Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=ifra20 Free Radical Research ISSN: 1071-5762 (Print) 1029-2470 (Online) Journal homepage: https://www.tandfonline.com/loi/ifra20 Conferring specificity in redox pathways by enzymatic thiol/disulfide exchange reactions Luis Eduardo S. Netto, Marcos Antonio de Oliveira, Carlos A. Tairum & José Freire da Silva Neto To cite this article: Luis Eduardo S. Netto, Marcos Antonio de Oliveira, Carlos A. Tairum & José Freire da Silva Neto (2016) Conferring specificity in redox pathways by enzymatic thiol/disulfide exchange reactions, Free Radical Research, 50:2, 206-245, DOI: 10.3109/10715762.2015.1120864 To link to this article: https://doi.org/10.3109/10715762.2015.1120864 Accepted author version posted online: 16 Nov 2015. Published online: 08 Jan 2016. Submit your article to this journal Article views: 727 View Crossmark data Citing articles: 24 View citing articles https://www.tandfonline.com/action/journalInformation?journalCode=ifra20 https://www.tandfonline.com/loi/ifra20 https://www.tandfonline.com/action/showCitFormats?doi=10.3109/10715762.2015.1120864 https://doi.org/10.3109/10715762.2015.1120864 https://www.tandfonline.com/action/authorSubmission?journalCode=ifra20&show=instructions https://www.tandfonline.com/action/authorSubmission?journalCode=ifra20&show=instructions http://crossmark.crossref.org/dialog/?doi=10.3109/10715762.2015.1120864&domain=pdf&date_stamp=2015-11-16 http://crossmark.crossref.org/dialog/?doi=10.3109/10715762.2015.1120864&domain=pdf&date_stamp=2015-11-16 https://www.tandfonline.com/doi/citedby/10.3109/10715762.2015.1120864#tabModule https://www.tandfonline.com/doi/citedby/10.3109/10715762.2015.1120864#tabModule FREE RADICAL RESEARCH, 2016 VOL. 50, NO. 2, 206–245 http://dx.doi.org/10.3109/10715762.2015.1120864 REVIEW ARTICLE Conferring specificity in redox pathways by enzymatic thiol/disulfide exchange reactions Luis Eduardo S. Nettoa, Marcos Antonio de Oliveirab, Carlos A. Tairumb and José Freire da Silva Netoc aDepartamento de Genética e Biologia Evolutiva, Instituto de Biociências, Universidade de São Paulo, São Paulo, Brazil; bDepartamento de Biologia, Universidade Estadual Paulista Júlio de Mesquita Filho, Campus do Litoral Paulista São Vicente, Brazil; cDepartamento de Biologia Celular e Molecular e Bioagentes Patogênicos, Faculdade de Medicina de Ribeirão Preto, Universidade de São Paulo, Ribeirão Preto, São Paulo, Brazil ABSTRACT Thiol–disulfide exchange reactions are highly reversible, displaying nucleophilic substitutions mechanism (SN2 type). For aliphatic, low molecular thiols, these reactions are slow, but can attain million times faster rates in enzymatic processes. Thioredoxin (Trx) proteins were the first enzymes described to accelerate thiol–disulfide exchange reactions and their high reactivity is related to the high nucleophilicity of the attacking thiol. Substrate specificity in Trx is achieved by several factors, including polar, hydrophobic, and topological interactions through a groove in the active site. Glutaredoxin (Grx) enzymes also contain the Trx fold, but they do not share amino acid sequence similarity with Trx. A conserved glutathione binding site is a typical feature of Grx that can reduce substrates by two mechanisms (mono and dithiol). The high reactivity of Grx enzymes is related to the very acid pKa values of reactive Cys that plays roles as good leaving groups. Therefore, although distinct oxidoreductases catalyze similar thiol–disulfide exchange reactions, their enzymatic mechanisms vary. PDI and DsbA are two other oxidoreductases, but they are involved in disulfide bond formation, instead of disulfide reduction, which is related to the oxidative environment where they are found. PDI enzymes and DsbC are endowed with disulfide isomerase activity, which is related with their tetra-domain architecture. As illustrative description of specificity in thiol– disulfide exchange, redox aspects of transcription activation in bacteria, yeast, and mammals are presented in an evolutionary perspective. Therefore, thiol–disulfide exchange reactions play important roles in conferring specificity to pathways, a required feature for signaling. ARTICLE HISTORY Received 20 August 2015 Revised 11 November 2015 Accepted 11 November 2015 Published online 14 December 2015 KEYWORDS Disulfide reductases; glutathione; glutathione peroxidase; peroxiredoxin; thiols; thioredoxin Introduction Many redox proteins rely on non-proteic cofactors (such as NAD+; FAD; heme; Cu, Fe, or other transition metals) for their activity. In contrast, other proteins use Cys residues for electron transfer processes (reviewed in Ref. [1]). The free amino acid cysteine presents low reactivity for redox transitions [2,3]. However, protein folding can generate environments in which cysteine residues become redox active. Thiol and selenol groups are the most easily oxidiz- able residues in proteins. Distinct Cys-based proteins are specifically oxidized by different agents. For instance, peroxiredoxins, thioredoxins, and methionine sulfoxide reductase are efficiently oxidized by hydroperoxides, proteins disulfides, and methionine sulfoxide, respect- ively [1]. Probably, folding in each protein is adapted to stabilize distinct transition states. One of the most common forms of oxidized cysteine in these proteins is the disulfide bond, which is generated from thiols by a two electron oxidation process from two sulfhydryl groups separated apart by distances in the 2.00 Å range: 2 Cys-SH! Cys�SS�Cysþ 2Hþ þ 2e� ðReaction 1Þ Disulfides are stable covalent bonds with dissociation energy of around 60 kcal/mole (251 kJ/mol), being 40% weaker than C–C and C–H bonds (reviewed in Ref. [4]). The disulfide bond length is about 2.05 Å and its stability is related to the C–S–S–C dihedral angles (Figure 1). Disulfides with dihedral angles close to 90� present low energy levels. In contrast, when the angle goes far from this value (0� or 180�), the disulfide is in a strain condition, becoming more unstable and consequently presenting significantly better oxidant properties [5]. For instance, five-membered cyclic disulfides that are strained with a C–S–S–C angle of around 30� are more CONTACT Dr Luis Eduardo S. Netto, PhD nettoles@ib.usp.br Departamento de Genética e Biologia Evolutiva, Instituto Biociencias – Universidade Sao Paulo, Genetica e Biologia Evolutiva, Sao Paulo 05508-090, Brazil � 2015 Taylor & Francis unstable and play a role as an oxidizing agent in bioenergetics processes. Accordingly, lipoamide is a biological five-membered cyclic disulfide that is involved in oxidizing substrates in bio-energetic pathways (sec- tion ‘‘Txn–T(SH)2 system’’). Taking advantage of the large number of structures whose coordinates were deposited into the protein data bank, we analyzed the dihedral angles of disulfides bonds in several proteins. The global analyses indicated that there is no evident correlation of proteins with reducing and oxidizing capacities (E0 values) and their dihedral angles (Table I). Other factors affect the redox properties of proteins and are further discussed below. In some proteins, disulfides play a role in increasing their overall stability, especially in the case of extracel- lular proteins that have to cope with the harsh condi- tions of the extra-cellular environment [29]. In contrast, disulfide bonds can be reversible interconverted into dithiols, a redox process that can play roles in enzymatic catalysis, transport of reducing or oxidizing equivalents, and antioxidant defense, among other processes. In some cases, dithiol–disulfide inter-conversion is one of the several steps of the overall enzymatic mechanism. For example, thiol–disulfide exchange reactions are involved in the regeneration of the reduced states in enzymes such as ribonucleotide reductase [30], methio- nine sulfoxide reductase [31,32], and Prxs [33]. In other cases, reduction of disulfide to dithiol is not only a step in the catalytical cycle per se, but also its occurrence enhances (or diminishes) the enzymatic activities, by an allosteric-like mechanism. This kind of mechanism is well described for several chloroplastic enzymes and the reductive power comes from photo- dependent processes (reviewed in Ref. [4]). Also within this category are the reductive reactions of disulfides in some proteins that are components of cell receptors. Reductions of these disulfides to dithiols activate the binding of agonists (reviewed in Ref. [4]). A third category of enzymes, whose activities are also affected by thiol-based redox processes comprise pro- teins that a sulfhydryl group is essential for their activities, but thiol–disulfide exchange per se is not part of the catalytical cycle. Examples are thiol proteases (papain and bromelain), b-ketoacylthiolase, creatine kinase, glyceraldehyde-3-phosphate dehydrogenase, and protein tyrosine phosphatase. These enzymes can be reactivated using reducing agents, such as dithio- threitol and TCEP. Therefore, thiols and disulfides can be intercon- verted back and forward in conditions that reflect environmental state. In a basal condition, with no oxidative insult, about 90% of the total sulfur atoms in mammalian cells are in the thiol (or thiolate) form; whereas 10% are engaged in disulfide bonds (0.1% in the protein-glutathione mixed disulfide). From the reduced pool, about 75% are proteic thiol, whereas 25% are low molecular weight thiols, mostly reduced glutathione [34]. Since inter-conversions of thiols into disulfides are central processes in biology, it is relevant to fully comprehend the fundamentals of the reaction that control these processes: thiol–disulfide exchange (also called thiol–disulfide interchange). Physico-chemical factors affecting thiol–disulfide exchange reaction for low molecular weight compounds Thiol–disulfide exchange reaction involves the reduc- tion of a disulfide bond by a thiolate anion (RS�), resulting in the formation of a new disulfide and a new thiolate (see the ‘‘Reaction 3’’ below). It is unique in the sense that reagents and products contain the same functional groups. Therefore, in this reaction, a strong covalent bond (the S–S bond; bond energy about 60 kcal/mol) is cleaved and another bond is formed. It occurs reversibly at physiological conditions, but the non-catalyzed process is very slow, with second order rate constants in the 0.1–10 M�1 s�1 range, in the case of aliphatic thiols (reviewed in Ref. [4]). The half-life for alkane thiols is about 2 h for a reaction containing thiol and disulfide at millimolar concentrations. Therefore, it is important to consider the equilibrium of thiolate protonation to thiol in aqueous solution, according to the Reaction 2. The corresponding acid- dissociation constants (the pKa values) vary among low molecular weight and protein thiols, affecting the reaction rates. RSH ! RS� þ Hþ ðReaction 2Þ Figure 1. Dihedral angle � – determined from the two planes containing a disulfide bond. When the dihedral angle is equivalent to 90� the corresponding disulfide bond is in the most stable form [5]. FREE RADICAL RESEARCH 207 Thiol–disulfide exchange is a second-order reaction: first order for the thiolate and first order for the disulfide (reviewed in Refs. [4,35,36]). It is a nucleo- philic substitution reaction of the SN2 type, in which as one bond is broken and the new bond is formed in one step, with the inversion of the symmetry. As expected for a SN2 reaction, there is a single transition state. In this specific case, the transition state complex is composed of an intermediate with three sulfur atoms (Figure 2), with a negative charge delocalized among them. At this point, it is convenient to name the three sulfur atoms for a general thiol–disulfide exchange reaction: (i) the attacking sulfur atom, corresponding to the thiolate anion is the nucleophilic sulfur (RSnuc), (ii) the central sulfur (RSc) as the one that will take part in the new disulfide, and (iii) the sulfur leaving group as RSlg. Therefore, for a general thiol–disulfide exchange reaction: RS�nuc þ RSc � SlgR! RSnuc � ScRþ RS�lg ðReaction 3Þ As mentioned before, for aliphatic thiols and disul- fides this is a very slow reaction (0.1–10 M�1 