Desalting Paper Spay Mass Spectrometry (DPS-MS) for Rapid Detection of Glycans and Glycoconjugates

Abstract

The detection of glycans and glycoconjugates has gained increasing attention in biological fields. Traditional mass spectrometry (MS)-based methods for glycoconjugate analysis are challenged with poor intensity when dealing with complex biological samples. We developed a desalting paper spray mass spectrometry (DPS-MS) strategy to overcome the issue of signal suppression of carbohydrates in salted buffer. Glycans and glycoconjugates (i.e., glycopeptides, nucleotide sugars, etc.) in non-volatile buffer (e.g., Tris buffer) can be loaded on the paper substrate from which buffers can be removed by washing with ACN/H₂O (90/10 v/v) solution. Glycans or glycoconjugates can then be eluted and spray ionized by adding ACN/H₂O/formic acid (FA) (10/90/1 v/v/v) solvent and applying a high voltage (HV) to the paper substrate. This work also showed that DPS-MS is applicable for direct detection of intact glycopeptides and nucleotide sugars as well as determination of glycosylation profiling of antibody, such as NIST monoclonal antibody IgG (NISTmAb). NISTmAb was deglycosylated with PNGase F to release N-linked oligosaccharides. Twenty-six N-linked oligosaccharides were detected by DPS-MS within a 5-minute timeframe without the need for further enrichment or derivatization. This work demonstrates that DPS-MS allows fast and sensitive detection of glycans/oligosaccharides and glycosylated species in complex matrices and has great potential in bioanalysis.

Graphical Abstract

Introduction

Glycosylation refers to the process of the addition of carbohydrate or glycan chains to proteins,¹ lipid molecules,² RNA,³ or other biomolecules,^4–5 which are referred to as glycoconjugates. This post-translational modification is critical for the function and characteristics of glycoconjugates, such as conformation, stability, pharmacokinetic and pharmacodynamic properties,^6–7 and protease resistance.⁸ Many proteins are glycosylated and glycan chains are linked to the nitrogen atoms of asparagine residue by an N-glycosidic bond (N-linked glycans) or to the oxygen atoms of serine/threonine residues by an O-glycosidic bond (O-linked glycans). Biosynthesis of N-glycans is a complex process in eukaryotic cells including mammals,⁹ which typically involves the catalysis of glycosyltransferases (GTs) using nucleotide sugar donors (e.g., monosaccharide building blocks activated by nucleoside mono- or diphosphate).¹⁰ As an important class of glycoproteins, monoclonal antibodies (mAb) are key therapeutics for a range of diseases including cancer, infections and autoimmune disorders.^11–14 The glycan composition impacts the stability and safety in the development of most antibody based biotherapeutics.^15–17 In particular, oligosaccharides on Asn²⁹⁷ of the Fc region play an essential role in Fc effector functions that are key mechanisms of therapeutic action^18–19, hence the influence of therapeutic efficacy drives an intense focus on glycoengineering of the attached glycoforms.²⁰ Therefore, due to the impact of glycosylation on the structures and functions of proteins or other biomolecules, structural characterization of glycans and glycoconjugates is critical.

Electrospray ionization mass spectrometry (ESI-MS) and matrix-assisted laser desorption-ionization mass spectrometry (MALDI-MS) are often used to characterize glycans.^21–24 Accurate molecular mass, along with tandem mass analysis, provides unprecedented capability for glycan structure characterization and sugar sequence assignment.^25–30 However, due to signal suppression by more readily ionized compounds (e.g., salts and other solutes like detergents, chaotropes that are often used in proteomics sample preparation step^31–32), it is still challenging to detect and analyze glycans/oligosaccharides in complex biological mixtures.³³ Therefore, efficient ionization and detection of carbohydrates require enrichment and desalting steps. Existing desalting methods include chromatography^34–44 and capillary electrophoresis,^45–49 which are time-consuming (time ranging from 50 mins up to 2 hrs) and have limitations such as low enrichment efficiency and nonspecific binding of oligosaccharides due to their high polarities. Additionally, derivatization is needed prior to MS analysis due to the low ionization efficiency of oligosaccharides, which also brings the limitation of laborious procedure and extensive experiment time. Thus, developing a fast and sensitive MS detection method for analysis of glycans/oligosaccharides in complex matrices is needed.

