Specific reactions of S-nitrosothiols with cysteine hydrolases: A comparative study between dimethylargininase-1 and CTP synthetase

Abstract

S-Transnitrosation is an important bioregulatory process whereby NO⁺ equivalents are transferred between S-nitrosothiols and Cys of target proteins. This reaction proceeds through a common intermediate R–S–N(O⁻)–S–R′ and it has been proposed that products different from S-nitrosothiols may be formed in protein cavities. Recently, we have reported on the formation of such a product, an N-thiosulfoximide, at the active site of the Cys hydrolase dimethylargininase-1 (DDAH-1) upon reaction with S-nitroso-l-homocysteine (HcyNO). Here we have addressed the question of whether this novel product can also be formed with the endogenously occurring S-nitrosothiols S-nitroso-l-cysteine (CysNO) and S-nitrosoglutathione (GSNO). Further, to explore the reason responsible for the unique formation of an N-thiosulfoximide in DDAH-1 we have expanded these studies to cytidine triphosphate synthetase (CTPS), which shows a similar active site architecture. ESI-MS and activity measurements showed that the bulky GSNO does not react with both enzymes. In contrast, S-nitrosylation of the active site Cys occurred in DDAH-1 with CysNO and in CTPS with CysNO and HcyNO. Although kinetic analysis indicated that these compounds act as specific irreversible inhibitors, no N-thiosulfoximide was formed. The reasons likely responsible for the absence of the N-thiosulfoximide formation are discussed using molecular models of DDAH-1 and CTPS. In tissue extracts DDAH was inhibited only by HcyNO, with an IC₅₀ value similar to that of the isolated protein. Biological implications of these studies for the function of both enzymes are discussed.

Nitric oxide (NO) is an important signaling and messenger molecule in many diverse biological processes like neuronal function, cardiovascular regulation, and immune defense (Colasanti and Suzuki 2000). It has been suggested that endogenous NO is stabilized and stored by carrier molecules that prolong its half-life and preserve its biological activity (Gaston 1999; Hogg 2002). Endogenous low molecular weight thiols (RSH) such as Cys, l-homocysteine (Hcy), and glutathione (GSH) exhibit such a carrier function through the formation of S-nitrosothiols (RSNO), i.e., S-nitroso-l-cysteine (CysNO), S-nitroso-l-homocysteine (HcyNO), and S-nitrosoglutathione (GSNO) (see Fig. 1A). These compounds are believed to be involved in many biological functions, including NO storage, transport, and delivery (Gaston 1999; Hogg 2002; Wang et al. 2002). S-Nitrosothiols release NO in the presence of Cu⁺, ascorbate, or thiols, and can undergo S-transnitrosation reactions in which the direct transfer of NO⁺ equivalents between RSNO and RSH takes place (Wang et al. 2002). The latter may represent an important mechanism for the regulation of protein function through modification of specific Cys residues in target proteins (Gaston 1999; Martínez-Ruiz and Lamas 2004). In a number of enzymes the active site Cys has already been reported to be S-nitrosylated by S-nitrosothiols resulting in a protein-SNO, e.g., creatine kinase (Konorev et al. 2000), glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (Padgett and Whorton 1995), or rhinovirus 3c protease (Xian et al. 2000c).

Figure 1.

Chemical structures and reaction mechanisms that appear in this study. (A) Structures of ADMA (substrate of DDAH-1) and Gln (substrate of EcCTPS) in comparison with the S-nitrosothiols CysNO, HcyNO, and GSNO. (B) The reaction mechanisms for S-transnitrosation and the proposed formation of an N-thiosulfoximide (4). The product 4 is formed from the common intermediate 1 occurring in S-transnitrosation reactions (Singh et al. 1996; Knipp et al. 2005).

Recent investigations toward the understanding of the mechanism of S-transnitrosation using low molecular weight thiols revealed that this equilibrium reaction proceeds through an intermediate R–S–N(O⁻)–S–R′ (see Fig. 1B; Singh et al. 1996^; Wong et al. 1998^; Houk et al. 2003^; Perissinotti et al. 2005). The existence of this intermediate in a protein structure has been demonstrated on the example of the reaction of Cys298, located at the active site of aldose reductase, with S-nitroso-N-acetyl-d,l-penicillamine (SNAP) and N-(β-d-glucopyranosyl)-N ²-acetyl-S-nitroso-d,l-penicillaminamide (glyco-SNAP). However, this intermediate was found to dissociate into two different products, i.e., a protein-SNO and a mixed disulfide (Srivastava et al. 2001). It has been proposed that in a protein binding site the formed intermediate R–S–N(O⁻)–S–R′ may undergo further chemical reactions yielding different product(s) (Houk et al. 2003). Previously, we have reported on the formation of such a novel product, an N-thiosulfoximide (DDAH-1—Cys273:Sγ-NH—(O)Sδ:Hcy) within the active site of the Cys hydrolase N ^ω,N ^ω-dimethyl-l-arginine dimethylaminohydrolase-1 (DDAH-1, EC 3.5.3.18) upon its reaction with HcyNO (Fig. 1B; Knipp et al. 2005^; Frey et al. 2006). DDAH-1 hydrolyzes the side-chain methylated derivatives of Arg, N ^ω-methyl-l-arginine (MMA) and N ^ω,N ^ω-dimethyl-l-arginine (ADMA) to l-citrulline (Cit) and CH₃NH₃ ⁺ or (CH₃)₂NH₂ ⁺, respectively (Ogawa et al. 1989). Both substrates are endogenous competitive inhibitors of all three nitric oxide synthase (NOS) isoforms (Vallance and Leiper 2002; Böger et al. 2003). Thus, DDAH through the control of ADMA and MMA levels acts as a regulator of NO production (MacAllister et al. 1996; Knipp 2006). DDAH-1 belongs to the family of guanidino-modifying enzymes (Murray-Rust et al. 2001; Shirai et al. 2001). These enzymes such as peptidylarginine deiminase 4 (PAD4) (Arita et al. 2004), Arg:Gly amidinotransferase (AT) (Humm et al. 1997), and the bacterial enzyme arginine deiminase (ADI) (Das et al. 2004) are involved in the modification of Arg or its derivatives. All of them share a Cys, His-(Glu/Asp) catalytic triad and exhibit a similar active site architecture.

