Plasma Membrane Proteomics of Human Embryonic Stem Cells and

Plasma Membrane Proteomics of Human Embryonic Stem Cells and Human Embryonal Carcinoma Cells Wilma Dormeyer,‡ Dennis van Hoof,§ Stefan R. Braam,§ Albert J. R. Heck,‡ Christine L. Mummery,§ and Jeroen Krijgsveld*,‡ Biomolecular Mass Spectrometry and Proteomics Group, Bijvoet Center for Biomolecular Research and Utrecht Institute for Pharmaceutical Sciences, Utrecht University, Sorbonnelaan 16, 3584 CA Utrecht, The Netherlands, and Hubrecht Institute, Uppsalalaan 8, 3584 CT Utrecht, The Netherlands Received January 23, 2008

Human embryonic stem cells (hESCs) are of immense interest in regenerative medicine as they can self-renew indefinitely and can give rise to any adult cell type. Human embryonal carcinoma cells (hECCs) are the malignant counterparts of hESCs found in testis tumors. hESCs that have acquired chromosomal abnormalities in culture are essentially indistinguishable from hECC. Direct comparison of karyotypically normal hESCs with hECCs could lead to understanding differences between their mechanisms of growth control and contribute to implementing safe therapeutic use of stem cells without the development of germ cell cancer. While several comparisons of hECCs and hESCs have been reported, their cell surface proteomes are largely unknown, partly because plasma membrane proteomics is still a major challenge. Here, we present a strategy for the identification of plasma membrane proteins that has been optimized for application to the relatively small numbers of stem cells normally available, and that does not require tedious cell fractionation. The method led to the identification of 237 and 219 specific plasma membrane proteins in the hESC line HUES-7 and the hECC line NT2/D1, respectively. In addition to known stemnessassociated cell surface markers like ALP, CD9, and CTNNB, a large number of receptors, transporters, signal transducers, and cell-cell adhesion proteins were identified. Our study revealed that several Hedgehog and Wnt pathway members are differentially expressed in hESCs and hECCs including NPC1, FZD2, FZD6, FZD7, LRP6, and SEMA4D, which play a pivotal role in stem cell self-renewal and cancer growth. Various proteins encoded on chromosome 12p, duplicated in testicular cancer, were uniquely identified in hECCs. These included GAPDH, LDHB, YARS2, CLSTN3, CSDA, LRP6, NDUFA9, and NOL1, which are known to be upregulated in testicular cancer. Distinct HLA molecules were revealed on the surface of hESCs and hECCs, despite their low abundance. Results were compared with genomic and proteomic data sets reported previously for mouse ESCs, hECCs, and germ cell tumors. Our data provides a surface signature for HUES-7 and NT2/D1 cells and distinguishes normal hESCs from hECCs, helping explain their ‘benign’ versus ‘malignant’ nature. Keywords: embryonic stem cells • embryonic carcinoma cells • plasma membrane proteins • surface marker • protein enrichment • LC–MS/MS

Introduction Embryonic stem cells (ESCs) and embryonal carcinoma cells (ECCs) are opposite sides of the same coin1 (Figure 1). ESCs are derived from the inner cell mass (ICM) of blastocyst-stage embryos and have the ability to self-renew indefinitely.2–4 They are called pluripotent because they can give rise to any type of adult cell.5,6 However, they are not equivalent to the ICM, since upon transfer from the embryo to culture, they become subject to selective pressures and adapt. Several reports have shown that they may remain diploid and stable for extended periods or may acquire chromosomal abnormalities strikingly similar to those seen in hECCs.7 These commonly affect chromosomes * Corresponding author: [email protected]. § Hubrecht Institute. ‡ Utrecht University.

2936 Journal of Proteome Research 2008, 7, 2936–2951 Published on Web 05/20/2008

12 and 17. ECCs are the undifferentiated stem cells of teratocarcinomas, a subset of malignant germ cell tumors most common in the testis of young adults and also associated with abnormalities on chromosome 12 and 17.1,8 Like their nonmalignant ESC counterparts, ECCs are capable of self-renewal and are pluripotent, although in culture they usually differentiate into only a limited range of cell types.8–10 It is believed that germ cell carcinogenesis represents a caricature of normal embryogenesis, and that ESCs and ECCs share biological mechanisms that regulate their proliferation and differentiation.1,11–17Since the karyotypic changes associated with culture-adapted hESC and hECC may provide them with a growth advantage, it is important to understand what the molecular basis of this advantage might be. Comparison of ESCs and ECCs is therefore an efficacious way to elucidate 10.1021/pr800056j CCC: $40.75

 2008 American Chemical Society

Plasma Membrane Proteomics of hESC and hECC

research articles

Figure 1. Similarities of embryonic stem cells (ESCs) and embryonal carcinoma cells (ECCs). ECCs are the stem cells of teratocarcinomas, and the malignant counterparts of ESCs which are derived from the inner cell mass of blastocyst-stage embryos. Both cell types are capable of self-renewal and can give rise to ectodermal, endodermal, and mesodermal cell lineages. ESCs and ECCs are believed to share the biological mechanisms that regulate their proliferation and differentiation during embryogenesis and carcinogenesis.

the pathways that regulate self-renewal, growth control, and lineage commitment of pluripotent cells. Additionally, comparative studies of ESCs and ECCs may help to determine the factors that determine the ‘benign’ or ‘malignant’ nature of these two types of pluripotent cells. This is of special interest for the safe use of pluripotent cells in cell transplantation therapies and for the development of drugs that specifically target pluripotent cancer initiator cells.8,17 The behavior of ESCs and ECCs is tightly controlled by extrinsic and intrinsic factors;1,8 the comparison of their plasma membranes can therefore shed light on pathways that translate external factors into internal signals for self-renewal and differentiation induction. The method of choice for such a comparative profiling at the protein level is proteomics, but plasma membrane proteomics is presently challenging. The representation of membrane proteins in proteomic studies is unexpectedly low in general, even though it was estimated that 20-35% of eukaryotic genomes encode membrane proteins.18 The characterization of integral plasma membrane proteins in particular lags behind the analysis of cytoplasmic and organellar proteins. This has two major reasons: the isolation of the plasma membranes and the sample preparation of plasma membrane proteins for MS is far from optimal.19,20 The plasma membrane is the link between a cell and its environment, protects the cellular interior from external damage and fluctuating environmental conditions, regulates the transport of biomolecules and mediates communication with neighboring cells.21,22 To facilitate this variety of physiological processes, the plasma membrane is compartmentalized into functional microdomains.23,24 Their appearance can range from short-lived clusters of certain lipids and proteins to stable patches with inner suborganization of the components. Accordingly, the number and type of plasma membrane proteins identified in proteomic studies is highly dependent on the isolation technique.25 This is reflected in the multitude of reported isolation methods and the dissenting results of subsequent proteomic analyses. The outcome of such analyses represents parts lists of certain plasma membrane microdomains rather than general plasma membrane protein profiles. Reported isolation methods are mainly based on density

gradients,26,27 differential ultracentrifugation,28,29 two-phase partitioning,30–32 and sequential detergent extraction,33,34 or on the affinity purification of in vivo or in vitro modified plasma membrane proteins.35–38 The individual procedures are elaborate and must be adapted to the cells of interest in a timeconsuming manner. Their success largely depends on the availability of sufficient starting material, the compatibility of the studied cells with the isolation protocol, the quality of the purification, and the prevention of contamination by nonplasma membrane proteins. Integral membrane proteins, such as plasma membrane proteins, are amphipathic, that is, they contain hydrophilic and hydrophobic regions. The hydrophilic regions protrude into the cytosol or the extracellular environment for interaction with soluble proteins, whereas the hydrophobic regions are embedded into the membranous lipid bilayer. It is the inaccessibility of the proteins in the lipid bilayer as well as the hydrophobicity of their transmembrane domains that makes them so difficult to study by MS.19,20 Common proteomics approaches based on 1D- or 2D-PAGE and subsequent tryptic in-gel digestion are less suitable for membrane proteins, as they are prone to precipitation and insolubility during several steps of the procedure.39–41 Precipitation leads to sample loss and incomplete separation during PAGE, and insolubility impairs the efficiency of the digest. To overcome these problems, various detergent additives and adjusted PAGE protocols have been tested.42–44 Alternatively, in-solution digest of membrane proteins followed by multidimensional separation of the resultant complex peptide mixtures has been performed.45,46 The latter approach was originally developed for whole cell lysates, but several membrane-specific alterations of the procedure have been reported.20,47–54 Here, we present a comparative analysis of the plasma membrane proteomes of human ESCs and ECCs. The technical aim of our study was to improve the identification of plasma membrane proteins by MS, because previously reported plasma membrane-directed approaches are less suitable for more challenging cell types such as hESCs. Traditionally, hESCs are grown as small, multilayered colonies on selected mouse embryonic fibroblast feeder cell (MEF) layers or biologically Journal of Proteome Research • Vol. 7, No. 7, 2008 2937

research articles active matrix materials with growth factor supplements. Depending on their growth conditions, they are passaged mechanically or enzymatically, which limits the numbers of cells that can readily be harvested for shotgun or fractionation approaches.55 Cell surface-directed approaches such as biotinylation of the extracellular part of plasma membrane proteins are also not applicable, either because the biotinylation reagent can reach only the top layer of hESC colonies, or it might react nonspecifically with the MEFs or matrix material. We therefore tested various sample preparation and digestion procedures, and evaluated these with respect to the efficiency and quality of the subsequent MS analysis. We assembled an optimized protocol which is applicable to low amounts of starting material and to challenging cell types, but does not require tedious cell fractionation procedures. The biological aim of our study was to qualitatively compare the plasma membrane proteomes of hESCs (line HUES-7) and the hECCs (line NT2/D1). Using our optimized protocol, we were able to identify 237 and 219 plasma membrane proteins in the HUES-7 cells and the NT2/ D1 cells, respectively. These numbers reflect the benefit of our method, since they outperform genomic and proteomic data sets of ESCs and ECCs reported previously. Our results provide a plasma membrane protein signature for HUES-7 and NT2/ D1 cells and extend knowledge on the similarities and differences between normal hESCs and hECCs which might explain their common and distinct behaviors.

