Proteomic analysis of platelet α-granules using mass spectrometry


Dawn M. Maynard, Medical Genetics Branch, NHGRI, NIH, 10 Center Drive, 10/10C103, MSC 1851, Bethesda, MD 20892-1851, USA.
Tel.: +1 301 402 6622 (direct); +1 301 496 9101 (lab): fax: 301 402 7290; e-mail:


Summary. Background: Platelets have three major types of secretory organelles: lysosomes, dense granules, and α-granules. α-Granules contain several adhesive proteins involved in hemostasis, as well as glycoproteins involved in inflammation, wound healing, and cell–matrix interactions. This article represents the first effort to define the platelet α-granule proteome using mass spectrometry (MS). Methods: We prepared a subcellular fraction enriched in intact α-granules from human platelets using sucrose gradient ultracentrifugation. α-Granule proteins were separated and identified using sodium dodecylsulfate polyacrylamide gel electrophoresis and liquid chromatography–tandem MS. Results: In the sucrose fraction enriched in α-granules, we identified 284 non-redundant proteins, 44 of which appear to be new α-granule proteins, on the basis of a literature review. Immunoelectron microscopy confirmed the presence of Scamp2, APLP2, ESAM and LAMA5 in platelet α-granules for the first time. We identified 65% of the same proteins that were detected in the platelet releasate (J. A. Coppinger et al. [Blood 2004;103: 2096–104]) as well as additional soluble and membrane proteins. Our method provides a suitable tool for analyzing the granule proteome of patients with storage pool deficiencies.


In bone marrow, megakaryocytes form under the influence of thrombopoietin. The megakaryocytic cytoplasm releases hundreds of platelets into the circulation to function in coagulation. Platelets contain three major secretory organelles: the lysosome, the dense granule, and the α-granule. Each organelle has a specific molecular composition, ultrastructural morphology, rate of exocytosis, and secretory response to different stimuli. The relatively scarce lysosomes contain acidic hydrolases. Dense granules, numbering three to eight per platelet [1], contain calcium, polyphosphates, adenosine diphosphate (ADP), adenosine triphosphate, and serotonin, and are readily detected by whole mount electron microscopy (EM) [2]. Release of dense granule contents, specifically ADP, stimulates the secondary aggregation response of platelets. α-Granules are the largest (200–500 nm) and most abundant secretory organelles in platelets, numbering approximately 80 per cell [3]. They contain soluble adhesive glycoproteins (GP) such as fibrinogen and von Willebrand factor (VWF), chemokines such as platelet factor IV (PF4) and platelet basic protein, and receptors such as αIIbβ3 integrin, P-selectin, and the glucose transporter Glut-3. Some of these proteins, active in hemostasis, also decorate the plasma membrane when the platelet is activated, reflecting the secretory process of α-granules [3,4].

Genetic deficiencies of α-granules or dense granules cause congenital bleeding disorders termed storage pool deficiencies (SPDs). For example, in Hermansky–Pudlak syndrome, platelet dense granules are absent, causing a bleeding diathesis, and melanosomes fail to mature within melanocytes [5–8], causing albinism; some patients also develop granulomatous colitis or a fatal pulmonary fibrosis [9].

The genes causing α-granule diseases are poorly defined, although some transcription factors and GPs may be involved [10–15]. In different types of gray platelet syndrome (GPS), there are deficient or absent α-granules, causing a gray appearance in Wright-stained blood films [16]. EM shows small α-granules or empty vacuoles the size of normal α-granules [16–19]. There are autosomal recessive, autosomal dominant and X-linked types of GPS, but the causative gene has been identified only for the X-linked form. Patients have relatively mild bleeding problems, but an occasional fatality has been observed, and myelofibrosis has been reported [20].

We began our investigation of the protein content of platelet organelles with the α-granule. Ultimately, we hope to determine whether individuals with SPDs, including GPS, have normal contingents of soluble and membrane-bound α-granule proteins or α-granule membrane ghosts. We performed platelet organelle fractionation on sucrose gradients followed by discovery-based proteomics using mass spectrometry (MS). We have identified known [4] and previously unknown proteins of the α-granule, providing a methodological framework and a basis for future comparisons with SPD platelets.



