Detection of Tyrosine Phosphorylated Peptides by Precursor Ion

Tri Hoang , Udo Roth , Karen Kowalewski , Christopher Belisle , Kerstin Steinert and Michael Karas. Analytical Chemistry 2010 82 (1), 219-228. Abstrac...
0 downloads 0 Views 180KB Size
Download PDF
Anal. Chem. 2001, 73, 1440-1448

Detection of Tyrosine Phosphorylated Peptides by Precursor Ion Scanning Quadrupole TOF Mass Spectrometry in Positive Ion Mode Hanno Steen,† Bernhard Ku 1 ster,‡ Minerva Fernandez,† Akhilesh Pandey,§,⊥ and Matthias Mann*,†,‡

Protein Interaction Laboratory at the Center for Experimental Bioinformatics, Department of Biochemistry and Molecular Biology, University of Southern Denmark, Campusvej 55, DK-5230 Odense M, Denmark, MDS Protana A/S, Staermosegaardsvej 6, DK-5230 Odense M, Denmark, Whitehead Institute for Biomedical Research, Cambridge, Massachusetts 02142, Department of Pathology, Brigham and Women’s Hospital, Boston, Massachusetts 02115

Phosphorylation is a common form of protein modification. To understand its biological role, the site of phosphorylation has to be determined. Generally, only limited amounts of phosphorylated proteins are present in a cell, thus demanding highly sensitive procedures for phosphorylation site determination. Here, a novel method is introduced which enables the localization of tyrosine phosphorylation in gel-separated proteins in the femtomol range. The method utilizes the immonium ion of phosphotyrosine at m/z 216.043 for positive ion mode precursor ion scanning combined with the recently introduced Q2-pulsing function on quadrupole TOF mass spectrometers. The high resolving power of the quadrupole TOF instrument enables the selective detection of phosphotyrosine immonium ions without interference from other peptide fragments of the same nominal mass. Performing precursor ion scans in the positive ion mode facilitates sequencing, because there is a no need for polarity switching or changing pH of the spraying solvent. Similar limits of detection were obtained in this approach when compared to triple-quadrupole mass spectrometers but with significantly better selectivity, owing to the high accuracy of the fragment ion selection. Synthetic phosphopeptides could be detected at 1 fmol/µL, and 100 fmol of a tyrosine phosphorylated protein in gel was sufficient for the detection of the phosphorylated peptide in the unseparated digestion mixture and for unambiguous phosphorylation site determination. The new method can be applied to unknown protein samples, because the identification and localization of the modification is performed on the same sample. As more and more genomes are sequenced completely and identification of gel-separated proteins is applied more routinely, * Corresponding author: Phone: +45-65502364. Fax: E-mail: [email protected]. † University of Southern Denmark. ‡ MDS Protana A/S. § Whitehead Institute for Biomedical Research. ⊥ Brigham and Women’s Hospital.

the next logical step is the analysis of protein modifications and the roles that such modifications play in vivo. One of the most commonly encountered functionally important protein modifications is phosphorylation. It is estimated that one-third of all proteins present in a mammalian cell are phosphorylated, and this estimate is underscored by the notion that 2-5% of the expressed genome may encode for kinases.1-4 Protein phosphorylation events are involved in a multitude of regulatory mechanisms which include metabolism, cell division, cell growth, and cell differentiation. Serine, threonine, tyrosine, histidine, arginine, lysine, cysteine, glutamic acid, and aspartic acid residues can be modified by a phosphate group. However, phosphorylation of hydroxyl groups of serine, threonine, or tyrosine residues are by far the most prevalent.2 Although tyrosine phosphorylation is less common than serine or threonine phosphorylation (relative occurrence: pS, ∼90%; pT, ∼10%; pY, ∼0.05%), it plays a crucial role in, for instance, receptor-mediated signaling pathways.4,5 To elucidate the physiological role of a particular phosphorylation event, the exact site of the modification has to be determined. The analysis of in vivo phosphorylation sites is hindered by the fact that the proteins involved are often present in only low copy numbers and that individual sites are often only partially phosphorylated. As a result, techniques used for the analysis of in vivo phosphorylation sites have to be sensitive in the low- to subpicomol range in order to be practical. The traditional way to localize phosphorylation sites is to label the proteins under investigation with radioactive phosphorus isotopes using [33P]- or [32P]-γ-ATP followed by protease digestion of the protein and separation of the peptide mixture by twodimensional thin-layer chromatography (TLC) or HPLC. Sitedirected mutagenesis or deletion experiments are subsequently performed so that the site of mutation may be correlated with the site of phosphorylation and the biological function. Although

+45-65933929.

