Anal. Chem. 1997, 69, 294-301
Detection of Modified Peptides in Enzymatic Digests by Capillary Liquid Chromatography/ Electrospray Mass Spectrometry and a Programmable Skimmer CID Acquisition Routine Paul T. Jedrzejewski* and Wolf D. Lehmann
Department of Central Spectroscopy, German Cancer Research Center (DKFZ), Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany
A method for the identification of multiple covalent protein modifications in enzymatic protein digests by specific marker ion signals in a single analysis is described. This method is based on the combined strengths of capillary liquid chromatography (µLC) to purify, concentrate, and resolve complex mixtures and electrospray mass spectrometry (ESI-MS) to selectively and sensitively detect ions. A variety of modification-specific marker ions can be generated using a programmable skimmer collisioninduced dissociation (sCID) acquisition routine, which allows for flexibility in the (i) number of marker ions monitored under single-ion monitoring conditions, (ii) selection of optimal polarity for both marker ions and molecular ions, (iii) use of variable dwell times for marker ions, and (iv) selection of optimal sCID offset. Using this combined method of µLC/ESI/sCID-MS, phosphorylated, sulfated, acrylamide-modified, and glycosylated peptides were identified in a model enzymatic digest at 200 fmol. The capability of reversed-phase LC to resolve isomeric compounds which cannot be identified by low-energy CID underscores the utility of this combined method. Further capabilities of this technique are demonstrated by the analysis of biologically important proteins. Characterization of protein modifications is an important aspect of the analysis of protein structure and function. Co- and posttranslational modifications such as phosphorylation, sulfation, glycosylation, and N-terminal modifications (e.g., myristoylation) are means to regulate the cellular distribution and function of proteins and are of particular importance in cell signaling (for reviews, see refs 1 and 2). In vivo, these modifications are due to specific enzymatic activities which are regulated on the basis of requirements of the cell. Thus, the precise qualitative elucidation of these modifications is of critical importance in the investigation of a protein’s function in a cell. To date, approximately 250 physiological covalent protein modifications are known,3,4 and nonphysiological alterations occurring during protein isolation and enrichment procedures, e.g., oxidation or
acrylamide addition, add to this variability. Among the various analytical techniques available for the characterization of covalent protein modifications, such as enzymatic or chemical cleavage (e.g., Edman degradation), isoelectric focusing, chromatography, radiolabeling, and mass spectrometry (MS), the latter technique is the most flexible and informative.5-9 Conventional protein sequencing methods (e.g., Edman degradation) are generally not suitable for the detection of such protein modifications because of the instability of the modifications under the conditions of analysis and/or the incompatibility of the method for their detection.10-12 Of the mass spectrometric methods for analysis of these modifications, off-line tandem mass spectrometric analysis suffers from the need for further sample manipulation (e.g., purification, transfer, loading); however, a more detrimental limitation of this approach is the lack of selectivity (due to “cross reactivity”) in analyzing complex mixtures containing peptides with multiple covalent modifications.13 Additionally, since these methods eliminate chromatographic separation, valuable data concerning the chromatographic behavior of the peptides, which aids in the overall characterization of the peptide composition, are not obtained. The combination of liquid chromatography (LC) with electrospray mass spectrometry (ESI-MS) is a well-established method for the identification and characterization of proteins because of its inherent selectivity, specificity, and sensitivity.5,6,14-19 The capabilities of this method (LC/ESI-MS) for mixture analysis are
* Corresponding author: Tel: +49(6221) 424560. Fax: +49(6221) 424561. E-mail:
[email protected]. (1) Wold, F. Annu. Rev. Biochem. 1981, 50, 783-814. (2) Methods in Enzymology; Wold, F., Moldave, K., Eds.; Academic Press: London, 1984; p 106. (3) Krishna, R. G.; Wold, F. In Methods in Protein Sequence Analysis; Imahori, K., Sakiyama, F., Eds.; Plenum Press: New York, 1993; pp 167-172.
