal cohorts.

■

Majaheem breed and three belonged to the white Wadha breed. Serum was stored at −80 °C. Chemicals and Reagents

Water used throughout this study was obtained from a Milli-Q Gradient A10 Elix system (Millipore, Bedford, MA) and was 18.2 MΩ or greater with a total organic carbon (TOC) content less than 5 parts per billion (ppb). Acetonitrile (Fisher Far UV gradient grade), solid-supported hydrazide as well as Protein A and Protein G affinity purification plates were obtained from Fisher Scientific (Dublin, Ireland). Solvinert filter plates were from Merck-Millipore (Darmstadt, Germany) and Acroprep Advance filter plates from Pall (NY, USA). All other chemicals used were purchased from Sigma-Aldrich (Dublin, Ireland) and were of the highest available quality. IgG Affinity Purification

IgG purification was carried out on the Hamilton Microlab STAR robotic liquid handler as described previously.31 Briefly, a protein A spin plate (Thermo Scientific) was preconditioned by washing three times with 500 μL of binding buffer (PBS) at room temperature. Camel serum samples (50 μL) were diluted with equal volume of binding buffer and applied to the Protein A plate. The plate assembly was shaken and incubated for 30 min for IgG binding. After incubation the plate was washed five times with washing buffer (500 μL, 0.1 M sodium phosphate, 0.15 M sodium chloride, 1% Triton-X, pH 7.4) to remove nonspecific proteins. This was followed by additional washing with binding buffer (500 μL per well). Next, IgG was eluted from the Protein A plate using 3 × 200 μL elution buffer (0.2 M glycinehydrochloride, pH 2.5). Eluates were collected in a 96 deep well plate and immediately neutralized to pH 7.0 with neutralization buffer (1 M Tris-hydrochloride, pH 9.0) to maintain the IgG stability. The purity of the isolated IgG was assessed using 10% reducing SDS-PAGE in an Xcell SureLock Mini-Cell (Invitrogen, Carlsbad, CA) according to the manufacturer. Precision Plus Protein All Blue Standard (BioRad, Hercules, CA) was used as the molecular weight marker. The gel was run at 50 mA for 45 min and stained with Gel Code Blue staining reagent (Pierce, Rockford, IL). IgG and Serum Glycoproteins Preparation

Denaturation, alkylation, glycan release, hydrazide-mediated glycan clean up, labeling and glycan solid phase extraction were performed on the Hamilton Microlab STAR robotic liquid handler as described previously31 with some modifications for serum samples.32 HILIC−UPLC−Fluorescence N-Glycan Profiling

The 2-AB labeled camel IgG and serum N-glycans were separated by ultraperformance liquid chromatography with fluorescence detection using a Waters BEH Glycan column (150 × 2.1 mm, 1.7 μm BEH particles), which was installed on a Waters H-Class instrument, consisting of a quaternary solvent manager, sample manager, and fluorescence detector. Solvent A was composed of a 50 mM ammonium formate buffer pH 4.4 and solvent B was acetonitrile. A 20 min high-throughput method was used for IgG separation (a linear gradient of 70−53% solvent B over 16.5 min at a flow rate of 0.56 mL/min) and a 30 min method for serum separation with a linear gradient of 70−53% solvent B at 0.56 mL/min in 24.81 min. The injection volume was 10 μL, and the sample was prepared in 70% v/v acetonitrile. Samples were kept at 5 °C prior to injection and the separation temperature was 40 °C. The fluorescence detection wavelengths were λex = 330 nm and λem = 420 nm with a data collection rate of

MATERIALS AND METHODS

Serum Samples

Serum samples from dromedary camels (Camelus dromedarius) were provided by the Camel Research Centre (King Faisal University, Saudi Arabia) and were derived from whole blood collected from the jugular veins of healthy female 7−12 years old camels from farms in Al-Ahasa Oasis (Saudi Arabia). Of the 16 serum samples provided, eight samples were from pregnant animals (gestation time, 2−4 months) and eight samples were from nonpregnant animals. For both, pregnant and nonpregnant animals, five of the respective eight animals belonged to the black 3256

DOI: 10.1021/acs.jproteome.6b00439 J. Proteome Res. 2016, 15, 3255−3265

Article

Journal of Proteome Research

dark. The reaction was quenched by adding 470 μL of water. The labeled samples were analyzed using reversed phase (RP)-UPLC on a Waters BEH C18 column (2.1 × 50 mm, 1.7 μm particle size) with fluorescence detection (λex, 343 nm; λem, 448 nm). Isocratic elution was performed with methanol/ACN/water (7:9:84) for 3.5 min.

