Simple Approach for De Novo Structural Identification of Mannose Trisaccharides
- 347 Downloads
Oligosaccharides have diverse functions in biological systems. However, the structural determination of oligosaccharides remains difficult and has created a bottleneck in carbohydrate research. In this study, a new approach for the de novo structural determination of underivatized oligosaccharides is demonstrated. A low-energy collision-induced dissociation (CID) of sodium ion adducts was used to facilitate the cleavage of desired chemical bonds during the dissociation. The selection of fragments for the subsequent CID was guided using a procedure that we built from the understanding of the saccharide dissociation mechanism. The linkages, anomeric configurations, and branch locations of oligosaccharides were determined by comparing the CID spectra of oligosaccharide with the fragmentation patterns based on the dissociation mechanism and our specially prepared disaccharide CID spectrum database. The usefulness of this method was demonstrated to determine the structures of several mannose trisaccharides. This method can also be applied in the structural determination of oligosaccharides larger than trisaccharides and containing hexose other than mannose if authentic standards are available.
KeywordsOligosaccharide Structure CID
Carbohydrates (or glycans) are one of the four major classes of biomolecules. They are commonly found on cellular surfaces and extracellular matrixes and are influential in all cellular recognition processes and many other biological processes [1, 2]. To determine the relationship between the chemical and biological properties of carbohydrates, the structures of carbohydrates must be identified. However, this remains a challenge  because of the heterogeneity of carbohydrates and only small quantities can be extracted from biological systems. Most importantly, the number of carbohydrate isomers for a given chemical formula is significantly high, which makes the structural identification of carbohydrates through a single analytical method very difficult. For example, the number of all possible oligosaccharide isomers containing six hexoses was estimated  to be higher than 1.05 × 1012. This high number, known as the isomer barrier , is a long-standing technological problem in the structural analysis of carbohydrates.
The structures of carbohydrates can be determined through nuclear magnetic resonance spectroscopy  and mass spectrometry (MS) . MS is widely applied in the structural analysis of carbohydrates because it requires only a small sample (typically from 100 μm to 10 nm) compared with other methods. However, the application of MS in the structural determination of carbohydrates has not been as successful as that in the structural determination of proteins, mainly because of the low ionization efficiency of carbohydrates in a mass spectrometer [7, 8], the large number of carbohydrate isomers , and the similarity of mass spectra between isomers.
The complete structural characterization of oligosaccharides requires the identification of the constituent monosaccharides, sequence, linkages, anomeric configurations, and branch location. Collision-induced dissociation (CID) tandem MS is a major method that employs mass spectrometry to determine the structure of carbohydrates . Several empirical fragmentation patterns of carbohydrate cations [10, 11, 12], anions [13, 14, 15, 16, 17, 18], and derivatized carbohydrates [19, 20, 21, 22, 23, 24] have been reported and used in structural determination. However, they can only determine part of the structure or a small set of well-characterized oligosaccharides in databases.
Few methods for the de novo structural identification of carbohydrates have been reported in the literature. Nagy and Pohl demonstrated a fixed-ligand kinetic method for the absolute configuration determination of hexose monosaccharides , but no application of this method for the structural determination of oligosaccharides has been reported. Bendiak et al. assigned the stereochemistry and anomeric configuration of monosaccharides in oligosaccharides by using anion m/z 221 [26, 27, 28]. However, the structure of two monosaccharides on the reducing side cannot be determined, and it is only suitable for identifying linear oligosaccharides.
N- and O-linked glycans are important compounds for biological systems, as many research findings have proven that these glycans are involved in immune responses, cell-to-cell communications, cell apoptosis, and congenital disorders. N-linked glycans share a common core structure containing mannose oligosaccharides. Recently, we demonstrated a new approach for the de novo structural determination of oligosaccharides . Through this method, the linkage, anomeric configuration, and branched locations of underivatized linear and branched oligosaccharides can be determined. The structural determination of mannose trisaccharides was demonstrated in this study. The approach is simple and the sensitivity level is high. Because the dissociation mechanism we used in the structural determination does not depend on the size of oligosaccharides and the orientations of hydroxyl groups, this method has the potential to be developed into a standard procedure for determining larger oligosaccharides and those containing hexoses other than mannose.
