Biochimica et Biophysica Acta 1830 (2013) 3217–3266

Contents lists available at SciVerse ScienceDirect

Biochimica et Biophysica Acta

journal homepage: www.elsevier.com/locate/bbagen

Review and the reaction mechanisms of glutathione-dependent ☆

Marcel Deponte ⁎

Department of Parasitology, Ruprecht-Karls University, Im Neuenheimer Feld 324, D-69120 Heidelberg, Germany article info abstract

Article history: Background: Glutathione-dependent catalysis is a metabolic adaptation to chemical challenges encountered Received 31 August 2012 by all life forms. In the course of evolution, nature optimized numerous mechanisms to use glutathione as the Accepted 25 September 2012 most versatile nucleophile for the conversion of a plethora of -, - or -containing electrophilic Available online 2 October 2012 substances. Scope of review: This comprehensive review summarizes fundamental principles of glutathione catalysis and Keywords: compares the structures and mechanisms of glutathione-dependent enzymes, including glutathione reductase, Catalysis glutaredoxins, glutathione , , glyoxalases 1 and 2, glutathione and Glutathione MAPEG. Moreover, open mechanistic questions, evolutionary aspects and the physiological relevance of gluta- thione catalysis are discussed for each enzyme family. Major conclusions: It is surprising how little is known about many glutathione-dependent enzymes, how often Electrophile reaction geometries and acid–base catalysts are neglected, and how many mechanistic puzzles remain unsolved despite almost a century of research. On the one hand, several enzyme families with non-related folds recognize the glutathione moiety of their substrates. On the other hand, the thioredoxin fold is often used for glutathione catalysis. Ancient as well as recent structural changes of this fold did not only significantly alter the reaction mechanism, but also resulted in completely different protein functions. General significance: Glutathione-dependent enzymes are excellent study objects for structure–function rela- tionships and molecular evolution. Notably, in times of systems biology, the outcome of models on glutathione metabolism and redox regulation is more than questionable as long as fundamental enzyme properties are neither studied nor understood. Furthermore, several of the presented mechanisms could have implications for drug development. This article is part of a Special Issue entitled Cellular functions of glutathione. © 2012 Elsevier B.V. Open access under CC BY-NC-ND license.

1. Introduction mechanistic models. I will approach the topic from two perspectives: In Section 2, I will start with a focus on the substrates. I will present Glutathione is the central redox agent of most aerobic organisms. theories on the origin and benefits of glutathione-dependent pro- Its reduced form (GSH≡γ-L-glutamyl-L-cysteinylglycine) serves as a cesses, summarize the properties of this extraordinary molecule and ubiquitous nucleophile in order to convert a variety of electrophilic provide an overview of the glutathione-dependent enzymes and substances under physiological conditions. Glutathione-dependent pathways. The mechanisms of glutathione-dependent enzymes and enzymes significantly accelerate most of these chemical reactions their physiological relevance will be subsequently discussed and com- in numerous metabolic pathways. Accordingly, tens of thousands pared in Sections 3–8. of articles on glutathione-dependent enzymes and pathways have been published since the disputed discovery of glutathione by 2. Theories on the benefits, functions and evolution of Hopkins as well as Hunter and Eagles in the 1920s [1]. It is therefore glutathione catalysis rather surprising that many fundamental mechanistic questions still remain to be solved in order to precisely understand the role of gluta- 2.1. Two chemical challenges for life thione metabolism at the cellular and organismic level. This review is a (doomed) attempt to summarize the knowledge on glutathione- Why do we need a glutathione system? Life as we know it has en- dependent catalysis and to outline the relevance of the current countered several chemical challenges in the course of evolution. In fact, countless “natural” chemicals—including electrophilic substances— are carcinogens, mutagens, teratogens and clastogens [2,3]. In addition ☆ This article is part of a Special Issue entitled Cellular functions of glutathione. ⁎ Tel.: +49 6221 56 6518; fax: +49 6221 56 4643. to xenobiotics, two of the presumably most important chemical chal- E-mail address: [email protected]. lenges are (i) the formation of reactive oxygen species (ROS) due to an

0304-4165 © 2012 Elsevier B.V. Open access under CC BY-NC-ND license. http://dx.doi.org/10.1016/j.bbagen.2012.09.018 3218 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 aerobic atmosphere, and (ii) the formation of 2-oxoaldehydes (2-OA) intracellular conditions, to avoid the formation of ROS, to detoxify due to glycolysis and other fundamental metabolic pathways. ROS, and to reverse or repair ROS-derived damage [8–10]. On the other hand, partially oxidizing conditions as well as appropriate 2.1.1. The formation of reactive oxygen species redox steady states in different cellular compartments became essen- Oxygenic photosynthesis most likely caused the first global “envi- tial for life. So-called occurs only when the balance ronmental pollution crisis”. As a consequence of anoxygenic and between the formation and the removal of ROS is disturbed, thereby oxygenic photosynthesis, the presumably reducing, sulfide- resulting in the accumulation of oxidized and damaged biomolecules enriched oceans and atmosphere changed to oxidizing, oxygen- [10]. Please note that precise mechanistic definitions of oxidative enriched habitats with two significant oxygenation boosts occurring stress at the molecular level are just beginning to emerge and seem approx. 2.5–2.2 and 0.8-0.5 billion years ago (Fig. 1) [4,5]. Under the to highly depend on the cell type or organism. present conditions, electrophilic ROS are expected to be easily formed in all aerobic organisms with the help of light, flavins, semiquinones 2.1.2. The formation of 2-oxoaldehydes as well as , and other ions (Fig. 2A) [6–9].H2O2 •− Glycolysis-dependent ATP-formation is an imperfect process. During and O2 can both react with selected containing Fe/S-clusters, “ ” – – 2+ an unwanted side reaction of the Emden Meyerhof Parnas pathway, liberating their iron ions. Free or complexed Fe reduces H2O2, yield- • phosphate is eliminated from the triosephosphates glyceraldehyde- ing OH which unspecifically modifies all kinds of biomolecules at a 3-phosphate (GAP) and dihydroxyacetone-phosphate (DHAP) (Fig. 2C) diffusion-limited rate. Hence, radicals, sulfenic acids, disulfides and [21–23]. The molecular architecture of the glycolytic enzyme triose- (hydro)peroxides are directly or indirectly formed by ROS (Fig. 2B). phosphate (TIM) stabilizes the enediolate intermediate These ROS-dependent modifications result in inactivated proteins, of the isomerization reaction and therefore significantly reduces this damaged membranes and mutations [8–10]. ubiquitous side reaction [24]. Nevertheless, the elimination product However, thiyl radicals, disulfides, sulfenic acids and ROS can (MG) is continuously generated at a low level. For exam- also fulfill vital functions: (i) Some ROS are not only involved in the ple, in human red blood cells about 0.1% of GAP and DHAP were estimat- defense against pathogens, but can also serve as signal mediators in ed to end up as MG [25].Evenarchaea—using the Entner–Doudoroff the redox regulation of metabolism and transcription. Accordingly, instead of the Emden–Meyerhof–Parnas pathway—have a functional there are several proteins and enzymes that either sense or even gen- TIM for gluconeogenesis [26] and were shown to produce MG [27]. erate ROS [7,11,12]. Excellent examples for the latter enzymes are MG and other structural analogs of glyoxal (OCHCHO≡ethanedial) myeloperoxidases, producing HOCl, and NADPH-oxidases, generating • − are 2-oxoaldehydes (2-OA). In addition to gylcolysis these compounds O [13]. (ii) Some cysteine-derived thiyl radicals, sulfenic acids and 2 are also formed during peroxidation as well as acetone, glyc- disulfides are pivotal intermediates during catalysis or could serve erol and metabolism [21,23,28,29]. Owing to the adjacent as signal mediators [7,11,14,15]. Of note, the reduction of ribonucleo- carbonyl groups, 2-OA are strong electrophiles that spontaneously tides is a peculiar example for a fundamental thiyl radical-dependent react with nucleophiles from proteins, and nucleic acids, as well as disulfide-dependent physiological process in all domains of thereby yielding so-called advanced glycation endproducts (AGEs) life [16,17]. (iii) The importance of protein disulfide bonds is further- (Fig. 2D). As a consequence, 2-OA are potentially cytotoxic and more underlined by the fact that bacteria and eukaryotes established mutagenic, and their removal by a detoxification system is benefi- non-related analogous machineries to stabilize secreted and intracel- cial [30–32].However, and other bacteria sometimes lular proteins in the periplasmic space, the endoplasmic reticulum even generate MG with the help of methylglyoxal synthase to me- and the mitochondrial intermembrane space [18–20]. tabolize DHAP under conditions of limited phosphate [21,28,33]. In summary, on the one hand, the ancestors of modern organ- As outlined in Section 7.4, 2-OA can be also involved in signal trans- isms had to develop numerous mechanisms to maintain reducing duction and cellular differentiation. Hence, the structures, cellular concentrations and effects of 2-OA highly depend on the often neglected biological context. In summary, 2-OA are ubiquitous elec- trophilic metabolites that are usually detoxifiedbutthatmightalso exert regulatory functions in analogy to the janus-faced hydroper- oxides [31].

2.2. One single solution: glutathione

2.2.1. Overview of glutathione metabolism and catalysis How are the chemical challenges outlined in Section 2.1 mastered? The glutathione system—together with the thioredoxin system— probably evolved very early in aerobic organisms (Fig. 1). Owing to the cysteine moiety of GSH, the whole system is based on common sulfur biochemistry (Fig. 3A). It therefore requires, (i) an electron Fig. 1. The evolution of aerobic life and glutathione metabolism. Oxygenic photosyn- relay, linking the universal reducing agent NADPH to thiol/disulfide- thesis resulted in an oxidation of the environment followed by a delayed increase of free oxygen in the atmosphere (during the so-called 1st and 2nd great oxidation metabolism, and (ii) a thiol-containing adapter molecule to transfer event highlighted in red). Several glutathione-dependent enzymatic activities are electrons to a set of different acceptors. Flavoproteins are widely found in contemporary eukaryotes as well as purple bacteria and cyanobacteria but used as electron relays [18]. Hence, it is not surprising that the reduc- seem to be absent in many other bacteria and archaea. Ondarza as well as Fahey and ing equivalents from NADPH enter the glutathione system either with colleagues therefore suggested that glutathione metabolism evolved together with oxygenic photosynthesis [86,549–551]. More recent in silico analyses revealed that the help of the FAD-dependent enzyme glutathione reductase (GR) the domains of some glutathione-dependent enzymes such as Grx and GST are found [34–36] or the thioredoxin reductase/thioredoxin couple (TrxR/Trx) in all kingdoms of life, including some archaea and all kinds of bacteria [203,479] [37–43]. The electrons are subsequently transferred to glutathione (Deponte, unpublished). Thus, a putative earlier evolution of glutathione-dependent disulfide (GSSG), yielding two molecules of GSH (Fig. 3B). GSH either enzymes and a subsequent loss or replacement in bacteria and archaea cannot be serves as a reducing agent for disulfides (Fig. 3C) and hydroperoxides fully excluded. Nevertheless, based on the current data, it seems more likely that the few encoding glutathione-dependent enzymes in archaea and bacteria originate (Fig. 3D), or is conjugated with 2-OA (Fig. 3E) and other electrophilic from horizontal transfers. substances (Fig. 3F). Alternatively, GSSG can also oxidize under M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3219

Fig. 2. Formation of ubiquitous electrophiles and subsequent modification of biomolecules. (A) Formation of ROS owing to flavin- and Fenton-chemistry as well as other catalyzed or spontaneous electron transfers. The chemical formula, the oxidation number and the Lewis structure of oxygen, superoxide anion, and hydroxyl radical are shown from the left to the right. (B) Representative modifications of molecules by ROS leading to the formation of low and high molecular weight peroxides, radicals, sulfenic acids, nitrosothiols and disulfides. (C) Formation of MG as a by-product of glycolysis due to the elimination of phosphate. (D) Exemplary modifications of arginine, lysine and guanine residues (Arg, Lys and Gua, respectively) by glyoxal, MG and other 2-OA. The modified biomolecules are often summarized as AGEs.

certain conditions (Fig. 3C) depending on thermodynamic and, in par- condition of the cell (stressed, apoptotic, etc.) were also reported to ticular, kinetic parameters as outlined in the next section. influence the GSH/GSSG ratio [46,47]. As a consequence, GSH is

In summary, disulfide-reducing GR and TrxR act as electron relays not only a potent nucleophile—despite a rather high thiol pKa value of to tap into the NADPH pool, GSH is a versatile adapter molecule, and approx. 9 [44,48]—but also an extremely flexible biological reducing the glutathione system serves in most aerobic cells and organisms agent [44,49]. as the central metabolic network to remove or modify endogenous What is more important for glutathione catalysis: the kinetics or electrophilic compounds and numerous xenobiotics. Accordingly, the thermodynamics? As emphasized by Flohé in this BBA issue the effects that are summarized in Fig. 2B,D are mastered with the [50], cells and organisms are open systems. Thus, metabolic fluxes help of GSH, demonstrating the versatility of glutathione-dependent are in transition or in regulated steady states, and isolated E′ values catalysis as an answer to different chemical challenges in the evolu- at equilibrium do not necessarily explain whether a reaction is of tion of life. physiological significance or not. It is the kinetics that determines whether a potential is utilized in a physiological context. So what is 2.2.2. The kinetics and thermodynamics of glutathione catalysis the relevance of measuring redox potentials and glutathione con- As depicted in Fig. 3 and as outlined in the following sections, centrations [50]? A controversy resulting from this valid question several glutathione-dependent reactions are catalyzed by a variety might be solved by considering theoretical studies on the general reg- of enzymes with different physiological concentrations as well as ulation of metabolic fluxes by Hofmeyr and Cornish-Bowden [51,52]: kcat and Km values. Some of these enzymes exert overlapping func- According to their model, control of metabolism can be understood in tions and/or exist in a variety of isoforms.1 Thus, the relevance and terms of elasticities of supply and demand. Each elasticity coefficient rates of the reactions in Fig. 3 highly depend on the overall enzyme rep- is the sum of a thermodynamic term (depending on the law of mass ertoire (V=−d[S]/dt=V1+V2…=kcat1[E1][S]/(Km1 +[S])+kcat2[E2] action) and a kinetic term (determined by the enzymatic repertoire [S]/(Km2 +[S])…). The thermodynamic parameter expressing the and its status). The thermodynamic term in the supply elasticity be- driving force of the redox reactions in Fig. 3 is the redox potential E′, comes negligible at conditions far from equilibrium but “completely which can be easily derived from the Gibbs energy. In contrast to swamps the kinetic term” near equilibrium [51,52]. In other words, many other physiological redox buffers, the redox potential of the glu- the relevance of the measured redox potentials and glutathione tathione system not only depends on the GSH/GSSG ratio, the tempera- concentrations depends on whether the analyzed flux is close to or ture and the pH, but also on the actual concentration of glutathione as far from equilibrium. exemplified by the Nernst equation in Fig. 4 [44,45].Theintracellular In summary, the GSH/GSSG couple is the redox buffer of the gluta- concentration of GSH is quite high and ranges from approx. 0.1 to thione system maintaining appropriate redox conditions from the 15 mM. The concentration of GSSG is usually several orders of magni- suborganellar to the organismic level. The glutathione-dependent tude lower. Both concentrations depend on the subcellular compart- reactions summarized in Fig. 3 highly depend on kinetic parameters ment (Fig. 4), the cell type and the organism. The cell cycle and the and the enzymatic repertoire. The relevance of measured redox potentials and glutathione concentrations for redox metabolism is fl 1 Please note that the term “isoform” is used for homologous proteins without im- controversial and probably depends on the metabolic ux and the plying that such proteins have similar functions or are even isozymes. distance from equilibrium. 3220 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266

2.2.3. GSH as a reducing agent for disulfides and the reduction of GSSG enzymatically or by glutaredoxins (Grx) (Fig. 3C) [14,53–55]. In addi- The roles of GSH as the major reducing agent for disulfides and of tion, the reduction of non-native and the formation of native protein GSSG as a major thiol-modifying agent are mediated either non- disulfide bonds in the endoplasmic reticulum depend on GSH,

Fig. 3. Overview and current models of glutathione metabolism. (A) Composition and redox conversion of GSH and GSSG. (B) NADPH-dependent regeneration of GSH by GR and/or by Trx and TrxR. Please note that the direct reaction of GSSG with Trx in vitro is rather slow, and the Trx-dependent reduction in vivo might therefore be indirect (Section 2.2.3). n.-e., non-enzymatic. (C) Reduction or oxidation of intra- or intermolecular disulfides or thiols by 2 GSH/GSSG (left panel). Deglutathionylation/glutathionylation of thiols by GSH/ GSSG (right panel). The reactions can occur either non-enzymatically or enzymatically with the help of Grx, PDI and some GST-isoforms. Reactants include proteins and low molecular weight compounds. (D) The GSH-dependent removal of H2O2 and other hydroperoxides is catalyzed by a variety of enzymes including specialized GPx, Prx, GST and a few Grx-isoforms. (E) The GSH-dependent conversion of 2-OA to 2-hydroxycarboxylic acids is catalyzed by the isomerase Glo1 and the thioesterase Glo2. (F) A variety of other electrophiles are modified by GSH with the help of GST and MAPEG. The products are either removed from the cell or are precursors for metabolites such as eicosanoids. M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3221

Fig. 4. Correlation between the half cell E′ and the percentage of oxidized glutathione. The equilibrium between GSSG and GSH can be calculated using the Nernst equation [45], resulting in sigmoidal E′-GSSG diagrams. E′ not only depends on the [GSH]/[GSSG] ratio but also on the indicated total concentration of glutathione (as em- phasized in the upper right version of the Nernst equation). An increase of glutathione—e.g. due to the de novo biosynthesis or uptake of GSH—shifts the curve to the left. The pro- tonation of both sulfur upon GSSG reduction depends on the pH value which therefore also affects E′. Please note that the pH at 25 °C is already considered in the presented diagrams and versions of the Nernst equation (E°′=EpH7(25°C) =−0.24 V). At a more alkaline pH all curves are shifted to the left: EpH =−240–59.16×(pH—7.0) mV, resulting in shifts of −24 and −59 mV at pH 7.4 and 8.0, respectively [45]. Please also note that the curves are based on calculated concentrations instead of the activities aGSH and aGSSG, neglecting the fact that salts/H+/OH− as well as side chains all interact with the thiol, amino and carboxylate groups of glutathione and therefore influence E′. Calculated redox potentials and glutathione ratios from different subcellular compartments in [552–554] and mammals [46,555–557] at estimated pH values are indicated for compar- ison. Most of the values should be interpreted with caution because the exact concentrations of GSH and GSSG in the compartments were often not determined (nd), and the parameters depend on the metabolic and developmental conditions as well as the chosen methodology [45,46,555]. Obviously, much more work is necessary to obtain reliable and comparable values for E′, pH, [GSH] and [GSSG] of all subcellular compartments.

