H2S and homocysteine control a novel feedback regulation of cystathionine beta synthase and cystathionine gamma lyase in cardiomyocytes

Hydrogen sulfide (H2S), a cardioprotective gas, is endogenously produced from homocysteine by cystathionine beta synthase (CBS) and cystathionine gamma lyase (CSE) enzymes. However, effect of H2S or homocysteine on CBS and CSE expression, and cross-talk between CBS and CSE are unclear. We hypothesize that homocysteine and H2S regulate CBS and CSE expressions in a dose dependent manner in cardiomyocytes, and CBS deficiency induces cardiac CSE expression. To test the hypothesis, we treated murine atrial HL1 cardiomyocytes with increasing doses of homocysteine or Na2S/GYY4137, a H2S donor, and measured the levels of CBS and CSE. We found that homocysteine upregulates CSE but downregulates CBS whereas Na2S/GYY4137 downregulates CSE but upregulates CBS in a dose-dependent manner. Moreover, the Na2S-treatment downregulates specificity protein-1 (SP1), an inducer for CSE, and upregulates miR-133a that targets SP1 and inhibits cardiomyocytes hypertrophy. Conversely, in the homocysteine-treated cardiomyocytes, CBS and miR-133a were downregulated and hypertrophy was induced. In vivo studies using CBS+/−, a model for hyperhomocysteinemia, and sibling CBS+/+ control mice revealed that deficiency of CBS upregulates cardiac CSE, plausibly by inducing SP1. In conclusion, we revealed a novel mechanism for H2S-mediated regulation of homocysteine metabolism in cardiomyocytes, and a negative feedback regulation between CBS and CSE in the heart.

. Dose-dependent effect of homocysteine or hydrogen sulfide donor Na 2 S or GYY4137 (H 2 S donor) on CBS and CSE levels in HL1 cardiomyocytes. (A) qPCR analyses of CBS and CSE after treatment with different dosages of homocysteine (Hcy), or (C) Na 2 S. (B) Western blot analyses of CBS and CSE after treatment with different doses of homocysteine (Hcy), or (D) Na 2 S. (E) Western blot analyses of CBS and CSE after treatment with different doses of GYY4137. Values are represented as mean ± SEM. "*" represents statistically significant value for CBS when compared to the untreated control, " §" represents statistically significant value for CSE when compared to the untreated control, and p < 0.05 is considered statistically significant. N = 3-5.
Scientific RepoRts | 7: 3639 | DOI: 10.1038/s41598-017-03776-9 Na 2 S ( Fig. 2A) corresponds to the decreased levels of CSE (Fig. 1D) in HL1 cardiomyocytes, suggesting that Na 2 S downregulates CSE directly by suppressing the activity of SP1 in cardiomyocytes. Conversely, we found a trend of increased SP1 activity at 25 µM dose of Hcy, which continued up to 75 µM and 100 µM doses (Fig. 2B), showing that HHcy induces SP1. To validate the SP1 activity on CSE promoter, we performed Chromatin immunoprecipitation (ChIP) assay using p-SP1 specific antibody and CSE forward and reverse primers. Our results demonstrated that higher (50 µM and 75 µM) doses of Hcy increased the copy number of CSE promoter with respect to the untreated cells (Fig. 2C). It suggests that activated p-SP1 recruitment to the CSE promoter was increased at higher doses of Hcy, which results in an increased CSE transcriptional activity. The input chromatin was used as a positive control and normal rabbit IgG isotype was used as a negative control for ChIP assay (Fig. 2C). To corroborate the effect of Hcy on SP1 activity, we performed electrophoretic mobility shift assay (EMSA) using CSE promoter having normal SP1 consensus binding sites (CSE WT) and the mutant SP1 consensus binding sequence (CSE MUT). In CSE MUT, the SP1 consensus binding sequence on CSE promoter was scrambled. We observed that with increasing doses of Hcy, there was higher CSE WT-SP1 complex formation and enhanced gel-shift with p-SP1 antibody (Fig. 2D(i) and (ii)) but there was neither complex formation nor gel-shift in case of CSE MUT (Fig. 2D(iii)) suggesting that HHcy upregulates CSE by inducing SP1 binding to the CSE promoter in cardiomyocytes. Altogether, our findings provide a concrete evidence that Hcy upregulates whereas Na 2 S downregulates CSE directly by inducing and suppressing, respectively, the SP1 activity on CSE promoter in cardiomyocytes.

MiR-133a targets CSE.
In silico analyses showed that CSE is a potential target for miR-133a (Fig. 3A).
