- Email: [email protected]
Site-Directed Mutagenesis on Human Cystathionine-γ-Lyase Reveals Insights into the Modulation of H2S Production
doi:10.1016/j.jmb.2009.11.058
J. Mol. Biol. (2010) 396, 708–718
Available online at www.sciencedirect.com
Site-Directed Mutagenesis on Human Cystathionine-γ-Lyase Reveals Insights into the Modulation of H2S Production Shufen Huang 1 , Jia Hui Chua 1 , Wen Shan Yew 2 , J. Sivaraman 3 , Philip K. Moore 4 , Choon-Hong Tan 1 and Lih-Wen Deng 2 ⁎ 1
Department of Chemistry, National University of Singapore, Singapore 2
Department of Biochemistry, National University of Singapore, Singapore 3
Department of Biological Sciences, National University of Singapore, Singapore 4
Pharmaceutical Sciences Division, King's College, University of London, UK Received 22 October 2009; received in revised form 24 November 2009; accepted 25 November 2009 Available online 1 December 2009
In recent years, increased interest has been directed towards hydrogen sulfide (H2S) as the third gasotransmitter and its role in various diseases. Cystathionine-γ-lyase (CSE) is one of the enzymes responsible for the endogenous production of H2S in mammals. With the aid of the crystal structures of human CSE and site-directed mutagenesis studies, we have identified several amino acid residues in CSE that are actively involved in the catalysis of H2S production. Contrary to reports suggesting that Tyr114 is required for substrate binding, our results reveal a significant increase in the production of H2S upon mutation of Tyr114 to phenylalanine. This is attributed to an increased rate of pyridoxal 5′-phosphate (PLP) regeneration due to weakened π-stacking interactions between Phe114 and PLP. Thr189 is also identified as a crucial residue where hydrogen bonding to Asp187 keeps the latter in an optimal position for hydrogen bonding to the pyridoxal nitrogen of PLP. Furthermore, mutation of Glu339 to lysine, alanine or tyrosine reveals the importance of the hydrophobicity of the 339th amino acid in determining the specificity of the enzyme for the catalysis of α,γ-elimination or α,β-elimination reaction. Our study also shows that the rate of H2S production is increased with increasing exogenous PLP concentration, hence supporting our hypothesis that apoCSE is formed during the catalysis of H2S production. Taken together, these findings suggest novel routes towards the design of activators or inhibitors that modulate the production of H2S; these modulators may also serve as lead compounds in the development of drugs or mechanistic probes in the study of various H2S-related diseases. © 2009 Elsevier Ltd. All rights reserved.
Edited by R. Huber
Keywords: hydrogen sulfide (H2S); cystathionine-γ-lyase (CSE); site-directed mutagenesis; pyridoxal 5′-phosphate (PLP); modulation of H2S production
Introduction Hydrogen sulfide (H2S), previously regarded primarily as an environmental hazard and toxic gas, has recently been recognized as the third gasotransmitter besides carbon monoxide and nitric oxide.1 The endogenous production of H2S in mammalian brain *Corresponding author. 8 Medical Drive, Block MD7, 02-09, Singapore 117597, Singapore. E-mail address: [email protected]. Abbreviations used: H2S, hydrogen sulfide; CSE, cystathionine-γ-lyase; PLP, pyridoxal 5′-phosphate; PAG, DL-propargylglycine; BCA, β-cyanoalanine; GST, glutathione S-transferase; CBL, cystathionine-β-lyase.
and heart tissues is attributed mainly to cystathionine-β-synthase and 3-mercaptopyruvate sulfur transferase, whilst that in mammalian liver, kidney, intestine and vascular smooth muscle cells is ascribed largely to cystathionine-γ-lyase (CSE; EC 4.4.1.1).2,3 Physiologically, H2S has been shown to be a potent vasorelaxant and regulator of blood pressure.4,5 CSEdeficient mice exhibit markedly reduced endogenous H2S levels and display pronounced hypertension and reduced endothelium-dependent vasorelaxation.4 Imbalances in endogenous H2S levels have also been associated with other diseases such as haemorrhagic shock, carrageenan-induced hindpaw oedema, acute pancreatitis and endotoxemia.6–9 In each of these diseases, H2S donors such as sodium hydrosulfide or GYY4137 (a newly characterized slow-
0022-2836/$ - see front matter © 2009 Elsevier Ltd. All rights reserved.
709
Modulation of H2S Production in CSE
releasing H2S compound), as well as inhibitors of H2S production such as DL-propargylglycine (PAG) or βcyanoalanine (BCA), have been found to exhibit therapeutic potential where the severity of the diseases was alleviated upon administration of these compounds.5–9 Currently, the two commercially available inhibitors of H2S production, PAG and BCA, possess low potency, poor selectivity and limited cell membrane permeability.10 For the development of more effective inhibitors of H2S production, a better understanding of the binding mode of the substrate and identification of key residues in the active site of the enzyme are crucial. CSE is an enzyme that is found predominantly in mammals and some fungi and is traditionally known for its role in the reverse transsulfuration pathway, where L-methionine is converted into L-cysteine through a series of metabolic interconversions.11 Specifically, the role of CSE in this reaction pathway is to convert L-cystathionine into L-cysteine whilst generating α-ketobutyrate and ammonia (Fig. 1). The reaction proceeds via an α,γ-elimination mechanism where the C–γ–S bond of L-cystathionine is specifically cleaved to yield L-cysteine.12 Defects in this metabolic pathway are associated with cystathioninuria, L-cysteine deficiency and subsequent impairment of glutathione metabolism, as well as higher plasma homocysteine concentrations.13–17 Besides its role in the conversion of L-cystathionine into L-cysteine, studies have also shown that CSE can utilize L-cysteine as a substrate for producing H2S via an α,β-elimination reaction (Fig. 1).18–20 However, to date, no reports have clearly demonstrated the residues that affect CSE-mediated H2S production. The three-dimensional structures of human and yeast CSE have recently been elucidated via X-ray crystallography.21,22 Structurally, the enzyme consists of four identical monomers of approximately 45 kDa, with a covalently bound pyridoxal 5′phosphate (PLP) cofactor in each monomer. However, some studies show evidence for differential PLP binding affinities among the monomers of CSE and transient dissociation or removal of PLP from the enzyme during the catalytic process.19,23 It was reported that addition of L-cysteine could inhibit the enzyme from the α,γ-elimination reaction by partially removing PLP from the holoenzyme.23 Additionally, our recent structural studies on human CSE showed that crystallization of human CSE apoen-
zyme was attained only when L-cysteine was added to the crystallization conditions, indicating the possible dissociation and rebinding of PLP to CSE during the production of H2S.21 Large conformational changes were also observed in the apo-CSE enzyme, where the tetramer exists in a much more open form compared to the holoenzyme, hence suggesting the possible role of PLP in maintaining the structure of the holoenzyme. In this report, we identified critical residues that are involved in the catalysis of H2S production, with the aid of structure-based site-directed mutagenesis of human CSE. We also explored the effects of varying exogenous PLP concentrations on catalysis (rate of H2S production) for both mutant and wildtype CSE enzymes. Taken together, this study provides insights into the catalytic importance of key active-site residues and affords clues towards the design of small molecules that activate or inhibit H2S production.
Results and Discussion Selection of amino acids for mutagenesis studies Based on the crystal structures of yeast and human CSE,21,22 as well as on sequence alignment with other transsulfuration enzymes (Supplementary Fig. 1), we identified several active-site residues that could be involved in catalyzing the α,β-elimination reaction, leading to H2S production. Studies by Messerschmidt et al. proposed that for the α,γelimination reaction catalyzed by yeast CSE, Tyr114 (the numbering of residues corresponds to that used in human CSE, unless otherwise stated) may activate the incoming substrate for subsequent transaldimination with Lys212 during substrate binding to the PLP cofactor, whilst Glu339 was hypothesized to be the key amino acid in determining the enzymatic specificity of CSE.22 These amino acids, together with some others that are either highly conserved across different CSE homologs (Tyr60, Arg62, Asp187, Thr189 and Arg375) or in close proximity to the PLP cofactor or substrate binding site of the enzyme (Ser209, Thr211 and Glu349), were selected for mutagenesis studies to determine their roles in the catalysis of H 2S production. The position of these residues in the
Fig. 1. Reactions catalyzed by CSE. CSE converts L-cystathionine into L-cysteine, α-ketobutyrate and ammonia in the reverse transsulfuration pathway via an α,γ-elimination reaction. This enzyme can also utilize L-cysteine as a substrate in an α,β-elimination reaction to produce H2S, pyruvate and ammonia.
710 active site of human CSE, as well as their binding interactions with each other or with the PLP cofactor, is shown in Fig. 2a. H2S production activities of wild-type and mutant CSE proteins An initial alanine-scanning mutagenesis of selected amino acids led to several interesting observa-
Modulation of H2S Production in CSE
tions (Fig. 2b). Most of these mutants displayed a complete loss of H2S production activity. Interestingly, the E339A mutant demonstrated an approximately 6-fold enhancement in H2S production. The effects of these mutations would be further discussed subsequently. A few other alanine mutants (S209A, T211A and E349A) displayed enzyme activities comparable to those for wild-type enzyme, suggesting that these residues are most probably not
Fig. 2. (a) Active site of the human CSE enzyme. The location of the residues (only side chains shown) studied by sitedirected mutagenesis is shown, and significant hydrogen bonds and polar interactions are depicted by dotted lines. Besides interacting with residues from the same subunit, the PLP cofactor is hydrogen bonded to Tyr60⁎ and Arg62⁎ from the adjacent subunit as well. The figure was produced by the program PyMOL.28 (b) H2S production by wild-type and various mutant CSE proteins. Each assay was performed in duplicate in the presence of 5 μg of the GST-tagged protein, 2.75 mM L-cysteine and 0.5 mM PLP. The results are displayed as the mean from three independent runs ± SD. ⁎P b 0.05 compared to wild-type GST-tagged CSE enzyme.
