Enhanced Ionization of Phosphorylated Peptides during MALDI TOF

Anal. Chem. 2004, 76, 1532-1536

Enhanced Ionization of Phosphorylated Peptides during MALDI TOF Mass Spectrometry Xiaofeng Yang,† Huaping Wu,† Tomoyoshi Kobayashi,‡ R. John Solaro,‡ and Richard B. van Breemen*,†

Department of Medicinal Chemistry and Pharmacognosy, College of Pharmacy, and Department of Physiology and Biophysics, College of Medicine, University of Illinois at Chicago, Chicago, Illinois 60612

Although r-cyano-4-hydroxycinnamic acid functions as an excellent matrix for the analysis of most peptides using matrix-assisted laser desorption/ionization time-of-flight (MALDI TOF) mass spectrometry, the ionization of phosphorylated peptides is usually suppressed by nonphosphorylated peptides. As an alternative matrix, 2′,4′,6′trihydroxyacetophenone (THAP) with diammonium citrate was found to overcome this problem for the MALDI TOF mass spectrometric analysis of proteolytic digests of phosphorylated proteins. Specifically, the abundances of phosphorylated peptides in tryptic digests of bovine β-casein and protein kinase C (PKC)-treated mouse cardiac troponin I were enhanced more than 10-fold using THAP during positive ion MALDI TOF mass spectrometry. The protonated molecules of phosphorylated peptides were sufficiently abundant that postsource decay TOF mass spectrometry was used to confirm the number of phosphate groups in each peptide. Finally, tryptic digestion followed by analysis using MALDI TOF mass spectrometry with THAP as the matrix facilitated the identification of a unique phosphorylation site in PKC-treated troponin I. Electrospray and matrix-assisted laser desorption/ionization (MALDI) time-of-flight (TOF) mass spectrometry are standard analytical tools for protein identification, peptide sequencing, and protein and peptide characterization. Both electrospray and MALDI TOF mass spectrometry are used frequently for the rapid identification of proteins in proteolytic digests of bands and spots from one- or two-dimensional gels1 and for locating sites of posttranslational modifications such as glycosylation or phosphorylation within peptides.2 However during positive ion electrospray and MALDI mass spectrometric analysis, signals for phosphorylated peptides are suppressed in protein digests containing abundant unphosphorylated peptides. * Corresponding author. Tel: (312) 996-9353. Fax: (312) 996-7107. E-mail: [email protected]. † Department of Medicinal Chemistry and Pharmacognosy. ‡ Department of Physiology and Biophysics. (1) Lim, H.; Eng, J.; Yates, J. R.; Tollaksen, S. L.; Giometti, C. S.; Holden, J. F.; Adams, M. W. W.; Reich, C. I.; Olsen, G. J.; Hays, L. G. J. Am. Soc. Mass Spectrom. 2003, 14, 957-70. (2) Raska, C. S.; Parker, C. E.; Dominski, Z.; Marzluff, W. F.; Glish, G. L.; Pope, R. M.; Borchers, C. H. Anal. Chem. 2002, 74, 3429-33.

