Analysis of Protein Tyrosine Phosphorylation by Nanoelectrospray

May 1, 2003 - A novel highly sensitive strategy is introduced for analysis of tyrosine phosphorylation in previously identified proteins channelling f...
0 downloads 12 Views 123KB Size
Download PDF
Anal. Chem. 2003, 75, 2724-2729

Analysis of Protein Tyrosine Phosphorylation by Nanoelectrospray Ionization High-Resolution Tandem Mass Spectrometry and Tyrosine-Targeted Product Ion Scanning Mogjiborahman Salek,† Angel Alonso,‡ R. Pipkorn,§ and Wolf D. Lehmann*,†

Central Spectroscopy Unit, Department for Cell Differentiation, and Central Peptide Synthesis Unit, German Cancer Research Center, Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany

A novel highly sensitive strategy is introduced for analysis of tyrosine phosphorylation in previously identified proteins channelling for this aim all analytical and sequence information available. Nanoelectrospray high-resolution MS/MS analysis is targeted to precalculated m/z values corresponding to phosphotyrosine-containing tryptic peptides. Identification of these peptides is supported by the occurrence of the phosphotyrosine immonium ion at m/z 216, neutral loss of 79.97/z () loss of HPO3), and similarity of the fragmentation patterns of phosphotyrosine-containing peptides with their nonphosphorylated analogues. This tyrosine-targeted tandem mass spectrometry strategy is demonstrated for epidermal growth factor receptor showing that phosphotyrosine-containing tryptic peptides invisible in the survey spectrum can be safely identified. Protein phosphorylation is an important covalent modification involved in many signal transduction processes controlling, for example, cell morphology, cell-cell interaction, cell proliferation, differentiation, and apoptosis.1 Disregulations of these essential cellular processes lead to many diseases including cancer.2 In eukaryotic cells, protein phosphorylation frequently occurs at threonine, serine, and tyrosine residues, which results in esterification of their hydroxyl function with a phosphate group. Although phosphorylation at tyrosine represents only ∼1% of these events, this subclass deserves special interest due to its outstanding function in cellular signaling cascades. Mass spectrometry is a core technology for characterization of protein phosphorylation,3-5 since it is capable of executing the most informative steps within the multistep analytical strategies developed for this purpose. These strategies all start with * Corresponding author. Tel: ++49-6221-424563. Fax: ++49-6221-424554. E-mail: [email protected]. † Central Spectroscopy Unit. ‡ Department for Cell Differentiation. § Central Peptide Synthesis Unit. (1) Hunter, T. Cell 2000, 100, 113-127. (2) Blume-Jensen, P.; Hunter, T. Nature 2001, 411, 355-365. (3) McLachlin, D. T.; Chait, B. T. Curr. Opin. Chem. Biol. 2001, 5, 591-602. (4) Zhou, H. L.; Watts, J. D.; Aebersold, R. Nat. Biotechnol. 2001, 19, 375378. (5) Mann, M.; Ong, S. E.; Gronborg, M.; Steen, H.; Jensen, O. N.; Pandey, A. Trends Biotechnol. 2002, 20, 261-268.