s�1) (reviewed in Refs. [4,36]). Anyway, it is important to understand the factors that affect the kinetics of these reactions, in order to understand how enzymes can accelerate them. Electrostatic effects Knowing that the thiolate and not the corresponding thiol is the nucleophilic agent in thiol–disulfide exchange reactions, sulfhydryl compounds with low pKa values would be more efficient as attacking sulfur compounds. This is because the fraction of total thiol group present as thiolate will be higher for those thiols having acidic pKa values. However, the picture is more complex, as it will be described below. Table I. List of some thiol–disulfide exchange enzymes that present 3D-structure in the disulfide form. Enzymes Organisms Dihedral angle (�) S–S distance (Å) PDB File pKa Redox potential (Eo, mV) Alkyl hydroperoxide reductase C (AhpC) S. typhimurium �93.20 2.04 1YEP 5.94 [6] �178 [7] Alkyl hydroperoxide reductase F (NTD domain) S. typhimurium 93.22 2.10 1ZYN 5.1 [8] �265 [8] Disulfide bond protein a (Dsba) E. coli 90.34 2.03 1A2M 3.5 [9,10] �125 [11] FTR-Ferredoxin thioredoxin reductase (chloroplast) S. oleracea (spinach) �84.93 2.05 1DJ7 – �320 [12] Glutaredoxin (Grx) E. coli 108.34 2.02 1EGO 5.0 [13] �198 [14] Glutaredoxin 1 (Grx1) S. cerevisiae 66.12 2.06 3D4M 3.2–4.0 [15] – Glutaredoxin 2 (Grx2) S. cerevisiae 71.37 2.08 3C1R 3.1–3.5 [15] – Glutathione reductase (GR) E. coli 127.41 2.05 1GER 8.5 [16] �243 [16] Glutathione reductase (GR) H. sapiens �133.43 2.10 3GRS 7.6 [16] �227 [16] Glutathione reductase (GR) S. cerevisiae �126.76 2.10 2HQM 7.4 [16] �237 [16] Lipoamide dehydrogenase E. coli 28.95 1.96 4JDR 6.7 [17] – Mercuric reductase P. aeruginosa �116.32 2.01 1ZX9 5.2 [18] �269 [19] Protein disulfide isomerase (PDI) H. sapiens 91.77 2.21 4EL1 4.81 [20] �162 to 169 [21] Thioredoxin A (TrxA) E .coli 81.02 2.09 2TRX 6.4–7.45 [22,23] �270 [24] Thioredoxin 1 (Trx1) S. cerevisiae 72.4 2.05 3F3Q – �275 [25] Thioredoxin 2 (Trx2) S. cerevisiae 84.17 2.10 2FA4 – �265 [25] Thioredoxin f (Trx f) S. oleracea (spinach) 74.18 2.03 1FAA – �290 [12] Thioredoxin m (Trx m) S. oleracea (spinach) 87.05 2.03 1FB6 – �300 [12] Thioredoxin reductase (TrxR) (FAD)-enzyme-(S)2/(SH)2 E. coli 73.90 2.01 1TDE � �254 [26,27] Thioredoxin reductase (TrxR) (FADH2)-enzyme-(S)2/(SH)2 E. coli 73.90 2.01 1TDE � �271 [26,27] Tryparedoxin (TXN) T. brucei 80.01 2.00 1O73 7.2 [28] �249 [28] Yeast activator protein 1-Redox Domain (Yap1-RD) (Cys 303–Cys 598) S. cerevisiae �67.20 2.02 1SSE – �155 [25] Yeast activator protein 1-Redox Domain (Yap1-RD) (Cys 310–Cys 629) S. cerevisiae 64.44 2.02 1SSE – �330 [25] The respective dihedral angles and disulfide sulfurs distance of each enzyme are listed. pKa and redox potential data were compiled from the discriminate references. The structure models were obtained from PDB database and determination of the distances and angles were performed using PyMol software (The PyMOL Molecular Graphics System, Version 1.7.4 Schrödinger, LLC.) Figure 2. Free energy diagram for a generic thiol disulfide exchange reaction. Ea¼Activation energy. There is a kinetic barrier, which can be alleviated by aprotic solvents and/or hydrophobic environments such as the active site pockets of the enzymes (reviewed in Ref. [4]). 208 L. E. S. NETTO ET AL. The acid dissociation for all three sulfur atoms interfere with the speed of this reaction, therefore, it is relevant to analyze the pKa values for all of them. Brønsted base coefficients are convenient parameters to correlate ionization state of the sulfur atoms and their effects on the reaction rates. The log of the real rate constants or rate constant pH-independent (kRS�) of a series of thiol–disulfide exchange reactions follows a Brønsted correlation as a function of their pKa values for the three sulfur atoms (reviewed in Ref. [4]). Before describing this equation, it is important to distinguish the real second order rate constant (pH independent) that is distinct from the apparent (kobs) rate constant (pH dependent). The fraction of available thiolate at a given pH, for a given thiol is calculated to account for its reactivity toward the electrophilic species [37,38]. Consequently, pH independent rate constant (kRS�) corresponds to the specific reactivity of the thiolate fraction of total sulfhydryl pool against the electrophile (in this case the disulfide) is given by: kobs ¼ ks�= 1þ 10ðpKa�pHÞ � � ð1Þ Therefore, the Brønsted relation for a thiol–disulfide exchange reaction is given by: log kRS � ¼ cþ �nucpKaðnucÞ þ �cpKaðcÞ þ �lgpKaðlgÞ ð2Þ where bnuc, bc, and blg are the Brønsted base coefficients for the nucleophile, central group, and leaving group, respectively [4,39,40]. For didactic reasons, let us initially consider only the first term of the overall Equation (2). log kRS � ¼ cþ �nucpKaðnucÞ ð3Þ Once formed, the reactivity of the thiolate is itself influenced by electrostatic effects. A plot of the log k against pKa of the pH-independent rate constants for a series of thiol–disulfide exchange reactions give raise to a straight line with slopes between 0.4 and 0.6 [40,41,76] that corresponds to the bnuc. This positive correlation implies that the enhancement of the thiol basicity (higher pKa) favor its nucleophilic ability to displace the sulfur atom (RSC) of the disulfide bond. This is a well- known phenomenon for organic chemists in other SN2 type of reactions, which indicates that those factors that favor the transfer of electron density from the thiolate to a proton (basicity), also favor the transfer of negative charge to other atoms (such as sulfur atoms in the transition state for thiol–disulfide exchange reaction). Therefore, the reactivity of the nucleophilic thiol (RSnuc) is determined by two opposite factors related with the pKa values [42,43]. Thiols with high pKa values are expected to be stronger nucleophiles. However, as the reacting species is believed to be the thiolate anion, acidic thiols (with low pKa values) would be favored at physiological pH, as a larger fraction of anionic thiolate would be available to react. From the balance between these two factors it emerges that the most reactive thiols (in terms kobs) are those with a pKa close to the pH of the solution. As it will be discussed later, thioredoxin enzymes appear to take advantage of this property to act as a catalyst in thiol–disulfide exchange reactions. Besides the attacking thiol (RSnuc), the ionizations of the RSC and of the RSlg also affect the rates of thiol– disulfide exchange reactions [39,40,44]. Implicit in Equation (1) is the concept that the substituents attached to the three sulfur atoms influence the reaction rate independently. However, analysis of thiol–disulfide exchange employing a redox fluorescent protein indi- cates that there is some dependence among the three sulfur atoms [44]. Moreover, there is considerable uncertainty in the values of the Brønsted coefficients, but it appears that bc and blg lie in the range of�0.5 to�0.3 [39,40]. Therefore, the more acidic pKa values for RSC and RSlg, higher the rates of the thiol–disulfide exchange reaction, indicating that thiolate is not only a strong nucleophile, but also a good leaving group, which is related with its high polarizability and low degree of solvation, among other factors [4]. Therefore, analysis of the ionization of the RSlg deserves further attention. For a set of similar thiol–disulfide exchange reactions, the second-order rate constant increases by a factor of four for each unit decrease in the pKa of the leaving group [29,40]. Furthermore, electrostatic factors that increase the electrophilicity of the disulfide formed by RSC and RSlg will lead to a decrease of the activation energy and thereby activate the reaction rate [36,39,44–47]. The limited data available indicated that electronic density is evenly distributed among the three sulfur atoms in the transition state [39 and references therein]. However, some theoretical studies appoint for higher absolute values of Brønsted coefficients for the RSnuc and RSlg, indicating that these atoms have the greater contribution to the overall reaction. Through quantum mechanical calculations, it was observed that peripheral sulfur atoms (RSnuc and RSlg) should have the most negative charge at the transition state, with zero or small negative charge at RSC. It appears that along the course of the reaction, the negative charge is transferred directly from RSnuc to RSlg without accumulating at RSC [43]. Solvent effects Hydrophobic environments accelerate the rate of thiol– disulfide exchange reactions [43,48]. Molecular dynamics FREE RADICAL RESEARCH 209 of a reaction between methylthiol and the disulfide of dithiothreitol (DTT) using a dielectric constant compat- ible with the hydrophobic pocket of ribonucleotide reductase produced a reaction rate 1000 times faster than the similar reaction in water [43]. Indeed, rates of dithiol–disulfide exchange reaction in polar aprotic solvents are 1000 times faster than similar reaction in polar protic solvents. In contrast, the nature of the counter cation has no effect on such reaction in DMSO [48,76]. It appears that polar solvents stabilize the reactants in relation to the transition state, increasing the activation energy barrier (Figure 2). This effect is related with the negative charge distribution in the transition state described above. Extensive hydrophobic portions in Mia40 and PDI are probably related with the ability of these enzymes to catalyze thiol–disulfide exchange reactions [43,48,49]. In fact, a common feature among thiol–disulfide oxidoreductases is the presence of hydro- phobic patches around the active site pocket (Figure 3), which may help to speed up the thiol–disulfide exchange reactions. Steric effects Transition state in thiol–disulfide exchange reaction is crowded and presents a linear geometry as expected for a SN2 type process. Therefore, bulky substituents in carbon a of all three sulfur atoms can profoundly affect reaction rates [4]. Alkyl substituents for the carbon b of all three sulfur atoms induced less pronounced effects. In enzymes, steric factors probably are associated with the selectivity towards their targets, as will be discussed later. It was proposed that steric factors might create a strain of a disulfide, making its dihedral angle distant from 90� (Figure 1). However, our analysis of dihedral angles in available structures indicated that is not the case (Table I). There is no clear apparent correlation among dihedral angles, S–S bond distance, E0 (mV) and pKa values (Table I). Probably the diversity of structural Figure 3. Molecular surfaces of thiol enzymes in reduced and oxidized states revealing hydrophobic patches. The proteins surfaces are colorized in light gray and hydrophobic amino acids in purple. E. coli Trx in reduced (A, PDB code: 1CL0) and oxidized state (B, PDB code: 1XOA); Grx from E. coli in reduced (C, PDB code: 1EGR) and oxidized state (D, PDB code: 1EGO) and Dsba from E. coli in reduced (E, PDB code: 1A2L) and oxidized state (F, PDB code: 1A2M). The figures were generated using the Chimera software (http:// www.cgl.ucsf.edu/chimera/). 