We have reported that Na 2 S upregulates miR-133a in HHcy cardiomyocytes 17 . To determine whether miR-133a targets CSE, we performed luciferase reporter assay using normal 3′ untranslated region (WT CSE UTR) and mutant 3′UTR (Mut CSE UTR) of CSE. In Mut CSE UTR, the miR-133a binding sequence on 3′UTR of CSE was deleted. The HEK293 cells were treated with either miR-133a mimic (miR-133a) or scrambled miRNA (scm) Values are represented as mean ± SEM. "*" represents statistically significant when compared to the untreated control. p < 0.05 is considered statistically significant. N = 5. MiR-133a targets cystathionine gamma lyase (CSE). (A) (i-iii) Potential binding sites for miR-133a on 3′ UTR of CSE through in silico analysis. (B) Luciferase reporter assay showing the relative expression of luciferase activity (humanized luciferase, hLuc: humanized renilla luciferase, hRLuc) in miR-133a mimic-, or scrambled miRNA-treated HEK 293 cells co-transfected with either CSE 3′ UTR or mutant CSE 3′ UTR (miR-133a binding site is deleted in CSE 3′ UTR). hRLuc is an internal control for luciferase activity. Values are represented as mean ± SEM. p < 0.05 is considered statistically significant ( p = 0.04). N = 8. (C) miRNA-mRNA electrophoretic mobility shift assay showing binding of LNA miR-133a* probes with WT or mutant CSE 3′ UTR mRNA probes incubated with increasing concentration of 3′ UTRs (2 μM or 4 μM). (D) Quantification of miR-133a*-CSE band intensity in 'C' incubated with 2 μM or 4 μM of WT or mutant CSE promoter. Values are represented as mean ± SEM. p < 0.05 is considered statistically significant. N = 3. "*" represents statistically significant value for WT CSE at 4 μM when compared to 2 μM, " §" represents statistically significant value for mutant CSE when compared to the WT CSE at 2 μM.
with either WT CSE UTR or Mut CSE UTR, and relative luciferase activities were measured. We found that miR-133a decreased relative luciferase activity in WT CSE UTR with respect to scm, suggesting that miR-133a targets WT CSE UTR. Moreover, there was no change in luciferase activity after miR-133a mimic treatment to Mut CSE UTR (Fig. 3B) supporting that miR-133a targets CSE 3′UTR. To confirm that miR-133a target CSE, we performed RNA-EMSA (miRNA-mRNA interaction analysis by EMSA) using WT CSE UTR and Mut CSE UTR, a 22-mer RNA sequence for the 3′UTR corresponding to CSE, and miR-133a and anti-miR-133a probes. We performed the binding assays using the LNA miRNA-133a* sequence as the miRNA probe. In RNA-EMSA, miR-133a displays one band (Fig. 3C, lane 3). The incubation of miR-133a with an equimolar concentration of anti-miR-133a reduces the mobility of miR-133a band due to formation of miR-133a-anti-miR complex (Fig. 3C, lane 4). Similarly, 1 μM WT CSE UTR RNA displays single band (Fig. 3C, lane 5), however, incubation of different doses (2 μM and 4 μM) of WT CSE UTR RNA with miR-133a reduces electrophoretic mobility and form a second band of miR-133a-WT CSE UTR complex (Fig. 3C, lanes 6 and 7). This second band corresponds to the specific binding of the miR-133a with WT CSE UTR RNA (Fig. 3C, lane 6). Moreover, there was increased intensity of second band with increasing doses of CSE (Fig. 3C, lane 6 and 7, and 3D). The upper bands in lanes 6 and 7 represents the specific binding between miR-133a and WT CSE UTR RNA. To corroborate the specificity of miR-133a binding to CSE UTR, we also used 2 μM and 4 μM Mut CSE UTR, which had 7 mismatches with the target site of miR-133a. We observed no gel-shift in Mut CSE UTR suggesting an absence of miR-133a-Mut CSE UTR complex formation. Further, there was no second band for the two doses (2 μM and 4 μM) of Mut CSE UTR (Fig. 3C, lanes 8 and 9), demonstrating the specificity of miR-133a binding to CSE 3′UTR. Altogether, our results showed that miR-133a targets CSE.

Dose-dependent effects of H 2 S and Hcy on miR-133a levels in cardiomyocytes.
