Electron Capture Dissociation and Infrared Multiphoton Dissociation


Electron Capture Dissociation and Infrared Multiphoton Dissociation...

0 downloads 189 Views 87KB Size

Anal. Chem. 2001, 73, 4530-4536

Electron Capture Dissociation and Infrared Multiphoton Dissociation MS/MS of an N-Glycosylated Tryptic Peptide To Yield Complementary Sequence Information Kristina Håkansson,† Helen J. Cooper,† Mark R. Emmett,† Catherine E. Costello,‡ Alan G. Marshall,† and Carol L. Nilsson*,‡,§

Center for Interdisciplinary Magnetic Resonance, National High Magnetic Field Laboratory, Florida State University, 1800 East Paul Dirac Drive, Tallahassee, Florida 32310, Mass Spectrometry Resource, Boston University School of Medicine, 715 Albany Street R-806, Boston, Massachusetts 02118-2526, and Institute of Medical Biochemistry, Go¨teborg University, Box 440, SE-405 30 Go¨teborg, Sweden

Glycoproteins are a functionally important class of biomolecules for which structural elucidation presents a challenge. Fragmentation of N-glycosylated peptides, employing collisionally activated dissociation, typically yields product ions that result from dissociation at glycosidic bonds, with little occurrence of dissociation at peptide backbone sites. We have applied two dissociation techniques, electron capture dissociation (ECD) and infrared multiphoton dissociation (IRMPD), in a 7-T Fourier transform ion cyclotron resonance mass spectrometer, in the investigation of an N-glycosylated peptide from an unfractionated tryptic digest of the lectin of the coral tree, Erythrina corallodendron. ECD provided c and z• ions derived from the peptide backbone, with no observed loss of sugars. Cleavage at 11 of 15 backbone amine bonds was observed. The lack of cleavage at sites located close to the glycosylated asparagine residue may result from steric blocking by the glycan. IRMPD provided abundant fragment ions, primarily through dissociation at glycosidic linkages. The monosaccharide composition and the presence of three glycan branch sites could be determined from the IRMPD fragments. The two types of spectra, obtained with the same instrument, thus provide complementary structural information about the glycopeptide. The current result extends the applicability of ECD for glycopeptide analysis to N-glycosylated peptides and to peptides containing branched, highly substituted glycans. Glycoproteins are an important group of biomolecules, comprising about half of all proteins from eukaryotic sources and having members in a wide range of functional classes, such as enzymes, hormones, cytokines, lectins, light transducers, receptors, and immunoglobulins. The characterization of protein glycoforms is desirable because glycosylation affects biological activity, * To whom correspondence should be addressed (Sweden). E-mail: [email protected]. † Florida State University. ‡ Boston University School of Medicine. § Go ¨teborg University.

4530 Analytical Chemistry, Vol. 73, No. 18, September 15, 2001

stability, and solubility. Differences in glycosylation may be observed in disease states, such as the spongiform encephalopathies.1 The structural elucidation of glycoproteins is a challenge. Whereas unmodified proteins can often be studied by X-ray crystallography or nuclear magnetic resonance spectroscopy, those methods may not provide satisfactory structural information about the saccharide portion of glycoproteins. Structural prediction of glycoproteins is difficult as well; for instance, no consensus sequence has been determined for O-glycosylation. A consensus sequence for N-glycosylation has been reported, Asn-X-Ser/Thr, in which X may be any amino acid except Pro;2 however, the occupancy of potential glycosylation sites and glycan structure(s) must be determined experimentally. All N-linked glycans are known to contain a pentasaccharide core structure, ManR3(ManR6)Manβ4GlcNAcβ4GlcNAc. Mass spectrometry is a valuable technique for structural analysis of glycoproteins. In particular, recent advances in mass spectrometry have greatly increased the mass resolution and mass accuracy of bioanalysis. Characterization strategies for a glycoprotein by mass spectrometry typically require several steps of analysis in combination with various glycosidase and protease treatments, because of the different chemical nature of the saccharide and peptide components. However, in investigations requiring structural elucidation of small amounts of glycoproteins isolated from biological sources, it is preferable to limit the number of sample manipulation steps. High-resolution, high-mass accuracy techniques enable the efficient identification of biological compounds. Smaller molecules (small peptides/enzymatic fragments for which the gene sequence is known) can often be identified based on accurate mass data, thus eliminating the need to perform MSn analysis. Accurate mass measurements are also important in determining posttranslational modifications of proteins. Currently, the highest performance (highest resolution, highest mass accuracy) mass analyzer is the Fourier transform ion cyclotron resonance (FTICR) mass spec(1) Rudd, P. M.; Endo, T.; Colominas, C.; Groth, D.; Wheeler, S. F.; Harvey, D. J.; Wormald, M. R.; Serban, H.; Prusiner, S. B.; Kobata, A.; Dwek, R. A. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 13044-13049. (2) Kornfeld, R.; Kornfeld, S. Annu. Rev. Biochem. 1985, 54, 631-664. 10.1021/ac0103470 CCC: $20.00