In view of the novel covalent product, an N-thiosulfoximide, formed in the reaction of DDAH-1 with HcyNO, we have addressed the question of whether such a product may also be formed in other classes of Cys hydrolases. In the search for a suitable candidate we have considered the following structural requirements of the active site of DDAH-1 (Murray-Rust et al. 2001; Frey et al. 2006): (1) an active site Cys acting as a nucleophile, (2) an activated water molecule participating in the hydrolysis of the reaction intermediate, and (3) a similar structure of the natural substrate to that of S-nitrosothiol(s). We found that cytidine triphosphate synthetase (CTPS) fulfills these requirements. CTPS (EC 6.3.4.2) belongs to the family of triad (Cys-His-Glu) glutamine amidotransferases (GATases) and catalyzes the formation of CTP from UTP and Gln in the presence of Mg²⁺, ATP, and GTP. At present, there are 16 members of GATases known, which are involved in the biosynthesis of nucleotides, amino acids, aminated sugars, coenzymes, and antibiotics (Massière and Badet-Denisot 1998). All GATases consist of a glutaminase domain and a synth(et)ase domain and show a high degree of sequence homology within the active sites of their glutaminase domains (Zalkin and Smith 1998). Akin to DDAH-1, in the crystal structure of CTPS the substrate Gln is anchored through its α-NH₃ ⁺/α-COO⁻ moieties in the active site pocket (Fig. 2A₁; Endrizzi et al. 2004^; Goto et al. 2004).

Figure 2.

Comparison of molecular models of DDAH-1 and EcCTPS in complex with CysNO and HcyNO. (A) Schematic representation of the active site of DDAH-1 in complex with the substrate ADMA (blue) and HcyNO overlaid (green) and (A₁) of the active site of the glutaminase domain of EcCTPS in complex with the substrate Gln (blue) and CysNO overlaid (green). The catalytic residues in A and A₁ are shown in red. Hydrogen bonds and salt bridges are indicated by dotted lines. Molecular models of DDAH-1 in complex with (B) CysNO and (C) HcyNO and of EcCTPS in complex with (B₁) CysNO and (C₁) HcyNO. Polar interactions in molecular models are indicated by dotted lines (blue). The ligands were docked into the active sites of the crystal structures of DDAH-1 (PDB: 2CI5) and EcCTPS (PDB: 1S1M) and the structures were energy minimized using the program CNS (Brunger et al. 1998). For a detailed description of docking and energy minimization see Supplemental material.

The aim of this study was to shed light onto the so far unique formation of an N-thiosulfoximide in the reaction of DDAH-1 with HcyNO. Therefore, the products of DDAH-1 upon reaction with CysNO, HcyNO, and GSNO were compared with those formed in CTPS. ESI-MS and enzyme kinetics were used in the characterization of the products and the determination of kinetic constants. The absence of the N-thiosulfoximide formation is discussed using molecular models of DDAH-1 and CTPS. Biological implications of these studies for the function of both enzymes are discussed.

Results

Molecular modeling of HcyNO and CysNO in complex with DDAH-1 and EcCTPS

To address the question whether CysNO, by analogy to HcyNO (Fig. 1A), could form the N-thiosulfoximide product at the active site of DDAH-1 and whether the formation of such a product is unique to DDAH-1, we have built molecular models of DDAH-1 and CTP synthetase from Escherichia coli (EcCTPS) with S-nitrosothiols. EcCTPS possesses an active site architecture similar to DDAH-1, and its substrate Gln is structurally related to CysNO (Figs. 1A, 2A₁). Molecular models in which HcyNO and CysNO were docked into the active site of mammalian DDAH-1 and EcCTPS were built based on their crystal structures and energy minimized (Fig. 2; Endrizzi et al. 2004^; Goto et al. 2004^; Frey et al. 2006). Modeling of GSNO was not attempted as it is too bulky to enter both active site clefts (for details, see the Supplemental material).

In DDAH-1 the residues involved in the catalytic reaction comprise the active site Cys273 and the cocatalytic residues His172 and Asp126. The model shows that both CysNO and HcyNO may form a hydrogen bridge between His172 and the nitroso oxygen of their S-NO group (Fig. 2B,C). However, the distance between the active site Cys273:Sγ and HcyNO:Nɛ of 2.9 Å is significantly smaller compared to that of CysNO:Nδ (3.8 Å). Therefore, with CysNO the formation of an R–S–N(O⁻)–S–R′ intermediate required for the enzyme modification should not occur. This is in agreement with the finding that the compound (S)-2-amino-4-(3-methylguanidino)-butanoic acid (4124W), a structural analog of MMA lacking one methylene group, is a competitive inhibitor (MacAllister et al. 1996).

In EcCTPS the catalytic triad consists of the active site Cys379 and the cocatalytic residues His515 and Glu517 (Endrizzi et al. 2004; Goto et al. 2004). The molecular model revealed that both CysNO and HcyNO may form hydrogen bond interactions with the main chain NH groups of Gly352 and Leu380, which form the oxyanion hole (Fig. 2B₁,C₁). Distances between Cys379:Sγ and CysNO:Nδ or HcyNO:Nɛ are 3.5 and 2.6 Å, respectively, and hence comparable to the distance between Cys379:Sγ and Gln:Cδ (3.1 Å). This model analysis suggests that CysNO and HcyNO are not excluded from reacting with the EcCTPS active site Cys residue.

Effect of CysNO, HcyNO, and GSNO on the enzymatic activity of DDAH-1 and EcCTPS

To compare the extent of DDAH-1 and EcCTPS inhibition by S-nitrosothiols, the enzymes were incubated for 30 min with 500 μM of CysNO, HcyNO, or GSNO and the residual enzymatic activity determined. As seen in Figure 3A,A₁ the enzymatic activity of both enzymes is affected by CysNO and HcyNO, but not by GSNO. The reaction of DDAH-1 with HcyNO inhibited the enzymatic activity by 96%. However, CysNO that lacks one methylene group compared to HcyNO reduced the activity only marginally (9%–15%). In marked contrast, the incubation of EcCTPS with CysNO and HcyNO inhibited the activity by 94% and 90%, respectively (Fig. 3A₁). For both enzymes, the absence of inhibition with GSNO is presumably due to its inability to enter the active sites.

Figure 3.