To determine the dose-dependent effects of Na 2 S on miR-133a levels, we treated HL1 cardiomyocytes with the same doses of Na 2 S that was used for determining the CSE levels (Fig. 1D). We found no significant change in the levels of miR-133a when the cardiomyocytes were treated with the below 25 µM dose of Na 2 S. However, miR-133a level was markedly increased when treated with above 50 µM dose of Na 2 S (Fig. 4A). These results further support that Na 2 S have an indirect role, via upregulation of miR-133a, to suppress CSE in cardiomyocytes. Since Na 2 S increased miR-133a in a dose-dependent manner (Fig. 4A), we determined whether Hcy has a dose-dependent effect on miR-133a levels in cardiomyocytes. For that, we measured the levels of miR-133a after treatment with increasing doses of Hcy (Fig. 4B). Our results demonstrated that above 25 µM dose of Hcy, the levels of miR-133a was significantly decreased (Fig. 4B). Based on these findings, we infer that elevated levels of Hcy attenuates miR-133a in cardiomyocytes in a dose-dependent manner. Overall, our results suggest that Hcy and H 2 S have an opposite effect on miR-133a levels in cardiomyocytes.
H 2 S mitigates homocysteine-mediated hypertrophy in cardiomyocytes. Since miR-133a is an anti-hypertrophic miRNA 18 , we sought to determine whether Na 2 S can attenuate Hcy-mediated hypertrophy in cardiomyocytes. For that, first we evaluated the dose-dependent effect of Hcy or Na 2 S on cardiomyocyte hypertrophy. We found that higher doses of Hcy induced cardiomyocyte hypertrophy but Na 2 S did not have effect on cardiomyocyte size ( Fig. 5A-C). Moreover, we also confirmed cardiomyocyte hypertrophy by using molecular markers for hypertrophy such as atrial natriuretic peptide (ANP) and beta-myosin heavy chain (β-MHC). These molecular markers-based findings demonstrated that 25 µM or higher doses of Hcy caused hypertrophy in cardiomyocytes whereas Na 2 S did not change cardiomyocytes size (Fig. 5D,E). To corroborate the effect of H 2 S on hypertrophy, we treated HL1 cardiomyocytes with an alternate H 2 S donor GYY4137 using the same doses and measured the levels of beta-myosin heavy chain (β-MHC), a molecular marker for cardiomyocyte hypertrophy, to U6 is an endogenous control. Values are represented as mean ± SEM. "*" represents significant value when compared to the untreated control. p < 0.05 is considered statistically significant. N = 5.
determine the cardiomyocyte hypertrophy. Our results showed that GYY4137 (same doses as Na 2 S, 5 µM-100 µM range) did not upregulates β-MHC, supporting that H 2 S donors do not induce cardiomyocyte hypertrophy. As a positive control, we also treated cardiomyocytes with increasing doses of Hcy, which showed upregulation of β-MHC at higher doses (above 25 µM) (Fig. 5F). These results provide a solid evidence that H 2 S have no effect on cardiomyocytes size in normal cardiomyocytes. However, it was unclear whether H 2 S will have same effect in HHcy cardiomyocytes. Therefore, we used the same molecular markers for hypertrophy and determined whether Na 2 S can mitigate Hcy-mediated cardiomyocyte hypertrophy. We treated HL1 cardiomyocytes with 100 µM of Hcy with or without 30 µM of Na 2 S, following our previously reported protocol 17 . Our results demonstrated that 100 µM of Hcy increased the protein (Fig. 6C) and cellular (Fig. 6A,B) levels of ANP but 30 µM of Na 2 S did not change the levels of ANP. Notably, Na 2 S blunted the upregulation of ANP in Hcy-treated cardiomyocytes ( Fig. 6A-C). We further evaluated the effect of Hcy and/or Na 2 S on hypertrophy of cardiomyocytes by measuring the surface area of cardiomyocytes (Fig. 7A,B), using a 3D surface plot for cardiomyocytes (Fig. 7C,D), and by Values are represented as mean ± SEM. "*" Represents statistically significant value for ANP when compared to the untreated control, and " §" represents statistically significant value for β-MHC when compared to the untreated control. p < 0.05 is considered statistically significant. Scale bar: 100 μm. N = 5.
determining expression and intensity of F-actin in cardiomyocytes (Fig. 7E,F). We found that elevated levels of Hcy induces cardiomyocyte hypertrophy, which is mitigated by Na 2 S. Altogether, these results demonstrated that HHcy upregulates hypertrophy of cardiomyocytes and although H 2 S donor Na 2 S may not have an effect on hypertrophy in normal cardiomyocytes, it attenuates Hcy-mediated hypertrophy of cardiomyocytes.