210 L. E. S. NETTO ET AL. features among Cys-based proteins is large enough to prevent a clear correlation between these parameters. We suspect that this kind of analysis should be performed within more homogenous set of proteins. Anyway, the more labile and unstable the disulfide, more oxidizing are the oxidoreductases such as DsbA [50]. Catalysis of thiol–disulfide exchange reaction In spite of some limitations, the Brønsted relationship describes relatively well thiol–disulfide exchange reac- tions for aliphatic low molecular weight thiols. However, little is known on the applicability of the Brønsted relationship to proteins. Part of the difficulties to expand this knowledge to proteins is related to the fact that two sequential thiol–disulfide exchange reactions are required to reduce a disulfide by an oxidoreductase, as it will be further discussed below. In an attempt to gain insights in this direction, reactions of peptides mimicking bovine pancreatic trypsin inhibitor (BPTI) and two oxidoreductases (DsbA and Trx) in the reduced state with low molecular weight disulfides were performed. As aliphatic thiols, peptide thiols also reacted slowly with disulfide reagents (GSSG, HED, cysteine, and cystamine), displaying kobs in the 0.1– 10 M�1 s�1 range [39]. Remarkably, rates for the same peptide vary depending on the disulfide reagent employed: GSSG (a negatively charged disulfide) dis- playing slowest rates, whereas cystamine (a positively charged disulfide) displaying fastest rates [39]. In some proteins, rates of thiol–disulfide exchange can attain higher values, such as the GSSG reaction with two of the six cysteine residues (Cys14 and Cys38) of BPTI. This phenomenon was attributed to positive charges nearby residues Cys14 and Cys38 of BPTI and other local structural factors related to the BPTI folding process [51]. Similarly, GSSG at micromolar concentrations can induce formation of inter-molecular disulfide in Thimet oligo- peptidase, a thiol-rich metallopeptidase, containing 15 Cys residues [52]. Indeed, GSSG can catalyze disulfide bonds isomerization in proteins, until the configuration that stabilizes a given protein in its most stable conformation is achieved [53]. Anyway, in all these cases, rates of thiol–disulfide exchange reactions involved in disulfide shuffling are very slow in the 0.1– 10 M�1 s�1 range. In contrast, oxidoreductases within the Trx superfamily (which are the most studied thiol– disulfide exchange catalyst so far) can attain rates of thiol–disulfide exchange reaction in the 106–107 M�1 s�1 range. Below is the description of the best studied oxidor- eductases and some of the factors that underlie their capacity to achieve extraordinary rates in thiol–disulfide exchange reactions. Trx system As mentioned before, thiol–disulfide exchange reactions are slow for aliphatic, low molecular thiols, but in living cells these reactions can be accelerated million times by oxidoreductases. Thioredoxin (Trx) was the first enzyme described as a catalyst for thiol–disulfide exchange reactions. This oxidoreductase was initially discovered in Escherichia coli as a hydrogen donor for ribonucleo- tide reductase [30]. One of the first descriptions of the thiol–disulfide exchange catalysis by Trx was the dem- onstration that this oxidoreductase accelerates insulin reduction by DTT or by dihydrolipoamide by a factor of 10,000 times [54]. Thioredoxin is a low molecular weight protein (12–13 kDa), whose redox activity is endowed by a C–X–X–C motif. All Trxs have very conserved three-dimensional structures that are compact, with 90% of its residues in a-helices, b-strands, or reverse turns [55]. The so-called Trx fold is also present in other oxidoreductases such as Grx, PDI, and DsbA and comprises a central four or five stranded b-sheet surrounded by a-helices, with part of the dithiols/disulfide motif protruding from the protein surface (Figure 4A). The most N-terminal Cys (Cys32 in EcTrx1) is more exposed to the solvent, whereas the second Cys is buried into the polypeptide chain (reviewed in Ref. [56]) (Figure 4B). Another important feature of the Trx fold is a cis-Pro loop that approaches a Pro residue (Pro76 in EcTrx1) to the C–X–X–C motif (Figure 4B), which is important for substrate recognition [57]. Remarkably, this folding has been highly conserved for 4 billion years or close to the origin of life [58]. Trxs present a conserved catalytic site (-Trp-Cys-Gly-Pro-Cys- Lys-) that undergoes reversible oxidation to the cystine disulfide (Trx-S2) through the transfer of reducing equivalents from the two catalytic Cys residues to a disulfide substrate. Besides these residues, other amino acids play a role in the catalytic activity of Trx, but their precise role is not clearly established. A Pro and an Asp (Pro76 and Asp26 in EcTrx1) are frequently implied in catalysis and are located in the proximity of the two catalytic Cys (Figure 4B). The oxidized Trx is reduced back to the Cys form [Trx-(SH)2] by thioredoxin reductase, which is a NADPH-dependent flavoprotein (Figure 5). In principle, the properties that govern the rates of thiol-disulfide exchange reactions for low molecular weight compounds [Equation (1)] should also apply to the reduction of protein disulfides by Trx. However, it is FREE RADICAL RESEARCH 211 clear that electrostatic (section ‘‘Electrostatic effects’’) and steric factors (section ‘‘Steric effects’’) among other features should confer specificity for the reduction of target disulfides by Trx. On the other hand, analysis of the EcTrx1 X-ray crystal structure found no evidence of strained dihedral or bond angles at the disulfide [55], which is in agreement with our analysis shown in Table I. Therefore, additional mechanisms may be considered. Possibly, conformational changes during substrate bind- ing that can generate geometries and structural features such as hydrophobic and complementary polar patches that may contribute to the catalysis of thiol-disulfide exchange reactions (reviewed in Ref. [56]) (Figure 6). To reduce a disulfide bond in a client protein, dithiolic Trx enzymes have to undertake two sequential thiol– disulfide exchange reactions (Figure 5). In the first one, the reactive Cys (the more N-terminal Cys and equivalent to Cys32 of EcTrx1) attacks the disulfide, generating a mixed disulfide between Trx and the target protein (Figure 5, reaction I). Then, in the second thiol-disulfide exchange reaction, the second Cys (more C-terminal and equivalent to Cys35 of EcTrx1) reduces the mixed disulfide, generating oxidized Trx, and dithiolic target protein (Figure 5, reaction II). A factor that may favor this second thiol–disulfide exchange reaction is that Cys35 from EcTrx1 is in a hydrophobic environment. As will be further discussed below Asp26 in EcTrx1 appears to be a key residue for the activation of Cys35 as a nucleophile for the mixed disulfide (Figure 3B), in the second thiol– disulfide exchange reaction. During these two reactions, conformational movements take place and appear to be important in nucleophilic activation of the C-terminal Cys. Therefore, besides the properties accounted by the Brønsted correlation, other structural factors should also be taken into account when analyzing enzymes (reviewed in Ref. [56]). As described previously, factors that affect ionization of attacking sulfur are critical for the generation of Figure 4. Thioredoxin structure and active site amino acids. (A) Crystallographic structure of thioredoxin from E. coli (PDB ID¼ 1XOB) showing the thioredoxin fold. Light green is the central five-stranded � sheet, which is surrounded by four � helices (dark red). Catalytic Cys residues and amino acids involved in catalysis are depicted as ball and sticks. (B) Closer view of the active site. The two catalytic Cys residues (Cys32¼ nucleophilic; Cys35¼ resolving) are depicted in orange; conserved Pro76 and Asp26 residues are represented in green in blue, respectively. (See color version of this figure in the online version.) Figure 5. Catalytic cycle of thioredoxins. Initially, the reactive cysteine attacks the target disulfide, resulting in a transient mixed disulfide between Trx and the biological target which is resolved by the second Trx cysteine residue (I) resulting in the formation of an intramolecular disulfide bond in Trx (II), which is reduced by the TrxR using reducing equivalents from NADPH (III). 212 L. E. S. NETTO ET AL. nucleophilic thiolate. Therefore, several attempts were performed to determine the pKa values of the two catalytic Cys, using methodologies, such as redox changes in fluorescence [59], UV absorption at 240 nm [9,60,61], Raman spectroscopy [62,63], NMR [64–66], and measurements of the pH dependence of the rate of alkylation [24]. For EcTrx1, pKa values for the thiolate of Cys32 varies from 6.3 to 10.0, whereas for the thiolate of Cys35 the range of the values are even higher (between 7 and 14), probably reflecting the fact that the second thiol group is buried into the polypeptide chain. Cheng and colleagues [56] performed a compre- hensive analysis of all these data and observed that in spite of all the variation the majority of the determin- ations of the pKa of Cys32-SH are below the value for free cysteine (8.5–9.0). The positive dipole charge of the N-terminal extremity of helix a2 (where Cys32 is located) is one of the factors that favor the deprotonation of this thiol group. Anyway, the variation of the determined pKa values is high and probably this is, at least, partially related to shortcomings of the distinct methods used to measure the pKa values [67,68]. Noteworthy, most measurements were performed in the absence of protein ligands, in a condition where the protein is static. Static measurements may not be informative about how dynamic processes affect the relevant pKa values during the docking of a target protein into Trx active site [69]. Conformational changes occur in Trx through molecular interactions during substrate recognition and binding, and these can generate a local hydrophobic environment and a juxta- position of titratable groups that affect the ionization of the Trx thiols [70,71]. Indeed, thiol proteins react with Trxs orders of magnitude faster than with low molecular weight thiols, possibly due to conformational changes and reorientation of reactive groups during protein– protein interactions that generate optimal geometry for SN2 type of reaction [9,64,68,72–76]. Indeed, 13C NMR experiments indicated that hydrogen bond rearrange- ments are involved in micro-millisecond range, compat- ible with Trx turnover rates [77]. Therefore, to account for the contribution of these dynamic structural factors, a correction factor (Fc) has been introduced into the Brønsted relation for a thiol-disulfide exchange reaction. Accordingly, to best describe reactions of Trx (SH)2 with proteins, Equations (1) and (2) were taken and Fc was added to yield Equation (4) [61,78]. kobs ¼ ðcþ �nucpKaðnucÞ þ �cpKaðcÞ þ �lgpKaðlgÞÞ � log 1þ 10ðpKaðnucÞ�pHÞ � � þ Fc ð4Þ Acid/base catalysis is a central feature in the mech- anism of thiol/disulfide oxidoreductases as expected for a reaction that depends on the ionization of sulfhydryl groups. The factors that influence the ionization of the two catalytic Cys are controversial and this issue has been also comprehensively reviewed by Cheng and coworkers [56]. As mentioned above, for the N-terminal Cys (Cys32 in EcTrx), the positive charge of the N-terminal extremity of helix a2 appears to favor the deprotonation of its thiol group. Assuming the pKa value for the most exposed residue as 7.0, probably the dipole effect of the helix a2 would be enough to reduce the pKa below the level of the free Cys (�8.5). This is because it was estimated that the interaction between the thiolate anion at the N-terminus of the a-helix in a model Figure 6. Factors underlying the efficiency and specificity in the reduction of biological targets by Trx. Trx recognize disulfide targets in proteins by a combination of factors such as polar (represented by ‘‘+’’ and ‘‘�’’ signals) and hydrophobic (represented by triangles) interactions, which allow (A) right orientation for the nucleophilic attack of the reactive Cys (Cys S�) over the target disulfide. (B) Formation of mixed disulfide Trx-S-S-Target and the nucleophilic attack over the mixed disulfide by the second Trx Cys residue. (C) Release of oxidized Trx (S–S) and the reduced biological target. During the thiol-disulfide exchange reactions, also topological changes occur between Trx and target proteins. FREE RADICAL RESEARCH 213 peptide provoked a decrease in the thiol pKa value by up to 1.6 pH units when compared to an ordinary thiol pKa value [79]. However, theoretical studies indicated that other features should also be taken into account (reviewed in Ref. [80]). More controversial than the macrodipole effect of helix a2 is the action of specific residues in acid–base catalysis. Several studies were performed to identify residues involved in the ionization of reactive Cys, but there are still uncertainties, which are probably related to differences among distinct Trxs. More challenging is the identification of the factors that affect the ionization of the thiol in the second Cys (Cys35 in EcTrx1), which is buried in the