Direct sampling ionization by ambient ionization MS has attracted more attention in the past decade.^50–52 Paper spray ionization (PSI) is one of the most popular ambient ionization techniques that has emerged in recent years.^53–55 Paper spray involves paper substrates cut to triangular shapes with sharp points, onto which samples are loaded. Next, spray solvent and high-voltage (HV) are applied to the paper for simultaneous elution and ionization. Since its introduction in 2010 by Ouyang, Cooks, and co-workers,⁵⁶ paper spray has been developed for various clinical,^57–59 food safety,⁶⁰ and forensic applications.^61–63 However, only few studies have been reported on large carbohydrate molecules. However, analysis of glycans from complex matrices using PSI-MS is still lacking, probably because the conventional PSI solvent (MeOH or ACN/H₂O 50/50 v/v) cannot elute target oligosaccharides from the paper due to their strong hydrophilic interaction with the fibrous paper.^64–66 Li et al. developed a hydrophobic polystyrene-impregnated paper substrate for protein and peptide analysis by first coating polystyrene microspheres followed by baking at a high temperature.⁶⁷ Riboni et al. designed a new setup named Solvent-Assisted Paper Spray Ionization (SAPSI) that integrates a power supply and fluidics system to prolong analysis time for protein, peptide and N-glycans.⁶⁸ Recently, we developed a desalting paper spray ionization (DPS) method for the rapid analysis of oligosaccharides in complex matrices or from enzymatic reactions, which employs a triangular filter paper for both sample desalting and ionization (desalting refers to removal of excess amount of non-volatile buffers and inorganic salts).⁶⁹ The rationale is that the paper substrate for spray ionization consists mainly of cellulose which could have strong interactions with oligosaccharides via hydrogen bonding and thus can serve as a stationary phase for desalting purpose.

In this study, we further extended the application of DPS-MS for detecting low amounts of various glycans with complex structures, as well as glycopeptides in the presence of non-volatile buffers. Our findings indicate that simple one-step desalination resulted in the elimination of signal suppression from non-volatile buffer without compromising the desirable qualities of paper spray, namely short analysis time, minimum sample pretreatment, accuracy, and simplicity. DPS-MS uses ACN/H₂O (90/10 v/v) to wash away unbound salts and other chemicals in sample mixtures before elution of glycans/oligosaccharides with higher water content (ACN/H₂O/FA, 10/90/1 v/v/v) for ionization. DPS-MS successfully detected various glycans/oligosaccharides, including sialylated and non-sialylated glycopeptides as well as nucleotide sugars in non-volatile buffers. We further extended the study to include N-glycan analysis from deglycosylated antibody (NISTmAb), which brings a new perspective for DPS-MS, as a valuable tool for clinical and pharmaceutical applications.

Experimental Section

Materials

Xyloglucan (XyG) heptasaccharide, 6³, 6⁴-α-D-Galactosyl-mannopentaose, 2³-(4-O-Methyl-α-D-Glucuronosyl)-xylotetraose, 6³-α-D-Glucosyl-maltotriosyl-maltotriose and 2³,3³-di-α-L-Arabinofuranosyl-xylotriose were obtained from Megazyme (Bray, Ireland). β-cyclodextrin, vancomycin hydrochloride and ammonium hydroxide were from Sigma-Aldrich (St. Louis, MO, United States). Sialylglycopeptide was from Tokyo Chemical Industry Co. (Tokyo, Japan). A UDP-glucuronate sample mixture in 2.5 mM Tris buffer (pH 7.4) with 50 μM MgCl₂ and 50 μM MnCl₂ was received from Sigma-Aldrich (St. Louis, MO, United States). Humanized IgG1κ monoclonal antibody (NISTmAb) was NIST reference material 8671 purchased from NIST. HPLC-grade acetonitrile, 1 M Tris-HCl buffer solution (pH 7.0), formic acid was from Fisher Scientific (Waltham, MA, United States). Deionized water was from EMD Millipore (Burlington, MA, United States). Whatman-grade 42 and 2.5 μm chromatography paper was from Whatman International Ltd. (Maidstone, England). Rapid PNGase F enzyme was purchased from New England Biolabs (Ipswich, MA, United States).

