Determining the Isomeric Heterogeneity of Neutral Oligosaccharide

Jan 16, 2015 - Department of Cell and Developmental Biology, Program in Structural Biology and Biophysics, University of Colorado, Health Sciences Cen...
0 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/ac

Determining the Isomeric Heterogeneity of Neutral OligosaccharideAlditols of Bovine Submaxillary Mucin Using Negative Ion Traveling Wave Ion Mobility Mass Spectrometry Hongli Li,† Brad Bendiak,‡ William F. Siems,† David R. Gang,§ and Herbert H. Hill, Jr.*,† †

Department of Chemistry, Washington State University, Pullman, Washington 99164, United States Department of Cell and Developmental Biology, Program in Structural Biology and Biophysics, University of Colorado, Health Sciences Center, Anschutz Medical Campus, Aurora, Colorado 80045, United States § Institute of Biological Chemistry, Washington State University, Pullman, Washington 99164, United States ‡

S Supporting Information *

ABSTRACT: Negative ions produced by electrospray ionization were used to evaluate the isomeric heterogeneity of neutral oligosaccharidealditols isolated from bovine submaxillary mucin (BSM). The oligosaccharide-alditol mixture was preseparated on an off-line highperformance liquid chromatography (HPLC) column, and the structural homogeneity of individual LC fractions was investigated using a Synapt G2 traveling wave ion mobility spectrometer coupled between quadupole and time-of-flight mass spectrometers. Mixtures of isomers separated by both chromatography and ion mobility spectrometry were studied. Tandem mass spectrometry (MS/MS) of multiple mobility peaks having the same mass-to-charge ratio (m/z) demonstrated the presence of different structural isomers and not differences in ion conformations due to charge site location. Although the oligosaccharide-alditol mixture was originally separated by HPLC, multiple ion mobility peaks due to structural isomers were observed for a number of oligosaccharide-alditols from single LC fractions. The collision-induced dissociation cells located in front of and after the ion mobility separation device enabled oligosaccharide precursor or product ions to be separated by ion mobility and independent fragmentation spectra to be acquired for isomeric carbohydrate precursor or product ions. MS/MS spectra so obtained for independent mobility peaks at a single m/z demonstrated the presence of structural variants or stereochemical isomers having the same molecular formula. This was observed both for oligosaccharide precursor and product ions. In addition, mobilities of both [M − H]− and [M + Cl]− ions, formed by adding NH4OH or NH4Cl to the electrospray solvent, were examined and compared for selected oligosaccharide-alditols. Better separation among structural isomers appeared to be achieved for some [M + Cl]− anions.

A

Thus, the physical isolation of isomeric carbohydrates from a complex mixture is highly preferred for structural analysis using collision-induced dissociation (CID) experiments. With liquid chromatography (LC) coupled to MS,10,13−16 several columns17−19 have been able to differentiate isobaric oligosaccharides. However, LC−MS can be time-consuming, both in establishing the method and performing the separation; sample derivatization is sometimes required prior to analysis. Frequently one specific column is only suitable for the separation of a certain class of glycan isomers. In addition, multidimensional orthogonal LC methods are often needed in order to purify structures to homogeneity as evaluated by NMR.20−22 Other techniques like gas chromatography (GC)23−25 or capillary electrophoresis (CE)26−28 typically

s one of the most predominant post-translational modifications in biological systems, glycosylation contributes to a vast functional diversity in nature.1−5 Carbohydrate compounds, with 80−90% of carbons being chiral in addition to having positional linkage isomers, often exist as a complex population displaying considerable isomeric heterogeneity among molecules.6,7 The capability to rapidly separate various oligosaccharide isomers from a mixture is an advance toward structural analysis of independent carbohydrate precursor isomers and furthers development of glycomics research. Mass spectrometry (MS) has become a sensitive and popular tool for the determination and identification of oligosaccharides isolated from glycoproteins.8−12 However, MS analysis of carbohydrate compounds has been challenging due to the frequent presence of large numbers of oligosaccharide isomers. MSn information is unreliable without the separation of isomeric mixtures as well as the establishment of unambiguous precursor−product ion relationships for specific structures. © 2015 American Chemical Society

Received: October 4, 2014 Accepted: January 16, 2015 Published: January 16, 2015 2228

DOI: 10.1021/ac503754k Anal. Chem. 2015, 87, 2228−2235

Article

Analytical Chemistry

oligosaccharide-alditols from a BSM mixture using IMMS in the positive ion mode.55 Multiple isomers were separated using ion mobility TOF MS. A specially home-built dual-gate ion mobility quadrupole ion trap MS was utilized to acquire mobility-separated tandem MS spectra for the precursor ions, where the isomeric heterogeneity of oligosaccharide product ions was not evaluated. Here we report a negative mode study using a traveling wave IMMS to evaluate the isomeric heterogeneity of neutral oligosaccharide-alditols from BSM. The unique capabilities of IMMS for isomer differentiation and its advantages in assessing stereostructural differences of glycans within biological carbohydrate mixtures were demonstrated. Individual LC fractions from the separation of an oligosaccharide-alditol mixture were used as target analytes; carbohydrate isomeric heterogeneity within single LC fractions was investigated. The analysis of oligosaccharide-alditols included isomer separation and acquisition of mobility-resolved fragmentation spectra for both precursor and product ions using multiple functions of the Synapt G2 instrument. While demonstrated here for complex sets of carbohydrate isomers, the general capabilities of the instrument for assessment and fingerprinting of complex sets of isomers based on IMS profiles of product ions and their dissociation patterns should be applicable to a wide variety of compounds.