Effect of S-nitrosothiols on the enzymatic activity of DDAH-1 and EcCTPS in the absence and presence of inhibitor or substrate. Inhibition of (A) DDAH-1 and (A₁) EcCTPS by HcyNO, CysNO, and GSNO (500 μM). The effect of CysNO and HcyNO (500 μM) on the activity of (B) Zn²⁺-free and Zn²⁺-containing DDAH-1and (B₁) EcCTPS in the presence of 1 mM or 10 mM Gln, respectively.

Localization of the site modified by CysNO and HcyNO within DDAH-1 and EcCTPS

We have previously reported that DDAH-1 is modified at the active site Cys273 by HcyNO (Knipp et al. 2005; Frey et al. 2006). To establish that the inhibition of DDAH-1 and EcCTPS by CysNO and HcyNO is due to the modification of the active site Cys, we have performed enzymatic activity measurements in the presence of an inhibitor or substrate.

In the case of DDAH-1 we have used Zn²⁺ to protect the active site Cys273 from modification by CysNO and HcyNO (Fig. 3B). The protection was also observed with the substrate MMA, but to a lesser extent due to its lower affinity (K _M = 7.0 × 10⁻⁵ M) (Knipp 2006) compared to Zn²⁺ (K_d = 4.2 × 10⁻⁹ M) (Knipp et al. 2001). Zn²⁺ is a reversible competitive inhibitor of DDAH-1 that is tightly bound to the active site Cys273 (Knipp et al. 2001; Frey et al. 2006; Stone et al. 2006). We have shown that Zn²⁺ binding to DDAH-1 protects the enzyme from S-nitrosylation by free NO (Knipp et al. 2003). The bound Zn²⁺ is efficiently removed by small chelating agents such as phosphate or imidazole, but its removal by bulky EDTA or 1,10-phenantroline is a very slow process (Knipp et al. 2001). Although the covalent modification of Cys273 by HcyNO was already shown, the role of Zn²⁺ in this reaction was not examined. The Zn²⁺-containing form of DDAH-1 was prepared as described by Knipp et al. (2001) and was then preincubated with CysNO or HcyNO in HEPES buffer. Subsequently, the enzymatic activity was determined by diluting the preincubation mixture into a large volume of imidazole buffer in the presence of 8.3 mM MMA (∼23 × K _M). Both CysNO and HcyNO at 500 μM concentrations had no influence on the enzymatic activity (Fig. 3B; Supplemental Fig. S1) indicating that the active site Cys273 of DDAH-1 is the target of modification by these S-nitrosothiols.

To identify the site of modification in EcCTPS, the enzyme was incubated with 500 μM CysNO and HcyNO in the presence of 1 mM (4 × K _M) or 10 mM Gln. The results revealed that, in the presence of 1 mM Gln, CysNO reduced the enzymatic activity by 88% and by 32% in the presence of 10 mM Gln. Similar studies performed with HcyNO resulted in reduction of the activity by 43% and 19%, respectively (Fig. 3B₁). The results suggest that the substrate Gln competitively protects the active site of EcCTPS from the modification with CysNO and HcyNO. This is in agreement with the active site Cys379 being the target of modification by these S-nitrosothiols.

Characterization of DDAH-1 and EcCTPS upon reaction with S-nitrosothiols

The products of the reaction of the enzymes DDAH-1 and EcCTPS with CysNO, HcyNO, and GSNO were characterized by ESI-MS. The deconvoluted mass spectra are shown in Figure 4. Upon incubation of both enzymes with GSNO, only the masses of the unmodified enzymes were detected (Fig. 4A, [DDAH-1 + H]⁺: m = 31,200.4 Da; calculated, 31,200.7 Da; Figure 4A₁, [CTP synthetase + H]⁺: m = 60,244.0 Da; calculated, 60,244.0 Da). The lack of enzyme modification by GSNO is in agreement with the preserved enzymatic activity of both enzymes (see above). At this point is should be noted that, as isolated, EcCTPS represents a mixture of two enzyme forms with mass peaks of 60,244.0 Da and 60,375.0 Da (Fig. 4A₁). The mass difference of 131.0 Da corresponds to a processed and unprocessed N-terminal Met1. Consequently, both forms showed the same mass peak pattern in the ESI-MS spectra (Fig. 4A₁–C₁).

Figure 4.

Deconvoluted nano-ESI-MS spectra of DDAH-1 and EcCTPS modified with S-nitrosothiols. DDAH-1 and EcCTPS were incubated with 500 μM of (A,A₁) GSNO, (B,B₁) CysNO, and (C,C₁) HcyNO. The resulting reaction products are depicted in the spectra. The mass peak at 60,375.0 Da observed in the spectrum (A₁) corresponds to the mass of the enzyme in which the N-terminal Met1 (Δm = 131.0 Da) is not processed.

The ESI-MS spectra obtained upon incubation of DDAH-1 with CysNO revealed a major mass peak of unmodified DDAH-1 and a second small mass peak (Δm = 29.1 Da) corresponding to the formation of a single S-nitroso adduct (Fig. 4B). The absence of an S-nitroso adduct when Zn²⁺-containing DDAH-1 was used indicates that modification of the active site Cys273 occurred (data not shown). In contrast, incubation of DDAH-1 with HcyNO results in a predominant mass peak (Δm = 163.9 Da) (Fig. 4C), originating from the previously characterized covalent product, an N-thiosulfoximide (Knipp et al. 2005).

Similar studies performed on EcCTPS revealed major mass peaks of 60,272.5 Da (Δm = 28.5 Da) upon incubation with CysNO (Fig. 4B₁) and 60,273.0 Da (Δm = 29.0 Da) upon incubation with HcyNO (Fig. 4C₁). Based on the effect of these S-nitrosothiols on the enzymatic activity presented above (Fig. 3B₁), S-nitrosylation of the active site Cys379 occurred. In general, HcyNO is chemically more stable than CysNO. The observed lower intensity mass peaks of double and triple S-nitrosylated species with CysNO (Fig. 4B₁) may be attributed to an unspecific S-nitrosylation of two Cys residues by free NO released from CysNO under the conditions of sample preparation. In contrast to the structure of DDAH-1, the structure of EcCTPS reveals two solvent accessible Cys (Cys228, Cys268) residues, which may account for the observed effect (see Supplemental material). However, in neither case was an N-thiosulfoximide formed.

