A Simple One-Step Deuterated Peracetylation-MALDI-MS/MS (OSDPM) Strategy for Structural Sequencing of Glycosaminoglycan Oligosaccharides

Abstract

Heparin and heparan sulfate (HP/HS), characterized by heterogeneous sulfation motifs, uronic acid epimerization (GlcA/IdoA), and variations in N-acetyl/N-sulfo groups, play pivotal roles in regulating physiological and pathological processes. This inherent structural complexity, resulting from sulfation heterogeneity and epimeric diversity, poses significant challenges for HP/HS structural analysis. Current labor-intensive multi-step derivatization methods coupled with LC-MS/MS or MSⁿ usually require substantial sample amount (>50 μg) and prolonged processing times, which severely limit their applicability to scarce HP/HS samples. Herein, we develop a one-step deuterated peracetylation-MALDI-MS/MS (OSDPM) analytical strategy that enables high-throughput structural characterization of HP/HS oligosaccharides with ng to sub-μg-level sensitivity. Additionally, this OSDPM approach simultaneously determines oligosaccharide composition, sulfation patterns, and distinguishes between GlcA2S/IdoA2S epimers. We successfully applied this method to characterize the structures of 17 natural HP/HS glycans separated from porcine intestinal mucosa, and their glycan microarray data further validated the structural accuracy of OSDPM-derived sequencing results. The simplicity, sensitivity, and efficiency of our OSDPM strategy provides a valuable solution for HP/HS structural characterization.

INTRODUCTION

Heparin and heparan sulfate (HP/HS), which belong to the glycosaminoglycan (GAG) family, are highly charged linear polysaccharides composed of repeating disaccharide units. Each unit contains a hexuronic acid [either iduronic acid (IdoA) or glucuronic acid (GlcA)] and glucosamine (GlcN), with varying degrees of sulfation at different positions.^{1, 2} HP/HS are ubiquitously expressed on mammalian cell surfaces and play crucial roles in physiological and pathological processes like blood coagulation, inflammatory responses, embryonic development, and pathogen infections.^3–7 Despite their shared backbone, HP/HS exhibit distinct structural features, including variations in sulfation patterns, chain lengths, and hexuronic acid epimerization.^8–11 This structural complexity confers distinct biological functions critical for mediating interactions with diverse proteins like growth factors, chemokines, and enzymes.^12–14 These interactions regulate important cellular processes, such as cell signaling, adhesion, migration, and differentiation.^15–18 The biological activities of HP/HS are highly dependent on their specific structural motifs.^19–21 Therefore, structural characterization is essential for understanding their roles in physiological and pathological processes. Precise determination of HP/HS structures is crucial not only for basic research but also for therapeutic development. For instance, low molecular weight heparins (LMWHs), clinically approved anticoagulants, rely on well-defined structural characteristics for efficacy and safety.^{22, 23} Furthermore, aberrant HS structures have been associated with various diseases, including cancer, inflammatory disorders, and neurodegenerative diseases,^24–27 underscoring the importance of accurate structural analysis for biomarker discovery and targeted therapy.

