Probing Proteomes Using Capillary Isoelectric Focusing

Additionally, a genome can only tell us what proteins can potentially be expressed, while providing no information on when, where, or at what level an...
0 downloads 0 Views 156KB Size
Download PDF
Anal. Chem. 1999, 71, 2076-2084

Probing Proteomes Using Capillary Isoelectric Focusing-Electrospray Ionization Fourier Transform Ion Cyclotron Resonance Mass Spectrometry Pamela K. Jensen, Ljiljana Pasˇa-Tolic´, Gordon A. Anderson, Julie A. Horner, Mary S. Lipton, James E. Bruce, and Richard D. Smith*

Environmental Molecular Sciences Laboratory, Pacific Northwest National Laboratory, P.O. Box 999, Richland, Washington 99352

Unlike the genome, the proteome is exquisitely sensitive to cellular conditions and will consist of proteins having abundances dependent upon stage in the cell cycle, cell differentiation, response to environmental conditions (nutrients, temperature, stress, etc.), or disease state(s). Therefore, the study of proteomes under well-defined conditions can provide a better understanding of complex biological processes and inference of protein function. Thus, much faster, more sensitive, and precise capabilities for the characterization of cellular constituents are desired. We describe progress in the development and initial application of the powerful combination of capillary isoelectric focusing (CIEF) and Fourier transform ion cyclotron resonance (FTICR) mass spectrometry for measurements of the proteome of the model system Escherichia coli. Isotope depletion of the growth media has been used to improve mass measurement accuracy, and the comparison of CIEF-FTICR results for the analysis of cell lysates harvested from E. coli cultured in normal and isotopically depleted media are presented. The initial studies have revealed 400-1000 putative proteins in the mass range 2-100 kDa from total injections of ∼300 ng of E. coli proteins in a single CIEF-FTICR analysis. With the growing availability of genomic databases, greater emphasis is being put on their exploitation in the “postgenomic era”. For sequenced genomes, a high percentage of the predicted proteins have no assigned functions,1 even after extensive homology searching using bioinformatic tools. DNA sequence alone gives us very little indication of a protein’s structure or physiological function, largely because it ignores co- and posttranslational modifications, such as phosphorylation, glycosylation, prenylation, ADP-ribosylation, and limited proteolysis and sheds little light upon covalent and noncovalent associations (e.g., protein-protein, protein-DNA). Additionally, a genome can only tell us what proteins can potentially be expressed, while providing no information on when, where, or at what level and how all of this is affected by changing the environment. (1) Cash, P. Anal. Chim. Acta 1998, 372, 121-145.

2076 Analytical Chemistry, Vol. 71, No. 11, June 1, 1999

At present, measurements of proteins are far less effective than measurements of DNA or RNA. Two rapid methods for assessing the activities of thousands of genes simultaneously by measuring mRNA levels have already been developed. These are the microarray assay technology2-4 and the serial analysis of gene expression (SAGE) techniques.5,6 However, it is still problematic to accurately predict the exact sequence of expressed proteins (especially for eukaryotes), and it has been shown that protein abundances cannot be accurately predicted from mRNA expression levels.7,8 Clearly, the understanding of complex cellular systems will be greatly advanced by the ability to make broad measurements of proteomes. Consequently, there is a broad interest in measurements of the “proteome”sthe entire complement of proteins expressed by a particular cell, organism, or tissue type at a given time, for a disease state or a specific set of environmental conditions. Current proteome studies rely almost completely on the use of two-dimensional polyacrylamide gel electrophoresis (2-D PAGE), which has undeniably assumed a major role due to its ability to provide detailed views of thousands of proteins expressed by an organism or cell type.1 In 2-D PAGE, proteins are first separated according to their isoelectric point (the pH where their net charge is zero) using isoelectric focusing (IEF) and then orthogonally separated by approximate mass using SDS-PAGE. Protein identification and characterization has classically involved the separate extraction and analysis of a visible protein spot, followed by N-terminal sequencing, internal peptide sequencing, or immunoblotting.9-11 More recently, methods have been developed based on matrix-assisted laser desorption/ionization (MALDI) or (2) Himmelreich, R.; Plagens, H.; Hilbert, H.; Reiner, B.; Herrmann, R. Nucleic Acids Res. 1997, 25, 701-712. (3) Schena, M.; Shalon, D.; Davis, R. W.; Brown, P. O. Science 1995, 270, 467-470. (4) Shalon, D.; Smith, S. J.; Brown, P. O. Genome Res. 1996, 6, 639-45. (5) Adams, M. D. Bioessays 1996, 18, 261-262. (6) Velculescu, V. E.; Zhang, L.; Zhou, W.; Vogelstein, J.; Basrai, M. A.; Bassett, D. E., Jr.; Hieter, P.; Vogelstein, B.; Kinzler, K. W. Cell 1997, 88, 243251. (7) Haynes, P. A.; Gygi, S. P.; Figeys, D.; Aebersold, R. Electrophoresis 1998, 19, 1862-1871. (8) Anderson, N. L.; Anderson, N. G. Electrophoresis 1998, 19, 1853-1861. 10.1021/ac990196p CCC: $18.00

