Anal. Chem. 2003, 75, 3751-3757
Capillary High-Performance Liquid Chromatography/Mass Spectrometric Analysis of Proteins from Affinity-Purified Plasma Membrane Yingxin Zhao,† Wei Zhang,† Michael A. White,‡ and Yingming Zhao*,†
Department of Biochemistry and Department of Cell Biology, University of Texas Southwestern Medical Center at Dallas, Dallas, Texas 75390-9038
Proteomics analysis of plasma membranes is a potentially powerful strategy for the discovery of proteins involved in membrane remodeling under diverse cellular environments and identification of disease-specific membrane markers. A key factor for successful analysis is the preparation of plasma membrane fractions with low contamination from subcellular organelles. Here we report the characterization of plasma membrane prepared by an affinity-purification method, which involves biotinylation of cell-surface proteins and subsequent affinity enrichment with strepavidin beads. Western blotting analysis showed this method was able to achieve a 1600-fold relative enrichment of plasma membrane versus mitochondria and a 400-fold relative enrichment versus endoplasmic reticulum, two major contaminants in plasma membrane fractions prepared by conventional ultracentrifugation methods. Capillary-HPLC/MS analysis of 30 µg of affinity-purified plasma membrane proteins led to the identification of 918 unique proteins, which include 16.4% integral plasma membrane proteins and 45.5% cytosol proteins (including 8.6% membrane-associated proteins). Notable among the identified membrane proteins include 30 members of ras superfamily, receptors (e.g., EGF receptor, integrins), and signaling molecules. The low number of endoplasmic reticulum and mitochondria proteins (∼3.3% of the total) suggests the plasma membrane preparation has minimum contamination from these organelles. Given the importance of integral membrane proteins for drug design and membrane-associated proteins in the regulation cellular behaviors, the described approach will help expedite the characterization of plasma membrane subproteomes, identify signaling molecules, and discover therapeutic membrane-protein targets in diseases. In the past few years, expression proteomics has evolved into a powerful method to reveal protein dynamics in response to changes in cellular environments.1,2 Comparative protein expres* Corresponding author. E-mail: [email protected]
. Fax: (214) 648-2797. Tel: (214) 648-7947. † Department of Biochemistry. ‡ Department of Cell Biology. (1) Aebersold, R.; Goodlett, D. R. Chem. Rev. 2001, 101, 269-295. 10.1021/ac034184m CCC: $25.00 Published on Web 06/24/2003
© 2003 American Chemical Society
sion analysis is traditionally performed by 2D-gel electrophoresis.3,4 Over 2500 unique protein spots can be separated in one gel, making this one of the most powerful single-separation methods.5 Quantification of protein expression is achieved by comparing staining intensities of the protein spots in 2D-gel images. Proteins of interest are subsequently identified by highly sensitive mass spectrometry methods. An alternative approach for quantification of protein expression is isotope-coded affinity tag (ICAT) introduced by Aebersold, Gygi, and colleagues.6 In this method, two pools of proteins, labeled with light and heavy biotin, respectively, are chemically identical and therefore serve as a good internal standard for accurate quantification. The method has been applied to quantification of microsomal proteins in differentiated versus undifferentiated HL60 cells and quantification of protein expression in rat myc-null cells versus myc-plus cells.7,8 Rapid evolution of 2D-gel methods and HPLC/MS methods in the past few years expands our arsenal to attack various problems of proteomics analysis. Nevertheless, the number of proteins expressed and the wide dynamic range of protein expression levels in a cell still exceed the performance of current proteomics methods. Therefore, current proteomics technologies have difficulty detecting low-abundant or even medium-abundant proteins when whole-cell protein lysate is used as starting material.2 To overcome the difficulties, several methods were proposed that aim to reduce complexity of proteomics and focus the analysis on subproteomes of interest.9-17 (2) Lewis, T. S.; Hunt, J. B.; Aveline, L. D.; Jonscher, K. R.; Louie, D. F.; Yeh, J. M.; Nahreini, T. S.; Resing, K. A.; Ahn, N. G. Mol. Cell. 2000, 6, 13431354. (3) Klose, J.; Kobalz, U. Electrophoresis 1995, 16, 1034-1059. (4) O’Farrell, P. H. J. Biol. Chem. 1975, 250, 4007-4021. (5) Celis, J. E.; Gromov, P. Curr. Opin. Biotechnol. 1999, 10, 16-21. (6) Gygi, S. P.; Rist, B.; Gerber, S. A.; Turecek, F.; Gelb, M. H.; Aebersold, R. Nat. Biotechnol. 1999, 17, 994-999. (7) Shiio, Y.; Donohoe, S.; Yi, E. C.; Goodlett, D. R.; Aebersold, R.; Eisenman, R. N. EMBO J. 2002, 21, 5088-5096. (8) Han, D. K.; Eng, J.; Zhou, H.; Aebersold, R. Nat. Biotechnol. 2001, 19, 946951. (9) Oda, Y.; Nagasu, T.; Chait, B. T. Nat. Biotechnol. 2001, 19, 379-382. (10) Zhou, H.; Watts, J. D.; Aebersold, R. Nat. Biotechnol. 2001, 19, 375-378. (11) Ficarro, S. B.; McCleland, M. L.; Stukenberg, P. T.; Burke, D. J.; Ross, M. M.; Shabanowitz, J.; Hunt, D. F.; White, F. M. Nat. Biotechnol. 2002, 20, 301-305. (12) Goshe, M. B.; Conrads, T. P.; Panisko, E. A.; Angell, N. H.; Veenstra, T. D.; Smith, R. D. Anal. Chem 2001, 73, 2578-2586.
