Analysis of Phosphorylated Peptides by Ion Mobility-Mass Spectrometry

Oct 5, 2004 - College Station, Texas 77843, Behavioral Neuroscience, Intramural Research Program, NIDA, NIH,. Baltimore, Maryland 21224, and Ionwerks,...
0 downloads 0 Views 242KB Size
Download PDF
Anal. Chem. 2004, 76, 6727-6733

Analysis of Phosphorylated Peptides by Ion Mobility-Mass Spectrometry Brandon T. Ruotolo,† Kent J. Gillig,† Amina S. Woods,‡ Thomas F. Egan,§ Michael V. Ugarov,§ J. Albert Schultz,§ and David H. Russell*,†

Laboratory for Biological Mass Spectrometry, Department of Chemistry, Texas A&M University, College Station, Texas 77843, Behavioral Neuroscience, Intramural Research Program, NIDA, NIH, Baltimore, Maryland 21224, and Ionwerks, Inc., 2472 Bolsover, Suite 255, Houston, Texas 77005

An ion mobility-mass spectrometry technique for rapid screening of phosphopeptides in protein digests is described. A data set of 43 sequences (ranging in mass from 400 to 3000 m/z) of model and tryptic peptides, including serine, threonine, and tyrosine phosphorylation, was investigated, and the data support our previously reported observation (Ruotolo, B. T.; Verbeck, G. F., IV; Thomson, L. M.; Woods, A. S.; Gillig, K. J.; Russell, D. H. J. Proteome Res. 2002, 1, 303.) that the drift time-m/z relationship for singly charged phosphorylated peptide ions is different from that for nonphosphorylated peptides. The data further illustrate that a combined data-dependent IM-MS/MS approach for phosphopeptide screening would have enhanced throughput over conventional MS/MSbased methodologies. Of the many challenges associated with structural biology, understanding the dynamics of functioning proteins, or sets of proteins, is among the most difficult. Changes in the secondary, tertiary, or quaternary structure of a protein, or protein complex, can be initiated by remote and complicated mechanisms involving posttranslational modifications, chemical features of amino acid sequences that are not predicted by genomic information.1 Phosphorylation, which occurs at serine, tyrosine, or threonine residues, is one such posttranslational modification, which have been implicated in numerous signaling pathways.2 Mass spectrometry (MS) has become a key technique for studies of posttranslational modifications.3 For example, Deshaies and coworkers have identified net 1 as an important phosphorylationstate-dependent signaling protein that acts as a control in cellular division (mitosis) in yeast.4 Generally, phosphorylated signaling proteins are in low abundance relative to other cell constituents; thus, the phosphoproteins must be detected in the presence of a large background of * Corresponding author. E-mail: [email protected]. † Texas A&M University. ‡ NIDA, NIH. § Ionwerks, Inc. (1) Yeagle, P. L.; Albert, A. D Biochemistry 2003, 42, 1365. (2) Herve, J. C.; Sarrouilhe, D. Biol. Cell 2002, 94, 423. (3) Mann, M.; Jensen, O. N. Nat. Biotechnol. 2003, 21, 255. (4) Straight, A. F.; Shou, W.; Dowd, G. J.; Turck, C. W.; Deshaies, R. J.; Johnson, A. D.; Moazed, D. Cell 1999, 97, 245. 10.1021/ac0498009 CCC: $27.50 Published on Web 10/05/2004