GSSG and protein disulfide (PDI). (The exact mechanisms abundant Trx-dependent hydroperoxidases—can also utilize GSH as of PDI in vivo still remain to be clarified [56,57] and are not discussed an electron donor [63–67] and are therefore discussed in Section 6. in this review.) Once a disulfide bond has reacted with GSH (or a thiol Noteworthy, in addition to specialized GPx- and Prx-isoforms, some has reacted with GSSG), the stability of the resulting glutathionylated Grx- and many glutathione transferases (GST) can also act as hydro- molecule can vary over several orders of magnitude (Fig. 3C). The peroxidases on their own. However, the rate constants of these stability depends on whether the mixed disulfide is an intermediate enzymes, if determined, were usually found to be significantly lower during catalysis, a species required for redox-mediated signal trans- than for or the canonical thiol/selenol-dependent hydro- duction, a protected cysteine residue under oxidizing conditions or peroxidases Prx and GPx [68–72]. In summary, there are numerous a biosynthetic product. These differences are highly important with proteins with a GSH-dependent hydroperoxidase activity. Their con- respect to the diversity of Grx-isoforms as described in Section 4.3. tribution and relevance are often unknown but seem to highly depend The glutathionylated compound can subsequently react with another on the type of organism and/or subcellular compartment. GSH molecule yielding a second (regenerated) thiol product and GSSG (Fig. 3C). Again, the thiol–disulfide exchange occurs either 2.2.5. GSH as a nucleophile for other electrophiles non-enzymatically or enzymatically (with the help of the same or Disulfides and peroxides are not the only compounds reacting another enzyme). GSSG is finally reduced by NADPH with the help with GSH. Other electrophiles are converted in a GSH-dependent of GR or the TrxR/Trx couple (Fig. 3B) [14,53–55]. Please note that manner by the glyoxalase pathway and by GST. In the glyoxalase the apparent second order rate constants for the direct reduction of pathway, GSH spontaneously reacts with electrophilic 2-OA to form − − GSSG by Trx were found to be lower than 103 M 1 s 1 [37].Thus,an a diastereomeric hemithioacetal (Fig. 3E). The latter substance is iso- efficient turnover at estimated nanomolar Trx and micromolar or even merized to a single thioester by glyoxalase 1 (Glo1) and subsequently nanomolar GSSG concentrations remains controversial (right side in hydrolyzed by glyoxalase 2 (Glo2) as outlined in Section 7.Thepathway Fig. 3B). An alternative explanation for the Trx/TrxR-dependent reduc- yields regenerated GSH and a non-toxic 2-hydroxycarboxylic acid such tion of GSSG in vivo [37–43] might be the GSSG-dependent formation as D- from MG. Thus, in contrast to most GST-dependent path- of glutathionylated/oxidized proteins (Fig. 3C)thataremoreefficient ways, GSH acts as a coenzyme and is not consumed in the over- substrates of the system. In such a scenario the reduction of GSSG all reaction of the glyoxalase pathway (Fig. 3E). Moreover, since by the thioredoxin system would be indirect. The latter hypothesis the conversion of MG and other 2-OA is an intramolecular redox is supported by a few in vitro studies, revealing for example that reaction, GSH does not act as a reducing agent in the overall glutathionylated human Grx2 and GSSG-treated Grx4 from E. coli can reaction [21,23,31,73,74]. be substrates of TrxR [58,59]. In addition to the reduction of peroxides and disulfides, the predom- inant function of the extremely heterogeneous families of GST-isoforms 2.2.4. GSH as a reducing agent for peroxides and non-related MAPEG (membrane-associated proteins with diver- In analogy to the reduction of disulfides, GSH also reduces a vari- gent functions in eicosanoid and glutathione metabolism) is the cata- ety of hydroperoxides (Fig. 3D). These irreversible reactions are cata- lytic conjugation of the sulfur of GSH to (carbon atoms of) a lyzed by a subgroup of glutathione peroxidases (GPx), yielding GSSG, large variety of electrophilic substances (Fig. 3F) [3,75–77].Thesesub- water and/or an [60–62] as outlined in Section 5. Alterna- strates do not necessarily contain disulfide or peroxide bonds, and the tively, selected peroxiredoxins (Prx)—which are usually highly conjugation reactions often result in a reduced toxicity and an increased 3222 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 solubility of the electrophiles. The glutathione-labeled substances can not restricted to amino acids as adjacent groups: In mycothiol— be subsequently metabolized and/or excreted. Alternatively, some which is the replacement for GSH in many actinobacteria (including GST-isoforms also use GSH for isomerizations [3,76]. All these reactions the important pathogen Mycobacterium tuberculosis)—the central are summarized in Section 8. cysteine residue forms amide bonds with acetate and a neutral amino sugar [85]. These modifications were also reported to slow 2.3. Further evolutionary and chemical aspects of glutathione catalysis down copper-catalyzed autoxidation [84]. In bacillithiol—a similar cysteine-containing compound from bacilli (including the model or- 2.3.1. The benefits of a single thiol compound ganism Bacillus subtilis)—only the carboxy group of cysteine is modi- The advantage of utilizing a single adapter molecule as a universal fied by a negatively charged amino sugar [93,94]. Thus, it is not really nucleophile instead of different compounds for each electrophile understood why GSH instead of other soluble cysteine derivates becomes obvious considering the numerous functions summarized became the central reducing agent in most organisms. In fact, even in Fig. 3: Instead of optimizing a large set of unrelated proteins for non-cysteine thiol/disulfide couples are able to exert similar func- (i) synthesizing different nucleophiles and for (ii) catalyzing the turn- tions: Coenzyme M (2-mercaptoethanesulfonate) and coenzyme B over of each nucleophile/electrophile couple, only one pathway for (7-mercaptoheptanoylthreonine phosphate) both facilitate the re- glutathione synthesis was required and rather moderate structural duction of methyl groups in CH4-producing archaea [85,95]. The changes of ancient protein scaffolds such as the thioredoxin fold thiolhistidines ergothioneine and ovothiols also possess were sufficient to generate novel enzymatic activities in the course properties as scavengers, but differ significantly from cysteine thiols of evolution (as outlined in Section 4.2 and as exemplified in all sub- due to the instability of their disulfides [44,95]. Although thiolhistidines sequent sections). Why has a thiol compound evolved as the univer- are found in many organisms at high concentrations, their functions are sal adapter molecule? Taking into account Pearson's HSAB theory, poorly understood and specific enzymes seem to be absent [85,88,95]. are quite hard bases and therefore far less versatile than In summary, the utilization of cysteine-based thiols as universal thiols [78]. In comparison with thiols, selenols are restricted due to nucleophiles for the modification or removal of diverse physiological the limited bioavailability of the trace element [79]. More- electrophiles is plausible. The chemical properties of GSH due to over, although selenols have much lower pKa values and are usually its composition/structure provide a sufficient condition for catalysis more reactive than thiols [79,80], the utilization of and the complex metabolic network depicted in Fig. 3, even though for biocatalysis (e.g. in TrxR or GPx) remains enigmatic [80,81].In alternative thiols exert analogous functions in archaea and many conclusion, owing to the bioavailability, size and electron config- bacteria. uration of sulfur, thiols instead of alcohols and selenols are pre- destined to catalyze such a variety of reactions under physiological 2.3.3. Mechanistic principles of glutathione catalysis conditions [79]. Before I discuss selected enzyme/substrate couples in detail, I want to end Section 2 with an overview of chemical principles of glu- 2.3.2. Comparison with alternative thiols as catalysts tathione catalysis that seem to be often ignored. Most of the reactions Why is the major reducing agent a cysteine-containing tripeptide? in Fig. 3 include one or multiple (predicted) nucleophilic substitu- First of all, the availability of the proteinogenic amino acids cysteine, tions, regardless whether a disulfide, a hydroperoxide or a sulfenic glycine and glutamate during very early evolution is a prerequisite acid is the electrophile (reactions with electrophilic carbon atoms for the success of GSH [82]. Second, in contrast to coenzyme A are outlined in Sections 7.2 and 8.3). Mechanistically, bimolecular

(containing cysteamine due to a decarboxylation), all components/ nucleophilic substitutions (SN2 reactions) are likely for several of amino acids of GSH can be directly salvaged [83], providing a poten- these pathways (Fig. 5), even though atomistic data on enzyme catal- tial advantage for the ancestors of modern organisms under limiting ysis are so far rather limited to a few examples such as Trx [96,97]. growth conditions. Third, GSH provides significant advantages over Please note that glutathione—as well as cysteine residues at the active unmodified cysteine: (i) Protein biosynthesis and other cysteine- site of a glutathione-dependent enzyme—can either play the role of utilizing anabolic processes can be separated from detoxification the nucleophile (GS−, Cys-S−) or the electrophile (GSSG, GSSR, and redox processes in the same cellular compartment. (ii) As I will Cys-SSR, Cys-SOH) in SN2 reactions, depending on the elementary outline below, the charged functional groups of the glycine- and the reaction (Fig. 5).

γ-glutamyl moiety are perfect electrostatic anchors for substrate rec- Before or during the first step of the SN2 reaction, the attacking ognition, resulting in substrate specificity. (iii) The modification of thiol (or selenol) group becomes deprotonated. As a consequence, the amino group of cysteine was suggested to prevent the intramolec- a negatively charged transition state is formed (Fig. 5). Thus, two ular transfer of acyl groups (yielding amides from thioesters) [84]. important aspects of glutathione catalysis are the generation of the However, whether the latter reaction could occur at a significant nucleophile by deprotonation and/or the stabilization of the negative rate in vivo has, to my knowledge, not been systematically studied. charge of the transition state (therefore lowering its Gibbs energy). (iv) Protection of the amino and of the carboxy group of cysteine While GSH deprotonation is more often considered in glutathione can furthermore decrease the metal-, salt- and pH-dependent autoxi- catalysis, sterical constraints are predominantly neglected [55].A dation rate [48,84–87]. Obviously, this protection is highly important SN2 reaction usually requires a trigonal bipyramidal transition state since thiols are not only but also sources for ROS with the entering and leaving groups in apical positions and substitu- (Fig. 1A) [48]. A tripeptide with cysteine in the middle is the smallest ents at the central atom in an angle of approx. 90°. As the cleavage of protected peptide and therefore a simple solution to this problem. a disulfide bond is thought to occur without essential participation of What could be the advantage of GSH in comparison to other 3d orbitals, a linear orientation also seems to be valid for sulfur atoms thiols? Some organisms employ glutathione precursors or derivates (Fig. 5) [98–100]. Thus, a central aspect of glutathione catalysis is to instead of GSH, e.g. γ-glutamyl-cysteine in halophilic archaea [86] align the electrophile and the nucleophile appropriately. Several en- and (T(SH)2) in kinetoplastid parasites [44,88–90]. zymes seem to master this challenge with the help of positively Even E. coli uses GSH and glutathionylspermidine which accumulates charged side chains that direct the glutathione substrate. Moreover, under anaerobic conditions [91] and oxidative challenge [92]. Please before the nucleophilic attack, the substituent of the central atom note that entropically favored monomeric T(SH)2 is a positively of the electrophile (e.g. the side chain of a cysteine residue RC-S) charged dithiol with a pKa value of approx. 7.4 and therefore differs could be stabilized in a position resembling the transition state (for ex- significantly from the negatively charged monothiol compound GSH ample, in a rather strained protein disulfide bond). As a result, the [44] (Fig. 3A). In addition, protective modifications of cysteine are reactivity of the electrophile could increase, and the activation energy M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3223

Fig. 5. Principles of thiol-dependent SN2 reactions. (A) Thiol–disulfide exchange reaction. (B) Thiol-dependent cleavage of electrophilic hydroperoxides. (C) Thiol-dependent cleav- age of electrophilic sulfenic acids. In all reactions an initial deprotonation generates the nucleophilic thiolate. Upon attack, a linear, negatively charged transition state (highlighted in brackets) is formed between the nucleophile and the electrophile. The properties of the leaving group can be altered owing to protonation. RN, residue of the nucleophile;

RC, residue of the central atom; RL, residue of the leaving group. Changes of the hybridization of the central atom and of the position of RC are indicated by red arrow heads. See Section 2.3.3 for further details.

ΔG* of the transition state could be lowered. In the last part of the SN2 addition–elimination reaction with a rather stable intermediate instead − − − reaction, the rather poor leaving group (RL-S >RL-O >OH )canbe of a SN2 reaction [102]. stabilized by protonation (Fig. 5). Whether this step occurs simulta- In summary, the SN2 reaction presented in Fig. 5 is the most likely neously or right after the bond is cleaved might depend on the enzyme mechanism for glutathione-dependent thiol–disulfide exchange reac- and the leaving group. tions. Principles including the deprotonation/activation of GSH as a

Are there alternatives to the mechanism outlined in Fig. 5?(i)ASN1 nucleophile, the correct substrate alignment via (positively charged) reaction with an electrophilic, positively charged sulfur atom as an binding sites, the stabilization of the (glutathionylated) transition intermediate is improbable [98,101], particularly under physiological state, and the stabilization/activation of a leaving group are of course conditions. (ii) A direct nucleophilic attack of one of the two free also applicable to other glutathione-dependent enzymes that do not electron pairs of the thiol group without deprotonation also seems un- catalyze thiol–disulfide exchange reactions (i.e. Fig. 5B,C). likely. First, thiols are rather poor nucleophiles. Second, the resulting uncharged transition state is acidic, and the simultaneous protonation 3. Glutathione reductase of the leaving group is therefore problematic. (iii) Under acidic condi- tions, the leaving group could be protonated before the nucleophilic 3.1. Pioneers of GR catalysis attack of the thiolate. Accordingly, a better leaving group is generated, and the orbital energy of the LUMO that accepts the incoming electrons Based on studies by Hopkins and several other groups between from the nucleophile is lowered [98,101]. However, even in a protein the 1930–50s, Racker purified GR from yeast in 1955 and confirmed environment, it is difficult to envision disulfide protonation by a strong NADPH as the electron donor [103]. In 1963, Mapson and Isherwood acid on the one hand, and proximal thiol deprotonation on the other. confirmed that GR from pea seedlings requires FAD and a thiol- (iv) In a variation of the reaction in Fig. 5, the transition state might group for activity. Their steady-state kinetics furthermore revealed be significantly stabilized. Thus, the mechanism would be an parallel lines in Lineweaver–Burk plots [104]. Two years later, Massey 3224 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 and Williams suggested a ping-pong mechanism for yeast GR [105]. (Fig. 6C), and therefore the enzyme is only functional as a homodimer In 1977, the first low resolution crystal structure of a GR-isoform [36]. The structure, both substrate-binding sites and even the overall was solved for the human enzyme from erythrocytes, followed by a amino acid sequence of different GR-isoforms are extremely con- key article on the structure at 3 Å resolution in 1978 by Schulz et al. served in the course of evolution. Biggest differences are found [36]. The exact amino acid sequence was obtained in the ensuing at the subunit interface. For example, the subunits of crystallized years, and, in 1981, Thieme et al. assigned the sequence to an X-ray human GR are linked by a cysteine disulfide bond [106,108,114] in data set with 2 Å resolution [106]. Owing to numerous additional contrast to the GR-isoforms from yeast [111], Plasmodium falciparum protein crystallographic studies, e.g. by Pai and Karplus, spectropho- [44,110] and E. coli [109]. Other poorly conserved cysteine residues— tometric analyses, e.g. by Williams, Arscott, Krauth-Siegel, Perham e.g. residue Cys3 at the flexible N-terminus of human GR or residue and Scrutton, as well as genetic screens, e.g. by Grant, GR is nowadays Cys239 of yeast GR—are often solvent exposed and might play a regu- one of the best understood enzymes and a reference protein for redox latory role [34,106,111]. Another potential for regulatory catalysis. molecules is a cavity at the dimer interface [44,110,114]. Functionally, GR is an NADPH:GSSG (previously 3.2. Structure and function of GR EC 1.6.4.2, now 1.8.1.7). The enzyme has actually three substrates (NADPH, H+ and GSSG) and two products (GSH and GSH), although GR (also termed GLR) is a flavoenzyme of the pyridine nucleotide- the is usually neglected as a substrate owing to the official disulfide oxidoreductase family that also includes the related mechanistic nomenclature. The enzyme adopts a central role in gluta- enzymes trypanothione reductase, dihydrolipoamide dehydrogenase, thione metabolism by linking the cellular NADPH-pool with the thiol/ mercuric ion reductase and the so-called high Mr type TrxR-isoforms disulfide-pool (Fig. 3B). Thus, GR helps to maintain a reducing intra- [44,107,108]. GR-isoforms from pro- and eukaryotes form stable cellular milieu owing to high GSH and low GSSG levels (Fig. 4). Note- homodimers of ~110 kDa with a large subunit interface of more worthy, different GR-isoforms are found not only in the cytosol than 3000 Å2 (Fig. 6A) [36,108–112]. Each subunit contains an FAD- but also in the mitochondrial matrix and in chloroplasts [115–120]. binding site that is formed by a Rossmann-fold. The isoalloxazine These proteins are often encoded by alternative in-frame start codons ring of FAD separates the distinguished substrate-binding sites for of the same gene, resulting in the presence or absence of an NADPH and GSSG (Fig. 6B). The NADPH-binding site of each subunit N-terminal targeting sequence [115,119–121]. The balance between is also formed by a typical Rossmann-fold and presumably originated the isoforms—at least in yeast—seems to be regulated by the translation from a of the ancestor encoding most of the FAD- initiation efficiency and therefore depends on the mRNA sequence binding site [113]. Each GSSG-binding site is formed by both subunits flanking the start codon [115].

Fig. 6. Structure of GR. (A) Front view of homodimeric GR with one FAD molecule bound per subunit. Based on the molecular 2-fold axis, the opposite sides of the flavins (si and re side) are shown. Both subunits are not fully symmetrical owing to slight structural deviations. (B) Top view along the 2-fold axis revealing a cleft at the re side of FAD where NADPH binds. (C) Zoom in at one of GR. An NADP+ molecule is bound at the re side in the back. The GSSG-binding site is composed of both subunits and is shown in the front.

Conserved residues that are important for substrate binding and catalysis are highlighted. The essential interchange and charge-transfer cysteine residues (Cysint and CysCTC, respec- tively) are located at the N-terminus of a long α-helix (presumably stabilizing thiolate anions due to its dipole). (D) Side view of one active site demonstrating the spatial separation of both substrate-binding sites by the flavin. Selected atoms of NADP+ and FAD are highlighted. See Section 3.2 for details. The images were generated using Swiss-Pdb viewer and the structure of GR from E. coli (PDB ID: 1GET [109]). M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3225

3.3. The enzymatic mechanism of GR upon GSSG reduction. This process might be assisted by TyrGSSG [108,122,127,129–132]. Once the first GSH molecule (GSHII) has left The ping-pong mechanism of GR is coupled to the spatial separa- the active site, the intermolecular disulfide bond is attacked at tion of the NADPH- and the GSSG-binding site and comprises a reduc- the sulfur atom of Cysint by the thiolate of CysCTC yielding GRox. The tive and an oxidative half-reaction (Fig. 7). First, the enzyme becomes thiolate leaving group of the second GSH molecule (GSHI) could reduced by NADPH. Then, the electrons are transferred to GSSG, again be protonated by His′ [35,108,123,127,132]. Considering the regenerating the oxidized enzyme. Catalysis is facilitated by several kinetics of the numerous steps, one of the protonations (presumably conserved key residues that are highlighted in Fig. 6C,D. yielding GSHII) was suggested to be rate-limiting during the oxidative half-reaction—which was furthermore reported to be slower than the 3.3.1. The reductive half-reaction of GR reductive half-reaction [122,130]. Accordingly, mutation of His′ was

Oxidized GR (GRox) contains two essential cysteine residues that form shown to have drastic effects on catalysis [123,126,131]. adisulfide bridge at the si side of the isoalloxazine ring. The disulfide bond is close to a residue which is furthermore hydrogen- 3.3.3. Properties of GR reaction intermediates in vitro and in vivo bonded to a glutamate residue (His′ and Glu′, Fig. 6C,D). Please note Reported macroscopic E°′ values for the reduction of the fully that His′ and Glu′ belong to the second subunit of the homodimer. A ty- oxidized enzyme GRox to the two-electron reduced form GRH2 are rosine residue (TyrNADPH)atthere side shields the FAD and acts as a gate- between −227 and −243 mV for the isoforms from human, yeast + keeper at the NADPH-binding site. Upon rapid NADPH binding, TyrNADPH and E. coli [127]. Using (i) an estimated NADPH:NADP ratio of 4.2 rotates away from the isoalloxazine ring and clamps the nicotinamide for unbound pyridine dinucleotides in erythrocytes [133], (ii) an E°′ moiety of the substrate [35,108,122,123]. A transfer from value of −317 mV, and (iii) the Nernst equation, the calculated E′ NADPH reduces the flavin to FADH- (Fig. 7) which subsequently shuttles value for NADPH is −335 mV. Thus, under physiological conditions an electron pair to the proximal cysteine residue (CysCTC). The thiolate (see also E′ values in Fig. 4), the concentration of GRox in the cytosol group of CysCTC forms a stable charge-transfer complex with the isoallox- or in the mitochondrial matrix is presumably low and the enzyme azine ring, whereas the reduced distal cysteine residue (Cysint)couldbe gets permanently reduced owing to the rapid reaction with NADPH protonated by His′ [35,108,112,124–127]. At the end of the reductive [127,132,134]. half-reaction, NADP+ dissociates from the two-electron reduced enzyme Is the enzyme also constantly saturated with substrates? Apparent species (GRH2) and is replaced by another molecule of NADPH [124,128]. and true Km values for NADPH in vitro were found to be usually between 3 and 20 μM [104,122,135,136]. These values were predom- 3.3.2. The oxidative half-reaction of GR inantly determined for GR from various species at a single fixed

