Methods In Shotgun Proteome Analysis

Trends

Trends in Analytical Chemistry, Vol. 26, No. 1, 2007

Advances in chromatographic techniques and methods in shotgun proteome analysis Mingliang Ye, Xiaogang Jiang, Shun Feng, Ruijun Tian, Hanfa Zou Shotgun proteomics is a high-throughput approach to proteome analysis whereby the protein mixture is digested and the peptides generated are separated by capillary liquid chromatography and sequenced by tandem mass spectrometry (MS). Due to the huge number of peptide species, separation prior to MS analysis plays an important role in shotgun proteomics. Overall sensitivity, dynamic range, throughput and general effectiveness of shotgun proteomic analysis largely depend on how well the peptide mixture is separated. In recent years, new separation techniques have been applied successfully to proteome analysis and have dramatically improved protein identification. We briefly review the recent development of chromatographic techniques and methods in shotgun proteome analysis, including the following three aspects: one-dimensional separation; multidimensional separation; and, automated proteome-analysis systems. ª 2006 Elsevier Ltd. All rights reserved. Keywords: Automated proteome-analysis system; Multidimensional One-dimensional separation; Separation technique; Shotgun proteomics

separation;

1. Introduction Mingliang Ye, Xiaogang Jiang, Shun Feng, Ruijun Tian, Hanfa Zou* National Chromatographic R&A Center, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, China

*

Corresponding author. Tel.: +86 411 84379610; Fax: +86 411 84379620; E-mail: [email protected]

80

Proteomics aims to understand complex biological systems by analyzing protein expression, protein function, protein modifications and protein interactions. Mass spectrometry (MS) is a central analytical technique for proteome analysis. According to the review paper published in Science in 2006 [1], there are in general four MSbased proteome-analysis strategies: (1) MS analysis of substantially purified proteins; (2) MS analysis of complex peptide mixtures; (3) comparative pattern analysis; and, (4) hypothesis-driven strategies. In all the above strategies, separation techniques play an important role. Proteins or peptides in proteomic samples need to be separated prior to MS analysis. While the latter two approaches are still in their infancy, the first two are widely applied in proteome research.

The first is exemplified by the classic proteomic approach: two-dimensional (2D) gel electrophoresis of proteins followed by MS identification of proteins in gel spots. Although 2D gel electrophoresis provides unprecedented separation power for proteins, this approach suffers several limitations, including the difficulties of resolving proteins with extreme size, pI or hydrophobicity, and the difficulties associated with automation and reproducibility. In the second approach, also referred to as shotgun proteomics, complex protein samples are digested, the resulting peptides are separated and then subject to tandem MS (MS2) analysis, and the proteins are finally identified by searching databases. Due to its good compatibility with on-line MS detection, reversed phase liquid chromatography (RPLC) is typically used to separate peptides in shotgun proteomics as that circumvents some limitations of 2D gel electrophoresis. A major advantage of this approach is that a large number of proteins could be identified in a high-throughput manner. Proteome analysis faces challenges because of the great complexity of protein species and the large dynamic range of protein levels. In shotgun proteomics, the samples are even more complex because tryptic digestion generates several dozen peptides per protein. For example, there are 52,816 protein entries in Human IPI database (version 2.23), and these proteins will generate 892,584 peptides after trypsin digestion in silico [2]. The number of species is increased 17-fold. Due to the limited scan rate of MS and ion-suppression effects in MS, efficient separation prior

0165-9936/$ - see front matter ª 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.trac.2006.10.012

Trends in Analytical Chemistry, Vol. 26, No. 1, 2007

to MS is required. However, to separate such a complex sample is a big challenge in separation science. In the past few years, some new separation techniques were developed for peptide separation and significantly improved the overall sensitivity, dynamic range, throughput and general effectiveness of shotgun proteomic analysis. We briefly review the contribution of separation techniques and methods to shotgun proteomics analysis, covering the following three aspects: onedimensional separation; multidimensional separation; and, automated proteome-analysis systems.

