Comprehensive proteomic analysis of breast cancer cell membranes reveals unique proteins with potential roles in clinical cancer

Proteins associated with cancer cell plasma membranes are rich in known drug and antibody targets as well as other proteins known to play key roles in the abnormal signal transduction processes required for carcinogenesis. We describe here a proteomics process that comprehensively annotates the protein content of breast tumor cell membranes and defines the clinical relevance of such proteins. Tumor-derived cell lines were used to ensure an enrichment for cancer cell-specific plasma membrane proteins because it is difficult to purify cancer cells and then obtain good membrane preparations from clinical material. Multiple cell lines with different molecular pathologies were used to represent the clinical heterogeneity of breast cancer. Peptide tandem mass spectra were searched against a comprehensive data base containing known and conceptual proteins derived from many public data bases including the draft human genome sequences. This plasma membrane-enriched proteome analysis created a data base of more than 500 breast cancer cell line proteins, 27% of which were of unknown function. The value of our approach is demonstrated by further detailed analyses of three previously uncharacterized proteins whose clinical relevance has been defined by their unique cancer expression profiles and the identification of protein-binding partners that elucidate potential functionality in cancer.     

In vivo fluorescence labeling of plasma membrane from plant tissue and cells

Eosin-5-maleimide (EMA) is a membrane impermeable thiol reagent and was used to label proteins of the outer surface of the plasma membrane from intact cells and plant tissue. The covalent reaction of the fluorescent dye with cysteine residues of proteins allowed the visualization of the labeled cell compartment by microscopy, as well as in membrane preparations of cell homogenates from corn, squash and carrots. Density gradients from a resuspended membrane pellet (1000–15 000 × g) showed a fluorescent band at 40–42% (w/w) sucrose in all specimens. This EMA-enriched zone was well separated from the protein maximum and from membranes, which could be classified as tonoplast, golgi apparatus and mitochondria by its marker enzymes. A conformity of the fluorescence maximum with maximum enzyme activity of the plasma membrane marker (UDPG)-steroyl-glucosyl transferase (SGT) could be observed in a few cases (e.g. corn), although SGT activity varied considerably (32–40% sucrose) according to the object and the tissue used.Microscopic analysis of the labeled cells detected a strong fluorescence in the region of the cell wall and the plasma membrane. The use of EMA as a tool for the labeling and identification of plant plasma membranes is discussed.     

The identification and characterization of membranome components

Analyses of proteins from a number of proteomic studies of cell membranes have demonstrated that a significant component of the identified proteins is not predicted to contain transmembrane regions. The presence of such proteins may arise as a result of contamination of the membrane preparations or through real associations. Our aim was to identify integral proteins as well as those that are intimately associated with the microsomal membranes of K562 cells. Isolated membranes were treated under conditions reported to remove noncovalently associated 'peripheral' proteins and the residual proteins were SDS-PAGE-separated and analyzed by LC-MS/MS. Tandem lectin affinity was also examined as a complementary approach for the enrichment of membrane glycoproteins. Approximately 41% of the isolated proteins were assigned as membrane proteins based on the presence of transmembrane regions or covalent post-translational modifications that could account for membrane association. Collectively, these results indicate that there is a significant component of non integral proteins that appear to be as closely associated with membranes as integral elements.     

Contamination of nuclear fractions with plasma membrane lipid rafts

Subcellular fractionation is central to a range of cell biological, biochemical and proteomic studies. Purification of nuclear-enriched fractions is critical for studies on nuclear structure and function. Here we show that detergent-based nuclear isolation methods cause the redistribution of proteins associated with plasma membrane lipid rafts into nuclear fractions. The glycosyl-phosphatidylinositol (GPI)-anchored prion protein (PrP(C)) and a GPI-anchored construct of angiotensin converting enzyme (GPI-ACE), as well as the lipid raft markers flotillin-1 and -2, were present in the nuclear fractions derived using three different subcellular fractionation protocols. Incubation of intact cells with bacterial phosphatidylinositol-specific phospholipase C (PI-PLC), which cleaves GPI-anchored proteins from the cell surface, significantly reduced the amount of PrP(C) and GPI-ACE in the nuclear fraction. Buoyant sucrose density gradient centrifugation in the presence of Triton X-100 of the nuclear fraction resulted in a significant proportion of the GPI-anchored proteins being recovered in the low density lipid raft fractions. These data indicate that the nuclear fraction isolated using such subcellular fractionation protocols is contaminated with components of plasma membrane lipid rafts and raises questions as to the integrity of the nuclear fraction isolated by such protocols for use in detailed cell biological studies and proteomics analysis.     

