Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2017 Feb 20;8:14151.
doi: 10.1038/ncomms14151.

Protein Disulfide Isomerase Secretion Following Vascular Injury Initiates a Regulatory Pathway for Thrombus Formation

Free PMC article

Protein Disulfide Isomerase Secretion Following Vascular Injury Initiates a Regulatory Pathway for Thrombus Formation

Sheryl R Bowley et al. Nat Commun. .
Free PMC article


Protein disulfide isomerase (PDI), secreted by platelets and endothelial cells on vascular injury, is required for thrombus formation. Using PDI variants that form mixed disulfide complexes with their substrates, we identify by kinetic trapping multiple substrate proteins, including vitronectin. Plasma vitronectin does not bind to αvβ3 or αIIbβ3 integrins on endothelial cells and platelets. The released PDI reduces disulfide bonds on plasma vitronectin, enabling vitronectin to bind to αVβ3 and αIIbβ3. In vivo studies of thrombus generation in mice demonstrate that vitronectin rapidly accumulates on the endothelium and the platelet thrombus following injury. This process requires PDI activity and promotes platelet accumulation and fibrin generation. We hypothesize that under physiologic conditions in the absence of secreted PDI, thrombus formation is suppressed and maintains a quiescent, patent vasculature. The release of PDI during vascular injury may serve as a regulatory switch that allows activation of proteins, among them vitronectin, critical for thrombus formation.

Conflict of interest statement

The authors declare no competing financial interests.


