Regulation of the Catalytic Activity and Structure of Human Thioredoxin 1 via Oxidation and S-Nitrosylation of Cysteine Residues*

The mammalian cytosolic/nuclear thioredoxin system, comprising thioredoxin (Trx), selenoenzyme thioredoxin reductase (TrxR), and NADPH, is the major protein-disulfide reductase of the cell and has numerous functions. The active site of reduced Trx comprises Cys32-Gly-Pro-Cys35 thiols that catalyze target disulfide reduction, generating a disulfide. Human Trx1 has also three structural Cys residues in positions 62, 69, and 73 that upon diamide oxidation induce a second Cys62–Cys69 disulfide as well as dimers and multimers. We have discovered that after incubation with H2O2 only monomeric two-disulfide molecules are generated, and they are inactive but able to regain full activity in an autocatalytic process in the presence of NADPH and TrxR. There are conflicting results regarding the effects of S-nitrosylation on Trx antioxidant functions and which residues are involved. We found that S-nitrosoglutathione-mediated S-nitrosylation at physiological pH is critically dependent on the redox state of Trx. Starting from fully reduced human Trx, both Cys69 and Cys73 were nitrosylated, and the active site formed a disulfide; the nitrosylated Trx was not a substrate for TrxR but regained activity after a lag phase consistent with autoactivation. Treatment of a two-disulfide form of Trx1 with S-nitrosoglutathione resulted in nitrosylation of Cys73, which can act as a trans-nitrosylating agent as observed by others to control caspase 3 activity (Mitchell, D. A., and Marletta, M. A. (2005) Nat. Chem. Biol. 1, 154–158). The reversible inhibition of human Trx1 activity by H2O2 and NO donors is suggested to act in cell signaling via temporal control of reduction for the transmission of oxidative and/or nitrosative signals in thiol redox control.

served from archaebacteria to man (1,2). The reduced or dithiol form Trx-(SH) 2 utilizes its exposed nucleophilic Cys N residue to attack a target protein disulfide and form a transient mixed disulfide, which is subsequently attacked by the Cys C residue to generate oxidized Trx-S 2 and the reduced protein (Reaction 1). Trx-S 2 is a substrate for NADPH and thioredoxin reductase (TrxR) (Reaction 2).

REACTIONS 1 AND 2
Escherichia coli Trx has only two Cys residues in the conserved active site (1), but mammalian cytosolic thioredoxins (Trx1) have additional structural cysteines. For human Trx1, they are located in positions 62, 69, and 73 (see Fig. 1). Post-translational modifications of Trx1 via these cysteine residues, such as thiol oxidation, glutathionylation, and S-nitrosylation have been proposed to regulate function and are implicated in signal transduction pathways (3,4). For instance, the loss of Trx activity and aggregation of the protein after oxidation was reported for the first time during purification (5), and it was later shown to be due to the formation of a second disulfide motif (6). Watson et al. (6) showed that under oxidizing conditions, besides the active site disulfide, a second disulfide may form between Cys 62 and Cys 69 with a considerable effect on Trx1 activity. These modifications are mediated by reactive oxygen species (ROS) and reactive nitrogen species (RNS), which are involved in numerous processes such as cell signaling (7,8). However, excessive ROS and RNS are also involved in pathogenesis of several diseases under oxidative or nitrosative stress (9 -12). The role of ROS in cell signaling, specially via the effect on mitogen-activated protein kinases and activation of nuclear transcription factors, is well known (13,14). For example, H 2 O 2 is involved in signaling pathways (15,16); however, it is not precisely understood how H 2 O 2 functions as a signaling molecule in vertebrates (17). Proteins containing thiol groups are one of the molecular targets of H 2 O 2 causing selective thiol oxidation (18). NO is also a reactive radical gas that is implicated in numerous physiological states (19 -22) as well as pathophysiologic conditions (23,24). NO exerts its effects via cGMP formation as well as cGMP-independent mechanisms such as nitration and nitrosylation of proteins (25).
