Analysis of sulfated sialic acids by liquid chromatography coupled with high-resolution mass spectrometry and fluorescence detection

Highlight

• •No method of modified sialic acids analysis is established except O-acetylated one by LC-MS.
• •The DMB-LC-MS platform for O-acetylated sialic acids is applied to O-sulfated ones.
• •8- and 9-O-sulfated sialyl isomers were successfully differentiated on EAD-analysis.
• •The presence of a novel sulfated and acetylated Neu5Gc was first demonstrated.

Abstract

Sialic acid is localized at the outermost glycan, and involved in regulatory roles as one of the most important glycotopes in various biological phenomena. Recently, we established a highly sensitive and selective method for the fine differentiation of 1,2-diamine-4,5-methylenedioxybenzene (DMB)-labeled sialic acids, including N-acetylneuraminic acid (Neu5Ac), N-glycolylneuraminic acid (Neu5Gc), deaminoneuraminic acid (Kdn), and their O-acetyl derivatives. This approach enables unambiguous identification of these three sialic acids based on characteristic C-5 diagnostic fragment ions obtained by high-resolution accurate-mass spectrometry and fluorescence detection. In the present study, we extended this analytical platform to sulfated sialic acids (SiaSs) to evaluate its applicability to sialic acids bearing an additional negatively charged sulfate group. We first assessed the liquid chromatographic separation of DMB-labeled SiaSs using authentic standards. Using a CAPCELL PAK ADME column, we successfully separated six authentic SiaS standards and three previously unreported SiaS species in sea urchins. We identified multiple O-sulfated substitutions on Neu5Ac and Neu5Gc species, which were further confirmed by characteristic collision-induced dissociation (CID) fragmentation. In addition, electron-activated dissociation (EAD) was employed to evaluate the fragmentation behavior of SiaS species. EAD analyses efficiently generated characteristic diagnostic ions that enable reliable discrimination of the sulfation site, allowing clear differentiation between 8-O- and 9-O-sulfated isomers. Liquid chromatography-based fragment analysis led to the identification of a novel O-sulfated and O-acetylated sialic acid, Neu5Gc8SOAc1, which had never been reported before. These findings underscore the complementary nature of fluorescence detection and high-resolution mass spectrometry techniques for the identification and quantification of DMB-labeled sialic acids.

1. Introduction

Sialic acids are negatively charged nine-carbon monosaccharides found at the non-reducing termini of glycans of glycoconjugates, mainly in deuterostomes and pathogenic microorganisms [1,2]. Depending on the substituents at the C5 position, three forms of sialic acid occur in nature: N-acetylneuraminic acid (Neu5Ac), N-glycolylneuraminic acid (Neu5Gc), and deaminoneuraminic acid (Kdn) [[1], [2], [3], [4]]. These forms also exhibit further diverse structures resulting from differences in modification (O-acetylation, O-sulfation, O-methylation, O-lactylation, O-phosphorylation, and so on) (Fig. 1A), linkage (α2,3, α2,4, α2,5, α2,6, α2,8, and α2,9), and degree of polymerization (DP, 2 to 400) [[5], [6], [7]]. Thus, the number of reported modified sialic acids exceeds 80 [1]. Of these diverse structures, very little is known about the occurrence, metabolism, and functions of modified sialic acids, with the exception of O-acetylation and O-sulfation.

Fig. 1.

Structure of sialic acids. (A) Presence of sialic acids and modified sialic acids on glycoconjugates (glycoproteins and glycolipids). Sialic acid and modified sialic acids on N-glycans, O-glycans, and glycolipids are shown. Blue star, modification; Red diamond, sialic acid; Yellow circle, galactose; Blue square, N-acetylglucosamine; Green circle, mannose; Red triangle, fucose. (B) Structure of sialic acid and SiaS. Three species of sialic acid—N-acetylneuraminic acid (Neu5Ac), N-glycolylneuraminic acid (Neu5Gc), and deaminoneuraminic acid (Kdn)—are shown with sulfated substitutions at C-8 (8-O-sialic acid, Sia8S) and C-9 (9-O-sialic acid, Sia9S). (C) Scheme of sulfated sialic acid (SiaS) derivatization.

O-Acetylated sialic acids are well-studied in humans and other vertebrates [8] and are involved as ligands in the infection of influenza virus group C [9]. Since transgenic mice of O-acetyl-esterase showed an embryonic lethal phenotype [10], the O-acetylation/de-O-acetylation system is considered important during development, although the underlying mechanism remains unknown. For structural and functional studies of O-acetylation of sialic acids, a key issue is to identify the synthetic enzyme gene or the sialate: O-acetyltransferase gene. In this regard, two enzyme genes, CASD1 (capsule structural domain containing 1) and NXPE1 (neurexophilin and PC-esterase domain family member 1), have recently been identified [[11], [12], [13]], representing a significant step toward understanding the functional importance of O-acetylation of sialic acid. Another key issue is to establish methods for accurately determining the diverse structures of O-acetylated sialic acids, in which mono-, di-, and tri-O-acetylation occur at the 4-, 7-, 8-, and 9-O-positions. For example, hydrolysis under milder conditions was introduced to efficiently release free sialic acid without removing the O-acetyl group, because the O-acetyl group is sensitive to alkaline and strong acid conditions [14]. In addition, we recently established an analytical protocol using an online liquid chromatography–mass spectrometry (LC-MS) system [15].

O-Sulfated sialic acids (SiaSs; Fig. 1B) are unique modified sialic acids whose biological functions have been well analyzed. SiaSs are identified in glycolipids [[16], [17], [18]] and glycoproteins [19,20] of sea urchin gametes, and their importance in sperm–egg interaction [21] and sperm motility [20] has been demonstrated. The occurrence of SiaSs using a specific probe for SiaSs [22] is shown to be ubiquitous in mammals [23], including bovine gastric mucosa [24,25]. For the synthetic enzyme, no sialic acid-modifying enzymes other than the aforementioned sialate: O-acetyltransferases had been identified until recently. In 2022, we first discovered the sialate: O-sulfotransferases, Wscd 1 and Wscd 2, in vertebrates [23], based on the homology of the sea urchin genome. To detect and understand the occurrence of these enzymes, we first developed an 8-O-sulfated Neu5Ac (Neu5Ac8S)-specific monoclonal antibody, 3G9, using sea urchin sperm as an immunogen [22,26,27]. This antibody was then used as a probe for detecting SiaS expression on cells transfected with the sialate: O-sulfotransferase genes. However, we do not know if 3G9 recognizes not only Neu5Ac8S but also Neu5Ac9S and other SiaSs, such as Neu5Ac4S, Neu5Gc8S, and Neu5Gc9S, because authentic SiaS glycosides are not available. SiaSs potentially have diverse structures, similar to O-acetylated Sias, and require appropriate methods to accurately determine these structures. We previously demonstrated that 8-O- and 9-O-sulfated Neu5Ac, Neu5Gc, and Kdn could be separated using fluorometric ultra-high-performance liquid chromatography (UHPLC) after derivatizing their compounds with the DMB fluorescent dye [26] (Fig. 1C). While the co-injection method can confirm which peak corresponds to which authentic SiaS sample, it is difficult to identify if a peak represents other types of SiaSs, e.g., 4-O-, 7-O-sulfated, or 8,9-di-O-sulfated, because those authentic sialic acids are not available.

