Top-Down Characterization of a Native Highly ... - ACS Publications

Jun 18, 2010 - Address: 460 W 12th Ave., 246 Biomedical Research Tower, Columbus, ... mainly occurred at the amide bonds from C-terminal to aspartic a...
0 downloads 0 Views 2MB Size
Anal. Chem. 2010, 82, 6079–6089

Top-Down Characterization of a Native Highly Intralinked Protein: Concurrent Cleavages of Disulfide and Protein Backbone Bonds Jianzhong Chen,*,†,‡ Pavel Shiyanov,† Liwen Zhang,‡ John J. Schlager,† and Kari B. Green-Church‡ Applied Biotechnology Branch, Air Force Research Laboratory, Dayton, Ohio 45433 and Mass Spectrometry and Proteomics Facility, The Ohio State University, Columbus, Ohio 43210 Top-down analysis of proteins has developed rapidly in recent years. However, its application to disulfide-bonded proteins is still limited. Using native chicken lysozyme as a model, we studied the characteristics of collisioninduced dissociation (CID) of disulfide-bonded proteins on an LTQ Orbitrap mass spectrometer with electrospray ionization (ESI) in positive mode. For low-charged protein precursor ions with no or limited mobile protons, product ions generated from CID correspond to the concurrent cleavages of disulfide and protein backbone bonds. Up to three disulfide bonds could be easily cleaved with four possible dissociation pathways for each disulfide bond. That led to modifications of the corresponding cysteine residues through addition or subtraction of a hydrogen atom or sulfhydryl group. The protein backbone cleavages mainly occurred at the amide bonds from C-terminal to aspartic acid residues (e.g., ion series of b18, b48, y10, and y28), N-Cr bonds from N-terminal to cysteine residues (e.g., c5, ion series of c29 and c63), and amide bonds from C-terminal to glutamic acid residues (e.g., ion series of b35). The characteristics of the top-down analysis for this highly knotted protein will help to understand the general dissociation pattern of disulfide-bonded proteins, which in turn will help to avoid time-consuming bottom-up procedures for the identification of proteins and their modifications. The identification of proteins and protein modifications is usually performed by the traditional bottom-up method. Using this approach, proteins are first digested with a protease (or proteases) and generated peptides are analyzed with mass spectrometry and/ or tandem mass spectrometry analysis, followed by database search for the identification of proteins. However, there are several disadvantages of this approach. Some artificial modifications such as deamidation could be introduced due to digestion conditions.1 The modification information may also get lost because of the * To whom correspondence should be addressed. Address: 460 W 12th Ave., 246 Biomedical Research Tower, Columbus, OH 43210. Fax: 614-292-5955. E-mail: [email protected]. † Air Force Research Laboratory. ‡ The Ohio State University. (1) Ren, D.; Pipes, G. D.; Liu, D.; Shih, L. Y.; Nichols, A. C.; Treuheit, M. J.; Brems, D. N.; Bondarenko, P. V. Anal. Biochem. 2009, 392, 12–21. 10.1021/ac1006766  2010 American Chemical Society Published on Web 06/18/2010

following two reasons:2 (1) usually only a small fraction of the proteolytic peptides of a protein can be detected and even less can be sequenced from the fragments of these peptides; (2) it is more difficult to distinguish a certain modified protein isoform with a small number of copies from the ones with high numbers. In addition, peaks that originate from protease self-digestion can mask peaks from the proteins of interest. Since the early 1990s, an approach called top-down proteomics analysis has been rapidly developing for the identification of proteins.3-5 The development of this approach has significantly accelerated with the recent fairly easy access to high-resolution mass spectrometers including LTQ Orbitrap and Fourier Transform ion cyclotron resonance (FTICR) mass spectrometers. With this approach, intact protein ions are directly fragmented in mass spectrometers without the need of enzymatic digestion. The generated product ions can then be used for the identification of proteins.2 The use of this approach will ease identification of different protein isoforms. With the minimum sample preparation and fast experiment analysis, this method can circumvent the problems of artificial modifications as well as interference from protease self-digestion while preserving the modification information. Proteins as large as 200 kDa have been analyzed with this approach.6 The top-down analysis has been applied to many proteins without disulfide bonds or those with relatively simple disulfide linkages. However, the application of top-down analysis to disulfidebonded proteins is still limited. Although electron capture dissociation (ECD) was reported to be able to cleave both disulfide bond and protein backbone bond, the dissociation efficiency is very low,7,8 while for CID it is usually difficult to generate product ions of proteins from backbone cleavages between residues enclosed by a disulfide bond because of the need to cleave an (2) Chait, B. T. Science 2006, 314, 65–66. (3) McLafferty, F. W.; Breuker, K.; Jin, M.; Han, X.; Infusini, G.; Jiang, H.; Kong, X.; Begley, T. P. FEBS J. 2007, 274, 6256–6268. (4) Kelleher, N. L. Anal. Chem. 2004, 76, 197A–203A. (5) Whitelegge, J.; Halgand, F.; Souda, P.; Zabrouskov, V. Expert Rev. Proteomics 2006, 3, 585–596. (6) Han, X. M.; Jin, M.; Breuker, K.; McLafferty, F. W. Science 2006, 314, 109–112. (7) Zubarev, R. A.; Haselmann, K. F.; Budnik, B.; Kjeldsen, F.; Jensen, F. Eur. J. Mass Spectrom. 2002, 8, 337–349. (8) 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, 2857– 2862.