Currently, the high structural heterogeneity and multiple sulfate groups of HP/HS pose a great analytical challenge.^{28, 29} Uronic acid (UA), exists as either GlcA or IdoA and differs only in the stereochemistry at the C-5 position. The other monosaccharide is GlcN, which can be either GlcNAc (acetylated amine), GlcNS (sulfated amine), or GlcN (free amine).^{30, 31} HP/HS exhibit extensive variability in sulfation patterns, with modifications predominantly occurring at the 6-O position of GlcN or the 2-O position of UA, and less frequently at the 3-O position of GlcN,^{29, 32, 33} which is reported as the final modification in the biosynthetic pathway.^34–36 All these variations give rise to numerous isomers, complicating structural resolution and identification. Overall, HP/HS sequencing requires the determination of composition, differentiation of epimeric monosaccharides (GlcA/IdoA), and recognition of sulfate positions. A major analytical challenge is the inherent lability of sulfate groups during mass spectrometry (MS).³⁷ To address this issue, researchers have developed multi-step chemical derivatization workflows, usually involving permethylation, desulfation, and methylation/acylation. These strategies aim to generate stable acyl/methyl derivatives which preserve and improve positional information for subsequent analysis, either through online liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) or offline MSⁿ.^{36, 38} Specifically, permethylation protects the −OH/NH groups, enhances MS ionization efficiency, and improves cross-ring cleavage signals, facilitating epimer discrimination.³⁹ Solvolytic desulfation followed by site-specific labeling by methylation/acylation further enables sulfate pattern identification by MS/MS.^39–42 Notably, the Sharp group developed multi-step chemical derivatization coupled with online LC-MS/MS strategies for analyzing complex mixtures of GAG oligosaccharides.^{37, 40–42} In contrast, multi-step chemical derivatization coupled with MSⁿ represents offline methods for sequencing purified GAG oligosaccharides.^{39, 43}However, offline multi-step derivatization methods are complicated, time-consuming, and require large amounts of starting materials (often >50 μg). These limitations significantly restrict their utility for analyzing minute quantities of GAG oligosaccharides, such as glycan fractions isolated from natural sources. Therefore, a simpler, faster, and more sensitive strategy is needed for the characterization of natural HP/HS glycans.

In this study, we present a simple, faster, and sensitive strategy termed one-step deuterated peracetylation-MALDI-MS/MS (OSDPM) for high-throughput structural analysis of HP/HS oligosaccharides (Fig 1). The OSDPM approach involves one-step peracetylation with deuterated acetic anhydride (Ac₂O-d₆) in pyridine, followed by matrix-assisted laser desorption/ionization-tandem mass spectrometry (MALDI-MS/MS) analysis. This method not only enables simultaneous determination of oligosaccharide composition and sulfation patterns, but also distinguishes between GlcA2S and IdoA2S epimers. Validation was performed using commercially available HP/HS oligosaccharides, and the method was further applied to characterize 17 natural HP/HS glycans purified from porcine intestinal mucosa. Moreover, glycan microarray data also supported the structural accuracy of OSDPM-based sequencing results.

Figure 1.

Overview of OSDPM strategy. means Anhydromannose (AHM).

EXPERIMENTAL SECTION

Materials.

Commercial heparan sulfate (HS) oligosaccharides containing a para-(6-azidohexanamido) phenyl (Az) tag was obtained from Glycan Therapeutics (Raleigh, NC, USA). Crude heparin sodium from porcine intestinal mucosa was obtained from Celsus Laboratories (Cincinnati, OH, USA). Arixtra sodium salt was obtained from Sigma-Aldrich (St. Louis, MO, USA). HS oligosaccharides containing a para-(6-azidohexanamido) phenyl (Az) tag were purchased from Glycan Therapeutics (Raleigh, NC, USA). Bio-GelP-10 resin was obtained from Bio-Rad (USA). Sodium cyanoborohydride (NaBH₃CN), dimethyl sulfoxide (DMSO), sodium nitrite (NaNO₂), acetic anhydride-d6 and pyridine were purchased from Sigma-Aldrich. AEAB was synthesized in our laboratory. Milli-Q water was used to prepare all aqueous solutions. SARS-CoV-2-Spike protein was obtained from Sino Biological (40589-V08B1). Alexa fluor 647 anti-His antibody was obtained from MBL (D291-A64).

Preparation of Natural HP/HS Oligosacchrides.

The partial NaNO₂ degraded heparin (porcine intestinal mucosa)/Arixtra^® was labeled with AEAB via reductive amination using AEAB·2HCl (88 mg/mL) and NaCNBH₃ (64 mg/mL) in DMSO/AcOH (7:3, v/v) at 65 °C for 2 h. Then the Heparin-AEAB was fractionated via Bio-Gel P-10 column chromatography (5×170 cm) using 0.2 M NH4HCO₃ as the eluent, with fractions collected based on UV absorption at 330 nm and lyophilized. Subfractionation of HDP4/6-AEAB isomers was performed by preparative HPLC (Luna C18(2) column, 50×250 mm) using a gradient of 15-25% ACN in 25 mM ammonium acetate with 15 mM di-n-butylamine and 0.3% acetic acid. Fractions were further purified by recycling HPLC under isocratic conditions (20% ACN) and analyzed for purity/quantification via HPLC (Sinochrom ODS-BP column, 4.6×250 mm) using a 2-15% ACN gradient with 25 mM ammonium acetate, 15 mM trimethylamine, and 0.3% acetic acid, monitored at UV 330 nm and fluorescence (Ex 330 nm/Em 420 nm).

