Phosphopeptide Screening Using Nanocrystalline Titanium Dioxide

Jan 12, 2010 - Hartmut Schlu¨ ter,§ Ingo Ebner,† and Michael W. Linscheid*,†. Department of Chemistry, Humboldt-Universitaet zu Berlin, Berlin, ...
0 downloads 0 Views 3MB Size
Download PDF
Anal. Chem. 2010, 82, 1047–1053

Phosphopeptide Screening Using Nanocrystalline Titanium Dioxide Films as Affinity Matrix-Assisted Laser Desorption Ionization Targets in Mass Spectrometry Marie-Luise Niklew,† Ulrike Hochkirch,† Anna Melikyan,† Thomas Moritz,†,‡ Sandra Kurzawski,§ Hartmut Schlu¨ter,§ Ingo Ebner,† and Michael W. Linscheid*,† Department of Chemistry, Humboldt-Universitaet zu Berlin, Berlin, Germany, Analytical Laboratories, Atotech GmbH, Berlin, Germany, and Department of Clinical Chemistry/Central Laboratories, University Medical Center, Hamburg-Eppendorf, Germany The use of nanocrystalline titanium dioxide films as affinity targets for the selective isolation and enrichment of phosphopeptides with subsequent analysis by matrixassisted laser desorption ionization (MALDI) mass spectrometry is described. A strong affinity of phosphopeptides to anatase titanium dioxide surfaces is observed, and a standard protocol for the selective isolation and enrichment of phosphopeptides on titanium dioxide films using a proteolytic digest of r- and β-casein was developed. All washing and elution procedures using these films can be processed directly on the MALDI target, thereby avoiding sample contamination and losses. In addition, the enrichment of the phosphopeptides was improved due to a considerable enlargement of the surface. Several film substrates compatible with routine inlet systems of mass spectrometers, as conductive glass, aluminum, and silicon, have been manufactured and tested. A biological application was examined by the human fibrinogenthrombin system. For a quantification and comparison of different expression levels of phosphoproteins in biological systems, the peptides were labeled with S-methyl thioimidate reagents. The capability of this method for high-throughput applications make the use of mesoporous titanium dioxide films as an affinity MALDI target a promising tool in phosphoproteomics. A combination of an amidation protocol showed that a quantification of phosphorylated peptides can easily be performed using TiO2 films. Post-translational modifications of proteins are the means of cell regulation and enzyme functionality. Among them the reversible phosphorylation of hydroxyl groups is a principle for information transfer.1,2 The fact that a hundred different protein kinases

act in cancer makes their results, the phosphorylated proteins, even more interesting.3 Therefore, the detection of phosphorylated proteins is a major challenge. The enzymatic digestion of a protein extract generally required for its analysis results mostly in the generation of numerous peptides. The overwhelming number of peptides in combination with substoichiometric amounts of phosphorylated peptides decreases the chance of their detection significantly. Beside the fact of phosphorylation, the nature of the modified amino acid within the sequence is of interest, and the assignment of the phosphorylation site is a major challenge in protein analysis. As most proteomics approaches use mass spectrometry as the analytical tool, it offers several possibilities in the detection of posttranslational modifications, too.4,5 After a successful separation of the peptides, tandem mass spectrometry experiments can be performed. With the use of collision induced dissociation (CID), a loss of 98 Da indicates one phosphorylation site. Analyzing the sequence ions by switching into different modes like MS3 or different dissociation techniques (e.g., electron capture dissociation; ECD) the phosphorylation sites can be determined exactly.6,7 Thus, their isolation from their unphosphorylated peptide species (isoforms) and unmodified peptides and the concurrent enrichment is one goal in separation sciences.8-11 It is even more important as the signal intensity of a phosphopeptide in a mass spectrum is often suppressed relative to its isoform. Nonchromatographic methods include chemical modifications of the (2) (3) (4) (5) (6) (7)

* To whom correspondence should be addressed. M. W. Linscheid, Department of Chemistry, Humboldt-Universitaet zu Berlin, Brook-Taylor-Str. 2, D-12489 Berlin, Germany. Phone: +49(0)30 2093-7575. Fax: +49(0)30 2093-6985. E-mail: [email protected]. † Humboldt-Universitaet zu Berlin. ‡ Atotech GmbH. § University Medical Center. (1) Graves, J. D.; Krebs, E. G. Pharmacol. Ther. 1999, 82, 111–121. 10.1021/ac902403m  2010 American Chemical Society Published on Web 01/12/2010

(8) (9) (10) (11)

Hunter, T. Cell 2000, 100, 113–127. Blume-Jensen, P.; Hunter, T. Nature 2001, 411, 355–365. Aebersold, R.; Mann, M. Nature 2003, 422, 198–207. Mann, M.; Jensen, O. N. Nat. Biotechnol. 2003, 21, 255–261. Larsen, M. R.; Trelle, M. B.; Thingholm, T. E.; Jensen, O. N. Biotechniques 2006, 40, 790–798. Shi, S. D.; Hemling, M. E.; Carr, S. A.; Horn, D. M.; Lindh, I.; McLafferty, F. W. Anal. Chem. 2001, 73, 19–22. Mann, M.; Ong, S. E.; Gronborg, M.; Steen, H.; Jensen, O. N.; Pandey, A. Trends Biotechnol. 2002, 20, 261–268. Morandell, S.; Stasyk, T.; Grosstessner-Hain, K.; Roitinger, E.; Mechtler, K.; Bonn, G. K.; Huber, L. A. Proteomics 2006, 6, 4047–4056. Collins, M. O.; Yu, L.; Choudhary, J. S. Proteomics 2007, 7, 2751–2768. Schmidt, S. R.; Schweikart, F.; Andersson, M. E. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2007, 849, 154–162.

