Abstract
Although hydrogen sulfide (H2S) is generally thought of in terms of a poisonous gas, it is endogenously produced in the brain. Physiological concentrations of H2S selectively enhance NMDA receptor-mediated responses and alter the induction of hippocampal long-term potentiation (LTP). Here we use cystathionine β-synthase (CBS) knock-out mice to clearly show that CBS produces endogenous H2S in the brain and that H2S production is greatly enhanced by the excitatory neurotransmitterl-glutamate, as well as by electrical stimulation. This increased CBS activity is regulated by a pathway involving Ca2+/calmodulin. In addition, LTP is altered in CBS knock-out mice. These observations suggest that H2S is produced by CBS in response to neuronal excitation and that it may regulate some aspects of synaptic activity.
Hydrogen sulfide (H2S) is a well known toxic gas, and most studies about H2S have been devoted to its toxic effects (Reiffenstein et al., 1992). However, relatively high endogenous levels of H2S have been measured in the brains of rats, humans, and bovine (Goodwin et al., 1989; Warenycia et al., 1989;Savage and Gould, 1990), suggesting that H2S may have a physiological function. Endogenous H2S can be formed from l-cysteine by pyridoxal-5′-phosphate-dependent enzymes, including cystathionine β-synthase (CBS) (Stipanuk and Beck 1982; Griffith, 1987; Swaroop et al., 1992). CBS is expressed in the brain, and the CBS inhibitors hydroxylamine and amino-oxyacetate suppress the production of H2S, whereas the CBS activatorS-adenosyl-l-methionine (SAM) enhances H2S production. Physiological concentrations of H2S specifically potentiate the activity of NMDA receptor and alter the induction of long-term potentiation (LTP) in the hippocampus (Abe and Kimura, 1996). cAMP-mediated pathways may be involved in the modulation of NMDA receptor by H2S (Kimura, 2000). H2S can also regulate the release of corticotropin-releasing hormone from the hypothalamus (Russo et al., 2000). Based on these observations, it has been proposed that CBS can produce H2S in the brain and that H2S may function as a neuromodulator (Abe and Kimura, 1996).
Two other gases, nitric oxide (NO) and carbon monoxide (CO), are endogenously produced by enzymes localized in the brain (Maines, 1988;Palmer et al., 1988; Verma et al., 1993). Both NO and CO have been proposed as retrograde messengers in hippocampal LTP, a synaptic model of learning and memory (O'Dell et al., 1991; Schuman and Madison, 1991; Haley et al., 1992; Bliss and Collingridge, 1993;Stevens and Wang, 1993; Zhuo et al., 1993). The activities of NO synthase are regulated by Ca2+/calmodulin, and NO is released when NMDA receptors are activated byl-glutamate (Garthwaite et al., 1988; Bredt and Snyder, 1990). The regulation of CO production by neuronal excitation is not understood (Baranano et al., 2001).
CBS knock-out mice have been established (Watanabe et al., 1995). Animals homozygous for the CBS mutant gene are born at the expected frequency from matings of heterozygotes, but a majority of them die within 5 weeks after birth. They have less body weight than the wild-type mice, but the weight and morphology of the brain is normal (Watanabe et al., 1995; our unpublished observation). We used the CBS knock-out mice to show that CBS produces the endogenous H2S in the brain. We also found a novel regulation for H2S production by Ca2+/calmodulin and determined the 19 amino acid calmodulin binding domain in CBS. In addition, it is shown that l-glutamate, as well as electrical stimulation, enhances the production of H2S from brain slices and that LTP is altered in CBS knock-out mice. These observations suggest that endogenous H2S is produced when CBS is activated by the Ca2+ influx, which occurs with neuronal excitation, and that H2S may function as a neuromodulator or transmitter (Baranano et al., 2001).
MATERIALS AND METHODS
Determination of genotype by PCR. CBS knock-out heterozygous mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Exons 3 and 4 were deleted and exchanged with the neomycin-resistant gene in knock-out mice (Watanabe et al., 1995). Genomic DNA was isolated from mouse tails and amplified by PCR with three primers: 5′-CGG ATG ACC TGC ATT CAT CT-3′; 5′-GAA GTG GAG CTA TCA GAG CA-3′; and 5′-GAG GTC GAC GGT ATC GAT A-3′.