EPAB labeled disaccharides GlcA2S-AHM-EPAB and IdoA2S-AHM-EPAB were obtained by complete degradation of commercial heparin using nitrous acid followed by EPAB labeling by reductive amination. After size exclusive chromatography, the EPAB tagged disaccharides are purified by preparative C18-HPLC (30 x 150 mm, 5 μm) using isocratic 12% acetonitrile/0.1% TFA as mobile phase.

NMR Spectrum for GlcA2S-AHM-EPAB.

¹H NMR (600 MHz, D₂O) δ 7.78 (d, J = 8.8 Hz, 2H, H-Ar), 6.73 (d, J = 8.8 Hz, 2H, H-Ar), 5.11 (s, 1H), 4.62 (d, J = 2.5 Hz, 1H, H-1), 4.28 – 4.21 (m, 3H), 4.17 – 4.15 (m, 1H), 4.13 – 4.07 (m, 4H), 4.00 (t, J = 3.7 Hz, 1H), 3.91 (t, J = 2.9 Hz, 1H), 3.54 (dd, J = 11.8, 4.4 Hz, 1H), 3.45 (dd, J = 11.7, 6.6 Hz, 1H), 3.42 – 3.37 (m, 2H), 3.26 – 3.24 (m, 2H), 1.26 (t, J = 7.2 Hz, 3H, CH₃). (Fig S7).

NMR Spectrum for IdoA2S-AHM-EPAB.

¹H NMR (600 MHz, D₂O) δ 7.89 – 7.79 (m, 2H, H-Ar), 6.85 – 6.75 (m, 2H, H-Ar), 5.08 (s, 1H), 4.67 (d, J = 7.9 Hz, 1H), 4.50 – 4.41 (m, 2H), 4.33 – 4.27 (m, 2H, OCH₂), 4.21 – 4.11 (m, 2H CH₂NH-Ar), 4.02 – 3.95 (m, 1H), 3.78 – 3.74 (m, 1H), 3.65 – 3.61 (m, 2H), 3.55 – 3.52 (m, 5H), 3.38 – 3.32 (m, 1H), 1.39 – 1.28 (m, 3H, CH₃). (Fig S8).

Structure Analysis of HP/HS Glycans by Deuterated Peracetylation.

0.5 to 5 nmol of the above separated dried HP/HS oligosaccharides were reacted with 10 μL of reaction reagents (acetic anhydride-d6/pyridine=1:2, v/v) at 65°C for 1 hour, then dried by speed vacuum concentrator.

MS Spectrometry.

The derivatized glycans were resuspended in 15 μL of 50% methanol. For MALDI-MS (Bruker Daltonics Ultraflex-II) analysis, 1 μL sample and 1 μL matrix (10 mg/mL 2,5-dihydrobenzoic acid in 80% methanol) were spotted on an Anchorchip target plate, air dried and analyzed by MALDI-MS, then target ions were selected for fragmentation analysis. The MALDI-MS and MS/MS spectra were analyzed by flexAnalysis software. LC-MS (Agilent, MA) analyses were conducted on a 1290 Infinity II LC coupled with a 6545XT AdvanceBio LC/Q-TOF system equipped with an Agilent Jet Stream source. 2 μL of derivatized glycan or underivatized glycan (0.2 nmol/μL) was separated on a ZORBAX Extend-C18 column (2.1×50 mm, 1.8 μm, Agilent, 727700-902) with a 20 min gradient, solvent B (ACN with 0.1% fomic acid), from 2% to 35% in 15 min. The column was re-equilibrated at initial conditions for 5 min. MS/MS acquisition was performed automatically. All LC-MS data were processed using Agilent MassHunter Qualitative Analysis 10.0 software. Interpretation of MS data sets were made manually.

