Cystathionine gamma-lyase overexpression inhibits cell proliferation via a H2S-dependent modulation of ERK1/2 phosphorylation and p21Cip/WAK-1.

Cystathionine gamma-lyase (CSE) is a key enzyme in the trans-sulfuration pathway. CSE uses L-cysteine as a substrate to produce hydrogen sulfide (H2S). The CSE/H2S system has been shown to play an important role in regulating cellular functions in different systems. In the present study, we used CSE stably overexpressed HEK-293 cells to explore the effect of the CSE/H2S system on cell growth and proliferation. The overexpression of CSE resulted in increases in CSE mRNA levels, CSE proteins, and intracellular H2S production rates, as well as the inhibition of cell proliferation and DNA synthesis. These effects were accompanied by a sustained ERK activation and up-regulation of the cyclin-dependent kinase inhibitor p21Cip/WAK-1. Blocking the action of ERK with U0126 inhibited the induction of p21Cip/WAK-1, suggesting that ERK activation functions upstream of p21Cip/WAK-1 activation to initiate the CSE overexpression-induced cell growth inhibition. The antiproliferative effect of CSE is likely mediated by endogenously produced H2S because the H2S scavenger methemoglobin (10 microm) significantly decreased the H2S production rate and reversed the antiproliferative effect afforded by CSE. Exogenous H2S (100 microm) also inhibited cell proliferation. However, the other CSE-catalyzed products, ammonium and pyruvate, failed to inhibit cell proliferation. Methemoglobin also abolished the inhibitory effect of exogenous H2S on cell proliferation. Moreover, exogenous H2S induced a sustained ERK and p21Cip/WAK-1 activation. These findings support the hypothesis that endogenously produced H2S may play a fundamental role in cell proliferation and survival.

The endogenous production of hydrogen sulfide (H 2 S) and its physiological functions, including membrane hyperpolarization and smooth muscle cell relaxation, place this gas in the family of gas transmitters, together with nitric oxide and carbon mon-oxide (1,2). Two pyridoxal-5Ј-phosphate-dependent enzymes, cystathionine ␤-synthase (EC 4.2.1.22) and cystathionine ␥-lyase (CSE) 1 (EC 4.4.1.1), are responsible for the endogenous production of H 2 S in mammalian tissues, which use L-cysteine as the main substrate (3)(4)(5). Cystathionine ␤-synthase is a predominant H 2 S-generating enzyme in the brain and nervous system (6), and CSE is mainly expressed in liver, kidney, and vascular smooth muscles (7,8). Cystathionine ␤-synthase and CSE are important for the metabolism of sulfur-containing amino acids (e.g. cystathionine), as well as the production of H 2 S, ammonium, and pyruvate from L-cysteine.
H 2 S inhibits cell proliferation (9) and induces cell death predominantly by an apoptotic mechanism in polymorphonuclear cells (10). H 2 S treatment has also been shown to lead to nasal lesions and olfactory epithelial necrosis (11). On the other hand, H 2 S induces serum-independent cell cycle entry in rat intestinal epithelial cells and increases the fraction of colonic mucosa cells in the S phase (12,13). However, little is known about the cellular consequences of an elevated CSE expression or about the associated increase in endogenously produced H 2 S. In the present study, we overexpressed the CSE gene using a highly effective expression system. The successful overexpression of CSE was confirmed by measuring the CSE protein contents, CSE mRNA expression level, and endogenous H 2 S production rate. Proliferation of the transfected HEK-293 cells was monitored. We determined whether the CSE overexpression-induced cellular changes were because of overproduced H 2 S. Finally, the effect of both the cyclin-dependent kinase (cdk) inhibitor p21 Cip/WAK-1 and the ERK/mitogen-activated protein kinase (MAPK) pathway on the CSE/H 2 S system was examined. Our findings support the hypothesis that endogenously produced H 2 S may play a fundamental role in cell proliferation and survival.