Experimental Procedures Cell Culture. hESCs, line HUES-756 (a kind gift from D. Melton, Harvard University), were cultured under feeder-free conditions on Matrix Growth Factor Reduced Matrigel (BD Biosciences)-coated flasks57 in MEF (strain CD1, 13.5 dpc)conditioned medium, which was prepared as described below, and passaged enzymatically using trypsin.58 For the preparation of MEF-conditioned medium, MEFs were cultured in D-MEM High Glucose (Invitrogen) supplemented with 10% fetal calf serum (FCS; BioWhittaker), 2 mM glutamine (Invitrogen), and 1% nonessential amino acids (Invitrogen). MEFs were mitotically inactivated for 2.5 h with 10 µg/mL mitomycin C (Sigma) at passage no. 5, and seeded at 6 × 104 cells/cm.2 Twenty-four hours after treating the MEFs with mitomycin C, their medium was replaced by D-MEM/F12 Glutamax (Invitrogen) supplemented with 15% KNOCKOUT Serum Replacement (Invitrogen), 100 µM β-mercaptoethanol (Invitrogen), 4 ng/mL basic fibroblast growth factor (basic FGF; Peprotech), and 1% nonessential amino acids. This MEF-conditioned hESC medium was collected every day for 7 consecutive days, and resupplemented with 4 ng/mL basic FGF after each collection. hECCs, line NT2/ D1, were cultured in D-MEM/F-12 Glutamax with 10% FCS (Invitrogen) and passaged enzymatically using trypsin-EDTA (Invitrogen) every 2-3 days. hESCs (passage 22-26) and hECCs (passage 39-41) were collected for plasma membrane protein extraction and Western blotting or fixed for immunofluorescence as described below. Preparation of the Membrane-Enriched Fraction of hESCs and hECCs. The hESCs (line HUES-7) and hECCs (line NT2/ D1) were washed twice with PBS and harvested by scraping. Harvested cells were pelleted at 1100g for 5 min at room temperature, resuspended in lysis buffer (50 mM Tris, pH 7.8, 250 mM sucrose, and 2 mM EDTA) with protease inhibitor cocktail (Roche Diagnostics, Switzerland), and incubated on ice for 10 min. Cells were lysed by 30 passes through a 301/2 gauge needle at 4 °C. Cell debris, unbroken nuclei, ER mem2938

Journal of Proteome Research • Vol. 7, No. 7, 2008

Dormeyer et al. branes, and mitochondrial membranes were removed by centrifugation at 1000g for 10 min at 4 °C. The postnuclear supernatant was layered onto a 60% sucrose cushion and centrifuged at 160 000g for 1 h at 4 °C in a MLS50 rotor (Optima, Beckman-Coulter, Netherlands). The membrane fraction on top of the sucrose cushion was collected, diluted 1:2 with cold 50 mM Tris, pH 7.8, and pelleted at 100 000g for 1 h at 4 °C in a TLA55 rotor (Optima, Beckman-Coulter, Netherlands). The supernant was discarded and the membrane pellet was washed for 1 h at 4 °C with 100 mM Na2CO3, pH 11.5, rinsed twice with cold H2O, and pelleted at 20 000g for 30 min at 4 °C. Digestion of the Membrane-Enriched Fraction of hESCs and hECCs. The four sets of experiments performed for the optimization of the sample preparation and digestion of plasma membrane proteins for MS are described in the text and figures. Washes were performed with 100 mM Na2CO3, pH 11.5, for 1 h at 4 °C. Trypsin predigests were performed overnight at 37 °C with 0.1 µg of trypsin per 1 × 105 cells in 25 mM NH4HCO3, pH 8.0. After washes or predigests, the membranes were rinsed twice with cold H2O and pelleted at 20 000g for 30 min at 4 °C. For deglycosylation membrane pellets were incubated for 10 min at 95 °C in denaturation solution (20 mM NH4HCO3, pH 8.0, 0.2% SDS, and 135 mM β-mercaptoethanol) and cooled down on ice. A total of 0.6 U PNGase F was added and deglycosylation was performed for 2 h at 37 °C. After deglycosylation, the membranes were washed with cold H2O and pelleted at 20 000g for 30 min at 4 °C. For delipidation, acetone/ methanol 8:1 was added to the membrane pellets and membrane proteins were precipitated for 2 h at -20 °C. After delipidation, the membrane proteins were washed with cold H2O and pelleted at 20 000g for 30 min at 4 °C. For tube gel digests, the membrane pellets were solubulized in 2% SDS and 25 mM NH4HCO3, pH 8.0, for 1 h at 4 °C. The solution was casted directly into a 10% tube gel. The gel was cut into pieces and thoroughly washed with 50% acetonitrile (ACN) and H2O, and dehydrated with ACN. After reduction and alkylation of reactive cysteines, in-gel digestion with trypsin was performed overnight at 37 °C. For the final comparison of 5 × 105 hESCs and hECCs, the membrane-enriched fractions were incubated for 10 min at 95 °C in denaturation solution (20 mM NH4HCO3, pH 8.0, 0.2% SDS, and 135 mM β-mercaptoethanol). After cooling down, 0.5 U PNGase F (Sigma-Aldrich, Germany) was added and deglycosylation was performed for 2 h at 37 °C. The deglycosylated protein pellet was washed with ice-cold H2O, pelleted at 20 000g for 30 min at 4 °C, and dissolved in 8 M urea in 50 mM NH4HCO3, pH 8.0. Reduction and subsequent alkylation were performed with 45 mM DTT for 30 min at 56 °C and 100 mM iodoacetamide for 30 min at room temperature in the dark, respectively. A total of 0.5 µg of endoproteinase Lys-C (RocheDiagnostics, Switzerland) was added and digestion was performed overnight at 37 °C. After dilution to 2 M urea with 50 mM 100 mM NH4HCO3, 0.5 µg of trypsin (Roche-Diagnostics, Switzerland) was added and digestion was performed for 8 h. The remaining membranes were sedimented at 20 000g for 30 min at 4 °C and the supernatant was stored at -80 °C. The pellet was redissolved in 80% ACN and redigested with 0.5 µg of trypsin overnight at 37 °C. This second digest was pooled with the first followed by desalting and concentration using Aqua-C18 material (Phenomenex, CA) packed into a ZipTip (Millipore, MA). The eluate was dried in a vacuum centrifuge and reconstituted in 20% ACN and 0.05% formic acid.

research articles

Plasma Membrane Proteomics of hESC and hECC Strong Cation Exchange (SCX) Fractionation. SCX fractionation was performed on a system consisting of two Zorbax BioSCX-Series II columns (i.d., 0.8 mm; l, 50 mm; particle size, 3.5 µm), a Famos autosampler (LC packings, The Netherlands), a Shimadzu LC-9A binary pump, and a SPD-6A UV-detector (Shimadzu, Japan). In the first 10 min after injection, unbound material was washed from the column with 100% solvent A (0.05% formic acid in 8/2 (v/v) water/ACN, pH 3.0). The subsequent linear gradient increased with 1.3%/min solvent B (500 mM NaCl in 0.05% formic acid in 25% ACN, pH 3.0) with a flow rate of 50 µL/min. One-minute fractions of 50 µL volume were manually collected, dried in a vacuum centrifuge, and reconstituted in 0.1% acetic acid. NanoLC-MS/MS. The digests of the four sets of experiments performed for the optimization of the sample preparation and digestion of plasma membrane proteins for MS were analyzed by nanoLC-LTQ-MS using a 45 min LC gradient (Thermo, CA). The SCX fractions of the digests performed for the final comparison of hESCs and hECCs were analyzed by nanoLCLTQ-Orbitrap-MS using a 120 min LC gradient (Thermo, CA). Both systems were run with the following setup. An Agilent 1100 series LC system was equipped with a 10 mm Aqua C18 (Phenomenex, Torrance, CA) trapping column (packed inhouse, i.d., 100 µm; resin, 5 µm) and a 254 mm ReproSil-Pur C18-AQ (Dr. Maisch GmbH, Ammerbuch, Germany) analytical column (packed in-house, i.d., 50 µm; resin, 3 µm). Trapping was performed at 5 µL/min for 10 min and elution was achieved with a gradient of 0-45% B in 45 or 120 min, respectively, 45-100% B in 1 min, 100% B for 4 min. The flow rate was passively split from 0.4 mL/min to 100 nL/min. Nanospray was achieved using a distally coated fused silica emitter (New Objective, Cambridge, MA) (o.d., 360 µm; i.d., 20 µm, tip i.d. 10 µm) biased to 1.8 kV. The mass spectrometer was operated in the data-dependent mode to automatically switch between MS and MS/MS. Survey full scan MS spectra were acquired from m/z 350 to m/z 1500 on the orbitrap with a resolution of R ) 60 000 at m/z 400 after accumulation to a target value of 500 000 with lock-mass. The two most intense ions were fragmented using collisionally induced dissociation at a target value of 10 000. Protein Identification. Spectra were processed with Bioworks 3.2 (Thermo, Bremen, Germany) using default settings, and the subsequent data analysis was carried out using the Mascot (version 2.1.0) software platform (Matrix Science, London, U.K.). Protein identification was performed by searching the IPI human database (release 3.24, 66 921 protein entries) with a precursor ion mass tolerance of 0.9 Da or 5 ppm for data generated by the LTQ linear ion trap or the LTQ Orbitrap, respectively. The fragment mass tolerance was set to 0.9 Da for both instruments. Carbamidomethylation of cysteines and oxidation of methionines were considered as fixed and variable modifications, respectively. For deglycosylated samples, deamidation of asparagines was allowed as variable modification. Fully proteolytic peptides produced by Lys-C, V8, trypsin, and/ or chymotrypsin with 1 maximum miscleavage were accepted. The ion score cutoffs were set to 25 for peptides and to 60 for proteins. Scaffold (version 01_07_00, Proteome Software, Inc., Portland, OR) was used to validate MS/MS based peptide and protein identifications. Peptide identifications were accepted if they could be established at greater than 90.0% probability as specified by the Peptide Prophet algorithm. Protein identifications were accepted if they could be established at greater than 90.0% probability and contained at least 2 identified

peptides. Proteins that contained similar peptides and could not be differentiated based on MS/MS analysis alone were grouped to satisfy the principles of parsimony. The rate of false positive identifications was estimated by searching the IPI human database with sequences reversed and were based on Mascot scores of 25 for peptides and 60 for proteins. For classification, the gene ontology (GO) symbols of the identified proteins were curated using the XRef database, and queried against the gene ontology database using the GoMiner tool (http://discover.nci.nih.gov/gominer/index.jsp). Prediction of the origin of proteins with unknown cellular localization was performed using the pTArget algorithm (http://bioinformatics. albany.edu/ptarget). Western Blotting. Cells were harvested from 75 cm2 polystyrene flasks at 80% confluency and resuspended in lysis buffer (8 M urea, 50 mM NH4HCO3 pH8.0, and protease inhibitor cocktail (Roche Diagnostics, Switzerland)). After addition of concentrated Laemmli buffer and denaturation for 10 min at 50 °C, equal amounts of ∼25 µg of protein were separated by SDS-PAGE on a 12.5% polyacrylamide gel. The separated proteins were transferred to PVDF membrane (Millipore) by Western blotting and the membrane was incubated with goat anti-FZD7 (1:200, Santa Cruz, sc-31061), rabbit anti-NPC1 (1: 200, Santa Cruz, sc-20152), rabbit anti-IGF1R (1:1000, Cell Signaling, #3027), rabbit anti-BST2 (1:1500, Abcam, Cambridge, U.K., ab-14694), mouse anti-Oct3/4 (1:500, Santa Cruz), or rabbit anti-CANX (1:1000, StressGen, Ann Arbor, MI). HRPcoupled secondary antibodies followed by ECL were used for visualization. Immunofluorescence. Immunofluorescence microscopy was conducted as described by Van Hoof et al.55 Briefly, cells were fixed for 30 min with 2% paraformaldehyde and incubated with mouse anti-Oc4 (Santa Cruz) 1:200, rabbit anti-BST2 (Abcam) 1:100, or mouse anti-Connexin43 (BD Transduction; a kind gift from Dr. MAG Van der Heyden, Medical Physiology, University Medical Center Utrecht, The Netherlands) 1:200, followed by DNA staining using TOPRO. The cells were imaged using a confocal laser scanning microscope type SP2 (Leica) to generate z-stacks and the separate layers were merged using Paint Shop Pro 9.