Formic acid (98%) was purchased from Fluka (Buchs SG, Switzerland). High-performance liquid chromatography (HPLC)-grade acetonitrile and methanol were purchased from Burdick & Jackson (Muskegon, MI, USA). Modified porcine trypsin was purchased from Promega (Madison, WI, USA). Larex UF Powder was purchased from Larex, Inc. (White Bear Lake, MN, USA). Complete, EDTA-free Protease Inhibitor Cocktail Tablets were purchased from Roche Applied Science (Indianapolis, IN, USA). All other chemicals were purchased from Sigma-Aldrich (St Louis, MO, USA). Mouse anti-VWF, goat anti-secretory carrier membrane protein 2 (Scamp2) and goat anti-amyloid-like protein 2 (APLP2) were obtained from Santa-Cruz Biotechnology Inc. (Santa Cruz, CA, USA); rabbit anti-calreticulin was obtained from ABR (Golden, CO, USA); mouse anti-prohibitin and mouse anti-P-selectin were obtained from Abcam, Inc. (Cambridge, MA, USA); mouse anti-CD63 was obtained from Hybridoma Bank (Iowa City, IA, USA). The anti-human GLUT-3 antibody was raised against a synthetic peptide corresponding to the last 12 amino acids in the C-terminus of the protein [21]. Mouse monoclonal anti-laminin α5 chain (LAMA5) directed to laminin α5 was obtained from Abnova Corporation (Taipei, Taiwan), and goat anti-endothelial cell-selective adhesion molecule (ESAM)-1 was obtained from R&D Systems (Minneapolis, MN, USA). The HPLC trapping column was a PepMap C18 (5 mm length × 300 μm internal diameter, 5-μm particle size, 100-Å pore size) purchased from LC Packings/Dionex (Sunnyvale, CA, USA). The reversed phase column was a PicoFrit BioBasic C18 column (10-cm length × 75 μm internal diameter, 15-μm tip, 5-μm particle size, 300-Å pore size) purchased from New Objective (Woburn, MA, USA).

Blood collection and platelet preparation

Whole blood was collected from normal, healthy donors at the NIH Department of Transfusion medicine under an institutional review board-approved protocol (99-CC-0168), after written informed consent had been obtained. Blood was collected into acid–citrate–dextrose and centrifuged at 800 × g for 3 min to collect platelet-rich plasma. The platelets, separated from plasma using a discontinuous arabinogalactan gradient [22], were washed with Tyrode’s I buffer (140 mm NaCl, 3 mm KCl, 47 mm NaH2PO4, 1.2 mm NaHCO3, 1 mg mL–1 glucose, adjusted to pH 6.5) supplemented with protease inhibitors. Residual red cells were lyzed with 1% ammonium oxalate.

Platelet lysis and organelle fractionation

Platelets were resuspended in 250 mm sucrose and lyzed by ultrasonication (Microson ultrasonic cell disruptor; Misonix, Farmingdale, NY, USA). Non-lyzed platelets were pelleted by centrifugation at 700 × g for 6 min. The supernatant containing platelet organelles was loaded onto six preformed, 10-mL linear sucrose gradients [23] (20–50% sucrose; 60% sucrose cushion). The gradients were centrifuged at 217 000 × g in a Beckman SW 41 Ti rotor for 16 h at 4 °C. Nine fractions were collected using careful pipetting from the top. Fractions from three gradients were combined, transferred to polycarbonate tubes, diluted in 6% sucrose, and centrifuged at 140 000 × g in a Beckman 70.1 Ti rotor for 1 h at 4 °C. The pellets were resuspended in 50 μL of Tyrode’s I buffer (pH 6.5) and frozen at −20 °C. Protein concentration was determined using the BioRad Protein Assay with bovine serum albumin (BSA) as standard. Fractions were also pelleted and fixed for conventional EM and immunoelectron microscopy (IEM).

Light microscopy

Platelet samples were smeared and dried on glass LM slides and stained with Wright–Giemsa stain. Images were acquired with a 100×/1.4 Plan-APOCHROMAT oil objective using a Zeiss axiovert 200M Inverted light microscope (Carl Ziess, Thornwood, NY, USA), equipped with Zeiss axiovision 4.4 software.

Electron microscopy

Sucrose gradient fractions (∼100 μL) were fixed in ∼100 μL of 0.1% glutaraldehyde in White’s Saline, and processed for conventional EM [24]. For IEM, the fractions were fixed in a mixture of 2% paraformaldehyde and 0.2% glutaraldehyde in 0.1 m sodium phosphate buffer (pH 7.4). After washing, samples were frozen, cryosectioned and immunogold labeled according to established procedures [25]. Sections were viewed in a Philips (Mahwah, NJ, USA) 301 and JEOL 1200CX electron microscope (JEOL, Tokyo, Japan). Resting platelets were obtained from freshly fixed whole blood that was collected directly from the vein into fixative containing 2.4% paraformaldehyde and 0.24% glutaraldehyde in 0.1 m sodium phosphate buffer (pH 7.4) (1:5 v/v).