EXPERIMENTAL PROCEDURES
Cell Culture and Measurement of H 2 S Production-HEK-293 cells (American Type Culture Collection, Manassas, VA) were cultured at 37°C in a humidified incubator with 95% air and 5% CO 2 in minimal essential medium supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Invitrogen), 100 units of penicillin, and 100 g of streptomycin/ml. The cultured cells were subjected to gene transfection when they had grown to 70 -80% confluence. The H 2 S production rate in CSE stably transfected HEK-293 cells was measured as described previously (8).
Cloning of CSE cDNA and Stable Transfection-PCR was used to amplify the open reading frame of CSE (GenBank TM accession number AY032875) from rat vascular tissues using the primers 5Ј-CGTCCCA-GCATGCAGAAGAA-3Ј and 5Ј-CAGTTATTCAGAAGGTCTGGCCC-3Ј. * This study was supported in part by the Natural Sciences and Engineering Research Council of Canada. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  The amplified open reading frame of CSE was subcloned into TA cloning vector (PCR4-TOPO). Positive clone containing CSE open reading frame insert was sequenced to confirm the accuracy of the inserted CSE sequence.
The constructs containing CSE cDNA were cleaved and subcloned into the mammalian expression vector pIRES2-EGFP (Clontech), which contained the human cytomegalovirus immediate early promoter/enhancers and the SV40 poly(A) signal. For stable transfection, the constructs were linearized with KpnI (MBI Fermentas) and then subjected to phenol-chloroform extraction and ethanol precipitation. Linearized constructs were mixed with a FuGENE 6 transfection reagent (Roche Applied Science) in a ratio of 1 g to 3 l in 100 l of FBS-free minimal essential medium (14). After incubating for 45 min at room temperature (20 -22°C), the mixture was added to HEK-293 cells in 2 ml of FBS-free minimal essential medium. After 48 h of transfection, the cells were trypsinized, counted, and replated at 1 ϫ 10 5 cells/plate in 35-mm plates, which contained 500 g/ml G418 for antibiotic selection. Mock (empty vector) transfection was also performed. In the pIRES2-EGFP vector, the EGFP gene (which encodes the enhanced green fluorescent protein) was expressed separately from the gene of interest and was used as a transfection marker. Non-transfected HEK-293 cells were included as the negative control for antibiotic selection. After 5 weeks of the antibiotic selective culturing, survival-transfected cells were harvested and grown to establish the sublines, which were subsequently examined for the presence of the CSE gene. Stable transfectants were used at passage numbers Ͻ15. Wild-type HEK-293 cells were maintained in identical conditions but without selection antibiotic G418 treatment.
mRNA Collection and Reverse Transcription-Total cellular RNA from wild-type, mock, or CSE-transfected cells was harvested after the cells were plated for 48 h. Monolayers were rinsed twice with phosphate-buffered saline (pH 7.6), and RNA was collected using TRIzol reagent (Molecular Research Center, Inc., Cincinnati, OH). Contaminated DNA was removed using the DNA-free kit (Ambion, Austin, TX), and total RNA (2 g) was reverse transcribed into cDNA with avian myeloblastosis virus reverse transcriptase using random hexamer primers according to the manufacturer's protocol (Roche Applied Science). Controls without reverse transcriptase were used to check for genomic DNA contamination in each sample.