Instrument setup

Mass analysis was performed by using a high-resolution Q Exactive Orbitrap mass spectrometer (Thermo Scientific, San Jose, CA, USA) with ion source removed to accommodate PSI, DPS and nano-ESI setups. The capillary temperature was set at 250°C. Spray voltage was 3.0 kV at positive ion mode and 3.5 kV in negative ion mode. Tandem mass spectra were acquired using collision-induced dissociation (CID).

Experimental Procedures

DPS-MS Method

The DPS method was slightly modified based on a previously described method.⁶⁹ The filter paper was cut into a triangular shape (15 × 10 mm, height × base) followed by sequential sonication-assisted cleaning in acetone, methanol, and methanol/water (50/50 v/v, 15 min each) and complete drying. Twenty μL of sample solution was dropped onto a paper triangle by loading 10 μL twice. The paper was placed over a Kimwipe to facilitate the absorption by capillarity as shown in Scheme 1a. The paper triangle was transferred onto a clean Kimwipe to perform desalting. Ten μL of ACN/H₂O solution (90/10 v/v) was applied to the paper to remove salts and other chemicals. The washed paper triangle was then placed 8 mm in front of the MS inlet (see picture shown in Scheme 1b). A high-voltage cable alligator clip was used to hold the paper triangle and 10 μL of ACN/H₂O/FA solution (10/90/1 v/v/v) was then applied to the paper triangle to elute the target compounds for ionization upon the application of high voltage.

Scheme 1.

a) Schematic showing the workflow and setup of our DPS-MS method. b) Picture showing the instrumental setup of DPS-MS.

PSI-MS Method

Filter paper was cut into a triangular shape (15 × 10 mm, height × base) followed by sequential sonication-assisted cleaning in acetone, methanol, and methanol/water (50/50 v/v, 15 min each) and complete drying. Sample solution (20 μL) was loaded onto the paper triangle. After total drying, the paper was held in front of the MS inlet with the same setup as for DPS (see previous section). 10 μL of ACN/H₂O/FA solution (10/90/1 v/v/v) was then added to the paper triangle for elution of the target compounds and ionization upon the application of high voltage.

Nano-ESI-MS Method

A laser puller (model P1000, Sutter Instrument Inc, USA) was used to produce a conical-tip (o.d. ~15 μm) fused-silica capillary (100 μm i.d., 200 μm o.d., Polymicro Technologies, Phoenix, AZ). A syringe loaded with sample solution was attach to the capillary and held in front of the MS inlet with the application of high voltage (+3.0 kV) for ionization. The flow rate was 2 μL/min. The distance between the tip of the capillary and the MS was 8 mm.

Deglycosylation by PNGase F treatment

An optimized one-step protocol that utilizes Rapid PNGase F for deglycosylation of proteins was developed. Each reaction mixture contained 100 μg of NIST IgG antibody (NISTmAb) mixed with 6 μL of water, 4 μL of Rapid PNGase F Buffer (5X, the buffer composition information is proprietary) and 1 μL of Rapid PNGase F (0.8 mg/mL). The mixture was incubated overnight at 37°C in a water bath to release N-linked glycans (Scheme 2).

Scheme 2.

Deglycosylation of antibody by PNGase F. PNGase F reaction N-glycosylation sites is shown.