polypeptide chain. Other complicating factor is that the mixed disulfide is formed only transiently [69,81]. In the specific case of GS-SG reduc- tion, the pKa of GS-S-Trx-SH intermediate may not be very relevant because the rate of mixed disulfide intermediate formation is slow relative to the rate of its resolution [82]. It should be kept in mind that the rate constants for reduction of low molecular weight disul- fides by Trx are several orders of magnitude lower than the reduction of protein disulfides (reviewed in Ref. [56]). An Asp residue (Asp26 in EcTrx) is frequently con- sidered as one important residue for ionization of the second Cys (Cys35 in EcTrx1), residues through a hydro- gen bond network mediated by a water molecule since these two residues are 5.6 Å distant apart [23,83,84] (Figure 4B). However, biochemical and theoretical studies indicated that the reduction of the mixed disulfide between Staphylococcus aureus arsenate reduc- tase and SaTrx is not affected by Asp23 (Asp26 in EcTrx). In contrast, for EcTrx1, considerable experimental data have been accumulated to support a role of Asp26 in acid/base catalysis [23,67,83,85,86]. Indeed: (i) Asp26 affects the microscopic pKa values of the active thiols [23] and (ii) the catalytic efficiency (kcat/KM) for the EcTrx1D26A mutant was only about 10% of the wild-type [67]. Remarkably, the effects of Asp26 in catalysis are greatly decreased at high pHs [77,83,85–87], further indicating that this residue play a role in the deprotona- tion of Cys35 sulfhydryl in EcTrx1. Differences in the mechanisms among distinct Trxs may help to explain these conflicting results related to the conserved Asp residue (reviewed in Ref. [56]). Although less efficient than the wild-type protein, EcTrx1D26A mutant is still quite reactive towards thior- edoxin reductase and ribonucleotide reductase [67], displaying (kcat/KM) that are several orders of magnitude faster than thiol–disulfide exchange reactions between low molecular weight compounds. Therefore, other residues and structural factors should play a role in catalysis by Trx. Indeed, for Chlamydomonas reinhardtii thioredoxin h (CrTrxh), besides Asp30 (Asp26 in EcTrx) and the dithiol–disulfide residues (Cys36–Cys39), Trp35, and Asp65 were also proposed to participate in the catalytic mechanism [66]. Accordingly, EcTrx1W31A mutant (Trp31 is equivalent to Trp35 in CrTrxh) decreased the rate of insulin reduction [88]. Lys57 appears also to play some role in acid/base catalysis as L57M mutation of EcTrx1 provoked signifi- cant increase of thiols pKa’s (similar to the effect of D26A mutation). However, Lys57 of EcTrx1 had a minor role in the reduction of insulin or in the oxidation of thioredoxin reductase in comparison with the D26A replacement [67]. It would be interesting to study replacements of Lys57 and Asp26 by other residues than Met and Ala. In yeast, it was proposed that Asp24 (Asp26 in EcTrx1) and Lys54 (Lys57 in EcTrx1) act in a same mechanism involved in the dynamics of the water cavity in the activity site pocket. This water cavity is surrounded by Cys33, Lys54, and the buried Asp24 and appears to play relevant roles in the protein dynamics that occurs during transitions back and for between reduced and oxidized states [89]. Probably this water cavity in Trx mediates proton transfer and recognition of the target protein [90]. Finally, Pro76 (Figure 4B) is another residue that may play relevant roles in catalysis, besides its involvement in the folding process of EcTrx1 [91,92]. As mentioned above, this residue is located in the cis-Pro loop, conserved in the Trx fold that is relevant for substrate recognition [57,93]. Indeed, Pro76 is even closer to the C– X–X–C motive than Asp26 so it is possible that this residue also has roles in the catalysis of thiol–disulfide exchange reaction [56]. Indeed, P76A mutant of EcTrx1 possessed diminished activity than the wild type oxidoreductase [91]. Possibly, Pro76 could be involved in a proton shuttle mechanism between the two catalytic Cys [56]. Probably multiple factors play a role in the acidity of the thiol groups in the C–X–X–C motif of Trx enzymes. Therefore, Trx enzymes share several common residues and motifs that are involved in catalysis of thiol–disulfide exchange reactions. Moreover, Trx enzymes possess a common overall tertiary structure. In spite of these similarities, the distribution of surface charges varies considerably among distinct Trx enzymes [94–98], which affects their specificity to interact with their targets (reviewed in Ref. [56]). This is well illustrated for CFBPase [99,100], an enzyme involved in gluconeogenesis and Calvin cycle. The redox- active cysteines of CFBPase that are responsible for its activation are located in a solvent-exposed loop (170’s loop) that is highly negatively charged [101,102]. Therefore, Trx enzymes that undergo thiol–disulfide exchange with CFBPase are expected to have a 214 L. E. S. NETTO ET AL. corresponding positively charged patch. Indeed, posi- tively charged residues of Trx-f (a Trx belonging to the ferredoxin–Trx system) and of Trx-m favor their capa- cities to activate CFBPase [99,103]. Apparently, these electrostatic interactions contribute to the formation of a non-covalent Trx–CBPase complex, in which the hydro- phobic patch that comprise the C–X–X–C motif would allow optimal orientation for thiol–disulfide exchange reactions between Trx and CBPase [95,104–106]. Indeed, as it will become evident throughout this review, electrostatic factors are main factors conferring specifi- city for enzymatic thiol–disulfide exchange reactions. Initially, it was considered that chloroplasts rely in a unique redox system, in which Trxs are reduced by a ferredoxin-dependent Trx reductase, which in turn is dependent on ferredoxin reduced by the photosynthetic electron transport chain and, thus on light. One of the substrates of this ferredoxin–Trx system is CFPBase as described above. However, later on, the thiol-based system NADPH–Thioredoxin Reductase C (NRTC) was described in chloroplasts. NTRC is a multi-domain protein, containing in a single polypeptide two portions corresponding to the Trx (C-terminal) and thioredoxin reductase (N-terminal). Remarkably, the targets for the two thiol based systems are distinct, indicating the importance of features such as electrostatic interactions described above to guarantee specificity. One of the main targets for NTRC is 2-Cys Peroxiredoxins group (reviewed in Ref. [107]). Similarly to plants, bacteria also contain in a single polypeptide (called AhpF), portions of both Trx (N- terminal), and thioredoxin reductase (C-terminal). Frequently, the AhpF gene is neighbor of the AhpC gene (encoding a 2-Cys Peroxiredoxin) [108]. This is consistent to the integrated function of AhpF and AhpC to reduce hydroperoxides at the expense of NADH or NADPH (reviewed in Ref. [108]). Bacteria have other thiol based systems and again their specificities are distinct, although there are some overlaps (reviewed in Ref. [109]). For instance, for another group of Prxs, such as BCP/PrxQ from the plant pathogen Xylella fastidiosa, the electron donor is XfTsnC (a Trx similar to EcTrx1) and not XfAhpF [110]. In Saccharomyces cerevisiae, electrostatic factors were also identified as important elements underlying the species-specificity in yeast Trxs–ScTrxR1 interactions. Remarkably, ScTrxR1 is highly specific in the reduction of yeast Trxs but not Trx from bacteria or mammals [96]. While ScTrx1R can reduce yeast Trxs with extraordinary rates (second-order rate constants in the 107 M�1 s�1 range), no reduction of bacterial and mammalian Trx was observed [96]. Several factors underlie this specifi- city including the complementary electrostatic surfaces between ScTrxR1 and Trx enzymes loops (Figure 7). Previously, NMR analysis of ScTrx1 and ScTrx2 structures revealed that these oxidoreductases contain three highly flexible loops, which were implicated in Trx–substrate interactions [111,112]. In fact, single amino acids substi- tutions are found between the ScTrx1 and ScTrx2 loops (ScTrx1, 27-Y-ATWCGPCK-35; ScTrx2, 27-F-ATWCGPCK-35 and ScTrx1, 68-AEVS-A-MP-74; ScTrx2, 69-AEVS-S-MP-75), which appear to be related to differences with their reduction by ScTrxR1, since reciprocal amino acids substitutions provoked the inversion of reduction rates by ScTrxR1, and probably is related with the species specificity of Trx–TrxR reduction [96]. As a matter of fact, interactions with TrxR appear to have shapped the Trx enzymes (reviewed in Ref. [56]). There are five distinct groups of TrxR with N- and C-terminal extensions to a common catalytic core composed of a NADPH and a FAD binding domains [96]. Clearly, the evolution processes of Trx and TrxR enzymes evolution are connected. A challenge in the field is the elucidation of more structures of TrxR–Trx complexes, to reveal novel molecular aspects of protein–protein interactions. As a matter of fact, there is only one single complex between type 1 TrxR–Trx available in PDB database [73]. Very recently, again the concept that electrostatic factors are central aspects in the specificity of oxidor- eductases was demonstrated by employing phosphoa- denylylsulfate (PAPS) reductase from Escherichia coli [113]. Complementary electrostatic surfaces and not reduction potentials (also called redox potentials) determined specificity and efficiency of the oxidoreduc- tases from distinct organisms as reductants for PAPS reductase. As mentioned above, electrostatic inter- actions contribute to the formation of non-covalent Trx-target protein complexes, in which the hydrophobic patches around the C–X–X–C motif would allow optimal orientation for thiol-disulfide exchange reactions between Trx and target proteins [95,104–106]. Indeed, the hydrophobicity along the C–X–X–C motif appears to be important for interaction with redox partners (Figure 3). Accordingly, mutation of Gly33 (first X in the C–X–X–C motif in Trx) to either positively (Lys) or negatively (Asp) charged residues impaired the redox activity of EcTrx1 in both enzymatic and in cellular systems [114,115]. Indeed, structural analysis surround- ing C–X–X–C motif of thiol–disulfide oxidoreductases indicated that their environment is hydrophobic (Figure 3). Employing single molecule force spectroscopy and molecular dynamics simulations in eight Trx enzymes from distinct organisms comprising all three domains of life, further evidence was given that a hydrophobic binding groove is important for specificity in the binding FREE RADICAL RESEARCH 215 to the substrate [116]. Noteworthy, the cis-Pro loop, a conserved feature of the Trx fold, is a constituent of this groove. According to this proposal of a hydrophobic binding groove, isothermal titration calorimetry indi- cated a universal entropy mechanism for Trx enzymes. Interestingly, Trx recognizes the oxidized form of its target proteins with exquisite selectivity, compared with their reduced counterparts [117]. Noteworthy, a protu- berance (containing residues Glu50; Arg146 and Cys170 of yeast Tsa1) present only in the oxidized form of 2-Cys Peroxiredoxin is involved in its recognition as a target by ScTrx1 [118]. Anyway, it is still difficult to predict what proteins might be a Trx target. Probably, a good model for recognition of target protein by Trx enzymes involves initially electrostatic interaction that direct target pro- teins to the hydrophobic environment surrounding the C–X–X–C motif (Figure 6). It is also important to mention here that some Trx enzymes contain additional Cys residues besides those that take part in the C–X–X–C motif that may affect substrate recognition. HsTrx1 contains five Cys residues: Cys32 and Cys35 are the two residues that belong to the CGPC motif, whereas Cys62, Cys69, and Cys73 are three non-conserved residues. Cys62 and Cys69 work together as an additional dithiol–disulfide center. When a disul- fide is formed between Cys62 and Cys69, the ability of HsTrx1 to bind a target is impaired [119]. Remarkably, this Cys62–Cys69 non-active site disulfide is not a substrate for thioredoxin reductase, it is reducible by GSH/Grx1 system, whereas the Cys62–Cys69 dithiol can be oxidized by Prx1 [120]. Furthermore, Cys73 is also subjected to redox modifications, such as glutathionyla- tion [121], and oxidation into an intermolecular disulfide bond [122,123]. Recently, a mixed disulfide between HsTrx1 and tissue factor involving Cys73 was reported and could be important in substrate recognition [124]. Possibly, these multiple oxidative states of HsTrx1 involving its five Cys residues play a role in the fine Figure 7. Molecular surfaces and complementary electrostatic interactions between ScTrxR1 and ScTrx1/ScTrx2. (A, C, E): Molecular surfaces of ScTrxR1, ScTrx1, and ScTrx2, respectively. (B, D, F): Electrostatic distribution on the surfaces of ScTrxR1, ScTrx1, and ScTrx2, respectively; positive and negative charges are represented by blue and red colors, respectively. (A) Specific residues involved in the interaction with ScTrx1/2 are colored as: Cys142 and Cys145¼ orange; Lys137¼ blue; Asp146¼ red. The yellow, black, and green squares in (A) and (B) highlight the position of the Lys137, Asp146 and Cys142/Cys145, respectively. Once again, the squares highlight the position of the residues Tyr27 (ScTrx1) or Phe27 (ScTrx1) (purple), Ala71 (ScTrx1) or Ser72 (ScTrx2) (red), and cysteine disulfides (orange) (green) that are involved in the interaction with ScTrxR1 (more details on the effects of mutations on Ref. [96]). (See color version of this figure in the online version.) 