Hydrogen sulphide reduces hyperhomocysteinaemia‐induced endothelial ER stress by sulfhydrating protein disulphide isomerase to attenuate atherosclerosis

Abstract Hyperhomocysteinaemia (HHcy)‐impaired endothelial dysfunction including endoplasmic reticulum (ER) stress plays a crucial role in atherogenesis. Hydrogen sulphide (H2S), a metabolic production of Hcy and gasotransmitter, exhibits preventing cardiovascular damages induced by HHcy by reducing ER stress, but the underlying mechanism is unclear. Here, we made an atherosclerosis with HHcy mice model by ApoE knockout mice and feeding Pagien diet and drinking L‐methionine water. H2S donors NaHS and GYY4137 treatment lowered plaque area and ER stress in this model. Protein disulphide isomerase (PDI), a modulation protein folding key enzyme, was up‐regulated in plaque and reduced by H2S treatment. In cultured human aortic endothelial cells, Hcy dose and time dependently elevated PDI expression, but inhibited its activity, and which were rescued by H2S. H2S and its endogenous generation key enzyme‐cystathionine γ lyase induced a new post‐translational modification‐sulfhydration of PDI. Sulfhydrated PDI enhanced its activity, and two cysteine‐terminal CXXC domain of PDI was identified by site mutation. HHcy lowered PDI sulfhydration association ER stress, and H2S rescued it but this effect was blocked by cysteine site mutation. Conclusively, we demonstrated that H2S sulfhydrated PDI and enhanced its activity, reducing HHcy‐induced endothelial ER stress to attenuate atherosclerosis development.

Hydrogen sulphide (H 2 S) is a novel gasotransmitter, dependent on cystathionine-gamma lyase (CSE), cystathionine beta synthase (CBS) and 3 mercaptopyruvate sulphur transferase (3MST). 4 In cardiovascular tissues, CSE is the major endogenous key enzyme for H 2 S generation. 5 CSE/H 2 S exhibits anti-atherogenic effects, evidenced by down-regulating aortic CSE/H 2 S in atherosclerotic mice, 6 global CSE knockout exacerbating 7 but the H 2 S donor NaHS or GYY4137 attenuating atherosclerosis development. 6,8,9 More interestingly, H 2 S donor can lower serum total Hcy level by sulfhydrating CSE to increase its activity, 10 and reduce heart and aorta oxidative stress injuries and ER stress in HHcy rats. 11 However, the molecular mechanism of H 2 S attenuating HHcy-induced ER stress is unclear.
Protein disulphide isomerase (PDI) is a redox-dependent chaperone protein and key enzyme of protein folding. PDI can transfer oxidizing equivalents to an unfolding substrate, facilitating protein folding; so oxidized PDI is activated PDI. 12 Inhibition PDI potentiated but overexpression PDI reduced ER stress and cellular toxicity by ox-LDL in human microvascular endothelial cell. 13 These studies also raise a question: whether H 2 S trigging PDI activity reduced HHcyinduced endothelial ER stress, contribution to genesis of atherosclerosis. The present study aims to investigate it.

| MATERIAL S AND ME THODS
The data that support the findings of this study are available from the corresponding author upon reasonable request.

| Plaque area measurement
Mice were anaesthetized by 1% sodium pentobarbital (50 mg/kg, ip); blood was collected from the angular artery and then perfused initially with 30 ml saline and then with 4% paraformaldehyde via the left ventricle. Entire aorta was fixed 10 minutes in 4% paraformaldehyde and then marinated 5 minutes in 60% isopropanol. Total plaque area was quantified by en-face analysis of aorta. Removed adventitial fat from the aorta and stained with Oil Red O, opened longitudinally, pinned the aorta and photographed en-face. The extent of lesion development was determined as a percentage of the total area of the aorta that was occupied by Oil Red O-positive atherosclerotic lesions. The total plaque area in total arterial surface area was analysed by ImageJ.
8μm frozen aortic root slices were obtained by cutting from the apex of the heart towards the origin of the aorta (sections were mounted from the point where all three aortic valve cusps became clearly visible). The slices were stained with haematoxylin-eosin staining. Plaque cross-sectional area was determined by quantifying the plaque area by images ImageJ software.

| Immunohistochemistry and immunofluorescence staining
Aortic root plaque slice (6 μm) was performed antigen retrieval, quenching endogenous peroxides and blocking (1% BSA, 1 hour), then incubated with anti-GRP78 bip antibody (ab21685, Abcam) overnight at 4℃. After washing, the slices were incubated with HRPconjugated secondary antibody (ab7083, Abcam) for 1 hour at room temperature; the brown colour was developed using DAB, and then, images were acquired. Using mouse IgG replaced the primary antibody and corresponding secondary antibody as a negative control.
HAEC cells were planted in 6-well plates and continuously cultured to 90% fusion, then changed the ECM without FBS, homocysteine (from

| Western blot
Total protein from cells or tissues was lysed by RIPA buffer, and the for 1 hr at room temperature. After washing, images of chemiluminescence using ECL (Thermo Fisher) were acquired. Grayscale semiquantitative analysed by ImageJ.