1532 Analytical Chemistry, Vol. 76, No. 5, March 1, 2004

Protein phosphorylation is intimately involved in the regulation of myocardial contraction and metabolism.3 Phosphorylation of either cardiac troponin I (cTnI) or cardiac troponin T (cTnT) reduces the maximum actomyosin ATPase rate in reconstituted systems and the maximum tension in skinned fiber bundles.4 The determination of phosphorylation sites in these phosphorylated proteins is an important step in the understanding of the molecular basis of these processes. For example, the phosphorylation of serines 22 and 23 of cTnI by cAMP-dependent protein kinase during β-adrenergic signaling5 decreases the affinity of cTnI for cardiac troponin C (cTnC) and decreases myofibrillar Ca2+ sensitivity.6,7 Consequently, the rate of Ca2+ dissociating from cTnC is increased,8 resulting in decreased pCa50 and decreased ATPase activity and force development.9 One solution to this problem is to carry out HPLC separation of the phosphorylated peptides from the other peptides prior to mass spectrometric analysis.10 However, chromatography introduces additional sample handling and preparation time prior to MALDI or electrospray mass spectrometry. Although nanoelectrospray may be used to enhance the abundance of phosphorylated peptide ions in a protein digest without chromatography,11 nanoelectrospray is available in fewer laboratories than MALDI and is considerably more difficult and expensive to operate. Another solution to the suppression of phosphorylated peptides is to switch to negative ion analysis during electrospray or MALDI. However, during negative ion electrospray or MALDI, the signals of the unphosphorylated peptides are often suppressed, which would necessitate a second analysis in positive ion mode. Not only does this take additional time, but the small sample size of bands isolated from gels might not permit such additional analysis. Since peptide mapping to identify sites of phosphorylation requires detecting as many peptides as possible, an ideal solution to this problem would be to overcome the ion suppression problem during a single analysis. (3) Raju, R. V. S.; Kakkar, R.; Radhi, J. M.; Sharma, R. K. Mol. Cell. Biochem. 1997, 176, 135-43. (4) Solaro, J. R.; Burkart, E. M. J. Mol. Cell. Cardiol. 2002, 34, 689-93. (5) Solaro, J. R.; Moir, A. J. G.; Perry, S. V. Nature 1976, 262, 615-7. (6) Liao, R.; Wang, C.-K.; Cheung, H. C. Biochemistry 1994, 33, 12729-34. (7) Reiffert, S.; Jaquet, K.; Heilmeyer, L. M. G.; Herberg, F. W. Biochemistry 1998, 37, 13516-25. (8) Kentish, C. J.; McCloskey, T. D.; Layland, J.; Palmer, S.; Leiden, M. J.; Martin, F. A.; Solaro, J. R. Circ. Res. 2001, 88, 1059-65. (9) Ward, D. G.; Ashton, P. R.; Trayer, H. R.; Trayer, I. P. Eur. J. Biochem. 2001, 268, 179-85. (10) Annan, R. S.; Carr, S. A. Anal. Chem. 1996, 68, 3413-21. (11) Wilm, M.; Neubauer, G.; Mann, M. Anal. Chem. 1996, 68, 527-33. 10.1021/ac035203v CCC: $27.50

© 2004 American Chemical Society Published on Web 01/31/2004

Here we report a solution to the problem of the suppression of phosphorylated peptides during positive ion MALDI TOF mass spectrometry by using the matrix 2′,4′,6′-trihydroxyacetophenone (THAP) with diammonium citrate. No HPLC separation is required, and both phosphorylated and unphosphorylated peptides can be measured in complex proteolytic digests. The utility of this approach is demonstrated using bovine β-casein and then is applied to identify a unique site of phosphorylation in cTnI that had been treated with a protein kinase C (PKC). EXPERIMENTAL SECTION Materials. The matrixes R-cyano-4-hydroxycinnamic acid (CHCA), 2,5-dihydroxybenzoic acid, sinapinic acid, and THAP were purchased from Aldrich (Milwaukee, WI). Sequencing grade modified trypsin was obtained from Promega Co. (Madison, WI), and endoprotease Lys-C was purchased from Wako Chemicals (Dallas, TX). Ammonium bicarbonate, trifluoroacetic acid (TFA), ammonium citrate dibasic, β-casein, and all other biochemicals were purchased from Sigma Chemical (St. Louis, MO). ZipTip C18 (spherical silica with a particle diameter of 15 µm and pore size of 200 Å) pipet tips from Millipore (Bedford, MA) were used to desalt samples prior to analysis. HPLC grade acetonitrile (Optima) was purchased from Fisher Scientific (Hanover Park, IL). Recombinant mouse cTnI, cTnT, and cTnC was purified from Escherichia coli and stored at -80 °C. Phosphorylation of Troponin I. Phosphorylation of cTnI by PKC was carried out in vitro. Briefly, the incubation contained 200 µL of 20 mM HEPES buffer (pH 7.5), 10 mM MgCl2, 100 mM CaCl2, 500 mM EGTA, 20 µg of phosphatidylserine, 4 µg of diolein, 100 mM NaCl, 0.4 mM ATP, 30 mM 2-mercaptoethanol, 5 mM troponin complex (1:1:1 of cTnI/cTnT/cTnC), and 10 µg of β- or -PKC. The reaction was initiated by the addition of the ATP and was carried out for 60 min at 30 °C. After phosphorylation, the mixture of troponins was separated using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and stained with Coomassie Brilliant Blue using standard procedures. Digestion of β-Casein and Phosphorylated Troponin I. β-Casein, 1 µM in 50 mM ammonium bicarbonate buffer at pH 8.3, was incubated overnight at 37 °C with trypsin at an enzyme/ substrate ratio of 1:100 to produce a proteolytic digest for MALDI TOF mass spectrometric analysis. The cTnI band was excised from the polyacrylamide gel, washed with 50% acetonitrile, and then dehydrated using 100% acetonitrile. The gel pieces were digested overnight at 37 °C in 10 µg/mL Lys-C solution. The peptides in the hydrolysate were then extracted with 50 µL of acetonitrile/water/TFA (47.5:47.5:5, v/v/v). To identify a unique phosphorylation site in cTnI, it was necessary to digest the Lys-C proteolytic digest of cTnI again using trypsin as described above for β-casein. Desalting of the Peptide Mixtures. Prior to analysis using MALDI TOF mass spectrometry, a pipet tip containing a C18 solidphase extraction sorbant (Ziptip) was used to remove salts and other impurities from the peptides. First, the peptide solutions were acidified by adding 5% aqueous TFA as an ion-pairing agent to enhance the retention of the peptides on the stationary phase of the C18 sorbant. Prior to sample application, the ZipTip was washed twice with 10 µL of acetonitrile/water (1:1, v/v) and then equilibrated with aqueous 0.1% TFA. Then, the acidified peptide solution was loaded onto the tip by aspirating and dispensing the