20 Hz. A dextran hydrolysate ladder was used to convert retention times into glucose unit (GU) values. Data were processed by Waters Empower 3 chromatography workstation software. Exoglycosidase Digestions

All enzymes were purchased from Prozyme, San Leandro, CA, USA. The 2-AB labeled glycans were digested in 10 μL of 50 mM sodium acetate buffer, pH 5.5 for 18 h at 37 °C, using arrays of the following enzymes: ABS, Arthrobacter ureafaciens sialidase (EC 3.2.1.18, releases α2−3,6,8 linked nonreducing terminal sialic acid), 1 U/mL; BKF, bovine kidney α-fucosidase (EC 3.2.1.51, releases α1−2,6 linked nonreducing terminal fucose residues more efficiently than α1−3,4 linked fucose; digests core α1−6 fucose), 1 U/mL; BTG, bovine testes β-galactosidase (EC 3.2.1.23, hydrolyses nonreducing terminal β1−4 and β1−3 linked galactose), 2 U/mL; and GUH, hexosaminidase cloned from Streptococcus pneumoniae expressed in E. coli (EC 3.2.1.30, releases GlcNAc residues but not a bisecting GlcNAc linked to MAN), 4 U/mL. After incubation, enzymes were removed by filtration using Pall spin filters 10 kDa (Pall Corporation). The N-glycans were then analyzed by HILIC−UPLC.

Calculation of Derived Glycosylation Traits

Derived glycosylation traits were calculated from the percentages of glycan pools (GP1-GP24) and combined the glycans with the same structural characteristics. GP10 and GP17 from IgG were excluded from the calculations as these glycan pools were considered as contaminants from serum. Derived traits for IgG and serum were defined as NeutraIIgG/Serum = SUM(GP1 − GP9 + GP11 + GP12)IgG/Serum G0IgG/Serum = [SUM(GP1 + GP2)IgG/Serum /NeutraIIgG/Serum]100

G0(nonfuc)IgG/Serum = [GP1IgG/Serum /NeutraIIgG/Serum]100 FG0IgG/Serum = [GP2 IgG/Serum /NeutraIIgG/Serum]100

Weak Anion Exchange High Performance Liquid Chromatography (WAX-HPLC)

G1IgG/Serum = [SUM(GP3 + GP4 + GP6 + GP7)IgG/Serum /NeutraIIgG/Serum]100

WAX-HPLC was performed using ProzymeGlycoSep C 7.5 mm × 75 mm column (Prozyme, Leandro,CA) on a 2795 Alliance Separation module with a 2475 fluorescence detector (Waters, Milford, MA). Solvent A was 20% (v/v) acetonitrile in water, and solvent B was 0.1 M acetic acid adjusted to pH 7.0 with ammonia solution in 20% (v/v) acetonitrile. Gradient conditions were as follows: 100% A for 5 min, then a linear gradient of 100% to 0% A for 15 min at a flow rate of 0.75 mL/min, followed by 0% A for 2.5 min, returning to 100% A for 1.5 min and then finishing with 100% A for 7 min. Samples were injected in water. A fetuin Nglycan standard was used for calibration.33

G1(nonfuc)IgG/Serum = [SUM(GP3 + GP4)IgG/Serum /NeutraIIgG/Serum]100 FG1IgG/Serum = [SUM(GP6 + GP7)IgG/Serum /NeutraIIgG/Serum]100 G2 IgG/Serum = [SUM(GP9 + GP11)IgG/Serum /NeutraIIgG/Serum]100

G2(nonfuc)IgG/Serum = [GP9IgG/Serum /NeutraIIgG/Serum]100 FG2 IgG/Serum = [GP11IgG/Serum /NeutraIIgG/Serum]100