Analytical Chemistry, Vol. 75, No. 15, August 1, 2003 3751
The plasma membrane provides a physical boundary between the cell and its environment, playing important roles in material transport, signal relay, and cell-cell interaction. Definition of the dynamic plasma membrane protein profiles under different biological contexts helps elucidate regulatory pathways controlling cell behavior, providing molecular leads for further biological characterization and new drug design. Integral membrane and associated peripheral membrane proteins are major foci for drug targeting. Included in this category are G-protein coupled receptors, receptors for growth factors and cytokines, and receptorassociated signaling proteins.18,19 Despite its biological importance, analysis of membrane proteomes has proven challenging. The current 2D-gel separation method often results in under-representation of membrane proteins, largely due to low solubility of hydrophobic membrane proteins and the difficulty of focusing large membrane proteins (>100 kDa) in a 2D-gel. In addition, integral membrane proteins and signal molecules associated with plasma membrane are usually present in low relative abundance and are usually undetected by either 2D-gel/mass spectrometry-based or LC/ mass spectrometry-based proteomics methods. One strategy to address this problem is to reduce sample complexity by subcellular fractionation. Enrichment of plasma membrane proteins should facilitate the identification of low-abundance integral membrane proteins and membrane-associated regulatory proteins. Recently, we described an affinity purification method, which combines cell-surface biotinylation and affinity purification by immobilized streptavidin beads, for the isolation of plasma membranes.20 Cell-surface biotinylation occurs at lysine side chains (instead of free cysteines in the conventional ICAT chemistry) in the cell-surface proteins. We demonstrated that our procedure reduced contamination of subcellular organelles, and the biotinylation reaction did not activate or suppress ligand-dependent activation of receptor tyrosine kinases or G-protein coupled receptors. Thus, membrane fractions prepared by this method should provide a good material for the studies of plasma membrane dynamics associated with signal transduction. To directly test contamination of subcellular organelles, we carried out western blotting analysis and exhaustive protein identification from enriched plasma membrane. Western blotting analysis showed a >1600-fold relative enrichment of plasma membrane versus mitochondria and ∼400-fold relative enrichment versus endoplasmic reticulum (ER). HPLC/mass spectrometric analysis of enriched plasma membrane resulted in identification of 918 proteins, of which ∼3.3% were from mitochondria and ER, the major contaminants associated with the plasma membrane fraction prepared by the conventional ultracentrigation method. Therefore, both western blotting analysis and mass spectrometric (13) von Haller, P. D.; Donohoe, S.; Goodlett, D. R.; Aebersold, R.; Watts, J. D. Proteomics 2001, 1, 1010-1021. (14) Scheffler, N. K.; Miller, S. W.; Carroll, A. K.; Anderson, C.; Davis, R. E.; Ghosh, S. S.; Gibson, B. W. Mitochondrion 2001, 1, 161-179. (15) Simpson, R. J.; Connolly, L. M.; Eddes, J. S.; Pereira, J. J.; Moritz, R. L.; Reid, G. E. Electrophoresis 2000, 21, 1707-1732. (16) Blonder, J.; Goshe, M. B.; Moore, R. J.; Pasa-Tolic, L.; Masselon, C. D.; Lipton, M. S.; Smith, R. D. J. Proteome Res. 2002, 1, 351-360. (17) Goshe, M. B.; Veenstra, T. D.; Panisko, E. A.; Conrads, T. P.; Angell, N. H.; Smith, R. D. Anal. Chem 2002, 74, 607-616. (18) Drews, J. Science 2000, 287, 1960-1964. (19) Gibbs, J. B. Science 2000, 287, 1969-1973. (20) Zhang, W.; Zhou, G.; Zhao, Y.; White, M. A.; Zhao, Y., submitted.