HP/HS Microarray.

Contact printing on Nexterion slide H (Schott Nexterion, az) was carried out using an Aushon’s 2470 microarrayer (Quanterix, MA) according to standard protocol for microarray printing of ECGC (Emory comprehensive glycomics core).⁵⁹ Briefly, AEAB labeled heparin glycan fractions were printed at 100 μM in 100 mM sodium phosphate (pH 8.5) in replicates of four. The printed slides were placed in a chamber at 55°C water bath for 1 hour, and were air dried at room temperature, then the slides were blocked in 50 mM ethanolamine in 50 mM sodium borate (pH 8.5) for 1 hour. The binding assay was started with blocking the arrayed slides with 5% BSA in TSMT buffer (20 mM Tris, 2 mM CaCl2, 2 mM MgCl2, 150 mM NaCl, 0.2% Tween 20, pH 7.4) at room temperature for 1 hour. Subsequently, different concentrations of FGF2 (0.001, 0.01, 0.1 μg/mL), SARS-CoV-2 spike protein (2.5, 10, 25, 50, 100 μg/mL), or UV treated SARS-CoV-2 (original neat titer was 1.8 x 107 PFU/mL before UV treatment) were applied. The slides were then washed and incubated with suitable detection reagent. For FGF2 protein, anti-FGF2 antibody (10 μg/mL) and subsequent AlexaFluor 647 goat anti-mouse lgG (10 μg/mL Jackson Immuno Research, 115-605-062) were used; for His-tagged SARS-CoV-2, and spike protein, AlexaFluor 647 anti-His (10 μg/mL) was used. The slides were scanned using a fluorescent scanner (InnoScan AL1100, Innopsis, IL). The images were analyzed by Mapix software (Innopsys, IL), and the data was analyzed with a home written Excel macro. The highest and the lowest value of the total fluorescence intensity of the replicate spots were removed, and the remaining values were used to provide the mean value and standard deviation.

RESULTS AND DISCUSSION

Development of the OSDPM Strategy for Structural Characterization of Heparin Oligosaccharides.

Matrix-assisted laser desorption/ionization-mass spectrometry (MALDI-MS) is widely used for glycan analysis due to the unique advantages of low sample consumption, simple sample preparation, high-throughput, rapid analysis, etc.⁴⁴ However, application of MALDI-MS in GAGs has been limited because of the easy decomposition of sulfate groups and low ionization efficiency of the highly anionic glycans.^{36, 45} Here, we applied peracetylation using deuterated acetic anhydride (Ac₂O-d₆)/pyridine to HP/HS oligosaccharides, followed by analysis by MALDI-MS/MS at positive mode. We found that this simple one-step deuterated peracetylation coupled with MALDI-MS/MS method (OSDPM) of HP/HS not only can increase the detection sensitivity, but also produce clear MALDI-MS and tandem MS (MS/MS) patterns of HP/HS oligosaccharides for structural identification. After deuterated peracetylation, we observed several very informative MALDI-MS patterns (Fig 2a): 1) As expected, peracetylation is very efficient. All hydroxyl group (−OH) and amino group (−NH) in N-sulfoglucosamine (−NHSO₃H) are acetylated while O-sulfate is not. Interestingly and importantly, all of the sulfate groups were easily lost during MALDI-MS at positive ion mode, resulting in a 45 Da mass loss per sulfate group. This essentially generates a clear desulfated and peracetylated (pAc) molecular ion. When this desulfated molecular ion is selected for collision induced dissociation (CID), glycosidic bond breakage produces a series of partially acetylated and desulfated fragments where a free −OH/NH indicates an O/N-sulfation before peracetylation. These fragments can be used to deduce the oligosaccharide sequence as well as sulfate number of each monosaccharide (Fig 2b, c and d). 2) A lactone forms between the hydroxyl groups at C3 or C2 (C3-OH or C2-OH) and the carboxyl group at C5 (C5-COOH) of GlcA, GlcA2S, and IdoA, but not IdoA2S. The formation of this lactone results in a 63 Da mass loss, which can be used to distinguish GlcA2S from IdoA2S. 3) The N-sulfoglucosamine group (−NHSO₃H) in GlcNS can be acetylated to form (−NHAc-d₃), causing a 3 Da mass difference with GlcNAc, which could make a distinction between GlcNS and GlcNAc.