Upon GSSG binding to GRH2, a tyrosine residue (TyrGSSG)is millimolar concentration of GSSG. Furthermore, different pH values repositioned in such a way that its hydroxyl group contacts the disul- were used, although this might be rather unproblematic since the fide bond of the substrate (Fig. 6C,D) [108]. In addition, GSSG is bound pH optimum of most GR-isoforms is rather broad (with a maximum by other conserved residues from both subunits, including four posi- around pH 7, except for some proteins from photosynthetic organ- tively and two negatively charged side chains that compensate isms) [105,136,137]. Apparent and true Km values for GSSG were the charges of the substrate (Fig. 3A) [35]. After substrate binding, often determined with 100 μM NADPH and usually ranged between

CysI of GSSG is attacked by the interchange residue Cysint of GRH2, 50 and 80 μM, though some isoforms with lower and higher values resulting in the formation of an intermolecular disulfide bond were also reported [34,104,105,122,134,136,138]. Do the Km values (Fig. 7). The nucleophilic attack could be accelerated owing to the for NADPH and GSSG correspond to the physiological substrate con- deprotonation of the Cysint thiol group by His′. The interaction of centrations? To my knowledge, there is surprisingly very limited in- the latter residue with Glu′ could facilitate the deprotonation in ana- formation on the concentration of NADPH in vivo. In erythrocytes, logy to serine proteases [35,108,123]. His′ was furthermore suggested the concentrations of protein-bound and free NADPH were reported to protonate the thiolate leaving group of CysII which is liberated to be 32 and 2 μM, respectively [133]. Considering the latter value

+ Fig. 7. Model of GR catalysis. Both subunits, FAD, the substrates NADPH, H and GSSG, as well as residues Cysint,CysCTC and His′ are highlighted. The NADPH-binding site is at the top, and the GSSG-binding site is at the bottom. Please note that the glutathione moieties GSI and GSII are not identical. The charge-transfer complex is highlighted in red. See Section 3.3 for details. 3226 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 and a micromolar or even nanomolar GSSG concentration in the cell, out strains are viable (as long as there is a functional TrxR/Trx app it is quite likely that the Km values for NADPH and GSSG are signifi- couple), but were suggested to be more sensitive to oxidants and to cantly lower under physiological conditions (because decreasing the have higher GSSG levels in the cytosol and in the mitochondrial app concentration of one substrate also decreases the Km value for the matrix [40,41,115,148]. Moreover, despite similar GR activities and second substrate of an enzyme with a ping-pong mechanism). For ex- concentrations in both subcellular compartments [115], removal ample, reevaluation of fluorimetric data on GR from peas at 0.3 μM of the mitochondrial but not of the cytosolic GR-isoform rendered app NADPH reveals a Km value for GSSG of approx. 1 μM—a value far yeast cells more sensitive to hyperoxia [149]. In contrast to yeast, below the true Km of 17 μM [104].Insummary,aslongasweneither E. coli GR knock-out strains lack a phenotype and do not have in- know the exact concentrations of the substrates nor the corresponding creased GSSG levels as long as there is an alternative electron donor app Km values under physiological conditions, it is difficult to estimate or system [150]. The GR from rodent malaria parasites was shown to to predict the degree of saturation of GR-isoforms in vivo. be essential for oocyst development in the mosquito midgut but not for the blood stage parasites in the vertebrate host [39,42]. In con- 3.3.4. Outlook on GR catalysis and mechanistic questions trast, blood stage cultures of the human malaria parasite P. falciparum There are still open questions concerning GR catalysis. (i) The fates were suggested to require GR for survival [151]. The two GR-isoforms and sources of several remain to be determined: What hap- from the plant are encoded by alternative genes. pens for example to the proton Hs that was transferred as a hydride A deletion of the cytosolic isoform did not result in a significant phe- ion from NADPH [108,131]? Is TyrGSSG really involved in acid–base notype, even though the in vivo redox potential for the glutathione catalysis [108,122,127,131]? Which of the candidates Cysint, CysI system increased by 45 mV owing to higher GSSG levels [43].In and CysII receives a proton from His′, and/or does His′ remain proton- contrast, a deletion of the dual targeted mitochondrial/chloroplast ated to stabilize the thiolate of CysCTC [108,112,125–127]? (ii) Does GR-isoform was lethal during embryo development as revealed by a the reductive half-reaction occur simultaneously or sequentially? genetic screen [152]. The human gene encoding the cytosolic and In contrast to previous kinetic studies on GR from E. coli and the mitochondrial GR-isoform ( p21.1 on the short arm of chro- P. falciparum [126,134], recent high resolution crystallographic stud- mosome 8) consists of 13 exons [120]. In addition to the full length ies on human GR suggested that the Hs hydride transfer of atom C4 transcript, two splice variants lacking either exon 8 or 9 seem to be from NADPH to atom N5 from the isoalloxazine on the one hand, present in various tissues. The physiological role of these variants is and the electron transfer from atom C4 of the isoalloxazine to CysCTC rather cryptic, in particular, because the predicted translation prod- on the other, are not separate steps, but occur in a simultaneous ucts are expected to be inactive [153]. Up-regulation of mitochondrial

1,2-addition reaction with respect to the flavin [139]. (iii) Is Cysint of GR was shown to increase the resistance of lung cells to exogenous some GR-isoforms predominantly glutathionylated in vivo as suggested hydroperoxides and hyperoxia in cell culture [154] but not in mice by Arscott and colleagues [132,134]? To my knowledge, an accumula- [155]. Noteworthy, patients with low or even absent GR activity in tion of oxidized Cysint has so far not been detected (using for example blood cells (that could not be compensated by FAD supplementation) quantitative redox proteomics in E. coli, Caenorhabditis elegans and were already reported in the 1960s and 1970s [156–158]. More yeast [140–142]). (iv) Is an NADH-dependent GR activity of any physi- recent genetic analyses revealed three rare underlying homo- and ological relevance? Some GR-isoforms were shown to utilize NADH heterozygous mutations resulting in either truncated/non-functional as an alternative electron donor in vitro. For example, the Vmax of GR or destabilized/short-lived GR [159]. from spinach with NADH was found to be 18% of the activity with In summary, functional GR is not a prerequisite for the survival of NADPH [135]. In addition, at a rather acidic pH, the activities of mam- several aerobic organisms including humans. Even though mitochon- malian GR with NADH and NADPH were reported to be similar [137]. drial and chloroplast GR-isoforms seem to be more important with (v) Kinetic studies indicate that the outlined mechanism might not be respect to oxidative challenges than cytosolic GR, most prokaryotes that simple. For example, mutation of TyrNADPH in E. coli GR switched and eukaryotes have alternative back-up systems that provide elec- the steady-state kinetics from ping-pong to sequential patterns trons at an adequate rate to maintain sufficient amounts of GSH and [123,143] in accordance with a hybrid ping-pong bi-bi/ordered bi-bi a physiologically acceptable GSH/GSSG ratio (Fig. 4). mechanism [137,144,145]. Moreover, do both reaction centers of GR function independently, or is there a synchronization of the catalytic 3.4.2. Medical relevance of GR catalysis cycle including subunit ? Studies by the Perman lab Owing to the central role that GR exerts in glutathione metabolism in the 1990s support both hypotheses. On the one hand, data on (Fig. 3B), the absent or mild phenotypes of GR knock-outs from differ- heterodimeric GR mutants from E. coli with one functional and one mu- ent organisms are surprising at first sight. Indeed, three patients tated reaction center favor an independent catalysis [143]. On the other with homozygous GR deficiency in blood cells were reported to be hand, steady-state kinetics of an E. coli GR mutant with a single amino in good health at ages 48, 54 and 58. To date, the only documented acid replacement at the dimer interface revealed subunit cooperativity clinical symptoms related to a GR deficiency are restricted to a higher at 0.1 mM NADPH that was lost with 0.4 mM NADPH [146,147].The susceptibility of erythrocytes to oxidative challenge (including hemo- crucial question is now, whether wild type GR also shows cooperativity lytic crisis after eating fava beans) and cataract development during at physiological substrate concentrations (Section 3.3.3). Furthermore, early adulthood [157,159]. Nevertheless, the numerous studies on is a potential cooperativity of human GR coupled to the stability of the catalytic mechanism of GR provide excellent lessons on rational the cysteine disulfide bond at the dimer interface [106,108,114]? drug development, and it is nowadays accepted that knowing as In summary, GR works via a ping-pong mechanism. The enzyme much as possible about a target enzyme is highly advantageous. For requires FAD, two essential cysteines, an activated histidine for acid– example, despite high overall sequence similarities, the GR-isoforms base catalysis as well as several other conserved residues for substrate from human and P. falciparum were shown to differ significantly binding. Although GR is one of the best understood enzymes, several with respect to their dimer interfaces as well as their kinetic and fundamental mechanistic aspects have not been unraveled yet. redox properties [44,110,134,138,160]. Accordingly, alternative strate- gies to exploit GR as a drug target have been developed: (i) Traditional 3.4. Physiological and medical relevance of GR catalysis approaches included the inhibition of the enzyme at the GSSG-binding site or at the dimer interface and its cavity [34,44,160,161].Withrespect 3.4.1. Physiological relevance of GR catalysis to irreversible inhibition of GR at the GSSG-binding site, highly reactive The physiological relevance of GR catalysis can be estimated from electrophiles such as gold-compounds and fluoronaphthoquinones a variety of GR mutants and knock-out organisms. Yeast GR knock- turned out to efficiently inactivate GR-isoforms in vitro, but to be of M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3227 limited suitability for in vivo applications [34,160,161].(ii)Amore donor for ribonucleotide reductase (RR) in crude extracts from an recent, alternative approach to kill malaria parasites aims to exploit E. coli strain lacking Trx [185] (which is the classic hydrogen donor functional instead of inactive GR in order to regenerate drugs that sub- for RR [17]). He therefore introduced the term “glutaredoxin”. Three sequently act as harmful redox cyclers. Based on in vitro experiments, years later, Holmgren reported the purification of the enzyme and such drugs—including naphthoquinones and methylene blue—were the reconstitution of the RR-assay in vitro [186,187]. During the therefore classified as “turncoat inhibitors” or “subversive substrates” following years, E. coli Grx1 (EcGrx1) became an excellent model [161,162]. Indeed, several synthetic naphthoquinones had a low toxici- protein: The amino acid sequence was determined [188], the corre- ty towards mammalian cells, a high activity with low nanomolar IC50 sponding gene was cloned, and two isoforms were successfully puri- values against P. falciparum blood stage cultures, and a moderate acti- fied [189,190]. Furthermore, between 1991 and 1994, Wüthrich vity in parasitized mice [161]. Phase II trials in Burkina Faso furthermore and colleagues determined the solution structure of EcGrx1 in the revealed that methylene blue can be useful in combination therapies oxidized, the reduced and the glutathionylated state by NMR- with fast acting antimalarials, even though the compound is not suited spectroscopy [191,192]. for monotherapy [163,164]. Recent studies on rodent malaria parasites Today, it is accepted that GSH-dependent transhydrogenases, suggested that the presence or absence of GR does not alter the activity thioltransferases and Grx from yeast, mammals and E. coli are of methylene blue [39,42] in contrast to the GR-dependent drug- isoforms of the same . Since their discovery, numerous activation hypothesis. Whether the latter results can be transferred to studies on the structural diversity, the enzymatic mechanism and the the human system awaits clarification. physiological functions of these ubiquitous proteins have been pub- In summary, attempts to exploit GR as a traditional drug target have lished [14,53–55,173,193,194]. Nonetheless, as outlined in the follow- failed to date, but the enzyme might be suited for the activation of sub- ing sections, the more we know about Grx, the more questions seem versive substrates. Related flavoenzymes of organisms with alternative, to arise. non-redundant redox systems—such as thioredoxin–glutathione reduc- tase in parasitic plathelminths [165] and trypanothione reductase in kinetoplastid parasites [44]—could be better suited for drug develop- 4.2. Structure of Grx and related glutathione-dependent proteins ment. Future studies on such enzymes could benefit from the experi- ences with GR. 4.2.1. Comparison of the catalytic core domains All Grx-isoforms possess a thioredoxin fold and are therefore mem- 4. Glutaredoxins bers of the thioredoxin superfamily. This fold of approx. 11–13 kDa is highly conserved in the course of evolution and is composed of four or 4.1. Pioneers of Grx catalysis five central β-strands surrounded by α-helices (Fig. 8A,B) [195]. Similar architectures are found in other (glutathione-dependent) enzymes A glutathione-dependent thiol:disulfide oxidoreductase activity such as GST (Fig. 8C,D), GPx (Fig. 8E,F) and Prx (Fig. 8G,H), supporting (Fig. 3C) was first described in crude enzyme preparations from the theory of a common ancestor for all these proteins [195].Please beef liver by Racker in 1955. In this study GSH/homocystine and note that the positions of the glutathione binding residues and of the homocysteine/GSSG were successfully used as redox couples, con- active site residues are often either interchanged, similar or even iden- firming the reversibility of the catalyzed thiol–disulfide exchange tical. Thus, selected mutations resulted in novel functions (see also reaction (Fig. 3C). The catalyst of the reaction was classified as a Section 2.3.1). “transhydrogenase” [166]. In the following years, several groups ana- Biggest structural differences between Grx- and Trx-isoforms are lyzed similar enzymatic activities in partially purified liver extracts found at the active site and at the N-terminus because of an additional from mammals including human. Most of these studies focused on β-strand in Trx (Fig. 8A). Furthermore, the N-terminus of many the GSH-dependent reduction of insulin disulfide bonds [167–169]. eukaryotic Grx-isoforms is modified by a targeting sequence, a mem- In retrospective, as already pointed out by Freedman in 1979, it brane anchor or additional domains [53,196–199]. Grx can be distin- is quite likely that canonical Grx were analyzed in these liver prepa- guished from Trx owing to their specificity for glutathione. Accordingly, rations, although other enzymes such as Trx, GST or PDI could also Grx possess moderately conserved polar as well as charged amino acid have contributed to the detected activities [170,171]. residues that interact with the carboxylate group(s) of glutathione as In 1968, Nagai and Black published the first characterization of an highlighted in Figs. 8Band9 [55,184,192,200–202]. However, despite isolated GSH:homocystine oxidoreductase. The 15 kDa protein was numerous alignment-based subgroup classifications [196,199],the purified from baker's yeast (). Moreover, boundaries between Grx- and Trx-isoforms nowadays become more the authors established a coupled spectrophotometric assay with GR and more blurred [203], and it is difficult to clearly separate both groups using different disulfide substrates including L- and D-cystine, several because of structural hybrid forms and overlapping or absent activities cystine-derivates as well as bis(2-hydroxyethyl)disulfide (HEDS) [55]. The same holds true for various Trx- or GSH/Grx-dependent GPx- [172]. The latter substance became an important model substrate and Prx-isoforms [60,61,204] (Sections 5 and 6). for Grx research [53–55,173]. In 1974, Mannervik's group introduced Was the ancestor of certain (sub)families of the thioredoxin su- the name “thioltransferase” instead of transhydrogenase, based on perfamily a glutathione-dependent protein? Considering the putative their mechanistic studies on partially purified rat liver extracts onset of glutathione metabolism (Fig. 1), it seems far more likely that [174]. Four years later, the group succeeded in purifying functional glutathione-independent members of the thioredoxin superfamily rat liver thioltransferase [175] which was further analyzed in numer- are more ancient. Nevertheless, recent in silico analyses suggest ous studies [176]. Alternative purification procedures for mammalian that Grx evolved rather early from one initial gene in the last common isozymes from calf thymus [177,178] and pig liver [179,180] were ancestor of all organisms [203]. The relatively low numbers of established in the following years by Luthman and Holmgren as well glutathione-dependent GPx- and Prx-isoforms (Sections 5 and 6) fur- as Gan and Wells. The sequences of these model isozymes were deter- thermore point to independent acquisitions of glutathione activities mined [181], the according genes were cloned [182],andthecrystal for different (sub)families in the course of evolution. In addition, it and NMR structures of oxidized and mutant glutathionylated mam- is also possible that various isoforms—including some Grx-like proteins malian thioltransferase were solved in 1995 and 1998, respectively or GST-isoforms—have secondarily lost their specificity for glutathione [183,184]. (see also Section 8.3.2). Obviously, the research field could become In parallel to the studies on mammalian thioltransferases, an eldorado for (bioinformatic) studies on the molecular evolution of Holmgren discovered in 1976 a heat-stable GSH-dependent hydrogen structure–function relationships. 3228 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266

Fig. 8. Structural comparison between members of the thioredoxin superfamily. Protein architectures are shown on the left side. Top and side views of the glutathione-binding site of representative proteins are shown on the right side. More or less conserved residues that were demonstrated or suggested to bind the glycine moiety (1) or the γ-glutamyl moi- ety (2) of glutathione are labeled. The ionic and hydrogen bonds with the substrate are predominantly formed by Lys/Arg and Asn/Gln residues. The structures were visualized using Swiss-Pdb viewer. (A) Overall architecture of Grx and Trx. The canonical thioredoxin fold is highlighted in black. This fold is identical to the architecture of the most simple

Grx-isoforms such as Grx1 and Grx3 from E. coli. Helices α1 and α5 (blue) are found in many other Grx-isoforms. Trx have an additional N-terminal β-strand (green) but lack helix

α5. Additional targeting sequences or domains at the N-terminus of various Grx-isoforms are omitted for clarity. The N-terminal cysteine residue in the CxxC/S-motif at the active site is highlighted by an asterisk. (B) NMR structure of glutathionylated human Grx1 with six glutathione conformations (PDB ID: 1B4Q [184]). (C) Common architecture of a single GST subunit without the C-terminal helical domain. The tyrosine residue at the active site is highlighted by an asterisk. (D) Crystal structure of the N-terminal domain of a P. falciparum GST subunit in complex with S-hexylglutathione (PDB ID: 2AAW [71]). (E) Common architecture of a single GPx subunit. Please note the insertion of additional struc- tural elements (blue). The selenocysteine (or cysteine) residue at the active site is highlighted by an asterisk. (F) Crystal structure of a mammalian GPx1 subunit with the selenocysteine residue at the active site in the ‘over’-oxidized seleninate state (PDB ID: 1GP1 [304]). The N-terminal part is omitted for clarity. (G) Common architecture of a single Prx subunit. Please note the insertion of additional structural elements (blue). A C-terminal arm/domain interacting with a second subunit is present in many Prx classes. The cys- teine residue at the active site is highlighted by an asterisk. (H) Crystal structure of a poplar D-Prx subunit with the peroxidatic residue (Cysp) at the active site (PDB ID: 1TP9 [558]). The precise glutathione-binding site (if any) is unknown. The N-terminal part is omitted for clarity, and a C-terminal domain is absent.