© 2001 American Chemical Society Published on Web 08/16/2001

trometer.3 Recent work has also shown that FTICR analysis can be performed on subfemtomole amounts of sample when microelectrospray4 is the ionization mode.5-7 For peptide primary structure investigation, FTICR MS offers several possible fragmentation techniques. In infrared multiphoton dissociation (IRMPD), energy from an infrared laser is transferred to trapped ions and used to fragment them.8 IRMPD fragmentation of peptides can be advantageous because there is no need to pulse gas into the mass analyzer, and therefore, a step that degrades the vacuum and thus resolution is avoided. IRMPD thereby permits faster fragmentation analysis and higher resolution than sustained off-resonance irradiation collisionally induced dissociation (SORI-CID).9 Both IRMPD and SORI-CID can cause fragmentation of labile bonds within side chains and posttranslational modifications. Electron capture dissociation (ECD)10-12 is a soft fragmentation technique that is unique to FTICR MS. This technique mainly induces fragmentation of the backbone of a peptide or protein, forming c and z• ions. However, modifications, even extremely thermolabile ones such as γ-carboxylation, Oglycosylation, or phosphorylation, are retained.13-16 ECD therefore provides important information complementary to that obtained with IRMPD or SORI-CID as dissociation techniques. The legume lectins are an extensively studied group of glycoproteins, bearing a high degree of amino acid sequence homology.17 In particular, detailed structural data from X-ray diffraction and mass spectrometry have been reported for the 28kDa (255 amino acid residues) lectin of the coral tree, Erythrina corallodendron, making it an excellent model glycoprotein for the development of new biophysical methods. In its native form, this lectin contains two identical glycoprotein subunits, each carrying two N-linked saccharides of the xylose type18 at amino acid residues 1717 and 113,19,20 ManR3(ManR6)(Xylβ2)Manβ4GlcNAcβ4(FucR3)GlcNAc.21 The two glycosylation sites are both fully occupied and homogeneous. (3) Marshall, A. G.; Hendrickson, C. L.; Jackson, G. S. Mass Spectrom. Rev. 1998, 17, 1-35. (4) Emmett, M. R.; Caprioli, R. M. J. Am. Soc. Mass Spectrom. 1994, 5, 605613. (5) Belov, M. E.; Gorshkov, M. V.; Udseth, H. R.; Anderson, G. A.; Smith, R. D. Anal. Chem. 2000, 72, 2271-2279. (6) Martin, S. E.; Shabanowitz, J.; Hunt, D. F.; Marto, J. A. Anal. Chem. 2000, 72, 4266-4274. (7) Quenzer, T. L.; Emmett, M. R.; Hendrickson, C. L.; Kelly, P. H.; Marshall, A. G. Anal. Chem. 2000, 73, 1721-1725. (8) Little, D. P.; Speir, J. P.; Senko, M. W.; O’Connor, P. B.; McLafferty, F. W. Anal. Chem. 1994, 66, 2809-2815. (9) Gauthier, J. W.; Trautman, T. R.; Jacobson, D. B. Anal. Chim. Acta 1991, 246, 211-225. (10) Zubarev, R. A.; Kelleher, N. L.; McLafferty, F. W. J. Am. Chem. Soc. 1998, 120, 3265-3266. (11) Axelsson, J.; Palmblad, M.; Hakansson, K.; Hakansson, P. Rapid Commun. Mass Spectrom. 1999, 13, 474-477. (12) Zubarev, R. A.; Horn, D. M.; Fridriksson, E. K.; Kelleher, N. L.; Kruger, N. A.; Lewis, M. A.; Carpenter, B. K.; McLafferty, F. W. Anal. Chem. 2000, 72, 563-573. (13) Kelleher, N. L.; Zubarev, R. A.; Bush, K.; Furie, B.; Furie, B. C.; McLafferty, F. W.; Walsh, C. T. Anal. Chem. 1999, 71, 4250-4253. (14) Mirgorodskaya, E.; Roepstorff, P.; Zubarev, R. A. Anal. Chem. 1999, 71, 4431-4436. (15) Stensballe, A.; Norregaard-Jensen, O.; Olsen, J. V.; Haselmann, K. F.; Zubarev, R. A. Rapid Commun. Mass Spectrom. 2000, 14, 1793-1800. (16) Shi, S. D.-H.; Hemling, M. E.; Carr, S. A.; Horn, D. M.; Lindh, I.; McLafferty, F. W. Anal. Chem. 2001, 73, 19-22. (17) Adar, R.; Richardson, M.; Lis, H.; Sharon, N. FEBS Lett. 1989, 257, 8185.