Results and Discussion

Optimization of DPS

First, we tested the optimal sample solution loading amount on the paper (10 × 5 mm, height × base) for the best signal intensity. In our previous DPS-MS workflow,⁶⁹ 10 μL was chosen as the sample loading volume. Optimization was achieved by increasing the sample loading volume on the paper triangle in series of 10 μL. Sample solution (10 μL) was loaded onto paper and allowed to be absorbed completely by the paper substrate (e.g., cellulose) before adding an additional 10 μL of sample, followed by a desalting step. The paper substrate was placed over a Kimwipe to facilitate capillarity as shown in Scheme 1a. For optimization, 50 μM β-cyclodextrin sample in 50 mM Tris-HCl (pH 7.0) buffer was used. The signal intensity of β-cyclodextrin, when loaded twice onto the paper triangle, was shown more than two times greater than when only loaded once (Figure S-1a, b, Supporting Information). Further additions of sample were also tested and our data indicated that the salt in the sample solution exceeded the desalting capacity of the paper triangle when a total of 30 μL of sample was used. The spectrum shows the reduced intensity of β-cyclodextrin peak due to the increased signal suppression resulting from the more abundant Tris clusters existing in the sample solution (Figure S-1c, Supporting Information). Thus, 20 μL of sample solution was selected as the optimal loading volume throughout this study.

Detection of oligosaccharides

Three standard oligosaccharides were tested in our previous work.⁶⁹ In this study, we investigated a broader variety of oligosaccharides with different structure features such as long chains and branches. Oligosaccharides were dissolved in non-volatile Tris-HCl buffer to mimic a biological matrix. Solutions of XyG heptasaccharide (X₃Glc₄), 6³, 6⁴-α-D-Galactosyl-mannopentaose (GGM5), 2³-(4-O-Methyl-α-D-Glucuronyl)-xylotetraose (XUXX), 6³-α-D-Glucosyl-maltotriosyl-maltotriose (GMH), and 2³,3³-di-α-L-Arabinofuranosyl-xylotriose (A^2,3XX), were prepared as 500 μM in 50 mM Tris-HCl buffer (pH 7.0). Scheme 3 shows the structure and chemical formula of these five oligosaccharides used in this experiment. DPS-MS was able to detect all of these oligosaccharides, namely, [Heptasaccharide +Na]⁺ at m/z 1085 (measured m/z 1085.33765, theoretical m/z 1085.33786, mass error: −0.19 ppm), [6³, 6⁴-α-D-Galactosyl-mannopentaose+Na]⁺ at m/z 1175 (measured m/z 1175.37109, theoretical m/z 1175.36955, mass error: 1.31 ppm), [2³-(4-O-Methyl-α-D-Glucuronyl)-xylotetraose+Na]⁺ at m/z (measured m/z 759.21649, theoretical m/z 759.21656, mass error: −0.09 ppm), [6³-α-D-Glucosyl-maltotriosyl-maltotriose+Na]⁺ at m/z (measured m/z 1175.36987, theoretical m/z 1175.36955, mass error: 0.27 ppm) and [2³,3³-di-α-L-Arabinofuranosyl-xylotriose+Na]⁺ at m/z (measured m/z 701.21124, theoretical m/z 701.21108, mass error: 0.23 ppm) (Figure 1). The protonated species of all five oligosaccharides were also detected but with intensity at 1 to 2 orders of magnitude less than the corresponding sodiated species (Figure 1).

Scheme 3.

Structure and chemical formula of five oligosaccharides detected by DPS-MS.

Figure 1.

DPS-MS spectra of a) 500 μM Heptasaccharide, b) 500 μM 6³, 6⁴-α-D-Galactosyl-mannopentaose, c) 500 μM 2³-(4-O-Methyl-α-D-Glucuronyl)-xylotetraose, d) 500 μM 6³-α-D-Glucosyl-maltotriosyl-maltotriose and e) 500 μM 2³,3³-di-α-L-Arabinofuranosyl-xylotriose in 50 mM Tris buffer.