Analytical Chemistry, Vol. 75, No. 15, August 1, 2003
analysis suggest the affinity-purified plasma membrane fraction is minimally contaminated with subcellular organelles, eliminating one of the major impediments for the analysis of plasma membrane proteome. EXPERIMENTAL SECTION Materials. Chemicals used in this study included ammonium bicarbonate and dithiothreitol (DTT) from Fisher Scientific (Fair Lawn, NJ); acetic acid from Aldrich Chemical Co., Inc. (Milwaukee, WI); Colloidal Blue staining kit from Invitrogen (Carlsbad, CA), Dynabeads M-280 streptavidin from Dynal Biotech ASA (Oslo, Norway), EZ-Link Sulfo-NHS-SS-Biotin [sulfosuccinimidyl2-(biotinamido)ethyl-1,3-dithiopropionate] (sulfo-NHS-SS-Biotin) from Pierce (Rockford, IL), HPLC-grade water, acetonitrile and methanol from EM Science (Gibbstown, NJ); protease inhibitor cocktail tablets from Roche Molecular Biochemicals (Indianapolis, IN); sequencing grade modified trypsin from Promega Co. (Madison, WI); trifluoroacetic acid from Fluka (Buchs, Switzerland);and Immobilon transfer membrane (PVDF) and ZipTipC18 from Millipore (Bredford, MA). Antibodies used in this study include the following: anti-NADH-ubiquinol oxydoreductase 39 monoclonal antibody from Molecular Probes (Eugene, OR); anticaveolin polyclonal antibody from Transduction Laboratories (Lexington, KY); anti-GRP94 polyclonal antibody from Santa Cruz Biotechnology (Santa Xruz, CA); and HRP conjugated anti-mouse and anti-rabbit IgG from Sigma-Aldrich (St. Louis, MO). Fusedsilica capillary tubing (75 µm i.d. × 190 µm o.d.) was purchased from Polymicro Technologies Inc. (Phoenix, AZ). Preparation of Affinity-Purified Plasma Membrane. Two dishes (15 cm) of A431 cells were grown in DMEM media supplemented with 10% fetal bovine serum and antibiotics until approaching confluency (∼95%), and the cells were washed with PBS buffer (37 °C) twice. To each dish was added 10 mL of PBS (37 °C) and 10 µL of sulfo-NHS-SS-Biotin stock solution (100 mg/ mL in dimethyl sulfoxide, 37 °C), and the resultant mixture was incubated at 37 °C for 5 min. The unreacted sulfo-NHS-SS-biotin was quenched by adding 100 µL of lysine solution (100 mg/mL) to the plate and incubating the plate at 37 °C for 10 min. PBS was removed, and the cells were washed twice with ice-cold PBS. The biotinylated cells were scraped into ice-cold PBS and centrifuged at 1000g for 5 min. The cell pellet was resuspended in 10 mL of ice-cold hypotonic buffer (10 mM HEPES, pH 7.5, 1.5 mM MgCl2, 10 mM KCl, 1× protease inhibitor cocktail, 1 mM Na3VO4, and 1 mM NaF) and incubated on ice for 15 min. The resulting cell suspension was dounce homogenized to lyse the cells (25 passes, Kontes Glass Co. (Vineland, NJ)). Unbroken cells and nuclei were removed by centrifugation of the cell homogenate at 1000g for 10 min at 4 °C. The supernatant was poured on the top of 50% sucrose (10 mM Tris-Cl, pH 7.5) and centrifuged at 46000g (Beckman, SW28 rotor) at 4 °C for 45 min. The crude plasma membrane band was collected from the centrifugation tube, washed with the hypotonic buffer, and pelleted again by centrifugation at 46000g (Beckman, SW28 rotor) for 15 min. The resulting membrane pellet was resuspended in 1 mL of hypotonic buffer and then incubated with 200 µL of Dynabeads M-280 streptavidin at 4 °C for 1 h. The beads were collected by magnetic plates (Polysciences Inc., Warrington, PA) and washed six times with hypotonic buffer to obtain the affinity-purified membrane fraction. The affinity-purified plasma membrane was
resuspended in ice-cold lysis buffer (7 M urea, 2 M thiourea, 4% CHAPS, 100 mM DTT, and 1× protease inhibitors cocktail) for 5 min and lysed on ice by 3-s sonication repeated three times. The lysate was pellet by TCA/acetone precipitation. Immunoblotting Analysis. The affinity-enriched membrane proteins were resolved in SDS-PAGE using 4-20% gradient minigel (Bio-Rad Laboratories, Hercules, CA). For immunoblotting analysis, the proteins were transferred to a PVDF membrane. The membrane was blocked with 5% dry milk in TBST (25 mM Tris pH 7.5, 150 mM NaCl, 0.1% Tween-20) for 1 h, incubated with the appropriate antibody in the blocking solution for 1 h, and followed by washing with TBST three times. The resulting membrane was incubated with horseradish peroxidase-conjugated secondary antibody at 1:5000 dilution in 5% dry milk/0.1% TBST at room temperature for 1 h and then washed with TBST three times. The blot was detected by Western Lighting Chemiluminescence reagent (Perkin-Elmer Life Sciences, Boston, MA). In-Gel Digestion of Proteins. The protein bands of interest were in-gel digested using a protocol modified from previous work.21,22 Briefly, the gel slice was cut into small particles (∼1 mm3) using a scalpel. The resulting gel particles were destained in 1 mL of water/methanol solution (50:50, v/v) containing 25 mM NH4HCO3 (pH 8.0) three times, with the solution changed every 10 min. The destained gel was washed in 1 mL of an acetic solution (acetic acid/methanol/water, 10:40:50, v/v/v) for 3 h, with the solution changed every 1 h. The resulting gel was soaked in 1 mL of water twice with the solvent changed every 20 min. The gel was then transferred into a 0.5-mL microcentrifuge tube and dehydrated by soaking the gel in 100% ACN until it became opaque white. The solution was removed and the gel was dried in a SpeedVac for 20-30 min. The dried gel was rehydrated with an adequate amount of trypsin digestion solution (10 ng of trypsin/µL in 50 mM NH4HCO3, pH 8.0). The solution should completely wet the entire gel, adding additional trypsin solution if necessary. The digestion was carried out at 37 °C overnight. To extract tryptic digest, the gel was soaked in 40 µL of extraction solution (ACN/ TFA/water, 50:5:45, v/v/v) for 60 min with a vortex and the extraction solution was carefully removed with a gel-loading pipet tip. Extraction was repeated once. The extracts were pooled and dried with a SpeedVac. The extracted peptides were cleaned with ZipTipC18 (Millipore, Bredford, MA) as described by the manufacturer. Briefly, the ZipTip was pretreated with 10 µL of solution I (ACN/TFA, 99.9: 0.1, v/v) and 10 µL of solution II (ACN/TFA/water, 50.0:0.1:49.9, v/v/v) and then equilibrated with 10 µL of solution III (TFA/ water, 0.1:99.9, v/v). The extracted peptides were redissolved in 8 µL of solution III and allowed to bind to ZipTip by aspiration of the peptide solution into the ZipTip 10 times. The ZipTip was then washed with 10 µL of solution III five times to remove salts. The peptides were then eluted with 8 µL of solution II. The resulting peptide elute was dried in a SpeedVac and redissolved in 4 µL of HPLC buffer A solution (water/acetonitrile/acetic acid, 97.9:2.0: 0.1 (v/v/v)). Capillary-HPLC/Mass Spectrometric Analysis for Protein Identification. HPLC/MS was carried out in a LCQ Deca XP (21) Shevchenko, A.; Wilm, M.; Vorm, O.; Mann, M. Anal. Chem. 1996, 68, 850-858. (22) Zhang, X.; Herring, C. J.; Romano, P. R.; Szczepanowska, J.; Brzeska, H.; Hinnebusch, A. G.; Qin, J. Anal. Chem. 1998, 70, 2050-2059.
(ThermoFinnigan, Palo Alto, CA) coupled on-line with a capillary HPLC system (Agilent 1100 capillary pump, Agilent Technologies, San Jose, CA) and nanospray source. The HPLC capillary column was made in-house by packing 5-µm Luna C18 resin (Phenomenex, St. Torrance, CA) into a 75 µm i.d. × 190 µm o.d. fusedsilica capillary tubing (Polymicro Technologies Inc.) by an air bomb. A 2-µL aliquot of the tryptic peptide solution was loaded on the C18 capillary column (∼5-cm length, 75-µm i.d.). The peptides were sequentially eluted from the HPLC column with a gradient of 5-80% buffer B (acetonitrile/water/acetic acid, 90:9.9:0.1, v/v/ v) in buffer A (acetonitrile/water/acetic acid, 2:97.9:0.1, v/v/v). The flow rate was set at 0.6 µL/min during the loading and washing stages and was reduced to 0.1 µL/min right before the peptide elution from the column. The eluted peptides were sprayed directly from the tip of the capillary column to the LCQ mass spectrometer for MS analysis.23 The LCQ was operated in a datadependent mode where the machine measured intensity of all the ions in the mass range from 350 to 1200 (mass-to-charge ratios) and isolated the two most intense ion peaks for collision-induced dissociation using a relative collision energy level of 35%. Each cycle of analysis (one MS and two MS/MS analysis) took ∼4 s. Dynamic exclusion was enabled with a repeat count of 1, a repeat duration of 0.5 min, and a 0.5-min duration time. The resulting spectra were searched for protein candidates in the NCBI nonredundant protein sequence database with the program Sonar search engine (Genomics Solution Inc., Lansing, MI). RESULTS Plasma Membrane Affinity Purification. To estimate contamination from mitochondria and ER in the affinity-purified membrane fraction, NADH-ubiquinol oxydoreductase 39 (a mitochondria-specific protein) and GRP 94 (an ER-specific protein) were examined in the affinity-purified membrane fraction (40 µg, 1× loading) and in whole-cell lysate using miniature dilution curves (Figure 1). NADH-ubiquinol oxydoreductase 39 was reduced by more than 80-fold and GRP 94 was reduced by more than 20-fold in the affinity-purified membrane fraction as compared to whole-cell lysate (Figure 1A, B). In contrast, caveolin 1, a highly enriched plasma membrane protein,24 was 20-fold more abundant in the affinity-purified membrane fraction as compared to wholecell lysate (Figure 1C). Together, these results suggest that our experimental method is able to achieve a 1600-fold enrichment of plasma membrane proteins relative to mitochondrial proteins and a 400-fold enrichment to ER proteins. Capillary-HPLC/Mass Spectrometric Analysis of AffinityPurified Plasma Membrane Fraction. To identify the protein components in the affinity-purified membrane, 30 µg of protein extract of affinity-purified membrane fraction was resolved in 4-20% SDS-polyacrylamide gradient minigel (Figure 1D). The protein sample was alkylated by iodoacetamide and preheated at 50 °C for 5 min prior to SDS-PAGE. It was shown that this procedure helps prevent the aggregation of membrane proteins.25 (23) Gatlin, C. L.; Kleemann, G. R.; Hays, L. G.; Link, A. J.; Yates, J. R., III. Anal. Biochem. 1998, 263, 93-101. (24) Rothberg, K. G.; Heuser, J. E.; Donzell, W. C.; Ying, Y. S.; Glenney, J. R.; Anderson, R. G. Cell 1992, 68, 673-682. (25) Wong, S. K.; Slaughter, C.; Ruoho, A. E.; Ross, E. M. J. Biol. Chem. 1988, 263, 7925-7928.