show higher resolution separations, but closely related isomers have some probability of comigration in any system; there is currently no universal single-separation method for resolution of carbohydrate isomers. Rapid, sensitive, efficient, and robust analytical methodologies are needed to address the analytical challenge of structural identification of isomers in carbohydrate analysis. Preferably, such methods should employ multiple orthogonal types of separations of high resolution, each with the capability of independently resolving many isomeric species. Ion mobility mass spectrometry29 (IMMS) is capable of rapidly differentiating isomers using carbohydrate standards30−33 and mixtures of glycans from biological sources.34−36 Separation is dependent on the drift velocities of gas-phase ions under an electric field in a counter flow of drift gas through an ion mobility spectrometry (IMS) drift tube prior to MS.37,38 Recently, a Synapt G2 high-definition mass spectrometer (HDMS), utilizing a traveling wave ion mobility instrument coupled to mass spectrometry,39−42 has been used in a wide range of applications.34,43−48 The G2 HDMS is a hybrid quadrupole/traveling wave ion mobility/orthogonal time-offlight (TOF) MS. Additionally, a trap cell is located in front of the traveling wave ion mobility separator and a transfer cell behind it, whereby ions can be dissociated at either location. This special design enables both mobility-resolved precursor ions to be fragmented and mobility-separated product ions to be dissociated. Multiple applications of traveling wave ion mobility mass spectrometry (TWIMMS) for carbohydrate analysis have been reported.34,49−52 Yamagaki and Sato49 found that the traveling wave ion mobility peak-width mass correlation can be used for structural analysis and isomer differentiation using cyclodextrin oligosaccharide standards as examples. Williams et al. demonstrated the capability of TWIMMS for the separation of isomeric oligosaccharide standards as well as glycan mixtures,34 where mobility-separated MS/MS spectra were also shown. The application of matrixassisted laser desorption ionization (MALDI)-MS/MS traveling wave ion mobility for structural analysis of N-linked glycans illustrated the value of MALDI tandem MS for profiling Nlinked glycans with the unique advantage of isomer detection when combined with TWIMS.51 However, the majority of separations of glycans by IMS have been performed for the precursor ions in the positive mode. Mass spectrometry of oligosaccharides in the negative ion mode also yields structural information that is valuable and often complementary to that obtained in the positive ion mode, with or without derivatization. For example, disaccharides dissociated in the negative ion mode yield one key product ion, a glycosylglycoaldehyde; it enables the stereochemistry and monosaccharide anomeric configuration to be assigned, which is not possible to obtain in the positive ion mode.53 More importantly, the isomeric heterogeneity of oligosaccharide product ions derived from biological mixtures has not been systematically evaluated yet in both positive and negative ion modes. Liu and Clemmer54 studied the mobility of negative product ions derived from simple oligosaccharide standards in 1997 by fragmenting precursor ions with high injection energy, although no tandem MS data were reported. Martensson et al.19 previously reported that neutral O-linked oligosaccharide-alditols isolated from bovine submaxillary mucin (BSM) contained a large number of isomers. Thus, BSM serves as a good paradigm for glycomics of O-linked oligosaccharides. The authors recently reported an application for the determination of isomeric heterogeneity among

■

EXPERIMENTAL SECTION Materials. All the solvents (methanol, water, and acetonitrile) were Optima LC/MS grade and supplied by Fisher Chemical (Thermo Fisher Scientific Inc.). All the chemicals (NaOH, NaBH4, NH4OH, and NH4Cl) plus the glycoprotein of bovine submaxillary mucin (BSM) were purchased from Sigma-Aldrich (St. Louis, MO) and used as received. An equal volume mixture of methanol and water was used to prepare the electrospray ionization (ESI) solvent. The preparation and separation of the neutral oligosaccharide-alditol mixture were described in detail by Martensson et al.19 The high-performance liquid chromatography (HPLC) elution profile of the BSM neutral oligosaccharide-alditol mixture was reported previously55 and is included in the Supporting Information (Figure S1). Also included in the Supporting Information are the mass spectra (Figure S1) and mass lists (Table S1) for different HPLC fractions in the negative ion mode. ESI solvent (1 mL) was added to each HPLC fraction (10 fractions in total, dried sample) to prepare stock solutions. Different dilution ratios were used for different fractions, resulting in a final solution having a total concentration of ∼0.03 mg·mL−1 carbohydrate for each fraction. To promote the formation of deprotonated and chloride adducted anions in the negative ion mode, 25 μL each of NH4OH and NH4Cl stock solutions (1 mM in ESI solvent) were added to 1 mL of individual diluted oligosaccharide-alditol LC fractions, resulting in concentrations of 25 μM NH4OH and 25 μM NH4Cl in the solution. Traveling Wave Ion Mobility Mass Spectrometry Measurements. All experiments were performed on a Synapt G2 high-definition mass spectrometer (HDMS) (WatersMicromass Corp., Milford, MA, U.S.A.). The instrument was composed of a quadrupole, a trap cell, a traveling wave ion mobility separator, a transfer cell, and a TOFMS in sequential order. Several operation modes were employed in this work: (1) MS mode, the system was used only as a mass analyzer to acquire the mass values of neutral oligosaccharide-alditols in the different LC fractions; (2) ion mobility MS mode, the ion 2229

DOI: 10.1021/ac503754k Anal. Chem. 2015, 87, 2228−2235

Article

Analytical Chemistry

Figure 1. Overlaid mobility profiles of individual oligosaccharides in multiple LC fractions. (a) Overlaid mobility profile of the m/z 623.2 precursor ions in LC fractions 4, 5, and 6. (b) Overlaid mobility profiles of the m/z 628.3 precursor ions in LC fractions 4, 5, and 6. (c) Overlaid mobility profile of the m/z 769.3 precursor ions in LC fractions 5, 6, and 7. (d) Overlaid mobility profiles of the m/z 895.3 precursor ions in LC fractions 8 and 9. (e) Overlaid mobility profiles of the m/z 972.4 precursor ions in LC fractions 8 and 9. (f) Overlaid mobility profiles of the m/z 1041.4 precursor ions in LC fractions 9 and 10.