© 1999 American Chemical Society Published on Web 05/04/1999

electrospray ionization (ESI) mass spectrometric analysis of ingel digested proteins that allow relatively rapid protein identification compared to conventional approaches.12,13 A major drawback is the relatively slow, labor-intensive, and cumbersome technology of 2-D PAGE, and the still extensive subsequent efforts for “one at a time” protein identification, that together make proteome characterization throughput much lower than for mRNA measurements. Moreover, useful sensitivity in 2-D PAGE is limited by the amount of a protein needed to visualize a spot (typically ∼10 fmol for conventional staining methods). Clearly considerable effort has been, and continues to be, directed toward addressing the limitations of 2-D PAGE technology as well as the development of new techniques for higher throughput proteome characterization. A promising high-throughput technique for proteome characterization is based upon capillary IEF (CIEF) in conjunction with ESI mass spectrometry.14,15 CIEF combines the resolving power of traditional gel IEF with the high efficiency, speed, and potential for automation of capillary electrophoresis. Separations of pI differences as small as 0.013 pH unit have been obtained in recent work at our laboratory.16 In CIEF, the fused-silica capillary has most commonly been coated with linear polyacrylamide to eliminate electroosmotic flow17 and reduce protein adsorption to the walls. CIEF employs a polyampholyte mixture in free solution which, when bridging high-pH and low-pH solutions, sets up a pH gradient in the capillary through which proteins migrate until they have zero net charge and are focused at their respective isoelectric points.18-20 After focusing, protein bands are mobilized from the capillary by one of several methods for detection by ESIMS. The CIEF-MS combination is thus analogous to a 2-D PAGE separation, providing information on pI and molecular mass, but with added potential advantages that include greater speed and sensitivity, ease of automation, and much more accurate mass measurements. The use of Fourier transform ion cyclotron resonance (FTICR) as the mass detector offers many advantages, such as unsurpassed mass resolution and mass measurement accuracy, greater sensitivity, and the potential to perform multistage MSn experiments. Yang et al. recently presented initial results obtained for CIEFFTICR of proteins.15 In this work, we demonstrate several advances in CIEF-FTICR, including improved CIEF resolution and the use of isotopic depletion for obtaining greater sensitivity and accuracy of molecular mass measurements from FTICR.21 We also demonstrate significantly enhanced mass spectral quality for the (9) Humphery-Smith, I.; Cordwell, S. J.; Blackstock, W. P. Electrophoresis 1997, 18, 1217-1242. (10) Patterson, S. D. Anal. Biochem. 1994, 221, 1-15. (11) Wilkins, M. R.; Pasquali, C.; Appel, R. D.; Ou, K.; Golaz, O.; Sanchez, J. C.; Yan, J. X.; Gooley, A. A.; Hughes, G.; Humphery-Smith, I.; Williams, K. L.; Hochstrasser, D. F. Bio/Technology 1996, 14, 61-65. (12) Shevchenko, A.; Jensen, O. N.; Podtelejnikov, A. V.; Sagliocco, F.; Wilm, M.; Vorm, O.; Mortensen, P.; Boucherie, H.; Mann, M. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 14440-14445. (13) Yates, J. R. I. J. Mass Spectrom. 1998, 33, 1-19. (14) Tang, Q.; Harrata, A. K.; Lee, C. S. Anal. Chem. 1997, 69, 3177-3182. (15) Yang, L.; Lee, C. S.; Hofstadler, S. A.; Pasa-Tolic, L.; Smith, R. D. Anal. Chem. 1998, 70, 3235-3241. (16) Shen, Y.; Smith, R. D., to be submitted. (17) Kilar, F.; Hjerten, S. Electrophoresis 1989, 10, 23. (18) Mazzeo, J. R.; Krull, I. S. Anal. Chem. 1991, 63, 2852-2857. (19) Foret, F.; Muller, O.; Thorne, J.; Gotzinger, W.; Karger, B. L. J. Chromatogr., A 1995, 716, 157-166. (20) Huang, T. L.; Shieh, P. C. H.; Cooke, N. Chromatographia 1994, 39, 543548.

