Protein Tyrosine-O-Sulfation Analysis by Exhaustive Product Ion

A new strategy for the detection of tyrosine-. O-sulfated peptides in total protein digests was designed based on exhaustive product ion scanning at t...
0 downloads 0 Views 127KB Size
Download PDF
Anal. Chem. 2004, 76, 5136-5142

Protein Tyrosine-O-Sulfation Analysis by Exhaustive Product Ion Scanning with Minimum Collision Offset in a NanoESI Q-TOF Tandem Mass Spectrometer Mogjiborahman Salek,† Sabine Costagliola,‡ and Wolf D. Lehmann*,†

Central Spectroscopy Unit, German Cancer Research Center, Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany, and Free University Brussels, I.R.I.B.H.M., 808 Lennik Street, B-1070 Brussels, Belgium

Tyrosine-O-sulfated peptides were studied by nanoESI Q-TOF mass spectrometry and were found to exhibit an abundant loss of SO3 in positive ion mode under the usually nonfragmenting conditions of survey spectrum acquisition. A new strategy for the detection of tyrosineO-sulfated peptides in total protein digests was designed based on exhaustive product ion scanning at the collision offset conditions typical for the recording of survey spectra (minimum collision offset). From these data, Q-TOF neutral loss scans for loss of 80/z and Q-TOF precursor ions scans were extracted. The specificity of this approach for analysis of tyrosine-O-sulfation was tested using a tryptic digest of bovine serum albumin spiked with sulfated hirudin (1:1 and 1000:1 molar ratio of BSA to sulfated hirudin, respectively) and using an in-solution digest of the recombinant extracellular domain of thyroid stimulating hormone receptor (ECD-TSHr). For both examples, the combination of in silico neutral loss scans for 80/z and subsequent in silico precursor ion scans resulted in a specific identification of sulfated peptides. In the analysis of recombinant ECD-TSHr, a doubly sulfated peptide could be identified in this way. Surprisingly, ∼1/4 of the product ion spectra acquired from the tryptic digest of ECD-TSHr at minimum collision offset exhibited sequence-specific ions suitable for peptide identification. Complementary ion pairs were frequently observed, which either were b2/y(max-2) pairs or were induced by cleavage N-terminal to proline. MS/MS analysis at minimum collision offset followed by extraction of neutral loss and precursor ion scans is ideally suited for highly sensitive detection of analyte ions which exhibit facile gas-phase decomposition reactions. The complexity of the DNA-encoded cellular proteome is enhanced substantially by posttranslational covalent modifications, which often alter protein structures, interactions, and functions. Reversible covalent protein modification represents a general principle for the physiological and pathophysiological regulation * Corresponding author. Tel.: ++49-6221-424563. Fax: ++49-6221-424554. E-mail: [email protected]. † German Cancer Research Center. ‡ Free University Brussels.