e.g., using [M − H]− ions as examples, m/z 425.2 was detected in fractions 2 and 3, m/z 587.2 and 628.3 appeared in fractions 4, 5 and 6, m/z 733.3 was seen in fractions 5, 6 and 7, m/z 895.3 and 936.4 were in both fractions 8 and 9, and m/z 1041.4 was observed in fractions 9 and 10. This complicated mass distribution prompted the following questions: (1) How would one know whether compounds having the same m/z in different fractions are isomers or not? (2) Is an oligosaccharide having a specific m/z from a single HPLC fraction one pure isomer? (3) How can the isomeric nature of oligosaccharide-alditols for both precursor and product ions be best evaluated? Evaluation of Oligosaccharide-Alditol Precursor Ions for Isomeric Homogeneity. All individual HPLC fractions from fractions 2 to 10 were analyzed individually by traveling wave ion mobility. The data was acquired in 3 min for each fraction. Figure 1 shows overlaid mobility profiles for the same oligosaccharide-alditol precursor ions found in different HPLC fractions. Figure 1a contains overlaid mobility spectra for the m/z 623.2 ions that were observed in fractions 4, 5, and 6. This is an [M + Cl] − ion that has the composition of Hex1HexNAc1HexNAc-ol. The proposed sugar compositions in this work are based on the previous study of BSM oligosaccharide-alditol standards19 and further validated by the m/z values and the fragmentation spectra. Two mobility peaks for m/z 623.2 were clearly observed with fractions 5 and 6 having identical mobility profiles. The predominant mobility peak (drift time 4.4 ms) in fractions 5 and 6 showed the same mobility as the minor peak in fraction 4, and vice versa for the peaks observed from fraction 4. The isomer(s) with the smaller collision cross section (shorter mobility drift time) eluted in fractions 5 and 6, while the isomer with the larger collision cross section (longer mobility drift time) eluted in fraction 4. When two mobilities are observed for the same m/z value, it is possible that the separate mobilities arise from a single structure that has been ionized in different locations. A number of BSM oligosaccharide-alditol standards purified by HPLC on multiple columns19 were analyzed previously using different ion mobility mass spectrometry systems in our laboratory.31,56 A general feature of the reduced alditol precusors was that they run as a single ion mobility peak. Interestingly, structures that have

mobility separation was utilized to evaluate the isomeric heterogeneity of oligosaccharide-alditol precursor ions with knowledge of precursor mass-to-charge ratio (m/z) values; (3) ion mobility−MS/MS mode, the quadrupole was set to allow precursor ions having a specific m/z to pass through the trap into the traveling wave for ion mobility separation and the mobility-separated precursor ions were dissociated by collision in the transfer cell to obtain their fragmentation spectra in the TOFMS; (4) MS/MS−ion mobility mode, precursor ions were filtered by the quadrupole and fragmented in the trap cell prior to IMS analysis. Isomeric product ions were evaluated by traveling wave IMS; (5) MS/MS−ion mobility−MS/MS mode, in addition to the process described in method 4, fragmentation of the mobility-resolved product ions were obtained in the transfer cell; this is referred to as time aligned parallel (TAP) fragmentation. The ion trap and the ion transfer cell were converted to collision-induced dissociation (CID) cells by increasing the collision energy (CE) in each cell. The detailed instrumental parameters are included in the Supporting Information. Masslynx 4.1 software (Waters-Micromass Corp., Milford, MA, U.S.A.) was used to acquire and process all data.

■

RESULTS AND DISCUSSION LC and MS Analysis of BSM Neutral OligosaccharideAlditols. The initial mixture and all the fractions were subjected to mass measurements on the Synapt G2 in the MS-only mode. Both deprotonated and chloride adducts were observed for all the oligosaccharide-alditols. [M − H]− ions were more intense than [M + Cl]− ions, which could be due to better initial formation of deprotonated ions during electrospray or possibly due to the decomposition of [M + Cl]− ions to [M − H]− ions with the neutral loss of HCl. Fraction 1 contained GalNAc-ol and was not examined.19 Two observations are worthy of note: (1) LC separation increased the sensitivity of low-abundant oligosaccharide-alditols due to a decrease in the presence of compounds causing ion suppression, enabling them to be analyzed by traveling wave IMS. (2) There were multiple oligosaccharide-alditols detected in the majority of individual HPLC fractions with the later fractions having higher mass molecules. Molecules having specific m/z values were observed in multiple HPLC fractions, 2230

DOI: 10.1021/ac503754k Anal. Chem. 2015, 87, 2228−2235

Article

Analytical Chemistry

Figure 2. Characterization of precursor ions with m/z 936.4. (a) Overlaid mobility profiles of the m/z 936.4 precursor ions in HPLC fractions 8 and 9. Further dissociation spectra are shown for valley-to-valley selected peaks 1, 2, and 3 as indicated for the blue profile of fraction 8. (b) Fragmentation pattern of the m/z 936.4 ion for mobility peak 1 in panel a, where the region of m/z 770−775 was magnified for 20-fold. (c) Fragmentation pattern of the m/z 936.4 ion for mobility peak 2 in panel a. (d) Fragmentation pattern of the m/z 936.4 ion for mobility peak 3 in panel a, where the region of m/z 400−415 was magnified for 3-fold.