S-Nitrosylation is a post-translational modification, involved in the regulation of structure or activity of certain proteins (26,27). Human Trx1 is such a protein, but the available data are controversial. In two studies, S-nitrosylation of hTrx1 at only one structural Cys residue was reported. Haendeler et al. (28) detected the S-nitrosylation of Cys 69 under basal conditions, which was asserted to be essential for anti-apoptotic and redox regulatory function of the protein and to increase Trx catalytic activity. But, Mitchell and Marletta (29) showed the oxidation of active site cysteines to a disulfide as well as the extensive nitrosylation of only Cys 73 after treating human Trx1 with 50 molar equivalents of S-nitrosoglutathione (GSNO) followed by the biotin switch method and matrix-assisted laser desorption ionization mapping. They could not detect any significant S-nitrosylation of Cys 62 and Cys 69 . S-Nitrosylation of Cys 73 was implicated in the reversible and specific transfer of a nitrosothiol between Trx (Cys 73 ) and caspase 3 (Cys 163 ) in vitro (29). This transnitrosation reaction was suggested to be involved in the regulation of apoptosis. On the other hand, in three other studies, the S-NO content of hTrx was increased after incubation with GSNO in a concentration-dependent manner to a maximum of about 2 mol of S-NO/mol of hTrx (30 -32). Sumbayev (30) showed that S-nitrosylation of Trx correlated directly with apoptosis signal-regulating kinase 1 activation. In another study, comparable with the endogenous S-NO content in endothelial cells that had been utilized by Haendeler et al. (28), only 0.3 Ϯ 0.09 mol of S-NO/mol of Trx was detected. However, under this condition, the cardio-protective effect of hTrx1 was markedly enhanced (31). Moreover, in a recent study regarding the crystal structure of human Trx1, only Cys 62 of GSNO-treated sample became stably nitrosylated in the crystal at neutral pH; Cys 69 could also be nitrosylated but required a more alkaline pH (33).
In this paper we studied the reaction of human Trx1 with diamide and H 2 O 2 as well as GSNO, which is a substrate for mammalian TrxR and was previously shown to inhibit the protein-disulfide reductase activity of the mammalian Trx system (34). We hypothesized that this effect could be due to the S-nitrosylation of Trx (28,29), inhibiting its activity rather than stimulating activity. Our results also suggest a mechanism for reversible inactivation of thioredoxin via H 2 O 2 generated from NADPH oxidases in signaling.

MATERIALS AND METHODS
Chemicals and Enzyme-Bovine serum albumin, diamide, DTNB, DTT, NADPH, and insulin were purchased from Sigma. The NAP-5 Sephadex G-25 column was purchased from Amersham Biosciences. NuPAGE Bis-Tris gels, NuPAGE LDS sample buffer (4ϫ), and running buffer were from Invitrogen. GSNO was prepared as previously reported (35) by the reaction of acidified sodium nitrite and reduced glutathione. The yield was calculated using the extinction coefficient of 920 M Ϫ1 cm Ϫ1 at 335 nm. E. coli Trx1 and TrxR (36), wild type (wt) hTrx1, and its mutants C62S, C32S/C35S,and C62S/C73S were prepared as described (37,38). E. coli P34H Trx was also utilized (39). Rat recombinant TrxR1 was kindly provided by Olle Rengby and Dr. Elias Arner from the Department of Medical Biochemistry and Biophysics of the Karolinska Institutet (40).
Reduction of Human Trx1 and Treatment with Oxidants-The protein (100 M in TE buffer (50 mM Tris-HCl, 1 mM EDTA, pH 7.4)) was incubated with 100 molar equivalents of DTT for 15 min at 37°C for reduction. Thereafter, DTT was removed by chromatography using a NAP-5 column equilibrated with TE buffer.
