An Improved Method for Fragmentation and Detection of Cross-Linked

Article pubs.acs.org/ac

Multiple and Sequential Data Acquisition Method: An Improved Method for Fragmentation and Detection of Cross-Linked Peptides on a Hybrid Linear Trap Quadrupole Orbitrap Velos Mass Spectrometer Elena L. Rudashevskaya,† Florian P. Breitwieser,† Marie L. Huber,†,‡ Jacques Colinge,† André C. Müller,† and Keiryn L. Bennett*,† †

CeMM Research Center for Molecular Medicine of the Austrian Academy of Sciences, Vienna, Austria Department of Dermatology, Division of Immunology, Allergy and Infectious Diseases, Medical University of Vienna, Vienna, Austria

‡

S Supporting Information *

ABSTRACT: The identification and validation of cross-linked peptides by mass spectrometry remains a daunting challenge for protein−protein cross-linking approaches when investigating protein interactions. This includes the fragmentation of cross-linked peptides in the mass spectrometer per se and following database searching, the matching of the molecular masses of the fragment ions to the correct cross-linked peptides. The hybrid linear trap quadrupole (LTQ) Orbitrap Velos combines the speed of the tandem mass spectrometry (MS/MS) duty circle with high mass accuracy, and these features were utilized in the current study to substantially improve the confidence in the identification of cross-linked peptides. An MS/MS method termed multiple and sequential data acquisition method (MSDAM) was developed. Preliminary optimization of the MS/MS settings was performed with a synthetic peptide (TP1) cross-linked with bis[sulfosuccinimidyl] suberate (BS3). On the basis of these results, MSDAM was created and assessed on the BS3-cross-linked bovine serum albumin (BSA) homodimer. MSDAM applies a series of multiple sequential fragmentation events with a range of different normalized collision energies (NCE) to the same precursor ion. The combination of a series of NCE enabled a considerable improvement in the quality of the fragmentation spectra for cross-linked peptides, and ultimately aided in the identification of the sequences of the cross-linked peptides. Concurrently, MSDAM provides confirmatory evidence from the formation of reporter ions fragments, which reduces the false positive rate of incorrectly assigned cross-linked peptides.

O

amenable to derivatization. BS3 or other succinimidyl suberates (N-hydroxysuccinimide esters, NHS esters) have proven to be very useful in identifying the interaction sites between different proteins,1,2 and also in structural studies of proteins.3−5 Among all the theoretically possible cross-linked peptide derivatives that can be generated,6 the most common and readily assessable with available software tools are peptides modified with only one molecule of the cross-linker. Cross-links are formed when a cross-linking reagent covalently binds two residues from different peptides either within a single protein or between two different proteins. Loop links are formed when the cross-linker reagents connects two residues in close proximity on the same peptide. Monolinks are generated when a peptide is modified with the cross-linker, but the second reactive group

ne of the single greatest challenges still remaining in protein−protein cross-linking coupled to mass spectrometry is the successful identification and validation of crosslinked peptides. Some of the major factors impeding the advancement of the technology include (i) appropriate fragmentation of the cross-linked peptides to produce a sufficient number of fragment ions to aid in the identification of the cross-linked species and (ii) the high number of theoretically possible combinations of sequences with a similar molecular mass to the cross-linked peptides. The approach is even further compounded when a sample is composed of a large number of different proteins, e.g., multicomponent protein or affinity-purified complexes. The advent of highspeed, high-resolution mass spectrometers now provides a means to address and solve these problems. Bis[sulfosuccinimidyl] suberate (BS3) is a commonly used and favored cross-linker for many applications. BS3 is a watersoluble cross-linker with a spacer distance of 11 Å. The crosslinker is highly reactive and specific for lysine residues that are primarily located on the surface of the proteins and thus © 2013 American Chemical Society

