Depletion of High-Abundant Proteins in Body Fluids Prior to Liquid Chromatography Fourier Transform Ion Cyclotron Resonance Mass Spectrometry Margareta Ramstro1 m,† Charlotte Hagman,‡ Jennifer K. Mitchell,§ Peter J. Derrick,§ Per Håkansson,‡ and Jonas Bergquist*,† Department of Chemistry, Analytical Chemistry, Uppsala University, P.O. Box 599, SE-751 24 Uppsala, Sweden, Division of Ion Physics, The Ångstro¨m Laboratory, Uppsala University, Uppsala, Sweden, and Chemistry Department, University of Warwick, Coventry, United Kingdom Received October 21, 2004
Today, proteomics is an exciting approach to discover potential biomarkers of different disorders. One challenge with proteomics experiments is the wide concentration range of proteins in various tissues and body fluids. The most abundant component in human body fluids, human serum albumin (HSA), is present at concentrations corresponding to approximately 50% of the total protein content in, e.g., plasma and cerebrospinal fluid (CSF). If this component could be selectively removed, then the chances of observing lower-abundance component of clinical interest would be greatly improved. There are today several approaches of varying specificity available for depletion. In this study, the properties of two commercially available kits, for the removal of HSA and HSA and immunoglobulin G (IgG), respectively, were compared, and the benefits of using depletion steps prior to on-line LC-FTICR MS were evaluated. Both methods were applied on plasma and CSF. To our knowledge, these are the first results reported for CSF. Also, the combination with electrospray LC-FTICR MS is novel. The proportion of depleted HSA and IgG was estimated using global labeling markers for peptide quantification. Both depletion-methods provided a significant reduction of HSA, and the identification of lower abundant components was clearly facilitated. A higher proportion of HSA was removed using the affinity-based removal kit, and consequently more proteins could be identified using this approach. Keywords: FTICR MS • LC • HSA • plasma • CSF • protein • QUEST-markers
Introduction Body fluids, such as plasma and CSF, are rich sources of biomolecules, and investigations of these proteomes may reveal correlations between protein expression and certain disorders. Characteristic changes in protein levels in plasma are indicative of many conditions including severe liver disease, hemolytic anemia and Downs syndrome.1 CSF is in direct contact with the extracellular space of the brain, and hence this body fluid has been analyzed when searching for biomarkers of neurodegenerative disorders. Relevant protein level alterations correlated with schizophrenia,2 Alzheimer’s disease,2-4 amyotrophic lateral sclerosis4 and Creutzfeldt-Jakob disease2,5 have been observed. The traditional proteomics methodology involves separation and visualization of proteins on twodimensional sodium dodecyl sulfate polyacrylamide gel electrophoresis (2D-PAGE),6,7 followed by mass spectrometric detection of the components. The dominating alternative approach is “shotgun” proteomics, which relies on liquid * To whom correspondence should be addressed. Phone: +46 18 471 3675. Fax: +46 18 471 3692. E-mail:
[email protected]. † Department of Chemistry, Analytical Chemistry, Uppsala University. ‡ Division of Ion Physics, The Ångstro¨m Laboratory, Uppsala University. § Chemistry Department, University of Warwick.