Real Time Quantitative PCR-Real time PCR was performed in an iCycler iQ apparatus (Bio-Rad) associated with the iCycler optical system software (version 3.1) using SYBR Green PCR Master Mix. All PCRs were performed in a 20-l volume using 96-well optical grade PCR plates and optical sealing tape. Negative controls for this experiment were samples without a template. The cycling conditions were 95°C for 90 s followed by 38 cycles of 95°C for 10 s and 60°C for 20 s. For quantification, the target gene was normalized to the internal standard gene ␤-actin. The primers of CSE (GenBank TM accession number AY032875) were 5Ј-AGCGATCACACCACAGACCAAG-3Ј (sense, position 432-453) and 5Ј-ATCAGCACCCAGAGCCAAAGG-3Ј (antisense, position 589 -609). These primers produced a product of 178 bp. The primers of ␤-actin (Ambion) produced a product of 295 bp. A standard curve was constructed with a series of dilutions of total RNA (Ambion) transcribed to cDNA using the protocol outlined above to confirm the same amplifying efficiency in the PCR. A standard melting curve analysis was performed using a thermal cycling profile that began at 95°C for 10 s, increased to 55°C for 15 s, and then ramped to 95°C in one-degree increments to confirm the absence of primer dimers. Product size was determined by running PCR products on a 1.8% agarose gel. Relative mRNA quantification was calculated by using the arithmatic formula "2 Ϫ⌬⌬CT " (16,17), where ⌬CT is the difference between the threshold cycle of a given target cDNA and an endogenous reference cDNA. Thus, this value yields the amount of the target normalized to an endogenous reference.
Measurement of Cellular Proliferation-Cell count was performed first to assay the cell growth curve in CSE-transfected HEK-293 cells. Briefly, an equal number of cells (2 ϫ 10 5 ) were seeded in 35-mm Petri dishes, and then cells were counted daily using a hemocytometer. The medium was changed every 3 days.
DNA synthesis was assessed by the level of [ 3 H]thymidine incorporation. An equal number of cells seeded in the 24-well plates (2 ϫ 10 4 cells/well) were cultured in growth medium for 48 h. They were washed three times with serum-free medium and then incubated in the same medium for 24 h for synchronization. After the cells were stimulated with 10% FBS in minimal essential medium for 6 h, 0.1 Ci/ml [ 3 H]thymidine was added to each well in growth medium for another 6 h. The medium was discarded, and the cells were washed three times with 0.5 ml of ice-cold phosphate-buffered saline containing 1 mM MgCl 2 and 1 mM CaCl 2 . The cells were precipitated with 0.5 ml of 5% ice-cold trichloroacetic acid for 10 min, and then the precipitates were lysed with 0.5 ml of 0.1% SDS, 0.1 N NaOH. Four-hundred microliters of cell solution were put into scintillation vials and mixed with 5 ml of scintillation fluid. Radioactivity was determined using a liquid scintillation counter (Beckman LS3801).
Reagents and Chemicals-H 2 S stock solution was freshly prepared by directly bubbling distilled water with pure H 2 S gas (Praxair) to make the saturated H 2 S solution (0.09 M at 30°C) (8). H 2 S stock solution was diluted to different concentrations in cell culture medium, and the pH of the medium was adjusted to 7.4.
The CSE antibody was kindly provided by Dr. N. Nishi (Kagawa Medical School). The MAPK antibody, ERK/MEK (mitogen-activated protein kinase/extracellular signal-regulated kinase kinase) inhibitor U0126, and p21 Cip/WAK-1 antibody were obtained from New England Biolabs (Camarillo, CA). Cyclin D1 antibody was from Lab Vision Corporation (Fremont, CA). [ 3 H]Thymidine was purchased from Amersham Biosciences. Methemoglobin, ammonium hydroxide, and sodium pyruvate were from Sigma. Horseradish peroxidase-conjugated goat anti-rabbit IgG antibody was from Bio-Rad.
Statistical Analysis-All data are expressed as means Ϯ S.E. and represent at least three independent experiments. Statistical comparisons were made using the Student's t test or one-way analysis of variance followed by a post hoc analysis (Tukey test) where applicable. The level of significance was set at p Ͻ 0.05.