The desalting power was further demonstrated by comparing DPS to conventional PSI and nano-ESI. Using 5 μM of 6³, 6⁴-α-D-Galactosyl-mannopentaose solution in 50 mM of Tris-HCl buffer (pH 7.0) as an example, the sodiated glycan ion at m/z 1175 was detected by DPS-MS with a signal intensity of 7.18E6 (Figure S-2a, Supporting Information). The signal intensity was reduced approximately 100-fold when PSI was used for detection (6.94E4, see Figure S-2b, Supporting Information), which was probably due to ion suppression by salts. By contrast, nano-ESI failed to detect galactosyl-mannopentaose from the sample (Figure S-2c, Supporting Information).

Detection of nucleotide sugars

Many chromatography-based methods such as ion exchange chromatography have been developed for the detection of nucleotide sugars.^70–72 A major limitation of such method is the ion suppression effect from buffer with elevated salt contents in the electrospray ionization.^{71, 73–74} In our previous work,⁶⁹ the detection of nucleotide sugars was not explored. In this study, we demonstrated DPS-MS analysis with uridine diphosphate-α-d-glucuronic acid (UDP-GlcA), a nucleotide sugar used as precursor for synthesis of many polysaccharides in animals and plants. A solution of 250 μM UDP-GlcA in 2.5 mM Tris-HCl buffer (pH 7.4) containing 50 μM MgCl₂ and 50 μM MnCl₂ was used to determine the capability of DPS to detect nucleotide sugars in non-volatile buffers. Ten μL of sample was loaded onto paper followed immediately by a desalting step before elution with 10 μL of ACN/H₂O/NH₄OH (90/10/1 v/v/v) as described in the Experimental Section. Ionization was in the negative ion mode at −3.5 kV. DPS was able to detect [UDP-GlcA-H]⁻ ion at m/z 579 (measured m/z 579.02732, theoretical m/z 579.02591, mass error 2.45 ppm; Figure 2c) with signal intensity 6.44E6. By comparison, nano-ESI detected UDP-GlcA at signal intensity 4.99E5 (Figure 2a) and the conventional paper spray method did not detect UDP-GlcA signal under the same conditions (Figure 2b).

Figure 2.

MS spectra of 250 μM UDP-glucuronic acid in 2.5 mM Tris-HCl buffer (pH 7.4), 50 μM MgCl₂ and 50 μM MnCl₂ acquired by a) nano-ESI, b) conventional paper spray and c) desalting paper spray.

Detection of glycopeptides

Vancomycin (VCM) is a class of glycopeptide antibiotics used to treat infections caused by bacteria.⁷⁵ Methods such as HPLC^76–77 and capillary electrophoresis⁷⁸ have been established and commonly used for the determination of VCM.⁷⁹ To identify a highly sensitive, low-cost method with high separation efficiency and short analysis time for glycopeptide detection, we evaluated the DPS-MS method for detecting intact VCM glycopeptide. Vancomycin hydrochloride solution at 5 μM was prepared in 50 mM Tris-HCl buffer (pH 7.0) for detection by three methods: DPS-MS, PSI and nano-ESI. [Vancomycin+H]⁺ was detected at m/z 1448.43518 by DPS-MS (Figure 3a) with a mass error of 1.58 ppm (theoretical m/z 1448.43748). With conventional PSI (Figure 3b), VCM was detected at m/z 1448.43652 (mass error 0.66 ppm), but with about a 5-fold reduction in intensity. Also, we observed salt clusters in the spectra, which might have suppressed the signal of the glycopeptide. On the other hand, VCM was not detected by nano-ESI at the level of 5 μM (Figure 3c). The identity of the VCM peak was further confirmed through MS/MS spectra using collision-induced dissociation (CID). VCM at m/z 1448 initially produced a fragment ion at m/z 1305. The fragment ion then produced a series of characteristic product ions at m/z 1143, 1115 and 1087. Fragmentation of the polypeptide occurred with the loss of a hexose residue (C₆H₁₀O₅, 162 Da) to produce a product ion at m/z 1143 and two consecutive CO losses to produce product ions at m/z 1115 and 1087 (Figure S3, Supporting Information). The observed pattern is consistent with the VCM fragmentation pathway.⁸⁰

Figure 3.