5136 Analytical Chemistry, Vol. 76, No. 17, September 1, 2004

of essential cellular functions including the development and outcome of disease.1 Several hundred protein posttranslational modifications have been described,2 which affect the majority of the 20 standard amino acid residues. These modifications range from common types such as kinase-catalyzed phosphorylation,3 estimated to affect ∼1/3 of human proteins, to extremely rare modifications, such as iodination, which is observed in only a single human protein, thyroglobulin.4 Tyrosine-O-sulfation is a relatively common reversible covalent protein modification. It was first observed in fibrinogen5 in the 1950s and it took about three decades until it was unraveled that numerous membrane and secreted proteins are tyrosine-Osulfated.6-9 Protein sulfation takes place in the Golgi compartment during protein transport toward the plasma membrane.10 It is effected by tyrosyl protein sulfotranferases, which use 3′-phosphoadenosine-5′-phosphosulfate as sulfate donor. This protein modification is a key modulator of protein-protein interactions.11 It is required for hormone receptor recognition.12 Chemokine receptor CCR5 tyrosine sulfation facilitates HIV entry in cells,13 and recently, it has been shown that tyrosine sulfation of human antibodies contributes to recognition of the CCR5 binding region of HIV-1 gp120.14 (1) Hunter, T. Cell 2000, 100, 113-127. (2) Collections of modified amino acid residues. http://www.unimod.org; http://pir.georgetown.edu/pirwww/dbinfo/resid.html. (3) Manning, G.; Whyte D. B.; Martinez, R.; Hunter, T.; Sudarsanam, S. Science 2002, 2982, 1912-1934. (4) van de Graaf S. A. R.; Ris-Stalpers, C.; Pauws, E.; Mendive, F. M.; Targovnik, H. M.; de Vijlder, J. J. M. J. Endocrinol. 2001, 170, 307-321. (5) Bettelheim, F. R. J. Am. Chem. Soc. 1954, 76, 2838-39. (6) Huttner, W. B. Nature 1982, 299, 273-276. (7) Huttner, W. B. Methods Enzymol. 1984, 107, 200-223. (8) Huttner, W. B. Annu. Rev. Physiol. 1988, 50, 363-76. (9) Niehrs, C.; Beisswanger, R.; Huttner, W. B. Chem. Biol. Interact. 1994, 92, 257-271. (10) Bauerle, P. A.; Huttner, W. B. J. Cell. Biol. 1987, 105, 2655-2664. (11) Kehoe, J. W.; Bertozzi, C. R. Chem. Biol. 2000, 7, R57-61. (12) Costagliola, S.; Panneels, V.; Bonomi M.; Koch, J.; Many, M. C.; Smits, G.; Vassart, G. EMBO J. 2002, 21, 504-513. (13) Farzan, M.; Mirzabekov, T.; Kolchinsky, P.; Wyatt, R.; Cayabyab, M.; Gerard, N. P.; Gerard, C.; Sodroski, J.; Choe, H. Cell 1999, 96, 667-676. (14) Choe, H.; Li, W.; Wright, P. L.; Vasilieva, N.; Venturi, M.; Huang, C. C.; Grundner, C.; Dorfman, T.; Zwick, M. B.; Wang, L.; Rosenberg, E. S.; Kwong, P. D.; Burton, D. R.; Robinson, J. E.; Sodroski, J. G.; Farzan, M. Cell 2003, 114, 161-170. 10.1021/ac0400414 CCC: $27.50

