Membranes of Human Neutrophils Secretory Vesicle Membranes And

Comparison of Proteins Expressed on Secretory Vesicle Membranes and Plasma Membranes of Human Neutrophils

This information is current as Silvia M. Uriarte, David W. Powell, Gregory C. Luerman, of September 25, 2021. Michael L. Merchant, Timothy D. Cummins, Neelakshi R. Jog, Richard A. Ward and Kenneth R. McLeish J Immunol 2008; 180:5575-5581; ; doi: 10.4049/jimmunol.180.8.5575

http://www.jimmunol.org/content/180/8/5575 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2008/04/01/180.8.5575.DC1 Material http://www.jimmunol.org/ References This article cites 44 articles, 25 of which you can access for free at: http://www.jimmunol.org/content/180/8/5575.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision by guest on September 25, 2021 • No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2008 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Comparison of Proteins Expressed on Secretory Vesicle Membranes and Plasma Membranes of Human Neutrophils1

Silvia M. Uriarte,* David W. Powell,*† Gregory C. Luerman,† Michael L. Merchant,* Timothy D. Cummins,* Neelakshi R. Jog,‡ Richard A. Ward,* and Kenneth R. McLeish2*†§

Secretory vesicles are neutrophil intracellular storage granules formed by endocytosis. Understanding the functional consequences of secretory vesicle exocytosis requires knowledge of their membrane proteins. The current study was designed to use proteomic technologies to develop a more complete catalog of secretory vesicle membrane proteins and to compare the proteomes of secretory vesicle and plasma membranes. A total of 1118 proteins were identified, 573 (51%) were present only in plasma membrane- enriched fractions, 418 (37%) only in secretory vesicle-enriched membrane fractions, and 127 (11%) in both fractions. Gene Ontology categorized 373 of these proteins as integral membrane proteins. Proteins typically associated with other intracellular organelles, including nuclei, mitochondria, and ribosomes, were identified in both membrane fractions. Ingenuity Pathway Knowl- Downloaded from edge Base analysis determined that the majority of canonical and functional pathways were significantly associated with proteins from both plasma membrane-enriched and secretory vesicle-enriched fractions. There were, however, some canonical signaling pathways that involved proteins only from plasma membranes or secretory vesicles. In conclusion, a number of proteins were identified that may elucidate mechanisms and functional consequences of secretory vesicle exocytosis. The small number of common proteins suggests that the hypothesis that secretory vesicles are formed from plasma membranes by endocytosis requires more critical evaluation. The Journal of Immunology, 2008, 180: 5575–5581. http://www.jimmunol.org/

irculating neutrophils are capable of undergoing a series Golgi network during neutrophil maturation (1). The intragranule of phenotypic changes that result in their transition from constituents of secretory vesicles include plasma proteins, result- C cells that are poorly responsive to proinflammatory stim- ing in the hypothesis that secretory vesicles are formed by endo- uli to become the primary effector cells of innate immunity. These cytosis and that functional changes from their exocytosis are due phenotypic changes involve the incorporation of proteins from the entirely to incorporation of new molecules into the plasma mem- membranes of intracellular storage granules into the plasma membrane (1, 5). Thus, to understand the changes in neutrophil func- brane and the release of proteins stored in granule matrix through tional capability induced by secretory vesicle exocytosis, a com- by guest on September 25, 2021 regulated exocytosis. Granule exocytosis contributes to enhanced prehensive catalog of membrane proteins is required. neutrophil tethering and adhesion to vascular endothelial cells at a Proteomic techniques, which include methods for protein ex- site of inflammation; enhanced migration across blood vessel traction and separation, protein identification and characterization, walls; chemotaxis to a site of microbial invasion; phagocytosis of and database analysis, provide an unbiased approach to identifying invading organisms; and microbicidal activity through a combina- proteins expressed in subcellular compartments (6–9). We re- tion of enzymatic degradation, reactive oxygen species generation, cently published a comprehensive proteomic analysis of neutrophil and release of microbicidal peptides into phagosomes. Neutrophils gelatinase, specific, and azurophil granules (10). Using protein contain a heterologous group of storage granules that have been separation by two-dimensional gel electrophoresis and two-dimen- classified into four subsets based on density and composition: sional HPLC coupled with MALDI-TOF-MS3 and ESI-MS/MS, azurophil (primary) granules, specific (secondary) granules, gela- 286 proteins were identified on one or more granule subsets, many tinase (tertiary) granules, and secretory vesicles (1). These granule of which had not been found previously on neutrophil granules. subsets undergo hierarchical stimulated exocytosis, with secretory The current study was designed to use similar proteomic technol- vesicles the most easily and completely mobilized (2–4). Gelati- ogies to provide a more complete identification of secretory vesicle nase, specific, and azurophil granules are formed from the transmembrane proteins and to compare those proteins with the proteins expressed on neutrophil plasma membranes. The ability to extract and solubilize membrane proteins is a major limitation to all pro- *Department of Medicine, †Department of Biochemistry and Molecular Biology, and teomic approaches. To overcome this limitation, proteins were ‡Department of Microbiology and Immunology, University of Louisville, Louisville, extracted from membranes using a recently described methanol § KY 40202; and Veterans Affairs Medical Center, Louisville, KY 40206 extraction procedure, followed by two-dimensional HPLC and Received for publication October 29, 2007. Accepted for publication February ESI-MS/MS (11). With this approach, we identified a number of 7, 2008. membrane spanning and membrane associated proteins and uncov- The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance ered significant differences between secretory vesicle-enriched and with 18 U.S.C. Section 1734 solely to indicate this fact. plasma membrane-enriched proteomes. 1 This work was supported by a Merit Review Grant from the Department of Veterans Affairs (to K.R.M.), National Institutes of Health Grants DK62389 (to R.A.W. and K.R.M.) and DK176743 (to D.W.P.), and the Office of Science Financial Assistance Program, Department of Energy (to D.W.P.). 3 Abbreviations used in this paper: MS, mass spectrometry; SCX, strong cation ex- 2 Address correspondence and reprint requests to Dr. Kenneth R. McLeish, Baxter I change; RP, reversed-phase; PAF, protein abundance factor; NCBI, National Center Research Building, University of Louisville, 570 South Preston Street, Louisville, KY for Biotechnology Information; IPKB, Ingenuity Pathways Knowledge Base; 40202. E-mail address: [email protected] SNARE, soluble NSF attachment receptor. www.jimmunol.org 5576 NEUTROPHIL SECRETORY VESICLE PROTEOME