To solve these problems, we explored the advanced use of fragmentation in DMB-SiaS analysis with the aim of discovering a potential tool for identifying the detailed structural characteristics of various sialic acids. We compared high-resolution mass spectrometry (HRMS) with CID- and EAD-fragmentation on DMB-SiaS to confirm its capabilities in sialomics analysis. The ZenoTOF7600 is the quadrupole time-of-flight (Q-TOF) instrument equipped with EAD cell capable for fast electron-based dissociation [28,29]. EAD analysis has been reported to provide additional fragmentation patterns from N-acetylglucosamine which were not observed by CID analysis, demonstrating its potential to give detailed structural information for monosaccharides [30].

The early-stage application of EAD revealed some limitations, making it difficult to discriminate between 8-O-substituted SiaS and 9-O-substituted SiaS. However, we have developed an improved workflow for SiaS analysis, combining high-resolution column separation, fluorescence detection with DMB, and the EAD system. This approach has improved our understanding of structural features and increased the reliability of structural identification in sialic acid analysis.

2. Results and discussion

2.1. Development of the liquid chromatography–fluorescence detection–mass spectrometry (LC-FLD-MS) analysis method for SiaSs using standards

Various analytical methods have been developed to detect and quantify sialic acid species. Among these, UHPLC following derivatization with DMB [31] has been widely used as one of the most effective methods for sialic acid analysis, owing to its high sensitivity when coupled with either fluorescent detection (FLD) [26] or mass spectrometry (MS) [15]. However, both detection methods have several limitations. Although FLD provides excellent sensitivity for DMB-labeled sialic acids, complete chromatographic separation of all molecular species using a single LC method remains challenging. Moreover, the identification of newly discovered or structurally modified sialic acids is difficult in the absence of authentic reference standards.

With respect to MS-based detection, detailed structural information, such as the position of modifications, is often difficult to obtain solely from conventional fragmentation analysis because DMB-labeled sialic acids exhibit highly similar fragmentation patterns dominated by the DMB core structure [15]. In our previous study, we therefore focused on the fine structural characterization of these species and developed a highly sensitive, high-resolution online dual-detection LC-FLD-MS platform for the analysis of O-acetylated sialic acid species. Using this approach, we identified characteristic diagnostic fragment ions that enable discrimination of C-5-dependent sialic acid structures, including Neu5Ac, Neu5Gc, and Kdn. We further identified several characteristic ions that facilitated the structural assignment of individual O-acetylated sialic acid species[15]. In the present study, we aimed to extend this analytical platform to the comprehensive detection and structural characterization of SiaSs. We first optimized the liquid chromatographic conditions for the separation of DMB-labeled SiaS standards. A panel of reference compounds, including Neu5Ac, Neu5Ac8S, Neu5Ac9S, Neu5Gc, Neu5Gc8S, Neu5Gc9S, Kdn, Kdn8S, and Kdn9S, was prepared by enzymatic or chemical synthesis as previously described [26] (Fig. 1B). These standards were subsequently derivatized with DMB, which selectively labels α-keto acid groups, prior to LC-FLD-MS analysis (Fig. 1C).

DMB-labeled SiaSs exhibited markedly reduced sensitivity in positive ion mode compared with non-SiaSs, due to the strong negative charge of the sulfate ester. Furthermore, during positive-mode electrospray ionization, sulfate-containing species were prone to in-source loss of the sulfate group, resulting in the formation of desulfated sialic acid ions (data not shown). The concomitant generation of these fragment ions reduced the abundance of precursor ions and consequently compromised the analytical sensitivity. All target species, including non-SiaS, were stably ionized and detected in negative-ion mode (data not shown). Therefore, negative-ion mode analysis was initially employed for the detection of both sulfated and non-SiaS. Each of the nine sialic acids was analyzed at 10 pmol per injection. All compounds were successfully resolved under the optimized chromatographic conditions using a CAPCELL PAK ADME column (3 μm, 2.1 mm i.d. × 150 mm) at 30°C, with 0.1% formic acid in water (solvent A) and acetonitrile/methanol (50:50, v/v; solvent B) delivered at 0.25 mL/min (Fig. 2). These conditions were carefully optimized to allow separation of closely related sialic acid species, including sulfated forms. By contrast, a conventional C18 column did not provide clear separation of SiaSs under the tested conditions (Fig. S1). Furthermore, the precursor ions of all DMB-labeled standard compounds were readily detected by full-scan MS. The extracted ion chromatograms (XICs) exhibited complete chromatographic resolution of each analyte, consistent with the corresponding fluorescence detection profiles (Fig. 2). Tandem mass spectrometry (MS/MS) analysis in negative-ion mode revealed characteristic fragmentation patterns for the DMB-labeled sialic acid species (Fig. 3). Based on the proposed fragment ion structures, product ions corresponding to the cyclic dehydrated species commonly observed under positive-ion CID conditions were not detected under negative-ion CID conditions. A fragment ion at m/z 203.05 [C₁₀H₇N₂O₃]^- was observed in all DMB-sialic acid species analyzed in this study (Table 1). This fragment appears to be common to both positive- and negative-ion modes. It is likely generated by cleavage of the C3-C4 bond of the DMB-labeled sialic acid, while the C1–C2–C3 moiety, including the α-keto carbon atoms (C1 and C2), remains attached to the DMB backbone. The resulting fragment retains the aromatic framework and is further stabilized by resonance, giving rise to the ion at m/z 203.0462. The corresponding positive-ion species is observed at m/z 203.0451 [C₁₀H₇N₂O₃]⁺, indicating that this fragment is a stable structural motif formed irrespective of the ionization polarity. Among the species-specific fragment ions reflecting differences in the C5 substituent, major product ions at m/z 286.08 [C₁₄H₁₂N₃O₄]^- and 100.04 [C₆H₆NO₂]^- were consistently observed for DMB-Neu5Ac and its sulfated derivatives. The identical fragmentation patterns obtained for Neu5Ac8S and Neu5Ac9S indicate that these ions retain the C5 moiety of Neu5Ac, consistent with the proposed fragment ions (Fig. 3A–C). Similar trends were observed for the Neu5Gc and Kdn derivatives. DMB-Neu5Gc and its sulfated derivatives produced characteristic fragment ions at m/z 302.08 [C₁₄H₁₂N₃O₅]^-and 116.04 [C₆H₆NO₃]^- (Fig. 3D–F), whereas DMB-Kdn and its sulfated derivatives generated a characteristic ion at m/z 245.06 [C₁₂H₉N₂O₄]^- (Fig. 3G–I). Comparison of the MS/MS spectra of the 8-O- and 9-O-SiaS derivatives revealed no unique product ions that directly distinguished the positional isomers. However, the relative abundances of several fragment ions differed reproducibly between the two isomers. To evaluate these differences quantitatively, the intensities of diagnostic fragment ions were normalized to the common product ion at m/z 203.05, which was consistently detected in all SiaS derivatives (Fig. 3). This normalization clearly highlighted isomer-dependent fragmentation patterns (Table 1).

