A novel enhancing mechanism for hydrogen sulfide-producing activity of cystathionine beta-synthase.

H2S is produced from cysteine by cystathionine beta-synthase (CBS) in the brain and functions as a neuromodulator. Although the production of H2S is regulated by Ca2+ and calmodulin in response to neuronal excitation, little is known about the molecular mechanism for the regulation in CBS activity. Here we show that four cysteine residues of CBS are involved in the regulation of its activity in the presence of Ca2+ and calmodulin. Sodium nitroprusside (SNP), a modifying agent for cysteine residues, enhances CBS activity, whereas N-ethylmaleimide, an alkylating agent for cysteine residues, completely abolished the effect of SNP. Site-directed mutagenesis of the 13 cysteine residues of CBS identified four cysteine residues that are involved in the regulation of CBS activity by SNP, and two of the four residues are involved in the regulation of the basal CBS activity. The enhancement of CBS activity by SNP is independent of nitric oxide production. In the presence of Staphylococcus aureus alpha-hemolysin, which permeabilizes the cell membrane, exogenously applied SNP enhances the activity of CBS in intact cells. The present study demonstrates a novel mechanism for the regulation of CBS activity and provides a possible therapeutic application of SNP for the diseases in which CBS activity is deficient.

H 2 S is produced from cysteine by cystathionine ␤-synthase (CBS) in the brain and functions as a neuromodulator. Although the production of H 2 S is regulated by Ca 2؉ and calmodulin in response to neuronal excitation, little is known about the molecular mechanism for the regulation in CBS activity. Here we show that four cysteine residues of CBS are involved in the regulation of its activity in the presence of Ca 2؉ and calmodulin. Sodium nitroprusside (SNP), a modifying agent for cysteine residues, enhances CBS activity, whereas N-ethylmaleimide, an alkylating agent for cysteine residues, completely abolished the effect of SNP. Site-directed mutagenesis of the 13 cysteine residues of CBS identified four cysteine residues that are involved in the regulation of CBS activity by SNP, and two of the four residues are involved in the regulation of the basal CBS activity. The enhancement of CBS activity by SNP is independent of nitric oxide production. In the presence of Staphylococcus aureus ␣-hemolysin, which permeabilizes the cell membrane, exogenously applied SNP enhances the activity of CBS in intact cells. The present study demonstrates a novel mechanism for the regulation of CBS activity and provides a possible therapeutic application of SNP for the diseases in which CBS activity is deficient.
Relatively high endogenous levels of H 2 S, which is well known toxic gas, have been found in the brains of rats, humans and bovine (1)(2)(3), suggesting that H 2 S may have a physiological function. Endogenous H 2 S in the brain is produced from L-cysteine by the pyridoxal 5Ј-phosphate-dependent enzyme, cystathionine ␤-synthase (CBS) 1 (4 -7). CBS is expressed in the brain, and a CBS activator, S-adenosylmethionine (AdoMet), enhances H 2 S production (6). These observations, together with our recent finding that endogenous H 2 S is under detectable levels in the brains of CBS knock-out mice, indicate that CBS is a major H 2 S-producing enzyme in the brain (7).
The production of H 2 S is regulated by a Ca 2ϩ -and calmodulin-mediated pathway that is activated in response to neuro-nal excitation (7). Physiological concentrations of H 2 S specifically potentiate the activity of the N-methyl-D-aspartate (NMDA) receptor, and hippocampal long term potentiation is altered in CBS knock-out mice (6,7). H 2 S can regulate the release of corticotropin-releasing hormone from the hypothalamus (8). In addition to the function in the brain H 2 S relaxes smooth muscle in synergy with nitric oxide (NO) by activating ATP-dependent potassium channels (9,10). Based upon these observations it has been proposed that H 2 S may function as a neuromodulator, as well as a smooth muscle relaxant (6,9).
Sodium nitroprusside (SNP), which is well known as an NO donor, activates guanylyl cyclase and relaxes smooth muscle (11,12). SNP has also been used to detect cysteine, cystine, acetophenone, and sulfite in a reaction known as the Legal and Boedeker reaction, as well as secondary amines (13)(14)(15). The reactivity of SNP on cysteine residues modifies biological activity of proteins in an NO-independent manner. For example, SNP modifies NMDA receptors, whereas other NO donors, S-nitroso-N-acetylpenicillamine (SNAP) and S-nitroso-L-glutathione do not (16). SNP evokes the release of calcitonin generelated peptide (CGRP) and substance P, but ferricyanide, a substance related structurally to SNP, also evokes the release of CGRP and substance P (17).
