Sodium hydrogen sulfite and sodium sulfite (NaHSO₃/Na₂SO₃, freshly mixed at 1:3 M ratio, pH 7.4) were used as SO₂ donors and purchased from Sigma (Zhang et al., 2021). Isoprenaline hydrochloride was purchased from Sigma (I5627). A chemically selective fluorescent probe SS-1 was provided by Professor Kun Li and Xiaoqi Yu. DAz-2 was used as a protein sulphenylation probe (13382, Cayman, Michigan, USA) to capture and enrich the sulphenylated protein. The primary antibodies in the present study included CypD (abcam, USA), AAT1 (Sigma, USA), AAT2 (Sigma, USA), β-actin (Zsbio, China), caspase9 (CST, USA), caspase3 (Beyotime, China), cleaved caspase3 (Beyotime, China), cytc (santa, USA), His (Zsbio, China), and β-tubulin (Zsbio, China). Human CypD wild type (WT), and C104S, C82S, C157S, and C203S mutant plasmids were constructed by Sangon Biotech. The information of three kinds of ATP plasmids used in this study are as follows: EcAT3.10 (Conley et al., 2017) was deposited at Addgene by Mathew Tantama (Addgene plasmid #107215); pm-iATPSnFR1.1 (Lobas et al., 2019) (Addgene plasmid #102549) and cyto-iATPSnFR1.0 (Lobas et al., 2019) (Addgene plasmid #102550) were deposited at Addgene by Baljit Khakh.

Culture of NMCMs: The kit for isolation of primary mouse cardiomyocytes was purchased from Thermo Fisher (Waltham, USA). In brief, the ventricular tissue parts from 1–2 days old C57BL/6J mouse neonates were subjected to modified enzymatic digestion. The enriched cardiomyocytes were cultured in DMEM with 1% penicillin-streptomycin solution (PS) and 10% fetal bovine serum (FBS). A cardiomyocyte growth supplement was added to inhibit the growth of the remaining fibroblasts.

NMCMs were divided randomly into control, ISO, ISO + SO₂, and SO₂ groups. For the control group, the cells were treated with the equivalent amount of saline for 48 h. For the ISO group, they were treated with 10 μmol/L ISO for 48 h. For the ISO + SO₂ group, the cells were given 100 μmol/L SO₂ donor for 30 min and 10 μmol/L ISO for 48 h (Ding et al., 2005). For the SO₂ group, cells were treated with 100 μmol/L SO₂ donor for 0.5 h and the equivalent amount of saline for 48 h.

H9c2 cell culture: H9c2 rat embryonic cardiomyocytes were purchased from the National Infrastructure of Cell Line Resource (China) and they were cultured in DMEM medium containing 10% FBS, 1% PS, and 1% glutathione in 5% CO₂ at 37°C. After a 6 h of synchronous treatment in serum-free medium, the cell treatment and grouping were similar to those in NMCMs, except that in H9c2 cells the ISO treatment was 200 μmol/L for 24 h (Deng et al., 2017).

SS-1 is a chemically selective fluorescent probe for in situ visualization of SO₂, developed and presented by Kun Li from Sichuan University, Sichuan, China (Yang et al., 2017). As mentioned in the previous study, SO₂ in NMCMs was tested with an SS-1 probe (Huang et al., 2021a). The cells were incubated with a 10 μM probe for 30 min at 37°C and then they were washed with PBS for removing the unlabeled probe. Subsequently, the cells were fixed with 4% paraformaldehyde solution. Finally, DAPI was added to stain nuclei. The green fluorescence detected by Olympus confocal laser scanning microscope (CK40, Olympus, Japan) is considered to be a positive signal.

AAT assay kit (Njjcbio, Nanjing, China) was used as described previously (Zhang et al., 2021). Firstly, cells were collected and lysed on ice with pre-cooled PBS for 30 min. The substrate solution and tested samples were added into the 96-well plate, and reacted at 37°C for 30 min. Then, 2,4-dinitrophenylhydrazine was added to each well, and the plate was placed at 37°C for 20 min. Finally, sodium hydroxide solution (200 μl, 0.4 mol/L) was added and gently mixed into each well. The plate was incubated at room temperature for 15 min. Finally, the optical density was measured at the absorption wavelength of 510 nm.