2724 Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

phosphoprotein isolation and proteolytic cleavage followed by either phosphopeptide recognition (spotting) or their specific enrichment and end with a tandem MS step. At the phosphopeptide level, enrichment may be performed by immobilized metal affinity chromatography, which is based on the interaction between the phosphate monoester group and a partially chelated, immobilized three-valent metal cation such as Fe3+ or Ga3+.6-8 Another method for enrichment utilizes the replacement of pSer or pThr residues by an affinity label.9,10 At the phosphoprotein level, pY phosphoproteins can be enriched using immobilized antiphosphotyrosine antibodies.11 The first indicator for a phosphopeptide in a protein digest is the observation of a 79.97/z () HPO3/z) mass difference in the survey spectrum, a criterion that cannot be applied for the (minority of) cases of a 100% phosphorylation stoichiometry. More sophisticated methods for phosphopeptide spotting rely on a specific, collision-induced fragmentation process. In the negative ion mode, phosphopeptides generate marker ions at m/z 96.97 () H2PO4-) and 78.96 () PO3-), which can be used for their recognition by precursor ion scanning12 or skimmer CID LC/ESIMS.13,14 These anions are highly specific for peptide phosphate esters; however, the poor amount of sequence information present in the MS/MS spectra of peptide anions is a serious drawback of this ionization mode. Since current mass spectrometers are designed neither for simultaneous detection of negative and positive ions nor for fast polarity switching, strategies utilizing exclusively the positive ion mode are strongly preferred. In the standard positive ion polarity, observation of a neutral loss of 97.97/z () H3PO4/z) is the method of choice for recognition of (6) Nuwaysir, L.; Stults, J. T. J. Am. Soc. Mass Spectrom. 1993, 4, 662-669. (7) Posewitz, M. C.; Tempst, P. Anal. Chem. 1999, 71, 2883-2892. (8) Ficarro, S. B.; McLeland, M. L.; Stukenberg, P. T.; Burke, D. J.; Ross, M. M.; Shabanowitz, J.; Hunt, D. F.; White, F. M. Nat. Biotechnol. 2002, 20, 301-305. (9) Oda, Y.; Nagasu, T.; Chait, B. T. Nat. Biotechnol. 2001, 19, 379-382. (10) Goshe, M. B.; Veenstra, T. D.; Panisko, E. A.; Conrads, T. P.; Angell, N. H.; Smith, R. D. Anal. Chem. 2002, 74, 607-616. (11) Pandey, A.; Podtelejnikov, A. V.; Blagoev, B.; Bustelo, X. R.; Mann, M.; Lodish, H. F. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 179-184. (12) Carr, S. A.; Huddleston, M. J.; Annan, R. S. Anal. Biochem. 1996, 239, 180192. (13) Huddleston, M. J.; Bean, M. F.; Carr, S. A. Anal. Chem. 1993, 65, 877884. (14) Jedrzejewski, P. T.; Lehmann, W. D. Anal. Chem. 1997, 69, 294-301. 10.1021/ac020657y CCC: $25.00

