Article pubs.acs.org/ac
Differential Chemical Derivatization Integrated with Chromatographic Separation for Analysis of Isomeric Sialylated N‑Glycans: A Nano-Hydrophilic Interaction Liquid ChromatographyMS Platform Fateme Tousi, Jonathan Bones,*,† William S. Hancock, and Marina Hincapie*,‡ Barnett Institute of Chemical and Biological Analysis, Northeastern University, Boston, Massachusetts 02115, United States S Supporting Information *
ABSTRACT: MS analysis of sialylated glycans is challenging due to their low ionization efficiency in positive ion mode as well as the possibility of in-source fragmentation. Chemical derivatization strategies have been developed to address this issue focused on removal of the labile acidic proton prior to MS analysis. Highly sialylated negatively charged glycans also exhibit high retention and unsatisfactory separation efficiency when analyzed by hydrophilic interaction liquid chromatography (HILIC) due to their high polarity. Here, we combined linkage specific derivatization of sialic acids by reaction with the condensation reagent 4-(4,6-dimethoxy-1,3,5-triazin-2-yl)4-methylmorpholinium chloride (DMT-MM) in methanol with nanoscale liquid chromatographic separation prior to accurate mass Orbitrap MS analysis. Coupling DMT-MM charge neutralization of sialic acids with nano-HILIC-Orbitrap-MS not only allows for linkage specific characterization of sialylated glycans directly from the precursor mass but also improves the preceding HILIC separation by increasing the hydrophobicity and altering the selectivity of the oligosaccharide analytes. We focused on the trisialylated N-glycan fraction from haptoglobin and human plasma, enriched using weak anion exchange chromatography, as this trisialylated fraction has been linked with cancer associated changes in the serum N-glycome. The developed methodology was applied to investigate whether structural alterations in this oligosaccharide pool, enriched from the sera of pathological stage and sex matched patients bearing lung, breast, ovarian, pancreatic, or gastric cancer, demonstrate any degree of cancer specificity or whether changes in expression levels are purely cancer associated. The results of this pilot study indicate limited degrees of cancer specificity, particularly for pancreatic cancer, based on alterations in the relative abundance of specific trisialylated isomers.
S
ionization efficiency when using positive ionization.9−11 Sialylated glycans are commonly analyzed using negative mode ionization, while mass spectra of neutral glycans are recorded in the positive mode.12,13 This impairs quantitative profiling of glycan mixtures containing both sialylated and neutral glycans. In MALDI-time of flight (TOF) instruments, sialylated glycans have been reported to undergo both in-source and post-source loss of sialic acid.12 The presence of sialic acid on oligosaccharides also complicates the interpretation of positive ion electrospray ionization (ESI)-MS spectra due to the presence of [M + H]+ as well as cationized adduct forms of the same glycan, e.g., [M + Na]+ or [M + K]+, etc.14 Methyl esterification of sialic acid residues resulting in neutralization of their negatively charged carboxylic acid groups has been used to address the aforementioned issues associated
ialic acids are a family of 9-carbon monosaccharides found as terminal residues on many glycan structures attached to glycoproteins.1 Specific glycosyltransferase enzymes located in the Golgi add terminal sialic acids to galactose residues in either an α(2,3) or α(2,6) linkage.2 Sialylated glycans play important roles in biological and pathological processes,3 including cell− cell adhesion, cell surface receptor recognition, and progression of human malignancies.4 Increased levels of α(2,6)-linked sialic acids have been shown to be associated with various cancer types.5−7 Moreover, sialyl LewisX, an α(2,3)-linked sialylated tetrasaccharide epitope found on many tumor cells, strongly correlates with tumor progression and metastasis in humans.8 Therefore, the ability to quantitatively analyze alterations in specific sialic acid linkage isomers could facilitate discovery of informative diagnostic and/or prognostic cancer biomarkers rather than analyzing total levels of sialylation alone. Analysis of sialylated glycans using mass spectrometry is challenging. Unlike neutral sugars, due to their inherent negative charge at pH >2.6, sialylated glycans exhibit low © 2013 American Chemical Society
Received: June 21, 2013 Accepted: July 31, 2013 Published: July 31, 2013 8421
dx.doi.org/10.1021/ac4018007 | Anal. Chem. 2013, 85, 8421−8428
Analytical Chemistry
Article