The glycan at Asn 17 is unusual because it provides excellent X-ray diffraction data.22 This glycan is important to establish the structure of the dimeric lectin, correct for complexing with lactose.22 The second glycan at Asn 113 is more mobile, yielding only faint electron density.23 Mass spectrometric techniques have elucidated important structural features of the lectin. LC-MS in a triple quadrupole instrument was used to resolve discrepancies between the protein and cDNA sequences and to demonstrate the presence of the second glycan.19 Tryptic digests of the protein have been separated by capillary electrophoresis followed by ESIMS/MS in order to study the glycosylation of the lectin.20,24 However, low-energy CID spectra of glycopeptides typically display fragment ions derived from the cleavage of glycosidic bonds, whereas little dissociation in the peptide backbone structure is observed. It would be desirable in mass spectrometric studies of glycopeptides to obtain fragment ions from both the peptide backbone and the glycan. We therefore have explored the use of IRMPD and ECD in an FTICR mass spectrometer to produce two complementary sets of fragment ion data from an N-glycosylated tryptic peptide present in the unfractionated digest from the lectin of E. corallodendron. EXPERIMENTAL SECTION Enzymatic Digestion. Approximately 10 nmol of the lectin from E. corallodendron (Sigma, St. Louis, MO) was dissolved in 500 µL of buffer containing 0.1 mM CaCl2 and 0.1 M NH4HCO3, to which 100 pmol of modified trypsin (Promega) had been added. Digestion of the protein proceeded at 38 °C for 4 h. The mixture was dried in a Speedvac (Savant, Inc., Holbrook, NY). Determination of Molecular Masses of Tryptic Peptides. The dried lectin digest was dissolved in 100 µL of HPLC grade water (J. T. Baker, Philipsburg, NJ) to produce a 100 µM stock solution. A 10-µL aliquot of the stock solution was placed in 80 µL of an electrospray solvent consisting of 50:50 methanol (Baker) and water with 2% acetic acid (Aldrich, Milwaukee, WI). A 10-µL aliquot of an electrospray calibration mixture (Agilent Technologies, Wilmington, DE) was also added, resulting in a final tryptic peptide concentration of 10 µM. A home-built, passively shielded, 9.4-T FTICR mass spectrometer25 equipped with microelectrospray ionization4 was used for accurate mass determination of the tryptic peptides. The peptides were infused at a flow rate of 500 nL/min through an electrospray emitter consisting of a 50-µm-i.d. fused-silica capillary which had been mechanically ground to a uniform thin-walled tip.26 In the (18) Kamerling, J. P. Pure Appl. Chem. 1991, 63, 465-472. (19) Young, N. M.; Watson, D. C.; Yaguchi, M.; Adar, R.; Arango, R.; RodriguezArango, E.; Sharon, N.; Blay, P. K. S.; Thibault, P. J. Biol. Chem. 1995, 270, 2563-2570. (20) Kelly, J. F.; Locke, S. J.; Ramaley, L.; Thibault, P. J. Chromatogr., A 1996, 720, 409-427. (21) Ashford, D.; Dwek, R. A.; Welply, J. K.; Atamayakul, S.; Homans, S. W.; Lis, H.; Taylor, G. N.; Sharon, N.; Rademacher, T. W. Eur. J. Biochem. 1987, 166, 311-320. (22) Shaanan, B.; Lis, H.; Sharon, N. Science 1991, 254, 862-866. (23) Elgavish, S.; Shaanan, B. J. Mol. Biol. 1998, 277, 917-932. (24) Bonneil, E.; Thibault, P.; Young, N. M.; Lis, H.; Sharon, N. 48th ASMS Conference on Mass Spectrometry and Allied Topics; Long Beach, CA, 2000; CD-ROM. (25) Senko, M. W.; Hendrickson, C. L.; Pasa-Tolic, L.; Marto, J. A.; White, F. M.; Guan, S.; Marshall, A. G. Rapid Commun. Mass Spectrom. 1996, 10, 1824-1828. (26) Quinn, J. P.; Emmett, M. R.; Marshall, A. G. 46th ASMS Conference on Mass Spectrometry and Allied Topics; Orlando, FL, 1998; pp 1388-1388.

