C18 Tip

Anal. Chem. 2005, 77, 5144-5154

Specificity of Immobilized Metal Affinity-Based IMAC/C18 Tip Enrichment of Phosphopeptides for Protein Phosphorylation Analysis Makiko Kokubu,†,‡ Yasushi Ishihama,† Toshitaka Sato,† Takeshi Nagasu,† and Yoshiya Oda*,†

Laboratory of Seeds Finding Technology, Eisai Co., Ltd., Tokodai 5-1-3, Tsukuba, Ibaraki 300-2635, Japan, and Graduate School of Medicine, The University of Tsukuba, Tennodai 1-1-1, Tsukuba, Ibaraki 305-8575, Japan.

We have developed a simple, highly specific enrichment procedure for phosphopeptides, by increasing the specificity of an immobilized metal affinity column (IMAC) without using any chemical reaction. The method employs a biphasic IMAC-C18 tip, in which IMAC beads are packed on an Empore C18 disk in a 200-µL pipet tip. Phosphopeptides are separated from non-phosphopeptides on the IMAC in an optimized solvent without any chemical reaction, then desorbed from the IMAC using a phosphate buffer, reconcentrated, and desalted on the C18 disk. The increase in selectivity was achieved by (a) using a strong acid to discriminate phosphates from carboxyl groups of peptides and (b) using a high concentration of acetonitrile to remove hydrophobic non-phosphopeptides. The entire procedure was optimized by using known phosphoproteins such as Akt1 kinase and protein kinase A. Although it was difficult to detect phosphopeptides in MALDI-MS spectra of tryptic peptide mixtures before enrichment, after the IMAC procedure, we could successfully detect phosphopeptides with almost no nonphosphopeptides. Next, we constructed an array of IMACIMAC/C18 tips, such that number of arrayed tips on a 96-well plate could easily be changed depending on the loading amount of sample. Applying this approach to mouse forebrain resulted in the identification of 162 phosphopeptides (166 phosphorylation sites) from 135 proteins using nano-LC/MS. Recent advances in analytical methods, particularly in the area of mass spectrometry (MS), have brought the field of proteomics to the forefront in biological science.1,2 Significant progress has been made in sequence analysis of peptides at the femtomole level, permitting the rapid and effective identification and quantification of a large number of proteins within a given biological sample.3-5 * Corresponding author. Phone: +81-29(847)7098. Fax: +81-29(847)7614. E-mail: [email protected]. † Eisai Co., Ltd. ‡ The University of Tsukuba. (1) Lin, D.; Tabb, D. L.; Yates, J. R., 3rd. Biochim. Biophys. Acta 2003, 1646, 1-10. (2) Yates, J. R., 3rd. Annu. Rev. Biophys. Biomol. Struct. 2004, 33, 297-316. (3) Aebersold, R.; Mann, M. Nature 2003, 422, 198-207. (4) Patterson, S. D.; Aebersold, R. Nat. Genet. 2003, 33 (Suppl.), 311-323. (5) de Hoog, C. L.; Mann, M. Annu. Rev. Genomics Hum. Genet. 2004, 5, 267293.

5144 Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

The identification of proteins, however, is not always sufficient to interpret biological function, because many of the naturally occurring proteins are posttranslationally modified.6,7 One such ubiquitous modification is phosphorylation, which affects a significant subset of the proteome and plays an especially important role in numerous regulatory pathways and cell cycle control in all living cells.8-10 Studies in protein phosphorylation may hold particular promise for dissecting signaling pathways, molecular classification of diseases, and profiling of novel kinase inhibitors.11-14 Although there is a great need for methods capable of accurately elucidating sites of phosphorylation, it is generally difficult to identify in vivo phosphorylation sites within proteins of interest. Selective detection of phosphopeptides from proteolytic digests is a challenging and critical task in many proteomics applications, because often cellular levels of protein phosphorylation are low and the detection sensitivity of phosphorylated peptides is reduced by ion suppression effects during the initial ionization process in MS.15-17 More importantly, phosphopeptides are buried among a huge number of non-phosphopeptide shadows.18 Therefore, selective isolation or enrichment of phosphopeptides is needed before identification. Strategies to preferentially enrich phosphoproteins include application of phospho-Ser/Thr-specific chemistry based on β-elimination,19-25 use of phosphoTyr-specific antibodies,26-29 (6) Mann, M.; Jensen, O. N. Nat. Biotechnol. 2003, 21, 255-261. (7) Jensen, O. N. Curr. Opin. Chem. Biol. 2004, 8, 33-41. (8) Graves, J. D.; Krebs, E. G. Pharmacol. Ther. 1999, 82, 111-121. (9) Bollen, M.; Beullens, M. Trends. Cell. Biol. 2002, 12, 138-145. (10) Manning, G.; Plowman, G. D.; Hunter, T.; Sudarsanam, S. Trends Biochem. Sci. 2002, 27, 514-520. (11) Honkanen, R. E.; Golden, T. Curr. Med. Chem. 2002, 9, 2055-2075. (12) Cohen, P. Nat. Rev. Drug Discovery 2002, 1, 309-315. (13) Fischer, P. M. Curr. Med. Chem. 2004, 11, 1563-1583. (14) Lu, K. P. Trends Biochem. Sci. 2004, 29, 200-209. (15) Carr, S. A.; Huddleston, M. J.; Annan, R. S. Anal. Biochem. 1996, 239, 180192. (16) McLachlin, D. T.; Chait, B. T. Curr. Opin. Chem. Biol. 2001, 5, 591-602. (17) Shou, W.; Verma, R.; Annan, R. S.; Huddleston, M. J.; Chen, S. L.; Carr, S. A.; Deshaies, R. J. Methods Enzymol. 2002, 351, 279-296. (18) Mann, M.; Ong, S. E.; Gronborg, M.; Steen, H.; Jensen, O. N.; Pandey, A. Trends Biotechnol. 2002, 20, 261-268. (19) Oda, Y.; Nagasu, T.; Chait, B. T. Nat. Biotechnol. 2001, 19, 379-382. (20) Adamczyk, M.; Gebler, J. C.; Wu, J. Rapid Commun. Mass Spectrom. 2001, 15, 1481-1488. (21) Qian, W. J.; Goshe, M. B.; Camp, D. G., 2nd.; Yu, L. R.; Tang, K.; Smith, R. D. Anal. Chem. 2003, 75, 5441-5450. (22) Knight, Z. A.; Schilling, B.; Row: R. H.; Kenski, D. M.; Gibson, B. W.; Shokat, K. M. Nat. Biotechnol. 2003, 21, 1047-1054. (23) Klemm, C.; Schroder, S.; Gluckmann, M.; Beyermann, M.; Krause, E. Rapid Commun. Mass Spectrom. 2004, 18, 2697-2705. 10.1021/ac050404f CCC: $30.25