216 L. E. S. NETTO ET AL. tune of the redox regulation of targets such as Ask1 (kinases), PTEN (phosphatases), and Ref-1 (transcription factor) (reviewed in Refs. [109,125]). Grx system Glutaredoxin (Grx) like Trx enzymes are heat stable thiol– disulfide oxidoreductases with the thioredoxin structural fold (Figure 8A) that rely on C–X–X–C motif (typically C- P-Y-C, but highly variable) to reduce disulfide bonds in the client proteins. Since Trx and Grx share a common structural fold, they probably also share the same ancestry [126,127]. ‘‘Glutaredoxin’’ (Grx) has a synonym- ous ‘‘Thioltransferase’’ name, reflecting the fact that these enzymes were independently discovered by two distinct groups. In one case, Grx was identified as the factor that still supports the reduction of ribonucleotide reductase (and thereby DNA replication) in strains of E. coli lacking thioredoxin [128]. This oxidoreductase was named thioltransferase by the authors that identified this enzyme by its ability to catalyze thiol–disulfide exchange reaction in rat liver, using [35S] GSH [129]. Grx denomination became more popular and will be used throughout this review. In spite the similarities with Trx proteins, Grx enzymes possess very distinct amino acid sequences, displaying a specific binding site for GSH (Figure 8) that enables them to efficiently participate in glutathionylation/deglu- tathionylation reactions (reviewed in Refs. [130,131]). As a matter of fact, this GSH binding site is relevant to understand the distinct mechanistic features between Trx and Grx enzymes. NMR and X-ray crystallographic structures of several Grxs in mixed disulfides with GSH [15,132–136] revealed conserved features of the GSH binding site (Figure 8B and C). This binding site comprises: (1) the active site itself, especially the N-terminal active site cysteine (colored in yellow in Figure 8); (2) the residues colored in green in Figure 8; and (3) the TVP motif colored in red (Figure 8). The Pro residue of this motif is the amino acid present in the cis-Pro loop characteristic of the Trx fold (reviewed in Ref. [131]). Like Trx, the most N-terminal Cys residue of the C–X–X–C motif in Grx is solvent exposed and it is Figure 8. ScGrx2 structure and GSH binding site. (A) Overall structure of the ScGrx2-SG mixed disulfide (PDB code: 3D5J) represented in cartoon and colorized in gray. The glutathione and Cys residues are represented in balls and sticks and colorized by CPK (C¼ gray, N¼ blue, O¼ red, and S¼ orange). GSH binding site residues are depicted in the cartoon: K24, blue; Q 63, purple; 32-CPYC-35, orange; 74-TVP-76; dark red; 88-NSD-90, dark green. (B) Detail of the ScGrx2 glutathione binding site residues and motifs represented by spherical dots, with the same colors presented before and the GSH molecule represented in spheres and colorized by CPK. (C) ScGrx2 molecular surface of the active site region containing a GSH molecule bound to Cys27. The hydrophobic residues are colored in purple, the Cys27 in orange and all the others amino acids residues in white gray. The glutathione molecule is represented in green. The figures were generated using the Pymol software (www.pymol.org). (See color version of this figure in the online version.) FREE RADICAL RESEARCH 217 responsible for the nucleophilic attack on the target disulfide. The most C-terminal Cys in the C–X–X–C motif is solvent inaccessible (also similarly with Trx). In contrast to the majority of Trx enzymes, this buried Cys for most Grx enzymes is not required for some catalysis of thiol– disulfide exchange reactions (reviewed in Ref. [137]), as this will be further discussed below. The description of enzymes belonging to the Grx family is increasing, with novel variants being described. It appears that the Grx family of enzyme is more heterogeneous than the Trx counterpart. For instance, besides the classical Grx enzymes that contain the C–X– X–C motif (called ‘‘dithiol Grx’’), monothiol Grx that contain a C-X-X-S motif were also described, in which the C-terminal Cys is replaced by a Ser. In some cases, Grx is a domain of multi-domain protein (reviewed in Ref. [130]). The variability among Grx types is very high, in terms of their ability to bind iron–sulfur clusters, oligomerization state, and enzymatic activities. For some Grx, it was not possible to observe catalysis of thiol–disulfide exchange reaction (reviewed in Refs. [130,137]). Noteworthy, the Trx family also displays heterogeneity with proteins containing extra domains, such as EcTrx2 that contain a zinc finger motif at the N-terminal extremity [138]. Glutaredoxin (Grx) enzymes can catalyze the reduc- tion of disulfide bond by two mechanisms, depending on the number of Cys residues involved in catalysis: one (monothiol) or two (dithiol). The dithiol mechanism is essentially identical to the oxidative half-reaction described for Trx (Figure 9B, steps I and II). It is composed of two sequential thiol–disulfide exchange reactions. In the first one, the most exposed Cys residue of Grx attacks the disulfide in the target protein, generating a mixed Grx–protein disulfide (Figure 9B, step I). Then the most buried Cys residue attacks the (A) (B) Figure 9. Mechanisms of disulfide bond reduction by the Grxs. (A) Monothiol mechanism. Reactive cysteine attacks the Target-SG disulfide, resulting in glutathionylated Grx (I), which is then reduced by GSH, generating GSSG (II) that is reduced by NADPH in a reaction catalyzed by the glutathione reductase (GR) (IV). Alternatively, an intramolecular disulfide bond in Grx can be generated by the nucleophilic attack of the second Cys residue to the mixed disulfide in glutathionylated Grx, releasing reducing GSH (III). (B) Dithiol mechanism. Reactive cysteine attacks the target protein disulfide, resulting in a mixed disulfide between Grx and the biological target, which is then resolved by the second Grx cysteine (I) resulting in the formation of an intramolecular disulfide bond in Grx (II), which is then reduced by GSH, resulting in reduced Grx and GSSG, which is in turn reduced back by the glutathione reductase enzyme (GR) using reducing equivalents from NADPH (IV). 218 L. E. S. NETTO ET AL. mixed disulfide, generating oxidized Grx and dithiolic target protein (Figure 9B, step II). The reductive half- reaction then differs from Trx enzymes, in which two GSH molecules (through two thiol/disulfide exchange reactions) reduce disulfide in Grx, generating reduced Grx and GSSG that is then reduced by glutathione reductase/NADPH (reviewed in Refs. [130,131]). The dithiol mechanism is rarely investigated for Grxs, possibly because there is no handy assay available. However, because ribonucleotide reductase from E. coli can be readily obtained by conventional expression and purification methods, the dithiol mechanism could be investigated in bacterial Grx [134]. Probably the same applies to PAPS reductase and OxyR (a bacterial redox regulated transcription factor), which are also reduced by bacterial Grxs via the dithiol mechanism (reviewed in Ref. [131]). The reduction of ribonucleotide reductase by bacterial Grx resembles in several features the catalysis by Trx enzymes: (1) mutation of the second Cys in the C– X–X–C motif provokes abolishment of the disulfide reductase activity [134,139], implying a dithiol mechan- ism; (2) electrostatic interaction appear to be responsible for specificity in enzyme–substrate interactions [140]; and (3) conformational changes are associated with ribonucleotide reductase binding to EcGrx1 [140]. The physiological significance of dithiol mechanism among Grx enzymes is not very well studied but recent studies with DrGrx2 (oxidoreductase from zebrafish and orthologous to mammalian Grx2) indicated that this enzyme is central to the homeostasis of the central nervous system. Indeed, silencing DrGrx2 expression resulted in loss of virtually all types of neurons by apoptotic cell death [141]. In both zebrafish and human cellular models for neuronal differentiation, Grx2 effects on vertebrate embryonic development are mediated by thiol–disulfide exchange reactions with collapsin, a central component of the semaphorin pathway [141]. Among eight Cys residues of collapsin, only two appear to be sensitive to hydrogen peroxide oxidation and this specific disulfide is target for DrGrx2 reductase activity through the dithiol mechanism [142]. Compared to the dithiol mechanism, the monothiol mechanism is more studied and many more targets were identified. It is generally related to the deglutathionyla- tion of substrates and involves a Grx-glutathione mixed disulfide intermediate [130,131,137]. In the first step, the most exposed Cys residue attacks the mixed disulfide between the substrate and glutathione, generating reduced substrate, and glutathionylated Grx (Figure 9A, step I). Then, a second GSH molecule reduces the mixed disulfide between Grx and glutathione, regenerating the reduced form of Grx (Figure 9A, step II). This is the rate limiting step of the overall catalytic cycle (reviewed in Ref. [130]). Noteworthy, in the monothiol mechanism, no mixed disulfide between Grx and target protein is formed. This might have consequences for the specificity in substrate binding, which remains to be elucidated. It is possible that other non-covalent factors (such elec- trostatic and hydrophobic interactions as described for Trx) between Grx and target proteins may provide specificity for enzyme–substrate interactions. Of note, mutation of the buried cysteine residue by serine in dithiol Grxs does not abolish the overall activity [135,139,143–145], thus validating the monothiol mech- anism. This is one of the features that distinguish Grx and Trx dithiol enzymes. The monothiol mechanism can be analyzed through a bi-substrate (GSH and glutathionylated target protein), steady-state kinetic approach, following NADPH oxida- tion in a NADPH-glutathione reductase coupled assay. For most substrates, parallel lines are obtained in Lineweaver–Burk plots, when the concentration of the second substrate is varied in between each line, which is characteristic of a ping–pong mechanism, i.e. with no ternary complex formation. The exception is the HEDS assay and this will deserve special attention below. Probably most Grx enzymes operate through an ‘‘encounter type mechanism’’, taking into account: (i) that the obtained KM and Vmax values by bi-substrate, steady state analyses are infinite (secondary reciprocal plots intercept at zero); and (ii) the lack of potent and specific inhibitors for these oxidoreductases, meaning that substrate and enzyme react without any binding (reviewed in Ref. [130]). The possible absence of enzyme–substrate complex is consistent with the fact that no mixed disulfide between Grx and target protein is formed in the monothiol mechanism. HsGrx2 may