Analytical Chemistry, Vol. 73, No. 18, September 15, 2001

4531

9.4-T instrument, the electrosprayed ions are transported into the mass spectrometer through a Chait-style atmosphere-to-vacuum interface27 and externally accumulated28 for 2 s in an octopole pretrap. After accumulation, the collected ions are transferred through multipole ion guides and captured by gated trapping in an open29 cylindrical cell. The ions were subjected to chirp excitation (72-320 kHz at 150 Hz/µs) and direct-mode broadband detection (1 Mword data points). Hanning apodization and one zero fill were applied prior to fast Fourier transform (FFT) followed by magnitude calculation. The experimental event sequence was controlled by a modular ICR data acquisition system (MIDAS).30 Internal frequency-to-m/z calibration31 was performed by use of three of the ions from the added calibration mixture. The presented spectrum represents a sum of 10 time-domain transients. Bioinformatics Tools. Experimental monoisotopic masses were matched against the theoretical masses generated from the mature protein sequence derived from the Swiss-Prot database entry (P16404), by use of the MS-Digest tool (http://www.prospector.ucsf.edu). A mass deviation of 3 ppm was allowed. Unmatched peptides were examined for the presence of glycosylation by use of the GlycoMod tool (http://www.expasy.ch) and other modifications using the FindMod tool at the same site. In both cases, the protein identity (P16404) and unmatched monoisotopic masses were entered, and a mass deviation of 3 ppm was tolerated. Electron Capture Dissociation. ECD was performed on a home-built, unshielded, 7-T FTICR mass spectrometer7 equipped with a heated-filament electron gun located behind the ICR cell. The lectin tryptic peptide stock solution was diluted 20-fold into an electrospray solvent consisting of 50:50 methanol and water with 0.25% acetic acid. This solution was microelectrosprayed at a flow rate of 300 nL/min by use of an electrospray configuration similar to that for the 9.4-T instrument. The ions were externally accumulated for 3 s. After accumulation, the ions were transferred through the hexapole ion guide and captured by gated trapping in an open orthorhombic cell. A transfer time of 1.5 ms was used to obtain maximum magnitude of the triply protonated molecular ion of the N-glycosylated peptide at m/z 1005.5. Stored waveform inverse Fourier transform (SWIFT)32,33 ejection was applied to isolate the glycopeptide under investigation. Two consecutive SWIFT excitation profiles were employed: first, a broadband SWIFT with an 80-Da-wide notch centered around the glycopeptide ion and, second, a heterodyne SWIFT within the 80-Da window. The isolated parent ion was irradiated with electrons for 30 s. A filament bias voltage of 3.1 V and a heating current of 3.8 A, resulting in a voltage drop of 4.3 V across the filament, was used. A voltage (-2 V) was applied to a collector plate located (27) Chowdhury, S. K.; Katta, V.; Chait, B. T. Rapid Commun. Mass Spectrom. 1990, 4, 81-87. (28) Senko, M. W.; Hendrickson, C. L.; Emmett, M. R.; Shi, S. D.-H.; Marshall, A. G. J. Am. Soc. Mass Spectrom. 1997, 8, 970-976. (29) Beu, S. C.; Laude, D. A., Jr. Int. J. Mass Spectrom. Ion Processes 1992, 112, 215-230. (30) Senko, M. W.; Canterbury, J. D.; Guan, S.; Marshall, A. G. Rapid Commun. Mass Spectrom. 1996, 10, 1839-1844. (31) Ledford, E. B., Jr.; Rempel, D. L.; Gross, M. L. Anal. Chem. 1984, 56, 27442748. (32) Marshall, A. G.; Wang, T.-C. L.; Ricca, T. L. J. Am. Chem. Soc. 1985, 107, 7893-7897. (33) Guan, S.; Marshall, A. G. Int. J. Mass Spectrom. Ion Processes 1996, 157/ 158, 5-37.