Analytical Chemistry, Vol. 82, No. 3, February 1, 2010

1047

phosphopeptides by phosphoamidation or β-elimination.12,13 Most separation procedures are based on the interaction of the phosphopeptides with the separation material, while nonphosphorylated compounds cannot bind and, therefore, can be easily removed. One of these strategies uses the biotin-streptavidin system for affinity chromatography, where the phosphate groups are linked to biotin.13,14 Since biotin must be covalently and, moreover, quantitatively bound, several chemical reactions have to follow each other. Thus, and according to the affinity chromatography, sample loss is unavoidable.12 Also anion exchange chromatography can be performed.15 A big field was opened with the specific affinity of metal oxide to phosphate. Different types of affinity chromatography were developed. Immobilized metal affinity chromatography (IMAC) is the most popular method for the selective enrichment of phosphopeptides, utilizing Fe3+, Ga3+, Al3+, or Zr4+ as complexing metal ions.16-20 Furthermore, zirconium dioxide, ZrO2, aluminum hydroxide, Al(OH)3, or titanium dioxide, TiO2 were used.21-25 Also the application of titania coated magnetic Fe3O4 nanoparticles was shown to be successful.26 The metal oxide particles exhibit a higher selectivity for enriching phosphopeptides than IMAC materials.21 Only acidic peptides show a strong affinity to the metal ions surface, too. They might compete with the phosphopeptides for the limited number of active places on the materials surface. This problem was successfully solved by a methylation of the peptide mixture before the chromatographic separation.27 Unfortunately, all these methods are time-consuming and a high-throughput analysis is hard to perform. Therefore, in this work, the development of titanium dioxide coated matrix-assisted laser desorption ionization (MALDI) targets is described. MALDI is known for its high-throughput approaches and the purification procedure can be performed directly on the target. Qiao et al. compared their MALDI approach with IMAC and achieved similar detection limits and selectivity.28 They spotted pretreated titanium dioxide nanoparticles onto and calcinated the MALDI target. Afterward the sample was pipetted on top. The captured phos(12) Zhou, H.; Watts, J. D.; Aebersold, R. Nat. Biotechnol. 2001, 19, 375–378. (13) Oda, Y.; Nagasu, T.; Chait, B. T. Nat. Biotechnol. 2001, 19, 379–382. (14) Qian, W. J.; Goshe, M. B.; Camp, D. G., 2nd; Yu, L. R.; Tang, K.; Smith, R. D. Anal. Chem. 2003, 75, 5441–5450. (15) Dai, J.; Jin, W. H.; Sheng, Q. H.; Shieh, C. H.; Wu, J. R.; Zeng, R. J. Proteome Res. 2007, 6, 250–262. (16) Corthals, G. L.; Aebersold, R.; Goodlett, D. R. Methods Enzymol. 2005, 405, 66–81. (17) Feng, S.; Pan, C.; Jiang, X.; Xu, S.; Zhou, H.; Ye, M.; Zou, H. Proteomics 2007, 7, 351–360. (18) Posewitz, M. C.; Tempst, P. Anal. Chem. 1999, 71, 2883–2892. (19) Li, Y.; Liu, Y.; Tang, J.; Lin, H.; Yao, N.; Shen, X.; Deng, C.; Yang, P.; Zhang, X. J. Chromatogr., A 2007, 1172, 57–71. (20) Feng, S.; Ye, M.; Zhou, H.; Jiang, X.; Jiang, X.; Zou, H.; Gong, B. Mol. Cell. Proteomics 2007, 6, 1656–1665. (21) Kweon, H. K.; Hakansson, K. Anal. Chem. 2006, 78, 1743–1749. (22) Wolschin, F.; Wienkoop, S.; Weckwerth, W. Proteomics 2005, 5, 4389– 4397. (23) Cantin, G. T.; Shock, T. R.; Park, S. K.; Madhani, H. D.; Yates, J. R., 3rd Anal. Chem. 2007, 79, 4666–4673. (24) Pinkse, M. W.; Uitto, P. M.; Hilhorst, M. J.; Ooms, B.; Heck, A. J. Anal. Chem. 2004, 76, 3935–3943. (25) Sano, A.; Nakamura, H. Anal. Sci. 2004, 20, 861–864. (26) Chen, C. T.; Chen, Y. C. Anal. Chem. 2005, 77, 5912–5919. (27) 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, 301–305. (28) Qiao, L.; Roussel, C.; Wan, J.; Yang, P.; Girault, H. H.; Liu, B. J. Proteome Res. 2007, 6, 4763–4769.