410
Journal of Proteome Research 2005, 4, 410-416
Published on Web 02/05/2005
separation of tryptic digested proteins in one or more dimensions followed by tandem mass spectrometry of the peptides.8,9 This method is generally faster, easier to automate, and the sample consumption is minor. In our laboratory, several proteomics studies have been performed on tryptic digested body fluids using on-line LC-FTICR MS.10-13 The high mass accuracy provided in FTICR MS permits the identification of proteins based on accurate mass measurement alone, and is beneficial when comparing protein patterns of two sample categories. A complication when doing proteomics, regardless of experimental approach, is the very wide concentration range in biological samples. The most abundant protein, human serum albumin (HSA), constitutes over 50% of the total protein content in plasma1 and around 45% in CSF.14,15 Removing this component from the fluid should make it possible to identify less abundant proteins. HSA-depletion would allow concentration of the remaining proteins before loading the sample onto the column. Also, ion suppression effects in the electrospray and signal suppression in the mass spectrometer would be significantly reduced.16 The challenge is to find a depletion method that removes HSA specifically but leaves the other proteins intact in the sample. Several possible strategies have 10.1021/pr049812a CCC: $30.25
2005 American Chemical Society
Depletion of High-Abundant Proteins in Body Fluids
been developed, of which the two most commonly used approaches are based on Cibacron Blue-Sepharose media,17 and monoclonal antibodies against HSA,18-20 respectively. Besides HSA, some antibody-based methods have been developed to simultaneously remove other highly abundant proteins, such as immunoglobulin G (IgG).18-20 Other solutions to the problem have been to apply iso-electric trapping,21 where proteins within a mass range close to that of HSA are excluded, and centrifugal ultrafiltration22,23 that separates molecules of molecular weight below and above a defined cutoff level. Neither of the latter approaches is specific against albumin and both deplete proteins with similar chemical properties to HSA. The aim of this study was to investigate the performance and reproducibility of two commercially available depletion kits combined with a LC-FTICR MS approach. The two kits have been developed for the removal of HSA or HSA and immunoglobulin G (IgG), respectively, from plasma and serum. Within the study, the methods were applied to plasma and CSF. To our knowledge, there are no results on CSF reported for either of these kits until now. The results were evaluated by comparing FTICR mass chromatograms from tryptic digests of depleted and native body fluid. The proportion of removed HSA and IgG was determined using Quantification-Using-EnhancedSignal-Tags (QUEST)-markers.24 This is a rather novel global labeling method for quantitative proteomics. It has previously been tested in combination with the FTICR MS on a standard protein digest,25 and its performance on complex mixtures, including CSF, was recently investigated.26
Materials and Methods Sample Handling. Plasma was obtained from a healthy blood donor. The CSF used in this study was taken from a pool consisting of >200 individual CSF samples drawn from patients in the age of 16-65 years. The samples were taken by lumbar puncture for diagnostic purposes, and the patients showed no signs of neurological or psychiatric disorders. Routine CSF analysis revealed no signs of inflammation or damage to the blood-brain barrier function.The study was approved by the Human Ethics Committee at the Faculty of Medicine, Sahlgrenska University Hospital, Go¨teborg, Sweden. Two commercially available depletion kits were investigated, the Montage Albumin Deplete Kit (Millipore Corporation, Bedford, USA) and the Albumin and IgG Removal Kit (Amersham Biosciences AB, Uppsala, Sweden). For simplicity, the two methods will be referred to as the HSA-depletion kit and the HSA/IgG-removal kit throughout this report. Both kits have been constructed for plasma and serum. The HSA-depletion kit consists of an affinity column placed in a spin tube. The composition of the affinity resin is proprietary, but known not to be based on monoclonal antibodies or Cibacron blue. In the experiments, 40 µL of plasma (containing ∼2.2 mg protein) was added to 160 µL of Buffer 1 (90% of the original HSA was removed in plasma and CSF. Around 95% of the immunoglobulin G in plasma was depleted using the HSA/IgG-removal kit, while not enough labeled pairs were found to determine the proportion of removed IgG in CSF, mostly due to the low abundance of IgG-fragments in the depleted samples. The proportion of IgG is lower in native CSF than in plasma, and hence the reduction of this protein resulted in a concentration below the detection limit. Journal of Proteome Research • Vol. 4, No. 2, 2005 413
research articles
Ramstro1 m et al.
Figure 2. Relative signal intensities of the identified tryptic peptides from selected proteins after affinity removal. The signals from mass chromatograms of the HSA- and the HSA/IgG-depleted samples are normalized to signals from the corresponding mass chromatograms of native CSF and plasma. The median values are given, error bars show the range of three experiments. Both depletion methods should improve the chances of detecting less abundant proteins in plasma and CSF.