After confirming the complete reduction of Trx1 by measuring the number of free thiol groups (as described below), it was incubated with different concentrations of either diamide for less than 2 min at room temperature or H 2 O 2 for 15 min at 37°C. For the latter, DTT-treated Trx1 was desalted on a NAP-5 Sephadex G-25 column equilibrated with 50 mM Tris-HCl, pH 7.4, or phosphate-buffered saline, pH 7.4 with or without 1 mM EDTA.
S-Nitrosylation Method-The fully reduced Trx1, prepared as described above, was incubated with 10 molar equivalents of GSNO at 37°C for 60 min, which was followed by a desalting step on a column of NAP-5 Sephadex G-25 equilibrated with TE buffer.
Spectrophotometric Analysis of GSNO-treated Thioredoxin-GSNO-treated Trx samples were analyzed spectrophotometrically to quantitate the number of nitrosothiols, using a molar extinction coefficient of 920 M Ϫ1 cm Ϫ1 at 335 nm.
Localization of Nitrosothiols-Using wt and different mutants of hTrx1 including C32S/C35S, C62S/C73S, and C62S, the localization of nitrosothiols was investigated. E. coli Trx was also utilized to investigate whether active site cysteines can be nitrosylated.
Determination of Total Thiol Groups-DTNB was used to determine the number of free thiols in reduced, oxidized, or nitrosylated thioredoxin samples (41). Briefly, 10 M of reduced, oxidized, or nitrosylated protein was incubated with 6 M guanidine hydrochloride and 1 mM DTNB in a final volume of 500 l of 0.2 M Tris-Cl, pH 8.0. A molar extinction coefficient of 13,600 M Ϫ1 cm Ϫ1 at 412 nm was used to calculate the number of free thiols.
For diamide-treated Trx1, because diamide had not been removed from the sample, the reaction between diamide and free thionitrobenzoate, which is the product of the reaction between thiols and DTNB, was also studied to rule out any bleaching effect of diamide and subsequent interference with the determination of free thiol content. Guanidine hydrochloride 6 M in 0.1 M Tris-HCl, pH 8.0, and 1 mM DTNB were mixed with 25 M DTT. DTT was excluded from the reference cuvette. After recording the initial absorbance at 412 nm, diamide with a final concentration of 100 M was added to both the reference and sample cuvettes, and changes at 412 nm were recorded. The decrease of absorbance with a rate of 0.040 min Ϫ1 indicated the oxidation of thionitrobenzoate with diamide (data not shown); however, it did not interfere with the determination of free thiol content because of the slow rate of the reaction.
Activity of Oxidized Thioredoxin as a Substrate for Thioredoxin Reductase-Oxidized or S-nitrosylated hTrx1 was tested as a substrate for TrxR by following the oxidation of NADPH (Reaction 2). The experiment was performed at 25°C using 100 M NADPH and either 20 M oxidized or nitrosylated Trx1 in a final volume of 500 l of TE buffer. Trx1 was excluded from the reference cuvette. The reaction was started by adding 10 nM rat TrxR, and the oxidation of NADPH was monitored by following A 340 nm .
Insulin Disulfide Bond Reduction Activity of Thioredoxin-Two assays were utilized to study the effects of oxidative posttranslational modifications of Trx1 on the protein disulfide bond reductase activity of the Trx system. First, as described before (42), we mixed 200 M NADPH and 160 M insulin in a final volume of 500 l of TE buffer. In 1-cm semi-micro quartz cuvettes, 2 M of Trx sample was added to the mixture at 25°C. Trx1 was excluded from the reference cuvette. After starting the reactions by the addition of 10 nM rat recombinant TrxR to all samples, A 340 nm was recorded to calculate and compare the velocity of the reactions.
Insulin precipitation assay was also utilized as described (43). Insulin disulfide bonds can be reduced by DTT leading to the precipitation of reduced insulin, and this reaction is catalyzed by Trx. This activity of Trx was compared between oxidized and nitrosylated molecules using 160 M insulin, 2 M of either oxidized or nitrosylated Trx, and 1 mM DTT in a final volume of 500 l of 50 mM potassium phosphate buffer, pH 7.0, and 2 mM EDTA. Increase in turbidity was followed by recording A 650 nm .
