Advertisement

Enhanced Mixture Separations of Metal Adducted Tetrasaccharides Using Frequency Encoded Ion Mobility Separations and Tandem Mass Spectrometry

  • Kelsey A. Morrison
  • Brad K. Bendiak
  • Brian H. Clowers
Focus: Emerging Investigators: Research Article

Abstract

Using five isomeric tetrasaccharides in combination with seven multivalent metals, the impact on mobility separations and resulting CID spectra were examined using a hybrid ion mobility atmospheric pressure drift tube system coupled with a linear ion trap. By enhancing the duty cycle of the drift tube system using a linearly chirped frequency, the collision-induced dissociation spectra were encoded in the mobility domain according to the drift times of each glycan isomer precursor. Differential fragmentation patterns correlated with precursor drift times ensured direct assignment of fragments with precursor structure whether as individual standards or in a mixture of isomers. In addition to certain metal ions providing higher degrees of separation than others, in select cases more than one arrival time distribution was observed for a single pure carbohydrate isomer. These observations suggest the existence of alternative coordination sites within a single monomeric species, but more interesting was the observation of different fragmentation ion yields for carbohydrate dimers formed through metal adduction. Positive-ion data were also compared with negative-ion species, where dimer formation did not occur and single peaks were observed for each isomeric tetrasaccharide-alditol. This enhanced analytical power has implications not only for carbohydrate molecules but also for a wide variety of complex mixtures of molecules where dissociation spectra may potentially be derived from combinations of monomeric, homodimeric, and heterodimeric species having identical nominal m/z values.

Graphical Abstract

Keywords

Ion mobility spectrometry Mass spectrometry Ion trapping Fourier transform Multiplexing Collision induced dissociation Glycan separation 

Introduction

Advances in soft ionization techniques suitable for biopolymers has fundamentally altered the focus of a range of analytical fields including metabolomics, proteomics, lipidomics, petroleomics, and glycomics [1, 2, 3, 4]. Of these approaches, glycomics remains particularly challenging owing to the high degree of structural similarity between glycans. More specifically, isomeric carbohydrates arise through variation in linkage position, stereochemistry, anomeric configuration, ring forms (pyranose versus furanose), ketose versus aldose, and enantiomers (D- versus L-forms) of individual sugars. All of these combinations produce an enormous number of potential isomeric variants and also help explain the vast roles carbohydrates occupy biologically [5]. Moreover, this wide range of isomeric structures confounds efforts to obtain truly unique single-stage mass spectral (MS) fragmentation patterns that unambiguously identify an unknown carbohydrate structure from a mixture [6, 7, 8]. In fact, some isomers are known to yield identical product ion m/z sets, which diminish the utility of single precursor-product transitions for glycan identification [9, 10]. Nonetheless, alternative fragmentation patterns that provide information via different cross-ring cleavages remain attractive as these fragments can directly aid in assigning glycan composition. Enhancement of cross-ring cleavages has been realized through the use of metal ion adducts [11, 12, 13, 14, 15, 16, 17], alternative dissociation techniques (e.g., IRMPD, UVPD, and ECD) [18, 19, 20, 21, 22], and chemical derivatization [23, 24, 25, 26, 27]. However, with complex mixtures of unknowns, the presence of more than one or even several precursor isomers renders interpretation of spectra difficult without establishing firm specific precursor-product relationships.

Traditional structural analysis of glycans involves multiple stages of high performance liquid chromatography (HPLC) combined with nuclear magnetic resonance (NMR), where fractions isolated from each liquid chromatography (LC) profile can be readily assessed for isomeric heterogeneity. Once reasonable purity (>95%) is obtained, efforts using NMR and MS yield results to which high levels of confidence are assigned [7, 28, 29]. Unfortunately, these separations are time-consuming, require comparatively large amounts of starting material, and more often than not, the collected fractions are mixtures of multiple isomers, which require additional chromatographic separation using different column types to yield reasonably pure molecules. Capillary electrophoresis (CE) provides an alternative, comparatively high resolution approach for separating isomeric glycans, but buffer systems incompatible with the electrospray (ESI) process have hindered its broad-scale coupling with mass spectrometry [30]. Additional approaches that may enhance the level of information obtained from complex glycan mixtures include chemical modification techniques that aid separation and, in some cases, enhance the yield of diagnostic glycan fragments using tandem MS [31].

Although new approaches to multidimensional chromatography have been demonstrated along with marked gains in high resolution m/z, the challenge of confidently identifying a glycan from an unknown mixture remains elusive. When chromatographic separations fail to resolve isomers and the mechanism of ion isolation for tandem MS experiments is incapable of discriminating on the basis of stereochemistry, alternative approaches to separation are necessary. As a post-ionization separation technique, ion mobility spectrometry (IMS) represents an attractive technique to enhance the level of information derived from a sample regarding the presence of isomeric species. Because most IMS separations occur on the millisecond timescale, it is also highly compatible with most LC approaches and provides a dimension of information that directly complements m/z measurements [32]. With broad-scale commercialization of ion mobility-mass spectrometry (IM-MS) instruments, reports leveraging mobility for ion separation of glycans has increased significantly in recent years [4]. Early yet separate works by Leavell et al. [33], Clowers et al. [10], and Zhu et al. [34] highlighted the potential of the technique to separate isomeric forms of carbohydrates. Zhu et al, in particular, indicated the utility of ion mobility separations coupled to a linear ion trap MS and using CID to perform mannose7 isomer separation and MS/MS analysis of an RNase B glycan mixture, verifying the importance of drift time separation in combination with tandem MS spectral data [34]. A key finding from the Leavell et al. and Clowers et al. works was that the different alteration of gas-phase conformation of carbohydrates using metals as coordinating charge carriers may hold promise when analyzing glycans. Diastereomers were separated with Zn adduction by Leavell et al. [33], while isomeric anomeric methyl glycosides were separated by the Hill group as proof of concept for cobalt adduction to increase isomer drift time separation [35]. Since then, metal adduction to glycans has been used with increasing frequency for sugar isomer separation; however, the focus and range of carbohydrates probed in any one report has been somewhat narrow. Deprotonation has also been utilized to achieve isomeric separations of glycans; however, the mechanism and resulting charge location does not necessarily translate across all glycans [36].

Although capable of providing higher resolution separation than low pressure IM-MS, signal-averaged, drift tube, atmospheric pressure IM-MS has historically suffered from a low duty cycle of ~1%, resulting in neutralization of the vast majority of ion produced [37, 38]. A compounding issue with this instrumental combination has been the delivery of comparatively low ion populations, which has minimized its impact when relying upon tandem MS results for assigning glycan composition. For drift tube IM-MS instruments using time of flight (TOF) MS systems, tandem MS results are possible; however, any fragmentation stages that slow ion transit times can severely degrade mobility peak shapes. Mobility separations prior to mass analysis using ion trapping systems, though possible, have historically been hindered by the low duty cycles of both instruments. Nevertheless, MS/MS information from mobility selected ions is possible [34, 39, 40], but significant enhancements in ion throughput are necessary to fully leverage the MSn capacity of such instruments [41, 42]. Recently, our group demonstrated two broadly applicable approaches to ion multiplexing that greatly improve ion throughput using a linearly chirped ion gate and the Fourier transform [43, 44]. Although the modes of operation differ depending on the type of mass analyzer employed (i.e., TOF MS versus ion trapping systems) the ion mobility duty cycle increases to levels between 25% and 50%. With such levels of ion flux, tandem MS experiments are tractable while still recording ion separations in the mobility domain even with ion trapping systems.