Purification of CBS from brain homogenates. For the measurement of H2S production, CBS was purified by calmodulin Sepharose 4B from brain homogenates. Brain homogenates in 3 vol of Tris-buffered saline (TBS) [50 mmTris-HCl, pH 7.4, 150 mm NaCl, 10% glycerol, 0.2 mm PMSF, 1 mm EGTA, and protease inhibitor cocktail (Roche, Basel, Switzerland)] were centrifuged at 10,000 × g for 60 min at 4°C, and the supernatant was recovered. Immediately before applying to calmodulin Sepharose column, CaCl2 was added to the final concentration of 4 mm. After washing with five bed volumes of 1 mm CaCl2containing TBS, CBS was eluted with 2 mm EGTA containing TBS. The resultant eluent was dialyzed with TBS. The protein concentrations of the eluent were estimated by Protein Assay (Bio-Rad, Hercules, CA).
Brain cell suspensions. Brain cell suspensions were prepared from the brain of 8-d-old mice by a modified method described byGarthwaite et al. (1988). The fresh tissues were minced and digested with 5 ml of 0.25 mg/ml trypsin in Ca2+/Mg2+-free basic salt solution [10 mm HEPES, pH7.2, 150 mm NaCl, 5 mm KCl, 11 mm glucose, and 0.75% BSA (type III; Sigma, St. Louis, MO)] at 37°C for 30 min. The reaction was terminated by adding 5 ml of Ca2+/Mg2+-free basic salt solution containing 40 μg/ml deoxyribonuclease I and 0.25 mg/ml soy bean trypsin inhibitor (Sigma). After washing, cells were resuspended at the density of 105cells/ml. After 1 hr of preincubation at 37°C in air, 100 μl of cell suspensions were transferred to a 1.5 ml microtube, and agonists or antagonists were applied. Stimulation was terminated by adding 2 μl of 10 m NaOH.
Measurement of H2S. The amounts of endogenous H2S in the brain and H2S produced by cell suspensions were measured by a gas chromatograph (Hoshika and Iida, 1977) (GC-14B; Shimazu, Kyoto, Japan). Briefly, 100 μl of homogenates consisting of 1 vol of brain and 3 vol of 10 mm NaOH in a 1.5 ml microtube were filled with N2 gas and sealed with parafilm (American National Can, Chicago, IL). H2S gas was released by adding 100 μl of 100% trichloro acetic acid to the tube with a 1 ml syringe and then incubated at 37°C for 30 min. Three hundred microliters of gas were removed from the reaction tube and applied to a gas chromatograph.
H2S produced by purified CBS was measured as follows: 100 μl of 50 mm Tris, pH 8.6, 2 mmpyridoxal 5′-phosphate, and 1 mml-cysteine, with 4.6 μg of total protein of purified CBS was incubated at 37°C for 30 min. Concentrations of Ca2+ in the reaction mixture were determined by an ion meter (F-23; Horiba, Kyoto, Japan). The procedures to measure the amounts of H2S released were the same as above. The quantitation of H2S was done using NaHS as a standard.
Measurement of free l -cysteine in the brain. The amounts of l-cysteine were measured by using a reverse-phase HPLC with fluorescence detection (Waters 2690 separation module and 474 scanning fluorescence detector; Waters, Milford, MA). The brain extracts were boiled and extracted with phenol–chloroform and then centrifuged at 15,000 × gfor 10 min. The supernatant was labeled by AccQ-Tag system (Waters) and applied to the HPLC. Quantitation was done with an external standard ofl-cysteine.
The electrical stimulation of brain slices and the induction of LTP. For the study of H2S production induced by electrical stimulation, slices of cerebral cortices (300 μm) were prepared from the 4-week-old mice and maintained in a chamber at 30°C in artificial CSF (ACSF) containing (in mm): 119 NaCl, 2.5 KCl, 2.5 CaCl2, 1.3 MgSO4, 1 NaH2PO4, 26.2 NaHCO3, and 11 glucose (bubbled with 95% O2–5% CO2). A bipolar stimulating electrode was placed at the white matter, and four 100 Hz pulses (100 μsec duration) at 200 msec intervals and 3 V of a stimulus intensity were applied for 30 sec or 1 min. After stimulation, each slice was transferred into a 1.5 ml microtube, and the amount of H2S was measured.