UPLC−HILIC-FLD−MS N-glycan characterization

UPLC−HILIC-FLD−mass spectrometry analysis of the 2-AB labeled camel IgG and serum was performed on a Waters Acquity instrument with fluorescence detection connected to a Waters XEVO G2 QTOF instrument. The Waters Acquity system consists of a binary solvent manager, sample manager, and fluorescence detector and was equipped with a Waters BEH Glycan column (150 × 1 mm, 1.7 μm BEH particles). The solvents, settings of the fluorescence detector, injection volume, sample preparation, and sample storage were the same as for UPLC−HILIC-FLD analysis. The column was 60 °C and a linear gradient of 72−57% solvent B over 31 min at a flow rate of 0.15 mL/min was used. Data were acquired in negative mode with the following conditions: 2 kV source capillary voltage, 100 °C source temperature, 450 °C desolvation temperature, 50 L/h cone gas flow, and 800 L/h desolvation gas flow. The collision energy ramp for mass fragmentation was 30−42 V. The mass range for MS was set to m/z 400−2000 and to m/z 100−2000 for MS/MS. Data were processed by Waters MassLynx software V4.1.

BG1IgG/Serum = [SUM(GP5 + GP8)IgG/Serum /NeutraIIgG/Serum]100

BG2 IgG/Serum = [GP12 IgG/Serum /NeutraIIgG/Serum]100 FucIgG/Serum = [SUM(GP2 + GP6 − 8 + GP11 + GP12)IgG/Serum /NeutraIIgG/Serum]100 FG0/G0IgG/Serum = [GP2 IgG/Serum /G0IgG/Serum]100 FG1/G1IgG/Serum = [SUM(GP6 + GP7)IgG/Serum /G1IgG/Serum ]100

FG2/G2 IgG/Serum = [GP11IgG/Serum /G2 IgG/Serum]100 FBG1/G1IgG/Serum = [GP8IgG/Serum /BG1IgG/Serum ]100

Sialyl IgG = SUM(GP13 − GP24)IgG

SialylSerum = SUM(GP10 + GP13 − GP24)Serum

Sialic Acid Speciation

Monosialyl IgG/Serum = [SUM(GP13 − GP16 + GP18

Sialic acids were released from the samples (approx 200 μg) by mild acid hydrolysis (2 M acetic acid, 80 °C, 2 h). Twenty μL of DMB (1,2-diamino-4,5-methylenedioxybenzene) labeling solution (DMB in mercaptoethanol and sodium dithionite, Ludger, Abingdon, U.K.) was added to a 10 μL aliquot of the release solution, and the mixture was incubated for 3 h at 50 °C in the

− GP19)IgG/Serum /Sialyl IgG/Serum]100

Disialyl IgG/Serum = [SUM(GP20 − GP24)IgG/Serum /Sialyl IgG/Serum]100 3257

DOI: 10.1021/acs.jproteome.6b00439 J. Proteome Res. 2016, 15, 3255−3265

Article

Journal of Proteome Research

Figure 1. (A) SDS-PAGE of purified camel IgG with three bands representing IgG1, IgG2, and IgG3. (B) HILIC-FLD−UPLC profiles of individual camel IgG subclasses.

chromatography (UPLC) using a C18 column in a fifth dimension.

AcetylSerum = [SUM(GP10 + GP13a + GP14 + GP16 + GP17 + GP24)Serum /SialylSerum]100