© 2005 American Chemical Society Published on Web 07/08/2005

Figure 1. MALDI-MS spectra of tryptic peptides of ovalbumin (5 µg) enriched by IMAC/C18 tip. (A) 0.1% acetic acid was used as loading and washing solvent. (B) 0.1% TFA was used as loading and washing solvent.

collection of the early-eluting fraction from a strong cation exchange (SCX) column for tryptic phosphopeptides,30,31 enrich(24) Jiang, X.; Wang, Y. Biochemistry 2004, 43, 15567-15576. (25) Vosseller, K.; Hansen, K. C.; Chalkley, R. J.; Trinidad, J. C.; Wells, L.; Hart, G. W.; Burlingame, A. L. Proteomics 2005, 5, 388-398. (26) Pandey, A.; Podtelejnikov, A. V.; Blagoev, B.; Bustelo, X. R.; Mann, M.; Lodish, H. F. Proc. Natl. Acad. Sci. U. S. A 2000, 97, 179-184. (27) Pandey, A.; Fernandez, M. M.; Steen, H.; Blagoev, B.; Nielsen, M. M.; Roche, S.; Mann, M.; Lodish, H. F. J. Biol. Chem. 2000, 275, 38633-3869. (28) Hinsby, A. M., Olsen, J. V., Bennett, K. L., and Mann, M. Mol. Cell. Proteomics 2003 2, 29-36. (29) Hinsby, A. M. Olsen, J. V.; Mann, M. J. Biol. Chem. 2004, 279, 4643846447. (30) Beausoleil, S. A.; Jedrychowski, M.; Schwartz, D.; Elias, J. E.; Villen, J.; Li, J.; Cohn, M. A.; Cantley, L. C.; Gygi, S. P. Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 12130-12135. (31) Ballif, B. A.; Villen, J.; Beausoleil, S. A.; Schwartz, D.; Gygi, S. P. Mol. Cell. Proteomics 2004, 3, 1093-1101.

ment by immobilized metal ion affinity chromatography (IMAC)32-40 and combination of IMAC with β-elimination chemistry.41-43 Among them, the IMAC method is a fast, easy-to-use, and (32) Andersson, L.; Porath, J. Anal. Biochem. 1986, 154, 250-254. (33) Figeys, D.; Gygi, S. P.; Zhang, Y.; Watts, J.; Gu, M.; Aebersold, R. Electrophoresis 1998, 19, 1811-1818. (34) Posewitz, M. C.; Tempst, P. Anal. Chem. 1999, 71, 2883-2892. (35) Li, S.; Dass, C. Anal. Biochem. 1999, 270, 9-14. (36) Cao, P.; Stults, J. T. Rapid Commun. Mass Spectrom. 2000, 14, 1600-1606. (37) Annan, R. S.; Huddleston, M. J.; Verma, R.; Deshaies, R. J.; Carr, S. A. Anal. Chem. 2001 73, 393-404. (38) Nuhse, T. S.; Stensballe, A.; Jensen, O. N.; Peck, S. C. Mol. Cell. Proteomics 2003, 2, 1234-1243. (39) Shu, H.; Chen, S.; Bi, Q.; Mumby, M.; Brekken, D. L. Mol. Cell. Proteomics 2004, 3, 279-286. (40) Hirschberg, D.; Jagerbrink, T.; Samskog, J.; Gustafsson, M.; Stahlberg, M.; Alvelius, G.; Husman, B.; Carlquist, M.; Jornvall, H.; Bergman, T. Anal. Chem. 2004, 76, 5864-5871.

Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

5145

which does not require any chemical reaction, but can increase the specificity of IMAC with simple procedures.

Figure 2. Effect of TFA concentration on IMAC/C18 tip. Tryptic peptides of ovalbumin (5 µg) were enriched using different concentrations of TFA in the loading and washing solvents. The bar graph shows peak intensities in MALDI spectra of phosphopeptides, and the line graph (9) shows total peak intensities of non-phosphopeptides.

economic procedure, but many non-phosphopeptides also bind to IMAC. Although the use of IMAC coupled with an SCX column allowed identification of a large number of phosphopeptides from yeast,44 this required a double enrichment strategy coupled with improvements in high-speed scanning, high-quality MS/MS spectra assisted with MS/MS/MS analysis, and high mass accuracy in Fourier transform ion cyclotron resonance MS. Another large-scale phosphoproteomics study using IMAC involved a double IMAC protocol on several levels, which enabled identification and characterization of phosphoproteins across a wide range of protein abundance and phosphorylation states.45 This procedure is relatively simple, but multiple processes result in loss of samples in general, and essentially, the selectivity of single IMAC was not improved. It is still of interest to decrease the binding of non-phosphopeptides to the IMAC column, because it would become much easier to detect phosphopeptides in target phosphoproteins and would decrease the likelihood of falsepositive identification. Although methyl esterification of carboxylate groups could decrease nonspecific binding of acidic peptides, this would require a chemical reaction step before the IMAC process.46-48 Here, we have developed an alternative approach, (41) Li, W.; Backlund, P. S.; Boykins, R. A.; Wang, G.; Chen, H. C. Anal. Biochem. 2003, 323, 94-102. (42) Thompson, A. J.; Hart, S. R.; Franz, C.; Barnouin, K.; Ridley, A.; Cramer, R. Anal. Chem. 2003, 75, 3232-3243. (43) Ahn, Y. H.; Park, E. J.; Cho, K.; Kim, J. Y.; Ha, S. H.; Ryu, S. H.; Yoo, J. S. Rapid Commun. Mass Spectrom. 2004, 18, 2495-2501. (44) Gruhler, A.; Olsen, J. V.; Mohammed, S.; Mortensen, P.; Faergeman, N. F.; Mann, M.; Jensen, O. N. Mol. Cell. Proteomics 2005, 4, 310-327. (45) Collins, M. O.; Yu, L.; Coba, M. P.; Husi, H.; Campuzano, I.; Blackstock, W. P.; Choudhary, J. S.; Grant, S. G. J. Biol. Chem. 2005, 280, 5972-5982. (46) Ficarro, S. B.; McCleland, M. L.; Stukenberg, P. T.; Burke, D. J.; Ross, M. M.; Shabanowitz, J.; Hunt, D. F.; White, F. M. Nat. Biotechnol. 2002, 20, 301-306.

5146

Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

MATERIALS AND METHODS Modified trypsin was obtained from Promega (Madison, WI). R-Cyano-4-hydroxycinnamic acid (CHCA) was purchased from Bruker Daltonics (Bremen, Germany). R-Casein (bovine), β-casein (bovine), ovalbumin (chicken egg), phosvitin (chicken egg), and Phos-Select were from Sigma (St. Louis, MO). Akt1 kinase (active form), protein kinase A (catalytic subunit), and myelin basic protein (MBP) were obtained from Upstate USA, Inc. (Charlottesville, VA). C18 Empore disk was from 3M (St. Paul, MN). 5-Cyclohexyl-1-pentyl β-D-maltoside (CYMAL-5) was obtained from Anatrace (Maumee, OH). Negative gel stain MS kit was obtained from Wako Pure Chemicals (Osaka, Japan). Gels with a thickness of 1.0 mm (Tris-HCl, 5-20% T) were obtained from DRC (Tokyo, Japan). All reagents were of sequence grade, and water was obtained from a Milli-Q system (Millipore, Bedford, MA). IMAC/C18 Tip. Custom-made biphasic chromatographic microcolumns used for enrichment of phosphopeptides prior to mass spectrometric analysis were prepared as follows. C18 Empore disk was packed into a 200-µL micropipet tip (C18 StageTip) with a length of ∼1 mm, as described in detail earlier,49 and then 30 µL of Phos-Select gel (IMAC-Fe3+type, approximately solvent/supporting material ) 1/1) was loaded onto the C18 Stage tip. The C18 Empore disk serves as a frit to retain the Phos-Select gel, and a chromatographic column to desalt and concentrate samples. Optimized Protocol for an IMAC/C18 Tip. Sample loading, washing, and elution were performed by centrifugation at 1200g for 1-3 min. Each tip was used only once to avoid contamination. 200 µL of acetonitrile/water/trifluoroacetic acid (TFA) ) 50/50/ 0.3 (TWA) was used to equilibrium an IMAC/C18 tip. Samples were dissolved in 200 µL of TWA or 20-fold diluted with TWA. After sample loading onto IMAC/C18 tips, samples were washed with 100 µL of TWA and then 200 µL of 1% phosphoric acid was added onto the IMAC/C18 tip to transfer samples from IMAC to a C18 Empore disk. After washing with 200 µL of acetonitrile/ water/TFA (5/95/0.1), 10 µL of acetonitrile/water/TFA (50/50/ 0.1) was added onto the IMAC/C18 tip. 0.5 µL of the eluate was deposited onto a MALDI-MS plate directly, and the remaining sample was dried in a Speedvac. Mass Spectrometry. MALDI-MS was performed using an ABI 4700 instrument (Applied Biosystems, Framingham, MA). The matrix solution was saturated with CHCA and was prepared by adding solid CHCA to acetonitrile, followed by the addition of 0.1% TFA. The mixture was thoroughly vortexed and centrifuged, leaving a clear working matrix solution. A 0.5-µL aliquot of a sample was deposited onto a MALDI plate, and then 0.5 µL of the working matrix solution was slowly put on the sample droplet. The sample-matrix droplet was allowed to air-dry at room temperature. Spectra were obtained in the positive linear mode using external calibration. An LTQ (Thermo Electron, CA) nano(47) Salomon, A. R.; Ficarro, S. B.; Brill, L. M.; Brinker, A.; Phung, Q. T.; Ericson, C.; Sauer, K.; Brock, A.; Horn, D. M.; Schultz, P. G.; Peters, E. C. Proc. Natl. Acad. Sci. U. S. A. 2003, 100, 443-448. (48) Brill, L. M.; Salomon, A. R.; Ficarro, S. B.; Mukherji, M.; Stettler-Gill, M.; Peters, E. C. Anal. Chem. 2004, 76, 2763-2772. (49) Rappsilber, J.; Ishihama, Y.; Mann, M. Anal. Chem. 2003, 75, 663-670.