CBS deficiency upregulates CSE by inducing SP1 in the mouse heart. In vitro studies showed that
Hcy downregulates CBS but upregulates CSE (Fig. 1A,B) whereas H 2 S donor Na 2 S upregulates CBS but downregulates CSE (Fig. 1C,D), indicating that when CBS is upregulated CSE is downregulated or vice versa. To determine whether CBS has a direct effect on CSE, we used CBS deficient (CBS+/−) and a sibling control (CBS+/+) mice. First, we validated these mice by genotyping (Fig. 8A), and mRNA expression in the heart (Fig. 8B). Then we measured the protein levels of CBS and CSE in the heart. We found that cardiac levels of CSE was upregulated but CBS was downregulated in CBS+/− as compared to CBS+/+ mice (Fig. 8C). These results showed that CBS deficiency induces CSE in the heart. We also measured the activity of SP1, which upregulates CSE 12,13 , and found that SP1 activity (ratio of p-SP1: SP1) was increased in CBS+/− hearts (Fig. 8D). It suggests that CBS deficiency upregulates CSE plausibly by inducing SP1 activity in the CBS+/− hearts. We also determined cardiac hypertrophy in these mice by measuring heart to body weight ratio (Fig. 8E) and determining β-MHC levels in the heart (Fig. 8F). Our results support that CBS deficiency induces cardiac hypertrophy in mice (Fig. 8E,F). Overall, these findings revealed a negative feedback regulation of CBS and CSE in cardiomyocytes/hearts, which can be influenced by HHcy or H 2 S donor Na 2 S/GYY4137. Values are represented as mean ± SEM. N = 5. p < 0.05 is considered statistically significant. "*" Represents statistically significant value when compared to the untreated control, and " §" represents statistically significant value when compared to the HHcy + Na 2 S treated cells. Scale bar: 100 μm.

Discussion
Although the physiological effect of H 2 S on the heart [20][21][22] , and the underlying molecular mechanisms are documented 10, 23-25 , the differential effect of H 2 S or Hcy on CBS and CSE expressions, the cross-talk/feedback regulation of CBS and CSE, and the underlying molecular mechanism by which H 2 S regulates CSE are unclear. The present study fill-in the gap of knowledge in the field of H 2 S and Hcy biology by answering these questions. We reveal a novel negative feedback regulation of CBS and CSE, where deficiency of CBS upregulates CSE, plausibly by inducing SP1. The elevated levels of Hcy or H 2 S donor Na 2 S/GYY4137 influence this regulation. HHcy suppresses CBS that results in upregulated CSE. HHcy also downregulates anti-hypertrophic miR-133a causing cardiomyocyte hypertrophy. On contrary, H 2 S suppresses CSE directly by inhibiting SP1 and indirectly by inducing miR-133a that targets CSE, which results in CBS upregulation. H 2 S also mitigates Hcy-mediated hypertrophy of cardiomyocytes by increasing the levels of miR-133a (Fig. 8G).
Hcy is biosynthesized from methionine, an essential amino acid, by two enzymes, s-adenosyl methionine and s-adenosyl homocysteine that transfer methyl groups from methionine to convert it into Hcy. Hcy remethylates to methionine by methyl tetrahydrofolate enzyme and folic acid cofactor, or metabolizes to cysteine and H 2 S by CBS and CSE enzymes through transsulfuration pathway 8 . If remethylation or transsulfuration pathway is impaired, Hcy is accumulated and circulating levels of Hcy is elevated (HHcy). As per The American Heart Association advisory statement, the normal homocysteine level in the blood ranges from 5-15 µmol L −1 , however, more than 9 µmol L −1 is associated with mortality of patients 3 . The increasing doses of Hcy can be classified showing the negative feedback regulation of CBS and CSE, and impact of homocysteine (Hcy) or hydrogen sulfide (H 2 S) on this feedback regulation. Hcy downregulates CBS and miR-133a. Reduced levels of miR-133a induces cardiac hypertrophy, which is reflected by increased levels of atrial natriuretic peptide (ANP) and betamyosin heavy chain (β-MHC), the molecular markers for cardiac hypertrophy. H 2 S inhibits hypertrophy by inducing miR-133a. MiR-133a also targets SP1, an inducer of CSE. Therefore, H 2 S indirectly downregulates SP1 by upregulating miR-133a. H 2 S also inhibits CSE directly by suppressing the activity of SP1. into mild HHcy, where Hcy level in the blood ranges from 9-14.9 3, 26, 27 , intermediate HHcy which ranges from 31-100 µmol L −1 , and severe HHcy where Hcy levels are more than 100 µmol L −1 and it is associated with inborn error in Hcy metabolism 28 . Previous studies on Hcy-mediated cardiomyocyte hypertrophy focused on higher dose of Hcy 17,29,30 . However, the dose-dependent effect of Hcy on cardiomyocyte hypertrophy was unclear. In the present study, we used different doses of Hcy and demonstrate that even 5 µmol L −1 of Hcy can increase size of cardiomyocytes, and there is a linear relationship with increasing doses of Hcy with cardiomyocyte hypertrophy when the Hcy dose is higher than 25 µmol L −1 (Fig. 5C). Since Hcy is transsulfurated into H 2 S 8 , and H 2 S induces anti-hypertrophy miR-133a 17 , we also measured the effect of H 2 S on cardiomyocyte hypertrophy. Interestingly, increasing doses of H 2 S donor Na 2 S did not have an effect on cardiomyocyte size and there was no hypertrophy in Na 2 S-treated cardiomyocytes ( Fig. 5C and E). It suggests that H 2 S may not influence hypertrophic pathway in normal cardiomyocytes. However, in Hcy-treated cardiomyocytes, we observed that even low levels of H 2 S mitigates hypertrophy of cardiomyocytes (Figs 6 and 7) implying that H 2 S acts as anti-hypertrophy when cardiomyocyte hypertrophy is instigated by pathological insult such as HHcy. These findings are consistent with previous report demonstrating that increased level of Hcy causes cardiomyocyte hypertrophy 17,29 , H 2 S blunts Hcy-mediated cardiomyocyte hypertrophy 17 , and cardiac hypertrophy is suppressed by miR-133a 18,31 .