Identification of Components in Depleted and Native Samples. The main focus of this study was to investigate how the removal of HSA and IgG would influence the chance of observing less abundant components in the fluids using an online LC-MS approach. LC-FTICR mass chromatograms were collected for tryptic digests of native plasma and CSF, and for the corresponding samples treated with the two depletion kits. In the experiments, the total protein content of all digests loaded on the LC-column was approximately the same. One way to compare the performances of the two depletion methods would be to compare the relative intensities of some components under study. The average peak intensities of some major components in plasma and CSF in the mass chromatograms of the depleted samples were normalized to the average intensities of the same proteins in native plasma and CSF (Figure 2). This should be regarded as a rough indication of the performance, as shown by the relatively large variation between samples, and so trends rather than specific numbers should be observed. The general trend is a reduction of the average signal intensity of HSA-fragments due to removal of this protein and an increase of the signal of other components. The average intensity of HSA is significantly reduced using the HSA/IgG-removal kit in the analysis of both samples, while the value for plasma using the HSA-depletion kit is close to unity. However, the depleted samples loaded onto the column corresponded to larger initial volumes of plasma and CSF, and hence a higher content of other proteins. This, together with reduced suppression from the abundant HSA-fragments16 improves the intensities of fragments from other proteins, such as transferrin and haptoglobin. The efficiency of the HSA/IgGremoval kit to remove IgG is also monitored in Figure 2. The number of identified proteins would be commonly regarded as the critical measure of the performance. Both plasma (or serum) and CSF represent very complex proteomes, and many powerful approaches have been tested for the analysis of these body fluids. Applying liquid chromatography combined with tandem mass spectrometry, Adkins et al.33 were able to identify 490 different proteins in serum. However, it should be noted that many of those proteins were identified from MS/MS data of a single peptide. The ambition in the present report was to get an indication of the possibility to observe proteins in plasma and CSF prior to and after deple414
Journal of Proteome Research • Vol. 4, No. 2, 2005
Table 3. Identified Proteins in Mass Chromatograms of Plasma, Plasma Depleted by the HSA-Depletion Kit and the HSA/IgG-Removal Kit, Respectivelya
a All sample categories were analyzed three times, and dark gray is used in the table to denote proteins observed in all three individual experiments, while light gray denotes the proteins identified in two of three runs. The white color indicates that the protein could not be identified in the sample. Both depletion methods clearly improved the possibility of detecting tryptic peptides from plasma proteins using LC-FTICR MS.
tion. For this purpose, experimental peptide masses were compared to theoretical masses from a small database of 150 human proteins. The depletion had been performed three times for each sample and hence three mass chromatograms were recorded for each sample category. The results are given in Tables 3 (plasma) and 4 (CSF). Proteins observed in three and two of the three individual runs are indicated by dark and light gray colors, respectively, while white cells denote that the protein could not be identified in the sample. For both plasma
research articles
Depletion of High-Abundant Proteins in Body Fluids Table 4. Proteins Identified in Native CSF, and CSF Treated with the HSA-Depletion Kit and the HSA/IgG-Removal Kit, Respectivelya
a
Dark gray denotes that the protein was identified in all three individual experiments of the same sample category, light gray denotes identification in two of three cases, while white means that the protein could not be identified. The HSA/IgG-removal kit significantly improved the chances of observing proteins in CSF, while the advantage of using the HSA-depletion kit was less pronounced.
and CSF, more proteins were identified in the depleted samples. The advantage was most apparent when analyzing the plasma samples. Generally, the HSA/IgG-removal kit seems to have more favorable performance than the HSA-depletion kit. As described earlier, this kit reduces the HSA content more efficiently and it also removes IgG, another abundant component in both fluids. Only a minor improvement was observed for CSF treated with the HSA-depletion kit. As described earlier, around 51% of the HSA was removed in this depletion-step. The reduced suppression effect from HSA peptides should be less apparent in this case. On the other hand, the result for the plasma sample depleted with the HSA-depletion kit is also very promising. The great majority of proteins found in native plasma and CSF were also identified in the depleted samples. Even though the albumin level was reduced, this molecule remains one of the significant hits in all samples. It is worth noticing that after 90% reduction, the concentration of HSA is on the same level as other abundant proteins, e.g., transferrin and R-1-antitrypsin in plasma and CSF. Twice as many proteins were observed in plasma and CSF after depletion with the HSA/IgG-removal kit, and many of the additional proteins found in these experiments are present at concentration levels one hundred times lower than those observed in native body fluid. The protein repertoire in CSF resembles that in plasma, even though the total protein concentration is 100-fold lower. As shown in this report, depletion kits developed for and tested on plasma and serum are also highly suitable for the treatment of CSF.