fraction 8. Fractions 8 and 9 separated by HPLC have different mobilities, demonstrating two different structural isomers. Moreover, there are at least two structural isomers in fraction 9, even not fully resolved by IMS, where the separation was not achieved by HPLC only. There were in total five different drift time values for chloride-adducted pentasaccharide-alditols, deoxyHex1Hex1HexNAc2HexNAc-ol, at m/z 972.4 overlaid from fractions 8 and 9 in Figure 1e. Three separated mobility peaks were detected in fraction 9, and partially separated mobility peaks were detected in fraction 8. Fractions 8 and 9 each contained multiple structural isomers after a single HPLC separation as evaluated by IMS; more isomers could be differentiated by IMS in 10 ms than were resolved by LC separation in 4 h. For the deprotonated hexasaccharide-alditols having the composition deoxyHex2Hex2HexNAc1HexNAc-ol at m/z 1041.4 from fractions 9 and 10 shown in Figure 1f, only one mobility peak was detected in fraction 9, while two were observed in fraction 10, showing that fraction 10 is not isomerically pure and that multiple structural isomers may exist. Several conclusions can be made based on these results: First, multiple mobility peaks for oligosaccharide-alditol negative ions represent different structural isomers in a sample and not differences due to charge site location. Second, IMS is an orthogonal separation technique compared to LC. Isomers that eluted earlier by LC can actually drift more slowly (Figure 1, parts a and b) using IMS. The separation principles, of course, are entirely different. Third, the majority of LC peaks contained isomers as assessed by IMS; thus, one should expect that every LC fraction of oligosaccharides derived from glycoproteins may harbor isomeric species. This has been well-established with glycoproteins where enough sample is available to analyze LC fractions by NMR. Fourth, LC separations are typically on the order of many minutes, sometimes hours, whereas TWIMS can evaluate isomeric heterogeneity in less than 10 ms. Fifth, the resolving power of TWIMS peaks in this study was measured with a full width at half-maximum (fwhm) of ∼0.2−0.3 ms; thus, coincidental comigration of some compounds, even after HPLC and ion mobility separation, is not unexpected. Hence

reducing sugars usually run in the IMS as more than one peak, or a pattern of nearby peaks,30 because reducing sugars can exist in solution as the α and β anomers of both pyranose and furanose hemiacetals, all of which may be present in the gas phase upon ionization. In addition, the isomers in different fractions were both separated by chromatography19 and ion mobility; it is suggested that what separated by IMS for reduced oligosaccharide-alditols in this study are isomers having different structures, not conformational isomers differing at charge locations or adduct positions. Otherwise, if there was only one isomer, their IMS profile from different HPLC fractions should have been identical. Similarly, Figure 1b displays overlaid mobility profiles of the [M − H]− ion at m/z 628.3 from fractions 4, 5, and 6 having the proposed molecular formula of HexNAc2HexNAc-ol. The mobility profiles had similar characteristics as those observed in Figure 1a but with slightly longer drift times of 4.8 and 5.1 ms. These data are consistent with the presence of at least two structurally different isomers as discussed above. Mobility profiles of chloride-adducted tetrasaccharide-alditols having the molecular formula of deoxyHex1Hex1HexNAc1HexNAc-ol at m/z 769.3 are shown in Figure 1c. Fractions 5 and 6 shared the same drift time distributions having two partially separated mobility peaks (drift time 5.3 and 5.7 ms) detected at different intensities. The single mobility peak observed (drift time 5.7 ms) in fraction 7 was partially overlapped with the less abundant mobility peaks observed in fractions 5 and 6 for m/z 769.3. Unlike parts a and b of Figure 1, Figure 1c shows that the isomer(s) that eluted later using HPLC now had a longer drift time using IMS. Like parts a and b of Figure 1, however, the data clearly supported the presence of at least two structural isomers and not differences due to different charge site locations. In Figure 1d, another example for the [M − H]− precursor ion at m/z 895.3 from HPLC fractions 8 and 9 is presented, having the proposed composition of deoxyHex1Hex2HexNAc1HexNAc-ol. Two mobility peaks were detected in fraction 9; the peak having a shorter drift time (6.3 ms) was overlapped with the single peak observed in 2231

DOI: 10.1021/ac503754k Anal. Chem. 2015, 87, 2228−2235

Article

Analytical Chemistry

Figure 3. Isomeric analysis of product ions derived from precursor ions having m/z 790.3 in BSM HPLC fraction 7. (a) Fragmentation pattern of the precursor ion m/z 790.3 in HPLC fraction 7. (b) TWIMS spectra of the precursor ions having m/z 790.3 and product ions of m/z 749.3 (blue trace) and m/z 569.2 (red trace) from HPLC fraction 7.

the number of IMS peaks detected for one m/z only denotes the minimum number of carbohydrate isomers that can be differentiated by IMS under one defined set of experimental conditions. Finally, a specific oligosaccharide-alditol can sometimes be found solely within a single LC peak, but it is possible to find it within two of them. When collecting peaks from valley to valley as shown in Supporting Information Figure S1, some structures may span parts of both peaks as the valleys do not always reach baseline. Moreover, online coupling HPLC to ion mobility is expected in future work to improve the sample analysis efficiency. Fragmentation Spectra of Mobility-Separated Precursor Ions. Mobility-resolved MS/MS spectra for isomeric oligosaccharide-alditol precursor ions were obtained by operating the Synapt G2 in the ion mobility−MS/MS mode as described in the Experimental Section. Figure 2a displays the overlaid mobility spectra of pentasaccharide-alditols having the composition deoxyHex1Hex1HexNAc2HexNAc-ol as [M − H]− ions at m/z 936.4 in fractions 8 and 9. Two isomeric mobility peaks were observed for fraction 9, while three isomeric mobility peaks were detected for fraction 8, labeled as 1, 2, and 3. The corresponding fragmentation spectra for mobility peaks 1, 2, and 3 in fraction 8 are displayed in Figure 2b−d, respectively; 45 V CE was applied in the transfer cell after ion mobility separation. All three tandem spectra were different: for mobility peak 1 (Figure 2b) the base fragment ion is 715.3; for peak 2 (Figure 2c), the base fragment is 247.1 with a second abundant fragment ion at 407.2; peak 3 (Figure 2d) also had a base peak at 247.1, but the second most abundant fragment ion is at 570.2. The fragments in Figure 2 resulted from dissociation at glycosidic linkages with losses of deoxyHex, Hex, HexNAc, and HexNAc-ol as well as cross-ring cleavages; the key neutral losses are annotated in the spectra. These data indicated that the three mobility peaks separated for m/z 936.4 in fraction 8 carry different stereostructure or linkage information as validated by their different MS/MS spectra, further validating the unique separation capability of ion mobility and its utility for discriminating between structural isomers. The information from MS/MS alone is not adequate to derive unambiguous structures. Without ion mobility separation of isomers, mass spectra of mass selected precursor ions do not enable a direct relationship to be established between a specif ic combination of product ions and individual isomers of precursor ions. A similar