be an exception, since there are some evidences that this oxidoreductase can bind GSH tightly [146,147], which might be related with the fact that this protein can form Fe/S clusters. Moreover, DrGrx2 can catalyze reversible S- glutathionylation of the NAD+-dependent protein dea- cetylase sirtuin 1, which regulates vascular development in zebrafish [148]. Remarkably, DrGrx2 can reduce substrates by the monothiol (sirtuin 1) or dithiol (collapsin) mechanism [141,148]. Recently, DrGrx2 was also implied in heart development [149]. Specifically in regard to the first step (Figure 9B, step I) of the monothiol mechanism, this reaction is quite fast relative to the second step and it is highly selective for the glutathionyl moiety of the mixed disulfide [143,150]. For instance, reduction of disulfides that do not contain glutathione are not catalyzed by HsGrx1 [150] or by HsGrx2 [143]. The exclusive product of this reaction is the glutathionylated Grx species (no mixed disulfide between Grx and the target protein is detected) as well FREE RADICAL RESEARCH 219 documented by means of mass spectrometry, among other approaches (reviewed in Ref. [130]). As a matter of fact, mutants of HsGrx1 and EcTrx1 that contain only the solvent exposed Cys residue revealed a remarkable difference: EcTrx1 (but not HsGrx1) gave rise to both glutathionyl-Grx and also to the mixed disulfide between this oxidoreductase and the target protein [135]. This is probably related to the existence of a GSH binding site in Grx. As emphasized above, no mixed disulfide is formed between Grx and the target protein in the monothiol mechanism and as a consequence Grx enzymes can display low requirements for specific protein inter- actions. Accordingly, at least for some substrates (such as acyl coenzyme A binding protein), the pKa values and accessibility of their cysteine residues are the two factors that dictate substrate discrimination by HsGrx1 [151]. Accordingly, for other proteins differences in KM and Vmax values were attributed to steric factors [150]. Reduction of glutathionylated-Grx by a second GSH molecule (Figure 9A, step II) allows turnover of Grx and is considered the rate limiting step. Not only GSH, but also other thiols can reduce Grx-glutathione mixed disulfide. Remarkably, the corresponding pH profiles of the reaction rates match the deprotonation of the corres- ponding low molecular weight sulfhydryl [143,152]. These findings are consistent with the fact that the step II is the rate limiting step of the monothiol mechanism (reviewed in Ref. [130]). The rates enhance- ments by Grx in relation to the non-catalyzed reactions are in the one thousand range and obey the Brønstead relationship [Equation (2)], except for GSH where an additional enhancement effect was observed [143,152]. The very low pKa of the active site Grx cysteine may help in the catalysis, since thiolate is a good leaving group in the second step of the monothiol mechanism (Figure 9A, step II). This may be another distinguishing feature between Trx and Grx enzymes. Accordingly, the pKa values of the corresponding Cys residue in Trx are four orders of magnitude higher (�7.0) and these oxidor- eductases are not endowed with monothiol activity. One of the reasons why the monothiol mechanism is well studied for Grx enzymes is the availability of a simple and inexpensive assay. The HEDS assay is based on the reduction of HEDS (the disulfide of 2-mercap- toethanol) by GSH (reviewed in Ref. [137]). The formation of GSSG (a product of this reaction) is then monitored spectrophotometrically in a coupled assay with gluta- thione reductase through the oxidation of NADPH. Puzzling, whereas reductions of most glutathionylated substrates occur through a ping–pong mechanism, the kinetic pattern observed in the HEDS assay is the sequential. Although the HEDS assay is a simple procedure, the interpretation of the results requires caution. The HEDS assay is composed of two sequential thiol–disulfide exchange reactions as described below: OH� CH2 � CH2 � S� S� CH2 � CH2 � OH HEDSð Þ þ GSH! OH� CH2 � CH2 � S� SGþ OH� CH2 � CH2 � SH ðReaction 4Þ OH� CH2 � CH2 � S� SGþ GSH! OH� CH2 � CH2 � SHþ GSSG ����������������������������������������������� ðReaction 5Þ OH� CH2 � CH2 � S� S� CH2 � CH2 � OH HEDSð Þ þ 2GSH! 2OH� CH2 � CH2 � SHþ GSSG ðReaction 6Þ In a standard HEDS assay, all the reagents, less Grx, are incubated for a couple of minutes for the formation of the mixed disulfide between GSH and HEDS. The reaction is then started by the addition of Grx [153]. It is classically assumed that reaction 4 is not catalyzed by Grx (Figure 10) since this oxidoreductase prefers glutathionylated disulfide substrates. To clarify contro- versies on the reaction mechanism in the HEDS assay, the mixed disulfide GSSEtOH was synthesized and purified and subsequently analyzed in an analogous glutathione reductase – NADPH coupled photometric assay [154]. As expected for other glutathionylated substrates, the results were consistent with a ping– pong mechanism. In the HEDS assay, probably Grx can also catalyze the formation of the mixed disulfide between GSH and 2-mercaptoethanol (Figure 10, reac- tion I), even though the HEDS disulfide does not contain a glutathionyl moiety [154]. It is possible that GSH binds to a high affinity site in Grx, which might provide better orientation and geometry for reduction of a non- glutathionylated disulfide, such as HEDS. In spite of these controversies, the HEDS assay has been widely used to characterize Grx oxidoreductases. Most of the conventional dithiol Grx are active in the HEDS assay. Moreover, mutational analyses of the second cysteine in the C–X–X–C motif do not abolish this enzymatic activity (reviewed in Refs. [130,137]). ScGrx8 is an exception, since its ability to catalyze the reduction of HEDS at the expense of GSH is lost when the second cysteine residue of the C–X–X–C motif is mutated by serine [155]. Intriguingly, native monothiol Grxs (such as ScGrx3-5, which have a C-X-X-S motif) are usually inactive in the HEDS assay. Possible explanations for the lack of enzymatic activity of these monothiol Grxs 220 L. E. S. NETTO ET AL. might be related with the lack of activation of the second GSH molecule as a nucleophile; and poor leaving group properties of the N-terminal cysteine in the C–X–X–C motif among other possibilities (reviewed in Ref. [137]). For dithiol Grxs, the glutathionylated Grx intermediate can have two outcomes: (1) reduction by a second GSH molecule, regenerating reduced Grx (Figure 9A, step II); or (2) intramolecular disulfide formation due to the attack of the buried Cys residue (Figure 9A, step III). Reaction in step II allows turnover of Grx. Therefore, in principle, reaction in step III detracts Grx-SSG from turnover, slowing down the overall reaction. Therefore, to the extent that this reaction occurs, it is inhibitory. Indeed, mutation of the C-terminal active-site Cys by Ser in HsGrx1 and HsGrx2 resulted in a two fold increase in specific activity [135,143,156], since this mutation makes the reaction in step III (Figure 9A) impossible to occur. However, for other dithiol Grxs (such as EcGrx1 and ScGrx2) mutation of the second Cys residue provoked the opposite effect, i.e. decrease in the HEDS reduction activity [15,139]. Therefore, there should be biochemical and/or structural differences among dithiol Grxs. ScGrx1 and ScGrx2 are dithiol oxidoreductases, pos- sessing the typical conserved Cys-Pro-Tyr-Cys motif. Although these enzymes are 85% similar in terms of their amino acid sequence, ScGrx2 is 50 times more efficient than ScGrx1 in the HEDS assay [15]. Crystal structures of ScGrx2 in the oxidized form (intramolecular disulfide) and as glutathionyl mixed disulfide were elucidated [15] and compared with similar ones for ScGrx1 [133]. As expected, all the four structures overlapped quite well. However, distinct side-chain conformations of Ser30 (Cys30 in corresponding wild-type proteins) were observed in the glutathionylated struc- tures of the C30S mutants of ScGrx1 and ScGrx2, which are related with the fact that the distances between Ser30 and the reactive Cys27 are markedly distinct (3.47 Å in ScGrx1C30S and 5.14 Å in ScGrx2C30S). Probably, anything that increases the distance between the two sulfur atoms of the C–X–X–C motif might slow down the reaction in step III (Figure 9A) and thereby favor the overall monothiol mechanism [157]. Accordingly, site- directed mutagenesis experiments are consistent with this hypothesis [15]. Equivalent to the ScGrx1/ScGrx2 pair, similar features are present in the EcGrx1/EcGrx3 pair. Into EcGrx3 glutathionylated structure [134], Ser14 (which replaces the C-terminal active-site cysteine in this structure) presents a more buried conformation similar to that of ScGrx2C30SGS. In all other Grx structures reported in the glutathionylated form (all obtained with mutations of the C-terminal active site cysteine to serine), including the EcGrx1 [132] the conformation of this Ser is similar to that found in ScGrx1GS. The more buried conformation of Ser30 in the ScGrx2GS structure appears to be related to the interaction with other residues, including a Glu residue [15]. In the wild-type ScGrx2 protein, this configuration would make Cys30 less prone to disulfide formation [15]. Remarkably, like ScGrx2, EcGrx3 pos- sesses similar conformation in the active site and higher monothiolic activity than EcGrx1, although to a much lower extent (2-fold) than the ScGrx1/ScGrx2 pair [134]. Besides this structural factor, other features are associated with the catalytic activity of Grx enzymes. Since reaction in step II is the rate-limiting step in the monothiol mechanism [130,158], the very low pKa values (3.0–4.0) of the N-terminal Cys residue (that functions as Figure 10. Schematic representation of the HEDS assay. (I) Thiol–disulfide exchange reaction between HEDS and GSH, producing a mixed disulfide between 2-mercaptoethanol and GSH with the release of 2-mercaptoethanol. (II) Reduction of mixed disulfide between 2-mercaptoethanol and Grx, releasing 2-mercaptoethanol and S-glutathionylated Grx. (III) Then, GSH reduces the S- glutathionylated Grx, (IV) resulting in reduced Grx and GSSG, which is reduced by NADPH in a reaction catalyzed by glutathione reductase (GR) (Adapted from Ref. [154]). FREE RADICAL RESEARCH 221 the leaving group in this reaction) might favor the catalysis. This is consistent with the physicochemical principles described above (in the ‘‘Physico-chemical factors affecting thiol–disulfide exchange reaction for low molecular weight compounds’’ section) for thiol– disulfide exchange reactions for aliphatic, low molecular weight thiols that follow the Brønsted relationship [Equation (2)] [29,40]. Therefore, for HsGrx1, the rate constant enhancement of the catalyzed reaction (in which the pKa value of the leaving group is 3.5) over the uncatalyzed reaction (taking into account the pKa value of an ordinary Cys residue, such as BSA-SH equal to 8.5), it is predicted to be 4DpKa, i.e. 45 (�1000-fold) [151]. For HsGrx2, the predicted rate enhancement is 44, which is again consistent with the 4DpKa model, since the pKa value of its catalytic cysteine is 4.6 [143]. In contrast, Trx enzymes (with the N-terminal catalytic Cys with pKa values around 7.0) exhibit very low deglutathionylating activity and no monothiol mechanism (reviewed in Ref. [130,137]). Anyway, the extremely low pKa of the Grx catalytic cysteine does not fully account for the observed rate enhancement in the catalyzed reactions [130]. That is, the difference in catalytic cysteine pKa (around 1 pH unit) accounts for only about half of the 10-fold lower specific activity of HsGrx2 compared with HsGrx1 [130]. Accordingly, dithiol yeast Grxs (ScGrx1 and ScGrx2) possess about the same pKa values for the reactive Cys, but display distinct activities (50-fold difference) in the HEDS assay [15]. Although it is intriguing to observe the extraordinary low pKa values for the nucleophilic Cys in Grx, the chemical basis for this phenomenon are still not completely understood. Similar to Trx enzymes, the location of the solvent exposed Cys residue at the positive pole (N-extremity) of helix a2, only