Time dependence of DDAH-1 and EcCTPS inhibition by CysNO and HcyNO

To assess the specificity of DDAH-1 and EcCTPS inactivation by CysNO and HcyNO, we have performed kinetic measurements. In the case of CysNO we found that DDAH-1 upon 30 min of incubation with CysNO is inhibited with an IC₅₀ of ∼5 mM (Supplemental Fig. S1). Because of the very low inhibitory efficiency no kinetic studies could be performed.

To examine the time-dependent DDAH-1 inactivation by HcyNO, we used the dilution assay method described by Kitz and Wilson (1962). This method is based on the incubation of the enzyme with an inhibitor and the measurement of the residual enzymatic activity as a function of time. In this case the enzyme-inhibitor mixture is diluted into a large volume of standard assay buffer containing an excess of the natural substrate. The obtained results reveal that DDAH-1 is inactivated by HcyNO in a time- and concentration-dependent pseudo first-order process, which is in agreement with the covalent modification of DDAH-1 (Figs. 5A, 4C). The apparent first-order rate constants (k _obs) were calculated from Figure 5A according to Equation 1 (see Materials and Methods) and plotted versus the inhibitor concentrations. Since no saturation kinetics were obtained in the concentration range used (0–500 μM), the data were plotted according to the method of Kitz and Wilson (1962). The observed nonzero y-intercept in the Kitz–Wilson plot (Fig. 5B) supports a two-step mechanism for the reaction between the enzyme (E) and the inhibitor (I), in which the formation of a Michaelis-like complex (E·I) is preceded by a slower inactivation event leading to the product E–I.

Figure 5.

Determination of the kinetic parameters for the inactivation of DDAH-1 by HcyNO and EcCTPS by CysNO. (A) Semilogarithmic plot of the time-dependent inhibition of DDAH-1 with 70 (•), 100 (▴), 150 (▪), 200 (○), 300 (△), and 500 (□) μM HcyNO. At different incubation times, aliquots were removed and the residual enzymatic activity was determined. (B) Kitz-Wilson plot of the time-dependent inactivation of DDAH-1 by HcyNO (Kitz and Wilson 1962). (A₁) Analysis of progress curves of EcCTPS inactivation by CysNO. The product formation was followed over 90 min in the presence of 0 (▪), 5 (•), 10 (▴), 19 (□), 38 (○), and 75 (△) μM CysNO. (B₁) The apparent k _obs values obtained from the plot (A₁) were multiplied by (1 + [S]/K _M) and plotted versus the inhibitor concentration.

From the Kitz–Wilson plot the kinetic parameters K _I = 0.69 mM and k _inact = 0.38 min⁻¹ were obtained according to Equation 2 (see Materials and Methods). Taken together, HcyNO represents a specific inhibitor of DDAH-1 that irreversibly modifies the active site Cys273.

The kinetic parameters for the inactivation of EcCTPS were obtained from the analysis of progress curves of the enzymatic reaction in the presence of inhibitor. Both CysNO and HcyNO showed a time- and concentration-dependent inhibition of EcCTPS activity (Fig. 5A₁; Supplemental Fig. S2A). In the case of CysNO the obtained saturation kinetic indicates a two-step inactivation process with a K _I = 0.027 mM and k _inact = 0.48 min⁻¹ (Fig. 5B₁), which is typical for specific irreversible inhibitors as seen also for DDAH-1 inactivation by HcyNO. In contrast, no saturation kinetic was obtained for EcCTPS in the reaction with HcyNO, which contains an extra methylene group compared to CysNO (Supplemental Fig. S2B). This behavior is sometimes observed for specific irreversible inhibitors for which the reversible formation of the Michaelis-like complex (E·I) is kinetically insignificant relative to the rate of inactivation (Kitz and Wilson 1962; Copeland 2000). The specificity of the inhibition by HcyNO is supported by the competition experiments with Gln (see above). The irreversibility of the EcCTPS inactivation by CysNO and HcyNO is supported by the fact that its activity could not be recovered after extensive dialysis of the sample and by the covalent enzyme modification seen in ESI-MS experiments (Fig. 4B₁,C₁).

The obtained inactivation parameters for the reaction of DDAH-1 with HcyNO and that of EcCTPS with CysNO and HcyNO are summarized in Table 1. These parameters are compared with those available for other Cys hydrolases. The second-order rate constants (k _inact/K _I) for the inactivation of DDAH-1 and EcCTPS by HcyNO of 551 and 650 M⁻¹min⁻¹, respectively, lie well in the range reported for arylamine N-acetyltransferase (Dairou et al. 2003), papain (Xian et al. 2000a), PTP-1B (Xian et al. 2000b), and rhinovirus 3C protease (Xian et al. 2000c), all inactivated by GSNO or SNAP. In all these cases the kinetic specificity reflected by k _inact and the binding specificity reflected by K _I are within the same order of magnitude. However, the inactivation of EcCTPS by CysNO revealed a significantly higher second-order rate constant (k _inact/K _I) of 17,778 M⁻¹min⁻¹, which results from a higher binding specificity of CysNO to EcCTPS reflected by the low K _I of 27 μM.

Table 1.

Kinetic parameters for the inactivation of Cys hydrolases by S-nitrosothiols

Reversibility of the DDAH-1 and EcCTPS inactivation

The activity of a number of enzymes modified by S-nitrosothiols could be restored with DTT (Konorev et al. 2000; Xian et al. 2000b; Leiper et al. 2002; Dairou et al. 2003). To examine the stability of the N-thiosulfoximide modification to reducing agents the covalently modified DDAH-1 was incubated with 5 mM of Cys, Hcy, GSH, DTT, or TCEP for 30 min at 37°C (Supplemental Fig. S3A). However, the enzymatic activity could not be restored with any of these compounds. This suggests that the covalent product N-thiosulfoximide will remain stable under the reducing conditions present in the cytosol. In contrast, the incubation of EcCTPS modified by CysNO and HcyNO with 5 mM DTT for 30 min at 37°C revealed that 88% and 97%, respectively, of the original activity could be recovered (Supplemental Fig. S3B).