The enzymatic activity of CBS has two metabolic outcomes (2,18). In addition to the production of H 2 S CBS catalyzes the reaction with the substrate homocysteine to produce cystathionine (4,18). Both metabolic pathways are dependent on pyridoxal 5Ј-phosphate and regulated by AdoMet (6,19). Loss of CBS activity causes homocystinuria, an autosomal recessive disease characterized by mental retardation, skeletal abnormalities, and vascular disorders with severe thromboembolic complications (18). Over 50 different missense mutations in CBS gene have been found in patients of homocystinuria (20). Pyridoxine (vitamin B 6 ), the precursor of pyridoxal 5Ј-phosphate, has been tried for the therapy of homocystinuria (21). Some patients respond to pyridoxine, whereas the remaining patients are non-responsive (22). A development of a new substance that enhances the activity of CBS has been needed for therapy. The present study shows that SNP enhances the activity of CBS and suggests that SNP may be useful for diseases in which CBS activity is deficient.
Transient Expression of CBS Mutants-COS-7 cells were transferred into 100-mm dishes 12 h before transfection. Immediately before the transfection, Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum was changed to serum-reduced medium (OPTI-MEM; Invitrogen), and cells were transfected with 10 g of plasmid DNAs using 60 g of LipofectAMINE reagent (Invitrogen) in 6.4 ml of medium for 5 h. After 48 h the cells were washed with phosphatebuffered saline and lysed at 4°C for 30 min with 1 ml of lysis buffer (150 mM NaCl, 20 mM Tris/HCl (pH 7.5), 1 mM EDTA, 1 mM EGTA, 1% Nonidet P-40, 10% glycerol, 0.2 mM phenylmethylsulfonyl fluoride, and protease inhibitor mixture) (Roche Molecular Biochemicals). Cell lysates were centrifuged at 15,000 ϫ g for 10 min, and the supernatant was assayed for CBS activity. The relative amount of CBS was measured by Western blot analysis using an antibody against CBS following SDS-7.5% PAGE.
In Vitro Assays for H 2 S-or Cystathionine-producing Activity of CBS-Rat brains were homogenized with a Teflon potter in lysis buffer followed by centrifugation at 15,000 ϫ g for 10 min. The assay for the H 2 S-producing activity of CBS was performed as described previously (7). Briefly, 33 g of brain extracts or 3 g of COS-7 cell extracts was incubated at 37°C for 30 min in 100 l of a reaction mixture containing 50 mM Tris (pH 8.6), 2 mM pyridoxal 5Ј-phosphate, and 1 mM L-cysteine in the presence of 0.6 mM CaCl 2 and 9.6 M calmodulin, and the tubes were filled with N 2 gas. The reaction was terminated by the addition of an equal volume of 100% trichloroacetic acid. H 2 S was measured by gas chromatography (Shimadzu, GC-14B gas chromatograph).
To determine the cystathionine-producing activity of CBS (23), 66 g of brain extract was incubated at 37°C for 30 min in 200 l of a reaction mixture containing 250 mM Tris (pH 8.6), 2 mM pyridoxal 5Ј-phosphate, 30 mM L-serine, and 15 mM L-homocysteine in the presence or absence of 0.6 mM CaCl 2 and 9.6 M calmodulin. After the reaction was terminated by the addition of a mixture of water-saturated phenol and chloroform (1:1), cystathionine was extracted. The amounts of cystathionine were measured by HPLC as described previously (7).