Fig. 2.

Liquid chromatography-mass spectrometry (LC-MS) chromatograms and negative ion mode mass spectra of standard samples. All standard samples were derivatized by 1,2-diamino-4,5-methylenedioxybenzene (DMB), and 10 pmol of each sample was injected into the LC-MS system. Samples were separated with a CAPCELL PAK ADME column (3 μm particle size, 2.1 mm internal diameter × 150 mm) using a Nexera ultra-high performance liquid chromatography (UHPLC) platform interfaced with a ZenoTOF 7600 mass spectrometer, as described in Materials and Methods. (A) Fluorescence detection was performed with an RF-20A XS fluorescence detector (Ex: 373 nm, Em: 448 nm). Extracted ion chromatograms (XICs) of each sample were acquired on the ZenoTOF 7600 instrument (B–G). (B) Neu5Ac (m/z: 424.136). (C) Neu5AcS1 (m/z: 504.093). (D) Neu5Gc (m/z: 440.131). (E) Neu5GcS1 (m/z: 520.088). (F) Kdn (m/z: 383.110). (G) KdnS1 (m/z: 463.066). Peak1: Kdn9S, Peak2: Kdn8S, Peak3: Kdn, Peak4: Neu5Gc9S, Peak5: Neu5Gc8S, Peak6: Neu5Ac9S, Peak7: Neu5Gc, Peak8: Neu5Ac8S, Peak9: Neu5Ac.

Fig. 3.

Negative ion mode liquid chromatography-tandem mass spectrometry (LC-MS/MS) fragmentation spectrum of sialic acids from standard samples. All standard samples were derivatized by DMB, and 10 pmol of each sample was injected into the LC-MS system. Samples were separated with a CAPCELL PAK ADME column (3 μm particle size, 2.1 mm internal diameter × 150 mm) using a Nexera UHPLC platform interfaced with a ZenoTOF 7600 mass spectrometer, as described in Materials and Methods. Yellow triangles indicate the presence of sulfate ions. Purple, blue, and green indicate the presence of ions at the 5-position: purple for Neu5Ac and Neu5AcS1; blue for Neu5Gc and Neu5GcS1; and green for Kdn and KdnS1. (A) MS/MS spectrum of Neu5Ac. (B) MS/MS spectrum of Neu5Ac8S. (C) MS/MS spectrum of Neu5Ac9S. (D) MS/MS spectrum of Neu5Gc. (E) MS/MS spectrum of Neu5Gc8S. (F) MS/MS spectrum of Neu5Gc9S. (G) MS/MS spectrum of Kdn. (H) MS/MS spectrum of Kdn8S. (I) MS/MS spectrum of Kdn9S.

Table 1.

Retention times and m/z values of DMB-labeled sialic acids (Sia and SiaS) in negative ion mode.

Sialic acid	RRT* (min)	Precursor ion	Characteristic fragment ions
		[M-H]-m/z	m/z	Molecular formula	Relative intensity‡ (%)	Type
Neu5Ac	1.00	424.14	286.0833	C14H12N3O4-	120	C5†, D#
			203.0462	C10H7N2O3-	100	D
			100.0404	C6H6NO2-	68	C5
			142.0510	C6H8NO3-	8.2	C5
			160.0615	C6H10NO4-	6.9	C5
Neu5Ac9S	0.73	504.09	286.0833	C14H12N3O4-	130	C5, D
			424.1351	C18H22N3O9-	110	C5, D
			203.0462	C10H7N2O3-	100	D
			100.0404	C6H6NO2-	80	C5
			300.0395	C8H14NO9S-	46	C5, S
			96.9601	HSO4-	32	S$
			198.9918	C4H7O7S-	20	S
			160.0615	C6H10NO4-	15	C5
			282.0289	C8H12NO8S-	12	C5, S
			142.0510	C6H8NO3-	11	C5
			138.9707	C2H3O5S-	4.9	S
Neu5Ac8S	0.82	504.09	286.0833	C14H12N3O4-	130	C5, D
			424.1351	C18H22N3O9-	102	C5, D
			203.0462	C10H7N2O3-	100	D
			96.9601	HSO4-	81	S
			100.0404	C6H6NO2-	76	C5
			300.0395	C8H14NO9S-	71	C5, S
			198.9918	C4H7O7S-	15	S
			160.0615	C6H10NO4-	15	C5
			282.0289	C8H12NO8S-	10	C5, S
			142.0510	C6H8NO3-	10	C5
			138.9707	C2H3O5S-	0.45	S
Neu5Gc	0.77	440.13	116.0353	C4H6NO3-	1500	C5
			302.0782	C14H12N3O5-	540	C5, D
			203.0462	C10H7N2O3-	100	D
Neu5Gc9S	0.54	520.09	116.0353	C4H6NO3-	2100	C5
			302.0782	C14H12N3O5-	700	C5, D
			440.1311	C18H22N3O10-	690	C5, D
			316.0344	C8H14NO10S-	380	C5, S
			96.9601	HSO4-	172	S
			198.9918	C4H7O7S-	160	S
			203.0462	C10H7N2O3-	100	D
			138.9707	C2H3O5S-	21	S
Neu5Gc8S	0.60	520.09	116.0353	C4H6NO3-	1900	C5
			302.0782	C14H12N3O5-	660	C5, D
			440.1311	C18H22N3O10-	600	C5, D
			316.0344	C8H14NO10S-	460	C5, S
			96.9601	HSO4-	443	S
			203.0462	C10H7N2O3-	100	D
			198.9918	C4H7O7S-	100	S
			138.9707	C2H3O5S-	2.8	S
Kdn	0.52	383.11	203.0462	C10H7N2O3-	100	D
			245.0568	C12H9N2O4-	95	C5, D
			275.0673	C13H11N2O5-	3.7	C5, D
Kdn9S	0.38	463.07	203.0462	C10H7N2O3-	100	D
			245.0568	C12H9N2O4-	92	C5, D
			383.1096	C16H19N2O9-	71	C5, D
			259.0129	C6H11O9S-	24	C5, S
			96.9601	HSO4-	22	S
			275.0673	C13H11N2O5-	4.4	C5, D
			138.9707	C2H3O5S-	1.9	S
Kdn8S	0.42	463.07	203.0462	C10H7N2O3-	100	D
			245.0568	C12H9N2O4-	91	C5, D
			383.1096	C16H19N2O9-	70	S
			96.9601	HSO4-	47	C5, S
			259.0129	C6H11O9S-	38	C5, D
			275.0673	C13H11N2O5-	4.1	S
			138.9707	C2H3O5S-	0.62	C5, D

^‡All intensities are normalized to m/z 203 (100%).