In situ cell death detection kit (Roche, Mannheim, Germany) was used as described previously (Du et al., 2018). Firstly, after gently rinsed with PBS, the cells were fixed in 4% paraformaldehyde for 1 h. Subsequently, they were permeabilized in 0.3% triton-X100 solution for 2 min. They were incubated with the TUNEL reaction mixture for 60 min at 37°C, preventing from light after washing with PBS. Finally, the nuclei were stained by a DAPI dye. The fluorescence images were captured with the help of a confocal laser scanning microscope (CK40, Olympus, Japan). The cell apoptosis was measured by using the percentage of TUNEL positive cells to the total DAPI positive cells (Du et al., 2018).

H9c2 cells were seeded and grown to 50% confluency, and then pmiATPSnFR1.1 and cyto-iATPSnFR1.0 were respectively transfected into H9c2 cells, which were replaced with a complete culture medium after a 6 h-transfection. The cells were imaged under a confocal laser scanning microscope (TCS SP8, Leica, Wetzlar, Germany) by the excitation at 473 nm and the emission at 525 nm.

The plasmid ecAT3.10 was transfected into 50% confluent H9c2 cells with a lipofectamine 3000 transfection reagent (Invitrogen, Carlsbad, CA, United States). After a 6 h-transfection, the freshly completed culture medium was replaced. The activity of the ecATeam sensor was measured by examining the fluorescence intensities in CFP, CFP-YFPFRET, and YFP channels as described in the previous study. The settings of the bandpass filter were identical to those previously reported (Conley et al., 2017). The ratio of CFP-YFPFRET to YFP is expressed as the activity of the ecATeam biosensor. The conformational change of the probe induced by ATP binding to the ecATeam probe increases the Förster resonance energy transfer (FRET) between the CFP donor and the YFP receptor (Imamura et al., 2009). Therefore, the higher the ratio of CFP-YFPFRET to CFP, the more extracellular ATP binding, which indirectly reflects the extracellular ATP content.

The enzymatic activities of caspase3 and caspase9 in H9c2 cells were detected with the commercial caspase3 and caspase9 activity kit, respectively (Applygen, Beijing, China) (Wang X. et al., 2019). Briefly, with the lysis buffer, the cells were lysed for 30 min at 4°C, and then centrifuged at 12,000 g for 5 min to harvest the supernatant. The concentration of protein was measured by the Bradford method. Then, 50 μg of cell lysate mixed with a reaction reagent was added to each well in a 96-well plate in order and then incubated for a period of 2 h at 37°C away from light. The activity of caspase3 or caspase9 was calculated from the absorbance value at 405 nm.

The mPTP opening was detected using the mPTP detection kit (Genmed, Shanghai, China) according to the instruction of the manufacture. The detection principle is that the fluorescence of calcein-AM is quenched when it leaks from mitochondria via the opening mPTP. Therefore, the strong green fluorescence represents the closed mPTP while the faint green fluorescence represents the opening mPTP. After completing the cell experiment, the cells were rinsed and incubated with the staining working solution for 20 min at 37°C in the dark. The remaining staining solution was removed by twice rinse with the cleaning solution. After the fixation with 4% paraformaldehyde, the fluorescence was observed by using a laser confocal microscopy (CK40, Olympus, Japan) with 488 nm excitation (Ex) and 505 nm emission (Em) settings (Song et al., 2018).

To determine the cytc subcellular localization, NMCMs were firstly incubated with the pre-warmed medium which contained 200 nmol/L MitoTracker (Life Technologies, USA) for 1 h. After the fixation with 4% paraformaldehyde, the cells were permeabilized in 0.1% Triton-X100 solution for 30 min, and then the cells were incubated with the cytc primary antibody at 4°C overnight. After the triplicate rinses with PBS buffer, an Alexa 594-conjugated secondary antibody (Life Technologies, USA) was added and then incubated for 90 min at 37°C in a dark container. Finally, images were captured with laser confocal microscopy (SP8-STED, Leica, Germany) (Wang X. et al., 2019).

Sulphenylation of CypD in the H9c2 cells and purified CypD protein was detected by the BSA method as previously reported (Huang et al., 2021a). The H9c2 cells were lysed in a non-denaturing lysis buffer (Applygen, Beijing, China) containing 5 mM DAz-2 for 20 min. The supernatant was collected by centrifuging at 16,000 g for 4 min at 4°C and gently shaken for 2.5 h at 37°C to label the DAZ-2. The DAz-2-labelled samples reacted with 250 μM p-biotin in a water bath for 2 h at 37°C. Then, the hyperlinked Neutral Affinity Protein™ was added at a volume ratio of 1–10 (Thermo Fisher Science) and incubated in a shaker for 4 h to capture the sulfenylated protein at 4°C. The sulphenylated proteins were mixed with a non-denaturing sample buffer and boiled for 10 min. Then, the supernatant was collected by centrifuging at 5,000 g for 10 min and it was subjected to western blot for the sulphenylation of CypD.