In an effort to demonstrate the utility of CID fragmentation combined with drift tube FT-IM-MS for glycan separations and analysis, seven metal ions have been evaluated as adducts with five isomeric tetrasaccharide-alditols as well as evaluation of the deprotonated glycans. Reduced alditols were used as previous results indicated the potential for the configuration of the reducing sugar (α- or β-anomers of pyranose or furanose forms) to complicate analyses by providing additional peaks for each reducing-end configuration [35]. In addition to reporting the impact of different adducted metals on the resolution of the carbohydrate isomers, we also explored the impact of these differing metal ions on the types and range of fragments produced. Along with demonstrating the benefits of this instrumental combination and mode of operation, our data illustrate the potential for such metal-glycan combinations to form both monomeric and dimeric species. The latter observations have direct implications for traditional approaches that leverage metal coordination to enhance cross-ring cleavages using mass spectrometry alone.

Experimental

Atmospheric Pressure Dual-Gate Drift Tube Ion Mobility Linear Ion Trap Mass Spectrometer

Drift times and mass spectra were measured using an atmospheric pressure dual-gate drift tube ion mobility spectrometry (MA3100; EXCELLIMS, Acton, MA, USA) coupled directly to a linear ion trap mass spectrometer (LTQ; Thermo Scientific, San Jose, CA, USA) without hardware modifications to either instrument. A schematic of the overall IM-MS system is shown in Figure S-1 (Supporting Information). In this system, the drift tube was designed with a rectangular electrode configuration with the homogeneous electric field generated via metal-plated ceramics. Temperatures for all experiments were maintained at 180 °C; the IM-MS system has an operating temperature range of ~30 to 250 °C. An electric field of approximately 470 V/cm was applied for all experiments using custom supporting electronics provided by ExcellIMS for the MA3100 drift tube. A 2 L/min flow of dry, purified nitrogen was introduced into the drift tube immediately prior to the LTQ inlet capillary, flowing counter-current to the ions produced at atmospheric pressure (~690 Torr in Pullman, WA) via electrospray ionization (ESI). Following ionization, the ions traverse a desolvation region that is approximately 6.25-cm long before encountering the first Bradbury-Nielson style gate (BN-gate). After traveling through the 10.56-cm long drift tube, the ions then reach a second BN-gate located immediately in front of the LIT-MS inlet, wherein the LIT-MS was used as the detector. All analyses were performed using the “normal” mass scan rate except for the spectra obtained for Figure 4, which were obtained with the “zoom” mass scan rate setting. The ion trap maximum inject times were 100 and 200 ms, with the longer trapping time corresponding to CID and “zoom” scan experiments. When CID was performed, a normalized collision energy of 15 was applied for each analysis, with a width of ±3 m/z units surrounding the precursor m/z used for all except the Ba and Sr adducts, which necessitated a width of ±5 m/z units due to their larger isotope distributions. The first and second gates were both pulsed (±60 and ±100 V, respectively) concurrently for the Fourier multiplexing experiments by connecting the two external gate triggers of the IMS to a pulse generator (NI BNC-2120; National Instruments, Austin, TX, USA) controlled via a function generation interface that was designed to accept custom pulse waveforms (IGOR Pro; WaveMetrics Inc., Lake Oswego, OR, USA). Fourier multiplexing of the ion mobility experiments was performed by generating frequency encoded spectra by pulsing the ion gates with increasing frequency until the linear sweep reaches the desired terminal frequency. Here, frequency chirps used were 5–10,005 Hz sweeps. The total time length for each sweep was maintained at 8 min, which was found previously to provide optimal signal-to-noise ratio (SNR) while still permitting experiment completion within a relatively short timeframe [44]. LIT-MS data acquisition and ion mobility system gates are synchronized to begin simultaneously by the trailing edge of a 1 Hz pulse of +5 V. Because the ion mobility separations were performed with a dual-gate system, the arrival time distributions shown relate to the space bracketed by the two gates, during which the ions remain as precursor ions, even if they are subsequently fragmented upon reaching the LIT-MS. As such, drift time profiles shown for any particular fragment actually reflect the arrival time distributions of the precursor ions that produce that same fragment m/z. An example segment of a 2D plot of m/z versus drift time obtained via FT-IM-CID-MS analysis of the full tetraitol mixture adducted with Co(II) can be seen in Figure 1a, with the Fourier ion current decays for two m/z values shown in Figure 1b and c, and the resultant drift time spectra in Figure 1d and e. These m/z values correspond to fragments produced by precursor ion isomers in the mixture; by obtaining the FT ion current decays for fragments of a given precursor m/z, the drift time profiles of the fragments’ precursor can permit tracing a fragment back to its appropriate isomer precursor. Metal-glycan adducts involving the divalent metals were found to adduct as monomers with the glycans as (Glycan - H + Metal)+, which had been observed previously with cobalt adducts by Konig and Leary in 1998 [12]. All theoretical glycan fragments identified in this article were identified using GlycoWorkbench 2 [45, 46].
Figure 1

With FT-IM-CID-MS, the drift times of all ions within the chosen m/z window are encoded in the ions’ ion current decay over the course of the experiment. (a) An example segment of a 2D plot of sweep time versus m/z, covering the mass range 540–570 for the full glycan isomer mixture (cellotetraitol, maltotetraitol, isomaltotetraitol, nigeran tetraitol, and the unnamed tetraitol) with added cobalt for adduction, demonstrates the simultaneous drift time encoding. Extracted ion currents over time for the precursors of specific cobalt-glycan fragments, at (b) m/z 544 and (c) m/z 564, convey how ions with differing arrival time distributions have contrasting Fourier decays. The resulting drift time spectra for the precursors of fragments at (d) m/z 544 and (e) m/z 564 reveal which precursors produce each fragment and to what extent

Chemicals and Reagents

Maltotetraose, α-D-Glcp-(1-4)-α-D-Glcp-(1-4)-α-D-Glcp-(1-4)-D-Glc was purchased from MP Biomedicals (Santa Ana, CA, USA); isomaltotetraose, α-D-Glcp-(1-6)-α-D-Glcp-(1-6)-α-D-Glcp-(1-6)-D-Glc, was from US Biological (Salem, MA, USA); nigeran tetrasaccharide from Aspergillus japonicas, α-D-Glcp-(1-3)-α-D-Glcp-(1-4)-α-D-Glcp-(1-3)-D-Glc, cellotetraose, β-D-Glcp-(1-4)-β-D-Glcp-(1-4)-β-D-Glcp-(1-4)-D-Glc, and an unnamed tetrasaccharide, α-D-Glcp-(1-6)-α-D-Glcp-(1-4)-α-D-Glcp-(1-4)-D-Glc, were from Sigma-Aldrich (St. Louis, MO, USA). Tetrasaccharides (10 mg) were reduced in a solution of aqueous 1M NaBH4 (1 mL) for 24 h. After acidification with 100 μL glacial acetic acid added dropwise with bubbling of H2, samples were diluted with 5 mL water and passed through columns of well-washed Dowex AG50W-X8, 100–200 mesh, H+ form (Bio-Rad; Hercules, CA, USA) and with 10 molar equivalents of H+ to the Na+ present. The column separation was then followed by a wash with five column volumes of water. Eluate from the column was rotary-evaporated to near dryness, and boric acid was removed as the volatile trimethylborate by five rotary evaporations from 1% glacial acetic acid (v/v) in reagent-grade methanol (5 mL) to near dryness, followed by three rotary evaporations from methanol (5 mL) alone to dryness. 1H-NMR (D2O, 500 MHz) showed complete reduction, with three anomeric doublets for each tetrasaccharide-alditol. A schematic of the structures following the glycan symbolism outlined by Varki et al. [47] is shown (Figure 2). Upon receipt of the alditols, each sample was solvated with 1 mL of HPLC grade water to produce concentrated alditol stock solutions. Dilutions of these stock solutions were then made such that the concentration of an individual tetraitol was 3 μM for a given solution, including both mixtures and individual standard runs, unless otherwise noted. All samples were brought to final dilution in HPLC grade methanol with 0.1% formic acid, except for the samples used for analysis of the deprotonated glycans, which were brought up in straight HPLC grade methanol. Metal chloride salts used include CoCl2 from Fischer Scientific (Waltham, MA, USA), CuCl2 and BaCl2 from Sigma-Aldrich (St. Louis, MO, USA), and MnCl2, SrCl2, and CaCl2 from J.T. Baker Co. (Center Valley, PA, USA). All metal-alditol samples were prepared with final metal chloride concentrations of 50 μM. Each individual glycan standard was analyzed with FT-IM-MS with a minimum of three replicates for the determination of the precision of measured ion drift times, reduced mobility values, and ion-neutral collision cross sections.
Figure 2