TrxR Activity Assay-In a final volume of 500 l of TE buffer, we mixed 5 mM DTNB, 0.2 mM NADPH, and varying concentrations of either H 2 O 2 or GSNO. The reactions were started by adding 5 nM rat TrxR to all samples, and A 412 nm was monitored for 10 min. TrxR was excluded from the reference cuvette.
SDS-PAGE-NuPAGE 12% Bis-Tris gels were utilized for SDS-PAGE electrophoresis to study the redox state of reduced and oxidized human Trx1 samples. Fifteen g of reduced Trx1 as well as samples treated with different concentrations of either diamide or H 2 O 2 were loaded on the gel, and electrophoresis was performed with 200 V for 45 min. DTT was excluded from the sample buffer, and a nonreducing gel was utilized.

RESULTS
Oxidation of Human Thioredoxin 1 by Diamide-Chemically reduced human Trx1, with five thiols, was incubated with different concentrations of diamide for 2 min at room temperature; thereafter, the number of free thiols was determined. A decrease in the number of free thiols was recorded by increasing the ratio between diamide and Trx1 (Fig. 1B). With 10 molar equivalents of diamide, all of the cysteine residues were oxidized, and essentially no free thiol was detected in the Trx1.
When the structure and redox state of diamide-treated human Trx1 was investigated using nonreducing SDS-PAGE electrophoresis, it was seen that oxidation by diamide led to a complex mixture of monomer, mainly dimers, but also oligomers. The reduced protein was as expected at 12 kDa. The effect of diamide on C62S/C73S Trx1, which has only one structural cysteine (Cys 69 ), showed a progressive dimerization (data not shown).
The effect of diamide on the activity of human Trx1 was measured using insulin disulfides as a substrate and recycling of Trx-S 2 by NADPH and TrxR (Reactions 1 and 2). With one and three molar equivalents of diamide, the initial activity of Trx1 was decreased to 57 and 6% of the control sample, respectively ( Fig. 2A). The inhibition of Trx catalysis was transient, and activity was regained after a lag phase as seen from the progress curves. The duration of the lag phase was a function of diamide concentrations. For the Trx1 C62S/C73S mutant protein (Fig.  2B), the activity was not affected by diamide, indicating that the dimers generated by Cys 69 oxidation were active. No further studies of this protein were made. The results are in agreement with those of Watson et al. (6), who, using His-tagged Trx1 showed that diamide generated a second structural disulfide Cys 62 -Cys 69 with a loss of activity. Obviously, diamide is a strong oxidant inducing profound changes in thioredoxin. In particular, the generation of multimers as observed in this study was not evident from the native gel electrophoresis by Watson et al. (6). The structure of reduced human thioredoxin 1 and its oxidation by diamide. A, the three-dimensional structure of reduced human Trx1 (Protein Data Bank code 1ERT). B, human Trx1 (100 M) was reduced after incubation with 10 mM DTT for 15 min at 37°C. After removing the excess DTT with chromatography on a NAP-5 column, reduced Trx1 in TE buffer was treated with different ratios of diamide; thereafter the number of free thiols was measured by using 1 mM DTNB in 6 M guanidine hydrochloride. The inset shows the redox state of diamide-treated Trx1 in nonreducing SDS gel electrophoresis.

Oxidation of Human Thioredoxin 1 by H 2 O 2 -
The oxidation of human Trx1 by H 2 O 2 was studied using the same method as described for diamide, except excluding EDTA from the buffer and changing the incubation time to 15 min and the temperature to 37°C. Similar to the effect of diamide, a progressive decrease in the number of sulfhydryl groups was observed (Fig.  3). However, the number of free thiols did not reach zero. Even with a 100-fold molar excess of H 2 O 2 , one free thiol was left in the structure of human Trx1.
