Lectin Affinity as an Approach to the Proteomic Analysis of Membrane Glycoproteins Dhiman Ghosh,†,‡ Oleg Krokhin,†,§ Mihaela Antonovici,†,‡ Werner Ens,†,§ Kenneth G. Standing,†,§ Ronald C. Beavis,†,| and John A. Wilkins*,†,‡ Manitoba Centre for Proteomics; Rheumatic Disease Research Laboratory, Department of Medicine; and Time-of-Flight Laboratory, Department of Physics and Astronomy, University of Manitoba, and Beavis Informatics Received March 12, 2004
The aim was to determine the proportion of membrane glycoproteins captured using concanavalin A or wheat germ agglutinin lectin affinity chromatography. Digests of the isolated proteins were separated by reversed-phase liquid chromatography and analyzed by matrix-assisted laser desorption tandem mass spectrometry. The two lectins identified different groups of proteins with a broad range of molecular mass and pI values, including a number of proteins that overlapped the two groups. Approximately 30% of the proteins were positively identified as containing domains that were predicted using standard bioinformatics methods to be characteristic of integral membrane proteins. This approach represents an effective method of surveying the membrane protein pool of mammalian cells for subsequent proteomic analysis. Keywords: membrane proteins • transmembrane • MALDI MS/MS • lectin affinity chromatography • Con A • WGA • glycoproteins
Introduction Integral membrane proteins are estimated to represent 2030% of all open reading frames in the eukaryotic genome. Our current state of knowledge regarding the expression and distribution of this important class of proteins is limited.1,2 Despite these limitations, membrane proteins and their ligands, especially components of the exofacial surface of the plasma membrane, represent the targets of 70% of new biotherapeutics currently in clinical trials.3 These observations highlight the need to develop better methods to examine mature membrane proteins in order to acquire a more complete understanding of membrane protein composition and function. There have been several recent advances in the characterization of membrane proteins that have addressed issues of solubility and separation.4-6 However, there are still several impediments to experimentally determining the correct identity and function of the membrane proteins. Conventional approaches to analyzing these proteins depend on the isolation of membranes with fractionation into their component pools.7,8 Even with surface biotinylation and subsequent membrane isolation these procedures often have low recoveries and the desired membranes may be contaminated with proteins from * To whom correspondence should be addressed. John A. Wilkins Ph.D., Manitoba Centre for Proteomics, RDU Research Laboratory, Room 805, John Buhler Research Centre, 715 McDermot Avenue, Winnipeg, Manitoba R3E 3P4, Canada. Phone: (204) 789-3835. Fax: (204) 789-3987. E-mail: jwilkin@ cc.umanitoba.ca. † Manitoba Centre for Proteomics. ‡ Rheumatic Disease Research Laboratory. § Time-of-Flight Laboratory, Department of Physics and Astronomy. | Beavis Informatics. 10.1021/pr049937f CCC: $27.50
2004 American Chemical Society
other membrane compartments or the cytosol.9 These requirements make it difficult, if not impossible, to study many normal cell types with such approaches because of the limited cell numbers. Thus, it was questioned if alternate strategies based on the properties of membrane proteins could be used to bypass some of the steps normally employed for membrane protein characterization. Membrane proteins are often extensively glycosylated, especially those that decorate the extracellular membrane.10-13 Similarly, those proteins that are components of re-circulating intracellular membrane pools such as endosomes and lysosmes are also heavily glycosylated.14-16 Affinity purification based on the properties of the glycoforms attached to these proteins has been used extensively to enrich membrane proteins. 17,18 However, there is relatively little information regarding the actual composition of these total glycoprotein pools. Lectins are proteins that specifically detect carbohydrate components.19,20 This specificity has provided the basis for probing different types of glycoconjugates in the cellular context.21,22 Lectin affinity chromatography has been used to purify glycoproteins and to examine their microheterogeneity to the level of a single sugar residue.23 In the present study, two different lectin affinity matrixes, concanavalin A (Con A) and wheat germ agglutinin (WGA) were used to capture glycoproteins with the ultimate goal of identifying cell surface proteins. The results demonstrate the utility of the approach in the capture of membrane proteins.