| Biotin switch assay for protein sulfhydration
The assay was performed as described by Mustafa et al. 14

| PDI activity assay
PDI activity assay was performed as described by Appenzeller-Herzog et al. 15 The proteins without denatured were separated by using non-reducing SDS-PAGE on 12% (w/v) native-polyacrylamide gels. To prevent protein denaturation, make sure the entire electrophoresis and transfer processes are performed on ice, and do not add any denaturants and reducing agents throughout the process.
The other steps are same as Western blot. The bands of oxidative-PDI (ox-PDI, active protein) and reduction-PDI (red-PDI, inactive protein) appear at 60 KD and 150 KD, respectively. The ratio of ox-PDI/red-PDI acts as PDI activity.

| Statistical analysis
Data are presented as means ± SD. Differences among groups were analysed by one-way ANOVA, then Student-Newman-Keuls test. P < 0.05 was considered statistically significant.

| H 2 S donor's treatment reduced HHcypromoted plaque size and ER stress
To investigate the effects of H 2 S on atherosclerotic development and ER stress of plaque with HHcy, we generated a mouse model by feeding ApoE knockout mice with Paigen diet and L-methionine in drinking water (1 g/kg) for 12 weeks as our previous study. 10 H 2 S donors NaHS and GYY4137 were administered by intraperitoneal injection (twice/day), with normal saline as a control. By aortic enface Oil Red O staining, the atherosclerotic plaque areas were decreased with H 2 S donor (NaHS and GYY4137) treatment ( Figure 1A).
Consistently, aortic root plaque size was also lowered using haematein and eosin staining ( Figure 1B). Here, we also measured the serum lipid profile and found that NaHS lowered serum total cholesterol (TC) and LDL-cholesterol (LDL-c) but slightly increased HDLcholesterol (HDL-c) ( Figure 1C). By immunohistochemical staining for ER stress marker GRP78, we confirmed that GRP78 dominantly expressed in lumen layer of the plaque, and which significantly | 3441 JIANG et Al.
reduced by H 2 S donors ( Figure 1D). These data indicated that H 2 S attenuated the atherosclerotic plaque development and ER stress response by HHcy.  Figure 2B). Both NaHS and GYY4137 treatments reduced the PDI co-localization in the CD31-positive cells and α-SMApositive cells in the plaque, especially (Figure 2A and B) association with GRP78 attenuation ( Figure 1D).
According to these data, we selected DL-Hcy (200 μM) treatment HAEC for 24 h in following experiments. In this cellular model, we confirmed that H 2 S donor lowered Hcy-induced PDI elevation and ER stress ( Figure 3C).
PDI activity-specific oxidoreductase activity tightly linked with misfold protein causing unfold protein response (UPR) and apoptosis. 13 Oxidized PDI is activated form of PDI, so we detected ratio of oxidized PDI/reduced PDI as PDI activity. Here, we found that DL-Hcy also reduced PDI activity in dose-dependent ( Figure 4A) and time-dependent ( Figure 4B) manner. The inhibition of PDI activity was reversed by H 2 S donor ( Figure 4C). These data highlight HHcy increased UPR, in part due to the PDI dysfunction, and which were reversed by H 2 S treatment.

| H 2 S sulfhydrated PDI to promote PDI activity
Protein sulfhydration at the cysteine residue is a novel translational modification attribution to protein functional regulation. 14 To investigate whether the PDI activity is dependent on its sulfhydration, we detect it by biotin switch assay. In vitro, cellular lysis incubated with NaHS, and the PDI sulfhydration was confirmed and which could be removed by DTT (sulfhydration remover, 100 mM). In cultured HAEC, NaHS (1 mM) or DTT (1 mM) treatment for 2 hours, endogenous PDI sulfhydration was also verified ( Figure 5A). In keeping with PDI sulfhydration, PDI activity also elevated by NaHS and GYY but reduced by DTT ( Figure 5B).
Since endogenous H 2 S generation in endothelium is majorly dependent on CSE, we found that overexpression or knockdown CSE using adenovirus enhanced or reduced PDI sulfhydration ( Figure 5C).
Accordingly, PDI activity also increased by overexpression of CSE or decreased by knockdown CSE (Figure 5D). Corresponding to PDI activity by Hcy, DL-Hcy also down-regulated endogenous PDI sulfhydration, which was reversed by GYY4137 ( Figure 5E). All of data indicate that sulfhydration is a novel functional post-translational modification of PDI by CSE/H 2 S, modulation its activity.  Figure 6B showed, all M1, M2 and M3 blocked PDI sulfhydration, indicating two CXXC domains could be sulfhydrated by H 2 S. In line with the sulfhydration blocking, H 2 S-promoted PDI activity was also blocked by these mutations ( Figure 6C). To verify the PDI sulfhydration association with ER stress, we treated these HEK-293 cells with DL-Hcy and GYY4137 for 24h. In wild-type PDI, GYY4137 lowered Hcy-induced GRP78, XBP1s and cleavedcaspase-12 expression, which was blocked by M1, M2 and M3 mutation ( Figure 6D). These data suggest that PDI sulfhydration at C53/57 and C397/400 sites partly contributes to reducing HHcyinduced ER stress effect of H 2 S.