MS spectra of 5 μM vancomycin in 50 mM Tris buffer acquired by a) desalting paper spray, b) conventional paper spray and c) nano-ESI compared with d) isotope simulation generated by Xcalibur.

Sialylation, a derivative of neuraminic acid, is one of the most common types of glycosylation observed in glycoproteins and is known to play a key role in protein stability, signal mediation, and cell-to-cell interactions, including adhesion, migration, and immune recognition.^81–82 Additionally, sialic acid is known to be enriched in tumor cell surfaces to drive cancer malignancy.⁸³ A number of MS-based methods are widely used for detection of protein sialyation. However, in the most common shotgun proteomics strategy, MS is operated in data-dependent mode (DDA), which favors the selection of high-abundance precursor ions over low-abundance species such as sialyl-glycopeptides (SGP).⁸⁴ To overcome this limitation, various enrichment techniques have been developed.^85–86 In this experiment, we sought to test the capability of DPS-MS for direct detection of intact SGP. A chemically synthesized SGP with eleven sugar constituents was prepared at a concentration of 50 μM in 50 mM Tris-HCl buffer (pH 7.0). The sample was loaded directly onto a paper triangle for DPS-MS detection. Abundant fragments of SGP with a sodium adduct carrying two charges was detected at m/z 1444.58655 (theoretical m/z 1444.58446, mass error: 1.45 ppm, Figure 4a). The same sample was also examined by conventional PSI and nano-ESI (Figure 4b, c), showing a decrease by 4-fold and 20-fold in signal intensity, respectively (DPS-MS: 9.42E6, PSI: 2.08E6, nano-ESI: 4.16E5). The second most abundant fragment was [SGP+H]⁺ detected at m/z 2866 by PSI and DPS, but was completely undetectable by nano-ESI. In other words, through the desalting step, the DPS-MS method was proven to be more tolerant for SGP in the presence of salted buffer. The potential utility of DPS-MS as a powerful peptide profiling tool was also demonstrated in Figure 4a. Multiple fragmentations were observed in the DPS spectra at m/z 2575, 2413, 2208, 2046, 1755 and 1593, corresponding to the natural dissociation pattern of SGP.^87–88

Figure 4.

Panel a) showing the structure of sialyglycopeptide. MS spectra of 50 μM of sialylglycopeptide in 50 mM Tris buffer by b) desalting paper spray, c) conventional paper spray, and d) nano-ESI. Fragmentation ions of SGP are also labeled in b).