Materials and Methods Cell Culture and Preparation of Cell Lysates. K562 cells were grown to saturation in RPMI 1640 medium containing Journal of Proteome Research 2004, 3, 841-850
841
Published on Web 06/09/2004
research articles 10% FBS (Gibco). Cell viability as assessed by trypan blue exclusion was >98%. Between108-109 cells were used for different lectin affinity isolations. Cells were collected by centrifugation (200 × g, 5 min) and washed 3× in ice cold PBS. The packed cells were suspended in 5 volume of lysis buffer containing 10mM Tris-HCl, pH-7.5, 150 mM NaCl, 1% Nonidet P-40 and protease inhibitor cocktail (Roche) for 30 min and sonicated (ARTEK systems Corporation, USA) on ice for a total of 2 min in 15 s pulses with 15 s cooling periods after each pulse. The lysate was centrifuged at 14 000 rpm for 30 min at RT, and the supernatant was collected and stored at -20 °C if not used immediately. Cell lysates were prefiltered, 0.45 µm filter (Millipore), prior to loading them on to columns. Con A Chromatography. Before filtration, the cell lysate was diluted 1:1 with 2× of column equilibration buffer A (10mM Tris-HCl, 150mM NaCl, 1mM CaCl2 and 1mM MnCl2) and the pre-clarified cell lysate was loaded onto a 700 µL Con A Sepharose column pre equilibrated with 15 mL of buffer A. The loaded column was washed with 35 mL of buffer A and then with 6 mL of buffer B (10mM Tris-HCl, 1M NaCl, 1mM CaCl2, and 1mM MnCl2), bound material was eluted with buffer C (10mM Tris-HCl, 1M NaCl, 1mM CaCl2, 1mM MnCl2, and IM R-methyl-D mannopyranoside) until the baseline returned to zero OD at 280 nm. The proteins were concentrated using 30 kDa cutoff Centricon (Amicon) tube as per the manufacturer’s instruction. WGA Chromatography. The lysate was diluted 1:1 with 2× column equilibration buffer A (20mM Tris, 200mM KCl, 1mM CaCl2, 1mM MgCl2, pH-7.8), filtered and loaded onto a 800 µL WGA-agarose (Sigma) column equilibrated with 15 mL buffer A. The unbound material was removed by washing with 40 mL of column equilibration buffer and 6 mL of buffer B (20 mMTris, 1mM CaCl2, 1mM MgCl2, pH-7.8 and 1M KCl) thereafter, glycoproteins were eluted slowly by 10 mL of buffer C (20 mM Tris, 1 mM CaCl2, 1 mM MgCl2, pH-7.8, 1 M KCl and 0.8 M N-acetyl D-glucosamine). Concentration of proteins and removal of sugar were done simultaneously using a 30 kDa cutoff Centicon membrane (Amicon). Tryptic Digestion of Proteins. The concentrated lectinaffinity purified proteins (20-40µg) were resuspended in 100mM ammonium bicarbonate buffer and reduced with 10mM DTT at 57.5 °C for 1 h. The reaction mixture was immediately cooled on ice for 2 min and alkylated with iodoacetamide in the dark for 30 min. The mixture was dialyzed using 7 kDa membrane mini dialysis units (Pierce) in 100 mM ammonium bicarbonate buffer with several changes for 6-8 h at RT. After dialysis the proteins were digested with 1:50 Trypsin:Protein (w/w) ratio with sequence grade modified porcine trypsin (Promega) for 16 h at RT. The digested protein mixtures were vacuum-dried and dissolved in 5µl of 0.1% TFA. RP-HPLC. Chromatographic separation of peptide mixtures was carried out using an Agillent 1100 Series micro LC system. The reversed-phase separation was performed using 5 µL of sample injected onto a 0.15 × 150 mm column (Vydac 218 TP, C18-5µL). Peptides were eluted with a linear gradient of 1-46.7% acetonitrile (0.1%TFA) over a period of 75 min and ramped from 46.7 to 99% acetonitrile (0.1%TFA) over the final 5 min. The column eluent (4µL/min) was mixed online with DHB matrix solution of 0.15 mg/mL @0.5µL/min and deposited by an ‘in house’ produced spotting device onto a movable MALDI target at 1 min intervals and 80 such fractions were collected over a total period of 80 min. 842
Journal of Proteome Research • Vol. 3, No. 4, 2004
Ghosh et al.
Mass Spectrometry. Mass spectra of the HPLC separated tryptic fragments were acquired using a MALDI QqTOF mass spectrometer, developed at the Time-of-Flight laboratory, operated in single-MS or MS/MS mode as described by Krohkhin et al.24 Protein Identification. Peak assignments were made using Knexus automation software (Proteometrics, Canada) and peak lists were analyzed using the Global Proteome Machine (an open source protein identification system that uses the Tandem,25 available at http://www.thegpm.org/) to search the ENSEMBL human genome protein translation (NCBI 34b). Default GPM settings were used for the analysis (see Supporting Information for details). All of the proteins reported as being “identified” were determined to have an expectation value E