1440 Analytical Chemistry, Vol. 73, No. 7, April 1, 2001

(1) Hunter, T. Cell 1987, 50, 823-829. (2) Hunter, T. Methods Enzymol. 1991, 200, 3-37. (3) Hubbard, M. J.; Cohen, P. Trends Biochem. Sci. 1993, 18, 172-177. (4) Hunter, T. Philos. Trans. R. Soc. London, Ser. B 1998, 353, 583-605. (5) Hunter, T.; Sefton, B. Proc. Natl. Acad. Sci. U.S.A. 1980, 77, 1311-1315. 10.1021/ac001318c CCC: $20.00

© 2001 American Chemical Society Published on Web 03/03/2001

this approach is very sensitive, it is also very tedious. A more direct way to determine phosphorylation sites is to elute the phosphorylated peptide from the TLC plate or to collect the radioactive HPLC fractions for Edman sequencing.6-8 However, problems can arise from the facts that the PTH derivative of phosphotyrosine is hardly soluble under standard Edman sequencing conditions9 and that phosphoesters of threonine and serine undergo β-elimination, giving undefined products in the following reactions under the conditions used for Edman sequencing. As a result, a peak representing a phosphoserine or phosphothreonine can be assigned only with ambiguities.10-12 The problems derived from Edman sequencing can be circumvented by using mass spectrometry for the subsequent determination of the peptide sequence.13 In vivo labeling is relatively inefficient because of the presence of endogenous ATP. To obtain detectable amounts of labeled peptides or proteins, large quantities of radioactivity are required, which might constitute a potential health hazard and, thus, require special laboratory safety precautions. Mass spectrometry is increasingly becoming the method of choice for the localization of phosphorylation sites, because no radioactivity is required for detection and because of the inherent sensitivity and capability of performing mixture analysis, thus eliminating the need for peptide separation. In addition, mass spectrometry is faster and less tedious when compared to the more traditional techniques. Several mass spectrometry-based techniques have been successfully employed for the mapping of phosphorylation sites. One of those techniques utilizes the relative instability of the phospho group. The partial loss of the phospho moiety can be observed as a metastable ion in MALDI-TOF mass spectrometry;14,15 however, the presence of metastable ions in a MALDI-TOF spectrum is not limited to phosphopeptides and is, therefore, not very specific when complex peptide mixtures are analyzed. Qin and Chait have observed similar fragmentation on a MALDI-iontrap mass spectrometer in which signals originating from phosphopeptides are accompanied by a satellite of -98 Da, corresponding to the loss of H3PO4.16 Whether or not the detected peptides are indeed phosphorylated can be confirmed by treating the sample with phosphatase.17-19 Mass spectra of a phospho(6) Hunter, T.; Sefton, B. M. In Methods in Enzymology; Abelson, J. N., Simon, M. I., Eds.; Academic Press: San Diego, 1991; Vol. 200 and 201. (7) van der Geer, P.; Hunter, T. Electrophoresis 1994, 15, 544-554. (8) Yan, J.; Packer, N.; Gooley, A.; Williams, K. J. Chromatogr. A 1998, 808, 23-41. (9) Aebersold, R.; Watts, J.; Morrison, H.; Bures, E. Anal. Biochem. 1991, 199, 51-60. (10) Meyer, H.; Hoffmann-Posorske, E.; Korte, H.; Heilmeyer, L. J. FEBS Lett 1986, 204, 61-66. (11) Campbell, D.; Hardie, D.; Vulliet, P. J. Biol. Chem. 1986, 261, 10489-10492. (12) Dedner, N.; Meyer, H. E.; Ashton, C.; Wildner, G. F. FEBS Lett. 1988, 236, 77-82. (13) Affolter, M.; Watts, J.; Krebs, D.; Aebersold, R. Anal. Biochem. 1994, 223, 74-81. (14) Annan, R. S.; Carr, S. A. Anal. Chem. 1996, 68, 3413-3421. (15) Schno ¨lzer, M.; Lehmann, W. Int. J. Mass Spectrom. Ion Proc. 1997, 169/ 170, 263-271. (16) Qin, J.; Chait, B. T. Anal. Chem. 1997, 69, 4002-4009. (17) Yip, T.-T.; Hutchens, T. W. FEBS Lett. 1992, 308, 149-153. (18) Wang, K. Y.; Liao, P.-C.; Allison, J.; Gage, D. A.; Andrews, P. C.; Lubman, D. M.; Hanash, S. M.; Strahler, J. R. J. Biol. Chem. 1993, 268, 1426914277. (19) Liao, P.-C.; Leykam, J.; Andrews, P. C.; Gage, D. A.; Allison, J. Anal. Biochem. 1994, 219, 9-20.

