Cysteine Sulfenylation of Protein Disulfide Isomerase Links Oxidative Stress to Thrombosis

Oxidative stress increases the risk of clinically significant thrombosis in the setting of inflammation, malignancy, infection, and dyslipidemia. Oxidants generated during oxidative stress are prothrombotic; however, the mechanisms by which oxidants induce thrombus formation is poorly understood. Protein disulfide isomerase (PDI) is a thiol isomerase that serves an important role in thrombus formation. It has a domain structure of a-b-b'-a', in which the aand a' domains are catalytic and the b and b' are substrate binding. A flexible x linker is between the b' and a'domains. PDI is uniquely sensitive to oxidative cysteine modification within the CGHC catalytic motif of its a and a'domains. Yet whether PDI cysteine modifications mediates thrombosis in an oxidative environment has not previously been studied. We hypothesized that oxidized PDI links oxidative stress to a prothrombotic phenotype through post-translational modification of proteins. We have previously shown that dyslipidemia promotes hydrogen peroxide generation from platelets and oxidizes proteins. Using hydrogen peroxide as a model oxidant, we found that oxidation of recombinant PDI promoted the loss of free thiols and inhibited reductase activity in a dose-dependent manner (IC50= 3.3 ± 0.12 µM). To understand how free thiols were lost, we used a benzothiazine-based nucleophilic probe, termed BTD, to label sulfenic acid. The transient cysteine sulfenic acid (S-OH) was detected by BTD labeling in wildtype PDI following exposure to hydrogen peroxide, but not in a PDI mutant in which all active site cysteines had been mutated to alanine. PDI exposure to oxidized lipoprotein particles (OxLDL) or 3-morpholinosydnonimine (SIN-1), a peroxynitrite donor, also promoted its sulfenylation. Using single cysteine mutants of PDI to identify the sulfenylation site, we found that Cys56 was preferentially sulfenylated as mutation of this cysteine prevented further labeling of the sulfenic acid probe. To assess whether interdomain interactions influenced sulfenylation, we evaluated BTD labeling in isolated fragments of PDI. Sulfenylation was identified in the a domain, consistent with Cys56 labeling, and was significantly enhanced in the ab fragment, implicating an interaction between the a and b domains in sulfenylation. No labeling was observed in an isolated b'xa' fragment. Further analysis of structural elements revealed that a conserved Arg120 within the a domain regulates Cys56 reactivity. Indeed, mutation of Arg120 to Ala prevented sulfenylation by peroxides compared to the wildtype control. Maleimide labeling showed that sulfenylated PDI is an intermediary that resolves into disulfided PDI. To determine whether this disulfided PDI promotes oxidase activity, we used a denatured and reduced RNAse that refolds when disulfides are transferred from oxidized PDI to the RNAse. PDI oxidized by hydrogen peroxide promoted RNAse refolding, whereas reduced PDI did not. A selective inhibitor of sulfenylation, arsenite, prevented RNAse refolding, indicating that hydrogen peroxide-induced oxidation of PDI proceeds through a sulfenic acid intermediate. Since PDI is secreted from cells, we evaluated whether activated cells release sulfenylated PDI using both human vein endothelial cells (HUVECs) and platelets. Sulfenylated PDI was increased in the cultured media of HUVECs stimulated with thrombin and in the releasate of platelets stimulated via PAR1 or glycoprotein VI. OxLDL exposure sensitizes platelets to activation by low doses of agonists. Inhibition of PDI using a neutralizing antibody inhibited augmentation of platelet aggregation by oxLDL, suggesting that oxidation of PDI by oxLDL promotes platelet activation. The ability of oxLDL to enhance the prothrombotic effect of PDI was tested in vivo in a murine model of laser-induced thrombus formation. Mice were infused with either control IgG or IgG directed at PDI, followed by the infusion of oxLDL. OxLDL significantly enhanced thrombus formation in this model. The effect of oxLDL was inhibited by anti-PDI IgG but not control IgG antibody or buffer, implicating PDI as an effector of oxLDL-augmented thrombus formation. In conclusion, our study demonstrates sulfenylation of PDI and shows that PDI sulfenylation retains PDI oxidase activity. Oxidation of PDI by oxidants such as oxLDL occurs in platelets and in endothelium and promotes thrombosis in the setting of oxidative stress.