Figure 3. MALDI-MS spectra of tryptic peptides of ovalbumin (5 µg) enriched by IMAC/C18 tips. (A) 0.1% TFA without organic solvent was used as loading and washing solvents. (B) 50% acetonitrile with 0.1% TFA was used as loading and washing solvents.

LC/MS system was used, equipped with laboratory-made nanospray ion source (spray voltage, 2.5 kV) and LC10 A gradient pumps (Shimazdu, Kyoto, Japan). The nano-LC column was a “stone-arch” column packed in-house with 100 mm × 0.1 mm C18 beads50 and was used at a flow rate of 200 nL/min after flow splitting. A linear gradient of B from 5 to 30% was run, using mobile phase A of 0.5% acetic acid and mobile phase B of 0.5% acetic acid/acetonitrile(20/80). The MS scan range was 350-1500 m/z, and LC/MS spectra were collected in a data-dependent mode in which the highest intensity peaks in each MS scan were chosen for collision-induced dissociation, and the isolation window was 3 Da (precursor m/z 1.5 Da) to select a precursor ion. Also, the dynamic exclusion option was selected with a repeat count of 2, a repeat duration of 0.5 min, and an exclusion duration of 1 min. (50) Ishihama, Y.; Rappsilber, J.; Andersen, J. S.; Mann, M. J. Chromatogr., A 2002, 979, 233-239.

Neutral loss of phosphoric acid from phospho-Ser/Thr peptides is a common feature of ion trap MS/MS spectra, and therefore, subsequent MS/MS/MS were automatically collected with respect to such neutral losses of phosphate peaks, which appear as losses of 49 and 32.6 amu ((0.7 Da.) from double- and triple-charged precursor phosphopeptides, respectively, observed as base peaks on MS/MS spectra. Samples were redissolved in 10 µL of acetonitrile/water/TFA (5/95/0.1), and then 3 µL was injected onto the LC/MS. LC/MS data were analyzed against NCBI nonredundant databases (Mus musculus database) using MASCOT (Matrix Science, London) with propionamide (Cys) as fixed modification allowing one miscleavage of trypsin. Oxidized methionine and phosphorylation (Ser/Thr or Tyr) were searched as variable modifications. Searches were done as peptide charges of 2+ and 3+ and the tolerance of mass measurement of 2 Da in MS mode and 0.8 Da for MS/MS ions. MS/MS spectra given Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

5147

Figure 4. Effect of acetonitrile concentration on IMAC/C18 tip. Tryptic peptides of ovalbumin (5 µg) were enriched using different concentrations of acetonitrile as loading and washing solvents. The bar graph shows peak intensities in MALDI spectra of phosphopeptides, and the line graph (9) shows total peak intensities of nonphosphopeptides.