sample 10 times. The adsorbed peptides were washed three times with 10-µL aliquots of aqueous 0.1% TFA, and the peptides were eluted using 3 µL of 0.1% TFA in acetonitrile/water (1:1, v/v). Matrix Preparation. The CHCA matrix was prepared by dissolving 10 mg of R-cyano-4-hydroxycinnamic acid in 1 mL of water/acetonitrile (50:50, v/v) containing 0.1% TFA. A modified CHCA matrix was prepared from the original CHCA solution by the addition of 50 mg/mL diammonium citrate (aq.) at a ratio of 9:1 (CHCA solution/diammonium citrate solution, v/v). A 10 mg/ mL solution of the THAP matrix was prepared by dissolving 2′,4′,6′-trihydroxyacetophenone monohydrate in water/acetonitrile (50:50, v/v) followed by the addition of 50 mg/mL diammonium citrate (aq) at a ratio of 9:1 (v/v). These matrix solutions were used for MALDI TOF mass spectrometric analysis. MALDI TOF Mass Spectrometry. In preparation for analysis using MALDI TOF mass spectrometry, a 0.5-µL aliquot of a matrix solution (described above) was mixed with 0.5 µL of peptide solution. Then, 0.5 µL of the matrix/peptide solution was spotted onto the MALDI sample stage and air-dried immediately prior to analysis. Positive or negative ion MALDI TOF mass spectra were acquired using an Applied Biosystems (Foster City, CA) Voyager DE-Pro mass spectrometer operated in reflectron mode. After time-delayed extraction, the ions were accelerated to 20 kV for TOF mass spectrometric analysis. A total of 200 laser shots were acquired and signal averaged per mass spectrum. After peak detection, the peptide mapping data were compared to the NCBI database using MS-Fit.12 The potential phosphorylated peptides were selected as precursor ions for additional analysis using positive ion MALDI TOF mass spectrometry with postsource decay (PSD). Precursor ions for PSD analysis were selected using an ion gate at a resolving power of ∼100. Each PSD mass spectrum was recorded in 12 segments under computer control. A total of 200 laser shots were accumulated per segment, and all segments were stitched together to form one MALDI TOF PSD mass spectrum. RESULTS AND DISCUSSION β-Casein. A tryptic digest of β-casein was used for the selection and optimization of matrixes for the characterization of phosphorylated peptides in peptide mixtures using MALDI TOF mass spectrometry. Since phosphorylated peptides are negatively charged in solution, mass spectrometric analyses of these molecules are usually carried out in negative ion mode. For example, negative ion MALDI TOF mass spectra of the tryptic digest of β-casein were obtained using either CHCA plus diammonium citrate or THAP with diammonium citrate (Figure 1). An abundant deprotonated molecule of m/z 2059.8 corresponding to the amino acid sequence FQpSEEQQQTEDELQDK was detected using either matrix. However, the deprotonated molecule of m/z 3119.3 corresponding to the tetraphosphorylated peptide with the sequence RELEELNVPGEIVEpSLpSpSpSEESITR was abundant only when using THAP with diammonium citrate (Figure 1B) but not CHCA with diammonium citrate (Figure 1A). Although phosphorylated peptides of β-casein produced negative ions with higher relative abundance than positive ions when using matrixes such as CHCA, the protein sequence coverage, which is critical for the reliable identification of an unknown protein or for the confident identification of sites of modification within a protein, was lower in negative ion mode. This occurred because nonphosAnalytical Chemistry, Vol. 76, No. 5, March 1, 2004