analysis was performed for the tetrasaccharide-alditols having the sugar composition deoxyHex1Hex1HexNAc1HexNAc-ol; the results are shown in Supporting Information Figure S2. Evaluation of Oligosaccharide-Alditol Product Ions. The presence of isomeric product ions generated from a specific mass-selected oligosaccharide-alditol precursor ion was demonstrated by collisionally dissociating a precursor ion followed by ion mobility separation of the product ions. Figure 3 shows the result of an experiment where a selected deprotonated tetrasaccharide-alditol, Hex1HexNAc2HexNAc-ol from LC fraction 7 (precursor ion m/z 790.3), was dissociated into its corresponding product ions prior to separation by traveling wave IMS. The fragmentation spectrum of m/z 790.3 is shown in Figure 3a using a 35 V CE in the trap cell; the extracted ion mobility spectra for the precursor ion of m/z 790.3, and for product ions of m/z 749.3 and 569.2, are presented in Figure 3b. Three isomeric mobility peaks were resolved for the precursor ion m/z 790.3 with mobility peaks at 5.48 and 5.63 ms being barely resolved. Two mobility peaks were fully separated for the product ion m/z 749.3 at drift times of 5.21 and 5.9 ms and partially resolved for the product ion m/ z 569.2 at drift times 4.35 and 4.83 ms; the mobility shoulder at 4.35 ms of low abundance was shown in the inserted window by expanding the intensity scale and confirmed by checking its reproducibility. Figure 3 demonstrates an example of isomeric precursor ions giving rise to isomeric product ions derived from a biologically relevant mixture of oligosaccharide-alditols that are not experimentally atypical. This is not surprising, since the structures of many carbohydrate product ions derived from isomeric precursor ions varying in their molecular structures would be expected to be different in many cases. It is also entirely possible that isomeric product ions may be derived from the same precursor ion, depending on its branching pattern. Fragmentation Spectra of Mobility-Separated Product Ions. To further demonstrate the isomeric heterogeneity of oligosaccharide product ions, fragmentation spectra of individual isomeric product ions were obtained. This was achieved by operating the Synapt G2 HDMS in MS/MS-ion mobility−MS/ MS mode as explained above. A representative example is shown here in the analysis of isomeric product ions generated from the precursor ion of m/z 1041.4 in HPLC fraction 9; the spectra are displayed in Figure 4. Figure 4a shows the MS/MS 2232

DOI: 10.1021/ac503754k Anal. Chem. 2015, 87, 2228−2235

Article

Analytical Chemistry

Figure 4. Analysis of isomeric product ions generated from precursor ions of m/z 1041.4 in HPLC fraction 9. (a) Fragmentation pattern of the m/z 1041.4 precursor ion in fraction 9. (b) TWIMS spectra of the precursor ion of m/z 1041.4 and the product ions of m/z 715.3 and m/z 551.2 from HPLC fraction 9. (c) Fragmentation pattern of product ion m/z 715.4 having a drift time of 4.88 ms derived from precursor ion m/z 1041.4. (d) Fragmentation pattern of product ion m/z 715.4 having a drift time of 5.48 ms derived from the precursor ion m/z 1041.4. (Note that the m/z 715.3 and m/z 715.4 refer to the same product ion, the mass difference is within the mass accuracy ±0.1 Da in this study.) Key neutral losses are indicated.

Figure 5. IMS spectra of both deprotonated and chloride adduct ions of oligosaccharide-alditols from BSM having specific m/z values illustrating the effect of different anions on isomer separation: (a) HexHexNAc1HexNAc-ol in HPLC fraction 4; (b) Hex1HexNAc2HexNAc-ol in HPLC fraction 7; (c) DeoxyHex2Hex2HexNAc1HexNAc-ol in HPLC fraction 10.

spectrum of m/z 1041.4. Fragmentation was carried out in the trap cell using 50 V CE, and the isomeric heterogeneity of all the ions was evaluated by traveling wave IMS. Figure 4b displays the mass-extracted mobility spectra for the precursor ion of m/z 1041.4 and selected isomeric product ions of m/z 715.3 and 551.2. Interestingly, only one mobility peak was detected for m/z 1041.4 at 7.32 ms, while two isomeric peaks were observed for its corresponding fragment ions at m/z 715.3 (4.88 and 5.48 ms) and 551.2 (4.01 and 4.50 ms). As mentioned previously, a single IMS precursor peak could still be an isomeric mixture, but it is also possible that more than one isomeric product ion may be derived from the same precursor. The separation that occurs in the ion mobility drift tube is based on an ion’s collision cross section. It is entirely feasible for isomeric precursor ions, where their overall geometry and differences in structural configuration are small, that they may coincidentally comigrate by IMS. However, when such oligosaccharide precursor ions break down to different substructures, differences in the stereochemistry and branching of product ions may give rise to cross-sectional areas that are very different. This can result in separable drift times for