2. Isolation of Antibody Subclasses from Camel Serum

Affinity chromatography on Protein A was used to isolate IgG from camel serum. The workflow was fully automatized using a robotic liquid handler as described previously.31 Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) of the IgG extracts resulted in three strong bands at 50 kDa, 46 kDa, and 43 kDa and two weak bands at 30 kDa and 80 kDa (Figure 1). According to Hamers et al.23 the strong bands represent the heavy chains of the three camel IgG isoforms (IgG1 = 50 kDa, IgG2 = 46 kDa and IgG3 = 43 kDa). The weak band at 30 kDa corresponds to the light chain which belongs to IgG1, whereas IgG2 and IgG3 are single-chain antibodies and therefore do not contain light chains. The band at 80 kDa was not further studied as most likely it represents IgG that was not fully reduced. We tested both, Protein A and Protein G resin to purify camel IgG, and our data were in agreement with previously described observations that Protein G does not efficiently purify IgG2.23 Therefore, Protein A was used for the efficient purification of the three IgG fractions from camel serum. These results point to the different binding properties of IgGs from camel and other species, such as human. All four human IgG subclasses bind efficiently to Protein G, but only three bind to Protein A. To compare the N-glycosylation of the different IgG subclasses the corresponding gel bands were excised and treated with recombinant peptide-N-glycosidase-F (PNGaseF). The released glycans were subsequently labeled with the fluorescent 2-AB and analyzed by HILIC−UPLC-FLD. N-glycan profiles for all three IgG fractions yielded very similar chromatograms (Figure 1B) with only minor quantitative differences observed. Thus, the mechanism for the glycosylation of single-chain and conventional antibodies in the camel ER and Golgi appear to be similar. Because of the qualitative similarity of the glycan profiles, the total camel IgG was used for the characterization of the IgG N-glycan structures.

G1S/G1Total IgG/Serum = [SUM(GP13 + GP14 + GP16)IgG/Serum /SUM(GP13 + GP14 + GP16 + GP4 + GP7)IgG/Serum ]100 G2S/G2Total IgG/Serum = [SUM(GP15 + GP18 + GP19 + GP21 − GP24)IgG/Serum /SUM(GP15 + GP18 + GP19 + GP21 − GP24 + GP9 + GP11)IgG/Serum ]100

Statistical Data Analysis

Two-way ANOVA with a cutoff p-value of 0.05 was performed for the glycan pools and derived glycosylation traits using R software (http://www.r-project.org/). Heat maps were as well plotted in R and included hierarchical sample clustering according to Ward’s minimum variance method.

■

RESULTS AND DISCUSSION

1. Overall Approach for the Detailed Characterization of the Camel N-glycome

Our strategy was inspired by a recent report showing different analytical workflows and levels of information obtained for the structural characterization of N-glycans.34 Aiming at precisely quantitative data and detailed structural information, we applied orthogonal techniques that allowed us to investigate five different dimensions. Weak anion exchange (WAX) enabled the first dimension separation of 2-AB fluorescently labeled glycans into pools of differently charged glycans. Complementary hydrophilic interaction liquid chromatography (HILIC) separation of each pool served as second dimension. This was followed by exoglycosidase digestions resulting in third-dimension information on the glycan sequence and sugar linkages. As a fourth dimension, electrospray ionization (ESI) mass spectrometry (MS) and fragmentation was used to confirm the glycan structures. In addition, to determine the proportions of Nglycolylneuraminic acid (Neu5Gc), N-acetylneuraminic acid (Neu5Ac) and their O-acetylated derivatives, sialic acids, were released from the N-glycans by acid hydrolysis, labeled with 1,2diamino-4,5-methylenedioxybenzene (DMB) and subsequently identified by reverse-phase (RP) ultraperformance liquid

3. Characterization of Camel IgG N-Glycans

To investigate and fully characterize the N-glycome of the dromedary camel species, we first focused on the IgG N-glycans. Total camel IgG from pregnant (P) and nonpregnant (NP) camels (n = 8 healthy female animals, respectively; n = 5 belonging to the Mejaheem breed and n = 3 belonging to the Wadheh breed, respectively) were automatically processed on our robotic platform in order to release and fluorescently label N3258

