Abstract

   S-sulfhydration is a signalling pathway of hydrogen sulfide (H[2]S),
   which is suggested as an anti-atherogenic molecule that may protect
   against atherosclerosis. The identification of S-sulfhydrated proteins
   by proteomic approach could be a major step towards understanding the
   mechanisms of H[2]S in response to atherosclerosis. The present study
   studied targeted S-sulfhydrated proteins using the modified biotin
   switch method followed by matrix-assisted laser desorption/ionisation
   time of flight tandem mass spectrometry identification. The results
   showed that H[2]S can protect against atherosclerosis by reducing body
   weight gain and alleviating aortic plaque formation. In addition, H[2]S
   treatment can increase aortic protein S-sulfhydration. Seventy targeted
   S-sulfhydrated aortic proteins were identified, mainly involved in
   metabolism, stimulus response and biological regulation, as determined
   by gene ontology database analysis. H[2]S also induced S-sulfhydration
   of glutathione peroxidase 1 and further reduced lipid peroxidation and
   increased antioxidant defence in the aorta by prompting glutathione
   synthesis. Our data suggest that H[2]S is a cardiovascular-protective
   molecule that S-sulfhydrates a subset of proteins that are mainly
   responsible for lipid metabolism and exerts its cytoprotective effects
   to clear free radicals and inhibit oxidative stress through cysteine
   S-sulfhydration.

Introduction

   Hydrogen sulfide (H[2]S) is one of the three molecules (along with
   nitric oxide and carbon monoxide) most recently identified as
   gasotransmitters [[26]1]. Despite its toxicity, it is recognised as an
   endogenously produced gaseous signalling molecule that induces a
   variety of beneficial reactions in cardiovascular tissues [[27]2].
   H[2]S activity upregulates such physiological properties as
   vasodilation, anti-inflammation, anti-oxidation and angiogenesis
   [[28]3]. In addition, accumulating evidence suggests an association
   between atherosclerosis and deficiency in vascular cystathionine
   γ-lyase (CSE), an enzyme responsible for endogenous H[2]S production.
   It was shown by using CSE knockout mice that CSE gene deletion led to a
   decrease in H[2]S production with accelerated atherosclerosis [[29]4].
   Atherosclerotic plaques could also be ameliorated by overexpression of
   CSE in apolipoprotein E knockout (ApoE KO) mice [[30]5] and H[2]S
   supplementation by exogenous H[2]S donors such as NaHS [[31]6] and
   GYY4137 [[32]7]. H[2]S is therefore suggested as an anti-atherogenic
   molecule. However, the mechanism by which H[2]S reduces atherosclerosis
   is still not clear.

   Recently, H[2]S signalling through S-sulfhydration has been reviewed,
   and S-sulfhydration has been recognised as a key mechanism of H[2]S
   physiology [[33]8]. Similar to S-nitrosylation, S-sulfhydration can
   cause post-translational modification of proteins [[34]9].
   S-sulfhydration takes place at reactive cysteine residues in proteins
   and results in the conversion of an–SH group of cysteine to an–SSH
   group. This conversion occurs in many proteins and modulates the
   functions of a targeted protein [[35]9]. For example, H[2]S was shown
   to protect gastric epithelial cells from ischemia-reperfusion injury by
   Keap1 sulfhydration and further Keap1/Nrf2 disassociation and
   enhancement of Nrf2 nuclear translocation, which eventually attenuated
   oxidative stress and protected against mucosal injury [[36]10].
   Recently, H[2]S was suggested to inhibit the development of
   atherosclerosis through upregulating protein S-nitrosylation, thereby
   inhibiting the proliferation and migration of vascular smooth muscle
   cells [[37]11]. Cystathionine β-synthase (CBS), another H[2]S-producing
   enzyme, was found to regulate endothelial function via protein
   S-sulfhydration, maintaining vascular health and function [[38]12]. All
   these observations suggest the potential value of investigating
   S-sulfhydration in the context of atherosclerosis.

   H[2]S signalling via S-sulfhydration is still not fully elucidated, in
   that only a few proteins have been identified to be S-sulfhydrated by
   H[2]S. To investigate H[2]S signalling in atherosclerosis,
   meta-analysis of S-sulfhydrated proteins in aortic proteins is the main
   step to be examined. Here we explore S-sulfhydrated proteins in aortic
   tissues of ApoE KO mice on an atherogenic diet and H[2]S treatment. We
   detected S-sulfhydration by modified biotin switch assay coupled with
   advanced proteomic identification strategies and analysed altered
   S-sulfhydration levels of proteins. Potential targets of protein
   S-sulfhydration were identified in aortic tissue. The difference in
   S-sulfhydration levels of proteins in aortic tissue indicates the
   possibility of functional alteration as well as disease-related
   consequences.

Materials and methods

Animals and experimental design

   Male wild-type C57BL/6J mice (WT) and ApoE KO mice were used in this
   study. All mice were housed in a controlled environment with free
   access to water and food on a 12-h light/dark cycle. WT mice were used
   as a reference of healthy mice without atherosclerosis. Atherosclerotic
   mice, ApoE KO mice were treated with or without H[2]S in the
   experiment. Briefly, they were fed a normal diet until 8 weeks of age
   and then switched to an atherogenic diet. After 4 weeks of the
   atherogenic diet, the ApoE KO mice were divided into two groups:
   vehicle control treatment with normal saline and H[2]S treatment with
   GYY4137 (Cayman Chemical Co., Ann Arbor, MI, USA), a novel H[2]S donor
   [[39]13]. The H[2]S supplement experiment consisted of daily IP
   injection of GYY4137 (133 μmol/kg body weight for 4 weeks of
   atherogenic diet feeding) in the ApoE KO mice. Control ApoE KO mice
   were injected with normal saline alone. Body weights of mice were
   recorded weekly. Food uptake per mouse per day was calculated as
   follows: [total food intake per cage]/[mice per cage]/[days of food
   consumption]. After 4 weeks of treatment, all mice were sacrificed by
   using slow fill carbon dioxide asphyxiation. Non-clotted blood was
   collected from each mouse. Aortic tissues were excised, immediately
   frozen in liquid nitrogen and stored at –80°C until further processing.
   All animal experiments were approved by the Animal Experimentation
   Ethics Committee (AEEC) of the Chinese University of Hong Kong (CUHK)
   and in accordance with the Animals (Control of Experiments) Ordinance
   (Cap. 340) licensed from the Hong Kong Government Department of Health.
   They were performed conform the university’s regulations on Use of
   Experimental Animals which incorporate The International Guiding
   Principles for Biomedical Research Involving Animals and The Hong Kong
   Code of Practice for Care and Use of Animals for Experimental Purposes.