Figure 1
Figure 1. PDI and its variants for mechanism-based kinetic trapping of substrates.
(a) Domains a and a′ (light grey) contain the catalytic motif (CGHC) whereas the b and b′ domains (dark grey) have no catalytic activity. ‘x' connects the b′ and a′ domains and there is a C-terminal extension c. CCCC, CACC, CCCA, CACA and AAAA indicate the amino acid residue in PDI at residues 53, 56, 397 and 400. FLAG is linked to the N-terminus and SBP (streptavidin binding peptide) to the C-terminus. (b) The reductase activities of PDI and the PDI variants were monitored by insulin reduction. PDI–CACC (inverted open triangle) and PDI–CCCA (open triangle) have impaired reductase activity compared with PDI–CCCC (black square) while PDI–CACA (open diamond) and PDI–AAAA (black circle) do not express reductase activity. The error bars represent standard deviation from the mean of 3 replicate experiments done in duplicate. (c) Kinetic trapping of PDI substrates was performed in platelet-rich plasma. PDI and PDI-associated proteins were isolated by streptavidin affinity chromatography. Samples were run under non-reducing conditions on 3–8 or 7% SDS–PAGE and proteins visualised by silver staining (Left) and by Western blot (Right) for PDI and PDI-substrate complexes (green), as blotted with fluorescently labelled anti-FLAG antibody. MW, molecular weight × 10−3. Upper panel, nonreduced; lower panel, reduced.
Figure 2
Figure 2. Western blot of trapped PDI substrates.
(a) 2D electrophoresis of PDI and variants complexed to substrates from platelet-rich plasma. PDI complexes were isolated and analysed by 2D SDS–PAGE and Western blotting. Samples were run under nonreducing conditions, the gel treated with DTT, newly formed thiols alkylated with NEM, and the proteins in the gel strip run in the second dimension. Left: silver stain. Right: Western blot: PDI (green) using anti-FLAG antibody and vitronectin (red) using anti-vitronectin antibody. (b) Kinetic trapping with PDI and PDI variants in platelet-rich plasma. Samples were analysed by SDS–PAGE, blotted for vitronectin (red) and PDI (green) and detected by immunofluorescence as in A. (c) Kinetic trapping experiments with ERp57 and ERp57 variants were performed in parallel in platelet-rich plasma. ERp57 (green) and vitronectin (red) were detected by immunofluorescence as in A. MW, molecular weight × 10−3.
Figure 3
Figure 3. PDI reductase activity cleaves two disulfide bounds on vitronectin.
PDI-CCCC, ERp57-CCCC or PDI–AAAA was added to platelet-poor plasma and disulfide reduction in vitronectin monitored by alkylation of new sulhydryls with 5,000 MW PEG-maleimide. Proteins were separated by SDS–PAGE on 3–8% gels and vitronectin probed by Western blotting. (a) Time dependence, from 0 to 5 min at 1 μM enzyme; (b) Dose dependence, 0.25 to 2.5 μM for 5 min. As above except alkylation of new sulhydryls was performed with 2,000 MW PEG-maleimide. (c) Time dependence, from 20 to 120 s at 1 μM enzyme; (d) Dose dependence, 0.25 to 2.5 μM for 5 min.
Figure 4
Figure 4. PDI cleaves Cys137-Cys161 and Cys 274–453 disulfide bonds in the hemopexin-like domains of vitronectin.
(a) The mixed disulfide complex between PDI and vitronectin followed by thiol alkylation with NEM and IAM identified PDI-cleaved vitronectin disulfide bonds by mass spectrometry. (b) The tryptic sequence containing 158–167 was observed as a singly-charged species (left). The mass difference between b3 and b4 ions shows that Cys 161 is alkylated with NEM. A singly-charged peptide containing 445–453 shows a mass increase between b8 and b9 ions consistent with NEM-alkylated Cys 453 (right). (c) Cys 137 is alkylated with IAM from the doubly charged peptide 120–149 (left). The carbamidomethyl cysteine is evident from the mass difference of b17 and b18 ions. The mass difference between b15 and b16 ions from the doubly-charged peptide containing 258–291 is consistent with carbamidomethylation at Cys 274 (right).
Figure 5
Figure 5. PDI enables binding to β3 integrins via the RGD motif.
(a) Platelet-free plasma was treated with PDI-CCCC, PDI-AAAA, ERp57-CCCC or BSA, and the reaction terminated with NEM. The treated plasma was added to immobilized αVβ3 or αIIbβ3. After washing, bound vitronectin was quantitated with anti-vitronectin. PDI-CCCC, black; PDI-AAAA, dark grey; ERp57-CCCC, light grey; BSA, white. Left, αVβ3; Right, αIIbβ3. The error bars represent standard deviation from the mean of 2 replicate experiments done in triplicate. (b) Vitronectin from plasma treated with PDI-CCCC binds to activated HUVECs in the presence of GRGESP whereas vitronectin from plasma treated with PDI-AAAA showed minimal binding. Binding of vitronectin from plasma treated with PDI-CCCC to activated cells pre-treated with GRGESP,GRGDSP, or αVβ3-function blocking LM609 antibody. Scale bar (white): 15 μm. Lane 1, DAPI; Lane 2, vitronectin; Lane 3, merge.
Figure 6
Figure 6. Vitronectin accumulates at the injury site during thrombus formation in vivo.