protein digest before and after phosphatase treatment are unchanged except for the signal of the phosphorylated peptide, which is shifted downward by 80 Da as the result of the removal of the phospho group. The combination of MALDI-TOFMS and phosphatase treatment has become popular, because it is fast and simple;20,21 however, no information of the exact phosphorylation site is obtained from differential mass measurements alone. Difficulties in this approach can arise from (i) presence of several potential phosphorylation sites within a peptide, (ii) occurrence of isobaric peptides, and (iii) analysis of unknown proteins. In MALDI mass spectra of complex peptide mixtures, phosphopeptides might go unnoticed because of poor ionization efficiency or low site occupancy. In such cases, phosphopeptides need to be enriched prior to mass spectrometric analysis. One effective technique for the partial purification of phosphopeptides from protein digests is immobilized metal affinity chromatography (IMAC). This method exploits the relatively high affinity of phosphate to FeIII or GaIII ions.22,23 IMAC has been interfaced to MS either off-line24 or on-line, coupled with electrospray ionization.25 However, because the metal cation in IMAC columns also acts as an electron acceptor (Lewis acid) in general, problems can arise from samples with high salt content, from copurification of peptides containing many acidic amino acids, or from peptides containing electron donor (Lewis base)-comprising residues (e.g., histidine or free cysteine), because such peptides also have the tendency to bind to the IMAC resin.26 Therefore, peptides isolated in this manner cannot a priori be expected to represent phosphopeptides, and thus, either phosphatase treatment or peptide sequencing experiments are necessary to confirm the presence of a particular phosphopeptide and to determine the site of phosphorylation. A more specific means of phosphopeptide detection in peptide mixtures utilizes the preferred loss of the phospho group upon low-energy collisional activation. In positive ion tandem MS, an intense neutral loss of 98 Da corresponding to H3PO4 is observed for serine and threonine phosphorylated peptides, whereas phosphotyrosine-containing peptides are generally stable under these conditions. Covey et al. have used neutral loss scanning on a triplequadrupole MS for the identification of phosphopeptides in an LC/MS/MS experiment.27 This method suffers from the fact that the value of the neutral loss depends on the charge state of the precursor ion such that only a limited set of phosphopeptides can be detected in one LC/MS/MS experiment. On the other hand, because all data are acquired in the positive ion mode, on-line sequencing of the phosphopeptide is at least theoretically possible. In negative ion tandem MS experiments, all phosphopeptides yield a fragment ion at m/z -79 corresponding to PO3-. In contrast (20) Zhang, X. L.; Herring, C. J.; Romano, P. R.; Szczepanowska, J.; Brzeska, H.; Hinnebusch, A. G.; Qin, J. Anal. Chem. 1998, 70, 2050-2059. (21) Cortez, D.; Wang, Y.; Qin, J.; Elledge, S. J. Science 1999, 286, 1162-1166. (22) Andersson, L.; Porath, J. Anal. Biochem. 1986, 154, 250-254. (23) Posewitz, M. C.; Tempst, P. Anal. Chem. 1999, 71, 2883-92. (24) Betts, J. C.; Blackstock, W. P.; Ward, M. A.; Anderton, B. H. J. Biol. Chem. 1997, 272, 12922-12927. (25) Nuwaysir, L. M.; Stults, J. T. J. Am. Soc. Mass Spectrom. 1993, 4, 662-669. (26) Chicz, R.; Regnier, F. In Methods Enzymol.; Deutscher, M. P., Ed.; Academic Press: San Diego, CA, 1990; Vol. 182, pp 392-421. (27) Covey, T.; Shushan, B.; Bonner, R.; Schro¨der, W.; Hucho, F. In Methods in Protein Sequence Analysis; Jo ¨rnvall, H., Ho¨o ¨g, J.-O., Gustavsson, A.-M., Eds.; Birkha¨user Verlag: Basel, 1991; pp 249-256.