Analytical Chemistry, Vol. 75, No. 15, August 1, 2003
Figure 1. Mitochondrial and ER contamination in affinity-purified plasma membrane. The protein lysate from affinity-purified plasma membrane and whole-cell lysate were separated in 4-20% SDS-PAGE, transferred to PVDF membrane, and probed with antibodies against organellespecific proteins: (A) anti-NADH-ubiquinol oxidoreductase 39 for mitochondria, (B) anti-GRP 94 for ER, and (C) anti-caveolin for plasma membrane. Lane WC, whole cell lystae; Lane PM, affinity-purified membrane fraction. The amounts of proteins loaded in each lane were indicated. (D) Thirty micrograms of proteins from affinity-purified plasma membrane was resolved in 4-20% minigel, and the gel was stained with Colloidal Blue.
Under our lysis conditions, biotinylated surface proteins would remain associated with the streptavidin beads, and the extracted proteins were not biotinylated. Forty-three gel slices with equal size in the mass range between 10 and 500 kDa were excised, in-gel digested, and extracted. Streptavidin beads were used to affinity-enrich the biotinylated membrane sheets; the membrane proteins were lysed by the lysis buffer. Tryptic peptides were cleaned in ZipTipC18 prior to capillary HPLC/mass spectrometric analysis. Under our experimental conditions, the sample-cleaning step helps increase the signal-to-noise ratios of peptides in mass spectrometry. Figure 2 shows HPLC/mass spectrometric analysis of tryptic peptides from gel slice 11. The flow rate of capillary HPLC was set at 0.6 µL/min during sample loading and washing and dropped to 0.1 µL/min when the peptides started eluting from the HPLC column. Low flow rate during sample elution helps maximize the number of MS/MS spectra while maintaining reasonable peak widths. Most of peptides were eluted within a window of ∼30 min. Theoretically, 30 min of elution time allows ∼700 MS/MS spectra under the conditions used. The resulting MS/MS data were searched against the NCBI nonredundant protein sequence database for protein identification using the Sonar search engine. Mass spectrometric analysis of gel slice 11 resulted in the identification of 10 proteins, which included 5 known integral plasma membrane proteins (EGF receptor, integrin β 4, CD 109, Na/K-ATPase catalytic subunit R, plexin B2), two plasma membrane associated proteins (clathin heavy chain, IQGAP1), an 3754
Analytical Chemistry, Vol. 75, No. 15, August 1, 2003
uncharacterized protein with sequence homology to the Na, K/Cl transporter, an uncharacterized protein with sequence homology to UDP-glucose ceramide glucosyltransferase 1, and one other unknown protein KIAA1522. Of these last three, both the UDPglucose ceramide glucosyltransferase 1 homologue and the Na, K,Cl transporter homologue contain putative transmembrane domains based on the Sosui prediction algorithm (http://sosui.proteome.bio.tuat.ac.jp/cgi-bin/sosui.cgi?/sosui_submit.html), suggesting they are most likely integral membrane proteins. The tryptic peptides from the other 42 gel slices were analyzed similarly. HPLC/MS/MS analysis of 43 gel slices led to identification of 918 proteins, of which 30% were found to match >2 MS/ MS spectra and the others matched only either 1 or 2 mass spectra, which were manually analyzed to ascertain the accuracy of protein identification. Summary of Identified Proteins. We analyzed the subcellular localizations of the identified proteins based on NCBI annotation. Prediction by sequence analysis and sequence comparison was not used to define subcellular localization due to limited accuracy. Included in 918 identified proteins are 230 integral membrane or membrane-associated proteins (25.0% of total, with 151 integral plasma membrane proteins and 79 membrane-associated proteins), 339 cytosol proteins (36.9%, excluding membrane-associated cytosolic proteins), 92 proteins from other organelles (10.0%, including 58 nuclear proteins, 17 ER proteins, 13 mitochondria proteins, and 4 proteins localized
Figure 2. Example of HPLC/MS/MS analysis. (A) Total ion current (TIC) chromatogram of a capillary-HPLC/MS of the tryptic digests of gel slice 11 (mass range 150-185 kDa in SDS-PAGE). MS (B) and MS/MS (D) analysis at retention time of 39.18 min identified the peptide LQQTYAALNSK, unique to IQ motif containing GTPase activating protein 1. MS (C) and MS/MS (E) analysis at retention time of 54.73 min identified the peptide NLQEILHGAVR from EGF receptor.