Received: August 30, 2012 Accepted: January 9, 2013 Published: January 9, 2013 1454

dx.doi.org/10.1021/ac302251f | Anal. Chem. 2013, 85, 1454−1461

Analytical Chemistry

Article

HCl, pH 8.0. The prepared sample was reduced, alkylated, and separated on a NuPage 4−12% Bis−Tris gel. The band at approximately 140 kDa that corresponded to the dimer of BSA was excised. After in situ digestion with trypsin, peptides were desalted on a C18 column, concentrated, and reconstituted in 5% FA for LC−MS/MS (for details, see the Supporting Information, p S-11). Static Nanoelectrospray Mass Spectrometry. The 5 pmol/μL desalted, cross-linked TP1 peptides in 50% methanol, 5% formic acid were loaded into a borosilicate nanoelectrospray emitter (Proxeon, Odense, Denmark), infused into the ion source of an LTQ Orbitrap Velos, and analyzed with collisioninduced dissociation (CID) and higher energy collision induced dissociation (HCD). ProteinXXX 1.0 (protein cross-linking function of GPMAW, Lighthouse data, Odense, Denmark) was used to calculate the theoretical masses of fragment ions of the cross-linked TP1 peptide. The masses of all theoretically possible fragments were tracked and assessed for the formation of the fragment ions from the cross-linked peptides at increasing normalized collision energy (NCE) (Supporting Information p S-11). The data were saved as .grf files and transferred to Excel for further analysis and plotting of fragment absolute ion intensity against NCE. As not all the theoretical fragments were generated in the experiment, or had very low intensities, only ions with an absolute intensity over 200 are presented (Figure 2 and Supporting Information Figures S2− S5). Liquid Chromatography−Mass Spectrometry. Mass spectrometry was performed on a hybrid LTQ Orbitrap Velos mass spectrometer (ThermoFisher Scientific, Bremen, Germany) using the Xcalibur version 2.1.0 coupled to an Agilent 1200 HPLC nanoflow system (dual-pump system with one precolumn and one analytical column) (Agilent Biotechnologies, Palo Alto, CA) via a nanoelectrospray ion source using liquid junction (formerly Proxeon, Odense, Denmark). Solvents and liquid chromatography were as described previously.12 Tandem Mass Spectrometry Analysis of Cross-Linked Peptides. The analyses of cross-linked TP1 were performed in a data-dependent acquisition mode using a top 6 HCD method for peptide identification (for details, see the Supporting Information, p S-11). Independent LC−MS/MS analyses at 30%, 35%, 40%, 45%, 50%, 60%, 70%, 80%, and 90% NCE were performed. The analyses of the cross-linked BSA homodimer were performed in a data-dependent acquisition mode with four different MS/MS methods for peptide identification: MSDAM1; MSDAM-2; CID 37.5, FTMS, top 8; CID 30, LTQ, top 15. MSDAM-1 (HCD 80/30/70/40/50, FTMS, top 3) is a top 3 HCD method. Each selected precursor ion was subjected to MS/MS with consecutive 80%, 30%, 70%, 40%, 50% NCE. MSDAM-2 (HCD 80/70/50/CID 35.7, FTMS, top 5) is a top 5 HCD method in combination with CID. Each selected precursor ion was subjected to MS/MS with consecutive 80%, 70%, 50% NCE for HCD and 35.7% NCE for CID. The third (CID 37.5, FTMS, top 8) and fourth methods (CID 30, LTQ, top 15) are top 8 and top 15, respectively, CID methods. Charge states 1+ and 2+ were rejected throughout, and all other settings were identical in all four methods (see the Supporting Information, p S-11). Data Analysis. The acquired data were converted into peak list files with msconvert (ProteoWizard Library v2.1.2708), merged using internally developed Perl scripts, and searched

is hydrolyzed and not conjugated to a residue, thus forming a dead end (Supporting Information Table S-1). It has been shown that cross-linked peptides release reporter ions when fragmented in the collision cell of a mass spectrometer7,8 (Supporting Information Table S-1). Several reporter ions are known for established and well-utilized crosslinkers that belong to the group of NHS esters, e.g., DSS, BS3, plus the shorter and longer analogues of these two cross-linking reagents.9 Unfortunately, diagnostic reporter ions are often very low in signal intensity with poor signal-to-noise ratios and are not routinely observed in every tandem mass spectrometry (MS/MS) spectrum generated from cross-linked peptides. Only approximately 22% of all tandem mass spectra assigned to cross-links or loop links contain reporter ions, and 53% of the spectra are assigned as monolinks.10 A higher collision energy (CE) is required to produce these fragment ions;8 however, the CE required for the generation of superior quality, intense reporter ion fragments is not conducive to the production of a consecutive series of b- or y-ions. Thus, identification of the peptide is compromised. Conversely, lower CE is favorable for larger peptide fragments and subsequent identification. Confident assignment of diagnostic reporter ions, however, is hampered. The aim of this study, therefore, was to combine the detection of reporter ions that confirm the presence of a crosslinked peptide with an improved sequencing of the cross-linked peptide for the same precursor ion. To achieve this goal, we propose an approach that utilizes the speed, sensitivity, and mass accuracy of the hybrid linear trap Orbitrap Velos.