© 2004 American Chemical Society Published on Web 08/05/2004

The standard method to detect protein tyrosine sulfation is radiolabeling using [35S]-sulfate followed by detection of [35S] on the protein, peptide, or amino acid.15 Despite their high sensitivity, these methods have certain limitations. First, the efficiency of in vivo protein labeling is compromised, since labeled sulfate is diluted by the presence of endogenous sulfate and since other sulfation pathways are effectively competing with protein sulfation on tyrosine, so that high initial amounts of radioactivity are required. In recent years, mass spectrometry (MS) and in particular tandem MS (MS/MS) in combination with low-energy collisioninduced dissociation has become a powerful tool for detection of covalent protein modifications. In MS/MS analysis, ring-modified tyrosine residues and phosphotyrosine residues exhibit a relatively high stability. Accordingly, the corresponding MS/MS spectra exhibit diagnostically useful low-mass ions, such as immonium ions and b2, b3, y1, or y2 ions, which still carry the modification.16-18 Unfortunately, these analytically useful features do not apply for tyrosine-O-sulfation. In positive ion mode, tyrosine-O-sulfated peptides exhibit a facile loss of SO3 by which the unmodified peptide is generated. This is the case for ionization by fast atom bombardment,19 matrix-assisted laser desorption/ionization,20-22 and electrospray ionization.23-25 In one of these ESI-MS studies,24 tyrosine-O-sulfate residues were spotted by neutral loss scanning using a combination of microelectrospray with a triple quadrupole analyzer. Since the combination of nanoelectrospray with a Q-TOF analyzer offers a higher sensitivity for protein analysis, we used this standard type of instrumentation for detection of protein tyrosine sulfation. For this purpose we describe in detail a novel strategy based on Q-TOF exhaustive product ion scanning, which takes advantage of the exceptional gas-phase instability of tyrosineO-sulfated peptides in positive ion mode. MATERIALS AND METHODS Reagents. Analytical grade solvents were purchased from E. Merck (Darmstadt, Germany) and Sigma-Aldrich (Taufkirchen, Germany). Sulfated hirudin was purchased from Sigma-Aldrich and modified trypsin, sequencing grade, was from Roche Diagnostics (Mannheim, Germany). Production of the Recombinant Extracellular Domain (ECD) of Human TSH Receptor. The ECD of the human thyroid stimulating hormone (TSH) receptor (Swissprot P16473; amino acid 22, Met, to amino acid 415, Lys) was cloned in the vector (15) Bundgaard, J. R.; Johnsen, A. H.; Rehfeld, J. F. Methods Mol. Biol. 2002, 194, 223-39. (16) Lehmann, W. D. Proceedings of the 32rd Annual Meeting of the German Mass Spectrometry Society, Oldenburg, 1999; p 112. (17) Steen, H.; Kuster, B.; Fernandez, M.; Pandey, A.; Mann M. Anal. Chem. 2001, 73, 1440-1448. (18) Salek. M.; Alonso, A.; Pipkorn, R.; Lehmann, W. D. Anal. Chem. 2003, 75, 2724-2729. (19) Gibson, B. W.; Cohen, P. Methods Enzymol. 1990, 193, 480-501. (20) Talbo, G.; Roepstorff, P. Rapid Commun. Mass Spectrom. 1993, 7, 201204. (21) Wolfender, J. L.; Chu, F.; Ball, H.; Wolfender, F.; Fainzilber, M.; Baldwin, M. A.; Burlingame, A. L. J. Mass Spectrom. 1999, 34, 447-454. (22) Seibert, C.; Cadene, M.; Sanfiz, A.; Chait, B. T. Sakmar, T. P. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 11031-11036. (23) Jedrzejewski, P. T.; Lehmann, W. D. Anal. Chem. 1997, 69, 294-301. (24) Severs, J. C.; Carnine, M.; Eguizabal, H.; Mock, K. K. Rapid Commun. Mass Spectrom. 1999, 13, 1016-1023. (25) Nemeth-Cawley, J. F.; Karnik, S.; Rouse, J. C. J. Mass Spectrom. 2001, 36, 1301-1311.