High-Performance Liquid Chromatography-Electrospray Ionization-MSn
All disaccharide CID spectra for the database were measured in the positive mode by using a linear ion trap mass spectrometer (LTQ XL, Thermo Fisher Scientific, Waltham, MA, USA) coupled with a Dionex Ultimate 3000 high-performance liquid chromatography (HPLC) system (Thermo Fisher Scientific). The Dionex chromatography mass spectrometry link was installed as an interface to control the Dionex chromatography system with Xcalibur (MS) (Thermo Fisher Scientific). Samples were prepared in ultrapure water at a concentration of 1 × 10−4 M. The 18O-labeled carbohydrates were prepared using a previously described method .
HPLC separations of all disaccharides were performed using a Hypercarb (100 × 2.1 mm) column with a particle size of 3 μm, operated at room temperature (25 °C). The mobile phase comprised (A) 0.1% (v/v%) aqueous formic acid with 1 × 10−4 M NaCl, and (B) HPLC-grade acetonitrile. Various mobile phase elution conditions were carried out to optimize a superior separation for different isomers. The mobile phase for isocratic elution was a mixture of A-B (97:3, v/v) for β-Man-(1→2)-Man and β-Man-(1→6)-Man, and a mixture of A-B (99:1, v/v) for α-Man-(1→3)-Man, α-Man-(1→4)-Man, α-Man-(1→6)-Man, and β-Man-(1→4)-Man. The solvent gradient employed for α-Man-(1→2)-Man, and β-Man-(1→3)-Man separation consisted of a linear increase in the amount of acetonitrile (B) in solvent A. The gradient mode conditions for α-Man-(1→2)-Man were as follows: t = 0 min, A: 100%, B: 0%; t = 1 min, A: 100%, B: 0%; t = 11 min, A: 92.5%, B: 7.5%. The gradient was slightly tuned to obtain a more efficient separation for β-Man-(1→3)-Man as follows: t = 0 min, A: 100%, B: 0%; t = 1 min, A: 100%, B: 0%; t = 30 min, A: 92.5%, B: 7.5%. The mobile phase flow rate was set to 300 μL/min and the volume of the sample injected was 10 μL for all mannobioses. The column eluate was directly infused into the electrospray ionization (ESI) source without any postcolumn addition. The MS conditions were optimized using the built-in semiautomatic tuning procedure in Xcalibur. The ESI source was operated at a temperature of 280 °C with 30 units of sheath gas flow and 10 units of auxiliary gas flow. The ion spray voltage was 4.00 kV, and the transfer capillary temperature was 280 °C. The capillary voltage was 80 V and the tube lens voltage was 150 V. Helium gas was used as the buffer gas for the ion trap as well as the collision gas in CID. The MSn experiments were carried out with a normalized collision energy ranging from 20% to 100%, an activation Q value of 0.25, and a 30-ms activation time. The number of ions regulated by automatic gain control was set to 1 × 105 and the precursor ion isolation width was set to 2 u.
All oligosaccharide MSn were obtained using the same mass spectrometer (without HPLC) under similar operating conditions for disaccharides, except the ESI source, which operated at 35 °C, and CID performed at only 30% of normalized collision energy. Samples were prepared at a concentration of 1 × 10−4 M in 50% (v/v%) of HPLC-grade methanol and ultrapure water. Sodium chloride was added to the sample solution at a concentration of 1 × 10−4 M. A total of 1 min of scans was accumulated for each spectral acquisition.
Results and Discussion
Low-energy CID and resonance excitation were used in this study. Sodiated oligosaccharides were dissociated into fragments through sequential low-energy CID and the structures of oligosaccharides were determined using these CID spectra. The spectra were measured according to a procedure that we built from an understanding of the saccharide dissociation mechanism. The linkages, anomeric configurations, and branch locations of oligosaccharides were then determined by comparing the CID spectra of oligosaccharides with the predicted fragmentation patterns and disaccharide database. In the following paragraphs, we first describe the predicted fragmentation patterns, followed by the disaccharide database, the procedure for CID spectrum measurement, and the applications utilized for the structural determination of oligosaccharides.
Predicted Fragmentation Patterns
In the resonance excitation of low-energy CIDs, only the selected precursor ions were resonance excited to obtain energy from collisions. The internal energy of precursor ions, accumulated through many collisions, increased slowly because the amount of transferred energy in each collision was small. Most ions dissociate through the dissociation channels with low-energy barrier heights before the further increase of internal energy. This ensures that the energy remaining in products after dissociation is insufficient to undergo a secondary dissociation. As product ions are not resonance excited, they do not undergo subsequent excitation and dissociation. The combination of low-energy CID and resonance excitation ensures that only one dissociation process takes place for most ions during each CID.