4.2.2. Comparison of Grx structures family. The numerous isoforms are traditionally subdivided into Several structures of Grx-isoforms in a variety of redox states are monothiol and dithiol Grx, depending on the number of cysteine res- shown in Fig. 9. Even though all Grx-isoforms share a solvent exposed idues in the CxxC/S-motif at the active site. For example, canonical active site cysteine residue (Cysa) at the N-terminus of helix α2 Grx are dithiol isoforms of the CPYC-type with the second cysteine (Fig. 8A,B, Fig. 9), they are an extremely heterogeneous protein residue being rather buried (Fig. 9). The aromatic amino acid in this M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3229

Fig. 9. Structural comparison between the substrate/ligand-binding sites of Grx-isoforms. The orientation is identical to the left panel in Fig. 8B. The N-terminal cysteine residue at the active site (Cysa) and the residues s1 and s2 in the s1-C-x-s2-C/S-motif at the N-terminus of helix α2 are highlighted. The predominantly positively charged glycine moiety-binding site (1) is formed by residues r1,r2 and r3 (or r3* at an alternative position). The often negatively charged γ-glutamyl moiety-binding site (2) at the N-terminus of helix α4 is formed by residues r4–r6. A conserved proline residue before strand β3 is shown in dark red at the center of each image. (A) Structures of S. cerevisiae Grx2 in the oxidized, glutathionylated and reduced state (from left to right, PDB IDs: 3CTF, 3D5J and 3CTG, respectively [206,215]). Structural rearrangements—including single side chains or the back bone of the active site motif—are indicated by black arrows. (B) Structures of S. cerevisiae Grx1 in the oxidized and glutathionylated state (PDB IDs: 3C1R and 3C1S, respectively [205]). (C) Structure of S. cerevisiae Grx6 in the glutathionylated state (PDB ID: 3L4N [216]). (D–F) Structures of human Grx2, human Grx5 and E. coli Grx4 in complex with an Fe/S-cluster (PDB IDs: 2HT9, 2WUL and 2WCI [201,209,210]). Only one subunit is displayed for clarity. An insertion between r1 and s1 is shown in pink and a WP-motif after helix α3 is highlighted in purple. The structures were visualized using Swiss-Pdb viewer.

motif (residue s2 in Fig. 9) seems to play an important structural role, of the γ-glutamyl moiety—seem to be rather variable (see also the since the γ-glutamylcysteinyl-moiety of glutathione is wrapped NMR-structures in Fig. 8B [184]). Moreover, the conformations of selected around it. A second tyrosine residue preceding Cysa (residue s1 in Grx side chains and of the peptide backbone around the active site are Fig. 9) has a surprisingly flexible side chain (Fig. 9A,B) [205,206] but quite flexible, indicating redox-dependent structural changes (Fig. 9) is often replaced, i.e. by serine or threonine (Fig. 9C,D,F). Please note [205,206,212–215]. As far as the quaternary structure is concerned, Grx that the type and the overall number of glutathione-binding residues are usually thought to be monomeric proteins. However, non-covalently

(r1–r7 in Fig. 9) differ significantly among Grx-isoforms. The basic linked dimers were detected for recombinant S. cerevisiae Grx6 and residue r1 after strand β1 seems to be the most conserved one. Inter- Grx7 [173,216], brucei 1-C-Grx1 [217], Populus tremula Grx estingly, some mono- and dithiol Grx-isoforms have an additional C4 [218] and EcGrx1 [219]. As outlined in Section 4.3.2,andasreviewed cysteine residue, replacing r4 after a GG-motif at the N-terminus of by Berndt and Lillig, several Grx-isoforms are furthermore able to bind helix α4 (C* in Fig. 9E). A CGFS-motif is common for the active site Fe/S-clusters with glutathione as a ligand (Fig. 9D–F). The association of many monothiol Grx-isoforms, but other variations such as CSYS with Fe/S-clusters can lead to the formation of dimers and tetramers [173,207] and CKYS are also found [208]. Additional structural alter- with a variety of alternative protein–protein contact sites in mono- and ations in many monothiol Grx-isoforms include an inserted loop be- dithiol Grx [173,201,209,210]. tween residue r1 and the two residues preceding Cysa (Fig. 9E,F), and the replacement of r3 in the loop connecting helix α3 and strand 4.3. Functions of Grx β3 by a WP-motif (Fig. 9E,F) [173,196,208–211]. Although the general orientation of Grx-bound glutathione is quite As described in the previous section, Grx-isoforms can be structur- similar for a variety of isoforms (Fig. 9), the conformations—in particular ally categorized, i.e. as monomeric or dimeric monothiol or dithiol Grx 3230 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266

with or without a WP-motif, an insertion after r1 or an Fe/S-cluster specialized function of this protein [55]. Noteworthy, the thiol- (Fig. 9). Grx can be furthermore grouped based on alternative bio- oxidizing agent diamide was less toxic in the absence of ScGrx1 and chemical properties such as enzymatic activities, subcellular localiza- ScGrx2 [55,226] (Section 4.4.3). Both proteins were furthermore tions or (putative) physiological functions [53–55,173,194,196,199]. reported to possess significant direct glutathione and GST

For many isoforms the functions and substrates of Grx seem to activities in vitro and in vivo [68,234] (with apparent kcat and kcat/Km −1 4 −1 −1 −1 overlap to a certain degree with Trx [220–222] (Section 6)orare values around 50 s and 3–5×10 M s for H2O2 and 1–13 s just beginning to emerge: Grx are officially classified as electron and 3–6×103 M−1 s−1 for the GST model substrate 1-chloro-2,4- donor for arsenate reductases (EC 1.20.4.1) producing arsenite and dinitrobenzene (CDNB) [234]). The latter results are quite surprising GSSG. Even though some isoforms have a high activity in the corre- and lead to the question whether such activities are either absent for sponding in vitro assay [223], this rather specialized function does other Grx-isoforms [208] or are just often overlooked. The functions of not reflect the general importance of Grx. Central physiological sub- ScGrx1 and ScGrx2 are not fully overlapping (as revealed by menadione strates of canonical dithiol Grx and Trx are the different isoforms of or paraquat treatment [55,226] and by expression analyses upon oxida- oxidized RR. Hence, Grx and Trx are crucial for DNA synthesis tive challenge and heat-shock [226]). This is plausible considering the [171,185,186,222,224]. Human Grx and Trx furthermore differen- different subcellular localizations and activities [215,228,230,231,234]. tially regulate apoptosis signal-regulating kinase 1 upon oxidative In summary, even though the exact substrates and metabolic networks challengeincellculture[221]. Moreover, a variety of Grx-isoforms remain to be unraveled, ScGrx1 and ScGrx2 have partially overlapping provide a biochemical platform for iron ion sensing and the delivery functions with Trx, GPx and GST and protect yeast cells from challenges of Fe/S-clusters (Fig. 9) and therefore play a central role in iron with oxidants and other electrophiles. homeostasis [173,193,196,198,201,209–211,225]. Further (potential) functions are outlined in Section 4.5 and were previously reviewed, 4.3.2. Enzymatic activities and functions of yeast monothiol Grx for example, by Mieyal et al. [14]. In order to provide a summary of Why do yeast cells have five monothiol Grx-isoforms (ScGrx3–7)? the functions of a complete Grx system in an organism, I will continue On the one hand, ScGrx3 and ScGrx4 as well as ScGrx6 and ScGrx7 with a comparison of the eight different Grx-isoforms from S. cerevisiae presumably originate from the aforementioned yeast genome dupli- [55,193,196–198,220]. cation event [227]. On the other hand, the proteins localize in a vari- ety of subcellular compartments. ScGrx3 and ScGrx4 are both found in 4.3.1. Enzymatic activities and functions of yeast dithiol Grx the cytosol and in the nucleus [193,196,235,236], and their additional Yeast has three dithiol Grx (ScGrx1/2/8) and five monothiol Grx Trx-like domain at the N-terminus was suggested to be a prerequisite (ScGrx3–7) [53,173,196,197]. The two dithiol isozymes ScGrx1 and for the nuclear localization [235]. In contrast, ScGrx5 is a mitochon- ScGrx2 possess a canonical KxxCPYC-motif at the active site and drial protein with an N-terminal matrix-targeting sequence [211]. share 64% sequence identity [226] (Fig. 9A,B). The high similarity ScGrx6 and ScGrx7 are the first Grx-isoforms that were identified in presumably originates from a yeast genome duplication event in the the secretory pathway of eukaryotes. Both proteins are N-terminally course of evolution [227]. The third dithiol Grx-isoform ScGrx8 has membrane-anchored facing the lumen of the cis-Golgi [197,198]. an unusual Trp14-type SWCPDC-motif at the catalytic center [55] In addition, tagged ScGrx6 was also detected in the endoplasmic retic- and is a bona fide candidate for a Grx/Trx hybrid (Section 4.2.1). ulum [198]. Estimated concentrations [232] of ScGrx3 and ScGrx4 ScGrx1 and ScGrx8 both lack a targeting signal and are therefore are approx. 1.1×104 and 7.8×103 molecules per cell, respectively. considered to be cytosolic proteins [55,196,228]. A GFP-fusion con- Based on these numbers and on an estimated compartment volume struct of ScGrx8 was indeed detected in the cytosol [229]. In con- of 25 fl the concentrations of ScGrx3 and ScGrx4 are roughly 0.4 trast, ScGrx2 is dual targeted to the cytosol and to the mitochondrial and 0.3 μM, respectively. The organellar concentrations of ScGrx5 matrix owing to alternative in-frame translation start codons. A sub- and ScGrx6 (6.3×103 and 1.6×103 molecules per cell, respectively) population of the unprocessed mitochondrial precursor was further- are presumably higher owing to the smaller volume of the cellular more suggested to localize to the outer mitochondrial membrane compartments. [228,230,231] (or could be in the intermembrae space). According First functional insights on monothiol Grx were gained during to a global protein analysis [232], ScGrx2 is far more abundant than the last decade by the Herrero lab: Loss of ScGrx5 resulted in growth ScGrx1 and ScGrx8 (approx. 3×104,3×103 and 6×102 molecules defects in minimal medium, hypersensitivity to external oxidants per cell, respectively). This estimation is in good agreement with and protein hypercarbonylation indicative for oxidative damage activity measurements in cell extracts from wild type and Grx- [237]. Like most other monothiol Grx, ScGrx5 was found to be mutant strains, suggesting that ScGrx2 accounts for the majority of inactive in the HEDS assay (although the protein was shown to the detected activity in the HEDS assay [226]. The GSH:disulfide deglutathionylate rat carbonic anhydrase III in vitro) [238]. Notably, oxidoreductase activity with HEDS was also confirmed for recombi- the absence of ScGrx5 resulted in the accumulation of iron ions and nant ScGrx1 and ScGrx2 [206,215] (with kcat and kcat/Km values in a reduced activity of Fe/S-cluster-containing enzymes [196,211]. from secondary plots of 17 s−1 and 2.8×103 M−1 s−1 for ScGrx1, Moreover, the growth phenotype was reversed by the overexpression −1 5 −1 −1 and 129 s and 1.4×10 M s for ScGrx2 [215]). Apparent kcat of the genes SSQ1 and ISA2 [211] which are both involved in the and kcat/Km values for ScGrx8 were approx. thousand fold lower mitochondrial biosynthesis and assembly of Fe/S-clusters [239]. [55]. Thus, the enzyme that was initially characterized by Nagai and Thus, mitochondrial ScGrx5 was the first Grx-isoform shown to par- Black [172] (Section 4.1) was most likely ScGrx2. ticipate in iron metabolism. The auxotrophy in the absence of The exact physiological substrates of ScGrx1/2/8 are (predomi- ScGrx5 was presumably based on the requirement of Fe/S-cluster- nantly) unknown. As far as the reduction of RR is concerned, Trx- containing enzymes for the biosynthesis of selected amino acids, isoforms seem to be more relevant electron donors than Grx [233]. whereas the increased susceptibility to oxidants could have been Yeast strains carrying a single, double or triple deletion of the genes caused by the oxidizing basal conditions [196,211] owing to Fenton encoding ScGrx1/2/8 were not only viable, but also grew with unaltered chemistry (Fig. 2A). The increased glutathionylation of proteins such rate on fermentable/non-fermentable carbon sources or on minimal as GAP-dehydrogenase [240] in a ScGrx5 mutant strain supports the medium [55,226]. However, single and double mutant strains of latter theory. Most important, the function of ScGrx5 in the assembly ScGrx1 and ScGrx2 were more susceptible to external hydroperoxides, or synthesis of Fe/S-clusters seems to be conserved in the course paraquat or iron chloride, and an overexpression of both genes in- of evolution [196] as indicated by studies on zebrafish [241] and creased the tolerance towards oxidants [55,68,220,226,234]. The dele- A. thaliana [242]. Further studies are now required to decipher the tion of ScGrx8 did not alter the growth phenotypes, suggesting a exact mode of action of ScGrx5. Recent analyses on ScGrx5 and its M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3231 homologues suggest a direct interaction between the complexed 178,186,187,238,252,259,261,262]. To my knowledge, there are no Grx-isoform and Fe/S-cluster-binding proteins in mitochondria and standardized kinetic assays for the Grx-dependent transfer or incorpo- chloroplasts [196,210,242,243]. ration of Fe/S-clusters, and it remains to be shown that such processes

ScGrx3 and ScGrx4 are also involved in iron metabolism [193,196]. obey typical with kcat and Km values. In fact, the in Both proteins interact with the iron-sensing transcription factor vitro transfer of an Fe/S-cluster from poplar GrxS14 to ferredoxin was Aft1 in the nucleus. Under conditions of limited iron ion availability, best fit by (non-enzymatic) second order kinetics yielding an apparent Aft1 activates the transcription of genes in an iron regulon. The rate constant of 3×102 M−1 s−1 [242]. Thus, I will focus in the follow- Grx-isoforms negatively regulate Aft1 as reflected by a constitutive ing sections on the thiol:disulfide oxidoreductase activities of Grx transcription in the absence of ScGrx3 and ScGrx4. In the presence and outline only structural aspects as far as Fe/S-cluster binding is of ScGrx3 and ScGrx4 the localization of Aft1 is shifted towards concerned. the cytosol. The redistribution is mediated by the Grx- and not by the Trx-domain of ScGrx3 and ScGrx4 in an iron-independent 4.4.1. The traditional model of Grx catalysis manner [244,245]. Analogously, the homologue of ScGrx4 from The traditional model of the catalytic mechanism of dithiol Grx Schizosaccharomyces pombe triggers the nuclear export of Php4, a is summarized in Fig. 10 [14,53,55]. It is based on studies on E. coli component of an iron-dependent transcription repressor complex Grx by Holmgren and co-workers [54,187,200] as well as on kinetic [246]. ScGrx4 was furthermore shown to be a substrate of the nuclear analyses on mammalian Grx (thioltransferase) by Yang and Wells kinase Bud32 in vitro and in vivo, but the physiological relevance [263,264] and, in particular, by Mieyal and colleagues [184,252, of the phosphorylation remains to be deciphered [236,247]. Apart 256,262,265]. According to the traditional model, dithiol Grx reduce from the interaction with Aft1, Fe/S-cluster-containing ScGrx3 and disulfide bonds of (i) glutathionylated substrates or (ii) protein disul- ScGrx4 were shown to play an important role in iron ion sensing fide substrates: and delivery in the cytosol [193]. The latter function probably explains the (redox-sensitive) phenotype in the absence of both (i) Deglutathionylation of substrates occurs via a monothiol proteins owing to free iron ions (Fig. 2A) [193,245,246]. Noteworthy, ping-pong (double displacement) mechanism. During the oxi- regulatory roles of monothiol Grx-isoforms containing a Trx-domain dative half-reaction of Grx, the reduction of a glutathionylated are not restricted to yeast [53,196]. For example, the human PICOT substrate starts with the nucleophilic attack of the thiolate protein also localizes to the cytosol and the nucleus of T lymphocytes group of Cysa (Fig. 10). The deglutathionylated first product is where it was reported to negatively regulate the activity of protein released, and a mixed disulfide between glutathione and Cysa kinase Cθ and, therefore, to influence the transcription factors AP-1 of Grx is formed (Grx-SSG). During the reductive half-reaction, and NFκB [248]. In summary, the monothiol Grx-isoforms ScGrx3 and one molecule GSH regenerates Grx, yielding dithiol Grx(SH)2 ScGrx4 significantly differ from mitochondrial ScGrx5. They exert and GSSG as the second product. Please note that the more overlapping regulatory functions with respect to iron-dependent gene C-terminal cysteine residue of the CxxC-motif at the active site expression in the nucleus and influence the availability of intracellular of dithiol Grx-isoforms is not required for glutathionylated sub- iron ions for biosynthetic processes. strates in this model. The formation of Grx disulfide (Grx(S2)) In contrast to ScGrx3-5 and most other monothiol Grx-isoforms, is in fact considered an unnecessary side reaction detracting ScGrx6 and ScGrx7 have a significant oxidoreductase activity in the from catalysis. HEDS assay in vitro [173,197,198,216]. Genetic experiments further- Steady-state kinetics with the substrates GSH and Cys-SSG (as more suggest a contribution of both proteins to redox homeostasis well as mass spectrometric analyses) support a ping-pong in the secretory pathway in vivo [197,198]. Metabolic challenges mechanism for dithiol Grx1 from human erythrocytes and rat lead to a differential up-regulation of the encoding genes depending liver [262], for recombinant human dithiol Grx2 [252] and on the Crz1-calcineurin pathway (ScGrx6) or the transcription factor for monothiol ScGrx7 [55,173]. Ping-pong patterns were also Msn2/4 (ScGrx7) [198]. Notably, ScGrx6—but not ScGrx7—binds a found for human dithiol Grx1 and Grx2 as well as poplar glutathione-stabilized Fe/S-cluster in vitro, resulting in a loss of monothiol GrxS12 using GSH and glutathionylated BSA (or he- the oxidoreductase activity [173]. Since ScGrx6 also binds iron ions moglobin) as substrates [252,262,265,266].Ofnote,replacing in vivo [198], it is tempting to speculate that ScGrx6 plays a role in the second cysteine residue in the CxxC-motif with serine did iron metabolism (i.e. as a sensor regulating a transporter-dependent not abolish the general activity of several dithiol Grx-isoforms uptake or the storage of iron ions). In summary, ScGrx6 and ScGrx7 with these substrates or with HEDS [184,200,252,257,263]. are enzymatically active monothiol Grx-isoforms with partially over- Thus, the monothiol mechanism was validated for small lapping functions in the secretory pathway. Their physiological sub- glutathionylated molecules and proteins and is utilized by strates or interaction partners are unknown. ScGrx6 might play mono- and dithiol Grx-isoforms. a role in iron metabolism in analogy to the monothiol Grx-isoforms (ii) In contrast to the monothiol mechanism, the enzyme species

ScGrx3–5. Grx(S2) is a central intermediate during the reduction of select- ed protein disulfide substrates via the dithiol mechanism 4.4. The enzymatic mechanism of Grx (Fig. 10). In this model, the dithiol Grx-isoform and the pro- tein disulfide substrate exchange the disulfide bond. Thus, an Many, but by far not all, Grx-isoforms catalyze the GSH-dependent intermolecular protein–protein disulfide bond between the sub- reduction of the model substrate HEDS in a standard GR-coupled strate and Grx is transiently formed before the C-terminal cyste- enzymatic assay [53–55,172,173]. A significant activity in the HEDS ine residue in the CxxC-motif attacks the sulfur atom of Cysa. assay (yielding GSSG and two molecules of 2-mercaptoethanol) was Grx(S2) is subsequently reduced in two steps with the help of detected for dithiol Grx-isoforms from E. coli [178,186,249,250],mam- two molecules of GSH yielding GSSG. The species Grx-SSG is an mals [59,178,249,251–255], P. falciparum [256],yeast[172,206,226] as intermediate of the latter regeneration. well as plants and algae [257–259]. In contrast, with very few excep- RR from E. coli is the traditional model substrate for the disulfide tions [173,198,216], all monothiol Grx-isoforms analyzed so far were mechanism. A mutant of EcGrx1 with a serine residue replacing found to be inactive in this assay [58,208,238,257,259,260]. Alternative the second cysteine residue in the CPYC-motif was still function- Grx-dependent assays include the reduction of L-cysteine-glutathione al in the HEDS assay but lost its activity with RR [200].TheNMR disulfide (Cys-SSG), dehydroascorbate, RR, 3′-phosphoadenylylsulfate structure of a mixed disulfide between mutant EcGrx1 and a reductase, insulin and glutathionylated model proteins [53,55,173, peptide of subunit B1 of RR furthermore suggested the 3232 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266

glutathionylated [58,208,259]. Moreover, in this model, the second cysteine residue in the CxxC-motif is dispensable for catalysis and can be replaced with another amino acid without (complete) loss of function (Fig. 11) in accordance with previ- ous findings on monothiol [173] and mutant dithiol Grx- isoforms [184,200,252,257,263]. The scaffold model could also explain the rather low affinity of some Grx-isoforms for GSH as a reducing agent, or why GSH can be replaced by other thiols during the reductive half-reaction in vitro [265]. Based on this model, the structures of glutathionylated Grx-isoforms in Fig. 9 would reveal the permanent binding site for the glutathione moiety of the first substrate (Fig. 11). Nevertheless, the gluta- thione scaffold model also has several limitations. Why are most monothiol Grx-isoforms inactive in the HEDS assay [58,208,238,257,259,260]? Why do mutations of the second cysteine residue in the CxxC-motif (or of residues that are equivalent to C* in Fig. 9) sometimes significantly reduce or even abolish the enzymatic activity [55,200,238,257]? These phenomena might be explained by the glutathione activator model and the cysteine resolving model. (ii) For the glutathione activator model, the interaction of Grx with GSH during the reductive half-reaction is most important for catalysis (Fig. 11) [55]. Kinetic studies, in particular by Srinivasan et al., indeed suggest that the reductive Fig. 10. Traditional model of Grx catalysis. The monothiol mechanism for glutathionylated half-reaction is critical and often rate-limiting for two reasons: fi substrates and the dithiol mechanism for protein disul de substrates are shown on the First, GS- nucleophile formation and, second, the stability of the left and right sides, respectively. Since Cysa is sufficient for the monothiol mechanism, the second cysteine residue in the CxxC-motif can be replaced with serine (X). A side leaving group [55,238,265,266]. The thiol group of Cysa in reaction of the monothiol mechanism for dithiol Grx is shown in the middle. See Figs. 9–11 is quite acidic with an (apparent) pKa value that usu- Section 4.4.1 for further details. ally ranges from 3.2 to 5.5 [208,215,238,252,255,266]. Hence,