Detection of N-linked glycans released from glycoproteins

Glycosylation is a heterogeneous modification where a multitude of variables such as the type of host cell-line, cell culture conditions and genetic perturbations can influence glycan diversity during the development of monoclonal antibodies.^89–90 A conserved N-linked glycan at Asn²⁹⁷ of the single-chain Fc (scFc) region of IgG1 is critical for the stability, conformation, aggregation, and effector function of therapeutic antibodies.⁹¹ Removing glycans prior to mass spectrometric analysis is a commonly used approach for glycan analysis. N-Glycosidase F (PNGase F), a recombinant glycosidase from Elizabethkingia miricola, is one of the most effective enzymes at cleaving N-linked oligosaccharides from glycoproteins. A scheme of NIST IgG1κ monoclonal antibody with the N-glycosylation sites labeled (NISTmAb) is shown above (Scheme 2). N-linked oligosaccharides were released from NISTmAb by Rapid PNGase F treatment. The digested sample was loaded directly onto a paper triangle for fast detection and analysis by DPS-MS within a 5 minutes timeframe. In Figure 5, five major ions were observed at m/z 1444.5, 1485.5, 1467.6, 1809.6 and 1971.7, corresponding to G1F-N, G0F, G1F, G2F and G2FGal oligosaccharides⁹² that are typical major glycans in IgGs. The most abundant glycans observed were G0F (measured m/z 1485.53188, theoretical m/z 1485.53365, mass error −1.19 ppm), G1F (measured m/z 1647.58462, theoretical m/z 1647.58648, mass error −1.13 ppm), and G2F (measured m/z 1809.63772, theoretical m/z 1809.63930, mass error −0.87 ppm). Remarkably, a total of 26 glycans were detected by DPS-MS. Table 1 summarizes the theoretical and the experimental monoisotopic mass assignments of these N-linked glycans. Among these N-glycans, 23 native glycoforms with sodium adducts in the positive ion mode carry one charge and 3 containing sialic acid with hydrogen adducts in the positive ion mode carry two charges. DPS-MS spectra of all 26 glycans are listed in Table S1, Supporting information. In contrast, no glycans were detected by either PSI-MS or nano-ESI method (Figure S4, Supporting information). Glycosylation of NISTmAb was previously characterized by three laboratories using HILIC with fluorescence detection of 2-AB-labeled N-glycans and collectively found 24 glycan peaks.^93–94 In the NIST inter-laboratory study conducted in 2020, one hundred and three reports on glycan distributions of NISTmAb identified a total of 116 glycan compositions. The number of glycan compositions reported by each laboratory ranged from 4 to 48.⁹⁵ By eliminating the necessity of time-consuming enrichment and derivatization approaches that involve extensive conditioning, washing and equilibrating steps in traditional glycoproteomics,⁹⁶ DPS significantly reduces the analysis time and reagent cost for protein glycosylation analysis.

Figure 5.

DPS-MS spectra of N-glycans released from NIST IgG treated with Rapid PNGase F

Table 1.

Glycans observed from deglycosylated IgG mixtures as detected by DPS-MS

Composition	Symbol	Ion	Theoretical m/z	Measured m/z	Mass error (ppm)
HexNAc(1)		[M+Na]+	244.07916	244.07900	−0.655525035
HexNAc(2)		[M+Na]+	447.15853	447.15829	−0.536722401
HexNAc(1)Fuc(1)		[M+Na]+	390.13707	390.13683	−0.615168407
HexNAc(2)Fuc(1)		[M+Na]+	593.21644	593.21669	0.421431341
HexNAc(2)Hex(3)		[M+Na]+	933.31700	933.31629	−0.760727598
HexNAc(2)Hex(3)Fuc(1)		[M+Na]+	1079.37491	1079.37398	−0.8616098
HexNAc(2)Hex(4)		[M+Na]+	1095.36982	1095.36899	−0.757734954
HexNAc(2)Hex(5)	M5	[M+Na]+	1257.42265	1257.42253	−0.095433306
HexNAc(2)Hex(6)	M6	[M+Na]+	1419.47547	1419.47305	−1.704855104
HexNAc(3)Hex(3)	G0-N	[M+Na]+	1136.39637	1136.39585	−0.457586819
HexNAc(3)Hex(3)Fuc(1)	G0F-N	[M+Na]+	1282.45428	1282.45428	0
HexNAc(3)Hex(4)		[M+Na]+	1298.44920	1298.44998	0.600716609
HexNAc(3)Hex(4)Fuc(1)	G1F-N	[M+Na]+	1444.50711	1444.50511	−1.384555317
HexNAc(3)Hex(6)		[M+Na]+	1622.55484	1622.55640	0.961446702
HexNAc(4)Hex(3)	G0	[M+Na]+	1339.47575	1339.47436	−1.037719421
HexNAc(4)Hex(3)Fuc(1)	G0F	[M+Na]+	1485.53365	1485.53188	−1.191491017
HexNAc(4)Hex(4)	G1	[M+Na]+	1501.52857	1501.52669	−1.252057428
HexNAc(4)Hex(4)Fuc(1)	G1F	[M+Na]+	1647.58648	1647.58462	−1.128924049
HexNAc(4)Hex(5)	G2	[M+Na]+	1663.58139	1663.58245	0.637179525
HexNAc(4)Hex(5)Fuc(1)	G2F	[M+Na]+	1809.63930	1809.63772	−0.87310217
HexNAc(4)Hex(6)Fuc(1)	G2FGal1	[M+Na]+	1971.69212	1971.69410	1.004213579
HexNAc(5)Hex(4)Fuc(1)	G1FB	[M+Na]+	1850.66585	1850.66458	−0.686239496
HexNAc(5)Hex(5)Fuc(1)		[M+Na]+	2012.71867	2012.71727	−0.695576595
HexNAc(4)Hex(4)Fuc(1)NeuAc(1)	G1S1F	[M+2H]2+	949.84833	949.84898	0.68431978
HexNAc(4)Hex(5)Fuc(1)NeuAc(1)	G2S1F	[M+2H]2+	1030.87474	1030.87275	−1.93039942
HexNAc(4)Hex(5)Fuc(0)NeuAc(2)	G2S2	[M+2H]2+	1103.39350	1103.39495	1.314127734

