Introduction

Glycosylation, one of the most important post-translational modification, plays key roles in protein stability, bioactivity, immunogenicity, and many other properties. Aberrant glycosylation is associated with various biological and pathological processes from cell adhesion, immune response, and signal transduction to cancers, heart failure, and congenital disorders [1–4]. Characterization of oligosaccharides is essential to understand the oligosaccharide functions in biological pathways and disease progression.

Because of the generation of simply charged ions and high tolerance against salts and buffers, matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) has become a powerful and widespread tool for oligosaccharide analysis [5]. However, sensitive analysis of native oligosaccharides by MALDI MS is often restricted by their low ionization efficiency [6, 7]. Derivatization is a common tool to overcome this problem, among which reductive amination is the most widely used method [8, 9]. However, the excess reductive reagents (e.g., sodium cyanoborohydride) seriously suppress the ion signals of analytes and therefore have to be removed prior to MALDI MS analysis [10]. To avoid the desalting process, which is tedious and inevitably leads to sample losses [11], several studies adopted non-reductive derivatization for oligosaccharide tagging using amino or hydrazine agents [12–16]. For instance, N-glycans can be detected in attomole levels by using 3-aminoquinoline as a derivative agent and co-matrix with cyano-4-hydroxycinnamic acid [17]. Although non-reductive derivatization is able to enhance oligosaccharide signals, the quantitative labeling is time-consuming and even difficult to achieve [15].

2-Hydrazinopyrimidine (2-HPM) possesses two electron-withdrawing imine nitrogens which are supposed to further increase the reactivity of the hydrazine group toward the reducing ends of oligosaccharides [18]. Meanwhile, the pyrimidine ring of 2-HPM contributes to the enhancement of MALDI MS signal of its target [19]. Hence, 2-HPM is proposed and explored as a derivative reagent for rapid and sensitive profiling of oligosaccharides.

We report here the optimization of reaction conditions during which the excess 2-HPM simultaneously acts as a co-matrix with 2,5-dihydroxybenzoic acid (DHB) to further increase the oligosaccharide signal of MALDI MS. Quantitative derivatization can be rapidly achieved within 15 min without the need for purification processes before and after derivatization. This method was comprehensively studied and proved to be ideal for the analysis of the oligosaccharides from glycopeptide, glycoprotein, and complex biological samples. The MALDI MS sensitivities of all the derivatives were increased over 10-fold in positive or negative mode. We also show that 2-HPM derivatives produce plenty of enhanced fragments which substantially facilitate the structural analysis of oligosaccharides.

Experimental

Chemicals and reagents

Maltoheptaose (DP7), maltodextrins, 2,5-dihydroxybenzoic acid (DHB), 2-aminobenzoic acid (2-AA), 2-aminobenzamide (2-AB), 2-aminopyridine (2-AP), 2-hydrazinopyridine (2-HP), RNase B, chicken ovalbumin, and bovine fetuin were purchased from Sigma-Aldrich (St. Louis, MO, USA). Peptide N-glycosidase (PNGase F) was obtained from New England Biolabs (Ipswich, MA, USA). HPLC-grade methanol, acetonitrile (ACN), acetic acid, and 2-hydrazinopyrimidine (2-HPM) were from Fisher Scientific (Fairlawn, NJ, USA). Nonporous graphitized carbon (Carbograph) solid-phase extraction (SPE) columns (150 mg/4 mL) were from Alltech Associates (Deerfield, IL, USA).

Optimization of reaction conditions

DP7 was employed as an oligosaccharide standard for the optimization of derivatization conditions, including acetic acid percentage, 2-HPM concentration, solvent type, and reaction temperature. Initially, 10 pmol of DP7 was dissolved in 10 μL of methanol–acetic acid solution containing 5 mg/mL of 2-HPM, and the reaction was carried out at 60 °C for 15 min. During the optimization, one parameter was varied while all other parameters were kept constant at their optimum. Derivatization rate was calculated by dividing peak intensity of 2-HPM-derivatized DP7 by peak intensities of both underivatized and 2-HPM-derivatized DP7. Average values and standard deviations were determined for each condition. For comparison, underivatized and an equal amount of derivatized oligosaccharides were dissolved in identical solvent and subjected to MALDI MS.