Cysa is deprotonated and highly reactive in vitro and in vivo. According to the glutathione activator model, a strong/perma- formation of a solvent-protected disulfide bond that cannot be nent interaction with the disulfide substrate is not necessarily re- easily attacked by an external nucleophile [212]. Sterical hin- quired as long as a correct orientation of the substrate is achieved drance might therefore explain the general relevance of dithiol forashortmomentduringtheoxidativehalf-reaction(Fig. 11). Grx for the reduction of non-glutathione disulfides. However, (The model in principle also allows the turnover of non- the dithiol mechanism for RR itself is not conserved throughout glutathionylated low molecular weight and protein disulfide sub- evolution. Studies on a mammalian Grx/RR couple indeed re- strates.) Binding and activation of the nucleophile GSH, however, vealed an efficient GSH-dependent reduction of oxidized RR re- presumably require a defined interaction (Fig. 11), for example, gardless whether the second cysteine residue in the CxxC-motif with a potential base or a thiolate-stabilizing residue such as of Grx was replaced by serine or not. Thus, RR was probably the positively charged r in Fig. 9. Distinct, yet unassigned, glutathionylated and subsequently reduced by Grx via the 1 GSH-activating interactions could therefore be responsible for monothiol mechanism [224]. Since the dithiol mechanism does the variable enzymatic activities of Grx-isoforms [55]. Some stud- not seem to be a prerequisite for the reduction of protein disul- ies indeed indicate a specific recognition of GSH via the fide bonds, why are dithiol Grx-isoforms found in so many bac- γ-glutamyl moiety, resulting in a significantly increased activity teria and all kinds of eukaryotes? To address this and other as compared to other thiols [252,265,267]. Moreover, the (appar- questions in more detail, we will have to extent and refine the ent) K value for GSH ranges from 0.3 to 4 mM for different traditional model of Grx catalysis. m Grx-isoforms [55,173,207,253,266], pointing to a highly variable affinity for the reducing agent. Regarding the glutathione moie- 4.4.2. Refined models of Grx catalysis ties in the Grx and GST structures in Fig. 8BandD,respectively, A central aspect of Grx catalysis is not addressed by the traditional similar molecule orientations but different positions of the gluta- model: The glutathionyl moieties of protein-bound RSSG and GSH are thione sulfur atoms become obvious. The position of the glutathi- not equal owing to the geometry of the transition state of S 2 reac- N one moiety in the GST structure in Fig. 8D could indeed resemble tions (Fig. 5). In fact, both moieties must adopt alternative positions the position of the nucleophile GS− in Grx. In the same line of during catalysis unless significant conformational changes occur thought, the thiolate of GS− could become activated/stabilized (Fig. 11) [55]. As a consequence, two non-exclusive mechanisms can by the basic side chain of r (explaining its conservation in Grx) be theoretically distinguished: (i) The “glutathione scaffold model” 1 which is equivalent to Tyr in GST. Of note, a basic residue at and (ii) the “glutathione activator model”. (iii) Furthermore, a combi- a the same position is also important for serine- and nation of the latter model with a concept named “cysteine resolving cysteine-dependent GST-isoforms (Section 8.3). model” might explain the relevance of dithiol Grx-isoforms [55]. As demonstrated for mammalian Grx1 and Grx2, the reductive

(i) For the glutathione scaffold model, the glutathionyl moiety of half-reaction also depends on the Cysa thiolate leaving group fi the disul de substrate occupies the glutathione-binding site [252,265]: The lower the pKa value of Cysa, the higher the reac- during the entire catalytic cycle (Fig. 11). Grx working with tivity of the Grx-isoform because of the stabilized leaving group

this mechanism would therefore provide an optimal geometry of the second SN2 reaction. This might explain the absent enzy- for the first SN2 reaction (the oxidative half-reaction). The matic activity of some Grx-isoforms with less acidic (predicted)

scaffold model could explain why even monothiol Grx- pKa values around 5–6 [55,208,238]. A comparison of ScGrx1 isoforms that lack an activity in the HEDS assay can be easily and ScGrx2, however, revealed that such a strict correlation M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3233

Fig. 11. Refined models of Grx catalysis. The glutathione scaffold and the glutathione activator model discriminate between the non-identical glutathione moieties of both Grx sub- strates (see also Fig. 5). The cysteine resolving model provides an alternative mechanistic explanation for the conservation of a proximal cysteine residue in Grx. The models are not exclusive. See Section 4.4.2 for further details.

between pKa value and activity is not always valid [215].Insum- reversed (Fig. 11). Based on these mechanistic assumptions, mary, the contribution of both factors—GS− nucleophile forma- some structures of glutathionylated Grx-isoforms (Fig. 9)might tion and leaving group stability—could be weighted differently resemble such a protein conformation. If the enzyme has a depending on the enzyme. CxxC-motif or a C* residue (Fig. 9), the trapped protein species Please note that the scaffold and activator models are not exclu- could be reintroduced in the catalytic cycle (Fig. 11). The role sive. For example, there could be two glutathione-binding sites, of the second cysteine would therefore be to maintain or to re- one for the disulfide substrate and one for GSH. Sequential generate the functional Grx-isoform. The cost for this resolving kinetic patterns with HEDS [173,255,262] might indeed be inter- functioncouldbethatsomedithiolGrxhaveareducedactivity preted this way as discussed in Section 4.4.4. In another scenario, owing to the additional reaction loop depicted in the resolving Grx-SSG could be formed according to the scaffold model model [55]. In conclusion, by considering the reaction geometry and subsequently undergo a conformational change (Fig. 9 and and potential conformational changes, the refined mechanistic Section 4.2.2). As a result, the glutathione-binding site becomes models could be useful to unravel Grx catalysis and the molecu- available for GSH and the reaction could proceed according to lar evolution of this highly versatile protein family. the activator model (Fig. 11) [55]. (iii) The cysteine resolving model extends the activator model and 4.4.3. Properties of Grx reaction intermediates in vitro and in vivo

provides an alternative explanation as to why so many Grx- As mentioned in Section 4.4.2, Cysa of reduced Grx is deprotonated isoforms retained a second cysteine residue in the course of evo- under physiological conditions; but how stable are the reduced and lution. After release of the first product, the reaction could either the oxidized protein species? In vitro, many monothiol isoforms continue without any difficulties, or the covalently bound gluta- are efficiently glutathionylated at the active site (and at additional thione moiety could undergo a conformational change and residues), and the resulting Grx-SSG species are often stable occupy the GSH activator-site (Fig. 11) in agreement with kinetic [55,205,208,216,238]. Interestingly, the proposed dead-end species studies on ScGrx7 and ScGrx8 [55]. For inactive (artificial) with respect to monothiol Grx-SSG in Fig. 11 (Section 4.4.2) fulfills monothiol Grx-isoforms the described side reaction could result the recently proposed central criteria for redox sensors in vivo [11]: in a dead-end complex unless the conformational change is First, Grx can have a high specificity and high reactivity with respect 3234 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 to a glutathionylated signal molecule. Second, if catalysis stops after The sequential kinetic patterns with the model substrate HEDS are the oxidative half-reaction in Fig. 11, specific signal transduction can another enigma of Grx-catalysis [173,255]: Experimental data sup- take place because of a significant structural change of the trapped port the theory that HEDS itself is not the substrate of Grx. It first sensor [55]. Thus, the low or absent activity of some Grx-isoforms has to non-enzymatically react with GSH yielding a mixed disulfide during the reductive half-reaction might be a desirable property in between GSH and 2-mercaptoethanol (EtOH-SSG) that is subse- vivo. In contrast, the Grx-SSG species of dithiol isoforms is usually quently attacked by Cysa [55,173,255,262]. Nevertheless, the non- instable owing to the rapid formation of Grx(S2) [55,262,266]. enzymatic formation of the substrate does not necessarily explain The reactivity/stability might be reflected by the redox potentials the sequential patterns [55,173], and other factors such as (i) de- which vary drastically among Grx-isoforms. However, the (macro- creased net concentrations of GSH in the assay (owing to EtOH-SSG scopic) E°′ values are seldom comparable because of multiple thiol formation), (ii) product inhibition (owing to 2-mercaptoethanol groups and alternative redox states (including the formation of formation during the non-enzymatic reaction), or (iii) a sequential intra- and intermolecular disulfide bonds). For example, ScGrx5 has mechanism reflecting simultaneous binding of small EtOH-SSG and a quite high E°′ value of −175 mV [238] as compared to −198 mV GSH cannot be excluded. Notably, in contrast to Grx, ping-pong and −233 mV for EcGrx3C66Y and EcGrx1, respectively [268]. The patterns at low HEDS concentrations were reported for bacterial pH-adjusted E°′ value of a chloroplast monothiol Grx-isoform is cysteine-containing GST B1-1 [269]. Thus, assay-dependent aspects even as low as −270 mV and therefore more similar to canonical (i) and (ii) seem less likely, and the sequential patterns might indeed Trx [259]. The redox potentials of EcGrx3C66Y and EcGrx1 were indicate simultaneous binding of two small substrates to Grx. A more assigned to the couple Grx(S2)/Grx(SH)2 with the typical CxxC disulfide detailed analysis of the HEDS assay could therefore reveal whether bond at the active site [268].Incontrast,E°′ of ScGrx5 reflects the redox there are two different glutathione-binding sites (with significant potential of the couple Grx(S2)/Grx(SH)2 with an atypical disulfide implications for the models in Fig. 11). In summary, much more bond between residue Cysa and residue C* (Section 4.2.2). Residue C* work is required to unravel the enzymatic mechanism(s) of Grx and is also found in some other “monothiol” Grx-isoforms—including to define which structure–function relationships precisely determine EcGrx3, EcGrx4 (Fig. 9F) [202] and the mentioned chloroplast protein the enzymatic properties in vitro and in vivo. [259]—as well as canonical dithiol Grx-isoforms such as mammalian Grx1 [183]. Of note, in the solution structure of reduced EcGrx4 and in 4.5. Physiological and medical relevance of Grx catalysis the crystal structures of oxidized and glutathionylated mammalian

Grx1, the thiol groups of Cysa and residue C* are separated by 9 to The physiological relevance of Grx catalysis depends on the iso- 14 Å [183,184,202].Thus,disulfide bond formation between these res- form and on the presence or absence of redundant redox systems. idues would require significant structural rearrangements in accor- For example, whether monothiol Grx that are involved in iron meta- dance with the resolving cysteine model (Fig. 11). In summary, bolism really have an enzymatic function in vivo awaits clarification. macroscopic and, in particular, microscopic redox potentials probably Furthermore, knock-outs of Grx-isoforms in yeast result in inconspi- differ considerably among alternative oxidized or glutathionylated cuous up to lethal phenotypes (Section 4.3), and Grx1 knock-out Grx species and conformations. It is therefore extremely diffult to corre- mice are not only viable, but also do not even have an increased sus- late E°′ values with the reactivity/stability of Grx. ceptibility to hyperoxia-mediated injury [270]. Of note, the distribu- How much Grx is reduced, glutathionylated or associated with tion of the human isoforms Grx1/2/3/5 has been recently reported Fe/S-clusters in vivo? Even though there is no quantification of in an impressive redox atlas, revealing tissue-specific localization Fe/S-cluster binding, redox proteomics revealed that approx. 30% of and expression patterns [271]. Such patterns might explain why residue C* of EcGrx3 are oxidized under aerobic steady-state condi- some tissues are more susceptible to genetic manipulation than tions in vivo. This value increased to roughly 65% upon oxidative others. For example, knock-down of dithiol Grx2 in zebrafish was challenge [140]. Cysa in ScGrx1 was even 80% oxidized under shown to impair neuron development, presumably owing to redox steady-state conditions [142]. Although Grx-SSG and Grx(S2) were regulation of collapsin response mediator protein 2 [225]. not discriminated in these quantitative studies, the results are quite Other reversibly glutathionylated components that were suggested surprising considering the low redox potential of the cytosol (Fig. 4) or sometimes even shown to be physiological Grx substrates include and the range of E°′ values mentioned above. It is therefore possible glycolytic enzymes [272,273], creatine kinase [274], carbonic anhydrase that Grx-SSG or Grx(S2) accumulate in vivo owing to the rather III [275], components of the respiratory chain [276,277],actin[278,279], slow reduction by GSH (see also Section 2.2.2). This interpretation membrane receptors, transporters and ion channels [280–284], several has important implications with respect to the physiological function: protein kinases and phosphatases [285–290] as well as other pro- Are Grx predominantly deglutathionylating enzymes [14], sensors, teins involved in signal transduction such as nuclear factor I [291],Ras or can they sometimes also glutathionylate selected proteins [252] [292] and NFκB [293,294]. Glutathionylation of these proteins (poten- by catalyzing the reverse reaction in Fig. 11? The latter possibility tially) alters enzymatic activity [273–275,285–290], ion and metabolite might, for example, explain why artificial diamide treatment of yeast transport [280,282–284], the formation of ROS [276,277], the cytoskel- cells (resulting in an increase of GSSG) is less toxic for Grx knock-out eton [279], signal transduction [285,289,291,292] and cell death strains [55], because the accumulation of (inactivated) glutathionylated [281,283,294]. Thus, as reviewed by Mieyal et al., perturbations in Grx enzymes could be slowed down. However, this hypothesis obviously content and activity could have consequences on the organellar, cellular requires experimental validation. In summary, the physiological redox and organismic level with implications for pathophysiological condi- states of several Grx-isoforms seem to lack an obvious correlation tions including diabetes, cardiovascular diseases, lung diseases and with E°′ values, and our understanding of Grx catalysis in vivo remains [14]. A compromised glutathione system was also suggested incomplete. to be detrimental to the human brain [295], and alterations of Grx contents and/or the accumulation of glutathionylated proteins due to 4.4.4. Outlook on Grx catalysis and mechanistic questions oxidative challenge might therefore influence the onset or the progres- To the best of my knowledge, it is neither known which structure– sion of Alzheimer's disease, Parkinson's disease, Huntington's disease, function relationships exactly determine whether a Grx-isoform has a amyothrophic lateral sclerosis and Friedreich's ataxia [14]. hydroperoxidase or GST activity, nor how such Grx-catalyzed reac- Importantly, the number and variety of proteins that are (de) tions exactly proceed. Furthermore, the potential proton acceptor(s) glutathionylated in vertebrates point to regulatory networks far or stabilizing residues with respect to Cysa and GSH activation—as beyond so-called oxidative stress (Section 2.1.1). The involvement well as the proton donor for the thiolate leaving group—are unknown. of Grx in such physiological redox networks is not restricted to M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3235 vertebrates, but seems to be an evolutionary conserved principle. For intramolecular disulfide bond between the active site and the Cys example, Grx-dependent redox networks are also found in plants block [316] in analogy to atypical Prx-isoforms (Section 6.2). Even [194,296] and begin to emerge for malaria parasites [297]. Neverthe- though the structural rearrangements upon disulfide formation were less, we are still far from understanding the precise physiological less drastic for a GPx-isoform from T. brucei [318], such conformational functions of Grx catalysis and their implications for health and changes are presumably a prerequisite for the transfer of electrons disease. from bulky protein substrates. In agreement with this theory, a (puta-

tive) resolving cysteine in helix α2 is found in the Trx-dependent 5. Glutathione peroxidases GPx-isoforms [204,312,314,316,318], whereas the GSH-dependent isoforms either do not have such a residue or (were suggested to) 5.1. Pioneers of GPx catalysis adopt alternative quaternary structures with a trapped conformation preventing disulfide bond formation (Fig. 12A,B) [60,204]. Of note, a

The term was introduced by Mills who dis- second/third cysteine residue in a FPCN-motif at the end of strand β2 covered the activity with H2O2 in enzyme preparations from mamma- is extremely conserved among all GPx classes (Fig. 12A,B). However, lian erythrocytes in 1957 [298]. In 1962, Neubert et al. described GPx this residue seems to be dispensable for GPx catalysis and its role as “contraction factor I” for swollen mitochondria [299]. (The authors remains unclear (Section 5.4) [314,315]. presumably observed the influence of GPx4 on the mitochondrial Another structural variation among the different GPx classes is a transition pore owing to the removal of mitochondria-derived hydro- so-called oligomerization loop after helix α3 at the top of the active peroxides.) The ping-pong mechanism of GPx1 [300] and the sig- site (Fig. 8E,F). This loop mediates the contact between two dimers nificant reactivity with a variety of hydroperoxides [301] were in crystallized homotetrameric bovine GPx1 and covers a part of characterized in 1972 by Flohé and co-workers. One year later, the active site (Fig. 12A) [304]. Accordingly, GPx-isoforms cannot based on studies with crystalline bovine GPx1, the tetrameric enzyme only be grouped based on their electron donor, active site residue became the first mammalian to be discovered [302], and sequence similarity, but also on the their quaternary structure: pointing to an explanation as to why some GPx have such a high Mammalian GPx4 (Fig. 12B) as well as GPx-isoforms from P. falciparum, activity in the absence of heme (see also Sections 2.3.1 and 5.3). In T. brucei and D. melanogaster were reported to be monomeric enzymes 1979 and 1983, Ladenstein and Epp et al. solved and refined the crys- [204,312,318], whereas poplar GPX5 forms unusual dimers that are tal structure of bovine GPx1 (even though the full sequence was still similar to Prx [316].Allofthesefive examples have no oligomeriza- unknown resulting in some reassignments) [303,304]. The analysis of tion loop in contrast to homotetrameric mammalian GPx1/2/3 rat liver GPx by Forstrom et al. revealed that selenium is incorporated [304,319,320] and related GPx5/6 [204]. This demonstrates that as selenocysteine [305]. This conclusion was later confirmed by the quaternary structure is indeed correlated with the oligomeriza- Günzler et al. who determined the amino acid sequence of bovine tion loop, whereas the GSH-dependency (e.g. of monomeric GPx4 GPx1 [306], and by Chambers et al. who showed that the seleno- and tetrameric GPx1) is not. To my knowledge, it is unclear whether cysteine residue in mouse GPx is encoded by a [307]. tetrameric mammalian GPx-isoforms evolved as a specialized subclass During that time GPx4—having a high peroxidase activity with of the already specialized class of GSH-dependent GPx, or whether con- phospholipid substrates—was discovered as a second mammalian vergent evolution is the cause for the GSH-dependency of monomeric selenoprotein by Ursini and colleagues [308,309]. In 1992, Rocher GPx4 and tetrameric GPx1. et al. experimentally confirmed that a selenocysteine to cysteine sub- Functionally, GSH-dependent GPx are GSH:hydroperoxide and/or stitution significantly reduces the GPx activity and that a serine GSH:lipid-hydroperoxide (EC 1.11.1.9 and/or EC mutant is completely inactive [310] (see also Section 2.3.1). Interest- 1.11.1.12) with a more or less pronounced specificity for the peroxide ingly, of the seven or eight mammalian GPx-isoforms only four or five substrate [61,62,301,309]. Since GPx3 is a secreted plasma protein have a selenocysteine at the active site, and two or three were shown [320]—and therefore cannot efficiently work at low GSH concentra- to possess a cysteine residue [60]. Furthermore, among the numerous tions (Fig. 4)—the protein has been suggested to be a good candidate GPx-isoforms that have been identified in all domains of life, non- as a redox sensor [60]. Mammalian GPx4 was furthermore shown to mammalian GPx usually have a cysteine residue at the active site have an additional moonlighting function as a structural protein of [311]. During the last decade, several of these cysteine-containing the mitochondrial capsule that is formed during spermatogenesis: GPx-isoforms from a variety of organisms turned out to be Trx- Redox-dependent capsule formation requires the covalent oligomeri- dependent enzymes as demonstrated by Sztajer et al., Jung et al., Tanaka zation of the protein via the active site and surface-exposed cysteine et al., Maiorino et al. and Schlecker et al. [204,312–315]. Thus, GSH- residues (Fig. 12B) [60,317,321]. GPx4 therefore provides an excellent dependent (mammalian) GPx-isoforms are the exception and seem to example for a rather recent event in the molecular evolution of be a rather recent evolutionary acquisition (Section 4.2.1). glutathione-dependent enzymes outlined in Sections 2.3 and 4.2.1. In contrast to GPx1 and intestinal GPx2 in the cytosol, and secreted 5.2. Structure and function of GPx GPx3 in the plasma, GPx4 has an extremely variable subcellular locali- zation in the cytosol, in the nucleus and in mitochondria (Section 5.4). GPx have a significantly altered thioredoxin fold containing several Furthermore, the protein exists in a free and a membrane-associated additional structural elements as highlighted in Fig. 8E,F. For example, form [309] in accordance with its preference for lipid hydroperoxides. helix α2—which seems to be specific for GPx and Prx (Fig. 8E–H)— Some Trx- or tryparedoxin-dependent GPx are also found in chloro- contributes to the monomer–monomer contact site in crystallized plasts or mitochondria [322,323]. homotetrameric GPx1 (Fig. 12A) [304]. This area is highly variable among the different GPx classes [204,316,317]. Moreover, helix α2 5.3. The enzymatic mechanism of glutathione-dependent GPx often contains a cysteine residue that is mechanistically relevant for Trx-dependent GPx-isoforms such as yeast Gpx2 [313], poplar GPX5 Numerous kinetic measurements clearly revealed that diverse [316], and isoforms from Drosophila melanogaster [204] and T. brucei GPx-isoforms act via a ping-pong mechanism [60,300,301]. Thus, [314,318]. This structural element was therefore termed “Cys block” there is no ternary complex between the enzyme, the hydroperoxide (Fig. 12A) [204]. The active site (seleno)cysteine residue is found at and GSH, and the reaction can be subdivided into an oxidative half- the N-terminal end of helix α1 (Fig. 8E,F). In the crystallized structure reaction and a reductive half-reaction. Even though the GSH- of oxidized poplar GPX5, partial unfolding of the N-terminus of helix dependent overall reaction somehow resembles the Grx-catalyzed

