Phosphorus-Based Absolutely Quantified Standard Peptides for Quantitative Proteomics Nico Zinn,† Bettina Hahn,† Ru ¨ diger Pipkorn,‡ Dominik Schwarzer,† and Wolf D. Lehmann*,† Molecular Structure Analysis, German Cancer Research Center, Heidelberg, Germany, and Central Peptide Synthesis Unit, German Cancer Research Center, Heidelberg, Germany Received June 05, 2009
Abstract: An innovative method for the production of absolutely quantified peptide standards is described. These are named phosphorus-based absolutely quantified standard (PASTA) peptides. As the first step, synthetic phosphopeptides are calibrated via a hybrid LC-(ICP+ESI)MS system. Quantification is achieved by ICP-MS detection of 31P, and identification is performed by ESI-MS. Generation of phosphopeptide standard solutions with this system is demonstrated to provide absolute concentrations with an accuracy better than 10%. The concept was extended to the production of peptide standards by subjecting a PASTA phosphopeptide to gentle and complete dephosphorylation to obtain the cognate PASTA peptide. It is demonstrated that both enzymatic hydrolysis by alkaline or antarctic phosphatase or chemical hydrolysis by hydrofluoric acid can be employed for this purpose. Further, the introduction of one or more stable isotope-labeled amino acids (preferably labeled by 13C, 15N) results in the production of a labeled PASTA peptide, which then can be employed as an internal standard for quantitative analysis by LC-ESI-MS. Using a 1:1 mixture of a stable isotope-labeled PASTA peptide/phosphopeptide pair as dual standard, a quantification between active and inactive recombinant MAP kinase p38R was performed by a combination of tryptic digestion and nanoLC-MS. Keywords: LC-ICP-MS • absolute protein quantification • phosphopeptides • phosphorylation degree • dephosphorylation
Introduction Systems biology attempts to provide a time-resolved and quantitative description of the cellular proteome or parts thereof in the response to internal and external stimuli. To extend the scope of mass spectrometry to systems biology studies, numerous quantitative methods for protein and peptide quantification have been developed. Currently, three different principles can be recognized, which are (i) label-free quantification based on absolute signal intensities1 including * Corresponding author. Tel.: ++49 - 6221 - 424563. Fax: ++49 - 6221 424561. † Molecular Structure Analysis. ‡ Central Peptide Synthesis Unit.
4870 Journal of Proteome Research 2009, 8, 4870–4875 Published on Web 08/08/2009
the semiquantitative spectral count approach,2 (ii) relative quantification by introduction of (labeled) tags via metabolic (e.g., SILAC), enyzmatic (18O water), or chemical (e.g iTRAQ, ICPL) steps,3,4 or (iii) the addition of calibrated stable isotopelabeled peptides as internal standards. Whereas the methods including a derivatization step are best suited but not confined to relative quantifications, the internal standard approach is taylored for absolute quantification and is connected with the acronym AQUA (for absolute quantification) peptides. To obtain absolute quantitative data in isotope dilution experiments, the internal standard has to be quantified on an absolute basis. AQUA peptides5 are labeled by 13C and/or 15N and quantified by amino acid analysis.6 Peptides are subjected to total hydrolysis, and a subset of free amino acids is quantified by a combination of LC with ultraviolet, fluorimetric, or mass spectrometric detection after pre- or postcolumn derivatization. Amino acid analysis is sensitive to salts and detergents used during peptide synthesis and of course to the presence of peptide impurities. Multiple unmodified peptide standards