Overexpression of CSE cDNA in HEK-293 Cells-HEK-293
cells were transfected with a CSE cDNA/pIRES2-EGFP construct or an identical empty vector lacking a cDNA insert as a control (mock). The transfection of HEK-293 cells with CSE cDNA or the control empty vector did not change the morphological characteristics of the cell. In HEK-293 cells transfected with CSE cDNA or the control empty vector, green fluorescence emitted by green fluorescent protein was observed. In contrast, green fluorescence was not detectable in wild-type HEK-293 cells (data not shown). The expression of CSE was verified by Western blot analysis, real time PCR analysis, and the H 2 S production rate. Western blot and real time PCR analysis revealed that transfection of CSE cDNA, but not of the mock cDNA, resulted in significant increases in both CSE protein expression and CSE mRNA levels compared with wild-type HEK-293 cells (Fig. 1, A and B). Transfection of HEK-293 cells with CSE cDNA also resulted in a marked increase in the H 2 S production rate (1.51 Ϯ 0.12 nmol/g/min) compared with mocktransfected cells (0.41 Ϯ 0.06 nmol/g/min, p Ͻ 0.05) (Fig. 1C).
Suppressed Proliferation and DNA Synthesis in CSE-transfected HEK-293 Cells-To determine whether overexpression of the CSE gene affects proliferation and DNA synthesis of HEK-293 cells, cell growth curves were constructed. Compared with mock-transfected or wild-type cells, HEK-293 cells transfected with CSE cDNA exhibited a significantly reduced cell growth rate on the second day. At the fourth day, the number of CSE-overexpressed cells was only 72.2 Ϯ 2.8% of the mocktransfected cells (Fig. 2A).
An antiproliferative effect of CSE overexpression was also evidenced by the extent of [ 3 H]thymidine incorporation. As shown in Fig. 2B, HEK-293 3A). Neither the phosphorylation of c-Jun NH 2 -terminal kinase nor total amounts of MAPK were affected. Incubating cells with 10 M U0126 or 20 M SB203580 (the p38 MAPK inhibitor) for 24 h significantly decreased the expression of phosphorylated ERK and p38 MAPK (Fig. 3B).
p21 Cip/WAF-1 is a cdk inhibitory protein, and it can prevent cell cycle progression by directly binding with cyclin and cdk complexes or binding to a proliferating cell nuclear antigen (18,19). Increased p21 Cip/WAF-1 expression is an important indicator for MAPK-dependent cell growth arrest (20 -22). As shown in Fig. 4, the expression of p21 Cip/WAF-1 increased significantly in CSE-transfected HEK-293 cells (ϳ3-fold that in mock-transfected or wild-type cells). To verify the correlation of MAPK activation with CSE overexpression-mediated p21 Cip/WAF-1 upregulation, we tested the effects of U0126 and SB203580 on the expression of p21 Cip/WAF-1 . After incubating cells with 10 M U0126, the expression of p21 Cip/WAF-1 in CSE-transfected HEK-293 cells decreased almost to the basal level, whereas SB203580 (20 M) had little effect. This suggests that p21 Cip/WAF-1 induction by CSE overexpression may result from activation of the ERK pathway.
Cyclin D1 is an activator of cdk4 and cdk6, which limit the rate of cell cycle reentry (23). p21 Cip/WAF-1 can inhibit cyclin D1 nuclear export by binding to Thr-286-phosphorylated cyclin D1, therefore preventing cyclin D1/CRM1 association (24). Our results showed that cyclin D1 expression was unchanged in all FIG. 1. Overexpression of CSE cDNA in HEK-293 cells. A, CSE protein expression was increased in CSE stably transfected cells. After being cultured for 48 h, the cells (3 ϫ 10 6 ) were lysed, and 30 g of protein were subjected to Western blot analysis using a monoclonal anti-CSE antibody. Results are representative of three individual experiments. B, the CSE mRNA level was higher in CSE stably transfected cells. The CSE mRNA level was normalized to that of an internal standard gene (␤-actin) using real time PCR. Experiments were repeated three times. *, p Ͻ 0.05 versus mock or wild-type cells. AU, arbitrary unit. C, the H 2 S production rate was significantly increased in CSE stably transfected cells. After being cultured for 48 h, the cells (3 ϫ 10 6 ) were collected and homogenized to measure the H 2 S production rate. Data represent four independent experiments. *, p Ͻ 0.05 versus mock or wild-type cells. kinds of cells, and U0126 and SB203580 had little effect on cyclin D1 expression (Fig. 4).