© 2004 American Chemical Society

nonphsophorylated peptides.5 Conventional methods of phosphoprotein detection and characterization involve radiolabeling (with either 32P or 33P) the protein of interest in combination with Edman sequencing and phosphoamino acid analysis,6 as well as modern MS-based methods, for the determination of the site of phosphorylation.7 MS has several advantages over traditional biochemical methods, including throughput and the ability to interrogate complex mixtures for the analysis of protein phosphorylation. For example, Nuwaysir and Stults pioneered the application of mass spectrometry to phosphoprotein analysis by coupling ion metal affinity chromatography (IMAC) with mass spectrometry detection.8 More recently, Hunt and co-workers employed esterification chemistry to minimize the influence of acidic side chains on IMAC separation and increase the phosphopeptide selectivity of the separation technique.9 Methods for confidently identifying the site(s) of phosphorylation are also commonly based on tandem mass spectrometry. For example, Carr and co-workers have developed several data-dependent, liquid chromatography (LC)-MS/MS strategies for the identification of phosphoproteins in complex mixtures that rely on the detection of marker ions (i.e., either m/z 79 or 63 in negative mode analysis) for phosphoprotein identification.10 Also, Zubarev and co-workers recently applied electron capture dissociation and collisional activation methods, using an LC-Fourier transform ion cyclotron resonance MS approach, to thoroughly characterize an extensively modified milk protein.11 In addition, recent work performed in our laboratory suggests that 193-nm photodissociation is also well suited for the analysis of phosphorylated peptides.12 A disadvantage of liquid chromatography combined with tandem mass spectrometry is LC separation times (hours) combined with the requirement for sequential MS analysis of each (5) Yan, J. X.; Packer, N. H.; Gooley, A. A.; Williams, K. L. J. Chromatogr., A 1998, 808, 23. (6) Coyler, J. In The Protien Protocols Handbook; Walker, J. M., Ed.; Humana Press Inc.: Totowa, NJ, 1996; pp 501-506. (7) McLachlin, D. T.; Chait, B. T. Curr. Opin. Chem. Biol. 2001, 5, 591. (8) Nuwaysir, L. M.; Stults, J. T. J. Am. Soc. Mass Spectrom. 1993, 4, 662. (9) 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. (10) Zappacosta, F.; Huddleston, M. J.; Karcher, R. L.; Gelfand, V. I.; Carr, S. A.; Annan, R. S. Anal. Chem. 2002, 74, 3221. (11) Kjeldsen, F.; Haselmann, K. F.; Budnik, B. A.; Sorensen, E. S.; Zubarev, R. A. Anal. Chem. 2003, 75, 2355. (12) Hettick, J. M. Ph.D. Dissertation, Texas A&M University, College Station, TX, 2003.