4532 Analytical Chemistry, Vol. 73, No. 18, September 15, 2001

Figure 1. 9.4-T FTICR mass spectrum obtained after microelectrospray ionization of an unfractionated tryptic digest of the lectin from E. corallodendron. Assignments of identified tryptic peptides are given. Two peaks assigned to N-glycosylated peptides are marked with stars. Calibrant peaks are marked with dots.

behind the filament.34 The ECD fragment ions were subjected to chirp excitation (43-540 kHz at 350 Hz/µs) and direct-mode broadband detection (512 Kword data points). Hanning apodization and one zero fill were applied prior to FFT followed by magnitude calculation. The experimental event sequence was controlled by an Odyssey data acquisition system (Finnigan Corp., Bremen, Germany). External calibration31,35 was performed by use of ECD fragment ions from doubly protonated substance P (Sigma). The presented spectrum represents an average of 100 time-domain transients. Infrared Multiphoton Dissociation. IRMPD was performed with the same 7-T FTICR instrument used for ECD. The electron gun can be moved off-axis from its preferred central position to allow infrared photons to enter the ICR cell on-axis. A 40-W, 10.6µm CO2 laser (Synrad, Mukilteo, WA) provided the photons. The laser beam is directed to the center of the cell through a BaF2 window. No optics were used to focus the beam. The same tryptic peptide solution as for the ECD experiment was microelectrosprayed under the same conditions. The triply protonated glycopeptide at m/z 1005.5 was SWIFT isolated by two consecutive ejection profiles, as described above. Photon irradiation was performed for 200 ms at 20% laser power. Excitation, detection, and data manipulation were performed in the same way as for the ECD experiment. The spectrum was internally calibrated by use of the triply protonated parent ion and the doubly protonated ion resulting from proton stripping of the parent. The presented spectrum represents an average of 100 time-domain transients. RESULTS AND DISCUSSION 9.4-T FTICR Mass Spectrometry of the Lectin Digest. The 9.4-T ESI FTICR mass spectrum of the unfractionated lectin digest is shown in Figure 1. The calibration peaks are marked with dots. The 19 most abundant components are listed in Table 1. Nine peaks, corresponding to six unique masses, matched the theoretical masses of unmodified peptides from the lectin at a mass accuracy of 2.6 ppm or better. The observed masses of two other peaks (marked with stars in Figure 1) were accurately matched (34) Hakansson, K.; Emmett, M. R.; Hendrickson, C. L.; Marshall, A. G. Anal. Chem. 2001, 73, 3605-3610. (35) Shi, S. D.-H.; Drader, J. J.; Freitas, M. A.; Hendrickson, C. L.; Marshall, A. G. Int. J. Mass Spectrom. 2000, 195/196, 591-598.

Table 1. Nineteen Most Abundant Ions Observed by 9.4-T FTICR Mass Spectrometry Following Microelectrospray Ionization of an Unfractionated Lectin Tryptic Digesta

a

obsd m/z

charge state

corresponding [M + H]+

theoretical [M + H]+

assignment

error (ppm)

622.0288 773.8583 857.8203 869.4594 875.4457 922.0103 936.0061 982.5031 992.0204 1005.457

+1 +2 +3 +3 +3 +1 +2 +3 +3 +3

622.0288 1546.709 2571.446 2606.364 2624.323 922.0103 1871.005 2945.495 2974.047 3014.357

622.0290 1546.707 2571.443

calibrant [37-50] [178-201]

-0.2 +1.5 +1.3

2624.318 922.0098 1871.002 2945.491

[51-73] calibrant [155-171] [74-99]

+1.8 +0.6 +1.6 +1.3

3014.351

+1.9

1184.524 1286.229 1312.665 1329.309 1339.641 1473.250 1521.971 1675.809

+2 +2 +2 +3 +3 +2 +1 +3

glycosylated [100-116]

2368.041 2571.450 2624.322 3985.912 4016.909 2945.493 1521.971 5025.411

2571.443 2624.318 3985.899

[178-201] [51-73] [117-151]

+2.6 +1.5 +3.3

2945.491 1521.972 5025.387

+0.5 -0.5 +4.8

1681.488

+3

[74-99] calibrant glycosylated [1-36]

5042.448

The spectrum was internally calibrated. The N-glycosylated peptide marked in boldface type was chosen for further analysis.

Figure 2. Electron capture dissociation FTICR mass spectrum obtained from the triply protonated N-glycosylated peptide of m/z 1005.5 (peptide segment 100-116) in the lectin digest (Figure 1). The y-axis is magnified 10×. Cleavages at 11 of 15 backbone amine positions (i.e., the bond between a backbone nitrogen and the adjacent R-carbon) are observed. The achieved c ions provide a peptide sequence tag of six amino acids. No fragmentation of the branched, N-linked heptasaccharide was observed.