Analytical Chemistry, Vol. 82, No. 14, July 15, 2010

6079

additional bond.6,8,9 Nevertheless, a few studies by Stephenson, McLuckey, and co-workers reported the cleavages of disulfide bonds along with backbone bonds in proteins with CID using a specific instrument design for ion/ion interactions to improve the resolution.10-12 The product ions assigned in these reports corresponded to a maximum number of only four cleavages of backbone bonds.8,10-12 In addition, the resolution was still not high enough to confidently assign many peaks. Different mechanisms have been proposed for the disulfide bond cleavage in peptides or proteins with different dissociation methods, which can essentially be divided into two types. One type, including ECD,8 electron detachment dissociation (EDD),13 and photodissociation,14 involves radicals, while the other type, including CID and infrared multiphoton dissociation (IRMPD) in either negative10,13,15 or positive mode,12,16-18 involves lone pairs of electrons. The cleavage site for the former type is proposed to happen at the S-S bond with possible hydrogen retention, hydrogen transfer, and loss of S or CH2S. On the other hand, the cleavage sites of disulfide bonds for CID and IRMPD are proposed to be at the C-S and S-S bonds.10,12,13,15-18 The mechanism for the negative mode of the second type of dissociation methods is more consistent between different reports: the cleavages were reported13,15 to result in four products, including addition or subtraction of H and SH, or implied a possibility to yield such products (although only three products were observed for each of the two polypeptides linked by a disulfide bond).10 In contrast, for the analysis in positive mode of the second type of dissociation, different numbers of products (three12,16-18 and six16) from disulfide bond cleavage have been reported. A special case of the second type of the dissociation mechanism, metal ionassisted disulfide bond dissociation, resulted in the loss of H2S2.19,20 Different fragmentation pathways have also been studied with cystine and its derivative small peptides.21 However, the characteristics of these small peptides are probably different from typical peptides or proteins, e.g., both their N and C termini are very close to the disulfide bonds. We report here that using an LTQ Orbitrap mass spectrometer, more than 10 backbone cleavages could be generated from the top-down analysis of chicken lysozyme, a 14 kDa protein containing 129 amino acid residues highly knotted with four disulfide bonds Cys6-Cys127, Cys30-Cys115, Cys64-Cys80, and Cys76Cys94 (Figure 1). Only five N-terminal residues and two C-terminal residues are not enclosed by any disulfide bond. Some amino acid (9) Ryan, C. M.; Souda, P.; Halgand, F.; Wong, D. T.; Loo, J. A.; Faull, K. F.; Whitelegge, J. P. J. Am. Soc. Mass Spectrom. 2010, 21, 908–917. (10) Chrisman, P. A.; McLuckey, S. A. J. Proteome Res. 2002, 1, 549–557. (11) Stephenson, J. L., Jr.; Cargile, B. J.; McLuckey, S. A. Rapid Commun. Mass Spectrom. 1999, 13, 2040–2048. (12) Wells, J. M.; Stephenson, J. L.; McLuckey, S. A. Int. J. Mass Spectrom. 2000, 203, A1–A9. (13) Kalli, A.; Hakansson, K. Int. J. Mass Spectrom. 2007, 263, 71–81. (14) Fung, Y. M. E.; Kjeldsen, F.; Silivra, O. A.; Chan, T. W. D.; Zubarev, R. A. Angew. Chem., Int. Ed. 2005, 44, 6399–6403. (15) Zhang, M. X.; Kaltashov, I. A. Anal. Chem. 2006, 78, 4820–4829. (16) Bean, M. F.; Carr, S. A. Anal. Biochem. 1992, 201, 216–226. (17) Tie, J. K.; Mutucumarana, V. P.; Straight, D. L.; Carrick, K. L.; Pope, R. M.; Stafford, D. W. J. Biol. Chem. 2003, 278, 45468–45475. (18) Wallis, T. P.; Huang, C. Y.; Nimkar, S. B.; Young, P. R.; Gorman, J. J. J. Biol. Chem. 2004, 279, 20729–20741. (19) Kim, H. I.; Beauchamp, J. L. J. Am. Chem. Soc. 2008, 130, 1245–1257. (20) Kim, H. I.; Beauchamp, J. L. J. Am. Soc. Mass Spectrom. 2009, 20, 157– 166. (21) Lioe, H.; O’Hair, R. A. J. J. Am. Soc. Mass Spectrom. 2007, 18, 1109–1123.