Modulation of H2S Production in CSE
catalytically involved in the production of H2S. Through CD studies, we ascertained that the loss of enzyme activity for most of the alanine mutants was not due to the disruption of overall protein structure, as the protein conformation of all mutants was maintained (Supplementary Fig. 2). A comparison of the H2S-synthesizing activities of glutathione S-transferase (GST)-tagged and GST-cleaved wildtype CSE revealed that the presence of the GST tag did not affect the enzyme activity. The GST tag was also unlikely to interfere with the assay, as no H2Ssynthesizing activity was detected in a control activity experiment with GST alone. Residues affecting the binding of PLP cofactor In the crystal structures of yeast and human CSE, several residues were found to contribute towards the binding of the PLP cofactor.21,22 From Fig. 2a, these residues include Tyr60 and Arg62 from the adjacent monomer; Tyr114, which πstacks with the pyridoxal ring of the cofactor; Asp187, which hydrogen bonds to the pyridoxal nitrogen; and Ser209 and Thr211, which provide hydrogen-bond contacts to the phosphate group. Studies have also shown that Lys212 serves as an important catalytic residue by forming a covalent bond to the PLP cofactor and facilitating proton transfer reactions during the α,γ-elimination reaction of L-cystathionine.22 In our study of the α,βelimination reaction of cysteine, mutation of the Lys212 residue to alanine decreased the production of H2S drastically by about 78 times in comparison to wild-type CSE (Fig. 3a). From our crystal structure of the human CSE holoenzyme, which had been previously determined, it is apparent that the PLP cofactor is covalently bound to the enzyme via a Schiff base formation to Lys212 (Fig. 3b, magenta). However, in the absence of PLP, the crystal structure
711 of the apoenzyme revealed that Lys212 was significantly displaced away from the PLP binding site (Fig. 3b, green). This may imply that the binding of PLP to the enzyme is crucial in maintaining the conformation and structure of the active site of CSE. Our observations also suggest that the impediment of PLP binding is likely to alter the productive geometry of the active site, resulting in a decrease in the enzyme activity of the mutant enzyme. Tyr114 is generally believed to be a critical residue in most transsulfuration enzymes.22,24 In the case of yeast CSE, the phenolic side chain of Tyr114 was suggested to activate the substrate by deprotonating L-cystathionine, hence allowing it to bind to the enzyme at the initial step of the α,γ-elimination reaction. However, no reports have hitherto explored the role of Tyr114 in the α,β-elimination reaction of L-cysteine. In our study, although mutation of this residue to alanine (Y114A) resulted in a loss of enzyme activity (Fig. 3a), a corresponding mutation of this residue to phenylalanine (Y114F) unexpectedly led to a 3.6-fold increase in H2S production compared to wild-type CSE (Fig. 3a). This suggests that the hydroxyl functionality of the phenolic side chain of Tyr114 may not be required during H2S production. In contrast to L-cystathionine, L-cysteine exists predominantly as an anion whose amino group is electrically neutral under physiological pH. Hence, activation of the L-cysteine substrate towards binding is not crucial, since the amino group would be sufficiently nucleophilic for a transaldimination reaction with Lys212 during substrate binding. In our CSE holoenzyme crystal structure, π-stacking interactions between the aromatic side chain of Tyr114 and the bound PLP exist (Fig. 3c, magenta). As with Lys212, we have also observed a large displacement in the Tyr114 residue of the apoenzyme structure (Fig. 3c, green). This may suggest that the π-stacking interactions between
Fig. 3. (a) H2S production by wild-type CSE and various mutant CSE proteins with mutagenic alterations to Lys212 and Tyr114. The results are displayed as the mean from three independent runs ±SD. ⁎P b 0.05 compared to wild-type GST-tagged CSE enzyme. #P b 0.05 compared to the corresponding alanine mutant protein. (b and c) Structural alignment of the CSE holoenzyme (magenta) and apoenzyme (green). Large displacements in polypeptide chains around Lys212 (b) and Tyr114 (c) are observed. The figure was produced by the program PyMOL.28
712
Modulation of H2S Production in CSE
Fig. 4. (a) H2S production by wild-type CSE and various mutant CSE proteins with mutagenic alterations to Tyr60 and Arg62. Results are displayed as the mean from three independent runs ±SD. ⁎P b 0.05 compared to wild-type GST-tagged CSE enzyme. #P b 0.05 compared to the corresponding alanine mutant protein. (b) The human CSE holoenzyme tetramer made up of a dimer of dimers. Interactions between PLP and Tyr60 and Arg62 from the adjacent monomer are shown for two of the enzymatic subunits C and D, which are shown in pale green and pink, respectively. The figure was produced by the program PyMOL.28
Tyr114 and PLP are required for preserving the active-site structure in the CSE holoenzyme. Thus, we speculate that the lack of H2S production from the Y114A mutant was most likely attributed to a disruption of the π-stacking interactions. Without these stabilizing interactions between Ala114 and PLP in the mutant enzyme, it is likely that the productive geometry of the active site is decreased, leading to the observed mutant enzyme inactivity. In the case of Y114F, we hypothesize that the increase in the production of H2S is due to weaker π-stacking interactions between the phenylalanine ring and PLP; this would be further elaborated on subsequently.
In this study, the effects of mutagenic alterations to Tyr60 and Arg62 on the catalysis of H2S production were also explored. Initial mutageneses to alanine (Y60A and R62A) showed that production of H2S was reduced compared to production of the wild-type enzyme (Fig. 4a). However, enzyme activity was not restored despite a subsequent mutation to threonine and lysine, respectively, even though these residues possessed similar sidechain properties (hydroxyl or basic for tyrosine and arginine, respectively) (Fig. 4a). From the crystal structure of the CSE holoenzyme (Fig. 4b), these two residues are extensively involved in hydrogen-bond contacts with the PLP cofactor from the neighboring
Fig. 5. (a) H2S production by wild-type CSE and various mutant CSE proteins with mutagenic alterations to Asp187 and Thr189. The results are displayed as the mean from three independent runs ±SD. ⁎P b 0.05 compared to wild-type GST-tagged CSE enzyme. #P b 0.05 compared to the corresponding alanine mutant protein. (b) Hydrogen-bonding network between Thr189, Asp187 and the pyridoxal nitrogen group of PLP. The figure was produced by the program PyMOL.28
713
Modulation of H2S Production in CSE
ionic interactions can be formed between the αcarboxyl group of L-cystathionine and Arg375.22 In the case of the α,β-elimination reaction of cysteine, we originally speculated that the guanidinium side chain of Arg375 of human CSE would play a similar role by forming ionic interactions with the carboxylate group of the L-cysteine substrate. Within expectations, the mutation of Arg375 to alanine indeed resulted in a loss of H2S production (Fig. 2b). However, a subsequent mutation to lysine (R375K) failed to restore any enzyme activity (data not shown), suggesting the importance of the fully conserved Arg375 residue rather than the positive charge of the side chain in maintaining enzyme activity. From structure-based sequence alignments between yeast CSE and bacterial or plant cystathionine-β-lyase (CBL), Messerschmidt et al. predicted that the specificity of the enzyme for the α,γelimination or α,β-elimination of L-cystathionine would depend on the hydrophobicity of the 339th residue.22 From the sequence alignment of CSE and CBL, this residue correlated with more hydrophobic residues such as valine and tyrosine in plant or bacterial CBL, respectively (Supplementary Fig. 1). In our study, mutation of the Glu339 residue in human CSE to a more hydrophobic residue such as alanine or tyrosine caused an increase in the α,βelimination reaction on L-cysteine. The net production of H2S was increased by about 6-fold for these two mutant proteins when compared to wild-type CSE (Table 1). Interestingly, the production of H2S was not compromised even upon mutation of the negatively charged Glu339 residue to the positively charged lysine amino acid (Table 1). This observation stresses that the extent of hydrophobicity— rather than the charge of this residue—determines the rate of H2S production by CSE. From Table 1, enzyme activity, as reflected by catalytic efficiency (kcat/Km), correlated with the logP values of various mutants (indicating the degree of hydrophobicity of the mutated residue). As the hydrophobicity of the mutated residue increased, a corresponding 1.8fold, 3.2-fold and 7.2-fold increase in the catalytic efficiency of E339K, E339A and E339Y, respectively, was observed in comparison to that of the wild-type enzyme. In addition, although the catalytic efficiency of E339A was lower than that of E339Y, the firstorder rate constant of product formation, as reflected by the turnover number (kcat) of E339A, was about twice that of E339Y. This was, however, offset by an increased Michaelis–Menten constant
monomer, thus forming stable dimers and the active-site interface between two monomers. Sizeexclusion chromatography and analytical ultracentrifugation experiments that had been previously performed also showed that the apoenzyme exists as a weak tetramer (dimer of dimers) in solution.21 Thus, it is likely that mutation of Tyr60 and Arg62 would disrupt the active-site interface and the interaction between adjacent monomers, leading to a loss in enzyme activity. Indeed, alignment of various CSE homologs and transulfuration enzymes showed that these two residues are fully conserved (Supplementary Fig. 1), hence reflecting their importance in maintaining enzyme activity. The crystal structure of the CSE holoenzyme revealed that the pyridoxal nitrogen group of the PLP cofactor is hydrogen bonded to the carboxylate side chain of Asp187 (Fig. 5b). This hydrogenbonding interaction has been proposed to be critical in stabilizing the positive charge of the pyridoxal nitrogen group and in increasing the electrophilic character of the pyridoxal ring.22 Indeed, we observed a complete loss of H2S production upon mutation of Asp187 to alanine (Fig. 5a). A corresponding mutation of Asp187 to the acidic glutamate residue did not restore enzyme activity (Fig. 5a). We attribute this lack of activity to the longer side chain that glutamate possesses, as an additional methylene group may render the carboxyl group in an unfavorable position for hydrogen bonding to PLP. Alanine mutation of Thr189, whose hydroxyl side chain is in turn hydrogen bonded to Asp187 (Fig. 5b), was also observed to drastically reduce H2S production (Fig. 5a). However, a corresponding mutation of Thr189 to serine was found to restore the enzyme activity by about 60% (Fig. 5a). This suggests that the hydroxyl side chain of Thr189 (or Ser189 in T189S) is crucial and possibly positions Asp187 in an optimal position for a stronger interaction with the pyridoxal nitrogen of the cofactor. Partial restoration in the enzyme activity of T189S suggests that the methyl group in threonine most probably conformationally restricts the hydroxyl group to an optimal direction for hydrogen bonding to Asp187, leading to a productive activesite geometry within the wild-type enzyme. Residues affecting the binding of substrate or the release of products Based on docking experiments of L-cystathionine into the active site of yeast CSE, it was proposed that
Table 1. Correlation between the logP values of residues at the 339th position and kinetic parameters for the production of H2S from wild-type and various CSE mutants
Increasing hydrophobicity ↓
Residue at the 339th position
logPa
kcat (s− 1)
Km (mM)
kcat/Km (mM− 1 s− 1)
Net production of H2S (nmol)
Glutamate (wild-type CSE) Lysine (E339K) Alanine (E339A) Tyrosine (E339Y)
−3.7 −3.1 −2.9 −2.3
0.12 0.34 1.44 0.76
1.9 3.3 7.7 1.8
0.06 0.11 0.19 0.43
8.6 15.5 50.2 51.3
The net production of H2S from each of these proteins was determined after 30 min of incubation of the proteins with the substrate at 37 °C. a Values obtained from http://www.ecosci.jp/amino/amino2j_e.html.27
714 (Km), suggesting that the affinity of E339A for the Lcysteine substrate may be much lower than the affinity of E339A for the other mutants. However, once L-cysteine is bound to the enzyme, the catalysis of H2S production by E339A will occur much readily to produce the desired products. Recently, a similar mutation had been performed on this residue in yeast CSE to explore its effect on the catalytic efficiency of the enzyme during the α,γ-elimination
Modulation of H2S Production in CSE
reaction of L-cystathionine.25 Interestingly, a 30-fold reduction in catalytic efficiency was observed for the alanine and tyrosine mutants. This inverse correlation between the α,β-elimination or α,γ-elimination reaction of CSE and the hydrophobicity of the 339th residue may imply that residues with a more hydrophobic side chain at the 339th position could possibly favor the α,β-elimination reaction over the α,γ-elimination reaction in CSE.