DOI: 10.1021/acs.jproteome.6b00439 J. Proteome Res. 2016, 15, 3255−3265

Article

Journal of Proteome Research glycans.31 A representative HILIC−UPLC profile of camel IgG N-glycans is shown in Figure 2. The chromatographic profiles

and then verified by a combination of exoglycosidase digestions and mass spectrometry. Exoglycosidase digestions are generally used in combination with HILIC−UPLC to identify structures with terminal sugars against which the exoglycosidase enzyme is active. The removal of the terminal residue results in a shift in GU value. The shift depends on the type of residue being hydrolyzed as well as the number of sugars removed. For example, BTG (bovine testes βgalactosidase) releases β(1−3/4) nonreducing terminal galactose residues. The removal of one galactose is indicated by a shift in 0.8 GU, whereas a shift of 1.6 GU indicates the loss of 2 galactose residues. The digestion with BKF (bovine kidney αfucosidase) which releases α(1−2/3/4/6) linked fucose enabled us to parse out key differences in core-fucosylation between camel and human IgG glycosylation. In contrast to humans, where the majority (∼96%) of the IgG N-glycome is corefucosylated,8 camels contain a more equal proportion of corefucosylated to nonfucosylated structures. The percentage of core-fucosylated to noncore fucosylated G0 (i.e., no outer arm galactosylation), G1 (i.e., one outer arm galactosylated), and G2 (i.e., two outer arms galactosylated) pools ranged from 43% to 56% for our cohorts (Table 1). Similar results were observed for feline IgG in a previous study.13 The amount of bisected Nglycans for camel IgG is negligible which is different to human IgG but similar to equine IgG. Another structural feature characteristic for IgG N-glycans from animals is the presence of five alpha-galactosylated structures in the camel IgG N-glycome, namely FA2Gal1, FA2G1Gal1, A2G2Gc1[6]Gal, FA2G2S[6]1Gal, and FA2G2Gc[6]1Gal. For our in-depth structural characterization, a weak anion exchange separation of IgG into pools of differently charged glycans was performed. This approach was complemented by HILIC−UPLC-FLD separation of each pool and further digestions with an α(2−3/6)-specific sialidases from Arthrobacter urefaciens (ABS) and an α(2,3)-specific sialidases from Streptococcus pneumoniae recombinant in E. coli (NAN1). This resulted in the annotation of sialic acid linkages. The majority of the sialic acids on IgG were α(2−6)-linked and a small proportion of disialylated structures contained a combination of α(2−3) and α(2−6)-linked sialic acids. Our results did not indicate any monosialylated structures containing exclusively α(2−3)-linked sialic acids. One of the main limitations of the HILIC−UPLC-FLD technique is its inability to differentiate between the various species of sialic acids. Therefore, we applied negative-mode LC− HILIC−MS to distinguish Neu5Ac and Neu5Gc on the basis of their mass difference. The overall proportion of Neu5Gc and Neu5Ac was determined by labeling of the hydrolytically released sialic acids with DMB and their subsequent identification by comparison to a reference panel when analyzed on a C18 column. Our data indicate that approximately 76% of the sialic acid content from camel IgG is Neu5Gc and 24% is Neu5Ac. Comparison with human IgG and IgG from different animal species shows a variable distribution of sialic acids. Human and chicken IgG N-glycans contain only Neu5Ac, whereas IgG from rhesus monkey, cow, sheep, goat, horse, and mouse contains glycans with exclusively Neu5Gc.9 Similar to camels, IgG from the dog, guinea pig, rat, and rabbit contain both Neu5Gc and Neu5Ac.9 In total, we identified 35 N-glycan structures on camel IgG, which are distributed over 24 glycan peak pools. Table 1 lists all glycans assigned for camel IgG, along with their GU, their observed and theoretical masses ([M-H]2−) as well as their

Figure 2. HILIC-FLD−UPLC chromatograms of N-linked glycans released from (A) camel IgG and (B) camel serum of a nonpregnant camel. Glycans are represented using the Oxford symbol nomenclature.57 Symbols encode the following monosaccharide structures: GlcNAc, filled square; mannose, open circle; galactose, open diamond; Neu5Ac, filled star; Neu5Gc, open star; fucose, diamond with a dot inside; beta linkage, solid line; alpha linkage, dotted line; unknown beta linkage (could not be confidently assigned), solid wave; unknown alpha linkage (could not be confidently assigned), dotted wave.

from all 16 camels were qualitatively consistent, representing the same glycan structures. Peak retention times were converted into glucose units (GU) values by comparison with a dextran hydrolysate ladder. Initial separation of the undigested sample yielded 24 chromatographic peaks with GU values between 5.38 and 10.86 in a 20 min separation run. Neutral smaller structures eluted first, and the large sialylated structures eluted later. The structures were preliminarily assigned based on their GU values according to the publicly available glycan library GlycoBase35,36 3259

