The alternative dissociation pathways initiated by ultraviolet photodissociation (UVPD) compared with collision-induced dissociation (CID) may provide useful diagnostic fragments for biomolecule identification, including glycans. However, underivatized glycans do not commonly demonstrate strong UV absorbance, resulting in low fragmentation yields for UVPD spectra. In contrast to UVPD experiments that leverage covalent modification of glycans, we detail the capacity of metal adduction to yield comparatively rich UVPD fragmentation patterns and enhance separation factors for an isomeric glycan set in a drift tube ion mobility system. Ion mobility and UVPD-MS spectra for two N-acetyl glycan isomers were examined, each adducted with sodium or cobalt cations, with the latter providing fragment yield gains of an order of magnitude versus sodium adducts. Furthermore, our glycan analysis incorporated front-end ion mobility separation such that the structural glycan isomers could still be identified even as a mixture and not simply composite spectra of isomeric standards. Cobalt adduction proved influential in the glycan separation by yielding an isomer resolution of 0.78 when analyzed simultaneously versus no discernable separation obtained with the sodium adducts. It is the combined enhancement of both isomeric drift time separation and isomer distinction with improved UVPD fragment ion yields that further bolster multivalent metal adduction for advancing glycan IM-MS experiments.
Glycans occupy primary roles in a number of vital biological processes, including but not limited to immune response, blood type determination, cancer cell differentiation, and cell recognition [1–4]. Currently, the combination of data from nuclear magentic resonance (NMR) and tandem mass spectrometric (MS) analyses are necessary to obtain an accurate assessment of the glycan structures present in a given sample . Development of comprehensive MS analytical methods alone would be ideal, but the large structural variety of glycans  and stereochemical blindness of mass spectrometry necessitates the incorporation of secondary separation schemes to infer carbohydrate identity even with tandem fragmentation approaches.
When MS is used for glycan determination, tandem MS remains essential for determining glycan composition. Fragmentation techniques for glycans often include collision-induced dissociation (CID), electron capture dissociation (ECD), and electron transfer dissociation (ETD), however, a priori knowledge of the glycan system is often needed for conclusive identification. Ultraviolet photodissociation is no exception to this trend; however, it has demonstrated the capacity to generate an abundance of informative fragments such as cross-ring cleavages [7–12].
Along with using alternative fragmentation methods to adjust fragment yield, the use of metal adduction has also been found to modify fragmentation variety. Konig and Leary found the use of cobalt adduction to aid in pentasaccharide linkage determination when complemented with CID-MS of the deprotonated species . Vartanian et al. investigated the use of alkaline earth and transition metal divalent cations as adducts with tetracycline antibiotics for IRMPD-MS analysis; however, when compared to the protonated species, these metal-tetracycline adducts yielded equivalent or poorer quality spectra than the protonated species’ IRMPD spectra . By contrast, later work from Adamson and Hakansson, and Zhou and Hakansson, applied divalent metal ions as charge carriers for glycan analysis with ECD-MS and IRMPD-MS with the conclusion that IRMPD worked well as a supplemental technique for glycan characterization with ECD-MS [15, 16]. These results suggest that the fragmentation of divalent metal adducts alone is not adequte to determine glycan composition. However, given the limited assessments made to date for such systems, it seems imprudent to assume that the technique should be relegated to obscurity. Rather, the work by Adamson and Hakansson, and Zhou and Hakansson, indicate the potential for the application of divalent metal adduction to glycan analysis using photon-based fragmentation [15, 16], To the best of our knowledge, the use of UV photons for the fragmentation of divalent metal-glycan adducts is absent from the published literature. This information gap and a desire to maximize informative fragment yields prompted the pursuit of UVPD-MS of metal-glycan adducts.
While UVPD has been applied to increase fragmentation variety for a range of chemical systems, there are relatively few reports integrating this approach with post ionization separations such as ion mobility spectrometry. Limited reporting for this instrumental combination may, in large part, be due to challenges coupling drift tube separation with ion trapping systems. However, notable developments in this domain now enable such experiments [11, 12, 17–20]. Selection of fragmentation mass spectra from a given drift time window reveals the fragments specific to a precursor of a particular drift time, excluding isomers of differing mobility. This drift time separation method was used by Zucker et al. with vacuum UVPD in comparison with CID for analysis of tetrasaccharide and trisaccharide isomeric pairs, with vacuum UVPD yielding more informative spectra overall . Similarly, Lee et al. used vacuum UVPD on seven isobaric disaccharides after dual-gate ion mobility separation, also finding improved fragment yield over CID, but with low fragment abundances characteristic of sodiated glycan photodissociation .