^⁎Relative retention time (RRT) is a relative value indicating the retention time of each sialic acid species in relation to that of Neu5Ac.

^†C5: fragment ions containing the C5 moiety;

^$S: sulfate-containing fragment ions;

^#D: fragment ions containing the DMB. Bold characters indicate diagnostic MS fragments of the sulfation position

Among the diagnostic ions, the sulfate ion at m/z 96.96 [HSO₄]^- was observed at higher relative abundance in the 8-O-sulfated derivatives (Fig. 3C, F, and I) than in the corresponding 9-O-sulfated derivatives (Fig. 3B, E, and H). This suggests that the sulfate ester at C8 undergoes cleavage more readily than that at C9. In contrast, the fragment ion at m/z 138.97 [C₂H₃O₅S]^- was substantially more abundant in the 9-O-sulfated derivatives (Fig. 3B, E, and H). This ion is proposed to arise from the cleavage of the C7-C8 bond, indicating that this fragmentation pathway is favored in the 9-O-sulfated isomers. The lower abundance of this fragment in the 8-O-sulfated derivatives (Fig. 3C, F, and I) may reflect the reduced propensity of the sulfate group to migrate from C8 to C9, thereby limiting the formation of the structural intermediate that facilitates C7-C8 bond cleavage. The relative intensity of the ion at m/z 138.97, normalized to the ion at m/z 203.05, differed by approximately 10-fold, 7.5-fold, and 3-fold between DMB-Neu5Ac8S/DMB-Neu5Ac9S, DMB-Neu5Gc8S/DMB-Neu5Gc9S and DMB-Kdn8S/DMB-Kdn9S, respectively (Table 1). Likewise, the relative abundance of the sulfate ion at m/z 96.96 was approximately 2.5-fold higher for the Neu5Ac and Neu5Gc 8-O-sulfated isomers (Fig. 3C and F) and 2-fold higher for the Kdn 8-O-sulfated isomer (Fig. 3I) than for their respective 9-O-sulfated counterparts (Fig. 3B, E, and H). Collectively, these results demonstrate that the ions at m/z 96.96 and 138.97 serve as complementary diagnostic fragment ions for distinguishing sulfation at the C8 and C9 positions.

Next, the quantitative performance of two detection systems was evaluated using calibration curves generated from authentic standards. Fluorescence detection allowed reliable quantification of DMB-Neu5Ac8S over the range of 1 fmol to 10 pmol, showing excellent linearity within this concentration range (Fig. 4B). Although a fluorescence signal was detectable at 10 amol, the quantitative performance at this concentration was not fully validated and was therefore excluded from the quantitative range (Fig. 4A). In LC-MS analysis, XIC-based quantification using MRM-HR analysis in the negative-ion mode achieved a detection limit of approximately 50 fmol (Fig. 4C), whereas the corresponding detection limit in positive-ion mode was 320 fmol [15]. These results demonstrate that negative-ion mode offers markedly greater sensitivity for the analysis of SiaSs. In a previous study, Saraswat et al. directly compared fluorescence and ESI-MS detection of various biomolecules and showed that the relative sensitivity of the two depend on the analyte. Similarly, the detector-dependent sensitivity of DMB-labeled sialic acid in the present study may reflect the balance between the fluorescence response of the DMB fluorophore and the ionization efficiency of the DMB-labeled sialic acid.

Fig. 4.

Calibration curves of standard Neu5Ac8S comparing fluorescence detection and negative ion mode MRM-HR mass spectrometry. Samples were serially diluted 10-fold, ranging from 10 pmol to 10 amol, and 2-fold, ranging from 100 fmol to 50 fmol. The diluted samples were subjected to the liquid chromatography-fluorescence-mass spectrometry (LC-FLD-MS) system. (A) Fluorescent detection (FLR) (Ex: 373 nm, Em: 448 nm) within the range of 10 pmol to 10 amol. (B) Fluorescent detection (FLR) (Ex: 373 nm, Em: 448 nm) within the range of 10 pmol to 1 fmol with calibration curve (C) XIC of Neu5Ac (m/z: 424.136) within the range of 10 pmol to 50 fmol with calibration curve.

2.2. Comparative analysis of fragmentation patterns by CID and EAD with sodium adducts

In our previous study on O-acetylated sialic acids, positive-ion MS/MS analysis predominantly produced DMB-core-centered fragment ions rather than characteristic, informative ions for the sialic acid backbone itself, by both CID and EAD analysis [15]. This fragmentation behavior limited the structural characterization of modified sialic acids in positive-ion mode. To overcome this limitation, we attempted to add sodium ions prior to electrospray, as they are known to enhance structural analysis of saccharides [32] and also coordinate more stable conformations with acidic functional groups like sulfate esters than protons or ammonium adducts [33]. In this study, we examined sodium adduct formation to alter charge localization, suppress sulfate loss, improve positive-ion mode MS sensitivity, and promote the generation of structurally informative fragment ions from DMB-SiaS. Using the same UHPLC system described above, sodium adduct formation was therefore employed to enable positive-ion mode collision-induced dissociation (CID) and electron-activated dissociation (EAD) analyses to investigate the diagnostic fragment ions for differentiating 8-O- and 9-O-SiaSs. First, to evaluate the effect of sodium adduction on fragmentation behavior, the CID and EAD fragmentation patterns of Neu5Gc, Neu5Gc8S, and Neu5Gc9S were compared (Fig. 5 and Table 2). In Table 2, fragment ion signals were evaluated using signal-to-noise ratios (S/N) calculated by the instrument software (SCIEX OS). The S/N ratio represents the peak height divided by the local baseline noise, providing a standardized measure of peak quality and relative intensity within each MS/MS spectrum. This table includes fragment ions with S/N ≤ 3 that demonstrated reproducible detection across replicate analyses. These results indicate that the method achieves high sensitivity and high resolution, enabling reliable detection of low-abundance fragment ions with excellent reproducibility even at marginal S/N levels. Sodium adduction dramatically altered fragmentation patterns compared with non-adducted species (Fig. 3D–F vs. Fig. 5A–C, upper panel), suggesting that sodium coordination substantially affects both ionization and dissociation pathways. Notably, in contrast to previous reports in which fragmentation of the DMB moiety predominated, sodium adduction led to an increased contribution of sialic acid-derived fragment ions. This behavior can be rationalized by the reduced affinity of charge coordination with the DMB group, leading to preferential charge localization on the sialic acid and consequent promotion of its fragmentation. In addition, the sodium cation showed strong coordination complexes with multiple hydroxyl groups on the sialic acid moiety compared to the DMB backbone, which may further disfavor the formation of DMB-derived fragment ions and shift dissociation toward the sialic acid moiety. CID generated relatively more fragment ions than EAD. This difference likely reflects the distinct dissociation mechanisms: CID preferentially induces cleavage at low-energy bonds through multiple collisions with inert gas molecules, whereas EAD promotes homolytic bond cleavage via electronic excitation. As a result, markedly different fragmentation patterns were observed (Fig. 5). Under the present experimental conditions using sodium adduction, the product ions mainly originated from C-C cleavage in the sialic acid moiety. Interestingly, there were no dehydrated cyclic product ions (intramolecularly bridged structures of sialic acids), which have been commonly observed in positive-ion mode without sodium adduction. Regarding species-specific characterization, Neu5Ac, Neu5Gc, and Kdn can each be readily distinguished based on their sodium adducts, specifically by the C5 substitution-specific fragment ions at m/z 244.08 [C₉H₁₉N₁O₅]⁺, m/z 260.07 [C₉H₁₉N₁O₆]⁺, and m/z 203.05 [C₆H₁₂Na₁O₆]⁺, respectively. These ions were consistently observed in both CID (Fig. S2) and EAD spectra (Fig. S3). Furthermore, in CID spectra, dehydration products derived from these structures, namely m/z 226.07 [C₉H₁₇N₁Na₁O₅]⁺ and m/z 242.06 [C₉H₁₇N₁Na₁O₆]⁺, were also characteristically observed for Neu5Ac and Neu5Gc, respectively (Fig. S2). These MS/MS fragment ions clearly reflect differences in the C5 substituents, demonstrating that distinct C5 modifications can be effectively discriminated based on their fragmentation patterns.