The five tetrasaccharide alditols examined here include cellotetraitol, maltotetraitol, unnamed tetraitol, nigeran tetraitol, and isomaltotetraitol. Each glycan is composed of glucose monomer units—shown as blue circles—with the appropriate linkage type and positions indicated on the figure. The reduced reducing end found on each sugar is indicated on the figure as a small white circle with two small, curved lines branching from the circle. Alditols were specifically chosen for use here so as to minimize any possibility of isomerization in the gas phase

Results and Discussion

As demonstrated previously by other groups [15, 16, 17, 35], the formation of metal cation adducts with isomeric glycans can alter the conformations of the glycans enough such that isomers that previously had the same drift times using one metal cation can have a greater arrival time distribution spread using another cation. The purpose of this particular study was to investigate the adduction of three alkaline earth metals and three transition metals to five isomeric glycans. Further, the intent was to evaluate the impact of different metal adduction on separation and fragmentation behavior for the target carbohydrates using FT-IM-CID-MS. To expand this comparison to more traditionally observed carbohydrate ions, sodium adducts and deprotonated ions were also included. More specifically, each of the glycan isomers examined was subjected to FT-IM-MS analysis as either a metal adduct (divalent metal cations: Co, Cu, Mn, Ca, Sr, Ba, monovalent metal cations: Na) or as a deprotonated ion for the purpose of both determination of Ko and ion-neutral collision cross-section, as well as giving an initial glimpse of which ion or ion adduct provides the best drift time separation of the five isomers. As shown in Figure 2, all five isomers are linked linearly without any branching and differ only by anomeric configuration and/or location. Cellotetraitol and maltotetraitol are both linked at (1→4) positions and thus provide an example of separation efficacy when isomers vary solely in their anomeric configurations. For evaluation of isomer separations based on differences in linkage positions, the α-linked glycans serve as examples based on positions that are increasingly dissimilar to maltotetraitol. The unnamed tetraitol differs from maltotetraitol in that its terminal residue is (1→6) linked, but with the remaining residues connected at the (1→4) positions. Nigeran tetraitol retains a (1→4) linkage between the second and third residues from the reduced reducing end, but has (1→3) linkages between the first and second residues, as well as between the third and fourth. Finally, isomaltotetraitol differs the most in linkage from maltotetraitol in that isomaltotetraitol contains only (1→6) linkages.

Ion Mobility Evaluation of Monomeric and Dimeric Precursor Ions

During the course of these initial experiments, it was realized that select divalent metal ions do not simply form single monomeric adducts with glycans. As shown in Figure 3, and in Figures S-2 to S-9 (Supporting Information), more than one drift time peak—between approximately 11 and 16 ms—was observed for many of the metal ion-tetrasaccharide-alditol combinations. These data illustrate that varying proportions of dimeric gas-phase species [2M + 2Metal - 2H]+2 are present. Central to Figure 3 are the 2D contour plots showing m/z versus drift time for isomaltotetraitol adducted with Mn(II) and maltotetraitol adducted to Ca(II) (Figure 3a and d, respectively). It should be noted that Ca is 97% natural abundance 40Ca (atomic mass 39.963) but does have other isotopes in small natural abundance, of which 44Ca is about 2%, and there is no reported Ca isotope having an atomic nominal mass of 41, while Mn is essentially 100% one isotope (atomic mass 54.938). Upon initial examination of the mass spectra, seen in Figure 3a showing a 2D plot of m/z versus drift time of isomaltotetraitol adducted with Mn(II) and Figure 3d with maltotetraitol and Ca(II), it was clear that a doubly charged species was present at the same nominal mass as the known singly charged monomeric adducts, as well as two drift time peaks (Figure 3b and e). Extraction of a narrow m/z window covering a range that would contain only the doubly charged peak showed that the drift time peak with a shorter arrival time was the only peak remaining (Figure 3c and f), indicating that the second peak was very likely a doubly charged dimer of the metal-glycan adduct. Because the doubly charged dimer appears at the same nominal mass as the monomer, it follows that the dimer is composed of a cluster of two glycans and two metal ions with the replacement of two protons. It should be noted that these dimers appear in this glycan system with divalent metal adducts, not with the monovalent sodium adducts or with the deprotonated glycans.
Figure 3

Shown here are illustrations of the underlying higher mobility, doubly charged species found to form with the metal-glycan adducts; (a)(c) present Mn-isomaltotetraitol adducts, whereas (d)(f) present Ca-maltotetraitol adducts. (a) A 2D plot of m/z versus drift time for manganese-isomaltotetraitol adducts illustrating the presence of the doubly charged dimer species present along with the monomeric [M – H + Mn]+ species. (b) Fourier transformation of the ion current decay extracted for the range m/z 721.75–722.25 yields an arrival time distribution revealing two drift time peaks corresponding to the dimer and the monomer species, whereas (c) transformation of the decay from the range m/z 722.25–722.75 produces only the shorter drift time dimer peak. Similarly, (d) the 2D plot of the calcium-maltotetraitol adducts shows the presence of the monomer and the dimer species, with the 706.75–707.25 m/z range showing both species as drift time peaks, whereas 707.25–707.75 reveals only the dimer drift time peak

Previous research by other groups has primarily focused on the formation of monomeric metal-glycan adducts [48, 49, 50], but as our data suggest, metal adduction to carbohydrates readily produces dimeric adducts and is particularly noticeable in IM-MS analysis. Our initial impression was that the dimer formation was a consequence of the 50 μM metal to 3 μM glycan concentration ratio used to ensure adduct formation, using a metal salt concentration that is roughly typical for metal-glycan adduction in ion mobility experiments [17, 18]. However, the analysis of solutions containing concentration ratios as low as 5 μM of metal cations to 3 μM of glycans did yield dimer formation (data not shown). Furthermore, while still lower metal to glycan ratios did result in the dimers’ disappearance, the abundance of the monomeric metal-glycan adduct dropped off to such a significant extent that the metal adduction method becomes unfeasible even without mobility separation. Not surprisingly, dimer formation here appears to be dependent on both the metal and glycan, with certain combinations of the two yielding greater dimer abundance than others. As a group, the alkaline earth metals are more prone to form dimers than the transition metals for this particular set of glycans. In fact, the dimer abundance with Sr and Ba can rival the abundance of the monomer, whereas the transition metals experience dimer formation roughly at half or even lower abundance compared with that of the transition metal-glycan monomeric adducts. Because of the presence of these additional species formed during metal adduction, there are subsequent implications regarding CID spectra of adducts. More specifically, these data highlight the challenge of interpreting fragmentation patterns of metal adducted carbohydrates. Using traditional approaches to m/z isolation and fragmentation in an ion trap, the presence of a doubly charged dimer presents a scenario in which any resultant fragmentation pattern is confounded by two related but distinctly different ion populations that may differ in proportions depending on precise sample concentrations and electrospray conditions. This experimentally observed scenario necessitates additional modes of separation to obtain fragmentation patterns derived from a single ion population. Although it is tempting to exploit potential fragmentation differences of the dimers between each glycan isomer as an additional identification metric, it should be noted that the aim of this study was to explore methods of isomeric glycan separation in a mixture. Because the possibility of heterodimer formation cannot be excluded without the use of isotope labeling, the focus of the data presented remains on the monomeric adducts. Furthermore, it should be emphasized that alditols were chosen specifically for this experiment because reduction of glycans at the reducing end eliminates the possibility of hydrolysis occurring in solution or isomerization in the gas phase at the anomeric carbon.