For the purpose of distinguishing isomers, high resolution ion mobility spectrometry (IMS) offers a tractable approach to isomer separation as it provides a rapid mechanism to distinguish isomers according to the ionized glycans’ mobility down a voltage gradient against a counter-current neutral gas flow. Mobility separation combined with UVPD-MS holds potential as a valuable technique for isomeric mixture analysis. Accordingly, we demonstrate enhanced UVPD fragmentation patterns and isomeric separation using IM-MS for two cobalt cation-adducted isomeric tetrasaccharide N-acetylated alditols.
The two tetrasaccharides in question were initially analyzed as sodium adducts, but the fragment yield was less than ideal at ≤1% of the precursor abundance. Along with evidence from Konig and Leary that cobalt adduction may improve the yield of glycan structural information, the addition of cobalt acetate adducted to glycans was found by Dwivedi et al.  to promote improved drift time separation for a simple mixture of disaccharides. Cobalt acetate was also found to be necessary here to produce two distinct drift time peaks for the isomers together as a mixture. Additionally, metal adduction via the addition of cobalt acetate to the glycan solutions was used to increase UV photon absorption to facilitate fragmentation. The isomeric glycans used here were found to have improved relative fragment abundance to that of the precursor when adducted to cobalt than when assessed as sodiated adducts.
Ion mobility experiments were conducted using an atmospheric pressure, dual-gate ExcellIMS MA3100 ion mobility spectrometer coupled directly to a Thermo Scientific LTQ linear ion trap mass spectrometer (Figure 1) [19, 22]. Ion trap injection times were 100 ms for CID experiments and 500 ms for UVPD experiments, while ion activation times were 30 ms for CID and 500 ms for UVPD. The longer injection and activation times were used with UVPD to permit a minimum of 1 laser pulse per mass scan. Ultraviolet photodissociation was performed with a Lambda Physik LPF-200 193 nm ArF excimer laser firing at a frequency of 3.3 Hz, with pulsing synchronized with MS trapping. Each laser pulse has an approximate energy output of 120 mJ/cm2. Dual-gate mobility separations with single-stage MS were performed on a 11-20 ms drift time range in 25 μs steps with 1000 ms dwell times, with each such experiment lasting 6 min. Drift time separation of the mixture with UVPD-MS was performed on a 14-20 ms range in 20 μs steps with 3 s dwell times, for a total analysis time of 15 min. The gate pulse width used for all mobility separations was 500 μs under a field strength of ~420 V/cm in order to maximize ion current to the linear ion trap.
The two isomeric tetrasaccharides used were O-glycans with one N-acetyl group on two residues per glycan, one fucose residue each, and were alditols with fully reduced reducing-ends. Both glycans were purified with high performance liquid chromatography (HPLC) and were characterized by nuclear magnetic resonance (NMR) . Structures of both glycans can be found in Figure 4 and Supplementary Material (Figure S-1). One of the glycans has residues linked in a linear fashion with no true branching, while the second tetrasaccharide does have a branched structure. For fragmentation experiments, solutions of each individual tetrasaccharide were prepared at 25 μM each for the mixture of both, and 50 μM each for individual standard analyses, with cobalt acetate added to each solution at a final concentration of 50 μM to ensure complete complexation. For drift time spectra obtained with single-stage MS, solutions were prepared with 10 μM each of the tetrasaccharides and 50 μM of cobalt acetate. This concentration of added cobalt salts for adduction was found to yield cobalt-glycan adduct ions at abundances that are approximately equal to that of the corresponding sodium adducts.
Results and Discussion
Initially, our aim was to utilize ion mobility separation in combination with UVPD for analysis of the two tetrasaccharides in question as sodiated adducts in a mixture. For the sake of comparison, CID fragmentation data were also obtained for the sodium-glycan adducts. However, these experiments involving the sodium adducts encountered a similar problem exhibited in previous reports regarding UVPD: the fragment yield in absolute and relative abundance to the precursor was poor and arguably too low [12, 24]. This behavior is common for glycans containing functional groups or modifications that do not possess any appreciable UV absorption cross sections. The attachment of chromaphores to target analytes has been arguably the most direct method to increase UV absorption . However, to counter the poor absorption, our method involved a technique previously used to identify glycan linkages with CID and IRMPD: adduction with cobalt cations was found to be effective in producing a greater relative abundance of UVPD fragments. A comparison of UVPD fragmentation spectra for both isomeric glycans as sodium and cobalt adducts can be found in Figure 2 as well as Figure S-2 (Supplementary Material). The sodiated glycans produced fragment yields consistent with carbohydrates’ typical low absorbance of UV photons, at relative abundances well below 1%. Furthermore, the sodiated glycan UVPD-MS spectra lacked significant fragment ion diversity. By contrast, a notable aspect of the cobalt-glycan adduct spectra is that the two isomers produce a broader, more distinct array of fragment ions between the isomers and can be distinguished accordingly.