can be generated in one step by expression of an artifical protein composed of concatenated proteotypic peptides (polySIS,7 QCAT,8 QconCAT9 concepts). Quantification of these concatenated standards is challenging since it requires quantification of the intact protein construct. Inductively coupled plasma mass spectrometry (ICP-MS) is classically used for quantification in inorganic chemistry. It offers a high linear dynamic range and sensitivity and is used for routine multielement quantification. The development of interfaces between LC and ICP10,11 has pushed the use of ICP for element quantification in biological samples.12 Several applications of LC-ICP-MS in the pharmaceutical and biological sciences have been described.13-15 In the field of analytical proteomics, detection of phosphorus by LC-ICP has been introduced for phosphopeptide analysis.16 In addition, sulfur detection by LC-ICP-MS has been demonstrated as a tool for protein quantification.17,18 The LC-ICP detection limit for phosphorus is about 50 fmol and that for sulfur about 2 pmol. When these elements are used as surrogates for protein/phosphoprotein or peptide/phosphopeptide quantification, less favorable detection limits are achieved compared to those of ESI- or MALDI-MS with typical detection limits in the lower femtomole to higher attomole region. Therefore, a number of methods concerning chemical tagging with metals19,20 of proteins and peptides for sensitivity enhancement of LC-ICP-MS are under investigation. This study introduces capillary LC-(ICP+ESI)-MS with phosphorus de10.1021/pr900494m CCC: $40.75
2009 American Chemical Society
technical notes
PASTA Peptides for Quantitative Proteomics
Chemicals and Enzymes. Water, acetonitrile, and 2-propanol (element analysis grade) were from Sigma-Aldrich (Munich, Germany). Trifluoroacetic acid (TFA) was from Biosolve (Valkenswaard, Netherland), and heptafluorobutyric acid (HFBA) was from Sigma-Aldrich (Munich, Germany). Phosphatases were from New England Biolabs, and active MAP kinase p38 was obtained from invitrogen. All synthetic peptides were produced in-house by solid-phase synthesis using fmoc amino acids, except for LIEDNEpYTAR which was from Sigma-Aldrich (Munich, Germany), and the four-component phosphopeptide standard mixture (MassPREP phosphopeptide standard enolase) was from Waters (Milford, MA). Bis(4-nitrophenyl)phosphate (BNPP) of 99% purity was purchased from SigmaAldrich (Munich, Germany). Hydrofluoric acid (48%, ultrapure grade) was purchased from Merck (Darmstadt, Germany).
tions of a nominal concentration of 10 nmol µL-1 were prepared gravimetrically. Solutions were prepared in water or water/ acetonitrile in the case of poorly water-soluble peptides. Standards were stored as stock solutions with a nominal concentration of 0.1 nmol µL-1. The true concentrations were determined by the method described below. Dephosphorylation of Peptides. Dephosphorylation was performed by alkaline phosphatase, antarctic phosphatase, or hydrofluoric acid, as indicated. For phosphatase-catalyzed dephosphorylation, phosphatase buffer was added to the peptide solution to obtain a final concentration of 50 mM TRIS-HCl, 0.1 mM EDTA, pH 8.5, and incubated for 1 h with about 0.5 units of phosphatase. Dephosphorylation by hydrofluoric acid21 was performed in Teflon tubes. After 1 h reaction at ambient temperature in aqueous 48% HF, the sample is dried by lyophilization. The dephosphorylated peptide was redissolved in water. Careful consideration of the corresponding safety precautions is recommended when working with HF since this acid is highly toxic upon contact and ingestion. Data Evaluation. Data evaluation was performed using Origin 8G (Northhampton, USA).