Materials and Methods acid and mobile phase B: 80% acetonitrile/0.1% formic acid). Spectra were Neutrophil isolation acquired with a LTQ linear ion trap mass spectrometer (Thermo Fisher Sci- entific). During LC-MS/MS analysis, the mass spectrometer performed data- Neutrophils were isolated from healthy donors using plasma-Percoll gradients dependent acquisition with a full MS scan between 300 and 2000 mass to as described by Haslett et al. (12). Trypan blue staining revealed that at least change ratio followed by five MS/MS scans (35% collision energy) on the five 97% of cells were neutrophils with Ͼ95% viability. After isolation, neutro- most intense ions from the preceding MS scan. Data acquisition was per- phils were suspended in Krebs-Ringer phosphate buffer (pH 7.2) at 4 ϫ 107 formed using dynamic exclusion with a repeat count of 30 anda1and3min cells/ml and treated with 5 mM diisopropyl fluorophosphate for 10 min on ice exclusion duration window. to inhibit proteases (13). Extensive evaluation using respiratory burst activity and granule exocytosis indicates that this isolation technique does not prime Mass spectral data interpretations neutrophils. The Human Studies Committee of the University of Louisville approved the use of human donors. The acquired mass spectrometric data were searched against a human protein database (human RefSeq) using the Sequest algorithm and a commercial com- Plasma membrane and secretory vesicle membrane enrichment putational platform (SequestSorcerer, Sage-N Research) assuming modifica- tions of oxidation of methionine (ϩ15.99) and carbamidomethylation of cys- Neutrophil plasma membrane and secretory vesicle membranes were enriched ϩ by centrifugation on a two-layer Percoll density gradient as described by Dahl- teine ( 57). High-probability protein identifications were assigned from the gren et al. (14). Briefly, isolated neutrophils from single donors (4 ϫ 107/ml) Sequest results using the BIGCAT (16) and ProteinProphet (17) statistical were incubated with diisopropyl fluorophosphate, pelleted by centrifugation, platforms (18, 19). Both of these programs eliminate redundant listing by and resuspended in disruption buffer containing 100 mM KCl, 1 mM NaCl, 1 grouping proteins with 100% identity and merging proteins with 100% shared MS/MS spectra. The BIGCAT filter uses Sequest Xcorr cut-offs of 1.5, 2, and mM ATPNa2, 3.5 mM MgCl2, 10 mM PIPES, and 0.5 mM PMSF. Cells were ϩ ϩ ϩ disrupted by nitrogen cavitation at 380 p.s.i. and 4°C. The cavitate was col- 2.5 for 1, 2, and 3 ions, respectively. Proteins were also ranked by relative abundance or enrichment using a protein abundance factor (PAF). The

lected, supplemented with 1.5 mM EGTA, and nuclei and intact cells were Downloaded from removed by centrifugation at 400 ϫ g for 5 min. The supernatant membrane PAF is defined as the total number of nonredundant spectra (spectral counts) that correlate significantly to each respective candidate protein normalized to suspension was aspirated, placed in a 50-ml conical centrifuge tube, and mixed ϫ 4 with an equal volume of a 1.12 g/ml Percoll gradient. A total of 10 ml of the the protein’s m.w. ( 10 ). Studies demonstrating linearity between the num- membrane suspension/Percoll gradient was layered under 5 ml of disruption ber of spectral counts and protein concentration provide the framework for this buffer in a 50-ml ultracentrifuge tube. A total of 10 ml of 1.05 g/ml Percoll type of label-free quantitative analysis from 2D-LC-MS/MS experiments gradient solution was layered under the membrane suspension, then 5 ml of (20–22). The PAF approach has been highly successful in the development of 1.12 g/ml Percoll gradient solution layered under the 1.05 g/ml solution. The statistical models based on 2D-LC-MS/MS experimental data (23–25). gradient was centrifuged at 37,000 ϫ g in a SS-34 fixed angle rotor in a Sorvall ProteinProphet gives each protein a ranked probability score, with 1.0 being http://www.jimmunol.org/ RC-5B centrifuge for 30 min at 4°C. Following centrifugation, successive the highest probability. Our results were fitted into a model where probability 1.5-ml fractions were collected from the top of the gradient. scores decreasing from 1.0 were correlated to a predicted false positive identification error rate. Proteins with a probability score greater than 0.65 pre- Protein extraction from plasma membrane and secretory vesicle dicted a false positive error of Ͻ10%, and those proteins were included in the membranes list of candidates. The proteins were further analyzed using the National Center for Biotech- Fractions obtained from Percoll gradients were analyzed for alkaline phos- nology Information (NCBI) database. Any proteins that were removed from phatase in the absence (nonlatent) or presence (latent) of Triton X-100 the database in the process of annotation or that were identified as a bacterial using p-nitrophenylphosphate as substrate. Briefly, 100:l aliquots of each protein were excluded from the final protein list. To identify integral mem- fraction were added to wells of a 96 flat-well plate with or without 0.3% brane proteins and to determine intracellular location and function, identified Triton X-100. Reactions were started by adding 200 ␮l of reaction buffer proteins were analyzed using the NCBI database, by Gene Ontology (26), and by guest on September 25, 2021 containing 5 mM p-nitrophenylphosphate in 100 mM 2-amino-2-methyl- by the Ingenuity Pathways Knowledge Base (IPKB) (Ingenuity Systems) (27). 1-propanol (pH 10.0). Following incubation for 30 min at 37°C, reactions The IPBK is a comprehensive knowledge base of biological findings for genes were stopped by addition of 150 ␮l of 0.04 N NaOH and absorbance was of human, mouse, and rat origin, which is used to construct pathways and read at 405 nm in a microplate reader. Fractions containing nonlatent, but define biological functions (28). The canonical pathways are well-character- not latent, alkaline phosphatase were pooled and analyzed as plasma mem- ized metabolic and cell signaling pathways that have been curated from spe- brane-enriched. Fractions containing latent alkaline phosphatase were cific journal articles, review articles, text books, and KEGG Ligand. The func- pooled and analyzed as secretory vesicle-enriched membrane. Percoll was tional analysis has three primary categories of functions: molecular and removed from membranes by ultracentrifugation at 100,000 ϫ g for 90 cellular functions; physiological system development and function; and dis- min. Membrane pellets were resuspended in 50 mM ammonium bicarbonate, eases and disorders. There are 85 high-level functional categories that are washed by centrifugation at 100,000 ϫ g, then resuspended in 60% methanol classified under these categories. The significance value of a given canonical in 100 mM ammonium bicarbonate to extract membrane proteins (11). Pro- pathway or functional analysis category is a measurement of the likelihood that teins were digested by incubation with 3 ␮g each of trypsin and chymotrypsin the pathway or function is associated with the data set by random chance. The at 37°C overnight. Following centrifugation at 100,000 ϫ g for 20 min at 4°C, value of p is calculated using the right-tailed Fisher Exact Test, and values of the supernatant was removed for peptide analysis. Peptides were lyophilized p Ͻ 0.05 were a priori assumed to be statistically significant. and prepared for mass spectrometry analysis using a desalting trap method. Briefly, peptides were resuspended in a 100 ␮l of 5% acetonitrile and 0.05% formic acid and applied to a peptide microtrap (Michrom BioResources) equil- Results ibrated with 1 ml of the same buffer. The trap was then washed twice with Neutrophils from four individual donors were subjected to mem- 100 ␮l of resuspension buffer and peptides eluted with 100 ␮l of 95% aceto- brane fractionation, and each set was analyzed separately by mass nitrile, 0.05% formic acid. Eluted peptides were dried in a speed vacuum and spectrometry. A total of 1118 proteins were identified, for which resuspended in 5–10 ␮l of 5% acetonitrile and 0.05% formic acid. there was a Ͻ10% probability that the identification was by chance 2D-LC-MS/MS and computer-assisted data analysis according to Protein Prophet, and which were confirmed to have A modified version of a previously described 2D-LC-MS/MS method was mammalian homologues according to the NCBI database. Of the applied (15). Trypsin/Chymotrypsin-generated peptides were loaded onto an 1118 proteins, 573 were present only in plasma membrane-en- analytical 2D microcapillary chromatography column packed with 3–4 cm of riched fractions, 418 only in secretory vesicle-enriched fractions, 5 ␮m (pore size) strong cation exchange (SCX) resin (Phenomenex) followed and 127 in both fractions. The 1118 proteins, together with their ␮ by 2–3 cm of 5 m (pore size) C18 reversed-phase (RP) resin (Phenomenex). probability score, the peptides used to identify them, the mem- This bi-phasic column was attached to an analytical RP chromatography column (100 ϫ 365 ␮m fused silica capillary with an integrated, laser pulled brane fraction from which they were identified, and the Gene On- emitter tip packed with 10 cm of Synergi 4 ␮m RP80A (Phenomenex)). Pep- tology analysis of membrane association, cellular location, and tides were eluted from SCX with seven step gradients of 5, 10, 15, 30, 50, 70, function, are listed in supplemental data Tables I and II.4 and 100% 500 mM ammonium acetate. Following each SCX elution step, peptides were ionized and sprayed into the mass spectrometer using the following linear RP gradient: 20 min: 0% B, 80 min: 40% B, 90 min: and 60% B at a flow rate of 200 nl/min (mobile phase A: 5% acetonitrile/0.1% formic 4 The online version of this article contains supplemental material. The Journal of Immunology 5577