Figure 2.

Structural characterization of peracetylated heparin oligosaccharide. (a) The MS change patterns of peracetylated HP/HS on MALDI-MS/MS at positive mode. (b) The MALDI-MS/MS profile of peracetylated NA 4-Mer. (c) The MALDI-MS/MS profile of peracetylated NS IdoA 6-Mer A. (d) The MALDI-MS/MS profile of peracetylated 3S 6-Mer. (d) The MALDI-MS/MS profile of peracetylated GlcA2S-AHM-EPAB and IdoA2S-AHM-EPAB. means Anhydromannose (AHM), means IdoA, means GlcN, means GlcA.

To establish the OSDPM method, we performed deuterated peracetylation on commercially available enzymatically synthesized HS oligosaccharides containing a para-(6-azidohexanamido) phenyl (Az) tag. All MS¹ MALDI data show multiple peaks of peracetylated and desulfated HS, with each peak corresponding to a distinct number of lactones (Fig 2b, c, d). The formation of different numbers of lactones are obviously due to the competition of acetylation of available hydroxyl group for lactone formation. Interestingly, when MS¹ peaks derived from the same HS were selected for secondary mass spectrometry (MS²) analysis, their resulting fragmentation patterns were very similar and matched the spectrum of the MS¹ peak with the highest number of lactones (Fig S1). This phenomenon suggests that the structural determination of HS using peracetylated MS¹ data can be reliably achieved by utilizing the MS¹ data corresponding to the most abundant lactone-containing species (the smallest molecular ion in MS¹ data). For instance, the molecular ions with the highest number of lactones at m/z 1308 ([M+Na]⁺, Fig 2b), 1811 ([M+Na]⁺, Fig 2c), and 1652 ([M+Na]⁺, Fig 2d) match the expected compositions for peracetylated and desulfated NA 4-Mer [(GlcA/IdoA)₂(GlcNAc)₂], NS IdoA 6-Mer [(GlcA/IdoA)₃ GlcNAc₁ (GlcNS)₂], and 3S 6-Mer [(GlcA/IdoA)₃(GlcNS)₃]with 5 O-sulfate groups, respectively. Moreover, the MS² data further characterize HS sequences and sulfation positions as GlcNAc-GlcA/IdoA-GlcNAc-GlcA/IdoA-Az, GlcNAc-GlcA/IdoA-GlcNS-GlcA/IdoA-GlcNS-GlcA/IdoA-Az, and GlcNS(S)-GlcA/IdoA-GlcNS3S6S-IdoA2S-GlcNS(S)-GlcA-Az, respectivety, showing high consistency with their known structures: GlcNAc-GlcA-GlcNAc-GlcA-Az (NA 4-Mer), GlcNAc-GlcA-GlcNS-IdoA-GlcNS-GlcA-Az (NS IdoA 6-Mer), and GlcNS6S-GlcA-GlcNS3S6S-IdoA2S-GlcNS6S-GlcA-Az (3S 6-Mer). Due to the instability of azido group (42 Da) during the MALDI-MS process, which readily reduces to amino group (16 Da) ⁴⁶, characteristic peak pairs with a 26 Da mass difference are consistently observed in the mass spectrum (Fig 2b and c). Moreover, Fig 2b, 2c and 2d demonstrated that GlcA and IdoA undergo lactonization during peracetylation, whereas IdoA2S does not. However, whether GlcA2S can form a lactone remains unclear. To address this, we prepared two disaccharides containing GlcA2S and IdoA2S respectively from heparin disaccharides with an EPAB (ethyl 4-aminobenzoate) tag (Fig 2e, Fig S7–S8 ). After deuterated peracetylation, the MALDI-MS data revealed that GlcA2S, like GlcA and IdoA, can form a lactone, in contrast to IdoA2S, which showed no detectable lactone formation. Furthermore, the 102 Da mass increase was identified as SO₃Na⁺ based on the MS² data (Fig S2).