For LTP experiments, hippocampal slices (400 μm) were prepared from the 12- to 16-d-old CBS knock-out mice and the wild-type litter mates and maintained in a chamber at 30°C in ACSF that contained 10 μm bicuculline to suppress inhibitory synaptic responses. A bipolar stimulating electrode was placed in the stratum radiatum in the CA1/CA2 border region, and the evoked EPSP was extracellularly recorded from the stratum radiatum in the CA1 region with a glass capillary microelectrode (1–5 MΩ) filled with 0.5 mNaCl. A single test stimulation (100 μsec) was applied at intervals of 10 sec. The initial EPSP slopes of 0.15–0.20 mV/msec were used. To induce LTP, five sets with 10 sec intervals of theta-burst stimulation (10 bursts of four pulses at 100 Hz with 200 msec interburst intervals) were applied. Changes in field potential were recorded with an Axopatch 200A amplifier (Axon Instruments, Foster City, CA).
RESULTS
CBS produces the endogenous brain H2S
Relatively high endogenous levels of brain H2S have been measured in rats, humans, and bovine (Goodwin et al., 1989; Warenycia et al., 1989; Savage and Gould, 1990), and H2S production froml-cysteine in brain homogenates is suppressed by CBS-specific inhibitors, such as hydroxylamine and aminooxyacetate, and is increased by the CBS activator SAM (Abe and Kimura, 1996). Based on these observations, we proposed that CBS is an enzyme that produces endogenous H2S in the brain. A critical experiment to support the hypothesis is to measure the endogenous levels of H2S in the brains of CBS knock-out mice. Although the homozygotes of CBS knock-out mice show growth retardation (Watanabe et al., 1995), the morphology and weights of their brains are normal (data not shown). The genotypes of CBS knock-out mice were determined by PCR and confirmed by Western blot analysis (Fig.1A,B). Because CBS knock-out mice have a high incidence of death during the third and fourth postnatal weeks (Watanabe et al., 1995), 2-week-old mice were used to measure endogenous brain H2S. H2S in the brains of the homozygous CBS knock-out mice was under detectable levels (Fig. 1C). The H2S level of the heterozygous mice (0.76 ± 0.04 nmol/mg protein; n = 5) was less than one-half of the wild-type mice (1.60 ± 0.32 nmol/mg protein;n = 5; p < 0.05 by the Student'st test). Because the above data could be attributable to differences in substrate concentration, the amounts ofl-cysteine in the brains of CBS knock-out mice were measured and compared with those of the wild type. Although the levels of l-cysteine in homozygous (68.0 ± 2.4 μm; n = 5) CBS knock-out mice are less than those of the wild-type (79.8 ± 8.8 μm; n = 5; p < 0.05 by the Student's t test) and heterozygous (83.1 ± 2.8 μm; n = 5;p < 0.001) mice, the lack of brain H2S in CBS knock-out mice cannot be attributable to the slightly lower level of l-cysteine (Fig.1D). These observations clearly show that CBS produces the endogenous H2S in the brain.
Brains of CBS knock-out mice do not contain endogenous H2S. A, Determination of the genotype of CBS knock-out mice by PCR. PCR amplification with three primers gave fragments of 500 bp for the wild type and 420 bp for the disrupted mutant. B, Determination of CBS levels in the brain by Western blot analysis. Protein (15 μg) obtained from the brain homogenates of homozygous (−/−) and heterozygous (+/−) CBS knock-out and the wild-type (+/+) mice were analyzed by Western blotting with antibodies against CBS and actin. C, Determination of endogenous H2S levels in the brain. Endogenous H2S levels in the brains of homozygous (−/−) and heterozygous (+/−) CBS knock-out mice and the wild type (+/+) were determined by gas chromatography. D, Endogenousl-cysteine levels in the brain. Endogenousl-cysteine levels in the brains used for Cwere determined by HPLC. Data in C and Drepresent the mean ± SEM of five experiments for the heterozygous mice and the wild type and three experiments for the homozygous CBS knock-out mice. *p < 0.05; **p< 0.001; Student's t test.