The clear presence of doubly charged species observable shorter drift times than the related singly charged species argue strongly for the formation of dimeric species. Stating that the faster-arriving species is entirely dimeric may be premature, but it is our assertion that it is unlikely for the ion-neutral collision cross-section of any singly charged monomeric metal-glycan complexes to take on such significantly more compact forms compared with the identifiably monomeric drift time peaks and have the same drift times as identifiable doubly charged dimers. Moreover, the shorter drift time peaks appear only when the mass spectra concurrently show the distinctive m/z pattern for doubly charged species. An additional observation was that the doubly charged metal-glycan dimers yield mass spectra that are variable in the dimers’ fragment contributions and abundances when included with the monomeric subpopulation of precursor ions. As a result, these spectra are accordingly less desirable when assigning composition using precursor isolation by m/z alone.

Differential Fragmentation Spectra for Mobility Selected Ion Populations

The use of high duty cycle multiplexed ion mobility separation readily permits the simultaneous application of collision induced dissociation during the course of the mobility separation, providing the corresponding fragmentation mass spectra of each ion at the drift time of its precursor. With the combination of drift time and fragmentation data, even isomers with overlapping drift time peaks have the potential to be distinguished, provided the fragmentation spectra differ sufficiently. A necessary first step is to determine the fragmentation patterns produced by each pure isomer and metal adduct combination, in addition to the CID spectra produced by the deprotonated glycan anions. Each individual metal-glycan standard was subject to direct electrospray and CID-MS analysis to determine the product ions produced by each metal-glycan adduct. However, the presence of the dimers indicated that there may exist contributions of fragments that differ in relative abundance or m/z compared with fragments specifically derived from the monomer. With this in mind, FT-IM-CID-MS analysis of each metal-glycan adduct standard pair was performed. As seen in Figure 4 for the Co adducts, the resulting CID mass spectra selected specifically according to the drift times of the singly charged metal-glycan adducts differed noticeably from the direct ESI-CID-MS analysis of the standards. Most noticeably for the Co adducts, the drift time selected fragmentation spectra are astonishingly simple compared with the direct ESI-CID-MS spectra, indicating that the dimer species formed with the same precursor m/z as the desired adducts may have a substantial impact on the spectra produced. Testing whether this observation holds for other complexes as well, the drift time isolated fragmentation spectra for each metal-glycan adduct was obtained and used for the analysis of each glycan mixture for a given metal ion, rather than simply the direct ESI-CID-MS spectra of each adduct. To show the variable impact of the dimer fragments on monomer spectra, we have included drift time selected CID spectra for the metal-glycan adducts which have been found to form dimers (Ca, Sr, Ba, Mn, and Co), including both monomer and dimer drift time selected fragmentation spectra, as well as the drift time isolated monomer spectra for those not forming dimers (Na and Cu adducts, deprotonated glycans), in Supporting Information Figures S-10 through S-17.
Figure 4

CID mass spectra for each isomeric glycan-cobalt adduct are shown with each characteristic primary fragment of the sugars indicated with a line and appropriate nominal mass; structures for each fragment can be seen in Table 1. Spectra shown above the x-axis were obtained by selection of the cobalt-glycan monomer according to drift time of the monomer at a width of 0.5 ms each, which largely encompasses each drift time peak. Reflected below the x-axis are CID mass spectra obtained by traditional means of direct electrospray of the cobalt-glycan samples. Notable differences between direct ESI and drift time selected fragmentation spectra include simplified spectra when drift time selected, and slightly differing peak ratios between the two methods

For the purpose of evaluating the ions and ion adducts with improved isomer separation compared with the sodium adduct, the focus here was on the drift-time-isolated CID spectra of the transition metal-glycan adduct monomers. Furthermore, because the adducts with cobalt produced spectra that were the simplest and easiest to interpret, with almost all fragments identifiable, the remaining results and discussion regarding CID spectra will focus on the cobalt adduct spectra while CID spectra from the other ions and ion adducts can be viewed in the included Supporting Information (Figures S-10 to S-17). In regard to the specific discussion of the cobalt-tetraitol adducts produced via CID, Table 1 shows that the major fragments returned include predominantly glycosidic bond fragments and not cross-ring cleavages. This result is largely expected, as single-stage CID typically does not input energy targeted to the locations necessary to produce cross-ring cleavages [51]. However, it was found from the drift time isolated spectra that there is a possibility of finding isomer specific fragmentation patterns among the tetraidols. Konig and Leary found similar potential for distinguishing linkage isomers based on CID of pentasaccharides as [M + Co – H]+ adducts [12]. Beginning with the one β-linked tetrasaccharide-alditol, cellotetraitol adducted with the cobalt cation primarily produced fragments at m/z 382 and 544, which corresponded to B2, B3/Y2, or C3/Z2 ions and the B3 ion, respectively, according to fragment nomenclature by Domon and Costello [52]. The α-linked analogue of cellotetraitol, maltotetraitol yielded a simpler drift time isolated CID spectrum, with only one major fragment at m/z 544, the B3 fragment. The unnamed tetraitol showed slightly more complex spectra with the addition of fragments at m/z 402 and 564 in addition to the B2, B3/Y2, or C3/Z2 ions and the B3 ion, found at m/z 382 and 544. At m/z 402, the corresponding unnamed tetraitol fragment is Y0, and m/z 564 corresponds to the Y2 ion. Intriguingly, isomaltotetraitol, which is α-linked at (1→6) positions for each glucose residue, does not appreciably produce any fragments at m/z 382 or 544 as an adduct with cobalt, but fragments at m/z 402 and 564 are the predominant product ions in isomaltotetraitol’s drift time isolated CID spectrum. In regard to linkage position, nigeran tetraitol stands out, in that it is the only tetraitol with any (1→3) linkage positions, which perhaps explains its particularly unique CID spectra. Specifically, the fragmentation spectrum of nigeran tetraitol is the most recognizable because of its five characteristic peaks; the particular fragments at the nominal m/z 364 and 384 are not noticeably abundant upon fragmentation of any of the other isomers. The m/z 364 peak is thought to be due to a B3/Z2 fragment, while the 384 fragment corresponds to a Z1 fragment ion. From here, the fragmentation patterns can be compressed into a rough assessment of how each tetrasaccharide interacts with the cobalt cation. With the assumption that the observed fragments retain the adducted metal cation, the fragments must then contain the particular glucose residues where the stronger coordination of the cation takes place.
Table 1

Fragments Produced for Each Cobalt-Glycan Isomer Monomeric Adduct as Determined with Drift Time Selected FT-IM-CID-MS