Analytical Chemistry, Vol. 73, No. 7, April 1, 2001

1441

to the neutral loss scan, the fragment at m/z -79 is independent of the charge state of the precursor ion, hence, giving a more comprehensive picture of the phosphopeptides present in a peptide mixture. Combining this approach together with m/z-dependent skimmer fragmentation in an LC/MS experiment on a single quadrupole MS, Huddleston et al.28 and, later, Ding et al.29 were able not only to localize the phosphopetide-containing fractions in an LC run, but also to measure the precursor mass of the phosphopeptides in the same experiment. A similar approach was taken later by Mann et al.,30,31 as well as Carr et al.,32 in which a precursor ion scan of m/z -79 on a triple-quadrupole MS in conjunction with nanoelectrospray ionization was employed to identify phosphopeptides in unseparated peptide mixtures. Subpicomol limits of detection were achieved for in-solution,32 and for in-gel, protein digests31 of model phosphoproteins when sprayed from solutions of pH 10. However, using nanoelectrospray negative ion MS has a number of disadvantages. Initiating a stable spray in the negative ion mode is frequently hampered by discharges from the tip of the spray capillary. In addition, peptide sequencing by negative ion MS/MS is not feasible, and thus, an additional positive ion tandem MS experiment is required for the localization of the phosphorylation site. This generally means either the rebuffering of the sample into a solution of pH 1 to pH 4 or the use of additional sample. The buffer change from basic to acidic conditions can sometimes be circumvented by spraying from neutral solutions; however, this compromises sensitivity in negative as well as in positive ion mode.30,32 Nonetheless, because of their superior sensitivity, nanoelectrospray phospho-precursor ion scanning experiments have found successful application in a number of studies for the localization of phosphorylation sites of gel-separated proteins.24,33,34 In this paper, an analytical approach is described for the sensitive detection and phosphorylation site determination of phosphotyrosine containing peptides from crude peptide mixtures. The method overcomes some of the current limitations of phosphorylation analysis, including difficulties associated with negative ion tandem mass spectrometry. Key features of the approach are: (i) detection of phosphorylated peptides by highaccuracy, high-resolution precursor ion scanning for the immonium ion of phosphotyrosine (m/z 216.043) in the positive ion mode on a quadrupole TOF mass spectrometer equipped with pulsed fragment ion injection and (ii) phosphopeptide sequencing in the same experiment for phosphorylation site determination. Analytical sensitivity in the femtomol range has been achieved for the analysis of synthetic peptides, and peptides from gelseparated phosphoproteins and in vivo phosphorylated signaling molecules. (28) Huddleston, M. J.; Annan, R. S.; Bean, M. F.; Carr, S. A. J. Am. Soc. Mass Spectrom. 1993, 4, 710-717. (29) Ding, J.; Burkhart, W.; Kassel, D. B. Rapid Commun. Mass Spectrom. 1994, 8, 94-98. (30) Wilm, M.; Neubauer, G.; Mann, M. Anal. Chem. 1996, 68, 527-533. (31) Neubauer, G.; Mann, M. Anal. Chem. 1999, 71, 235-242. (32) Carr, S. A.; Huddleston, M. J.; Annan, R. S. Anal. Biochem. 1996, 239, 180192. (33) Weijland, A.; Williams, J. C.; Neubauer, G.; Courtneidge, S. A.; Wierenga, R. K.; Superti-Furga, G. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 3590-3595. (34) Verma, R.; Annan, R. S.; Huddleston, M. J.; Carr, S. A.; Reynard, G.; Deshaies, R. J. Science 1997, 278, 455-460.

1442

Analytical Chemistry, Vol. 73, No. 7, April 1, 2001

MATERIALS AND METHODS Chemicals and Compounds. Chemicals were obtained from Sigma (St. Louis, MO). High-purity solvents used for nanoelectrospray experiments were purchased from Labscan (Dublin, Ireland). The peptide TNLSEQ(pY)ADVYR was custom-made by Sigma-Genosys (Pampisford, UK). The unphosphorylated counterpart was prepared by dephosphorylation using alkaline phosphatase (Roche Diagnostics, Mannheim, Germany) in 50 mM NH4HCO3 at 37 °C for 1 h. Single-strand DNA binding protein (SSB) was purchased from Stratagene; human transferrin was from Sigma, and recombinant His-tagged RrmA was prepared inhouse. Activated MAP-kinase 2 (MAPK) was purchased from Upstate Biotechnology (Waltham, MA). For in-gel digests, concentrations of 0.1-1 pmol MAPK were loaded onto 4-12% NuPage gels (Novex, San Diego, CA) and visualized either by colloidal Coomassie Blue staining (Colloidal Blue Staining Kit, Novex) or silver staining. In-gel reduction, alkylation, and tryptic digestion were performed as described previously.35 Cell Culture and Immunoprecipitation. For immunoprecipitation experiments, a total of 7 × 107 HeLa cells were grown in Dulbecco’s modified Eagle medium (DMEM) containing 10% fetal bovine serum (FBS). Cells were grown to 80% confluence and then cultured for an additional 15 h without serum. Cells were either untreated (control) or treated (experiment) with 1 µg/mL of epidermal growth factor (EGF, Upstate Biotechnology) for 5 min and subsequently lysed in 50 mM Tris-HCl pH 7.6, 150 mM NaCl, 1% Nonidet P-40, and 1 mM sodium orthovanadate in the presence of protease inhibitors. Cleared cell lysates were incubated overnight at 4 °C with a mixture of antiphosphotyrosine antibodies: 30 µg of 4G-10 monoclonal antibody coupled to agarose beads (Upstate Biotechnology) and 10 µg of biotinconjugated RC20 monoclonal antibody bound to streptavidinagarose beads (Transduction Laboratories, Lexington, KY). Precipitated immune complexes were washed three times with lysis buffer and then eluted twice with 100 mM phenyl phosphate in 1× PBS at 37 °C. Proteins from the control and the experiment were separated by SDS-PAGE under reducing conditions. After visualizing by silver-staining, bands of interest were excised and subjected to in-gel reduction, alkylation, and tryptic digestion, as previously described.35 Mass Spectrometry. All experiments were performed on a QSTAR Pulsar quadrupole time-of-flight tandem mass spectrometer (AB/MDS-Sciex, Toronto, Canada) equipped with a nanoelectrospray ion source (MDS Protana, Odense, Denmark). Precursor ion scanning experiments were acquired using a dwell time of 50 ms at a step size of 0.5 Da and with the Q2 pulsing function turned on. Nitrogen was used as the collision gas. The Q0-voltage, which determines the collision energy, was set to a value corresponding to one-tenth of the m/z value of the precursor ion. Serial dilutions of synthetic peptides were prepared at the appropriate concentrations and without additional purification in 5% formic acid/60% methanol. Protein digests were desalted and concentrated using a double column of POROS R2 and OLIGO R3 material (Perseptive Biosystems, Framingham, MA) packed into GELoader tips (Eppendorf, Hamburg, Germany), as described (35) Shevchenko, A.; Wilm, M.; Vorm, O.; Mann, M. Anal. Chem. 1996, 68, 850-858.