Fig. 6. (a) Graphs of the initial reaction velocity V (expressed as U/mg purified human CSE, where 1 U = 1 μmol of H2S produced per minute) against the concentration of exogenous PLP determined under various concentrations of L-cysteine substrate. (b) Proposed mechanism for the catalysis of H2S production from L-cysteine by human CSE. The L-cysteine substrate first binds to the holoenzyme via a transaldimination reaction with Lys212 (Step 1). Lys212 then abstracts the αproton of the bound substrate (Step 2). This causes an influx of electron density into the pyridoxal ring, which in turn destabilizes the π-stacking interaction between Tyr114 and the pyridoxal ring. The C–β–S bond of the bound substrate is then cleaved, and H2S is released via a β-elimination process (Step 3). The release of the aminoacrylate intermediate (Step 4) for subsequent hydrolysis (Step 5a) is likely to be aided by a subsequent rebinding of free PLP so that the holoenzyme can be regenerated (Step 5b). (c) Comparison of the initial reaction velocity against the concentration of exogenous PLP between GST-tagged Y114F mutant and wild-type CSE enzyme in the presence of 1.5 mM L-cysteine substrate. Results are displayed as the mean from three independent runs ± SD.
Modulation of H2S Production in CSE
Catalysis of H2S production from wild-type CSE in the presence of varying PLP concentrations Previously, in the development of the H2S detection assay from rat liver or kidney tissue homogenates, the inclusion of exogenous PLP had been found to be crucial for maximal enzymatic activity.18 This was also observed in our purified human CSE assay where the rate of production of H2S was found to increase with an increase in exogenous PLP concentrations for all of the tested L-cysteine substrate concentrations (Fig. 6a). According to Yamagata et al., in addition to binding to the enzyme, Lcysteine may also bind to the exogenous PLP cofactor to form thiazolidines.26 This may offer an explanation for the subsequent decrease in reaction rates beyond an optimum concentration of exogenous PLP (approximately 50 μM; Fig. 6a). Under high concentrations of PLP, the probability of the L-cysteine substrate being bound to the exogenously added cofactor to form thiazolidines becomes higher. This, in turn, would decrease the effective concentration of the substrate, leading to a lower production of H2S. The extent of decrease in reaction rates was found to be greater under lower concentrations of L-cysteine, consistent with our expectations of the competitive nature of PLP–L-cysteine interactions vis-à-vis enzyme–substrate binding; when the amount of Lcysteine was increased, the extent of the competing side reaction relative to catalysis decreased, leading to the observed increase in reaction rates. The initial increase in reaction rates as exogenous PLP concentration increased was most probably not attributed to the enzyme being purified in apo form, since the reaction rates had been observed to increase progressively beyond an exogenous PLP concentration of 1.5 μM (which was the molar concentration of PLP required to fully saturate the amount of enzyme utilized in the assay) for each concentration of Lcysteine that was assayed. Previously, we have shown that addition of L-cysteine to the crystallization condition triggered the dissociation of PLP from our purified CSE enzyme, hence leading to crystallization of the enzyme in apo state.21 In addition, spectrophotometric experiments have indicated that the bound cofactor can be removed when L-cysteine is incubated with the enzyme, and that the holoenzyme can be regenerated upon addition of PLP to the apoenzyme.21 These findings are consistent with the observed increase in initial reaction rates as exogenous PLP concentration was increased in this study. It is hence likely that, during catalysis (H2S production from L-cysteine), the PLP cofactor is displaced from the active site, and that free PLP would have to reassociate with the enzyme active site for subsequent catalysis after a single turnover or a few turnovers (Fig. 6b, Step 5b). As the exogenous PLP concentration was increased, the proportion of PLPbound CSE would increase, leading to an increased rate of H2S production. The release of the cofactor may be triggered upon abstraction of the α-proton after binding of the L-cysteine substrate (Fig. 6b, Step 2). Although the mechanistic origin for the release of
715 the PLP cofactor remains unknown, it is observed that the rate of the α,β-elimination reaction of Lcysteine is much slower than the rate of the α,γelimination reaction of L-cystathionine. Hence, we postulate that a consequence of the β-elimination step and subsequent release of H2S (Fig. 6b, Step 3) could be the generation of a possible catalytically ineffective aminoacrylate intermediate. The displacement of this intermediate may be aided by an incoming molecule of PLP, which would establish more effective π-stacking interactions with Tyr114 so as to regenerate the catalytically competent holoenzyme (Fig. 6b, Steps 4 and 5b). Catalysis of H2S production by Y114F mutant CSE We sought further evidence for the displacement of the PLP cofactor during the catalytic cycle by assaying the Y114F mutant enzyme in the presence of varying concentrations of exogenous PLP. In comparison to wild-type CSE, the Y114F mutant enzyme produced more H2S under all assay conditions (Fig. 6c). An inspection of the side chains of Tyr114 (in wild-type CSE) and Phe114 (in Y114F) revealed a plausible explanation for the increase in H2S production: the former is more electron rich than the latter due to the resonance electrondonating effects of the para-hydroxyl group in tyrosine; thus, the initial π-stacking interactions between the electron-rich aromatic side chain and the electron-deficient pyridoxal ring would be stronger for the wild-type enzyme. As a consequence of the stronger π-stacking interactions between Tyr114 and the pyridoxal ring, the αproton of the bound L-cysteine would be less acidic, and electron delocalization into the pyridoxal ring would be less favored (Fig. 6b, Step 2). In contrast, the Y114F mutant would exhibit a weaker πstacking interaction that would facilitate electron delocalization into the pyridoxal ring. As such, the catalytic production of H2S by the Y114F mutant occurs at a rate higher than that observed for the wild-type enzyme (Fig. 3a). Due to electron delocalization into the pyridoxal rings, the π-stacking interactions could be weakened, and the dissociation of the aminoacrylate intermediate might be facilitated in the proceeding step; as a result, the much weakened π-stacking interactions of Phe114 with the pyridoxal ring of the aminoacrylate intermediate could translate to a faster dissociation of the aminoacrylate intermediate for the mutant enzyme. Such π-stacking interactions may hence also affect the rate of PLP reassociation with the apoenzyme subsequently. Indeed, a greater increase in reaction rates as exogenous PLP concentrations increased was observed for Y114F relative to wildtype CSE, corroborating our hypothesis (Fig. 6b). Implications to the modulation of endogenous H2S production Recent studies have shown that a number of diseases are associated with imbalances in endoge-
716 nous H2S levels. More importantly, H2S donor compounds and inhibitors of H2S production have been found to exhibit therapeutic potentials in various animal disease models,10 reinforcing the need for novel strategies in the design and development of activators and inhibitors of CSE towards modulating the in vivo production of H2S. In this study, elucidation of the roles of various active-site residues of CSE has been accomplished through sitedirected mutagenesis and PLP-dependent H2S production assays. Specifically, mutation of Tyr114 to phenylalanine, and mutation of Glu339 to alanine, tyrosine and lysine, respectively, have been found to increase H2S production. Our observations provide a novel route towards the screening of activators of CSE, whereupon compounds can be identified based on their ability to weaken the π-stacking interactions between Tyr114 and PLP, or to increase the hydrophobicity of Glu339, resulting in an increase in H2S production. Such compounds may include intercalating agents that possess electron-rich or electron-poor aromatic rings that disrupt the πstacking interactions between PLP and Tyr114, or specific alkylating agents that target the side chain of Glu339. With regard to the design of the inhibitors of H2S production, our observation that the PLP cofactor is displaced from the enzyme during the course of catalysis (H2S production) possibly affords a new strategy for designing an inhibitor that competes with the rebinding of PLP to the apo-CSE enzyme. Unlike PAG and BCA, which inhibit the enzyme by binding to PLP itself and would thus inhibit all other PLP-dependent enzymes (in an unwarranted nonspecific manner), such a new inhibitor may be more specific, since it targets and binds reversibly to the PLP cofactor binding site rather than to the cofactor itself. Since the mode of inhibition is centered on binding to the cofactor site upon displacement of PLP during the catalytic cycle, this new inhibitor would not directly inhibit the α,γ-elimination reaction of CSE (where displacement of the cofactor is unlikely to occur). Although the enzyme would likely be inactive towards the α,γ-elimination reaction after inhibition of the α,β-elimination reaction, this inhibition can be reversed by a competitive rebinding of the PLP cofactor. The selective and reversible inhibition of the α,βelimination reaction of CSE thus holds therapeutic potential, as endogenous H2S levels are lowered physiologically, whilst any indirect inhibition of the reverse transsulfuration pathway can be reversed and L-cysteine can still be produced by the enzyme. Hydrogen-bonding or ionic interactions that are afforded through the side chains of Tyr60, Arg62, Asp187 and Thr189 suggest an additional route of inhibitor design, through the incorporation of various functional groups into the inhibitor to enhance its binding to the enzyme. Taken together, our findings present novel strategies for the future design and development of potent and selective drug lead compounds that modulate the endogenous production of H2S.