partially explains the high acidity of this thiolate. Indeed, studies with helical synthetic peptides, in which the thiolate anion is located at the N-terminus (positive pole) of a-helices resulted in only 1.6 units decreases in com- parison with corresponding thiol pKa measured in an unfolded control peptide [79]. Since pKa values for thiols in the reactive Cys residues in Grxs (�3.5) are five units below the pKa value of free Cys (�8.5), it is evident that other factors should operate. One possibility is that H- bonding and ion pair of N-terminal Cys thiolate with other residuescan be responsible to decrease pKa of reactive Cys residues in Grxs by 5 pH units [159]. Computational approaches for HsGrx1 predicted that Cys22–S� ��� +H3N–Lys19 the ion par might be relevant for the extraordinary low pKa, due to the proximity of the opposite charges. However, site directed mutagenesis provoked only slight changes in pKa values (less than 1 pH unit) [159]. Apart from the already mentioned dipole effect of helix a2, hydrogen bonds with the Thr21 hydroxyl group may also contribute to the thiolated stabilization [159], although this is also controversial [80]. Through the analysis of the HsGrx1 reduced structure, it was shown that the extraordinary low pKa of the Cys22 thiolate was likely due to stabilization by a local region of substantial positive potential resulting from the composite local environment rather than from a single interaction [160]. For HsGrx1, the overall dipole moment of the whole polypeptide extends directly through the active site pocket [160] (Figure 11A). We performed similar analyses in the electrostatic surfaces of ScGrx2 (Figure 11B) and EcGrx1 (Figure 11C), but in these cases no such dipole was observed. Since ScGrx2 and EcGrx1 Figure 11. Charge distribution on the molecular surfaces charge of Grx enzymes in reduced form. Surface representation of the H. sapiens Grx1 (A) (PDB code: 1JHB), S. cerevisiae Grx2 (B) (3CTG) and E. coli Grx1 (C) (1EGO). The position of sulfur atom (S�) of the reactive Cys is assigned by the ‘‘S’’ character in the yellow circle. The surfaces are represented in charge gradient and colorized as follow: blue¼ positive, white¼ neutral and red¼ negative. The coulombic surface color range from +5 (blue) to�5 (red) (kcal mol�1). (See color version of this figure in the online version.) 222 L. E. S. NETTO ET AL. also have extraordinary low pKa values, other still elusive factors should be involved in stabilizing the nucleophilic Cys in the thiolate form. Anyway, taking into account the Brønsted linear free energy relationship [Equation (2)], it was evident that the extraordinary low pKa of the Grx catalytic Cys does not fully account for the observed rate enhancement in the presence of GSH (reviewed in Ref. [130]). Indeed, when GSH is used as the second substrate, second-order rate constants are further increased for HsGrx1 [151] and HsGrx2 [143] over rates with other thiol substrates. Therefore, it is a challenge to reveal the Grx mechanisms that underlie the enhancement of GSH nucleophilicity as a second substrate for attack and turnover of the Grx-SSG intermediate (the rate limiting step of the overall monothiol pathway), especially considering the lack of kinetically relevant substrate-binding modes (see earlier). Possible mech- anisms for this enhanced nucleophilicity are acid/base catalysis of proton abstraction from the attacking GSH by Grx residues or by the glutathionyl moiety of the Grx-SSG mixed disulfide. Other hypothesis is the stabilization of the incipient thiolate of the second GSH substrate by positively charged basic groups on the enzyme. Indeed, mutation of the Lys (K19Q or K19L) close to the catalytic cysteine of HsGrx1, caused little change in the pKa of the catalytic cysteine but resulted in substantially lower specific activity [159], suggesting that Lys19 might enhance the nucleophili- city of GSH [130]. NMR characterization of the Grx- SSG mixed disulfide indicated that the Lys19 residue is also relevant in the stabilization of the adducted glutathionyl moiety [135]. Structure–activity relations in which distinct residues are mutated may yield important information about their roles in catalysis [130]. The physiological meanings of the ability of Grx to catalyze reversible S-glutathionylation has been extensively discussed elsewhere [161–163] and will not be further reviewed here. The amount of information in this subject is huge, reflecting the fact that more than 200 mammalian proteins are post-translationally mod- ified with the glutathionyl moiety (reviewed in Ref. [131]). Also, very relevant are the new findings on the complexes between Grx enzymes and Fe–S clusters that were reviewed recently [164] and impact biology and medicine. Therefore, Grxs mediate a myriad of processes such as DNA synthesis; regulation of gene expression; energy production; protein folding; protection against apop- tosis; sulfate assimilation, cytoskeleton arrangement. The mechanisms underlying these processes include de- glutathionylation (through the monothiol mechanism) or the disulfide reduction (through the dithiol mechan- ism) or binding to Fe–S clusters, which involve hydro- phobic and electrostatic interactions in the recognition of protein substrates. Txn–T(SH)2 system Another important example of specificity in thiol–disul- fide exchange reactions comes from organisms belong- ing to the Kinetoplastea order that includes Trypanosome and Leishmania genera. These protozoa (many of them parasites) present an unique thiol– disulfide network that include trypanothione [T(SH)2] an abundant low molecular weight thiol (reviewed in Refs. [165,166]). T(SH)2 is formed by the conjugation of two glutathione molecules with spermidine [N1,N8-(bis)-glu- tathionylspermidine], generating a dithiol [167]. The positively charged amino group in the spermidine bridge confers to the thiol groups of T(SH)2 a pKa value of 7.4 that coincides with the physiological pH and, therefore, is more acidic than the corresponding value of GSH [166]. As expected, formation of an intra- molecular disulfide is kinetically favored for T(SH)2, when Figure 12. The hydroperoxide decomposition by the system composed by TxnPx, Txn, T (SH2), and TR. The hydroperoxide decomposition is carried out by the Tryparedoxin peroxidase (TxnPx) releasing an alcohol and a water molecule. The disulfide oxidized cysteines from TxnPx are reduced by the enzyme tryparedoxin (Txn) using reducing equivalents from trypanothione T(SH2) which is oxidized to TS2 which is reduced by the Trypanothione reductase (TR) a flavoprotein that use reducing equivalents from NADPH. FREE RADICAL RESEARCH 223 compared with the intermolecular oxidation of two molecules of GSH [29,168]. T(SH)2 is the direct reductant of tryparedoxin (Txn), a disulfide reductase with homology to Trx but displaying a characteristic C–X–X–C (residues 40–45) and higher molecular weight (�16 kDa), among other distinct fea- tures [169]. T(SH)2 and Txn play central roles in the metabolism of trypanosomatids such as the synthesis of DNA precursors and hydroperoxide reduction [165,166]. Regarding, hydroperoxide metabolism, several thiol– disulfide exchange reactions take place and they confer specificity towards the overall system (Figure 12). Txn is the predominant oxidoreductase in trypanoso- matids [166]. Txn is reduced by T(SH)2 and oxidized by TxnPx and both of these thiol–disulfide exchange reactions display high selectivity that are mediated by electrostatic factors, as will be described below. Txn enzymes are endowed with a WCPPCR active site motif and form a distinct molecular clade within the super- family of thioredoxin-type proteins [170]. Regarding the reduction of Txn by T(SH)2, an unique mechanism was observed, in which electrostatic forces induce an unusual T(SH)2 conformation [169,171]. Structure of Txn2C44S in complex with mono-glutathio- nyl-spermidine [T(SH)] (that probably mimicks several features of the real substrate T(SH)2), displays Txn Arg129 interacting to the carboxyl group of T(SH) and a negative charge in Txn attracting the basic amine of spermidine (Figure 13A), whereas Trp40 and Pro43 perform hydro- phobic interactions with T(SH) (Figure 13B). Furthermore, a positive patch containing the Arg129 is probably stabilizing the structure of the mixed disulfide between Tnx2 and T(SH) (Figure 13C–E). Besides, modelling studies were also carried out and suggested that Arg45 (44), Glu73 (72), Ile110 (109), cis-Pro111 (110)-bond and Arg129 (128) residues are involved in the binding of CfTxn2 (CfTxn1) to T(SH)2, the two Arg residue would interact with the glutathionyl carboxyl groups of the dithiol [169]. In addition, the substitution of the Pro43 by Tyr may interfere with the Arg129 orientation [172]. Figure 13. Crystallographic structure C. fascicula Txn2 bound to glutathionyl-spermidine. (A) Cartoon representation of the CsTxn2C44S mutant (colorized in light blue) containing a glutathionyl-spermidine molecule making a disulfide bond with Cys41 (PDB code: 1I5G). The side chains of the residues related to substrate binding or catalysis as also the glutathionyl-spermidine are represented by sticks with carbon atoms colorized in light blue (CsTxn2) or dark green (glutathionyl-spermidine). Colors of other atoms follow: N¼ blue, O¼ red and P¼ dark orange. (B) Molecular interactions between CsTxn2C44S and the ligand. The enzyme is represented in ribbon and the amino acids involved in substrate binding are represented by ball and stick. Hydrophobic interactions are represented by the green dashed lines and polar ones by pink dashed lines. The interactions and representations were performed using the RCSB PDB ligand explorer 4.2.0 (http://www.rcsb.org/). (C) CsTxn2 with the cysteines oxidized to disulfide; (D) CsTxn2 bounded to T(SH); (E) The same as in D but with the ligand omitted to facilitate the observation of the binding site. The CsTxn2C44S surface was colorized as follow: blue¼ positive, white¼ neutral and red¼ negative. The glutathionyl-spermidine is represented by spheres and the atoms colorized as follow: C¼ dark green, N¼ blue, O¼ red, and P¼ dark orange. The coulombic surface color range from +10 (blue) to�10 (red) (kcal mol�1). (See color version of this figure in the online version.) 224 L. E. S. NETTO ET AL. Mutations in Txn were performed in an attempt to make its active site more similar to Trx or Grx, but they did not result in thioredoxin or glutaredoxin-like activities [172]. Probably, specificity in the Txn–T(SH)2 is achieved by complementary electrostatic interactions that are far away from the active site [169,171]. Following the same trend described before, Tryparedoxin peroxidase (TxnPx) enzymes are specific- ally reduced by TbTxn, less efficiently by Trx, and not reduced at all by GSH or T(SH)2 [173]. Therefore, the molecular aspects underlying the specificity of the reduction of TbTxnPx by TbTxn were evaluated by site- directed mutagenesis and by structural biology, includ- ing electron microscopy. Negatively charged residues Glu72, Asp76, Asp79, and Arg128 from TbTxn could establish saline bonds with the Arg92, Lys93, Arg94, and Glu171 from TbTxnPx (Figure 14). The structure of TxnPx from Trypanosoma cruzi is decameric [174], similarly to the structure of other 2-Cys Prx (Figure 15A–C). Considering this structure and electron microscopy images, a complex of decameric TcTxnPx with 10 molecules of TcTxn, one in each TcTxnPx active site [173] was modelled (Figure 15D). Due to the specificities of the thiol–disulfide exchange reaction, enzymes involved in the hydroperoxide metab- olism dependent on T(SH)2 and Txn can be good candidates for the design of inhibitors with the potential to the development of drugs. Other thiol–disulfide oxidoreductases In this item, some general aspects for disulfide bond formation protein A (DsbA) and protein disulfide isom- erase (PDI) will be described. DsbA and PDI act mainly as oxidants, in contrast to Grx and Trx, which react as reductants towards client proteins. The number of thiol disulfide oxidoreductases is extensive and specific reviews are available for each one. At the end of this topic, the use of lipoamide in thiol–disulfide