Cell Suspensions and Permeabilization with ␣-Hemolysin-A published method was used for the permeabilization of cells (24). Briefly, cells were detached with trypsin/EDTA and equilibrated in suspension in a Ca 2ϩ -free basic salt solution (10 mM Hepes (pH 7.2), 150 mM NaCl, 5 mM KCl, 2 mM MgCl 2 , 11 mM glucose, 0.75% bovine serum albumin) for 1 h. Cells were permeabilized by the addition of Staphylococcus aureus ␣-hemolysin, with an effective concentration of 50 or 100 units/ 10 6 cells/100 l of Ca 2ϩ -free basic salt solution at 37°C for 5 min. The permeabilized cells (10 6 cells/100 l) were exposed to SNP at 37°C for 5 min in the presence or absence of 2 mM Ca 2ϩ and then sonicated after adding 2 l of 10 M NaOH to terminate the reaction. The lysates were used to measure endogenous H 2 S by gas chromatography (7), and for Western blot analysis using an antibody against CBS. Western blot analysis was performed according to the standard protocol described previously using horseradish peroxidase-conjugated anti-rabbit Ig antibody (Amersham Biosciences) as a secondary antibody and a chemiluminescence reagent (PerkinElmer Life Sciences) for detection (7). Protein concentrations were estimated by Coomassie dye binding (Bio-Rad) using bovine serum albumin as a standard. (25). Because Ca 2ϩ and calmodulin regulate CBS activity, the effect of NO on CBS activity in the presence of Ca 2ϩ and calmodulin was examined (7). It was asked whether the NO donors, SNP, SIN-1, and SNAP ( Fig. 1A) modify the H 2 Sproducing activity of CBS in cell lysates. Initially the effect of NO donors on CBS activity in the presence of Ca 2ϩ and calmodulin was examined in the standard non-reducing conditions (7). In the presence of Ca 2ϩ and calmodulin, 300 M SNP increases the production of H 2 S 3-fold (410 Ϯ 42 pmol H 2 S/mg protein/min, n ϭ 3) over control (139 Ϯ 7 pmol H 2 S/mg protein/ min, n ϭ 3) (Fig. 1B). In contrast, the enhancing effect of SNP is not observed in the absence of Ca 2ϩ and calmodulin. The other NO donors, SIN-1 and SNAP, do not have any effect on the H 2 S-producing activity of CBS, even though they release NO (data not shown) (Fig. 1B). Compounds related structurally to SNP, ferricyanide and ferrocyanide, also have no effect on CBS activity (Fig. 1, A and B). These observations suggest that only SNP enhances H 2 S-producing activity of CBS and that this effect of SNP is independent of NO.  ). B, SNP enhances H 2 S production. Thirty-three g of brain extracts were incubated with or without (Ϫ) 300 M SIN-1, SNAP, SNP, ferricyanide, or ferrocyanide in the presence (ϩ) or absence (Ϫ) of 0.6 mM Ca 2ϩ and 9.6 M calmodulin at 37°C for 30 min. C, SNP enhances the CBS activity in a dosedependent manner. Thirty-three g of brain extracts were incubated with varying concentrations of SNP (q), SIN-1 (E), or SNAP (Ⅺ) in the presence of Ca 2ϩ and calmodulin. H 2 S produced from L-cysteine was measured by a gas chromatography. All data represent the mean Ϯ S.E. of three experiments.

Enhancement of H 2 S Production from CBS
The dose-response curve of SNP was examined by measuring the production of H 2 S. The enhancing effect of SNP on H 2 S production is increased in a dose-dependent manner and reaches the maximum level at 300 M (Fig. 1C). The enhancing effect is decreased at concentrations greater than 300 M. Neither SIN-1 nor SNAP has any effect on CBS activity in the range of concentrations tested (Fig. 1C).
The enzymatic activity of CBS has two metabolic outcomes (2,18). In addition to the production of H 2 S from cysteine, CBS catalyzes the reaction with substrate homocysteine to produce cystathionine. To examine whether SNP also enhances the cystathionine-producing activity of CBS, the effect of SNP on the production of cystathionine was measured in the presence or absence of Ca 2ϩ and calmodulin. 300 M SNP enhances the production of cystathionine 4-fold more greatly in the presence of Ca 2ϩ and calmodulin (1400 Ϯ 70 pmol cystathionine/mg protein/min, n ϭ 3) than in the absence of Ca 2ϩ and calmodulin (260 Ϯ 10 pmol cystathionine/mg protein/min, n ϭ 3) ( Fig. 2A).