Fig. 5.

Comparison of collision-induced dissociation (CID) and electron-activated dissociation (EAD)-derived LC-MS/MS fragmentation spectra of standard samples in positive ion mode. Each DMB-derivatized sample (10 pmol) was subjected to the LC-MS system with Na adduct. CID spectra (blue line) are shown in the upper panel, and EAD spectra (pink line) are shown in the lower panel. (A) Neu5Gc. (B) Neu5Gc8S. (C) Neu5Gc9S.

Table 2.

Retention times and m/z values of DMB-labeled sialic acids (Sia and SiaS) in positive ion mode with sodium adduct.

Sialic acid	RRT*	Precursor ion		Characteristic fragment ions			Fragmentation method
		m/z					CID	EAD
	(min)	[M+Na]+	[M+2Na-H]+	m/z	Molecular formula	Type	S/N ratio	S/N ratio
Neu5Ac	1	449.14	-	244.0792	C8H15NNaO6	C5†	400	62
				226.0686	C9H17NNaO4	C5	110	2.1
Neu5Ac9S	0.73		550.07	430.1221	C18H21N3NaO8	C5, D#	6.2	0.5
				346.0179	C8H14NNa2O9S	C5, S$	120	150
				328.0074	C8H12NNa2O8S	C5, S	160	0.5
				244.9702	C4H7Na2O7S	S	41	13
				184.9491	C2H3Na2O5S	S	12	2.1
				142.9385	HNa2O4S	S	17	8.3
Neu5Ac8S	0.82		550.07	430.1221	C18H21N3NaO8	C5, D	21	21
				346.0179	C8H14NNa2O9S	C5, S	149	21
				328.0074	C8H12NNa2O8S	C5, S	180	0.9
				244.9702	C4H7Na2O7S	S	37	4.7
				184.9491	C2H3Na2O5S	S	0.5	0
				142.9385	HNa2O4S	S	59	8.6
Neu5Gc	0.77	465.14		260.0741	C8H15NNaO7	C5	340	91
				242.0635	C8H13NNaO6	C5	56	3.8
Neu5Gc9S	0.54		566.07	446.1170	C18H21N3NaO9	C5, D	5.9	0.60
				362.0128	C8H14NNa2O10S	C5, S	150	58
				344.0023	C8H12NNa2O9S	C5, S	120	1.6
				244.9702	C4H7Na2O7S	S	38	13
				226.9597	C4H5Na2O6S	S	62	5.2
				184.9491	C2H3Na2O5S	S	8.5	1.3
				142.9385	H1Na2O4S	S	20	4.1
Neu5Gc8S	0.6	-	566.07	446.117	C18H21N3NaO9	C5, D	8.3	23
				362.0128	C8H14NNa2O10S	C5, S	71	14
				344.0023	C8H12NNa2O9S	C5, S	69	0.5
				244.9702	C4H7Na2O7S	S	19	3.7
				226.9597	C4H5Na2O6S	S	20	5.1
				184.9491	C2H3Na2O5S	S	0	0
				142.9385	HNa2O4S	S	12	6.4
Kdn	0.52	407.11	-	203.0526	C6H12NaO6	C5	450	96
				173.0420	C5H10NaO5	C5	0	17
Kdn9S	0.38	-	509.05	389.0955	C16H18N2NaO8	C5, D	2.2	0.60
				371.085	C16H16N2NaO7	C5, D	0.8	0
				304.9914	C6H11Na2O9S	C5, S	160	38
				274.9808	C5H9Na2O8S	C5, S	1.3	11
				244.9702	C4H7Na2O7S	S	43	9.4
				226.9597	C4H5Na2O6S	S	2.1	11
				184.9491	C2H3Na2O5S	S	3.5	2.2
				142.9385	HNa2O4S	S	27	8.3
Kdn8S	0.42	-	509.05	389.0955	C16H18N2NaO8	C5, D	6.8	1.5
				371.085	C16H16N2NaO7	C5, D	3.5	0.8
				304.9914	C6H11Na2O9S	C5, S	110	28
				274.9808	C5H9Na2O8S	C5, S	1.2	14
				244.9702	C4H7Na2O7S	S	25	13
				226.9597	C4H5Na2O6S	S	8.6	23
				184.9491	C2H3Na2O5S	S	0.2	0
				142.9385	H1Na2O4S	S	37	10

^⁎Relative retention time (RRT) is a relative value indicating the retention time of each sialic acid species in relation to that of Neu5Ac.

^†C5: fragment ions containing the C5 moiety;

^$S: sulfate-containing fragment ions;

^#D: fragment ions containing the DMB. Bold characters indicate diagnostic MS fragments of the sulfation position

When C8 is modified by a sulfate group, formation of the C4-C8 cyclic structure is expected to be prohibited. Moreover, C9 sulfation may similarly hinder cyclization, presumably due to steric constraints and/or altered sodium coordination. This would limit access to cyclic product-ion pathways for both 8-O- and 9-O-sulfated species. Thus, sodium adduction clearly modified the fragmentation behavior and increased the contribution of sialic acid-derived ions. However, comparison of CID spectra suggested that this alone was not sufficient to provide a complete set of diagnostic ions for unambiguous positional assignment of sulfation sites.

