|Home | About | Journals | Submit | Contact Us | Français|
Mass spectrometry (MS) has the potential to revolutionize structural glycobiology and help in the understanding of how post-translation events such as glycosylation affect protein activities. Several approaches to determine the structure of glycopeptides have been used successfully including fast atom bombardment, matrix-assisted laser desorption ionization, and electrospray ionization with a wide variety of mass analyzers. However, the identification of glycopeptides in a complex mixture still remains a challenge. The source of this challenge is primarily due to the poor ionization efficiency and rapid degradation of glycopeptides. In this report we describe the use of a chip-based infusion nanoelectrospray ionization technique in combination with a recently developed linear ion trap for identification and characterization of glycosylation in complex mixtures. Two standard synthetic glycans were analyzed using multiple-stage fragmentation analysis in both positive and negative ionization modes. In addition, the high mannose type N-glycosylation in ribonuclease B (RNase B) was used to map the glycosylation site and obtain the glycan structures. We were able to map the glycosylation site and obtain the glycan structures in RNase B in a single analysis. The results reported here demonstrate that the fully automated chip-based nanoelectrospray linear ion trap platform is a valuable system for oligosaccharide analyses due to the unique MS/MS and MSn capability of the linear ion trap and the extended analysis time provided by the ionization technique.
Glycosylation of proteins plays a vital role in biological processes such as molecular recognition and intra- and intercellular signaling. The carbohydrates in proteins can also significantly alter protein conformation and consequently impact protein functions.1 Characterizing the detailed structure of glycoproteins provides insight to aid in biomedical research and drug discovery. Oligosaccharides can be linked to serine or threonine (O-glycosylation) or to asparagine (N-glycosylation).
N-linked glycosylation is well documented in the literature and the structures of N-linked glycans have been studied extensively.2 Glycosylation occurs in the endoplasmic reticulum and Golgi apparatus of the cell and involves a complex series of reactions. N-linked core oligosaccharides are added to proteins in the endoplasmic reticulum and further trimmed by glycosidases and mannosidases. The spectrum of glycoforms remains rather uniform until the glycoproteins reach the medial stacks of the Golgi apparatus, where structural diversification is introduced by a series of nonuniform modifications. Generally, all N-linked oligosaccharides can be divided into three classes; the high mannose type, the complex type, and the less common hybrid type. The type of sugar attached to a protein will depend on the cell in which the protein is expressed and on the physiological status of the cell.3,4
O-linked glycosylation is initiated in the Golgi by the addition of a single sugar, typically N-acetylgalactosamine in mammals, to serine or threonine. O-oligosaccharides vary in size from a single N-acetylgalactosamine residue to oligosaccharides comparable in size to N-linked glycans. O-linked glycans are divided into domains, namely the core and the antenna. Since the antenna of O- and N-linked glycans is synthesized in a similar manner, they often carry the same terminal structures. There are at least seven O-oligosaccharide core structures, four of which (core types 1, 2, 3, and 4) are particularly widespread in mammals.5
The analysis and characterization of the glycan structures are difficult primarily due to branching, different configurations, isoforms, and heterogeneity. Mass spectrometry (MS) has been a powerful tool for the characterization of glycosylation including fast atom bombardment, matrix-assisted laser desorption ionization, and electrospray ionization (ESI) using a wide variety of mass analyzers.5 Generally, the intact glycoprotein is isolated and the glycan structure can be characterized either after its release from the protein or while still linked to the protein. Many strategies have been developed to analyze the structure of the glycans including derivatization techniques of the reducing ends and protection of the functional groups.6–9 The development of separation techniques for glycan and glycopeptides has also been an important part of structural glycobiology. The use of porous graphitic carbon columns for the separation of oligosaccharides has been especially successful.10–12 Other approaches include the combination of MS with capillary electrophoresis and reversed-phase high-performance liquid chromatography.13–17
Although ESI-tandem MS remains widely recognized as an effective means for oligosaccharide characterization, the hydrophilic nature of these compounds has often been cited as the cause for poor ESI sensitivity requiring either derivatization18 or nanoESI19 to improve ionization efficiency. NanoESI also offers low sample consumption due to its low flow rate, which provides extended analysis time for completing multistage fragmentation analyses on oligosaccharides.19 However, the disadvantages of nanoESI have included its low sample throughput due to tedious single tip alignment procedures, potential sample carryover if the tip is used to analyze multiple samples, and poor reproducibility of the relative intensities of analytes due to the variable shape of the spray tip and the capillary-sampling cone orifice distance for each repeat analysis.20
In this study, we used a chip-based fully automated nanoESI system21–24 in combination with a recently developed linear ion trap25 for the analysis of glycopeptide and oligosaccharide samples without prior separation or modification of the isoforms. We report the benefits realized by using MSn capability of the linear ion trap with extended analysis times to enable the detailed examination and mapping of glycosylation sites and glycan structures.