1048

Analytical Chemistry, Vol. 82, No. 3, February 1, 2010

phopeptides could be measured with MALDI-mass spectrometry (MS). Torta et al. used pulsed laser deposition (PLD) to build the TiO2 substrate onto a MALDI target.29 The layer in that work is more then 200 µm thick and is made of long, thin needles. This method lacks behind, as PLD is not available for everyone. The here described approach uses TiO2 nanoparticles synthesized via the sol-gel route for producing a mesoporous film mainly consisting of anatase titania. The chemistry to synthesize them is straightforward and nontoxic. The preconcentration protocol is optimized but can be adapted individually for every single separation problem. One important feature of these films is the increased internal surface area compared to a flat surface of the same diameter. With an increase in the internal surface area, an increased number of available binding sites for the phosphopeptides are created. The advantage of this methodology is that all sample treatments can be carried out directly on the MALDI target. Thereby, sample losses are reduced and no additional chemical reaction steps are necessary. The production of the affinity targets is reproducible, is low cost, and can be easily performed. EXPERIMENTAL SECTION Materials. Titanium(IV)isopropoxide (97%), ammonium dihydrogenphosphate, ammonium hydrogen carbonate, sodium hydrogen carbonate, tris(hydroxymethyl)-aminomethane, iodomethane, thioacetamide, R-cyano-4-hydroxycinnamic acid (HCCA), dithiothreitol (DTT, DL 99,5%) the phosphorylated proteins β-casein (bovine), R-casein (bovine), serum albumin (bovine, BSA) and human fibrinogen, protease bovine trypsin, elastase, and human thrombin were purchased from Sigma-Aldrich and were used without further purification. Thio-propionamide was supplied by TCI America. Acetonitrile, fetal serum fetuin (bovine), trifluoroacetic acid (TFA), acetic acid, formic acid, potassium-morpholinoethanesulphonic acid (MES), concentrated nitric acid (concentrated HNO3) and phosphoric acid were obtained from Merck. Isopropanol (p.a.) was purchased from J. T. Baker, and alkaline phosphatase was obtained from Roche. Iodacetamide (98%) was supplied by Acros. Synthesis of the Titanium Dioxide Nanoparticles. A volume of 47 mL of titanium tetraisopropoxide and 10 mL of isopropanol were mixed in a dropping funnel. The mixture was allowed to drop into a round-bottom flask containing 5 mL of concentrated HNO3 and 250 mL of water. The reaction was carried out under strong mixing at 0 °C during a time of 40 min. The resulting amorphous gel was slowly heated, isopropanol was removed by a dehydrator, and the gel was refluxed for 8 h. A total of 40 mL of the resulting titanium dioxide gel was put into a 60 mL Teflon inlet and was autoclaved in a circulating air oven for 18 h at 180 °C. After it was cooled down, the gel was treated for 10 min in an ultrasonic bath and was stored with mixing in the dark. The size of the nanoparticles was 5-20 nm, and the crystalline phase was anatase with traces of brookite. Nanocrystalline Titanium Dioxide Films. The film was structured using screen-printing and spraying techniques on SnO2:F conductive glass (Asahi Glass Co., Ltd.). Therefore, a fixation was required and done in two different ways: (I) The (29) Torta, F.; Fusi, M.; Casari, C. S.; Bottani, C. E.; Bachi, A. J. Proteome Res. 2009, 8, 1932–1942.

TiO2 suspension was spread on the glass slide with a mask made out of an adhesive tape and was smeared with a flexible spatula. (II) The homemade airbrush device was used to generate a fine spray of TiO2 suspension diluted with water (1:1); an alumina mask with well-ordered pores was used to fabricate 1 mm film spots.30 The adhesive tape or alumina mask was removed after drying for 30 min, and the substrate slide was calcinated at 400 °C for 60 min. Afterward the nanoparticles were fixed on the surface resulting in a mesoporous film with spot sizes of 1-3 mm. The characterization was performed by scanning electron microscopy (SEM, Zeiss, S360) and interference spectroscopy (Atos, Micromap 512). Digestion of Proteins. For proteolysis, 100 µL of the protein stock solutions (β-casein (4.2 mg/mL), R-casein (3.3 mg/mL)) were mixed with 100 µL of trypsin dissolved in 5 mM ammonium hydrogen carbonate to give a mass ratio of 20:1. Human fibrinogen (20 mg/mL) was digested with human thrombin (5 units) dissolved in 50 mM MES. A total of 100 µL of bovine fetuin stock solution (0.3 mg/mL) was mixed with an equal volume of 20 mM DTT solution. After 2 h reduction time, 50 µL of 100 mM iodacetamide solution was added. Reaction mixtures were allowed to stand at ambient temperature for 2 h in the darkness. Digestions were incubated for 24 h at 37 °C. R-Casein was treated with elastase (1 mg/mL) in a substrate-enzyme ratio of 40:1 at 47 °C for 18 h. For all experiments aliquots of each digest solution were stored at -20 °C. Dilution was done with 5 mM ammonium hydrogen carbonate. Preconcentration Protocol. Typically, 2 µL of a 15 pmol/µL tryptic digest were spotted on the affinity target. After 20 min, the unbound peptides were removed with three to five washing steps using 100 mM acetic acid. The incubation time varied from 15 to 45 min. The washing procedure was adapted to the individual analytical problem. Afterward the target was rinsed with water. To prevent the evaporation of the droplet for longer incubation times, the target was stored in a box with a saturated water atmosphere. Alternatively to the droplet adsorption, the target was immersed in the digest solution. The immersion time was varied from 15 min to several hours. Phosphopeptide Desorption Protocol. After the washing procedure, 1.5 µL of 100 mM ammonium dihydrogenphosphate solution was added on the titanium dioxide film. After release of the phosphopeptides from the titanium dioxide surface, the matrix solution (HCCA; 10 mg/mL in 1/1 acetonitrile/1% formic acid) was added. Alternativly the matrix solution contained phosphoric instead of formic acid, then the additional treatment with ammonium dihydrogenphosphate is not necessary. After the target was dried, it was introduced into the mass spectrometer. On Target Alkaline Phosphatase Treatment. After the surface was washed three times using 100 mM acetic acid, 2 µL of an alkaline phosphatase solution (10 mg/mL, diluted 1:1000 in 50 mM ammonium dihydrogen carbonate) was spotted on the target and was stored 18 h in a moisture chamber. Afterward the indigested, still phosphorylated tryptic peptides were released according to the desorption procedure. Amidation. Synthesis of S-methyl thioacet imidate, S-methyl thiopropion imidate, and amidation of tryptic peptides were (30) Miliotis, T.; Kjellstrom, S.; Nilsson, J.; Laurell, T.; Edholm, L. E.; MarkoVarga, G. Rapid Commun. Mass Spectrom. 2002, 16, 117–126.

