New Tools for Quantitative Phosphoproteome Analysis

Biochemical and Biophysical Research Communications 290, 885– 890 (2002) doi:10.1006/bbrc.2001.6275, available online at http://www.idealibrary.com on

BREAKTHROUGHS AND VIEWS New Tools for Quantitative Phosphoproteome Analysis Thomas P. Conrads, Haleem J. Issaq, and Timothy D. Veenstra 1 Analytical Chemistry Laboratory, SAIC–Frederick, National Cancer Institute at Frederick, Frederick, Maryland 21702

Received December 11, 2001

Recent advances in analytical methods, particularly in the area of mass spectrometry, have brought the field of proteomics to the forefront in biological science. The ultimate goal of proteomics—to characterize proteins expressed within a cell under a specific set of conditions—is daunting due to the complexity and dynamic nature the of protein population within the cell. While much of the effort has focused on developing methods to identify expressed proteins, the identification of posttranslational modifications is equally important for comprehensive proteome characterization. Of all the known posttranslational modifications, phosphorylation arguably plays the largest role in the context of cellular homeostasis. This review discusses some of the recent progress made in the development of techniques not only to identify, but also to quantitatively determine sites of phosphorylation. Key Words: proteomics; phosphorylation; phosphopeptide; mass spectrometry; stable-isotope labeling; affinity tag.

The success of the various genome sequencing projects (1) and the development of techniques to measure differences in gene expression at the transcription level (2) have shifted considerable focus toward the development of advanced methods aimed at proteomics—the characterization of gene expression at the protein level (3). Most of the initial efforts in proteomics have focused on methods to effectively identify large numbers of proteins rapidly (4). Significant advances in sample preparation and analytical methods now enable effective identification and determination of relative protein abundances (5). Delineation of protein function solely from abundance changes, however, provides a limited view of the proteome since numer1 To whom correspondence and reprint requests should be addressed at SAIC–Frederick, National Cancer Institute at Frederick, P.O. Box B, Building 469, Room 160, Frederick, MD 21702. Fax: 301-846-6037.

ous vital activities of proteins are modulated by posttranslational modifications (PTMs) that may not be accurately reflected solely by changes in protein abundance. One of the most important posttranslational protein modifications used to modulate protein activity and propagate signals within cellular pathways and networks is phosphorylation (6). Cellular processes ranging from cell cycle progression, differentiation, development, peptide hormone response, and adaptation are all regulated by protein phosphorylation. Quite often regulation of protein function by phosphorylation occurs without a change in the protein’s abundance (7). While mass spectrometric (MS)-based methods have played a leading role in the identification of phosphorylated proteins, they are generally best suited for single proteins or simple mixtures of phosphoproteins. Due in large part to their typical low abundance, combined with the relatively narrow dynamic range of present analytical technologies, the detection of phosphopeptides in a complex mixture containing all other types of peptides is a difficult task. The ability to enrich for phosphopeptides in complex samples would greatly simplify the identification of these species. The predominant method used to study protein phosphorylation employs protein radiolabeling with 32P inorganic phosphate ( 32P i). To measure differences in relative abundances of phosphorylation, 32P-labeled proteomes are resolved by two dimensional polyacrylamide gel electrophoresis (2-D PAGE) and the relative spot intensities are compared (8). The use of 32P i to label proteins does not lend itself to high-throughput proteome-wide analysis due to safety issues with handling radioactive compounds and the associated contamination of analytical instrumentation. Fortunately, new methods that involve modifying phosphoproteins with affinity tags in combination with stable isotope incorporation (9 –11) have been developed which specifically enrich for phosphopeptides and allow for their subsequent identification and quantitation by MS. The principles underlying these new methods and their po-