The redox state of Trx1 after treatment with H 2 O 2 was investigated using nonreducing SDS-PAGE electrophoresis as described. Remarkably, H 2 O 2 -treated Trx1 even after treatment with 100 molar equivalents of H 2 O 2 was still in the monomeric form. The same result was obtained with and without EDTA present; EDTA presence decreased the oxidation rate. Thus, H 2 O 2 gives rise to clean two-disulfide monomers.
To study the effect of H 2 O 2 on Trx activity, 100 M human Trx1 was also incubated with 10, 50, or 100 molar excess of H 2 O 2 for 15 min at 37°C and then desalted on a NAP-5 column, and the activity was measured. A progressive inhibition of Trx1 activity was recorded by increasing the concentration of H 2 O 2 (Fig. 4), which was reversible, and Trx1 regained its activity after a lag phase. The duration of lag phase was again a function of the H 2 O 2 concentration.
S-Nitrosylation of Human Thioredoxin 1 by GSNO-Reduced Trx1 was treated with 10 molar equivalents of GSNO for 60 min at 37°C. After chromatography on a NAP-5 column to remove excess GSNO, the number of thiols and nitrosothiols were measured in both reduced and GSNO-treated samples ( Table 1). The spectrum of GSNO-treated Trx1 showed an extra peak of absorbance at 335 nm, indicating the S-nitrosylation of the protein (Fig. 5). The number of nitrosothiols was  Fig. 1) for 2 min at room temperature, which was followed by a chromatographic step to remove the excess diamide, the insulin-disulfides reductase activity of the Trx system was studied. The oxidation rate of 200 M NADPH was followed at 340 nm in the presence of 160 M insulin, 10 nM rat recombinant TrxR, and 2 M of either untreated or diamide-treated Trx1 in TE buffer.   Fig. 3) followed by the removal of the excess H 2 O 2 using chromatography on a NAP-5 column, the insulin-disulfide reductase activity of Trx1 was studied as described for Fig. 2. After the initial decrease of absorbance, which is because of NADPH oxidation, the increase of absorbance in two of samples is due to the insulin precipitation that happens after the reduction of insulin. calculated using a molar extinction coefficient of 920 M Ϫ1 cm Ϫ1 at 335 nm (Table 1). There was close to two nitrosothiols in nitrosylated hTrx1. To find the localization of these nitrosothiols, different mutants of hTrx1 as well as the wt protein were utilized (Table 1). Furthermore, E. coli Trx was treated by GSNO to investigate whether cysteines in the active site can be nitrosylated. As shown in Table 1, nitrosylated human C32S/ C35S and C62S mutants as well as wt Trx displayed close to two nitrosothiols, whereas there was only one nitrosothiol in the structure of nitrosylated C62S/C73S Trx, and E. coli Trx was not nitrosylated but oxidized. Therefore, cysteines 69 and 73 became modified by GSNO, and cysteines 32 and 35 in the active site were oxidized to a disulfide.
The Effect of Nitrosylation of Trx1 on Its Catalytic Activity-The effect of nitrosylation on the activity of Trx as a substrate for TrxR was studied. Oxidation of NADPH by 20 M of either oxidized or nitrosylated Trx1 in the presence of 10 nM rat TrxR was recorded (Fig. 6). As expected the Trx-S 2 was rapidly reduced, oxidizing a stoichiometric amount of NADPH (Reaction 1). The rate of NADPH oxidation with nitrosylated hTrx1 was very slow; but interestingly, the rate was increased after a lag phase reaching finally up to 60 M NADPH oxidized. The interpretation of this is that active Trx-(SH) 2 molecules can reduce the nitrosothiol groups of Trx1. Therefore, nitrosothiols may be regarded as reduced by an autocatalytic reaction in which fully active Trx-(SH) 2 was the catalyst. The NADPH oxidation result showing two nitrosothiols in nitrosylated Trx1 corroborated our spectral data regarding the number of nitrosothiols (Table 1).