Protein separation and western blotting

Proteins were reduced with dithiothreitol (DTT) (10× NuPage Sample Reducing Agent, Invitrogen, Carlsbad, CA, USA), solubilized in 2× sodium dodecylsulphate (SDS) Protein Gel Solution (Quality Biological, Inc., Gaithersburg, MD, USA) and heated at 70 °C for 3 min in a water bath. Samples were loaded on a 4–12% Novex Tris–glycine gel (Invitrogen, Carlsbad, CA, USA) with NuPage antioxidant in the inner chamber and separated at 125V for 1.5 h. Gels were stained with Coomassie Brilliant Blue R-250 Staining Solution (Bio-Rad, Hercules, CA, USA) and destained with Coomassie Brilliant Blue R-250 Destaining Solution. For western blotting, proteins were blotted onto Protran BA83, 0.2-μm nitrocellulose membranes (Whatman, Sanford, ME, USA) using 1× Tris–glycine–methanol buffer (Bio-Rad Hercules, CA, USA) for 2 h at 25V at 4 °C. Primary antibodies and secondary antibodies conjugated to horseradish peroxidase were diluted in phosphate-buffered saline (PBS) + 0.05% Tween-20 (PBS-T), supplemented with 0.5% BSA. Membranes were blocked with PBS-T, supplemented with 5% non-fat milk, washed, and incubated overnight at 4 °C or for 2 h at room temperature. Proteins were detected using the enhanced chemiluminescence Western blot detection kit (GE Healthcare, Piscataway, NJ, USA).

In-gel digestion and peptide extraction

Gel bands were processed as follows. Bands were excised, destained in 50:50 methanol/0.1 m ammonium bicarbonate (NH4HCO3), washed in 0.1 m NH4HCO3, and washed in 50:50 acetonitrile/0.1 m NH4HCO3. Bands were shrunk in acetonitrile and dried in an Eppendorf Vacufuge (Eppendorf North America, Westbury, NY, USA). Proteins were reduced with 10 mm DTT in 25 mm NH4HCO3 for 1 h at 56 °C, and then alkylated with 55 mm iodoacetamide in 25 mm NH4HCO3 for 45 min in the dark at room temperature. Bands were washed with 25 NH4HCO3 and dehydrated with acetonitrile. Gel pieces were reswelled with 10 ng μL–1 trypsin in 25 mm NH4HCO3. Digestions were performed overnight in a 37 °C water bath. Trypsin digestion cocktail was transferred from each tube into a corresponding labeled tube, and sequential peptide extractions were performed on each band using solutions of 5%, 50% and 80% acetonitrile in water, supplemented with 0.1% formic acid. Samples were dried and resuspended in HPLC mobile phase A for liquid chromatography (LC)-MS/MS analysis.

LC-MS/MS analysis

The fully automated HPLC-MS system consisted of the Ultimate Nano-HPLC system (LCPackings/Dionex, Sunnyvale, CA, USA) with Famos microautosampler, Switchos micro switching module, and Ultimate pump module connected online to a ThermoElectron LTQ NSI-Ion Trap mass spectrometer (San Jose, CA, USA). The reverse phase separation was conducted using mobile phase A (2:97.9:0.1 acetonitrile/water/formic acid) and mobile phase B (80:19.9:0.1 acetonitrile/water/formic acid). Samples were dissolved in 6.3 μL of mobile phase A, injected using the Famos autosampler, and trapped on a PepMap trapping column. Peptides were desalted with mobile phase A, and then eluted and separated on a BioBasic C18 column at approximately 400 nL min–1 using the following gradient program: 5% B (5 min), a linear gradient of 5–43% B (50 min), 43–95% B (5 min), and 95% B (15 min). The LTQ was operated in positive ion mode with dynamic exclusion set to repeat count = 1, repeat duration = 30, exclusion duration = 60, exclusion mass width low = 0.5 amu, exclusion mass width high = 1.0 amu, and expiration count = 2. Spectra were acquired in a data-dependent manner with the top five most intense ions in the MS scan selected for MS/MS. A threshold of 3000 was required to trigger MS/MS.

Data searching

Raw MS/MS files were submitted to the NIH mascot [26] Cluster using mascot daemon. Data were searched against the uniprot-sprot + uniprot-trembl database (updated on 17 December 2005), using a restricted taxonomy of Homo sapiens (human), enzymatic cleavage = trypsin, fixed modification = carbamidomethyl (C), variable modification = oxidation (M), monoisotopic mass, peptide tolerance = 1.5 Da, MS/MS tolerance = 0.8 Da, one missed cleavage, charge state = 1+, 2+, and 3+, instrument = ESI-TRAP.

mascot output analysis using dbparser and mass sieve

dbparser [27] version 3.0, a perl program that retrieves output from mascot flat files, parses it into a mysql relational database, and uses it to generate html output reports, was used for data analysis. mass sieve (D. Slotta et al., unpubl. data; NIMH, NIH, Bethesda, MD, USA) was used to calculate percentage coverage for each protein identification.