Influence of S-nitrosothiols on DDAH-1 activity in tissue extract

In hyperhomocyst(e)inemia, a risk factor predisposing to the development of cardiovascular diseases, elevated levels of NOS inhibitors ADMA/MMA and a decreased NO production were observed (Böger et al. 2000). Although different underlying mechanisms for the adverse action of Hcy in hyperhomocyst(e)inemia are discussed, the reason for this effect remains unknown (Lentz et al. 2003). The previous cell culture studies suggested that the enzymatic activity of DDAH is inhibited by Hcy (Stühlinger et al. 2001; Selley 2004). Based on the incubation of DDAH from Pseudomonas aeruginosa (PaDDAH) with biotinylated Hcy in vitro followed by immunochemical analysis, it has been concluded that PaDDAH inhibition occurred through the formation of a mixed disulfide between Hcy and the active site Cys249 (Stühlinger et al. 2001). In view of the covalent enzyme inhibition by HcyNO, we have examined the effect of Hcy on DDAH-1 activity upon its preincubation with 100 μM Hcy at 37°C for 30 min at pH 7.3. However, in our studies no changes in DDAH-1 activity could be observed in the presence and absence of air O₂. In addition, no covalent product was detected in ESI-MS upon incubation of DDAH-1 with up to 5 mM Hcy under the same conditions (data not shown). Thus, we conclude that no significant amounts of a mixed disulfide can be formed even in the presence of air O₂. The obtained results are in agreement with the crystal structure of mammalian DDAH-1 in complex with Hcy where no formation of a mixed disulfide was observed (Frey et al. 2006). However, the irreversible inhibition of DDAH-1 by HcyNO may be a possible explanation for the observed adverse effect of elevated Hcy levels. In this case endogenously formed HcyNO would have to reduce DDAH-1 activity even in the presence of GSH, which represents the main pool of intracellular free thiols (Meister 1974). To adress this question we have examined the inhibition of DDAH activity by HcyNO in tissue extract.

The first characterization of DDAH-1 has been reported for the enzyme isolated from rat kidney, where it was found highly abundant (∼0.11 mg/g tissue; Ogawa et al. 1989). Therefore, we used tissue extract of bovine kidney to assess the inhibitory potential of S-nitrosothiols on the activity of DDAH. To avoid thiol oxidation the cells were opened in an argon atmosphere. Subsequently, the concentration of free thiols was determined. The obtained value of 20 ± 1 nmol/mg of protein corresponds to 2.1 mM of free thiols, which constitute mainly the GSH concentration. To examine the effect of S-nitrosothiols on the enzymatic activity of DDAH, bovine kidney tissue extracts were incubated with increasing concentrations (0–300 μM) of CysNO, HcyNO, and GSNO. The results show that both GSNO and CysNO were without effect on DDAH activity (Fig. 6). In marked contrast, HcyNO inhibited the enzymatic activity of DDAH in a concentration-dependent manner with an IC₅₀ of 75 ± 11 μM (Fig. 6). A similar IC₅₀ value was also obtained in independent measurements using isolated DDAH-1, suggesting that HcyNO is stable in tissue extract and preferentially inhibits DDAH in the presence of millimolar GSH concentrations (Fig. 6).

Figure 6.

Effect of S-nitrosothiols on the enzymatic activity of isolated DDAH-1 and in tissue extracts of kidney. Tissue extracts of bovine kidney were preincubated with 0–300 μM of CysNO (▴), HcyNO (•) and GSNO (▪) for 30 min at 37°C in 50 mM HEPES/NaOH (pH 7.3), 150 mM KCl, 5 mM EDTA prior to activity measurements. (○) Concentration-dependent inhibition of isolated DDAH-1 by HcyNO under the same experimental conditions.

Discussion

The obtained data provide insights into the reaction of the three S-nitrosothiols CysNO, HcyNO, and GSNO with the Cys hydrolases DDAH-1 and EcCTPS. In these enzymes the active site Cys residues carry out a nucleophilic attack on the C atom of the guanidino group (ADMA and MMA for DDAH-1) or the C atom of the amide (Gln for EcCTPS), respectively (Figs. 1A, 2). The subsequent hydrolysis reaction leads to the formation of a carbamide (Cit) or carboxylate (Glu) and the release of an amine or ammonia, respectively. A key common feature of both enzymes is the presence of a long, channel-like substrate cavity and the substrate anchoring at its entrance. As a result, the site of reaction is buried inside the protein structure. These structural features are responsible for the absence of inhibitory effect by the bulky GSNO. However, CysNO and HcyNO were found to react with both enzymes. In the case of DDAH-1 inhibition by CysNO an IC₅₀ of ∼5 mM was observed, indicating a very low inhibitory potency. This is in agreement with our molecular model where the CysNO:Nδ is located too far away for nucleophilic attack by DDAH-1-Cys273:Sγ. At present, the mechanism of NO-transfer from CysNO remains unclear.

An N-thiosulfoximide was formed only in the reaction of DDAH-1 with HcyNO. Previously, we have suggested that the structural and chemical features of HcyNO can be used in the development of an inhibitor for DDAH-1 and related enzymes such as ADI and PAD4 (Knipp et al. 2005). At present, there are only a few inhibitors for DDAH-1 known, but these have not been kinetically characterized (Knipp et al. 2005; Rossiter et al. 2005; Vallance et al. 2005; Knipp 2006). The only exception is 2-chloroacetamidine, which was tested with PaDDAH and PAD4 and revealed second-order rate constants of 387 M⁻¹min⁻¹ and 35 M⁻¹min⁻¹, respectively (Stone et al. 2005). In the present study, the second-order rate constant of 551 M⁻¹min⁻¹ for the irreversible inactivation of DDAH-1 by HcyNO was obtained. This may represent a further benchmark for the future design and evaluation of DDAH-1 inhibitors.

In contrast to the N-thiosulfoximide product formed at the active site Cys273 in the reaction of DDAH-1 with HcyNO, the active site Cys379 of EcCTPS was modified by CysNO and HcyNO through S-transnitrosation. The analysis of the kinetic data showed that CysNO acts as a specific irreversible inhibitor of EcCTPS in a two-step process with a high second-order inactivation rate of 17,778 M⁻¹min⁻¹. In contrast, the obtained kinetic data for HcyNO would be in agreement with a single-step inactivation process. This is usually observed for nonspecific irreversible inhibitors reacting with residues on the enzyme surface whereby no enzyme-inhibitor complex (E·I) is formed (Aldrige 1950; Tipton 1989). However, based on the substrate protection studies, HcyNO reacts at the active site (Fig. 3B₁). Single-step inactivation kinetics is sometimes seen also for specific inhibitors such as small molecule affinity labels, which react with active site residues. In this case the formation of the enzyme-inhibitor complex (E·I) is kinetically insignificant relative to the rate of inactivation (Copeland 2000). For example, this has been demonstrated by Kitz and Wilson (1962) for the compound methanesulfonyl fluoride that irreversibly inhibits the enzyme acetylcholinesterase.