Comparison of the polypeptide composition of Escherichia coli outer membranes prepared by two methods.

Escherichia coli outer membranes were prepared by centrifugation to equilibrium in sucrose gradients and then treated with Sarkosyl in the presence of ethylenediaminetetraacetate. The polypeptide profiles of the two outer membrane preparations were compared by two-dimensional polyacrylamide gel electrophoresis. The patterns obtained were not identical, and Sarkosyl removed several minor proteins from the outer membrane.     

Location of an F-pilin pool in the inner membrane.

Polyacrylamide gel analysis of [35S]methionine-labeled membrane preparations from Escherichia coli has revealed the presence of five polypeptides present only in the membranes of cells containing the conjugative plasmid F. In addition to the previously reported product of traT, polypeptides migrating with apparent molecular weights of 100,000, 23,500, 12,000, and 7,000 were resolved. Membrane preparations from F traJ mutants lacked these polypeptides, indicating that all of these proteins are tra gene products. The 7,000-molecular-weight polypeptide comigrated with unlabeled purified F-pilin protein. About 4 to 5% of the total radioactive label in whole membrane preparations was present in this polypeptide, indicating the existence of a substantial pool of membrane-associated F-pilin. The polypeptide could be extracted from whole membrane preparations with Triton X-100 and was found in the inner membrane fraction of membranes separated by sucrose density centrifugation.     

Proteins Specified by Herpes Simplex Virus IX. Contiguity of Host and Viral Proteins in the Plasma Membrane of Infected Cells

Artificial mixtures of plasma membrane vesicles produced by microcavitation from infected and uninfected cells band at the same density on isopycnic centrifugation in sucrose density gradient. However, after reaction with antiviral antibody, the density of the infected cell plasma membrane vesicles increases, and the infected and uninfected cell membranes are quantitatively separable on isopycnic centrifugation. Plasma membrane vesicles prepared from cells doubly labeled before and after infection with radioactive amino acids and reacted with antibody banded at a high density. Polyacrylamide gel electropherograms show that the vesicles reacted with antibody consist of both host- and virus-specific membrane proteins. Microcavitation does not disrupt viral envelopes since infectivity is not affected by this procedure. We conclude that viral and cellular proteins in the plasma membrane preparations are contiguous.     

Proteins Specified by Herpes Simplex Virus VI. Viral Proteins in the Plasma Membrane

Previous papers in the series have shown that the surface membranes of herpesvirus-infected cells acquire new immunological specificities and that purified infected cell membrane preparations, characterized by their physical properties rather than topology in the cell, contain new glycoproteins genetically determined by the virus. In this study, we prepared purified plasma membrane identified by its 5' nucleotidase, fucose, and reduced nicotinamide adenine dinucleotide-diaphorase content. Analysis of the membrane proteins and glycoproteins by electrophoresis in acrylamide gels indicated the following. (i) Purified plasma membranes from infected cells contained two sets of proteins, i.e., host proteins were present both before and after infection and viral proteins were present only after infection. (ii) After infection, no appreciable selective or nonselective loss of host proteins from membranes was demonstrable. However, no new host proteins were made. (iii) Electropherograms of plasma membrane proteins from infected cells indicated the presence of at least 12 virus-specific proteins ranging in molecular weight from 25 x 10(3) to 126 x 10(3) daltons. Of these, at least nine were glycosylated. Proteins and glycoproteins with similar electrophoretic mobilities but in somewhat different ratios were also present in preparations of highly purified virions.     