Peptide-level stringency  For each peptide, mascot reports a probability-based ion score, which is defined as – 10*log10(P), where P is the absolute probability that the observed match between the experimental data and the database sequence is a random event. The significance threshold for inclusion of each peptide in the dbparser output file is the individual ion score meeting or exceeding its mascot identity score threshold (< 0.05) [28]. Peptides with ion scores below their identity scores were rejected.

Protein-level stringency  Proteins were collated by dbparser’s parsimony analysis [27], which generates a concise record of the protein sequence database records that account for all of the observed peptides. Putative protein identifications required at least two peptides from at least two out of three gel experiments.


Separation of platelet organelles

Platelet preparations were greater than 95% pure by light microscopy (Supplementary Fig. S1A), and the platelet ultrastructure showed little signs of activation by EM (Supplementary Fig. S1B). Sucrose gradient sedimentation revealed several distinct bands throughout the gradient (Fig. 1A), which was divided into fractions (1–9) for pelleting by ultracentrifugation. Portions of fractions 2, 4 and 6 were examined to determine which fraction was enriched in α-granules. The remainder was frozen for further investigations.

Figure 1.

 (A) Picture of the sucrose gradient following ultracentrifugation. Fractions were collected by pipetting from the top. Fraction 6 was saved and prepared for transmission electron microscopy, immuno-electron microscopy and subsequent mass spectrometric analysis. (B) Transmission electron micrograph of fraction 6 from the sucrose gradient. Fraction 6 consisted predominantly of alpha granules, i.e., vesicles with a gray matrix and dark nucleoid. Approximately 10% of the vesicles had a much darker matrix. Scope magnification × 33 000. (C–E) Immuno-electron micrographs of fraction 6 using three different antibody labels. C: Anti-Glut-3; D: Anti-CD63; E: Anti-GPIb. Glut-3 antibodies identified the majority of fraction 6 organelles as alpha granules [30] (Fig. 1C), whereas immunolabeling with anti-CD63 occasionally decorated the lumenal internal membranes of alpha granules (Fig. 1D). Occasional multivesicular body-like membranes and putative dense granules were encountered in this fraction (data not shown). GPIb showed occasional labeling [42] in fraction 6 (Fig. 1E). a, alpha granule; m, mitochondria. Size bar, 200 nm (all figures).

EM of sucrose fractions 2, 4 and 6

Transmission EM revealed that fraction 6 contained a population of membranes that consisted predominantly of α-granules, i.e. vesicles with a gray matrix and a dark nucleoid (Fig. 1B). Fraction 2 was mainly composed of membrane sheets and irregular vesicles, whereas fraction 4 contained a heterogeneous population of mostly smaller membranes (Supplementary Fig. S2). IEM analysis using the α-granule marker Glut-3 confirmed that fraction 6 was highly enriched in intact α-granules (Fig. 1C) [29]. Immunolabeling with anti-CD63 occasionally decorated luminal membranes of the α-granules (Fig. 1D). Occasional multivesicular bodies and putative dense granules were encountered in this fraction (data not shown). GPIb was predominantly associated with the membrane sheets in fraction 2 (Supplementary Fig. S2D), representing membranes of the cell surface and open canalicular system (OCS). Occasional labeling of GPIb was found in fraction 6 (Fig. 1E).

Western blot analysis

Western blotting revealed the soluble α-granule protein VWF in fractions 5–7 (Fig. 2A). The membrane-bound protein P-selectin was found primarily in fraction 6, but was observed faintly in fractions 2, 5 and 7 (Fig. 2B). Prohibitin, a mitochondrial marker, was found in four fractions but mainly in fractions 4 and 5 (Fig. 2C). Calreticulin, an established endoplasmic reticulum (ER) marker, was found in most fractions but mainly in fractions 3 and 4 (Fig. 2D). CD63, a marker for lysosomes and dense granules, resides mostly in fractions 2 and 3, and to some extent in fractions 4 and 6 (Fig. 2E). As fraction 6 was enriched in VWF and P-selectin, it was used for further investigations.

Figure 2.

 Western blots of sucrose fractions. Numbers above each column indicate the sucrose gradient fraction applied to the gel. (A) Antibody to vWF, an alpha granule marker (>207 kDa). (B) Antibody to P-selectin (CD62P), an alpha granule and dense granule marker (∼140 kDa). (C) Antibody to prohibitin, a mitochondrial marker (∼30 kDa). (D) Antibody to calreticulin, an ER marker (∼60 kDa). (E) Antibody to CD63, a lysosome and dense granule marker (∼50 kDa). Fraction 6 contained the highest expression of the soluble protein vWF and the membrane protein P-selectin, indicating enrichment of alpha granules, therefore, fraction 6 was used for subsequent mass spectrometric analysis.