Mass Spectrometry
Results and Discussion
Hybrid LC-(ICP+ESI)-MS. The HPLC separations were performed on a capLC system (Waters, Milford, MA) equipped with a 300 µm × 150 mm capillary column (Waters, Milford, MA) packed with C18 particles of 5 µm and a pore size of 300 Å. The capLC was coupled online to a sector field ICP-MS type Element2 (Thermo, Bremen, Germany) through a low-flow microconcentric nebulizer CEI-100 (CETAC, Omaha, USA) and to a ion trap type LCQ Deca XP Plus (Thermo, Bremen, Germany) at a flow rate of 5 µL min-1. Splitting of the LC flow was achieved by a nano-Y-piece connected to two fused silica capillaries (Upchurch Scientific, Oak Harbor, USA) of 25 µm i.d. with 1.64 m length and 50 µm i.d. with 0.18 m length, respectively. As a result of this set up, a flow of about 400 nL min-1 is introduced into the ion trap, while the major flow of 4.6 µL min-1 is directed to the ICP source. Samples were loaded directly on the analytical column. After a washing period of 10 min with 97% A (98% water, 1.5% acetonitrile, and 0.5% 2-propanol with an additional 0.05% TFA and 0.01% HFBA), a linear gradient from 3% B (98% acetonitrile, 1.5% water, and 0.5% 2-propanol with an additional 0.05% TFA and 0.01% HFBA) to 40% B between 20 and 60 min was applied. The ICP-MS was operated with the following conditions: sample gas, 1.0 L min-1; auxiliary gas, 0.8 L min-1; plasma power, 1300 W. 31P and 32S were monitored with a mass spectrometric resolution of 1500 or 4000, respectively. Each selected isotope was recorded by a set of 20 data points with a dwell time of 10 msec each. NanoLC-MS. NanoLC-MS was performed with a nanoLC system type Ultimate (Dionex, Sunniyvale, CA) coupled to an LTQ-Orbitrap XL system (Thermo, Bremen, Germany). NanoESI-MS. NanoESI spectra were recorded on a QToF2 (Waters Micromass, Manchester, UK) using spray tips prepared in-house. The instrument was operated in positive mode with a capillary voltage between 1 and 2 kV. Preparation of the BNPP Standard. Stock solutions of BNPP were prepared gravimetrically in water at about 50 nmol µL-1 and stored at -20 °C until used. Mixtures between BNPP and peptides were prepared immediately prior to analysis. Preparation of Phosphopeptide Solutions. Synthetic phosphopeptides were obtained as lyophilized powder, and solu-
Hybrid LC-(ICP+ESI)-MS. Application of ICP-MS to proteomic analysis requires the coupling to LC so that the method becomesapplicabletomixtureanalysis.However,anLC-ICP-MS analysis can only provide the retention time of a peptide as an additional analytical parameter to the absolute amount of phosphorus and/or sulfur in a chromatographic peak. To add another level of specificity, the LC flow was divided at a ratio of about 9:1 using a y-split. The actual flow rates after the y-split were determined using a microliter syringe. The larger portion of the LC flow was directed to the ICP instrument and the minor portion to an ion trap ESI mass spectrometer. The hybrid LC-(ICP+ESI)-MS system thus established offers the possibility for phosphopeptide quantification and identification within a single analytical run (see Figure 1). For interfacing with ICP, capillary LC with typical flow rates of 1-10 µL/min was selected since it fits best with respect to the typical picomole sample amounts required for quantification of peptides and phosphopeptides. In addition, the typical capLC flow rates are technically well suited for a direct interfacing with the inductively coupled plasma (ICP) source of an element mass spectrometer. Retention Time Matching in a Hybrid LC-(ICP+ESI)-MS System. Since the LC flow is split into capillaries of different length, an analyte reaches the ICP and ESI ion source at different times. To correlate the signals of the two MS techniques, the retention times of the ICP signals were plotted as a function of their corresponding ESI signals. The linear regression analysis of the data is described by the function, retention timeESI ) 3.4 min ((0.4 min) + retention timeICP, which directly implicates a constant retention time shift of +3.4 min for the ESI data in the setup used. For a certain split ratio and flow rate, a single analysis of any standard peptide is sufficient for measurement of the retention time shift. LC Gradient Correction. What favors the application of ICP-MS for quantitative analyses in protein and peptide research is its analyte-independent response, allowing quantification to be performed relative to any internal standard containing the element of interest. However, using a water/ acetonitrile gradient, an increasing acetonitrile content leads to an increased ICP response.22 To account for this gradient
tection as a novel and generally applicable tool for the preparation of absolutely quantified standard peptide and phosphopeptide solutions for quantitative proteomic analyses by ESI- or MALDI-MS.
Experimental Details
Journal of Proteome Research • Vol. 8, No. 10, 2009 4871
technical notes
Zinn et al.
Figure 1. Analysis of an equimolar four-component phosphopeptide mixture using the hybrid LC-(ICP+ESI)-MS system. Upper panel: 31 P trace recorded by ICP-MS showing the four phosphopeptides and the internal standard BNPP. Lower trace: positive ion mode ESI-MS extracted ion chromatogram for the corresponding molecular ion signals.