The human purified CypD protein (Abnova) was randomly divided into 3 groups: control, SO₂, and SO₂+DTT groups. The amount of purified protein used in each group was 0.1 μg. The protein was incubated for 2 h at 37°C. After the termination of the incubation experiment, the purified protein was divided into two portions to detect the sulphenylation of CypD and total CypD, respectively. The CypD sulphenylation was detected according to the abovementioned protocol in the cell experiment.

A sulphenylated protein cell-based detection kit (Cayman, USA) with a DAz-2-based fluorescence probe was used to visualize the sulphenylated proteins in cells (Song et al., 2020). The cells were treated with 100 μM SO₂ for 1 h or SO₂ plus 200 μM DTT for 15 min. The sulphenylation of CypD in H9c2 cells was observed by the co-localization of the fluorescent signals indicating CypD and sulphenylated proteins. Nuclei were stained with DAPI dye. The sulphenylated protein, CypD protein, and nuclei exhibited green, red, and blue fluorescence, respectively, with confocal laser-scanning microscope (Leica, Germany).

The pcDNA3.1 vector for human CypD-WT, CypD-C82S, CypD-C104S, CypD-C157S, and CypD-C203S mutant plasmids was constructed by Sangon Biotech (Shanghai, China). When H9c2 cells grew to approximately 50–60% confluence, the cells were treated with jetPEI TM reagent (Polyplus-transfection, France) before treatment.

The protein expression of AAT1, AAT2, cleaved caspase3, caspase3, cleaved caspase9, caspase9, CypD, and His in NMCMs and H9c2 cells was detected by a standard western blotting analysis as described in the previous study (Huang, et al., 2021a). Briefly, the cells were lysed using protein lysis buffer for 20 min at 4°C. The total protein was harvested and quantitated by the BCA method. The proteins were subjected to electrophoresis and they were subsequently transferred to nitrocellulose membranes. The protein bands were treated with primary antibodies at 4°C overnight, respectively. Then, at room temperature, they were incubated with the corresponding horseradish peroxidase-coupled secondary antibody for 1 h. Finally, the protein bands were incubated by using the enhanced chemiluminescent western blotting substrate kit (GE, Pittsburgh, PA, United States) in a FluorChem M MultiFluor System (Protein Simple, San Francisco, CA, United States).

A dataset containing a total of 658 SO₂-mediated sulfenylated vascular smooth muscle cell (VSMC) proteins was from Huang’s research (Huang et al., 2021a). The gene ontology cellular component analysis of the 658 SO₂-mediated sulfenylated VSMC proteins was performed by g: Profiler (https://biit.cs.ut.ee/gprofiler) (version e104_eg51_p15_3922dba). The parameters were set as follows: organism (Rattus norvegicus); statistical domain scope (only annotated genes); significance threshold (g:SCS threshold); and p values <0.05 indicated statistical significance.

The statistical analysis was conducted using SPSS 17.0 (IBM, USA) and Graphpad Prism 8.0 (GraphPad Software Inc., San Diego, CA, United States). The data are expressed as mean ± standard error. A comparison among multiple groups was performed by ANOVA followed by LSD post-doc analysis if the data were normally distributed, while followed by Dunnett T3 post-doc analysis if the data were not normally distributed. A statistical significance is set by p < 0.05.

ISO-stimulated NMCMs showed a marked decrease in SO₂ content as compared with the control group as evidenced by a significant decrease in SO₂-specific green fluorescence, and an obvious decrease in AAT1 and AAT2 protein expression, and AAT activity in the ISO-stimulated NMCMs (Figures 1A–C), accompanied by a significant increase in cell apoptosis, as evidenced by an increase in apoptotic cells, the cleaved-caspase3/caspase3 ratio, and caspase3 activity demonstrated by colorimetric assay (Figures 1D–F). The supplementation of SO₂ donors in ISO-stimulated NMCMs restored SO₂ content while significantly reduced the apoptotic cells, lowered the cleaved-caspase3/caspase3 ratio, and inhibited caspase3 activity (Figure 1). Therefore, the abovementioned data suggested that ISO stimulation downregulated the endogenous SO₂/AAT pathway and promoted cardiomyocyte apoptosis; whereas adequate endogenous SO₂ inhibited the apoptosis of cardiomyocytes.