Figure 3. Pie diagram of the subcellular locations of 918 identified proteins. The subcellular locations were based on NCBI annotation.
in other organelles), and 257 unclassified proteins (28.0%, Figure 3). (1) Integral Membrane Proteins and Cytosolic Proteins. Integral membrane proteins contribute to 16.4% of the proteins identified. Included in the list are receptors (EGF receptor, protein tyrosine phosphatase, transferrin receptor, fas antigen, CD 44 receptor),
cell-cell adhesion proteins (integrins, cell adhesion molecules (CAMs), cadherins, metalloproteinases, tight junction proteins), channels (ion pump proteins), cell-surface markers (cell surface antigens, MHC proteins), traffic proteins (transporters, syntaxins, vesicle proteins), and ras family proteins. Thirty one members of Ras protein superfamily were identified in this analysis (Table 1). Without transmembrane domains, Ras family proteins are anchored to membranes by carboxy-terminal prenylation, often accompanied by palmitoylation or poly-basic domains.26 Ras family proteins play important roles in a variety of dynamic biological processes, including mitogenic responses.27 Interestingly, a number of Ras protein regulators (i.e., IQ GAP1, Rab6 GAP, Rab GDI, and Rho GDI) and effectors (i.e., Raf1, RalGDS, and AF6) were also identified. These proteins are generally cytosolic but are recruited to the plasma membrane upon mitogen stimulation and Ras activation.27 The detection of these (26) Reuther, G. W.; Der, C. J. Curr. Opin. Cell Biol. 2000, 12, 157-165. (27) Shields, J. M.; Pruitt, K.; McFall, A.; Shaub, A.; Der, C. J. Trends Cell Biol. 2000, 10, 147-154.
Analytical Chemistry, Vol. 75, No. 15, August 1, 2003
Table 1. Identified Proteins from Gel Slice 11a identified proteins epidermal growth factor receptor Na,K-ATPase subunit R CD109 integrin, β 4 plexin B2 clathrin heavy chain IQ GAP1 similar to UDP-glucose ceramide glucosyltransferase-like 1 similar to solute carrier family 12 KIAA1522 protein
no. of peptides
no. of TMD
134.2 112.8 161.6 202.2 205.3 191.5 189.1 177.1
17 1 1 5 3 1 9 1
NP_005219 NP_000692 NP_598000 NP_000204 BAA21571 A40573 NP_003861 NP_064505
plasma membrane plasma membrane plasma membrane plasma membrane plasma membrane membrane associated membrane associated unknown
2 9 2 2 2 0 0 1
a The protein names, theoretical masses, numbers of peptides identified, NCBI protein accession numbers, subcellular locations, and numbers of predicated transmembrane domains (TMDs) are listed.