DPS sensitivity and reproducibility of peptide analysis

To evaluate the detection sensitivity of DPS-MS, samples containing a constant concentration of Tris-HCl buffer (50 mM) was spiked with 0.1, 0.5, 5, 50, and 100 μM of VCM which was selected as a test sample. As the Tris-HCl concentration to VCM ratio increased a thousand times from 5 × 10² to 5 × 10⁵, the VCM signal at m/z 1448 was detected at intensities of 7.42E3, 3.00E4, 1.33E5, 1.75E6, and 3.17E6, respectively (Figure S5, Supporting information). Furthermore, a high correlation was observed when plotting the signal intensity of VCM against concentration (Figure S6, Supporting information). A linear trend with R² = 0.9973 indicates that DPS-MS can be applicable for relative peptide quantitative analysis.

Reproducibility and signal stability were also examined in this study. With the aid of a three-axis XYZ platform (picture shown above in Scheme 1b), we were able to reproduce the results with high accuracy. To assess repeatability, we evaluated 5 replicates of VCM in 50 mM Tris-HCl buffer at the levels of 5 μM and 50 μM (Figure S7, Supporting information) to show the analysis of intensity repeatability. Five replicates were taken on the same day. Standard deviation and coefficient of variation of signal intensity for 5 μM and 50 μM VCM was 1.54E4, 9.0% and 2.13E5, 9.6% respectively. Five replicates were within 2 standard deviations (95%) (Figure S7a and b, Supporting information). The boxes represent the inter-quartile range, while the whiskers represent the full range of observed values (Figure S7c and d, Supporting information).

Conclusions

In summary, we have demonstrated the application of DPS-MS to detect different types of complex glycans and glycoconjugates (e.g., glycopeptides, nucleotide sugars). In contrast to other common methods for carbohydrate analysis, DPS-MS is simple to implement and can achieve high-throughput analysis that is free from more complex and time-consuming enrichment methods. Not only is our method accurate and sensitive, but it also overcomes common limitations of conventional paper spray related to salt-induced signal suppression. By using our approach, we successfully detected oligosaccharides, nucleotide sugars, glycopeptides, sialylated glycopeptides and mAb N-linked glycans in complex matrices. We believe rapid analysis of post-translational modifications such as N-linked glycosylation of antibodies demonstrated that DPS-MS technique can also be extended to other glycoproteins. Further improvements in instrumentation, robustness as well as potential quantitation of glycopeptides with incorporation of internal standard bring promising foreground to this technique in biological and clinical applications in glycoproteomics.