The distinct lactone formation patterns of GlcA, GlcA2S, IdoA, and IdoA2S upon acetylation can be attributed to their unique steric configurations (Scheme 1). When C5-COOH is activated, only hydroxyl groups positioned on the same side of the ring can participate in intramolecular attack, leading to bridge lactone formation. This reaction is feasible for the C3-OH in GlcA/GlcA-2S and the C2-OH in IdoA. However, in IdoA-2S, the C2-OH is blocked by sulfation, and the C3-OH is positioned opposite to C5-COOH, making lactone formation unfavorable under these conditions. Therefore, this OSDPM method distinguishes sulfated GlcA/IdoA epimers (GlcA2S vs. IdoA2S) via lactone formation, non-sulfated IdoA and GlcA cannot be resolved due to identical lactone masses.

Scheme 1.

Bridge lactone formation of GlcA, GlcA2S, IdoA and IdoA2S. Red and blue bonds represents the opposite sides of the sugar ring.

To verify the reliability of the OSDPM method for HP/HS structural analysis, we extended our validation to incorporate more complex HP/HS structures. Four 8-Mer HS standards were peracetylated using acetic anhydride-d₆/pyridine, and then analyzed by MALDI-MS/MS. The MS¹ data present three dominant peaks at m/z 2313 (M+Na)⁺, 2376 (M+Na)⁺, and 2440 (M+Na)⁺, corresponding to the 4-lactone, 3-lactone, and 2-lactone states of the pAc NS 8-Mer [(GlcA/IdoA)₄(GlcNS)₄], respectively (Fig 3). The MS² data of the precursor ion at m/z 2313 show fragment ions of Y2-Y6, B3, B5, B6, C2, C4, and C6, which determine the locations of 4 GlcNS and 4 lactones (GlcA/IdoA), closely matching the known NS 8-Mer structure. Moreover, three other 8-Mer HS standards were also analyzed by the OSDPM (Fig S3). All MALDI-MS/MS data confirmed most of their structural features, except for distinguishing between GlcA and IdoA.

Figure 3.

Structural characterization of peracetylated NS 8-Mer.

Application of the OSDPM Strategy for Structural Characterization of Natural HP/HS.

As mentioned above, the structural characterization of natural HP/HS remains challenging because of their low available amount. To address this, we applied our OSDPM method to sequence HP/HS oligosaccharides prepared from porcine intestinal mucosa. To further validate the approch, we first used the commercial heparin pentasaccharide fondaparinux sodium (Arixtra^®) as a model compound. After partial nitrous acid (NaNO₂) degradation and N-(aminoethyl)-2-aminobenza-mide (AEAB) conjugation, we isolated the pentasaccharide-AEAB for deuterated peracetylation. Fig 4a shows the MS profile of peracetylated Arixtra-AEAB. The dominant molecular ion at m/z 1410 ([M+Na]⁺) matches the expected peracetylated and desulfated Hdp5-AEAB [(GlcA/IdoA)₂(GlcNS)₂AHM-AEAB] with 5 O-sulfate groups. The MS² profile of the precursor ion at m/z 1410 ([M+Na]⁺), which is consistent with the known structure GlcN6S-GlcA-GlcN3S6S-IdoA2S-AHM6S-AEAB, shows 5 major fragment ions: B1 (m/z 297.10, [M+Na]⁺ ) indicating one sulfation, and one GlcNS; Y₁ (AHM-AEAB) (m/z 483.56, [M+Na]⁺) indicating one sulfation; Y₂ (m/z 704.63, [M+Na]⁺) indicating two sulfations, and one IdoA2S; Y₃ (m/z 889.82, [M+Na]⁺) indicating four sulfations, one IdoA2S, and one GlcNS; Y₄ (m/z 1113.24, [M+Na]⁺) indicating four sulfations, one IdoA2S, one GlcNS, and one lactone (GlcA/IdoA); C₃ (m/z 889.82, [M+Na]⁺) indicating three sulfations, one GlcNS and one lactone (GlcA/IdoA). We also compared the MALDI-MS/MS data with LC-MS/MS data (Fig S4). The ESI-LC-MS¹ analysis of peracetylated Arixtra-AEAB presents clusters of peaks due to partial desulfation; the MS² data on LC-MS/MS were very consistent with the MS² data on MALDI-MS/MS. Importantly, we found even with an initial amount of material as low as 0.5 nmol (<1 μg), a very clear MALDI-MS profile was obtained for structural characterization (Fig 4b). Therefore, this OSDPM approach, which is faster, simpler and more sensitive than multi-step derivatization, is suitable for structural characterization of natural HP/HS glycans separated from porcine intestinal mucosa, ranging from trisaccharides to heptasaccharides with sulfation number varying from 2 to 9 (Fig 4e and Fig S5). Fig 4c and 4d showed two example MALDI-MS/MS profiles of one hexasaccharide and one heptasaccharide respectively. Interestingly, most of the identified natural HS/HP structures contain GlcA2S, which is reported as a less abundant monosaccharide in HS (~<5% of total saccharides)⁴⁸.