exchange is also briefly discussed as an example of another oxidizing thiol (distinct than Cys residues discussed so far). According to the above sections, the mechanisms by which DsbA, PDI, Grx and Trx enzymes increase rates of thiol–disulfide exchange reactions are distinct, although they share the same structural fold and the C–X–X–C motif. [175]. Protein disulfide isomerase (PDI) in mammalian organisms and DsbA in bacteria are two oxidoreductases that catalyze reversible thiol-disulfide exchange reac- tions and are involved in disulfide formation in target proteins. It is important to realize that in spite of their biological role, oxidoreductases can catalyze reactions in both directions. This is particularly relevant for thiol– disulfide exchange reactions that are highly reversible [35]. Therefore, as any enzyme, DsbA and PDI accelerate thiol–disulfide exchange reactions in both directions and, as a consequence, the equilibrium is achieved at shorter times [35,130]. Anyway, as described for other thiol–disulfide oxidoreductases, electrostatic factors and acid/base catalysis also underlie the catalytic mechanism of PDI [176]. Protein disulfide isomerase (PDI) and DsbA also possess the Trx fold (Figure 16), but they present structural differences (reviewed in Ref. [57]). DsbA are single domain proteins (E. coli DsbA presenting a molecular weight of 23 kDa) with the characteristic C-P- H-C motif (Figure 16A) [177]; whereas PDI enzymes contain four Trx domains named a, a0, b, and b0 arranged in a U-shape form (Figure 16B) [178]. Domains b and b0 form a rigid base in the U-shaped tetra- domain molecule with respect to the more flexible a and a0 domains [57]. Only domains a and a0contain the typical C-G-H-C motif (containing the Cys53 and Cys56; Cys397 and Cys400, respectively – the numbering referring to the human isoform) found in PDI enzymes (Figure 16B). In turn, the domains b and b0 are related Figure 14. Complementary charge residues on the surfaces of TbTxnPx and TbTxn. (A) Surface model of TxnPx with each monomer of the homodimer colorized in green and yellow. The positive cluster represented by Arg92, Lys93, and Arg94 is represented in blue and the Glu171 in red. (B) Molecular surface of monomers of TbTxn colorized in rose with the acidic cluster (Glu72, Asp76, and Asp79) represented in red and the Arg128 in blue. Cys residues in both enzymes are in magenta. (See color version of this figure in the online version.) FREE RADICAL RESEARCH 225 to the substrate specificity of the PDI enzymes (reviewed in Refs. [178,179]). The reduction potentials of DsbA and PDI enzymes are more oxidizing (as example, EcDsbA has E�0¼�0.09 to�0.11 V; [10,11]) than Trx and Grx enzymes (as an example EcTrx1 has E�0 ¼�0.27 V [61]), which are in line with their roles in thiol–disulfide redox status (reviewed in Refs. [35,56]). Reduction potentials (also called redox potentials) refer to equilibrium conditions and it means by definition that electrons will tend to flow from the conjugated redox pair of lower reduction potential to the pair of higher reduction potential. Therefore, the more negative redox potential, the greater will be the trend to donate electrons. Reduction potentials are thermodynamic parameters that are correlated with Kox (for oxidation constant), which are particular redox equilibrium constants (Keq) for redox reactions as is the case of thiol–disulfide exchange reactions (For more details see Refs. [35,36]). The variations in the reduction potential described above (�0.15V) represent huge thermodynamics changes in these structurally related thiol–disulfide oxidoreductases, since they correspond to variation of Kox in the 105 range [35]. One of the factors determining this immense variation in the reduction potentials among thiol– disulfide oxidoreductases is the composition of the two amino acids in between the two Cys residues of a C–X–X–C motif. Remarkably, replacement of the two amino acids in between the two Cys of a reducing enzyme with those from a more oxidizing one causes the redox potential of the enzyme to become more oxidizing, and vice versa [10,24,64,88,180,181]. Indeed, in the case of E. coli DsbA (EcDsbA), whose active site motif is C–P–H–C, mutations of the Pro and His residues provoked changes in the equilibrium constant higher than 1000 thousand fold [10], which was correlated with the pKa of the thiolate in its reactive Cys [182]. Then the C–X–X–C active-site motif of thiol– disulfide oxidoreductases was proposed to act as a Figure 15. Crystallographic structure T. cruzi TcTxnPx and model of TcTxbP–Txn complex. Cartoon representation of the TxnPx from Typanossoma cruzi (TcTxnPx) homodimer (A) and decamer (B) (PDB code: 4LLR). In both representations, one monomer is colorized in yellow and the other in green. The Cys residues are represented by spheres and the carbon atoms in a color tone similar to the cartoon monomer. The sulfur atoms are in orange. (C) TcTxnPx surface colorized by charge (red¼ negative, blue¼ positive, and white¼ neutral). (D) Model of the Txn–TxnPx complex. The homo dimers of the decameric structure TcTxnPx are colorized in green and yellow. Monomers of TcTxn are colorized in red or rose. (See color version of this figure in the online version.) 226 L. E. S. NETTO ET AL. redox rheostat, the sequence of which determines its reduction potential properties [182]. Although the two X–X residues in the C–X–X–C are important for the reduction potential, other factors appear to affect this thermodynamic parameter. Indeed, as mentioned before, proteins with the thioredoxin fold possess a loop containing a cis-proline, which closely approaches the C–X–X–C motif [183]. The residue that precedes the cis-proline can take part in hydrogen bond with the catalytic Cys residues and thereby affect reduction potentials [183]. Mutational studies indicated that the cis-proline loop is also involved in stabilization of EcDsbA mixed disulfide intermediates with substrates in vivo [184], metal binding by the active site Cys residues [185]; and binding and positioning of the disulfide present in the substrate [93,186]. Studies with archeal Trxs are in an apparent contrast with the rheostat model [187], reinforcing the idea that there are multiple factors governing the reduction potentials of thiol–disulfide oxidoreductases, besides the compos- ition of the amino acids in the C–X–X–C motif. In any case, the reduction potentials are not neces- sarily determinants of the function of these proteins as oxidants or reductants. In general, kinetics predominates over thermodynamics in controlling reactions pathways in living organisms [188]. For instance, mutations at position 33 (first X in the C–X–X–C motif of Trx) do not affect significantly their redox potential but do affect their activities in vitro and in vivo [114,115]. Moreover, EcGrx1 catalyzes protein disulfide formation 30-fold faster than rat PDI, although thermodynamically rat PDI is a 600-fold better oxidizing agent than EcGrx1 [172]. Also, EcTrx1 can also act as an oxidant in vivo under certain condition, promoting disulfide bond formation in the cytoplasm [189,190]. The most important factor to determine if an oxidoreductase will favor disulfide reduction or forma- tion is the redox environment where the protein is located (reviewed in Ref. [191]). The redox environment of a cellular compartment is influenced by various redox couples that interact with each other in a dynamic way, under the influence of metabolic and anabolic pro- cesses. Since glutathione is by far the most abundant low molecular thiol in cells, the couple GSH/GSSG should be frequently considered in order to analyze the function of a thiol-disulfide oxidoreductase, taking into account that in cells kinetics rather than thermo- dynamics prevail (reviewed in Ref. [188]). Therefore, disulfide bond formation in proteins takes place at the appropriate cell compartments, where the concentrations of GSSG are higher. In the case of eukaryotic organisms, PDI is located within the endo- plasmic reticulum (reviewed in Ref. [192]), whereas in the case of bacteria, DsbA is located within the periplasm of Gram-negative (reviewed in Ref. [193]). These are oxidizing environments that favor oxidation of dithiols into disulfides before protein secretion. Studies using ribonuclease A as a model gave support to the postulate that, at least for small globular proteins, the native structure is determined only by the protein’s amino acid sequence [194]. However, PDI and DsbA can accelerate this process, since disulfide bond formation is often the rate-limiting step in the overall folding process [195]. After introducing a disulfide into a client protein, PDI is reduced. To turnover, PDI needs to be re-oxidized (Figure 17A and B). The classical pathway involves the Figure 16. Overall structures of the DsabA and PDI. (A) Overall crystallographic structure of the one – domain EcDsbA with Cys in reduced form (Cys30 and Cys33) (PDB code: 1A2L). (B) Structure of the reduced HsPDI (PDB code: 3ELI) revealing the U shape conferred by the spatial organization of the four domains (a, a0, b and b0), all of them off with the Trx fold. � sheets are represented in light green, the � helices in dark red, and the turns/random coils in yellow orange. Catalytic Cys residues are depicted as spheres and colorized as follow: C¼white, N¼ blue, O¼ red, and S¼ orange. (See color version of this figure in the online version.) FREE RADICAL RESEARCH 227 flavin–thiol/disulfide containing protein named Ero1 (Figure 17A). After oxidizing PDI, ERO1 gets reduced and also needs to be re-oxidized to turnover. This happens through a sequence of reaction, whose final electron acceptor is molecular oxygen, giving rise to hydrogen peroxide (Figure 17A). Therefore, formation of disulfide bond in the ER has a potential oxidative threat to the cells. Mammalian Prx4 was recently discovered to have two roles in the ER: (i) as hydrogen peroxide removing enzyme and (ii) as PDI oxidizing (Figure 17B) [196,197]. In less complex eukaryotes such yeast, there is only one gene for ERO1 and no Prx to oxidize PDI, whereas in mammals there are two ERO1 (ERO1a and ERO1b) and Prx4. ERO1a is widespread in diferent cell types, whereas ERO1b is more restricted to the pancreas [36]. Then, it was predicted that the triple mutant Ero1a/ ERO1b/Prx4 mice would present a very severe pheno- type, since PDI re-oxidation would be severly affected [198]. Surprisingly, that was not the case. The triple mutant mice displayed only mild phenotype with scurvy-like features. In analogy with classical scurvy, it was hypothesized that the scurvy-like phenotype was associated with reduced ascorbate levels. Oxidation of ascorbate by sulfenic acids, as described by 1-Cys Prx [199] is one possible pathway to divert this vitamin from enzymes involved in the synthesis of collagen. According to their hypothesis, deletion of all three PDI oxidases lead to accumulation of sulfenates (Cys-SOH) and reduction in the steady-state levels of ascorbate. These data indicated that the redox pathway involving ascor- bate and sulfenic acids may be widespread, being kinetic characterization required to validate this proposal. Anyway, several other redox reactions take place in the ER during oxidative protein folding, such as those mediated by Gpx7 and Gpx8, which may represent alternative ways to oxidize PDI (reviewed in Ref. [200]). Protein disulfide isomerase (PDI) has a second activity that is rearrangement of disulfide bonds in proteins, which is called disulfide isomerization (Figure 17D), which is driven by energy minimization during the folding process of the client protein [57]. Disulfide isomerization activity requires PDI in the reduced state to start a sequence of thiol-disulfide exchange reactions that will lead to disulfide shuffling (Figure 13D). Two thiol disulfide exchanges (each one involving two Figure 17. Redox pathways involved in the disulfide bond formation and disulfide bond isomerization activities of PDI. (A) Disulfide bond formation results in reduced PDI, which is re-oxidized by Ero1, which in turn is re-oxidized by molecular oxygen (O2), releasing hydrogen peroxide. (B) Disulfide bond format