to the two known glycopeptide structures by the GlycoMod tool. No other modifications could be matched to the remaining peptides according to Find Mod. Three of the experimental masses were calibrants. In total, 73% protein sequence coverage was obtained. The amino acid residue numbers used in Table 1 and Figure 1 correspond to the amino acid sequence of the mature protein product. Electron Capture Dissociation of the Glycopeptide at m/z 1005.5 (3+). The ECD fragment spectrum obtained from the N-glycosylated peptide of m/z 1005.5 (residues 100-116) in the lectin digest is shown in Figure 2, and the corresponding mass assignments are listed in Table 2. The amino acid sequence and glycan structure of this peptide, and sites of dissociation in ECD, are shown in Figure 3 (top). Mainly N-terminal c-type ions are observed in the ECD spectrum. All c ions were identified at a

Figure 3. Peptide sequence and glycan structure of the investigated N-glycosylated peptide. Top: Dissociation sites for ECD. Bottom: Major dissociation sites for IRMPD.

mass accuracy of 6.6 ppm or higher. A major product is also the charge-reduced species, [M + 3H + e]2+•. The ion labeled z3 is observed at a mass 1.007 ( 0.001 Da higher than expected for the radical z3• ion. That mass can be explained by the formation of an even-electron z fragment (which contains one extra hydrogen atom), accompanied by a radical c• fragment, either by direct hydrogen atom capture at the R-carbon or by H rearrangement.12 Two species of mass greater than the parent ion were observed in the ECD spectrum. Neither could be assigned to the lectin but may derive from a contaminant coisolated during SWIFT. The low-magnitude second harmonic observed at twice the frequency of the parent ion could arise from an imperfect on-axis alignment of the ions, imperfectly balanced detection circuitry, or both. Analytical Chemistry, Vol. 73, No. 18, September 15, 2001

4533

Table 2. Fragment Ions Observed after ECD of the Triply Protonated N-Glycosylated Peptide of m/z 1005.5 in the Lectin Digest. obsd m/z

charge state

corresponding [M + H]+

333.1762 335.1819 401.2507 502.7267 529.3106 586.3329 749.3993 754.3392 806.4127 969.4824 1005.446 1082.562 1325.655 1342.274 1399.615 1443.140 1456.149 1470.158 1477.132 1486.667 1499.668 1508.139

1+ 9+ 1+ 6+ 1+ 1+ 1+ 4+ 1+ 1+ 3+ 1+ 3+ 3+ 2+ 2+ 2+ 2+ 2+ 2+ 2+ 2+

333.1762

332.1696

401.2507

401.2512 second harmonic 529.3098 586.3313 749.3946 3014.352 806.4161 969.4794 3014.352 1082.564

529.3106 586.3329 749.3993 3014.335 806.4127 969.4824 3014.323 1082.562 3974.951 4024.808 2798.222 2885.272 2911.291 2939.309 2953.253 2972.326 2998.328 3015.271

theoretical [M + H]+

assignment z3 third harmonic c4

error (ppm) see text -1.3

c5 c6 c7 [M + 4H]4+ c8 c9 [M + 3H]3+ c10

+1.5 +2.7 +6.2 -5.5 -4.2 +3.0 -9.5 -1.6

2798.241 2885.273 2911.301

c152+ c162+ z•162+

-6.6 -0.1 -3.4

2972.341 2998.333 3015.360

[M -CH3CO + 3H + e]2+• [M - NH3 + 3H+e]2+• [M + 3H + e]2+•

-5.2 -1.8 -7.8

Table 3. Fragment Ions Observed after IRMPD of the Triply Protonated Glycopeptide of m/z 1005.5 in the Lectin Digesta

a

obsd m/z

charge state

corresponding [M + H]+

theoretical [M + H]+

204.0891 366.1427 528.1957 690.2485 750.7072 804.7224 858.7369 902.7543 907.4091 912.7560 922.4747

1+ 1+ 1+ 1+ 3+ 3+ 3+ 3+ 3+ 3+ 2+

204.0891 366.1427 528.1957 690.2485 2250.107 2412.153 2574.196 2706.248 2720.213 2736.253 1843.942

204.0867 366.1395 528.1923 690.2451 2250.093 2412.146 2574.199 2706.241 2720.257 2736.252 1843.934

951.4385 956.7688 961.4409 1002.215 1005.455 1006.972 1015.002

3+ 3+ 3+ 4+ 3+ 4+ 2+

2852.301 2868.292 2882.308 4005.839 3014.351 4024.866 2028.998

2852.299 2868.294 2882.309

[GlcNAc′ + H]+ [ManGlcNAc′ + H]+ [Man2GlcNAc′ + H]+ [Man3GlcNAc′ + H]+ [M - Man3XylFuc + 3H]3+ [M - Man2XylFuc + 3H]3+ [M - ManXylFuc + 3H]3+ [M - ManFuc + 3H]3+ [M - ManXyl + 3H]3+ [M - FucXyl + 3H]3+ [M - Man3XylFuc GlcNAc2 + 2H]2+ [M - Man + 3H]3+ [M - Fuc + 3H]3+ [M - Xyl + 3H]3+