In addition, reduction of insulin disulfide bonds (Reactions 1 and 2) was used to investigate the effect of nitrosylation. Fig. 7 shows that the fully nitrosylated Trx was almost inactive, although it started to regain activity after a lag phase. When the experiments were done using nitrosylated double mutant of human Trx (C62S/C73S) with a single nitrosothiol on Cys 69 (Table 1), the initial protein-disulfide reductase activity of   The insulin-disulfide reductase activity of human Trx1 was studied as described for diamide and H 2 O 2 (see Fig. 2), and it was compared between untreated and nitrosylated samples. The decrease of absorbance in untreated sample is followed by an increase that is because of the insulin precipitation and the subsequent turbidity of the sample.  (28). Preincubation of reduced hTrx1 with varying concentrations of GSNO for 30 min at 37°C was also used to study the effect of nitrosylation on the reductase activity. Fig. 8 shows a progressive inhibition of Trx activity with higher concentrations of GSNO and under no conditions an increase in activity. Also in assays of thioredoxin activity using DTT and insulin precipitation (43), there was a lag phase consistent with the initial inhibition of activity.
Protection of Active Site Cysteines from S-Nitrosylation by the Pro 34 Residue-Protein-disulfide isomerase, a Trx fold protein, which has two active sites (-Trp-Cys-Gly-His-Cys-) with homology to the active site of Trx, is known to be nitrosylated on both thioredoxin domains (44). The nitrosylation of the active site Cys residues in protein-disulfide isomerase inhibits catalytic activity (44). We hypothesized that the Pro 34 of Trx has a protective effect against the S-nitrosylation of the Cys residues in the active site. To test this hypothesis, we used E. coli Trx1 P34H, a mutant with increased isomerase activity (39). P34H Trx1 was reduced and treated with 10 molar equivalents of GSNO for 60 min at 37°C. S-Nitrosylation was investigated spectrophotometrically after removing the excess GSNO with NAP-5 chromatography. In contrast to wt E. coli Trx, the P34H mutant showed an extra peak of absorbance at 335 nm corresponding to one nitrosothiol (data not shown).
Rat TrxR Was Not Affected by either H 2 O 2 or GSNO-Reduced rat TrxR obtained after NADPH incubation was treated with H 2 O 2 or GSNO. No effect was seen even with 5000 molar equivalents of the oxidants (data not shown).

DISCUSSION
The three structural Cys residues in human cytosolic Trx1 (Fig. 1A) are of critical importance for the structure and activity of the protein. There was a major difference in the outcome of oxidation by diamide and H 2 O 2 . Both oxidants induced a second disulfide Cys 62 -Cys 69 (6) with profound effects on the structure, because Cys 62 and Cys 69 are 17 Å apart in the threedimensional structure (Fig. 1A). However, only H 2 O 2 gave rise to a monomeric two-disulfide form that was inactive as a substrate for TrxR and NADPH. Diamide also inactivated Trx1, but in contrast gave rise to covalent dimers and multimers, and is a strong nonphysiological oxidant, originally devised to be GSH-specific. Diamide has played a major role in the development of redox signaling particularly regarding transcription factors. As outlined in Fig. 9, H 2 O 2 -dependent oxidation will block the catalytic activity of Trx, and it may be a signal for secretion of Trx and potentially generation of truncated thioredoxin (Trx80). In receptor-mediated signal transduction employing NADPH oxidases, the H 2 O 2 generated may act on Trx1 to inhibit the activity to allow temporal inhibition of phosphotyrosine phosphatases. The effect of H 2 O 2 on Trx1 will inhibit its activity as an electron donor for peroxiredoxins, potentially increasing the concentration of H 2 O 2 . Inhibition of Trx will also block its repair activity of the oxidized active site Cys residue (sulfenic acid) in the phosphotyrosine phosphatases like PT1B (45).