Plasmodium lophurae: membrane proteins of erythrocyte-free plasmodia and malaria-infected red cells

Plasma membranes of normal duckling erythrocytes were prepared by blender homogenization and nitro-en decompression. Surface membrane vesicles of red cells infected with the avian malaria Plasmodium lophurae were produced by nitrogen decompression. Membranes of erythrocyte-free malaria parasites were removed from cytoplasmic constituents by Dounce homogenization. These membranes were collected by centrifugation in a sucrose step gradient and purified on a linear sucrose gradient. Red cell membranes had a buoyant density of 1.159 g/cm3, whereas plasmodial membranes banded at 2 densities: 1.110 g/cm3 and 1.158 g/cm3. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis of the isolated red cell membranes revealed 7 major protein bands with molecular weights (MW) ranging from 230, 000 to 22,000, and 3 glycoprotein bands with MW of 160,000, 88,000 and 37,000. Parasite membranes also had 7 major bands with MW ranging from 100,000 to 22,000. No glycoproteins were identifiable in these membranes. The proteins of the surface membranes from infected red cells had MW similar to those from normal red cells; however, there was some evidence of a reduction in the amount of the high MW polypeptides. The red cell membrane contained 79 nmoles sialic acid/mg membrane protein, whereas plasmodial membranes had 8 nmoles sialic acid/mg membrane protein. The sialic acid content of the surface membranes of infected red cells was significantly smaller than that of normal cells. Lactoperoxidase-glucose oxidase-catalyzed iodination of intact normal and malaria-infected erythrocytes labeled 7 surface components. Although no observable differences in iodinatable proteins were seen in these preparations, there was a striking reduction in the iodinatability of erythrocytic membranes obtained from P. lophurae-infected cells. Erythrocyte-free plasmodia bound very little radioactive iodine; the small amount of radioactivity was distributed among 3 major bands with MW of 42,000, 32,000 and 28,000. It is suggested that the alterations of the surface of the P. lophurae-infected erythrocyte do not occur by a wholesale insertion of plasmodial membrane proteins into the red cell plasma membrane, but rather that there are parasite-mediated modifications of existing membrane polypeptides.     

Preparation of Synaptic Plasma Membrane and Postsynaptic Density Proteins Using a Discontinuous Sucrose Gradient

Neuronal subcellular fractionation techniques allow the quantification of proteins that are trafficked to and from the synapse. As originally described in the late 1960’s, proteins associated with the synaptic plasma membrane can be isolated by ultracentrifugation on a sucrose density gradient. Once synaptic membranes are isolated, the macromolecular complex known as the post-synaptic density can be subsequently isolated due to its detergent insolubility. The techniques used to isolate synaptic plasma membranes and post-synaptic density proteins remain essentially the same after 40 years, and are widely used in current neuroscience research. This article details the fractionation of proteins associated with the synaptic plasma membrane and post-synaptic density using a discontinuous sucrose gradient. Resulting protein preparations are suitable for western blotting or 2D DIGE analysis.     

ATPase activities in peroxisome-proliferating yeast.

Preliminary studies on yeast peroxisomes have suggested that the membrane of these organelles may contain a proton-pumping ATPase. It has been reported that peroxisome-associated activity is similar to the F0-F1 mitochondrial type ATPase in its sensitivity to azide at pH 9.0, but characteristics of the plasma membrane type ATPase are also evident in peroxisomal preparations in that they exhibit pH 6.5 activity that is sensitive to vanadate. A comparative study of the prominent organellar ATPase activities was undertaken as a probe into the existence of an enzyme that is unique to the peroxisome, and biochemical properties of yeast mitochondrial, plasma membrane, together with peroxisomally-associated H(+)-ATPases are presented. Enzyme marker analysis of sucrose gradient fractions revealed a high degree of correlation between the amount of azide-sensitive pH 9.0 ATPase activity and that of the mitochondrial membrane marker, cytochrome c oxidase, in peroxisomal preparations. Purified mitochondrial and peroxisomally-associated activities were highly sensitive to the presence of sodium azide, N,N' -dicyclohexylcarbodiimide (DCCD) and venturicidin when measured at pH 9.0. Comparisons of peroxisomal activities with those of the purified plasma membrane at pH 6.0 in the presence of azide showed similar sensitivity profiles with respect to inhibitors of yeast plasma membrane ATPases such as vanadate and p-chloromercuriphenyl-sulfonic acid (CMP). Purified peroxisomal membranes, furthermore, reacted with antibody to the mitochondrial F1 subunit (as revealed by Western blot analysis), and [35S] methionine-labeled, glucose-grown cells processed with unlabeled methanol-grown cells, yielded sucrose gradient fractions that were radioactive in bands that were also recognized by F1 antibody. Isolated fractions in these experiments had similar ratios of cpm:pH 9.0 ATPase activities, suggesting that this activity is mitochondrial in origin. The data presented for the characteristics of the peroxisomally-associated activity strongly suggest that the majority of the ATPase activity found in peroxisomal preparations is derived from other organelles.     