characterization of complex protein mixtures (i.e., cell lysates). These improvements have resulted in the observation of considerably larger numbers of proteins from the Escherichia coli proteome than obtained previously. Finally, we demonstrate several of the mass spectrometric approaches that can be used to facilitate protein identification. EXPERIMENTAL SECTION Sample Preparation. The JM109 strain of E. coli was purchased from Promega and cultured on plates containing LB media. Innoculation cultures were grown overnight in 3 mL of isotopically depleted (∼99.95% 12C, ∼99.99% 14N, and >99.995% 1H) Bioexpress media (Cambridge Isotope Laboratories, Andover, MA). Two cultures, each containing 100 mL of either normal or isotopically depleted Bioexpress media were grown at 37 °C with shaking at 225 rpm, until the OD600 nm was ∼1. Cells were harvested by centrifugation at 10000g for 30 s. The cells were resuspended in 200 µL of PBS buffer and lysed by mechanical agitation with 0.1-mm-diameter zirconium/silica beads (Biospec Products, Inc., Bartlesville, OK) using a Mini-Beadbeater (Biospec Products, Inc.) operated at 4600 rpm for 4 min. The cell lysate was separated from the beads and centrifuged at 10000g for 5 min to remove the cellular debris. Salts and high-molecular-mass components were removed using a modified dual-microdialysis22 system composed of a 300-kDa membrane in the first microfabricated stage, now followed by a second microfabricated stage containing an 8-kDa membrane. Protein concentration was measured using the Bradford assay.23 CIEF-FTICR MS. On-line coupling of CIEF with ESI-FTICR used 30-cm-long (50 µm i.d., 192 µm o.d.) fused-silica capillaries coated with linear polyacrylamide17 which were mounted within the electrospray probe of a standard Finnigan MAT (San Jose, CA) ESI source utilizing a coaxial liquid sheath flow configuration.24 The capillary was filled with a sample solution (total protein concentration of ∼0.5 mg/mL) containing 0.5% Pharmalyte 3-10 (Amersham Pharmacia Biotech, Piscataway, NJ). For the CIEF separation, the inlet reservoir contained 20 mM phosphoric acid as the anolyte and the outlet reservoir contained 20 mM sodium hydroxide as the catholyte during the focusing step. Isoelectric focusing of cell lysates was performed at a constant voltage of 8 kV for 10 min using a high-voltage power supply (Glassman HighVoltage, Inc., Whitehouse Station, NJ). Once the focusing was complete, the capillary tip (emitter) was fixed to protrude ∼1 mm outside the electrospray (sheath) needle. A sheath liquid composed of 50% methanol, 49% water, and 1% acetic acid (v/v/v) at pH 2.6 was delivered at a flow rate of 2 µL/min using a Harvard Apparatus 22 syringe pump (South Natick, MA). Gravity mobilization was applied concurrently with cathodic mobilization by raising the inlet reservoir 5 cm above the level of the electrospray needle. The ESI mass spectra were acquired with a 7-T ESI-FTICR mass spectrometer equipped with an Odyssey data system (Finnigan FTMS, Madison, WI).25 Ions were transferred to the trap from the ESI interface through an rf-quadrupole for collisional (21) Marshall, A. G.; Senko, M. W.; Li, W. Q.; Li, M.; Dillon, S.; Guan, S. H.; Logan, T. M. J. Am. Chem. Soc. 1997, 119, 433-434. (22) Liu, C. L.; Hofstadler, S. A.; Bresson, J. A.; Udseth, H. R.; Tsukuda, T.; Smith, R. D.; Snyder, A. P. Anal. Chem. 1998, 70, 1797-1801. (23) Bradford, M. M. Anal. Biochem. 1976, 72, 248-254. (24) Smith, R. D.; Wahl, J. H.; Goodlett, D. R.; Hofstadler, S. A. Anal. Chem. 1993, 65, A574-A584.