α1 and complete unwinding of helix α2 allowed the formation of an reduction of disulfide substrates (Fig. 3C,D), the mechanism for the 3236 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266

Fig. 12. Structures of GSH-dependent GPx. Helices and strands are labeled according to Fig. 8E,F. (A) View along the 2-fold axis of two subunits of homotetrameric GPx1. Two sub- units interacting with the so-called oligomerization loops are omitted. The selenocysteine residue in the ‘over’-oxidized seleninate state is highlighted and further structural ele- ments are labeled. (B) Similar view on monomeric GPx4 with the catalytic tetrad highlighted (the selenocysteine at the N-terminus of helix α1 is mutated to cysteine). GPx4 lacks the oligomerization loop. Moreover, helix α2 adopts an alternative position and lacks a cysteine residue. Three additional surface-exposed cysteines are shown on the left side. (C) Zoom in at the active site of GPx4 with the catalytic tetrad and a highly conserved cysteine residue. (D) Zoom in at the predicted substrate-binding site of GPx1. Residues that were suggested to bind GSHI and/or GSHII are highlighted. The orientation is identical to the left panel in Fig. 8F. The oligomerization loop is shown in dark red. The tryptophane and asparagine residue of the tetrad are highlighted in purple and green, respectively. (E) Zoom in at the predicted substrate-binding site of GPx4. The images were generated using Swiss-Pdb viewer and the structures of bovine GPx1 and human GPx4U47C (PDB IDs: 1GP1 and 2OBI, respectively [304,317]). See Sections 5.2 and 5.3 for details.

GPx-catalyzed reduction of hydroperoxides is very different (Fig. 13). After the highly reactive anionic nucleophile is formed (Fig. 13), (i) The hydroperoxide is not glutathionylated before it is attacked the ROOH substrate is irreversibly turned over as soon as it adopts by GPx. (ii) A highly reactive selelenic/sulfenic acid is formed as an in- the required SN2 geometry (Fig. 5B). Thus, based on the high reacti- termediate. (iii) The regeneration of the modified enzyme requires vity of the active site residue, there is probably not a real binding two steps during the reductive half-reaction. site for the hydroperoxide substrate, explaining the rather low substrate-specificity [60,301]. Nevertheless, there are significant dif- ferences among the GPx-isoforms regarding the accessibility of the 5.3.1. The oxidative half-reaction of glutathione-dependent GPx active site, for example, owing to the presence or absence of the olig- The position of the selenocysteine/cysteine residue in GPx is omerization loop (Fig. 12D,E). These differences could explain why identical to Grx (Fig. 8A,E), however, apart from the helix dipole, monomeric GPx4 has such a high activity with lipid hydroperoxides the activation seems to differ significantly. In principle, the deproton- [60,309,317]. The oxidative half-reaction ends with the release of a ation and orientation of the active site selenol/thiol group could be molecule of water or alcohol, depending on whether R in ROOH is a identical for all GPx classes (regardless of the reducing agent) because hydrogen atom or not (Fig. 13). of the separation of the oxidative and reductive half-reaction. Activa- tion of the nucleophile was suggested to depend on the side chains of 5.3.2. The reductive half-reaction of glutathione-dependent GPx three proximal residues: a glutamine residue in a QE-motif in the loop After formation of the instable selenenic/sulfenic acid, the first after strand β2, and a tryptophane and asparagine residue in a WNF- GSH molecule (GSHI) enters the active site and is turned over during motif after the oligomerization loop preceding strand β4 (Fig. 8F, the first reductive SN2 reaction, yielding a water molecule as the Fig. 12C) [60]. Even though mutational analyses indicate that the second product (Fig. 13). (Alternatively, the selenenic acid was glutamine and asparagine residues are more important than the tryp- suggested to initially react within the catalytic tetrad yielding a tophane [60,314], the precise role of these residues is all but clear. selenylamide which subsequently reacts with GSH. This postulated First, not only the nucleophile in this potential catalytic tetrad [60] “parked” enzyme species could decrease the probability of further but also the glutamine residue is sometimes variable [316]. Second, ‘over’-oxidation in analogy to 2-Cys Prx [60,61] described in the conformation of the crystallized mammalian GPx-isoforms might Section 6.2.). Since GPx1 is highly specific for GSH [60,61], there has not be representative/universally valid for the activated enzyme. In to be a defined substrate-binding site. However, a structure of a fact, NMR and mutational analyses of a T. brucei GPx-isoform pointed glutathionylated GPx intermediate has not been solved to date, to an alternative conformation, and a conserved lysine residue at the which is not surprising considering that the second reductive SN2 end of strand β3 was suggested to play an important role for peroxide reaction directly follows in the presence of GSHII, yielding GSSG as reduction [318]. the third product (Fig. 13). It is therefore all but understood which M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3237

Fig. 13. Model of GPx catalysis. The active site either contains a deprotonated selenocysteine or cysteine residue. A conformational change could occur at the end of the oxidative half-reaction. The reductive half-reaction comprises two parts. The existence and composition of a glutathione-binding site highly depend on the GPx-isoform and therefore both predicted sites are labeled with a question mark (alternatively, GSHI and GSHII might also compete for some binding residues). See Section 5.3 for further details. glutathione molecule occupies which surface area during GPx cataly- 5.3.3. Properties of GPx reaction intermediates in vitro and in vivo sis, even though crystal structures and molecular models support the The oxidative half-reaction of GPx1 and other selenocysteine- theory that different basic residues (residues r1–r4 in Fig. 12D) bind containing GPx-isoforms is very efficient with second order rate 7 −1 −1 GSHI and GSHII via their glycine and γ-glutamyl carboxylate groups constants around 1–5×10 M s . For these enzymes there is vir- [60,61,304,317]: tually no enzyme–hydroperoxide complex, and, accordingly, the kcat For bovine GPx1, Epp et al. suggested that the glycine carboxylate and Km values for several GPx-isoforms obtained from secondary group interacts with r1, and the γ-glutamyl carboxylate group binds plots are infinite [60,300]. The ping-pong patterns and an apparently to r2 (without clear discrimination between GSHI and GSHII) [304]. absent inhibition by GSSG furthermore indicate that the enzyme– Notably, this pattern results in a significant bending of the glutathi- substrate complexes with GSHI and GSHII cannot be saturated. GSH one backbone, and the binding mode is therefore very different binding and the adoption of a correct conformation were therefore from the one seen in Fig. 9. According to a model in ref. [60],itis suggested to be rate-limiting for GPx1 catalysis [60]. In vivo, GPx- the glycine carboxylate group of GSHI that interacts with r1, whereas isoforms are expected to be more or less fully reduced. As a conse- the γ-glutamyl carboxylate group was suggested to bind to r4 instead quence, irreversible hydroperoxide removal was suggested to only of r2. In this model, r2 binds the glycine carboxylate group of GSHII, depend on (i) the concentrations of GPx and hydroperoxide, (ii) the and the two γ-glutamyl carboxylate groups of both glutathione encounter of both molecules, and (iii) the second order rate constant molecules compete for r4 and a lysine residue from a second sub- of the oxidative half-reaction. Thus, under physiological conditions, unit. To increase the number of permutations—and the confusing the exact concentration of GSH or the redox potential (see also possibilities—a recent model for human GPx1 suggested that the Section 2.2.2) could be irrelevant for GPx catalysis [60,300]. glycine moiety of GSHI interacts with r3 (instead of r1) while the γ-glutamyl moiety interacts with r1 (instead of r2 or r4) [61].With 5.3.4. Outlook on GPx catalysis and mechanistic questions respect to GPx4, residue r1* is located at an alternative position The high catalytic efficiency of GPx-isoforms usually results in a whereas residue r5 is found instead of r2–r4 (Fig. 12E). Thus, one focus on the activation of the nucleophile. However, the poor leaving − − of the two predicted GSH binding sites could be absent (Fig. 13), group of the first SN2 reaction (RO or OH ) also requires significant and these structural alterations were suggested to be responsible activation, e.g. by simultaneous protonation. A network of hydrogen for the lower specificity of GPx4 for GSH as a reducing agent [60,61]. bonds between the conserved glutamine and asparagine residues Interestingly, the area corresponding to the glycine moiety-binding might fulfill this task [60]. Nevertheless, as partially outlined above, site in the structures of Grx and GST is blocked by additional loops in how the thiol/selenol is really deprotonated and the leaving group mammalian GPx1 but not in GPx4 (please compare the panels in the is protonated still remains to be clarified. The situation gets even middle of Fig. 8 as well as Fig. 9 and Fig. 12) [304]. Thus, if this surface more confusing regarding the end of the oxidative half-reaction and area is utilized by either GSHI or GSHII, GPx1 has to undergo significant the whole reductive half-reaction: (i) The existence of conformation- conformational changes. al changes and/or the formation of a selenylamide obviously require Of note, the positions of the two glutathione sulfur atoms and experimental validation. (ii) Analogous to the situation described the selenium atom in ref. [60] are not in agreement with the SN2 for Grx (Fig. 11), we still do not know which of both glutathione geometry described in Fig. 5A, and an efficient reaction between moieties exactly binds to which protein area. (iii) How are both

GSHII and glutathionylated GPx without structural rearrangements GSH molecules activated as a nucleophile and what is the fate of the is therefore questionable: As the catalytic cycle comprises three con- protons (Fig. 13)? Owing to the position (Fig. 12D,E), and based on secutive SN2 reactions, either (i) the reaction geometry is not in mutational studies [60,314], the conserved glutamine residue might accordance with the chemical principle (for example, owing to the also play a role in deprotonation during the reductive half-reaction usage of other selenium orbitals), (ii) the positions of the five (regardless of the nature of the electron donor that is utilized by the involved atoms (SII–SI–Se/Sa–O–O) adopt a linear orientation; an various GPx-isoforms). (iv) Another important aspect that could be unlikely scenario owing to the limited space at the active site—or addressed in far more detail is the relevance of the active site residue (iii) there is a conformational change as suggested in Fig. 13.As as a leaving group. The advantage of selenocysteine-dependent outlined in the next section, a conformational change (with or with- GPx-isoforms could be that the selenolate is a better leaving group out selenylamide formation) at the end of the oxidative half-reaction than the thiolate group (see also Section 4.4.2 on Grx activity). is in agreement with the kinetic data and could indeed explain Thus, the relevance of this residue could be the second part of why productive GSH binding is rate-limiting [60,300]. Finally, the the reductive half-reaction instead of the oxidative half-reaction. suggested competition of the two γ-glutamyl carboxylate groups of (v) What is the role of the highly conserved cysteine residue in the both glutathione molecules for r4 [60] might explain the efficient FPCN-motif (Fig. 12)? It appears rather unlikely that the residue has release of GSSG at the end of the reaction. been maintained in almost all GPx-isoforms without being somehow 3238 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 relevant. Even though mutational analysis on a T. brucei GPx was studies revealing no time-dependent import of sGPx4 [332]. Neverthe- inconspicuous [314], initial studies on a GPx-isoform from Chinese less, a trapping mechanism in the intermembrane space [330] could cabbage suggested the formation of alternative disulfide bonds theoretically result in the mitochondrial import of sGPx4. The highly [315]. Whether these bonds and the residue have a regulatory, pro- conserved cysteine residue in the FPCN-motif and residue Cys108 of tective or structural function—i.e. for protein trapping in the mito- human GPx4 (in a well conserved IxVNG-motif) are indeed quite chondrial intermembrane space (Section 5.4)—remains to be studied. close to each other, and I hypothesize that they could form a trapped disulfide bond in the oxidizing intermembrane space. Notably, of the 5.4. Physiological and medical relevance of GPx catalysis three yeast GPx-isoforms, only one protein has a cysteine residue in the IxVNG-motif, and only this protein (GPx1) is annotated to Homozygous GPX1 and GPX2 single and double knock-out mice interact with the outer mitochondrial membrane in accordance with are all viable [324,325], presumably due to overlapping functions the hypothesis. Importantly, a (somatic) mitochondrial sGPx4 species with cytosolic Prx-isoforms (Section 6). GPX1 knock-out mice lack a might also explain why the selective knock-out of (testicular) lGPx4 phenotype under normal growth conditions, but are more susceptible had no effect on apoptosis [333], because the final protein in the to oxidative challenges in accordance with a function for peroxide intermembrane space should be identical regardless of the import removal [325]. The GPX1/GPX2 double knock-out mice were shown pathway. Thus, the reported lipid hydroperoxidase-dependent anti- to have inflammatory bowel disease with a high incidence of mucosal apoptotic effects of mitochondrial GPx4 in vitro and in vivo inflammation in the ileum and colon but not in the jejunum [335,339–341] could be achieved by (up-regulation) of sGPx4 and/or [324]. The phenotypes become explainable considering the similar/ lGPx4. complementary activity levels of both enzymes in the mucosal epithelium [326]. Whether mucosal inflammations in the human 6. Peroxiredoxins digestive tract are also correlated with decreased GPx activities needs clarification. In contrast to GPX1 and GPX2, GPX4 is essential, 6.1. Pioneers of Prx catalysis and knock-out mice were reported to die in utero [327]. Which of the different GPx4 forms and functions (Section 5.2) are required As reviewed previously, the first Prx was described in 1968 by Harris for survival? owing to the peculiar ring-shaped quaternary structure of a decamer- In rodents, all different GPx4 forms are encoded by one single forming isoform [61]. Much later, in 1994, Rhee and colleagues showed gene with eight exons. The N-termini of the short form, sGPx4, and for the first time that a yeast Prx-isoform—which was previously of the long form, lGPx4, are encoded by the same exon 1A. The discovered in the Stadtman lab as thiol-specific antioxidant protein expression of both variants probably not only depends on the transla- (Tsa1)—is a Trx-dependent hydroperoxidase [342,343]. Since then, tion initiation efficiency at alternative in-frame start codons, but also Prx have been identified in all domains of life. Their reduction by flavo- on alternative transcription initiation sites in exon 1A [328,329].Asa proteins in bacteria and by Trx or closely related thiol-containing result lGPx4 has an additional N-terminal bipartite presequence proteins in eukaryotes and other bacteria has been studied in much consisting of a positively charged amphipathic helix and a hydro- detail by the groups of Rhee, Poole, Flohé, Karplus and many others phobic sorting signal [329]. Proteins with such signals are usually [61,67,343–347]. In contrast, only little is known about the GSH- inserted into the inner mitochondrial membrane facing the inter- dependent reduction of Prx. membrane space, and the mature protein is often released into the latter compartment after proteolysis of the transmembrane segment 6.2. Structure and function of glutathione-dependent Prx [330]. This import pathway and a localization of GPx4 in the inter- membrane space are in accordance with subfractionation and Based on their primary structure Prx-isoforms can be grouped into in organello import assays using rat mitochondria [331,332]. Of note, five major classes [67,343,346,347]. However, such Prx classifications specific deletion of lGPx4 in mice allowed normal embryogenesis are often complicated by the fact that a few amino acid substitutions and postnatal development, but caused male infertility owing to can completely change the enzymatic mechanism and/or the quater- impaired mitochondrial capsule formation during spermatogenesis nary structure [61,63,345,348] which often ranges from monomeric (Section 5.2) [333]. The same phenotype was observed for condi- to dimeric and (do)decameric protein species (Fig. 14A,B) [67,346]. tional GPX4 knock-out mice [334]. Hence, the structural function of Thus, a mechanistic classification seems to be more appropriate GPx4 provides an explanation for the crucial role of the trace element [61,67,343–345]: All Prx-isoforms have the peroxidatic cysteine resi- selenium for male fertility in mammals [321,333]. In addition to its due Cysp in common (Fig. 8H, Fig. 14). This residue is found in a con- structural role, a peroxidase activity-based anti-apoptotic function served Px3Tx2C-motif and reacts with the hydroperoxide substrate, of lGPx4 has been reported for cell cultures [335] in agreement with resulting in an instable (easily ‘over’-oxidizable) sulfenic acid in ana- the initial observations by Neubert et al. [299] (Section 5.1). However, logy to GPx (Section 5.3). The mechanistic classification of Prx- deregulation of apoptosis was not observed in the lGPx4 knock-out isoforms into 2-Cys and 1-Cys Prx now depends on the fate of the mice under normal growth conditions [333]. sulfenic acid: For 1-Cys Prx only one subunit is mechanistically The nuclear form, nGPx4, from rat testis was reported to be required, and the sulfenic acid is expected to directly react with the chromatin-associated [336]. Accordingly, nGPx4 has an alternative, often unknown reducing substrate. For typical 2-Cys Prx a resolving arginine-rich N-terminus which is encoded by exon 1B (located after cysteine residue (CysR) from a second subunit attacks the sulfenic exon 1A). This form is generated by transcription initiation at an exon acid, and an intermolecular disulfide bond is formed. The disulfide- 1B-specific promoter and not by [328,337].Selective bridged homodimer subsequently reacts with the reducing agent deletion of nGPx4 in mice neither affected viability nor fertility, but the yielding the regenerated enzyme. In contrast, CysR from atypical peroxidase activity was suggested to contribute to chromatin stability 2-Cys Prx forms an intramolecular disulfide bond within the same [338]. The lethal effect of the GPX4 deletion in mice therefore has to subunit before the reaction with the reducing agent [61,67,343–345]. be attributed to the short form sGPx4 [333]. In fact, up-regulation of Functionally, Prx are thiol-containing-reductant:hydroperoxide human sGPx4 in the GPX4 knock-out background rescued the mice— oxidoreductases (EC 1.11.1.15). In vitro hydroperoxidase assays even though male mice were infertile—whereas up-regulation of revealed that some Prx-isoforms do not only accept electrons from a lGPx4 did not reverse the lethal phenotype [339].Interestingly, coupled Trx/TrxR donor system, but also work with a coupled Grx/ a major fraction of sGPx4 was reported to associate with the inner GSH/GR system. Such activities were detected for various Prx- mitochondrial membrane [339], in contrast to in organello import isoforms from plants, algae, cyanobacteria [67,349], Gram-negative M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3239 bacteria including Haemophilus influenzae [350] and E. coli [351], and to follow a ping-pong mechanism analogous to Trx-dependent Prx non-related eukaryotes including yeast [64,66] and P. falciparum catalysis. Ping-pong patterns were indeed reported for recombinant [63,352]. Most of these GSH-dependent enzymes seem to belong to GST/GSH-dependent PrxVI using variable phospholipid hydroperoxide the mechanistically heterogeneous classes of Prx5-type and Prx6- concentrations, however, the essential (?) GST P1-1 was absent in these type isoforms [346] and are found in the cytosol, in mitochondria measurements [355]. [66,67], in plastids [67,119] or in the nucleus [352]. A physiological The oxidative half-reaction of Prx (Fig. 15A) is similar to cysteine- relevance of GSH/Grx-dependent hydroperoxide removal by Prx is containing GPx (Fig. 13), although other amino acids are involved supported by the natural H. influenzae Prx–Grx hybrid protein [61,345]. Formation or stabilization of the thiolate of Cysp was sug- (Fig. 14B) which cannot be reduced by the Trx/TrxR system in vitro gested to depend on two residues that are highly conserved in most [350]. Nevertheless, considering the redundancy and high abundance Prx-isoforms regardless of the mechanistic class [61,345,346]: a thre- of different peroxidase systems in subcellular compartments, the onine (or sometimes serine) residue in the Px3Tx2C-motif at the exact roles of the GSH-dependent pathways remain predominantly N-terminus of helix α1 and an arginine residue at the N-terminus of unclear, in particular, for the Prx-isoforms accepting electrons from strand β6 (Fig. 14C). Please note the similar positions of several active both Grx/GSH and Trx. Interestingly, one of the six mammalian site residues in GPx (Fig. 12C) and Prx (Fig. 14C), pointing to a con- Prx-isoforms, the 1-Cys isoform PrxVI, was shown to reduce phos- vergent evolution (Section 4.2.1). Replacement of the cysteine resi- pholipid hydroperoxides with the help of GSH and the pi class GST due or of the threonine hydroxyl group was demonstrated to P1-1 [65,353–355]. The enzyme was also reported to have a moon- inactivate a variety of Prx-isoforms, whereas replacement of the lighting function as a phospholipase A2 [353]. So far, the combination arginine residue did not always result in a completely inactive en- of these properties is unique, suggesting that they have evolved rath- zyme [61]. Notably, a histidine residue before the proline in the er recently during the speciation in mammals (see also Sections 2.3 Px3Tx2C-motif is relatively conserved in (1-Cys and 2-Cys) Prx6-type and 4.2.1). isoforms and could also interact with the Cysp thiolate (Fig. 14C) and/or protonate the first leaving group [348,356].