2. One-dimensional separation RPLC coupled on-line with electrospray MS2 is typically used for shotgun proteomic analysis because of the good compatibility of the mobile phase with MS detection. The 75-lm · 12-cm capillary column packed with 5-lm porous C18 particles was often used for nanoflow RPLCMS2analysis. Although relatively complex mixtures can be separated well in RPLC, the analysis of mixtures in shotgun proteomic experiments containing thousands of peptides, which is extremely complex, has given rise to urgent demand for development of an RP capillary column with higher resolution. Reducing the diameter of chromatographic packing materials and increasing the column length are the most effective ways to achieve highly efficient separations. To take advantage of long columns packed with small particles, ultra-performance liquid chromatography (UPLC) has been developed by several research groups [3–5]. Highly efficient analysis of BSA digest by UPLC-MS2 was demonstrated by Tolley et al. [6] using a 22-cm · 150-lm column packed with 1.5-lm C18 nonporous particles with applied pressures varied from 790 bar (11,500 psi) to 930 bar (13,500 psi). Shen et al. reported using an ultra-long column in UPLC for proteome analysis [7,8]. Fused silica capillaries with length of 87 cm were packed with 3-lm C18 porous silica particles for separation of a proteolytic digest of soluble yeast proteins. Peak capacity up to about 1000 was achieved at a back pressure of about 10,000 psi. LC efficiencies achieved with one-dimensional separation in UPLC provided good proteome coverage relative to the use of moderate-efficiency LC involving two-dimensional separations. With improved separation, the co-eluted peptides decrease, so the ionization suppression in MS is alleviated and that leads to improvement of detection sensitivity. Another effective way to improve detection sensitivity is to reduce the inner diameter of the separation column as an electrospray mass spectrometer is a concentrationdependent detector. Proteome analysis using a narrowi.d. separation column has been shown to increase the

Trends

detection sensitivity significantly [7–10]. For example, Haskins et al. [9] used a 25-lm i.d. capillary LC column with ion trap MS2 to enable identification of peptides at the 60-amol level, and Shen et al. [10] used a 15-lm i.d. capillary LC column to enable identification and confirmation of six tryptic peptides from only 7 amol of a BSA tryptic digest sample. The use of ultra-long and ultra-narrow packed C18 capillary columns offers the advantages of high peak capacity, high detection sensitivity, and low sample and mobile-phase consumption. However, special highpressure pumping equipment is required to operate these systems, and it is extremely difficult to pack capillary columns with i.d. less than 30 lm. Monolithic columns have attracted a great deal of interest because of their ease of preparation, reliable performance, good permeability and versatile surface chemistry [11,12]. A 10-cm · 20-lm i.d. polystyrenedivinylbenzene monolithic column was reported to provide 10-amol sensitivity demonstrated by the detection of three peptides from a bovine catalase tryptic digest [13]. Although polymer-based monoliths have the advantages of good biocompatibility and wide application range of pH values, they also undergo shrinking or swelling in organic solvents and may contain micropores that adversely affect column efficiency and peak symmetry. Silica-based monolithic columns have demonstrated high efficiencies and low back pressure as their bimodal pore structures can be controlled independently (through pores and mesopores) [14,15]. Luo et al. [16] reported the preparation of extra-long octadecylated silica-based monolithic capillary columns for proteome analysis. The monolithic column with dimension of 70 cm · 20 lm i.d. could be operated at a mobile-phase pressure of 5000 psi providing a separation peak capacity of 420 and the detection sensitivity of 15 amol. By connecting with a replaceable emitter, monolithic capillary columns can be readily interfaced with electrospray ionization (ESI)-MS. Although using a separate emitter is convenient and flexible, the connection of the tapered ESI emitter via a union will increase extra-column volume and decrease separation efficiency. To circumvent this problem, we constructed a silica-based monolithic column with an integrated emitter [17]. Extremely sharp peaks were obtained for the separation of tryptic digest of BSA, and that indicated the high efficiency of the integrated column. This good performance was further demonstrated by using a 60-cm monolithic capillary column to analyze yeast proteome as it resulted in the identification of 1323 proteins. Recently, Luo et al. [18] developed 10-lm i.d. silicabased monolithic capillary columns providing even more sensitive proteomics measurements. The nL/min flow rates from the 10-lm i.d. monolithic columns minimize compound-to-compound variations in MS response and http://www.elsevier.com/locate/trac

81

Trends

Trends in Analytical Chemistry, Vol. 26, No. 1, 2007

provide the basis for more quantitative measurements from label-free analyses.