© 2003 American Chemical Society Published on Web 05/01/2003

pS- and pT-phosphopeptides,15,16 whereas for pY peptides, a characteristic loss of m/z 79.97 () HPO3) is observed17 and a specific pY immonium ion at m/z 216.04 is formed. This pY marker ion was first documented in an ESI-MS/MS spectrum of a pYcontaining tetrapeptide but was erroneously interpreted as sequence ion.18 Then, a signal at m/z 216 was observed in the MALDI PSD spectrum of a pY peptide from in vitro phosphorylated calmodulin and suggested to represent the immonium ion of pY.19 In a following triple quadrupole ESI-MS/MS study using synthetic pY peptides, a signal at m/z 216 was also observed with high abundance. It was identified as the pY immonium ion by its isotopic distribution and was used for specific detection of tyrosine phosphorylation in a mixture of a pY and pS peptide using precursor ion scanning.20 High-resolution precursor ion scanning using a Q-TOF instrument was subsequently used for specific detection of pY peptides in digests of phosphoproteins enriched by the use of immobilized anti-pY antibodies.21-23 In the MS-based strategies, the final step generally is the recording of a corresponding product ion spectrum. In this study, we describe a tyrosine-targeted strategy that channels all available information to achieve both optimal sensitivity and optimal confidence for identification of tyrosine phosphorylation sites in previously identified proteins. EXPERIMENTAL SECTION Reagents. All solvents were purchased from E. Merck (Darmstadt, Germany). Synthetic peptides were prepared in-house. Activated MAP kinase ERK1 was purchased from CalbiochemNovabiochem (San Diego, CA). Modified trypsin sequencing grade was from Roche Diagnostics (Mannheim, Germany). All other reagents were from Sigma-Aldrich (Taufkirchen, Germany). Isolation of EGF Receptor. HaCaT cells containing ∼106 EGF receptors/cell were cultured in DMEM containing 10% fetal calf serum and penicillin/streptomycin until 90% confluency and then serum starved for 24 h. Human recombinant EGF (Roche) was added to 50 or 100 ng/mL, and the cells were first incubated for 10 min. Cells were then washed twice with ice-cold PBS and covered with ice-cold buffer (800 µL/10-cm-diameter Petri dish) 30 mM Tris, pH 7.4, 150 mM NaCl, 1 mM EDTA, 0.5% Triton X-100, 0.5% sodium deoxycholate, 10 mM NaF, and protease and phosphatase inhibitors (Sigma) for 10 min. The cells were then scraped into a cold Eppendorf tube and centrifuged for 10 min at 14.000 rpm at 4 °C. The supernatant was saved, and the protein content was measured using the dc-protein assay system of BioRad. (15) Hunter, A. P.; Games, D. E. Rapid Commun. Mass Spectrom. 1994, 8, 559570. (16) Schlosser, A.; Pipkorn, R.; Bossemeyer, D.; Lehmann, W. D. Anal. Chem. 2001, 73, 170-176. (17) Tholey, A.; Reed, J.; Lehmann, W. D. J. Mass Spectrom. 1999, 34, 117123. (18) Hoffmann, R.; Wachs, W. O.; Berger, R. G.; Kalbitzer, H. R.; Waidelich, D.; Bayer, E.; Wagner-Redeker, W.; Zeppezauer, M. Int. J. Pept. Protein Res. 1995, 45, 26-34. (19) Annan, R. S.; Carr, S. A. Anal. Chem. 1996, 68, 3413-3421. (20) Lehmann, W. D. in Proceedings of the 32rd Annual Meeting of the German Mass Spectrometry Society, Oldenburg, 1999; p 112. (21) Steen, H.; Kuster, B.; Mann, M. J. Mass Spectrom. 2001, 36, 782-790. (22) Steen, H.; Kuster, B.; Fernandez, M.; Pandey, A.; Mann, M. Anal. Chem. 2001, 73, 1440-1448. (23) Steen, H.; Pandey A.; Andersen JS.; Mann M. Sci. STKE 2002, 154, 16.

Table 1. Charge-State Selection for the m/z List Generation of pY-Containing Peptides Used in Tyrosine-Targeted MS/MS Analysis. mol wt range

selected charge state

300-600 600-800 800-2000 2000-3000 >3000

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

For immunoprecipitation, monoclonal antibodies to human EGF receptor produced in our laboratory were used. The antibodies recognize an epitope of the intracellular part of the receptor (unpublished results). Immunoprecipitation of ∼1 mg of protein was performed using 1 mL of hybridoma cell supernatant in the presence of 200 µL of protein G bound to magnetic beads (Miltenyi, Bergisch-Gladbach, Germany) for 45 min at 4 °C. After washing thoroughly, the receptors were eluted using 1% SDS and the proteins were separated by gel electrophoresis. Sample Preparation for nanoESI-MS. Proteins were separated by standard one-dimensional SDS-polyacrylamide gel electrophoresis using NuPAGE Bis- Tris gels (4-12%, MES). Proteins were stained by Coomassie blue (Simply Blue Safe Stain, Invitrogen, Karlsruhe, Germany). In-gel digestion was performed by trypsin as previously described omitting the reduction and alkylation step. Supernatants obtained after tryptic digestion were pooled and desalted with pipet tips packed with reversed phase C18 material (ZipTip, Millipore, Bedford, MA). NanoESI Mass Spectrometry. Mass spectra were recorded using a hybrid Q-TOF mass spectrometer type Q-TOF 2 (Micromass, Manchester, U.K.). Using the automated MS to MS/MS switching option, tandem mass spectra of a list of preprogrammed m/z values were acquired. Spray capillaries were manufactured in-house using a micropipet puller type P-87 (Sutter Instruments, Novato, CA) and coated with a semitransparent film of gold in a sputter coater type SCD 005 (BAL-TEC AG, Balzers, Liechtenstein). For each m/z value, tandem MS spectra were recorded using five collision energies to ensure formation of both sequenceand composition-specific fragment ions. A single scan of 24-s duration was performed for each individual collision offset value. To allow fragmentation of precursor ions with very low signal-tonoise ratio the intensity threshold was set to zero. Charge states of the tyrosine-containing tryptic peptides were selected according to the scheme given in Table 1. RESULTS AND DISCUSSION General Information. Tyrosine phosphorylation is characterized by a low degree of phosphorylation, making the mass difference criterion particularly valuable for the detection of pY phosphorylation. As an example, Figure 1 shows the survey spectrum of a tryptic digest of activated EGF receptor, from which all three autophosphorylation sites (Tyr1173 in T1202+, Tyr1148 in T1182+, Tyr1068 in T1152+, T ) tryptic fragment) can be recognized on the basis of this characteristic mass difference. In detail, the measured mass differences in the order as listed were 79.974, 79.966, and 79.965 Da (average of three values using the three main isotopes of each isotopic cluster), which is in good agreement with the calculated mass difference of 79.965 Da Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