Analytical Chemistry, Vol. 71, No. 11, June 1, 1999

2077

focusing (at ∼200 mTorr) and two subsequent sets of rf-only quadrupoles (∼750 kHz, ∼500 Vpp) located in higher vacuum regions. Mass spectra were obtained utilizing standard experimental sequences (i.e., ion injection and accumulation, pumpdown, excitation/detection) with a total sequence time of ∼3.5 s (i.e., the maximum spectrum acquisition rate). Ion accumulation was facilitated by ∼10-5 Torr of N2 injected into the trap via a piezoelectric pulse valve (Lasertechniques Inc., Albuquerque, NM). Background pressure in the ICR trap was maintained at ∼10-9 Torr by a custom cryopumping assembly that provides effective pumping speeds of ∼105 L/s and thus allows rapid transition between in-trap ion accumulation (i.e., 10-5 Torr) and high-performance ion excitation/detection (i.e., 10-9 Torr) events. The Odyssey data station provided ICR trap control, ion excitation (i.e., broad-band chirp excitation over a 100-kHz bandwidth with a 35 Hz/µs sweep rate), data acquisition (256K data points at 270 kHz), and storage. CIEF-FTICR MS/MS. On-line CIEF-FTICR MS/MS experiments utilize an ancillary PC that controls the instrument through an Ethernet link, with the Odyssey data station providing most of the common experimental parameters, similar to that described previously for automated dynamic range enhancement.26 Automated CIEF-FTICR MS/MS employs selective ion accumulation (SIA) with single-frequency quadrupolar excitation (QE),27 followed by stored-waveform inverse Fourier transform (SWIFT) ion isolation and sustained off-resonance irradiation-collisionally induced dissociation (SORI-CID, typically ∼15 Vpp at a frequency 1 kHz lower than the reduced ICR frequency of the selected ions). QE and dipolar excitation (DE) waveforms (i.e., SIA and SWIFT/ SORI) were generated by use of a PC board (PCIP-AWFG, 5 MHz, 12 bit, Keithley Metrabyte Co., Taunton, MA) based on the most abundant signal detected in the prior MS sequence. Data Analysis. Analysis of the large and complex data sets arising from CIEF-FTICR experiments and the interpretation of results in the context of available genomic databases were done using software being developed at our laboratory to assist protein identification, MS/MS data analysis, and database searching.28 This process is fully automated, at present requiring ∼4.5 h on a 266-MHz PC to process a CIEF-FTICR experiment containing 250 files and resulting in the detection of ∼1500 masses from putative proteins. Use of faster computers or multiple processors will significantly reduce the time required for data processing in the future. Data analysis can be started after only a few minutes of user interaction with the analysis software, and the processed data can be presented in several formats, including the visualization of these large data sets in the form of familiar 2-D gel displays. RESULTS AND DISCUSSION In the present work with E. coli, increased separation resolution in CIEF has been achieved, allowing for the detection of more putative protein masses. This advancement, together with the (25) Winger, B. E.; Hofstadler, S. A.; Bruce, J. E.; Udseth, H. R.; Smith, R. D. J. Am. Soc. Mass Spectrom. 1993, 4, 566-577. (26) Bruce, J. E.; Anderson, G. A.; Smith, R. D. Anal. Chem. 1996, 68, 534541. (27) Bruce, J. E.; Anderson, G. A.; Hofstadler, S. A.; Van Orden, S. L.; Sherman, M. S.; Rockwood, A. L.; Smith, R. D. Rapid Commun. Mass Spectrom. 1993, 7, 914-919. (28) Anderson, G. A.; Bruce, J. E.; Pacific Northwest National Laboratory: Richland, WA, 1995.