Release of H[2]S in plasma and tissues after GYY4137 treatment

   To detect H[2]S concentration in vivo after GYY4137 treatment, tissues
   (liver and kidney) and blood were collected at timed intervals (0 to 8
   hours). H[2]S level in tissue homogenates was measured by zinc acetate
   assay as described previously with some modifications [[40]13].
   Briefly, liver and kidney were homogenized in ice-cold 100 mM potassium
   phosphate buffer (pH 7.4). L-cysteine (10 mM; 20 μl), pyridoxal
   5’-phosphate (2 mM; 20 μl) and saline (30 μl) were added into the
   homogenate (430 μl). After incubation at 37°C for 30 min, zinc acetate
   (1% w/v, 250 μl) was injected to trap-generated H[2]S followed by
   trichloroacetic acid (10% w/v, 250 μl) to precipitate protein and thus
   stop the reaction. Subsequently, N,N-dimethyl-p-phenylenediamine
   sulfate (20 μM; 133 μl) in 7.2 M HCl was added followed by FeCl[3] (30
   μM; 133 μl) in 1.2 M HCl. After 15 min, absorbance at 670 nm was
   measured with a microplate reader (Bio-Tek Instrument INC., Rockville,
   MD, USA). Aliquots (500 μl) of mouse plasma were tested for H[2]S using
   the same procedure. All standards and samples were assayed in
   duplicate. The calibration curve of absorbance versus H[2]S
   concentration was obtained by using NaHS solution of varying
   concentrations (10 to 160 μm).

Quantitation of atherosclerotic plaque

   Quantitation of atherosclerotic plaque was carried out as previously
   reported [[41]5]. In brief, atherosclerotic lesions in the aortic root
   were examined at three location levels, and nine serial sections were
   prepared from each location. Sections from three locations were
   selected and stained with oil red O and counterstained with Mayer’s
   hematoxylin. The images were then produced using an inverted microscope
   (Olympus IX-83, Tokyo, Japan). Histological staining in all images was
   quantified using ImageJ, and the mean staining within the aortic root
   of each mouse was obtained.

Detection of S-sulfhydration by modified biotin switch assay

   The assay was carried out with modifications according to Mustafa et
   al. [[42]8]. Control and H[2]S-treated ApoE KO mice aortas were
   homogenised in HENS buffer [250 mM HEPES-NaOH (pH 7.7), 1 mM EDTA, and
   0.1 mM neocuproine] supplemented with 100 μM deferoxamine and
   centrifuged at 13,000 g for 30 min at 4°C. Aorta homogenates were added
   to blocking buffer [HENS buffer adjusted to 2.5% sodium dodecyl sulfate
   (SDS) and 20 mM methyl methanethiosulfonate (MMTS)] at 50°C for 30 min
   with rotation. MMTS was removed by acetone, and the proteins were
   precipitated at –20°C for 30 min. After acetone removal, the proteins
   were resuspended in HENS buffer with 1% SDS and then incubated for 3 h
   at room temperature with biotin-HPDP in dimethyl sulfoxide without
   ascorbic acid, centrifuged, and resuspended in HENS buffer. The
   biotinylated and S-sulfhydrated proteins were pulled down by
   streptavidin beads. To detect biotinylated proteins, samples were
   separated on 12% SDS-PAGE gels, transferred to a polyvinylidene
   difluoride (PVDF) membrane, blocked with 5% non-fat milk, and incubated
   with streptavidin-peroxidase diluted 1:2000 for 1 h. The biotinylated
   proteins were also subjected to western blotting analysis using
   anti-glutathione peroxidase 1 (anti-GPx1) antibody (Abcam, Cambridge,
   UK).

In-gel tryptic digestion

   The biotinylated proteins were separated on a 12% SDS polyacrylamide
   gel (SDS-PAGE) and stained with coomassie blue. The gel lanes were cut
   into three slices. Each slice was further cut into 1x1 to 2x2-mm
   pieces, which were then placed into a 1.5 mL tube. The gel pieces were
   destained two to three times with destaining solution (50%
   acetonitrile, 25 mM ammonium bicarbonate). After in-gel reduction (10
   mM dithiothreitol, 45 min, 56°C) and alkylation (11 mM iodoacetamide,
   45 min in dark), gel pieces were shrunk in acetonitrite and dried down.
   Each gel slice sample was rehydrated by 0.3 μg trypsin in 25 mM
   ammonium bicarbonate and incubated overnight at 37°C. After that, the
   reaction was stopped with 1% trifluoroacetic acid. Peptides were
   extracted from gel pieces with incubation in extraction buffer (50%
   acetonitrile, 1% trifluoroacetic acid). The supernatant from all the
   fractions of the same sample were pooled and dried using a Speedvac.

Mass spectrometry analysis and protein identification

   The peptides were dissolved in a solution of 1% formic acid in water
   and injected onto an LC/MS system consisting of a Dionex Ultimate 3000
   nano-LC system (Thermo Fisher Scientific Inc., Germany) associated with
   fraction collector, Proteineer fcII (Bruker Daltonics, Bremen, Germany)
   to fractionate and spot fractions onto MALDI target plates for protein
   identification for matrix-assisted laser desorption/ionisation. Mass
   spectrometry analyses were carried out on an UltrafleXtreme MALDI
   TOF/TOF analyser (Bruker Daltonics; Bremen, Germany). The database
   search was done using the Mascot search engine (Matrix Science Mascot
   2.4.1) on the SwissProt databases (545536 sequences). The following
   parameters were used: up to 1 miss cleavages; MS tolerance 50 ppm; MSMS
   tolerance 0.5 Da; full tryptic peptides; partial modifications
   carbamidomethylation (C), oxidation (M); MudPIT scoring; significance
   threshold P < 0.05. Protein identifications were validated only if they
   satisfied three requirements: (i) their score was significant (P <
   0.05) with cut-off criteria; (ii) they were identified with three
   peptides with score >70; and (iii) they were identified in at least two
   out of the three runs.