Regulation of CBS activity by Ca2+/calmodulin
CBS is dependent on pyridoxal 5′-phosphate and heme, and its activity is enhanced by SAM (Finkelstein et al., 1975; Kery et al., 1994). No other regulators for this enzyme have been found. CBS contains a consensus sequence conserved in calmodulin binding proteins (Rhoads and Friedberg, 1997) (Fig.2A). Therefore, the potential interaction between CBS and calmodulin was examined by immunoprecipitation assays with brain extracts. Because calmodulin binding is Ca2+ dependent, brain extracts were immunoprecipitated with an antibody against CBS in the presence or absence of 1 mmCa2+. Calmodulin coimmunoprecipitated with CBS in the presence of 1 mmCa2+ but not in the absence of Ca2+ (Fig. 2B). CBS was not coimmunoprecipitated with an unrelated antibody against focal adhesion kinase in the same experiment (data not shown). This observation shows that CBS interacts with calmodulin in the presence of Ca2+.
H2S production by CBS is regulated by Ca2+/calmodulin. A, A comparison of a consensus sequence of calmodulin binding domain of CBS with other calmodulin binding proteins. CaM-KII, Calmodulin-dependent kinase II; CaM-KI, calmodulin-dependent kinase I; MARCKS, myristoylated alanine-rich protein kinase C substrate; HSP84, heat shock protein 84 kDa. B, Immunoprecipitation assay for calmodulin binding to CBS. Ten milligrams of brain homogenate were immunoprecipitated with an antibody against CBS and analyzed by Western blotting with antibodies against CBS and calmodulin. Actin in the lysate served as a control. C, H2S production from purified CBS is regulated by Ca2+and calmodulin. Protein (4.6 μg) of CBS purified by calmodulin Sepharose 4B column was incubated with 1 mm cysteine and 2 mm pyridoxal 5′-phosphate in the presence or absence of 6 μm Ca2+ and/or 1 μmcalmodulin, and H2S production was measured.RM, Reaction mixture alone; EF, enzyme fraction alone. *p < 0.01; **p< 0.001; Student's t test. D,E, Ca2+ (D) or calmodulin (E) concentration-dependent H2S production from purified CBS. Protein (4.6 μg) of CBS purified by calmodulin Sepharose 4B was incubated with 1 mmcysteine and 2 mm pyridoxal 5′-phosphate in the presence or absence of 1 μm calmodulin (D) or 6 μm Ca2+ (E), and the production of H2S with different concentrations of Ca2+ (D) or calmodulin (E) was measured. F, Inhibition of H2S production by calmodulin inhibitors. Purified CBS (4.6 μg) was incubated with 1 mm cysteine and 2 mmpyridoxal 5′-phosphate in the presence of 1 μm calmodulin and 6 μm Ca2+, and the effects of trifluoroperazine (●) or W-13 (○) on H2S production were examined. All data from C–F represent the mean ± SEM of five experiments.
Because CBS interacts with calmodulin, H2S production by CBS could be regulated by Ca2+/calmodulin. To examine this possibility, H2S production by CBS purified from brain homogenates by calmodulin Sepharose 4B chromatography was investigated. CBS was purified 45-fold as determined by the ratio of activity to protein relative to the crude brain homogenates. H2S was then measured in the presence of 1 mml-cysteine and 2 mm pyridoxal 5′-phosphate, plus or minus 0.6 μmCa2+ or 1 μm calmodulin. In the presence of Ca2+ and calmodulin, CBS produced H2S at a rate 3.5 times greater than those without Ca2+ and calmodulin (Fig.2C) (p < 0.01 by the Student'st test). In the presence of SAM, the enhanced CBS activity by Ca2+ and calmodulin was potentiated (Fig. 2C) (p < 0.001). Calmodulin or SAM alone very weakly activated the production of H2S, but Ca2+ alone did not have any effect on CBS activity. These observations show that CBS is regulated by Ca2+/calmodulin.
The minimal Ca2+ concentration required for the maximal activation of CBS was determined with purified CBS in the presence of 1 μm calmodulin. Ca2+ potentiates H2S production in the presence of calmodulin, with an ED50 value of 290 nm, whereas H2S production remained at the basal level in the absence of calmodulin (Fig. 2D). The dose–response curve of calmodulin required for CBS activity was also determined in the presence of 6 μmCa2+. Calmodulin potentiates H2S production with the ED50 of 140 nm (Fig.2E). To confirm that H2S production by CBS requires Ca2+/calmodulin, the effect of calmodulin inhibitors on H2S production from purified CBS was examined. A potent calmodulin inhibitor, trifluoroperazine, suppressed H2S production, with IC50 value of 8 μm (Fig.2F). A specific calmodulin inhibitor, W-13, suppressed H2S production, with an IC50 value of 51 μm (Fig.2F). Those observations confirm that H2S production by CBS is regulated by Ca2+/calmodulin.