2078 Analytical Chemistry, Vol. 71, No. 11, June 1, 1999

increased mass measurement accuracy gained by the use of isotope depletion methods, has resulted in a substantial improvement over previous work in the application of CIEF-FTICR to cell lysates.15 E. coli grown in LB media were harvested, lysed, and dualmicrodialyzed, as described above. A solution of E. coli proteins was prepared for CIEF-FTICR measurements by adding Pharmalyte 3-10 to a concentration of 0.5%. A total of 300 ng of protein (0.5 mg/mL × 0.59 µL capillary volume) was loaded in the CIEF capillary. A typical reconstructed ion electropherogram is shown in Figure 1A and exhibits numerous peaks attributed to proteins. Parts B-D of Figure 1 show single-scan mass spectra for the designated CIEF peaks. Resolution of the isotopic distributions obtained from ESI-FTICR facilitated rapid mass determinations from a single charge state (as shown for the cases of Figure 1B and C). Although isotopic resolution was not obtained for proteins with Mr > 30 kDa, an accurate average molecular mass could still be obtained from “deconvolution” of the charge-state distributions, as demonstrated in Figure 1D. Depletion of rare isotopes provides an effective strategy to increase both sensitivity and molecular mass measurement accuracies, compared to the use of natural-abundance isotopic distributions (∼98.89% 12C, ∼99.63% 14N, and 99.985% 1H), as recently shown by Marshall et al.21 For larger molecules, mass accuracy can be limited by the broad isotopic envelope for molecular ions and the statistical limitations on accurate fitting with the corresponding theoretical distributions. Depletion of rare isotopes allows detection of the monoisotopic species (without 13C, 15N, etc.) for significantly heavier species and thus avoids the 1-2-Da uncertainties that can result from discrepancies in fitting of the isotopic envelopes. Additionally, better sensitivity and detection limits (i.e., increased signal-to-noise ratios) are achievable since the signal is “concentrated” in fewer peaks. The “narrower” protein isotopic distributions substantially improve mass spectral quality, especially for the present application to complex protein mixtures (e.g., cell lysates). Isotope depletion further simplifies experiments by allowing easier isolation of the parent ion (i.e., for MS/MS studies) and more sensitive detection and identification of fragment ions. Thus, isotope depletion significantly extends achievable accuracy, sensitivity, dynamic range, and detection limit of mass measurements. For the present studies, we grew E. coli in isotopically depleted media. The reconstructed ion electropherogram obtained from CIEF-FTICR analysis of an isotopically depleted E. coli lysate (Figure 2A) appeared very similar to that for E. coli grown in the standard media (Figure 2B). This observation indicates there was no major impact on the E. coli culture due to the difference in growth media, as might be expected because of the small changes in isotopic composition and the fact that only rare isotopes were absent. In fact, most of the putative protein masses detected were observed in both samples. However, while the monoisotopic peak is easily identified in the isotopic envelope of isotopically depleted proteins, it is often undetectable for normal proteins (i.e., proteins with rare isotopes present at their natural abundances). The much narrower rare isotope-depleted protein isotopic distributions also clearly resulted in improved mass spectrum quality. In general, the use of isotopically depleted media enabled the detection of a greater number of proteins. Taking the isotopically

Figure 1. (A) CIEF-FTICR electropherogram for a lysate from E. coli grown in normal media. The positive ESI mass spectra in (B)-(D) are from single scans under their respective peaks. The insets of (B) and (C) show the high-resolution FTICR spectra for a single charge state.

Figure 2. Electropherograms for CIEF-FTICR positive ESI mass spectra. Shown are single scans under the indicated peaks (1) for lysates from E. coli grown in (A) isotopically depleted media and (B) normal media, along with insets showing high resolution of the indicated charge state.

depleted E. coli sample as an example, a wealth of information can be obtained from a single CIEF-FTICR experiment, as demonstrated in Figure 3. Even relatively low intensity peaks in the total ion electropherogram can reveal contributions due to several proteins, as indicated in Figure 3A. Figure 3 also shows the large dynamic range achievable by combining high-resolution CIEF and high-performance FTICR. As can be seen in Figure 3B, more than 10 proteins can sometimes be detected in one