Impact of Different Metals on Isomeric Mixture Separation

Even though some of the cation adduction methods to be presented did yield improvement in isomer separation, none has proven to be capable of even partially resolving every one of the five isomeric glycans in a mixture. With this in mind, differentiation of the isomers as a mixture necessitates additional compound information beyond drift time and, hence, further justifies the role of FT-IM-CID-MS in isomer separations. The fortunate circumstance of the dimers appearing at noticeably shorter drift times than the monomeric adducts due to higher charge permits not only the possibility of cleaner MS/MS spectra but also the capacity to explore the separation of monomeric species in samples containing each of the five glycan isomers. Each carbohydrate was analyzed in solution with each of the metals examined here as well as without added metals for the analysis of deprotonated glycan anions, all of which are shown in Figures S-2 through S-9 (Supporting Information). All but a few metal-glycan adduct combinations yield only one drift time peak in the monomers’ drift time window; those with multiple monomer drift time peaks have peak centroids determined for the most predominant peak. From the drift times obtained, the reduced mobility (Ko) and ion-neutral CCS could be determined for this system of glycans as each of the aforementioned ions or ion adducts (Table 2). Based on these ion-neutral CCS and Ko values obtained, it is apparent that the monomeric glycans adducted with transition metals form gas-phase coordination complexes that have the more compact structures relative to those adducted with either sodium or the alkaline earth metals. Although this is not surprising when considering the smaller ionic radii of transition metals [53], it is interesting that the transition metal clusters with glycans are even more compact than the lower mass and hypothetically smaller deprotonated glycan anions. Additionally, there is little variation in ion-neutral CCS and Ko among the three transition metals, even though the three cations have differing ionic radii. However, the behavior of the larger alkaline earth metal-glycan coordination adducts would indicate that larger ionic radii differences are necessary to yield noticeable changes in ion-neutral CCS and Ko. The alkaline earth metals used here encompass a larger range of ionic radii than the transition metals, and yet the alkaline earths have only subtle drift time increases with increasing ionic radius. This behavior indicates that the three transition metals do not differ substantially in their ability to separate this system of isomers. Similar work by Huang and Dodds involving divalent metal adduction for sugar separation has also explored glycan separation by adduction with Ca and Ba, in addition to some other divalent cations and a number of monovalent metal cations [16, 17]. In contrast to the glycan isomers evaluated here, Huang and Dodds assessed cation adduction effects on glycans with linear, homogeneous-residue tetrasaccharides as well with noticeably more elaborate structures. In one instance, two linear trisaccharide isomers were found to have roughly the same ion-neutral CCS when adducted with Ca, but another separation attempted with a pair of tetrasaccharides composed of heterogeneous residues showed potential for partial drift time separation. With this in mind, their results represent what can occur when the same metal adduction method is extended to substantially different systems. Moreover, this highlights how no one cation species is likely to be capable of providing universal metal-glycan adduct isomer separation, meaning that even metals providing poor separation as adducts herein may hold the potential to improve the resolution of other sets of isomers.
Table 2

Ko and Ion-Neutral CCS Values for Each Metal/Glycan Adduct Combination and Each Deprotonated Glycan Anion

 

Co

Cu

Mn

Ca

Sr

Ba

Na

–H

Ko (cm2/V-s)*

 Cellotetraitol

0.842 ± 0.007

0.840 ± 0.006

0.840 ± 0.006

0.795 ± 0.013

0.788 ± 0.007

0.791 ± 0.021

0.797 ± 0.006

0.783 ± 0.006

 Nigeran tetraitol

0.806 ± 0.006

0.809 ± 0.005

0.802 ± 0.006

0.796 ± 0.014

0.780 ± 0.010

0.771 ± 0.009

0.793 ± 0.007

0.815 ± 0.006

 Maltotetraitol

0.819 ± 0.006

0.826 ± 0.006

0.819 ± 0.005

0.804 ± 0.016

0.797 ± 0.013

0.799 ± 0.030

0.812 ± 0.009

0.835 ± 0.007

 Isomaltotetraitol

0.801 ± 0.006

0.805 ± 0.006

0.803 ± 0.007

0.790 ± 0.006

0.783 ± 0.006

0.777 ± 0.005

0.802 ± 0.005

0.819 ± 0.006

 Unnamed tetraitol

0.815 ± 0.005

0.818 ± 0.006

0.814 ± 0.006

0.786 ± 0.006

0.783 ± 0.008

0.779 ± 0.099

0.804 ± 0.005

0.820 ± 0.006

Ion-Neutral CCS (Å2)*

 Cellotetraitol

199 ± 2

199 ± 1

199 ± 2

211 ± 3

212 ± 2

212 ± 6

210 ± 2

214 ± 2

 Nigeran tetraitol

208 ± 2

207 ± 1

209 ± 2

211 ± 4

215 ± 3

217 ± 3

211 ± 2

206 ± 2

 Maltotetraitol

205 ± 1

203 ± 1

205 ± 1

208 ± 4

210 ± 3

210 ± 8

206 ± 2

201 ± 2

 Isomaltotetraitol

209 ± 2

208 ± 2

209 ± 2

212 ± 2

214 ± 2

215 ± 1

209 ± 1

205 ± 1

 Unnamed tetraitol

205 ± 1

205 ± 1

206 ± 2

213 ± 2

214 ± 2

215 ± 27

209 ± 1

205 ± 1

*Error values listed are population standard deviations of the arrival time distributions, from which a relative standard deviation was calculated and used to find error associated with Ko and ion-neutral CCS values. Although measurements made were internally consistent (mean %RSD <0.5%, median %RSD <0.15%), we acknowledge that the true associated errors stand to be higher than the measurement precision based on our experience with this system and generally accepted accuracy of drift time measurements [44, 55].

With the end goal of examining unknown mixtures of glycans, it is prudent to also explore isomer separation effects when the glycan standards are run together as a mixture. While composite spectra (i.e., equivalent to total ion chromatograms for fragment spectra) are a precursor to conclusive identification based upon drift time and fragment ion intensities, evaluation of the mixture can provide a wealth of information regarding the instrumental and analytical challenges facing the glycan analysis community. Figure 5 demonstrates how mixtures differ from the separation of individual standards, with lines overlaid on the mixture spectra positioned at the average drift time peak centroids for each isomer in the mixtures of the assessed metal adducts and deprotonated glycans. These vertical lines illustrate that sufficiently higher resolution is required for baseline separation. Nevertheless, combining this information with specific fragments can aid in the identification of partially resolved ion populations. Moreover, the additive effect of ion current from the individual fragments will cause isomers with overlapping drift times, and the same m/z to demonstrate less apparent separation when run as a mixture. This is a consequence of the mode of ion detection as a current, in which drift time peaks will sum—rather than simply overlap—whenever multiple overlapping isomeric species are analyzed simultaneously as opposed to simple overlap as from the separate analysis of single species. This phenomenon complicates sample mixture analysis beyond that of separated pure standards by reducing the apparent peak separation as the region of overlap exhibits the summed intensity from the multiple isomers, regardless of concentration. Another point worthy of note is the appearance of minute mobility features that appear for select metal-glycan complexes that may confound identification as mixtures. A close examination of the Mn-glycan spectra in Figure 5 and those for the individual standards (Figure S-5, Supplementary Material) illustrate the presence of stable, lower mobility features with a diminished intensity. Considering that some ligated conformations, even if they are derived from a single species, can adopt slightly different gas-phase conformations, direct quantitation may prove to be difficult without verification of individual standards. The separation selectivity induced by metal adduction can yield positive results, but care must be taken when transitioning this approach to quantitative efforts. Stated differently, in a mixture the possibility exists that additive ion current of a specific m/z at a given drift time can produce peak intensities that may not directly reflect total concentrations. As the Mn example illustrates, careful selection and characterization of each glycan-metal pair is essential as small yet noticeable shoulders or lower intensity peaks were observed for Na, Co, Ba, and Sr.
Figure 5

Arrival time distributions of mixtures containing all five tetraitol isomers (cellotetraitol, maltotetraitol, isomaltotetraitol, nigeran tetraitol, and the unnamed tetraitol) for each metal adduct or glycan ion examined here, which include cobalt adducts, manganese adducts, copper adducts, deprotonated glycans, sodium adducts, calcium adducts, strontium adducts, and barium adducts. Lines indicate mean peak centroid locations found for each individual metal-glycan adduct or deprotonated glycan standards, with symbols demarcating the identity of each isomeric glycan

As noted previously, all three of the transition metal adduct mixtures arrive at the shortest drift times of the mixtures analyzed. Additionally, the transition metals result in similar drift time separation and order of arrival times for the isomers, indicating that this glycan system experiences comparable interaction with the three transition metals under study. With each of the three transition metal-glycan mixtures, the sole β-linked tetrasaccharide—cellotetraitol—consistently has the shortest drift time and, thus, likely forms the most compact metal-glycan cluster. Following the β-linked, (1→4) positioned sugar, the analogous (1→4) but α-linked glycan maltotetraitol is consistently next in order of drift time, and is followed by the unnamed tetraitol, which has two α-(1→4) linkages and one α-(1→6). Nigeran tetraitol and isomaltotetraitol have alternately the longest drift times as adducts with the transition metals, thus forming transition metal-glycan adducts with the largest ion-neutral collision cross-sections.