Sample preparation and derivatization

Sialyglycopeptides (SGP) were purified from egg yolks according to our previous report [20]. Human serum samples were collected from a healthy volunteer. Briefly, after collection of the whole blood, the blood was allowed to clot by leaving it undisturbed at room temperature for 30 min, and the clot was then removed by centrifuging at 1500×g for 10 min in a refrigerated centrifuge. The resulting supernatant is designated serum.

For the preparation of standard N-glycans from glycoproteins or glycopeptides, RNase B (50 μg), OVA (50 μg), SGP (10 μg), and bovine fetuin (50 μg) were dissolved in 100 μL of 100 mM ammonium bicarbonate (ABC) buffer (pH 8.0), respectively. After the addition of 10 U of PNGase F to each sample, the mixture was incubated at 37 °C overnight. The buffer of collected human serum (100 μL) was changed to 100 μL of 100 mM ABC containing 1 % SDS and 20 mM DTT by using an Amicon Ultra (0.5 mL, 3 K cutoff) centrifuge tube, and then the sample was denatured at 100 °C for 5 min. The solution was cooled to room temperature and 10 U of PNGase F was added; the reaction was then carried out overnight at 37 °C.

For derivatization, 10 μL of RNase B, OVA, SGP, bovine fetuin, and human serum obtained from their PNGase F digestions were dried using a SpeedVac and then incubated with 10 μL of derivatization solution under the optimized conditions, respectively. The derivatized samples were centrifuged at 10,000×g for 2 min, and the supernatants were ready for MALDI analysis.

MALDI mass spectrometric measurement

One microliter of sample solution was deposited on the target plate (Bruker Daltonics, MTP 384 polished steel) with the same volume of DHB (50 mg/mL in 50:50 ACN–water) by a dried-droplet method for MALDI MS analysis. MS spectra were recorded on an ultrafleXtreme MALDI TOF-TOF mass spectrometer (Bruker Daltonics, Bremen, Germany) equipped with 1 kHz Smartbeam-II laser in reflectron mode, operated through an autoXecute method in flexControl (v3.4, Bruker Daltonics) and processed in flexAnalysis. External calibration was performed using maltodextrins. For positive mode, operation parameters were ion source 1, 25.15 kV; ion source 2, 22.75 kV; lens, 8.32 kV; reflectron 1, 26.66 kV; and reflectron 2, 13.48 kV; detector gain, ×4.0. Negative mode operation parameters were ion source 1, 20.00 kV; ion source 2, 17.95 kV; lens 6.50 kV; reflectron 1, 10.75 kV; and reflectron 2, 19.00 kV; detector gain, ×4.0. A total of 1000 shots were accumulated for each spectrum in positive or negative mode. LIFT MS/MS operation parameters were ion source 1, 7.50 kV; ion source 2, 6.75 kV; lens 3.50 kV; reflectron 1, 29.50 kV; reflectron 2, 13.95 kV; LIFT 1, 19.00 kV; LIFT 2, 3.40 kV; detector gain, ×10.0. The isolation window for precursor ions was set to 0.5 % of the precursor m/z value, and 1000 shots were accumulated for each spectrum in LIFT mode. Resulting spectra were interpreted manually assisted by the GlycoWorkbench software (Euro-CarbDB) [21].