previously.31,36 Columns were eluted in three steps (20%/40%/60% methanol in 5% formic acid, respectively) directly into nanospray needles (MDS Protana), and each fraction was subjected to MS analysis. Proteins were identified by searching peptide sequence tags,37 derived from fragment ion spectra of selected peptides, against the nonredundant protein database maintained and updated regularly at the European Bioinformatics Institute (EBI, Hinxton, UK) using the program PepSea (MDS Protana). RESULTS AND DISCUSSION Phosphotyrosine-Specific Precursor Ion Scanning on Quadrupole TOF Instruments. The use of the immonium ion of phosphotyrosine [(Im(pY), m/z 216.043)] was evaluated for the detection of phosphotyrosine-containing peptides by precursor ion scanning on a quadrupole TOFMS. This ion was described by Hoffmann et al. as a characteristic fragment ion for phosphotyrosine-containing peptides,38 and Lehmann suggested the use of this fragment as a “reporter ion” for tyrosine phosphorylated peptides in precursor ion scans on triple-quadrupole instruments.39 However, the application of triple-quadrupole MS for this experiment is not sufficiently specific because many a-, b-, and y-type peptide fragment ions give rise to signals at the same nominal mass of 216 Da (Table 1, nomenclature: Roepstorff and Fohlmann,40 as modified by Biemann41). The relatively low resolution of quadrupoles precludes the differentiation of these ions, and thus, false positives are often encountered in respective precursor ion mass spectra. Close inspection of the masses in Table 1 reveals that the mass of the phosphotyrosine immonium ion is at least 250 ppm smaller than those of the closest interfering peptide fragments. This mass difference can be easily resolved by quadrupole TOF mass spectrometers which provide resolution in excess of 5000 (fwhm) and mass accuracy of better than 50 ppm. Because high resolution is available in both MS and MS/MS mode, precursor ion scans with very accurate fragment ion selection can be obtained by scanning the first quadrupole and fragmenting all precursor ions in the collision cell in turn. Subsequently appropriate data extraction from the continuously acquired product ion TOF spectra generates a precursor ion spectrum with the user-defined fragment ion accuracy and resolution.42,43 However, until recently, the applicability of the precursor ion function was compromised by a much lower transmission of small m/z species into the orthogonal TOF when compared to ion transmission in triplequadrupole instruments. This low transmission originates from the requirement that the continuous ion beam, created by electrospray ionization, has to be converted into ion packages in (36) Gobom, J.; Nordhoff, E.; Mirgorodskaya, E.; Ekman, R.; Roepstorff, P. J. Mass Spectrom. 1999, 34, 105-116. (37) Mann, M.; Wilm, M. Anal. Chem. 1994, 66, 4390-4399. (38) Hoffmann, R.; Wachs, W. O.; Berger, R. G.; Kalbitzer, H.-R.; Waidelich, D.; Bayer, E.; Wagner-Redeker, W.; Zeppezauer, M. Int. J. Peptide Protein Res. 1995, 45, 26-34. (39) Lehmann, W. D., Proceedings of the 32nd Annual Meeting of the German Mass Spectrometry Society, Oldenburg, 1999; 112. (40) Roepstorff, P.; Fohlmann, J. Biomed. Mass Spectrom. 1984, 11, 601. (41) Biemann, K. Biomed. Environ. Mass Spectrom. 1988, 16, 99-111. (42) Bateman, R.; Carruthers, R.; Hoyes, J.; Gilbert, A.; Langridge, J.; Malone, K.; Bordoli, R. Proc. 46th ASMS Conf. Mass Spectrom. Allied Top., Orlando, FL, 1998; 42. (43) Borchers, C.; Parker, C.; Deterding, L.; Tomer, K. J. Chromatogr. A 1999, 854, 119-130.