1533

Figure 1. Negative ion MALDI TOF mass spectra of a tryptic digest of β-casein using the matrixes (A) CHCA with diammonium citrate and (B) THAP with diammonium citrate. Note that the monophosphorylated peptide of m/z 2060 (/) was highly abundant in both matrixes; however, the tetraphosphorylated peptide of m/z 3120 (////) was much more abundant when THAP was used as the matrix compared with CHCA with ammonium salts as the matrix.

Figure 2. Positive ion MALDI TOF mass spectra of a tryptic digest of β-casein using the matrixes (A) CHCA, (B) CHCA with diammonium citrate, and (C) THAP with diammonium citrate. Note that the monophosphorylated peptide of m/z 2062 (/) and the tetraphosphorylated peptide of m/z 3122 (////) were much more abundant when THAP with diammonium citrate was used as the matrix compared with commonly used matrix CHCA and the modified matrix CHCA containing diammonium citrate. The other peptides detected in these mass spectra were not phosphorylated. The number of phosphate groups on each phosphorylated peptide was determined using PSD during positive ion MALDI TOF mass spectrometry (see Figure 3).

phorylated tryptic peptides contain a positively charged lysine or arginine at the C-terminus as well as a positively charged amino terminus. Since a single sample prepared for MALDI may be analyzed in positive mode and then again using negative mode, a more effective matrix for the detection of phosphorylated peptides using either negative or positive ion MALDI mass spectrometry is required. Commonly used matrixes for peptide and protein analysis were evaluated for the analysis of proteolytic digests of phosphorylated proteins using positive ion mode. For example, 2,5-dihydroxybenzoic acid and sinapinic acid produced low responses for the phosphorylated peptides in the unpurified β-casein tryptic digest (data not shown). Furthermore, the total of all peptides that were detected using these matrixes represented less than 50% coverage of the original protein. More complete peptide mapping coverage 1534

Analytical Chemistry, Vol. 76, No. 5, March 1, 2004

of the protein sequence of β-casein (>50% at a mass accuracy tolerance of 50 ppm) was obtained using either CHCA or CHCA containing diammonium citrate as the MALDI matrix (see the positive ion MALDI TOF mass spectra in Figure 2). However, the two phosphoserine-containing tryptic peptides of β-casein, a monophosphorylated peptide of m/z 2061.8 and a tetraphosphorylated peptide of m/z 3121.3, elicited low signal responses when CHCA or CHCA with diammonium citrate was used as the MALDI matrix. Although the signals for the phosphorylated peptides obtained using CHCA with diammonium citrate remained weak (see Figure 2B), the addition of diammonium citrate to the CHCA provided better signal responses than the use of CHCA alone (Figure 2A). Next, THAP mixed with diammonium citrate was used as the matrix for positive ion MALDI, and the results are shown in Figure

Figure 3. Positive ion MALDI TOF mass spectra with PSD of the phosphorylated peptides of (A) m/z 2062 and (B) 3122 from β-casein using THAP with diammonium citrate as the matrix. The sequential losses of 98 u from the protonated molecule indicate the number of phosphate groups in each phosphorylated peptide.