6080

Analytical Chemistry, Vol. 82, No. 14, July 15, 2010

Figure 1. Crystal structure of chicken lysozyme with the disulfide bond linkages indicated. The image was generated by using online Jmol software (version 11.6) based on coordinates in file PDB ID 2lym obtained from ref 37 in the RCSB Protein Data Bank (http:// www.rcbs.org/pdb/).

residues of chicken lysozyme are enclosed by as many as three (amino acid sequence segment 65-75 and 81-93) and even four disulfide bonds (amino acid sequence segment 77-79). By CID of the low charge state of lysozyme cations, product ions were observed corresponding to the cleavages of 1-3 of the four disulfide bonds along with the protein backbone bonds. The cleavage of each disulfide bond apparently modified the cysteine residues by adding or subtracting a hydrogen atom (H) or sulfhydryl group (SH). It led to a generation of two peak clusters (referred to as ion series in the following discussion). Each cluster consisted of four products, i.e., F - SH, F - H, F + H, and F + SH, where F represents either N or C terminal fragments (b, c, or y ions) generated by homolytic cleavage of the disulfide bond. The mass differences between the sequential neighboring peak pairs (F - SH and F - H, F - H, and F + H, and F + H and F + SH) were 32, 2, and 32 Da, respectively. To explain the generation of these products, we propose four dissociation pathways (Scheme 1, see details in the section Proposed Fragmentation Pathways for the Dissociation of Disulfide Bonds). There were more products for a peak cluster with multiple disulfide bonds cleavages (see Results and Discussion for details). The corresponding protein backbone cleavages mainly occurred at the C terminal to aspartic acid residue (Asp-Xxx) and resulted in product ions including the b18, b48, y10, and y28 ion series. Ions corresponding to other cleavages are also discussed. MATERIALS AND METHODS Chicken lysozyme was purchased from Sigma (St. Louis, MO). HPLC-grade water, acetonitrile, and acetic acid were purchased from ThermoFisher Scientific (Waltham, MA). Ten or twenty micromolar chicken lysozyme solution was prepared in aqueous 2% acetic acid. The analysis was performed on an LTQ Orbitrap XL mass spectrometer (ThermoFisher Scientific, Waltham, MA) with direct infusion at a flow rate of 10 µL/min. Electrospray

Scheme 1

The monoisotopic neutral masses for the product ions were typically determined either from the monoisotopic peaks (if their intensities were high) or the most abundant peaks (if the monoisotopic peaks were of very low intensity or undetectable). The latter can be determined from the mass differences between the most abundant peaks and the monoisotopic peaks for proteins of the same molecular weights with average compositions (averagine).22 The mass differences, for convenience, were manually calculated based on the number of units (integer Da values, which can be determined from the theoretical distribution of averagine) between the most abundant peaks and those of the corresponding monoisotopic peaks, assuming the unit being the same as the