with positive ion MS of sialylated glycans.15 Chemical derivatization resulting in the charge neutralization of sialylated glycans generates associated ionization efficiency similar to that of neutral glycans, thereby facilitating more quantitative profiling.16 Recently, derivatization strategies using the condensation reagent 4-(4,6-dimethoxy-1,3,5-triazin-2-yl)-4methylmorpholinium chloride (DMT-MM) in either methanol or ammonium based reaction media have been described, which facilitates the derivatization of sialic acids in a linkage specific manner.16,17 However, reported methods have predominantly been employed prior to MALDI-TOF MS with no online separation or sample fractionation using liquid chromatography. In this work, we introduce an enhanced chromatographic separation using nano-hydrophilic interaction liquid chromatography (HILIC)-ESI-MS combined with linkage specific derivatization as an improved approach for the characterization of isomeric and highly sialylated N-glycans. Coupling DMTMM derivatization of sialic acids with nanoscale liquid chromatographic separation prior to MS analysis offers numerous advantages for the analysis of sialylated glycans. First, and most simply, as the complexity of the glycan mixture entering the mass spectrometer is reduced, ion suppression effects are minimized, thereby improving the reliability of quantitation. Second, as previously noted by Wheeler et al.,16 due to the regiospecific nature of the DMT-MM reaction, sialic acids present in a specific linkage react differently. For example, when the reaction is performed in methanol, α(2,6) linked sialic acids present on the glycan become methylated whereas α(2,3) linked sialic acids form cyclic lactones, resulting in the addition of 14 Da or the loss of 18 Da depending on the linkage type, respectively.16 These differences can be easily measured using mass spectrometry, facilitating structural annotation of glycans with linkage specificity. Third, when hydrophilic interaction liquid chromatography (HILIC) is employed, the use of DMT-MM derivatization chemistry prior to nanoLCHILIC-MS aids in addressing a fundamental problem associated with HILIC based separations of large polar molecules. In HILIC, as the gradient progresses and the percentage of aqueous modifier in the mobile phase increases, associated chromatographic efficiency drops due to increased dispersion of the aqueous partition layer surrounding the stationary phase particles. Therefore, peaks corresponding to large highly retained tri- and tetrasialylated oligosaccharides become broad, meaning that HILIC is unable to separate linkage specific subpopulations of these complex N-glycans. The regiospecific nature of the DMT-MM derivatization increases the hydrophobicity of these complex sialylated oligosaccharides, thereby potentially altering the selectivity of the resulting HILIC separation, which should result in earlier elution of the derivatized glycans in a region of the chromatogram wherein the chromatographic efficiency is higher, facilitating separation of these linkage specific subpopulations prior to their entry into the mass spectrometer. The combination of DMT-MM charge neutralization derivatization in a linkage specific manner with nano-HILICMS for improved speciation and quantitation of sialylated glycans is described herein. For initial experiments, trisialylated glycans enriched from the haptoglobin released N-glycan pool using weak anion exchange chromatography were employed. Following optimization, the developed DMT-MM nanoHILIC-MS platform was applied to the analysis of trisialylated glycans enriched from the total N-glycan pools released from
plasma of advanced stage III lung, breast, ovarian, pancreatic, and gastric cancer patients. We specifically chose the trisialylated glycan fraction for analysis based upon the consistent reporting of alterations in this fraction in previous cancer studies.17−22 In the current pilot study, we investigated whether differences exist in the subpopulations of isomeric trisialylated glycans that are cancer specific rather than previous reports of these alterations being cancer associated. Albeit with limited patient numbers, no differences in the relative abundance of trisialylated linkage isomers were observed in the glycan pools of patients bearing lung, breast, gastric, or ovarian carcinomas, suggesting that alterations observed in previous studies were indeed cancer associated rather than cancer specific. However, distinct alterations in the relative abundances of trisialylated linkage isomers were observed in the glycan pools of the pancreatic cancer patients when compared to the other cancer groups, suggesting a limited degree of cancer specificity in the identified oligosaccharide structures. The developed DMT-MM nanoLC-HILIC-MS platform facilitates the generation of highly informative data, although verification of suggested cancer specific alterations in the glycome will be necessary in a larger patient set to evaluate the potential of any identified structure as a candidate marker.
■
EXPERIMENTAL SECTION Chemicals and Reagents. LC-MS grade water, acetontrile, methanol, and formic acid used throughout were obtained from Thermo Fischer Scientific (Waltham, MA). 2-Aminobenzamide (2-AB), sodium cyanoborohydride (reagent grade, 95%), acetic acid, ammonia, dimethyl sulfoxide, sodium bicarbonate, and 4(4,6-dimethoxy-1,3,5-triazin-2-yl)-4-methylmorpholinium chloride (DMT-MM) were received from Sigma-Aldrich (St. Louis, MO) and were of the highest available purity. Blood Plasma Samples. Plasma samples collected from female race matched patients diagnosed with stage III cancer were obtained from ProteoGenex (ProteoGenex, Culver City, CA). A total number of 19 plasma samples from cancer patients composed of lung (n = 4), breast (n = 4), gastric (n = 4), pancreatic (n = 4), and ovarian (n = 3) cancer types were used in this study. N-Glycan Release. Haptoglobin (50 μg, Sigma-Aldrich, MO), dissolved in 96 μL of digestion buffer (20 mM sodium bicarbonate, pH 7), was subjected to enzymatic