pSGHV0.26 The final construct was then subcloned in the pEFIN vector27 between the Xho1 and Xba1 restriction sites and sequenced. K562 cells were maintained at 37 °C and 5% CO2 in DMEM containing 10% FCS. The clone was adapted to a serumfree medium (ISF-1, In Vivo Biotech, Hennigsdorf, Germany). The secreted protein was then produced in a miniPERM bioreactor (Catalog No. IV 76001052, Vivascience, Hannover, Germany). Conditioned medium was centrifuged, filtered, concentrated on a Vivaspin 15 concentrator (Catalog No. 02VSRH21, Vivascience), and dialyzed overnight against 50 mM Tris-HCl, pH 7.5, 0.5 M NaCl, 10 mM imidazole. A 1-mL aliquot of nickelnitrilotriacetic acid resin (Qiagen, Valencia, CA) was placed in a column and equilibrated in the same buffer. The dialyzed medium was loaded at 0.25 mL/min at 4 °C onto the resin and then washed with 10 bed volumes of 10 mM Tris-HCl pH 8, 0.5 M NaCl, 20 mM imidazole. The bound fusion protein was eluted with 10 mM Tris-HCl, pH 8, 50 mM NaCl, 70 mM imidazole. Fusion protein-containing fractions eluted from the Ni-NTA column were pooled and concentrated. After addition of EDTA (500 µM) and DTT (1 mM), the protein was incubated with TEV protease (Invitrogen, Merelbeke, Belgium) (100 units/mg of protein) overnight at 30 °C. The TEV protease and the uncleaved fusion protein were removed after loading on the Ni-NTA resin. Sample Preparation for NanoESI-MS. In-solution digestion was performed with 1 µg of total recombinant extracellular domain of thyroid stimulating hormone receptor (ECD-TSHr) partially deglycosylated. The protein was reduced and alkylated by 10 mM DTT at 56 °C for 45 min and with 50 mM iodoacetamide at room temperature in the dark for 30 min in 100 mM NH4HCO3, pH 8. Then in-solution digestion was performed with 0.5 µg of trypsin in 100 mM NH4HCO3, pH 8, at 37 °C overnight using the same buffer. The protein was concentrated with pipet tips packed with reversed-phase C18 material (ZipTip, Millipore, Bedford, MA), eluted with 50% acetonitrile and 2% formic acid. NanoESI Mass Spectrometry. Mass spectra were recorded using a hybrid Q-TOF mass spectrometer type Q-TOF 2 (Micromass, Manchester, U.K.). Spray capillaries were manufactured inhouse using a micropipet puller type P-87 (Sutter Instruments, Novato, CA) and coated with a semitransparent film of gold in a sputter unit type SCD 005 (BAL-TEC AG, Balzers, Liechtenstein). Data were acquired by exhaustive product ion scanning as follows. The Q-TOF instrument continuously performed product ion scans with the following settings: range m/z 400-1300, step width 3 Da, and data acquisition time of 20 s/step. The collision offset values were 10 V for m/z 400-1000 and 12 V for m/z 1000-1300. From this set of exhaustive product ion spectra, Q-TOF neutral loss scans and Q-TOF product ion scans were generated by data postprocessing. Neutral loss scans were generated with a wide mass window (2 Da width) and precursor ion scans with a narrow mass window (0.1 Da width). The standard setting of the collision offset for detection of survey specta (single-stage MS) is 10 V. (26) Leahy, D. J.; Dann, C. E., III; Longo, P.; Perman, B.; Ramyar, K. X. Protein Exp. Purif. 2000, 20, 500-506. (27) Costagliola, S.; Morgenthaler, N. G.; Hoermann, R.; Badenhoop, K.; Struck, J.; Freitag, D.; Poertl, S.; Weglohner, W.; Hollidt, J. M.; Quadbeck, B.; Dumont, J. E.; Schumm-Draeger, P. M.; Bergmann, A.; Mann, K.; Vassart, G.; Usadel, K. H. J. Clin. Endocrinol. Metab. 1999, 84, 90-97.