Hcy is involved in pathological cardiac remodeling [32][33][34][35] , and miR-133a prevents pathological remodeling by suppressing cardiac hypertrophy 18, 31 and fibrosis 16,36 . Although H 2 S upregulates miR-133a in cardiomyocytes 17 , the dose-dependent effect of Hcy or H 2 S on miR-133a is unclear. In the present study, we demonstrate that above 5 µM of Hcy, the increasing doses of Hcy is associated with decreasing levels of miR-133a in cardiomyocytes (Fig. 4B). On contrary, above 50 µM of Na 2 S, increasing doses of Na 2 S is associated with increasing levels of miR-133a in cardiomyocytes (Fig. 4A). These results elicit the dose-dependent effect of Hcy or H 2 S on miR-133a levels in cardiomyocytes, which may be important for future studies on Hcy or H 2 S-mediated cardiac remodeling. We have demonstrated that elevated levels of H 2 S increases miR-133a (Fig. 4A), whereas HHcy reduces miR-133a (Fig. 4B). Since miR-133a is anti-hypertrophic to the cardiomyocytes/heart 19, 37, 38 , we proposed that H 2 S upregulates miR-133a, whereas Hcy downregulates miR-133a, and differentially regulate cardiomyocytes hypertrophy (Fig. 8G). The effect of H 2 S on upregulation of miR-133a is not novel and is previously reported by our group 17 and others 39 . The present results support the fact that H 2 S-mediated upregulation of miR-133a reduces cardiomyocyte hypertrophy (Figs 5A-7F). Moreover, our group has shed light on the underlying molecular mechanism by which H 2 S mitigates Hcy-mediated downregulation of miR-133a in cardiomyocytes 17 . In the present study, we also observed that H 2 S blunts the effect of Hcy on cardiomyocytes hypertrophy (Figs 5A-7F).
Hcy is converted into H 2 S by CBS and CSE enzymes. CBS converts Hcy into cystathionine, which is then converted into cysteine by CSE that ultimately biosynthesize H 2 S 8 . However, whether Hcy or H 2 S has a dose-dependent effect on either CBS or CSE is unclear. Our results demonstrate that above 25 µM, the increasing doses of Hcy upregulates CSE and downregulates CBS (Fig. 1A,B). It suggests that HHcy suppresses CBS but induces CSE. On contrary, higher (25-75 µML −1 ) doses of H 2 S downregulates CSE but upregulates CBS (Fig. 1C,D). To our knowledge, this is the first report showing the yin-yang effect of Hcy versus H 2 S on CBS and CSE levels in cardiomyocytes.
H 2 S is a cardioprotective gaseous molecule 10,21,[40][41][42][43][44] , and is emerging as a novel therapeutic target for heart failure 45 (clinicaltrials.gov; No. NCT02180074). A wide range (15-300 µM) of physiological levels of H 2 S levels has been reported in vivo, perhaps due to variable detection methods 10 . In the present study, we used a range of 0-100 µML −1 , which is within the in vivo levels of H 2 S documented in the literature 10 . H 2 S is also emerged as a signaling molecule 24 , and biosynthesis of H 2 S from CBS and CSE enzymes is established 8, 10, 24 . However, whether H 2 S has any effect on CBS or CSE is poorly understood. In the present study, we reveal that lower levels (0-25 µM) or high level (100 µML −1 ) of Na 2 S may not have a significant effect on CBS or CSE levels in normal cardiomyocytes (Fig. 1C,D). The effective dose range for Na 2 S to have an impact on CBS or CSE expression in normal cardiomyocytes is between >25 µML −1 and <100 µML −1 (Fig. 1C,D). Although these results are encouraging, this concept needs further evaluation and validated at in vivo level.