To further evaluate positional isomer discrimination, we compared the CID fragmentation patterns of sodium-adducted DMB-labeled 8-O- and 9-O-SiaS derivatives, including Neu5Ac, Neu5Gc, and Kdn (Fig. 6; Table 2). In CID analysis, a fragment ion at m/z 184.95 [C₂H₃Na₂O₅S]⁺ was observed as a diagnostic fragment for all 9-O-sulfated species. This fragment structure corresponds to the m/z 138.05 ion detected in negative-ion mode with the addition of two sodium ions. In negative-ion mode, the m/z 138.05 fragment was observed for both 8-O- and 9-O-sulfated species, although with higher intensity in 9-O-sulfated species (Table 1). Importantly, this differential detection pattern was observed prominently for Neu5Ac and Neu5Gc, whereas it was weak for Kdn, indicating a structure-dependent fragmentation behavior. These differences are likely attributable to the nature of the C5 substituent, which may influence sodium coordination, charge localization, and consequently the preferred fragmentation pathways. Because sodium-adducted CID alone provided only limited discrimination of the sulfate positional isomers, we next investigated EAD under identical sodium-adduct conditions. In EAD analysis, a fragment ion at m/z 430.12 [C₁₈H₂₁N₃Na₁O₈]⁺ was detected as a dominant peak specific to Neu5Ac8S, while m/z 446.12 [C₁₈H₂₁N₃Na₁O₉]⁺ was specific to Neu5Gc8S (Fig. 6A and B, Neu5Ac8S and Neu5Ac9S, EAD). These fragments correspond to losses of one sulfate group, one sodium ion, and one proton from the precursor ions. Consistent with observations in negative-ion mode CID, the sulfate ester at the 9-position appears more stable than that at the 8-position, explaining the preferential detection of these fragments in 8-O-sulfated species (Fig. 6, CID). Although these ions were also detected in positive-ion CID spectra without sodium adduction, CID induced extensive desulfation across all SiaS species, resulting in the detection of desulfated products in both isomers and limiting their diagnostic utility (Fig. S2 and Fig. S3). In contrast, EAD provided more characteristic fragmentation features that were less obscured by non-selective desulfation, thereby enhancing positional discrimination. For Kdn derivatives, a corresponding fragment at m/z 389.10 [C₁₆H₁₈N₂Na₁O₈]⁺ was observed specifically for Kdn8S; however, it showed very low intensity in EAD, making it difficult to utilize as a reliable diagnostic fragment. In addition, a Kdn8S-characteristic fragment at m/z 371.09 [C₁₆H₁₆N₂Na₁O₇]⁺was observed in CID spectra, although its intensity was weak, similar to that of the m/z 184.95 ion (Table 2). This fragment is consistent with a consecutive loss of sulfate and water. The absence, or markedly reduced formation, of the analogous dehydration product in Kdn9S can be explained by the lower propensity for desulfation in Kdn9S; consequently, the pathway leading to subsequent water loss would be suppressed, and the corresponding fragment would not be generated efficiently. This reduced fragmentation efficiency may result from differences in charge distribution arising from the distinct C5 substituent and its effect on sodium coordination. Collectively, these results demonstrate that sodium adduction effectively redirects fragmentation from the DMB moiety toward the sialic acid backbone, thereby increasing structural information obtained in positive-ion mode. Although sodium-adducted CID enabled partial discrimination of the sulfate positional isomers through the detection of 9-O-specific product ions, complete differentiation required EAD. In particular, sodium-adducted EAD generated additional diagnostic fragments for the 8-O-sulfated Neu5Ac and Neu5Gc derivatives. Thus, whereas negative-ion CID remains the most sensitive approach for SiaS analysis, the complementary use of sodium-adducted positive-ion CID and EAD provides a robust strategy for distinguishing C8- and C9-sulfated positional isomers.

Fig. 6.

Comparison of LC-MS/MS fragmentation spectra of standard Sia8S and Sia9S by CID and EAD in positive ion mode. Sia8S spectra (blue line) are shown in the upper panel, and Sia9S spectra (pink line) are shown in the lower panel. DMB-derivatized samples (10 pmol each) were subjected to the LC-MS system with Na adduct. (A) Neu5Ac8S and Neu5Ac9S. (B) Neu5Gc8S and Neu5Gc9S. (C) Kdn8S and Kdn9S.

2.3. Comprehensive analysis of SiaS species in sea urchin sperm and eggs

To evaluate the applicability of the newly established method to biological specimens, SiaSs were analyzed using known crude samples: sea urchin sperm and eggs [[34], [35], [36]]. Neu5Ac8S has previously been reported as the predominant SiaS in sea urchin sperm, whereas Neu5Gc8S has been identified in sea urchin eggs. Sialic acid fractions were purified from both samples prior to LC-FLD-MS analysis. Using DEAE-Sephadex A25 anion-exchange chromatography, singly charged sialic acids (e.g., Neu5Ac and acetylated sialic acids) were eluted with 100 mM NaCl, doubly charged sialic acids (e.g., Neu5Ac8S) with 150 mM NaCl, and triply charged fractions (e.g., Neu5AcS2) with 200 mM NaCl. In the mono-charged fraction from sea urchin sperm, Neu5Ac and Neu5Gc were detected, with Neu5Ac at approximately 45-fold higher abundance than Neu5Gc (Fig. 7A). In the di-charged fraction, both Neu5Ac8S and Neu5Gc8S were detected (Fig. 7B), demonstrating that the developed workflow is directly applicable to complex biological samples. Moreover, their MS/MS fragmentation patterns were identical to those from authentic standards. In sea urchin eggs, Neu5Gc was exclusively detected in the singly charged fraction (Fig. 8A), whereas both Neu5Gc8S and Neu5Gc9S were detected in the doubly charged fraction (Fig. 8B). No sialic acid was detected in triply charged fractions of sea urchin eggs and sperm (data not shown). Unlike sperm, which showed only 8-O-sulfated species, eggs contained both sulfated isomers, suggesting the presence of an egg-specific sulfotransferase responsible for 9-O-sulfation of sialic acid. Neu5Gc8S was approximately 70-fold more abundant than Neu5Gc9S. Additionally, a previously unreported structure, Neu5GcS1Ac1, was suggested by a combined IDA and MRM-HR approach. Three peaks corresponding to Neu5GcS1Ac1 were separated using a UHPLC system, suggesting the presence of three isomers differing in the substitution sites of the acetyl and sulfate groups. The MS/MS spectra from peak 1 and peak 2 clearly showed the presence of a mass shift of 42 Da (O-acetyl ester) on the characteristic fragment ions, including de-sulfated ion at m/z 482.14 and sulfated ion at m/z 358.94, which were identified by comparison with authentic mono-sulfated Neu5Gc. The fragment ion at m/z 422.11 was also assigned as the de-O-acetylated product of the ion at m/z 482.14. To confirm the presence of an acetyl group in Neu5GcS1Ac1, alkaline treatment was performed. O‑acetyl esters are known to be easily saponified under alkaline conditions [37], whereas sulfate esters remain stable. The dominant peak (peak 2 at 161–165 min) was collected by UHPLC, dried, and incubated either in water or in 6.5 N aqueous ammonia at 37°C for 30 min before being subjected to LC-MS analysis. The water‑treated fraction retained the Neu5GcS1Ac1 structure, whereas aqueous ammonia treatment resulted in the detection of Neu5Gc8S (Fig. 9), supporting the presence of an acetyl substituent on Neu5Gc8S. Although the exact position of the acetyl group remains to be determined, alkaline treatment confirmed the presence of an additional alkali-labile acetyl modification on Neu5Gc8S. A sulfated and acetylated sialic acid, Neu5Ac4Ac8S, has previously been reported based on GC-MS analysis [41]. In this study, three potential Neu5GcS1Ac1 isomers were detected by LC-MS, and the presence of O-acetylated Neu5Gc8S was confirmed. To our knowledge, this is the first report of a both sulfated and acetylated Neu5Gc for which the structure was validated by high resolution MS/MS fragmentation analysis in an LC-MS system. This structure may confer resistance to sulfatases and esterases, thereby potentially enhancing the stability of sulfated and/or acetylated sialic acids against degradation by endogenous enzymes. Although this has not yet been established in sea urchins, studies in mammals have shown that O-acetylation of sialic acids occurs at the level of CMP-sialic acid in the Golgi apparatus, whereas sulfation takes place on sialic acids after their incorporation into glycoconjugates. Based on these observations, it is suggested that sialic acid sulfotransferases in sea urchins may recognize not only unmodified sialic acids but also acetylated sialic acids as substrates.