In contrast to adduct formation with transition metal cations, the drift time spreads of alkaline earth metal and glycan adduct mixtures were not improved over the separation seen with sodium adducts. Furthermore, the order of glycan isomer drift times for the alkaline earth metal adducts is less consistent than that of the transition metals. However, the adduct drift times for each alkaline earth metal adduct mixture overall did consistently increase as the ionic radii of the alkaline earth metal cations increase. Another noticeable difference between the transition metal adducts and the alkaline earth metal adducts is the higher level of noise as shown in Figures S-2 through S-9 (Supporting Information). It is suspected that the adduction differences may be partially rooted in the ionic radii of the cations. Even in the presence of the other metal cations, the sodiated adduct is consistently formed at a high abundance. With an effective ionic radius of ~0.99 Å for the sodium cation, the Ca cation and transition metal cations are approximately at or below sodium’s ionic radius [53], in contrast to the Sr and Ba cations, which are approximately 0.19 and 0.36 Å greater than the sodium cation’s ionic radius, respectively. The larger ionic radii of Ba and Sr correlate with weaker coordination by the glycans, resulting in the lower ion current as well as noisier drift time and mass spectra for the adducts. Alkali metal adduction investigated by Cancilla et al. followed a similar trend according to ionic radii, although all quasimolecular ions of the metal and glycans were formed for pentaose and hexaose, some of the larger ionic radii cations failed to coordinate well enough with maltose, triose, and tetraose for their corresponding quasimolecular ions to appear [11]. In the Ca adduct mixture alone, the presence of a higher-mobility feature was noticeable adjacent to the monomeric adduct peak that has not been found in any of the single isomer Ca adduct arrival time distributions. Due to the higher baseline found in the Ca adduct drift time profile, definitive conclusions regarding the peak’s origin remain tenuous. However, the presence of higher mobility features may be due to the ability of the multivalent metals to coordinate differently within a single glycan.

Also shown in Figure 5 is the arrival time distribution for the deprotonated ions of the glycan mixture. In comparison to the metal adducts, the two isomers cellotetraitol and maltotetraitol are essentially completely separated as anions, whereas the other three α-linked positional isomers have drift times that are statistically identical. IM-MS of negatively charged glycan ions has not received as much attention as protonated glycans or cation adducts have over the years [54], but recent work on the use of negative ion monitoring for glycan analysis has brought more attention to anionic glycan separation. Although in many cases glycan anion separation has proven to be successful for isomer distinction in arrival time distributions, accounting for the expansive range of glycan structures suggests that anion separation may be optimal for some systems, but not others. An example of this is shown in Figure 5, where the tetraitols with the same linkage positions but different anomeric configurations—cellotetraitol and maltotetraitol—underwent the most substantial separation as deprotonated ions than as any of the metal adducts. However, the other isomers included in this experiment—isomaltotetraitol, nigeran tetraitol, and unnamed tetraitol—were closer together in drift time than any set of three glycan-metal adducts. Consequently, the separation of deprotonated glycans was useful here for separating the α-linked and β-linked isomers, but not for differentiation between the other glycans with α-linkage.

For additional perspective on how the combination of metal adduction, mobility separation, and CID has the potential to provide simplified isomer separation, we have included 3D waterfall plots of FT-IM-CID-MS data in Figure 6 that correspond to spectra presented in Figure 4 and fragments outlined in Table 1. Figure 6a shows the resultant plot from the full five-glycan mixture with added cobalt. Here, four of the five isomers, including all but the unnamed tetraitol, can be clearly identified based on both their characteristic fragments and corresponding drift times. The six primary fragments produced by the five tetraitols as adducts with cobalt (Figure 4, positive MS spectra) are shown in waterfall IMS-MS mode as a mixture in Figure 6. Cellotetraitol is identified as the first glycan in the arrival time distribution based on the appearance of its fragments at m/z 544 and 382. Following cellotetraitol in drift time is maltotetraitol, which is presented as a singular peak at m/z 544. Nigeran tetraitol is readily identified following the maltotetraitol drift time peak based on the appearance of nigeran tetraitol’s unique combination of fragments at m/z 564, 544, 402, 384, and 364. Isomaltotetraitol is the last glycan visible among the mixture as indicated by fragment peaks at m/z 564 and 402 at a drift time following the nigeran tetraitol fragment peaks. Out of the five glycans, the unnamed tetraitol is not readily determined among the mixture; this outcome is not particularly surprising as unnamed tetraitol has consistently yielded metal-glycan adducts and deprotonated ions at ion abundances that are roughly one order of magnitude lower than any of the four other glycans. Figure 6b serves to further demonstrate the separation capability of the cobalt-glycan adducts by showing a 3D waterfall plot of the FT-IM-CID-MS analysis of the three easily separated cobalt-glycan adducts, which includes cellotetraitol, maltotetraitol, and isomaltotetraitol. Again, the identifiable fragments for these three cobalt-glycan isomers fall in line with the drift time isolated CID spectra in Figure 4 and fragments from Table 1, indicating the reproducibility of the approach. Furthermore, Figure 6 demonstrates how drift time-specific fragment separation promotes noise reduction and more facile isomer recognition because some product ions (such as m/z 382 and 402) are only found in specific subsets of isomeric precursor fragmentation patterns. When considering that the exclusion of extraneous dimeric peaks can be combined with distinctive fragmentation and isomer drift time separation, the viability of the FT-IM-CID-MS approach to attain species-specific fragmentation patterns is apparent.
Figure 6

Three-dimensional waterfall plots showing the fragments produced in cobalt-glycan mixtures with their separation according to each isomeric glycan-cobalt adduct precursor drift times. A full mixture of all five tetraitol isomers (cellotetraitol, maltotetraitol, isomaltotetraitol, nigeran tetraitol, and the unnamed tetraitol) is shown in (a), whereas a mixture of the three isomers with the broadest drift time separation—cellotetraitol, maltotetraitol, and isomaltotetraitol—are shown in (b). The characteristic fragments produced are indicated with lines along the drift time axis at the appropriate m/z, with the nominal masses of fragments given in red text

Conclusions

We present here a systematic evaluation using divalent metal cations for the drift time separation of five isomeric tetrasaccharide-alditols. For the purpose of isomeric glycan mixture separation, it is becoming readily apparent that the various techniques to improve drift time resolution cannot be applied as one-size-fits-all but need to be correlated with the structures under investigation. In select cases, substantial utility appears when separating glycan isomers by deprotonation; however, transition metal adduction yielded noticeable improvements over glycan anion separation for the particular isomer system examined here.

Furthermore, what has been successfully shown with the FT-IM-CID-MS spectra of metal-glycan adducts is that multiple precursor subpopulations may be present (monomeric and dimer species) that would simply not be separated when using direct electrospray of analytes selecting precursor ions by m/z alone, and that a typical dissociation spectrum may not be representative of the target analyte or even be very reproducible. Thus, for the analysis of specific isomers or even a given peptidoglycan, the direct fragmentation approach may not provide the entire picture as to which precursor species contribute to each product ion fragment. When tandem mass spectrometry is used in conjunction with a high duty cycle, high resolution mobility separation, fragmentation patterns can be isolated that are specifically associated with a precursor ion’s drift time. Correspondingly, ion mobility separation prior to all biomolecular mass spectrometric analyses, or even all isomeric analyses, may prove to be vital in discerning the appropriate precursor. The dimers found here would never have been distinguished from monomers in terms of CID spectra without assessment with IMS, so to assume that other systems of complex mixtures would solely include monomeric species appears unlikely. Furthermore, in an automated peak-picking scenario where only the major peaks are recorded, the fragment contributions from the dimer and monomer could easily be mingled without ion mobility separation. In the event of possible protomers and/or isomers confounding mass spectra produced, ion mobility separation remains an important technique capable of rapidly assessing the presence of multiple precursor subpopulations for a given m/z. Notably, the broader use of ion mobility separation may enable better reproducibility in biomolecule fragmentation since concentration and ionization conditions may very well alter the ratios of species formed and contribute to spectral fluctuation, whereas isolation of precursor subpopulations in ion mobility windows can yield mass spectra of selected ionic states. It is also worthy of note that dimeric species having the same m/z as monomers but different dissociation spectra complicate spectra interpretations when looking solely at ratios of ion yields.