(4) Ken Mitchelhill, John Holt Protein Structure Laboratory, St. Vincent’s Institute of Medical Research, 41 Victoria Parade, Fitzroy 3065, Victoria, Australia http://www.medstv.unimelb.edu.au/wwwdocs/svimrdocs/MassSpec/deltamassV2.html. (5) Carr, S. A.; Hemling, M. E.; Bean, M. F.; Roberts, G. D. Anal. Chem. 1991, 63, 2802-2844. (6) Kouach, M.; Belaiche, D.; Jaquinod, M.; Couppez, M.; Kmiecik, D.; Ricart, G.; VanDorssellaer, A.; Sautie`re, P.; Briand, G. Biol. Mass Spectrom. 1994, 23, 283-294. (7) Patterson, S. D.; Aebersold, R. Electrophoresis 1995, 16, 1791-1814. (8) Medzihradszky, K. F.; Burlingame, A. L. Methods: a companion to Methods in Enzymology; Academic Press: London, 1994, Vol. 6, pp 284-303. (9) Biemann, K.; Scoble, H. A. Science 1987, 237, 992-998. (10) Roach, P. J.; Wang, Y. In Methods in Enzymology; Hunter, T., Sefton, B. M., Eds.; Academic Press: London, 1991; pp 201, 200-224. (11) Meyer, H. E.; Hoffmann-Posorske, E.; Heilmeyer, L. M. G., Jr. In Methods in Enzymology; Hunter, T., Sefton, B. M., Eds.; Academic Press: London, 1991; pp 201, 169-185. (12) Cohen, P.; Gibson, B. W.; Holmes, C. F. B. In Methods in Enzymology; Hunter, T., Sefton, B. M., Eds.; Academic Press: London, 1991; pp 201, 153-168. (13) Wilm, M.; Neubauer, G.; Mann, M. Anal. Chem. 1996, 68, 527-533. (14) Davis, M. T.; Stahl, D. C.; Hefta, S. A.; Lee, T. D. Anal. Chem. 1995, 67, 4549-4556.
294 Analytical Chemistry, Vol. 69, No. 3, February 1, 1997
S0003-2700(96)00661-0 CCC: $14.00
© 1997 American Chemical Society
particularly powerful because of the combined strengths of LC to purify, concentrate, and resolve complex mixtures (e.g., enzymatic digests) and ESI-MS for mass discrimination and determination. Typical protocols for protein identification or for characterization of modifications begin with the enzymatic digestion of the protein sample, resulting in a complex mixture of modified and unmodified peptides (e.g., a tryptic digest of bovine serum albumin, a 69 kDa protein, results in an 82-component peptide mixture). Thus, the separation of these complex mixtures by chromatographic means prior to mass determination facilitates the task of spectra interpretation. To further improve the power of the LC/ESI-MS method in protein analysis, we have developed a new acquisition scheme which allows for the simultaneous determination of multiple peptide modifications as well as peptide mass. The technique utilizes skimmer collision-induced dissociation (sCID) to fragment peptides in the ES ion source, yielding specific marker ions indicative of phosphorylation, sulfation, acrylamidation, glycosylation, myristoylation, and other forms of modification. The use of sCID to generate specific marker ions for the detection of phosphorylation, sulfation, and glycosylation was introduced by Carr and Huddleston.15-17 These workers identified modified peptides by employing a single-polarity (negative) scanning technique and detecting a single type of modification per analysis. This approach restricts the detection of the molecular ion to the same polarity as employed for the fragment ions.18,19 More recently, we and others have simultaneously demonstrated the acquisition of marker ion signals and molecular weight information irrespective of polarity.20,21 However, as in the previous studies, these analyses detected only a single type of modification per LC/ MS analysis. Furthermore, all these prior analyses were performed on low picomole samples (25-50 pm). In this study, we have extended this technique in four aspects: (i) the simultaneous detection of multiple fragment ions within a single scan cycle, (ii) the use of variable dwell times for individual marker ions, (iii) the inclusion of a polarity change during a scan cycle, and (iv) the selection of an optimum sCID offset. These additional features considerably increase the flexibility of this sCID method when applied to the detection of protein modifications. In this paper, we will call this method programmable sCID to unify the various terms used so far in the literature.15-21 The utility of this acquisition method for the characterization of femtomole amounts of protein is presented. EXPERIMENTAL SECTION Materials. HPLC-grade solvents were obtained from Merck (Darmstadt, Germany). Asialofetuin, fetuin, cholecystokinin 2533 fragment peptide (RDYsTGWIDF, sulfotyrosine t Ys), and (15) Huddleston, M. J.; Bean, M. F.; Carr, S. A. Anal. Chem. 1993, 65, 877884. (16) Carr, S. A.; Huddleston, M. J.; Bean, M. F. Protein Sci. 1993, 2, 183-196. (17) Huddleston, M. J.; Annan, R. S.; Bean, M. F; Carr, S. J. Am. Soc. Mass Spectrom. 1993, 4, 710-717. (18) Ding, J.; Burkhart, W.; Kassel, D. B. Rapid Commun. Mass Spectrom. 1994, 8, 94-98. (19) Hunter, A. P.; Games, D. E. Rapid Commun. Mass Spectrom. 1994, 8, 559570. (20) Allen, M.; Anacleto, J.; Bonner, R.; Shushan, B. Proceedings of the 43rd ASMS Conference on Mass Spectrometry and Allied Topics, Atlanta, GA, May 2126, 1995; p 646. (21) Jedrzejewski, P.; Schnolzer, M.; Lehmann, W. D. Proceedings of the 43rd ASMS Conference on Mass Spectrometry and Allied Topics, Atlanta, GA, May 21-26, 1995; pp 632-633.