performed according to Beardsley.31 Amidation was performed by mixing β-casein digest aliquots and solutions of the reagents S-methyl thioacet imidate or S-methyl thiopropion imidate dissolved in 250 mM tris(hydroxymethyl)-aminomethane. Mass Spectrometry. MALDI time-of-flight (TOF) mass spectrometry was performed on a Voyager DE (Applied Biosystems, Darmstadt, Germany). For MS/MS experiments, a LCQ Deca XP (Thermo Finnigan, Bremen, Germany) with an atmospheric pressure (AP) MALDI ion source (GSG Mess-and Analysegera¨te, Bruchsal, Germany) was used. The MALDI targets of both instruments were modified for the substrates covered with the mesoporous titanium dioxide films. The MALDI parameters were standard peptide analysis parameters (voltage 20 kV, grid 94%, extraction delay time 50 ns, laser intensity 1500) or for AP MALDI the laser frequency was 10 Hz, capillary voltage 2.7 kV, and the transfer capillary temperature 270 °C. RESULTS Characterization of the Nanoparticle Titanium Dioxide. The synthesis of the titanium dioxide nanoparticles was done by means of sol-gel chemistry using the method of Anderson et al. and the modification by Zaban et al.32,33 The nanocrystalline films were produced in two different ways: the screen-printing or spraydeposition technique. Both methods resulted in a homogeneous deposition of the nanoparticles and gave no significant difference regarding the mass spectrometric results. The spraying-techique enables a faster production of the TiO2 films. The spots were characterized using scanning electron microscopy and interference spectroscopy. Figure 1A shows a SEM picture of the nanocrystalline titanium dioxide film coated SnO2-F glass surface. Obviously, the surface is inhomogenous and shows an enhanced area per square unit, as particles in the third dimension are visual. Figure 1B shows the unmodified surface of the substrate (SnO2-F). A height profile scan by profilometer is seen in Figure 1C. It shows a 4 µm thick TiO2 film spot produced using the screen-printing technique (d ) 1 mm) on a silicon substrate. The average layer thickness is about 3.64 µm (maximum 6.5, minimum 1.4 µm) and it is expected that this roughness leads to a better adsorption capacity. Analysis of the Phosphorylation Sites of Bovine β-Casein. As the detection of phosphopeptides out of numerous peptides is requested, a tryptic digest of the well characterized protein bovine β-casein with five phosphorylation sites was chosen. The tryptic digest was spotted, incubated, and washed following the optimized procedure (see Experimental Section). To release the adsorbed phosphorylated peptides, an addition of ammonium dihydrogenphosphate (NH4H2PO4) was necessary. No phosphopeptides were detected when only the matrix solution was used. When spotting the matrix solution on the films before the addition of NH4H2PO4 solution, the color changed to deep yellow indicating the formation of chelate complexes of the matrix molecules with Ti4+ centers. A variation of the pH of the NH4H2PO4 buffer solution from 3 to 10 had no significant effect on the surface desorption efficiency. The inorganic phosphate (31) Beardsley, R. L.; Reilly, J. P. J. Proteome Res. 2003, 2, 15–21. (32) Xu, Q. Y.; Anderson, M. A. J. Mater. Res. 1991, 6, 1073–1081. (33) Zaban, A.; Ferrere, S.; Sprague, J.; Gregg, B. A. J. Phys. Chem. B 1996, 101, 55–57.