Identification of α-granule proteins by MS

One-dimensional SDS–polyacrylamide gel electrophoresis gels (Fig. 3) were used to simplify the complex protein mixture of fraction 6 prior to LC-MS/MS analysis. The MS data acquired from three gel experiments using normal blood reinforced the EM and western blotting data showing that fraction 6 is enriched in α-granules.

Figure 3.

 SDS-PAGE gel of fraction 6 from the sucrose gradient. This gel represents 1 of 3 gel experiments performed. Each gel was loaded with 40 μg total protein and subsequently stained with Coomassie Brilliant Blue R-250 Staining Solution. The gels were sliced from top to bottom, then proteins were reduced, alkylated, and digested with trypsin. Resulting peptides were detected by LC-MS/MS.

Supplementary Table S1 lists proteins identified using strict criteria (see Methods). Obvious mitochondrial proteins (65) were deleted on the basis of gene ontology (GO) subcellular location information. In total, 219 putative protein identifications were made, 36 of which were reviewed by Rendu et al. [4] as being α-granule (and sometimes plasma membrane) proteins. Of the 219 non-mitochondrial proteins that we detected, 50 were also identified in the platelet releasate [30], indicating an approximate 65% overlap with the 81 releasate proteins.

We report 44 potentially new α-granule proteins in Table 1, derived from supplementary Table S1 and based on a critical literature review of the proteins. Twenty-six of the 44 proteins were also found in fraction 2 (plasma membrane fraction) by MS; of these, 17 are either cytoskeletal or membrane-bound. We found no overlap between the 44 new α-granule proteins and those proteins found in the platelet releasate.

Table 1.   New putative α-granule proteins identified from sucrose gradient fraction 6 by LC-MS/MS. This list contains 44 proteins identified as potentially new α-granule proteins by inference from primary literature. Proteins were selected from the supplementary Table S1 list and were identified by at least two peptides from at least two out of three gel experiments; all peptides used for protein identifications had mascot ion scores that met or exceeded their individual identity scores. Whether these proteins were also found in sucrose gradient fraction 2 (membrane fraction, data not shown) is indicated. We also provide literature references and evidence for inclusion of proteins in this list
Protein numberProtein nameAccession numberGene ontology subcellular locationAlso found in fraction 2 (membrane fraction?)Evidence
  1. ER, endoplasmic reticulum; APP, amyloid precursor protein; APLP, amyloid-like protein; PS1, presenilin 1; OCS, open canalicular system; NAP1, Nck-associated protein 1; MCI, mild cognitive impairment.