Table 2. List of Ras Super Family Proteins Identified in Affinity-Purified Plasma Membranea identified proteins
no. of peptides
G protein, R 11 K-ras N-ras RAP1A RAP1B RAP2B RAS/TC21 RAL A RAL B Rac1 Rho12 RhoG RAB1B similar to RAB13 RAB1A RAB9-like protein RAB18, RAB35 Rab5 Rab7 Rab2 RAB5C RAB8 RAB5B Rab14 RAB11A RAB11B rab GDI ARF4 ARF3
42.1 21.4 21.2 21 20.8 20.5 23.3 24 23.4 23.4 21.7 21.2 22.1 22.1 22.6 22.7 22.9 23 23.4 23.4 23.5 23.5 23.6 23.7 23.9 24.4 24.5 50.6 20.5 20.6
>2 2 2 >2 >2 >2 >2 >2 1 1 1 2 >2 >2 >2 1 2 >2 1 2 1 1 >2 2 1 2 2 1 2 2
NP_002058 AAH13572 NP_002515 NP_002875 NP_056461 NP_002877 TVHUC2 NP_005393 NP_002872 NP_061485 NP_001655 AAM21121 NP_112243 AAH09227 NP_004152 NP_057454 NP_067075 NP_006852 F34323 AAA86640 B34323 NP_004574 NP_005361 NP_002859 NP_057406 NP_004654 XP_058232 BAA03095 NP_001651 NP_001650
a b b b b b b b b c c c d d d d d d d d d d d d d d d d e e
a The protein names, theoretical masses, numbers of peptides identified, NCBI protein accession numbers, and the subfamily of ras proteins were indicated. b Notes: a, heterotrimeric G protein subfamily; b, ras subfamily; c, Rho subfamily; d, Rab subfamily; and e, Arf subfamily.
proteins in the plasma membrane fraction suggests that this protocol has the sensitivity that will be required for analysis of stimulus-dependent plasma membrane protein dynamics. The cytosolic proteins are another major group of proteins identified in affinity-purified plasma membrane, which include (1) 87 proteins involved in protein synthesis, 43 of which are ribosomal proteins, (2) 15 proteosome subunits, (3) 16 heat shock proteins, and (4) 62 signaling proteins. Seventy-nine of the proteins are membrane-associated. We did not have disrupted actin bundle 3756 Analytical Chemistry, Vol. 75, No. 15, August 1, 2003
in our experimental procedure. Many cytolic proteins could be anchored to the plasma membrane fraction through the actin bundle. (2) Unclassified Proteins. Two groups of proteins, those without documented information in the literature and NCBI annotation, and having complex expression patterns (e.g., localized in two or more organelles), were considered as unclassified proteins. Out of 257 unclassified proteins, 26 were predicted to contain transmembrane domain. Therefore, they are likely to be integral membrane proteins. Transmembrane domains were predicted by the Sosui prediction algorithm. (3) Proteins from Other Organelles. We believe that nuclei can be efficiently removed by one of the three steps described in our experimental protocol: (1) centrifugation at 1000g after homogenization, (2) ultracentrifugation at 46000g in 50% sucrose, and (3) affinity enrichment using streptavidin beads. It is unlikely for the nuclei to overcome all three purification steps and retain the plasma membrane fraction. We suppose the nuclear proteins identified in the plasma membrane preparation might come from leakage of the proteins from nuclei to cytosol during the homogenization step. Alternatively, the proteins could be dual localized in both the nucleus and cytosol, and existing literature only documented them as nuclear proteins. ER and mitochondria are the major contaminants in the plasma membrane prepared by the widely used ultracentrifugation method.15,28,29 In our experiment, only ∼3.3% of the total proteins identified in the plasma membrane preparation were shown to be either ER or mitochondria proteins. Affinity-purified plasma membrane seems to contain few other subcellular organelles. We only identified two Golgi proteins, one lysosome protein, and one peroxisome protein in our plasma membrane fraction. DISCUSSION Targeted enrichment of subcellular organelles or specific groups of proteins (e.g., phosphorylated proteins or glycosylated proteins) simplifies the complexity of a cellular proteome and reduces the dynamic range of protein expression. A major (28) Arnott, D., Kishiyama, A., Mohtashemi, I., Shillinglaw, W., Henzel, W., Stults, J. T., Eds. Proteomics via multidimensional separations: which dimensions to use?; American Society of Mass Spectrometry: Orlando, FL, 2002. (29) Lee, J.-s.; Park, Y.; Jeon, H.-K.; Lee, T. H.; Yoon, J.-B. Identification of follicular dendritic cell membrane proteome by LC/MS/MS; American Society of Mass Spectrometry: Orlando, FL, 2002.