Analytical Chemistry, Vol. 82, No. 3, February 1, 2010

1049

Figure 1. Characterization of TiO2 films produced by the screen-printing technique. (A) SnO2-F glass coated with nanocrystalline TiO2 film. (B) SnO2-F glass (parts A and B with a magnification of 50000). (C) Profile scan of one spot by interference spectroscopy and illustrating the average layer thickness of that spot.

Figure 2. MALDI TOF mass spectra of 1.5 pmol of β-casein after the enrichment and desorption procedure on TiO2 films produced by the screen-printing technique (phosphopeptides are marked with an *). (A) The monophosphorylated peptides at m/z 2063 (48FQpSEEQQQTEDELQDK-63), at m/z 2558 (-KIEKFQpSEEQQQTEDELQDK-63), and the 4-fold phosphorylated peptide at m/z 3124 (16-RELEELNVPGEIVEpSLpSpSpSEESITR-40) could be detected without further treatment. (B) After the enzymatic reaction with phosphatase, the detected m/z value of the monophosphorylated peptides shifted about 80 Da (m/z 2063 to m/z 1983 and m/z 2558 to m/z 2477), and the signal at m/z 3124 shifted to m/z 2924 (2 × -80 Da) indicating the loss of one or two phosphoric acid sites. Arrows show the loss of HPO3.

is strongly bound and replaces all other ligands of the titanium centers, as phospho- and acidic peptides. Adding the matrix solution after debonding of the phosphopeptides with the NH4H2PO4 solution no color change was observed. This indicates that after the release of phosphopeptides from the titanium films, all active surface sites are occupied by phosphate anions. Attempts to reactivate the membrane failed due to the strong affinity of phosphate to titanium. Figure 2A shows the spectrum of 1.5 pmol of β-casein digest after the specific adsorption of the phosphopeptides on the titanium dioxide affinity target. Three major peaks with m/z 2063, 2558, and 3124 were observed. The calculated average masses for the β-casein phosphopeptides are m/z 2063 for 48-FQpSEEQQQTEDELQDK-63, m/z 2562 for 1050

Analytical Chemistry, Vol. 82, No. 3, February 1, 2010

44-KIEKFQpSEEQQQTEDELQDK-63, and m/z 3124 for the 4-fold phosphorylated 16-RELEELNVPGEIVEpSLpSpSpSEESITR-40, respectively. The ions at m/z 2558 and m/z 3124 can be identified as phosphopeptides immediately due to their metastable loss of phosphoric acid (- 98 Da). In addition, two minor peaks at m/z 3469 and 3501 were found. For further proof an on-target alkaline phosphatase treatment was performed. The enzymatic reaction was done before the desorption of the phosphopeptides from the titanium dioxide surface. The enzyme is expected to have less access to strongly bound phosphopeptides, and the digested peptides occur beside their phosphorylated species. This is due to the equilibrium of the adsorption process, some peptides are released and can be digested subsequently. After phosphatase treatment (Figure 2B), the peak at m/z 2063 shifted to m/z 1983 (- 80 Da) and the peak at m/z 2558 shifted to 2477 (- 80 Da) indicating monophosphorylated peptides as one loss of HPO3 occurred. The peak at m/z 3124 shifted to m/z 2924 (- 160 Da) indicating a diphosphorylated peptide. The theoretical loss of 360 Da corresponding to four phosphorylation sites was not observed. A complete dephosphorylation is probably blocked by the strong attachment of the four neighboring phosphate groups to the titanium surface. No peak shifts were observed for m/z 3468 and m/z 3499. This indicated no phosphorylation within these peptides. They probably contain acidic amino acids interacting with the titanium surface and could not be removed during the washing process. For the detection of phosphopeptides in the subpicomole range using nanocrystalline titanium dioxide films, 2 µL of 1.5 and 0.15 pmol/µL of the solution were spotted on the affinity target and treated following the standard protocol. The monophosphopeptide with m/z 2063 was detectable down to 300 fmol (data not shown). Phosphopeptide Screening in Complex Mixtures. To prove the screening capacity of the developed nanocrystalline titanium dioxide films, a mixture of the tryptic digest of BSA and R- and β-casein were investigated using MALDI-MS. A volume of 1 µL of the mixture containing 15 pmol of BSA digest and 4 pmol of each, the R- and β-casein digests were spotted on the titanium dioxide film. After the preconcentration protocol, the complexity of the spectrum is reduced dramatically (see Figure 3A). The phosphopeptides (in the spectrum marked with an *) of R-casein with average masses of 1951 and 2926 Da and β-casein with their average masses of 2061 and 2557 Da were enriched

Figure 3. MALDI TOF spectra of a peptide mixture after the enrichment and desorption procedure on TiO2 films produced by the spraydeposition technique (phosphopeptides are marked with an *). (A) Tryptic digest containing BSA (15 pmol), R-casein (4 pmol), and β-casein (4 pmol). The phosphopeptides of R-casein at m/z 1952 and 2927 and β-casein at m/z 2062 and 2558 were selectively enriched. Some acidic not phosphorylated peptides of BSA could be observed as well. (B) Tryptic fragments of bovine fetuin (0.3 mg/mL) after 12 h enrichment. The monophosphorylated peptide p(HTFSGVASVESSSGEAFHVGK) was detected at m/z 2201 and its diphosphorylated isoform at m/z 2281 in low intensity. (C) Proteolysis of fibrinogen with thrombin after 2 h incubation (60 pmol). The signal at m/z 1617 represent the phosphofibrinopeptide A (1-ADpSGEGDFLAEGGGVR-16) from the AR chain of human fibrinogen. Peaks of not phosphorylated fibrinopeptide B and des-Glu1fibrinopeptide B (m/z 1553 and m/z 1442) from the Bβ chain are observed as well.