Results and discussion

Optimization of derivatization conditions

The hydrazone bond formation between the hydrazine group of 2-HPM and the reducing end of oligosaccharides is illustrated in Scheme 1. We optimized the derivatization conditions by using DP7 as a model oligosaccharide. As shown in Fig. 1, each average and standard deviation was calculated from three repeat experiments. Hydrazone formation is catalyzed by a weakly acidic environment, and the use of 10 % acetic acid proved to be optimal. It was reported that high acidity can induce the hydrolysis of hydrazine/hydrazide derivatives of oligosaccharides [22, 23]. Here, slight hydrolysis of the DP7 derivative was also noted with the treatment of 20 % acetic acid (Fig. 1a). With the concentration of 5 mg/mL of 2-HPM, a quantitative derivatization was obtained (Fig. 1b). Interestingly, when the concentration of 2-HPM was raised to 10 mg/mL, the sensitivity for the DP7 derivative was maximal (Figs. 1b, 2). Excess 2-HPM forms a binary matrix with DHB, resulting in a more homogenous crystallization which will be completely formed within 5 min at room temperature or rapidly at an elevated temperature. Among different organic solvents tested, methanol was favored as it promoted the best crystallization behavior (Fig. 1c). As shown in Fig. 1d, reaction temperature is another important parameter, and 60–70 °C is optimal for the derivatization. Because the pyrimidine group of 2-HPM functions as a better electron donor for the nucleophilic activity of the hydrazine group, the derivatization of a reducing oligosaccharide with 2-HPM is much faster than traditional reductive amination, non-reductive amination, and hydrazone formation which always require 1–3 h to achieve complete tagging [15, 24, 25]. As shown in Fig. S1 in the Electronic Supplementary Material (ESM), DP7 can be completely converted to the 2-HPM derivative within 15 min. In contrast, the derivatization rate of 2-AA, 2-AB, 2-AP, or 2-HP with DP7 was quite inefficient even after a 30-min reaction time. Therefore, 2-HPM derivatization is simple and rapid under the optimized conditions of 10 mg/mL of 2-HPM in a methanol–acetic acid (90:10, v/v) solution at 65 °C for 15 min.

Scheme 1
scheme 1

2-HPM derivatization of oligosaccharides

Full size image
Fig. 1
figure 1

Derivatization rates obtained by variation of reaction conditions. Influential parameters were percentage of acetic acid (a), 2-HPM concentration (b), type of organic solvent (c), and reaction temperature (d). Error bars are given as the standard deviation (n = 3)

Full size image
Fig. 2
figure 2

MALDI-TOF mass spectra of a DP7 derivatized with 0 mg/mL of 2-HPM, b DP7 derivatized with 5 mg/mL of 2-HPM, and c DP7 derivatized with 10 mg/mL of 2-HPM in positive ion mode. The amount of DP7 is 1 pmol per spot

Full size image

Enhanced detection of neutral oligosaccharides by 2-HPM derivatization

It was reported that the pyrimidine ring of 2-HPM can improve the MALDI MS sensitivities of its derivatives [19]. We compared the signal intensities as well as signal/noise (S/N) ratios of DP7 in positive mode before and after derivatization. The predicted increase in molecular weight of 92 Da was observed, and no signals of the underivatized oligosaccharides remained, demonstrating complete conversion of DP7 into 2-HPM-DP7 (Fig. 2b, c). The signal as well as S/N ratio of DP7 was significantly increased after derivatization (Fig. 2a, b). As shown in Fig. 2c, the excess of 2-HPM facilitates the crystallization and ionization of its derivative. So the desalting step prior to MS analysis is not required. The reacted DP7 can immediately be subjected to MALDI analysis or stored at −20 °C until use, benefiting from its excellent stability and signal reproducibility (ESM Fig. S2). To further confirm whether 2-HPM can act as co-matrix, we purified and analyzed native and derivatized DP7, and only a 2-fold increase of signal intensity was observed in the derivatized sample (ESM Fig. S3). We treated the reduced DP7 with and without 2-HPM, and the signal intensity was raised over 5-fold in the 2-HPM-treated sample (ESM Fig. S4). These results indicate the dual roles of 2-HPM as a derivative agent and co-matrix for improved analysis of oligosaccharides. In addition, 2-HPM significantly enhances the sample homogeneity, indicating the promising application of 2-HPM derivatization in quantitative profiling of oligosaccharides. More than 10 mg/mL of 2-HPM showed no further improvement of detection, but crystallization was impaired. Both the signal and S/N ratio were increased over 10-fold using 2-HPM as a derivative agent as well as co-matrix. [M+Na]+ was the dominant ion of oligosaccharide derivatives in positive mode (Fig. 2b, c), resulting in relative simplicity of spectral interpretation. A mass spectrum can be acquired from about 10 fmol of sample with an S/N ratio greater than 3 (Fig. 3b). The curves of sample amount versus signal intensity and derivative agent amount versus signal intensity are presented in Fig. S5 in the ESM. Compared to its native form (Fig. 3a), the detection limitation of DP7 was increased by 10-fold after derivatization.