2725

Figure 1. NanoESI survey spectrum of a tryptic digest of activated EGF receptor. The m/z region covering the tryptic peptides T120 (GSTAENAE-Y1173-LR), T118 (GSHQISLDNPD-Y1148-QQFFDPK), and T115 (YSSDPTGALTEDSIDDTFLPVPE-Y1068-INQSVPK) is shown containing the three autophosphorylation sites as indicated in italic type. For all three peptides, both the nonphosphorylated and the phosphorylated forms are detected. Sample amount estimated to 1-5 pmol.

(maximum deviation + 8 mDa) for dephospho/phospho ion pairs. In case sufficient sample amount is available (∼1 pmol or more), the mass difference is a good criterion to direct the subsequent MS/MS analysis to selected phosphopeptide candidate ions. In view of the complexity of protein digests, a tandem MS step is generally required for a reliable phosphopeptide identification and is necessary for pinpointing of the phosphorylated site(s). For low amounts of analyte, a low phosphorylation stoichiometry, or both, the mass difference criterion becomes less effective, since most often only the nonphosphorylated peptide is detected. This is because ESI has a lower ionization efficiency toward phosphopeptides compared to their nonphosphorylated analogues,24 so the detection of tyrosine-phosphorylated peptides in a survey spectrum is further compromised by the typically low degree of phosphorylation at this residue. Nevertheless, under optimized conditions, a meaningful ESI-MS/MS spectrum of a mixture component can be recorded even if the corresponding molecular ion signal is not visible among the background signals. In the following, we outline the channeling of all available information for the setup of a tyrosine-targeted strategy for highly sensitive analysis of protein tyrosine phosphorylation. pY Immonium Ion Formation. This pY immonium ion is specific for phosphotyrosine and was observed for all pYcontaining peptides analyzed in our laboratory so far. However, pronounced intensity variations as a function of collision offset are observed. As an example, Figure 2a shows the pY immonium ion formation for the nonapeptide D-pY-QQDFFPK and gives the dependence of its relative abundance on the collision offset. Since typically immonium ion formation requires more than one bond cleavage, relatively high offset values are required for optimal immonium ion abundance. As shown in Figure 2, the maximum intensity of the pY immonium ion is observed between 50 and 80 V offset. For offset values higher than the optimum value, the ion at m/z 216.043 is further fragmented and gives rise to two fragment ions at m/z 136.076 and 118.066 (data not shown). These ions correspond to the neutral loss from the pY immonium of H3PO4 and HPO3, respectively. (24) Carr, S. A.; Huddleston, M. J.; Annan, R. S. Anal. Biochem. 1996, 239, 180192.

2726 Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

Figure 2. Offset and sequence dependency of pY immonium ion formation from pY peptides: (a) low-mass region of the MS/MS spectrum of D-pY-QQDFFPK at 60 V offset; (b) intensity of the pY immonium ion displayed in (a) as a function of the collision offset; (c) optimal pY immonium ion intensity as a function of the position of the pY residue within the nonapeptide sequence. The intensity is given relative to the molecular ion abundance at the optimal collision offset. The immonium ion is most effectively formed from a position near the N-terminus and less effectively from a position near the Cterminus.