Isolation of adenylate cyclase-enriched membranes from mammalian cells using concanavalin A.

Plasma membrane vesicles containing adenylate cyclase and beta-adrenergic receptors were prepared from 1321N1 human astrocytoma cells by a procedure involving the use of concanavalin A to stabilize the plasma membrane to fragmentation and vesiculation upon cell lysis. Treatment of cells with concanavalin A causes these plasma membrane markers to sediment to a higher density of sucrose and in a narrower band than observed with untreated cells. Upon treatment of the heavy membrane fragments with alpha-methylmannoside to remove bound concanavalin A, the enzyme markers again sediment a lower densities of sucrose. This reversible change in sedimentation behavior has been used to obtain preparations of plasma membranes enriched 14- to 21-fold (recovery 25%) in adenylate cyclase activity and about 12-fold (recovery 16%) in beta-adrenergic receptor density, as compared to lysates. The adenylate cyclase of purified membranes responded normally to isoproterenol and prostaglandin E1. Experiments with S49 and YAC mouse lymphoma cells and human skin fibroblasts indicate that this procedure may be adaptable to the isolation of plasma membranes from a variety of cultured cell lines.     

Identification of pathways associated with invasive behavior by ovarian cancer cells using multidimensional protein identification technology (MudPIT)

Proteomic profiling has emerged as a useful tool for identifying tissue alterations in disease states including malignant transformation. The aim of this study was to reveal expression profiles associated with the highly motile/invasive ovarian cancer cell phenotype. Six ovarian cancer cell lines were subjected to proteomic characterization using multidimensional protein identification technology (MudPIT), and evaluated for their motile/invasive behavior, so that these parameters could be compared. Within whole cell extracts of the ovarian cancer cells, MudPIT identified proteins that mapped to 2245 unique genes. Western blot analysis for selected proteins confirmed the expression profiles revealed by MudPIT, demonstrating the fidelity of this high-throughput analysis. Unsupervised cluster analysis partitioned the cell lines in a manner that reflected their motile/invasive capacity. A comparison of protein expression profiles between cell lines of high (group 1) versus low (group 2) motile/invasive capacity revealed 300 proteins that were differentially expressed, of which 196 proteins were significantly upregulated in group 1. Protein network and KEGG pathway analysis indicated a functional interplay between proteins up-regulated in group 1 cells, with increased expression of several key members of the actin cytoskeleton, extracellular matrix (ECM) and focal adhesion pathways. These proteomic expression profiles can be utilized to distinguish highly motile, aggressive ovarian cancer cells from lesser invasive ones, and could prove to be essential in the development of more effective strategies that target pivotal cell signaling pathways used by cancer cells during local invasion and distant metastasis.     

Proteomic screening of salt-stress-induced changes in plasma membranes of Synechocystis sp. strain PCC 6803