deglycosylation by adding 4 μL of PNGase F solution (New England Biolabs, Ipswich, MA) equivalent to 2000 units. In the case of plasma samples, to 20 μL of plasma from each patient, 56 μL of digestion buffer and 4 μL of PNGase F solution were added. Samples were incubated at 37 °C overnight (16 h). The released N-glycans were purified using 10 kDa molecular weight cutoff filters (PALL Life Sciences, Ann Arbor, MI) and reduced to dryness via vacuum centrifugation. Twenty microliters of 1% formic acid solution was added to the dried glycans to promote hydrolysis of the released glycosylamines to the corresponding reducing sugar followed by subsequent evaporation to dryness via vacuum centrifugation. Fluorescent Labeling with 2-Aminobenzamide. Liberated glycans from haptoglobin, plasma, and standard sialyl lactose 3′ and 6′ isomers (Sigma-Aldrich, MO) were labeled with 2-aminobenzaminde (2-AB). Details of the 2-AB labeling procedure can be found in the Supporting Information. Exoglycosidase Digestions. To demonstrate the specificity of the DMT-MM reaction, 2-AB labeled 3′-sialylactose was digested with α(2,3)-neuraminidase (New England Biolabs, 8422
dx.doi.org/10.1021/ac4018007 | Anal. Chem. 2013, 85, 8421−8428
Analytical Chemistry
Article
Ipswich, MA) and ABS (Arthrobacter ureafaciens sialidase) (Prozyme, Hayward, CA) as detailed in the Supporting Information. Weak Anion Exchange Separation of N-Glycans. Weak anion exchange (WAX) separation of fluorescently labeled Nglycan pool was performed on a GlycoSep C (Prozyme, Hayward, CA (75 mm × 7.5 mm)) 10 μm particle packed DEAE functionalized column using a Shimadzu Prominence HPLC system equipped with a binary gradient solvent delivery unit, a sample injection valve, and a fluorescence detector as described previously.23 UPLC-HILIC-FLR Profiling of N-Glycans. Fluorescently labeled N-glycans were separated on a Waters BEH Glycan hydrophilic interaction chromatography column (100 mm × 2.1 mm), 1.7 μm amide functionalized BEH particles (Waters Corporation, Milford, MA). The ultra performance liquid chromatography system used was a Waters Acquity UPLC instrument consisting of a binary solvent manager, sample manager, column manager, and FLR fluorescence detector controlled by Empower 3 Chromatography workstation software. A linear gradient of 30−47%, 50 mM ammonium formate pH 4.5 over 16.5 min at the flow rate of 0.56 mL/min, followed by 2 min re-equilibration with 70% acetonitrile, was employed for oligosaccharide separation; total run time was 19 min. Samples were prepared in 80:20 acetonitrile/water and maintained at 5 °C in the thermostatted autosampler prior to injection; a volume of 20 μL was injected on the column. The separation temperature was 40 °C. The fluorescence detection settings were λex = 330 nm and λem = 420 nm with a data collection rate of 20 Hz. DMT-MM/MeOH Derivatization of Sialylated N-Glycans. 2-AB labeled glycans were subjected to linkage specific derivatization of sialic acid by dissolution in 50 μL of 0.1 M solution of DMT-MM in methanol, followed by incubation at 80 °C for 1 h. Samples were cooled to room temperature and reduced to dryness using a vacuum centrifuge. Samples were dissolved in 25 μL of the appropriate solvent (80:20 mobile phase B/mobile phase A) for the subsequent chromatographic analysis. Nano-HILIC-Orbitrap-MS. A 0.075 × 250 mm capillary column (New Objective, Woburn, MA) was packed in house with a 5 μm amide bonded silica-based stationary phase (GlycoSep N, Prozyme, Hayward, CA). Mobile phases A and B were 0.1% FA in H2O and 0.1% FA in acetonitrile, respectively, prepared using LC-MS grade solvents and formic acid. Five μL of each sample was loaded on the column in 95% mobile phase B using an autosampler at a loading flow of 300 nL/min for 30 min. The LC method consisted of a steep gradient of mobile phase A from 5% to 20% over 10 min followed by a shallow gradient from 20% to 50% mobile phase A over 60 min and finally an isocratic hold at 5% mobile phase A. The flow rate was 300 nL/min over 60 min. LC-MS was performed using a Dionex UltiMate 3000 liquid chromatography system (ThermoFisher Scientific, Waltham, MA) interfaced with either a LTQ Orbitrap XL or a LTQ Orbitrap Elite mass spectrometer (ThermoFisher Scientific, Waltham, MA) for the analysis of haptoglobin and plasma glycans, respectively. The mass spectrometer was tuned in positive mode and operated in data dependent mode; each full MS scan over the range of 400−2000 m/z was followed by data-dependent MS/MS CID fragmentation events on the four most abundant precursor ions present in the full MS spectrum. MS parameters were as follows: resolution, 60 000 at 400 m/z;
isolation width, 3; normalized collision energy, 35%; spray voltage, 1.6 kV; capillary temperature, 210 °C; tube lens offset and capillary voltage were automatically tuned using 2-AB labeled glycans from haptoglobin. Mass Spectrometry Data Analysis. Structural annotation of glycan MS and MS2 data was performed manually. GlycoWorkbench24 was used for drawing glycan structures as well as confirmation of MS/MS fragment annotation. Statistical Analysis. Nonparametric Kruskal−Wallis H and Mann−Whitney U tests were performed, and box-plots were generated using IBM SPSS Statistics version 21 (IBM, NY, USA) to evaluate the differences in each experimental parameter across different cancer types. P-values