scores of 35 or more were manually validated, and data-dependent MS/MS/MS spectra allowed us to confirm or reject the putative phosphopeptides;30,44,51,52 as a result, more than half the data seemed to be false positive. In-Gel Digestion. After SDS-PAGE, target bands were excised, and in-gel digestion was carried out as described previously.53 Briefly, the excised gel plugs were vigorously washed with methanol/water/acetic acid several times, neutralized with 50 mM ammonium bicarbonate, dehydrated with acetonitrile, and dried by vacuum centrifugation. A few microliters of trypsin (1030 ng) dissolved in 50 mM ammonium bicarbonate with 0.1% CYMAL-5 was added to the dry gel pieces and incubated until the gel recovered to its original size. If the recovery was insufficient, additional digestion buffer without trypsin was added and the digestion was continued at 37 °C for 4-18 h. The digested peptides were extracted with 50% acetonitrile containing 0.1% TFA twice in a 300-W sonicator. Mouse Forebrain Sample. All mice were treated ethically according to the rules of the Eisai Co., Ltd. Animal Use and Care Committee. Male C57BL/6 mice were used at eight weeks old. These mice were sacrificed, and the brains were removed and frozen immediately. The extraction buffer was typically Tris-HCl (50 mM, pH 7.3) supplemented with EDTA (1 mM), NaF (1 mM), Na3VO4 (2 mM), and protease inhibitor cocktail (Roche Diagnostics). Ice-cold extraction buffer was added to frozen brains, and proteins were extracted for 15 min using a Teflon homogenizer on ice. The soluble fraction was obtained by centrifugation at 100000g+ for 60 min at 4 °C. Protein concentration of the extract was determined using the Pierce micro BCA reagent kit. The protein solution was adjusted at pH 8.6 by adding 0.5 M Tris-HCl (51) Jin, W. H.; Dai, J.; Zhou, H.; Xia, Q. C.; Zou, H. F.; Zeng, R. Rapid Commun. Mass Spectrom. 2004, 18, 2169-2176. (52) Olsen, J. V.; Mann, M. Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 1341713422. (53) Katayama, H.; Tabata, T.; Ishihama, Y.; Sato, T.; Oda, Y.; Nagasu, T. Rapid Commun. Mass Spectrom. 2004, 18, 2388-2394.

5148

Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

buffer, and urea was dissolved in that sample until saturation. For reduction and alkylation of cysteine/cystine residues, 1 mg of dithiothreitol was added into 1 mg of protein solution and the mixture was incubated at 37 C for 1 h. A 1.1-mg portion of acrylamide was added into that mixture. After 1 h incubation at 37 °C, the sample was 4-fold diluted with water and then digested with 20 µg of trypsin for each 1 mg of proteins. Tryptic peptides mixtures were acidified with 2% of TFA (For a 0.1 M Tris-HCl buffered solution, equal volume is sufficient.). A C18-Stage tip was conditioned with 5 µL of methanol and then was equilibrated with 10 µL of acetonitrile/water/TFA (5/95/0.1). Acidified sample was loaded onto the C18-Stage tip up to 10 µg of protein per one C18Stage tip to remove hydrophilic phosphate-related substances such as nucleotides. After washing with 10 µL of acetonitrile/water/ TFA (5/95/0.1), the sample was eluted with 10 µL of acetonitrile/ water/TFA (50/50/0.1). The eluate was directly served for an IMAC/C18 tip to enrich phosphopeptides or dried in a Speedvac for non-phosphopeptide analysis. SCX Empore disk was packed as an SCX-Stage tip. The dried sample for non-phosphopeptide analysis was redissolved in 0.1% acetic acid and then was loaded onto the SCX-Stage tip preconditioned with in 200 µL of 0.1% acetic acid. The sample was eluted with a five-step ammonium acetate gradient (25-500 mM), and then these fractions were desalted by using a C18-Stage tip as mentioned above prior to LC/MS analysis.54 RESULTS AND DISCUSSION Optimization of IMAC/C18 Tip. After enrichment on IMAC and competitive elution of the phosphopeptides with phosphate buffer, the standard approach used in proteomics is to desalt and concentrate the eluted peptides using RP chromatographic material,55 such as ZipTips (Millipore) and C18 Stage-Tips49 prior to MS detection. Therefore, we constructed a biphasic IMAC-C18 tip by loading IMAC onto 200-µL tip microcolumns packed with a C18 Empore disk to allow enrichment, desalting, and concentration of small amounts of phosphopeptides. IMAC suffers from many nonspecific interactions, which increase the chance of false-positive results. One of the major nonspecific binders is acidic peptides. Methyl esterification of carboxyl groups of peptides improved the selectivity of IMAC for phosphopeptides by eliminating the acidic bias.46-48 However, the chemical reaction is sometimes incomplete, especially in multiple Asp/Glu-containing peptides, the solubility of peptides is decreased after esterification, and the attachment of chemical tags and removal of excess reagents can lead to sample losses. Here we examined an alternative approach to decrease nonspecific ionic interactors. Almost all IMAC protocols employ acetic acid in the buffer system, because it is believed that strong acids such as TFA prevent the interaction of phosphopeptides with IMAC by the protonation of phosphate groups. However, the pKa value of phosphate groups (pK1 of phosphoric acid is 2.15) is lower than that of carboxyl groups (pKa of β-COOH of Asp is 3.65 and γ-COOH of Glu is 4.25). We compared four different acids, TFA, hydrochloric acid, formic acid, and acetic acid, as additives at (54) Ishihama, Y.; Sato, T.; Tabata, T.; Miyamoto, N.; Sagane, K.; Nagasu, T.; Oda, Y. Nat. Biotechnol. 2005, 23, 617-621. (55) Wu, X.; Ranganathan, V.; Weisman, D. S.; Heine, W. F.; Ciccone, D. N.; O’Neill, T. B.; Crick, K. E.; Pierce, K. A.; Lane, W. S.; Rathbun, G.; Livingston, D. M.; Weaver, D. T. Nature 2000, 405, 477-482.