Images of vitronectin (green) and platelets (red) were monitored by intravital microscopy following laser-induced vessel wall injury using platelet-specific anti-CD42b antibody conjugated to Dylight 649 (0.1 μg g−1 body weight) and rat anti-mouse vitronectin monoclonal antibody conjugated to Dylight 488 (0.5 μg g−1 body weight) infused into a mouse 5–10 min before injury. (a) Images depicting platelets (red) and vitronectin (green) were monitored following injury. (b) Images depicting platelets (red) and an irrelevant isotype-matched antibody (green) were monitored following injury. (c) Eptifibatide (10 μg g−1 mouse) was infused, and reinfused every 20 min. Images depicting platelets (red) and vitronectin (green) were monitored following injury. (d) Kinetics of vitronectin binding to the endothelium and the developing thrombus. Median integrated fluorescence from at least 28 thrombi in each group is shown in the presence of (a) anti-vitronectin and anti-platelet antibodies; (b) control irrelevant antibody of the same isotype as anti-vitronectin and anti-platelet antibodies; (c) anti-vitronectin antibody and anti-platelet antibodies following the infusion of eptifibatide. (e) As in D the kinetics of platelet accumulation is shown. Scale bar (cyan): 20 μm. (f) Confocal intravital reconstructions of a thrombus following vessel wall injury: The kinetics of the appearance of vitronectin and platelets within the context of the vessel wall are shown before vessel wall injury (Time 0), at 10 s following vessel wall injury (Time 10 s) and at 60 s following vessel wall injury (Time 60 s). In the upper panel, rat anti-mouse vitronectin monoclonal antibody, anti-CD42b antibody conjugated to Dylight 649 (0.1 μg g−1 body weight), and anti-CD31 antibody conjugated to Alexa 561 (1 μg g−1 body weight) were infused into a mouse 5 min before injury. These antibodies detected vitronectin (green), platelets (red) and PECAM (bue) respectively. In the lower panel, an isotype control IgG conjugated to Dylight 488 (1 μg g−1 body weight) (green) was substituted for the anti-vitronectin antibody.
Figure 7
Figure 7. Vitronectin null mice have defective thrombus formation.
(a,b) Vitronectin null mice have reduced platelet accumulation (a) and fibrin deposition (b) after laser injury in the cremaster arterioles, as visualised by Dylight 649 labelled anti-CD42b and Alexa-488 labelled 59D8 antibody, respectively. The medium fluorescent intensity over time from at least 30 thrombi in each group is shown. (c,d) Statistical analysis of area under the curve (AUC) for platelets (c) and fibrin (d) from all thrombi collected in each group is shown. Asterisk indicates statistical significance on Mann Whitney test. (e,f) Vitronectin null mice have defective thrombus formation induced by FeCl3 in the carotid artery. (e) Representative images of platelet accumulation (red) in the carotid artery, as visualised by Dylight 649 labelled anti-CD42b, are shown. (f) The time to complete vessel occlusion from at least 6 mice in each group was monitored after applying 6% FeCl3 on the carotid artery for 1 min.
Figure 8
Figure 8. Inhibition of PDI by quercetin-3-rutinoside blocks vitronectin accumulation on the endothelium following vascular injury.
(a) Infusion of quercetin-3-rutinoside (0.5 μg g−1 body weight) 5 min before laser injury (upper panel) or vehicle (lower panel). The kinetics of the median integrated fluorescence of vitronectin (b) and platelets (c) at the injury site plotted versus time for at least 28 thrombi in 3–4 mice mice. Scale bar (cyan): 20 μm.
Figure 9
Figure 9. Inhibitory antibodies to PDI block vitronectin accumulation in the developing thrombus.
Specific antibodies to PDI, ARP48150 (anti-PDI #1, N-terminal epitope) and H-160 (anti-PDI #2, b′-domain epitope), demonstated inhibition of vitronectin accumulation on the vessel wall after infusion of the antibodies (3 μg g−1 body weight). The kinetics of the median integrated fluorescence of platelets (upper panel) and vitronectin (lower panel) at the injury site are plotted versus time from at least 30 thrombi in 3–4 mice.

Similar articles

See all similar articles

Cited by 26 articles

See all "Cited by" articles


    1. Furie B. & Furie B. C. The molecular basis of blood coagulation. Cell 53, 505–518 (1988). - PubMed
    1. Cho J., Furie B. C., Coughlin S. R. & Furie B. A critical role for extracellular protein disulfide isomerase during thrombus formation in mice. J. Clin. Invest. 118, 1123–1131 (2008). - PMC - PubMed
    1. Jasuja R., Furie B. & Furie B. C. Endothelium-derived but not platelet-derived protein disulfide isomerase is required for thrombus formation in vivo. Blood 116, 4665–4674 (2010). - PMC - PubMed
    1. Reinhardt C. et al. . Protein disulfide isomerase acts as an injury response signal that enhances fibrin generation via tissue factor activation. J. Clin. Invest. 118, 1110–1122 (2008). - PMC - PubMed
    1. Kim K. et al. . Platelet protein disulfide isomerase is required for thrombus formation but not for hemostasis in mice. Blood 122, 1052–1061 (2013). - PMC - PubMed

Publication types

MeSH terms