Modulation of H2S Production in CSE
Materials and Methods Preparation of mutant CSE clones Mutant DNA strand synthesis was performed using the QuikChange site-directed mutagenesis kit (Stratagene) in accordance with the manufacturer's protocol. The sequence of the mutated DNA strands was then confirmed by automated DNA sequencing. Expression and purification of GST-tagged CSE proteins Mutant and wild-type GST-tagged CSE proteins were expressed and purified based on previously described procedures,21 with slight modifications. After bacterial expression of the GST-tagged fusion protein, the protein was purified with glutathione Sepharose beads and subsequently eluted with 1 mM reduced glutathione (Sigma) in 20 mM Tris (pH 8.0), 50 mM NaCl and 1 mM dithiothreitol. The protein was estimated to be at least 90% pure from SDS-PAGE analysis. Protein concentration was then determined by Bradford protein assay or with the aid of a NanoDrop ND-1000 spectrophotometer. H2S production assay The production of H2S from the GST-tagged mutant or wild-type CSE proteins was assayed using a procedure similar to that described by Stipanuk and Beck, with some modifications.18 Each reaction mixture (100 μL) consisted of 5 μg of the GST-tagged protein, sodium chloride (20 mM), PLP (0.5 mM), L-cysteine (2.75 mM) and sodium phosphate (pH 8.2) buffer (50 mM). After the incubation of parafilmed tubes at 37 °C for 30 min, a 100-μL mixture of 0.85% wt/vol zinc acetate and 3% wt/vol NaOH was added to each tube via a needle to trap evolved H2S gas. Enzymatic reactions were terminated by the addition of 100 μL of 10% wt/vol trichloroacetic acid. This was followed by the addition of N,N-dimethyl-p-phenylenediamine-dihydrochloride dye and FeCl3 to a final concentration of 2.5 mM and 3.3 mM, respectively, for the development of methylene blue and subsequent absorbance measurements at 670 nm. The amount of H2S produced was calculated against a calibration curve of sodium hydrosulfide (0–800 μM) and compared to that for control tubes without addition of the protein. For kinetic assays, the amount of H2S produced was determined at 5min intervals over a period of 20 min at 37 °C. The Michaelis–Menten constant (Km) and maximal velocity (Vmax) values were obtained by plotting the reciprocal of the initial reaction velocity V0 and the L-cysteine substrate concentration [S] according to the Lineweaver–Burk equation. The graphical software SigmaPlot (Systat) was used for all curve-fitting and regression analyses. Catalysis of H2S production under varying exogenous PLP concentrations The production of H2S was monitored at 5-min intervals over 20 min at 37 °C using a protocol similar to that described above. Each reaction mixture (100 μL) consisted of 3.88 μg of the purified human CSE enzyme, sodium chloride (20 mM), PLP (0–300 μM), L-cysteine (0.75–5 mM) and sodium phosphate (pH 8.2) buffer (50 mM). For comparison of enzyme activities between wild-type and
717
Modulation of H2S Production in CSE Y114F mutant proteins under varying exogenous PLP concentrations, the reaction mixture consisted of 3.88 μg of the GST-tagged enzyme, sodium chloride (20 mM), PLP (0–30 μM) and L-cysteine (1.5 mM). The amount of H2S produced at each time point was calculated against the calibration curve and subsequently plotted against time to determine the initial reaction velocity V0 under each Lcysteine or PLP concentration. The graphical software SigmaPlot (Systat) was used for all curve-fitting and regression analyses. Statistical analysis Data are shown as mean ± SD. Statistical analysis was performed using a two-tailed paired Student's t test.
8.
9.
10. 11.
12.
Acknowledgements This work was supported by Interfaculty Research and by grants from Office of Deputy President (Research and Technology)/Yong Loo Lin School of Medicine (R-183-000-240-101 and R-183-000-240720), Academic Research Fund (R-154-000-438-112), Ministry of Education (R183-000-195-112) and Biomedical Research Council A⁎STAR, Singapore (WBS R-143-000-350-305). S. Huang is a graduate scholar of the Ministry of Education, Singapore.
13.
14.
15.
16.
Supplementary Data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/ j.jmb.2009.11.058
References 1. Wang, R. (2002). Two's company, three's a crowd: can H2S be the third endogenous gaseous transmitter? FASEB J. 16, 1792–1798. 2. Kamoun, P. (2004). Endogenous production of hydrogen sulfide in mammals. Amino Acids, 26, 243–254. 3. Shibuya, N., Tanaka, M., Yoshida, M., Ogasawara, Y., Togawa, T., Ishii, K. & Kimura, H. (2009). 3-Mercaptopyruvate sulfurtransferase produces hydrogen sulfide and bound sulfane sulfur in the brain. Antioxid. Redox Signal. 11, 703–714. 4. Yang, G., Wu, L., Jiang, B., Yang, W., Qi, J., Cao, K. et al. (2008). H2S as a physiologic vasorelaxant: hypertension in mice with deletion of cystathionine gamma-lyase. Science, 322, 587–590. 5. Li, L., Whiteman, M., Guan, Y. Y., Neo, K. L., Cheng, Y., Lee, S. W. et al. (2008). Characterization of a novel, water-soluble hydrogen sulfide-releasing molecule (GYY4137): new insights into the biology of hydrogen sulfide. Circulation, 117, 2351–2360. 6. Mok, Y. Y., Atan, M. S., Cheong, Y. P., Wang, Z. J., Bhatia, M., Moochhala, S. & Moore, P. K. (2004). Role of hydrogen sulphide in haemorrhagic shock in the rat: protective effect of inhibitors of hydrogen sulphide biosynthesis. Br. J. Pharmacol. 143, 881–889. 7. Bhatia, M., Sidhapuriwala, J., Moochhala, S. M. & Moore, P. K. (2005). Hydrogen sulphide is a mediator
17.
18.
19. 20.
21.
22.
23.
of carrageenan-induced hindpaw oedema in the rat. Br. J. Pharmacol. 145, 141–144. Bhatia, M., Wong, F. L., Fu, D., Lau, H. Y., Moochhala, S. M. & Moore, P. K. (2005). Role of hydrogen sulfide in acute pancreatitis and associated lung injury. FASEB J. 19, 623–625. Collin, M., Anuar, F. B., Murch, O., Bhatia, M., Moore, P. K. & Thiemermann, C. (2005). Inhibition of endogenous hydrogen sulfide formation reduces the organ injury caused by endotoxemia. Br. J. Pharmacol. 146, 498–505. Szabó, C. (2007). Hydrogen sulphide and its therapeutic potential. Nat. Rev. 6, 917–935. Rose, W. C. & Wixom, R. L. (1955). The amino acid requirements of man. XIV. The sparing effect of cystine on methionine requirement. J. Biol. Chem. 216, 763–773. Flavin, M. (1971). Trans-sulfuration reactions: introduction. Methods Enzymol. 17B, 416–417. Uren, J. R., Ragin, R. & Chaykovsky, M. (1978). Modulation of cysteine metabolism in mice: effects of propargylglycine and L-cysteine-degrading enzymes. Biochem. Pharmacol. 27, 2807–2814. Vina, J., Vento, M., Garcia-Sala, F., Puertes, I. R., Gasco, E., Sastre, J. et al. (1995). L-Cysteine and glutathione metabolism are impaired in premature infants due to cystathionase deficiency. Am. J. Clin. Nutr. 61, 1067–1069. Wang, J., Huff, A. M., Spence, J. D. & Hegele, R. A. (2004). Single nucleotide polymorphism in CTH associated with variation in plasma homocysteine concentration. Clin. Genet. 6, 483–486. Wang, J. & Hegele, R. A. (2003). Genomic basis of cystathioninuria (MIM 219500) revealed by multiple mutations in cystathionine gamma-lyase (CTH). Hum. Genet. 112, 404–408. Zhu, W., Lin, A. & Banerjee, R. (2008). Kinetic properties of polymorphic variants and pathogenic mutants in human cystathionine gamma-lyase. Biochemistry, 47, 6226–6232. Stipanuk, M. H. & Beck, P. W. (1982). Characterization of the enzymatic capacity for cysteine desulphydration in liver and kidney of the rat. Biochem. J. 206, 267–277. Braunstein, A. E. & Goryachenkova, E. V. (1984). The β-replacement-specific pyridoxal-P-dependent lyases. Adv. Enzymol. 56, 1–89. Chiku, T., Padovani, D., Zhu, W., Singh, S., Vitvitsky, V. & Banerjee, R. (2009). H2S biogenesis by cystathionine gamma-lyase leads to the novel sulfur metabolites, lanthionine and homolanthionine, and is responsive to the grade of hyperhomocysteinemia. J. Biol. Chem. 284, 11601–11612. Sun, Q., Collins, R., Huang, S., Holmberg-Schiavone, L., Anand, G. S., Tan, C. H. et al. (2009). Structural basis for the inhibition mechanism of human cystathionine-gamma-lyase: an enzyme responsible for the production of H2S. J. Biol. Chem. 284, 3076–3085. Messerschmidt, A., Worbs, M., Steegborn, C., Wahl, M. C., Huber, R., Laber, B. & Clausen, T. (2003). Determinants of enzymatic specificity in the Cys-Metmetabolism PLP-dependent enzyme family: crystal structure of cystathionine γ-lyase from yeast and intrafamiliar structure comparison. Biol. Chem. 384, 373–386. Yamagata, S., Yasugahira, T., Okuda, Y. & Iwama, T. (2003). Conversion of the aminocrotonate intermediate limits the rate of γ-elimination reaction catalyzed
718 by L-cystathionine γ-lyase of the yeast Saccharomyces cerevisiae. J. Biochem. 134, 607–613. 24. Clausen, T., Huber, R., Prade, L., Wahl, M. C. & Messerschmidt, A. (1998). Crystal structure of Escherichia coli cystathionine γ-synthase at 1.5 Å resolution. EMBO J. 23, 6827–6838. 25. Farsi, A., Lodha, P. H., Skanes, J. E., Los, H., Kalidindi, N. & Aitken, S. M. (2009). Interconversion of a pair of active-site residues in Escherichia coli cystathionine γ-synthase, E. coli cystathionine βlyase, and Saccharomyces cerevisiae cystathionine γlyase and development of tools for the investigation
Modulation of H2S Production in CSE of their mechanisms and reaction specificity. Biochem. Cell Biol. 87, 445–457. 26. Yamagata, S., Isaji, M., Yamane, T. & Iwama, T. (2002). Substrate inhibition of L-cysteine α,β-elimination reaction catalyzed by L-cystathionine γ-lyase of Saccharomyces cerevisiae. Biosci. Biotechnol. Biochem. 66, 2706–2709. 27. Data Book of Molecules: Home Page for Learning Environmental Chemistry (n.d.), retrieved on April 28, 2009 from http://www.ecosci.jp/ec.html. 28. DeLano, W. L. (2002). The PyMOL Molecular Graphics System. DeLano Scientific LLC, Palo Alto, CA.