Fig. 3
figure 3

MALDI-TOF mass spectra of a 100 fmol native DP7 and b 10 fmol 2-HPM-DP7 in positive ion mode

Full size image

To demonstrate the feasibility of our method on complex mixtures, neutral N-linked glycans released from two glycoproteins of RNase B and ovalbumin were subjected to 2-HPM derivatization, respectively. Initially, N-glycans released from 0.5 μg of RNase B or ovalbumin without any purification were directly analyzed by MALDI MS. Only 4 N-glycans of RNase B and 7 N-glycans of ovalbumin were observed with quite low intensities and S/N ratios, and some obvious peaks of impurities from PNGase F digestion were detected in these two samples (Fig. 4a, c). However, compared with the underivatized samples, 5 N-glycan derivatives of RNase B and 12 N-glycan derivatives of ovalbumin were detected, and the signal intensities as well as S/N ratios of 2-HPM-labeled N-glycans were both improved over 10-fold (Fig. 4b, d and ESM Table S1–2). Moreover, signals of impurities were suppressed after 2-HPM labeling, indicating that 2-HPM can selectively ionize oligosaccharides and suppress the signal of other molecules. To test whether the 2-HPM method is still applicable to complex samples, human serum were treated with PNGase F, dried, and subjected to derivatization. Owing to the enhancement of ionization efficiency of glycans and the suppression of endogenous impurities by 2-HPM derivatization, a total of 24 glycans were successfully identified (ESM Fig. S6) in positive mode from 1 μL of serum sample. The compositions and proposed structures were determined from the literature and checked with Glycoworkbench (ESM Table S3) [26, 27]. No purification or enrichment before and after derivatization is necessary. Therefore, 2-HPM derivatization is an ideal method for rapid and sensitive analysis of oligosaccharides from biological samples.

Fig. 4
figure 4

MALDI-TOF analysis of a underivatized and b derivatized N-linked glycans from RNase B; c underivatized and d derivatized N-linked glycans from ovalbumin in positive mode. All the marked ions are [M+Na]+. Green circles, yellow circles, and blue squares represent mannose, galactose, and GlcNAc residues, respectively

Full size image

Enhanced detection of sialylated oligosaccharides by 2-HPM derivatization

Particular attention is being paid to sialylated glycans owing to their involvement in many important biological processes, including cell–cell adhesion and cell–pathogen interaction [28, 29]. The ionic signal suppression and the labile nature of sialic acids are the major issues in the analysis of sialylated glycans by MS [30]. To explore the applicability of 2-HPM derivatization to acidic oligosaccharides, we released and derivatized the sialylated N-glycans from 0.1 μg of SGP which are glycopeptides extracted from egg yolk. As shown in Fig.  5a, three [M−H]− ions of sialylated N-glycans were detected by MALDI in negative mode. Because sialic acids are labile and easy to cleave from glycans, desialylation often happens during glycan derivatization. To test the stabilities of sialylated glycans during 2-HPM labeling, we treated SGP N-glycans under the optimized conditions with (bottom panel, Fig.  5a) and without (middle panel, Fig.  5a) the addition of 2-HPM. Compared to control (top panel, Fig.  5a), there is no desialylation observed during the derivatization, and both the signals and S/N ratios of sialylated glycans were increased over 10-fold after derivatization. We also tested the feasibility of our method by using sialylated N-glycans released from bovine fetuin. Significant increase of signal sensitivities and S/N ratios (greater than 10-fold) of glycans was observed in the 2-HPM-derivatized sample (Fig.  5b). Thus, this method is suitable for the analysis of sialylated oligosaccharides by enhancing its signal sensitivities as well as S/N ratios.