Results Organelle Fractionation and Plasma Membrane Preparation. A crucial issue of the comparison of the plasma membrane proteomes of hESCs (line HUES-7) and hECCs (line NT2/D1) is to adapt existing protocols to these cell types which have complex culture requirements and yield limited amount of material. We tested various organelle fractionation techniques that were based on the different solubilities or densities of the cell components. However, they resulted in the identification of low numbers of proteins in general and of membrane proteins in particular. For example, after solubility-based fractionation of 5 × 106 hECCs, only 93 proteins were identified, among which were 17 GO-annotated membrane proteins (Supplemental Figure 1A). Density-based fractionation on sucrose gradients suffered from high contamination with the endoplasmic reticulum (ER) as shown by Western blot analysis against the ER marker protein calnexin and the plasma membrane marker protein Na+/K+-ATPase (Supplemental Figure 1B). From one of the ATPase-containing sucrose gradient fractions, only 85 proteins were identified, among which were 6 annotated membrane proteins. We also tried to shave off the extracellular part of the plasma membrane proteins by incuJournal of Proteome Research • Vol. 7, No. 7, 2008 2939

research articles bating the intact cells in trypsin-containing buffers. For instance, after 5 min incubation of 5 million hECCs in trypsincontaining PBS and reincubation of the supernatant for 4 h at 37 °C, 40 proteins were identified, among which were 29 annotated membrane proteins (Supplemental Figure 1C). However, the cell shaving procedure could not be reproduced with hESCs since the cells lysed immediately in the trypsincontaining buffers, indicating a higher sensitivity for trypsin and thereby a noteworthy difference in the cell surface structure of hESCs and hECCs. We did not test biotinylation of the extracellular part of plasma membrane proteins because of the introductorily mentioned incompatibility of this approach with hESCs. Sample Preparation and Digestion of Plasma Membrane Proteins for MS. Since the plasma membrane-specific techniques failed to identify high numbers of proteins, especially from the hESCs, we focused on a less restricted approach. We used total membrane fractions prepared by soft mechanical lysis and ultracentifugation, and optimized the digest conditions for plasma membrane proteins within this crude membrane mixture. The comparisons were based on the absolute number and the relative number (percentage) of identified common membrane and plasma membrane proteins, respectively. However, since we sought to identify as many plasma membrane proteins of the hESCs and hECCs as possible, the optimized protocol was assembled with regard to the absolute number of identified plasma membrane proteins, only disregarding the commonly identified membrane proteins and their percentage. First, we tested the effect of the digestion buffer (Figure 2A). Four total membrane fractions of 6 × 105 NT2/D1 cells each were prepared and washed using 100 mM Na2CO3 and H2O. The digests were performed in solution with Lys-C in 8 M urea and subsequently by trypsin in 2 M urea (sample 1), in solution with trypsin in 2 M urea alone (sample 2), in solution with trypsin in 25 mM NH3HCO4 (sample 3), and in a tube-gel59 with trypsin in 25 mM NH3HCO4 (sample 4). The highest number of annotated plasma membrane proteins was identified after in-solution digest with Lys-C in 8 M urea and trypsin in 2 M urea (sample 1, 17 plasma membrane proteins) closely followed by tube-gel digestion with trypsin in 25 mM NH3HCO4 (sample 4, 13 plasma membrane proteins). In these two samples, the plasma membrane proteins accounted for 13% and 10%, respectively, of all identified proteins that could be annotated to cell components, while the absolute numbers of membrane proteins commonly identified were similar (40 and 41 membrane proteins). In contrast, the in-solution digests with trypsin in 2 M urea and trypsin in 25 mM NH3HCO4 were less effective, and led to the identification of a low number of proteins in general. Second, we examined the effect of the second enzyme in the sequential in-solution digestion (Figure 2B). Four total membrane fractions of 6 × 105 NT2/D1 cells each were prepared and washed using 100 mM Na2CO3 and H2O. As internal control to the preceding set of experiments, the first sequential digest was again performed in solution with Lys-C in 8 M urea and trypsin in 2 M urea (sample 5). The other three samples were digested in solution with trypsin in 2 M urea and V8 in 2 M urea (sample 6), with V8 in 8 M urea and trypsin in 2 M urea (sample 7) and with V8 in 4 M urea and trypsin in 2 M urea (sample 8). Once again, the highest number of annotated plasma membrane proteins was identified after digestion with Lys-C in 8 M urea and trypsin in 2 M urea (sample 5, 2940


Dormeyer et al. 45 plasma membrane proteins). Digestion with Lys-C and V8 or with V8 and trypsin, respectively, led to the identification of a significantly lower number of proteins. However, the plasma membrane proteins accounted for approximately 25% of the annotated cell component proteins in all samples. Third, we tested the effects of the Na2CO3-wash before the digestion, and of various solvents during the digestion (Figure 2C). Six total membrane fractions of 6 × 105 NT2/D1 cells each were prepared and either washed using 100 mM Na2CO3 and H2O (sample 9, 10, 14) or shortly predigested using trypsin in 25 mM NH3HCO4 (sample 11, 12, 13). As internal control to the two preceding sets of experiments, the first digest was again performed sequentially with Lys-C in 8 M urea and trypsin in 2 M urea (sample 9). The other two NH3HCO4-washed membrane fractions were digested with a combination of trypsin and chymotrypsin in 60% methanol (sample 10) or Lys-C in 6 M urea and trypsin in 2 M urea (sample 14). The three predigested membrane fractions were digested with a combination of trypsin and chymotrypsin in 60% methanol (sample 11), in 2 M urea (sample 12), or 80% ACN (sample 13), since digestion efficiency has been shown to be increased in organic solvents.54,60,61 Once more, the highest number of annotated plasma membrane proteins was identified after the NH3HCO4wash and the sequential digest with Lys-C and trypsin. Notably, the concentration of urea during the Lys-C digestion could be lowered from 8 to 6 M without loss of identifications (sample 9 and 14, 19 plasma membrane proteins). In these two samples, the plasma membrane proteins accounted for approximately 17% of the annotated cell component proteins, while they accounted for 13-38% (sample 10 to 13) after simultaneous digestion with trypsin and chymotrypsin in the various solvents. The plasma membrane proteins and peptides identified after digestion in 60% methanol or 80% ACN differed significantly from those identified after digestion in 2 M urea, indicating that the solubilization of membranes in different solvents leads to the identification of distinct data sets. Unfortunately, the total numbers of proteins identified after digestion in methanol, ACN, and 2 M urea were generally low (9 to 10 plasma membrane proteins). Furthermore, in a direct comparison, the NH3HCO4-wash perfomed better than the predigest in removing nonplasma membrane proteins but at the expense of the identification of a small number of plasma membrane proteins (six plasma membrane proteins in sample 10 versus nine in sample 11). Fourth, we examined the effects of deglycosylation or delipidation before the sequential digestion and after a third digestion step (Figure 2D). Eight total membrane fractions of 6 × 105 NT2/D1 cells each were prepared and washed using 100 mM Na2CO3 and H2O. As an internal control to the three preceding sets of experiments, the first digest was again performed sequentially with Lys-C in 6 M urea and trypsin in 2 M urea (sample 15 double). The second membrane fraction was digested in the same way, but a third digestion step with trypsin in 80% ACN was added (sample 15 triple). The other six membrane fractions were either deglycosylated, delipidated, or both, and subsequently double-digested using Lys-C in 6 M urea and trypsin in 2 M urea or triple-digested using Lys-C in 6 M urea, trypsin in 2 M urea and trypsin in 80% ACN. Both the deglycosylation and the third digestion step resulted in an increased number of annotated plasma membrane proteins (27 plasma membrane proteins in sample 15 triple and 16 double). In addition, they led to the detection of more peptides per protein, probably because formerly carbohydrate-protected


research articles

Figure 2. Optimization of the sample preparation and sequential digestion of plasma membrane proteins. hECCs were digested using the various protocols indicated in the table to test the effects of (A) the digestion buffer, (B) the enzyme combination in a sequential double digest, (C) the removal of nonmembrane proteins by washes and the solubilization of membrane proteins by solvents, and (D) the deglycosylation or delipidation of the proteins prior to sequential triple digestion. Note that samples 1, 5, 9, and 15d were prepared using the same protocol as internal control between the four sets of experiments. Each protocol was evaluated with regard to the absolute number (left axis, bars) and relative number (right axis, lines) of identified general membrane proteins (green) and plasma membrane proteins (blue). Relative number ) percentage of membrane or plasma membrane proteins of all GO-annotated proteins identified in the sample.

regions of the proteins were accessible for the proteases after deglycosylation, and because ACN was more effective than urea on the solubilization of certain proteins. Therefore, deglyco-

sylation and third digestion increased the sequence coverage and thereby the confidence of the identifications. In contrast, delipidation and the combination of deglycosylation and Journal of Proteome Research • Vol. 7, No. 7, 2008 2941

research articles

Dormeyer et al.