Analytical Chemistry, Vol. 76, No. 17, September 1, 2004

5137

RESULTS AND DISCUSSION Detection Strategy for Sulfated Peptides. Whereas both tyrosine-O-phosphorylated (pTyr) and tyrosine-O-sulfated (sTyr) peptides form stable molecular ions in negative ion mode, a different situation is observed in the positive ion mode. Here only pTyr-peptides ion are stable whereas sTyr-peptides decompose by loss of SO3. A low molecular ion stability such as observed for sTyr-peptides in the positive ion mode is normally considered as a disadvantage for mass spectrometric detection. However, since spontaneous fragmentation in ESI-MS is rare for peptides and covalently modified peptides in general, we designed a new analytical strategy for specific detection of sTyr-peptides comprising the following steps: exhaustive Q-TOF product ion scanning at the conditions of survey spectrum acquisition (minimum collision offset) followed by three levels of data evaluation, (i) neutral loss scanning for 80/z, if a peak is detected, a check for identical charge state between the molecular ion (sTyr-peptide candidate) and its fragment ion (unmodified analogue sTyr-peptide candidate) is performed; (ii) from a pair of peaks that have passed this test, the lower m/z value is selected to generate a precursor ion scan. Signals found in both this precursor ion scan and in the neutral loss scan are identified as peptides carrying tyrosine-Osulfate. (iii) Product ion spectra are recorded for the molecular ions of the sulfated and the nonsulfated molecule for peptide identification via sequence-specific fragment ions. In the following, this strategy is tested by analyzing model peptide mixtures and a digest of ECD-TSHt, a protein, in which the occurrence of sTyr residues had been proposed based on biochemical evidence.12 Selective Detection of Tyrosine-O-Sulfation in Model Mixtures. For a first test of the analytical method outlined above, we spiked a total tryptic digest of bovine serum albumin with an equimolar amount of sulfated hirudin and performed exhaustive nanoESI product ion scanning over the range from m/z 400 to 1000. The evaluation of this set of MS/MS spectra is displayed in Figure 1. In the survey spectrum of the spiked digest, the molecular ion of sulfated hirudin is observed only with minor relative abundance (spectrum not shown). In the neutral loss scan for 80/2 ) 40 performed with the data set obtained by exhaustive product ion scanning, two peaks are observed, as shown in Figure 1a. These two peaks are caused by two doubly charged ion pairs (m/z 706.3, 746.3 and 717.3, 757.3). The results of the subsequent precursor ion scans for m/z 706.3 and 717.3 are given in Figure 1. To further evaluate the specificity of our analytical strategy, we spiked a digest mixture of 1 nmol of trypsin and 1 nmol of chymotrypsin with 1 pmol of sulfated hirudin and analyzed this sample in the same way as described in Figure 1. The results given in Figure 2 demonstrate that detection of a sulfated peptide is possible even in the presence of a roughly 1000-fold molar excess of other peptides. However, as shown in Figure 2a, a second signal is observed in the neutral loss scan for m/z 40 (m/z 771.4, marked by an asterisk). Inspection of the corresponding product ion spectrum at m/z 771 revealed the presence of a singly charged ion at m/z 729.4, which is thus not caused by an sTyr-peptide. This ion signal was assigned as the y6 fragment of the chymotrypsin peptide IVNGEEAV-PGSWPW generated by cleavage of the V-P bond. 5138 Analytical Chemistry, Vol. 76, No. 17, September 1, 2004

Figure 1. NanoESI Q-TOF mass spectrometric analysis of a tryptic digest of BSA (1 pmol) spiked with an equimolar amount of sulfated hirudin. (a) Neutral loss scan for m/z 40; (b) precursor ion scan for m/z 706.3; (c) precursor ion trace for m/z 717.31. The neutral loss scan identifies only sTyr-peptide signals. Precursor ion scanning for unmodified hirudin shows a signal for this analyte and for its sulfated analogue (see text).

This finding is consistent with a report that the V-P peptide bond is particularly prone to fragmentation. 28 Proline-directed fragment ions were frequently observed in the Q-TOF product ion spectra recorded at minimum collision offset (see below, Table 1) and are highly useful for peptide identification. Their misinterpretation in terms of indicating a certain neutral loss from the molecular ion is highly improbable, since an accidental fit with respect to mass difference to the molecular ion m/z value and to its charge state would have to occur. (28) Breci, L. A.; Tabb, D. L.; Yates, J. R.; Wysocki, V. H. Anal. Chem. 2003, 75, 1963-1971.

Table 1. Peptides Identified from the NanoESI MS/MS Data Acquired with in Extenso Product Ion Scanning at Minimal Collision Offset of the Tryptic Digest of ECD-TSHra

a

sequence

MS/MS sequence ions

MFPDLTK LI-ETHLR LD-AVYLNK AD-LSY-PSHCac I-PSL-PPSTQTLK DAFGGVYSGPSLLDVSQTSVTAL-PSK

a1, a2, b2; y5, y6; neutral loss of MF a1, a2, b2, b3; y4, y5; neutral loss of LI a1, a2, b2, b3, b4, b5; y1, y2, y3, y4, y5, y6, y7; neutral loss of LD b2, a3, b3, a4, b4, a5, b5; y1, y4, y5, y6, y7; neutral loss of AD b2, b3, b4; y8; neutral loss of I and IPSL b22 2+, b23 2+; y3, y4;

Complementary ion pairs generated by backbone cleavage N-terminal to P or b2/y(max-2) ion pairs are given in boldface type (ac, acetamido).