Calmodulin binding domain in CBS
The above observations show that CBS produces endogenous H2S in the brain and that CBS is regulated by Ca2+/calmodulin. To determine the Ca2+/calmodulin regulatory domain within CBS, deletion mutants of CBS was prepared by transfecting COS-7 cells with expression plasmids containing mutant CBS cDNAs. The mutant (1–415), which has the C-terminal 141 amino acids of the wild-type CBS deleted (Kery et al., 1998), contains a consensus sequence for calmodulin binding (Fig. 2A), but the mutant (1–396), lacking the C-terminal 160 amino acids, is deficient in the consensus sequence. These mutants produced by COS-7 cells are shown in Figure 3A.
Determination of the calmodulin binding domain of CBS. A, Western blot analysis of deletion mutants of CBS. COS-7 cells were transiently transfected with expression plasmids containing the wild-type CBS (amino acid 1–556), mutant (1–415), and mutant (1–396). The cell lysates were analyzed by Western blotting with the antibody against CBS.B, Immunoprecipitation assay for calmodulin (CaM) binding to CBS mutants. Lysates (1.5 mg) described in A were immunoprecipitated with an antibody against CBS and analyzed by Western blotting with antibodies against CBS and calmodulin. Actin in the lysates served as control. Note that CBS mutants (1–415) and (1–396) appeared with a strong band of IgG heavy chain. C, Ca2+/calmodulin-dependent production of H2S from the wild-type and mutant CBS. Protein (1.5 μg) obtained from each cell lysate described in A was incubated with 1 mml-cysteine and 2 mm pyridoxal 5′-phosphate in the presence or absence of 6 μm Ca2+/1 μm calmodulin, and the production of H2S was measured. The data represent the mean ± SEM of five experiments. *p < 0.01; Student's t test.
To determine whether or not the 19 amino acid sequence that contains the calmodulin binding consensus sequence interacts with calmodulin, an immunoprecipitation experiment was performed with lysates of COS-7 cells containing the wild-type CBS, mutant (1–415) or mutant (1–396). Lysates of COS-7 cells were immunoprecipitated with the antibody against CBS, and Western blot analysis was performed with the antibody against calmodulin. Both the wild type and mutant (1–415) coimmunoprecipitated with calmodulin, but the mutant (1–396) did not (Fig. 3B). These data show that the 19 amino acid sequence of CBS is required for the interaction with calmodulin.
Because CBS interacts with calmodulin at a 19 amino acid consensus sequence, it was asked whether this consensus sequence is critical for the regulation of H2S production. Lysates of COS-7 cells containing the wild-type or CBS mutants were incubated with 1 mml-cysteine and 2 mm pyridoxal 5′-phosphate in the presence or absence of 0.6 μmCa2+/1 μm calmodulin, and H2S production was determined. The mutant (1–396), which is deficient in 19 amino acid consensus sequence, produced H2S at almost the same rate as the wild-type enzyme, even in the absence of Ca2+/calmodulin (Fig. 3C) (p < 0.01 by the Student's t test). The wild type and mutant (1–415) produced only a basal rate of H2S in the absence of Ca2+/calmodulin. These observations suggest that the 19 amino acid sequence suppresses the CBS activity in the absence of Ca2+/calmodulin. Once calmodulin binds to the sequence, CBS is released from the suppressed state to become active. A similar model has been proposed for the regulation of CBS activity by SAM (Shan et al., 2001).
H2S production is enhanced by l-glutamate, Ca2+ ionophore, and electrical stimulation
Because H2S production by CBS is regulated by Ca2+/calmodulin, H2S production may be controlled by neuronal activity. To examine this possibility, we prepared brain cell suspensions (Garthwaite et al., 1988) and measured H2S production induced by the application ofl-glutamate and its analogues. The production of H2S was greatly enhanced by stimulation withl-glutamate (p < 0.001 by the Student's t test), NMDA (p < 0.01), or AMPA (p < 0.001) in the presence of 2 mm Ca2+ (Fig.4A). The Ca2+-dependent activation by NMDA requiresl-glycine but is suppressed in the presence of Mg2+. The effect of NMDA was inhibited by an NMDA-specific inhibitor, AP-5, and that of AMPA was inhibited by an AMPA-specific inhibitor, CNQX (Fig. 4A). These observations indicate that H2S is produced when Ca2+ enters into the cells by the activation of at least two classes of ionotropic glutamate receptors.