Table I. Distribution of proteins by membrane association and cell locationa

Membrane Association Location Membrane-Associated Membrane-Independent Unknown Total

Cytoplasm 0 151 1 152 Cytoskeleton 0 65 1 66 Endoplasmic reticulum 27 13 4 44 Endosome/lysosome/peroxisome 9 6 1 16 Extracellular 0 44 0 44 Golgi 6 4 3 13 Mitochondria 24 31 4 59 Nucleus 9 196 8 213 Plasma membrane 268 0 0 268 Ribosome 0 25 0 25 Secretory granule 16 17 1 34 Unknown 14 0 170 184 Total 373 552 193 1118

a All identiﬁed proteins were categorized by Gene Ontology as integral membrane proteins (membrane-associated), as not associated with membranes (membrane-independent), and proteins whose membrane association was unknown and then listed by cellular location. Downloaded from

Recovery of membrane-associated proteins proteins were primarily localized to the endoplasmic reticulum In the past, protein extraction methods used in proteomic analysis (27), mitochondria (24), and secretory granules (16). The majority have provided poor identification of transmembrane proteins. To of proteins not associated with a cellular membrane were localized address this limitation, we used a recently described protein ex- in the cytoplasm (151), nucleus (196), or with the cytoskeleton traction method in which 60% methanol was added to 100 mM (65). The majority of proteins for which membrane association was http://www.jimmunol.org/ ammonium bicarbonate, as recently described by Fischer et al. unknown also had an unknown cellular location (169 of 193 proteins). (11). To determine the effectiveness of methanol extraction in the recovery of membrane proteins, each protein was analyzed for Cellular location of proteins identified in plasma membrane- known association with cellular membranes using the Gene On- enriched and secretory vesicle-enriched membrane fractions tology database. Of the 1118 proteins identified, 373 were catego- Table II shows the cellular location of proteins identified in plasma rized as integral membrane proteins, 552 were not associated with membrane-enriched and secretory vesicle-enriched fractions. any cellular membranes, and the membrane association of 193 Based on the Gene Ontology database, 25% of proteins were pre-

proteins was unknown (Table I). A number of proteins with single dicted to be from mitochondria, ribosomes, or nuclei. Of the re- by guest on September 25, 2021 transmembrane spanning regions were identified, including maining 821 proteins, over half were proteins known to be asso- Fc␥RIIa and IIIa, integrins ␣2b, ␣3, ␣4, ␣D, ␣M, ␣X, ␤1, and ␤2; ciated with the cytoskeleton or a cellular membrane compartment, TLRs 2 and 8; and complement receptors 1 and 2. Numerous pro- including plasma membrane, secretory granules, endoplasmic re- teins with multiple transmembrane spanning regions were also ticulum, endosomes, or golgi. Thus, the fractions enriched for identified, including seven transmembrane spanning receptors for plasma membrane and secretory vesicle membranes also contain leukotiene B4, formyl peptides, IL-8, C5a, and a number of ion proteins from a number of other cellular compartments. Another channels and transporters. Table I also shows the cellular location recent analysis reported that proteins from endoplasmic reticulum, of proteins that were integral to membranes, independent of mem- mitochondria, and golgi coisolated with neutrophil secretory ves- branes, or where membrane association was unknown. As ex- icle-enriched membranes and enriched plasma membranes sepa- pected, all 268 plasma membrane proteins were categorized as rated by free flow electrophoresis (29). Of the 1118 proteins iden- membrane-associated. The remainder of the membrane-associated tified, 51% were found in plasma membrane-enriched fractions,