13-Hydroxyacyl-CoA dehydrogenase type 2Q99714Cytoplasm; mitochondria; ER; plasma membraneNoInteracts with amyloid-β; may contribute to the neuronal dysfunction associated with Alzheimer disease
2Actin-like protein 3P61158CytoskeletonYesMajor constituent of the ARP2/3 complex
3Actin-related protein 2/3 complex subunit 1B (p41)O15143Cytoskeleton; actin polymerizationYesMay function as p41 subunit of the human ARP2/3 complex
4Actin-related protein 2/3 complex subunit 4P59998Cytoskeleton; actin polymerizationYesExact function unknown
5Alcadein α1 (calsyntenin 1)Q5UE58MembraneNoThought to interact with APP (A4_human) and APLP1 and APLP2 among others
6Amyloid-like protein 2Q06481MembraneNoAs a substrate for PS1, it may be in α-granules
7Angiopoietin-1 precursorQ15389SecretedNoThrombin induces the release of angiopoietin-1 from platelets
8Antithrombin III precursorP01008SecretedNoAntithrombin III antigen in human platelets
9β2-Glycoprotein I (GPI)P02749Plasma/secretedNoCan activate platelets in complex with anti-β2-GPI antibodies in a dysregulated manner via GPIb-IX-V; detected in two mass spectrometry papers on platelets
10β-ParvinQ9HBI1Cytoplasmic side of focal adhesionYesInteracts with integrin-linked kinase; cell–substrate interaction through integrins; forms a complex with ss-affixin in platelets
11Coiled-coil domain containing 109AQ8NE86Not reportedNoUnknown function
12Connective tissue growth factorP29279SecretedNoRelease from the storage granules. Suspected to be endocytosed
13Endothelial cell-selective adhesion molecule precursorQ96AP7Membrane; cell junctionYes (1)Strongly expressed by megakaryocytes and platelets. Upregulated on surface of platelets by thrombin activation; probably stored in granules
14Endothelin-converting enzyme 1Q9UPM4MembraneYes (1)Possible role in atherosclerosis, including platelet aggregation
15Erythrocyte band 7 integral membrane proteinP27105Membrane; melanosome; platelet α-granulesYesMajor lipid-raft component of platelet α-granules
16F-actin capping protein subunit α1P52907Cytoskeleton (F-actin capping)Yes (1)Member of the F-actin capping protein α-subunit family. A calcium-insensitive capping protein in resting and activated platelets
17Fibrocystin LQ86WI1MembraneNoA homolog of the autosomal recessive polycystic kidney disease gene; unlikely that it is ARPKD-associated
18G6fO95869MembraneYes (1)A novel platelet Grb2-binding membrane adapter
19Garpin; leucine-rich repeat-containing protein 32Q14392Plasma membraneNoHas structural similarities to the human GPIbα and GPV platelet proteins. Detected in platelets by mass spectrometry
20Guanine nucleotide-binding protein α13 subunitQ14344Membrane; melanosomeYesLack of G-α-13 severely reduced the ability of thrombin, thromboxane A2 and collagen to induce platelet shape changes and aggregation in vitro
21Guanine nucleotide-binding protein G(i), α1 subunitP63096Not reportedYesInvolved in platelet activation
22Inter-α-trypsin inhibitor heavy chain H4Q14624ExtracellularNoMay be involved in acute-phase reactions
23Intercellular adhesion molecule 2 precursorP13598MembraneYesPresent on plasma membrane and OCS of resting platelets
24Junctional adhesion molecule AQ9Y624Membrane; cell junction; plateletYesExpressed in megakaryocytes, platelets
25KIAA0068 proteinQ14467CytoplasmYesSRA1 and NAP1 are essential components of a WAVE2- and ABI1-containing complex linking RAC1 to site-directed actin assembly
26Laminin α5 chain precursorQ8TDF8ExtracellularNoMegakaryocytic cells synthesize and platelets secrete α5-laminins, and the endothelial laminin isoform laminin 10 (α5β1γ1) strongly promotes adhesion but not activation of platelets
27Latent transforming growth factor-β-binding protein 1 isoform LTBP-1S variantQ59HF7ExtracellularYes (1)May play critical roles in controlling and directing the activity of tgfb1. May have a structural role in the extracellular matrix. Found in platelets
28MoesinP26038CytoplasmYesAppears to function as crosslinkers between plasma membranes and actin-based cytoskeletons. Localized to filopodia. Interacts with PECAM-1 (CD31).
29Nck-associated protein 1Q9Y2A7Membrane; lamellipodiumNoPossible relation to Alzheimer’s disease (AD). Part of lamellipodial complex that controls rac-dependent acting remodeling
30OTTHUMP00000016748. Activated T-cell marker CD109Q5SYA8MembraneNoFunction in platelets is largely unknown
31PKM2 proteinQ504U3Not reportedYesPlays a significant role in the development of AD from MCI.
32Platelet-type lipoxygenase 12P18054CytoplasmYesInvolved in vascular disease and inflammation. Increased in AD: possible involvement in brain oxidative stress
33Protein NipSnap3AQ9UFN0CytoplasmNoNIPSNAP3A belongs to a family of proteins with putative roles in vesicular transport
34RAB11B, member of RAS oncogene familyQ5U0I1MembraneNoFunctions in sorting endosomes
35Ras-related C3 botulinum toxin substrate 1. p21-Rac1Q3Y4D3Membrane; melanosomeYesEssential for platelet lamellipodia formation and aggregate stability under flow
36Ras-related protein Rab-27BO00194MembraneYesLocalizes to α-granules and dense granules in megakaryocytes. May participate in platelet synthesis and coordinate proplatelet formation with granule transport. Griscelli syndrome
37Ras-related protein Rap-1b precursorP61224Cytoplasm; membrane; plateletYesRegulates crosstalk between platelet integrin α2β1 and integrin αIIbβ3
38Scamp2 proteinQ6IBK3MembraneYes (1)Partially colocalizes with SNAP-23 and syntaxin 4. Localized to intracellular organelles in mast cells
39SERPINE2 proteinQ5D0C4SecretedNoIntracellular trafficking in mammalian cells. Inhibitor of thrombin and urokinase. Platelet-bound form
40Small inducible cytokine A5P13501Secreted/also on alpha granules in literatureNoPlatelet chemokines and chemokine receptors: linking hemostasis, inflammation, and host defense
41T-complex protein 1 subunit etaQ99832CytoplasmYesMolecular chaperone; assist the folding of proteins upon adenosine triphosphate hydrolysis. Neurodegenerative disease
42TransketolaseP29401CytosolYesMutations have clinical implications for neurode generative diseases, diabetes and cancer. Found in platelets
43Trem-like transcript-1Q86YW5Membrane; plateletsYes (1)Abundant platelet-specific, type I transmembrane receptor. A soluble fragment found in the α-granules of resting platelets, and is translocated to the platelet surface following activation by thrombin
44Unc-13 homolog DQ70J99Peripheral membraneYesMUNC13-4 bound GTP-RAB27A and GTP-RAB27B in vitro and enhanced platelet secretion. Regulates the dense core granule secretion