ammonium bicarbonate were obtained from Sigma (Deisenhofen, Germany). Modified trypsin sequence grade was obtained from Promega (Madison, WI). Trifluoroacetic acid (TFA) was obtained from Pierce (Bruchsal, Germany). Fused silica capillary was obtained from Composite Metal Services Ltd. (Worcester, U.K.). Packing material was obtained from Vydac (The Separations Group, Hesperia, CA). Mass Spectrometric Analysis. Electrospray mass spectra were acquired on a Finnigan MAT TSQ7000 triple-quadrupole mass spectrometer (Palo Alto, CA) equipped with an ESI ion source. The standard spray needle assembly was replaced with a fiber-optic holder (Melles Girot, Darmstadt, Germany) which held the titanium union (Valco, Schenkon, Switzerland) (the location of HV connection) that joined the column transfer line (50 µm i.d.) and fused silica capillary spray needle (50 µm i.d.). The spray needle was positioned on-axis 1-2 mm from the heated capillary orifice. Solutions were sprayed with a potential of about (1.8 kV. Micro-HPLC System. The capillary columns employed in all studies were prepared in-house from fused silica capillary (0.2 mm × 300 mm). Capillary columns were packed with Vydac 5 µm, 300 Å pore, C18 packing material using a slurry packing procedure.22,23 A flow rate of 0.2-0.5 µL/min through the column was established by a precolumn split from a delivery flow rate of 50 µL/min. All chromatography was performed on an Applied Biosystems (San Jose, CA) Model 140B dual-syringe pump system equipped with a 75 µL mixer. Samples were eluted using a linear gradient from 5% B to 50% B in 100 min. The solvents used were solvent A (0.065% aqueous TFA) and solvent B (80% acetonitrile in 0.05% aqueous TFA). Infusion Analysis. Infusion samples (model peptides) were delivered at a flow rate of 150 nL/min with a syringe pump (Model 22, Harvard Apparatus, South Natick, MA) and sprayed from a fused silica needle (0.025 mm × 150 mm). In this configuration, the high voltage ((2 kV) was applied to the stainless steel needle of the 10 µL syringe (Hamilton, Darmstadt, Germany) upstream from the fused silica needle. Sample Preparation. Hirudin 54-65 fragment peptide (GDFEEIPEEYpLQ, phosphotyrosine t Yp), synthetic peptide (myrGDAAAAK, N-terminal myristoylated glycine ≡ myrG), and synthetic peptide (CaNTPSLSSLLGT, acrylamidocysteine t Ca) were prepared in house. Modified trypsin (50:1 protein:enzyme ratio, w/w) was incubated with asialofetuin and fetuin samples in digest buffer (10% acetonitrile, 50 mM NH4HCO3, pH 8.5) overnight at 37 °C. Model mixture consisted of hirudin 54-65 fragment peptide, cholecystokinin 25-33 fragment peptide, acrylamide-modified peptide, and asialofetuin tryptic digest. PKA CR was prepared by a published protocol and digested with trypsin in digest buffer.24-26 PKA Cβ2 was purified by gel electrophoresis and blotted onto PVDF membrane.24-26 Membrane digests were performed overnight at 37 °C with trypsin (10:1 protein:enzyme ratio, w/w). (22) Davis, M. T.; Lee, T. D. Protein Sci. 1992, 1, 935-944. (23) Moritz, R. L.; Reid, G. E.; Ward, L. D.; Simpson, R. J. Methods: A Companion to Methods in Enzymology; Academic Press: London, 1994; Vol. 6, pp 213226. (24) Kinzel, V.; Ku ¨ bler, D. Biochem. Biophys. Res. Commun. 1976, 71, 257264. (25) Ku ¨ bler, D.; Gagelmann, M.; Pyerin, W.; Kinzel, V. Hoppe Seyler’s Z. Physiol. Chem. 1979, 360, 1421-1431. (26) Nelson, N. C.; Taylor, S. S. J. Biol. Chem. 1981, 256, 3743-3750.