885

0006-291X/02

Vol. 290, No. 3, 2002

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

tential usefulness in characterizing phosphorylated proteins is the subject of this minireview. IDENTIFICATION AND QUANTITATION OF PHOSPHORYLATED PEPTIDES One of the primary difficulties in identifying phosphopeptides in complex mixtures revolves around the difficulties associated with enriching samples for these species. While phospho-specific antibodies and metalaffinity columns are widely used, they typically result in isolation of nonphosphorylated species along with the phosphopeptides of interest. In addition, both of these methods suffer from the inability to quantitatively determine the relative phosphorylation states of proteins isolated from different sources. Fortunately two new methods have been developed that provide for the specific enrichment and quantitation of phosphopeptides. Both methods incorporate stable isotopes to differentially label the samples to be compared and employ subsequent MS analysis for the identification and quantitation of the enriched phosphopeptide mixture. The first strategy to isolate and quantitate phosphopeptides we will discuss was developed concurrently, and independently, by two groups (9, 10). While there are subtle differences in the specific procedures, the overall approaches of both methods are similar. The generalized reaction scheme illustrating the labeling of the phosphoseryl (pSer) and phosphothreonyl (pThr) residues of phosphoproteins by this methodology is outlined in Fig. 1. The first step involves blocking reactive thiolates of cysteinyl residues via reductive alkylation or performic acid oxidation. In the next step, phosphate moieties are removed via hydroxide ionmediated ␤-elimination from the pSer and pThr residues, resulting in their conversion to dehydroalanyl and ␤-methyl-dehydroalanyl residues, respectively. The newly formed ␣,␤-unsaturated double bond renders the ␤-carbon in each of the newly chemically modified residues sensitive to nucleophilic attack. The next modification involves a Michael-type addition of the bifunctional reagent 1,2-ethanedithiol (EDT). The addition of EDT to either the dehydroalanyl or ␤-methyldehydroalanyl residues results in the creation of a free thiolate in place of what was formerly a phosphate moiety, which can now serve as a reactive site susceptible to covalent modification by iodoacetyl-PEO-biotin. The end result is the covalent modification of phosphoryl residues with a linker molecule that contains a terminal biotin group. The stable isotopic labeling that enables relative quantification is achieved by using commercially available sources of either light (HSCH 2CH 2SH, EDT-D 0) or heavy (HSCD 2CD 2SH, EDT-D 4) isotopomeric version of EDT. Upon modification, the samples are subsequently digested with tryp-

sin (or another proteolytic enzyme or chemical cleavage methodology) and the modified peptides are specifically extracted using immobilized avidin chromatography, taking advantage of the high affinity biotin/avidin interaction. These extracted peptides are analyzed using reversed-phase liquid chromatography (LC) coupled directly on-line with MS. Two experimental MS strategies (MS and tandem MS) are used to identify and quantify the phosphorylation state of the phosphopeptides. In the MS mode, the masses of the intact peptide pairs are measured and the relative signal intensities provide a direct measure of the peptide’s phosphorylation status. The MS signals originating from the modified versions of the phosphopeptides are easily recognizable since they occur as pairs separated by the mass difference between the EDT-D 0 and EDT-D 4 labels (i.e., 4.0 Da). The mass spectrum of the phosphorylated peptide FQS PEEQQQTEDELQDK from ␤-casein in which the pSer residue has been modified with EDT (D 0 or D 4) and iodoacetyl PEO-biotin is shown in Fig. 2A. The 2.01 m/z difference between the doubly charged [M ⫹ 2H] 2⫹ ions at 1236.96 and 1238.97 m/z, corresponds to a 4.02-Da mass difference. Another key attribute of this isotopic chemical modification strategy to identify phosphopeptides is that the modification remains attached to the residue during tandem MS fragmentation of the peptide. During tandem MS the intact peptide is subjected to collisional-induced dissociation (CID) where the ion selected collides with an inert gas (nitrogen or helium, for example) that causes it to fragment into smaller ions. The MS spectrum of these fragment ions typically provides partial sequence information that can be used in conjunction with commercially available computer algorithms to identify the peptide. In a typical MS experiment, however, the phosphate group dissociates from the phosphorylated residues during the tandem MS analysis (indeed, often even in MS analyses as well) preventing the site-specific assignment of the phosphate modification. The isotope labeling strategy described above, however, allows the exact phosphorylation site to be determined by tandem MS as shown in Fig. 2B. In this example the CID spectra for the EDT-D 0 and D 4 modified versions of a peptide are 2⫹ 2⫹ shown. The mass difference between the y 13 and y 14 daughter ions is equal to the mass of a seryl residue modified as described above. While there is at least one other possible site of phosphorylation within this peptide, the CID spectrum clearly identifies it as the seryl residue. One of the novel attributes of the labeling strategy described above is its ability to quantify the relative phosphorylation of a peptide from two different samples (9). As an illustration, several stoichiometric amounts of ␤-casein were labeled as described in Fig. 1. As shown for a modified phosphopeptide in Fig. 3, the

886

Vol. 290, No. 3, 2002

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

FIG. 1. Isotope affinity strategy for isolating and quantify phosphopeptides. Proteins containing phosphoseryl (X ⫽ H) or phosphothreonyl (X ⫽ CH 3) residues are modified with reagents containing both an isotopically labeled linker and biotin group. After proteolytic digestion, these modified peptides are isolated from using immobilized avidin affinity chromatography. Light (L ⫽ H, EDT-D 0) and heavy isotopic versions (L ⫽ D, EDT-D 4) of 1,2-ethanedithiol (EDT) are used to quantitate the relative phosphorylation state of phosphopeptides extracted from two different sources.