To investigate the dependency of the pY immonium ion formation on the sequence, we studied a set of peptides derived from the sequence DQQDFFPK, to which pY was added in all positions from 1 to 9. When fragmenting this set of peptides, we observed an optimal m/z 216 intensity when pY was placed at the N-terminus or at the directly adjacent position. This corresponds to the observation that the N-terminal position generally favors immonium ion formation, since in this case a particular one-step mechanism may contribute to their formation.25 The generally high abundance of the b2 ion may also favor immonium ion formation from the second position from the N-terminus, since immonium ions are formed by a multistep fragmentation.25 Therefore, those parts of a peptide present in the most abundant sequence ions will also contribute most to the formation of immonium ions. It is generally observed that basic sites such as arginine, lysine, and histidine block fragmentation of the peptide (25) Ambihapathy, K.; Yalcin, T.; Leung, H. W.; Harrison, A. G. J. Mass Spectrom. 1997, 32, 209-215.

backbone.26 According to the mobile proton model, this effect is explained by the trapping of protons thus being unavailable to catalyze cleavage of the peptide backbone amide bond.27 This model may also explain the observation shown in Figure 1c. Here the pY immonium ion intensity decreases as the pY residue is moving toward the basic residue at the N-terminus. For the case of nontryptic peptides, pY immonium ion formation is most favorable for positions near the peptide ends, in case these are carrying only acidic or neutral amino acids (data not shown). In summary, we conclude that the relative abundance of the pY immonium ion may vary considerably with respect to peptide sequence. Low relative intensities are to be expected when the pY residue is located in a region that is only poorly fragmented, as generally observed for the vicinity of basic residues. Larger peptides (>1500 Da) often show poorly fragmenting regions in MS/MS analyses. Thus, it can be concluded that the occurrence of the pY immonium ion is a useful marker ion with respect to its specificity, but that it cannot serve as a universal indicator, since the efficiency of its formation is heavily dependent on peptide size and sequence. MS/MS Spectra of pY Peptides and Their Nonphosphorylated Analogues. Due to the facts discussed above, criteria additional to mass difference and pY immonium ion formation are useful, in particular to support a highly sensitive detection of tyrosine phosphorylation. The situation that phosphorylation at tyrosine residues is characterized by a low degree of phosphorylation carries a certain advantage in that the nonphosphorylated analogue of a pY peptide is normally present in excess. This means, that the MS/MS spectrum of this nonphosphorylated analogue normally can be recorded with a better signal-to-noise ratio. This spectrum can serve as “reference” for interpretation of the fragment ions of the pY peptide, since the overall shape of the MS/MS spectra of peptide pairs differing only by phosphorylation at tyrosine is very similar. This favorable situation is due to the observation that the exchange of a neutral residue against an acidic one in a peptide chain generally has a minor effect on its fragmentation pattern in the positive ion mode. Thus, the influence of phosphorylation at tyrosine on the MS/MS spectrum is mainly a mass shift of those fragment ions carrying the additional phosphate group. Figure 3 demonstrates this effect by comparison of the MS/MS spectra of the T24 and the pT24 (Tyr 204) peptide of activated MAP kinase, recorded with an identical set of offset values. In the low-mass region, the T24 fragment shows only the Y immonium ion at m/z 136, whereas the pT24 fragment shows both the Y and the pY immonium ion. Besides this difference, both spectra are highly identical in the m/z region below the triply charged molecular ion, at least with respect to the more abundant fragment ions. The y ions in the higher m/z range are shifted by +80 Da for the pT24 fragment, but the relative abundances of the y ions are retained. Over all, the pT24 peptide shows a lower fragmentation efficiency. This might be caused by its slightly higher molecular weight or an increased stability due to additional intramolecular ionic interactions. Despite these minor differences, the gross structure of the MS/MS spectrum of the T24 peptide is preserved after phosphorylation at Tyr, and this is (26) Willard, B. B.; Kinter, M. J. Am. Soc. Mass Spectrom. 2001, 12, 12621271. (27) Wysocki, V. H.; Tsaprailis, G.; Smith, L. L.; Breci, L. A. J. Mass Spectrom. 2000, 35, 1399-1406.