Analytical Chemistry, Vol. 76, No. 22, November 15, 2004 6727

eluting species.13 Consequently, numerous laboratories are working to develop techniques that more efficiently utilize both the mass spectrometer, which can acquire data in microseconds to milliseconds, and more effectively interrogate the ions produced, i.e., data-dependent acquisition methods.14 For example, Salomon et al. have used recently introduced linear ion trap techniques that significantly improve detection of low-abundant phosphorylated peptides derived from tryptic digestion of phosphorylated proteins.15 Alternatively, we recently reported increased specificity for detection of phosphopeptides using a combination of matrixassisted laser desorption/ionization (MALDI) ion mobility and mass spectrometry, as well as new ion mobility (IM) instrument designs that significantly increase sensitivity and limits of detection of IM-MS.16 Several groups have demonstrated that IM separation coupled with MS greatly enhances sample throughput17 and that the technology is compatible with data-dependent acquisition modes of operation.18 A major driving force behind the development of IM-MS is the speed of IM separation; ion elution time ranges from 100 µs to 10 ms. In addition, ion mobility separation of ions is achieved on the basis of ion-neutral collision cross sections;19 thus, separation of large biomolecules is dependent upon conformational (2°/3°) preference(s) of the analyte ion.20,21 For both singly and multiply charged peptide ions, there is a high degree of correlation between the collision cross section (mobility) and the mass of an ion; i.e., most peptide ions prefer a closepacked (random coil) conformation in the gas phase.22 Thus, plots of drift time versus m/z for a mixture of peptides (m/z values ranging from 500 to ∼3000) can be fit to a linear relationship with a high correlation coefficient (r >0.98). Outliers (∼10% deviation) from the best-fit line are due to different conformational preferences (i.e., helical structures versus random coils), whereas lesser deviations (∼3%) may indicate more subtle structural differences (i.e., differences related to packing efficiency).23-26 In preliminary experiments designed to evaluate a MALDIIM-MS for phosphopeptide screening, we observed two distinct (13) Wang, H.; Hanash, S. J. Chromatogr., B 2003, 787, 11. (14) Bruegger, B.; Erben, G.; Sandhoff, R.; Wieland, F. T.; Lehmann, W. D. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 2339. (15) 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. (16) . Gillig, K. J.; Ruotolo, B. T.; Stone, E. G.; Russell, D. H. Rev. Sci. Instrum., submitted. (17) (a) Valentine, S. J.; Counterman, A. E.; Clemmer, D. E. J. Am. Soc. Mass Spectrom. 1999, 10, 1188. (b) Gillig, K. J.; Ruotolo, B.; Stone, E. G.; Russell, D. H.; Fuhrer, K.; Gonin, M.; Schultz, A. J. Anal. Chem. 2000, 72, 3965. (18) (a) Hoaglund-Hyzer, C. S.; Li, J.; Clemmer, D. E. Anal. Chem. 2000, 73, 2737. (b) Stone, E. G.; Ruotolo B.; Gillig, K. J.; Russell, D. H.; Fuherer K.; Gonin, M.; Shultz, A. J. Anal. Chem. 2001, 73, 2233. (19) Mason, E. A.; McDaniel, E. W. Transport Properties of Ions In Gases; John Wiley and Sons: New York, 1988; p 560. (20) Hoaglund-Hyzer, C. S.; Counterman, A. E.; Clemmer, D. E. Chem. Rev. 1999, 99, 3037. (21) Jarrold, M. F. Annu. Rev. Phys. Chem. 2000, 51, 179. (22) Valentine, S. J.; Counterman, A. E.; Clemmer, D. E. J. Am. Soc. Mass Spectrom. 1999, 10, 1188. (23) Ruotolo, B. T.; Gillig, K. J.; Stone, E. G.; Russell, D. H. J. Chromatogr., B 2002, 782, 385. (24) (a) Kinnear, B. S.; Hartings, M. R.; Jarrold, M. F. J. Am. Chem. Soc. 2001, 123, 5660. (b) Kinnear, B. S.; Hartings, M. R.; Jarrold, M. F. J. Am. Chem. Soc. 2002, 124, 4422. (25) Ruotolo, B. T.; Verbeck, G. F.; Thomson, L. M.; Gillig, K. J.; Russell, D. H. J. Am. Chem. Soc. 2002, 124, 4214. (26) Ruotolo, B. T.; Tate, C. C.; Russell, D. H. J. Am. Soc. Mass Spectrom,, submitted.