product ions that may, in combination with their dissociation spectra, more effectively differentiate two precursor ions that happen to comigrate by IMS. These additional criteria might be especially powerful when combined with the information obtained from pure standards, where precursor mobility, precursor dissociation patterns, product mobility patterns, and mobility-selected product ion dissociation patterns could all be rapidly recorded and stored as unique database criteria for individual compounds. Product ion (m/z 715.3) was dissociated to yield mobilityselected product ion dissociation patterns (Figure 4, parts c and d). The ion is derived from the neutral loss of deoxyHex1Hex1 from the precursor ion m/z 1041.4. Fragmentation of a product ion can be carried out in the transfer cell (36 V CE was applied) in addition to fragmentation of a selected precursor ion in the trap cell, which is termed time aligned parallel fragmentation (TAP). Parts c and d of Figure 4 display the MS/ MS spectra of m/z 715.3, corresponding to the two isomeric mobility peaks at 4.88 and 5.48 ms, respectively. The two spectra were similar but some features were quite different; the abundance of product ion m/z 205.0 was much higher in Figure 2233

DOI: 10.1021/ac503754k Anal. Chem. 2015, 87, 2228−2235

Article

Analytical Chemistry

dioxide, helium, argon)50,57,58 or alternate ion adducts,30,59 the equivalent of days of work using multiple HPLC columns.

4c in comparison to Figure 4d with the same CE used. Some fragments were shared by both mobility peaks, albeit with different relative intensities; some fragments were unique. These data unambiguously indicate that the two product ion mobility peaks for m/z 715.3 represent at least two stereochemical and/or branch isomers possibly differing in anomeric configuration. This is the first report of individual fragmentation patterns for ion mobility-separated oligosaccharide product ions, further demonstrating the value of IMS through its capability of discriminating between isomeric substructures derived from (potentially) isomeric precursor ions. Anion Effects on the Mobility Separation. Using electrospray ionization, Dwivedi et al.30 reported that the nature of cations markedly affected the mobility separation of carbohydrate isomers, but the effects of anions on the separation of oligosaccharide isomers by ion mobility has not been widely reported. Herein, the mobilities of chloride adducts of oligosaccharide-alditols from BSM were measured and compared with corresponding [M − H]− ions. Figure 5 displays the IMS spectra of both [M − H]− and [M + Cl]− ions for four different oligosaccharide-alditols from different HPLC fractions. They are (a) trisaccharide-alditols of sugar composition Hex1HexNAc1HexNAc-ol in fraction 4, (b) tetrasaccharide-alditols of composition Hex1HexNAc2HexNAc-ol in fraction 7, and (c) hexasaccharide-alditols of composition DeoxyHex2Hex2HexNAc1HexNAcol in HPLC fraction 10. In Figure 5a, no separation of isomers was observed based on the mobility profile of the [M − H]− ion at m/z 587.2, while the [M + Cl]− ion at m/z 623.2 showed two mobility peaks for the same species. There were in total three mobility peaks detected for both [M − H]− and [M + Cl]− ions of the oligosaccharide-alditols shown in Figure 5b with similar resolution among them. However, the mobility profiles were dramatically different. Among the three isomeric mobility peaks (Figure 5b), the first two partially separated isomeric mobility peaks having respective drift times of 5.4 and 5.7 ms were both at high abundance for [M − H]− ions at m/z 790.3, but only one structural isomer was present as the major peak for [M + Cl]− ions at m/z 826.3 with a drift time of 5.8 ms. Apparently, some carbohydrates preferentially lose a proton, while others preferentially adduct an anion. Figure 5c displays the ion mobility separation of isomeric hexasaccharidealditol species. Two partially separated mobility peaks were observed for [M − H]− at m/z 1041.4 and [M + Cl]− at m/z 1077.4 with similar drift times and resolution. While the relative intensities of the mobility peaks remained opposite for [M − H]− and [M + Cl]− species, this may be explained that different isomeric structures have different preference for different anions. In general, [M + Cl]− ions had slightly longer drift times and better resolution than [M − H]− ions in traveling wave IMS. However, no specific rules can be concluded and more experiments using purified oligosaccharide standards in the negative ion mode are needed to fully understand how different anions (such as F−, Cl−, Br−, NO3−, CH3COO−, or several others) influence the mobility of carbohydrate isomer ions. A second separation of isomeric mixtures of oligosaccharides using orthogonal HPLC columns normally requires that each column be set up, solvent systems prepared, pumps and injectors washed, and columns equilibrated prior to analysis on the second system. With ion mobility, the isomeric homogeneity of a single IMS peak that might actually harbor more than one isomer may be rapidly assessed by alternating different ion mobility parameters such as drift gases (carbon