complication for analysis of the plasma membrane subproteome is the heterogeneity of the plasma membrane preparation. Previous studies showed that the membrane fractions purified by ultracentrifugation methods are heavily contaminated with proteins from mitochondria and ER. This is most likely due to the density overlap between the organelles and plasma membrane.15,28,29 The protocol described here allows 1600- and 400-fold relative enrichment of plasma membrane versus mitochondria and ER, respectively, significantly reducing the contamination of the organelles. HPLC/MS/MS analysis of 30 µg of plasma membrane proteins identified 918 proteins, which include 151 integral plasma membrane proteins and 79 plasma membrane-associated proteins. Notable among the identified membrane proteins include 30 members of ras superfamily, receptors, G protein, cell-cell adhesion molecules, and transporters. Only 3.3% of total proteins from affinity-purified membrane fraction were found to be known ER or mitochondria proteins. Gibson and colleagues analyzed the mitochondrial proteome from SH-SY5Y neuroblastoma cell line by 2D-gel separation and subsequent protein identification by MALDI-TOF mass spectrometry.14 We compared the 918 proteins with the 15 most abundant proteins (HSP75, HSP 60, ATP synthase β chain, M69039, TAP 30, prohititin, ATPase subunit d, elongation factor Tu, fumarase, NAD(P), citrate synthease, malate dehydrogenase precursor, VDAC, D-prohibitin, 3-ketoacyl-CoAthiolase β subunit) from the mitochondria proteome14 and found that only HSP 75, HSP 60, and Tat-associated protein were present in the list of proteins identified in our plasma membrane fraction. HSP 75 and HSP 60 were shown dual localized in both cytosol and mitochondria.30,31 Tat protein was originally identified in HIV-1 virus and is involved in both transcriptional initiation and elongation.32 The Tatassociated protein (also called hyaluronic acid-binding proteins or gC1qR) has multiple functions and was shown to be located in the mitochondria, nucleus, and cytosol.32-34 Since 12 of the 15 most abundant proteins were not identified in our plasma membrane and the remaining 3 proteins were localized in cytosol, (30) Sconzo, G.; Amore, G.; Capra, G.; Giudice, G.; Cascino, D.; Ghersi, G. Biochem. Biophys. Res. Commun. 1997, 234, 24-29. (31) Gupta, S.; Knowlton, A. A. Circulation 2002, 106, 2727-2733. (32) Jeang, K. T.; Xiao, H.; Rich, E. A. J. Biol. Chem. 1999, 274, 28837-28840. (33) Robles-Flores, M.; Rendon-Huerta, E.; Gonzalez-Aguilar, H.; MendozaHernandez, G.; Islas, S.; Mendoza, V.; Ponce-Castaneda, M. V.; GonzalezMariscal, L.; Lopez-Casillas, F. J. Biol. Chem. 2002, 277, 5247-5255. (34) Krainer, A. R.; Mayeda, A.; Kozak, D.; Binns, G. Cell 1991, 66, 383-394. (35) Sun, H. Q.; Yamamoto, M.; Mejillano, M.; Yin, H. L. J. Biol. Chem. 1999, 274, 33179-33182. (36) He, W.; Melia, T. J.; Cowan, C. W.; Wensel, T. G. J. Biol. Chem. 2001, 276, 48961-48966. (37) Castle, J. D. In Current Protocols in Protein Science; Coligan, J. E., Dunn, B. M., Ploegh, H. L., Speicher, D. W., Wingfield, P. T., Eds.; John Wiley and Sons Inc.: New York, 1998; pp 4.2.42-44.42.43.
mitochondria, or nuclei, it is reasonable to assume that mitochondria contamination in our plasma membrane preparation is negligible. Since biotinlyation reaction does not induce or suppress the signal transduction pathways involved in G-protein coupled receptors and receptor tyrosine kinases,20 the reaction will most likely not interfere with other signaling pathways involving plasma membrane receptors. The described experimental procedure does not disrupt protein/protein interaction; therefore, natural interaction between peripheral membrane proteins and integral membrane proteins is not interrupted. We believe the affinity-purified membrane fractions will provide an excellent starting material for the analysis of membrane dynamics during the course of ligand challenge or for the identification of signaling molecules. If integral membrane proteins are the main target of the membrane proteomcis analysis, the protein mixture of affinitypurified plasma membrane can be further simplified by (1) depolymerization of the actin bundle with gelsolin,35 (2) striping peripheral membrane proteins by high-salt wash,36 and (3) striping the peripheral membrane proteins by alkaline buffer wash.37 This treatment will remove most of the cytosolic proteins including ribosomal proteins, heat shock proteins, proteosome components, and most structural proteins and make subsequent analysis much simpler (data not shown). In summary, we report an improved strategy for the analysis of plasma membrane proteins, which combines affinity purification of biotinylated plasma membrane, SDS-PAGE separation of the plasma membrane proteins, and subsequent HPLC/mass spectrometry for protein identification. A total of 918 proteins was uniquely identified, including many integral plasma membrane proteins that were not reported by previous proteomics analysis. HPLC/mass spectrometry analysis of the affinity-purified plasma membrane confirmed the results from western blotting analysis, suggesting the plasma membrane preparation has minimum ER and mitochondria contamination. Given the importance of integral membrane proteins and their associated proteins for drug design and membrane-associated proteins for the regulation of cellular behaviors, the described protocol would help expedite the discovery of therapeutic membrane protein targets in diseases and delineation of signal transduction pathways. ACKNOWLEDGMENT Supported by The Robert A. Welch Foundation (I-1550, I-1414) and NIH (CA 85146). Received for review February 24, 2003. Accepted May 12, 2003. AC034184M
Analytical Chemistry, Vol. 75, No. 15, August 1, 2003