Given the arduous nature of glycan purification and structure verification, the development of a predictive capacity to aid in the spectral interpretation of low abundance species is highly desired. Although the computational framework exists to place computed structure in the context of the mobility separation, integration of the different metal ions along with their potential coordination sites remains challenging. As the range of verified glycan standards examined by IMS expands, the range of input structures will surely increase. Integration of molecular simulations with experimental values, especially for the metal-glycan complexes, remains an area of active research within our laboratory and the broader community.

Notes

Acknowledgments

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.

Supplementary material

13361_2016_1505_MOESM1_ESM.docx (1.5 mb)
ESM 1 (DOCX 1580 kb)

References

  1. 1.
    Kaddis, C.S., Loo, J.A.: Native protein MS and ion mobility. Anal. Chem. 79, 1779–1784 (2007)CrossRefGoogle Scholar
  2. 2.
    Armenta, S., Alcala, M., Blanco, M.: A review of recent, unconventional applications of ion mobility spectrometry (IMS). Anal. Chim. Acta 703, 114–123 (2011)CrossRefGoogle Scholar
  3. 3.
    Hall, Z., Politis, A., Bush, M.F., Smith, L.J., Robinson, C.V.: Charge state-dependent compaction and dissociation of protein complexes: insights from ion mobility and molecular dynamics. J. Am. Chem. Soc. 134, 3429–3438 (2012)CrossRefGoogle Scholar
  4. 4.
    Gray, C.J., Thomas, B., Upton, R., Migas, L.G., Eyers, C.E., Barran, P.E., Flitsch, S.L.: Applications of ion mobility mass spectrometry for high throughput, high resolution glycan analysis. Biochim. Biophys. Acta. 1860, 1688–1709 (2016)Google Scholar
  5. 5.
    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)CrossRefGoogle Scholar
  6. 6.
    Zhu, M., Bendiak, B., Clowers, B., Hill, H.H.: Ion mobility-mass spectrometry analysis of isomeric carbohydrate precursor ions. Anal. Bioanal. Chem. 394, 1853–1867 (2009)CrossRefGoogle Scholar
  7. 7.
    Mårtensson, S., Levery, S.B., Fang, T.T., Bendiak, B.: Neutral core oligosaccharides of bovine submaxillary mucin use of lead tetraacetate in the cold for establishing branch positions. Eur. J. Biochem. 258, 603–622 (1998)CrossRefGoogle Scholar
  8. 8.
    Kailemia, M.K., Ruhaak, L.R., Lebrilla, C.B., Amster, I.J.: Oligosaccharide analysis by mass spectrometry: a review of recent developments. Anal. Chem. 86, 196–212 (2014)CrossRefGoogle Scholar
  9. 9.
    Williams, D., Young, M.K.: Analysis of neutral isomeric low molecular weight carbohydrates using ferrocenyl boronate derivatization and tandem electrospray mass spectrometry. Rapid Commun. Mass Spectrom. 14, 2083–2091 (2000)CrossRefGoogle Scholar
  10. 10.
    Clowers, B.H., Dwivedi, P., Steiner, W.E., Hill, H.H., Bendiak, B.: Separation of sodiated isobaric disaccharides and trisaccharides using electrospray ionization-atmospheric pressure ion mobility-time of flight mass spectrometry. J. Am. Soc. Mass Spectrom. 16, 660–669 (2005)CrossRefGoogle Scholar
  11. 11.
    Cancilla, M.T., Penn, S.G., Carroll, J.A., Lebrilla, C.B.: Coordination of alkali metals to oligosaccharides dictates fragmentation behavior in matrix assisted laser desorption ionization Fourier transform mass spectrometry. J. Am. Chem. Soc. 118, 6736–6745 (1996)CrossRefGoogle Scholar
  12. 12.
    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)CrossRefGoogle Scholar
  13. 13.
    Smith, G., Kaffashan, A., Leary, J.A.: Influence of coordination number and ligand size on the dissociation mechanisms of transition metal monosaccharide complexes. Int. J. Mass Spectrom. 183, 299–310 (1999)CrossRefGoogle Scholar
  14. 14.
    Harvey, D.J.: Ionization and collision-induced fragmentation of N-linked and related carbohydrates using divalent cations. J. Am. Soc. Mass Spectrom. 12, 926–937 (2001)CrossRefGoogle Scholar
  15. 15.
    Huang, Y., Dodds, E.D.: Ion mobility studies of carbohydrates as group i adducts: isomer specific collisional cross section dependence on metal ion radius. Anal. Chem. 85, 9728–9735 (2013)CrossRefGoogle Scholar
  16. 16.
    Huang, Y., Dodds, E.D.: Discrimination of isomeric carbohydrates as the electron transfer products of group II cation adducts by ion mobility spectrometry and tandem mass spectrometry. Anal. Chem. 87, 5664–5668 (2015)CrossRefGoogle Scholar
  17. 17.
    Huang, Y., Dodds, E.D.: Ion-neutral collisional cross sections of carbohydrate isomers as divalent cation adducts and their electron transfer products. Analyst 140, 6912–6921 (2015)CrossRefGoogle Scholar
  18. 18.
    Adamson, J.T., Håkansson, K.: Electron capture dissociation of oligosaccharides ionized with alkali, alkaline earth, and transition metals. Anal. Chem. 79, 2901–2910 (2007)CrossRefGoogle Scholar
  19. 19.
    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)CrossRefGoogle Scholar
  20. 20.
    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)CrossRefGoogle Scholar
  21. 21.
    Racaud, A., Antoine, R., Dugourd, P., Lemoine, J.: Photoinduced dissociation of heparin-derived oligosaccharides controlled by charge location. J. Am. Soc. Mass Spectrom. 21, 2077–2084 (2010)CrossRefGoogle Scholar
  22. 22.
    Brown, D.J., Stefan, S.E., Berden, G., Steill, J.D., Oomens, J., Eyler, J.R., Bendiak, B.: Direct evidence for the ring opening of monosaccharide anions in the gas phase: photodissociation of aldohexoses and aldohexoses derived from disaccharides using variable-wavelength infrared irradiation in the carbonyl stretch region. Carbohydr. Res. 346, 2469–2481 (2011)CrossRefGoogle Scholar
  23. 23.
    Domon, B., Mueller, D.R., Richter, W.J.: Tandem mass spectrometric analysis of fixed-charge derivatized oligosaccharides. Org. Mass Spectrom. 29, 713–719 (1994)CrossRefGoogle Scholar
  24. 24.
    Cheng, H.L., Her, G.R.: Determination of linkages of linear and branched oligosaccharides using closed-ring chromophore labelling and negative ion trap mass spectrometry. J. Am. Soc. Mass Spectrom. 13, 1322–1330 (2002)Google Scholar
  25. 25.
    Saba, J.A., Shen, X., Jamieson, J.C., Perreault, H.: Investigation of different combinations of derivatization, separation methods and electrospray ionization mass spectrometry for standard oligosaccharides and glycans from ovalbumin. J. Mass Spectrom. 36, 563–574 (2001)CrossRefGoogle Scholar
  26. 26.
    Konda, C., Londry, F.A., Bendiak, B., Xia, Y.: Assignment of the stereochemistry and anomeric configuration of sugars within oligosaccharides via overlapping disaccharide ladders using MSn. J. Am. Soc. Mass Spectrom. 25, 1441–1450 (2014)CrossRefGoogle Scholar
  27. 27.