IIEM of novel α-granule proteins

On the basis of our MS findings and a literature review, we assessed the subcellular location of four potential α-granule proteins using IEM (Fig. 4). Fixed cryosections were prepared from isolated resting platelets. Scamp2 was found in α-granules mainly associated with the membrane, with occasional labeling on the platelet plasma membrane (Fig. 4A). LAMA5 was predominantly localized to the intraluminal region of the platelet α-granules (Fig. 4B), whereas APLP2 was found both on membrane and intraluminal regions of the α-granules (Fig. 4C). ESAM-1 was predominantly localized to the α-granule membranes, as well as being detected on the cell surface (data not shown).

Figure 4.

 Frozen thin-sections of resting platelets immediately fixed in whole blood. Immunolabeling with primary antibodies, followed by protein A conjugated to 10 nm gold. Shown is the subcellular localization of 3 selected proteins as indicated on the figures. Scamp2 was found predominantly in alpha granules with occasional labeling on the platelet plasma membrane (A). LAMA5 was predominantly localized to the intracellular region of platelet alpha granules (B). APLP2 was found both associated with the membrane and in the intra-lumenal region of the alpha granules (C). All proteins were originally identified by mass spectrometry in the alpha granule-enriched sucrose fraction. α, alpha granules. Scale bars, 200 nm.


Proteomic analysis of platelets [31–33] thus far has been performed on non-stimulated [34–37] and activated platelets [38], the platelet releasate [30,39], platelet membranes [40], and platelet-derived microparticles [41]. Here we investigated the proteome of individual isolated α-granules, including both soluble and membrane-bound proteins.

After sucrose gradient ultracentrifugation, fraction 6 was enriched in intact α-granules, as demonstrated by western blot and IEM analysis. The presence of P-selectin in fraction 6 supported this finding, but fraction 2 also showed low expression of P-selectin, along with the dense granule marker CD63 [1]. The dense granules may have lost their contents and their density during fractionation. VWF also marked fraction 6 as being enriched in α-granules, and its abundance may have caused spillover into fractions 5 and 7. These results were confirmed using IEM, where fraction 6 contained primarily α-granules labeled with Glut-3. Fraction 2 contained high expression levels of GPIb, representing cell surface and OCS membranes, whereas fraction 4 contained small membrane structures possibly corresponding to membranes of the dense tubular system or other intracellular structures. Our finding of a small portion of GPIb decorating α-granules is consistent with previous reports by Berger et al. [42] and is supported by our MS results. On the basis of these results, we used fraction 6 for MS analysis.

To identify a protein in fraction 6, we required at least two of its peptides to be detected from at least two out of three gel experiments. Among the 284 fraction 6 proteins identified in this manner were peptides from 36 known α-granule proteins [4], including VWF, thrombospondin, FV, multimerin, and GPIIbIIIa. Several reported α-granule proteins, such as protease inhibitors, coagulation FVII and FIX, and cellular mitogens, were not detected in fraction 6, perhaps because they are present on α-granules only transiently, or they are heavily glycosylated or otherwise modified, or they are not readily digested or extracted from the gel, or they are present in low abundance.

We generated a list of 44 potentially new α-granule proteins, including several involved in Alzheimer’s disease (AD), such as 3-hydroxyacyl-CoA dehydrogenase type 2, alcadein α-1 (calsyntenin 1), amyloid-like protein 2, C3 g fragment, clusterin, Nck-associated protein 1 (NAP 1), platelet-type lipoxygenase 12, PKM2 protein, and transketolase. Of these, only C3 g and clusterin (secreted proteins) were found in the platelet releasate. All others are membrane-bound proteins, reflecting the advantage of our methodology. The use of platelets as a diagnostic tool for differentiating AD from other forms of dementia has been reviewed by Cattabeni et al. [43], who proposed the use of a platelet amyloid precursor protein (APP) ratio. It follows that other proteins identified in platelet α-granules by MS may merit further study in AD. In addition, our α-granule-enriched fraction may be useful in immunoprecipitation studies to find AD protein interacting partners or lower-abundance proteins. Other interesting proteins in the α-granule fraction include those involved in vesicle transport (Rab11B, Rab27B, and NipSnap3A), cell shape and motility (F-actin capping protein, guanine nucleotide-binding protein α13 subunit) and atherosclerosis (endothelin-converting enzyme 1) (see also Supplementary Table S1).