Figure 2. Absolute quantification of a phosphopeptide solution by capLC-ICP-MS and phosphorus detection. The signal areas of the phosphopeptide (early eluting peak) and of the internal standard (BNPP, late eluting peak) are subjected to gradient correction (right panel). The experimental amount of the AQUA phosphopeptide of 4.6 pmol is in agreement with the specified value of 5 pmol (1 pmol µL-1, injection volume 5 µL).
LC matrix effect on the 31P signal intensity, several approaches have been used. For instance, postcolumn addition of a constant amount of acetonitrile was used23 to reduce the effects of an acetonitrile gradient. Postcolumn addition of a mirrored LC gradient was found to remove the gradient effects,24 since this approach mimics isocratic elution and thus ensures constant operating conditions for the ICP source. To limit the instrumental complexity in the study presented here, the gradient-induced change in the phosphorus ionization efficiency was equalized by a correction function. Since this function is gradient-dependent, it has to be determined individually for each LC gradient. This can be achieved either by measuring the 31P signal intensity with solvents containing an identical phosphorus spike or by monitoring the phosphorus background in a blank run using unspiked solvents. Using spiked solvents, the 31P signal levels are considerably higher compared to the background signals; however, both methods are performed by an identical calculation method as follows: 4872
Journal of Proteome Research • Vol. 8, No. 10, 2009
During the LC gradient, the signal observed during the first 10 min period at constant initial eluent composition is used as the basic signal level to which all subsequent signal intensities observed at later times are normalized. This results in a continuous time- and gradient-dependent correction function. Both methods using spiked solvents or blank runs result in an identical correction function. Therefore, the blank run method was preferentially applied since it does not require the preparation of phosphate-spiked solvents and avoids an additional charging of the LC-MS system with phosphate. Preparation of a Phosphopeptide Standard. The procedure starts with the synthesis of the phosphopeptide of interest containing a stable isotope-labeled (13C, 15N) amino acid. For MS1 detection, the labeled amino acid can be introduced at any position; however, for MSn detection, the distribution of the label over the fragment ions should be considered in addition. For calibration of a phosphopeptide solution, an aliquot is mixed with a known amount of internal standard
PASTA Peptides for Quantitative Proteomics
technical notes
Figure 3. Quantification of phosphorus-containing compounds by capLC-ICP-MS and 31P detection. Left panel: calibration curve of BNPP peak areas showing a good linear correlation over almost 2 orders of magnitude, from 2.5 pmol to 100 pmol injected. Right panel: relative quantification of four different phosphopeptides using BNPP as the internal standard over the same range of injected sample amounts, showing that all results are within an error of 10% as indicated by the dotted lines.
Figure 4. Enzymatic dephosphorylation of the phosphopeptide HTDDEMpTGpYVATR by antarctic phosphatase as controlled by nanoESI-MS. Upper panel before treatment, lower panel after treatment. Full dephosphorylation without side reactions is achieved.