    Zaikin, V., Halket, J.M.: A handbook of derivatives for mass spectrometry. IM Publications, Chichester, UK (2009)Google Scholar
  28. 28.
    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 oligosaccharides. Eur. J. Biochem. 171, 643–654 (1988)CrossRefGoogle Scholar
  29. 29.
    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)CrossRefGoogle Scholar
  30. 30.
    Maxwell, E.J., Chen, D.D.Y.: Twenty years of interface development for capillary electrophoresis-electrospray ionization-mass spectrometry. Anal. Chim. Acta 627, 25–33 (2008)CrossRefGoogle Scholar
  31. 31.
    Costello, C.E., Contado-Miller, J.M., Cipollo, J.F.: A glycomics platform for the analysis of permethylated oligosaccharide alditols. J. Am. Soc. Mass Spectrom. 18, 1799–1812 (2007)CrossRefGoogle Scholar
  32. 32.
    Wyttenbach, T., Bowers, M.T.: Gas-phase conformations: the ion mobility/ion chromatography method. In: Schalley, C.A. (ed.) Modern Mass Spectrometry, pp. 207–232. Springer, Berlin (2003)CrossRefGoogle Scholar
  33. 33.
    Leavell, M.D., Gaucher, S.P., Leary, J.A., Taraszka, J.A., Clemmer, D.E.: Conformational studies of Zn-ligand-hexaose diasteromers using ion mobility measurements and density functional theory calculations. J. Am. Soc. Mass Spectrom. 13, 284–293 (2002)Google Scholar
  34. 34.
    Zhu, F., Lee, S., Valentine, S.J., Reilly, J.P., Clemmer, D.E.: Mannose7 glycan isomer characterization by IMS-MS/MS analysis. J. Am. Soc. Mass Spectrom. 23, 2158–2166 (2012)CrossRefGoogle Scholar
  35. 35.
    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)CrossRefGoogle Scholar
  36. 36.
    Li, H., Bendiak, B., Siems, W.F., Gang, D.R., Hill, H.H.: Determining the isomeric heterogeneity of neutral oligosaccharide-alditols of bovine submaxillary mucin using negative ion traveling wave ion mobility mass spectrometry. Anal. Chem. 87, 2228–2235 (2015)CrossRefGoogle Scholar
  37. 37.
    Henderson, S.C., Valentine, S.J., Counterman, A.E., Clemmer, D.E.: ESI / ion trap / ion mobility / time-of-flight mass spectrometry for rapid and sensitive analysis of biomolecular mixtures. Solutions 71, 291–301 (1999)Google Scholar
  38. 38.
    Clowers, B.H., Ibrahim, Y.M., Prior, D.C., Danielson, W.F., Belov, M.E., Smith, R.D.: Enhanced ion utilization efficiency using an electrodynamic ion funnel trap as an injection mechanism for ion mobility spectrometry. Anal. Chem. 80, 612–623 (2008)CrossRefGoogle Scholar
  39. 39.
    Zhu, F., Glover, M.S., Shi, H., Trinidad, J.C., Clemmer, D.E.: Populations of metal-glycan structures influence MS fragmentation patterns. J. Am. Soc. Mass Spectrom. 26, 25–35 (2014)CrossRefGoogle Scholar
  40. 40.
    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)CrossRefGoogle Scholar
  41. 41.
    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)CrossRefGoogle Scholar
  42. 42.
    Kurulugama, R.T., Valentine, S.J., Sowell, R.A., Clemmer, D.E.: Development of a high-throughput IMS-IMS-MS approach for analyzing mixtures of biomolecules. J. Proteome 71, 318–331 (2008)CrossRefGoogle Scholar
  43. 43.
    Clowers, B.H., Siems, W.F., Yu, Z., Davis, A.L.: A two-phase approach to Fourier transform ion mobility time-of-flight mass spectrometry. Analyst 140, 6862–6870 (2015)CrossRefGoogle Scholar
  44. 44.
    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)CrossRefGoogle Scholar
  45. 45.
    Ceroni, A., Maass, K., Geyer, H., Geyer, R., Dell, A., Haslam, S.M.: GlycoWorkbench: a tool for the computer-assisted annotation of mass spectra of glycans. J. Proteome Res. 7, 1650–1659 (2008)CrossRefGoogle Scholar
  46. 46.
    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)CrossRefGoogle Scholar
  47. 47.
    Varki, A., Cummings, R.D., Aebi, M., Packer, N.H., Seeberger, P.H., Esko, J.D., Stanley, P., Hart, G., Darvill, A., Kinoshita, T., Prestegard, J.J., Schnaar, R.L., Freeze, H.H., Marth, J.D., Bertozzi, C.R., Etzler, M.E., Frank, M., Vliegenthart, J.F.G., Lutteke, T., Perez, S., Bolton, E., Rudd, P., Paulson, J., Kanehisa, M., Toukach, P., Aoki-Kinoshita, K.F., Dell, A., Narimatsu, H., York, W., Taniguchi, N., Kornfeld, S.: Symbol nomenclature for graphical representations of glycans. Glycobiology 25, 1323–1324 (2015)CrossRefGoogle Scholar
  48. 48.
    Harvey, D.J., Sobott, F., Crispin, M., Wrobel, A., Bonomelli, C., Vasiljevic, S., Scanlan, C.N., Scarff, C.A., Thalassinos, K., Scrivens, J.H.: Ion mobility mass spectrometry for extracting spectra of n-glycans directly from incubation mixtures following glycan release: application to glycans from engineered glycoforms of intact, folded HIV gp120. J. Am. Soc. Mass Spectrom. 22, 568–581 (2011)CrossRefGoogle Scholar
  49. 49.
    Li, H., Bendiak, B., Kaplan, K., Davis, E., Siems, W.F., Hill, H.H.: Evaluation of ion mobility-mass spectrometry for determining the isomeric heterogeneity of oligosaccharide-alditols derived from bovine submaxillary mucin. Int. J. Mass Spectrom. 352, 9–18 (2013)CrossRefGoogle Scholar
  50. 50.
    Pu, Y., Ridgeway, M.E., Park, M.A., Lin, C., Costello, C.E.: Separation and identification of isomeric glycans by trapped ion mobility spectrometry-Fourier transform mass spectrometry. Anal. Chem. 88, 3–4 (2016)Google Scholar
  51. 51.
    Weiskopf, A.S., Vouros, P., Harvey, D.J.: Characterization of oligosaccharide composition and structure by quadrupole ion trap mass spectrometry. Rapid Commun. Mass Spectrom. 11, 1493–1504 (1997)CrossRefGoogle Scholar
  52. 52.
    Domon, B., Costello, C.E.: A systematic nomenclature for carbohydrate fragmentations in FAB-MS/MS spectra of glycoconjugates. Glycoconj. J. 5, 397–409 (1988)CrossRefGoogle Scholar
  53. 53.
    Shannon, R.D.: Revised effective ionic radii and systematic studies of interatomic distances in halides and chaleogenides. Acta Cryst 32, 751–767 (1976)CrossRefGoogle Scholar
  54. 54.
    Čmelík, R., Štikarovská, M., Chmelík, J.: Different behavior of dextrans in positive-ion and negative-ion mass spectrometry. J. Mass Spectrom. 39, 1467–1473 (2004)CrossRefGoogle Scholar
  55. 55.
    Crawford, C.L., Hauck, B.C., Tufariello, J.A., Harden, C.S., McHugh, V., Siems, W.F., Hill, H.H.: Accurate and reproducible ion mobility measurements for chemical standard evaluation. Talanta 101, 161–170 (2012)CrossRefGoogle Scholar

Copyright information

© American Society for Mass Spectrometry 2016

Authors and Affiliations

  • Kelsey A. Morrison
    • 1
  • Brad K. Bendiak
    • 2
  • Brian H. Clowers
    • 1
  1. 1.Department of ChemistryWashington State UniversityPullmanUSA
  2. 2.Department of Cell and Developmental Biology and Program in Structural Biology and BiochemistryUniversity of Colorado Health Sciences CenterAuroraUSA

Personalised recommendations