Predicted Fragmentation Patterns of Trisaccharides Obtained Using Low-Energy CID. Fragments Produced in MS2 and MS3 are in Red and Green, Respectively
Specially Prepared Disaccharide Database
The process of structural determination includes the dissociation of oligosaccharides into disaccharides through a mass spectrometer and the determination of these disaccharide structures by subsequent CID spectra. In practical applications, the CID spectra of various disaccharides are measured in advance and compiled in a database. The CID spectra of the disaccharides produced from the dissociation of oligosaccharides are compared with those in the disaccharide database to achieve structural identification.
A major difference between the CID spectra of the disaccharide with 1→3 linkage and the disaccharide with 1→4 linkage is the relative intensity of ion m/z 347, representing the elimination of H2O from the reducing end. The difference can be explained using the dissociation mechanism. High-level quantum chemistry calculations [8, 35] revealed that the role of sodium cation in the water elimination of sodiated carbohydrates is the promotion of H atom transfer. The mechanism involves transferring the H atom from the O2 atom of the reducing sugar to the O1 atom of the same sugar, as illustrated in Scheme 1. The barrier height of the O1 and O2 atoms in the cis configuration is substantially lower than that in the trans configuration. Therefore, the CID spectrum with a high intensity of ion m/z 347, which represents the large branching ratio of the dehydration reaction, can be assigned to β anomers. For exmple, comparing the CID spectra of these two peaks illustrated in Figure 1n and o, the spectrum with a higher intensity of ion m/z 347 (Figure 1o) was assigned to α-Man-(1→4)-β-Man, whereas the other (Figure 1n) was assigned to α-Man-(1→4)-α-Man. Similar assignments can be employed for the remaining disaccharides, except for the disaccharides with 1→2 and 1→6 linkages, where dehydration is a minor dissociation channel.
A comparison of Figure 1n and q revealed that the major difference in the CID spectra between α-Man-(1→4)-α-Man and β-Man-(1→4)-α-Man is the relative intensities of ion m/z 203 and 305. The ion m/z 203 (Y1) represents glycosidic bond cleavage, and the ion m/z 305 (0,2A2) represents cross-ring dissociation through the elimination of C2H4O2 from the reducing sugar. The difference in spectra can again be explained by the dissociation mechanism due to the cis and trans configurations. The glycosidic bond cleavage mechanism of sodiated carbohydrates is analogous to that of H2O elimination, as illustrated in Scheme 1. First, the H atom on the O2 atom of the mannose on the nonreducing side of the glycosidic bond is transferred to the O1 atom of the same mannose, and then the C1–O1 glycosidic bond cleavage occurs. The O1 and O2 atoms of the nonreducing mannose in β-Man-(1→4)-α-Man are in the cis configuration, whereas these two O atoms are in the trans configuration in α-Man-(1→4)-α-Man. β-Man-(1→4)-α-Man has a lower barrier height than α-Man-(1→4)-α-Man; consequently, the intensity of ion m/z 203 of β-Man-(1→4)-α-Man is higher than that of α-Man-(1→4)-α-Man. A similar argument can be applied to explain the difference in the CID spectra of other disaccharides.
The measurement of CID spectra for databases was repeated several times to ensure the stability of the relative intensities. The normalized collisional energy (NCE) of CID used in this study ranges from 20% to 100%. The relative ion intensities in each CID spectrum do not change significantly in these measurements. Spectrum similarities were calculated to identify the stability of the repeated measurements, independence of NCE, and to differentiate the linkages and anomeric configurations. Details of the calculations are displayed in the Supplementary Information.
Procedure of CID Spectrum Measurement
In the second step, the disaccharide fragment ions in region B represent the disaccharides from both the reducing and nonreducing ends of the linear trisaccharides or the disaccharides from both branches of the branched trisaccharides. The CID spectrum of these ions, namely MS3(B), is the sum of the CID spectra from these two disaccharides. If the structure of one disaccharide is identified, the structure of the other disaccharide can be determined from MS3(B) by subtracting the CID spectrum of the determined disaccharide.