Gene ontology (GO) and enrichment analysis in the protein data set

   To determine GO annotations for selected proteins, we used the Database
   for Annotation, Visualization and Integrated Discovery (DAVID)
   ([43]https://david.ncifcrf.gov/; version 6.8) to access a relational
   database of gene functional annotations [[44]14,[45]15]. For the
   enrichment analysis, Protein ANalysis THrough Evolutionary
   Relationships (PANTHER) classification ([46]http://geneontology.org;
   version 11.1) was performed to investigate the enrichment of molecular
   function, biological process and protein class associated with each
   identified protein [[47]16]. Enriched pathways were analysed by KEGG
   pathway analysis using WEB-based Gene SeT AnaLysis Toolkit (WEBGESTALT)
   ([48]http://www.webgestalt.org/option.php) [[49]17]. Enriched KEGG
   pathways and GO terms were determined based on their corresponding
   P-values for over-representation against the entire mouse genome. P
   values were calculated using a hypergeometric test and corrected for
   multiple testing with a Benjamini-Hochberg (GO terms) or Bonferroni
   (KEGG pathway) correction. A cut-off of 0.05 was applied.

Western blotting

   Briefly, protein samples were separated by 12% SDS-PAGE and then
   transferred into polyvinylidene difluoride membranes. The membranes
   were blocked with 5% non-fat milk and then probed overnight at 4°C with
   anti-GPx1 (1:1000; Abcam). This step was followed by exposure to
   secondary antibody for 1 h with a 1:2000 dilution of goat anti-rabbit
   secondary antibody (Cell Signaling Technology, USA). After washing, the
   membrane was developed with an ECL kit (Amersham Pharmacia
   Biotechnology, UK) and exposed on a ChemiDoc imaging system (Bio-Rad
   Laboratories, Hercules, CA). Intensity of bands was quantified using
   the public domain software, ImageJ software with linear signal tool.
   Each GPx1 band of each mouse was first normalized with its actin
   intensity and the SH-GPx1 was then calculated (SH-GPx1/Input-GPx1).

Immunofluorescence staining

   Immunofluorescence staining of aortic root sections was performed using
   rabbit anti-mouse GPx1 antibody to observe the intensity of GPx1 in the
   aortic roots. Briefly, air-dried cryostat sections (6 μm thickness) of
   aortic root were fixed with the pre-cooled acetone for 10 min and
   rinsed with 1× phosphate buffered saline (PBS) three times.
   Non-specific staining was blocked by incubation with 5% goat serum in
   1× PBS for 30 min followed by incubation of anti-GPx1 antibody
   overnight at 4°C in a 1:200 dilution. After rinsing with 1× PBS,
   sections were incubated with Alexa 488 anti-rabbit secondary antibody
   (Invitrogen Molecular Probes, UK) for 1 h while nuclei were stained
   with DAPI in the mounting agent (Invitrogen Molecular Probes).
   Microscopic examination was performed on an Olympus IX-83 inverted
   fluorescence microscope (Olympus, Tokyo, Japan). Images were acquired
   by a monochrome CCD camera using CellSens software (Olympus, Tokyo,
   Japan).

Glutathione activities determination

   A commercially available glutathione peroxidase (GPx) assay kit (Cayman
   Chemical Company) was used to measure plasma GPx activity according to
   the manufacturer’s instructions. Glutathione (GSH) and total GSH were
   measured using GSH colorimetric assay (BioVision Incorporated, CA,
   USA). Oxidised glutathione/glutathione disulfide (GSSG) was calculated
   by subtraction of GSH from total GSH.

Lipid peroxidation determination

   Malondialdehyde (MDA) is a product of lipid peroxidation. Therefore,
   lipid peroxidation was determined by measuring MDA concentration in
   mice plasma using a thiobarbituric acid reactive substances (TBARS)
   assay kit (Cayman Chemical Company).

Statistical analysis

   All data are presented as mean ± SEM. The non-parametric Mann Whitney
   test was used to compare variables between two groups. Statistical
   significance was determined at a level of P < 0.05. Statistical
   analysis was performed with GraphPad Prism 5.0 software.

Results

H[2]S reduced body weight gain and aortic plaque area in ApoE KO mice

   All ApoE KO mice were first fed with an atherogenic diet for 4 weeks.
   Their body weight gain decreased in the first week but subsequently
   returned to normal and remained stable. The mice were then divided into
   the vehicle-control group and H[2]S-treated group. The body weight gain
   in the H[2]S group was significantly reduced after H[2]S treatment,
   whereas the control group body weight remained similar over the whole
   period of treatment ([50]Fig 1A). The H[2]S-treated mice had about 5%
   body weight loss after 4 weeks of H[2]S treatment ([51]Fig 1B). To
   demonstrate whether GYY4137 treatment could account for food
   consumption and thus body weight gain, food uptake by both group mice
   was measured. It found that GYY4137 treatment did not affect the food
   consumption ([52]S2 Table). In addition, after the H[2]S treatment, the
   aortic roots were sectioned and stained with lipids using oil red O,
   which stains atherosclerotic plaques a red colour. There was no red
   staining in WT mice, and little staining seen in the atherogenic
   diet-fed ApoE KO mice before treatment ([53]Fig 1C). After treatment,
   comparing the size of plaques in the two groups, the control group of
   mice had a larger plaque size in the aortic root ([54]Fig 1D) and the
   oil red O staining was widely observed in the histological result
   ([55]Fig 1C). In contrast, the plaque size of H[2]S-treated mice was
   significantly diminished by 30% ([56]Fig 1C and 1D), indicating the
   retardation of the atherosclerotic response in the mice by H[2]S
   treatment.

Fig 1. H[2]S administration in ApoE KO mice can reduce body weight gain and
atherosclerotic plaque area.