selectively. Several not phosphorylated peptides originating from BSA could be observed. These peptides, e.g., with the average mass of 2105 Da (TVMENFVDCCAADK) contain acidic amino acids such as glutamic and aspartic acid. These have also a strong affinity to the titanium dioxide surface.27 This has been also observed using other titanium dioxide based approaches. Converting the carboxylic acids to their methyl esters was successful to improve the selectivity with IMAC. Preliminary experiments showed that this also prevents acidic peptides from nonspecific binding on nanocrystalline titanium dioxide films. Furthermore, lactic acid, which is too small to be detected, could be used to result in an enhanced selectivity of these methods. The reason might be a competition of the small organic acid with the acidic peptides for the free binding places on the titanium dioxide surface. Analysis of Bovine Fetuin. Beside the releasing procedure by the addition of NH4H2PO4, phosphoric acid can also be added to the matrix solution. This provides for easier handling and was demonstrated for bovine fetuin.34,35 For the enrichment of phosphopeptides from protein digests with low phosphorylation grades, an optimized phosphopeptide enrichment protocol was developed (see Experimental Section). According to this protocol, desorption of the phosphopeptides from the titanium dioxide surface was realized by spotting 1-2 µL of the matrix solution containing 1% phosphoric acid. This was to replace the bound phosphopeptides with inorganic phosphate and to enhance the detection of phosphopeptides in MALDI-MS. For this purpose, bovine fetuin with an unknown phosphorylation grade was investigated. The sites of phosphorylation of bovine fetuin are well characterized.36 There are four phosphorylation sites; two of them in the same peptide. For the mass spectrometric analysis the tryptic digest of reduced and alkylated bovine fetuin was allowed to equilibrate for 12 h at 20 °C on the TiO2 film. Mainly the (34) Kjellstrom, S.; Jensen, O. N. Anal. Chem. 2004, 76, 5109–5117. (35) Stensballe, A.; Jensen, O. N. Rapid Commun. Mass Spectrom. 2004, 18, 1721–1730. (36) Wind, M.; Gosenca, D.; Kubler, D.; Lehmann, W. D. Anal. Biochem. 2003, 317, 26–33.

monophosphorylated peptide p(HTFSGVASVESSSGEAFHVGK) was detected. The diphosphopeptide with the identical amino acid sequence was only observed in low intensity. Other phosphopeptides were not observed (see Figure 3B). In this study it was demonstrated that the phosphopeptides enrichment is a time dependent equilibration process. In samples with small amounts of phosphopeptides, the peptides containing acidic amino acids interact with the TiO2 surface first and block the adsorption sites for phosphate groups. With adsorption time up to several hours, the phosphopeptides increasingly replace the unspecifically bound peptides on the TiO2 surface. Human Fibrinogen as a Biologically Relevant Application. Fibrinogen plays a central role in blood clotting and a variety of other processes in the extracellular space. It is composed of three peptide chains AR, Bβ, and γ, which occur paired in fibrinogen (AR)2(Bβ)2γ2. The six chains are covalently linked near their N-terminals through disulfide bonds. The A and B portions of the AR (66 kDa) and Bβ (52 kDa) chains comprise the fibrinopeptides A and B, respectively. Thrombin hydrolyses fibrinogen at four arginine-glycine bonds between the fibrinopeptide and the A and B portions of the protein during clotting.37 Thrombin-mediated release of the fibrinopeptides generates fibrin monomers, which spontaneously aggregate forming the fibrin clot. The covalent modifications of fibrinogen are well characterized: the phosphorylations are localized at serine-3 in fibrinopeptide A and serine345 of the AR chain.37,38 This model system with a phosphorylation grade of 20-30% was chosen for the detection of low level phosphopeptides as an application of the nanocrystalline TiO2 affinity targets.36,37 Schlu¨ter et al. described an approach for the analysis of fibrinogen phosphopeptides using a new mass spectrometry assisted enzyme screening system.39,40 For the detection of the phosphorylation in fibrinogen, specifically in fibrinopeptide A as the interactive peptide to thrombin, (37) Maurer, M. C.; Peng, J. L.; An, S. S.; Trosset, J. Y.; Henschen-Edman, A.; Scheraga, H. A. Biochemistry 1998, 37, 5888–5902. (38) Mosesson, M. W.; Siebenlist, K. R.; Meh, D. A. Ann. N.Y. Acad. Sci. 2001, 936, 11–30.