Figure 4.

Structural characterization of natural HP/HS oligosaccharides prepared from porcine intestinal mucosa. (a) MALDI-MS/MS profile of peracetylated Arixtra. (b) MALDI-MS profiles of peracetylated Arixtra with different starting amounts. (c) MALDI-MS/MS profile of peracetylated Hdp6-9. (d) MALDI-MS/MS profile of peracetylated Hdp7-1. (e) Structures of natural HP/HS glycans identified via OSDPM strategy.

Glycan Microarrays for Structural Validation and Functional Glycomics.

To further validate the structures of the natural HS/HP glycans sequenced by the OSDPM-based method, we developed a heparin oligosaccharide microarray using the above-mentioned prepared natural HP/HS library. AEAB-labeled heparin glycans were printed onto N-hydroxysuccinimide (NHS)-activated glass slides following MIRAGE guidelines⁴⁹ (Table S1). We first examined the binding specificity of fibroblast growth factor-2 (FGF2), a well-characterized heparin-binding protein known to recognize IdoA2S.⁵⁰ As shown in Fig 5a, almost all of the HP/HS glycans identified with two IdoA2S demonstrated stronger FGF2 binding than those identified only with a single IdoA2S, confirming the structural validity of our sequencing results. Arixtra, Hdp3-1, Hdp7-1, and Hdp7-2 which contain a GlcNS(S)-IdoA2S terminal disaccharide structure showed obviously higher binding affinity to FGF2 than Hdp4-1, 2, and 3, which also contain one IdoA2S. This indicated that the GlcNS(S)-IdoA2S terminal motif had higher binding affinity to FGF2 than a single terminal/middle IdoA2S.

Figure 5.

Binding analysis of HP/HS oligosaccharides to different glycan binding proteins. (a) FGF2, (b) SARS-CoV-2, (c) Spike protein. Neat denotes that before UV treatment the original titer is 1.8×10⁷ PFU/mL, and 2×neat denotes that before UV treatment the original titer is 3.6×10⁷ PFU/ mL.

It is well known that HP/HS play a key role in coronavirus 2 (SARS-CoV-2) infection, and the structures of HP/HS including the degrees of polymerization (DP) and sulfation patterns are critical for viral entry into host cells.^51–57 Therefore, we next applied our HP/HS microarray to analyze the binding specificity of SARS-CoV-2 (Fig 5b). UV-inactivated SARS-CoV-2 was initially screened on the nCFG glycan microarray ⁵⁸ (Fig S6). No apparent binding was observed, indicating that SARS-CoV-2 does not bind the glycans included in the nCFG glycan library which contains no heparin structures. In contrast, SEC-separated dp2, 4, 6, 8, 10, 12, and high-dp HP/HS glycans exhibited length-dependent binding to SARS-CoV-2. Notably, hexasaccharides/heptasaccharides with ≥7 sulfate groups showed the strong interaction, suggesting that both chain length (dp ≥6) and sulfation intensity (≥7 sulfates) are essential for SARS-CoV-2 binding.