3014.352

[M + 3H]3+

-0.1

2029.003

-2.8

1024.007

2+

2047.008

2047.014

[M - Man3XylFuc GlcNAc - H2O + 2H]2+ [M - Man3XylFuc GlcNAc + 2H]2+

1054.567 1097.035 1125.543 1206.575 1287.605 1368.614 1434.664 1507.679

1+ 2+ 2+ 2+ 2+ 2+ 2+ 2+

1054.567 2193.063 2250.079 2412.142 2574.204 2736.222 2868.320 3014.351

2193.072 2250.093 2412.146 2574.199 2736.252 2868.294 3014.352

assignment

[M - Man3Xyl GlcNAc + 2H]2+ [M - Man3XylFuc + 2H]2+ [M - Man2XylFuc + 2H]2+ [M - ManXylFuc + 2H]2+ [M - XylFuc + 2H]2+ [M - Fuc + 2H]2+ [M + 2H]2+

error (ppm) +12 +8.8 +6.4 +4.9 +6.2 +2.7 -1.0 +2.7 -16 +0.7 -4.3 +0.7 -0.7 -0.5

-2.9 -4.0 -6.1 -1.5 +1.8 -11 +9.1 -0.1

GlcNAc′ ) dehydroGlcNAc.

The observed fragment ions correspond to cleavages at 11 of 15 peptide backbone amine bonds (i.e., the bond between a backbone nitrogen and the adjacent R-carbon), counting the z3 ion. The N-terminal side of the proline residue is not considered because cleavage at that site is not observed in ECD due to the 4534

Analytical Chemistry, Vol. 73, No. 18, September 15, 2001

cyclic structure of proline.10 All sites that are not cleaved, except the C-terminal side of proline, are located close to the glycosylated asparagine residue. A possible explanation for the absence of cleavage at those sites is that the bulky glycan sterically hinders access to the backbone carbonyl oxygens of the corresponding

amino acid residues. The carbonyl oxygens are involved in the proposed ECD cleavage mechanism,36 in which, after electron capture, an energetic hydrogen atom is released and subsequently captured by a site of high hydrogen atom affinity, such as a carbonyl oxygen. Lack of ECD cleavage due to steric blocking of a carbonyl oxygen has also been suggested to account for the minor cleavage observed on the C-terminal side of tryptophan in human luteinizing hormone releasing hormone.34 The C-terminal side of tryptophan has been reported elsewhere to be a preferred site of cleavage.37 In the ECD spectrum of the lectin glycopeptide, three glycosylated fragment ions are observed. All three fragments contain the entire, complex glycan structure; no carbohydrate loss due to cleavage of glycosidic bonds is observed. The signal-to-noise ratio (S/N) of the highest magnitude glycosylated c ion is ∼20:1; i.e., the probability of glycoejection is 10% or less, assuming a detection limit of S/N 2:1. Similar behavior has been reported previously for O-glycosylated peptides.14 However, the previous study was limited to peptides containing glycans with one or two sugar residues. The current result extends the applicability of ECD for glycopeptide analysis to N-glycosylated peptides and to peptides containing branched, highly substituted glycans. A peptide sequence tag38 containing six amino acid residues obtained from the achieved ECD fragment ion series can be used to retrieve the protein from the database. Infrared Multiphoton Dissociation of the Glycopeptide at m/z 1005.5 (3+). The IRMPD fragment ion spectrum obtained from the N-glycosylated peptide of m/z 1005.5 in the lectin digest is shown in Figure 4, and the corresponding mass assignments are listed in Table 3. A complex fragmentation pattern is observed. Mostly doubly protonated ions are observed in the spectral region, 1100 < m/z < 1600 (Figure 4, bottom). Several such ions can be identified as the parent glycopeptide with loss of one or more sugars, probably through a charge-remote fragmentation mechanism. The ion observed at m/z 1507.7 corresponds to the doubly protonated fully glycosylated peptide, resulting from proton stripping of the isolated triply protonated parent ion. In the 900 < m/z < 1100 range, both doubly and triply protonated ions are seen. Several correspond to the parent glycopeptide with loss of one or more sugars. Finally, in the 200 < m/z < 900 range, triply protonated ions corresponding to the parent glycopeptide with loss of several sugars are seen. Also, singly protonated, dehydrated sugar ions are observed. As in the ECD spectrum, two species of mass greater than the parent ion were observed in the IRMPD spectrum. Neither could be assigned to the lectin, but may derive from a contaminant coisolated during SWIFT. The fragmentation sites are illustrated in Figure 3 (bottom). Dissociation at each glycosidic bond was observed, as well as the loss of the entire glycan. The extensive monosaccharide losses are consistent with the presence of multiple branch points in the structure of the glycan. Observation of the doubly protonated fragment at m/z 1097.0 ([M - Man3XylGlcNAc + 2H]2+) made it possible to specify the site of fucosylation as the inner GlcNAc (36) Zubarev, R. A.; Kruger, N. A.; Fridriksson, E. K.; Lewis, M. A.; Horn, D. M.; Carpenter, B. K.; McLafferty, F. W. J. Am. Chem. Soc. 1999, 121, 28572862. (37) Kruger, N. A.; Zubarev, R. A.; Carpenter, B. K.; Kelleher, N. L.; Horn, D. M.; McLafferty, F. W. Int. J. Mass Spectrom. 1999, 182/183, 1-5. (38) Mann, M.; Wilm, M. Anal. Chem. 1994, 66, 4390-4399.