The plasma membrane of a cyanobacterial cell is crucial as barrier against the outer medium. It is also an energy-transducing membrane as well as essential for biogenesis of cyanobacterial photosystems and the endo-membrane system. Previously we have identified 57 different proteins in the plasma membrane of control cells from Synechocystis sp. strain PCC6803. In the present work, proteomic screening of salt-stress proteins in the plasma membrane resulted in identification of 109 proteins corresponding to 66 different gene products. Differential and quantitative analyses of 2-DE profiles of plasma membranes isolated from both control and salt-acclimated cells revealed that twenty proteins were enhanced/induced and five reduced during salt stress. More than half of the enhanced/induced proteins were periplasmic binding proteins of ABC-transporters or hypothetical proteins. Proteins that exhibited the highest enhancement during salt stress include FutA1 (Slr1295) and Vipp1 (Sll0617), which have been suggested to be involved in protection of photosystem II under iron deficiency and in thylakoid membrane formation, respectively. Other salt-stress proteins were regulatory proteins such as PII protein, LrtA, and a protein that belongs to CheY subfamily. The physiological significance of the identified salt-stress proteins in the plasma membrane is discussed integrating our current knowledge on cyanobacterial stress physiology.     

Plasmodium lophurae: Membrane Proteins of Erythrocyte-free Plasmodia and Malaria-Infected Red Cells*

SYNOPSIS. Plasma membranes of normal duckling erythrocytes were prepared by blender homogenization and nitrogen decompression. Surface membrane vesicles of red cells infected with the avian malaria Plasmodium lophurae were produced by nitrogen decompression. Membranes of erythrocyte-free malaria parasites were removed from cytoplasmic constituents by Dounce homogenization. These membranes were collected by centrifugation in a sucrose step gradient and purified on a linear sucrose gradient. Red cell membranes had a buoyant density of 1.159 g/cm³, whereas plasmodial membranes banded at 2 densities: 1.110 g/cm³ and 1.158 g/cm³. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis of the isolated red cell membranes revealed 7 major protein bands with molecular weights (MW) ranging from 230,000 to 22,000, and 3 glycoprotein bands with MW of 160,000, 88,000 and 37,000. Parasite membranes also had 7 major bands with MW ranging from 100,000 to 22,000. No glycoproteins were identifiable in these membranes. The proteins of the surface membranes from infected red cells had MW similar to those from normal red cells; however, there was some evidence of a reduction in the amount of the high MW polypeptides. The red cell membrane contained 79 nmoles sialic acid/mg membrane protein, whereas plasmodial membranes had 8 nmoles sialic acid/mg membrane protein. The sialic acid content of the surface membranes of infected red cells was significantly smaller than that of normal cells. Lactoperoxidase-glucose oxidase-catalyzed iodination of intact normal and malaria-infected erythrocytes labeled 7 surface components. Although no observable differences in iodinatable proteins were seen in these preparations, there was a striking reduction in the iodinatability of erythrocytic membranes obtained from P. lophurae-infected cells. Erythrocyte-free plasmodia bound very little radioactive iodine; the small amount of radioactivity was distributed among 3 major bands with MW of 42,000, 32,000 and 28,000. It is suggested that the alterations of the surface of the P. lophurae-infected erythrocyte do not occur by a wholesale insertion of plasmodial membrane proteins into the red cell plasma membrane, but rather that there are parasite-mediated modifications of existing membrane polypeptides.     

Calmodulin-binding proteins in granule and plasma membranes from bovine chromaffin cells

Calmodulin-binding proteins in chromaffin granule membrane and chromaffin cell plasma membranes have been investigated and compared. Chromaffin granules were purified by centrifugation over a 1.7 M sucrose layer. Plasma membranes were obtained in a highly purified form by differential and isopycnic centrifugation. Enzymatic determinations of 5'-nucleotidase, a generally accepted plasma membrane marker, showed a 40-50-fold enrichment as compared to the cell homogenate. Marker enzyme studies demonstrated only minimal contamination by other subcellular organelles. After solubilization with Triton X-100, calmodulin-binding proteins were isolated from chromaffin granule membranes and plasma membranes by affinity chromatography on a calmodulin/Sepharose 4B column. On two-dimensional polyacrylamide gelelectrophoresis a prominent protein (Mr = 65,000, pI ranging from 5.1 to 6) consisting of multiple spots, was present in the calmodulin-binding fraction from chromaffin granule membranes as well as from plasma membranes. Besides this 65 kDa protein both fractions had at least four groups of proteins in common. Also, proteins typical for either preparation were observed. In the calmodulin-binding protein preparations from chromaffin granule membranes a prominent spot with Mr = 80,000 and a pH ranging from 5.0 to 5.7 was present. This protein was enzymatically and immunologically identified as dopamine-beta-monooxygenase.     