In the presence of Ca 2ϩ and calmodulin 300 M SNP enhances the production of cystathionine 2.4-fold more greatly than that without SNP (590 Ϯ 98 pmol cystathionine/mg protein/min, n ϭ 3). SIN-1, SNAP, ferricyanide, or ferrocyanide did not enhance the production of cystathionine (Fig. 2A). The doseresponse curve of SNP was also examined by measuring the production of cystathionine. The enhancing effect of SNP on cystathionine production is increased in a dose-dependent manner up to 800 M (Fig. 2B). Neither SIN-1 nor SNAP has any effect on CBS activity in the range of concentrations tested (Fig. 2B). These observations indicate that SNP enhances both the H 2 S-and cystathionine-producing activities of CBS, although the concentrations for the maximal responses are different between both pathways.
SNP Interacts with Cysteine Residues of CBS to Enhance Its Activity-Because the enhancement of CBS activity by SNP requires Ca 2ϩ and calmodulin, there are two possible targets for SNP. One possibility is that SNP interacts with calmodulin and modifies the CBS activity. Alternatively, SNP interacts with CBS and enhances its activity. To address this problem the effect of SNP on a mutant form of CBS (1-396), which lacks the calmodulin binding domain but is constitutively active without calmodulin, was examined (7). The activity of the mutant CBS (1-396) in the absence of Ca 2ϩ and calmodulin was enhanced further by the application of 300 M SNP (8565 ϩ 1245 pmol H 2 S/mg protein/min, n ϭ 3) to a level similar to that of the wild-type CBS in the presence of Ca 2ϩ and calmodulin (8227 Ϯ 50 pmol H 2 S/mg protein/min, n ϭ 3) (Fig. 3). These observations indicate that SNP does not interact with calmodulin to enhance CBS activity.
Because SNP reacts with the thiol groups of sulfur-containing amino acids, especially cysteine (14), it is possible that the enhancement of CBS activity by SNP is induced by modifying the cysteine residues of CBS. To examine this possibility the effect of NEM, which alkylates thiol of cysteine residues, on the enhancement of CBS activity by SNP was investigated by measuring the production of H 2 S in the presence of Ca 2ϩ and calmodulin. The enhancement of CBS activity by SNP (368.4 Ϯ 13.7 pmol H 2 S/mg protein/min, n ϭ 3) was suppressed by NEM in a dose-dependent manner, and 1 mM NEM completely suppressed the effect of SNP (119.7 Ϯ 2.5 pmol H 2 S/mg protein/ min, n ϭ 3) (Fig. 4, A and B). The basal CBS activity, however, was not changed by NEM (Fig. 4A). These observations suggest that NEM masks cysteine residues to hinder SNP from interacting with cysteine residues to suppress the enhancing effect of SNP on CBS activity.
Identification of Cysteine Residues Modified by SNP-To identify cysteine residues of CBS that interact with SNP, the effect of SNP on 13 CBS mutants was examined. Each of the 13 cysteine residues of CBS was replaced with serine residue by site-directed mutagenesis. The mutants were expressed in COS-7 cells, and the expression levels of each mutant were examined by Western blot analysis (Fig. 4C). There were no significant differences in the expression levels of each mutant, and endogenous CBS in COS-7 cells was under the detectable level (Fig. 4C). The enhancement of CBS activity by SNP was then examined by measuring the amounts of H 2 S produced by each CBS mutant. CBS activity enhanced by SNP was suppressed in C49S, C162S, C367S, and C476S mutants by 37, 33, 33, and 30% of the wild-type CBS, respectively (Fig. 4C). All the mutants, however, show similar basal H 2 S-producing activity to the wild-type CBS in the absence of SNP (Fig. 4C). These observations suggest that at least four cysteine residues in CBS are involved in the modification by SNP to enhance CBS activity.