After determining that the cobalt-glycan adducts were more ideal for UVPD fragmentation of the two isomers, the UVPD-MS spectra were to be compared to CID fragments of the same adducts. While UVPD is generally regarded as producing more information-rich spectra of cross-ring cleavages, the effect of the cobalt complexation with the glycans on more traditional CID also produced favorable results. Figure 3 shows the spectral comparison for the tetrasaccharide alditols using both CID and UVPD. CID and UVPD both produced essentially the same major fragments per Co-glycan adduct. Of those major fragment ions (shown in Figure 3), a difference between UVPD and CID becomes apparent in regard to ion abundance. CID produces major fragments in a greater abundance than UVPD, but this is not necessarily indicative of UVPD having lower capacity for fragment production. Rather, while UVPD yields less of the major fragments, minor fragments are in far larger abundance with UVPD than with CID. Furthermore, the minor peaks seen here with UVPD appear with visibly low noise, even at 20x magnification, confirming that UVPD can yield improved structural information about the glycans in question over CID. Among the fragment ions produced from UVPD are those that correspond to the composition of cross-ring cleavages, which appear in the CID spectra for neither the linear nor branched glycans. Although not all of the numerous UVPD fragments for the two isomers could be correlated with a particular composition, many of the higher abundance fragments, with signal-to-noise ratios (SNRs) greater than or equal to three, could be identified according to theoretical fragment compositions and are shown in Figures S-3 through S-6 (Supplementary Material). UVPD produced 17 branched glycan and 15 linear glycan identifiable fragment m/z values, in contrast to the 13 branched glycan fragments and the 9 linear glycan fragments yielded by CID. Clearly, UVPD may indeed provide appreciable fragment variety, and with the improved fragment abundance from cobalt adduction, UVPD can applied to isomers separated by ion mobility despite the lower ion current relative to no prior separation.
To exemplify the combination of UVPD-MS and ion mobility separation, the performance of the hybrid system was evaluated when the two isomers were separated in a mixture. The adduction of cobalt to glycans has previously been demonstrated as one possible method of improving isomeric glycan drift time separation, which was conveniently exploited here . As shown in Figure 4, the two tetrasaccharides are largely separated according to their individual drift time peaks. Furthermore, the two run simultaneously in a single sample as a standard mixture yielded two distinct drift time regions where a single cobalt-glycan isomer adduct can be found, with centroid fits showing a resolution of 0.78 for the glycans run together. This result is an improvement compared to the broad peak resulting from analysis of the sodiated glycans assessed as a mixture, as shown in Figure S-7 (Supplementary Material). As can be seen in Figures 3, the relative peak heights for the two isomers’ cobalt adducts differ by just over a factor of two. Because the concentrations of the glycans were equal, this difference in abundance can likely be attributed to the isomers’ relative ionization efficiencies as cobalt adducts and correspondingly their affinities for cobalt cations.
Drift time selected UVPD-MS of the mixture produced mass spectra that, when averaged across appropriate drift time peaks, align well with UVPD-MS spectra from the individual standards. Fragment peaks considered to be matched between the mixture and corresponding individual standards’ spectra had to be of the same nominal m/z, same charge state, have SNRs exceeding 3, and also have a corresponding theoretical fragment composition generated by GlycoWorkBench 2 . Both glycans had no fewer than eight fragments corresponding between the individual standard spectra and mixture spectra. UVPD spectra that were drift time selected from the mixture, as well as individual UVPD-MS spectra of the two glycans, can all be viewed for comparison in Figure 5.