Rapid preparation of nuclei-depleted detergent-resistant membrane fractions suitable for proteomics analysis

Background

Cholesterol-rich membrane microdomains known as lipid rafts have been implicated in diverse physiologic processes including lipid transport and signal transduction. Lipid rafts were originally defined as detergent-resistant membranes (DRMs) due to their relative insolubility in cold non-ionic detergents. Recent findings suggest that, although DRMs are not equivalent to lipid rafts, the presence of a given protein within DRMs strongly suggests its potential for raft association in vivo. Therefore, isolation of DRMs represents a useful starting point for biochemical analysis of lipid rafts. The physicochemical properties of DRMs present unique challenges to analysis of their protein composition. Existing methods of isolating DRM-enriched fractions involve flotation of cell extracts in a sucrose density gradient, which, although successful, can be labor intensive, time consuming and results in dilute sucrose-containing fractions with limited utility for direct proteomic analysis. In addition, several studies describing the proteomic characterization of DRMs using this and other approaches have reported the presence of nuclear proteins in such fractions. It is unclear whether these results reflect trafficking of nuclear proteins to DRMs or whether they arise from nuclear contamination during isolation. To address these issues, we have modified a published differential detergent extraction method to enable rapid DRM isolation that minimizes nuclear contamination and yields fractions compatible with mass spectrometry.

Results

DRM-enriched fractions isolated using the conventional or modified extraction methods displayed comparable profiles of known DRM-associated proteins, including flotillins, GPI-anchored proteins and heterotrimeric G-protein subunits. Thus, the modified procedure yielded fractions consistent with those isolated by existing methods. However, we observed a marked reduction in the percentage of nuclear proteins identified in DRM fractions isolated with the modified method (15%) compared to DRMs isolated by conventional means (36%). Furthermore, of the 21 nuclear proteins identified exclusively in modified DRM fractions, 16 have been reported to exist in other subcellular sites, with evidence to suggest shuttling of these species between the nucleus and other organelles.

Conclusion

We describe a modified DRM isolation procedure that generates DRMs that are largely free of nuclear contamination and that is compatible with downstream proteomic analyses with minimal additional processing. Our findings also imply that identification of nuclear proteins in DRMs is likely to reflect legitimate movement of proteins between compartments, and is not a result of contamination during extraction.     

ISOLATION OF RAT LIVER PLASMA MEMBRANES : Use of Nucleotide Pyrophosphatase and Phosphodiesterase I as Marker Enzymes

Nucleotide pyrophosphatase and phosphodiesterase I of rat liver have been found to be localized primarily in cell particulates highly enriched with respect to the most commonly accepted plasma membrane marker, 5'-nucleotidase, and therefore should themselves be assigned a plasma membrane localization. The observation that plasma membranes sediment in isotonic sucrose with both nuclear and microsomal fractions was exploited to obtain plasma membrane preparations from each fraction. Both preparations are similar in chemical and enzymic composition. Moreover, the preparative method developed in this study appears to give the best combination of yield, purity, and reproducibility available. The question of the possible identity of nucleotide pyrophosphatase and phosphodiesterase I is considered, and evidence is presented suggesting that these activities may be manifestations of the same enzyme.     

HeLa cell plasma membranes. Changes in membrane protein composition during the cell cycle.

HeLa cells grown in suspension culture were synchronized by amethopterin block and thymidine reversal. In some cases an additional Colcemid block was used to obtain mitotic cells. From the various phases of the cell cycle, cells were harvested and the plasma membranes isolated. The membrane proteins were solubilized in sodium dodecyl sulphate and separated by gel electrophoresis in the presence of sodium dodecyl sarcosinate. About 35 protein bands, five of which were stained with periodic acid-Schiff reagent, appeared. Most of the bands were identical in all membrane preparations, but a few minor bands seemed to be associated with limited periods of the cell cycle. In particular, the cells in mitosis apparently contained plasma membrane proteins which did not occur in other phases. Amino acid analyses of the plasma membranes revealed no significant cell cycle-dependent changes in the amino acid composition.