J. Mol. Biol. (2010) 396, 708–718
Available online at www.sciencedirect.com
Site-Directed Mutagenesis on Human Cystathionine-γ-Lyase Reveals Insights into the Modulation of H2S Production Shufen Huang 1 , Jia Hui Chua 1 , Wen Shan Yew 2 , J. Sivaraman 3 , Philip K. Moore 4 , Choon-Hong Tan 1 and Lih-Wen Deng 2 ⁎ 1
Department of Chemistry, National University of Singapore, Singapore 2
Department of Biochemistry, National University of Singapore, Singapore 3
Department of Biological Sciences, National University of Singapore, Singapore 4
Pharmaceutical Sciences Division, King's College, University of London, UK Received 22 October 2009; received in revised form 24 November 2009; accepted 25 November 2009 Available online 1 December 2009
In recent years, increased interest has been directed towards hydrogen sulfide (H2S) as the third gasotransmitter and its role in various diseases. Cystathionine-γ-lyase (CSE) is one of the enzymes responsible for the endogenous production of H2S in mammals. With the aid of the crystal structures of human CSE and site-directed mutagenesis studies, we have identified several amino acid residues in CSE that are actively involved in the catalysis of H2S production. Contrary to reports suggesting that Tyr114 is required for substrate binding, our results reveal a significant increase in the production of H2S upon mutation of Tyr114 to phenylalanine. This is attributed to an increased rate of pyridoxal 5′-phosphate (PLP) regeneration due to weakened π-stacking interactions between Phe114 and PLP. Thr189 is also identified as a crucial residue where hydrogen bonding to Asp187 keeps the latter in an optimal position for hydrogen bonding to the pyridoxal nitrogen of PLP. Furthermore, mutation of Glu339 to lysine, alanine or tyrosine reveals the importance of the hydrophobicity of the 339th amino acid in determining the specificity of the enzyme for the catalysis of α,γ-elimination or α,β-elimination reaction. Our study also shows that the rate of H2S production is increased with increasing exogenous PLP concentration, hence supporting our hypothesis that apoCSE is formed during the catalysis of H2S production. Taken together, these findings suggest novel routes towards the design of activators or inhibitors that modulate the production of H2S; these modulators may also serve as lead compounds in the development of drugs or mechanistic probes in the study of various H2S-related diseases. © 2009 Elsevier Ltd. All rights reserved.
Edited by R. Huber
Keywords: hydrogen sulfide (H2S); cystathionine-γ-lyase (CSE); site-directed mutagenesis; pyridoxal 5′-phosphate (PLP); modulation of H2S production
Introduction Hydrogen sulfide (H2S), previously regarded primarily as an environmental hazard and toxic gas, has recently been recognized as the third gasotransmitter besides carbon monoxide and nitric oxide.1 The endogenous production of H2S in mammalian brain *Corresponding author. 8 Medical Drive, Block MD7, 02-09, Singapore 117597, Singapore. E-mail address: [email protected]. Abbreviations used: H2S, hydrogen sulfide; CSE, cystathionine-γ-lyase; PLP, pyridoxal 5′-phosphate; PAG, DL-propargylglycine; BCA, β-cyanoalanine; GST, glutathione S-transferase; CBL, cystathionine-β-lyase.
and heart tissues is attributed mainly to cystathionine-β-synthase and 3-mercaptopyruvate sulfur transferase, whilst that in mammalian liver, kidney, intestine and vascular smooth muscle cells is ascribed largely to cystathionine-γ-lyase (CSE; EC 4.4.1.1).2,3 Physiologically, H2S has been shown to be a potent vasorelaxant and regulator of blood pressure.4,5 CSEdeficient mice exhibit markedly reduced endogenous H2S levels and display pronounced hypertension and reduced endothelium-dependent vasorelaxation.4 Imbalances in endogenous H2S levels have also been associated with other diseases such as haemorrhagic shock, carrageenan-induced hindpaw oedema, acute pancreatitis and endotoxemia.6–9 In each of these diseases, H2S donors such as sodium hydrosulfide or GYY4137 (a newly characterized slow-
0022-2836/$ - see front matter © 2009 Elsevier Ltd. All rights reserved.
709
Modulation of H2S Production in CSE
releasing H2S compound), as well as inhibitors of H2S production such as DL-propargylglycine (PAG) or βcyanoalanine (BCA), have been found to exhibit therapeutic potential where the severity of the diseases was alleviated upon administration of these compounds.5–9 Currently, the two commercially available inhibitors of H2S production, PAG and BCA, possess low potency, poor selectivity and limited cell membrane permeability.10 For the development of more effective inhibitors of H2S production, a better understanding of the binding mode of the substrate and identification of key residues in the active site of the enzyme are crucial. CSE is an enzyme that is found predominantly in mammals and some fungi and is traditionally known for its role in the reverse transsulfuration pathway, where L-methionine is converted into L-cysteine through a series of metabolic interconversions.11 Specifically, the role of CSE in this reaction pathway is to convert L-cystathionine into L-cysteine whilst generating α-ketobutyrate and ammonia (Fig. 1). The reaction proceeds via an α,γ-elimination mechanism where the C–γ–S bond of L-cystathionine is specifically cleaved to yield L-cysteine.12 Defects in this metabolic pathway are associated with cystathioninuria, L-cysteine deficiency and subsequent impairment of glutathione metabolism, as well as higher plasma homocysteine concentrations.13–17 Besides its role in the conversion of L-cystathionine into L-cysteine, studies have also shown that CSE can utilize L-cysteine as a substrate for producing H2S via an α,β-elimination reaction (Fig. 1).18–20 However, to date, no reports have clearly demonstrated the residues that affect CSE-mediated H2S production. The three-dimensional structures of human and yeast CSE have recently been elucidated via X-ray crystallography.21,22 Structurally, the enzyme consists of four identical monomers of approximately 45 kDa, with a covalently bound pyridoxal 5′phosphate (PLP) cofactor in each monomer. However, some studies show evidence for differential PLP binding affinities among the monomers of CSE and transient dissociation or removal of PLP from the enzyme during the catalytic process.19,23 It was reported that addition of L-cysteine could inhibit the enzyme from the α,γ-elimination reaction by partially removing PLP from the holoenzyme.23 Additionally, our recent structural studies on human CSE showed that crystallization of human CSE apoen-
zyme was attained only when L-cysteine was added to the crystallization conditions, indicating the possible dissociation and rebinding of PLP to CSE during the production of H2S.21 Large conformational changes were also observed in the apo-CSE enzyme, where the tetramer exists in a much more open form compared to the holoenzyme, hence suggesting the possible role of PLP in maintaining the structure of the holoenzyme. In this report, we identified critical residues that are involved in the catalysis of H2S production, with the aid of structure-based site-directed mutagenesis of human CSE. We also explored the effects of varying exogenous PLP concentrations on catalysis (rate of H2S production) for both mutant and wildtype CSE enzymes. Taken together, this study provides insights into the catalytic importance of key active-site residues and affords clues towards the design of small molecules that activate or inhibit H2S production.
Results and Discussion Selection of amino acids for mutagenesis studies Based on the crystal structures of yeast and human CSE,21,22 as well as on sequence alignment with other transsulfuration enzymes (Supplementary Fig. 1), we identified several active-site residues that could be involved in catalyzing the α,β-elimination reaction, leading to H2S production. Studies by Messerschmidt et al. proposed that for the α,γelimination reaction catalyzed by yeast CSE, Tyr114 (the numbering of residues corresponds to that used in human CSE, unless otherwise stated) may activate the incoming substrate for subsequent transaldimination with Lys212 during substrate binding to the PLP cofactor, whilst Glu339 was hypothesized to be the key amino acid in determining the enzymatic specificity of CSE.22 These amino acids, together with some others that are either highly conserved across different CSE homologs (Tyr60, Arg62, Asp187, Thr189 and Arg375) or in close proximity to the PLP cofactor or substrate binding site of the enzyme (Ser209, Thr211 and Glu349), were selected for mutagenesis studies to determine their roles in the catalysis of H 2S production. The position of these residues in the
Fig. 1. Reactions catalyzed by CSE. CSE converts L-cystathionine into L-cysteine, α-ketobutyrate and ammonia in the reverse transsulfuration pathway via an α,γ-elimination reaction. This enzyme can also utilize L-cysteine as a substrate in an α,β-elimination reaction to produce H2S, pyruvate and ammonia.
710 active site of human CSE, as well as their binding interactions with each other or with the PLP cofactor, is shown in Fig. 2a. H2S production activities of wild-type and mutant CSE proteins An initial alanine-scanning mutagenesis of selected amino acids led to several interesting observa-
Modulation of H2S Production in CSE
tions (Fig. 2b). Most of these mutants displayed a complete loss of H2S production activity. Interestingly, the E339A mutant demonstrated an approximately 6-fold enhancement in H2S production. The effects of these mutations would be further discussed subsequently. A few other alanine mutants (S209A, T211A and E349A) displayed enzyme activities comparable to those for wild-type enzyme, suggesting that these residues are most probably not
Fig. 2. (a) Active site of the human CSE enzyme. The location of the residues (only side chains shown) studied by sitedirected mutagenesis is shown, and significant hydrogen bonds and polar interactions are depicted by dotted lines. Besides interacting with residues from the same subunit, the PLP cofactor is hydrogen bonded to Tyr60⁎ and Arg62⁎ from the adjacent subunit as well. The figure was produced by the program PyMOL.28 (b) H2S production by wild-type and various mutant CSE proteins. Each assay was performed in duplicate in the presence of 5 μg of the GST-tagged protein, 2.75 mM L-cysteine and 0.5 mM PLP. The results are displayed as the mean from three independent runs ± SD. ⁎P b 0.05 compared to wild-type GST-tagged CSE enzyme.