A comparison of the molecular models may provide an insight into the absence of an N-thiosulfoximide upon the reaction of EcCTPS with CysNO and HcyNO. In the mechanism accounting for the formation of an N-thiosulfoximide in DDAH-1, the release of OH⁻ from the (R–S–N(OH)–SR′) intermediate 2 (Fig. 1B) has been suggested (Knipp et al. 2005). Therefore, the initially formed intermediate 1 (R–S–N(O⁻)–S–R′) must become protonated to form 2. This would enable OH⁻ release, resulting in the intermediate 3 (R–S–N=S⁺–R′). In our model of DDAH-1 in complex with HcyNO, the calculated distance of 2.7 Å between the cocatalytic His172 and HcyNO:Oζ would be consistent with the formation of a salt bridge between the imidazolium ion His172 and the oxyanion 1 allowing the formation of 2 (Fig. 2C). Although the active sites of DDAH-1 and EcCTPS show a high degree of similarity, an important difference between both enzymes is the oxyanion hole in EcCTPS (Endrizzi et al. 2004; Goto et al. 2004). In the hydrolysis of Gln the oxyanion hole stabilizes the formed oxyanion of the acyl intermediate (Fig. 2A₁). In the reaction with CysNO and HcyNO this oxyanion hole would also stabilize the intermediate 1 formed with Cys379 of EcCTPS. In contrast to DDAH-1, the oxyanion hole would prevent the protonation of oxyanion 1, giving rise to the observed product EcCTPS-Cys379:Sγ-NO. Hence, it would appear that the special architecture of the DDAH-1 active site and the different reaction mechanism of this enzyme compared to those of most other Cys hydrolases are likely responsible for the formation of a so far unique N-thiosulfoximide. Because of similar active site architecture of DDAH-1 with those of other members of the family of guanidino-group modifying enzymes, e.g., ADI and PAD4, the formation of an N-thiosulfoximide upon the reaction with S-nitrosothiols is likely.

Our data show the formation of an S-nitrosothiol at the active site Cys379 of EcCTPS incubated with CysNO and HcyNO. S-Nitrosylation of critical Cys residues in proteins is an important mechanism for the protein function and enzyme regulation (Gaston 1999). It has been suggested that the metabolic fate of Arg might be controlled by S-nitrosylation of several functionally related enzymes (Hess et al. 2005). This has already been demonstrated for the enzymes DDAH-1 (Leiper et al. 2002; Knipp et al. 2003), argininosuccinate synthetase (Hao et al. 2004), ornithine decarboxylase (Bauer et al. 2001), S-adenosylmethionine decarboxylase (Hillary and Pegg 2003), and methionine adenosyltransferase (Pérez-Mato et al. 1999). In this context the family of GATase enzymes represents a so far unrecognized target of S-nitrosylation. GATases are involved in the synthesis of many biomolecules, including Arg, the substrate of NOS. In addition, the regulation of the arginine-NO pathway by Gln has also been shown in cell culture studies (Meininger and Wu 1997; Kakoki et al. 2006). Since all members of GATases share a high degree of sequence homology in their glutaminase active sites (Zalkin and Smith 1998; Chittur et al. 2001), we propose that besides EcCTPS other members of GATases are likely targets for an S-transnitrosation reaction.

Hcy is known as an independent risk factor for cardiovascular diseases and was found associated with endothelial dysfunction. The underlying mechanism is still unclear and discussed controversially (Lentz et al. 2003). However, in both humans and cell cultures a clear correlation between elevated Hcy and ADMA/MMA levels accompanied by reduced levels of NO has been reported (Böger et al. 2000; Böger 2001; Stühlinger et al. 2001, 2003; Yoo and Lee 2001; Selley 2004). This observation was linked to the direct inhibition of DDAH activity through the formation of a mixed disulfide between the active site Cys and Hcy (Stühlinger et al. 2001). However, in our experiments we were unable to confirm the formation of such a product. Thus, the question arose of whether covalent inhibition of DDAH-1 by HcyNO may be responsible for this effect. Considering that under normal physiological conditions the cytosolic HcyNO concentrations are in the nanomolar range (Tsikas et al. 1999), an inhibition of DDAH-1 by HcyNO is unlikely. However, in hyperhomocyst(e)inemia plasma Hcy levels are markedly increased (in severe cases >100 μM) (Stanger et al. 2003), which likely results in increased intracellular Hcy levels. In the cytosol, NO is bound mainly to GSH, forming a large pool of GSNO (Zhang and Hogg 2004). However, both the fast exchange of NO⁺ between Hcy and GSNO (Zhang and Hogg 2004) and its preferential binding to Hcy (Tsikas et al. 1999) suggest that, under the pathophysiological conditions of hyperhomocyst(e)inemia, increased intracellular levels of HcyNO may exist. The stability of HcyNO in the cytosol is supported by our data showing the inhibition of DDAH-1 by HcyNO in the tissue extract of bovine kidney even in the presence of millimolar GSH concentrations. In addition, our observation that HcyNO inhibited the enzymatic activity with a IC₅₀ value of ∼75 μM similar to that obtained for isolated DDAH-1 is consistent with the preferential inhibition of this enzyme (Fig. 6). Thus, the inactivation of DDAH by HcyNO may represent a mechanism for the impairment of the NOS pathway in patients with elevated levels of Hcy. Taken together, the obtained results on the reaction of DDAH-1 and EcCTPS with S-nitrosothiols deepen our understanding of the function of these enzymes under physiological and pathophysiological conditions.