spectrum. With so many proteins eluting at the same time, achievable dynamic range can clearly become limited due to the finite useful total trap charge capacity. Improved CIEF separations and faster FTICR spectrum acquisition rates will thus significantly extend the effective dynamic range achievable, and both improvements are presently being pursued at our laboratory. CIEF-FTICR experiments result in large data sets that are impractical to analyze manually and complicated to visualize. One way of visualizing the data is in the form of familiar 2-D PAGE displays, as illustrated in Figure 4. The data set shown, obtained from a single CIEF-FTICR experiment, reveals ∼900 “spots”, i.e., unique putative protein masses, which are plotted as molecular mass vs scan number (which can be correlated to pI by spiking the CIEF separation with standard proteins of known pI) to obtain the 2-D display. The spectral intensity in this display is related to ion abundance through the different spot sizes. The 2-D display of the data provides a quick and easy way to visualize the large quantity of proteomic information obtained in a single analysis and allows simpler comparison across many different analyses. However, this display does not convey the quality of the mass measurements and can obscure less abundant species (as can also happen in actual 2-D PAGE separations). The complexity of the samples can also be seen by comparing consecutive scans from a CIEF-FTICR analysis of E. coli, as shown in Figure 5. Currently, our duty cycle is ∼3 s, limited predominantly by the in-trap ion accumulation and associated pump-down time. However, in the time of only one scan, the proteomic information can change drastically, thus the need for faster FTICR duty cycles. If some proteins are only present at a detectable level during one scan of the FTICR, then there are probably proteins going undetected during the time between scans as well. With Analytical Chemistry, Vol. 71, No. 11, June 1, 1999

2079

Figure 3. (A) High-resolution FTICR spectra for charge states representing different molecular masses present in a single scan. (B) Mass spectrum containing peaks corresponding to at least 10 putative protein masses. The positive ESI mass spectra were taken from under the indicated peaks in the CIEF-FTICR electropherogram (bottom) for a lysate from E. coli grown in isotopically depleted media.

external accumulation of ions, the duty cycle can be significantly reduced, as demonstrated by Senko et al.,29 resulting in significant improvements in achievable resolution and sensitivity. Figure 6 shows the mass spectra containing the highest and lowest masses observed in these experiments. Although typically detected proteins are in the mass range of 3-60 kDa, we were able to detect proteins with Mr in excess of 90 kDa, despite use of the dual-microdialysis technique for sample preparation. The 300-kDa nominal molecular mass cutoff membrane in the first stage of this technique clearly discriminates against higher Mr species and has an “apparent” high-molecular-mass cutoff of ∼60 kDa. The practical molecular mass cutoffs for both the high end

(first stage) and low end (second stage) are fairly broad and not clearly defined, and complete characterization is beyond the scope of this study. The use of this dual-microdialysis technique results in the detection of more lower mass proteins and fewer high mass proteins when compared to 2-D PAGE. The differences can also be attributed to the increasing sensitivity of 2-D PAGE staining for larger proteins, to the greater sensitivity of ESI-FTICR for smaller proteins (due primarily to their smaller isotopic envelopes), and to the fact that detergents, urea, etc., which are typically used to stabilize large hydrophobic or membrane proteins for 2-D PAGE, cannot be used here due to their incompatibility with either CIEF or ESI-MS. Once the proteins in a particular cell lysate are detected, there is obviously a need to identify them, correlating them to the known genomic data. It is important to recognize that additional steps for protein identification are likely to be unnecessary in most cases once this has been accomplished. The reason for this is that the combination of pI and a highly accurate molecular mass measurement is extremely distinctive. Identification of a protein based on molecular mass alone is insufficient due to possible posttranslational modifications and proteolytic events, which play an important role in protein activity but are difficult to predict from genomic data.30 There are several different approaches to protein identification. We have investigated several of the possible techniques that can be used for protein identification, although proteome-wide identification of proteins from E. coli was not the goal of this study. Figure 7 gives an example of identification based solely on a protein’s accurate molecular weight and position in the 2-D display. Figure 7A shows the mass spectrum for a protein with Mr ) 40 966

(29) Senko, M. W.; Hendrickson, C. L.; Emmett, M. R.; Shi, S. D. H.; Marshall, A. G. J. Am. Soc. Mass Spectrom. 1997, 8, 970-976.

(30) Link, A. J.; Robison, K.; Church, G. M. Electrophoresis 1997, 18, 12591313.

Figure 4. 2-D display of the CIEF-FTICR analysis for a lysate from E. coli grown in isotopically depleted media.