Cellohexaose was purchased from Seikagaku Corp. (Toyko, Japan), the biantennary N-linked core pentasaccharide was purchased from V-LABS, Inc. (Covington, LA, USA), and bovine ribonuclease B was purchased from Sigma (St. Louis, MO, USA). Modified trypsin was from Promega (Madison, WI, USA). All other chemical reagents, unless otherwise noted, were obtained from Aldrich (Milwaukee, WI, USA).
One milligram of ribonuclease B (Sigma, St. Louis, MO, USA) was reconstituted in 100 μL of solution containing 20 mM Tris-HCl, pH 7.8, 6 M guanidine-HCl and 10 mM dithiothreitol. The mixture was incubated for 30 min at 50°C. To the above solution, 50 μL of 0.2 M iodoacetamide and 50 μL of 0.2 M ammonium bicarbonate, pH 7.8, were added. The mixture was incubated at room temperature in the dark for 2 h. The alkylated solution was dialyzed against 20 mM ammonium bicarbonate, pH 7.8, overnight at 4°C using a Slide-A-Lyzer Mini Dialysis Unit from Pierce (Rockford, IL, USA). The dialyzed protein was digested by adding 0.52 mg/mL trypsin (Promega) at a 1:60 ratio (w/w) in 50 mM ammonium bicarbonate, pH 7.8, and incubated at 37°C overnight.
Samples were reconstituted at a concentration of 1 pmol/μL in 50% methanol with 0.1% formic acid for positive ion mode and in 50% methanol with 0.1% NH4OH for negative ion mode. The NanoMate 100 (Advion BioSciences, Ithaca, NY, USA) was coupled to the front of the Finnigan LTQ via a mounting bracket as previously described for the Finnigan LCQ.24 Five-microliter samples were analyzed via fully automated chip-based nanoESI using spray voltages and sample delivery pressures of 1.55 kV and 0.2 psi for positive ion mode and 1.75 kV and 0.7 psi for negative ion mode. The estimated flow rate was approximately 100 nL/min.
The sample was analyzed in the tune mode of Xcalibur software using a Finnigan LTQ mass spectrometer. The capillary temperature was set to 150°C; collision energies were set to 20–25% for MSn experiments; the maximum scan time was set to 50 ms; and 2–3 micro-scans were summed for each scan.
Two standard oligosaccharides were analyzed by chip-based infusion nanoESI-MS as described in Materials and Methods to demonstrate the capability of the system for structural analysis. The full-scan mass spectrum of the biantennary N-linked core pentasaccharide showed sodium and ammonium adduct ions (Fig. 1A1A).). Figures 1B–E depict the systematic fragmentation of the oligosaccharide and the resulting detailed information about the branching of the analyte. As shown in Figure 1B1B the most intense fragment produced by the MS2 spectrum of the doubly sodiated core pentasaccharide ion (m/z 933.18) arises from cleavage of the glycosydic bond between the N-acetylglucosamine and the mannose sugar. The MS3 spectrum of the sodiated cationized remnant tetrasaccharide ion (m/z 730.09; Fig. 1C1C)) revealed a wealth of information confirming the structure of the tetrasaccharide. The ion fragment at m/z 478 was consistent with a3,5A3 intrasaccharide cross-ring fragment, indicating the 1–3 linkage between the mannoses. The3,5A2 cross-ring cleavage of the reducing-terminal mannose was also observed at m/z 275 in the MS4 spectrum of the sodium cationized trisaccharide ion (Fig. 1D1D).