Fig. 7.

Detection of sialic acids from sea urchin sperm by LC-FLD-MS and negative ion mode LC-MS/MS. Sialic acids released using acid solution were pooled as a flow-through fraction using the C18 column. Free sialic acids were then separated by anion exchange chromatography and eluted with 100 mM NaCl and 150 mM NaCl, pooled as singly charged and doubly charged fractions, respectively. The fractions were desalted by a Sephadex G-10 column. Purified sialic acids were derivatized by DMB and subjected to the negative mode LC-FLD-MS system (upper panels) and further analyzed by MS/MS (lower panels) in negative ion mode. (A) Singly charged fraction (100 mM fraction), containing sialic acids: Neu5Ac and Neu5Gc. (B) Doubly charged fraction (150 mM fraction), containing SiaSs: Neu5Ac8S and Neu5Gc8S. *Unknown peak.

Fig. 8.

Detection of sialic acids from sea urchin eggs by LC-FLD-MS and negative ion mode LC-MS/MS. Sialic acids released using acid solution were pooled as a flow-through fraction using the C18 column. Free sialic acids were then separated by anion exchange chromatography and eluted with 100 mM NaCl and 150 mM NaCl, pooled as singly charged and doubly charged fractions, respectively. The fractions were desalted by a Sephadex G-10 column. Purified sialic acids were derivatized by DMB and subjected to the negative mode LC-FLD-MS system (upper panels) and further analyzed by MS/MS (lower panels). (A) Singly charged fraction (100 mM fraction), containing sialic acid: Neu5Gc. (B) Doubly charged fraction (150 mM fraction), containing SiaSs: Neu5Gc8S, Neu5Gc9S, and Neu5GcS1Ac1. *Unknown peak.

Fig. 9.

Deacetylation of Neu5GcS1Ac1 from sea urchin egg (as shown in Figure 8) to confirm the structure. DMB-Neu5GcS1Ac1 was fractionated from a sea urchin egg-derived sample, treated with water (neutral conditions) or aqueous ammonia (final 6.5 N, basic conditions), and subjected to an LC-FLD-MS system. (A) Treatment of DMB-Neu5GcS1Ac1 by water. (B) Treatment of DMB-Neu5GcS1Ac1 by aqueous ammonia.

3. Conclusions

In conclusion, we established an LC–FLD-MS–based analytical strategy for the quantitative and qualitative characterization of SiaSs following DMB derivatization. The negative-ion mode provides highly sensitive quantification, whereas sodium-adducted positive-ion MS/MS, particularly when combined with EAD, enables structural discrimination of positional sulfation isomers through characteristic fragmentation patterns. Together, these complementary analytical approaches provide a robust platform for the sensitive detection and structural characterization of SiaSs. Furthermore, this newly established method was successfully applied to biological samples, leading to the identification of a previously unreported sialic acid structure, Neu5GcS1Ac1, in sea urchin eggs. These findings facilitate future investigations into their biological roles and sialic acid diversity.

4. Materials and methods

4.1. Materials

Neu5Ac and trifluoroacetic acid (TFA) were purchased from NACALAI TESQUE (Kyoto, Japan). Sodium hydrosulfite, 2-mercaptoethanol, Tris(hydroxymethyl)aminomethane (Tris), NH₃, and propionic acid were purchased from FUJIFILM Wako (Osaka, Japan). Neu5Gc was purchased from Tokyo Chemical Industry (Tokyo, Japan). DMB was purchased from Dojindo (Kumamoto, Japan). Methanol, 0.1% (v/v) formic acid in acetonitrile (ACN), and 0.1% (v/v) formic acid in water were purchased from KANTO CHEMICAL (Tokyo, Japan). 4-MU-8-O- and 9-O-sulfated SiaS (Neu5Ac, Neu5Gc, and Kdn) were prepared as previously described [26]. DEAE-sephadex A-25 and Sephadex G-10 columns were purchased from GE Healthcare (Buckinghamshire, UK).

4.2. Synthesis of sulfated Neu5Ac

The 8-O-sulfated Neu5Acα(2,3)Gal was synthesized by following the reported procedure [38]. The 9-O-sulfated Neu5Acα (2,3)Gal was synthesized from the known protected Neu5Acα(2,3)Gal derivative [39]. The derivative was first subjected to de-O-acetylation under Zemplén condition to give a 4,7,8,9-tetraol derivative. Next, the 9-OH was selectively masked with a levlinoyl group, followed by re-acetylation of the 4-,7- and 8-OH. The protected disaccharide derivative was converted into a trichloroacetimidate donor via debenzylation, benzoylation, cleavage of p-methoxyphenyl group and trichloroacetimidoylation. The obtained disaccharide donor was glycosidated with an amino-terminated aglycon moiety. Finally, the 9-OH was selectively deprotected by treatment with hydrazine acetate and subsequently sulfonated with SO₃-pyridine complex, followed by global deprotection to deliver 9-O-sulfated Neu5Acα(2,3)Gal. The details of the synthesis will be reported elsewhere.

4.3. DMB derivatization of sialic acids

Standard samples (Neu5Ac, Neu5Ac8S, Neu5Ac9S, Neu5Gc, Neu5Gc8S, Neu5Gc9S, Kdn, Kdn8S, and Kdn9S) were dissolved in 20 µL of 7 mM DMB solution in 5 mM TFA containing 1 M 2-mercaptoethanol and 18 mM sodium hydrosulfite, then mixed with 20 µL of 0.01 N TFA, and incubated at 50°C for 2 h in the dark [26].