H2S production is enhanced byl-glutamate, depolarization, and Ca2+ionophore. A, l-Glutamate and its analogs enhance the production of H2S. Cells at 1 × 104 in 100 μl of cell suspensions obtained from mouse cerebral cortices were incubated with glutamate analogs and their inhibitors at 37°C for 5 min, and the production of H2S was measured by gas chromatography. *p < 0.01; **p < 0.001; Student's t test.B, Membrane depolarization enhances the production of H2S. K+ at 30 or 60 mm was applied to brain cell suspensions in the presence or absence of 2 mm Ca2+ at 37°C for 5 min, and the production of H2S was measured. *p < 0.05; **p < 0.001; Student's ttest. C, H2S production in brain cell suspensions induced by Ca2+ influx. H2S produced in brain cell suspensions by the application of A23187 was measured. *p < 0.001; **p < 0.01; ***p < 0.05; Student's ttest. All data in A–C represent the mean ± SEM of five experiments.
When the neuronal membrane is depolarized, voltage-activated Ca2+ channels are opened and Ca2+ enters into the cells (Llinas, 1988). Because high concentrations of K+depolarize the membrane, the effect of high concentrations of K+ on the H2S production was examined using brain cell suspensions. In the presence of 2 mm Ca2+, 30 (p < 0.05 by the Student's t test) and 60 (p < 0.001) mmK+ greatly enhanced H2S production (Fig. 4B). To confirm that H2S production from brain cell suspensions was induced by Ca2+ influx, the effect of the Ca2+ ionophore A23187 on H2S production was examined. Concentrations up to 5 μmA23187 dose-dependently potentiate H2S production (p < 0.01 by the Student's t test), whereas 10 μmA23187 showed weaker potentiation (Fig.4C) (p < 0.05). These observations suggest that H2S production is induced by Ca2+ entry after the depolarization of the membrane.
Based on the above observations, it was asked whether H2S is produced from slices of cerebral cortices by l-glutamate and electrical stimulation.l-Glutamate at 100 μm enhances H2S production three times above the basal level (Fig. 5A) (p < 0.05 by the Student's t test). Electrical stimulation for 0.5 and 1 min causes H2S production at approximately twice (1.62 ± 0.40 nmol/mg protein; n = 3; p < 0.05) and three (2.18 ± 0.50 nmol/mg protein; n = 3; p < 0.05) times the basal level (0.76 ± 0.04 nmol/mg protein; n = 3), respectively, whereas longer stimulation for 2 min did not effectively increase H2S production (p < 0.01) (Fig. 5B). These observations show that H2S is produced when neurons in slices are excited by electrical stimulation.
The production of H2S and the induction of LTP in brain slices. A, B, The production of H2S induced by l-glutamate and electrical stimulation. H2S produced in the slices of cerebral cortices by the application of 100 μml-glutamate (A) and by electrical stimulation for 0, 0.5, 1, and 2 min (B) was measured. Data are represented as the mean ± SEM of three experiments. *p < 0.05; **p < 0.01; Student's t test. C, LTP is altered in the absence of H2S. Five sets of theta-burst stimulation (10 trains of 4 pulses of 100 μsec each at 200 msec intervals) applied at 10 sec intervals to hippocampal slices of CBS knock-out mice (●) and the wild-type mice (○). The field EPSP slopes were expressed as the percentage of baseline values before stimulation. Representative records at the times denoted by thenumbers are shown as insets. The mean field EPSP slope (166.1 ± 10.1%; n = 9) 60 min after stimulation in the slices of CBS knock-out mice is significantly different (p < 0.037; Student's t test) from those in the wild-type mice (132.1 ± 9.3%; n = 6).
LTP is altered in CBS knock-out mice
Because exogenously applied H2S modifies the induction of LTP in hippocampal slices (Abe and Kimura, 1996), we examined whether or not LTP is altered in CBS knock-out mice. Theta-burst stimulation was applied to induce LTP (Parent et al., 1998), and changes in the slopes of EPSPs were measured. After stimulation, the augmented field EPSP slope in CBS knock-out mice gradually decreased and reached a plateau of 166.1 ± 10.1% (n = 9) of that before stimulation (Fig.5C). In contrast, in slices of wild-type mice, the field EPSP slope reached a plateau of 132.1 ± 9.3% (n= 6) of that before stimulation (Fig. 5C). The statistical difference between EPSP slopes at 60 min after stimulation in CBS knock-out mice and the wild-type mice is significant (p < 0.037 by the Student's ttest). These observations show that LTP is altered in the absence of H2S and suggest the involvement of H2S in synaptic activity.