Inhibitory Effects of CSE Overexpression on Cell Proliferation Mainly Due to Endogenously Produced H 2 S-To determine the mechanism by which CSE overexpression inhibited cell growth, we assessed the effects of H 2 S, ammonium, and pyruvate, three products of CSE-catalyzed cysteine degradation, on cell growth. When confluent serum-starved wild-type HEK-293 cells were stimulated with 10% FBS in the presence of H 2 S (100 M), DNA synthesis decreased significantly. However, 100 M ammonium hydroxide and 100 M sodium pyruvate failed to inhibit DNA synthesis (Fig. 5A). Furthermore, we tested the effect of the H 2 S scavenger methemoglobin (1, 25) on CSE overexpression-mediated cell growth inhibition. As shown in Fig. 5B, methemoglobin at 10 M partly but significantly reversed the antiproliferative effect of CSE (p Ͻ 0.05). Pretreating wild-type HEK-293 cells with 10 M methemoglobin for 1 h prior to adding 100 M H 2 S significantly abolished the antiproliferative effect of H 2 S (Fig. 5A). Decreased H 2 S production in CSE-overexpressed cells by methemoglobin also provided evidence that methemoglobin scavenged the endogenous H 2 S (Fig.  5C). To assess the role of exogenous H 2 S on MAPK and cell cycle protein expression, we tested the status of ERK and p21 Cip/WAF-1 in wild-type HEK-293 cells after exposure to 100 M H 2 S. The expression of ERK and p21 Cip/WAF-1 increased after incubating wild-type HEK-293 cells with H 2 S for 2 h (p Ͻ 0.05) (Fig. 6), indicating that H 2 S likely mediates the antiproliferative effect of CSE.

DISCUSSION
Our present study demonstrated that the CSE expression level is an important regulatory element for cell growth. By increasing the endogenous production of H 2 S, the overexpression of CSE significantly inhibited cell proliferation. In this cascade of signaling transduction, ERK and p21 Cip/WAF-1 were consequentially activated by H 2 S, leading to cell growth inhibition.
CSE is a key enzyme of the trans-sulfuration pathway, which interconverts L-methionine and L-cysteine. It also uses L-cysteine as an alternative substrate to form H 2 S (4). As a pyridoxal phosphate-dependent enzyme, CSE is expressed in a range of mammalian cells and tissues, and it seems to be the main H 2 S-forming enzyme in the liver, kidney, and cardiovascular system (7,8). Deficiency of H 2 S-producing enzymes results in some disorders such as homocystinuria, which are characterized by mental retardation, skeletal abnormalities, increased urine homocysteine, increased risks of thromboembolism, and early onset of atherosclerosis (9,25,26).
Many previous studies on H 2 S focused on the toxicological profile rather than the physiological function of the gas. Recent studies, however, have demonstrated the biological functions of H 2 S, including hyperpolarization of cell membranes, relaxation of smooth muscle cells, and decreased neuronal excitability (1, 2, 6,8,27,28). Little is known about the modulatory effect of endogenous CSE/H 2 S on cell growth and proliferation. In the present study, we used CSE stably transfected HEK-293 cells to explore the effects of CSE overexpression on cell growth and proliferation. Our results indicate that the CSE overexpression increased intracellular H 2 S production, inhibition of cell proliferation, and DNA synthesis (Figs. 1 and 2). CSE induced the expression of cdk inhibitor p21 Cip/WAF-1 following sustained ERK activity, suggesting a cell cycle arrest. The antiproliferative effect of CSE is likely mediated via the release of H 2 S because the H 2 S scavenger methemoglobin (1,25) partly and significantly reversed the antiproliferative effect of CSE (Fig.  5B). Exogenous H 2 S alone also inhibited cell proliferation, evidenced by decreased [ 3 H]thymidine incorporation. The other CSE products, ammonium and pyruvate, failed to inhibit cell proliferation. Pretreatment of wild-type HEK-293 cells with methemoglobin for 1 h prior to the addition of H 2 S abolished the antiproliferative effect of H 2 S (Fig. 5A). Moreover, exogenous H 2 S induced sustained ERK and p21 Cip/WAF-1 activation (Fig. 6). These findings support the hypothesis that endogenously produced H 2 S plays an important role in cell proliferation and survival.