Figure 3. Strategy for plasma membrane proteomics of low amounts of hESCs and hECCs. Schematic representation of the optimized protocol used for the comparative identification of plasma membrane proteins from crude membrane fractions of hESCs and hECCs. See text for detailed description of the procedure.

delipidation showed no further positive effect. The highest number of plasma membrane proteins was identified after deglycosylation and sequential triple-digestion (35 plasma membrane proteins in sample 16 triple). To test the reproducibility of the method and the validity of our comparison, we prepared duplicates of the sequentially triple-digested sample with and without prior deglycosylation (samples 15 triple and 16 triple) and analyzed them on a mass spectrometer with high mass accuracy (LTQ-Orbitrap). The reproducibility of the identifications in the duplicates was high on both the protein and peptide level (Supplementary Table 1). Thirty-five and 36 plasma membrane proteins were identified without deglycosylation, and 44 and 45 with deglycosylation, and the identifications were based on the same sets of peptides. For example, the Voltage-dependent anion-selective channel protein 2 (VDAC2) was identified with the same 8 unique peptides in the duplicates performed with deglycosylation, resulting in a sequence coverage of 31%. Without deglycosylation, VDAC2 was identified with 8 or 9 unique peptides in the duplicates leading to a sequence coverage of 33% or 38%, respectively. Finally, we assembled an optimized protocol by selecting the conditions that in the four sets of experiments described above had proven beneficial for the digestion of plasma membrane proteins (Figure 3): the cells are washed with cold PBS and harvested by slow centrifugation. After a mild mechanical lysis, intact nuclei and ER are depleted by slow centrifugation, and the cellular membranes are harvested by ultracentrifugation onto a sucrose cushion. Soluble proteins that are loosely attached to the membranes are removed by washes with cold Na2CO3 and H2O, and deglycosylation is performed by hydrolysis of N-glycan chains by PNGase F. After reduction and alkylation of reactive cysteine residues, the proteins are cleaved by sequential triple digestion using Lys-C in 6 M urea, trypsin in 2 M urea, and finally trypsin in 80% ACN. Identification of Plasma Membrane Proteins in hESCs and hECCs by MS. The optimized protocol (Figure 3) was used for the analysis of 5 × 105 hESCs (HUES-7) and 5 × 105 hECCs (NT2/D1). The resulting complex peptide mixtures were separated by SCX fractionation, and 20 SCX fractions were analyzed by LC-MS/MS on a LTQ-Orbitrap-MS. Detailed information on the recorded MS/MS spectra and the Mascot database search results, including precursor mass and charge, protein and peptide scores, number of total and unique peptides per protein, peptide sequences, percent sequence coverage, and all MS/MS spectra with sequence annotation are avail2942


able at https://bioinformatics.chem.uu.nl/supplementary/ dormeyer_2008_JPR/. In total, 1077 and 1261 proteins were identified in hESCs and hECCs, respectively (Supplemental Table 2), using a Mascot protein score of at least 60, or a Peptide Prophet cutoff of 0.9 with a minimum of 2 peptides per protein. False positive rates were 1.2% and 2.3%, respectively, as determined by reversed database searches (Supplemental Table 2A and https://bioinformatics.chem.uu.nl/supplementary/dormeyer_2008_JPR/). Five hundred eighty-two of the proteins identified in the hESCs and 530 of the proteins identified in the hECCs were GO annotated as general membrane proteins of which 198 and 186 proteins, respectively, originated from the plasma membrane (Supplemental Table 2B,C). For the proteins that did not carry a GO annotation, we predicted the subcellular localization using their full-length sequences and the pTarget prediction algorithm. Thirty-nine and 33 proteins were predicted to localize to the plasma membrane of hESCs and hECCs, respectively (Supplemental Table 2B,C). No localization could be predicted for 9 in the hESCs and 3 proteins in the hECCs. The remaining 447 and 695 proteins identified in hESCs and hECCs, respectively, were classified as nonmembranous either via their GO annotation or by the localization prediction; that is, they originated from the nuclei or various cell organelles such as the ER and the mitochondria. In total, hESCs and hECCs had 725 proteins in common, 255 of which were annotated general membrane proteins, and 152 annotated and predicted plasma membrane proteins (Figure 4). They shared no protein of unknown localization and 318 that were classified as nonmembraneous (Figure 4 and Supplemental Table 3A,B). 129 membrane proteins and an additional 85 plasma membrane proteins were uniquely identified in hESCs (Supplemental Table 3C), while 89 of the proteins were uniquely identified in the hECCs originating from general membranes and 67 from the plasma membrane (Figure 4 and Supplemental Table 3D). To assess the similarities and differences between hESCs and hECCs further, we compared the functional and biological profiles of their plasma membranes by curating the molecular function and the biological process of the identified annotated plasma membrane proteins. We found that hESCs and hECCs dedicate the same proportion of proteins to molecular functions such as transport, signal tranduction, and various enzymatic activities and to biological processes such as metabolism, cell communication, and organism development (Figure 5). These functional and biological profiles of the plasma mem-


Figure 4. Comparison of the protein profiles of hESCs and hECCs. The Venn diagrams show unique and common proteins of hESCs (left) and hECCs (right). Given are the numbers of all proteins identified (gray), of those that were GO annotated as general membrane proteins (green), and those either GO-annotated or predicted as plasma membrane proteins (blue).

branes of hESCs and hECCs are not in agreement with the ‘average’ profiles of cells that are based on proteome studies of whole cell lysates. Unfortunately, the detailed picture of the ‘average’ plasma membrane proteome remains unknown due to the dependence on the plasma membrane isolation method, mentioned earlier and is, therefore, not available for comparison with the profiles of hESCs and hECCs determined here. However, even though hESCs and hECCs dedicate the same proportions of proteins to these functions and processes, they do not always employ the same proteins (Table 1). For instance, 30% of the plasma membrane proteins are involved in signal transduction in the two cell types, and most of these proteins of this functional class could be sorted to certain signaling pathways. This revealed, for instance, that hESCs and hECCs differ in the FGF receptors (FGFRs) expressed. Different FGFRs are known to differ in their ligand binding affinities, tissue distribution and biological impact, and serve to activate different signaling cascades. Several FGFs can signal through multiple FGFRs; for example, the most well-known self-renewal factor FGF2 can signal through FGFR1, FGFR3, and FGFR4.62 FGFR1, which was identified both in hESCs and hECCs, binds acidic and basic FGFs, is involved in limb induction, and chromosomal aberrations involving the FGFR1 gene are associated with stem cell myeloproliferative disorder and stem cell leukemia lymphoma syndrome. The FGFR substrate 2 (FRS2) was identified in hESCs only. This lipid raft-associated FGFR signaling adaptor is known to play a pivotal role in FGFinduced recruitment and activation of phosphatidylinositol 3-kinase (PI3-kinase) and links FGFRs to the mitogen-activated protein kinase (MAPK) signaling pathway.63,64 Several components of the PI3-kinase signaling pathway (ITPR1, ITPR2, ITPR3) were detected in both hESCs and hECCs. On the other hand, FGFR3 and FGFR4 were identified in hECCs only. However, earlier real time PCR analyses had already indicated that FGFR1, 2, 3, and 4 are all expressed in hESCs.65 From these

research articles results, we conclude that, even though hESCs and hECCs resemble each other in their general functional as well as biological profiles, they show subtle, yet important differences in their protein profiles. Validation of Plasma Membrane Proteins Identified in hESCs and hECCs. Since we performed a qualitative study here where peptide counts at best provide semiquantitative information, we evaluated the results of our proteomic comparison of hESCs and hECCs by performing Western blot analysis and immunofluorescence confocal microscopy. We used antibodies against plasma membrane proteins that were identified uniquely in hESCs [frizzled-7 (FZD7), Niemann-Pick C1 (NPC1)] or in both cell types [insulin-like growth factor 1 receptor (IGF1R), bone marrow stromal antigen 2 (BST2)]. FZD7 and NPC1 were identified by MS with 8 and 6 peptides, respectively, in hESCs but not in hECCs. Concordantly, Western blot analyses showed that FZD7 and NPC1 are highly expressed in hESCs but not or only marginally in hECCs (Figure 6A). The low signal intensity can be ascribed to the poor quality of the FZD7 and NPC1 antibodies, since higher loading concentrations and longer or more sensitive probing and developing did not improve the results. IGF1R was identified with 22 peptides in hESCs and 12 peptides in hECCs. Western blot analysis proved the high abundance of IGF1R in both cells types but did not indicate a difference in protein abundance. In fact, the IGF1R signal showed similar intensities in hESCs and hECCs on Western blot. BST2 was identified with 7 and 8 peptides in hESCs and hECCs, respectively. The Western blot analysis indicated BST2 expression in both cell types; however, the band was much stronger in hESCs than in hECCs. The undifferentiated status of hESCs and hECCs was verified by probing against the POU domain class 5 transcription factor 1 (Oct4), a nuclear stem cell protein benchmark. Equal sample loading was confirmed by coomassie-staining (data not shown) and by probing against calnexin (CANX), an ER protein. Confirmatory results from immunofluorescence confocal microscopy could only be achieved for BST2 and Connexin43 (GJA1, IPI00218487) due to the poor quality and the nonspecific nuclear staining of the other plasma membrane protein antibodies. As expected, BST2 did not co-localize with Oct4 in the nucleus. Instead, BST2 accumulated at the plasma membrane of both hESCs and hECCs where it was not evenly distributed but enriched on the surface of the cell filopodia (Figure 6B). This confirmed BST2 as a plasma membrane protein expressed in both cell types. Connexin43 was identified in both cell types, albeit with an almost 4-fold higher Mascot score and 3-fold higher spectral count in hESCs than hECCs (Supplemental Table 2). This suggests that Connexin43 is expressed at higher levels in hESCs compared with hECCs. The immunofluorescence images show that this is indeed the case; in contrast to hECCs, the plasma membrane of hESCs is highly enriched for Connexin43 (Figure 6C). In summary, for these proteins, Western blot analysis and immunofluorescence confocal microscopy confirmed our MS results on the differential expression of plasma membrane proteins in hESCs and hECCs. Performance Comparison with Genomic and Proteomic Data Sets Reported Previously. The optimized protocol (Figure 3) was superior to all other protocols tested throughout our study with respect to both the absolute number and the confidence of the plasma membrane protein identifications. Using only 5 × 105 cells, we were able to identify 384 and 344 common membrane proteins, and an additional 237 and 219 plasma membrane proteins in hESCs and hECCs, respectively. Journal of Proteome Research • Vol. 7, No. 7, 2008 2943

research articles

Dormeyer et al.

Figure 5. The plasma membranes of hESCs and hECCs have similar functional and biological profiles. The circle diagrams depict the functional and biological classification of the plasma membrane proteins that were identified in hESCs (inner circle) and hECCs (outer circle). The proteins were classified according to their GO annotation.