DOI: 10.1021/acs.jproteome.6b00439 J. Proteome Res. 2016, 15, 3255−3265

Article

Journal of Proteome Research Table 1. Summary of N-Glycan Structural Composition Determined for Camel IgGa glycan peak

GU

Neutral Structures GP1 5.38 GP2 5.83 GP3 6.21 GP4 6.31 GP5 6.39 GP6 6.62 GP7 6.74 GP8 GP9 GP11

6.98 7.13 7.52

GP12 7.74 Sialylated Structures GP10‡ 7.40 GP13 GP14 GP15 GP16 GP17* GP18

7.95 8.16 8.32 8.40 8.60 8.80

GP19 GP20

9.19 9.62

GP21

9.99

GP22

10.38

GP23 GP24

10.44 10.86

m/z observed

m/z theor

717.29 790.33 798.32 798.32 900.35 871.35 871.35 871.34 972.88 879.35 952.38 952.38 1053.92

717.27 790.30 798.30 798.30 899.84 871.33 871.33 871.33 972.86 879.32 952.35 952.35 1053.89

HexNAc4Hex3 dHex1HexNAc4Hex3 HexNAc4Hex4 HexNAc4Hex4 HexNAc5Hex4 dHex1HexNAc4Hex4 dHex1HexNAc4Hex4 dHex1HexNAc4Hex4 dHex1HexNAc5Hex4 HexNAc4Hex5 dHex1HexNAc4Hex5 dHex1HexNAc4Hex5 dHex1HexNAc5Hex5

1045.90 1212.46 1016.90 951.87 1024.90 1024.90 * 1097.93 1032.90 1105.94 1113.94 1170.46 1170.46 1178.97 1186.96 1243.50 1243.50 1178.47 1178.47 1251.50 1251.50 1186.45 1259.48

1045.88 1212.43 1016.87 951.84 1024.87 1024.87 * 1097.90 1032.87 1105.90 1113.89 1170.42 1170.42 1178.93 1186.92 1243.45 1243.45 1178.42 1178.42 1251.45 1251.45 1186.41 1259.44

HexNAc4Neu5Ac1Hex5Ac1 HexNAc4Neu5Ac2Hex5Ac2 dHex1HexNAc4Neu5Ac1Hex4 HexNAc4Neu5Gc1Hex4 HexNAc4Neu5Ac1Hex5 dHex1HexNAc4Neu5Gc1Hex4 * dHex1HexNAc4Neu5Ac1Hex5 HexNAc4Neu5Gc1Hex5 dHex1HexNAc4Neu5Gc1Hex5 HexNAc4Neu5Gc1Hex6 HexNAc4Neu5Ac2Hex5 HexNAc4Neu5Ac2Hex5 dHex1HexNAc4Neu5Ac1Hex6 dHex1HexNAc4Neu5Gc1Hex6 dHex1HexNAc4Neu5Ac2Hex5 dHex1HexNAc4Neu5Ac2Hex5 HexNAc4Neu5Ac1Neu5Gc1Hex5 HexNAc4Neu5Ac1Neu5Gc1Hex5 dHex1HexNAc4Neu5Gc1Neu5Ac1Hex5 dHex1HexNAc4Neu5Gc1Neu5Ac1Hex5 HexNAc4Neu5Gc2Hex5 dHex1HexNAc4Neu5Gc2Hex5

composition

% area P (sd)

% area NP (sd)

p-value

A2 FA2 A2[6]G1 A2[3]G1 A2BG1 FA2[6]G1 FA2[3]G1 FA2Gal1 FA2BG1 A2G2 FA2G2 FA2G1Gal FA2BG2

2.94 (0.09) 3.28 (0.12) 2.00 (0.07) 9.11 (0.29) 0.62 (0.07) 2.40 (0.08) 9.81 (0.36)

6.26 (0.22) 5.68 (0.35) 3.27 (0.12) 13.05 (0.43) 0.54 (0.06) 2.86 (0.12) 9.54 (0.20)