Since it is sometimes difficult to ionize glycans in positive ion mode, or the separation technique used requires negative ion mode, we analyzed a cellohexaose as a second standard glycan to demonstrate the negative ion mode capability of the instrumentation. The full-scan mass spectrum of cellohexaose is shown in Figure 2a2a.. The spectrum shows an ion at m/z 989.27, which corresponds very well with the molecular weight of the compound minus one proton. The doubly charged ion of cellohexaose (minus two protons) was visible at m/z 494.18. Figures 2B and CC depict the systematic fragmentation of the oligosaccharide and the resulting detailed information about the composition of the analyte. As shown in Figure 2B2B,, the MS2 spectrum of the singly charged ion at m/z 989 yielded a number of fragment ions that conform to familiar oligosaccharide fragmentation patterns. The spectrum is dominated by peaks resulting from cleavage at glycosidic bonds, giving the Y ion series (m/z 827.27, 665.18, 503.9, and 341.09) and a slightly less intense, 18-Da lower B ion series (m/z 809.27, 647.18, 485.09, and 323.00). In addition, the cross-ring fragment ions at m/z 928.36 and m/z 869.14 (corresponding to loss of 60 and 120 Da) are observed as 0,2A6 and2,4A6 structures respectively. A similar result was obtained in the MS3 spectrum of the selected Y5 ion at m/z 827.27 (Fig. 2C2C).). Cellohexaose was also analyzed using positive ion mode resulting in higher signal intensity allowing systematic fragmentation up to MS to the 4th (data not shown).
The tryptic digest of bovine pancreatic ribonuclease B (1 pmol/μL) was analyzed by chip-based infusion nanoESI-MS as described in the Materials and Methods. Tandem mass spectrometry was performed on all the major peaks and additional fragmentation (MSn) was recorded. The full scan mass spectrum (Fig. 3A3A)) shows the complex peptide mass fingerprint of ribonuclease B. Several peptides could be identified using fragmentation information from MS/MS spectra and the automated search program BioWorks 3.1. The software correlates theoretical MS/MS data from a database with the actual data for identification. The identified peptides and the protein coverage are shown in Table 11.. Nine out of fourteen possible tryptic peptides were identified, resulting in protein coverage of over 87%. The software could not identify several major ions, although the fragmentation pattern of those ions appeared to be of high quality. The unidentified doubly charged ions showed a difference of 81 Da, which is the typical pattern of a high mannose type glycopeptides. Collision-induced fragmentation of the unidentified ions (Figs. 3B–F) generated doubly charged fragment ions that differ by 81 Da, confirming the presence of a high mannose type glycosylation. Additionally, the presence of N-acetyl-d-glucosamine could be easily detected in the tandem mass spectra based on the marker ion of 204 Da. Assuming complete tryptic digestion, the only possible N-linked glycopeptide of ribonuclease B is NLTK, which fits the glycosylation sequence motif NXS/T. The molecular weight information of the unidentified peaks, together with the molecular weight of the amino acid sequence and the fragmentation information, confirmed the presence of five different high mannose-type glycopeptides in ribonuclease B.
As shown in Figures 4A–C, the MSn capability can be used to analyze the glycosylation in more detail. Tandem mass spectra of up to MS to the 5th showed high-quality spectra and resulted in useful information for the characterization of the glycosylation. Since glycopeptides tend to fragment only at their oligosaccharide structure in tandem mass spectra, it is often difficult to obtain amino acid sequence information. The advantage of the linear ion trap, which can perform MSn up to the 10th, becomes obvious when looking at the MS4 spectrum (Fig. 55).). The amino acid sequence of the glycopeptide can be easily determined manually or by the automated de novo sequencing software DeNovoX after the oligosaccharide is systematically removed from the peptide in the ion trap.
The new Finnigan LTQ ion trap mass spectrometer is a powerful tool for the analysis of glycoproteins and oligosaccharides.
The MSn capability of the LTQ, which also offers high mass accuracy and high resolution, overcomes the significant limitations often encountered with traditional triple quadrupole techniques. Using the unique MS/MS and MSn functions, combined with BioWorks 3.1 protein identification software, the glycosylated peptide of ribonuclease B was identified and the sugar structure determined. The MS4 spectra provided the additional information of the residue of sugar attachment. The fully automated nanoESI-MS platform used in this study delivered stable nanoESI with extended infusion time for completing various MS/MS and MSn modes in a single analysis, consuming only a small amount of sample.
We would like to thank Dave Bajkowski for the initial installation of the NanoMate for the Finnigan LTQ. We also thank Professor Jack Henion and Dr. Colleen Van Pelt for reviewing this manuscript and for providing helpful comments.