■

EXPERIMENTAL SECTION Materials. The following were used: test peptide 1 (TP1) Ac-TRTESTDIKRASSREADYLINKER, Mr 2880.47 (Creative Molecules, Canada); BS3, Pierce LTQ Velos ESI positive ion calibration solution (ThermoScientific, Austria, Vienna); methanol, acetonitrile (Fisher Scientific, Austria, Vienna); bovine serum albumin (BSA), isopropyl alcohol, tris(hydroxymethyl)aminomethane (Tris), triethylammonium bicarbonate (TEAB) (Sigma-Aldrich, Austria, Vienna); 2-[4-(2hydoxyethyl)-1-piperazinyl]ethanesulfonic acid (HEPES), formic acid 98−100% (FA Suprapur), trifluoroacetic acid (TFA, Uvasol for spectroscopy), water (LiChrosolv), hydrochloric acid (HCl) (Merck, Austria, Vienna); ammonium hydrogen carbonate (NH4HCO3) (AppliChem, Germany, Darmstadt); trypsin, sequencing grade modified, frozen (Promega, Madison, WI); high-performance extraction discs, 47 mm, Empore (3M, Austria, Perchtoldsdorf). General precautions such as using a fume hood, wearing gloves, and eye protection should be considered when working with concentrated acids and organic solvents. Cross-Linking of TP1. An amount of 10 nmol TP1 was diluted in 100 mM TEAB, pH 8.5, reacted with BS3 in a molar ratio of 1:1, incubated for 30 min at 23 °C, and quenched for 15 min with 50 mM NH4HCO3. Digestion with 5 μg/mL trypsin was performed overnight at 37 °C. Prior to mass spectrometric analysis, the cross-linked peptide was desalted on a C18 reversed-phase column,11 concentrated, and stored at 10 pmol/μL in 0.4% FA (for details, see the Supporting Information, p S-11). Cross-Linking of the Bovine Serum Albumin Homodimer. BSA (0.5 mg/mL) was cross-linked with BS3 in 20 mM HEPES buffer, pH 8.0 at a BS3/BSA ratio of 1000:1, incubated on ice for 2 h, and quenched for 15 min with 50 mM Tris− 1455

dx.doi.org/10.1021/ac302251f | Anal. Chem. 2013, 85, 1454−1461

Analytical Chemistry

Article

Figure 1. Dependence of HCD-generated reporter ions from BS3-modified TP1 on NCE: (A) reporter ions from cross-linked peptides; (B) reporter ions from monolinked peptides. The absolute intensity of the fragment ions (y-axis) is plotted against the NCE (x-axis).

Figure 2. Generation of HCD fragment ions from cross-linked peptides of BS3-modified TP1 at different NCE. Fragmentation of the triply charged precursor ion ([M + 3H]3+) at m/z 779.7308: (A) a-chain, b-ions; (B) a-chain, y-ions; (C) b-chain, b-ions; (D) b-chain, y-ions. Bold lines: fragment ion with the highest intensity at NCE below 60%. Regular lines: fragment ion with the highest intensity at NCE above 60%. Graded green, 1+ charged ions; blue, 2+ charged ions; red, 3+ charged ions. Hatched lines indicate fragment ions containing the cross-linker modification. Fragment ions (y-axis) with an absolute intensity over 200 are plotted against NCE (x-axis).