   [57]Fig 1
   [58]Open in a new tab

   ApoE KO mice were first fed an atherogenic diet for 4 weeks before
   H[2]S treatment and then randomly divided into two groups: (i) ApoE KO
   + saline (0.1 mL per day, i.p.) for 4 weeks and (ii) ApoE KO + H[2]S
   donor (133 μmol per kg per day, i.p.) for 4 weeks. (A) During the
   treatment, the body weight of mice was measured weekly and the weight
   gain per week was calculated. Body weight was lost in mice treated with
   H[2]S in the first week, whereas the control group had no significant
   change in weight. Data represent mean ± SEM (n = 6–12). (B) The body
   weight (± SD) was compared before treatment and after 4-weeks of
   treated in both groups of mice (n = 6–12). H[2]S-treated ApoE KO mice
   were shown to have significant body weight loss after the H[2]S
   treatment. (C) After treatment, the aortic root was sectioned and
   stained with oil red O for lipids. H[2]S-treated ApoE KO mice showed
   reduced lesion area (appearing as red colour in the aortic root) as
   compared to the saline control group. It was also compared to the ApoE
   KO mice before H[2]S treatment and the WT C57BL/6J mice, which showed
   few lesions and no lesions in the aortic root, respectively. (D) The
   plaque size was quantified as percentage area of red staining in the
   aortic root. Values are means ± SEM, n = 6. The groups were
   significantly different from each other according to the Mann Whitney
   test, P < 0.05.

Identification and characterisation of S-sulfhydrated proteins in aorta

   The modified approach for the biotin switch method for detecting
   endogenous and exogenous H[2]S revealed a total of 70 significant
   proteins (P values < 0.05), corresponding to different S-sulfhydrated
   proteins identified using tandem mass spectrometry ([59]Table 1). For
   the 70 identified and S-sulfhydrated proteins, GO analysis showed that
   38.6% of the identified proteins were involved in oxidation-reduction
   processes. Other biological processes involved transport (17.1%),
   metabolic process (17.1%), tricarboxylic acid cycle (TCA; 14.3%), lipid
   metabolic process (12.9%), fatty acid metabolic process (11.4%),
   response to oxidative stress (8.6%) and negative regulation of
   apoptotic process (8.6%) ([60]Fig 2A). In regard to molecular function,
   the identified proteins are responsible for protein binding (35.9%),
   ion binding (25.0%), nucleotide binding (23.4%), lipid binding (12.5%),
   enzyme regulator activity (3.1%) and also antioxidant activity (1.6%)
   ([61]Fig 2B).

Table 1. Identified S-sulfhydrated proteins from aorta of ApoE KO mice
treated with H[2]S.