6.3. The enzymatic mechanism of glutathione-dependent Prx After formation of the Cysp sulfenic acid and the release of the first product, an alternative conformation was demonstrated for several To the best of my knowledge, a comprehensive kinetic analysis Prx-isoforms (Fig. 15A) [346], including the GSH-dependent Prx– supporting a precise mechanistic model for Grx/GSH-dependent Prx Grx hybrid protein from H. influenzae [357]. Analogous to some has not been published to date. Nonetheless, in the models depicted GPx-isoforms (Section 5.2), helix α1 is partially unwound and the in Fig. 15, the reactions of all GSH-dependent Prx-isoforms are assumed sulfenic acid is exposed in a loop (left subunit in Fig. 14B). However,

Fig. 14. Structures of Prx. Helices and strands are labeled according to Fig. 8G,H. (A) Structure of human PrxVI with an extended β-sheet along the so-called B-type dimer interface formed between helices α4 and α4′. (B) Structure of H. influenzae Prx–Grx hybrid protein with a so-called A-type dimer interface formed between helices α2 and α2′. The orien- tation of the subunit on the right side is similar to the subunit on the left side in panel A. The positions of both subunits are switched owing to the dimer interface on the opposite side. The two other subunits of the tetramer are missing in the PDB file. (C) Zoom in at the active site of PrxVI in its “closed” conformation with the highlighted. A histidine residue proximal to the thiol group of Cysp is predominantly found in Prx6-type isoforms. The orientation is analogous to GPx in Fig. 12C to facilitate comparison. (D) The active site of PrxVI is poorly accessible in the “closed” conformation. The sulfur atom (which is oxidized to a sulfenic acid) points to a small window where the peroxide group can enter. The second subunit is labeled in blue. (E) Comparison of two active sites of crystallized H. influenzae Prx–Grx adopting alternative conformations. Images were generated using Swiss-Pdb viewer and the structures of human PrxVI and H. influenzae hyPrx5 (PDB IDs: 1PRX and 1NM3, respectively [356,357]). See Sections 6.2 and 6.3 for details. 3240 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266

Fig. 15. Models of Prx catalysis. Please note that Prx exist in a monomer–dimer equilibrium or can also form higher oligomers even if only one subunit is shown. (A) Universal ox- idative half-reaction resulting in three different Prxox species depending on the mechanistic class. (B) Potential mechanisms for the reductive half-reaction of GSH/Grx-dependent Prx-isoforms with 1-Cys Prx on top, atypical 2-Cys Prx in the middle, and typical 2-Cys Prx at the bottom. Potential first and second reducing agents are shown below (a–c). A con- formational change of reduced Prx at the end of the reaction and the reduction of the R'SSR disulfides by GSH are omitted. (C) Suggested mechanism for the reductive half-reaction of GSH/GST-dependent PrxVI. A potential shortcut is labeled with a question mark. See Section 6.3 for further details.

in contrast to GPx (Section 5.3), the conformational change is already describe the reductive half-reaction for Grx- and GST-dependent widely accepted to be a general prerequisite for Prx catalysis. This isoforms separately. theory is plausible considering that the active sites of crystallized Prx-isoforms with a “closed” conformation are not accessible for large 6.3.1. The mechanism of Grx-dependent Prx physiological reducing agents (Fig. 14D,E) [346]. Whether the sulfenic During the reductive half-reaction, the electron donor could acid forms a Prx disulfide bond before reduction by the second substrate either directly attack the sulfenic acid, or reduce the disulfide bond depends on the mechanistic class as outlined in Section 6.2 (Fig. 15A). of a typical or atypical 2-Cys Prx (or of a 1-Cys Prx dimer that is

The reduction of the oxidized enzyme (Prxox) seems to differ signifi- disulfide-bridged by two Cysp residues). These theoretical possi- cantly among various GSH-dependent Prx-isoforms, and I will therefore bilities, illustrated in Fig. 15B, are furthermore complicated M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3241 considering two potential reducing agents for Grx/GSH-dependent half-reaction that was later refined (Fig. 15C) [65,353,354]: First, Prx-isoforms: a) A dithiol Grx might simply replace Trx as the elec- a complex of GST P1-1 and GSH replaces a subunit of dimeric tron donor for Prxox. In this mechanistic scenario, GSH is not a Prx PrxVI-SOH, resulting in the formation of a GST/Prx heterodimer. substrate but just reduces Grx(S2) after Prx catalysis. b) Alternatively, Heterodimerization was reported to also occur in the presence of the sulfenic acid/disulfide of Prxox might be attacked by a monothiol S-methylglutathione and therefore seems to be independent of the Grx as a first reductant. The resulting Grx–SS–Prx intermediate could thiol group of GSH. In contrast, in the general absence of glutathione, subsequently react with GSH as the second reductant, yielding heterodimerization was less efficient [354]. During the second step glutathionylated Grx which is regenerated by another GSH molecule. of the reductive half-reaction, GST-bound GSH becomes activated c) Vice versa, the sulfenic acid/disulfide of Prxox might be initially (Section 8.3.1) and reduces the exposed sulfenic acid of PrxVI attacked by GSH as a first reductant, and the (monothiol) Grx could (Fig. 15C). Thus, GST acts as glutathionylating enzyme with its active subsequently deglutathionylate Prx-SSG. Of note, in the latter sce- site facing the sulfenic acid at the heterodimerization interface [354]. nario, Grx can either act as a true third substrate or as an enzyme. Notably, residue Cys47 of GST P1-1 was found to be essential for dimer Which of these alternative mechanisms are supported by experimen- stabilization, and a disulfide-bridged heterodimer was detected by tal data? non-reducing SDS-PAGE. Based on these results, Cys47 was suggested

Rouhier et al. determined the in vitro activity of a Prx5-type to attack Cysp of PrxVI-SSG yielding GSH and a GST–SS–Prx interme- isoform from poplar at variable Grx-concentrations using a set of diate (Fig. 15C). GST P1-1 therefore also provides a resolving residue cysteine mutants. Saturation kinetics and SDS-PAGE analyses sup- and was suggested to act as an unusual deglutathionylating enzyme. app ported a role of Grx as a true substrate with a Km of approx. In accordance with this theory, other GST-isoforms that lack the 2.5 μM. The authors therefore suggested that the enzyme is directly resolving cysteine cannot activate PrxVI [65,354]. In the last part of reduced by Grx either via a monothiol or a dithiol mechanism with the reaction, the GST–SS–Prx intermediate was suggested to react respect to Grx [349]. This mechanism is in accordance with path- with two GSH molecules yielding GSSG, and PrxVI adopts its original ways a) and b) in Fig. 15B. Glutathionylation of Cysp of the same closed conformation (Fig. 15C) [65,353,354]. Prx-isoform was, however, later shown to lead to a dissociation of Please note that the reduction of PrxVI–SSG could be significantly the homodimeric enzyme, and an alternative mechanism was there- simplified/accelerated (Fig. 15): Residue Cys47 of GST P1-1 could direct- fore suggested [358] in accordance with pathway c) in Fig. 15B. More- ly react with the glutathionyl moiety of PrxVI–SSGviaaGrx-like“usual” over, although a second cysteine residue of this enzyme was not deglutathionylation yielding reduced PrxVI. Glutathionylated GST could essential for catalysis (therefore leading to a mechanistic classifica- subsequently react with the second GSH molecule. Thus, the release and tion as a 1-Cys Prx), mutation of this residue significantly reduced binding steps for the first GSH molecule become obsolete (Fig. 15C). the activity [349]. Thus, the true Grx/GSH-dependent mechanism of This variation of the original model by Fisher and colleagues is based the poplar enzyme remains to be unraveled. on the assumption that the proposed GST–SS–Prx intermediate accu-

Steady-state kinetics of the (α2)2 tetrameric H. influenzae Prx–Grx mulates as a thermodynamically stabilized product under oxidative hybrid protein [357]—which is also a Prx5-type isoform [346]— conditions in vitro. The activity of a serine mutant of GST P1-1 has not revealed a hyperbolic activity plot with respect to the peroxide sub- been analyzed in the enzymatic assay because a stable heterodimer strate concentration and a sigmoidal dependency regarding the could not be purified [354], however, this does not exclude that a tran-

GSH concentration [350]. In addition, glutathionylation of Cysp was sient dimer could be active. detected after protein purification from E. coli [350]. Pauwels et al. therefore suggested a mechanism in accordance with pathway c) in 6.3.3. Properties of Prx reaction intermediates in vitro and in vivo

Fig. 15B with pathway b) distracting from catalysis at lower GSH con- Prx generally seem to have an activated Cysp, and a thiol pKa value centrations. Based on glutathionylation studies, pathway c) was also around 6 has been reported for different isoforms [61,345,351]. Thus, favored for the mitochondrial Prx6-type isoform from yeast (Prx1) the reduced enzyme is also deprotonated in vivo. Whether the with either Grx or TrxR as the second reducing agent [64,66]. enzyme stays fully reduced under physiological conditions (as Adefined binding site for Grx or GSH, if there is any, awaits iden- suggested for the highly efficient GPx-isoforms, Section 5.3.3) could tification for most Prx-isoforms. In the crystal structure of the not only depend on the mechanistic class, but—for those Prx that H. influenzae hybrid protein, the Prx domain has a negatively charged are less efficiently reduced—also on the cellular redox state. For ex- surface that is complementary to the positively charged surface on ample, quantitative mass spectrometry indicated that approx. 15% the Grx domain of another (α2) dimer (Fig. 14B) [357]. Kim et al. of the active sites of the cytosolic Prx-isoform Ahp1 from yeast are therefore proposed that such complementary areas could be absent oxidized under steady-state conditions in vivo. This value increased or partially replaced by hydrophobic patches in Trx-specific enzymes. to 50% upon peroxide treatment [142]. Tpx from E. coli was even Furthermore, a molecular model suggested that GSH is bound in a found to be 35% oxidized under steady-state conditions and more cleft between the Grx- and Prx-domain where GSH is predominantly than 60% oxidized upon oxidative challenge [140]. On the one hand, associated with the Grx-domain [357]. Of note, the positively charged these values suggest that the reductive half-reaction becomes rate- residues on the complementary surface of the dithiol Grx-domain limiting and/or that a trapped oxidized species accumulates in vivo.

(Fig. 14B) correspond to residues r1 and r2 in Fig. 9. In summary, the On the other hand, the accumulation of oxidized enzyme species is mechanistic data on Grx/GSH-dependent Prx-isoforms are neither in accordance with a function of Prx as redox sensors [11]. sufficient nor consistent. The suggested catalytic mechanisms in In order to draw further conclusions on reaction intermediates Fig. 15 remain to be tested experimentally and presumably have to and catalytic properties of GSH-dependent Prx in vivo, a comprehen- be determined individually for each enzyme. sive analysis of the kinetics in vitro would be extremely helpful. This

includes the determination of the (apparent) kcat and Km values for all 6.3.2. The mechanism of GST-dependent Prx potential substrates. So far, the characterization of most of the The physiological reducing agent of homodimeric PrxVI was ques- GSH-dependent Prx-isoforms is rather preliminary. The best studied tionable for a long time (and this is still the case for other Prx6-type examples are the H. influenzae Prx–Grx hybrid protein and human enzymes [63,344,348]). However, in 2004, Manevich et al. discovered PrxVI: (i) The H. influenzae Prx–Grx hybrid protein has the advantage a GST P1-1-dependent activity for heterodimeric PrxVI using impure of an immanent fixed Grx concentration, and apparent kcat and kcat/Km enzyme preparation from bovine lung [65]. Based on co-purification, values for H2O2 at a single GSH concentration were determined to −1 6 −1 −1 reconstitution, oligomerization and glutathione-labeling studies, the be 11 s and 5×10 M s , respectively. The apparent kcat and authors proposed an unusual catalytic mechanism for the reductive Km values for GSH at a single tert-butyl hydroperoxide (tBOOH) 3242 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266

−1 app concentration were 9 s and 3 mM, respectively [350].TheKm tissue-specific localization and expression patterns [271]. GSH- value for GSH in vivo could be much lower because of a lower phys- dependent PrxVI was usually among the highly abundant proteins iological concentration of the first substrate (see also Section 3.3.3). and had a (peri-)nuclear localization (i.e. in astroglia cells in the hip- In accordance with the rather high apparent catalytic efficiency, the pocampus and in all kinds of epithelial cells) [271,361]. Human PrxVI enzyme was reported to be of physiological relevance for growth is considered to play a protective role against oxidative membrane under aerobic conditions (Section 6.4). (ii) Km and k1 values of re- damage in vivo. In fact, although the encoding gene was non- combinant PrxVI around 120 μMand5×106 M−1 s−1 were estimat- essential, the lungs of knock-out mice were more susceptible to ed from secondary plots for phospholipid hydroperoxides [355]. hyperoxia-mediated injury, and the enzyme was suggested to be However, because catalysis was analyzed in the absence of GST more important for this organ than the lipid hydroperoxidase GPx4 P1-1, the high activity of the recombinant enzyme in this study re- [353,361]. mains enigmatic. Nevertheless, similar values were later reported Gram-negative H. influenzae—a major cause of pneumonia and in the presence of GST P1-1 [65,354]. non-epidemic meningitis in children—is a glutathione auxotroph

The apparent kcat and kcat/Km values of mitochondrial PrxIIF from pathogen that lacks the canonical bacterial Prx1-type alkyl hydroper- poplar for tBOOH were much lower (0.5 s−1 and 3×104 M−1 s−1, oxide reductase AhpC. Thus, the peroxidase activities of the Prx–Grx respectively) [67]. This was also the case for chloroplast PrxIIE hybrid protein and catalase are expected to play an important role (0.9 s−1 and 1×105 M−1 s−1) [67] and the versatile BCP/PrxQ-type upon infection of the human respiratory tract. In accordance with this −1 3 −1 −1 enzyme from E. coli (0.1 s and 7×10 M s for H2O2) [351]. theory, a knock-out of Prx–Grx was reported to be lethal for stationary Since the apparent values were only determined for the peroxide H. influenzae cell cultures under aerobic (but not under anaerobic) con- substrate—at just one GSH and Grx concentration—the true kcat and ditions [362]. Almost nothing is known about GSH-dependent isoforms the catalytic efficiency could be much higher. Alternatively, a high form other organisms and pathogens, and therefore—also considering enzyme concentration could compensate the lower activity. Several the aspects summarized in Section 6.3.3—it currently does not make Prx are in fact among the most abundant proteins (explaining why sense to further speculate on the physiological relevance of GSH- they often show up in proteomic studies) [67,343,347]. In this scenar- dependent Prx catalysis. io, oxidized Prx might accumulate owing to a rate-limiting reductive half-reaction in accordance with the in vivo results. However, the 7. Glyoxalase 1 and glyoxalase 2 physiological relevance of a GSH-dependent peroxidase activity will remain a matter of debate for most Prx-isoforms as long as the kinetic 7.1. Pioneers of Glo catalysis parameters (and the cellular concentrations) have not been analyzed in more detail. Some Prx-isoforms could, for example, predominantly Exactly 100 years ago, in 1913, Neuberg as well as Dakin and act as redox sensors [11] or could have a moonlighting function as Dudley independently discovered the conversion of MG to lactic chaperones [359,360]. acid in a variety of tissue and cell extracts and introduced the name “glyoxalase” for the unknown catalyst [363,364]. Two decades 6.3.4. Outlook on Prx catalysis and mechanistic questions later, Lohmann as well as Jowett and Quastel discovered that GSH is As outlined in the previous sections, there are far more open ques- a coenzyme of the reaction [365,366]. During that period the Emb- tions than answers regarding GSH-dependent Prx catalysis. (i) Similar den–Meyerhof–Parnas pathway was unraveled (1927–1942). Many sci- to GPx catalysis, the exact fate of the proton of the Cysp thiol group as entists therefore lost their interest in MG and questioned the well as the proton donor for the leaving group of the hydroperoxide physiological relevance of Glo catalysis, and, as a consequence, substrate require further consideration. (ii) Regarding the reductive glyoxalases are still neglected in most textbooks. Based on previous ob- half-reaction, the activation of the GSH thiol group has not been servations by Yamazoye as well as Hopkins and Morgan, Racker sepa- addressed for the Grx-dependent Prx-isoforms. Maybe some Grx rated two yeast proteins in 1951, and showed that Glo1 and Glo2 are also serve as activators that deprotonate GSH for the subsequent reac- both necessary for the conversion of MG to lactic acid [367]. He correct- tion with Prx-SOH in analogy to GST P1-1. (iii) In that context, do GSH ly predicted the activity of Glo2 as a thioesterase and demonstrated the and Grx bind simultaneously or one after the other? Is Grx the first or irreversibility of the overall reaction. Racker furthermore assigned a second reductant, or is it just a (de)glutathionylating enzyme that has condensation reaction to the activity of Glo1, but it took the following to be present in catalytic amounts? Maybe the alternative mechanisms decades and two key publications by Vander Jagt [368,369] to depicted in Fig. 15B are all valid and are used by different subclasses adequately address the one- vs. two-substrate hypotheses and to with so far unknown mechanistic principles and structure–function show that Glo1 is actually an isomerase converting a variety of gluta- relationships. (iv) The question why many Prx6-type isoforms are thione hemithioacetals [73]. Notably, some of the observed discrepan- neither efficiently reduced by the Trx/TrxR electron donor system cies from these decades might become explainable considering nor the GST or Grx/GSH/GR system in vitro also awaits further clarifica- that different classes of Glo1-isoforms from yeast and mammals tion [63,344,348]. (v) Moreover, how does GST P1-1 exactly work? The (Section 7.2.1) were studied, yielding alternative kinetic patterns [370]. rate-limiting step and the reaction between the postulated GST–SS–Prx In 1973, Ekwall and Mannervik confirmed that Glo1 specifically forms intermediate and GSH have not been characterized to date. (vi) Last but S-D-lactoylglutathione from MG [371], and Uotila reported the first thor- not least, what is the role of the reported monomer formation of ough characterization of purified Glo2 [372]. Ten years later, studies glutathionylated Prx-isoforms [358]? Even though the homodimer from the Creighton lab indicated that Glo1 is not stereospecificfor dissociation might be a prerequisite for the interaction with Grx, an its substrate [373]. This theory was later confirmed by Landro, Rae interaction of the dimer cannot be excluded to date. In summary, and co-workers [374,375].Thefirst crystal structures of Glo1 GSH-dependent Prx catalysis is a rather young research field with a and Glo2 were reported by Cameron et al. in 1997 and 1999, respec- great potential for multiple future discoveries. tively [376,377]. Since the initial studies, numerous Glo1- and Glo2-isozymes have been studied from prokaryotes [378] and 6.4. Physiological and medical relevance of Prx catalysis eukaryotes including human [23,73,372,376,377,379,380], S. cerevisiae [379,381–384], A. thaliana [385–388], P. falciparum [31], Onchocerca vol- The physiological relevance of Trx-dependent Prx-isoforms for vulus [389] and kinetoplastid parasites [390–394]. Nevertheless, as I will (potential) antioxidant defense and redox regulation has been outline separately for Glo1 and Glo2 in the following sections, the enzy- reviewed before [11,67,220,343,344,347]. The distribution of the six matic mechanisms and catalytic functions of glyoxalases still have sev- mammalian Prx-isoforms has been recently reported, revealing eral surprises in store. M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3243