H 2 S has an inhibitory effect on CSE at higher doses ( Fig. 1C-E), however, the underlying molecular mechanism is unclear. Here, we demonstrate that H 2 S downregulates CSE by directly suppressing SP1 activity ( Fig. 2A), which is an inducer of CSE [12][13][14] . On contrary, Hcy induces SP1 activity at higher doses by increasing its binding to CSE promoter (Fig. 2B-D). It is documented that exogenous treatment with H 2 S donor Na 2 S upregulates miRNAs 46, 47 , a non-coding regulatory RNA that has emerged as a therapeutic target for cardiovascular diseases 48 . Na 2 S treatment upregulates miR-21 that targets CSE 46 . The other miRNAs involved in regulation of CSE are miR-30 that directly inhibits CSE 47 , and miR-22 that inhibits SP1 49 . However, effect of miR-133a, the most abundant miRNA in the heart 50 , on CSE was unknown. In the present study, we reveal that miR-133a targets CSE (Fig. 3A-D). Notably, H 2 S may indirectly reduce CSE by upregulating miR-133a (Fig. 4A). Hcy has an opposite effect on miR-133a levels and increasing doses of Hcy downregulates miR-133a (Fig. 4B). Altogether, we show a novel mechanism for H 2 S-mediated regulation of CSE in cardiomyocytes, where it can directly inhibit CSE by suppressing SP1 and indirectly reduce CSE levels by upregulating miR-133a that targets CSE.
Although Hcy and H 2 S have opposite effects on CBS and CSE expression in cardiomyocytes (Fig. 1A-E), it was unclear whether CBS directly influences CSE expressions in the heart. Our results demonstrate that deficiency of CBS upregulates cardiac CSE in CBS+/− mice (Fig. 8A-C). Moreover, we reveal the underlying molecular mechanism for CBS-mediated regulation of CSE in the heart. Deficiency of CBS induces SP1 activity (Fig. 8D), an inducer for CSE that results in elevated levels of CSE in the heart (Fig. 8C). It is a novel mechanism for CBS-mediated regulation of CSE. Based on these findings, we propose a novel negative feedback regulation between CBS and CSE in the heart (Figs 1A-E and 8C). CBS deficient mice cannot transsulfurate Hcy into H 2 S that elevates Hcy levels. Lack of CBS induces cardiac hypertrophy (Fig. 8E,F); plausibly via HHcy, that reduces anti-hypertrophic miR-133a (Fig. 4B).
Scientific RepoRts | 7: 3639 | DOI:10.1038/s41598-017-03776-9 Cardiac hypertrophy is associated with pathological cardiac remodeling and involves several signaling pathways 51 . MiR-133a mitigates cardiac hypertrophy by targeting RhoA, a cardiac hypertrophy regulating protein, and cdc42, a kinase involved in hypertrophy 18 . We have reported that Hcy-treatment reduces the levels of miR-133a in cardiomyocytes and induces c-fos, an early marker for hypertrophy, in HL1 cardiomyocytes 17 . Hcy also regulates ERK pathway 30 and ATP7a 29 to control cardiac hypertrophy. In the present study, we demonstrate that Hcy-induced cardiomyocyte hypertrophy, reflected by the morphology of cardiomyocytes and molecular markers of hypertrophy, such as ANP and β-MHC, can be normalized by Na 2 S treatment (Figs 5A-7F).
H 2 S is a volatile gaseous molecule with a short half-life 10 whereas Hcy is a stable amino acid. Hence, there is a possibility that H 2 S levels are not maintained in the culture medium for 24-hour treatment period but Hcy levels maintained for 24-hour treatment period. Therefore, in the experimental group where HL1 cardiomyocytes were treated with both Na 2 S and Hcy, the effect of Na 2 S may be ephemeral but that of Hcy is prolonged. It is documented that even 30 minutes of H 2 S donor pre-treatment is able to mitigate phenylepinephrine-mediated hypertrophy in cardiomyocytes by upregulating anti-hypertrophic miR-133a 39 . MiR-133a transcription is regulated by myocyte enhancer factor-2C (Mef2c) 52 . We have reported that Hcy inactivates Mef2c by promoting the binding of Mef2c with HDAC1 whereas Na 2 S activates Mef2c by releasing Mef2c from HDAC1 in HL1 cardiomyocytes 17 . For this reason even the short-term presence of H 2 S may be adequate to induce miR-133a transcription in Hcy-treated cardiomyocytes, and high levels of miR-133a is able to inhibit cardiomyocyte hypertrophy even in the absence of H 2 S.