Modulation of H2S Production in CSE
catalytically involved in the production of H2S. Through CD studies, we ascertained that the loss of enzyme activity for most of the alanine mutants was not due to the disruption of overall protein structure, as the protein conformation of all mutants was maintained (Supplementary Fig. 2). A comparison of the H2S-synthesizing activities of glutathione S-transferase (GST)-tagged and GST-cleaved wildtype CSE revealed that the presence of the GST tag did not affect the enzyme activity. The GST tag was also unlikely to interfere with the assay, as no H2Ssynthesizing activity was detected in a control activity experiment with GST alone. Residues affecting the binding of PLP cofactor In the crystal structures of yeast and human CSE, several residues were found to contribute towards the binding of the PLP cofactor.21,22 From Fig. 2a, these residues include Tyr60 and Arg62 from the adjacent monomer; Tyr114, which πstacks with the pyridoxal ring of the cofactor; Asp187, which hydrogen bonds to the pyridoxal nitrogen; and Ser209 and Thr211, which provide hydrogen-bond contacts to the phosphate group. Studies have also shown that Lys212 serves as an important catalytic residue by forming a covalent bond to the PLP cofactor and facilitating proton transfer reactions during the α,γ-elimination reaction of L-cystathionine.22 In our study of the α,βelimination reaction of cysteine, mutation of the Lys212 residue to alanine decreased the production of H2S drastically by about 78 times in comparison to wild-type CSE (Fig. 3a). From our crystal structure of the human CSE holoenzyme, which had been previously determined, it is apparent that the PLP cofactor is covalently bound to the enzyme via a Schiff base formation to Lys212 (Fig. 3b, magenta). However, in the absence of PLP, the crystal structure
711 of the apoenzyme revealed that Lys212 was significantly displaced away from the PLP binding site (Fig. 3b, green). This may imply that the binding of PLP to the enzyme is crucial in maintaining the conformation and structure of the active site of CSE. Our observations also suggest that the impediment of PLP binding is likely to alter the productive geometry of the active site, resulting in a decrease in the enzyme activity of the mutant enzyme. Tyr114 is generally believed to be a critical residue in most transsulfuration enzymes.22,24 In the case of yeast CSE, the phenolic side chain of Tyr114 was suggested to activate the substrate by deprotonating L-cystathionine, hence allowing it to bind to the enzyme at the initial step of the α,γ-elimination reaction. However, no reports have hitherto explored the role of Tyr114 in the α,β-elimination reaction of L-cysteine. In our study, although mutation of this residue to alanine (Y114A) resulted in a loss of enzyme activity (Fig. 3a), a corresponding mutation of this residue to phenylalanine (Y114F) unexpectedly led to a 3.6-fold increase in H2S production compared to wild-type CSE (Fig. 3a). This suggests that the hydroxyl functionality of the phenolic side chain of Tyr114 may not be required during H2S production. In contrast to L-cystathionine, L-cysteine exists predominantly as an anion whose amino group is electrically neutral under physiological pH. Hence, activation of the L-cysteine substrate towards binding is not crucial, since the amino group would be sufficiently nucleophilic for a transaldimination reaction with Lys212 during substrate binding. In our CSE holoenzyme crystal structure, π-stacking interactions between the aromatic side chain of Tyr114 and the bound PLP exist (Fig. 3c, magenta). As with Lys212, we have also observed a large displacement in the Tyr114 residue of the apoenzyme structure (Fig. 3c, green). This may suggest that the π-stacking interactions between
Fig. 3. (a) H2S production by wild-type CSE and various mutant CSE proteins with mutagenic alterations to Lys212 and Tyr114. The results are displayed as the mean from three independent runs ±SD. ⁎P b 0.05 compared to wild-type GST-tagged CSE enzyme. #P b 0.05 compared to the corresponding alanine mutant protein. (b and c) Structural alignment of the CSE holoenzyme (magenta) and apoenzyme (green). Large displacements in polypeptide chains around Lys212 (b) and Tyr114 (c) are observed. The figure was produced by the program PyMOL.28
712
Modulation of H2S Production in CSE
Fig. 4. (a) H2S production by wild-type CSE and various mutant CSE proteins with mutagenic alterations to Tyr60 and Arg62. Results are displayed as the mean from three independent runs ±SD. ⁎P b 0.05 compared to wild-type GST-tagged CSE enzyme. #P b 0.05 compared to the corresponding alanine mutant protein. (b) The human CSE holoenzyme tetramer made up of a dimer of dimers. Interactions between PLP and Tyr60 and Arg62 from the adjacent monomer are shown for two of the enzymatic subunits C and D, which are shown in pale green and pink, respectively. The figure was produced by the program PyMOL.28
Tyr114 and PLP are required for preserving the active-site structure in the CSE holoenzyme. Thus, we speculate that the lack of H2S production from the Y114A mutant was most likely attributed to a disruption of the π-stacking interactions. Without these stabilizing interactions between Ala114 and PLP in the mutant enzyme, it is likely that the productive geometry of the active site is decreased, leading to the observed mutant enzyme inactivity. In the case of Y114F, we hypothesize that the increase in the production of H2S is due to weaker π-stacking interactions between the phenylalanine ring and PLP; this would be further elaborated on subsequently.
In this study, the effects of mutagenic alterations to Tyr60 and Arg62 on the catalysis of H2S production were also explored. Initial mutageneses to alanine (Y60A and R62A) showed that production of H2S was reduced compared to production of the wild-type enzyme (Fig. 4a). However, enzyme activity was not restored despite a subsequent mutation to threonine and lysine, respectively, even though these residues possessed similar sidechain properties (hydroxyl or basic for tyrosine and arginine, respectively) (Fig. 4a). From the crystal structure of the CSE holoenzyme (Fig. 4b), these two residues are extensively involved in hydrogen-bond contacts with the PLP cofactor from the neighboring
Fig. 5. (a) H2S production by wild-type CSE and various mutant CSE proteins with mutagenic alterations to Asp187 and Thr189. The results are displayed as the mean from three independent runs ±SD. ⁎P b 0.05 compared to wild-type GST-tagged CSE enzyme. #P b 0.05 compared to the corresponding alanine mutant protein. (b) Hydrogen-bonding network between Thr189, Asp187 and the pyridoxal nitrogen group of PLP. The figure was produced by the program PyMOL.28
713
Modulation of H2S Production in CSE
ionic interactions can be formed between the αcarboxyl group of L-cystathionine and Arg375.22 In the case of the α,β-elimination reaction of cysteine, we originally speculated that the guanidinium side chain of Arg375 of human CSE would play a similar role by forming ionic interactions with the carboxylate group of the L-cysteine substrate. Within expectations, the mutation of Arg375 to alanine indeed resulted in a loss of H2S production (Fig. 2b). However, a subsequent mutation to lysine (R375K) failed to restore any enzyme activity (data not shown), suggesting the importance of the fully conserved Arg375 residue rather than the positive charge of the side chain in maintaining enzyme activity. From structure-based sequence alignments between yeast CSE and bacterial or plant cystathionine-β-lyase (CBL), Messerschmidt et al. predicted that the specificity of the enzyme for the α,γelimination or α,β-elimination of L-cystathionine would depend on the hydrophobicity of the 339th residue.22 From the sequence alignment of CSE and CBL, this residue correlated with more hydrophobic residues such as valine and tyrosine in plant or bacterial CBL, respectively (Supplementary Fig. 1). In our study, mutation of the Glu339 residue in human CSE to a more hydrophobic residue such as alanine or tyrosine caused an increase in the α,βelimination reaction on L-cysteine. The net production of H2S was increased by about 6-fold for these two mutant proteins when compared to wild-type CSE (Table 1). Interestingly, the production of H2S was not compromised even upon mutation of the negatively charged Glu339 residue to the positively charged lysine amino acid (Table 1). This observation stresses that the extent of hydrophobicity— rather than the charge of this residue—determines the rate of H2S production by CSE. From Table 1, enzyme activity, as reflected by catalytic efficiency (kcat/Km), correlated with the logP values of various mutants (indicating the degree of hydrophobicity of the mutated residue). As the hydrophobicity of the mutated residue increased, a corresponding 1.8fold, 3.2-fold and 7.2-fold increase in the catalytic efficiency of E339K, E339A and E339Y, respectively, was observed in comparison to that of the wild-type enzyme. In addition, although the catalytic efficiency of E339A was lower than that of E339Y, the firstorder rate constant of product formation, as reflected by the turnover number (kcat) of E339A, was about twice that of E339Y. This was, however, offset by an increased Michaelis–Menten constant
monomer, thus forming stable dimers and the active-site interface between two monomers. Sizeexclusion chromatography and analytical ultracentrifugation experiments that had been previously performed also showed that the apoenzyme exists as a weak tetramer (dimer of dimers) in solution.21 Thus, it is likely that mutation of Tyr60 and Arg62 would disrupt the active-site interface and the interaction between adjacent monomers, leading to a loss in enzyme activity. Indeed, alignment of various CSE homologs and transulfuration enzymes showed that these two residues are fully conserved (Supplementary Fig. 1), hence reflecting their importance in maintaining enzyme activity. The crystal structure of the CSE holoenzyme revealed that the pyridoxal nitrogen group of the PLP cofactor is hydrogen bonded to the carboxylate side chain of Asp187 (Fig. 5b). This hydrogenbonding interaction has been proposed to be critical in stabilizing the positive charge of the pyridoxal nitrogen group and in increasing the electrophilic character of the pyridoxal ring.22 Indeed, we observed a complete loss of H2S production upon mutation of Asp187 to alanine (Fig. 5a). A corresponding mutation of Asp187 to the acidic glutamate residue did not restore enzyme activity (Fig. 5a). We attribute this lack of activity to the longer side chain that glutamate possesses, as an additional methylene group may render the carboxyl group in an unfavorable position for hydrogen bonding to PLP. Alanine mutation of Thr189, whose hydroxyl side chain is in turn hydrogen bonded to Asp187 (Fig. 5b), was also observed to drastically reduce H2S production (Fig. 5a). However, a corresponding mutation of Thr189 to serine was found to restore the enzyme activity by about 60% (Fig. 5a). This suggests that the hydroxyl side chain of Thr189 (or Ser189 in T189S) is crucial and possibly positions Asp187 in an optimal position for a stronger interaction with the pyridoxal nitrogen of the cofactor. Partial restoration in the enzyme activity of T189S suggests that the methyl group in threonine most probably conformationally restricts the hydroxyl group to an optimal direction for hydrogen bonding to Asp187, leading to a productive activesite geometry within the wild-type enzyme. Residues affecting the binding of substrate or the release of products Based on docking experiments of L-cystathionine into the active site of yeast CSE, it was proposed that
Table 1. Correlation between the logP values of residues at the 339th position and kinetic parameters for the production of H2S from wild-type and various CSE mutants
Increasing hydrophobicity ↓
Residue at the 339th position
logPa
kcat (s− 1)
Km (mM)
kcat/Km (mM− 1 s− 1)
Net production of H2S (nmol)
Glutamate (wild-type CSE) Lysine (E339K) Alanine (E339A) Tyrosine (E339Y)
−3.7 −3.1 −2.9 −2.3
0.12 0.34 1.44 0.76
1.9 3.3 7.7 1.8
0.06 0.11 0.19 0.43
8.6 15.5 50.2 51.3
The net production of H2S from each of these proteins was determined after 30 min of incubation of the proteins with the substrate at 37 °C. a Values obtained from http://www.ecosci.jp/amino/amino2j_e.html.27
714 (Km), suggesting that the affinity of E339A for the Lcysteine substrate may be much lower than the affinity of E339A for the other mutants. However, once L-cysteine is bound to the enzyme, the catalysis of H2S production by E339A will occur much readily to produce the desired products. Recently, a similar mutation had been performed on this residue in yeast CSE to explore its effect on the catalytic efficiency of the enzyme during the α,γ-elimination
Modulation of H2S Production in CSE
reaction of L-cystathionine.25 Interestingly, a 30-fold reduction in catalytic efficiency was observed for the alanine and tyrosine mutants. This inverse correlation between the α,β-elimination or α,γ-elimination reaction of CSE and the hydrophobicity of the 339th residue may imply that residues with a more hydrophobic side chain at the 339th position could possibly favor the α,β-elimination reaction over the α,γ-elimination reaction in CSE.