SNP Enhances CBS Activity in Intact Cells in the Presence of ␣-Hemolysin-Loss of CBS activity causes homocystinuria (18). Pyridoxine (vitamin B6) has been used for therapeutic trials of homocystinuria and has improved some cases, but the remaining patients are not pyridoxine-responsive (21,22). Because   FIG. 2. SNP enhances the cystathionine-producing activity of CBS. A, SNP enhances the cystathionine-producing activity of CBS in the presence of Ca 2ϩ and calmodulin. Sixty-six g of brain extracts were incubated with or without (Ϫ) SNP, SNAP, SIN-1, ferricyanide, or ferrocyanide in the presence (ϩ) or absence (Ϫ) of 0.6 mM Ca 2ϩ and 9.6 M calmodulin at 37°C for 30 min. B, SNP enhances the production of cystathionine in a dose-dependent manner. Sixty-six g of brain extracts was incubated with varying concentrations of SNP (q), SIN-1 (E), or SNAP (Ⅺ) in the presence of Ca 2ϩ and calmodulin. Cystathionine produced from L-serine, and L-homocysteine was measured using HPLC. All data represent the mean Ϯ S.E. of three experiments.

Enhancement of H 2 S Production from CBS
SNP enhances CBS activity, SNP can possibly be used for therapeutic purposes. As an NO donor SNP can be applied outside of the cells, but as a cysteine residue-modifying agent SNP must enter into the cells to interact with CBS. To examine the effect of exogenously applied SNP on CBS activity in intact cells, H 2 S production from cells expressing CBS was measured. S. aureus ␣-hemolysin, which makes pores with ϳ2 nm diameter in the plasma membrane without the loss of intracellular proteins (24), was used to introduce SNP into the cells. H 2 S production was ϳ3-fold greater in cells incubated with SNP and 50 units of ␣-hemolysin (67.3 Ϯ 6.5 nmol/10 6 cells, n ϭ 3) greater than that in cells without the application of SNP (20.4 Ϯ 2.0 nmol/10 6 cells, n ϭ 3) (Fig. 5A). SNP enhances CBS activity most effectively at 1 M (Fig. 5A). 50 units of ␣-hemolysin alone also induced Ca 2ϩ influx and increased the production of H 2 S. This is probably because of the activation of CBS by increased intracellular Ca 2ϩ interacting with endogenous calmodulin (Fig. 5A) and is consistent with our previous finding that H 2 S production is increased in the presence of Ca 2ϩ ionophore A23187 (7). These observations show that SNP enhances Ca 2ϩ -dependent CBS activity in the presence of ␣-hemolysin in intact cells.
To confirm the results obtained in vitro that the four cysteine residues are involved in the regulation of CBS activity by SNP, the endogenous H 2 S levels of cells that transiently express the four CBS mutants were measured. In the presence of ␣-hemolysin the application of SNP to cells expressing CBS mutant C49S, C476S, C162S, or C367S decreased the endogenous H 2 S levels by 61 (26.5 Ϯ 5.0 nmol/10 6 cells, n ϭ 3), 69 (20.9 Ϯ 5.6 nmol/10 6 cells, n ϭ 3), 96 (2.8 Ϯ 0.5 nmol/10 6 cells), or 90% (6.4 Ϯ 0.2 nmol/10 6 cells) of the wild-type CBS (67.3 Ϯ 6.5 nmol/10 6 cells), respectively. These observations confirmed the result of the in vitro assay and indicate that the exogenously applied SNP enters into the cells in the presence of ␣-hemolysin and enhances the CBS activity by modifying the cysteine residues. The basal endogenous H 2 S levels in cells expressing C49S and C476S mutants were similar to those in cells expressing the wild-type CBS, whereas the basal H 2 S levels in cells expressing C162S and C367S mutants were much less than those in cells expressing the wild-type CBS (Fig. 5B).

DISCUSSION
The present study demonstrates that SNP enhances the H 2 S-producing activity of CBS by modifying cysteine residues of CBS via an NO-independent mechanism. This effect of SNP requires Ca 2ϩ and calmodulin, but SNP does not interact directly with calmodulin (Fig. 3). We proposed recently a possible mechanism for the potentiation of the H 2 S-producing activity of CBS by Ca 2ϩ and calmodulin that is similar to that proposed for the potentiation of CBS by AdoMet (7,26). In this model the carboxy-terminal domain covers the catalytic domain of CBS in the absence of Ca 2ϩ and calmodulin in the basal state. When Ca 2ϩ /calmodulin binds to the calmodulin binding consensus sequence, the catalytic domain is exposed by opening of the carboxy-terminal domain, and CBS becomes active (7). Taking this model into account, it is possible that SNP may not be able to access cysteine residues in the wild-type CBS in the absence of Ca 2ϩ and calmodulin. This model is also supported by our observation that the CBS mutant (1-396), which lacks the carboxy-terminal domain to hinder the access of SNP, is activated further by SNP (Fig. 3).