Conclusions Conducting proteomics on body fluids is a great challenge due to the wide concentration range of proteins. This study evaluates the performances of two commercially available kits for the depletion of HSA, and HSA and IgG, in combination with on-line LC-FTICR MS. The depletion-methods are rather quick and easy to perform. Little sample preparation needs to be done, small volumes are processed, and the eluate can easily be combined with tryptic digestion. The integration with liquid separation-based proteomics is therefore straightforward. Both kits provided a significant reduction of HSA. The HSA-depletion was, however, more efficient using the antibody-based HSA/ IgG-removal kit. In quantitative proteomics, the goal is often to identify changes in protein expression related to, e.g., a disease state. A recent study from our laboratory, where the LC-FTICRmethodology has been applied, deals with the comparison of the CSF proteome pattern of patients suffering from amyotrophic lateral sclerosis and healthy controls.13 In such experiments, it should be advantageous to deplete higher abundant components. When performing quantitative proteomics, it is of high importance that a depletion step, if included, is reproducible. The experiments indicate a high reproducibility when treating three aliquots of the same sample with the powerful depletion kits investigated in this study. However, there is always a risk of unspecific losses of components other than the desired ones, and one must consider that an extra source of variation has been introduced.
Acknowledgment. The authors want to thank Rita Persson, Sahlgrenska University hospital, Go¨teborg, for skillful assistance with the total protein determination. Financial support from Knut and Alice Wallenberg, the Swedish Foundation for Strategic Research, the Swedish Society for Medical Research and the Swedish Research Council (Grant 621-20025261, 629-2002-6821 (J.B.), K-1618/1999 (P.H.)) is gratefully acknowledged. Jonas Bergquist has a senior research position at the Swedish Research Council (VR). Jennifer Mitchell was visiting Uppsala through a Marie Curie Fellowship. References (1) Anderson, N. L.; Anderson, N. G. Mol. Cell Prot. 2002, 1, 845867. (2) Rohlff, C. Electrophoresis 2000, 21, 1227-1234. (3) Blennow, K.; Hampel, H. Lancet Neurol. 2003, 2, 605-613. (4) Rosengren, L.; Karlsson, J.-E.; Karlsson, J.-O.; Persson, L. I.; Wikkelsø, C. J. Neurochem. 1996, 67, 2013-2018. (5) Otto, M.; Wiltfang, J.; Tumani, H.; Zerr, I.; Lantsch, M.; Kornhuber, J.; Weber, T.; Kretzschmar, A.; Poser, S. Neurosci. Lett. 1997, 225, 210-212. (6) Klose, J. Humangenetik 1975, 26, 231-243. (7) O’Farrell, P. H. J. Biol. Chem. 1975, 250, 4007-4021. (8) Hancock, W. S.; Wu, S.-L.; Shieh, P. Proteomics 2002, 2, 352359. (9) MacCoss, M. J.; McDonald, W. H.; Saraf, A.; Sadygov, R.; Clark, J. M.; Tasto, J. J.; Gould, K. L.; Wolters, D.; Washburn, M.; Weiss, A.; Clark, J. I.; Yates, J. R. PNAS 2002, 99, 7900-7905. (10) Wu, S.-L.; Choudhary, G.; Ramstro¨m, M.; Bergquist, J.; Hancock, W. S. J. Proteome Res. 2003, 2, 383-393. (11) Ramstro¨m, M.; Palmblad, M.; Markides, K. E.; Håkansson, P.; Bergquist, J. Proteomics 2003, 3, 184-190. (12) Nilsson, S.; Ramstro¨m, M.; Palmblad, M.; Axelsson, O.; Bergquist, J. J. Proteome Res. 2004, 3, 884-889. (13) Ramstro¨m, M.; Ivonin, I.; Johansson, A.; Askmark, H.; Markides, K. E.; Zubarev, R. A.; Håkansson, P.; Aquilonius, S.-M.; Bergquist, J. Proteomics 2004, 4, 4010-4018. (14) Seyfert, S.; Kunzmann, V.; Schwertfeger, N.; Koch, H. C.; Faulstich, A. J. Neurol. 2002, 249, 1021-1026.