Table 1. List of Exact Masses and Amino Acid Compositions of All Possible a-, b- and y-Type Fragment Ions of Unmodified Peptides Giving Rise to Signals at a Nominal Mass of 216 Da upon Low-Energy CIDa

a

fragment ion type

amino acid composition

m/z (Da)

Im(pY) b2 b2 b3 b3 a2 a2 a3 a3 b2 (13C) a2 a2 (13C) a3 (13C) a3 (13C) a4 (13C) a2 b2 a3 a3 a2 (13C) b2 (13C) y2 (13C) a2 a2 (13C) a3( 13C)

NT QS AGS GGT EN DQ ADG EGG DV GW NQ AGN GGQ AGGG DK KS AAT GSV EL LT PV RS NK GGK

216.043 216.098 216.098 216.098 216.098 216.098 216.098 216.098 216.098 216.106 216.113 216.118 216.118 216.118 216.118 216.135 216.135 216.135 216.135 216.143 216.143 216.143 216.146 216.154 216.154

Loss of ammonia or water is not considered

order to be compatible with a TOF analyzer. Conversion is achieved by orthogonally injecting “slices” of the ion beam into the TOFMS. The limitation of this configuration for precursor ion scanning is that only a small fraction of the generated ions are actually transmitted into the TOF part. The duty cycle (i.e., the fraction of ions injected into the TOF part out of all ions leaving the quadrupole section) is mass-dependent and decreases with decreasing m/z, so that only about 5% of the smaller ions reach the detector.44 Typically, small m/z species are used for precursor ion scanning; thus, the overall sensitivity of this experiment is rather low. In comparison, fragment ion transmission in precursor ion scanning experiments on triple-quadrupole instruments is close to 100%. The problem of low transmission of small m/z species has now been solved by the introduction of the Q2-pulsing function on quadrupole TOF instruments (QSTAR Pulsar, AB/MDS-Sciex). In this configuration, fragment ions are trapped in the collision cell for a user-defined period of time. Subsequently, ions are released into the quadrupole TOF interface region. This ion release is synchronized with the orthogonal injection device such that after a defined delay time, fragment ions of a user-defined m/z range are pulsed into the TOF mass analyzer. Because the continuous ion beam is now axially transformed into pulsed ion packages, transmission for the m/z region of interest can be greatly enhanced45. The gain in ion transmission is more pro(44) Chernushevich, I. V.; Ens, W.; Standing, K. G. Anal. Chem. 1999, 71, 452A461A. (45) Whitehouse, C. M.; Gulcicek, E.; Andrien, B.; Banks, F.; Mancini, R. Proc. 46th ASMS Conf. Mass Spectrom. Allied Top. Orlando, FL, 1998; 39.

Analytical Chemistry, Vol. 73, No. 7, April 1, 2001

1443

nounced for small m/z species than for larger ions, resulting in a duty cycle of >90% for ions of m/z 10 could be acquired in a reasonable period of time (15 scans, 1 min/scan, 0.5 Da step size, 50-ms dwell time), as compared to the spraying time of the nanoelectrospray experiment. Although the peak at m/z 464.5 contains only 23 events at the detector, the very good S/N in this spectrum results

Figure 4. Analytical sensitivity of the 216.04 (( 0.02 Da) precursor ion scan. 100 fmol MAPK was loaded onto a 1D SDS gel. The silver stained band was excised and in-gel digested with trypsin. The digest was desalted and concentrated on a POROS R2 column. (A) MS1 spectrum of the fraction eluted with 1 µL 25% methanol/5% formic acid. (B) Precursor-of-216.04 (( 0.02 Da) scan. In the spectrum shown, 36 scans were accumulated (50 s/scan). Apart from the triply charged doubly phosphorylated peptide at m/z 768.7, the triply charged monophosphorylated and the quadruply charged doubly phosphorylated species are also observed (both marked with a bullet).

in the high-accuracy fragment ion selection that is employed in the precursor ion scan (width ( 0.02 Da), so that the average background signal is < 1 event. This limit of detection is comparable to the value reported by Wilm et al.30 for a precursorof-(-79) scan in the negative ion mode on a triple-quadrupole mass spectrometer. In analytical practice, phosphoproteins are available in only limited quantities and are contaminated with other proteins. In contrast to “limit of detection” of a single phosphopeptide, the analytical question is how much of a phosphoprotein needs to be Analytical Chemistry, Vol. 73, No. 7, April 1, 2001

1445

Figure 5. Sequencing of the phosphotyrosine peptide derived from the MAPK digest. The triply charged, doubly phosphorylated peptide at m/z 768.7, detected by precursor ion scanning for 216.04 (( 0.02 Da, Figure 4B), was sequenced by tandem MS. Several fragment ion series corresponding to different charge states and partial loss of phosphate are observed: +, y1+ fragment ions; 9, y-pi1+ fragment ions; 2, y2+ fragment ions; b, y-pi2+ fragment ions; f, y-pi3+ fragment ions; V, bn+ and b-pin+ fragment ions. The sequence of 16 (underlined) out of 19 amino acids including the two phosphorylation sites could be derived from the MS/MS spectrum.