Figure 2. NanoESI mass spectrometric analysis of a mixture of trypsin (1 nmol) and chymotrypsin (1 nmol) spiked with 1 pmol of sulfated hirudin. (a) Neutral loss scan for m/z 40; (b) precursor ion scan for m/z 706.3. The signal at m/z 746.3 is found in both spectra and thus is identified as [M + 2H]2+ ion of sulfohirudin. The peak in (a) marked by an asterisk is caused by the chymotryptic peptide IVNGEEAVPGSWPW (see text).

Selective Detection of Tyrosine-O-Sulfation in Recombinant ECD-TSHr. In the next step, the strategy outlined above was applied to a sample of the recombinant extracellular domain of the thyroid stimulating hormone receptor ECD-TSHr. Figure 3 summarizes the results obtained for the different scan modes. In the neutral loss scan for 80/3 in Figure 3b, a set of three signals at m/z 1184.5, 1211.5, and 1238.5 is observed. The corresponding precursor ion scan for the signal with the lowest m/z value at 1184.5 given in Figure 3c exhibits signals at m/z 1211.5 and 1238.5, which are thus identified as sulfated peptides. The major ions detected in the MS/MS spectra of all three ions (m/z 1184.5, 1211.5, and 1239.5) were essentially identical and could be assigned to originate from the tryptic T36 peptide of ECD-TSHr with the sequence NPQEETLQAFDSH-Y385-DY387-TI-Cac-GDSEDMV-Cac-TPK. The MS/MS spectrum of

Figure 3. Spotting of tyrosine-O-sulfated peptides in a total in-solution tryptic digest of ECD-TSHr by nanoESI Q-TOF neutral loss scan at 12 V offset. (a) Survey mass spectrum; (b) neutral loss scan for m/z 80/3; (c) precursor ion scan for m/z 1184.5. The ion signals at m/z 1211.5 and 1238.5 are observed in both (b) and (c), so that they are identified as singly and doubly sulfated analogues of T36 (charge state 3+).

the nonsulfated peptide, generated by MS/MS of m/z 1184.5 is shown in Figure 4. Analytical Chemistry, Vol. 76, No. 17, September 1, 2004

5139

Figure 4. Product ion spectrum of the nonsulfated peptide at m/z 1184.5 (collision offset 45 V) which allows its identification as the T36 peptide of ECD-TSHr (NPQEETLQAFDSH-Y385-D-Y387-TI-Cac-GDSEDMV-Cac-TPK). The presence of carbohydrate-specific ions (indicated by asterisks) in some product ion spectra is possibly caused by glycosylation of ECD-TSHr.

Based on the analysis steps described above, the ion signals observed at m/z 1211.5 and 1239.5 were assigned as the monoand disulfated analogues of the T36 peptide. Since loss of SO3 forms the first and very effective step in the fragmentation of protonated sTyr-peptides, the corresponding MS/MS spectra do not indicate the distribution of the sulfate groups between the two sulfation sites, Y-385 and Y-389. Moreover, no information about the degree of sulfation can be extracted from the positive ion ESI mass spectra. Negative Ion NanoESI Analysis. In the negative ion mode, molecular ions of sTyr-peptides show the same characteristic loss of SO3 but exhibit a greater stability. Additional protons probably catalyze the fragmentation of sTyr residues. This explanation is confirmed by the fact that [Na]+ adducts or interaction of the negatively charged sulfate ester group with organic cations stabilizes the sTyr residue.29 Therefore, negative ion ESI MS probably provides a more realistic insight into the degree of sulfation than analyses in the positive ion mode. Figure 5 gives a comparison of the molecular ion region of the T36 peptide recorded in positive and negative ion modes, respectively. As expected, strongly different ion abundances are observed. The data in Figure 5 show that the major species of T36 in ECD-TSHr is represented by the disulfated peptide (roughly 2/3) accompanied by the unmodified peptide (roughly 1/3). The monosulfated peptide appears to be of minor abundance. These data may be somewhat biased by a preferential ionization of sulfated peptides versus their nonsulfated analogues, but we assume that they provide a broadly correct insight into the true molar abundances. Sensitivity of MS/MS Scans at Minimum Collision Offset. As shown above in Figure 2, Q-TOF neutral loss and precursor ion scanning at minimum collision offset provides a high sensitivity for the detection of tyrosine-O-sulfation. Figure 6 demonstrates this effect by presenting a set of original data, namely, two original (29) Yagami, T.; Kitagawa, K.; Aida, C.; Fujiwara, H.; Futaki, S. J. Pept. Res. 2000, 56, 239-249.