Reduced thioredoxin is implicated in a great number of physiological functions such as direct binding and inhibition of apoptosis signal-regulating kinase-1 or defense against oxidative stress by supplying electrons to peroxiredoxins and controlling the activity of numerous transcription factors like NF-B or p53.
As observed initially with GSNO (34), Trx1 is also a denitrosylating agent, and recent data show that thioredoxin catalyzes denitrosylation of low molecular mass and protein S-nitrosothiols (46). This activity controlling the S-NO content in proteins coupled with the fact that Trx1 itself is subject to S-nitrosylation (28 -31) and inactivation as shown in this study may explain the widely different results published on the subject of Trx1 and NO.
Regarding S-nitrosylation of human Trx1, our data show that the fully reduced protein was nitrosylated by GSNO on both Cys 69 and Cys 73 in vitro. This finding is in accordance with previous reports (28,29). However, in each of these studies the S-nitrosylation of one of these Cys residues was reported. On the other hand, three other studies corroborate our data regarding the number of nitrosothiols (30 -32). The different results may be derived by the kinetics of nitrosylation and by the different conditions including the redox state of the starting material. Because human Trx1 has five cysteines that can be modified under oxidative or nitrosative stress in different ways, affecting both the structure and activity, the redox state of the starting material is critical for the outcome of the experiment. For instance, under relatively mild oxidizing conditions, the second disulfide will form between Cys 62 and Cys 69 (6); this may include storage of the protein in frozen form. Homodimerization via an intermolecular disulfide bond between two Cys 73 residues may occur. The oxidized protein with two disulfides is strongly prone to noncovalent aggregation (5,42). Therefore, we strongly believe that any study regarding the redox state of Trx1 or modification of its Cys residues in vitro must be started using a fully reduced sample or control of the physicochemical state. In none of the studies about the nitrosylation of human Trx1, it is clear that the starting material was a fully reduced monomeric sample.
In this study, we also show that Cys 32 and Cys 35 , which are in the active site of Trx1, were not nitrosylated by GSNO but became oxidized to a disulfide bond. This finding is in accordance with previous reports (33). In addition, we show that Pro 34 is involved in protection of the active site against S-nitrosylation because the E. coli P34H Trx1 was sensitive to S-nitrosylation. Interestingly, it has recently been shown that another conserved proline residue (Pro 75 in human Trx1) prevents the formation of metal binding by the reactive thiolate-based active site (47).
The previous claim (28) that nitrosylation of Cys 69 increases the catalytic activity of human Trx1 could not be observed in our experiments. The data on activity measurements in Ref. 28 utilized E. coli thioredoxin reductase, which has no activity with human Trx1 (1,2,36), adding to the uncertainty of the results.
We used a double mutant of Trx1 C62S/C73S to investigate the effect of nitrosylation on Trx1 activity. After confirming the nitrosylation of Cys 69 , the protein-disulfide reductase activity was studied, showing no increase in activity.
In addition, we show that when reduced Trx1 was treated with different ratios of GSNO, the disulfide redox activity was progressively lost without evidence of activation. The idea that the inactivation of Trx system depends on the ratio between GSNO and Trx1, and subsequently on the nitrosylation of both Cys 69 and Cys 73 , can be reinforced by the fact that the addition of hTrx1 stimulates the initial NADPH oxidation rate of GSNO by TrxR severalfold (34) but is accompanied by progressive inactivation of the Trx system by increasing the concentration of GSNO (34). In our studies, we could not detect any increase in redox regulatory activity of human Trx1 after nitrosylation under any conditions.