Overall, we demonstrate a novel yin-yang effect of H 2 S versus Hcy on CBS and CSE in the heart using in vitro and in vivo approaches. It was interesting to note that dose-dependent effect of Hcy and H 2 S are unique for CBS, CSE, and miR-133a, and the dose effect of H 2 S in normal cardiomyocytes and Hcy-treated cardiomyocytes are different. We also revealed a novel molecular mechanism for H 2 S-mediated inhibition of CSE in cardiomyocytes, where H 2 S directly inhibits SP1, an inducer for CSE, to suppress CSE and indirectly upregulates miR-133a that targets CSE. The upregulation of miR-133a by H 2 S also mitigates Hcy-mediated hypertrophy of cardiomyocytes. Our results from CBS+/− and CBS+/+ hearts demonstrate a negative feedback regulation mechanism between CBS and CSE in the heart. These findings are important for understanding the underlying molecular mechanism of Hcy-, or H 2 S-mediated cardiac remodeling, including hypertrophy, which could be crucial for developing a H 2 S-based therapeutic strategy.
These results need further validation at in vivo levels by treating CSE+/− mice with a H 2 S donor such as Na 2 S or GYY4137, and treating pathological hearts such as pressure-overload or volume overload hearts with a H 2 S donor. Since the effect of H 2 S concentration on cellular activity is poorly understood 24 and a high dose of H 2 S could be toxic to the cell, treatment with a slow H 2 S-releasing donor such as SG1002 can be a better approach for in vivo experiments to assess the effect of H 2 S on CBS, CSE, miR-133a, and cardiac remodeling in the pathological hearts. It is unclear that how much Hcy or H 2 S enters into the cardiomyocytes after treatment with Hcy or Na 2 S. Examining the levels of intracellular Hcy or H 2 S and correlating that with the expression of CSE, CBS, and miR-133a will provide more clarity to the cross-talk among these molecules. Even though we have elucidated the underlying mechanism for H 2 S-, and Hcy-mediated regulation of CBS, CSE, and miR-133a, further molecular studies are required to uncover the regulatory signaling mechanisms.

Materials and Methods
We procured CBS+/− male and female mice from the Jackson Laboratory (Bar Harbor, ME USA, Stock # 002853) and bred these mice in the animal facility of the University of Nebraska Medical Center to obtain CBS+/+ mice. Mice were kept in the animal facility in a room with temperature 22-24 °C, humidity 30-40% humidity, and 12 hours' light/dark cycle, and food and water were supplied to them ad libitum. For experiments, twelve-week male mice were used. All the experimental protocols on mice were approved by the Institutional Animal Care and Use Committee of the University of Nebraska Medical Center, and all methods were performed in accordance with the relevant guidelines and regulations.
Genotyping of CBS+/− mice. Genomic DNA was extracted from the ear punch tissue of mice at the age of four to six week and was amplified by polymerase chain reaction (PCR) using the protocol provided by the Jackson Laboratory. The forward primer sequence which is common for CBS+/+ and CBS+/− was 5′GATTGCTTGCCTCCCTACTG3′. The reverse primer sequence for CBS+/+ was 5′AGCCAACTTAGCCCTTACCC3′, and for CBS+/− was 5′CGTGCAATCCATCTTGTTCA3′, respectively. PCR product for CBS+/+ mice had one band at 308 bp whereas PCR product for CBS+/− had two bands at 308 bp and 450 bp (Fig. 8A).
Measurement of cardiomyocyte surface area. The surface area of a cardiomyocyte was measured using the Image J (NIH) software that generates a 3-dimentional (3D) surface plot on a phase contrast image of a cardiomyocyte. In 3D image, the shadow spikes were transformed into lookup table (LUT) which correspond to cardiomyocytes arbitrary height. Moreover, individual cardiomyocyte surface area was also measured by using Alexa Fluor ® 595 Phalloidin staining and the color intensity was quantified by the Image J software.