Materials and Methods

Protein purification

DDAH-1 was purified from bovine brain and the Zn²⁺-containing and Zn²⁺-free enzyme were prepared as previously described (Knipp et al. 2001). EcCTPS was purified from E. coli grown in LB medium to three-quarter log phase and purified according to the method described by Anderson (1983) with the following modifications. Both hydrophobic interaction chromatography steps were omitted, and the protein was further purified by size exclusion chromatography on a HiLoad 26/60 Superdex 200 column (Pharmacia) equilibrated with 0.1 M KH₂PO₄/K₂HPO₄ (pH 7.5), 1 mM EDTA, 4 mM Gln, and 2 mM DTT. The active fractions were pooled and applied to a Mono Q HR 5/5 column (Pharmacia) equilibrated with 0.02 M KH₂PO₄/K₂HPO₄ (pH 7.7), 1 mM EDTA, 4 mM Gln, 2 mM DTT, and eluted with a linear gradient from 0.02 to 0.5 M KH₂PO₄/K₂HPO₄ (pH 7.7). The final purification step was performed on a HiLoad 26/60 Superdex 75 column equilibrated with 60 mM HEPES/KOH (pH 8.0), 1 mM ATP, 0.2 mM GTP, 1 mM UTP, 1 mM EDTA, and 2 mM DTT. The isolated EcCTPS was found >90% pure as judged by SDS-PAGE and ESI-MS. The protein was stored in 60 mM HEPES/KOH (pH 8.0), 1 mM ATP, 0.2 mM GTP, 1 mM UTP, 1 mM EDTA, 2 mM DTT, and 20% (v/v) glycerol. Prior to the experiments DTT was removed by dialysis against argon-saturated 0.2 M KH₂PO₄/K₂HPO₄ (pH 7.3), 1 mM EDTA.

Enzymatic activity assays

Prior to activity measurements, DDAH-1 samples were saturated with N₂. All other solutions were rendered metal free by passing over a Chelex 100 column (Bio-Rad) and O₂ free by three freeze–pump–thaw cycles on a vacuum line. All further steps were carried out in an N₂-atmosphere. Activity measurements were performed in polystyrene 96-microwell plates and the colorimetric product formed upon the conversion of Cit determined at 532 nm using a microplate reader as previously described (Knipp and Vašák 2000).

If not otherwise stated EcCTPS activity was measured in argon saturated 60 mM HEPES/KOH (pH 7.3) buffer containing 10 mM MgCl₂, 5 mM EDTA, 10 mM Gln, 1 mM ATP, 0.2 mM GTP, and 1 mM UTP in a final volume of 1 mL. If not otherwise stated 10 μL of EcCTPS (1–10 μM) were added to the assay mixture and preincubated at 37°C for 3 min. Subsequently, the conversion of UTP to CTP (Δ ɛ₂₉₁ = 1338 M⁻¹cm⁻¹) was recorded in a quartz cuvette in a spectrophotometer at 291 nm as previously described (Long and Pardee 1967).

Preparation of l-homocysteine

Hcy was prepared from l-homocysteine thiolactone·HCl according to the method of Duerre and Miller (1966) in a N₂-atmosphere. Briefly, 1 mmol (153 mg) of l-homocysteine thiolactone·HCl (Fluka) was dissolved in 1 mL of 5 M NaOH and the solution was incubated for 5 min at 37°C. Subsequently, the pH was adjusted to neutral by the addition of 2 M HCl. Finally, 50 mM TES/NaOH (pH 7.4) was added to a final volume of 5 mL (Büdy et al. 2001). The product was characterized by MS and NMR. The yield (96%) was estimated through the quantification of thiols using the method of Grasetti and Murray (1967).

Preparation of S-nitrosothiols

S-Nitrosothiols of Cys, Hcy, and GSH were always freshly prepared using the protocol given by Feelisch and Stamler (1996). Briefly, 2 mL of 0.2 M thiols were mixed with 2 mL 0.2 M NaNO₂ and acidified by the addition of 1 mL 2 M HCl. Subsequently, 0.5 mL 1 M HEPES and 0.1 mL 0.5 M Na₄-EDTA were added and the pH was adjusted to 7.3 through dropwise addition of 5 M KOH. Excess of NO was removed by saturating the solution with argon. The concentration of CysNO, HcyNO, and GSNO was determined photometrically using the molar extinction coefficients ɛ₅₄₄ = 14.9 M⁻¹cm⁻¹, ɛ₅₄₅ = 16.7 M⁻¹cm⁻¹, and ɛ₅₄₄ = 17.2 M⁻¹cm⁻¹, respectively (Feelisch and Stamler 1996). The products were characterized by ESI-MS, UV-vis, and NMR and found stable for at least 90 min under the experimental conditions.

Influence of S-nitrosothiols on the activity of DDAH-1 and EcCTPS

DDAH-1 (30 μM) was incubated with 500 μM of CysNO, HcyNO, or GSNO. Incubations were performed in 50 mM HEPES/NaOH (pH 7.3), 150 mM KCl, and 5 mM EDTA for 30 min at 37°C. Subsequently, 1 μL of the protein solution was diluted with 59 μL of 50 HEPES/NaOH (pH 7.3), 150 mM KCl, 5 mM EDTA, and 8.3 mM MMA·HOAc and incubated for 30 min at 37°C. The amount of Cit formed was determined as described (Knipp and Vašák 2000).

EcCTPS (0.2 μM) was incubated with 500 μM of CysNO, HcyNO, or GSNO for 30 min at 37°C in 60 mM HEPES/KOH (pH 7.3), 10 mM MgCl₂, 5 mM EDTA in the presence of 1 mM ATP, 0.2 mM GTP, and 1 mM UTP. Subsequently 50 μL of preincubation mixture were diluted with 950 μL of assay buffer, and the enzymatic activity was followed as described above. To assess the irreversibility of inactivation, a control and EcCTPS inactivated with 500 μM CysNO or HcyNO were dialyzed for 18 h against argon-saturated 60 mM HEPES/KOH (pH 7.3), 10 mM MgCl₂, and 5 mM EDTA at 4°C. The residual activity was determined using the standard assay method (see above).

Inhibitor and substrate protection experiments

Zn²⁺-DDAH-1 (45 μM) was incubated with 500 μM of CysNO or HcyNO in 250 mM imidazole/HCl (pH 7.3), 5 mM EDTA or 250 mM HEPES/NaOH (pH 7.3), 5 mM EDTA, respectively. After 30 min at 37°C, 1μL of the sample was mixed with 59 μL of 250 mM imidazole/HCl (pH 7.3), 8.3 mM MMA·HOAc and the enzymatic activity determined (see above). It may be noted that the applied imidazole concentration generates the active Zn²⁺-free form of DDAH-1 in situ (Knipp et al. 2001).