Table II. Distribution of proteins by cell locationa

Plasma Common to PM Location Membrane Secretory Vesicle and SV Total

Cytoplasm 70 68 14 152 Cytoskeleton 30 22 14 66 Endoplasmic reticulum 25 9 10 44 Endosome/lysosome/peroxisome 4 9 3 16 Extracellular 23 16 5 44 Golgi 5 7 1 13 Mitochondria 33 15 11 59 Nucleus 113 81 19 213 Plasma membrane 159 76 33 268 Ribosome 1 22 2 25 Secretory granule 10 15 9 34 Unknown 100 78 6 184 Total 573 418 127 1118

a All proteins identiﬁed were categorized by the membrane fraction from which they were identiﬁed, plasma membrane (PM), secretory vesicle (SV) membrane, or present in both membrane fractions (common to PM and SV). These proteins were then listed by cell location based on Gene Ontology analysis. 5578 NEUTROPHIL SECRETORY VESICLE PROTEOME

Table III. Distribution of proteins by functional categorya

Plasma Common to PM Function Membrane Secretory Vesicle and SV Total

Adhesion 31 14 6 51 Chaperone/protein folding 7 4 6 17 Chromosome organization 1 0 0 1 Cytoskeletal regulation 29 27 16 72 Enzyme/metabolism 28 23 8 59 GTPase 13 20 5 38 Immune/inflammatory response 8 4 5 17 Kinase/phosphatase 33 15 3 49 Membrane trafficking 30 22 5 57 Protein Modification/targeting 7 1 2 10 Proteolysis 23 22 5 50 Receptor/signal transduction 49 27 7 83 Redox 5 0 4 9 Structural proteins 2 7 2 11 Translation/transcription 9 8 2 19 Transport 33 16 3 52 Miscellaneous 12 11 5 28 Unknown 108 79 11 198 Downloaded from Total 426 300 95 821

37% in secretory vesicle-enriched fractions, and 11% were found icle, proteins localized to mitochondria, nuclei, and ribosomes http://www.jimmunol.org/ in both fractions. When proteins were analyzed according to their were eliminated from the analysis. All proteins identified from cellular location, several patterns were observed. Proteins local- plasma membrane-enriched fractions and secretory vesicle-en- ized to the endoplasmic reticulum, extracellular space, mitochon- riched fractions, however, are listed in supplementary Table II. dria, nucleus, plasma membrane, and whose location was unknown Table III shows the distribution of the remaining 821 proteins from were more likely to be present in plasma membrane-enriched frac- plasma membrane-enriched and secretory vesicle-enriched fractions. Proteins localized from endosome/lysosome/peroxisome tions according to protein function. Functional categories were as- compartments, golgi, ribosomes, and secretory granules were more signed based on analysis by the Gene Ontology database. Where likely to be present in secretory vesicle-enriched fractions. Proteins multiple functions were assigned to a protein, the function most defined as cytoplasmic or cytoskeletal were fairly equally distrib- likely to be pertinent to neutrophils was chosen. The majority of by guest on September 25, 2021 uted between plasma membrane-enriched and secretory vesicle- proteins were classified into the following functional categories: enriched fractions. Surprisingly, the number of proteins present in adhesion, cytoskeletal regulation, enzyme/metabolism, kinase/ both fractions was low for all Gene Ontology cell locations, rang- phosphatase, membrane trafficking, proteolysis, receptor/signal ing from 3% for proteins of unknown location to 26% for proteins transduction, and transport. For a given functional class, plasma in secretory granules. These results indicate that there was effec- membrane-enriched fractions and secretory vesicle-enriched frac- tive separation of cellular membranes based on their density, and tions contained either an approximately equal number of proteins the data suggest that proteins from all cellular compartments seg- or there were more proteins in the plasma membrane-enriched regate into distinct membrane compartments of different densities. fractions. Only a limited number of proteins in each functional category were found in both fractions, ranging from 6% of trans- Functional categorization of proteins port proteins to 22% of cytoskeletal regulatory proteins. To obtain a more accurate assessment of the functions of proteins In addition to identification of 57 proteins involved in mem- most likely associated with plasma membrane and secretory ves- brane trafficking, a total of 38 GTPases or their regulatory proteins

Table IV. Pattern of functions of plasma membrane and secretory vesicle proteinsa

Secretory Vesicle Proteins Secretory Vesicle and Common Only Proteins Plasma Membrane, Secretory Vesicle, and Common Proteins

Gene expression Protein degradation Cell signaling Carbohydrate metabolism Protein folding Protein trafficking Cellular movement Cell death Free radical scavenging Cell morphology Molecular transport Nucleic acid metabolism Tissue development Immune system function Cellular development Immunologic disease Cell growth and proliferation Post-translational modification Infectious disease Mineral metabolism Immune response Lipid metabolism Cellular assembly and organization Metabolic disease Cell-to-cell signaling and interaction Hematological disease Inflammatory disease

a Proteins, excluding those identified as nuclear, mitochondrial, or ribosomal, were categorized by the membrane fraction from which they were identified, plasma membrane (PM), secretory vesicle (SV) membrane, or present in both membrane fractions (common to PM and SV). Each of these categories was analyzed by Ingenuity Pathway Knowledge Base to determine the likelihood that functional pathways were associated with the dataset by random chance. Only those functional pathways significantly associated with proteins from one or more membrane fractions are listed. The Journal of Immunology 5579