Overall, the branched tetrasaccharide produced a larger variety of fragment ions in comparison to the linear glycan, but with the linear glycan yielding a more distinctive UVPD fragment pattern. As an example of its distinctive spectral pattern, the linear tetrasaccharide produced an easily recognizable cluster of four fragments at the nominal m/z values 425, 443, 466, and 484, with all appearing in roughly similar proportions relative to one another. Although the branched glycan does also produce each of these same four fragment ions, the unique feature of each appearing at nearly equal abundances in the linear tetrasaccharide facilitates distinguishing the two isomers. An interesting feature of the branched tetrasaccharide is the higher abundance of the 774 nominal m/z fragment, which is due to a loss of water, due to CID more so than UVPD, as well as the same fragment ion from the linear glycan having a comparatively muted presence. Additionally, the branched glycan has the 628 and 646 m/z fragments appearing at abundances that are more equivalent than from the linear tetrasaccharide, and the m/z 569 ion is a feature produced solely by the branched glycan, all of which can serve to distinguish the branched isomer from the linear isomer.
While gas-phase mobility separations theoretically possess the capacity to separate isomeric species prior to mass analysis, the feasibility of such approaches are only fundamentally limited by the underlying resolution of the system. When the resolution of an instrumental configuration is near its maximum, alternative approaches to differentially altering the gas-phase conformations of isomers are needed to realize separation prior to mass analysis. For the limited system examined in this work of two isomeric glycans, metal adduction provides a mechanism to realize gas-phase separations of isomers while having the added benefit of enhancing tandem mass spectrometry yields and variety. UVPD is known for providing feature-rich fragmentation spectra, but often at the cost of ion abundance relative to the precursor if unmodified glycans are the target analyte. Our results demonstrate that transition metal adduction may be yet another technique to raise the relative fragment yield in a fashion analogous to chromophore attachment . The distinguishing feature of metal adduction, however, is in the simplicity of the technique as well as avoiding derivatization and subsequent purification protocols. While unknown isomeric glycan mixture determination remains extremely challenging with current technology, the combination of mobility separation with high fragment variety dissociation methods may be a route by which the carbohydrate isomer barrier can be circumvented .
Robinson, L.N., Artpradit, C., Raman, R., Shriver, Z.H., Ruchirawat, M., Sasisekharan, R.: Harnessing glycomics technologies: integrating structure with function for glycan characterization. Electrophoresis 33, 797–814 (2012)
Zaia, J.: Mass spectrometry of oligosaccharides. Mass Spectrom. Rev. 23, 161–227 (2004)
Hart, G.W.: Myriad roles of glycans in biology. J. Mol. Biol. 428, 3147–3149 (2016)
Anugraham, M., Jacob, F., Nixdorf, S., Everest-Dass, A.V., Heinzelmann-Schwarz, V., Packer, N.H.: Specific glycosylation of membrane proteins in epithelial ovarian cancer cell lines: glycan structures reflect gene expression and DNA methylation status. Mol. Cell. Proteomics 13, 2213–2232 (2014)
Breg, J., Van Halbeek, H., Vliegenthart, J.F., Klein, A., Lamblin, G., Roussel, P.: Primary structure of neutral oligosaccharides derived from respiratory-mucus glycoproteins of a patient suffering from bronchiectasis, determined by combination of 500-MHz 1H-NMR spectroscopy and quantitative sugar analysis. 2. Structure of 19 oligosaccha. Eur. J. Biochem. 171, 643–654 (1988)
Laine, R.A.: Invited Commentary: a calculation of all possible oligosaccharide isomers both branched and linear yields 1.05 × 1012 structures for a reducing hexasaccharide: the Isomer Barrier to development of single-method saccharide sequencing or synthesis systems. Glycobiology 4, 759–767 (1994)
Devakumar, A., Thompson, M.S., Reilly, J.P.: Fragmentation of oligosaccharide ions with 157 nm vacuum ultraviolet light. Rapid Commun. Mass Spectrom. 19, 2313–2320 (2005)