present in a gel band after separation by SDS-PAGE in order to obtain information on its phosphorylation site. This question was investigated by loading defined amounts of MAPK (1 pmol to 100 fmol) onto a polyacrylamide gel, followed by processing the protein for future phosphopeptide analysis. Protein bands were cut from the gel and in-gel digested with trypsin. Peptides were extracted from the gel plugs and concentrated in a speedvac. Subsequently, peptides were desalted on a POROS R2 microcolumn.36,46 Peptides were eluted in three steps (1µL of 25, 40, and 60% methanol solutions in 5% formic acid) directly into nanoelectrospray capillaries for MS analysis. Figure 4A shows the MS1 spectrum of the 25% methanol fraction of the tryptic digest of 100 fmol MAPK loaded on gel. All of the major peaks correspond to trypsin autolysis peaks. The precursor-of-216.04 experiment is shown in Figure 4B. After 36 scans (50 s/scan, 50 ms dwell time, 0.5 Da step size) a clear peak at m/z 769 was observed, which corresponds to the triply charged doubly phosphorylated peptide T398-416 from the MAPK. Two further signals (marked by bullets) in the precursor ion scan could be correlated with the triply charged singly phosphorylated peptide T398-416 and the quadruply charged doubly phosphorylated peptide T398-416. These species were also observed in the spectrum of the in-solution digest (see Figure 2C). Close inspection of the MS1 spectrum revealed a triply charged peak at m/z 768.65. Using the same sample solution, a product ion spectrum of this peptide was acquired (Figure 5). The spectrum is rather complex, but all (46) Wilm, M.; Mann, M. Anal. Chem. 1996, 68, 1-8.

1446 Analytical Chemistry, Vol. 73, No. 7, April 1, 2001

of the major peaks could be explained by standard fragmentation mechanisms, revealing both phosphorylation sites and most of the sequence of this 2.3 kDa peptide. The interpretation of the product ion spectrum is somewhat complicated by the fact that most of the peaks are accompanied by a satellite of -98 Da due to the loss of H3PO4 from the phosphothreonine residue and the presence of series of singly, doubly, and triply charged fragment ions. This experiment demonstrates that selective phosphotyrosine detection in a peptide mixture and subsequent sequencing of the tyrosine phosphorylated peptide can be performed without splitting the same sample, even in the range of 100 fmol of protein loaded on a gel. Previous state-of-the-art methods for the detection of phosphopeptides from in-gel digested proteins by the -79 precursor ion scan method was in the 250 fmol range; however, phosphorylation site determination was not possible at this level in the previous study.31 Application to Tyrosine Phosphorylation Mediated Receptor-Signaling Pathways. Regulatory proteins are typically of low natural abundance within a cell. Furthermore, in vivo signaling through reversible phosphorylation of proteins is often already activated at low phosphorylation stoichiometries. From an analytical point of view, these two factors dictate that the analytical strategy employed must work on a low to subpicomol scale. The data shown in the previous sections demonstrate that it is now possible to obtain meaningful phosphorylation data at this concentration. To test the general applicability of the new method, an immunoprecipitation experiment was performed to isolate

Figure 6. Localization of a tyrosine phosphorylation site on the EGF receptor isolated by SDS-PAGE. (A) MS 1 spectrum of the 20% methanol/ 5% formic acid elution of the POROS R2 column. (B) Precursor ion scan for 216.04 (( 0.02 Da), showing a single peak at m/z 772.5. Close inspection of the MS 1 spectrum in (A) revealed a triply charged peptide at m/z 772.68 Da (arrow). (C) Product ion spectrum of the triply charged phosphopeptide at m/z 772.68. The sequence of 16 out of 19 consecutive amino acids and the site of tyrosine phosphorylation could be unambiguously identified from the spectrum [GSHQISLDNPD(pY)QQDFFPK].

tyrosine phosphorylated proteins involved in the epidermal growth factor (EGF)-signaling pathway.47 One-half of a set of HeLa cells was induced by the addition of EGF; the second was not induced and served as a control experiment. Following lysis of the cells, the cleared lysates were incubated with anti-phosphotyrosine antibody immobilized on agarose beads. Subsequent to washing, the proteins bound to the immobilized antibody were eluted with phenyl phosphate. After acetone precipitation, the experiment and control samples were separated by SDS-PAGE and silver staining, and bands that were clearly present in the induction experiment and absent in the control lane were excised and processed for subsequent MS analysis as described under materials and methods. To reduce the complexity of the peptide mixture in the following mass spectrometric analysis, protein digests were loaded onto a “tandem column” consisting of a POROS R2 and an OLIGO R3 column in sequence according to the procedure described by Neubauer et al. The use of such an arrangement ensures that as few peptides as possible are lost during desalting, because all of the small and hydrophilic peptides that are not retained by the POROS R2 column are trapped by the OLIGO R3 material.31 Each column was then step-eluted with 20, 40, and 60% methanol containing 5% formic acid, and each fraction was subjected to nanoelectrospray analysis. Figure 6A shows the MS1 spectrum of the 20% methanol fraction of the R2 column. The precursors-of-m/z-216.04 spectrum of this fraction (Figure 6B), exhibits one clear peak at m/z 773, which indicates the presence of a tyrosine phosphorylated peptide. Close inspection of the MS1 spectrum revealed a minor triply