Table V. Pattern of proteins in canonical pathwaysa

PM and SV Proteins and/or Common Proteins SV Proteins PM Proteins

Actin cytoskeleton signaling cAMP-mediated signaling Phosphatase and tensin homolog signaling Integrin signaling G-protein coupled receptor signaling Apoptosis signaling Leukocyte extravasation signaling Fibroblast growth factor signaling NK cell signaling Calcium signaling PI3K/AKT signaling TLR signaling Complement and coagulation cascade Inositol metabolism WNT/␤-catenin signaling Vascular endothelial growth factor Fc⑀R signaling TGF-␤ signaling signaling Ag presentation pathway NF-␬B signaling Protein ubiquitination pathway Oxidative stress response Epidermal growth factor signaling Platelet-derived growth factor signaling ERK/MAPK signaling JNK signaling

categories was analyzed by Ingenuity Pathway Knowledge Base to determine the likelihood that canonical pathways were associated with the dataset by Downloaded from random chance. Only those canonical pathways significantly associated with proteins from one or more membrane fractions are listed. were identified. Thus, 95 proteins were identified that are recog- included G-protein coupled receptor signaling, fibroblast growth nized to participate in regulation of events leading to membrane factor signaling, PI3K/AKT signaling, and Fc␧R signaling. Ca- trafficking. These proteins were equally distributed between nonical pathways containing proteins only identified in plasma

plasma membrane-enriched and secretory vesicle-enriched frac- membrane-enriched fractions included phosphatase and tensin ho- http://www.jimmunol.org/ tions, and only 10 proteins were present in both fractions (Rap1B, molog signaling, TLR signaling, TGF-␤ signaling, and NF-␬B sig-

Rab1B, Rab5C, Rab7, G␣i2, dynein 8, kinesin 27, reticulon 3c, naling. Thus, plasma membrane-enriched and secretory vesicle- testilin, and secretory carrier membrane protein 2). enriched fractions contained proteins common to a number of functions and signaling pathways. There were, however, protein Functional and canonical signaling pathways associated with components of signaling pathways unique to plasma membrane proteins in plasma membrane and secretory vesicle-enriched and to secretory vesicles. membrane fractions To better understand the functional roles of plasma membrane and Discussion secretory vesicle proteins, the 821 proteins, excluding those local- Understanding the functional consequences of secretory vesicle by guest on September 25, 2021 ized to mitochondria, nuclei, and ribosomes, were subjected to exocytosis in neutrophils requires knowledge of the membrane IPKB analysis for canonical pathways and functional pathways. proteins of this organelle. Previous studies using primarily Ab- Canonical pathways are defined as well-characterized metabolic based methods identified individual proteins from secretory vesicle and cell signaling pathways, whereas functional pathways contain membranes, including cytochrome b558 oxidase, CD11b/CD18 ad- three primary categories of functions: molecular and cellular func- hesion molecules, complement receptor 1 (CR1 or CD35), formyl tions; physiological system development and function; and dis- peptide receptors, Fc(RIIIa) (CD16), membrane metalloendopep- eases and disorders. Of the 821 proteins, 743 could be mapped tidase (CD10), aminopeptidase N (CD13), the TLR complex mol- using the IPKB, 368 proteins from the plasma membrane-enriched ecule CD14, the transmembrane protein tyrosine phosphatase fractions, 282 proteins from the secretory vesicle-enriched frac- CD45, V-type Hϩ-ATPase, the soluble NSF attachment receptor tions, and 93 protein identified in both fractions. Table IV lists the (SNARE) protein VAMP-2, and the metalloproteinase leukolysin functional pathways that demonstrated a significant association (1). Subcellular fractionation, combined with mass spectrometry- with proteins identified in each membrane fraction. The vast ma- based proteomics, represents a powerful approach to unbiased jority of functions, including cell signaling; cellular movement; identification of the protein composition of intracellular organelles hematologic, infectious, immunologic, and inflammatory diseases; (6–9). The current study applied these techniques to develop a and molecular transport, were significantly associated with pro- more complete catalog of neutrophil secretory vesicle membrane teins in all three groups, plasma membrane-enriched fractions, se- proteins and to compare the proteomes of secretory vesicle mem- cretory vesicle-enriched fractions, and proteins common to both branes and plasma membranes. Over 1100 proteins were identi- fractions. Thus, proteins performing these functions were not seg- fied, the majority of which segregated to either the plasma mem- regated into any particular membrane fraction. Table V lists the brane-enriched fractions or the secretory vesicle-enriched canonical pathways identified by the presence of multiple proteins membrane fractions; only 11% of identified proteins were present from each pathway in the membrane fraction. Three patterns were in both membrane fractions. observed. The largest number of canonical pathways contained Two significant problems affect the interpretation of our data. proteins present in both plasma membrane-enriched and secretory First, the sensitivity of proteomic analysis makes the ability to vesicle-enriched fractions, and/or proteins which were common to obtain highly purified intracellular organelles the limiting factor in both membrane fractions. This group of pathways included actin establishing the proteome of a specific organelle. Based on the cytoskeletal signaling, integrin signaling, leukocyte extravasation classification of our proteins by cell location and function using signaling, protein ubiquitination, oxidative stress, ERK or JNK Gene Ontology, proteins typically associated with other intracel- signaling, and growth factor (platelet-derived growth factor and lular organelles, notably nuclei, mitochondria, and ribosomes, epidermal growth factor) signaling. Canonical pathways contain- were identified in both secretory vesicle-enriched and plasma ing proteins only identified in secretory vesicle-enriched fractions membrane-enriched fractions. One reason for the presence of those 5580 NEUTROPHIL SECRETORY VESICLE PROTEOME contaminating proteins is that nitrogen cavitation partially disrupts proteins in secretory vesicle-enriched fractions, whereas only 7 a number of intracellular compartments, including mitochondria, proteins (10%) were common to both fractions. Several explana- golgi, endoplasmic reticulum, and nuclei. In the case of neutro- tions were considered to account for the low percentage of proteins phils, this disruption is reported to extend to intracellular storage that were common to both membrane fractions. First, secretory granules (10, 29–31). Soluble proteins released from these gran- vesicles are not endocytic vesicles derived from the plasma mem- ules likely associate and cosediment with membranes from gran- brane, as has been previously postulated (1, 5). Second, the protein ule-free fractions, consistent with our finding myeloperoxidase, extraction and/or identification techniques failed to identify pro- elastase, and lactoferrin in plasma membrane and secretory vesicle teins common to the two membranes. Third, separation of plasma membrane fractions. It is also likely that membranes released by membrane and secretory vesicle membrane by density results in partial disruption of other organelles cosediment in plasma mem- identification of proteins that segregate according to membrane brane-enriched and secretory vesicle-enriched membrane frac- density, rather than by organelle. It is unlikely that the sensitivity tions. Jethwaney et al. (29) identified proteins derived from mito- of proteomic techniques would vary between the two membrane chondria and endoplasmic reticulum in secretory vesicle-enriched fractions, resulting in the failure to identify proteins common to fractions derived by free-flow electrophoresis, whereas proteins both. The localization of latent alkaline phosphatase to the mem- from these and other organelles were distributed in both membrane brane fraction with higher density suggests that proteins associated fractions in the current study. Whether the presence in both mem- with secretory vesicles were limited to one group of membrane brane fractions of proteins from a broader range of organelles in fractions. Our data do not allow us to distinguish between the our study reflects differences in membrane enrichment or protein possibility that secretory vesicles are not endocytic vesicles or that identification techniques between the two studies cannot be deter- all membranes contain domains of lighter and heavier density to Downloaded from mined. No direct comparison of free-flow electrophoresis and den- which distinct sets of proteins localize. However, the data suggest sity-gradient centrifugation enrichment of secretory vesicle-en- that the hypothesis that secretory vesicles are formed from plasma riched and plasma membrane-enriched fractions has been membranes by endocytosis requires more critical evaluation. performed. Although the presence of latent alkaline phosphatase The cell functional classification based on the Gene Ontology activity indicates that both methods obtain fractions containing database revealed a number of proteins that may elucidate mech-