In relative numbers, ∼30% of the proteins identified in hESCs and hECCs were annotated common membrane proteins, and an additional ∼20% were annotated or predicted plasma membrane proteins. The somewhat low enrichment of plasma membrane proteins reflected by the relative number of identified plasma membrane proteins is negligible because of the high absolute number of plasma membrane proteins identified with our method, and the ability of modern mass spectrometers to record hundreds of thousands peptide spectra during a single LC run. We categorized the remaining ∼50% of the identified proteins as nonmembranous, since they were annotated or predicted to originate from other cell organelles, including the nucleus, ER, mitochondria, peroxisomes, and lysosomes. However, these organelles all comprise extensive membrane structures, such as envelopes, cisternae networks, and inner and outer membrane loops. A large portion of the proteins classified as nonmembranous might originate from these organellar membrane structures. Thus, the true relative number of identified general membrane proteins might actually be higher than given here. Furthermore, it cannot be excluded that proteins originating from organellar membrane structures also embeded into the plasma membrane, and that the relative number of identified plasma membrane proteins given here is also too low. Our strategy led to the identification of over 200 plasma membrane proteins from only 5 × 105 hESCs and hECCs, respectively, and can, therefore, compete with studies of plasma membrane proteomes reported previously, which employed elaborate organelle fractionation or protein affinity purification. For example, Rahbar et al. identified 214 (40%) plasma membrane proteins in a human breast cancer cell line after plasma membrane isolation by colloidal silica using 40 µg of protein material,66 Zhao et al. identified 526 (67%) plasma membrane proteins in a human lung cancer cell line by purification of biotinylated plasma membrane sheets using over 60 µg of protein material,67 Schindler et al. identified 197 (42%) plasma membrane proteins from one rat cerebellum using two-phase affinity partitioning,68 and Foster et al. identified 148 (32%) membrane proteins (among which are plasma membrane proteins) in mesenchymal cells differentiating toward osteoblasts using 50 µg of protein.69 Of course, no direct comparison of these numbers is possible because different cells, sample amounts, and mass spectrometers were used. In addition, we 2944


realized that the classification of identified proteins as plasma membrane proteins often remains ambiguous. To compare the plasma membrane proteomes of human and mouse ESCs, we matched our HUES-7 data set with a mouse ESC (mESC; line D3) data set reported by Nunomura et al.70 The latter was obtained using 4.8 × 109 cells as starting material (104 times more than here), biotinylation, and subsequent affinity-based purification of cell surface proteins, and resulted in the identification of 324 proteins. Employing the SOSUI prediction algorithm and signal sequence analyses, the authors had classified 235 of the identified proteins as transmembrane region-containing proteins, and had suggested that these must be cell surface proteins due to the selectivity of the protein labeling reaction. However, the SOSUI algorithm is known to be susceptible to predicting membrane localization signal sequences as transmembrane regions.71 Indeed, many proteins identified by Nunomura et al. were predicted to contain such signal sequences. Using the selectivity of the protein labeling reaction as plasma membrane localization criterium also seems questionable, as the list of proteins identified by Nunomura et al. includes many proteins from other cell organelles, such as nuclear histones and ribosomal proteins. However, using our classification strategy based on GO annotation, that is, on published experimental results, sequence similarity and GO curator judgment, it turned out that only 63 of the proteins identified by Nunomura et al. were annotated as common membrane proteins, and an additional 118 as plasma membrane proteins (Supplemental Table 4A). Thus, even though their specific approach resulted in a higher relative number of plasma membrane protein identifications (36% in mESCs versus 22% in hESCs), the absolute number of plasma membrane proteins identified with our approach was significantly higher (118 proteins in mESCs versus 237 proteins in hESCs). Biological Comparison with Genomic and Proteomic Data Sets Reported Previously. The comparison of the hESC (HUES7) data set here with Nunomura’s mESC (D3) data set revealed interesting similarities and differences in the plasma membrane proteomes of human and mouse ESCs (Supplemental Figure 2). In total, hESCs and mESCs had 116 proteins in common, 25 of which were annotated general membrane proteins, and 57 annotated plasma membrane proteins (Supplemental Table 4B-D). Identified uniquely in the hESCs were 359 membrane and 180 plasma membrane proteins, whereas, of those proteins

research articles


Table 1. Detailed Functional Classification of the Plasma Membrane Proteins Identified in hESCs and hECCsa process

Signal transduction

function

EGF signaling

unique hESC

ERBB3

Ephrin signaling

EFNB1 EPHB3

FGF signaling

FRS2

Hedgehog signaling IGF signaling

NPC1

Insulin signaling Integrin signaling

unique hECC

FGFR3 FGFR4

ADAM9 ITGA1 ITGA2 ITGA3 ITGB4

PSEN1

TGFB signaling

INSR CD47

JAG1 NOTCH2 TGFBR1

TNF signaling VEGF signaling WNT/FZD signaling

NGFR

G protein coupled receptor signaling

RGS19

Receptor activity

CACNA2D1 PIGR

Peptide receptor activity

CCR5 GRPR SSTR2

FZD7

KDR FZD2 FZD6 LRP6 ADCY9 AKAP7 GNG12 RHOB

P2RX4

Lipoprotein receptor activity

Kinase activity

Phosphatase activity

EGFR ERBB2 EPHA1 EPHA2 EPHB4 FGFR1

IGF1R IGF2R

NF-kappa B signaling Notch signaling

common hESC and hECC

AXL MET TEK PTPRK PTPRA

ALK ROR2

ADAM10 ITGA5 ITGA6 ITGAV ITGB1 ITGB5 BST2 GJA1 ADAM17 APP NCSTN ACVR1 SDC2

CTNNB1

GNA11 GNAI2 GNAQ GNG5 HRAS PRKAR2A BCAM CD46 CD59 CXADR HNRPM IFNGR1 ITPR1 ITPR2 ITPR3 M6PR PGRMC1 SCARB2 SLC20A2 SORL1 TFRC XPR1

LDLR LRP1 LRP4 LRP8 LSR VLDLR LCK MARCKS PTK7 ROR1 F11R PTPRD

Journal of Proteome Research • Vol. 7, No. 7, 2008 2945

research articles

Dormeyer et al.

Table 1. Continued process

Cell communication

Transport

function

unique hECC

Cell-cell adhesion

ALCAM CLDN12 DST FAT PCDH18 PCDH7 PCDHB2 PCDHB3 PVRL2 TPBG

DSG1 JAM2 TRO

Cell-cell signaling

CRB3 GJA7 HLA-C NCAM1

CD164 CNTNAP1 MSN TJP1

SLC1A1 SLC7A8

SLC12A7

Adenine transporter activity Amino acid transporter activity

Anion transporter activity Cation transporter activity

Electron transporter activity Folic acid transporter activity Peptide transporter activity Protein transporter activity

CLCN7 ATP1B1 ATP2A3 TCIRG1

PTPRF PTPRG PTPRS PTPRZ1 CD44 CLDN6 DCBLD2 DSC2 DSG2 FLOT1 FLOT2 PCDH1 PKP2 PVRL1 ROBO1 HLA-A MICA MME NEO1 OCLN RAB13 SLC25A4 SLC25A5 SLC1A5 SLC7A1 SLC7A5 SLC12A2 SLC12A6 VDAC3 ATP1A1 ATP1A2 ATP1A3 ATP1B3 ATP2A2 ATP2B1 ATP2B4 PKD2 SLC19A1

SLC15A1 RAB22A TIMM23

STX6

Transporter activity

SLC6A8 SLC25A11 PLP2 TM9SF2



CYB561

Secretory protein transporter activity

Various

2946

unique hESC

CAV1 CD200 CD81 DEGS1 FER1L3 ITM2B PEX3 PLD2 PLSCR1

AP2A2 AP2B1 AP2M1 SDCBP AP2A1

ABCC5 SLC5A6 SLC6A11 SLC6A6 SLC25A12 SLC29A1 SLC31A1 ATRN CD55 EBP EPB41L2 GPC1 GPC3 LARS MFI2 MMP15

CDC42 CLTC

BCAP31 CD9 RAB14 STX7 STX8 TMED10 ABCC1 ABCC4 SLC25A3 SLC25A13 SNAP23 SYPL1 TMED2 BASP1 CD63 CLPTM1 DAG1 DYSF FADS2 GOLPH2 GPC4 LAMP1

research articles

Plasma Membrane Proteomics of hESC and hECC Table 1. Continued process

function

unique hESC

unique hECC


PLXNB1 PRNP SGCB SILV ST14 STIM1 TMEM8 TSPAN9

PDPN SDC4 SIRPA STEAP2 TSPAN31

LETM1 MCAM MMP14 PHB PLEC1 PODXL PROM1 PRSS8 STOM TACSTD1 TEGT THY1 VAPB

a The plasma membrane proteins were grouped into signal transduction pathways, cell communication events, and transport functions according to their GO annotation. Given are the gene symbols of the proteins that were identified uniquely in hESCs, uniquely in hECCs or in hESCs as well as hECCs.

uniquely identified in mESCs, 38 originated from general membranes and 61 from the plasma membrane. This data may broaden our knowledge about the protein profile of distinct ESCs; however, our additional comparison with hECCs sheds further light on the subtle and more obvious differences between hESCs, mESCs, and ECCs. The three cell types had 46 plasma membrane proteins in common (Supplemental Figure 3). These included a variety of receptors, such as the insulin-like growth factor receptors IGF1R and IGF2R; the ephrin receptors EPHA1, EPHA2, and EPHAB4; and several integrin receptors ITGA5, ITGA6, ITGAV, ITGB1, and ITGB5 (Supplemental Table 5A). Also included were receptors for lipoproteins, like LDLR and LRP1, for nucleotides like ADT1 and ADT2, for carbohydrates like MPRD, and for ions like ATP1A1, ATP2A2, ATP1B3, SLC25A3, and VDAC3. We also screened the results for known pluripotency markers. However, not many of these are plasma membrane proteins.72–74 Commonly used cell surface markers for pluripotent cells are the stage-specific antigens (SSEAs) which are carbohydrate epitopes on closely related glycosphingolipids. SSEA-1 is expressed by mESCs, whereas SSEA-3 and SSEA-4 are expressed by hESCs. The SSEAs were reported to be involved in cell-cell adhesion, and our study revealed that hESCs, mESCs, and hECCs express different sets of cell-cell adhesion proteins. Twenty of the plasma membrane proteins identified in hESCs are involved in cell-cell adhesion, and PVRL3, STIM1, CLDN12, and the protocadherins PCDH7, 18, B2, and B3 were uniquely identified in hESCs (Supplemental Table 5B). In mESCs, 11 cell-cell adhesion plasma membrane proteins were identified, and only two of them, CELSR1 and ICAM1, were mESC-unique. The hECCs expressed 15 plasma membrane cell-cell adhesion proteins, 11 of which they had in common with hESCs. With regard to the general similarity of the functional and biological profiles shown above, we therefore conclude that hESCs and hECCs have similar cell-cell adhesion protein profiles which differ from those of mESCs. Potential new cell surface marker proteins for hESCs, mESCs, and hECCs might fall into this functional group, and be among those proteins differentially identified in our study. Other cell surface markers specific for human pluripotent cells are the TRA 1-60 and 1-81 antigens, which are carbohydrate epitopes on a yet unknown glycoprotein also recognized by the antibody GCTM2. Recently, it was reported that this unknown TRA epitope carrier protein is a large variant of podocalyxin.75 Strikingly, a Podocalyxin-like protein 1 (PODXL)

was identified in hESCs and hECCs as well as in mESCs. Thus, PODXL may be present in mESCs without exposing the specific TRA-1-61 and TRA-1-81. Additional human and mouse pluripotency cell surface markers are CD9, alkaline phosphatase (ALPL), and β-Catenin (CTNNB1), which were also identified in all three cell types, but ALPL was classified as general membrane protein in our study.