Figure 3. NanoESI MS/MS spectra of two tryptic fragments of activated MAP kinase ERK1, MS/MS spectra of T24 and pT24 showing the similar fragmentation behavior of these two peptides. Table 2. Occurrence of Tyrosine in Selected Proteins with Tyrosine Phosphorylation Sitesa

protein name DYRK 1A PDPK MAPK ERK1 EGF receptor EGF receptor intracellular domain

no. of residues

no. of Y residues

no. of tryptic peptides

764 556 379 1186 542

29 () 3.8%) 23 () 4.1%) 18 () 4.7%) 36 () 3.0%) 20 () 3.7%)

73 44 39 121 57

no. of tryptic peptides with Y 18 () 25%) 17 () 39%) 11 () 28%) 27 () 22%) 16 () 28%)

a The absolute and relative abundances of tyrosine residues and tyrosine-containing tryptic peptides are listed.

even true after introduction of a second phosphorylation at threonine (Thr202, data not shown). Thus, although this similarity is a “soft” criterion in the identification of a tyrosine phosphorylation site, it is a valuable support in the interpretation of lowintensity MS/MS spectra, as will be shown below. Tyrosine-Targeted Phosphorylation Analysis. The two points addressed above, marker ion detection and MS/MS spectrum similarity, both refer to data interpretation. In the following we introduce a strategy for improving the quality of the MS/MS spectra of pY peptides by optimization of the spectra acquisition time. The signal-to-noise ratio of a spectrum increases in proportion to the square root of this time. In tyrosine phosphorylation analysis of a previously identified protein, optimal sensitivity can be approached by directing the available tandem MS analysis time only to those m/z values, which represent candidate phosphotyrosine peptides. Since tyrosine is a relatively low-abundant amino acid residue (∼4%), this strategy avoids the fact that the continuous m/z range is selected for fragmentation, as required, for example, for precursor ion scanning. Table 2 summarizes the result of such a tyrosine-targeted selection for three phosphoproteins phosphorylated at tyrosine, showing that only between 20 and 40% of all tryptic peptides are selected. The m/z values for programming of the tandem mass spectrometer can be calculated by considering only tyrosine-containing tryptic peptides and then adding a phosphate ester group for each Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

2727

Figure 4. NanoESI MS/MS analysis of activated EGF receptor by tyrosine-targeted precursor ion scanning as described in the text: (A) expanded view of the survey spectrum showing the m/z range covering the ions of T118, T120 and their phosphorylated analogues; (B) selected ion chromatogram for m/z 216.04 ( 0.02 Da extracted from the pool of MS/MS spectra. Estimated sample amount, significantly less than 1 pmol.

tyrosine residue. We used the EGF receptor protein as an example, for which we calculated all possible tryptic pY peptides by considering the m/z values of doubly and triply charged ions (see Experimental Section). According to the general experience that methionine residues are often partially or completely oxidzed to their sulfoxides, the corresponding m/z values were also included in the mass list. For the EGF receptor, we considered only the intracellular domain, which carries the signal transduction-correlated tyrosine phosphorylation sites. Taking into account a precursor ion mass window of 3 Da, all relevant candidate pY phosphopeptides are covered with an MS/MS analysis within only 8.5% of the total mass range from m/z 300 to 1600. Thus, compared to a scan covering the complete mass range (e.g., a normal precursor ion scan), the measuring time is increased by more than 1 order of magnitude, resulting in an improvement in the signal-to-noise-ratio between 3 and 4. Using the list of precalculated m/z values for MS/MS experiments does not necessarily represent the most effective use of the available analysis time, since a complete sequence coverage is usually not observed. This limitation of the introduced tyrosine-targeted analytical strategy can be overcome as follows; however, this would probably require appropriate software support. At first, the survey spectrum is checked for the occurrence of tryptic peptides. Using this information, the MS/MS analysis is directed only to those tryptic pY peptide m/z values, for which the nonphosphorylated counterparts were observed. To achieve a complete sequence coverage, the same procedure then can be repeated using a different protease, such as chymotrypsin, or endoproteinases Asp-N or GluC. 2728 Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

Figure 5. NanoESI-MS/MS spectra of the T56 peptide from activated EGF receptor: (A) spectrum of pT120 generated by tyrosine-directed MS/MS analysis as visualized in Figures 4 and 5; (B) fragmentation pattern of T120 under identical experimental conditions. The comparison of the two MS/MS spectra confirms the phosphorylation at Tyr1173.