secretory vesicles, a comparative study is needed to determine sim- anisms of secretory vesicle exocytosis. Examination of the 38 http://www.jimmunol.org/ ilarities and differences between the two methods. GTPases, the majority of which were identified in secretory vesi- A total of 73 proteins were assigned by Gene Ontology to endo- cle-enriched fractions, revealed a number of Rab proteins known somes, golgi, or endoplasmic reticulum. The possible localization of to regulate membrane trafficking. Secretory vesicle-enriched frac- those proteins to plasma membranes or secretory vesicles, rather than tions contained Rab11a, Rab14, Rab15, and Rab35; plasma mem- other intracellular membrane compartments, was not examined fur- brane-enriched fractions contained Rab5b and Rab31, and Rab1b, ther in this study. A recent report, in which proteomic analysis of Rab5c, and Rab7 were common to both plasma membrane and phagosomes was performed, determined that endoplasmic reticulum secretory vesicle-enriched fractions. Rab5 was reported to play a fuse with maturing phagosomes in macrophages, suggesting one significant role in chemoattractant receptor endocytosis and fusion mechanism by which proteins may be “shared” by different intracel- of intracellular granules with phagosomes in human neutrophils by guest on September 25, 2021 lular compartments (9). Thus, it is likely that some proteins that localize (32–34). Rab5a was shown to undergo a significant translocation to cellular components, such as endoplasmic reticulum or endosomes, from endosomes and secretory vesicles to the plasma membrane may also be associated with secretory vesicles or plasma membrane. with stimulation of human neutrophils (35), and Rab11 was re- The second problem with application of proteomic techniques to ported to regulate endocytic-exocytic cycling of integrin molecules identification of proteins in secretory and plasma membranes is the (36). The roles of Rab 14, Rab15, Rab31, and Rab35 have not been difficulty extracting and identifying transmembrane proteins contain- examined in neutrophils. A total of 57 proteins were identified that ing ␣ helices. Identifying transmembrane proteins is made difficult by play a role in membrane trafficking, 30 were in the plasma mem- the absence of sites for tryptic cleavage in transmembrane regions, brane, 22 in secretory vesicles, and 5 were common to both. The variability in the size of exposed hydrophobic regions, low abundance present study found two SNARE proteins, VAMP-3 and VAMP-8, of transmembrane proteins, poor separation by two-dimensional gel on secretory vesicle-enriched membranes. VAMP-1, Ϫ2, Ϫ7, and electrophoresis of integral membrane proteins, and poor solubility of Ϫ8 were reported previously to be present in human neutrophils, hydrophobic peptides. Fischer et al. (11) recently reported that tryptic and VAMP-2 was identified on secretory vesicles (37–40). digestion of bacterial membrane proteins extracted in 60% methanol VAMP-3 has not been found previously in human neutrophils, increased identification of integral membrane proteins from 20 to 50% although it has been identified in human platelets and human of the total proteins identified. Our results suggest that this approach plasma cells (41, 42). Mollinedo et al. (39, 40) reported that is useful in membranes from mammalian cells, as one-third of the VAMP-1 and VAMP-2 mediated exocytosis of specific granules, proteins identified in the present study were classified as integral whereas VAMP-1 and VAMP-7 mediated azurophil granule exo- membrane proteins by Gene Ontology. cytosis. The SNARE proteins involved in secretory vesicle exo- The observation that only 11% of proteins identified were com- cytosis have not been determined. In addition to SNARE proteins, mon to both plasma membrane-enriched and secretory vesicle-en- the exocyst complex is a group of eight proteins involved in ves- riched fractions suggests that Percoll density-gradient centrifuga- icle targeting and docking at the plasma membrane (43, 44). Two tion effectively enriched two different membrane populations. components of the exocyst complex were identified in the current Combined with the assays for total and latent alkaline phosphatase, study: exocyst complex component 2 (Sec3) was identified from our findings suggest that there are greater differences in the protein plasma membrane-enriched fractions, and exocyst complex com- content of secretory vesicles and plasma membrane than previ- ponent 5 (Sec10) was identified from secretory vesicle-enriched ously appreciated. Jethwaney et al. (29) used free-flow electro- fractions. Boyd et al. (45) reported that Sec3p localized to the phoresis to separate plasma membrane from secretory vesicles, plasma membrane of Saccharomyces cerevisiae, while Sec10p was separated proteins by SDS-PAGE, and identified proteins using found on vesicles. The exocyst complex is an effector for multiple HPLC-MS/MS. Similar to our results, these authors identified 30 GTPases, including cdc42 and Rab11 (46, 47). Both of these proteins present in the plasma membrane-enriched fractions and 36 GTPases were identified, cdc42 from plasma membrane-enriched The Journal of Immunology 5581 fractions and Rab11A from secretory vesicle-enriched fractions. 