various concentrations (0.1-10%) using tryptic peptides of ovalbumin. Typical MALDI-TOF/MS spectra are shown in Figure 1. Relative intensities of phosphopeptides versus non-phosphopeptides on ovalbumin increased in the following order: TFA g hydrochloric acid > formic acid > acetic acid. This is similar to the order of acid strength. Strong acids protonated carboxyl groups, but still dissociated phosphates, which allows the discrimination of phosphoamino acids from acidic residues on IMAC. We optimized the concentration of TFA in the buffer system for IMAC/C18 tips as shown in Figure 2. Low concentrations of TFA were insufficient for the protonation of carboxyl groups, and higher concentrations resulted in loss of negative charge on phosphates, which is ineffective for IMAC. We found that 0.11% TFA in buffer was suitable for IMAC/C18 tips. We also explored the possibility of adding salt to the buffer system, because varying salt concentration is commonly used for elution of samples from ion exchange columns. Interestingly, up to 1 M sodium chloride in the buffer and in the sample had neither a positive nor a negative effect on IMAC/C18 tips. Another major factor in nonspecific binding on affinity columns is hydrophobic interaction. Although various detergents are used in biochemistry to decrease nonspecific interactions, they can block elution of peptides from an RP column, because they are distributed in both the liquid phase and the solid phase. Therefore, we tested four different organic solvents, acetonitrile (ACN), methanol (MeOH), ethanol (EtOH), and acetone, at 20-80% concentration to decrease hydrophobic interactions between nonphosphopeptides and IMAC. Among them, only ACN was very effective in eliminating non-phosphopeptides as shown in Figure 3. In ion exchange chromatography, a little ACN (5-20%) is sometimes added to the mobile-phase buffer to increase recovery of peptides, but ACN has not been used in IMAC buffer in most reported studies and Haydon et al. reported that increasing the ACN concentration to 25% (v/v) had no siginificant effect in reducing nonspecific binding to IMAC.56 We found that low ( 70%) concentrations of ACN were fairly ineffective, but 40-60% ACN dramatically improved the specificity on IMAC, as shown in Figure 4. The use of 50% ACN with 0.3% TFA not only enhanced the specific enrichment on IMAC but also washed away non-phosphopeptides from the C18 Empore disk due to higher ACN. Other four examples (R- and β-casein, MBP, phosvitin) are shown in Figure S-1 (A-D, Supporting Information). We also found that, once non-phosphopeptides became bound to IMAC, it was very difficult to separate them from phosphopeptides even with the optimized solvent (TWA) as a washing solvent. Although the reason for this is not clear, samples should be dissolved in the optimized solvent or diluted more than 20 times with the solvent. Recovery from IMAC/C18 Tip. We prepared synthetic phosphopeptides and non-phosphopeptide counterparts (the sequences are shown in Table 1) to measure recoveries from the IMAC/C18 tip with the optimized protocol. After IMAC/C18 tip enrichment, the non-phosphopeptides were added in the final sample, and then MALDI-MS analysis was conducted. As the reference sample, the phosphopeptide was directly added in to a MALDI sample without IMAC/C18 tip concentration, and peak (56) Haydon, C. E.; Eyers, P. A.; Aveline-Wolf, L. D.; Resing, K. A.; Maller, J. L.; Ahn, N. G. Mol. Cell. Proteomics. 2003, 2, 1055-1067.

Table 1. Recovery of Synthetic Phosphopeptides Using IMAC/C18 Tipa sequenceb GPETLpTYDAR GPETLTYDAR (reference) APESLpSYDAR APESLSYDAR (reference) TRDIpYETDYYRK TRDIYETDYYRK (reference)

no. of repeats

recovery (%)

RSD (%)

5

43.0

24.2

5

43.4

8.2

5

24.9

23.9

a 100 pmol of each amount of phosphopeptide was used. b Each reference peptide was 20 pmole.