The SARS-CoV-2 spike (S) protein, a known HP/HS-binding factor facilitating viral attachment,⁵¹ was further tested on the HP/HS microarray (Fig 5c). Compared to SARS-CoV-2, the same binding pattern of dp2, 4, 6, 8, 10, 12, and high dps was observed, indicating that both SARS-CoV-2 and S protein bind HP/HS in a length-dependent manner. Unlike the whole virions, which require dual conditions (dp ≥6 and ≥7 sulfates) for binding, the S protein depends solely on sulfation intensity (≥ 4), as evidenced by the strong binding of a trisaccharide (Hdp3-1) containing 4 sulfate groups.

CONCLUSIONS

Conventional methods for structural characterization of HP/HS glycans, usually involving multi-step derivatization prior to MS analysis, often demand substantial sample amount (>50 μg).^{39, 43} Such requirements are impractical for natural HP/HS glycans, which are often isolated in sub-μg amounts despite being sufficient for glycan microarray applications. To address this issue, we developed a streamlined one-step deuterated peracetylation-MALDI-TOF-MS/MS (OSDPM) strategy. Unlike multi-stage methods, relying on chemical derivatizations and desulfation, the OSDPM method simply requires single-step deuterated peracetylation (heating at 65°C for 1 h followed by solvent evaporation), significantly reducing the time (<2 h), sample amount (<1 μg), and side reactions, resulting a possibly automatable method. Importantly, acetylation serves as a code for free −OH’s and O-sulfation serves as a protecting group. The desulfation during MALDI ionization enables positive mode MS analysis while keeping the sulfation information. The formation of intramolecular bridge lactone during peracetylation on GlcA, IdoA, and GlcA2S can be used to distinguish GlcA2S and IdoA2S easily. As the result, this OSDPM method provides rich information on the sequence, sulfation pattern and GlcA2S/IdoA2S isomer identification. The data analysis relies only on abundant glycosidic bond fragmentation without involving cross-ring cleavage, thus resembling peptide sequence analysis, and does not directly determine either the sulfation position (GlcNS3S and GlcNS6S) or the GlcN/GlcNS differentiation. However, the sulfation pattern of different classes of GAGs are well established and can be applied to most purified HP/HS glycans. This limitation highlights the need for advanced MS techniques (e.g., MS³) or complementary chemical modifications to resolve sulfation isomers, as demonstrated in prior multi-step approaches.^{37, 40} Although the structural analysis in this report is carried out manually, we strongly believe this could provide a general route for future GAG automatic sequencing as more structures are analyzed using this method and the algorithm is developed into software. Currently, this method is primarily applied to purified glycans rather than complex mixtures due to incomplete derivatization, which produces chimeric MS spectra. In the future, it could potentially be extended to glycan mixtures if the derivatization yield, particularly lactone formation efficiency, is further improved. Additionally, further studies may enable the differentiation of GlcA and IdoA based on potentially distinct reactivity of C3-C5 lactone in GlcA and C2-C5 lactone in IdoA.

In summary, we have developed a simple, fast, and sensitive OSDPM approach for the structural characterization of HP/HS glycans. We have validated this method using commercial HP/HS oligosaccharides, and applied it to characterize 17 natural HP/HS glycans purified from porcine intestinal mucosa. The glycan microarray data further demonstrated the structural validity of the sequencing results by the OSDPM. We believe this OSDPM approach will serve as a unique valuable tool for future structural characterization of HP/HS glycans.

Supplementary Material

The MS² analysis of pAc-NA 4-Mer, pAc-NS IdoA 6-Mer A, pAc-3S 6-Mer, pAc-GlcA2S-AHM-EPAB and pAc-IdoA2S-AHM-EPAB. Structural characterization of pAc-2S 8-Mer B, pAc-3S 8-Mer, and pAc-2S 8-Mer C. The LC-MS/MS profile of peracetylated Arixtra. MALDI-MS/MS profiles of all 17 peracetylated natural HP/HS glycans separated from porcine intestinal mucosa. The binding specificity of SARS-CoV-2 on nCFG microarray. NMR spectrum for GlcA2S-AHM-EPAB and GlcA2S-AHM-EPAB.

All raw MS spectra data are available on the GlycoPost public repository. (URL: https://glycopost.glycosmos.org/preview/11798744756830abef41b33 , PIN CODE:6125).