■

CONCLUSIONS Using O-linked oligosaccharide-alditols isolated from bovine submaxillary mucin as a model mixture of isomeric carbohydrate molecules in the negative ion mode, this study demonstrated the value of TWIMS in differentiating carbohydrate stereo- and/or branched isomers. Previous studies with several pure oligosaccharide-alditols showed that these precursor molecules run as single peaks31,56 and that reduction of the sugars was important to avoid the effects of multiple anomers and/or ring forms that can occur with reducing oligosaccharides. Many isomeric precursor ions were observed when individual HPLC fractions were analyzed by TWIMS. Isomeric mobility peaks were observed for the majority of oligosaccharide-alditols ranging from disaccharide-alditols to hexasaccharide-alditols. Many isomeric product ions can also be separated by TWIMS. Also, either m/z and mobility-selected precursor ions or product ions can be dissociated to yield mass spectra that differ between isomeric species. Specific product ions derived from a single selected oligosaccharide-alditol can be resolved into more than one independent mobility peak. For the first time, mobility-resolved product ions were demonstrated to yield different characteristic mass spectra. Some mobility peaks were partially or barely resolved; thus, higher resolving power and possibly orthogonal gas-phase separations may be needed in order to completely separate all the structural variants within a single oligosaccharide mixture. However, with appropriate carbohydrate standards the possibility now exists to define more exacting criteria for databases: (1) the precursor ion mobility drift time or collision cross section under defined conditions, (2) the dissociation pattern of the precursor ion, (3) the mobility drift times or collision cross sections of all product ions and their intensity ratios, or drif t time patterns, derived from a precursor ion, as shown in Figures 3 and 4, and (4) the product ion dissociation patterns of all product ions derived from a given precursor, with mobility selection of those product ions when they resolve from other isomeric product ions. Together, these criteria significantly increase the probability for identification of a structure provided these patterns can be reproduced from lab to lab. Moreover, similar criteria may be used for analysis of complex mixtures of isomers from many sources, such as samples analyzed in metabolomics,47,48 proteomics,45,46 and petroleum products, among many others.

■

ASSOCIATED CONTENT

S Supporting Information *

Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: 509-335-5648. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported in part by a National Institutes of Health Grant (5R33RR020046). 2234

DOI: 10.1021/ac503754k Anal. Chem. 2015, 87, 2228−2235

Article

Analytical Chemistry

■

(37) Borsdorf, H.; Eiceman, G. A. Appl. Spectrosc. Rev. 2006, 41, 323−375. (38) Borsdorf, H.; Mayer, T.; Zarejousheghani, M.; Eiceman, G. A. Appl. Spectrosc. Rev. 2011, 46, 472−521. (39) Giles, K.; Pringle, S. D.; Worthington, K. R.; Little, D.; Wildgoose, J. L.; Bateman, R. H. Rapid Commun. Mass Spectrom. 2004, 18, 2401−2414. (40) Shvartsburg, A. A.; Smith, R. D. Anal. Chem. 2008, 80, 9689− 9699. (41) Giles, K.; Williams, J. P.; Campuzano, I. Rapid Commun. Mass Spectrom. 2011, 25, 1559−1566. (42) Zhong, Y. Y.; Hyung, S. J.; Ruotolo, B. T. Analyst 2011, 136, 3534−3541. (43) Kim, H. I.; Kim, H.; Pang, E. S.; Ryu, E. K.; Beegle, L. W.; Loo, J. A.; Goddard, W. A.; Kanik, I. Anal. Chem. 2009, 81, 8289−8297. (44) Ridenour, W. B.; Kliman, M.; McLean, J. A.; Caprioli, R. M. Anal. Chem. 2010, 82, 1881−1889. (45) Santos, L. F.; Iglesias, A. H.; Pilau, E. J.; Gomes, A. F.; Gozzo, F. C. J. Am. Soc. Mass Spectrom. 2010, 21, 2062−2069. (46) Shliaha, P. V.; Bond, N. J.; Gatto, L.; Lilley, K. S. J. Proteome Res. 2013, 12, 2323−2339. (47) Wickramasekara, S. I.; Zandkarimi, F.; Morre, J.; Kirkwood, J.; Legette, L.; Jiang, Y.; Gombart, A. F.; Stevens, J. F.; Maier, C. S. Metabolites 2013, 3, 701−717. (48) Paglia, G.; Williams, J. P.; Menikarachchi, L.; Thompson, J. W.; Tyldesley-Worster, R.; Halldorsson, S.; Rolfsson, O.; Moseley, A.; Grant, D.; Langridge, J.; Palsson, B. O.; Astarita, G. Anal. Chem. 2014, 86, 3985−3993. (49) Yamagaki, T.; Sato, A. J. Mass Spectrom. 2009, 44, 1509−1517. (50) Fasciotti, M.; Sanvido, G. B.; Santos, V. G.; Lalli, P. M.; McCullagh, M.; de Sa, G. F.; Daroda, R. J.; Peter, M. G.; Eberlin, M. N. J. Mass Spectrom. 2012, 47, 1643−1647. (51) Harvey, D. J.; Scarff, C. A.; Crispin, M.; Scanlan, C. N.; Bonomelli, C.; Scrivens, J. H. J. Am. Soc. Mass Spectrom. 2012, 23, 1955−1966. (52) Gelb, A. S.; Jarratt, R. E.; Huang, Y.; Dodds, E. D. Anal. Chem. 2014, 86, 11396−11402. (53) Fang, T. T.; Bendiak, B. J. Am. Chem. Soc. 2007, 129, 9721− 9736. (54) Liu, Y.; Clemmer, D. E. Anal. Chem. 1997, 69, 2504−2509. (55) Li, H.; Bendiak, B.; Kaplan, K.; Davis, E.; Siems, W. F.; Hill, H. H., Jr. Int. J. Mass Spectrom. 2013, 352, 9−18. (56) Clowers, B. H.; Dwivedi, P.; Steiner, W. E.; Hill, H. H.; Bendiak, B. J. Am. Soc. Mass Spectrom. 2005, 16, 660−669. (57) Asbury, G. R.; Hill, H. H. Anal. Chem. 2000, 72, 580−584. (58) Matz, L. M.; Hill, H. H.; Beegle, L. W.; Kanik, I. J. Am. Soc. Mass Spectrom. 2002, 13, 300−307. (59) Jiang, Y. J.; Cole, R. B. J. Am. Soc. Mass Spectrom. 2005, 16, 60− 70.