Analytical Chemistry, Vol. 82, No. 3, February 1, 2010

1051

human fibrinogen was incubated with human thrombin for 2 h. Afterward the digest solution was spotted on TiO2 film and was treated according to the enrichment protocol (see Figure 5). After application of the enrichment strategy, the phosphorylated fibrinopeptide A (1-ADpSGEGDFLAEGGGVR-16) with m/z 1617 was observed as the main peak. Signals of the not phosphorylated fibrinopeptide B and des-gln1-fibrinopeptid B (m/z 1553 and m/z 1442) from Bβ chain were observed. These peptides, e.g., desgln 1-(QGVNDNEEGFFSAR)-14 at m/z 1442 contained several aspartic and glutamic acids residues, which interact with the titanium surface. The same experiment was done with the tryptic digest solution. Here also the main peak at m/z 1617 was identified as phosphofibrinopeptide A by ions resulting from the loss of phosphoric acid (- 98 Da) (data not shown). With the fibrinogenthrombin system it could be shown that the enrichment strategy with nanocrystalline TiO2 films can be used for the analysis of peptides in complex biological systems. Localization of the Phosphorylation Sites. For the localization of the phosphorylation sites MS/MS experiments were performed using an AP MALDI ion source coupled to a LCQ ion trap.41,42 The resulting tandem mass spectrum of a R-casein digest with elastase after the enrichment procedure can be seen in Figure 4. As elastase is a less specific enzyme than trypsin smaller peptides can be obtained. This is requested due to the smaller mass range of the ion trap compared to the MALDI TOF. The fragmentation was performed by collision induced dissociation (CID). The main product ion results out of the loss of phosphoric acid at m/z 1488. The characteristic sequence fragments are y

and b ions which are indicated in the MS/MS spectrum.43 The phosphopeptide was identified by de novo sequencing as (123LELVPPNpSAEERLH-136) with a phosphorylation at the serine 130. Thus, by the combination of a selective enrichment of phosphopeptides using titanium affinity MALDI targets and peptide sequencing by tandem mass spectrometry, a localization of the phosphorylation site is possible. Quantification of Amidated Phosphopeptides. Quantification by MALDI-MS requires the production of differentially labeled peptides from two sets of proteins.44 The difference in amounts can be determined by measuring the relative intensities of the MS signals of the differently labeled peptides. Various strategies have already been described.45 Beardsley et al. demonstrated that a differential amidation of N-termini and ε-amino groups of lysine residues performed with S-methyl thioimidate reagents provides the basis for a novel approach to protein quantification using MALDI-MS.31 The amidation of lysine increases its basicity and ionization yields by MALDI TOF MS. Combining this technique with the here described preconcentration strategy on the TiO2 films results in an efficient approach for relative quantification of phosphoproteins using the two closely chemically related reagents S-methyl thioacetimidate (SMTA) and S-methyl thiopropionimidate (SMTP) for labeling. In order to get reliable results, it had to be shown that the amidation of the phosphopeptides will not affect the enrichment behavior on the TiO2 film. Internal standardization techniques are difficult to apply to this enrichment method because phosphopeptides added as an internal standard compete with the phosphopeptides to be determined for the limited number of binding sites. This may result in suppression of their adsorption.14,46-48 However, with the strategy applied here, the comparison of phosphoprotein levels in two different states is possible, since both derivatives bind with the same probability. Thus, the difference in signal intensities reflects the amount of differentially labeled tryptic peptides. To achieve the homogeneous disposition of nanoparticles within a film spot and increase the signal reproducibility on the sample spot surface, the spray technique was used. Thus, the relative standard deviation of the signal intensity between spots could be reduced to 5-14% (n ) 10). Amidation was performed by mixing β-casein digest aliquots and solutions of the reagents S-methyl thioacetimidate or S-methyl thiopropionimidate (for the chemical reaction see Figure 5A). The tagging occurred at the amino-terminus of peptides and at lysine as indicated by the addition of 41 or 55 Da for each amino group, respectively (see Figure 5B-D). Those peptides containing a lysine reacted twice, and the masses increased by 82 or 110 Da. The reagents are water-soluble, and the reactions are rapid and quantitative under mild conditions. Aliquots of the acidified reaction mixtures and, for comparison, an unmodified tryptic digest solution of β-casein were spotted on the nanocrystalline

(39) Schlu ¨ ter, H.; Jankowski, J.; Rykl, J.; Thiemann, J.; Belgardt, S.; Zidek, W.; Wittmann, B.; Pohl, T. Anal. Bioanal. Chem. 2003, 377, 1102–1107. (40) Schlu ¨ ter, H.; Rykl, J.; Thiemann, J.; Kurzawski, S.; Gobom, J.; Tepel, M.; Zidek, W.; Linscheid, M. Anal. Chem. 2007, 79, 1251–1255. (41) Laiko, V. V.; Taranenko, N. I.; Berkout, V. D.; Musselman, B. D.; Doroshenko, V. M. Rapid Commun. Mass Spectrom. 2002, 16, 1737–1742. (42) Laiko, V. V.; Taranenko, N. I.; Berkout, V. D.; Yakshin, M. A.; Prasad, C. R.; Lee, H. S.; Doroshenko, V. M. J. Am. Soc. Mass Spectrom. 2002, 13, 354– 361.