Highly Sensitive Tyrosine Phosphorylation Analysis of EGF Receptor. A sample of immunoprecipitated activated EGF receptor was separated by 1D gel electrophoresis. A very weakly Coomassie-stained spot was in-gel digested with trypsin, the total digest was subjected to nanoESI-MS/MS, and tyrosine-targeted phosphorylation analysis was performed as described above. A result from this analysis is displayed in Figure 4. Figure 4B shows a single ion trace of m/z 216.04 extracted from all MS/MS spectra acquired over the complete tandem MS analysis time of ∼1 h, which was performed using a list with the m/z values of all tryptic Y- and pY-containing peptides. It is evident that the pY marker ion is only found in two MS/MS spectra corresponding to the molecular ions of the monophosphorylated peptides pT118 and pT120 containing two of the three autophosphorylation sites of the EGF receptor. The expanded view of the survey spectrum (Figure 4A) shows that the pT118 and pT120 peptides are not or only poorly visible, and that their nonphosphorylated counterparts are visible but of low abundance. Both MS/MS spectra of the molecular ion of pT1182+ (at m/z 772) and of pT1202+ (m/z 645) provide sufficient sequence information for their identification. As an example, Figure 5 shows the MS/MS spectra of the T120 of the pT120 peptide. The MS/MS spectrum of the pT120 peptide given in Figure 5A is of very low abundance, which is not surprising, since it is not visible in the survey spectrum of the EGF receptor digest. However, comparison with the MS/MS spectrum of the more abundant nonphosphorylated T120 peptide

Figure 6. NanoESI-MS/MS spectrum of the pT115 peptide from activated EGF receptor (YSSDPTGALTEDSIDDTFLPVPE-pY1068-INQSVPK). Phosphorylation at Tyr1068 is evident by the proline-directed fragment y10 and y12 ions, which are both accompanied by ions indicating a loss of HPO3. A pY immonium ion is not detected in this MS/MS spectrum.

given in Figure 5B allows a confident identification of this phosphopeptide and a pinpointing of Tyr1173 as the phosphorylation site. The third autophosphorylation site is present in pT115, a phosphopeptide with 31 amino acid residues (∼3.5 kDa). This tryptic fragment cannot be recognized as a pY-containing peptide by the pY immonium ion (see Figure 4B) due to its virtual absence. The MS/MS spectrum of pT115 is dominated by the y2, y10, and y12 fragment ions generated by proline-directed fragmentation, as shown in Figure 6. The presence of pY and its location at Y1068 is recognized by the accurate m/z values of these ions and is further confirmed by the loss of HPO3 from both the y10 and y12 fragments. This loss can be identified by the accurate mass differences of 79.962 Da relative to the y10 ion and 79.981 Da relative to the y12 ion, corresponding to a deviation of - 4 and +15 mDa from the theoretical value, respectively. In addition, the

MS/MS spectrum of the nonphosphorylated T115 peptide shows an identical fragmentation pattern (data not shown). In conclusion, it has been demonstrated that optimized MS/ MS conditions established by tyrosine-targeted product ion scanning result in highly sensitive detection of protein tyrosine phosphorylation by high-resolution MS and MS/MS. Further, it is shown that an increased level of confidence in the identification of pY-containing peptides is established by using the MS/MS spectra of the nonphosphorylated analogues as reference. In this way, partly redundant information is created, which is highly welcome for any mass spectrometric analysis of covalent protein modification at an extreme level of sensitivity. Received for review October 22, 2002. Accepted March 11, 2003. AC020657Y

Analytical Chemistry, Vol. 75, No. 11, June 1, 2003

2729