in the preceding section were submitted to xQuest.13 The search in xQuest was performed with ±4 ppm and ±0.025 m/z mass tolerance for HCD fragmentation, and ±4 ppm and ±0.3 m/z for CID fragmentation. Two missed tryptic cleavage sites were allowed. The mass range of the fragment ions was 100− 2000 m/z. Carbamidomethyl cysteine was set as a fixed modification. For cross-linked TP1 peptides, the data were searched against a simulated protein sequence. The amino acid sequence of BSA was modified by manual insertion of the two cross-linked peptides of TP1 (Supporting Information Figure S-1). For cross-linked peptides of the BSA homodimer, oxidized methionine was also set as a variable modification. The data

against the SwissProt database version v2011.06_20110609 (35 683 sequences plus BSA appended) with the search engine Mascot (v2.2.03, MatrixScience, London, U.K.). Submission to Mascot was via a Perl script that performs an initial search with relatively broad mass tolerances on both the precursor and fragment ions (±10 ppm and ±0.6 Da, respectively). Highconfidence peptide identifications were used to recalibrate all precursor and fragment ion masses prior to a second search with narrower mass tolerances (±4 ppm and ±0.025 Da for HCD; and ±4 ppm and ±0.3 Da for CID fragmentation). For details, see the Supporting Information, p S-11. Analysis of BS3-Modified Peptides. To search for modified peptides, recalibrated .mgf files obtained as described 1456

dx.doi.org/10.1021/ac302251f | Anal. Chem. 2013, 85, 1454−1461

Analytical Chemistry

Article

of NCE where each of the fragment ions can be generated (Supporting Information Figure S-3), while the HCD optimum for each fragment ion is narrower. This was true for both the cross-linked peptide (Figure 2, Supporting Information Figures S-2 and S-3) and the monolink peptides (Supporting Information Figures S-4 and S-5). At the same time, HCD generates a larger total amount of different fragments, especially for cross-linked peptides. The advantages of HCD fragmentation over CID were shown recently with a higher number of peptide to spectrum matches for linear peptides14 and also with an improved score and mass accuracy in the identification of phosphopeptides.15,16 On the basis of the data generated from the static nanoelectrospray experiments, the cross-linked TP1 peptide mixture was analyzed by online liquid chromatography− tandem mass spectrometry (LC−MS/MS). To confirm the observations from the nanoelectrospray experiments, the effectiveness of the different fragmentation methods was assessed with the software xQuest.13 The xQuest score was used to estimate the efficiency of the peptide identification and thus to optimize the NCE. Figure 3 shows the xQuest score

were searched against the BSA sequence (>sp|P02769| ALBU_BOVIN), and the reversed database was allowed. The false discovery rate (FDR) was 20 Å24) or bear a bulky affinity group.23 Conversely, methods also exist for cross-linkers that utilize CID cleavage of the disulfide bond within the cross-linker to obtain characteristic ions25,26 Nevertheless, cross-linkers such as BS3 and analogues of differing length (e.g., DSG) are still very well utilized and favored in structural studies of proteins. Combining the data generated from cross-linkers of different length will provide more precise information on the distance between two crosslinked amino acid residues. Thus, MSDAM shows improved and confident identification of cross-linked peptides. The method favors succinimidyl suberates, but can be extrapolated to other cross-linkers for generic use in the field of protein covalent modification. Although MSDAM was developed for an LTQ Orbitrap Velos, the method can be readily extended to the next generation of mass spectrometers, i.e., Orbitrap Elite, Orbitrap Velos Pro, Exactive Plus Orbitrap, Q Exactive. With the development of appropriate computational data analysis, additional benefits are envisaged in the future.

fragment ions are generated from the sequence of the crosslinked peptides. Additionally, specific reporter ions are also produced if the peptide was modified with the cross-linker. As a perspective, the presence of a reporter ion in a block of scans could potentially be used to filter the .mgf file to extract the corresponding scans from the entire data set. This would enable simplification of the .mgf file and facilitate data processing with software programs. A feature such as this can be considered especially critical as current mass spectrometers produce huge data files containing large quantities of information. An additional challenge is the analysis of a mixture of proteins such as cross-linked protein complexes. These together would be a limitation in the study of multicomponent protein complexes. From our experience (data not shown), the analysis of cross-linked data with xQuest was compromised with respect to speed when data was searched against a set of several proteins. Also, many precursor ions that were selected, fragmented, and showed reporter ions indicative of the crosslinker did not match peptides due to poor signals from the both the precursor and fragment ions. For further data analysis it would be highly appealing to combine all spectra generated within one block of an MSDAM experiment into a single, merged .mgf file. In theory, such an approach should lead to (i) a higher number of fragment ion matches, (ii) an improved peptide to spectrum match as reflected by higher scores in xQuest, and ultimately (iii) more cross-linked peptide identifications. Such an approach has been utilized to merge CID and HCD spectra to improve phosphopeptide quantitation.18−20 HCD-generated information on reporter ions was combined with sequence information from CID-fragmented peptides. The improved identification of linear peptides was also reported from complementary low- and high-mass HCD fragmentation.21 At the same time, it was shown that merging CID and electron-transfer dissociation (ETD) spectra resulted in poorer phosphopeptide matching compared to individual searches.22 Our preliminary data showed that this is not as straightforward a procedure as initially envisaged. Even when the number of fragment ions matching the sequence is improved, the score can actually decrease. We postulated that the observed decrease in the score arises from both the increased noise and number of background ion fragments that occurs from merging several files. An approach such as this has high potential but obviously requires improvements either in the algorithm for merging many files and/or postmerge spectral “cleaning” to remove unwanted noise due to different types of fragment ions and/or protein modifications. In addition, scoring algorithms may require adjustment to cope with the complexity of several merged files. Complementary information from different types of fragmentation has the potential to aid in further validation of cross-linked peptides following identification using available cross-linking software and also enlarge the number of identifications. At present, complementary scans aid in manual validation of modified peptides.