3. Multi-dimensional separation Typical proteomic samples are too complex to achieve sufficient fractionation using any of the current onedimensional separation techniques, so various combinations of separation methods have been employed to obtain multi-dimensional peptide-separation systems with sufficient separation power for comprehensive proteome analysis. These include size-exclusion chromatography followed by reversed-phase liquid chromatography (RPLC) [19], RPLC followed by capillary electrophoresis [20], strong cation-exchange chromatography (SCX) followed by RPLC [21–24], SCX followed by avidin affinity chromatography to select specifically biotinylated peptides, and followed by RPLC [25,26], and isoelectric focusing (IEF) followed by RPLC [27–31]. Of these techniques, SCX followed by RPLC is most widely used. The SCX-RPLC steps can be either carried out in tandem on-line [21,23] or with off-line fraction collection between the dimensions [22,24,26]. The online mode has the advantages of low sample consumption and complete automation, while the off-line mode has the advantages of high resolution and high sample capacity. Using a biphasic column is a reliable, convenient way to realize SCX-RPLC 2D separation [21,23,32]. The biphasic column was prepared by packing C18 and SCX particles sequentially into a fused-silica nanospray tip. The acidified peptide mixture was bound onto the SCX section. Discrete fractions of peptides were displaced from the SCX section directly onto the RP section, and then separated and eluted from the RP column into the mass spectrometer. A major advantage of this separation technology is that the entire system is coupled directly online with MS, enabling a large number of peptides to be directly identified in a high-throughput manner. Even though SCX is effective in separating complex peptide mixtures, it is not an ideal match with RPLC for 2D separations because SCX separations partially depend on peptide hydrophobicity [24,33], so SCX is not completely orthogonal to RPLC and the peak capacity of the SCX-RPLC combination is limited. Unlike SCX, IEF bases the separation of peptides only on the peptideÕs pI value, which is completely orthogonal to the hydrophobicity used in RPLC, so IEF is a good alternative to SCX as the first dimension in a 2D separation system for shotgun proteomics. Using IEF for separation of peptides has attracted a great deal of interest in recent years. Shen et al. [34] performed IEF separations in fused-silica capillaries and found that peptides with pI-value differences as small as 0.01 could be resolved. 82

http://www.elsevier.com/locate/trac

Capillary IEF coupled online with capillary RPLC was used for proteome analysis of yeast-cell lysate [29,30]. It was found that the overall peak capacity of this system was significantly improved over SCX-RPLC due to the completely orthogonal separation mechanisms of IEF and RPLC. While high resolution has been achieved, the practical application of capillary IEF in proteome research is seriously limited by the small loading capacity. Immobilized pH gradient (IPG) gels, typically used for IEF of proteins, have also been applied for the fractionation of peptides as the first dimensional separation in shotgun proteomics [27,28]. IEF with IPG provides highresolution separation, but its disadvantages are the tedious post-IEF sample processing that requires cutting the IPG gel strip into sections and extracting the peptides from gel matrix. In recent years, a series of solution-based IEF techniques, including a multi-chamber IEF device [31,33– 35], off-gel electrophoresis [36,37] and free flow electrophoresis [38,39], have been developed for the separation of peptides. These techniques circumvented the limitations of gel-based IEF and demonstrated the high resolution for the fractionation of peptides. A significant advantage of IEF over SCX for fractionation of peptides is that accurate peptide pI information could be obtained. It was reported that increased proteome coverage and improved peptide identification confidence could be achieved when the pI information was used to sort the results of database searching [27,37,38,40]. While multi-dimensional separations at the peptide level are very effective in solving the complexity problem, sample preparations at the protein level are more effective in solving the dynamic range problem. Several highabundance proteins in the proteome sample will generate hundreds of peptides after digestion, and that seriously suppresses the detection of peptides from low-abundance proteins. In order to identify more low-abundance proteins by the shotgun proteomics approach, depletion of these high-abundance proteins before proteolysis is highly recommended. To detect low-abundance proteins in plasma or serum, a series of affinity-based approaches was developed specifically to remove high-abundance plasma proteins, such as albumin, immunoglobulins, antitrypsin, transferrin and haptoglobin [41,42]. Besides affinity chromatography, multi-dimensional chromatographic prefractionation at the protein level is also effective in decreasing the interference of high-abundance proteins with the detection of low-abundance proteins, as these high-abundance proteins were only clustered in a few fractions after the fractionation. As an example, it was reported that a total of 1292 distinct proteins were identified from plasma by prefractionation of the plasma proteins by SCX-RP 2D chromatography followed with RPLC-MS2 analysis of the tryptic digest of each protein fraction [43].

Trends in Analytical Chemistry, Vol. 26, No. 1, 2007

Instead of removing high-abundance proteins, another effective approach to detecting low-abundance proteins is to enrich the low-abundance proteins. For example, affinity chromatography with immobilized p-aminobenzamidine (ABA), an inhibitor of trypsin-like serine proteases, was applied to enrich trypsin-like serine proteases in human plasma for shotgun proteome analysis [44].

