Deciphering the human platelet sheddome

Abstract

Activated platelets shed surface proteins, potentially modifying platelet function as well as providing a source of bioactive fragments. Previous studies have identified several constituents of the platelet sheddome, but the full extent of shedding is unknown. Here we have taken a global approach, analyzing protein fragments in the supernate of activated platelets using mass spectroscopy and looking for proteins originating from platelet membranes. After removing plasma proteins and microparticles, 1048 proteins were identified, including 69 membrane proteins. Nearly all of the membrane proteins had been detected previously, but only 10 had been shown to be shed in platelets. The remaining 59 are candidates subject to confirmation. Based on spectral counts, protein representation in the sheddome varies considerably. As proof of principle, we validated one of the less frequently detected proteins, semaphorin 7A, which had not previously been identified in platelets. Surface expression, cleavage, and shedding of semaphorin 7A were demonstrated, as was its association with α-granules. Finally, cleavage of semaphorin 7A and 12 other proteins was substantially reduced by an inhibitor of ADAM17, a known sheddase. These results define a subset of membrane proteins as sheddome candidates, forming the basis for further studies examining the impact of ectodomain shedding on platelet function.

Figure 1

Proteomic strategies to identify and validate shed proteins of the platelet sheddome. Washed human platelets were activated with 10μM PMA at 37°C for 1 hour and then analyzed using the 2 strategies shown. Details are given in “Methods” and supplemental data.

Figure 2

Data mining of the platelet sheddome. (A) Subcellular localization of the 1048 proteins identified by MS. Cellular locations were assigned using the Swiss-Prot and TrEMBL databases. (B) Spectral counts for the 69 membrane proteins that were identified. Sema7A is highlighted in green.

Figure 3

Shedding of GPIbα and GPVI during platelet activation. Platelets were activated with 10μM PMA at 37°C for the times indicated. Representative Western blots of platelet supernatant for (A) GPIbα and (B) GPVI. Summary plots from 3 experiments (mean ± SEM).

Figure 4

Sema7A is present in human platelets. (A) Western blot of human platelet and brain lysates. (B) Confocal microscopy of activated human platelets allowed spreading on a fibrinogen-coated surface showing sema7A on the surface and colocalizing with VWF in α-granules.

Figure 5

Sema7A is cleaved and shed during platelet activation. (A-B) Shed sema7A in platelet supernatants with increasing time after (A) 10μM PMA or (B) 10μM SFLLRN activation. (C-D) Surface expression of sema7A determined by flow cytometry after platelet activation with (C) PMA or (D) SFLLRN activation. Alexa Fluor-488–conjugated goat IgG was used as the isotype control. (E-F) Representative Western blots of sema7A on platelet membranes.

Figure 6

Shedding of sema7A is inhibited by metalloproteinase inhibitors. Western blots of sema7A in the supernate of platelets activated with (A) PMA or (B) SFLLRN in the presence of the metalloprotease inhibitor GM6001 or the ADAM17-selective inhibitor BMS-561392 (mean ± SEM, N = 3).

Figure 7

Comparisons with other proteomes. The Venn diagrams in the figure indicate how many of the 69 membrane proteins identified in the present study were detected in (A) a study of proteins in which the starting material was isolated in platelet membranes and (B) 3 studies of either isolated platelet α-granules or platelet releasate. In the releasate studies, platelets were incubated with an agonist for 5 minutes or less, considerably shorter than the 60-minute incubation adopted here to allow time for delayed shedding to proceed to or close to completion.