   No. Accession ID Protein MW [kDa] pI Score No. of Peptides Coverage [%]
   1 [62]P60710 ACTB Actin, cytoplasmic 1 41.7 5.3 871.6 13 43.2
   2 [63]P63268 ACTH Actin, gamma-enteric smooth muscle 41.8 5.3 702.6 13
   43.9
   3 [64]P10605 CATB Cathepsin B 37.3 5.6 520.1 8 20.9
   4 [65]Q8BWT1 THIM 3-ketoacyl-CoA thiolase, mitochondrial 41.8 8.3 508.7
   10 27
   5 [66]Q9DBJ1 PGAM1 Phosphoglycerate mutase 1 28.8 6.7 447.4 7 32.3
   6 [67]Q9DCW4 ETFB Electron transfer flavoprotein subunit beta 27.6 8.2
   442.5 6 34.5
   7 [68]Q8BFZ3 ACTBL Beta-actin-like protein 2 42.0 5.3 420.7 7 20.5
   8 [69]P17742 PPIA Peptidyl-prolyl cis-trans isomerase A 18.0 7.7 366.5
   7 43.9
   9 [70]P07356 ANXA2 Annexin A2 38.7 7.5 348.5 5 22.4
   10 [71]P16015 CAH3 Carbonic anhydrase 3 29.3 6.9 314.6 6 30.8
   11 [72]Q05816 FABP5 Fatty acid-binding protein, epidermal 15.1 6.1 310
   5 52.6
   12 [73]P05064 ALDOA Fructose-bisphosphate aldolase A 39.3 8.3 309.1 6
   20.3
   13 [74]P62962 PROF1 Profilin-1 14.9 8.5 304.3 5 40.0
   14 [75]P35486 ODPA Pyruvate dehydrogenase E1 component subunit alpha,
   somatic form, mitochondrial 43.2 8.5 295.8 6 16.9
   15 [76]O88569 ROA2 Heterogeneous nuclear ribonucleoproteins A2/B1 37.4
   9.0 294.4 4 19.3
   16 [77]Q9CZU6 CISY Citrate synthase, mitochondrial 51.7 8.7 294.1 6
   16.8
   17 [78]P04117 FABP4 Fatty acid-binding protein, adipocyte 14.6 8.5
   284.8 4 26.5
   18 [79]P01942 HBA Hemoglobin subunit alpha 15.1 8.0 277.8 4 34.5
   19 [80]P97429 ANXA4 Annexin A4 35.9 5.4 271.9 6 22.9
   20 [81]P19783 COX41 Cytochrome c oxidase subunit 4 isoform 1,
   mitochondrial 19.5 9.3 271.2 6 28.4
   21 [82]Q8K1Z0 COQ9 Ubiquinone biosynthesis protein COQ9, mitochondrial
   35.1 5.6 270.5 3 18.2
   22 [83]Q9WVA4 TAGL2 Transgelin-2 22.4 8.4 257.2 4 28.1
   23 [84]Q01768 NDKB Nucleoside diphosphate kinase B 17.4 7.0 257.0 6
   42.1
   24 [85]P54071 IDHP Isocitrate dehydrogenase [NADP], mitochondrial 50.9
   8.9 253.9 6 15.3
   25 [86]Q9D6R2 IDH3A Isocitrate dehydrogenase [NAD] subunit alpha,
   mitochondrial 39.6 6.3 248.2 6 17.2
   26 [87]P70404 IDHG1 Isocitrate dehydrogenase [NAD] subunit gamma 1,
   mitochondrial 42.8 9.2 245.1 4 14.0
   27 [88]P51174 ACADL Long-chain specific acyl-CoA dehydrogenase,
   mitochondrial 47.9 8.5 244.3 6 17.0
   28 [89]P20152 VIME Vimentin 53.7 5.1 234.3 5 16.1
   29 [90]Q9D051 ODPB Pyruvate dehydrogenase E1 component subunit beta,
   mitochondrial 38.9 6.4 228.5 5 25.6
   30 [91]Q8BH95 ECHM Enoyl-CoA hydratase, mitochondrial 31.5 8.8 219.3 3
   15.2
   31 [92]P14152 MDHC Malate dehydrogenase, cytoplasmic 36.5 6.2 214.2 5
   18.9
   32 [93]Q7TMF3 NDUAC NADH dehydrogenase [ubiquinone] 1 alpha subcomplex
   subunit 12 17.1 9.4 213.9 5 45.5
   33 [94]P48036 ANXA5 Annexin A5 35.7 4.8 212.5 5 22.3
   34 [95]P10649 GSTM1 Glutathione S-transferase Mu 1 26.0 7.7 208.6 5
   25.2
   35 [96]Q99LY9 NDUS5 NADH dehydrogenase [ubiquinone] iron-sulfur protein
   5 12.6 9.1 207.3 4 46.2
   36 [97]P08249 MDHM Malate dehydrogenase, mitochondrial 35.6 8.9 205.3 4
   16.0
   37 [98]O08756 HCD2 3-hydroxyacyl-CoA dehydrogenase type-2 27.4 8.5
   201.2 4 19.9
   38 [99]Q61425 HCDH Hydroxyacyl-coenzyme A dehydrogenase, mitochondrial
   34.4 8.8 196.2 4 10.8
   39 [100]Q9CZ13 QCR1 Cytochrome b-c1 complex subunit 1, mitochondrial
   52.8 5.8 195.7 4 14.6
   40 [101]P10107 ANXA1 Annexin A1 38.7 7.0 193.0 4 13.6
   41 [102]P09671 SODM Superoxide dismutase [Mn], mitochondrial 24.6 8.8
   176.3 3 19.8
   42 [103]P02088 HBB1 Hemoglobin subunit beta-1 15.8 7.1 156.4 3 21.8
   43 [104]P99024 TBB5 Tubulin beta-5 chain 49.6 4.8 156.1 5 11.9
   44 [105]Q9CR61 NDUB7 NADH dehydrogenase [ubiquinone] 1 beta subcomplex
   subunit 7 16.3 8.4 153.1 3 38.0
   45 [106]O08807 PRDX4 Peroxiredoxin-4 31.0 6.7 149.7 3 11.7
   46 [107]Q9Z2I9 SUCB1 Succinyl-CoA ligase [ADP-forming] subunit beta,
   mitochondrial 50.1 6.6 148.0 4 9.7
   47 [108]Q9D0M3 CY1 Cytochrome c1, heme protein, mitochondrial 35.3 9.2
   142.9 3 15.4
   48 [109]P13707 GPDA Glycerol-3-phosphate dehydrogenase [NAD(+)],
   cytoplasmic 37.5 6.8 138 4 12.0
   49 [110]O35855 BCAT2 Branched-chain-amino-acid aminotransferase,
   mitochondrial 44.1 8.6 137.8 4 14.0
   50 [111]P14131 RS16 40S ribosomal protein S16 16.4 10.2 133.2 4 24.7
   51 [112]Q99JY0 ECHB Trifunctional enzyme subunit beta, mitochondrial
   51.4 9.4 133.0 4 9.9
   52 [113]P67778 PHB Prohibitin 29.8 5.6 131.5 3 16.2
   53 [114]Q9DC69 NDUA9 NADH dehydrogenase [ubiquinone] 1 alpha subcomplex
   subunit 9, mitochondrial 42.5 9.7 131.2 3 8.0
   54 [115]P09528 FRIH Ferritin heavy chain 21.1 5.5 128.5 3 15.4
   55 [116]Q9CQI6 COTL1 Coactosin-like protein 15.9 5.2 128 4 23.9
   56 [117]P42125 ECI1 Enoyl-CoA delta isomerase 1, mitochondrial 32.2 9.1
   127.1 3 9.3
   57 [118]Q9R0P5 DEST Destrin 18.5 8.1 124.4 3 18.2
   58 [119]P15105 GLNA Glutamine synthetase 42.1 6.6 123.5 3 12.3
   59 [120]Q8BFR5 EFTU Elongation factor Tu, mitochondrial 49.5 7.2 117.7
   3 10.0
   60 [121]P16110 LEG3 Galectin-3 27.5 8.5 115.5 3 12.1
   61 [122]P14206 RSSA 40S ribosomal protein SA 32.8 4.8 113.8 4 15.3
   62 [123]Q9DCJ5 NDUA8 NADH dehydrogenase [ubiquinone] 1 alpha subcomplex
   subunit 8 20.0 8.8 110.1 3 18.6
   63 [124]Q9JII6 AK1A1 Alcohol dehydrogenase [NADP(+)] 36.6 6.9 97.3 3
   6.2
   64 [125]P97351 RS3A 40S ribosomal protein S3a 29.9 9.8 92.5 3 14.8
   65 [126]P45952 ACADM Medium-chain specific acyl-CoA dehydrogenase,
   mitochondrial 46.5 8.6 91.7 3 5.9
   66 [127]P59999 ARPC4 Actin-related protein 2/3 complex subunit 4 19.7
   8.5 87.4 3 22.6
   67 [128]O88844 IDHC Isocitrate dehydrogenase [NADP] cytoplasmic 46.6
   6.7 85.5 3 9.4
   68 [129]P35700 PRDX1 Peroxiredoxin-1 22.2 8.3 81.0 3 22.6
   69 [130]P61205 ARF3 ADP-ribosylation factor 3 20.6 6.8 80.5 3 14.9
   70 [131]P11352 GPx1 Glutathione peroxidase 1 22.3 6.7 71.4 3 17.4
   [132]Open in a new tab

Fig 2. Gene ontology (GO) analysis of S-sulfhydrated aortic proteins.