4. Automated proteome-analysis system To avoid labor-intensive operations and to obtain highly reproducible results, it is necessary to automate the proteome-analysis system. Although the advantages of nanoflow RPLC-MS2 are apparent for shotgun proteome analysis, automation of sample introduction onto the analytical column remains a big challenge. In typical cases, the proteomic sample size ranges from a few microliter to 100 microliter. Due to the nanoliter (nL) flow rate adopted for separation of samples on the analytical column, it will take a long time if the sample is loaded directly onto the analytical column. To reduce the sampleloading time, a shorter, larger i.d. trap column may therefore be used. A large-volume peptide sample is first loaded onto the trap column at a fast rate in a short time, then the trapped peptides are eluted from the trap column to a reversed phase analytical column. The automation of this system can easily be realized by directly connecting the trap column with a switching valve [45–47]. However, even using a nanoflow switching valve, the dead volume attributed by the valve and the transfer lines is still significant and degrades the performance of the chromatography separation. A specially vented column system, where trap column and analytical column were directly connected via a microcross with an open/close switching valve, has automated sample introduction [48,49]. The dead volume significantly decreased due to the fact that the mobile phase for nanoflow RPLC separation does not pass through the switching valve. While a C18 trap column has been used in most automated sample-introduction systems, it was found recently that separation performance could be improved when an SCX trap column was used [50]. Because the peptide sample is retained on the entrance end of the separation C18 capillary column before the gradient started for separation, the dead volume before the capillary column hardly affected the separation adversely. The SCX trap column system was demonstrated to be a good substitute for the system using C18 trap column for automation of nanoflow RPLC-MS2. The use of a biphasic column is the simplest way to automate SCX-RP 2D separations for large-scale proteome analysis [32]. The system requires only one quaternary pump. Two channels of the pump generate a binary organic solvent (water/acetonitrile) gradient for

Trends

RP chromatography and another two channels generate a salt step gradient for SCX chromatography. Use of a volatile buffer permits the high salt solution to be flushed through the capillary biphasic column, which is coupled directly with MS. While the 2D separation of the system is automated, manual sample introduction is required, and that is labor intensive. The SCX trap-column system developed to automate sample introduction for nanoflow RPLC- MS2 can also be performed in 2D for large-scale analysis [50], which is fully automated, including sample introduction. The automation of SCX-RP 2D separations can also be achieved by column switching using separate pump systems for providing salt gradient [51]. The advantages of the column-switching approach are rapid sample loading, injection of large-volume samples and no introduction of salt into the mass spectrometer. Furthermore, column switching allows the independent control and optimization of the two dimensions. Besides separation, the automated proteome-analysis system should include sample-processing steps, such as sample clean-up and protein digestion. Using the immobilized enzyme reactor is a good way to realize the automation of protein digestion. It was reported that trypsin-based monolithic reactor with dimension of 50-mm · 4.6-mm i.d. was coupled on-line with an LC-MS2 system for automated protein digestion and protein identification [52]. However, the large size of the reactor is not suitable for use in a nanoflow LC-MS2 system. Recently, nL-scale monolithic microreactors were developed for rapid digestion of minute samples [53,54]. The microreactor coupled on-line with the nanoflow LC-MS2 system allowed fast proteome analysis of a minute, complex protein sample.

5. Conclusion In recent years, a series of new separation techniques has been applied successfully to proteome analysis and has provided dramatic improvement in protein identification, although complete coverage of proteins for any organism has not yet been accomplished. Analytical challenges remain in trying to resolve the problems of complexity and dynamic range. Besides the development of high-resolution separation techniques, the development of information-rich separation techniques, where physical properties of the peptides, such as pI values and hydrophobicity, could be obtained precisely, is preferable, as these properties could be used to sort the results of database searching to increase the confidence in protein identifications and to improve the coverage of proteomes. New specific approaches to remove high-abundance proteins and new approaches to enrich low-abundance proteins efficiently need to be developed to increase further the ability to detect more low-abundance proteins in http://www.elsevier.com/locate/trac