21. Old, W. M., K. Meyer-Arendt, L. Aveline-Wolf, K. G. Pierce, A. Mendoza, These findings suggest that the exocyst complex plays a role in teth- J. R. Sevinsky, K. A. Resing, and N. G. Ahn. 2005. Comparison of label-free methods for quantifying human proteins by shotgun proteomics. Mol. Cell. Pro- ering secretory vesicles to the plasma membrane before SNARE-me- teomics 4: 1487–1502. diated membrane fusion. Consistent with our previous studies which 22. Wienkoop, S., E. Larrainzar, M. Niemann, E. M. Gonzalez, U. Lehmann, and W. Weckwerth. 2006. Stable isotope-free quantitative shotgun proteomics combined showed large amounts of actin associated with both neutrophil plasma with sample pattern recognition for rapid diagnostics. J. Sep. Sci. 29: 2793–2801. membranes and secretory vesicle membranes (4), 72 cytoskeletal and 23. Fleischer, T. C., C. Weaver, K. J. McAfee, J. L. Jennings, and A. J. Link. 2006. cytoskeletal regulatory and binding proteins were equally distributed Systematic identification and functional screens of uncharacterized proteins as- in both membrane fractions. Two groups of actin assembly factors sociated with eukaryotic ribosomal complexes. Genes Dev. 20: 1294–1307. 24. Powell, D. W., C. Weaver, J. C. Jennings, K. J. McAfee, Y. He, P. A. Weil, and were identified. Formins act in conjunction with profilin to drive nu- A. J. Link. 2004. Cluster analysis of mass spectrometry data reveals a novel cleation, but not branching, of actin filaments (48). Formin 1 and 2 component of SAGA. Mol. Cell. Biol. 24: 7249–7259. were identified from plasma membrane-enriched fractions, and pro- 25. Sanders, S. L., J. Jennings, A. Canutescu, A. Link, and P. A. Weil. 2002. Pro- teomics of the eukaryotic transcription machinery: identification of proteins as- filin 1 was common to both fractions. The Arp2/3 complex acts as a sociated with components of yeast TFIID by multidimensional mass spectrom- nucleation and branching factor, and members of this complex were etry. Mol. Cell. Biol. 22: 4723–4738. identified from secretory vesicle-enriched fractions. Although it was 26. http://www.geneontology.org 27. www.ingenuity.com suggested previously that actin and actin-associated proteins reflect 28. Calvano, S., W. Xiao, D. R. Richards, R. M. Felcano, H. V. Baker, R. J. Cho, cytoskeletal contamination of these membrane fractions (29), the high R. O. Chen, B. H. Brownstein, J. P. Cobb, S. K. Tschoeke, et al. 2005. A network- content of cytoskeletal proteins and the importance of actin reorgani- based analysis of systemic inflammation in humans. Nature 437: 1032–1037. 29. Jethwaney, D., M. R. Islam, K. G. Leidal, D. B.-V. de Bernabe, K. P. Campbell, zation in exocytosis of secretory granules make it more likely that W. M. Nauseef, and B. W. Gibson. 2007. Proteomic analysis of plasma mem- cytoskeletal proteins are associated with plasma and secretory vesicle brane and secretory vesicles from human neutrophils. Proteome Sci. 5: 12. Downloaded from membranes (4, 10). 30. Kjeldsen, L., H. Sengelov, K. Lollike, M. H. Nielsen, and N. Borregaard. 1994. Isolation and characterization of gelatinase granules from human neutrophils. Blood 83: 1640–1649. Disclosures 31. Borregaard, N., J. M. Heiple, E. R. Simons, and R. A. Clark. 1983. Subcellular The authors have no financial conflict of interest. localization of the b-cytochrome component of the human neutrophil microbicidal oxidase: translocation during activation. J. Cell Biol. 97: 52–61. References 32. Perskvist, N., K. Roberg, A. Kulyte, and O. Stendahl. 2001. Rab5a GTPase regulates fusion between pathogen-containing phagosomes and cytoplasmic or- http://www.jimmunol.org/ 1. Faurschou, M., and N. Borregaard. 2003. Neutrophil granules and secretory ves- ganelles inhuman neutrophils. J. Cell Sci. 115: 1321–1330. icles in inflammation. Microbes Infect. 5: 1317–1327. 33. Rose, J., J. F. Foley, P. M. Murphy, and S. Venkatesan. 2004. On the mechanism 2. Sengelov, H., L. Kjeldsen, and N. Borregaard. 1993. Control of exocytosis in and significance of ligand-induced internalization of human neutrophil chemo- early neutrophil activation. J. Immunol. 150: 1535–1543. kine receptors CXCR1 and CXCR2. J. Biol. Chem. 279: 24372–24386. 3. Sengelov, H., P. Follin, L. Kjeldsen, K. Lollike, C. Dahlgren, and N. Borregaard. 1995. Mobilization of granules and secretory vesicles during in vivo exudation of 34. Prada-Delgado, A., E. Carrasco-Marin, C. Pena-Macarro, E. del Cerro-Vadillo, human neutrophils. J. Immunol. 154: 4157–4165. M. Fresno-Escudero, F. Leyva-Cobian, and C. Alvarez-Dominguez. 2005. Inhi- Listeria monocytogenes 4. Jog, N., M. J. Rane, G. Lominadze, G. C. Luerman, R. A. Ward, and bition of Rab5a exchange activity is a key step for sur- Traffic K. R. McLeish. 2007. The actin cytoskeleton regulates exocytosis of all neutro- vival. 6: 252–265. phil granule subsets. Am. J. Physiol. 292: C1690–C1700. 35. Vita, F., M. R. Soranzo, V. Borelli, P. Bertoncin, P. Bertoncin, and G. Zabucchi. 5. Borregaard, N., L. Kjeldsen, K. Rygaard, L. Bastholm, M. H. Nielsen, 1996. Subcellular localization of the small GTPase Rab5a in resting and stimulated human neutrophils. Exp. Cell Res. 227: 367–373.