■

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: +43-1-4016070010. Fax: +43-1-40160-970000.

■

Author Contributions

The manuscript was written through contributions from all authors. All authors have given approval to the final version of the manuscript. E.L.R. developed the method, conducted the experiments, analyzed and interpreted the data, and wrote the manuscript. M.L.H. provided technical support with 1D SDS− PAGE and sample preparation for LC−MS and assisted in writing sections of the manuscript. A.C.M. contributed intellectually and technically to the method development and

CONCLUSION Combining the advantages of high speed and sensitivity of the hybrid LTQ Orbitrap Velos, we developed a new method for analyzing cross-linked peptides. This approach was successfully applied to a model protein homodimer, BSA, to identify known and new cross-linked peptides. The method outlined here used the cross-linker BS3, but the approach can be applied to 1460

dx.doi.org/10.1021/ac302251f | Anal. Chem. 2013, 85, 1454−1461

Analytical Chemistry

Article

(19) Kocher, T.; Pichler, P.; Schutzbier, M.; Stingl, C.; Kaul, A.; Teucher, N.; Hasenfuss, G.; Penninger, J. M.; Mechtler, K. J. Proteome Res. 2009, 8 (10), 4743−4752. (20) Zhang, Y.; Ficarro, S. B.; Li, S.; Marto, J. A. J. Am. Soc. Mass Spectrom. 2009, 20 (8), 1425−1434. (21) Chi, H.; Sun, R. X.; Yang, B.; Song, C. Q.; Wang, L. H.; Liu, C.; Fu, Y.; Yuan, Z. F.; Wang, H. P.; He, S. M.; Dong, M. Q. J. Proteome Res. 2010, 9 (5), 2713−2724. (22) Kim, M. S.; Zhong, J.; Kandasamy, K.; Delanghe, B.; Pandey, A. Proteomics 2011, 11 (12), 2568−2572. (23) Petrotchenko, E. V.; Serpa, J. J.; Borchers, C. H. Mol. Cell. Proteomics 2011, 10 (2), M110 001420. (24) Liu, F.; Wu, C.; Sweedler, J. V.; Goshe, M. B. Proteomics 2012, 12 (3), 401−405. (25) King, G. J.; Jones, A.; Kobe, B.; Huber, T.; Mouradov, D.; Hume, D. A.; Ross, I. L. Anal. Chem. 2008, 80 (13), 5036−5043. (26) Calabrese, A. N.; Good, N. J.; Wang, T.; He, J.; Bowie, J. H.; Pukala, T. L. J. Am. Soc. Mass Spectrom. 2012, 23 (8), 1364−1375.

MS analysis, contributed to the writing of the manuscript, and provided valuable feedback in the review process. F.P.B. and J.C. performed the processing of raw MS/MS data and prepared the mgf. files for data analysis. F.P.B. assisted in writing sections of the manuscript. K.L.B. was responsible for project supervision, data interpretation, and manuscript writing. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors thank all the members of the Bennett laboratory for helpful input and discussion. Work in our laboratory is supported by the Austrian Academy of Sciences, the Austrian Federal Ministry for Science and Research (Gen-Au projects APP-III and BIN-III), the Austrian Science Fund FWF, and the Central Bank of the Republic of Austria. E.L.R. is supported by the GenAu APP-III program (no. 820965), and M.L.H. is supported by the Central Bank of the Republic of Austria (no. 14252).