charged peak at m/z 772.68 (indicated by an arrow in Figure 6A). The product ion spectrum of this species is shown in Figure 6C. Except for the three most N-terminal amino acids, the complete sequence of this 2.3 kDa peptide could be deduced from the yand b-fragment ion series present in the MS/MS-spectrum. A peptide sequence tag was constructed from the spectrum, and a database query allowing for the presence of phosphorylation identified the protein as the EGF receptor (SWISS-PROT: P00533, 130 kDa). The retrieved sequence GSHQISLDNPD(pY)QQDFFPK contains a single tyrosine residue and the data shown in Figure 6C unambiguously confirmed the known phosphorylation site at Tyr1172. Many other members of the EGF signaling pathway were analyzed in the same manner, and the results of that study are reported separately.48 The m/z 216.04 (( 0.02) precursor ion scan can be acquired while other MS/MS spectra are interpreted for “on-the-fly” protein identification. Therefore, it is easily possible to identify a protein and to check for a tyrosine phosphorylation site in the same experiment. Furthermore, the sequence information obtained from a single tyrosine phosphorylated peptide might often be sufficient for the identification of the protein and localization of the modification at the same time.

(47) Pandey, A.; Podtelejnikov, A. V.; Blagoev, B.; Bustelo, X. R.; Mann, M.; Lodish, H. F. Proc. Nat’l. Acad. Sci. U.S.A. 2000, 97, 179-184.

(48) Pandey, A.; Fernandez, M. M.; Steen, H.; Blagoev, B.; Nielsen, M. M.; Roche, S.; Mann, M.; Lodish, H. F. J. Biol. Chem. 2000, 275, 38633-38639.

CONCLUSIONS AND PERSPECTIVES Whenever the exact mass of a characteristic fragment ion differs from other fragment ions of the same nominal mass such that this difference can be resolved by a quadrupole TOF mass spectrometer, highly specific precursor ion experiments can be

Analytical Chemistry, Vol. 73, No. 7, April 1, 2001

1447

performed without the risk of interference from other species. This concept was used in this study to identify phosphotyrosinecontaining peptides within complex peptide mixtures in a highly specific manner by using the exact mass of the immonium ion of phosphotyrosine (216.04 Da) as a reporter ion in a precursor ion scan. Unlike on triple-quadrupole instruments, no unspecific signals from a-, b-, or y-type fragment ions with the same nominal mass were detected. With the recently introduced Q2-pulsing function for quadrupole time-of-flight tandem mass spectrometers, detection limits comparable to triple-quadrupole mass spectrometers can be obtained. For synthetic peptides, the limit of detection is in the low fmol/µL range, whereas 100 fmol of a phosphoprotein in gel was still sufficient to identify the phosphotyrosine-containing peptide and to localize the site of modification. Furthermore, the method was successfully applied to the analysis of in vivo phosphorylation sites of the EGF receptor. One of the main advantages of using the immonium ion of phosphotyrosine to detect the tyrosine phosphorylated peptides when compared to the “traditional” m/z -79 precursor ion scan is that it is performed at low pH in the positive ion mode, so that protein identification, phosphopeptide detection, and sequencing can conveniently be done in the same experiment, thus reducing the amount of sample and time needed for a successful analysis. The method presented here is in principle also applicable to mapping phosphorylation on serine and threonine residues by scanning for the immonium ions of the respective phosphoamino acid. However, owing to the inherent lability of the alkyl phosphoesters of serine and threonine, the corresponding immonium ions are formed with a very low yield, so that the sensitivity of such experiments is low.

1448

Analytical Chemistry, Vol. 73, No. 7, April 1, 2001

The presented method has proven very successful in analytical practice. In addition, it can be envisaged that the use of precursor ion scans with highly accurate fragment ion selection will be a generally applicable tool for the analysis of modified peptides with a relatively high incidence of mass deficient atoms. Specific detection of glycosylation, iodination, bromination, and nitration without the interference of common peptide fragment ions are all within reach, provided that a sufficiently stable fragment ion can be detected. ACKNOWLEDGMENT The authors thank W. D. Lehmann and R. Pipkorn (German Cancer Research Institute, Heidelberg) for providing a synthetic tyrosine phosphorylated peptide (Figure 3), Lene Jacobsen and Janne Crawford (MDS Protana) for technical assistance, and Jørgen Petersen (Department of Biochemistry and Molecular Biology, University of Southern Denmark, Odense, Denmark) for the recombinant His-tagged RrmA. We also acknowledge Nick Morrice at the Department of Biochemistry, University of Dundee (U.K.), and all of the members of the Protein Interaction Laboratory, University of Southern Denmark, for fruitful discussions. Work in M.M.’s laboratory is funded by a generous grant from the Danish National Research Foundation to the Center for Experimental Bioinformatics.

Received for review November 10, 2000. Accepted January 22, 2001. AC001318C