Journal of Proteome Research • Vol. 4, No. 2, 2005 415
research articles (15) Laretta, J.; Keep, R.; Betz, L. A.; Goldstein, G. W. In Basic Neurochemistry. Molecular, cellular and medical aspects; Siegel, G. J., Agranoff, B. W., Albers, R. W., Fisher, S. K., Uhler, M. D., Eds.; Lippincott Williams & Wilkins, 1999. (16) Sterner, J. L.; Johnston, M. V.; Nicol, G. R.; Ridge, D. P. J. Mass Spectrom 2000, 35, 385-391. (17) Travis, J.; Bowen, J.; Tewksbury, D.; Johnson, D.; Pannell, R. Biochem. J. 1976, 157, 301-306. (18) Steel, L. F.; Trotter, M. G.; Nakajima, P. B.; Mattu, T. S.; Gonye, G.; Block, T. Mol. Cell Prot. 2003, 2, 262-270. (19) Wang, Y. Y.; Cheng, P.; Chan, D. W. Proteomics 2003, 3, 243248. (20) Greenough, C.; Jenkins, R. E.; Kitteringham, N. R.; Pirmohamed, M.; Park, B. K.; Pennington, S. R. Proteomics 2004, 4, 3107-3111. (21) Herbert, B.; Righetti, P. G. Electrophoresis 2000, 21, 3639-3648. (22) Tirumalai, R. S.; Chan, K. C.; Prieto, D. A.; Issaq, H. J.; Conrads, T. P.; Veenstra, T. D. Mol. Cell Prot. 2003, 2, 1096-1103. (23) Georgiou, H. M.; Rice, G. E.; Baker, M. S. Proteomics 2001, 1, 1503-1506. (24) Beardsley, R. L.; Reilly, J. P. J. Proteome Res. 2003, 2, 15-21. (25) Hagman, C.; Ramstro¨m, M.; Håkansson, P.; Bergquist, J. J. Proteome Res. 2004, 3, 587-594.
416
Journal of Proteome Research • Vol. 4, No. 2, 2005
Ramstro1 m et al. (26) Hagman, C.; Ramstro¨m, M.; Jansson, M.; James, P.; Håkansson, P.; Bergquist, J. J. Proteome Res. 2004, accepted for publication. (27) Palmblad, M.; Håkansson, K.; Håkansson, P.; Feng, X.; Cooper, H. J.; Giannakopulos, A. E.; Green, P. S.; Derrick, P. D. Eur. J. Mass Spectrom. 2000, 6, 267-275. (28) Nilsson, S.; Wetterhall, M.; Bergquist, J.; Nyholm, L.; Markides, K. E. Rapid Commun. Mass Spectrom. 2001, 15, 1997-2000. (29) Lollo, B. A.; Harvey, S.; Liao, J.; Stevens, A. C.; Wagenknecht, R.; Sayen, R.; Whaley, J.; Sajjadi, F. G. Electrophoresis 1999, 20, 854859. (30) Rothemund, D. L.; Locke, V. L.; Thomas, T. M.; Wasinger, V.; Rylatt, D. B. Proteomics 2003, 3, 279-287. (31) Chromy, B. A.; Gonzales, A. D.; Perkins, J.; Choi, M. W.; Corzett, M. H.; Chang, B. C.; Corzett, C. H.; McCutchen-Maloney, S. L. J. Proteome Res. 2004, 3, 1120-1127. (32) Sobek, S.; Adam, P. J. Neurol. 2002, 250, 371-372. (33) Adkins, J. N.; Varnum, S. M.; Auberry, K. J.; Moore, R. J.; Angell, N. H.; Smith, R. D.; Springer, D. L.; Pounds, J. G. Mol. Cell Prot. 2002, 1, 947-955.
PR049812A