2080 Analytical Chemistry, Vol. 71, No. 11, June 1, 1999

Figure 5. Positive ion mode ESI mass spectra for two sets of consecutive scans for the indicated peaks in the CIEF-FTICR electropherogram for a lysate from E. coli grown in isotopically depleted media.

Figure 6. Positive ion mode ESI mass spectra for the indicated peaks in the CIEF-FTICR electropherogram (right) for a lysate from E. coli grown in normal media.

Da (an average mass calculated from the unresolved charge states of the isotopically depleted protein), identified as phosphoglycerate kinase by searching the experimentally obtained mass against a database of masses for all theoretical E. coli proteins. Figure 7B shows the location of the spots on a 2-D gel (SWISS-2DPAGE at http://www.expasy.ch/ch2d/, Swiss Institute of Bioinformatics)

attributed to phosphoglycerate kinase. Figure 7C shows a 2-D display generated from the CIEF-FTICR data. A good correlation in position can be seen between our 2-D display and the 2-D PAGE, but due to the differences in isoelectric focusing conditions, the accurate molecular weight is by far the most important information for identification. Analytical Chemistry, Vol. 71, No. 11, June 1, 1999

2081

Figure 7. (A) Positive ion mode ESI mass spectrum for phosphoglycerate kinase obtained from CIEF-FTICR analysis of the lysate from E. coli grown in isotopically depleted media. (B) The 2-D PAGE for E. coli reproduced from SWISS-2D PAGE, Swiss Institute of Bioinformatics (http://www.expasy.ch/ch2d/). (C) Our 2-D display of the CIEF-FTICR results for a lysate of E. coli grown in isotopically depleted media.

Figure 8. (A) Positive ion mode ESI mass spectrum from direct infusion of a lysate from E. coli grown in isotopically depleted media. (B) Selective-ion accumulation and SWIFT isolation of the peak indicated by (1) in (A). (C) Spectrum obtained after SORI-CID, providing mass fragments for identification of the cold shocklike protein (CspC).

However, as stated before, pI and accurate molecular mass measurements will be insufficient for identification in many cases. In those cases, one can apply multistage MS/MS (MSn) capabilities,31,32 as illustrated in Figure 8. To demonstrate this capability, a sample of isotopically depleted E. coli lysate was mixed 50/50 (v/v) with sheath liquid and directly infused in the FTICR (Figure 8A). A peak of interest was trapped using selective-ion accumulation and the mass determined to be 7266.6 Da (Figure 8B). Although this mass correlated well with the theoretical mass of 2082 Analytical Chemistry, Vol. 71, No. 11, June 1, 1999

the cold shocklike protein CspC in the E. coli database, MS/MS was performed using SORI-CID to verify the identity of the protein using its fragment mass “tags”. The resulting MS/MS spectrum is shown in Figure 8C, indicating the fragments detected, which provide partial sequence information and allowed for the unambiguous identification of the protein as CspC. Ideally, this MS/MS approach will be carried out on-line during the course of the CIEF mobilization. Initial experiments aimed at demonstrating this technique were applied to the separation of a mixture of two standards, horse heart myoglobin and bovine carbonic anhydrase II (Sigma, St. Louis, MO). Figure 9A shows the reconstructed ion electropherogram for the CIEF-FTICR analysis. The mass spectrum for a single scan taken for the myoglobin peak is shown in Figure 9B and provides a mass of 16 591 Da. The most abundant ion (or other programmed selection criteria) in each mass spectrum (e.g., Figure 9B) can be used to synthesize the waveform(s) that will be used during the SIA, SWIFT isolation, and SORI-CID events in the subsequent MS/ MS experiment. The reconstructed ion electropherogram for the CIEF-FTICR MS/MS experiment is shown in Figure 9C. The sequence tags derived from the MS/MS spectrum of the selected ion (Figure 9D) were used to verify the identity of the myoglobin peak. At present, the addition of the MS/MS procedure can extend the total sequence time to more than 10 s, allowing for less frequent spectrum acquisition and thereby decreasing the effective separation resolution for CIEF. Additionally, since base pressure in the trap increases with the high frequency of sequence repetition, especially during long CIEF-FTICR MS/MS experi(31) Hofstadler, S. A.; Wahl, J. H.; Bakhtiar, R.; Anderson, G. A.; Bruce, J. E.; Smith, R. D. J. Am. Soc. Mass Spectrom. 1994, 5, 894-899. (32) Senko, M. W.; Speir, J. P.; McLafferty, F. W. Anal. Chem. 1994, 66, 28012808.