Discussion Our strategy illustrates that plasma membrane proteomics can be performed without tedious organelle fractionation by carefully optimizing the sample preparation and protein digestion for MS. The multistep sample preparation and in-solution digestion protocol described here can be used for any cell type and for low amounts of starting material. Using only 5 × 105 cells, we were able to identify 582 and 530 membrane proteins of hESCs and hECCs, 237 and 219 of which were plasma membrane proteins, respectively. These numbers are similar to plasma membrane-specific approaches reported previously, and outperform published plasma membrane proteome data sets of ESCs using orders of magnitude more starting material. Our comparative study confirms for the first time on the plasma membrane proteome level the high similarity between hESCs and hECCs, and their distinction from mESCs. Those plasma membrane proteins identified in more than one of the three cell types may be members of signaling cascades or cell communication events that are relevant for pluripotency in general. In contrast, plasma membrane proteins that were identified in hESCs and mESCs but not hECCs could be involved in the ‘benign’ nature of the stem cells as opposed to the ‘malignant’ nature of the cancer cells. Cancer is increasingly being viewed as a stem cell disease in general, and in particular under the aspect of the roles of Hedgehog and Wnt signaling pathways in both stem cell self-renewal and cancer growth.76 Strikingly, we found distinct Hedgehog- and Wnt-pathway members in hESCs and hECCs. Activation of the Hedgehog pathway is initiated by binding a hedgehog ligand to the patched receptor (PTCH). We did not find PTCH in hESCs or hECCs, even though it was identified in mESCs. Instead, in hESCs, we identified Nieman Pick type C1 protein (NPC1), which is a PTCH-related plasma membrane protein with hedgehog receptor activity. The presence of NPC1 in hESCs and its absence in hECCs was confirmed by Western blot analysis (Figure 6). Regulation of NPC1 expression might be Journal of Proteome Research • Vol. 7, No. 7, 2008 2947

research articles

Dormeyer et al.

Figure 6. Western blot analysis and immunofluorescence confocal microscopy of proteins identified in hESCs and hECCs. (A) Whole cell lysates of hECCs (left lane) and hESCs (right lane) were subjected to SDS-PAGE and Western blotting, and were probed for the presence of proteins identified uniquely in hESCs (FZD7, NPC1) or in both cell types (IGF1R, BST2). The numbers under the panels indicate the number of unique peptides identifying the protein by MS. Stemness was assessed by probing against the nuclear marker Oct4, and equal loading was confirmed by probing against the ER marker CANX. (B) Fixed hECCs and hESCs were incubated with primary antibodies as indicated, and secondary antibodies fluorescently labeled with FITC (green) or Cy3 (red). Simultaneous visualization of Oct4 (red) and BST2 (green) showed that BST2 is enriched in the filopodia of the plasma membrane of hESCs and hECCs. (C) Costaining of the nucleus with TOPRO (blue) showed that Connexin43 (green) is highly expressed on the plasma membrane of by hESCs, but not on hECCs. Scale bars, 20 µm. 2948


research articles

Plasma Membrane Proteomics of hESC and hECC involved in the control of hedgehog signaling during stem cell self-renewal and cancer formation. Activation of the Wnt pathway is initiated by binding Wnt ligands to their FZD and low-density lipoprotein receptor-related protein (LRP) receptors leading to the relocalization of CTNNB from the degradation complex to the nucleus. The FZD7 receptor was uniquely identified in hESCs, whereas FZD2, FZD6, and LRP6 were only identified in hECCs. FZD2 was also reported in mESCs, together with FZD10. Identified in all three cell types were CTNNB and the Wnt modulator semaphorin SEMA4D, which was classified as general membrane protein in our study. Other identified semaphorins were SEMA6A in both hESCs and hECCs, and SEMA7A uniquely in hECCs. Exemplarily for these Wnt-pathway members, the expression of FZD7 in hESCs but not in hECCs was confirmed by Western blot analysis (Figure 6). Hence, our data discloses the receptors FZD2, FZD6, FZD7, and LRP6 and the modulator SEMA7A as interesting candidates for studies of the differential regulation of Wnt signaling in ‘benign’ and ‘malignant’ pluripotent cells. Further research on plasma membrane proteins identified uniquely in hECCs but not hESCs could help elucidate the mechanisms leading to germ cell tumor formation in more detail. It is known that the most common testicular germ cell tumors are characterized genetically by having one or more copies of an isochromosome of the short (p) arm of chromosome 12 or other forms of 12p amplification.77 Accordingly, the altered expression of 12p genes in testicluar cancer has been investigated intensively. In our study, we have identified 26 proteins whose genes are located on chromosome 12p (Supplemental Table 6). Fourteen of these 12p proteins were uniquely identified in hECCs, whereas only one was uniquely identified in hESCs and 11 were identified in both hESCs and hECCs. Strikingly, several of the genes coding for the 12p proteins identified in our study are known to be differentially expressed in testicular cancer and/or ECCs. For example, GAPDH, LDHB, YARS, CD9, MGST1, and GOLT1B reportedly show elevated DNA copy numbers and gene expression levels in testicular germ cell tumors and various ECC lines including NT2/D1 cells.78 Differential regulation of GAPDH, LDHB, CD9, and GOLT1B was also observed in testicular seminoma tumors, together with altered expression of DDX47, KRAS, M6PR, PHB2, SLC2A3, CLSTN3, CSDA, LRP6, NDUFA9, and NOL1.79 In our study, GAPDH, LDHB, YARS, CLSTN3, CSDA, LRP6, NDUFA9, and NOL1 were identified uniquely in hECCs, while CD9, GOLT1B, MGST1, DDX47, KRAS, M6PR, PHB2, and SLC2A3 were identified in both hESCs and hECCs. In addition, CD9 and M6PR were also identified in mESCs. Considering the genomic and proteomic data sets mentioned earlier, we conclude from our study that GAPDH, LDHB, YARS2, CLSTN3, CSDA, LRP6, NDUFA9, and NOL1 might play an essential role in testicular germ cell tumor formation. Further analyses are needed to disclose the functions and interactions of these 12p proteins in the membranes of hECCs. At present, great hope lies in the potential use of hESCs as an unlimited supply of transplantable cells to replace or regenerate damaged or diseased tissues in numerous conditions. One major problem in addition to the potential for teratoma formation, is the immunological rejection of transplanted cells due to MHC class I antigen (HLA) incompatibility between donor and recipient. Little is known about the regulation of HLA genes in hESCs, and recently, it was reported that HLA are not, or only weakly, exhibited on the surface of hESCs.80,81 The lack or loss of HLA molecules is probably due

to the down-regulation of the components of the antigenprocessing machinery, a mechanism frequently found during the immune-evasion of tumors.82 Despite the low copy number per cell, we identified HLA-A in both hESCs and hECCs, HLA-B uniquely in hECCs, and HLA-C uniquely in hESCs. In addition, the transporter associated with antigen processing 1 (TAP1), a member of the antigen-processing machinery, was identified in hESCs. The identification of these low-abundance HLA molecules underscores the potential of our simple strategy to identify plasma membrane proteins from low amounts of cell material and confirms the quality of our comparison of the plasma membrane proteomes of hESCs and hECCs. Abbreviations. ECC, embryonal carcinoma cell; ER, endoplasmic reticulum; ESC, embryonic stem cell; FCS, fetal calf serum; FGF, fibroblast growth factor; GO, gene ontology; hECC, human embryonal carcinoma cell; hESC, human embryonic stem cell; ICM, inner cell mass; Lys-C, endoproteinase Lys-C; MEF, mouse embryonic fibroblast feeder cell; mESC, mouse embryonic stem cell; MS, mass spectrometry; PNGase F, N-Glycosidase F; SCX, strong cation exchange.

Acknowledgment. The authors thank Dr. D. Melton (Howard Hughes Medical Institute, Harvard Stem Cell Institute, Harvard University, MA) for the hESC line HUES-7 and Dr. MAG Van der Heyden (Medical Physiology, University Medical Center Utrecht, The Netherlands) for the mouse anti-Connexin43 antibody. This work was supported by The Netherlands Proteomics Centre. Supporting Information Available: Supplemental Figure 1, plasma membrane fractionation techniques tested during our study; Supplemental Figure 2, comparison of the protein profiles of hESCs and mECCs; Supplemental Figure 3, comparison of the plasma membrane protein profiles of hESCs, mESCs and mECCs; Supplemental Table 1, reproducibility of mass spectrometric analysis; Supplemental Table 2, detailed information on the proteins identified in hESCs and hECCs; Supplemental Table 3, detailed information on the proteins differentially identified in hESCs and hECCs; Supplemental Table 4, proteins differentially identified in hESCs and mESCs; Supplemental Table 5, proteins differentially identified in hESCs, mESCs and hECCs; Supplemental Table 6, 12p proteins differentially identified in hESCs and hECCs. This material is available free of charge via the Internet at http://pubs.acs.org. References (1) Andrews, P. W.; Matin, M. M.; Bahrami, A. R.; Damjanov, I.; Gokhale, P.; Draper, J. S. Embryonic stem (ES) cells and embryonal carcinoma (EC) cells: opposite sides of the same coin. Biochem. Soc. Trans. 2005, 33 (Pt. 6), 1526–30. (2) Kaufman, M. H.; Robertson, E. J.; Handyside, A. H.; Evans, M. J. Establishment of pluripotential cell lines from haploid mouse embryos. J. Embryol. Exp. Morphol. 1983, 73, 249–61. (3) Martin, G. R. Isolation of a pluripotent cell line from early mouse embryos cultured in medium conditioned by teratocarcinoma stem cells. Proc. Natl. Acad. Sci. U.S.A. 1981, 78 (12), 7634–8. (4) Thomson, J. A.; Itskovitz-Eldor, J.; Shapiro, S. S.; Waknitz, M. A.; Swiergiel, J. J.; Marshall, V. S.; Jones, J. M. Embryonic stem cell lines derived from human blastocysts. Science 1998, 282 (5391), 1145–7. (5) Reubinoff, B. E.; Pera, M. F.; Fong, C. Y.; Trounson, A.; Bongso, A. Embryonic stem cell lines from human blastocysts: somatic differentiation in vitro. Nat. Biotechnol. 2000, 18 (4), 399–404. (6) Keller, G. Embryonic stem cell differentiation: emergence of a new era in biology and medicine. Genes Dev. 2005, 19 (10), 1129–55. (7) Baker, D. E.; Harrison, N. J.; Maltby, E.; Smith, K.; Moore, H. D.; Shaw, P. J.; Heath, P. R.; Holden, H.; Andrews, P. W. Adaptation to culture of human embryonic stem cells and oncogenesis in vivo. Nat. Biotechnol. 2007, 25 (2), 207–15.