Electrophoretic mobility shift assay (EMSA). We treated HL1 cells separately with 0, 5, 25, 50, 75, or 100 μmL-1 of Hcy for 24 h in incomplete Claycomb medium. After that, cells were washed with ice-cold phosphate-buffered saline (PBS) and harvested using trypsin. They were pelleted in 1 ml ice-cold PBS by centrifugation at 1000 × g for 5 min at 4 °C, suspended in 500 μl of lysis buffer (50 mM KCl, 0.5% NP40, 25 mM HEPES PH 7.8, and 125 μm DTT supplemented with protease inhibitor cocktail; Sigma, USA), and kept for 15 min at 4 °C to rupture the cell membrane. The ruptured cells were then centrifuged at 10000 × g for 1 min at 4 °C to separate nuclei from the cytoplasmic fraction. The supernatant was removed and the pellet was suspended in 500 μl of washing buffer (50 mM KCl, 25 mM HEPES PH 7.8, 125 μm DTT supplemented with protease inhibitor cocktail), and centrifuged at 10000 × g for 1 min at 4 °C to collect the pellet that contains nuclear fraction. The nuclear pellet was suspended in 100 μl of nuclear extraction buffer (500 mM KCl, 25 mM HEPES pH 7.8, 50% glycerol and 125 μm DTT supplemented with protease inhibitor cocktail) and incubated at 4 °C for 15 min with constant rocking. The nuclear lysate obtained by this process was centrifuged at 1000 × g at 4 °C and the supernatant was collected as nuclear extract. We estimated the protein concentration of nuclear extract by BCA protein assay kit (Thermo Scientific Inc., USA) and nuclear protein was immediately processed for EMSA. We used 5 μg of nuclear protein and incubated it with mouse CSE promoter-specific SP1 response elements (WT 5′GAGGCGGGGC3′ and mutant 5′GATTCGGGGC3′ as per the previous report 54 for 30 min at room temperature. The EMSA probes used were; WT 5′GCCACTGGGAGGCGGGGCAGGAACGATC3′ and Mutant 5′GCCACTGGGATTCGGGGCAGGAACGATC3′ and its complementary oligonucleotides (IDT, USA). We used fluorescence-based EMSA Kit (cat # E33075, Thermo Scientific Inc., USA) and labelled the oligonucleotides with SYBR green EMSA nucleic acid gel stain. For CSE-SP1 complex retardation, we used p-SP1 antibody (cat # ab59257, Abcam, USA). Reaction mixtures were loaded on a 6% polyacrylamide gel and gel electrophoresis was performed in 0.5X TBE buffer (pH 8) at 200 V for 1 h. The SYBR stained gel was scanned in a Chemidoc (ChemiDoc, Image Lab 4.1, Bio-Rad Laboratories, USA), using SYBR green filter with UV trans-illumination.
Chromatin Immunoprecipitation (ChIP) assay. We treated HL1 cells separately with 0, 5, 25, 50, 75, or 100 μmL −1 of Hcy for 24 h in incomplete Claycomb medium. Cells were harvested using trypsin and then washed twice with ice-cold PBS followed by centrifugation at 1000 × g for 5 min at 4 °C. Cells were cross-linked by adding formaldehyde to a final concentration of 1% (v/v) to 1 ml cell suspension and incubated for 7 min at room temperature with gentle rotation. The cross-linked cells were prepared for ChIP assay using Zymo-Spin ChIP kit (cat # D5209, Zymo Research, USA) following the manufacturer's instructions. In brief, cells were resuspended in nuclear pellet preparation buffer containing protease inhibitors. Nuclear pellets were sonicated on ice to shear chromatin into 200 to 300 bp in size using a sonication (Bioruptor, USA) in chromatin shearing buffer. The sheared chromatin supernatant was diluted with chromatin dilution buffer and incubated with p-SP1 (cat # ab59257, Abcam, USA), or normal rabbit IgG polyclonal antibody for overnight at 4 °C with constant rotation. An aliquot of the sheared chromatin was set aside for use as an input control. The chromatin-antibody complexes are precipitated with Zymomag Protein A beads, and chromatins were eluted, reversed cross-linked, and processed for ChIP DNA purification. The target CSE promoter fragment containing SP1 binding sites was amplified by qPCR using mouse CSE-specific oligonucleotide primers. The primer sequences used were: forward 5′CGGTACCTCTGTGCCACTGGGAG3′ and reverse 5′GAAGCTTGAGTGCGAGGTGTTGCT3′ as per published report 55 . The amplification of CSE promoter copy number was quantified by qPCR and was normalized by using control cells without Hcy treatment. The input was used as a positive control for the ChIP assay.

Statistical Analysis.
To compare the mean for more than two groups, we used one-way analysis of variance (ANOVA). It was followed by Tukey test for pairwise comparison. To compare the difference of mean between two groups, we used Student's t-test. Values are expressed as mean ± SEM. A p-value less than 0.05 was considered statistically significant.