DISCUSSION
The enzymatic activity of CBS has two metabolic outcomes (Mudd et al., 1989; Warenycia et al., 1989). Most studies have been devoted to a pathway in which CBS catalyzes the reaction with substrate homocysteine to produce cystathionine (Mudd et al., 1989). In another pathway, CBS catalyzes the reaction with l-cysteine as a substrate to produce H2S and pyruvate (Stipanuk and Beck, 1982). The latter reaction had not been studied in the nervous system until we proposed that CBS can produce endogenous H2S in the brain (Abe and Kimura, 1996). Because the activities of CBS in both metabolic pathways are regulated by SAM (Finkelstein et al., 1975; Abe and Kimura, 1996), a model for CBS regulation has been proposed in which the C-terminal domain of CBS bends to and covers its own catalytic domain, suppressing enzymatic CBS activity. Once SAM binds to the regulatory domain of CBS, a conformational change occurs that frees the catalytic domain, and CBS becomes active (Shan et al., 2001). Our present observations suggest that a similar mechanism may also function in the regulation of CBS by Ca2+/calmodulin. In the absence of Ca2+/calmodulin, the C-terminal domain may cover the catalytic domain, and CBS activity remains at a basal level. When Ca2+/calmodulin binds to the 19 amino acid sequence, the catalytic domain is exposed by opening of the C-terminal domain, and CBS becomes active. This model is supported by our observation that the CBS mutant (1–396), which is deficient in the 19 amino acid Ca2+/calmodulin binding sequence, is constantly active, even in the absence of Ca2+/calmodulin (Fig. 3C).
Physiological basal concentrations of H2S applied exogenously with a weak tetanic stimulation, which by itself does not induce LTP, facilitate the induction of LTP (Abe and Kimura, 1996). The present study shows that LTP is augmented in CBS knock-out mice (Fig.5C). Physiological basal concentrations of H2S enhance the NMDA receptor-mediated responses, whereas higher concentrations of H2S specifically suppress EPSPs (Abe and Kimura, 1996). When H2S is applied by superfusion, NMDA receptors on the postsynaptic membrane may be activated before the suppression of EPSPs, resulting in the facilitated induction of LTP. In contrast, electrical stimulation might produce H2S at nerve endings, which could suppress EPSPs before diffusing across the synaptic cleft to activate postsynaptic NMDA receptor. Because there is no endogenous H2S in CBS knock-out mice, LTP must be augmented.
Although H2S is a toxic gas, most toxicology work has been done with whole animals, and less is known about its direct effect on cells (Reiffenstein et al., 1992). Because the neuronal excitation is local and lasts only for a short time, the increase in the concentrations of H2S might not be toxic. For example, higher concentrations of H2S than the basal level suppress EPSPs, but this suppression is reversible (Abe and Kimura, 1996). In addition, H2S in the brain is tightly regulated to maintain endogenous concentrations at less than the toxic levels. For example, concentrations >10 μmA23187 and the electrical stimulation longer than 2 min did not efficiently enhance H2S production (Figs.4C, 5B).
In conclusion, endogenous H2S in the brain is produced by CBS, and the production of H2S by CBS is regulated by Ca2+/calmodulin. The production of H2S is greatly enhanced by the activation of glutamate receptors, as well as by electrical stimulation, and the loss of H2S alters LTP. These observations suggest that H2S may regulate some aspects of synaptic activity.
Footnotes
This work was supported by a grant from National Institute of Neuroscience/National Center of Neurology and Psychiatry (H.K.) and National Institutes of Health Grant 5R21GM57504. We thank Dr. J. P. Kraus for a CBS cDNA plasmid and Dr. N. Maeda for the information of three PCR primers to determine the genotype of CBS knock-out mice. We also thank Dr. D. Schubert for the critical reading of this manuscript.
Correspondence should be addressed to Dr. Hideo Kimura, National Institute of Neuroscience, National Center of Neurology and Psychiatry, 4-1-1 Ogawahigashi, Kodaira, Tokyo 187-8551, Japan. E-mail:kimura{at}ncnp.go.jp.