research articles (8) Gudjonsson, T.; Magnusson, M. K. Stem cell biology and the cellular pathways of carcinogenesis. APMIS 2005, 113 (11-12), 922–9. (9) Andrews, P. W. From teratocarcinomas to embryonic stem cells. Philos. Trans. R. Soc. London, Ser. B Biol. Sci. 2002, 357 (1420), 405–17. (10) Pierce, G. B. Teratocarcinoma: model for a developmental concept of cancer. Curr. Top. Dev. Biol. 1967, 2, 223–46. (11) Sell, S.; Pierce, G. B. Maturation arrest of stem cell differentiation is a common pathway for the cellular origin of teratocarcinomas and epithelial cancers. Lab. Invest. 1994, 70 (1), 6–22. (12) Draper, J. S.; Pigott, C.; Thomson, J. A.; Andrews, P. W. Surface antigens of human embryonic stem cells: changes upon differentiation in culture. J. Anat. 2002, 200 (Pt 3), 249–58. (13) Henderson, J. K.; Draper, J. S.; Baillie, H. S.; Fishel, S.; Thomson, J. A.; Moore, H.; Andrews, P. W. Preimplantation human embryos and embryonic stem cells show comparable expression of stagespecific embryonic antigens. Stem Cells 2002, 20 (4), 329–37. (14) Sell, S. Stem cell origin of cancer and differentiation therapy. Crit. Rev. Oncol. Hematol. 2004, 51 (1), 1–28. (15) Al-Hajj, M.; Clarke, M. F. Self-renewal and solid tumor stem cells. Oncogene 2004, 23 (43), 7274–82. (16) Liu, B. B.; Qin, L. X.; Liu, Y. K. Adult stem cells and cancer stem cells: tie in or tear apart. J. Cancer Res. Clin. Oncol. 2005, 131 (10), 631–8. (17) Newman, M. B.; Misiuta, I.; Willing, A. E.; Zigova, T.; Karl, R. C.; Borlongan, C. V.; Sanberg, P. R. Tumorigenicity issues of embryonic carcinoma-derived stem cells: relevance to surgical trials using NT2 and hNT neural cells. Stem Cells Dev. 2005, 14 (1), 29–43. (18) Ahram, M.; Litou, Z. I.; Fang, R.; Al-Tawallbeh, G. Estimation of membrane proteins in the human proteome. In Silico Biol. 2006, 6 (5), 379–86. (19) Marmagne, A.; Salvi, D.; Rolland, N.; Ephritikhine, G.; Joyard, J.; Barbier-Brygoo, H. Purification and fractionation of membranes for proteomic analyses. Methods Mol. Biol. 2006, 323, 403–20. (20) Wu, C. C.; Yates, J. R., III. The application of mass spectrometry to membrane proteomics. Nat. Biotechnol. 2003, 21 (3), 262–7. (21) Marguet, D.; Lenne, P. F.; Rigneault, H.; He, H. T. Dynamics in the plasma membrane: how to combine fluidity and order. EMBO J. 2006, 25 (15), 3446–57. (22) Ikeda, M.; Kihara, A.; Igarashi, Y. Lipid asymmetry of the eukaryotic plasma membrane: functions and related enzymes. Biol. Pharm. Bull. 2006, 29 (8), 1542–6. (23) Laude, A. J.; Prior, I. A. Plasma membrane microdomains: organization, function and trafficking. Mol. Membr. Biol. 2004, 21 (3), 193–205. (24) Cheng, Z. J.; Singh, R. D.; Marks, D. L.; Pagano, R. E. Membrane microdomains, caveolae, and caveolar endocytosis of sphingolipids. Mol. Membr. Biol. 2006, 23 (1), 101–10. (25) Raimondo, F.; Ceppi, P.; Guidi, K.; Masserini, M.; Foletti, C.; Pitto, M. Proteomics of plasma membrane microdomains. Expert Rev. Proteomics 2005, 2 (5), 793–807. (26) Chen, P.; Li, X.; Sun, Y.; Liu, Z.; Cao, R.; He, Q.; Wang, M.; Xiong, J.; Xie, J.; Wang, X.; Liang, S. Proteomic analysis of rat hippocampal plasma membrane: characterization of potential neuronal-specific plasma membrane proteins. J. Neurochem. 2006, 98 (4), 1126–40. (27) Durr, E.; Yu, J.; Krasinska, K. M.; Carver, L. A.; Yates, J. R.; Testa, J. E.; Oh, P.; Schnitzer, J. E. Direct proteomic mapping of the lung microvascular endothelial cell surface in vivo and in cell culture. Nat. Biotechnol. 2004, 22 (8), 985–92. (28) Watarai, H.; Hinohara, A.; Nagafune, J.; Nakayama, T.; Taniguchi, M.; Yamaguchi, Y. Plasma membrane-focused proteomics: dramatic changes in surface expression during the maturation of human dendritic cells. Proteomics 2005, 5 (15), 4001–11. (29) Cutillas, P. R.; Biber, J.; Marks, J.; Jacob, R.; Stieger, B.; Cramer, R.; Waterfield, M.; Burlingame, A. L.; Unwin, R. J. Proteomic analysis of plasma membrane vesicles isolated from the rat renal cortex. Proteomics 2005, 5 (1), 101–12. (30) Morre, D. M.; Morre, D. J. Aqueous two-phase partition applied to the isolation of plasma membranes and Golgi apparatus from cultured mammalian cells. J. Chromatogr., B: Biomed. Sci. Appl. 2000, 743 (1-2), 377–87. (31) Cao, R.; Li, X.; Liu, Z.; Peng, X.; Hu, W.; Wang, X.; Chen, P.; Xie, J.; Liang, S. Integration of a two-phase partition method into proteomics research on rat liver plasma membrane proteins. J. Proteome Res. 2006, 5 (3), 634–42. (32) Everberg, H.; Peterson, R.; Rak, S.; Tjerneld, F.; Emanuelsson, C. Aqueous two-phase partitioning for proteomic monitoring of cell surface biomarkers in human peripheral blood mononuclear cells. J. Proteome Res. 2006, 5 (5), 1168–75.

2950


Dormeyer et al. (33) McCarthy, F. M.; Burgess, S. C.; van den Berg, B. H.; Koter, M. D.; Pharr, G. T. Differential detergent fractionation for non-electrophoretic eukaryote cell proteomics. J. Proteome Res. 2005, 4 (2), 316–24. (34) Cox, B.; Emili, A. Tissue subcellular fractionation and protein extraction for use in mass-spectrometry-based proteomics. Nat. Protoc. 2006, 1 (4), 1872–8. (35) Guo, L.; Eisenman, J. R.; Mahimkar, R. M.; Peschon, J. J.; Paxton, R. J.; Black, R. A.; Johnson, R. S. A proteomic approach for the identification of cell-surface proteins shed by metalloproteases. Mol. Cell. Proteomics 2002, 1 (1), 30–6. (36) Goshe, M. B.; Blonder, J.; Smith, R. D. Affinity labeling of highly hydrophobic integral membrane proteins for proteome-wide analysis. J. Proteome Res. 2003, 2 (2), 153–61. (37) Elortza, F.; Nuhse, T. S.; Foster, L. J.; Stensballe, A.; Peck, S. C.; Jensen, O. N. Proteomic analysis of glycosylphosphatidylinositolanchored membrane proteins. Mol. Cell. Proteomics 2003, 2 (12), 1261–70. (38) Man, P.; Novak, P.; Cebecauer, M.; Horvath, O.; Fiserova, A.; Havlicek, V.; Bezouska, K. Mass spectrometric analysis of the glycosphingolipid-enriched microdomains of rat natural killer cells. Proteomics 2005, 5 (1), 113–22. (39) Santoni, V.; Molloy, M.; Rabilloud, T. Membrane proteins and proteomics: un amour impossible. Electrophoresis 2000, 21 (6), 1054–70. (40) Rabilloud, T.; Blisnick, T.; Heller, M.; Luche, S.; Aebersold, R.; Lunardi, J.; Braun-Breton, C. Analysis of membrane proteins by two-dimensional electrophoresis: comparison of the proteins extracted from normal or Plasmodium falciparum-infected erythrocyte ghosts. Electrophoresis 1999, 20 (18), 3603–10. (41) Bunai, K.; Yamane, K. Effectiveness and limitation of twodimensional gel electrophoresis in bacterial membrane protein proteomics and perspectives. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2005, 815 (1-2), 227–36. (42) Adessi, C.; Miege, C.; Albrieux, C.; Rabilloud, T. Two-dimensional electrophoresis of membrane proteins: a current challenge for immobilized pH gradients. Electrophoresis 1997, 18 (1), 127–35. (43) Chevallet, M.; Santoni, V.; Poinas, A.; Rouquie, D.; Fuchs, A.; Kieffer, S.; Rossignol, M.; Lunardi, J.; Garin, J.; Rabilloud, T. New zwitterionic detergents improve the analysis of membrane proteins by two-dimensional electrophoresis. Electrophoresis 1998, 19 (11), 1901–9. (44) Rabilloud, T. Use of thiourea to increase the solubility of membrane proteins in two-dimensional electrophoresis. Electrophoresis 1998, 19 (5), 758–60. (45) Peng, J.; Elias, J. E.; Thoreen, C. C.; Licklider, L. J.; Gygi, S. P. Evaluation of multidimensional chromatography coupled with tandem mass spectrometry (LC/LC-MS/MS) for large-scale protein analysis: the yeast proteome. J. Proteome Res. 2003, 2 (1), 43–50. (46) Washburn, M. P.; Ulaszek, R.; Deciu, C.; Schieltz, D. M.; Yates, J. R., III. Analysis of quantitative proteomic data generated via multidimensional protein identification technology. Anal. Chem. 2002, 74 (7), 1650–7. (47) Wu, C. C.; MacCoss, M. J.; Howell, K. E.; Yates, J. R., III. A method for the comprehensive proteomic analysis of membrane proteins. Nat. Biotechnol. 2003, 21 (5), 532–8. (48) Blonder, J.; Goshe, M. B.; Moore, R. J.; Pasa-Tolic, L.; Masselon, C. D.; Lipton, M. S.; Smith, R. D. Enrichment of integral membrane proteins for proteomic analysis using liquid chromatographytandem mass spectrometry. J. Proteome Res. 2002, 1 (4), 351–60. (49) Blonder, J.; Goshe, M. B.; Xiao, W.; Camp, D. G., II; Wingerd, M.; Davis, R. W.; Smith, R. D. Global analysis of the membrane subproteome of Pseudomonas aeruginosa using liquid chromatography-tandem mass spectrometry. J. Proteome Res. 2004, 3 (3), 434–44. (50) Zhang, N.; Li, N.; Li, L. Liquid chromatography MALDI MS/MS for membrane proteome analysis. J. Proteome Res. 2004, 3 (4), 719– 27. (51) Klammer, A. A.; MacCoss, M. J. Effects of modified digestion schemes on the identification of proteins from complex mixtures. J. Proteome Res. 2006, 5 (3), 695–700. (52) Strader, M. B.; Tabb, D. L.; Hervey, W. J.; Pan, C.; Hurst, G. B. Efficient and specific trypsin digestion of microgram to nanogram quantities of proteins in organic-aqueous solvent systems. Anal. Chem. 2006, 78 (1), 125–34. (53) Fischer, F.; Poetsch, A. Protein cleavage strategies for an improved analysis of the membrane proteome. Proteome Sci. 2006, 4, 2. (54) Fischer, F.; Wolters, D.; Rogner, M.; Poetsch, A. Toward the complete membrane proteome: high coverage of integral membrane proteins through transmembrane peptide detection. Mol. Cell. Proteomics 2006, 5 (3), 444–53.