Fig. 6. (a) Graphs of the initial reaction velocity V (expressed as U/mg purified human CSE, where 1 U = 1 μmol of H2S produced per minute) against the concentration of exogenous PLP determined under various concentrations of L-cysteine substrate. (b) Proposed mechanism for the catalysis of H2S production from L-cysteine by human CSE. The L-cysteine substrate first binds to the holoenzyme via a transaldimination reaction with Lys212 (Step 1). Lys212 then abstracts the αproton of the bound substrate (Step 2). This causes an influx of electron density into the pyridoxal ring, which in turn destabilizes the π-stacking interaction between Tyr114 and the pyridoxal ring. The C–β–S bond of the bound substrate is then cleaved, and H2S is released via a β-elimination process (Step 3). The release of the aminoacrylate intermediate (Step 4) for subsequent hydrolysis (Step 5a) is likely to be aided by a subsequent rebinding of free PLP so that the holoenzyme can be regenerated (Step 5b). (c) Comparison of the initial reaction velocity against the concentration of exogenous PLP between GST-tagged Y114F mutant and wild-type CSE enzyme in the presence of 1.5 mM L-cysteine substrate. Results are displayed as the mean from three independent runs ± SD.
Modulation of H2S Production in CSE
Catalysis of H2S production from wild-type CSE in the presence of varying PLP concentrations Previously, in the development of the H2S detection assay from rat liver or kidney tissue homogenates, the inclusion of exogenous PLP had been found to be crucial for maximal enzymatic activity.18 This was also observed in our purified human CSE assay where the rate of production of H2S was found to increase with an increase in exogenous PLP concentrations for all of the tested L-cysteine substrate concentrations (Fig. 6a). According to Yamagata et al., in addition to binding to the enzyme, Lcysteine may also bind to the exogenous PLP cofactor to form thiazolidines.26 This may offer an explanation for the subsequent decrease in reaction rates beyond an optimum concentration of exogenous PLP (approximately 50 μM; Fig. 6a). Under high concentrations of PLP, the probability of the L-cysteine substrate being bound to the exogenously added cofactor to form thiazolidines becomes higher. This, in turn, would decrease the effective concentration of the substrate, leading to a lower production of H2S. The extent of decrease in reaction rates was found to be greater under lower concentrations of L-cysteine, consistent with our expectations of the competitive nature of PLP–L-cysteine interactions vis-à-vis enzyme–substrate binding; when the amount of Lcysteine was increased, the extent of the competing side reaction relative to catalysis decreased, leading to the observed increase in reaction rates. The initial increase in reaction rates as exogenous PLP concentration increased was most probably not attributed to the enzyme being purified in apo form, since the reaction rates had been observed to increase progressively beyond an exogenous PLP concentration of 1.5 μM (which was the molar concentration of PLP required to fully saturate the amount of enzyme utilized in the assay) for each concentration of Lcysteine that was assayed. Previously, we have shown that addition of L-cysteine to the crystallization condition triggered the dissociation of PLP from our purified CSE enzyme, hence leading to crystallization of the enzyme in apo state.21 In addition, spectrophotometric experiments have indicated that the bound cofactor can be removed when L-cysteine is incubated with the enzyme, and that the holoenzyme can be regenerated upon addition of PLP to the apoenzyme.21 These findings are consistent with the observed increase in initial reaction rates as exogenous PLP concentration was increased in this study. It is hence likely that, during catalysis (H2S production from L-cysteine), the PLP cofactor is displaced from the active site, and that free PLP would have to reassociate with the enzyme active site for subsequent catalysis after a single turnover or a few turnovers (Fig. 6b, Step 5b). As the exogenous PLP concentration was increased, the proportion of PLPbound CSE would increase, leading to an increased rate of H2S production. The release of the cofactor may be triggered upon abstraction of the α-proton after binding of the L-cysteine substrate (Fig. 6b, Step 2). Although the mechanistic origin for the release of
715 the PLP cofactor remains unknown, it is observed that the rate of the α,β-elimination reaction of Lcysteine is much slower than the rate of the α,γelimination reaction of L-cystathionine. Hence, we postulate that a consequence of the β-elimination step and subsequent release of H2S (Fig. 6b, Step 3) could be the generation of a possible catalytically ineffective aminoacrylate intermediate. The displacement of this intermediate may be aided by an incoming molecule of PLP, which would establish more effective π-stacking interactions with Tyr114 so as to regenerate the catalytically competent holoenzyme (Fig. 6b, Steps 4 and 5b). Catalysis of H2S production by Y114F mutant CSE We sought further evidence for the displacement of the PLP cofactor during the catalytic cycle by assaying the Y114F mutant enzyme in the presence of varying concentrations of exogenous PLP. In comparison to wild-type CSE, the Y114F mutant enzyme produced more H2S under all assay conditions (Fig. 6c). An inspection of the side chains of Tyr114 (in wild-type CSE) and Phe114 (in Y114F) revealed a plausible explanation for the increase in H2S production: the former is more electron rich than the latter due to the resonance electrondonating effects of the para-hydroxyl group in tyrosine; thus, the initial π-stacking interactions between the electron-rich aromatic side chain and the electron-deficient pyridoxal ring would be stronger for the wild-type enzyme. As a consequence of the stronger π-stacking interactions between Tyr114 and the pyridoxal ring, the αproton of the bound L-cysteine would be less acidic, and electron delocalization into the pyridoxal ring would be less favored (Fig. 6b, Step 2). In contrast, the Y114F mutant would exhibit a weaker πstacking interaction that would facilitate electron delocalization into the pyridoxal ring. As such, the catalytic production of H2S by the Y114F mutant occurs at a rate higher than that observed for the wild-type enzyme (Fig. 3a). Due to electron delocalization into the pyridoxal rings, the π-stacking interactions could be weakened, and the dissociation of the aminoacrylate intermediate might be facilitated in the proceeding step; as a result, the much weakened π-stacking interactions of Phe114 with the pyridoxal ring of the aminoacrylate intermediate could translate to a faster dissociation of the aminoacrylate intermediate for the mutant enzyme. Such π-stacking interactions may hence also affect the rate of PLP reassociation with the apoenzyme subsequently. Indeed, a greater increase in reaction rates as exogenous PLP concentrations increased was observed for Y114F relative to wildtype CSE, corroborating our hypothesis (Fig. 6b). Implications to the modulation of endogenous H2S production Recent studies have shown that a number of diseases are associated with imbalances in endoge-
716 nous H2S levels. More importantly, H2S donor compounds and inhibitors of H2S production have been found to exhibit therapeutic potentials in various animal disease models,10 reinforcing the need for novel strategies in the design and development of activators and inhibitors of CSE towards modulating the in vivo production of H2S. In this study, elucidation of the roles of various active-site residues of CSE has been accomplished through sitedirected mutagenesis and PLP-dependent H2S production assays. Specifically, mutation of Tyr114 to phenylalanine, and mutation of Glu339 to alanine, tyrosine and lysine, respectively, have been found to increase H2S production. Our observations provide a novel route towards the screening of activators of CSE, whereupon compounds can be identified based on their ability to weaken the π-stacking interactions between Tyr114 and PLP, or to increase the hydrophobicity of Glu339, resulting in an increase in H2S production. Such compounds may include intercalating agents that possess electron-rich or electron-poor aromatic rings that disrupt the πstacking interactions between PLP and Tyr114, or specific alkylating agents that target the side chain of Glu339. With regard to the design of the inhibitors of H2S production, our observation that the PLP cofactor is displaced from the enzyme during the course of catalysis (H2S production) possibly affords a new strategy for designing an inhibitor that competes with the rebinding of PLP to the apo-CSE enzyme. Unlike PAG and BCA, which inhibit the enzyme by binding to PLP itself and would thus inhibit all other PLP-dependent enzymes (in an unwarranted nonspecific manner), such a new inhibitor may be more specific, since it targets and binds reversibly to the PLP cofactor binding site rather than to the cofactor itself. Since the mode of inhibition is centered on binding to the cofactor site upon displacement of PLP during the catalytic cycle, this new inhibitor would not directly inhibit the α,γ-elimination reaction of CSE (where displacement of the cofactor is unlikely to occur). Although the enzyme would likely be inactive towards the α,γ-elimination reaction after inhibition of the α,β-elimination reaction, this inhibition can be reversed by a competitive rebinding of the PLP cofactor. The selective and reversible inhibition of the α,βelimination reaction of CSE thus holds therapeutic potential, as endogenous H2S levels are lowered physiologically, whilst any indirect inhibition of the reverse transsulfuration pathway can be reversed and L-cysteine can still be produced by the enzyme. Hydrogen-bonding or ionic interactions that are afforded through the side chains of Tyr60, Arg62, Asp187 and Thr189 suggest an additional route of inhibitor design, through the incorporation of various functional groups into the inhibitor to enhance its binding to the enzyme. Taken together, our findings present novel strategies for the future design and development of potent and selective drug lead compounds that modulate the endogenous production of H2S.