(BNPP) and subjected to capLC-(ICP+ESI)-MS. The chromatogram obtained is normalized to uniform phosphorus response with the experimentally determined correction function, the amount of peptide is calculated, and error propagation calculations are performed. At least three replicates for each quantification were conducted. As an example, Figure 2 shows the original and corrected 31P traces and the correction function for a quantitative analysis of the phosphopeptide LIEDNEpYTAR using BNPP as internal standard. The investigated phosphopeptide represented an AQUA peptide standard (specified 1 nmol, error unknown) dissolved at a concentration of 1 pmol µL-1. The experimentally determined concentration of 0.92 ( 0.08 pmol µL-1 was found to be identical to the specified value within the error of the analysis. To further characterize the stability and accuracy of the quantification, a calibration curve from 2.5 pmol to 100 pmol of the internal standard BNPP was acquired, showing good linear behavior of the absolute signal intensity over the complete range. Reliable quantification of phosphopeptides relative to BNPP is also achieved with an error
of less than 10% independent of the injected amount over almost 2 orders of magnitude (see Figure 3). Preparation of a Peptide Standard. A two-step procedure is introduced as an innovative method for generation of a calibrated peptide solution. First, a calibrated solution of a phosphopeptide is created as described above which then is subjected to a quantitative dephosphorylation step. The perfection of the dephosphorylation step was checked by nanoESI MS. In this way, it was observed that both phosphatases used (alkaline and antarctic phosphatase) were highly effective for removing the phosphate group from phosphopeptides without detectable side reactions. A typical example is shown in Figure 4 for dephosphorylation of the doubly phosphorylated peptide HTDDEMpTGpYVATR by the use of antarctic phosphatase. If complete dephosphorylation is achieved as proven in Figure 4, the molar peptide concentration should be identical to that of the calibrated phosphopeptide standard. Preparation of a Dual Peptide/Phosphopeptide Standard. The preparation of a dual standard containing both the phosphorylated and the nonphosphorylated form of a peptide Journal of Proteome Research • Vol. 8, No. 10, 2009 4873
technical notes
Zinn et al.
Figure 5. Control of the phosphatase inhibition by analysis of the dual 1:1 standard containing HTDDEMTGYVATR and HTDDEMpTGpYVATR by LC-ICP-MS and 32S detection. Before analysis, the dual 1:1 standard was stored for 24 h at ambient temperature. The 32S chromatogram proves the sustained inactivation of the phosphatase used for preparation of the dephosphorylated form. Table 1. Peptide Recovery after Treatment of Phosphopeptides with Alkaline Phosphatase as Determined by capLC-ICP-MS and 32S Detection
sequence
recovery after dephosphorylation by alkaline phosphatase
HTDDEMpTGpYV-[13C3,15N-A]-TR YRSVpTPCDM YRpSVTPCDM
97% ( 3% (S.D., n ) 3)) 93% ( 6% (S.D., n ) 3) 100% ( 3% (S.D., n ) 3)
can be achieved simply by mixing the two standards described above in any desired ratio. Such a dual standard is highly useful for determination of a site-specific phosphorylation degree. To ensure that the phosphopeptide in this mixture is stable, it is essential that the phosphatase used for generation of the nonlabeled analogue is irreversibly inactivated prior to mixing. Both alkaline and antarctic phosphatase can be inactivated by acidification25 to pH < 3. Alternatively, antarctic phosphatase can be irreversibly inactivated by heating26 for 5 min to 65 °C or prolonged exposure to 55 °C. To control the stability of phosphopeptides in dual standards generated as described