5140 Analytical Chemistry, Vol. 76, No. 17, September 1, 2004

Figure 5. Partial nanoESI survey spectra of a tryptic digest of ECDTSHr showing the molecular ion region of the T36 peptide: (a) positive ion mode; (b) negative ion mode. In the negative ion mode, the molecular ion of doubly sulfated T36 is the most abundant ion, whereas in the positive ion mode, the signal of nonsulfated T36 prevails (see text).

product ion scans of the ECD-TSHr analysis recorded at high (Figure 6b) and minimum (Figure 6c) collision offset, respectively. For comparison, the corresponding ESI MS survey spectrum is also given (Figure 6a), and all spectra presented were recorded over the same time interval and are displayed with identical intensity scale. As expected, the ion abundances in the survey MS spectrum are significantly higher than in the MS/MS spectra. The MS/ MS spectrum at 45 V offset of m/z 1184, representing the triply charged molecular ion of the nonsulfated T36 peptide in Figure 6b, shows ion signals of very low abundance. In contrast, the MS/ MS spectrum at 12 V offset of m/z 1238, representing the triply charged molecular ion of doubly sulfated T36, clearly shows two fragment ions, generated by the loss of one or two molecules of SO3. We see two major reasons for the high sensitivity, which is achieved for the detection of sTyr residues at minimum collision offset: (i) loss of SO3 is practically the only and very effective fragmentation process for the disulfated T36 peptide under these conditions, so that all products of fragmentation are concentrated in two ions; (ii) the fragment ions formed at minimum collision offset are probably transmitted to the detector with higher efficiency than those formed by more energetic collisions.

Figure 6. Visualization of the high sensitivity achieved by product ion scanning at minimum collision offset: (a) survey positive ion nanoESI MS of the tryptic digest of ECD-TSHr; (b) product ion scan of m/z 1184 at 45 V offset; (c) product ion scan of m/z 1238 at 12 V offset. All three spectra were acquired for the same time period and are displayed with the same intensity scale. The product ion scan at 12 V is highly specific and selective for detection of sTyr (see text).

Peptide/Protein Identification from MS/MS Spectra with Minimum Collision Offset. Inspection of the set of MS/MS spectra acquired for the detection of sTyr-peptides in the tryptic digest of ECD-TSHr showed that an unexpectedly large number of these MS/MS spectra contain sequence-specific fragment ions (about 1/4 of the 290 MS/MS spectra). Six MS/MS spectra could be assigned to peptides of the ECD-TSHr protein, which resulted in a sequence coverage of ∼16%. Table 1 summarizes these assignments and the sequence-specific fragment ions, which could be extracted. As can be inferred from Table 1, peptide backbone fragmentation at minimum collision offset is particularly abundant at the N-terminal site of proline and at the second peptide bond from the N-terminus. Both types of fragmentation often result in the formation of complementary b/y pairs of fragment ions. These complementary ion pairs are of high significance, since they can be used for an independent and accurate determination of the molecular weight of the fragmented peptide. In the Q-TOF MS/ MS analysis of peptides of low abundance, this is of particular value in those cases where the molecular weight of the fragmented peptide cannot be reliably determined from the survey spectrum. For all doubly charged peptides listed in Table 1, neutral loss of amino acid residues (one, two, or four) was observed, thus providing valuable N-terminal sequence information as described recently.30 Inspection of the survey spectrum of the ECD-TSHr tryptic digest showed that the more abundant fragment ions were already present there. This is reasonable, since Q-TOF instruments are normally operated with the collision gas switched on, so that the fragment ions generated without additional offset voltage will appear in the survey spectrum and thus mimic a proteolytic peptide molecular ion. Selection of such a pseudomolecular ion as precursor ion in subsequent MS/MS analysis has different consequences depending whether the selected ion is a b- or y-type ion. If a b ion is selected, the MS/MS spectrum (30) Salek, M.; Lehmann, W. D. J. Mass Spectrom. 2003, 38, 1143-1149.