The disaccharide produced in the MS3(C) and MS3(D) is the disaccharide at the nonreducing side of the linear trisaccharides. The CID spectra of this disaccharide, MS4(C) and MS4(D) (Figure 2) can be used to determine the linkage, the anomeric configuration of the glycosidic bond, and the anomeric configuration of the reducing sugar if the linkage is 1→3 or 1→4 linkage. Notably, the anomeric configuration of the reducing sugar (labeled with 2) of the disaccharide fragment ions produced in MS3(C) and MS3(D) is equivalent to the anomeric configuration of the glycosidic bond between sugars 1 and 2.
The complete structural determination of trisaccharides was obtained by combining the structural information from each step, as illustrated in Figure 2b. One advantage of this method is that most of the structural assignments can be determined through multiple approaches, thus increasing the accuracy of the results.
Applications in Trisaccharides
Ions m/z 365 produced from 527→ 509(B3)→ fragments (Figure 4b) and 527→ 437(0,3A3)→ fragments (Figure 4c) represent the disaccharide on the nonreducing end of the linear trisaccharide. The CID spectra of these disaccharide fragment ions, 527→ 509(B3)→ 365(C2/B3)→ fragments and 527→ 437(0,3A3)→ 365(C2/0,3A3)→ fragments, are presented in Figure 4e and f, respectively. A comparison of the spectra in Figure 4e and f with that in Figure 1 suggests that this disaccharide is α-Man-(1→2)-Man. The anomeric configuration of the reducing sugar of disaccharide α-Man-(1→2)-Man cannot be determined due to the low dehydration intensity of the disaccharide with 1→2 linkage. Consequently, the combination of Figure 4a and e or f can only determine the trisaccharide as of α-Man-(1→2)-Man-(1→3)-Man. The linkages of both glycosidic bonds and the anomeric configuration of one glycosidic bond were identified, but the determination of the other anomeric configuration required a different approach.
The disaccharides produced from 527→ 365(C2 or Y2) (Figure 4d) include the disaccharides on both the reducing and nonreducing sides. As the disaccharide on the nonreducing side was determined to be α-Man-(1→2)-Man in the aforementioned approach (Figure 4e and f), subtracting Figure 4e or f from d generated the CID spectrum of the disaccharide on the reducing side (Figure 4g or h), which can be identified as α-Man-(1→3)-Man from the comparison with Figure 1. Therefore the structure of this trisaccharide can be completely determined to be α-Man-(1→2)-α-Man-(1→3)-Man by approach II through combining Figure 4a, d, and either g or h.
The analogous method of approach II for α-Man-(1→2)-α-Man-(1→3)-Man does not achieve desirable results for α-Man-(1→6)-α-Man-(1→6)-Man because the linkages of both glycosidic bonds are the same (1→6). The identification of the anomeric configuration of the glycosidic bond on the reducing end requires another approach.
The CID spectrum of 527→365→fragments (Figure 8c) is the sum of the spectra of disaccharides with 1→3 and 1→6 linkages. Subtraction of the spectrum with 1→6 linkage (Figure 1t, u, w, or x) from that in Figure 8c yields a spectrum with 1→3 linkage. In the subtraction process, ion intensities in Figure 1t, u, w, or x must be multiplied by a factor such that the spectrum after subtraction exhibits a near-zero intensity for ions m/z 245 and 305. This is because the spectrum after subtraction yields a spectrum of disaccharide with 1→3 linkage, which requires a near-zero intensity for ions m/z 245 and 305, as illustrated in the disaccharide database of Figure 1. If Figure 1w or x (β-Man-(1→6)-Man-1 or β-Man-(1→6)-Man-2) is used in subtraction, the ion intensity m/z 203 in the spectrum after subtraction displays negative values. Consequently, only Figure 1t and u (α-Man-(1→6)-Man-1 or α-Man-(1→6)-Man-2) can be used for subtraction. The subtraction of Figure 1t or u from Figure 8b is shown in Figure 8d and e, respectively, which indicates that the disaccharide is α-Man-(1→3)-α-Man through the comparison of the spectrum in Figure 8d and e with that in Figure 1. The combination of these CID spectra determined the trisaccharide as α-Man-(1→3)-[α-Man-(1→6)]-Man.