above, sulfur-containing peptide/phosphopeptide pairs were prepared. Peptides containing sulfur in the form of Cys and/ or Met are chosen for this purpose, since sulfur is an ICPdetectable tag that is not affected by the phosphorylation status. The synthetic peptide HTDDEMpTGpYVATR was selected for a first test for preparation of a dual standard. The phosphopeptide solution was calibrated by capLC-ICP-MS, an aliquot was dephosphorylated by antarctic phosphatase, the phosphatase was inactivated, and a molar 1:1 mixture of the peptide/phosphopeptide pair was prepared. The corresponding solutions were analyzed by LC-ICP-MS and 32S detection. Figure 5 shows the 31P and 32S LC traces of this 1:1 phosphopeptide/peptide dual standard after 24 h incubation at room temperature in the presence of inactivated phosphatase. The preservation of the 1:1 ratio proves the irreversible inactivation of the phosphatase. In summary, the data in Figure 5 demonstrate the dependability of the proposed strategy. The completeness of the dephosphorylation reaction was further studied by capLC-ICP-MS and 32S detection using three different methionine-containing phosphopeptides. The recoveries were found to be close to 100%, as summarized in Table 1. The work flow for preparation of phosphorus-based absolutely quantified standard (PASTA) peptides outlined above is finally summarized schematically in Figure 6. Application of Dual PASTA Peptides. The pair of standards generated from a single synthetic phosphopeptide may be mixed at any desired ratio for the preparation of a dual PASTA peptide/phosphopeptide mixture. Such a dual standard is useful for the determination of the phosphorylation degree since peptides and phosphopeptides differ in their relative ionization efficiencies in ESI. In the positive ion mode, phosphopeptides often exhibit a reduced ionization efficiency compared to their nonphosphorylated analogues.27 However, the reverse situation is also observed, in particular, for highly basic phosphopeptides.28 Using a standard mixture containing both the phosphorylated and the nonphosphorylated species in an isotopically labeled form, these differences are not of concern. As an example, the relative abundance of the active versus the inactive form of the MAP kinase p38R was determined. The active form is represented by the doubly phosphorylated tryptic peptide HTDDEMpTGpYVATR, whereas the inactive form is connected with the corresponding nonphosphorylated peptide. The two PASTA peptides were assembled
Figure 6. Work flow for the production of phosphorus-based absolutely quantified peptide/phosphopeptide standards (PASTA peptides). First, a phosphopeptide solution is calibrated using capLC-(ICP+ESI)-MS and phosphorus detection with simultaneous control of the peptide molecular weight by ESI-MS. Then, an aliquot is completely dephosphorylated, and a corresponding peptide standard is obtained. 4874
Journal of Proteome Research • Vol. 8, No. 10, 2009
technical notes
PASTA Peptides for Quantitative Proteomics
phosphopeptide standard mixtures since these can be generated from a single synthetic labeled phosphopeptide using a single quantification step.
Acknowledgment. We are indebted to O. Acuto and M. Salek for valuable discussions and to U. Korf for providing recombinant kinases. References
Figure 7. Quantification by nanoLC-MS of the molar ratio within the tryptic peptide pair HTDDEMpTGpYVATR and HTDDEMTGYVATR from MAP kinase p38R using a dual PASTA peptide mixture as the internal standard. An equimolar mixture of the labeled analogues (35 ( 4 fmol) was added as internal standards, carrying [13C3, 15N]-labeled alanine. The display is normalized to the corresponding PASTA peptide signals. As a result, the signal areas of the MAP kinase peptides directly represent their molar ratio, which is 6.3:1.
as 1:1 mixture, added to a tryptic digest of MAP kinase p38R, and analyzed by nanoLC-MS. The signals of the peptide and phosphopeptide molecular ions and of their PASTA analogues are displayed as extracted ion chromatograms in Figure 7. The signals were normalized relative to the internal standards set as unity. In this way, the signal intensities of the protein peptides directly represent their molar ratio. In the described example, the doubly phosphorylated active form is 6.3 times more abundant than the inactive form. Dual PASTA peptide/ phosphopeptide mixtures are ideal for determination of a sitespecific degree of phosphorylation. This strategy can also be combined with a phosphopeptide enrichment step since both species of interest in the sample are accompanied by an isotopically labeled internal standard. The strategy may also be used for the accurate determination of extreme degrees of phosphorylation by customizing the ratio of the dual PASTA peptides to the expected ratio of the analytes.