will only contain b ions. A subsequent protein database search will not generate a correct hit, since search engines usually rely on y ions and since the m/z value of the b ion will not fit to a peptide molecular weight in the database. On the other hand, if a y ion is selected, the MS/MS spectrum in principle could lead to a correct peptide identification; however, in a database search including a proteolytic specificity, a positive hit is highly improbable. Roughly 10 MS/MS spectra in the set of spectra recorded in the analysis of the ECD-TSHr digest at minimum collision offset showed abundant series of fragment ions with a distance of 162 or 162/2. They can be easily recognized by extraction of neutral loss scans for m/z 162 or 81, respectively. These singly or doubly charged fragment ion series consistently originate from precursor ions of the composition [C6H10O5]n (n ranging from 2 to ∼15). The sources of these oligosaccharide ions are probably modified peptide ions present in the digest of the highly glycosylated protein ECD-TSHr. We assume that these glycopeptides partially decompose under the conditions of minimum collision offset and thus give rise to the observed series of oligosaccharide fragment ions. We observed these ions also in digests of other glycosylated proteins (for instance, EGF receptor). They are ideally suited for internal recalibration, since their m/z values are distributed evenly over a wide range of the mass spectrum and since their accurate mass values show a mass deficiency relative to peptide ions,31 facilitating their assignment. However, these carbohydrate signals increase the background level considerably and may interfere with a straightforward interpretation of peptide MS/MS data. Thus, complete protein deglycosylation may further improve the sensitivity of protein tyrosine sulfation analysis by mass spectrometry. In conclusion, we have described exhaustive Q-TOF product ion scanning with minimal collision offset as a sensitive and specific method for detection of tyrosine-O-sulfation in protein (31) Lehmann, W. D.; Bohne, A.; von der Lieth, W. J. Mass Spectrom. 2000, 35, 1335-1341.

Analytical Chemistry, Vol. 76, No. 17, September 1, 2004

5141

digests. The unique instability of tyrosine-O-sulfate in the positive ion mode provides the basis for the high specificity and sensitivity achieved. This approach appears to be promising also for the analysis of other organic sulfate esters, such as hormone or drug conjugates32 as far as they can be detected in positive ion mode. Surprisingly, ∼1/4 of the product ion spectra recorded from a total protein digest at minimum collision offset contained sequencespecific ion series useful for protein identification. With this information, a number of ions in the survey spectrum could be assigned as fragment ions instead of peptide molecular ions. MS/ MS analysis at minimum collision offset generates product ion

spectra with extremely low background. Therefore, exhaustive product ion scanning is ideally suited for analytes exhibiting a facile gas-phase fragmentation reaction.

(32) Strott, C. A. Endocrinol. Rev. 2002, 23, 703-732.

AC0400414

5142

Analytical Chemistry, Vol. 76, No. 17, September 1, 2004

ACKNOWLEDGMENT We are indebted to V. Panneels, University of Heidelberg, for valuable support and we thank Daniel J. Leahy, Howard Hughes Medical Institute, Baltimore, for supplying the vector pSGHV0. Received for review March 2, 2004. Accepted June 11, 2004.