research articles

Plasma Membrane Proteomics of hESC and hECC (55) Van Hoof, D.; Mummery, C. L.; Heck, A. J.; Krijgsveld, J. Embryonic stem cell proteomics. Expert Rev. Proteomics 2006, 3 (4), 427–37. (56) Cowan, C. A.; Klimanskaya, I.; McMahon, J.; Atienza, J.; Witmyer, J.; Zucker, J. P.; Wang, S.; Morton, C. C.; McMahon, A. P.; Powers, D.; Melton, D. A. Derivation of embryonic stem-cell lines from human blastocysts. N. Engl. J. Med. 2004, 350 (13), 1353–6. (57) Denning, C.; Allegrucci, C.; Priddle, H.; Barbadillo-Munoz, M. D.; Anderson, D.; Self, T.; Smith, N. M.; Parkin, C. T.; Young, L. E. Common culture conditions for maintenance and cardiomyocyte differentiation of the human embryonic stem cell lines, BG01 and HUES-7. Int. J. Dev. Biol. 2006, 50 (1), 27–37. (58) Braam, S. R.; Denning, C.; Van den Brink, S.; Kats, P.; Hochstenbach, R.; Passier, R.; Mummery, C. L. Improved genetic manipulation of human embryonic stem cells. Nat. Methods 2008, 5 (5), 389–92. (59) Lu, X.; Zhu, H. Tube-gel digestion: a novel proteomic approach for high throughput analysis of membrane proteins. Mol. Cell. Proteomics 2005, 4 (12), 1948–58. (60) Blonder, J.; Conrads, T. P.; Yu, L. R.; Terunuma, A.; Janini, G. M.; Issaq, H. J.; Vogel, J. C.; Veenstra, T. D. A detergent- and cyanogen bromide-free method for integral membrane proteomics: application to Halobacterium purple membranes and the human epidermal membrane proteome. Proteomics 2004, 4 (1), 31–45. (61) Russell, W. K.; Park, Z. Y.; Russell, D. H. Proteolysis in mixed organic-aqueous solvent systems: applications for peptide mass mapping using mass spectrometry. Anal. Chem. 2001, 73 (11), 2682–5. (62) Brandenberger, R.; Wei, H.; Zhang, S.; Lei, S.; Murage, J.; Fisk, G. J.; Li, Y.; Xu, C.; Fang, R.; Guegler, K.; Rao, M. S.; Mandalam, R.; Lebkowski, J.; Stanton, L. W. Transcriptome characterization elucidates signaling networks that control human ES cell growth and differentiation. Nat. Biotechnol. 2004, 22 (6), 707–16. (63) Hadari, Y. R.; Gotoh, N.; Kouhara, H.; Lax, I.; Schlessinger, J. Critical role for the docking-protein FRS2 alpha in FGF receptor-mediated signal transduction pathways. Proc. Natl. Acad. Sci. U.S.A. 2001, 98 (15), 8578–83. (64) Ong, S. H.; Hadari, Y. R.; Gotoh, N.; Guy, G. R.; Schlessinger, J.; Lax, I. Stimulation of phosphatidylinositol 3-kinase by fibroblast growth factor receptors is mediated by coordinated recruitment of multiple docking proteins. Proc. Natl. Acad. Sci. U.S.A. 2001, 98 (11), 6074–9. (65) Dvorak, P.; Dvorakova, D.; Koskova, S.; Vodinska, M.; Najvirtova, M.; Krekac, D.; Hampl, A. Expression and potential role of fibroblast growth factor 2 and its receptors in human embryonic stem cells. Stem Cells 2005, 23 (8), 1200–11. (66) Rahbar, A. M.; Fenselau, C. Unbiased examination of changes in plasma membrane proteins in drug resistant cancer cells. J. Proteome Res. 2005, 4 (6), 2148–53. (67) Zhao, Y.; Zhang, W.; Kho, Y.; Zhao, Y. Proteomic analysis of integral plasma membrane proteins. Anal. Chem. 2004, 76 (7), 1817–23. (68) Schindler, J.; Lewandrowski, U.; Sickmann, A.; Friauf, E.; Nothwang, H. G. Proteomic analysis of brain plasma membranes isolated by affinity two-phase partitioning. Mol. Cell. Proteomics 2006, 5 (2), 390–400. (69) Foster, L. J.; Zeemann, P. A.; Li, C.; Mann, M.; Jensen, O. N.; Kassem, M. Differential expression profiling of membrane proteins

(70)

(71) (72) (73) (74)

(75) (76) (77) (78)

(79)

(80)

(81)

(82)

by quantitative proteomics in a human mesenchymal stem cell line undergoing osteoblast differentiation. Stem Cells 2005, 23 (9), 1367–77. Nunomura, K.; Nagano, K.; Itagaki, C.; Taoka, M.; Okamura, N.; Yamauchi, Y.; Sugano, S.; Takahashi, N.; Izumi, T.; Isobe, T. Cell surface labeling and mass spectrometry reveal diversity of cell surface markers and signaling molecules expressed in undifferentiated mouse embryonic stem cells. Mol. Cell. Proteomics 2005, 4 (12), 1968–76. Lao, D. M.; Okuno, T.; Shimizu, T. Evaluating transmembrane topology prediction methods for the effect of signal peptide in topology prediction. In Silico Biol. 2002, 2 (4), 485–94. Avery, S.; Inniss, K.; Moore, H. The regulation of self-renewal in human embryonic stem cells. Stem Cells Dev. 2006, 15 (5), 729– 40. Allegrucci, C.; Young, L. E. Differences between human embryonic stem cell lines. Hum. Reprod. Update 2007, 13 (2), 103–20. Koestenbauer, S.; Zech, N. H.; Juch, H.; Vanderzwalmen, P.; Schoonjans, L.; Dohr, G. Embryonic stem cells: similarities and differences between human and murine embryonic stem cells. Am. J. Reprod. Immunol. 2006, 55 (3), 169–80. Schopperle, W. M.; DeWolf, W. C. The TRA-1-60 and TRA-1-81 human pluripotent stem cell markers are expressed on podocalyxin in embryonal carcinoma. Stem Cells 2007, 25 (3), 723–30. Beachy, P. A.; Karhadkar, S. S.; Berman, D. M. Tissue repair and stem cell renewal in carcinogenesis. Nature 2004, 432 (7015), 324– 31. Reuter, V. E. Origins and molecular biology of testicular germ cell tumors. Mod. Pathol. 2005, 18 (Suppl. 2), S51–60. Skotheim, R. I.; Autio, R.; Lind, G. E.; Kraggerud, S. M.; Andrews, P. W.; Monni, O.; Kallioniemi, O.; Lothe, R. A. Novel genomic aberrations in testicular germ cell tumors by array-CGH, and associated gene expression changes. Cell Oncol. 2006, 28 (5-6), 315–26. Korkola, J. E.; Houldsworth, J.; Chadalavada, R. S.; Olshen, A. B.; Dobrzynski, D.; Reuter, V. E.; Bosl, G. J.; Chaganti, R. S. Downregulation of stem cell genes, including those in a 200-kb gene cluster at 12p13.31, is associated with in vivo differentiation of human male germ cell tumors. Cancer Res. 2006, 66 (2), 820–7. Drukker, M.; Katz, G.; Urbach, A.; Schuldiner, M.; Markel, G.; Itskovitz-Eldor, J.; Reubinoff, B.; Mandelboim, O.; Benvenisty, N. Characterization of the expression of MHC proteins in human embryonic stem cells. Proc. Natl. Acad. Sci. U.S.A. 2002, 99 (15), 9864–9. Li, L.; Baroja, M. L.; Majumdar, A.; Chadwick, K.; Rouleau, A.; Gallacher, L.; Ferber, I.; Lebkowski, J.; Martin, T.; Madrenas, J.; Bhatia, M. Human embryonic stem cells possess immuneprivileged properties. Stem Cells 2004, 22 (4), 448–56. Cabrera, C. M.; Nieto, A.; Cortes, J. L.; Montes, R. M.; Catalina, P.; Cobo, F.; Barroso-Del-Jesus, A.; Concha, A. The low rate of HLA class I molecules on the human embryonic stem cell line HS293 is associated with the APM components’ expression level. Cell Biol. Int. 2007, 31 (9), 1072–78.

PR800056J


Plasma Membrane Proteomics of Human Embryonic Stem Cells and

Recommend Documents