Fig. 5
figure 5

MALDI-TOF analysis of a SGP and b bovine fetuin N-glycans before and after 2-HPM derivatization. Top spectra native N-glycans dissolved in methanol–acetic acid (90:10, v/v); middle spectra native N-glycans dissolved in methanol–acetic acid (90:10, v/v) and incubated for 15 min at 65 °C; bottom spectra 2-HPM-derivatized N-glycans. Purple diamonds represent sialic acid residues

Full size image

MS/MS analysis of derivatized oligosaccharides

The emerging glycomics projects aim to characterize all forms of oligosaccharides from different sources. Tandem mass spectrometry (MS/MS) is a key experimental methodology for high-throughput identification and characterization of oligosaccharides [31, 32]. Derivatization affords enhanced signals and diverse fragments of oligosaccharides analyzed by MALDI MS/MS, facilitating the structural determination of oligosaccharides [33]. Herein, we chose DP7 as a standard oligosaccharide, and Fig. 6a shows the MS2 spectrum of underivatized DP7. Fragments are designated according to the nomenclature proposed by Domon and Costello [34]. As cleavage products from both the reducing and the non-reducing end are able to be charged during the fragmentation process, it is difficult to distinguish B/Y fragments from C/Z fragments resulting in a complication of oligosaccharide characterization. After derivatization, the corresponding precursor ion [M+Na]+ generates a highly abundant 0,2A7-type ion at m/z 1115.92, indicating the close-ring formation of derivatives and immediate loss of tag (Fig. 6b) [15]. The B-type and Y-type ions of derivatized DP7 were easy to recognize with higher signal sensitivity than underivatized DP7. Moreover, a series of cross-ring fragments from 0,2A7-type ions were observed in Fig. 6b. The cross-ring ions resulting from internal cleavages provide additional information that is important to reveal the type of linkage between monosaccharides. Additionally, we demonstrated the MS/MS analysis of neutral N-glycan of ovalbumin; the enhanced B-, Y-, and A-type ions provide distinct and sufficient information for the determination of glycan composition and sequence (ESM Fig. S7). However, 0,2A7-type ions were only observed in GlcNAc moieties of the reducing end but not interchain monosaccharides of the N-glycan.

Fig. 6
figure 6

MALDI-TOF MS/MS analysis of a native DP7 ([M+Na]+ precursor, m/z 1175.38) and b 2-HPM derivatized DP7 ([M+Na]+ precursor, m/z 1267.60)

Full size image

For the MS/MS analysis of sialylated oligosaccharides with 2-HPM derivatization, the [M−H]− ion at m/z 1860.75 was presented as a model. The underivatized sialylated glycans show insufficient fragments and low signal intensities (Fig. 7a). On the contrary, fragmentation of 2-HPM-labeled glycans benefits from its enhanced ionization, and thereby produced plenty of fragments including Y- and B-type ions (Fig. 7b). Both signal intensities and S/N ratios were increased approximately 10-fold after derivatization. Therefore, 2-HPM derivatization allows more efficient structural interpretation of both neutral and sialylated oligosaccharides through abundantly enhanced and unambiguous fragments.

Fig. 7
figure 7

MALDI-TOFMS/MS analysis of a underivatized SGP N-glycan ([M−H]− precursor, m/z 1768.66) and b derivatized SGP N-glycan ([M−H]− precursor, m/z 1860.75) with a composition of (Hex)4(HexNAc)4(NeuAc)1

Full size image

Conclusions

We developed a rapid and sensitive approach for improved analysis of oligosaccharides with MALDI-TOF MS using 2-HPM as a derivative agent and a co-matrix. The use of toxic reagents for reduction is avoided. The sample and time losses caused by tedious enrichment and purification were minimized. This method is ideal for the analysis of various oligosaccharides released from glycopeptides, glycoproteins, and biological samples. After derivatization, significant increase of MALDI signal sensitivities and S/N ratios was observed for all the tested neutral and sialylated oligosaccharides. Tandem mass spectra of derivatized oligosaccharides also provided plentiful and enhanced fragments, facilitating spectrum interpretation for oligosaccharide characterization. The various possibilities offered by this new derivatization approach should be of great benefit for oligosaccharide analysis. Moreover, this rapid and sensitive method has potential for high-throughput analysis of glycosylation in the study of glycan-related biomarkers, and thereby provides a better understanding of the structure–function relationships of oligosaccharides.