| D ISCUSS I ON
In the present study, we first demonstrate that CSE/H 2 S sulfhydrates PDI at C53/57, C397/400 sites facilitating its protein folding activity, then reducing unfold protein response (final process of ER stress) by HHcy, attenuating endothelial dysfunction and atherosclerosis development with HHcy ( Figure 6E). This study also highlights a novel molecular mechanism of H 2 S on anti-cardiovascular injuries of HHcy, except for scavenging reactive oxygen species 11 and lowering Hcy by sulfhydrating CSE. 10 F I G U R E 1 H 2 S donor treatment lowered plaque area and ER stress in atherosclerosis with HHcy mice. A, Aorta plaque area was measured by en-face Oil-red O staining. B, Aortic root plaque size was detected by H&E staining (bar = 500 μm). C, Serum total cholesterol, LDL-cholesterol (LDL-c), HDL-cholesterol (HDL-c) changes after H2S donor treatment. D, ER stress marker protein-GRP78 immunohistochemical staining was assayed in aortic root plaque (bar = 20 μm). Small images in the left corner are overall perspective pictures Hyperhomocysteinaemia is an independent risk factor of cardiovascular disease, especial atherosclerosis. 1 Although B type vitamins or folic acid interventions can lower total Hcy, these therapies did not effectively prevent cardiovascular events by serial meta-analyses, [16][17][18][19] in part due to these intervention failures to repair the existing damage by HHcy. H 2 S is a novel protective gasotransmitter for atherosclerosis. 20 H 2 S could reduce HHcyinduced cardiac, aortic, VSMC and endothelial oxidative stress Elevated Hcy caused endothelial dysfunction including ER stress-induced apoptosis, similar to other ER stress inducers tunicamycin and thapsigargin. 3 However, the underlying mechanism of HHcy-inducing ER stress is almost unknown. PDI is a prototypic thiol isomerase that catalyses the formation and cleavage of thioldisulphide bone. 28 PDI also functions as a molecular chaperone that prevents misfolded protein aggregation. 29 About 30% proteins required PDI to form disulphide bone. Oxidized PDI binding to unfold protein-free sulfhydryl facilitates protein folding with endoplasmic reticulum oxidoreductin 1. 12 So, the oxidized PDI is activated form of PDI. Ox-LDL inhibited PDI causing ER stress and endothelial apoptosis to promote atherosclerosis. 13 Here, we found that HHcy dramatically elevated PDI protein expression in endothelial-derived cells of plaque, association with ER stress. In HAECs, HHcy also up-regulated PDI and ER stress markers, but down-regulated PDI activity. These data highlight that PDI activity inhibition is the molecular mechanism of HHcy-inducing endothelial UPR. Furthermore, we found that H 2 S donor significantly enhanced PDI activity per se, and reversed the HHcy effects on PDI expression and activity. This finding also indicates that PDI activity is a regulatory target of H 2 S. is also nitrosylated at CXXC domains, leading to accumulation of polyubiquitinated proteins and activation of UPR. 38 Here, we identified a new post-translational modification-sulfhydration of PDI at the same sites of nitrosylation or glutathionylation. Sufhydrated PDI inhibited ER stress induced by HHcy, preventing endothelial dysfunction.

F I G U R E 4 H 2 S rescued
Collectively, our present study highlights a new molecular mechanism of H 2 S preventing endothelial function under HHcy, attenuating atherosclerosis. We also demonstrated a new post-translational modification, sulfhydration of PDI, which activate PDI to reduce ER stress. Some H 2 S-releasing donors such as SG1002 for heart failure are in phase I clinical trials 39 ; ATB-346 for anti-inflammation in a phase II clinical trial 40 showed some beneficial effects. Our work offers a further possibility of these drugs for atherosclerosis with HHcy therapy.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.