Several proteins identified in fraction 6 by MS were localized to the α-granules by IEM for the first time. Scamp2, a conserved, four-transmembrane-spanning protein associated with recycling vesicular carriers, was recently suggested to function in membrane fusion. To our knowledge, Scamp2 has not been shown in platelet α-granules. Rather, Guo et al. [44] demonstrated that Scamp2 may function as a late-acting inhibitor of exocytosis in mast cells; it is also present in the plasma membrane, colocalizing with SNAP-23 and syntaxin 4. Therefore, Scamp2 may function in the exocytosis of platelet α-granules.

Laminin α-5 chain (also termed LAMA5, LNα5, laminin 10/α5β1γ1) is a ∼350-kDa protein that promotes the adhesion of vascular endothelial cells, hematopoietic stem cells, lymphocytes, and neutrophils. Mice lacking LNα5 die in late embryogenesis [45]. Nigatu et al. [45] found that LAMA5 is secreted by activated platelets and participates in adhesion via integrin α6β1. They suggest that LAMA5 is in α-granules and may provide a ‘provisional’ extracellular matrix upon secretion in order to assist in wound healing after injury. Our study confirms that LAMA5 is present in platelet α-granules.

Proteolysis of APP by presenilins leads to accumulation of βA4, the major component of extracellular amyloid plaques in AD. Presenilins are also essential for proteolytic processing of other membrane proteins, including Notch, TrkB, and APLP2 [46]. With the use of IEM, APP was found in α-granules [47], and presenilin 1 (PS1) colocalized with Glut-3 in the membranes of α-granules [46]. Being a substrate for PS1, it follows that APLP2 is also localized in α-granules. This is precisely what we found by IEM. Walsh et al. [48] suggested that upregulation of APLP2 might provide a therapeutic intervention for AD, and platelets may offer a non-invasive model with which to study APLP2.

ESAM is related to the junctional adhesion molecules and contains V-type and C2-type IgG domains, a hydrophobic signal sequence, a single transmembrane region, and a cytoplasmic domain [49]. Nasdala et al. [50] reported that ESAM is strongly expressed on megakaryocytes and activated platelets, that it is associated with endothelial tight junctions in blood capillaries of mouse brain and muscle, and that it supports homotypic adhesion when transfected in CHO cells. The authors propose that this molecule may participate in platelet aggregation. In support of these findings, we also found that ESAM is associated with the platelet α-granule membrane (data not shown).

Another measure of the α-granule content of fraction 6 is its protein composition as compared with the 81 proteins of the platelet releasate, proteins released from platelets upon activation with α-thrombin, identified by Coppinger et al. [30]. We found 65% of the same proteins as in the releasate. In addition, 19 of our 44 potential α-granule proteins are membrane-bound proteins that were not identified in the releasate proteome. Dense granules, lysosomes, multivesicular bodies and cell surface ectodomain shedding may have contributed to the releasate proteome but not to our fraction 6 [25,30].

Among the 284 proteins identified in fraction 6 by MS are 65 known mitochondrial proteins, consistent with results of western blotting (Fig. 2) and IEM (Fig. 1E). This is also consistent with previous reports of mitochondria found in multiple sucrose gradient fractions [23]. Mitochondrial proteins identified in fraction 6 are well documented in the literature and were removed from the final list (Supplementary Table S1). Twelve other proteins were reported as ER proteins by SwissProt and were also found in human melanosomes by MS [51]. The ER marker calreticulin was found in all fractions in our western blot, but to a lesser extent in fraction 6, which may be contaminated with relatively smaller dense tubular system membranes derived from the megakaryocyte smooth ER. However, comparison of IEM images of fractions 2 and 4 with those of fraction 6 showed distinct differences in their overall membrane structures.

This work establishes a methodology for the separation and proteomic analysis of intact platelet organelles. It also identifies a variety of proteins within the α-granule fraction that were not previously recognized to reside within platelets or α-granules. Finally, our investigations lay the groundwork for the analysis of platelet granule proteins in SPD syndromes and related disorders.


The authors acknowledge P. Merryman of the Hematology Service of the Department of Laboratory Medicine, NIH Clinical Center for helpful suggestions and advice on platelet isolation. The authors are grateful to W. Westbroek for help in establishing the platelet IEM collaboration, H. Dorward for assistance in obtaining light microscope images, M. D. Krumwiede for help with transmission EM preparations, and S. van Dijk for all the IEM work. Finally, the authors thank M. Gunay-Aygun for her expert advice and many helpful suggestions.

Disclosure of Conflict of Interests

This work was supported by NHGRI, IRB support 99-CC-0168.