   [133]Fig 2
   [134]Open in a new tab

   The proteins were analysed in terms of (A) GO biological process and
   (B) GO molecular function of proteins.

Enrichment analysis of S-sulfhydrated proteins in aorta

   In the enrichment analysis, biological processes categorised as lipid
   metabolism (e.g., fatty acid metabolism and oxidation) or energy
   production and conversion (e.g., TCA cycle) were mainly enriched in the
   identified protein set ([135]Fig 3A). For the molecular function,
   proteins involved in antioxidant activity were enriched the most
   against the entire mice genome, followed by peroxidase activity
   ([136]Fig 3B). In terms of protein class, hydratase was augmented the
   most in the identified proteins (~ 50-fold). Others included peroxidase
   (~38-fold), actin-related proteins (~33-fold) and dehydrogenase
   (~23-fold) ([137]Fig 3C).

Fig 3. Gene ontology (GO) enrichment analysis for S-sulfhydrated proteins in
mice aorta.

   [138]Fig 3
   [139]Open in a new tab

   Enriched GO terms for (A) biological process, (B) molecular function
   and (C) protein class were determined based on their corresponding P
   values for over-representation against the mice entire genome.

   Furthermore, the enrichment of identified S-sulfhyrated proteins was
   observed in 17 molecular pathways by the pathway enrichment analysis
   using KEGG (WEBGESTALT) ([140]Fig 4A). Out of 70 sulfhydrated proteins,
   5 were involved in lipid metabolism pathways, including
   hydroxyacyl-coenzyme A dehydrogenase (HADH), acyl-CoA dehydrogenases,
   long chain (ACADL), enoyl CoA hydratase, short chain, 1, mitochondrial
   (ECHS1) acyl-CoA dehydrogenase, medium chain (ACADM) and enoyl-CoA
   delta isomerase 1 (ECI1). Also found to be enriched were proteins
   involved in valine/leucine/isoleucine (e.g., ACADM and
   branched-chain-amino-acid aminotransferase, mitochondrial, BCAT2) and
   GSH metabolism (e.g., GPx1 and glutathione S-transferase Mu 1, GSTM1).
   The peroxisome and peroxisome proliferator-activated receptors (PPARs)
   signalling pathway was also found to be augmented and could play a role
   in redox signalling and lipid homoeostasis. Enrichment of identified
   S-sulfhydrated proteins was also observed in the TCA cycle, pyruvate
   metabolism and glycolysis ([141]Fig 4A). Most of the enriched pathways
   were metabolic pathways involved in energy production and conversion.
   Other pathways of lipid metabolism and amino acid metabolism were also
   observed to be elevated ([142]Fig 4B), indicating that H[2]S appears to
   mitigate atherosclerosis through regulation of lipid metabolism and
   anti-oxidative response to stimulus.

Fig 4. Pathway enrichment analyses for S-sulfhydrated proteins in mice aorta.

   [143]Fig 4
   [144]Open in a new tab

   (A) 19 KEGG pathways were found to be enriched, and the ratio of
   enrichment of each pathway was revealed in the identified protein set
   of S-sulfhydration with their corresponding P values. (B) The enriched
   KEGG pathways were categorised and presented as percentage that the
   largest group of the proteins were constituted by global metabolism.
   Lipid metabolism and amino acid metabolism were also major
   contributors, whereas a number of pathways were classified as ‘others’.

Protein S-sulfhydration of aortic proteins

   H[2]S signalling through S-sulfhydration involves H[2]S altering
   functions of proteins by modification on cysteine residues in proteins
   through the formation of a persulfide (–SSH) bond [[145]18]. A modified
   biotin switch method, a way to detect S-sulfhydrated proteins, was used
   to determine the pattern of S-sulfhydration of aortic proteins in
   control and H[2]S-treated mice. The results indicated that H[2]S
   treatment induced an increase in S-sulfhydration of aorta proteins as
   compared to the controls ([146]Fig 5A). The S-sulfhydrated proteins
   appeared as bands on western blotting from low-molecular-weight
   proteins to high-molecular-weight proteins ([147]Fig 5B), including the
   band corresponding to GPx1, which is involved in cellular response to
   stress during atherosclerosis. Western blot analysis of aorta lysate
   revealed a band corresponding to GPx1 at 22 kDa as a putative target
   for S-sulfhydration ([148]Fig 5C). This was confirmed by
   immunoprecipitation of the aorta proteins with anti-GPx1 antibody on
   both control and H[2]S-treated aortas. We observed that endogenous GPx1
   S-sulfhydration increased by 1.4-fold upon H[2]S treatment ([149]Fig
   5C), suggesting that H[2]S can S-sulfhydrate GPx1 proteins and that
   S-sulfhydration was enhanced after H[2]S treatment.

Fig 5. H[2]S effects on S-sulfhydrated proteins and GPx1 S-sulfhydration.

   [150]Fig 5
   [151]Open in a new tab

   (A) Gel staining of biotinylated proteins in control and H[2]S-treated
   ApoE KO mice. (B) Detection of S-sulfhydrated proteins by western blot
   analysis in control and H[2]S-treated ApoE KO mice. There was an
   increase of S-sulfhydration upon H[2]S treatment. (C) Representative
   blot image of immunoprecipitated GPx1 from control and H[2]S-treated
   mice aorta was extracted from the blot of first mice pair (control and
   H[2]S-treated mice). H[2]S sulfhydrated GPx1 overexpressed in the aorta
   lysate of H[2]S-treated mice and the band intensities were quantified
   using ImageJ. SH–GPx1, S-sulfhydrated GPx1; input-GPx1, total GPx1.
   Comparison between groups by Mann Whitney test, P < 0.05; n = 4 for
   each group.