Conclusion It is concluded that the concept of PASTA peptides is capable to deliver absolutely quantified phosphopeptide and peptide standards containing at least one residue of serine, threonine, or tyrosine. The hybrid LC-(ICP+ESI)-MS system established allows a simultaneous element- and analyte-specific detection of phosphopeptides and therefore delivers highly dependable and analyte-specific quantitative results. By principle, the method is well suited for the preparation of dual peptide/
(1) Higgs, R. E.; Kniermann, M. D.; Gelfanova, V.; Butler, J. P.; Hale, J. E. J. Proteome Res. 2005, 4, 1442–1450. (2) Wienkoop, S.; Larrainzar, E.; Niemann, M.; Gonzalez, M. E.; Lehmann, U.; Weckwerth, W. J. Sep. Sci. 2006, 29, 2793–2801. (3) Ong, S. E.; Mann, M. Nat. Chem. Biol. 2005, 1, 252–262. (4) Bantscheff, M.; Schirle, M.; Sweetman, G.; Rick, J.; Kuster, B. Anal. Bioanal. Chem. 2007, 389, 1017–1031. (5) Kirkpatrick, D. S.; Gerber, S. A.; Gygi, S. P. Nat. Methods 2005, 35, 265–273. (6) Cooper, C. Amino Acid Analysis Protocols; Humana Press: Clifton, NJ, 2001. (7) Anderson, L.; Hunter, C. L. Mol. Cell. Proteomics 2006, 5, 573– 588. (8) Beynon, R. J.; Doherty, M. K.; Pratt, J. M.; Gaskell, S. J. Nat. Methods 2005, 2, 587–589. (9) Pratt, J. M.; Simpson, D. M.; Doherty, M. K.; Rivers, J.; Gaskell, S. J.; Beynon, R. J. Nat. Protoc. 2006, 1, 1029–1043. (10) Montes-Bayo´n, M.; DeNicola, K.; Caruso, J. A. J. Chromatogr. A 2003, 1000, 457–476. (11) Schaumlo¨ffel, D.; Ruiz-Encicar, J.; Lobinski, R. Anal. Chem. 2003, 75, 6837–6842. (12) Sanz-Medel, A.; Montes-Bayo´n, M.; del Rosario Ferna´ndez de la Campa, M.; Ruiz Encinar, J.; Bettmer, J. Anal. Bioanal. Chem. 2008, 390, 3–16. (13) Pereira Navaza, A.; Ruiz Encinar, J.; Sanz-Medel, A. Angew. Chem., Int. Ed. 2007, 46, 569–571. (14) Marshall, P. S.; Leavens, B.; Heudi, O.; Ramirez-Molina, C. J.Chromatogr. A 2004, 1056, 3–12. (15) Brouwers, E. E. M.; Tibben, M.; Rosing, H.; Schellens, J. H. M.; Beijnen, J. H. Mass Spectrom. Rev. 2008, 27, 67–100. (16) Wind, M.; Edler, M.; Jakubowski, N.; Linscheid, M.; Wesch, H.; Lehmann, W. D. Anal. Chem. 2001, 73, 29–35. (17) Wind, M.; Wegener, A.; Eisenmenger, A.; Kellner, R.; Lehmann, W. D. Angew. Chem., Int. Ed. 2003, 42, 3425–3427. (18) Zinn, N.; Kru ¨ger, R.; Leonhard, P.; Bettmer, J. Anal. Bioanal. Chem. 2008, 391, 537–543. (19) Bettmer, J.; Jakubowski, N.; Prange, A. Anal. Bioanal. Chem. 2006, 386, 7–11. (20) Prange, A.; Pro¨frock, D. J. Anal. Atom. Spectrom. 2008, 23, 432– 459. (21) Kuyama, H.; Toda, C.; Watanabe, M.; Tanaka, K.; Nishimura, O. Rapid Commun. Mass Spectrom. 2003, 17, 1493–1496. (22) Allain, P.; Jaunault, L.; Mauras, Y.; Mermet, J. M.; Delaporte, T. Anal. Chem. 1991, 63, 1497–1498. (23) Navaza, A. P.; Ruiz Encinar, J.; Carrascal, M.; Abian, J.; Sanz-Medel, A. Anal. Chem. 2008, 80, 1777–1787. (24) Pereira, A. S.; Schelfaut, M.; Lynen, F.; Sandra, P. J. Chromatogr. A 2008, 1185, 78–84. (25) Pigretti, M. M.; Milstein, C. Biochem. J. 1965, 94, 106–113. (26) Kobori, H.; Sullivan, C. W.; Shizuya, H. Proc. Natl. Acad. Sci. U.S.A. 1984, 81, 6691–6695. (27) Gropengiesser, J.; Varadarajan, B. T.; Stephanowitz, H.; Krause, E. J. Mass Spectrom. 2009, 44, 821–831. (28) Steen, H.; Jebanathirajah, J. A.; Rush, J.; Morrice, N.; Kirschner, M. W. Mol. Cell. Proteomics 2006, 5, 172–181.
PR900494M
Journal of Proteome Research • Vol. 8, No. 10, 2009 4875