ratios in MALDI spectra were calculated by comparison of the reference with the IMAC/C18 tip processed sample to evaluate recovery. As shown in Table 1, recovery was 43% for phosphorSer/Thr. Although the recovery of phosphopeptides from IMAC is dependent upon their sequences,56 we found that 10-15% of phosphopeptides was lost in the flow-through fraction from IMAC, 10-20% was still retained on IMAC, and 10-20% was lost during the C18 desalting step. We also confirmed that the incubation time (0-60 min) during sample loading onto the IMAC/C18 tip did not affect the enrichment efficiency. The loading capacity of this system was approximately a few micrograms of phosphopeptides. When ovalbumin was loaded on SDS-PAGE followed by in-gel digestion, the detection limit of the tryptic phosphoovalbumin enriched by IMAC/C18 tip was a few hundred femtomoles on SDS-PAGE in MALDI-MS analysis and less than 100 fmol in LC/ MS/MS analysis. The phospho-Tyr-containing peptide gave a lower recovery (25%) on the IMAC/C18 tip, probably because different phosphopeptides (sequences) give different recoveries56 or phospho-Tyr has a higher pKa value than phosphor-Ser/Thr and there may be steric hindrance on IMAC. From Small-Scale to Large-Scale Analysis. For small-scale analysis, purified Akt1 kinase and protein kinase A without any kinase reactions were each in-gel digested with trypsin after SDSPAGE separation, and the phosphopeptides were enriched with IMAC/C18 tips. It was difficult to find phosphopeptides in tryptic peptide mixtures on MALDI-MS spectra, but we could successfully detect phosphopeptides with almost no non-phosphopeptides by using the optimized solvent for the IMAC/C18 tip, as shown in Figure 5A (Akt1 kinase) and B (protein kinase A). In general, an optimized protocol for a small column is different from that for a large column, so a different amount of sample requires a suitable size of column with appropriate conditions. A robust procedure, which is independent of the analytical scale, is desirable to enrich phosphopeptides from peptide mixtures. We have constructed an array type of IMAC/C18 tips, where 96 IMAC/C18 tips were inserted onto a 96-well plate (Figure 6), and multiple 96-well plates were set in a centrifuge. Each IMAC/C18 tip on 96-well plate was exactly same as a single IMAC/C18 tip. Therefore, the protocol for large-scale analysis was the same as that for the small scale. We can easily adjust the number of arrayed tips on a 96-well plate depending on the loading amount (sample size). After enrichment of phosphopeptides, samples were collected by a multipipet, they were brought together in one, and then the final sample was analyzed by LC/MS. Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

5149

Figure 5. (A) MALDI-MS spectra of tryptic Akt1 kinase, (upper spectrum) before IMAC/C18 tip enrichment, and (lower spectrum) after IMAC/ C18 tip enrichment. 700 ng of Akt1 kinase was loaded and separated in SDS-PAGE. After in-gel digestion, phosphopeptides were enriched. (B) MALDI-MS spectra of tryptic protein kinase A, (upper spectrum) before IMAC/C18 tip enrichment, and (lower spectrum) after IMAC/C18 tip enrichment. The 500 ng of protein kinase A was loaded and separated in SDS-PAGE. After in-gel digestion, phosphopeptides were enriched.

Our approach was successfully applied to the detection and identification of phosphopeptides from mouse brain extracts. When protein extracts (3 mg) from mouse brain were processed by using 192 IMAC/C18 tips (2 × 96-well plates) after tryptic digestion, fractions from the IMAC/C18 tips were combined and analyzed by LC/MS (the base peak mass chromatogram is shown 5150

Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

in Figure S-2, Supporting Information). As a result, 1654 probable phosphopeptides were detected (MS/MS/MS was performed) and 96 non-phosphopeptides were identified. Mascot interpretation followed by manual inspections of MS/MS and MS/MS/MS spectra led to the identification of 166 sites of Ser/Thr phosphorylation (162 phosphopeptides) on a total of 135 different proteins

Table 2. List of Phosphopeptides from Mouse Brain Extract Identified in This Study

Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

5151

Table 2 (Continued)

5152

Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

Table 2 (Continued)

as shown in Table 2 (Typical 10 sets of MS/MS and MS/MS/ MS spectra are shown in Figure S-3 (A-J), Supporting Information.). There were 4 doubly phosphorylated peptides, and no phospho-Tyr-containing peptides, and 150 phosphorylation sites were on serine residues (90.3%). In 30 phosphopeptides out of 162 phosphopeptides, there was ambiguity regarding the exact

Figure 6. Array type of IMAC/C18 tips. Up to 96 IMAC/C18 tips were inserted onto a 96-well plate. Each IMAC/C18 tip on the 96well plate was exactly the same as a single IMAC/C18 tip.

phosphorylation sites, because fragment ions in the spectra were insufficient to identify the exact phosphoamino acids. Although 1654 probable phosphopeptides were detected, where the neutral loss of phosphoric acid was observed as the most intense fragment ions in the MS/MS spectra, a majority of the spectra was difficult to interpret, even with the assistance of MS/MS/MS spectra (Typical examples of MS/MS and MS/MS/MS spectra are shown in Figure S-4 (A, B), Supporting Information.). We were at least able to confirm MS/MS spectra in the essential step of falsepositive identification by manual evaluation,30 but it was very difficult to find false-negative spectra. Development of efficient and reliable search engines for phosphopeptides is still a major issue for phosphoproteomics. We compared our results with previously reported SCX-based results,31 as shown in Table 3; both analyses were performed using ion trap MS (Thermo Electron). The hydrophobicity values of identified peptides were similar in the three cases, which means that our approach did not enrich peptides based upon hydrophobicity, because we used 50% acetonitrile to decrease hydrophobic interaction between non-phosphopeptides and IMAC. Tryptic phosphopeptides have a lower charge at pH 2.7 than other tryptic peptides, so phosphopeptides are accumulated in the first fraction of SCX. In fact, the charge state of phosphopeptides enriched by SCX was -0.3, which is in agreement with the expected value. Although the states of the mouse brains examined might have been very different from each other, 80.2% of the identified phosphopeptides in our list did not overlap with those in the SCXAnalytical Chemistry, Vol. 77, No. 16, August 15, 2005