Modulation of H2S Production in CSE
Materials and Methods Preparation of mutant CSE clones Mutant DNA strand synthesis was performed using the QuikChange site-directed mutagenesis kit (Stratagene) in accordance with the manufacturer's protocol. The sequence of the mutated DNA strands was then confirmed by automated DNA sequencing. Expression and purification of GST-tagged CSE proteins Mutant and wild-type GST-tagged CSE proteins were expressed and purified based on previously described procedures,21 with slight modifications. After bacterial expression of the GST-tagged fusion protein, the protein was purified with glutathione Sepharose beads and subsequently eluted with 1 mM reduced glutathione (Sigma) in 20 mM Tris (pH 8.0), 50 mM NaCl and 1 mM dithiothreitol. The protein was estimated to be at least 90% pure from SDS-PAGE analysis. Protein concentration was then determined by Bradford protein assay or with the aid of a NanoDrop ND-1000 spectrophotometer. H2S production assay The production of H2S from the GST-tagged mutant or wild-type CSE proteins was assayed using a procedure similar to that described by Stipanuk and Beck, with some modifications.18 Each reaction mixture (100 μL) consisted of 5 μg of the GST-tagged protein, sodium chloride (20 mM), PLP (0.5 mM), L-cysteine (2.75 mM) and sodium phosphate (pH 8.2) buffer (50 mM). After the incubation of parafilmed tubes at 37 °C for 30 min, a 100-μL mixture of 0.85% wt/vol zinc acetate and 3% wt/vol NaOH was added to each tube via a needle to trap evolved H2S gas. Enzymatic reactions were terminated by the addition of 100 μL of 10% wt/vol trichloroacetic acid. This was followed by the addition of N,N-dimethyl-p-phenylenediamine-dihydrochloride dye and FeCl3 to a final concentration of 2.5 mM and 3.3 mM, respectively, for the development of methylene blue and subsequent absorbance measurements at 670 nm. The amount of H2S produced was calculated against a calibration curve of sodium hydrosulfide (0–800 μM) and compared to that for control tubes without addition of the protein. For kinetic assays, the amount of H2S produced was determined at 5min intervals over a period of 20 min at 37 °C. The Michaelis–Menten constant (Km) and maximal velocity (Vmax) values were obtained by plotting the reciprocal of the initial reaction velocity V0 and the L-cysteine substrate concentration [S] according to the Lineweaver–Burk equation. The graphical software SigmaPlot (Systat) was used for all curve-fitting and regression analyses. Catalysis of H2S production under varying exogenous PLP concentrations The production of H2S was monitored at 5-min intervals over 20 min at 37 °C using a protocol similar to that described above. Each reaction mixture (100 μL) consisted of 3.88 μg of the purified human CSE enzyme, sodium chloride (20 mM), PLP (0–300 μM), L-cysteine (0.75–5 mM) and sodium phosphate (pH 8.2) buffer (50 mM). For comparison of enzyme activities between wild-type and
717
Modulation of H2S Production in CSE Y114F mutant proteins under varying exogenous PLP concentrations, the reaction mixture consisted of 3.88 μg of the GST-tagged enzyme, sodium chloride (20 mM), PLP (0–30 μM) and L-cysteine (1.5 mM). The amount of H2S produced at each time point was calculated against the calibration curve and subsequently plotted against time to determine the initial reaction velocity V0 under each Lcysteine or PLP concentration. The graphical software SigmaPlot (Systat) was used for all curve-fitting and regression analyses. Statistical analysis Data are shown as mean ± SD. Statistical analysis was performed using a two-tailed paired Student's t test.
8.
9.
10. 11.
12.
Acknowledgements This work was supported by Interfaculty Research and by grants from Office of Deputy President (Research and Technology)/Yong Loo Lin School of Medicine (R-183-000-240-101 and R-183-000-240720), Academic Research Fund (R-154-000-438-112), Ministry of Education (R183-000-195-112) and Biomedical Research Council A⁎STAR, Singapore (WBS R-143-000-350-305). S. Huang is a graduate scholar of the Ministry of Education, Singapore.
13.
14.
15.
16.
Supplementary Data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/ j.jmb.2009.11.058
References 1. Wang, R. (2002). Two's company, three's a crowd: can H2S be the third endogenous gaseous transmitter? FASEB J. 16, 1792–1798. 2. Kamoun, P. (2004). Endogenous production of hydrogen sulfide in mammals. Amino Acids, 26, 243–254. 3. Shibuya, N., Tanaka, M., Yoshida, M., Ogasawara, Y., Togawa, T., Ishii, K. & Kimura, H. (2009). 3-Mercaptopyruvate sulfurtransferase produces hydrogen sulfide and bound sulfane sulfur in the brain. Antioxid. Redox Signal. 11, 703–714. 4. Yang, G., Wu, L., Jiang, B., Yang, W., Qi, J., Cao, K. et al. (2008). H2S as a physiologic vasorelaxant: hypertension in mice with deletion of cystathionine gamma-lyase. Science, 322, 587–590. 5. Li, L., Whiteman, M., Guan, Y. Y., Neo, K. L., Cheng, Y., Lee, S. W. et al. (2008). Characterization of a novel, water-soluble hydrogen sulfide-releasing molecule (GYY4137): new insights into the biology of hydrogen sulfide. Circulation, 117, 2351–2360. 6. Mok, Y. Y., Atan, M. S., Cheong, Y. P., Wang, Z. J., Bhatia, M., Moochhala, S. & Moore, P. K. (2004). Role of hydrogen sulphide in haemorrhagic shock in the rat: protective effect of inhibitors of hydrogen sulphide biosynthesis. Br. J. Pharmacol. 143, 881–889. 7. Bhatia, M., Sidhapuriwala, J., Moochhala, S. M. & Moore, P. K. (2005). Hydrogen sulphide is a mediator
17.
18.
19. 20.
21.
22.
23.
of carrageenan-induced hindpaw oedema in the rat. Br. J. Pharmacol. 145, 141–144. Bhatia, M., Wong, F. L., Fu, D., Lau, H. Y., Moochhala, S. M. & Moore, P. K. (2005). Role of hydrogen sulfide in acute pancreatitis and associated lung injury. FASEB J. 19, 623–625. Collin, M., Anuar, F. B., Murch, O., Bhatia, M., Moore, P. K. & Thiemermann, C. (2005). Inhibition of endogenous hydrogen sulfide formation reduces the organ injury caused by endotoxemia. Br. J. Pharmacol. 146, 498–505. Szabó, C. (2007). Hydrogen sulphide and its therapeutic potential. Nat. Rev. 6, 917–935. Rose, W. C. & Wixom, R. L. (1955). The amino acid requirements of man. XIV. The sparing effect of cystine on methionine requirement. J. Biol. Chem. 216, 763–773. Flavin, M. (1971). Trans-sulfuration reactions: introduction. Methods Enzymol. 17B, 416–417. Uren, J. R., Ragin, R. & Chaykovsky, M. (1978). Modulation of cysteine metabolism in mice: effects of propargylglycine and L-cysteine-degrading enzymes. Biochem. Pharmacol. 27, 2807–2814. Vina, J., Vento, M., Garcia-Sala, F., Puertes, I. R., Gasco, E., Sastre, J. et al. (1995). L-Cysteine and glutathione metabolism are impaired in premature infants due to cystathionase deficiency. Am. J. Clin. Nutr. 61, 1067–1069. Wang, J., Huff, A. M., Spence, J. D. & Hegele, R. A. (2004). Single nucleotide polymorphism in CTH associated with variation in plasma homocysteine concentration. Clin. Genet. 6, 483–486. Wang, J. & Hegele, R. A. (2003). Genomic basis of cystathioninuria (MIM 219500) revealed by multiple mutations in cystathionine gamma-lyase (CTH). Hum. Genet. 112, 404–408. Zhu, W., Lin, A. & Banerjee, R. (2008). Kinetic properties of polymorphic variants and pathogenic mutants in human cystathionine gamma-lyase. Biochemistry, 47, 6226–6232. Stipanuk, M. H. & Beck, P. W. (1982). Characterization of the enzymatic capacity for cysteine desulphydration in liver and kidney of the rat. Biochem. J. 206, 267–277. Braunstein, A. E. & Goryachenkova, E. V. (1984). The β-replacement-specific pyridoxal-P-dependent lyases. Adv. Enzymol. 56, 1–89. Chiku, T., Padovani, D., Zhu, W., Singh, S., Vitvitsky, V. & Banerjee, R. (2009). H2S biogenesis by cystathionine gamma-lyase leads to the novel sulfur metabolites, lanthionine and homolanthionine, and is responsive to the grade of hyperhomocysteinemia. J. Biol. Chem. 284, 11601–11612. Sun, Q., Collins, R., Huang, S., Holmberg-Schiavone, L., Anand, G. S., Tan, C. H. et al. (2009). Structural basis for the inhibition mechanism of human cystathionine-gamma-lyase: an enzyme responsible for the production of H2S. J. Biol. Chem. 284, 3076–3085. Messerschmidt, A., Worbs, M., Steegborn, C., Wahl, M. C., Huber, R., Laber, B. & Clausen, T. (2003). Determinants of enzymatic specificity in the Cys-Metmetabolism PLP-dependent enzyme family: crystal structure of cystathionine γ-lyase from yeast and intrafamiliar structure comparison. Biol. Chem. 384, 373–386. Yamagata, S., Yasugahira, T., Okuda, Y. & Iwama, T. (2003). Conversion of the aminocrotonate intermediate limits the rate of γ-elimination reaction catalyzed
718 by L-cystathionine γ-lyase of the yeast Saccharomyces cerevisiae. J. Biochem. 134, 607–613. 24. Clausen, T., Huber, R., Prade, L., Wahl, M. C. & Messerschmidt, A. (1998). Crystal structure of Escherichia coli cystathionine γ-synthase at 1.5 Å resolution. EMBO J. 23, 6827–6838. 25. Farsi, A., Lodha, P. H., Skanes, J. E., Los, H., Kalidindi, N. & Aitken, S. M. (2009). Interconversion of a pair of active-site residues in Escherichia coli cystathionine γ-synthase, E. coli cystathionine βlyase, and Saccharomyces cerevisiae cystathionine γlyase and development of tools for the investigation
Modulation of H2S Production in CSE of their mechanisms and reaction specificity. Biochem. Cell Biol. 87, 445–457. 26. Yamagata, S., Isaji, M., Yamane, T. & Iwama, T. (2002). Substrate inhibition of L-cysteine α,β-elimination reaction catalyzed by L-cystathionine γ-lyase of Saccharomyces cerevisiae. Biosci. Biotechnol. Biochem. 66, 2706–2709. 27. Data Book of Molecules: Home Page for Learning Environmental Chemistry (n.d.), retrieved on April 28, 2009 from http://www.ecosci.jp/ec.html. 28. DeLano, W. L. (2002). The PyMOL Molecular Graphics System. DeLano Scientific LLC, Palo Alto, CA.