Because H 2 O 2 -treated and S-nitrosylated Trx1 can regain the catalytic and redox regulatory function, the physiological importance of these modifications is, as yet, relatively unknown. However, the lag phase in the redox activity of oxidatively modified Trx1 could be a mechanism by which transient inhibition of Trx1 activity provides more time for the transmission of oxidative and/or nitrosative signals. It means that ROS and RNS, which are neutralized and eliminated by Trx system, can in turn modify the activity of the system via reversible inhibition of Trx activity to avoid decomposition. This interaction is probably crucial for the regulation of the multitude of ROS/RNS-mediated cell signaling. As noted above, this effect is comparable with the reversible inactivation of protein-tyrosine phosphatases by ROS such as H 2 O 2 that have important roles in cell signaling and redox homeostasis (48,49). The inhibition of protein-tyrosine phosphatases by ROS is mediated by the oxidative modifications of cysteine residues located in the active site of these proteins. This modification is firstly mediated by the formation of a sulfenic acid via the reaction between a cysteine residue and H 2 O 2 , which subsequently leads to the formation of a disulfide bond after the nucleophilic attack of another cysteine residue to the sulfenic FIGURE 9. The schematic presentation of oxidation and S-nitrosylation of human cytosolic thioredoxin by hydrogen peroxide and GSNO, respectively. Under oxidizing conditions, two disulfides may form in the structure of Trx1: one in the active site and the second one between structural cysteines 62 and 69. This modification of Trx1 probably leads to conformational changes which are involved in the secretion of Trx1 or the formation of Trx80. Moreover, it is shown that this form of Trx1 can be nitrosylated on Cys 73 ; thereafter, it is implicated in transnitrosation processes with caspase 3 (29). Following nitrosative stress, a disulfide forms in the active site, and the two structural cysteines 69 and 73 become nitrosylated. This form of Trx1 is also probably involved in transnitrosation processes. The formation of two disulfides or two nitrosothiols leads to the inhibition of catalytic activity of the Trx system. This effect is reversible, and Trx activity is regained via an autocatalytic mechanism.
acid (45,49) or to glutathionylation via GSH (50). Interestingly, in a recent study, it is reported that inhibition of protein-tyrosine phosphatases by mild oxidative stress is dependent on S-nitrosylation (51).
On the other hand, because oxygen can expedite S-nitrosylation, as shown for some proteins (52,53), nitrosylation of Trx1 may provide a protective mechanism against irreversible oxidation of thiol groups or inactivation of the protein via disulfide formation between structural cysteines during oxidative and nitrosative stress. It has been shown that intermolecular disulfide formation between Cys 73 of two Trx molecules lead to the inactivation of the redox regulatory function (38,54). A similar preserving role for the intramolecular two disulfide formation between Cys 62 and Cys 69 is also possible.
A recent study showed that hTrx1 expression reduces the tissue content of nitrotyrosine, which is an indirect determinant of in vivo ONOO Ϫ production and an index of nitrative stress (55). In another study it is suggested that hTrx1 can be nitrated on Tyr 49 , which leads to irreversible inhibition of its redox regulatory activity (56). Because nitration of proteins during nitrative stress is mediated by ONOO Ϫ , which is shown to be formed via a reaction between NO and superoxide, nitrosylation of hTrx1 might be a protective effect against nitrative stress by removing NO, leading to the inhibition of ONOO Ϫ formation. This mechanism would be considered more important when we know that most, if not all, superoxideinitiated tissue injury is mediated by ONOO Ϫ , and not by superoxide itself or H 2 O 2 (57).
In this study, we also treated rat TrxR with hydrogen peroxide and GSNO, but no effect on the activity of the enzyme was recorded, which is in agreement with previous studies (58). Thus, the redox regulatory functions are associated with thioredoxin. To understand the effects of intramolecular disulfide formation via structural cysteine residues as well as their S-nitrosylation on other biological activities of Trx1, such as binding to other proteins via formation of disulfide bridges between SH groups, regulation of transcription factors, apoptosis, secretion of Trx1, etc., further studies are required. Particularly challenging will be to understand the role of Trx1 as a denitrosylating agent (59,60) that also can be nitrosylated itself and then transnitrosylate.