2C. Like CHCA, the use of the THAP matrix during positive ion MALDI TOF mass spectrometry produced more than 50% coverage of the protein sequence of β-casein in a single mass spectrum of the proteolytic digest. However, the signal responses for the two phosphorylated peptides from β-casein were enhanced more than 10-fold when THAP was used compared to CHCA. Based on these promising results, THAP containing diammonium citrate was used for all subsequent analyses. THAP is an unconventional matrix for peptide mapping but is used often for the negative ion MALDI TOF mass spectrometric analysis of oligonucleotides.13 When samples containing sodium contaminants (which are often present in oligonucleotide or peptide preparations) are prepared for MALDI using diammonium citrate and THAP, disodium citrate probably precipitates first during the cocrystallization procedure, thereby decreasing the sodium ion concentration in the sample solution. In addition, the ammonium cations probably form complexes with the phosphate groups and facilitate their ionization during positive ion MALDI TOF mass spectrometry.14 In the case of peptide analysis, both phosphorylated and unphosphorylated ions are detected as protonated molecules due to the loss of ammonia from the complexes during MALDI TOF mass spectrometry. The abundances of the protonated phosphorylated peptides in the MALDI mass spectrum of the tryptic digest of β-casein shown in Figure 2C were high enough for PSD mass spectrometric analysis (Figure 3). MALDI TOF mass spectrometry with PSD may be used to obtain structurally significant fragment ions of peptides through the analysis of metastable fragmentation.15 The PSD mass spectra of the β-casein phosphorylated peptides in Figure 3 were dominated by [MH - 98]+ ions, which correspond to elimination of H3PO4. The next most abundant product ions, [MH - 80]+, were formed by loss of HPO3. Although the facile loss of H3PO4 and HPO3 prevented the formation of amino acid sequence ions, the presence of these abundant PSD fragment ions may be used to confirm which peptide ions are phosphorylated. In addition, the number of phosphate groups on a particular peptide may be determined using this approach. For example, the positive ion PSD MALDI TOF mass spectrum of (12) (13) (14) (15)

http://prospector.ucsf.edu. Last accessed on October 10, 2003. Papac, D. I.; Wong, A.; Jones, A. J. S. Anal. Chem. 1996, 68, 3215-23. Asara, J. M.; Allison, J. J. Am. Soc. Mass Spectrom. 1999, 10, 35-44. Spengler, B.; Kirsch, D.; Kaufmann, R.; Jaeger, E. Rapid Commun. Mass Spectrom. 1992, 6, 105-8.

Figure 4. A) Positive ion MALDI TOF mass spectrum of the Lys-C digest of PKC-treated cTnI using THAP with diammonium citrate as the matrix showing a monophosphorylated peptide of m/z 2791 (/) corresponding to RPTLRRVRISADAMMQALLGTRAK. Three potential sites of phosphorylation are indicated in boldface type. (B) Positive ion MALDI TOF PSD mass spectrum of the peptide of m/z 2791 detected in (A) confirming that this peptide is monophosphorylated. Note the loss of a single molecule of H3PO4, [MH - H3PO4]+, at m/z 2693.

the monophosphorylated peptide from β-casein in Figure 3A shows the loss of only one H3PO4 molecule. In contrast, the elimination of one, two, three, and four H3PO4 molecules was observed in the PSD mass spectrum shown in Figure 3B, which corresponds to a tetraphosphorylated peptide. Therefore, the use of THAP with diammonium citrate and PSD during MALDI TOF mass spectrometry facilitates peptide mapping of phosphorylated proteins, the identification of phosphorylated peptides using PSD, and the determination of the number of phosphate groups on each peptide. Cardiac Troponin I. Following phosphorylation and SDSPAGE, in-gel enzymatic cleavage by the endoproteinase Lys-C was carried out. Lys-C was selected because it is a highly specific proteolytic enzyme that was expected to produce peptides of cTnI of appropriate size and number for MALDI TOF mass spectrometric analysis. The proteolytic digest was analyzed using positive ion MALDI TOF mass spectrometry and THAP with diammonium citrate as the matrix. Signals for 13 cTnI peptides were identified in the MALDI TOF mass spectrum, which covered 54% of the protein sequence (Figure 4A). The peak at m/z 2791.5 corresponds to a monophosphorylated peptide consisting of amino acids 142-165 (RPTLRRVRISADAMMQALLGTRAK; the possible sites of phosphorylation are indicated in boldface type). Subsequent analysis of the ion of m/z 2791.5 using positive ion MALDI TOF mass spectrometry with PSD (Figure 4B) indicated that this peptide was monophosphorylated, since there was a single abundant fragment ion of m/z 2693.5 corresponding to [MH H3PO4]+. However, there were three possible phosphorylation sites in this peptide, 144T, 151S, and 162T. To resolve the ambiguity concerning the site of phosphorylation in the monophosphorylated peptide 142-165 of cTnI, a second proteolytic digestion was carried out using trypsin. The choice of trypsin for a second proteolytic digestion was based on the presence of multiple arginine residues in the peptide 142-165. In retrospect, it would have been more efficient to hydrolyze cTnI with trypsin alone and omit the digestion with Lys-C. However, many of the tryptic peptides were too small for detection using MALDI and essential information might have been missed if only Analytical Chemistry, Vol. 76, No. 5, March 1, 2004