We performed a GO-enriched cellular component (CC) reanalysis of the SO₂-affected redox proteomic dataset and showed that differential proteins regulated by SO₂ were significantly enriched in mitochondria (p = 2.829 × 10^–13) (Figure 2A). Furthermore, we compared the ATP fluorescence intensity in the cytoplasm, membrane, and extracellular matrix of H9c2 cells and found that ATP production was significantly reduced in H9c2 cells in the ISO group in comparison with the control group, and the supplementation with SO₂ significantly restored the ATP content (Figures 2B,C), suggesting that mitochondria might be an important target for the cytoprotective effect of SO₂ on cardiomyocytes.

The mPTP opening is an important event in the development of cardiomyocyte apoptosis under a variety of cardiac pathological conditions (Ahmad et al., 2019). Therefore, the present study analyzed the mitochondrial mPTP opening and downstream events. The results showed that ISO-stimulated NMCMs showed an increased mitochondrial mPTP opening, an increased cytc leakage from mitochondria, and an enhanced cleaved caspase9/caspase9 ratio and caspase9 activity in comparison with the control group; while SO₂ inhibited ISO-induced mPTP opening, suppressed the cytc release and reduced the ratio of cleaved caspase9/caspase9, thereby inhibiting apoptosis (Figures 2D–F). The results suggested that SO₂ significantly blocked ISO-opened mPTP, which might be involved in the mechanisms by which SO₂ inhibited cardiomyocyte apoptosis.

CypD, a peptidyl-prolylcis-trans isomerase (PPI) present in the mitochondrial matrix, is an important redox-regulation-dependent mPTP regulator (Lam et al., 2015). Therefore, we tested if CypD was a candidate target of SO₂ to explore the possible mechanism by which SO₂ blocked mitochondrial mPTP opening.

Firstly, there were no significant effects of SO₂ on the protein level of CypD either in NMCMs or in H9c2 cells as shown in Figures 3A,B (p > 0.05). However, the sulphenylation of CypD in human purified CypD protein in the SO₂ group was markedly higher than that of the control group, which was blocked by a sulfhydryl reducing agent DTT treatment (p all<0.05) (Figure 3C).

Furthermore, the quantitative analysis and in situ visualization of CypD sulphenylation in the H9c2 cells were performed. The results revealed that in comparison with the control, the sulphenylation of CypD was increased in H9c2 cells of the SO₂ group, which was also successfully blocked by DTT (p all<0.05) (Figure 3D). The multi-color confocal images showed that the co-localization of the fluorescent signals indicating CypD and sulphenylated proteins was strong in H9c2 cells of the SO₂ group but weak in that of the control group and SO₂+DTT group (Figure 3E). Correspondingly, the mPTP opening and cell apoptosis were inhibited in the cells of the ISO + SO₂ group in comparison with the ISO group, while DTT treatment reversed the above protective effect of SO₂ (Figures 3F,H).

The above facts indicate that SO₂ can directly sulphenylate CypD protein, which might be associated with SO₂-inhibited mPTP opening and cell apoptosis.

The human purified CypD protein has four cysteine sites: Cys82, Cys104, Cys157, and Cys203 (Linard et al., 2009). Homology analysis of CypD protein sequences among diverse species including human, goat, rat, mouse, pig, rabbit, chick and zebrafish showed that these four cysteine sites are all highly conserved among these different species (Figure 4A). Therefore, we mutated the four cysteine sites Cys82, Cys104, Cys157, and Cys203 to serine in the CypD protein, respectively, to clarify the precise site on which SO₂ affected CypD and then inhibited the cardiomyocyte apoptosis. The sulphenylation site screening results showed that the sulphenylation of CypD-His could be induced by SO₂ treatment in the H9c2 cells transfected with CypD-WT, C82S, C157S, and C203S plasmids, except in the cells transfected with CypD C104S plasmid (Figure 4B). The results suggested that the Cys104 in the CypD protein might be the sulphenylation site of SO₂.

Furthermore, the mPTP opening and the apoptosis of H9c2 cells was examined to identify if the CypD Cys104 might be the target of SO₂ affecting the cardiomyocyte apoptosis. The results showed that in the H9c2 cells transfected with CypD-WT plasmid, the mPTP opening in the ISO-stimulated cells was increased and the percentage of apoptotic cells was decreased by the SO₂ treatment, while DTT could reverse the effect of SO₂. In contrast, in the H9c2 cells transfected with CypD-C104S plasmid, SO₂ treatment failed to affect the ISO-stimulated mPTP opening and the apoptosis (Figures 4C,D). The result further confirmed that SO₂ sulphenylated CypD cysteine at Cys104, thereby inhibiting cardiomyocyte apoptosis.