6728

Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

classes of phosphopeptides: class I and class II.27 Class I phosphorylated peptide sequences have collision cross sections that are similar to their nonphosphorylated analogues, but the 80-Da difference between the two peptides shifts the trend line (slope of the drift time-m/z plot) for class I phosphopeptides relative to nonphosphorylated peptides. We attribute this behavior to structural difference between phosphorylated and nonphosphorylated class I peptides, which appears to be due to intramolecular interactions between the partial negative charge carried on the phosphate and basic sites (i.e., arginine, lysine, and histidine side chains) elsewhere on the peptide. Class II peptides are sequences that exhibit larger collision cross sections when compared to nonphosphorylated peptides; consequentially, the trend line does not exhibit a shift from that for nonphosphorylated peptides; i.e., the ion signals for the phosphorylated peptides fall on the same trend line as the nonphosphorylated peptide ions and lack strong intramolecular interactions. Similar behavior has been reported for alkali-adducted peptide ions. For example, Bowers and co-workers first reported that sodiated peptide ions ([M + Na]+ ions) have collision cross sections similar to their protonated analogues and suggested that the gas-phase conformation of the sodium-bound peptide is different from that of the protonated peptide; e.g., the peptide acts as a multidentate ligand, creating a more compact charge-solvated structure.28 Analogous structures have been proposed to explain the fragmentation chemistry of peptide [M + Na]+ ions.29 Here, we present data for a larger set (43 sequences, including serine, threonine, and tyrosine phosphorylation) of phosphorylated peptides in order to verify the dominance of class I behavior for singly charged, phosphorylated peptide ions. The phosphopeptide data set includes tryptic peptides, as well as other protein fragments, and ranges in mass from 400 to 3000 m/z. The data support the utility of a combined data-dependent screening-IM approach for phosphoprotein screening, as the majority of phosphopeptides investigated have similar collision cross sections compared to the same sequence without the phosphate group attached. EXPERIMENTAL SECTION Sample Preparation. Phosphorylated peptide samples having the sequences KRpTLRR (m/z 910.0, p1), pYpYEEP (m/z 763.7, p2), RRApSVA (m/z 739.8, p3), DDE(Nle)pTGpYVATR (m/z 1399.5, p4), F(Nle)(Nle)pTPpYVVTR (m/z 1368.5, p5), AcRRLIEDAEpYAARG-NH2 (m/z 1641.7, p6), and Ac-EPQpYEEIPIYL-NH2 (m/z 1515.6, p7) were acquired from the University of Michigan Protein Structure Facility (Ann Arbor, MI). The phosphopeptides pYMAPYDNY (m/z 1117.1, p8), TSTEPQpYQPGENL (m/z 1544.5, p9), pYVPML (m/z 702.8, p10), LKRApYLNH2 (m/z 901.5, p11), pYVNV (m/z 574.6, p12), KRREILSRRPSpYRK (m/z 1926.2, p13), ENDpYINASL (m/z 1119.1, p14), LKRApTLG-NH2 (m/z 838.0, p15), LRRApSLG (m/z 852.9, p16), SVLpYTAVQPNE (m/z 1300.4, p22), and NpYISKGSTFL (m/z 1209.3, p23) were acquired from Anaspec Inc. (San Jose, CA). (27) Ruotolo, B. T.; Verbeck, G. F., IV; Thomson, L. M.; Woods, A. S.; Gillig, K. J.; Russell, D. H. J. Proteome Res. 2002, 1, 303. (28) Wyttenbach, T.; von Helden, G.; Bowers, M. T. J. Am. Chem. Soc. 1996, 118, 8355-8364. (29) Russell, D. H.; McGlohon, E. S.; Mallis, L. M. Anal. Chem. 1988, 60, 1818. (b) Teesch, L. M.; Orlando, R. C.; Adams, J. J. Am. Chem. Soc. 1991, 113, 3668.

Table 1. Results from IM-MS Analysis of Phosphorylated Sequences sequence (peptide)a

massb

diffc

classc

ELEELNVPGEIVEpSLpSpSpSEESITR (β-cas)d FQpSEEQQQTEDELQDK (β-cas) VPQLEIVPNpSAEER (R-cas s1) YKVPQLEIVPNpSAEER (R-cas s1) KRpTLRR (p1) pYpYEE (p2)e RRApSVA (p3) DDE(Nle)pTGpYVATR (p4)d F(Nle)(Nle)pTPpYVVTR (p5)d Ac-RRLIEDAEpYAARG-NH2 (p6) Ac-EPQpYEEIPIYL-NH2 (p7) pYMAPYDNY (p8) TSTEPQpYQPGENL (p9) pYVPML (p10) LKRApYLG-NH2 (p11) pYVNV (p12) KRREILSRRPSpYRK (p13) ENDpYINASL (p14) LKRApTLG-NH2 (p15) LRRApSLG (p16) RRApSPVA (p17) KRpTIRR (p18) DRVpYIHPF (p19) RRREEEpSEEEAA (p20) RRREEEpTEEEAA (p21) SVLpYTAVQPNE (p22) NpYISKGSTFL (p23) GpSSGDLKKDD (p24) CAHPNDLpYVELPENIPFY (p25) CAHPNDLYVELPENIPFpY (p26) GEREEpTEEEEEEEDEN (p27) SAQEpSQGNT (p28) Ac-EPQpYEEIP (p7) Ac-EPQpYEE (p7) Ac-EPQpYE (p7) Ac-EPQpY (p7) PQpYE (p7) PQpY (p7) TEPQpYQP-H2O (p9) EILSRRPSpYRK (p13) ENDpY (p14) DpYIN -or- NDpYI (p14) SVLpYTAVQ (p22)