■

REFERENCES

(1) Chang, Z.; Kuchar, J.; Hausinger, R. P. J. Biol. Chem. 2004, 279 (15), 15305−15313. (2) Chen, Z. A.; Jawhari, A.; Fischer, L.; Buchen, C.; Tahir, S.; Kamenski, T.; Rasmussen, M.; Lariviere, L.; Bukowski-Wills, J. C.; Nilges, M.; Cramer, P.; Rappsilber, J. EMBO J. 2010, 29 (4), 717−726. (3) Young, M. M.; Tang, N.; Hempel, J. C.; Oshiro, C. M.; Taylor, E. W.; Kuntz, I. D.; Gibson, B. W.; Dollinger, G. Proc. Natl. Acad. Sci. U.S.A. 2000, 97 (11), 5802−5806. (4) Kalkhof, S.; Haehn, S.; Paulsson, M.; Smyth, N.; Meiler, J.; Sinz, A. Proteins 2010, 78 (16), 3409−3427. (5) Leitner, A.; Joachimiak, L. A.; Bracher, A.; Monkemeyer, L.; Walzthoeni, T.; Chen, B.; Pechmann, S.; Holmes, S.; Cong, Y.; Ma, B.; Ludtke, S.; Chiu, W.; Hartl, F. U.; Aebersold, R.; Frydman, J. Structure 2012, 20 (5), 814−825. (6) Schilling, B.; Row, R. H.; Gibson, B. W.; Guo, X.; Young, M. M. J. Am. Soc. Mass Spectrom. 2003, 14 (8), 834−850. (7) Iglesias, A. H.; Santos, L. F.; Gozzo, F. C. J. Am. Soc. Mass Spectrom. 2009, 20 (4), 557−566. (8) Iglesias, A. H.; Santos, L. F.; Gozzo, F. C. Anal. Chem. 2010, 82 (3), 909−916. (9) Santos, L. F.; Iglesias, A. H.; Gozzo, F. C. J. Mass Spectrom. 2011, 46 (8), 742−750. (10) Seebacher, J.; Mallick, P.; Zhang, N.; Eddes, J. S.; Aebersold, R.; Gelb, M. H. J. Proteome Res. 2006, 5 (9), 2270−2282. (11) Rappsilber, J.; Ishihama, Y.; Mann, M. Anal. Chem. 2003, 75 (3), 663−670. (12) Bennett, K. L.; Funk, M.; Tschernutter, M.; Breitwieser, F. P.; Planyavsky, M.; Ubaida Mohien, C.; Muller, A.; Trajanoski, Z.; Colinge, J.; Superti-Furga, G.; Schmidt-Erfurth, U. J. Proteomics 2011, 74 (2), 151−166. (13) Rinner, O.; Seebacher, J.; Walzthoeni, T.; Mueller, L. N.; Beck, M.; Schmidt, A.; Mueller, M.; Aebersold, R. Nat. Methods 2008, 5 (4), 315−318. (14) Frese, C. K.; Altelaar, A. F.; Hennrich, M. L.; Nolting, D.; Zeller, M.; Griep-Raming, J.; Heck, A. J.; Mohammed, S. J. Proteome Res. 2011, 10 (5), 2377−2388. (15) Nagaraj, N.; D’Souza, R. C.; Cox, J.; Olsen, J. V.; Mann, M. J. Proteome Res. 2010, 9 (12), 6786−6794. (16) Jedrychowski, M. P.; Huttlin, E. L.; Haas, W.; Sowa, M. E.; Rad, R.; Gygi, S. P. Mol. Cell. Proteomics 2011, 10 (12), M111 009910. (17) Huang, B. X.; Kim, H. Y.; Dass, C. J. Am. Soc. Mass Spectrom. 2004, 15 (8), 1237−1247. (18) Dayon, L.; Pasquarello, C.; Hoogland, C.; Sanchez, J. C.; Scherl, A. J. Proteomics 2010, 73 (4), 769−777. 1461

dx.doi.org/10.1021/ac302251f | Anal. Chem. 2013, 85, 1454−1461