(43) Roepstorff, P.; Fohlman, J. Biomed. Mass Spectrom. 1984, 11, 601. (44) Delcourt, N.; Jouin, P.; Poncet, J.; Demey, E.; Mauger, E.; Bockaert, J.; Marin, P.; Galeotti, N. Mol. Cell. Proteomics 2005, 4, 1085–1094. (45) Moritz, B.; Meyer, H. E. Proteomics 2003, 3, 2208–2220. (46) Mayya, V.; Rezual, K.; Wu, L.; Fong, M. B.; Han, D. K. Mol. Cell. Proteomics 2006, 5, 1146–1157. (47) Gutierrez, J. A.; Dorocke, J. A.; Knierman, M. D.; Gelfanova, V.; Higgs, R. E.; Koh, N. L.; Hale, J. E. Biotechniques 2005, 13–17 (Suppl.). (48) Chakraborty, A.; Regnier, F. E. J. Chromatogr., A 2002, 949, 173–184.

Figure 4. AP MALDI QIT MS/MS spectra of the molecular ion of the R-casein phosphopeptide (123-LELVPNpSAEERLH-136) at m/z 1586 after the enrichment on the TiO2 film target produced by the spray-deposition technique. The main product ion at m/z 1488 [M + H - 98]+ is due to the loss of phosphoric acid. The sequence ions are indicated within the peptide sequence: b ions below and y ions above the letter code.

1052

Analytical Chemistry, Vol. 82, No. 3, February 1, 2010

Figure 5. Amidation of tryptic peptides and MALDI TOF analysis of β-casein tryptic peptides after enrichment on the TiO2 film produced by the spray-deposition technique (phosphopeptides are marked with an *). (A) Chemical formula of the amidation process for SMTA labeling. (B-D) Phosphopeptides MALDI TOF analysis (B) after labeling with SMTP to give a mass shift of 55 per free amino group, resulting signal at m/z 2173 (# indicates the labeling site). (C) Without labeling at m/z 2063. (D) After labeling with SMTA to give a mass shift of 41 per free amino group, resulting signal at m/z 2145 (§ indicates the labeling site). (E) Average intensity ratio of 1:1, 1:2, 1:3, 1:4, and 1:5 mixtures of labeled peptides from tryptic digest of β-casein (n ) 10). The intensity ratio for SMTP/SMTA-labeled peptides of β-casein is plotted as a function of the SMTP/ SMTA-labeled peptides concentration ratio. The relative standard deviations were 15%, 35%, 5%, 20%, and 12%, respectively.

titanium dioxide film. Figure 5B-D shows the MALDI TOF mass spectra after treating the affinity target according to the enrichment protocol. The signal for the β-casein peptide FQpSEEQQQIEDELQDK containing both lysine residues and amino terminus shifted after the modification reaction with SMTP from m/z 2062 to m/z 2172 (Figure 5B) and for the labeling with SMTA to m/z 2145 (Figure 5D). The amidation has no influence on the adsorption behavior of the phosphopeptides. Since no unreacted peptides were detected, the derivatization reaction is quantitative. Beardsley et al. already described that the ionization efficiencies of methyl- and ethyl-coded peptides are very similar and there is no significant bias between differentially amidated peptides. Then, the similar intensities are consistent with comparable sample handling losses for methyl- and ethyl-coded peptides. Five SMTP and SMTA modified β-casein samples were mixed in ratios of 1:1 to 1:5 to generate a calibration curve. A total of 10 spectra of 1 single spot were acquired for each sample with 100 laser shots/spectrum. The intensity ratio for SMTP/ SMTA-labeled peptides of β-casein was plotted as a function of SMTP/SMTA-labeled peptides concentration ratio (see Figure 5E). The data points represent the determined ratios of methylversus ethyl-coded peptides while the line displays the expected trend. The relative standard deviations were 15%, 35%, 5%, 20%, and 12%, respectively. It is evident that the quantification of phosphopeptides using the amidation method and the enrichment on TiO2 affinity targets is a good realizable combination.

CONCLUSIONS In this study it was shown that mesoporous TiO2 films are excellent tools for the selectively immobilization and enrichment of phosphopeptides in proteolytic digests. They can be produced for low cost by sol-gel synthesis of nanocrystalline anatase titania with a subsequent calcination step. The films showed an excellent physical stability, high internal surface area, and a typical thickness of 4 µm. By the means of model proteins such as R- and β-casein and bovine fetuin, the robustness of the developed preconcentration protocol was shown. Also the low level phosphorylated products of the fibrinogen-thrombin interaction were investigated. It could be shown that phosphorylated peptides can be enriched on TiO2 films after various reaction times and at low concentration levels down to 300 fmol on target. Printing the TiO2 films on specifically made silicon substrates, they could be converted into standard MALDI target formats of robust design and small spot sizes offering superior detection capabilities. It was also shown that TiO2 MALDI affinity targets are compatible with the amidation method for relative quantification of different phosphorylation states. The next step will be the use of amidelabeled phosphopeptides for a comparable determination of the relative expression levels of phosphoproteins in complex biological systems. Received for review December 22, 2009.

October

23,

2009.

Accepted

AC902403M

Analytical Chemistry, Vol. 82, No. 3, February 1, 2010

1053