Analytical Chemistry, Vol. 69, No. 3, February 1, 1997
295
Figure 1. Effect of sCID offset on modification-specific marker ions intensities. Legend: 0, sulfatation (SO3-, m/z 79.96); [, phosphorylation (PO3-, 78.96); 9, acrylamidation ([S - CH2 - CH2 - CO NH2]+, 104.62), *, myristoylation ([C13H27 - CO]+, 211.21), b, glycosylation (HexNAc+, 204.09); 4, sialization ([SA - H2O]+, 274.08).
RESULTS Optimization of Marker Ions by Skimmer CID. Marker fragment ions can be generated by the collision of the unseparated ion beam with gas molecules in the free-jet expansion region between the skimmer and the first ion focusing element (e.g., octapole in TSQ7000) of the ESI ion source. The extent of sCID is most effectively controlled by modifying the voltage offset between the skimmer and the octapole. To characterize fragmentation behavior of specific covalent peptide modifications and to maximize sensitivity for their detection, model modified peptides were studied in infusion experiments over a range of sCID offset values, the results of which are shown in Figure 1. For phosphorylated peptides, the specific marker ion is the phosphite anion (PO3-) at m/z 78.96 resulting from an ester bond cleavage from phosphorylated Ser, Thr, or Tyr amino acids.17 The hirudin 54-65 fragment peptide (phosphorylated at Tyr) was used as a model compound to determine the sCID offset optimization plot for the phosphorylation marker ion (Figure 1A). In this case, a broad maximum at an offset voltage of about 100 V was obtained. Corresponding investigations on peptides phosphorylated at Ser gave offset optimization plots very similar in shape and optimum offset value. Although a peptide phosphorylated at Thr was not studied here, it is expected that results consistent with the other phosphorylated amino acids would be obtained.17,18 Sulfated peptides may be distinguished from phosphorylated peptides by the sulfite anion (SO3-) at m/z 79.96. The sCID offset 296 Analytical Chemistry, Vol. 69, No. 3, February 1, 1997
optimization plot for the formation of this fragment from the model peptide cholecystokinin fragment 25-33, sulfated at tyrosine, is given in Figure 1A. Relative to the curve describing the PO3production, the maximal production of SO3- ions occurred at a lower offset potential of about 60 V. Acrylamide derivatization of cysteine has been proposed as an alkylation step preceeding protein analysis.27 In addition, this reaction frequently occurs as a side reaction when proteins with unprotected cysteine residues are separated on polyacrylamide gels.28-30 Acrylamidation was therefore studied as an example of a nonbiological modification. This modification can be recognized by a positively charged marker ion at m/z 104.62 having the structure [S - CH2 - CH2 - CO - NH2]+. As shown in Figure 1B, the optimum sCID offset for the generation of this marker ion from a synthetic peptide with an N-terminal acrylamidemodified cysteine is about 140 V. Myristoylation of proteins containing an N-terminal glycine is a frequent covalent lipid modification.4,31,32 Fragmentation of the peptide myrGNAAAAK (data not shown) yielded three positive fragment ions containing the fatty acid chain at m/z 211.21, 240.23, and 268.23 with the compositions [C13H27 - CO]+, [C13H27 - CO - NH - CH2]+, and [C13H27 - CO - NH - CH2 - CO]+, respectively. Because of its general specificity for myristoylation irrespective of amino acid backbone, the ion at m/z 211.21 was used in the remainder of these studies. The optimum sCID offset voltage for this marker ion was determined to be 90 V (Figure 1B). Glycosylation may be identified by a number of marker fragment ions which occur at m/z 163.05 (Hex+), 204.09 (HexNAc+), and 366.14 (Hex-HexNAc+). The Hex+ fragment ion, which was characterized by an optimum offset voltage