83

Trends

Trends in Analytical Chemistry, Vol. 26, No. 1, 2007

complex proteome samples. A fully automated proteomeanalysis system integrated with sample processing, peptide separation and detection, and data processing need to be established and optimized to improve the throughput of shotgun proteome analysis. Acknowledgements This work was supported by National Natural Sciences Foundation of China (No. 20327002, 20675081), the China State Key Basic Research Program Grant (2005CB522701), and the Knowledge Innovation program of DICP to H.Z. and National Natural Sciences Foundation of China (No. 20605022) to M.Y. are gratefully acknowledged. References [1] B. Domon, R. Aebersold, Science (Washington, DC) 312 (2006) 212. [2] H. Zhang, W. Yan, R. Aebersold, Curr. Opin. Chem. Biol. 8 (2004) 66. [3] J.A. Lippert, B.M. Xin, N.J. Wu, M.L. Lee, J. Microcol. Sep. 11 (1999) 631. [4] J.E. MacNair, K.C. Lewis, J.W. Jorgenson, Anal. Chem. 69 (1997) 983. [5] J.E. MacNair, K.D. Patel, J.W. Jorgenson, Anal. Chem. 71 (1999) 700. [6] L. Tolley, J.W. Jorgenson, M.A. Moseley, Anal. Chem. 73 (2001) 2985. [7] Y. Shen, R.J. Moore, R. Zhao, J. Blonder, D.L. Auberry, C. Masselon, L. Pasa-Tolic, K.K. Hixson, K.J. Auberry, R.D. Smith, Anal. Chem. 75 (2003) 3596. [8] Y. Shen, R. Zhao, S.J. Berger, G.A. Anderson, N. Rodriguez, R.D. Smith, Anal. Chem. 74 (2002) 4235. [9] W.E. Haskins, Z. Wang, C.J. Watson, R.R. Rostand, S.R. Witowski, D.H. Powell, R.T. Kennedy, Anal. Chem. 73 (2001) 5005. [10] Y. Shen, N. Tolic, C. Masselon, L. Pasa-Tolic, D.G. Camp 2nd, K.K. Hixson, R. Zhao, G.A. Anderson, R.D. Smith, Anal. Chem. 76 (2004) 144. [11] I. Gusev, X. Huang, C. Horvath, J. Chromatogr., A 855 (1999) 273. [12] H. Zou, X. Huang, M. Ye, Q. Luo, J. Chromatogr., A 954 (2002) 5. [13] A.R. Ivanov, L. Zang, B.L. Karger, Anal. Chem. 75 (2003) 5306. [14] H. Minakuchi, K. Nakanishi, N. Soga, N. Ishizuka, N. Tanaka, J. Chromatogr., A 762 (1997) 135. [15] K. Nakanishi, H. Shikata, N. Ishizuka, N. Koheiya, N. Soga, J. High Resolut. Chromatogr. 23 (2000) 106. [16] Q. Luo, Y. Shen, K.K. Hixson, R. Zhao, F. Yang, R.J. Moore, H.M. Mottaz, R.D. Smith, Anal. Chem. 77 (2005) 5028. [17] C. Xie, M. Ye, X. Jiang, W. Jin, H. Zou, Mol. Cell. Proteomics 5 (2006) 454. [18] Q. Luo, K. Tang, F. Yang, A. Elias, Y. Shen, R.J. Moore, R. Zhao, K.K. Hixson, S.S. Rossie, R.D. Smith, J. Proteome Res. 5 (2006) 1091. [19] G.J. Opiteck, J.W. Jorgenson, Anal. Chem. 69 (1997) 2283. [20] K.C. Lewisa, G.J. Opitecka, J.W. Jorgenson, D.M. Sheeley, J. Am. Soc. Mass Spectrom. 8 (1997) 495. [21] A.J. Link, J. Eng, D.M. Schieltz, E. Carmack, G.J. Mize, D.R. Morris, B.M. Garvik, J.R. Yates 3rd, Nat. Biotechnol. 17 (1999) 676. [22] Y. Shen, J.M. Jacobs, D.G. Camp 2nd, R. Fang, R.J. Moore, R.D. Smith, W. Xiao, R.W. Davis, R.G. Tompkins, Anal. Chem. 76 (2004) 1134.