4.4. Analysis of standard sialic acids using LC-MS

LC–MS analyses of DMB-labeled sialic acids were carried out using a Nexera UHPLC platform (Shimadzu, Kyoto, Japan) interfaced with a ZenoTOF 7600 mass spectrometer (SCIEX, Framingham, MA, USA). The chromatographic system consisted of an LC-40D binary pump, a DGU-405 degasser, a SIL-40C autosampler, and an SCL-40 system controller [15]. Separation was performed on a CAPCELL PAK ADME column (3 μm particle size, 2.1 mm internal diameter × 150 mm; Osaka Soda, Japan) or a CAPCELL CORE C18 column (2.7 μm particle size, 2.1 mm internal diameter × 150 mm; Osaka Soda, Japan). The column temperature was maintained at 30°C in a CTO-40S column oven. The mobile phases were 0.1% (v/v) formic acid in water (solvent A) and 0.1% (v/v) formic acid in acetonitrile/methanol (50:50, v/v) (solvent B), delivered at a constant flow rate of 0.25 mL/min. For each analysis, standard samples corresponding to 10, 1, 0.1, 0.05, 0.025, and 0.01 pmol were injected. Elution was achieved using the following gradient program: solvent B was held at 3% from 0 to 100.0 min, increased from 3% to 7% between 100.0 and 170.0 min, ramped from 7% to 95% from 170.0 to 170.1 min, maintained at 95% from 170.1 to 175.0 min, returned to 3% from 175.0 to 175.1 min, and then re-equilibrated at 3% from 175.1 to 180.0 min, resulting in a total run time of 180 min. Fluorescence detection was performed with an RF-20A XS fluorescence detector using excitation and emission wavelengths of 373 and 448 nm, respectively [15]. To generate sodium adducts, 0.5 mM sodium acetate was introduced post-column after fluorescence detection at a flow rate of 0.05 mL/min. Mass spectrometric measurements were acquired on the ZenoTOF 7600 instrument. Both information-dependent acquisition (IDA) and MRM-HR methods were employed for MS/MS analysis. Fragmentation was induced by either CID or EAD. In the initial stage of analysis, full-scan TOF MS and IDA-triggered MS/MS experiments were used to determine the retention times of the analytes of interest. MRM-HR analysis was performed using both CID and EAD fragmentation to acquire qualitative MS/MS transitions at the retention times of each analyte. Quantitative measurements were performed using extracted ion chromatograms in MRM-HR experiments as well as fluorescence detection. The instrument was operated in negative ion mode and positive ion mode. Source and gas conditions were set as follows: curtain gas, 40 psi; ion source gas 1, 60 psi; ion source gas 2, 60 psi; collision-activated dissociation gas, 7; source temperature, 300°C; and ion spray voltage, −4500 V for CID negative mode and 5500 V for CID and EAD positive ion mode. The TOF MS scan range was m/z 100–1000, whereas the TOF MS/MS scan range was m/z 80–1000 for negative mode. Accumulation times were set to 0.10 s for TOF MS, 0.025 s for CID, and 0.50 s for EAD. For EAD measurements, the electron beam current was 7000 nA, the electron kinetic energy was 8 eV, and the collision energy was 10 V. The reaction time for EAD was 45 ms. For CID measurements, collision energies were adjusted according to analyte class, with 30 V applied to non-SiaS and 40 V to SiaS; the reaction time was 30 ms. All data were processed and analyzed using SCIEX OS software (version 3.4; SCIEX).

4.5. Sialic acid detection in sea urchins

Sea urchin egg and sperm lysate (100 μg) [36,40] was adjusted to 100 μL with water, and 100 μL of 4 N propionic acid was added. Acid hydrolysis was performed at 80°C for 2 h. After hydrolysis, the samples were dried in a centrifugal vacuum concentrator (SpeedVac, Thermo Fisher Scientific, Waltham, MA, USA). The dried samples were dissolved in 100 μL water and applied to an activated C18 Monospin column (GL, Sciences Inc., Tokyo, Japan), eluted with 100 μL water and 300 μL 10% ACN, and the flow-through and eluted samples were dried by SpeedVac. These samples were loaded onto an anion exchange chromatography column (DEAE-Sephadex A-25, 1 mL) and eluted with 100 mM NaCl in Tris-HCl (pH 8.0) and 150 mM NaCl in 20 mM Tris-HCl (pH 8.0). Each fraction was collected and dried by lyophilization. The samples were then desalted using a Sephadex G-10 column (20 mL) and dried by lyophilization. The desalted samples were labeled with DMB as described above. DMB-derivatized samples were desalted using an HPLC system equipped with a CAPCELL PAK C18 MG column (5 µm, 4.6 mm i.d. × 250 mm, OSAKA SODA, Japan), and fractions between 5 and 45 minutes were collected. The separation was performed using a mobile phase consisting of 0.01% (v/v) TFA, ACN, and MeOH (7:7:86, v/v/v) at an isocratic flow of 1.0 mL/min. Finally, the samples were evaporated, water was added, and they were subjected to the LC/MS system as described above (1 μg protein/injection).

4.6. Alkaline treatment for the deacetylation of Neu5GcS1Ac1 from sea urchin eggs

DMB-derivatized Neu5GcS1Ac1 was prepared as described above. It was fractionated using a UHPLC system equipped with a CAPCELL PAK ADME column, as described above, and 161- to 165-min fractions were collected. The samples were evaporated using SpeedVac. The dried sample was divided into two equal aliquots and incubated with water or aqueous ammonia (final 6.5 N) at 37°C for 30 min. The incubated samples were dried using SpeedVac. Finally, water was added to the samples, and they were subjected to the LC/MS system as described above.

Footnotes

CID	collision-induced dissociation
EAD	electron-activated dissociation
Kdn	deaminoneuraminic acid
DMB	1,2-diamine-4,5-dimethylenedioxybenzene
UHPLC	ultra-high-performance liquid chromatography
MS	mass spectrometry
HRAM	high-resolution, accurate-mass
HRMS	high-resolution and high-accuracy mass spectrometer
MRM-HR	high-resolution multiple reaction monitoring
Neu5Ac	N-acetylneuraminic acid
Neu5Gc	N-glycolylneuraminic acid

Ethics declaration

This study was conducted in accordance with the following guidelines for animal welfare and/or reporting: Nagoya University. Ethics committee approval was not required. The animal we used was sea urchin and experiments on sea urchin are not required for the acceptance of animal experiment committee in Nagoya University.

CRediT authorship contribution statement

Sota Yamashita: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Takayuki Omoto: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Nao Yamakawa: Writing – review & editing, Writing – original draft, Visualization, Validation, Software, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Naoko Komura: Writing – review & editing, Resources. Hiromune Ando: Writing – review & editing, Software. Di Wu: Writing – review & editing, Validation, Investigation. Ken Kitajima: Writing – review & editing, Writing – original draft, Validation, Resources. Masaya Hane: Writing – review & editing, Validation, Investigation, Funding acquisition. Chihiro Sato: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.