1535

Figure 5. (A) Positive ion MALDI TOF mass spectra of PKC-treated cTnI after sequential digestion using Lys-C and then trypsin. THAP with diammonium citrate was used as the matrix. (B) Positive ion MALDI-TOF PSD mass spectrum of the protonated peptide of m/z 878 (/) obtained using THAP with diammonium citrate as the matrix. The loss of a single molecule of H3PO4 confirmed that this peptide was monophosphorylated. Based on peptide mapping of the Lys-C and trypsin digested cTnI, this peptide was determined to correspond to the amino acid sequence RPpTLRR.

tryptic digestion had been used. Therefore, we still recommend a limited initial proteolysis using enzymes such as Lys-C or ArgC. The positive ion MALDI TOF mass spectrum of Lys-C and trypsin-digested cTnI is shown in Figure 5A. A phosphorylated hexapeptide 142-147 (RPpTLRR) was detected at m/z 878.47, and PSD analysis of this peptide confirmed that it was monophosphorylated (Figure 5B). Since 144T is the only possible phosphorylation site in peptide 142-147, the site of monophosphorylation of cTnI by PKC was identified. CONCLUSIONS THAP with diammonium citrate was shown to be an efficient matrix for either negative or positive ion MALDI mass spectrometric analyses of phosphorylated peptides in mixtures. Compared

1536 Analytical Chemistry, Vol. 76, No. 5, March 1, 2004

to CHCA, CHCA plus diammonium citrate, and other frequently used matrixes for peptide and protein analyses, the signal-to-noise ratio for phosphorylated peptides using THAP with diammonium citrate as the MALDI matrix was enhanced >10-fold while the peptide mapping coverage of the digested protein was maintained at >50%. Therefore, THAP with diammonium citrate helps prevent the suppression of phosphorylated peptide ions that occurs during positive ion MALDI mass spectrometric analysis of digests of phosphorylated peptides. Usually used for the analysis of oligonucleotides, the THAP diammonium citrate matrix probably facilitates the ionization of phosphorylated as well as unphosphorylated peptides by sequestering sodium cations as sodium citrate and by forming an ammonium adduct followed by elimination of ammonia during MALDI. Therefore, phosphate groups probably are analyzed as undissociated phosphoric acids, and an additional proton is added to each peptide to form a protonated molecule. The high abundances of the protonated phosphorylated peptides in the MALDI TOF mass spectra obtained using the THAP diammonium citrate matrix facilitated the use of PSD to obtain structurally significant fragment ions. During PSD, the phosphoric acid groups on each phosphorylated peptide were easily eliminated, which confirmed the numbers of phosphate groups on each peptide. Finally, the utility of this matrix for the mapping of phosphorylated peptides and identification of sites of phosphorylation was demonstrated on digests of the model protein β-casein and PKC-treated cTnI. The latter investigation resulted in the identification of a unique phosphorylation site in cTnI. ACKNOWLEDGMENT This work was supported by Grant S10RR14686 from the National Institutes of Health. Received for review December 28, 2003. AC035203V

October

10,

2003.

Accepted