REFERENCES

(1) An, H. J.; Kronewitter, S. R.; de Leoz, M. L.; Lebrilla, C. B. Curr. Opin. Chem. Biol. 2009, 13, 601−607. (2) Brownlee, M. Annu. Rev. Med. 1995, 46, 223−234. (3) Lis, H.; Sharon, N. Eur. J. Biochem. 1993, 218, 1−27. (4) Rudd, P. M.; Elliott, T.; Cresswell, P.; Wilson, I. A.; Dwek, R. A. Science 2001, 291, 2370−2376. (5) Shental-Bechor, D.; Levy, Y. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 8256−8261. (6) Maley, F. Fed. Proc. 1981, 40, 1556−1556. (7) Horton, D. Adv. Carbohydr. Chem. Biochem. 2009, 62, xi−xiii. (8) Broberg, S.; Nord, L. I.; Kenne, L. J. Mass Spectrom. 2004, 39, 691−701. (9) Dell, A.; Azadi, P.; Tiller, P.; Thomas-Oates, J.; Jennings, H. J.; Beurret, M.; Michon, F. Carbohydr. Res. 1990, 200, 59−76. (10) Karlsson, N. G.; Wilson, N. L.; Wirth, H. J.; Dawes, P.; Joshi, H.; Packer, N. H. Rapid Commun. Mass Spectrom. 2004, 18, 2282−2292. (11) Weiskopf, A. S.; Vouros, P.; Harvey, D. J. Rapid Commun. Mass Spectrom. 1997, 11, 1493−1504. (12) Williams, D.; Lee, T. D.; Dinh, N.; Young, M. K. Rapid Commun. Mass Spectrom. 2000, 14, 1530−1537. (13) Ahn, Y. H.; Yoo, J. S. Rapid Commun. Mass Spectrom. 1999, 13, 855−859. (14) Kurihara, T.; Min, J. Z.; Toyo’oka, T.; Fukushima, T.; Inagaki, S. Anal. Chem. 2007, 79, 8694−8698. (15) Ridlova, G.; Mortimer, J. C.; Maslen, S. L.; Dupree, P.; Stephens, E. Rapid Commun. Mass Spectrom. 2008, 22, 2723−2730. (16) Xu, Y. T.; Wang, X. J.; Su, X. L.; Xu, J. L.; Chen, H. M.; Chen, J. J.; Yan, X. J. Chin. J. Anal. Chem. 2011, 39, 1798−1804. (17) Ninonuevo, M.; An, H.; Yin, H.; Killeen, K.; Grimm, R.; Ward, R.; German, B.; Lebrilla, C. Electrophoresis 2005, 26, 3641−3649. (18) Gennaro, L. A.; Harvey, D. J.; Vouros, P. Rapid Commun. Mass Spectrom. 2003, 17, 1528−1534. (19) Martensson, S.; Levery, S. B.; Fang, T. T.; Bendiak, B. Eur. J. Biochem. 1998, 258, 603−622. (20) Bendiak, B. Carbohydr. Res. 1999, 321, 139−139. (21) Duus, J. O.; Gotfredsen, C. H.; Bock, K. Chem. Rev. 2000, 100, 4589−f. (22) Thureau, P.; Thevand, A.; Ancian, B.; Escavabaja, P.; Armstrong, G. S.; Mandelshtam, V. A. ChemPhysChem 2005, 6, 1510−1513. (23) Nilsson, B.; Zopf, D. Methods Enzymol. 1982, 83, 46−58. (24) Montilla, A.; van de Lagemaat, J.; Olano, A.; del Castillo, M. D. Chromatographia 2006, 63, 453−458. (25) Song, L. X.; Bai, L. J. Phys. Chem. B 2009, 113, 9035−9040. (26) Lamari, F.; Militsopoulou, M.; Gioldassi, X.; Karamanos, N. K. Fresenius’ J. Anal. Chem. 2001, 371, 157−167. (27) Joucla, G.; Brando, T.; Remaud-Simeon, M.; Monsan, P.; Puzo, G. Electrophoresis 2004, 25, 861−869. (28) Kamoda, S.; Ishikawa, R.; Kakehi, K. J. Chromatogr. A 2006, 1133, 332−339. (29) Kanu, A. B.; Dwivedi, P.; Tam, M.; Matz, L.; Hill, H. H. J. Mass Spectrom. 2008, 43, 1−22. (30) Dwivedi, P.; Bendiak, B.; Clowers, B. H.; Hill, H. H. J. Am. Soc. Mass Spectrom. 2007, 18, 1163−1175. (31) Zhu, M. L.; Bendiak, B.; Clowers, B.; Hill, H. H. Anal. Bioanal. Chem. 2009, 394, 1853−1867. (32) Fenn, L. S.; McLean, J. A. Phys. Chem. Chem. Phys. 2011, 13, 2196−2205. (33) Li, H. L.; Bendiak, B.; Siems, W. F.; Gang, D. R.; Hill, H. H. Anal. Chem. 2013, 85, 2760−2769. (34) Williams, J. P.; Grabenauer, M.; Holland, R. J.; Carpenter, C. J.; Wormald, M. R.; Giles, K.; Harvey, D. J.; Bateman, R. H.; Scrivens, J. H.; Bowers, M. T. Int. J. Mass Spectrom. 2010, 298, 119−127. (35) Harvey, D. J.; Sobott, F.; Crispin, M.; Wrobel, A.; Bonomelli, C.; Vasiljevic, S.; Scanlan, C. N.; Scarff, C. A.; Thalassinos, K.; Scrivens, J. H. J. Am. Soc. Mass Spectrom. 2011, 22, 568−581. (36) Yamaguchi, Y.; Nishima, W.; Re, S. Y.; Sugita, Y. Rapid Commun. Mass Spectrom. 2012, 26, 2877−2884. 2235

DOI: 10.1021/ac503754k Anal. Chem. 2015, 87, 2228−2235