The mitogen-activated protein kinase cascade plays a crucial role in transducing extracellular signals into responses governing growth and differentiation. The ERK pathway is involved in the stimulation of cellular proliferation (29 -31), although MAPK-induced growth inhibition has also been reported (32,33). Here, we provide evidence that a sustained increase in ERK activity induced by CSE overexpression leads to inhibited cell growth (Fig. 3A). The seemingly conflicting results can partially be explained by the fact that the MAPK pathway plays roles in both progression and inhibition of cell proliferation (21,34). The final cellular response (i.e. cell cycle arrest or cellular proliferation) to activation of the ERK/MAPK pathway depends on the strength and duration of the MAPK signal. Transient or cyclical activation may contribute to cell cycle progression, whereas sustained high levels of ERK may lead to cell growth inhibition. Because active MAPK accumulates in the nucleus, it has been suggested that the duration and magnitude of MAPK activation will direct qualitative changes in gene expression, which in turn will determine whether a cell re-enters the cell cycle, undergoes cell cycle arrest, or remains quiescent. Although p38 MAPK was also increased in CSEtransfected cells, it may not be involved in CSE overexpressioninduced p21 Cip/WAF-1 up-regulation (Figs. 3B and 4) because the inhibition of p38 activity by SB203580 did not change the expression of p21 Cip/WAF-1 .
p21 Cip/WAF-1 is one of the cdk inhibitory proteins, and it plays an important role in growth arrest, cellular differentiation, DNA repair, cell senescence, and apoptosis (35,36). Increased p21 Cip/WAF-1 expression is also an important indicator for MAPKdependent cell growth arrest (20,21,37,38). Our data provide evidence that the inhibition of growth by CSE overexpression in HEK-293 cells was accompanied by an induction of the cdk inhibitor p21 Cip/WAF-1 , which was dependent on the ERK activation (Fig. 4). Correlation between p21 Cip/WAF-1 induction, prolonged ERK activation, and the inhibition of p21 Cip/WAF-1 by U0126 confirmed that p21 Cip/WAF-1 induction occurs after prolonged ERK activation. In the progression of the cell cycle, activation of cdks has been demonstrated (23). The kinase activity of cdks is negatively regulated by cdk inhibitors such as p21 Cip/WAF-1 . p21 Cip/WAF-1 has been shown to directly bind with cyclin, cdk complexes, and DNA polymerase ␦ cofactor, a proliferating cell nuclear antigen (18). Blocking the activation of ERK blocked the induction of p21 Cip/WAF-1 expression, suggesting that ERK activation is correlated with CSE overexpression-induced up-regulation of p21 Cip/WAF-1 and cell growth inhibition. Several studies indicate that sustained ERK activation allows for the induction of cyclin D1 (20,21). p21 Cip/WAF-1 can inhibit cyclin D1 activation and thereby prevent cyclin D1/CRM1 association (24). However, our study demonstrated that cyclin D1 was not involved in CSE overexpression-induced cell growth inhibition. Although ERK and p21 Cip/WAF-1 likely play roles in CSE overexpression-mediated cell growth inhibition, it is also possible that some other molecules that were not tested here are involved. Furthermore, the linkage between sustained ERK activation and expression of p21 Cip/WAF-1 needs to be identified.
In conclusion, this study provides evidence for the first time that CSE regulates cell proliferation through a H 2 S-mediated phosphorylation of ERK and p21 Cip/WAK-1 . Collectively, the ability of the CSE/H 2 S system to inhibit cell growth suggests that genetic approaches to manipulate CSE expression and H 2 S production may provide a novel therapeutic avenue in the treatment of CSE/H 2 S disorder-related diseases.