To compare the effect of CysNO and HcyNO on the activity of EcCTPS, the enzyme (0.02 μM) was incubated with 500 μM of CysNO or HcyNO in the presence of either 1 mM Gln or 10 mM Gln, respectively. After 60 min the amount of CTP formed was determined and compared with the control.

ESI-MS analysis of DDAH-1 and EcCTPS incubated with S-nitrosothiols

For MS analysis 4.5 μL of DDAH-1 (10–30 μM) were incubated with 0.5 μL of 5 mM S-nitrosothiol solution for 20 min at 37°C in a N₂-atmosphere. Subsequently, the samples were measured either immediately or kept at −80°C until measurement. Prior to MS the samples were desalted using a C₄-ZipTip (Millipore) and eluted with 8 μL of H₂O:CH₃CN:HCOOH (30:70:0.1). Afterward, 4 μL of the sample were diluted into 16 μL of H₂O:CH₃CN:HCOOH (50:50:0.1). Nano-ESI-MS analyses of the protein solutions were performed on a Q-TOF Ultima API mass spectrometer. The solutions were infused through a fused silica capillary (ID 75 μm) at a flow rate of 0.5 μL/min. Electrospray PicoTips (ID 30 μm) were obtained from New Objective. Mass spectra were acquired by scanning an m/z range from 700 to 2000 with a scan duration of 1 sec and an interscan delay of 0.1 sec. Spray voltage was set to 2.1 kV, cone voltage to 50 V, and RF lens 1 energy to 50 V. Mass spectra were deconvoluted using the MaxEnt 1 software. For MS analysis 4.5 μL of EcCTPS (10–20 μM) were incubated with 0.5 μL of 5 mM S-nitrosothiol solution for 20 min at 37°C in a N₂-atmosphere. Prior to MS the samples were desalted using a C₄-ZipTip (Millipore) and eluted with 8 μL of H₂O:CH₃CN:HCOOH (25:75:0.2). MS measurements were performed as described for DDAH-1 (see above). All mass spectra were obtained at least in triplicate.

Enzyme inhibition kinetics

Time-dependent inhibition measurements of DDAH-1 modified with HcyNO were performed by the dilution assay method described by Kitz and Wilson (1962). DDAH-1 (8.4 μM) in 50 mM HEPES/NaOH (pH 7.3), 150 mM KCl, 5 mM EDTA was preincubated with 0, 70, 100, 150, 200, 300, and 500 μM HcyNO at 37°C. In intervals of 3 or 5 min, 1 μL of the sample was diluted with 59 μL of 50 mM HEPES/NaOH (pH 7.3), 150 mM KCl, 5 mM EDTA, 8.3 mM MMA·HOAc, 10% (v/v) glycerol and the residual enzymatic activity was determined as described (Knipp and Vašák 2000). The kinetic parameters k _inact and K _I were obtained by fitting the data to

and

where E represents the residual activity after time t, E ⁰ is the initial activity, k _obs is the observed first-order rate constant, k _inact is the first-order rate constant for the formation of the irreversibly inhibited enzyme, K _I = k ₁/k ₋₁ is the dissociation constant for the Michaelis-like complex formed, and [I] the inhibitor concentration.

Inactivation of EcCTPS by CysNO and HcyNO was analyzed using the progress curves method (Copeland 2000). EcCTPS (0.01 μM) in 1 mL argon-saturated 60 mM HEPES/KOH (pH 7.3) containing 10 mM MgCl₂, 5 mM EDTA, 1 mM Gln, 1 mM ATP, 0.2 mM GTP, and 1 mM UTP was incubated with CysNO (0, 5, 10, 19, 38, 75 μM) or HcyNO (0, 50, 100, 150, 300, 500 μM). The product formation was followed for 90 min and the obtained progress curves were fitted to

where A₂₉₁ is the absorbance at 291 nm, v _i is the initial velocity, k _obs is the apparent first-order rate constant. The apparent first-order rate constant was multiplied by (1 + [S]/K _M) to obtain the true k _obs (Copeland 2000). The Michaelis constant (K _M) was determined under the assay conditions used in this study (Supplemental Fig. S4). The inactivation parameters k _inact and K _I = k ₁/k ₋₁ were obtained from Equation 2.

Influence of S-nitrosothiols on DDAH activity in tissue extract

Fresh bovine kidney was obtained from the slaughterhouse and stored at −80°C in 0.3 M sucrose. Pieces of tissue were mixed 1:2 (w/v) with 50 mM HEPES/NaOH (pH 7.3), 150 mM KCl, 5 mM EDTA (argon saturated) and homogenized under argon using an Elvejheim-Potter homogenizer. The tissue extract was centrifuged for 10 min at 10,000g at 4°C. The supernatant was centrifuged for another 30 min at 21,000g at 4°C. The protein concentration of the samples was determined by the Bio-Rad protein assay. The concentration of free thiols was determined by the method of Grasetti and Murray (1967) after protein precipitation with 5% (w/v) trichloroacetic acid (TCA). To determine the enzymatic activity of DDAH-1, 30 μL of tissue extract were mixed with 55 μL of 50 mM HEPES/NaOH (pH 7.3), 150 mM KCl, 5 mM EDTA (argon saturated) and incubated with 0, 40, 70, 100, 150, 200, and 300 μM of S-nitrosothiols for 30 min at 37°C. Subsequently, 10 μL of 80 mM MMA·HOAc were added and the samples were incubated for another 30 min at 37°C. The reaction was stopped by mixing with 5 μL 100% (w/v) TCA and the precipitated protein was removed by centrifugation for 5 min at 2900g. The supernatant (60 μL) was transferred to a 96-well microplate and the activity determined as described (Knipp and Vašák 2000).

Electronic supplemental material

The Supplemental material includes a detailed description of the molecular modeling and four Supplemental figures. Figure S1 provides a closer examination of DDAH-1 inhibition by CysNO in the presence and absence of Zn²⁺. Figure S2 presents the time-dependent inactivation of EcCTPS by HcyNO. Figure S3 shows the effect of reducing agents on modified DDAH-1 and EcCTPS, and Figure S4 the determination of the K _M value for Gln hydrolyzed by EcCTPS under the experimental conditions used in this study.