average value 1.00235.23 This average was found to change the mass accuracies less than 0.3 ppm compared to the accurate distance between the monoisotopic peak and the most abundant peak for an averagine. The averagine and its distribution were generated with software ICR2Ls (downloaded at http://omics. pnl.gov/software/ICR2LS.php from Pacific Northwest National Laboratory’s website). In the case of the isotopic peaks of two species overlapping with each other, the second isotopic peaks were utilized for calculation of the monoisotopic peaks. Once the monoisotopic neutral masses were determined, they were input in ICR2Ls’ tool “Protein Fragmentation Generator” to determine the identities of these product ions. The search was performed against the theoretical fragments including possible modifications for chicken lysozyme amino acid sequence. The modifications were set to include six types: water loss, ammonia loss, SH loss, H loss, H gain, and SH gain. The last four modifications corresponded to the cleavage of disulfide bonds and were set for cysteine residues only. The maximum number for each of these four modifications was set as 1 since there are altogether four disulfide bonds (see details in the section Proposed Fragmentation Pathways for the Dissociation of Disulfide Bonds). The maximum numbers of water and ammonia losses were typically set as 1. The mass of cysteine residue was set to its oxidized form mass (102.00136 Da). The mass accuracy was typically set to be within 6 ppm. For product ions with multiple matches, the assignments were determined by taking into

(22) Senko, M. W.; Beu, S. C.; Mclafferty, F. W. J. Am. Soc. Mass Spectrom. 1995, 6, 229–233.

(23) Horn, D. M.; Zubarev, R. A.; McLafferty, F. W. J. Am. Soc. Mass Spectrom. 2000, 11, 320–332.

ionization and positive mode detection were applied. The instrument parameters were as follows: source voltage at 5 kV, capillary voltage at 50 V, tube lens voltage at 125 V, and capillary temperature at 225 °C. The resolution was set at 30 000. The scan range was m/z 435-2000 for MS analysis, while for MS/MS analysis, high m/z was 2000 and low m/z was automatically set. The isolation width for MS/MS analysis was set as 10, which provided good isolation efficiency and a high signal-to-noise ratio. The normalized collision energy was 20-30. The acquisition time was 20 min for the precursor ion with 9+ charges and 2-6 min for the precursor ions with other charge states. DATA ANALYSIS

Analytical Chemistry, Vol. 82, No. 14, July 15, 2010

6081

Figure 2. Spectra of b182+ (a) and y10+ (b) product ion series with the monoisotopic peaks labeled. The spectra were expanded from the MS/MS spectrum of chicken lysozyme precursor ion (M + 9H)9+. The normalized collision energy applied was 30. The product ions resulted from the cleavages of one disulfide bond (Cys6-Cys127) and one Asp-Xxx amide bond (Asp-Asn and Asp-Val for the b182+ and y10+ ion series, respectively). Cleavage of the disulfide bond resulted in four different products (shown as subtraction or addition of H or SH). Product ions with additional ammonia loss for the y10+ ion series are also shown. The theoretical isotopic distributions of the b182+ and y10+ ions are shown as the insets. The isotopic distributions for other ions of these two series are similar to those for the two ions.

consideration the following factors: (1) labile cleavage of Cterminal to aspartic acid residue in a nonmobile protein system,24 (2) generation of specific c and z ions with cleavage from N-terminal to cysteine residues,10,12 (3) general fragmentation characteristics of peptides,25 and (4) correlation between the number of disulfide bond cleavages and the combination of modifications for cysteine residues in the fragments. RESULTS AND DISCUSSION Cleavage of One Asp-Xxx Bond and One Disulfide Bond. CID tandem mass spectrometry analysis was performed on lysozyme ions with different charge states. The MS/MS spectra for protein ions with relatively low charge states (especially (M + 9H)9+) provided the richest information. Most of the product ions for these low charge state protein ions did not match the cleavages between the free amino acid residues (not enclosed by any disulfide bond) and could only be assigned by considering the cleavages of both disulfide bonds and the protein backbone bonds. More information obtained at such low charge state is probably related to the lack of mobile protons. In the case of 9+ lysozyme ions, all or at least most of the charges are sequestered by the 11 arginine residues of the lysozyme molecule.26 The lack of mobile protons increases the protein (24) Gu, C.; Tsaprailis, G.; Breci, L.; Wysocki, V. H. Anal. Chem. 2000, 72, 5804–5813. (25) Huang, Y.; Triscari, J. M.; Tseng, G. C.; Pasa-Tolic, L.; Lipton, M. S.; Smith, R. D.; Wysocki, V. H. Anal. Chem. 2005, 77, 5800–5813. (26) Wysocki, V. H.; Tsaprailis, G.; Smith, L. L.; Breci, L. A. J. Mass Spectrom. 2000, 35, 1399–1406.

6082

Analytical Chemistry, Vol. 82, No. 14, July 15, 2010

dissociation energy,26 which becomes sufficient for cleavage of disulfide bonds. At the same time, it renders the Asp-Xxx bond cleavages dominant24 and thus simplifies the data analysis in this study. The analysis for even lower charge state (