Figure 9. (A) CIEF-FTICR electropherogram of horse heart myoglobin and bovine carbonic anhydrase II. (B) Positive ion mode ESI mass spectrum for the myoglobin peak in (A). (C) CIEF-FTICR MS/MS electropherogram obtained after applying SIA, SWIFT and SORI-CID to the preselected parent ions. (D) Positive ion mode ESI mass spectrum for the myoglobin peak in (C), showing the distinctive high-resolution dissociation product spectrum that can be used for protein identification.

ments, the resolution obtained is decreased. These problems will be alleviated in future work by the addition of external ion accumulation methods (decreasing the required pump down time in the trap) and infrared multiphoton laser-induced dissociation33 in place of SORI-CID. It is anticipated that implementation of these steps will reduce the entire sequence for both stages of analysis to less than 3 s. Finally, it is important to emphasize that protein identification will generally only need to be done once for any proteome. The high mass accuracy of FTICR mass spectrometry provides accurate mass “tags” that in most cases will be sufficient for protein identification in subsequent studies of the same proteome. The advantage of this is that many studies of proteome “perturbations” resulting from environmental changes can then be conducted on a high-throughput basis. CONCLUSION CIEF-FTICR has been shown here to be an effective tool in proteome characterization, providing a means for measurement of a large number of soluble proteins with greater sensitivity, speed, and mass accuracy than possible with 2-D PAGE methods. Use of the CIEF approach avoids many of the problems associated with in-gel digestion procedures used in 2-D PAGE and the time required for subsequent “one at a time” MS analysis. Other advantages of CIEF include the capability to handle extremely small sample sizes, enhanced speed and resolution, ease of automation, and high sensitivity arising from the natural concentration effect (a factor of 50-100-fold) associated with the focusing (33) Little, D. P.; Speir, J. P.; Senko, M. W.; Oconnor, P. B.; McLafferty, F. W. Anal. Chem. 1994, 66, 2809-2815.

step. In fact, the amount of total protein loaded in the CIEF capillary was typically ∼300 ng, whereas 2-D PAGE can commonly involve 100 µg to >1 mg of total protein. Very often protein extracts from as many as 20 2-D PAGE separations need to be pooled for analysis by conventional MS methods. In addition, a single CIEF-FTICR analysis can be done in only a few hours, from sample preparation to data visualization, compared to 2-D PAGE, which commonly takes a few days. Furthermore, the improvement in separation resolution in CIEF has allowed for the detection of significantly more putative protein masses than accomplished by previous work,15 and recent studies suggest that even more substantial improvements in CIEF separations are obtainable,16 a development that would further enhance nearly all aspects of this approach. The utilization of FTICR provides the exceptional resolution, higher mass measurement accuracy, and greater sensitivity than feasible with conventional mass spectrometers, (e.g., quadrupole, ion trap, time-of-flight instruments). The employment of isotope depletion further improved sensitivity and accuracy of molecular mass measurements in FTICR, as well as significantly enhanced mass spectral quality for the characterization of these complex protein mixtures. Additionally, the nondestructive nature of FTICR potentially allows for multistage MS experiments on the same ion population, making unambiguous identification of proteins easier by generating partial sequence information with greater sensitivity. The initial efforts to incorporate on-line MS/MS techniques are encouraging, and future improvements in the FTICR instrumentation should make protein identification considerably more rapid. These features of FTICR, coupled with the advantages of CIEF, provide the basis for the high-throughput studies of proteomes, Analytical Chemistry, Vol. 71, No. 11, June 1, 1999

2083

with greater speed, resolution, and sensitivity than feasible with methods based on alternative technologies.

Pacific Northwest National Laboratory is operated by Batelle Memorial Institute for the U.S. Department of Energy under Contract DE-AC06-76RLO 1830.

ACKNOWLEDGMENT We thank the United States Department of Energy Office of Biological and Environmental Research and internal Laboratory Directed Research and Development for support of this research.

2084 Analytical Chemistry, Vol. 71, No. 11, June 1, 1999

Received for review February 18, 1999. Accepted April 9, 1999. AC990196P