84

http://www.elsevier.com/locate/trac

[23] M.P. Washburn, D. Wolters, J.R. Yates 3rd, Nat. Biotechnol. 19 (2001) 242. [24] J. Peng, J.E. Elias, C.C. Thoreen, L.J. Licklider, S.P. Gygi, J. Proteome Res. 2 (2003) 43. [25] S.P. Gygi, B. Rist, T.J. Griffin, J. Eng, R. Aebersold, J. Proteome Res. 1 (2002) 47. [26] D.K. Han, J. Eng, H. Zhou, R. Aebersold, Nat. Biotechnol. 19 (2001) 946. [27] B.J. Cargile, J.L. Bundy, T.W. Freeman, J.L. Stephenson Jr., J. Proteome Res. 3 (2004) 112. [28] B.J. Cargile, D.L. Talley, J.L. Stephenson Jr., Electrophoresis 25 (2004) 936. [29] J. Chen, B.M. Balgley, D.L. DeVoe, C.S. Lee, Anal. Chem. 75 (2003) 3145. [30] J. Chen, C.S. Lee, Y. Shen, R.D. Smith, E.H. Baehrecke, Electrophoresis 23 (2002) 3143. [31] Z. Xiao, T.P. Conrads, D.A. Lucas, G.M. Janini, C.F. Schaefer, K.H. Buetow, H.J. Issaq, T.D. Veenstra, Electrophoresis 25 (2004) 128. [32] D.A. Wolters, M.P. Washburn, J.R. Yates 3rd, Anal. Chem. 73 (2001) 5683. [33] J. Peng, S.P. Gygi, J. Mass Spectrom. 36 (2001) 1083. [34] Y. Shen, S.J. Berger, G.A. Anderson, R.D. Smith, Anal. Chem. 72 (2000) 2154. [35] A. Tan, A. Pashkova, L. Zang, F. Foret, B.L. Karger, Electrophoresis 23 (2002) 3599. [36] M. Heller, P.E. Michel, P. Morier, D. Crettaz, C. Wenz, J.D. Tissot, F. Reymond, J.S. Rossier, Electrophoresis 26 (2005) 1174. [37] M. Heller, M. Ye, P.E. Michel, P. Morier, D. Stalder, M.A. Junger, R. Aebersold, F. Reymond, J.S. Rossier, J. Proteome Res. 4 (2005) 2273. [38] J. Malmstrom, H. Lee, A.I. Nesvizhskii, D. Shteynberg, S. Mohanty, E. Brunner, M. Ye, G. Weber, C. Eckerskorn, R. Aebersold, J. Proteome Res. 5 (2006) 2241. [39] H. Xie, N.L. Rhodus, R.J. Griffin, J.V. Carlis, T.J. Griffin, Mol. Cell. Proteomics 4 (2005) 1826. [40] J. Krijgsveld, S. Gauci, W. Dormeyer, A.J. Heck, J. Proteome Res. 5 (2006) 1721. [41] H.J. Lee, E.Y. Lee, M.S. Kwon, Y.K. Paik, Curr. Opin. Chem. Biol. 10 (2006) 42. [42] H. Wang, S. Hanash, Mass Spectrom. Rev. 24 (2005) 413. [43] W.H. Jin, J. Dai, S.J. Li, Q.C. Xia, H.F. Zou, R. Zeng, J. Proteome Res. 4 (2005) 613. [44] R. Tian, X. Jiang, X. Li, X. Jiang, S. Feng, S. Xu, G. Han, M. Ye, H. Zou, J. Chromatogr., A 1134 (2006) 134. [45] D. Lin, D.L. Tabb, J.R. Yates 3rd, Biochim. Biophys. Acta 1646 (2003) 1. [46] J. Masuda, D.M. Maynard, M. Nishimura, T. Ueda, J.A. Kowalak, S.P. Markey, J. Chromatogr., A 1063 (2005) 57. [47] E. Nagele, M. Vollmer, P. Horth, J. Chromatogr., A 1009 (2003) 197. [48] L.J. Licklider, C.C. Thoreen, J. Peng, S.P. Gygi, Anal. Chem. 74 (2002) 3076. [49] E.C. Yi, H. Lee, R. Aebersold, D.R. Goodlett, Rapid Commun. Mass Spectrom. 17 (2003) 2093. [50] X.G. Jiang, S. Feng, R.J. Tian, G.H. Han, X.N. Jiang, M.L. Ye, H.F. Zou, Proteomics, accepted for publication. [51] G. Mitulovic, C. Sting, M. Smoluch, R. Swart, J.P. Chervet, I. Steinmacher, C. Gerner, K. Mechtler, Proteomics 4 (2004) 2545. [52] E. Calleri, C. Temporini, E. Perani, A. De Palma, D. Lubda, G. Mellerio, A. Sala, M. Galliano, G. Caccialanza, G. Massolini, J. Proteome Res. 4 (2005) 481. [53] S. Feng, M. Ye, X. Jiang, W. Jin, H. Zou, J. Proteome Res. 5 (2006) 422. [54] M. Ye, S. Hu, R.M. Schoenherr, N.J. Dovichi, Electrophoresis 25 (2004) 1319.