Currently, the identification of the linkage positions, anomeric configurations, and branching location of oligosaccharides is a major obstacle to carbohydrate research. In this study, we demonstrated a simple approach for the structural determination of underivatized mannose trisaccharides. The structural determination followed a logical procedure built from the understanding of the saccharide dissociation mechanism. This logical procedure can be easily extended to larger oligosaccharides and those containing hexoses other than mannose. The ion intensities of sodiated carbohydrates are high and the entire measurement of tandem MS can take less than 1 min. This method can be ultimately developed into a standard procedure for the structural identification of oligosaccharides and complex carbohydrates, a goal that remains a great challenge in glycomic analysis.
This work was financially supported in part by the Ministry of Science and Technology, Taiwan (103-2113-M-001-011-NY3).
- 1.Varki, A., Cummings, R.D., Esko, J.D., Freeze, H.H., Stanley, P., Bertozzi, C.R., Hart, G.W., Etzler, M.E. (eds.): Essentials of Glycobiology, 2nd edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor (2009)Google Scholar
- 3.National Research Council: Transforming Glycoscience: A Roadmap for the Future. The National Academies Press, Washington, DC (2012)Google Scholar
- 4.Laine, R.A.: A Calculation of All Possible Oligosaccharide Isomers Both Branched and Linear Yields 1.05x10(12) Structures for a Reducing Hexasaccharide - the Isomer-Barrier to Development of Single-Method Saccharide Sequencing or Synthesis Systems. Glycobiology. 4(6), 759–767 (1994)CrossRefGoogle Scholar
- 8.Chen, J.L., Lee, C., Lu, I.C., Chien, C.L., Lee, Y.T., Hu, W.P., Ni, C.K.: Theoretical investigation of low detection sensitivity for underivatized carbohydrates in ESI and MALDI. J. Mass Spectrom. 51(12), 1180–1186 (2016)Google Scholar
- 12.Vijayakrishnan, B., Issaree, A., Corilo, Y.E., Ferreira, C.R., Eberlin, M.N., Peter, M.G.: MSn of the six isomers of (GlcN) 2(GlcNAc) 2aminoglucan tetrasaccharides (diacetylchitotetraoses): Rules of fragmentation for the sodiated molecules and application to sequence analysis of hetero-chitooligosaccharides. Carbohydr. Polym. 84(2), 713–726 (2011)Google Scholar
- 21.Xue, J., Song, L., Khaja, S.D., Locke, R.D., West, C.M., Laine, R.A., Matta, K. L.: Determination of linkage position and anomeric configuration in Hex-Fuc disaccharides using electrospray ionization tandem mass spectrometry. Rapid Commun. Mass Spectrom. 18(17), 1947–1955 (2004)Google Scholar
- 22.Mendonca, S., Cole, R.B., Zhu, J., Cai, Y., French, A.D., Johnson, G.P., Laine R.A.: Incremented alkyl derivatives enhance collision induced glycosidic bond cleavage in mass spectrometry of disaccharides. J. Am. Soc. Mass. Spectrom. 14(1), 63–78 (2003)Google Scholar
- 26.Fang, T.T., Zirrolli, J., Bendiak, B.: Differentiation of the anomeric configuration and ring form of glucosyl-glycolaldehyde anions in the gas phase by mass spectrometry: isomeric discrimination between m/z 221 anions derived from disaccharides and chemical synthesis of m/z 221 standards. Carbohydr. Res. 342(2), 217–235 (2007)Google Scholar
- 27.Fang, T.T., Bendiak, B.: The stereochemical dependence of unimolecular dissociation of monosaccharide-glycolaldehyde anions in the gas phase: a basis for assignment of the stereochemistry and anomeric configuration of monosaccharides in oligosaccharides by mass spectrometry via a key discriminatory product ion of disaccharide fragmentation, m/z 221. J. Am. Chem. Soc. 129(31), 9721–9736 (2007)Google Scholar
- 29.Ni, C.K.: Simple methods for de novo structural determination of glucose-containing underivatized oligosaccharides. In the division of carbohydrate chemistry, 254th American Chemical Society National Meeting & Exposition August 20-24, 2017, Washington, DC USAGoogle Scholar
- 35.Chen, J.L., Nguan, H.S., Hsu, P.J., Tsai, S.T., Liew, C.Y., Kuo, J.L., Hu, W.P., Ni, C.K.: Collision-induced dissociation of sodiated glucose and identification of anomeric configuration. Phys. Chem. Chem. Phys. 19, 15454–15462 (2017)Google Scholar