ratios integrated for each mass spectrum correlate well with the stoichiometric concentrations ratios of the 2:1, 5:1, and 10:1 used in the labeling experiment. The two protein samples are pooled after the isotopic labeling step, therefore any sample loss occurring after this step is shared equally, removing many of the variables associated with separate handling of samples. A second labeling method to isolate and quantify phosphopeptides has been developed in the laboratory of Reudi Aebersold (11). The sequence of chemical reactions for selectively isolating phosphopeptides from a peptide mixture consists of six steps as shown in Fig. 4. To eliminate potential intra- and intermolecular condensation, the peptide amino groups are protected using t-butyl-dicarbonate (tBoc) chemistry (12). Follow-

ing this the carboxylate and phosphate groups are modified via a carbodiimide catalyzed condensation reaction to form amide and phosphoramidate bonds. The phosphoramidate bonds are then hydrolyzed via a brief acid wash to deprotect the phosphate group and cystamine is attached to the regenerated phosphate group via another carbodiimide-catalyzed condensation reaction. A free sulfhydryl group is generated at each phosphate group by reduction of the internal disulfide of cystamine, which allows the peptides to be attached to iodoacetyl groups immobilized on glass beads. The covalent attachment of the peptides allows stringent washing conditions to be used, thereby reducing the number of nonspecifically bound components being recovered with the phosphopeptides of interest. The

887

Vol. 290, No. 3, 2002

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

FIG. 2. (A) Mass spectra of EDT-D 0/D 4-labeled ␤-casein peptide. The enriched mixture of biotinylated phosphopeptides was analyzed by capillary reversed-phase liquid chromatography coupled directly online to a PE Sciex API QStar Pulsar hybrid quadrupole-TOF mass spectrometer. The [M ⫹ 2H] 2⫹ ion pair corresponds to the mass of the derivatized ␤-casein phosphopeptide FQS*EEQQQTEDELQDK, where S* has the modified side chain -CH 2-SCL 2CL 2S-acetyl-PEO-biotin and L is either H (EDT-D 0 label) or D (EDT-D 4 label). (B) Tandem mass spectrometry identification of ␤-casein phosphopeptides. The tandem MS/MS spectra of a phosphopeptide modified and affinity isolated using the (i) light and (ii) heavy isotopic versions of EDT and iodoacetyl-PEO-biotin are shown. Both labeled versions of the phosphopeptide were identified in a single LC data-dependent MS/MS analysis of the enriched mixture.

Vol. 290, No. 3, 2002

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

FIG. 3. Stoichiometric isotopic labeling of ␤-casein phosphopeptides. Samples of ␤-casein containing ratios of (A) 1:1, (B) 5:1, and (C) 10:1 were labeled with EDT-D 0:EDT-D 4, combined, biotinylated, affinity isolated using immobilized avidin, and analyzed by LC-TOF–MS. The integrated reconstructed ion chromatograms for each species, were used to calculate the D 0:D 4 ratio. (D) These measured isotope ratios were plotted against the molar ratios of the ␤-casein sample labeled with either EDT-D 0 or EDT-D 4.

phosphopeptides are recovered by cleavage of phosphoramidate bonds using trifluoroacetic acid at a concentration that also removes the tBoc protection group, thus regenerating peptides with free amino and phosphate groups. The carboxylate groups, however, remain blocked from step 2. While the chemistry involved is more complex, this method potentially provides greater enrichment since the phosphopeptides are covalently linked to a solid support during processing. Zhou et al. noted that this method yielded mixtures highly enriched in phosphopeptides with minimal contamination from other peptides. Since this strategy does not require the removal of the phosphate group it is equally applicable to phosphoseryl, phosphothreonyl and phosphotyrosyl-containing peptides. The CID spectra of the modified phosphopeptides were of high enough quality to allow the peptides to be identified using sequence database searching. The CID spectra could discriminate between pSer/pThr and pTyrcontaining peptides since pSer and pThr lose a H 3PO 4 group on MS/MS (13, 14), allowing these residues to be identified via a fragment ion corresponding to the loss of 98 Da. Phosphotyrosyl residues are more stable and do not lose their phosphate group during fragmentation. While this strategy does not provide a direct method to quantify changes in phosphorylation state between peptides from two different samples, the blocking of the carboxylates using either normal isotopic abundance or deuterated ethanolamine (i.e., ethanolamine-d 4) would allow for incorporation of sta-