H. Sengelov, O. W. Bjerrum, and A. H. Johnsen. 1992. Stimulus-dependent se- by guest on September 25, 2021 cretion of plasma proteins from human neutrophils. J. Clin. Invest. 90: 86–96. 36. Fabbri, M., S. D. Meglio, M. C. Gagliani, E. Consonni, R. Molteni, J. R. Bender, 6. Casey, T., J. L. Meade, and E. W. Hewitt. 2007. Organelle proteomics: identi- C. Tacchetti, and R. Pardi. 2005. Dynamic partitioning into lipid rafts controls the fication of the exocytic machinery associated with the natural killer cell secretory endo-exocytic cycle of the aL/B2 integrin, LFA-1, during leukocyte chemotaxis. lysosome. Mol. Cell. Proteomics. 6: 767–780. Mol. Biol. Cell 16: 5793–54803. 7. Brunet, S., P. Thibault, E. Gagnon, P. Kearney, J. J. M. Bergeron, and 37. Brumell, J., A. Volchuk, H. Sengelov, N. Borregaard, A.-M. Cieutat, M. Desjardins. 2003. Organelle proteomics: looking at less to see more. Trends D. F. Bainton, S. Grinstein, and A. Klip. 1995. Subcellular distribution of dock- Cell Biol. 13: 629–638. ing/fusion proteins in neutrophils, secretory cells with multiple exocytic com- 8. Garin, J., R. Diez, S. Kieffer, J. F. Dermine, S. Duclos, E. Gagnon, R. Sadoul, partments. J. Immunol. 155: 5750–5759. C. Rondeau, and M. Desjardins. 2001. The phagosome proteome: insight into 38. Logan, M., P. Lacy, S. O. Odemuyiwa, M. Steward, F. Davione, H. Kita, and phagosome functions. J. Cell Biol. 152: 165–180. R. Moqbel. 2006. A critical role for vesicle-associated membrane protein-7 in 9. Gagnon, E., S. Duclos, C. Rondeau, E. Chevet, P. H. Cameron, O. Steele-Motimer, exocytosis from human eosinophils and neutrophils. Allergy 61: 777–784. J. Palement, J. J. M. Bergeron, and M. Desjardins. 2002. Endoplasmic reticulum- 39. Mollinedo, F., B. Martin-Martin, J. Calafat, S. M. Nabokina, and P. A. Lazo. mediated phagocytosis is a mechanism of entry into macrophages. Cell 110: 2003. Role of vesicle-associated membrane protein-2, through Q-soluble N-eth- 119–131. ylmaleimide-sensitive factor attachment protein receptor/R-soluble N-ethylmale- 10. Lominadze, G., D. W. Powell, G. C. Luerman, A. J. Link, R. A. Ward, and imide-sensitive factor attachment protein receptor interaction, in the exocytosis of K. R. McLeish. 2005. Proteomic analysis of human neutrophil granules. Mol. specific and tertiary granules in human neutrophils. J. Immunol. 170: 1034–1042. Cell. Proteomics 4: 1503–1521. 40. Mollinedo, F., J. Calafat, H. Janssen, B. Martin-Martin, J. Canchado, S. M. Nabokina, 11. Fischer, F., D. Wolters, M. Roegner, and A. Poetsch. 2006. Toward the complete and C. Gajate. 2006. Combinatorial SNARE complexes modulate the secretion of membrane proteome. Mol. Cell. Proteomics 5: 444–453. cytoplasmic granules in human neutrophils. J. Immunol. 177: 2831–2841. 12. Haslett, C., L. Guthrie, J. Johnston, and P. Henson. 1985. Modulation of multiple 41. Polgar, J., S. H. Chung, and G. L. Reed. 2002. Vesicle-associated membrane neutrophil functions by preparative methods of trace concentrations of bacterial protein 3 (VAMP-3) and VAMP-8 are present in human platelets and are required lipopolysaccharide. Am. J. Pathol. 119: 101–110. for granule secretion. Blood 100: 1081–1083. 13. Amrein, P. C., and T. P. Stossel. 1980. Prevention of degradation of human poly- 42. Reales, E., F. Mora-Lopez, V. Rivas, A. Garcia-Poley, J. A. Brieva, and morphonuclear leukocyte proteins by diisopropylfluorophospate. Blood 56: 442–447. A. Campos-Caro. 2005. Identification of soluble N-ethylmaleimide-sensitive fac- 14. Dahlgren, C., S. R. Carlsson, A. Karlsson, H. Lundqvist, and C. Sjolin. 1995. The tor attachment protein receptor exocytic machinery in human plasma cells: lysosomal membrane glycoproteins lamp-1 and lamp-2 are present in mobilizable SNAP-23 is essential for antibody secretion. J. Immunol. 175: 6686–6693. organelles, but are absent from the azurophil granules of human neutrophils. 43. Lipschutz, J. H., and K. E. Mostov. 2002. Exocytosis: the many masters of the Biochem. J. 311: 667–674. exocyst. Curr. Biol. 12: R212–R214. 15. Link, A. J., J. Eng, D. M. Schieltz, E. Carmack, G. J. Mize, D. R. Morris, B. M. Garvik, and J. R. Yates. 1999. Direct analysis of protein complexes using 44. Munson, M., and P. Novick. 2006. The exocyst defrocked, a framework of rods mass spectrometry. Nat. Biotechnol. 17: 676–682. revealed. Nat. Struct. Mol. Biol. 13: 577–581. 16. http://bigcat.mc.vanderbilt.edu/BIGCAT/index.php 45. Boyd, C., T. Hughes, M. Pypaert, and P. Novick. 2004. Vesicles carry most 17. http://tools.proteomecenter.org/software.php exocyst subunits to exocytic sites marked by the remaining two subunits, Sec3p 18. McAfee, K. J., D. T. Duncan, M. Assink, and A. J. Link. 2006. Analyzing pro- and Exo70p. J. Cell Biol. 167: 889–901. teomes and protein function using graphical comparative analysis of tandem mass 46. Zhang, X., E. Bi, P. Novick, L. Du, K. G. Kozminski, J. H. Lipschutz, and spectrometry results. Mol. Cell. Proteomics 5: 1497–1513. W. Guo. 2001. Cdc42 interacts with the exocyst and regulates polarized secretion. 19. Nesvizhskii, A. I., A. Keller, E. Kolker, and R. Aebersold. 2003. A statistical J. Biol. Chem. 276: 46745–46750. model for identifying proteins by tandem mass spectrometry. Anal. Chem. 75: 47. Zhang, X. M., S. Ellis, A. Sriratana, C. A. Mitchell, and T. Rowe. 2004. Sec15 4646–4658. is an effector for the Rab11 GTPase in mammalian cells. J. Biol. Chem. 279: 20. Liu, H., R. G. Sadygov, and R. J. Yates. 2004. A model for random sampling and 43027–43034. estimation of relative protein abundance in shotgun proteomics. Anal. Chem. 76: 48. Chhabra, E. S., and H. N. Higgs. 2007. The many faces of actin: matching as- 4193–4201. sembly factors with cellular structures. Nat. Cell. Biol. 9: 1110–1121.