Racaud, A., Antoine, R., Joly, L., Mesplet, N., Dugourd, P., Lemoine, J.: Wavelength-Tunable Ultraviolet Photodissociation (UVPD) of heparin-derived disaccharides in a linear ion trap. J. Am. Soc. Mass Spectrom. 20, 1645–1651 (2009)
Wilson, J.J., Brodbelt, J.S.: Ultraviolet photodissociation at 355 nm of fluorescently labeled oligosaccharides. Anal. Chem. 80, 5186–5196 (2008)
Ko, B.J., Brodbelt, J.S.: 193 Nm ultraviolet photodissociation of deprotonated sialylated oligosaccharides. Anal. Chem. 83, 8192–8200 (2011)
Zucker, S.M., Lee, S., Webber, N., Valentine, S.J., Reilly, J.P., Clemmer, D.E.: An ion mobility/ion trap/photodissociation instrument for characterization of ion structure. J. Am. Soc. Mass Spectrom. 22, 1477–1485 (2011)
Lee, S., Valentine, S.J., Reilly, J.P., Clemmer, D.E.: Analyzing a mixture of disaccharides by IMS-VUVPD-MS. Int. J. Mass Spectrom. 309, 161–167 (2012)
Konig, S., Leary, J.A.: Evidence for linkage position determination in cobalt coordinated pentasaccharides using ion trap mass spectrometry. J. Am. Soc. Mass Spectrom. 9, 1125–1134 (1998)
Vartanian, V.H., Goolsby, B., Brodbelt, J.S.: Identification of tetracycline antibiotics by electrospray ionization in a quadrupole ion trap. J. Am. Soc. Mass Spectrom. 9, 1089–1098 (1998)
Adamson, J.T., Håkansson, K.: Electron capture dissociation of oligosaccharides ionized with alkali, alkaline Earth, and transition metals. Anal. Chem. 79, 2901–2910 (2007)
Zhou, W., Håkansson, K.: Electron capture dissociation of divalent metal-adducted sulfated N-glycans released from bovine thyroid stimulating hormone. J. Am. Soc. Mass Spectrom. 24, 1798–1806 (2013)
Clowers, B.H., Hill Jr., H.H.: Mass analysis of mobility-selected ion populations using dual gate, ion mobility, quadrupole ion trap mass spectrometry. Anal. Chem. 77, 5877–5885 (2005)
Gaye, M.M., Valentine, S.J., Hu, Y., Mirjankar, N., Hammoud, Z.T., Mechref, Y., Lavine, B.K., Clemmer, D.E.: Ion mobility-mass spectrometry analysis of serum N-linked glycans from esophageal adenocarcinoma phenotypes. J. Proteome Res. 11, 6102–6110 (2012)
Morrison, K.A., Bendiak, B.K., Clowers, B.H.: Enhanced mixture separations of metal adducted tetrasaccharides using frequency encoded ion mobility separations and tandem mass spectrometry. J. Am. Soc. Mass Spectrom. (2016). doi:10.1007/s13361-016-1505-y.
Liu, Y., Clemmer, D.E.: Characterizing oligosaccharides using injected-ion mobility/mass spectrometry. Anal. Chem. 69, 2504–2509 (1997)
Dwivedi, P., Bendiak, B., Clowers, B.H., Hill, H.H.: Rapid resolution of carbohydrate isomers by Electrospray Ionization Ambient Pressure Ion Mobility Spectrometry-Time-of-Flight Mass Spectrometry (ESI-APIMS-TOFMS). J. Am. Soc. Mass Spectrom. 18, 1163–1175 (2007)
Morrison, K.A., Siems, W.F., Clowers, B.H.: Augmenting ion trap mass spectrometers using a frequency modulated drift tube ion mobility spectrometer. Anal. Chem. 88, 3121–3129 (2016)
Bendiak, B., Harris-Brandts, M., Michnick, S.W., Carver, J.P., Cumming, D.A.: Separation of the complex asparagine-linked oligosaccharides of the glycoprotein fetuin and elucidation of three triantennary structures having sialic acids linked only to galactose residues. Biochemistry 28, 6491–6499 (1989)
Devakumar, A., Mechref, Y., Kang, P., Novotny, M.V., Reilly, J.P.: Identification of isomeric N-Glycan structures by mass spectrometry with 157 nm laser-induced photofragmentation. J. Am. Soc. Mass Spectrom. 19, 1027–1040 (2008)
Damerell, D., Ceroni, A., Maass, K., Ranzinger, R., Dell, A., Haslam, S.M.: The glycanbuilder and glycoworkbench glycoinformatics tools: updates and new developments. Biol. Chem. 393, 1357–1362 (2012)
Support for K.A.M. was provided in part by the Defense Threat Reduction Agency under Grant Award Number HDTRA-14-1-0023. The authors acknowledge support from the New Faculty Seed Grant from Washington State University.
Electronic supplementary material
Below is the link to the electronic supplementary material.
(DOCX 2011 kb)
About this article
Cite this article
Morrison, K.A., Clowers, B.H. Differential Fragmentation of Mobility-Selected Glycans via Ultraviolet Photodissociation and Ion Mobility-Mass Spectrometry. J. Am. Soc. Mass Spectrom. 28, 1236–1241 (2017). https://doi.org/10.1007/s13361-017-1621-3
- Ion mobility
- Ultraviolet photofragmentation
- Isomer separations