ble isotope tags that later can be differentiated and quantified by MS. While both of the labeling strategies described above were developed using model systems, they have also been assessed with cell lysates since their greatest utility will be in the their application to proteome-wide identification and quantitation of phosphopeptides. In the case of the strategy proposed by Aebersold et al. (11), phosphopeptides were isolated from a Saccharomyces cerevisiae cell lysate and analyzed by LC–MS/ MS, with CID spectra recorded and searched against the sequence database. The total ion intensity recorded with respect to retention time on the column is shown in Fig. 4A. An expansion of this chromatogram between 24.7 and 26.5 min (Fig. 4B) shows the m/z values obtained over this time window. The major peptide peaks labeled with an asterisk (*) showed a loss of 98 Da during MS/MS indicating that a majority of the peptides detected were phosphorylated. In addition, ⬎80% of the useful MS/MS spectra identified phosphopeptides. For example, the MS/MS spectrum of the ion at m/z 1032.7 provided a definitive fragment ion series, and a major signal at m/z 983.8, corresponding to the doubly charged parent ion that has undergone the loss of H 3PO 4. The peptide was identified as being from enolase. While this peptide contained three potential pThr, the parent ion mass indicated that the peptide contained a single phosphate group. While the exact site of phosphorylation could not be determined, the MS/MS spectrum did indicate that the phosphate was not on the N-terminal threonyl residue.

889

Vol. 290, No. 3, 2002

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

ACKNOWLEDGMENTS This project has been funded in whole or in part with Federal funds from the National Cancer Institute, National Institutes of Health, under Contract NO1-CO-12400. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government.

REFERENCES

FIG. 4. (A) LC–MS chromatogram of phosphopeptides isolated from a tryptic digest of whole yeast cell lysate. (B) Integrated m/z spectrum of ions eluting from the LC column with retention times between 24.7 and 26.5 min, as indicated in A. Major ion peaks exhibiting a loss of 98 Da on MS/MS, indicating that they are phosphopeptides, are annotated with an asterisk (*). (C) The MS/MS spectrum recorded for the peptide peak at m/z 1032.7 was sufficient to identify the phosphopeptide as TAGIQIVADDLT*VT*NPAR from enolase.

CONCLUSIONS As the direction of biological sciences experiences an increasing shift in focus from the study of isolated systems to more global cellular components, the development of methods that effectively characterize posttranslational modifications is critical. The strategies described above provide useful methods to enrich mixtures for phosphopeptides, identify the site of phosphorylation, and quantify the relative phosphorylation state of proteins from two different sources. Due to the importance of reversible phosphorylation, this last feature is essential if the goal to understand the effects of a perturbation on the entire cell system is ever to be achieved.

1. Broder, S., and Venter, J. C. (2001) Sequencing the entire genomes of free-living organisms: The foundation of pharmacology in the new millennium. Annu. Rev. Pharmacol. Toxicol. 40, 97– 132. 2. Schulze, A., and Downward, J. (2001) Navigating gene expression using microarrays—A technology review. Nat. Cell Biol. 3, E190 –E195. 3. Lee, K. H. (2001) Proteomics: A technology-driven and technology-limited discovery science. Trends Biotechnol. 19, 217– 222. 4. Washburn, M. P., Wolters, D., and Yates, J. R., III (2001) Largescale analysis of the yeast proteome by multidimensional protein identification technology. Nat. Biotechnol. 19, 242–247. 5. Patterson, S. D. (2000) Proteomics: The industrialization of protein chemistry. Curr. Opin. Biotechnol. 11, 413– 418. 6. Cohen, P. (2000) The regulation of protein function by multisite phosphorylation—A 25 year update. Trends Biochem. Sci. 25, 596 – 601. 7. Huang, C., Ma, W. Y., Young, M. R., Colburn, N., and Dong, Z. (1998) Proc. Natl. Acad. Sci. USA 95, 156 –161. 8. Mason, G. G., Murray, R. Z., Pappin, D., and Rivett, A. J. (1998) Phosphorylation of the ATPase subunits of the 26S proteasome. FEBS Lett. 430, 269 –274. 9. Goshe, M. B., Conrads, T. P., Panisko, E. A., Angell, N. A., Veenstra, T. D., and Smith, R. D. (2001) Phosphoprotein isotope coded affinity tag approach for isolating and quantitating phosphopeptides in proteome-wide analyses. Anal. Chem. 73, 2578 – 2586. 10. Oda, Y., Nagasu, T., and Chait, B. T. (2001) Enrichment analysis of phosphorylated proteins as a tool for probing the phosphoproteome. Nat. Biotechnol. 19, 317–318. 11. Zhou, H., Watts, J. D., and Aebersold, R. (2001) A systematic approach to the analysis of protein phosphorylation. Nat. Biotechnol. 19, 375–378. 12. Bodanszky, A. B. M. (Ed.) (1984) The Practice of Peptide Synthesis, Vol. 21, Springer-Verlag, New York. 13. Jonscher, K. R., and Yates, J. R. (1997) Matrix-assisted laser desorption ionization/quadrupole ion trap mass spectrometry of peptides. Application to the localization of phosphorylation sites on the P protein from Sendai virus. J. Biol. Chem. 272, 1735– 1741. 14. Qin, J., and Chait, B. T. (1997) Identification and characterization of posttranslational modifications of proteins by MALDI ion trap mass spectrometry. Anal. Chem. 69, 4002– 4009.

890