2967.8

0.97561

I

2063 1661.8 1953.1 910 763.7 739.8 1399.5 1368.5 1641.7 1474.8 1117.1 1544.5 702.8 901.5 574.6 1926.2 1119.1 838 852.9 836.9 910 1127.2 1571.5 1585.5 1300.4 1209.3 1102 2215.4 2215.4 2062.8 1001.9 1177 972 858.7 730.1 601.8 472 906.3 1484.7 602.15 586.19 942.2

0 1.41844 3.85439 0 1.97008 4.02485 1.50189 -0.9295 1.40383 1.43789 1.65944 2.96793 1.97008 2.23817 2.20715 2.20745 1.08712 4.19882 -0.6047 1.40845 2.64901 -1.1696 3.06122 3.06122 1.9802 5.39007 -0.6084 0.90407 0 0.99461 2.02847 2.16563 1.17724 -0.6451 0 0 -0.793 0.6237 0 -1.442 -2.9469 2.94985

I I II I I II I I I I I I I I I I I II I I I I II II I II I I I I I I I I I I I I I I I I

a Amino acid sequences are listed (using abbreviations discussed in Experimental Section), peptide ID/origin is noted in parentheses. ISD fragment ions are indicated with *. b Mass of the phosphorylated peptide is given. c Difference and class are computed as discussed in the Experimental Section. d Peptides contain multiple phosphorylation sites; % difference is given for the loss of 1 phosphate from the completely phosphorylated analogue; all losses were observed to be class I peptides. e Same as 4; however, peptide was observed to be class II.

The phosphopeptides GpSSGDLKKDD (m/z 1102.0, p24), CAHPNDLpYVELPENIPFY (m/z 2215.4, p25), CAHPNDLYVELPENIPFpY (m/z 2215.4, p26), GEREEpTEEEEEEEDEN (m/z 2062.8, p27), and SAQEpSQGNT (m/z 1001.9, p28) were all acquired from the Johns Hopkins University School of Medicine Synthesis and Sequencing Facility (Baltimore, MD). In the above sequences, the abbreviation Nle stands for the nonstandard amino acid norleucine, the notation Ac- indicates N-terminal acetylation, the notation -NH2 indicates C-terminal amidation, and the lower case p denotes the phosphorylation of the amino acid to its right (toward the C-terminus). For a complete analysis of the current phosphopeptide database, data from previous studies is also included here. Results with the sequences RRApSPVA (m/z 836.9,

Figure 1. Drift time vs m/z for the peptide p14 and in-source fragments. All phosphorylated sequences are marked with an *. The signals for the [M + H]+ and the [M + H - PO3H]+ ions are indicated.

p17), KRpTIRR (m/z 910.0, p18), DRVpYIHPF (m/z 1127.2, p19), RRREEEpSEEEAA (m/z 1571.5, p20), and RRREEEpTEEEAA (m/z 1585.5, p21) are revisited in Table 1.27 All peptides will be referred to in the text using the code provided above (i.e., pn). Samples were mixed in a 100:1 matrix-to-analyte ratio with R-cyano-4-hydroxycinnamic acid for MALDI-IM-MS analysis. All peptides were used without further purification or preparation. In all experiments, no additional alkali was added to the sample to promote [M + Na]+, [M + K]+, or [M + N a + K - H]+ ion formation (see Figure 1), and additional experiments have been performed that confirm that the in-source fragmentation observed is due to the protonated species rather than an alkali adduct. Protein digest samples of chicken egg white lysozyme, bovine β-casein phosphorylated, bovine R-casein phosphorylated, and bovine R-casein dephosphorylated (Sigma, St. Louis MO.) were prepared as described previously.30 Briefly, micromolar concentrations of protein were heated to 90 °C for 30 min, allowed to cool at ∼15 °C for 5 min (to quench the thermal denaturation process), then mixed in a 40:1 substrate-to-enzyme ratio with sequencing grade-modified trypsin (Promega, Madison WI.) and allowed to react for 4-5 h. Samples were mixed in a 100:1 matrix-to-analyte ratio with R-cyano-4-hydroxycinnamic acid for MALDI-IM-MS analysis. For dephosphorylation reactions involving β-casein, the protein was pretreated with rabbit muscle protein phosphatase (Sigma) in 50 mM Tris buffer (pH 6.8). The protein was then digested according to the protocol above. In mixture experiments involving phosphorylated peptides and peptides resulting from a tryptic digest of lysozyme, approximately equimolar amounts (∼100 pmol each) of the three peptides (p7, p9, p14) were added to ∼100 pmol of lysozyme tryptic digest peptides, utilizing R-cyano4-hydroxycinnamic acid (in a matrix-to-analyte ratio of. 100:1) for MALDI analysis. Ion Mobility-Mass Spectrometry.. Measurements of the cross-sectional difference between phosphorylated and nonphosphorylated peptides were made on a MALDI-IM-orthogonal timeof-flight (o-TOF) mass spectrometer built in-house and described in detail elsewhere.16,31 Briefly, the instrument consists of a MALDI (30) Park, Z.-Y.; Russell, D. H. Anal. Chem. 2001, 73, 2558. (31) Gillig, K. J.; et. al., manuscript in preparation.

Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

6729

werks). Contour plots shown here were produced using Transform and related programs produced with the IDL software language (Research Systems, Boulder CO.). Differences (% diff) in collision cross sections were calculated by comparing the centroid of the mobility arrival time distribution for the phosphorylated peptide to that for the nonphosphorylated (or less phosphoyrlated in some cases) and normalizing the measured difference to an average drift time. The distinctions between class I or II peptides are arbitrary and do not have any chemical or biological significance.27 The sequences RRREEEpSEEEAA (p20) and RRREEEpTEEEAA (p21) appear to have similar conformations (difference of ∼3%) in either phosphorylated or nonphosphorylated forms, which suggest that the structure arises due to a salt bridge interaction between glutamic acids near the C-terminus and N-terminal arginine residues. This value was used as a cutoff for distinguishing class I and II behavior in the larger data set presented here. For reference, the typical difference determined by the average mobility trend line is ∼3-4% for a mass difference of 80 under the two sets of IM conditions used in this study (∼35 V cm-1 Torr-1 for measurements of p1-p16 and higher pressure conditions (50% of all signals can be disregarded prior to data-dependent MS/MS interrogation, thus greatly enhancing sequencing throughput. We feel that this is a conservative estimate of the potential throughput enhancement offered by an IM-MS approach to phosphopeptide screening. For example, a tryptic digest of β-casein produces a few number of phosphorylated ion signals, and essentially no nonphosphorylated ion signals, that have a negative deviation from the peptide trend line. For this favorable example, >90% of all the signals could be disregarded prior to data-dependent MS/MS, enhancing the overall throughput for phosphopeptide identification by orders of magnitude. Overall, the data presented here support the utility of IM-MS as a rapid confidence level enhancement technique for the detection of phosphorylated peptides contained in a complex biological mixture when combined with other (chemical- or MSbased) methods of phosphopeptide analysis. It is apparent that the efficiency of MS/MS-based data-dependent screening, where the peptide ions are fragmented to identify phosphopeptides, could be significantly enhanced through the use of IM separation. We have shown that 193-nm photodissociation of phosphopeptides yields diagnostic fragment ions (i.e., loss of PO3H) as well as abundant b and y fragment ions that can be used to determine the specific site of phosphorylation.34 We have also demonstrated that this experiment can be combined with IM-MS, and it is relatively straightforward to design data-dependent experiments where the only ions subjected to 193-nm irradiation are those having negative deviations from the average drift time versus m/z for peptide ions. ACKNOWLEDGMENT Funding for this research is provided by the National Science Foundation (CHE-9629966), NIH-National Center for Research Resources (1 RO1 RR019587-01), the Texas Advance Research Program/Technology Development/Transfer (TARP/TDT, 0103660064-2001), and the Department of Energy Division of Chemical Sciences, BES (DE-FG03-95ER14505).

Received for review February 4, 2004. Accepted August 23, 2004. AC0498009

Analytical Chemistry, Vol. 76, No. 22, November 15, 2004

6733