Under reducing conditions NO interacts with the heme group, and the CBS activity is reduced. However, the inhibition

Enhancement of H 2 S Production from CBS
of CBS activity by NO is not observed in standard non-reducing conditions that were used in the present study (see also Refs. 23 and 25). In addition to our observations, there are several examples in which SNP exerts its effects independently of NO. The release of CGRP and substance P from dorsal horn is evoked by SNP (17). The release is also induced by photoinactivated SNP, as well as ferricyanide, which is structurally similar to SNP except that it lacks the coordinated NO, suggesting that its three-dimensional structure is necessary to evoke the release of CGRP and substance P. The binding of MK-801, an agonist of NMDA receptors, to NMDA receptors is inhibited by SNP and ferrocyanide but not by other NO donors such as SNAP and S-nitroso-L-glutathione (16). In contrast to the above data, the enhancing effect of SNP on CBS activity is not induced by ferricyanide or ferrocyanide (Fig. 1). These observations suggest that the coordinated NO of SNP may bind to the cysteine residues of CBS, and the remaining part of SNP may change the three-dimensional structure of CBS and modify CBS activity. Although both NEM and SNP modify cysteine residues, the specificity and the effect of modification is apparently different. NEM alkylates all cysteine residues, whereas SNP specifically modifies in a functional manner only four cysteine residues of CBS, Cys-49, Cys-162, Cys-367, and Cys-476 (Fig. 4). SNP enhances CBS activity, whereas NEM does not have any effect on basal CBS activity.
Of the four cysteine residues modified by SNP, Cys-162 and Cys-367 in rat CBS correspond to Cys-165 and Cys-370 in human CBS, respectively, and both cysteine residues are changed to tyrosine by the mutations found in homocystineurea patients (20). Mutants Cys-49 and Cys-476 have basal CBS activity similar to the wild-type CBS, whereas mutants Cys-162 and Cys-367 have a much lower basal CBS activity (Fig. 5). The human C52S mutant shows a reduced basal cystathionineproducing activity (27), whereas the corresponding rat C49S has an H 2 S-producing activity similar to the wild-type CBS for H 2 S production (Fig. 4C). This discrepancy may be because the present study was performed in the presence of Ca 2ϩ and calmodulin, and H 2 S production was assayed rather than cystathionine production. In addition, to measure the amount of H 2 S, the enzyme assay was performed in anaerobic conditions in which the CBS activity is enhanced relative to the aerobic conditions (23).
The CBS activity is reduced severely in homocystinuria (21).
Although the two mutants at Cys-162 and Cys-367 do not respond to SNP, approximately 50 other mutants have the intact cysteine residues and can respond to SNP (20). This observation suggests that there may be an endogenous substance that interacts with Cys-162 or Cys-367 to regulate CBS activity and that both CBS mutants found in homocystinuria patients can not function properly because of a lack of the interaction with the possible endogenous regulator. Enhancement of CBS activity in vitro requires 300 M SNP, whereas in intact cells the maximal effect is obtained at 1 M (see Fig. 1C and Fig. 5A). The difference of the potency of SNP between in vitro and in intact cells may be because the in vitro experiments was performed in the presence of 1 mM free cysteine, which compete with cysteine residues of CBS, whereas there is much less endogenous cysteine in intact cells. Alternatively, there may be a synergy between SNP and the endogenous regulator to enhance the CBS activity in intact cells. The requirement of the much lower dosage in vivo, however, may be an advantage for therapeutic applications. Another possible example of the disease in which the CBS activity may be deficient is Alzheimer's disease (AD). Plasma levels of homocysteine, which is a substrate for CBS to produce cystathionine, is high in AD patients (28), and the level of AdoMet, a CBS activator, is low in AD brains (29,30). AdoMet improves cognitive decline in patients with AD (31). Despite these observations little attention has been paid to CBS activity in AD. We have shown recently that brain H 2 S levels and CBS activities are decreased severely in AD patients (30). The present observations may provide a new approach for the therapy for these diseases in which CBS activity is deficient.