5153

Table 3. Comparison of Physicochemical Properties of Mouse Brain Phosphopeptides Enriched by IMAC/C18 Tip and SCX IMAC/C18 tipa average no. of identified phosphopeptides MH+d hydrophobicitye charge (pH 3)f no. of His-containing peptides no. of internal Lys/Arg-containing peptides

162 1699.4 17.0 0.3 23 (14.2%) 40 (24.7%)

SCX approachb

SD

average

430.3 3.1 0.7

460 2190.4 19.1 -0.3 17 (3.7%) 44 (9.6%)

SD

non-phosphopeptidesc average

SD

2103*3 648.2 4.2 0.4

1504.6 18.1 1.2 334 (15.9%) 164 (7.8%)

520.7 4.1 0.5

a This study (total 162 phosphopeptides). b Data from ref 31 (total 460 phosphopeptides). c 300 µg of the extracted mouse brain proteins was digested with trypsin after alkylation of cysteines and then fractionated by SCX into five fractions, which were analyzed by LC/MS/MS (LTQ devise): 2103 different non-phosphopeptides were identified. d These values were calculated as “without phosphates”. e These values were calculated using the following paper: Sakamoto, Y,; Kawakami, N.; Sasagawa. T, J. Chromatogr. 1988, 442, 69-79. f These values are positive charge states of peptides at pH 3 including phosphate groups if the peptides were phosphopeptides.

based approach. Interestingly, our approach contained quite a lot of miscleaved tryptic peptides (internal Lys/Arg-containing peptides), but the average length of our phosphopeptides was shorter than that of SCX-based phosphopeptides and the length of nonphosphopeptides from mouse brain was closer to that of our phosphopeptides. The reason for our result to contain a lot internal Lys/Arg-peptides is that it may become difficult for trypsin to recognize Lys/Arg near an acidic residue such as a phosphate. The SCX-based strategy misses the enrichment of some phosphopeptides, including those with His residues and miscleaved tryptic phosphopeptides (internal Lys/Arg residues).30 Also, the SCX approach does not work for other protease-digested peptides (not tryptic peptides), and the early-eluting fraction from SCX columns contains N-terminally blocked peptides (60-90% of known N-termini of eukaryotic proteins is modified57-60) and C-terminal peptides of proteins. Although the SCX-based approach has elucidated a huge number of phosphorylation sites, Gruhler et al. reported that the majority of identified peptides from the (57) Brown, J. L.; Roberts, W. K. J. Biol. Chem. 1976, 251, 1009-1014. (58) Brown, J. L. J. Biol. Chem. 1979, 254, 1447-1449. (59) Tsugita, A.; Kawakami, T.; Uchida, T.; Sakai, T.; Kamo, M.; Matsui, T.; Watanabe, Y.; Morimasa, T.; Hosokawa, K.; Toda, T. Electrophoresis 2000, 21, 1853-1871. (60) Gevaert, K.; Goethals, M.; Martens, L.; Van Damme, J.; Staes, A.; Thomas, G. R.; Vandekerckhove, J. Nat. Biotechnol. 2003, 21, 566-569. (61) Oda, Y.; Huang, K.; Cross, F. R.; Cowburn, D.; Chait, B. T. Proc. Natl. Acad. Sci. U. S. A. 1999, 96, 6591-6596. (62) Bonenfant, D.; Schmelzle, T.; Jacinto, E.; Crespo, J. L.; Mini, T.; Hall, M. N.; Jenoe, P. Proc. Natl. Acad. Sci. U. S. A. 2003, 100, 880-885. (63) Blagoev, B.; Ong, S. E.; Kratchmarova, I.; Mann, M. Nat. Biotechnol. 2004, 22, 1139-1145. (64) Ballif, B. A.; Roux, P. P.; Gerber, S. A.; MacKeigan, J. P.; Blenis, J.; Gygi, S. P. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 667-672.

5154 Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

SCX-based approach consisted of non-phosphopeptides.44 Overall, we believe that the IMAC/C18 tip method and the SCX-based approach will be complementary techniques for phosphoproteomics. In conclusion, we have developed a simple and a highly phosphopeptide-specific enrichment procedure, without any chemical reaction or expensive antibodies, that is suitable for both smallscale phosphoprotein analysis and large-scale phosphoproteomics. The IMAC/C18 tip approach in combination with stable isotope labeling of proteins allows direct comparison of the levels of phosphorylation between different states of samples.44,61-64 This technique, alone or in combination with other methods, should be useful in acquiring phosphorylation maps to elucidate signal transduction pathways in a variety of biological systems. ACKNOWLEDGMENT We thank Nikkyo Technos Co., Ltd. for providing the automatic robotic system to produce C18 Stage-Tips, Mitsui Knowledge Industry Co., Ltd. and Mr. Tsuyoshi Tabata for assistance with MS data management, and Dr. Takashi Seiki for animal care. This work was supported by funds from the New Energy and Industrial Technology Development Organization (NEDO), Japan. SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review March 8, 2005. Accepted June 3, 2005. AC050404F