S-sulfhydration reduced lipid peroxidation and enhanced GSH synthesis

   To confirm the effect of S-sulfhydration on aortic tissue, we examined
   the expression and activity of GPx1. GPx1 was abundantly expressed as
   green fluorescence signal in the aortic root of H[2]S-treated mice,
   whereas only minimal green fluorescence was observed in the control
   aortic sections ([152]Fig 6A). In addition, the plasma GPx1 activity
   also revealed a higher level in H[2]S-treated mice, that was about
   2.4-fold higher than the activity observed in control mice ([153]Fig
   6B). These results show that S-sulfhydration can enhance the level of
   GPx1 and thus its activity. It is known that GPx is an enzyme family
   with peroxidase activity whose main biological role is to protect the
   organism from oxidative damage. It converts toxic lipid peroxides into
   inert alcohols and reduces free hydrogen peroxide to water by using
   reduced GSH. Thus, we continued to examine the effect of H[2]S on MDA
   and GSH contents to determine lipid peroxides production and GSH
   synthesis. As shown in [154]Fig 6C, the level of MDA, the by-product of
   lipid peroxidation, was increased in the mice without H[2]S treatment
   but was significantly reduced by 43% after H[2]S administration
   ([155]Fig 6C), suggesting that H[2]S facilitates clearance of reactive
   oxygen species (ROS) such as free radicals and thus protects the cells
   from damage. In parallel, to further investigate the level of GSH in
   relation to H[2]S supplementation, plasma concentration of total GSH,
   reduced GSH (GSH) and oxidised GSH (GSSG) were measured in both groups
   of mice ([156]Fig 6D). The total GSH did not differ much in either
   group of mice, although H[2]S-treated mice had slightly higher total
   GSH than the control mice. Control mice had a higher GSSG level and
   lower GSH level, indicating accumulation of GSSG and exposure of
   oxidative stress in cells. However, this relationship was reversed by
   H[2]S treatment, as revealed by a decreased level of GSSG and an
   elevation in GSH level by about two times in H[2]S-treated mice in
   comparison to its level in control mice ([157]Fig 6D). Consistently,
   results showed that the GSH/GSSG ratio was increased by 2.3-fold in
   H[2]S-treated mice, as compared to that in the control group ([158]Fig
   6E), demonstrating protection of cells from oxidative damage upon H[2]S
   treatment.

Fig 6. S-sulfhydration enhances anti-oxidative stress responses.

   [159]Fig 6
   [160]Open in a new tab

   (A) Immunofluorescence staining of GPx1 in aortic root of ApoE KO mice
   with or without H[2]S treatment. It showed overexpression of GPx1 in
   aortic root lesions and intima of aortic root sections after 4 weeks of
   H[2]S administration. GPx1 appeared as green fluorescence, whereas cell
   nuclei are represented by blue fluorescence of DAPI. The images are
   shown at 10× magnification. The GPx1 signals in aortic roots were
   quantified by ImageJ software. (B) H[2]S treatment significantly
   increased plasma GPx activity (oxidation of NADPH to NADP^+),
   suggesting inhibition of oxidative damage. (C) Lipid peroxidation was
   determined by measurement of malondialdehyde (MDA) in plasma. MDA level
   was reduced in H[2]S-treated mice. (D) Plasma glutathione (GSH) assay
   significantly showed increase of total GSH and GSH but decreased GSSG,
   indicating the cells were protected from oxidative damage. (E) Higher
   ratio of GSH/GSSG in H[2]S treatment group also indicated antioxidant
   effects of H[2]S. GPx, glutathione peroxidase; GSH, glutathione; GSSG,
   oxidised glutathione. Values are means ± SEM, n = 6. The groups are
   significantly different from each other by Mann Whitney test, P < 0.05.

Discussion

   H[2]S administration with GYY4137 ameliorated atherosclerosis in the
   present study. GYY4137 is a slow-releasing donor of H[2]S. It releases
   H[2]S slowly both in vitro and in vivo and it just offers about 4–5
   nmol per 25 min [[161]13]. In this study, plasma H[2]S released from
   GYY4137 was increased at the first 30 minutes and then returned to
   normal after 8 hours ([162]S1 Fig). In addition, H[2]S was accumulated
   in liver and kidney very quickly upon GYY4137 treatment and maintained
   about 2 hours then started to decline to normal after 8 hours of
   injection, indicating no residue of H[2]S in liver and kidney ([163]S2
   Fig). All these showed that GYY4137 is a stable H[2]S donor and it has
   low toxicity which could be used to study regulatory role of H[2]S in
   animal model with acute diseases.

   H[2]S treatment was characterised by loss of body weight and reduction
   of plaque formation, both of which are consistent with our previous
   study using CSE overexpression transgenic mice as a model [[164]5], and
   also agrees with other studies [[165]4,[166]6]. We have identified 70
   proteins that are S-sulfhydrated to a significant degree from aortic
   tissue of ApoE KO mice. Functional annotation of the S-sulfhydration of
   aortic proteins reveals a main pathway of lipid metabolism in the
   development of atherosclerosis. Lipid metabolism is closely linked to
   the metabolism of carbohydrates, which may be converted to fats. The
   major aspect of lipid metabolism is involved with fatty acid oxidation
   to produce energy or the synthesis of lipids. The first step in lipid
   metabolism is the hydrolysis of the lipid in the cytoplasm to produce
   glycerol and fatty acids, which then further undergo glycolysis and
   β-oxidation of fatty acid, respectively. Here we show that H[2]S may
   regulate glycolysis and fatty acid metabolism through protein
   S-sulfhydration as revealed in our global analysis of S-sulfhydrated
   proteins in mouse aortic tissues. Pathways of glycolysis (e.g.,
   aldolase A, fructose-bisphosphate, ALDOA and phosphoglycerate mutase 1,
   PGAM1) and fatty acid metabolism/oxidation (e.g., ACADL and ACADM) were
   augmented in the over-representation analysis.