Figure 4. Infrared multiphoton dissociation FTICR mass spectrum (displayed in three segments) from the triply protonated N-glycosylated peptide of m/z 1005.5 in the lectin digest (Figure 1). The y-axis is magnified 2× (low- and high-mass regions) and 3× (intermediatemass region). Extensive fragmentation of the glycan is seen. The achieved fragmentation enabled partial determination of the glycan structure, including the presence of three branching sites. No ions corresponding to cleavage at the peptide backbone are observed.

residue. The IRMPD spectrum provides no information about the peptide sequence. CONCLUSION Mass spectrometric studies of glycoproteins typically entail either the release of glycans followed by mass spectrometric analysis or low-energy CID of glycopeptides. In the first case, information about glycosylation site occupancy and heterogeneity may be lost. In low-energy CID of glycopeptides, the underlying peptide sequence information is generally insufficient. Here, we have devised a method for obtaining abundant structural data on Analytical Chemistry, Vol. 73, No. 18, September 15, 2001

4535

an N-glycosylated peptide, based on complementary fragmentation techniques in an FTICR instrument. The method requires neither enzymatic deglycosylation nor extensive wet chemistry. Strictly complementary behavior was observed for the two peptide fragmentation techniques, electron capture dissociation and infrared multiphoton dissociation, in the fragmentation of an N-glycosylated peptide present in the unfractionated digest from the lectin of E. corallodendron. ECD provided c and z• ions resulting from cleavage of the peptide backbone with no observed loss of sugars. The fragment ions from the ECD investigation yielded a peptide sequence tag of six amino acid residues. In contrast, IRMPD resulted mainly in cleavage of glycosidic linkages, thus revealing structural information about the glycan. The presence of three branch sites could be established. A particular advantage to the method is that no release of the glycans is necessary in order to obtain structural information about both the peptide and the saccharide. The present approach can thus facilitate determination of glycopeptide heterogeneity and eliminate extra sample manipulation steps and sample loss. The present results show that ECD constitutes a valuable complementary tool for the analysis of N-glycosylated peptides. Also, the presented ECD spectrum is, to our knowledge, the first

4536

Analytical Chemistry, Vol. 73, No. 18, September 15, 2001

example of backbone fragmentation of a peptide containing a multiply branched glycan without any observed cleavage of the glycosidic bonds. Finally, because ECD and IRMPD can be performed with the same instrument, the analysis of glycopeptides can be streamlined and performed without extensive sample handling and dispersal. ACKNOWLEDGMENT We thank Christopher L. Hendrickson for helpful discussions and valuable suggestions. This work was supported by the NSF National High Field FT-ICR Facility (CHE-99-09502), Florida State University, the National High Magnetic Field Laboratory, the Swedish Foundation for International Cooperation in Research and Higher Education (STINT), NIH (P41-RR10088), Foundation for Strategic Research (Infection and Vaccinology), Swedish Society for Medical Research, Knut and Alice Wallenberg Foundation, Swedish Society of Medicine, Swedish Society of Clinical Chemistry, and Wilhelm och Martina Lundgrens Stiftelse in Go¨teborg. Received for review March 23, 2001. Accepted July 2, 2001. AC0103470