   Enhanced glycolysis and pentose phosphate pathway could result in
   increased antioxidant defence, as shown in the human osteosarcoma
   SAOS-2 cell line with expression of TAp63α [[167]19] and in
   vanadium-treated diabetic rats [[168]20] as a result of the generation
   of reduced GSH, a major scavenger of ROS during the process. A recent
   study also indicated that the CSE/H[2]S pathway plays an important role
   in the regulation of glucose production through S-sulfhydration of
   pyruvate carboxylase contributing to gluconeogenesis in the liver
   [[169]21]. In addition, an increase in fatty acid oxidation has been
   observed to reduce triglyceride content and inflammatory levels,
   improve insulin sensitivity in adipocytes and reduce endoplasmic
   reticulum stress and ROS damage in macrophages [[170]22]. Another fate
   of fatty acids after oxidation is the production of shortened fatty
   acids such as propanoate and acetate, which can then enter the
   circulation, where they transit directly to the liver and serve in
   energy production [[171]23]. These shortened fatty acids may play an
   important role in health and disease, in that they help to reduce the
   risk of inflammatory diseases, type 2 diabetes, obesity, heart disease
   and other conditions [[172]24]. Moreover, shortened fatty acids have
   also been shown to aid in control of body weight and insulin
   sensitivity [[173]25]. Therefore, augmented propanoate metabolism in
   the enrichment analysis suggests a lowering of cholesterol level
   [[174]26] and protection against diet-induced obesity [[175]27].

   Further, the PPAR signalling pathway was also elevated in the
   S-sulfhydrated aortic proteins we studied, and is also involved in
   lipid homeostasis. Fatty acid binding protein 4, adipocyte (FABP4) and
   fatty acid binding protein 5, epidermal (FABP5) were identified as
   S-sulfhydrated proteins in the PPAR pathway. These two proteins
   involved in lipid metabolism by binding to fatty acids through
   activation of PPARs, which comprise three subtypes: PPARα, PPARδ/β and
   PPARγ. These receptors are transcription factors that control the
   peroxisomal β-oxidation pathway of fatty acids through regulation of
   the acyl-CoA oxidase. Thus, PPARs play an important role in lipid
   metabolism. Cheng’s group demonstrated that PPARδ activated fatty acid
   oxidation in cultured cardiomyocytes by regulating key genes of fatty
   acid oxidation [[176]28]. PPARδ was shown to act as a regulator of fat
   burning and was identified as a potential target in the treatment of
   obesity and its associated disorders [[177]29]. In addition, PPAR was
   sulfhydrated after H[2]S treatment [[178]30] and its activation could
   improve lipid profiles and reduce adiposity [[179]31]. In line with
   this observation, recent evidence has shown that a PPARδ agonist
   inhibited aortic inflammation and attenuated the progression of
   pre-established atherosclerosis through reduction of plasma lipids,
   particularly very low-density and intermediate-density lipoproteins
   [[180]32]. Taken together, all these studies reveal that H[2]S is
   highly associated with reduced obesity by regulation of glucose and
   lipid homoeostasis, as supported by a recent study [[181]33] and a
   review [[182]34].

   In the enrichment analysis, antioxidant activity and peroxidase
   activity were shown to be enriched. Therefore, GPx1 can be considered
   as a putative target of S-sulfhydration in atherosclerosis. GPx1 is a
   selenium-dependent enzyme that is ubiquitously expressed and protects
   cells against oxidative damage by reducing hydrogen peroxide and a wide
   range of organic peroxides with reduced GSH. As an antioxidant enzyme,
   GPx1 has been used as a biomarker for cardiovascular risk
   [[183]35,[184]36]. It has also been proposed to be a therapeutic target
   for chronic obstructive pulmonary disease [[185]37]. In
   atherosclerosis, increased oxidative stress is a result of the
   generation of ROS and increased free-radical production, which may play
   an important role in the pathogenesis and progression of
   atherosclerotic plaques. Increased lipid peroxidation also appears to
   be associated with oxidative stress in atherosclerosis [[186]38] and
   diabetes [[187]39]. In the present study, H[2]S S-sulfhydrated GPx1 and
   stimulated its activity in cells and in plasma. With enhanced GPx1
   activity and enriched pathways such as GSH metabolism and glycolysis,
   GSH biosynthesis upregulation plays a critical role in the cellular
   defence against oxidative stress. GPx1 exerts its antioxidant effect to
   protect the cells from oxidative damage by reduction of lipid
   peroxidation and clearance of free radicals in the progression of
   atherosclerosis. Furthermore, enhanced GSH biosynthesis was found to be
   closely connected with increased fatty acid oxidation and synthesis
   [[188]40] as well as reduction of body weight, as shown in our previous
   study [[189]5] and another study [[190]41].

   In conclusion, the present study identified S-sulfhydrated proteins in
   several metabolic pathways. The S-sulfhydration level of altered
   proteins in aorta indicates significant protein modifications that may
   lead to structural or functional alterations. Protein S-sulfhydration
   in atherosclerosis targets particular enzymes participating in
   glycolysis, TCA cycle, and fatty acid oxidation, indicating that this
   post-translational modification may regulate metabolism and
   mitochondrial bioenergetics. Our results showed that H[2]S has such
   cardiovascular protective effects as prevention of body weight gain and
   reduction of atherosclerotic plaque formation through augmentation of
   lipid metabolism. H[2]S is also mediated by its antioxidant effect via
   GPx1 S-sulfhydration, which causes decreased lipid peroxidation and
   enhanced GSH biosynthesis. The present proteomics study provides a
   comprehensive analysis of H[2]S signalling via S-sufhydration in
   response to atherosclerosis.

Supporting information

   S1 Fig. Plasma concentration of H[2]S in mice administered GYY4137.

   Blood was collected at timed intervals after GYY4137 administration and
   plasma H[2]S concentration was assayed. Values are means ± SEM, n = 3.

   (TIF)
   [191]Click here for additional data file.^ (78.2KB, tif)
   S2 Fig. H[2]S concentrations in liver and kidney after GYY4137
   administration.

   Tissues were collected at timed intervals after administration and
   H[2]S concentration was assayed. Values are means ± SEM, n = 3.

   (TIF)
   [192]Click here for additional data file.^ (354.5KB, tif)
   S1 Table. A complete list of data set in this study.

   (PDF)
   [193]Click here for additional data file.^ (178.9KB, pdf)
   S2 Table. Food uptake per mice per day.

   (PDF)
   [194]Click here for additional data file.^ (169.4KB, pdf)

Data Availability

   All relevant data are within the paper and its Supporting Information
   files.

Funding Statement

   JYWL received the a Direct Grant [2015.1.085] from The Chinese
   University of Hong Kong. The funder had no role in study design, data
   collection and analysis, decision to publish, or preparation of the
   manuscript.

References