7.2. Glo1 catalysis glutathione-binding residues (Fig. 16E). Two essential glutamate resi- dues per active site exert the acid–base catalyzed isomerization of

7.2.1. Structure and function of Glo1 both diastereoisomers to a single thioester (residues GluI and GluII′ in Glo1 (also termed GloI or GlxI) belongs to the superfamily of vicinal Fig. 16D) [370,383,401,407–409]. The substrate seems to be pre- oxygen chelate enzymes, containing an ancient βαβββ-motif required dominantly bound by basic/polar residues via the γ-glutamyl and gly- for metal ion binding (Fig. 16) [395,396]. Different Glo1-isoforms can cine moiety (Fig. 16E) [370,376,378,380], explaining the broad range be roughly subdivided into two major classes: (i) small homodimeric of 2-OA that are efficiently isomerized [73,368,369]. Two additional and (ii) large monomeric enzymes (Fig. 16A–C). (i) The small isoforms residues, usually a histidine and a glutamine, are associated with the

(~20 kDa) are more common and have been structurally characterized metal ion (residues HisM and GlnM′ in Fig. 16D). In bacterial Glo1- e.g. from human [376], E. coli [397] and the kinetoplastid parasite Leish- isoforms and the trypanothione-dependent L. major enzyme, GlnM′ is mania major [390]. Due to an early gene duplication event these isoforms replaced by a second histidine residue, and the overall geometry is al- contain two βαβββ domains per subunit (Fig. 16B) [395,396]. The func- tered owing to the different metal ion [378,390,397]. The following tional dimeric proteins consist of four βαβββ-domains with the two catalytic models—which are predominantly based on structural N-terminal domain of one subunit interacting with the C-terminal do- analyses on human Glo1 [376,380,401]—have been proposed for the main of the other subunit [376,390,397]. Thus, the two structurally iden- proton transfer between atoms C1 and C2 in both diastereomeric sub- tical active sites A and A′ are formed between the swapped domains of strates (Fig. 17). the homodimer (Fig. 16A–C). (ii) Although a crystal structure of a large (i) According to a model that is supported by computations by monomeric Glo1 (~35 kDa) has not been solved yet, the overall architec- Himo and Siegbahn [408],afirst and a second proton transfer ture is probably similar to the homodimeric isoforms (except for a link- can be distinguished (Fig. 17A). After substrate binding, atom er area between the two inner domains) [370,383,398].Monomeric C becomes deprotonated, resulting in a first cis-enediolate. Glo1 arose from a second gene duplication event [398] that might 1 This step requires either Glu (Glu172 in the human enzyme) have occurred independently in non-related organisms such as fungi, I for the (S)-diastereomer or Glu ′ (Glu99′ in human Glo1) for plants and apicomplexan parasites [370].Asconsequence,thetwoac- II the (R)-diastereomer. Reprotonation of the enediolate yields tive sites A and B in monomeric Glo1-isoforms are structurally (and the same cis-enediol intermediate for both diastereomers. functionally) non-identical (Fig. 16C) [370,399,400].Analternative Atom O is then deprotonated by Glu which subsequently classification of Glo1-isoforms is based on the metal ion-dependency: 1 I transfers the proton to atom C of the second cis-enediolate, Some Glo1, including homodimeric human Glo1 as well as the mono- 2 giving the final product (Fig. 17A) [408]. As a result, a single meric enzymes from yeast and P. falciparum,preferoneZn2+ (or (R)-diastereomer is obtained (e.g. S-D-lactoylglutathione from Fe2+) at each active site [379,383,398,401]. In contrast, Glo1 from sev- MG). Please note that residue Glu is required for the turnover eral bacteria such as E. coli [402] and the important pathogens Yersinia I of both diastereomeric substrates. In contrast, residue Glu ′ is re- pestis, Pseudomonas aeruginosa and Neisseria meningitidis [403]—but II quired only for deprotonation of C of the (R)-diastereomer also the protist L. major [390,394]—are optimally activated in the pres- 1 (Fig. 17A) [408]. ence of Ni2+ (or Co2+). (ii) An alternative model for the turnover of the (S)-diastereomer Regarding the enzymatic function, Glo1 is officially classified was proposed by Creighton and Hamilton [407] and computed as a lactoylglutathione (EC 4.4.1.5). However, based on mecha- by Richter and Krauss [409] (Fig. 17B). One difference is that nistic studies [368,369], Glo1 is a metal ion-dependent isomerase residue Glu (partially) dissociates from the metal ion upon the (EC 5) converting various glutathione hemithioacetals to glutathione I initial protonation. In the second step, Glu directly protonates thioesters [23,73,74]. The diastereomeric hemithioacetals are formed I atom C (instead of O ). This step is coupled to a simultaneous pro- during a non-enzymatic reaction between GSH and the unhydrated 2 2 ton transfer from atom O to residue Glu ′ that subsequently pro- 2-OA (Fig. 3E) [23,73,74,368]. Both diastereoisomers are then con- 1 II tonates atom O , yielding the same product. Noteworthy, atom O verted by Glo1 to a single thioester [373–375]. With respect to the 2 2 is permanently associated with Gln ′ in this model [407,409].In subcellular localization, Glo1 is regarded to be a cytosolic protein; a M summary, the biggest differences to the model depicted in plausible scenario considering the major source of MG (glycolysis), Fig. 17A are the early protonation of C by Glu , the interchanged and the absence of targeting sequences in the Glo1-isoforms analyzed 2 I final positions of the protons, and the participation of residue so far [400]. Moreover, traditional Glo1 purification protocols were Glu ′ in the turnover of the (S)-diastereomer (Fig. 17B). performed with erythrocytes which lack membrane-bound organ- II elles [404–406]. However, there are probably exceptions to this rule. Both models are in good agreement with base- and metal-catalyzed For example, Glo1-isoforms from L. major and T. cruzi were suggested chemical studies supporting an cis-enediol(ate) mechanism [73,407] to be dual targeted to the cytosol and to mitochondria [391], and and with site-directed mutagenesis studies [370,383,401,407]. Proton several uncharacterized monomeric isoforms from fungi such as abstraction from atom C1 was suggested to be a rate-limiting step— Ashbya gossypii (gi|45188159), Kluyveromyces lactis (gi|50310681) in contrast to biochemical isotope exchange analyses [73]. These dis- and Verticillium dahliae (gi|83267732) [370] carry N-terminal mito- crepancies could be explained by conformational changes upon cataly- chondrial targeting sequences with high bioinformatic prediction sis. In fact, the absent isotope effects, biphasic kinetic patterns, the scores (Deponte, unpublished). Furthermore, the N-terminus of a inactivation of Glo1-mutants in phosphate buffer as well as proteolytic so far non-functional Glo1-like protein from P. falciparum guided a susceptibility analyses suggest that Glo1 adopts a protected, closed GFP-fusion construct to the apicoplast [119], and the uncharacterized conformation in the absence of substrate and during catalysis, whereas monomeric Glo1-isoform from A. thaliana (gi|21537360) [370] con- a kind of lid opens during substrate binding and product release tains a predicted chloroplast transit peptide (Deponte, unpublished). [73,370,407,410]. Conformational changes and catalysis at the two In some of these cases additional non-cytosolic MG-detoxification active sites of monomeric Glo1 from P. falciparum were furthermore processes cannot be excluded, however, considering the lack of ex- shown to be allosterically coupled [31,370,399,400]. perimental data, the functions of these (insular?) targeted proteins are completely unknown [31]. 7.2.3. Outlook on Glo1 catalysis and mechanistic questions The capacity of Glo1 to convert both diastereomers as substrates 7.2.2. The enzymatic mechanism of Glo1 was estimated to result in a 3- to 6-fold advantage in the steady- The two active sites of homodimeric and monomeric Glo1-isoforms state rate in vivo [411]. Nevertheless, none of the mechanistic models contain highly conserved metal-binding residues (Fig. 16D) as well as sufficiently explains the formation of a single (R)-diastereomeric 3244 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266

Fig. 16. Structure of Glo1. (A) Top view of homodimeric Glo1 along the 2-fold axis. Each quadrant contains one of the four βαββ(α)β domains. The metal ion complex and a glu- tathione transition state analog (GTSA) are highlighted at both active sites. (B) Same view with one subunit omitted. Domains I and II are highlighted and contribute to one half of each active site. (C) Schematic illustration of the protein architectures of small homodimeric and large monomeric Glo1-isoforms. (D) Zoom in at one active site of Glo1. Conserved residues that are important for metal ion coordination and acid–base catalysis are highlighted. (E) Illustration of conserved residues that were demonstrated or suggested to bind 37 101 103 122 the glycine moiety (1) or the γ-glutamyl moiety (2) of glutathione. Residues r1′ (Arg ′), r2′ (Thr ′) and r3′ (Asn ′) are highly conserved, whereas r4 (Arg ) is often replaced by 150 156 Gln. Positively charged residues at position r5 (Lys ) and r6 (Lys ) are also found in most homologues. The images were generated using Swiss-Pdb viewer and the structure of human Glo1 (PDB ID: 1QIN [380]). See Section 7.2.1 for details. product from both diastereomers. It is obvious that the reaction has to metallohydrolases [414,415]. Glo2-isoforms are highly variable with be controlled by the asymmetric protein environment at the active respect to metal ion binding: A variety of isoforms were shown to site [407–409], but how this is exactly driven remains to be shown. contain either two ions or one zinc/iron ion pair per protein This central enigma is of course related to the question which of molecule [377,385,392,416–418]. Other promiscuous recombinant both models in Fig. 17 is correct. Specific isotope labels and NMR- isoforms were reported to either contain a Zn2+/Zn2+ center, a analyses might reveal the fate of the transferred protons for both Fe3+/Zn2+ center, a Fe3+/Fe2+ center or even a Mn2+/Mn2+ center substrates. Another puzzling aspect in Glo1-catalysis is the structural [265,387,413,419]. The composition of the center seems to depend on variability of the enzyme. Why are there monomeric and homo- the medium and growth conditions [413,419]. With respect to the dimeric Glo1-isoforms (Fig. 16C) as well as Zn2+-andNi2+-dependent quaternary structure, human Glo2 [377] and mitochondrial Glx2-5 enzymes? Are both active sites in homodimeric Glo1 really identical from A. thaliana [387] were reported to be monomeric, even though and do they work independently? Several studies indeed suggest that alternative monomer–monomer contact sites were detected in their metal binding to Ni2+-dependent homodimeric Glo1-isoforms is not crystal structures (Fig. 18A–D). A recent study on cytosolic Glo2 identical [378,394,412]. Furthermore, the structure–function relation- from P. falciparum furthermore revealed a monomer–dimer equilibri- ships that determine allosteric effects in monomeric P. falciparum um in solution [400], and Glo2 from human erythrocytes indeed ran Glo1 are unclear and several hypotheses with respect to alternative in one and two bands after denaturing and non-denaturing PAGE, substrates/regulators and metabolic adaptations remain to be addressed respectively [420]. Notably, the monomer–monomer contact site of

[31,370]. crystallized human Glo2 is formed by the C-terminal helix α8 which carries two important glutathione-binding residues (Fig. 18B). In con- 7.3. Glo2 catalysis trast, the active sites of crystallized A. thaliana Glx2–5 are completely blocked (Fig. 18D) [31,377,387]. 7.3.1. Structure and function of Glo2 Regarding the enzymatic function, Glo2 is a thioesterase that cat- Glo2 (also termed GloII or GlxII) is composed of an N-terminal alyzes the hydrolysis of S-(2-hydroxyacyl)glutathione (EC 3.1.2.6). A β-lactamase domain with a conserved zinc/metal-binding motif at variety of other thioesters including S-acylglutathione substrates— the active site and a C-terminal domain with five α-helices but usually no non-glutathione thioesters—are also hydrolyzed in (Fig. 18A) [377,387,393,413]. Owing to the β-lactamase fold the pro- vitro [23,73,74,372,386,420,421]. On the one hand, Glo2 catalyzes tein is a member of the structurally diverse group of binuclear the last step of the glyoxalase pathway, regenerating GSH and M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3245

Fig. 17. Two models of Glo1 catalysis. Depending on the stereochemistry of the substrate, alternative acid–base catalyzed isomerization mechanisms have been suggested. (A) Sym- metrical reaction pathway with a first and second proton transfer for both diastereomeric hemithioacetal substrates. (B) Alternative shortened, non-symmetrical mechanism for the (S)-diastereomer. See Section 7.2.2 for further details.

yielding D-lactic acid from MG in the cytosol (Fig. 3E). On the other 7.3.2. The enzymatic mechanism of Glo2 hand, Glo2-isoforms are not only found in the cytosol. In fact, mito- The active site of Glo2 can be subdivided into (i) the metal binding site chondrial and even plastid Glo2 are either encoded by the same (Fig. 18E) forming the catalytic center, and (ii) the substrate-binding site gene like the cytosolic enzyme (due to an alternative exon usage) (Fig. 18F). (i) As their name implies, binuclear metallohydrolases [422] or by separate nuclear-encoded genes [381,382,386,387,400,423]. bind two metal ions that generate a nucleophile (e.g. an activated water Some of these isoforms seem to act independently from Glo1. This molecule or hydroxide ion) and/or coordinate an oxygen atom of the theory is not only supported by the turnover of alternative substrates hydrolyzable substrate [415,417]. Metal ion 1 is bound by three highly

[23,73,74,372,386,420,421], but also by the existence of insular Glo2 conserved histidine residues (HisM1A,B,C in Fig. 18E), whereas metal ion in mitochondria from rat [424],human[422],yeast[381,382] and 2 is bound by one aspartate (AspM2) and two (HisM2A,B). A. thaliana [386,387]. Another apparently insular Glo2-isoform was Both ions are bridged by an aspartate residue (AspBri)oppositetothe found in the apicoplast from P. falciparum [400] (the unusual Glo1-like substrate-binding site (Fig. 18E) [377]. (ii) The substrate-binding protein in this organelle is a non-functional partner for standard sub- site contains five well conserved residues among diverse eukaryotic strates [31,423]). The parasite T. brucei even lacks Glo1 activity [421] and prokaryotic Glo2-isoforms (r1–r5 in Fig. 18F), indicating that the and has just one (trypanothione-dependent [392]) glyoxalase (GLXII). enzyme predominantly interacts with the glutathione-moiety of the This enzyme was furthermore shown not to contribute to the detoxifica- substrate [377,378,387,413,417] (which explains the efficient turnover tion of MG [421]; a plausible result considering that these parasites could of a variety of glutathione substrates [23,73,74,372,386,420,421]). be protected from MG owing to membranous organelles (glycosomes) Residues r2 and r3 are aromatic, with the hydroxyl group of residue r3 where a major part of glycolysis takes place [425].Insummary,Glo2 pointing at the sulfur atom of the substrate (Fig. 18F). Accordingly, app are not only a part of the , but can also exist as insular mutation of r3 was shown to increases the Km value [423,426].Resi- isoforms with individual, yet unknown, physiological functions. dues r1,r4 and r5 are basic and interact with the carboxylate groups of 3246 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266

Fig. 18. Structures of Glo2. (A) Top view of two subunits of human Glo2 in the asymmetric crystal unit. The metal ion complex and GSH or a glutathione transition state analog

(GTSA) are highlighted at both active sites. The N-terminal β-lactamase domain and the C-terminal α-helical domain are circled. (B) Front view of human Glo2 along helix α8 with subunit interactions highlighted in the middle. Residues Arg249 and Lys252 bind the glycine-moiety of glutathione on the other side of the helix. (C) Structure of dimeric Glo2–5 from A. thaliana with the left subunit in the same orientation as in panel A and the second subunit in front. (D) Top view on Glo2–5 revealing the proximity of the metal ion complex to the dimer interface (left side). Accordingly, both active sites are hidden and blocked in the crystallized dimer by the subunit interface (highlighted on the right 54 56 110 side). (E) Zoom in at the highly conserved catalytic center. Metal ion 1 is bound by the three histidines HisM1A,B,C (His , His and His in human Glo2). Metal ion 2 is bound 58 59 173 134 by AspM2 and residues HisM2A,B (Asp , His and His ). Both ions are bridged by AspBri (Asp ). Substrate binding at the back side is indicated by an arrow. (F) Zoom in at the substrate-binding site. Residues that were demonstrated or suggested to interact with the glycine moiety (1), the γ-glutamyl moiety (2) or the sulfur atom of the substrate 143 145 175 are highlighted. A positively charged residue at position r1 (Lys ), an aromatic amino acid at position r2 (Tyr ), a tyrosine residue at position r3 (Tyr ) and the 249 252 glycine-binding residues r4 (Arg ) and r5 (Lys ) are highly conserved. The catalytic center is below the GTSA molecule. The images were generated using Swiss-Pdb viewer and the structures of human Glo2 and Glo2–5 from A. thaliana (PDB IDs: 1QH5 and 1XM8, respectively [377,387]). See Section 7.3.1 for details.

glutathione (Fig. 18F). The glycine-interacting residues r4 and r5 are part contrast, GSH is presumably still associated with metal ion 2 and rath- of helix α8 described above (Fig. 18B) [378,380] and seem to form the er tightly bound via the glycine carboxylate group when the next most important part of the substrate-binding site (Fig. 18F) [417]. water molecule enters the active site at metal ion 1 (Fig. 19) [417]. Glo2 works via acid–base catalysis (Fig. 19). Owing to the special Product inhibition patterns (obtained for cytosolic Glo2 from kinetics of the enzyme, this property usually seems to be masked P. falciparum [417] and Glx2-2 from A. thaliana [418]) and viscosity

[417]. Of note, the metal ion center was suggested to lower the pKa effects (observed for human Glo2 [428]) support the proposed value of water by almost ten orders of magnitude to approx. 6 “hit-and-run” Theorell-Chance bi-bi mechanism with OH− as the [417]. Thus, the much more potent nucleophile OH− can be generated first true substrate and an extremely instable ternary complex under physiological conditions (Fig. 19). As soon as the product GSH between the enzyme, OH− and the thioester [417]. In addition, rate- from the previous catalytic cycle is replaced by the next thioester limiting replacement of GSH by the next thioester substrate perfectly app substrate, the nucleophilic attack takes place and a tetrahedral transi- explains not only the pH- and salt-sensitivity of the kcat value of Glo2, tion state is formed [417,427]. Metal ion 1 probably stabilizes the but also why the metal-ion composition has a rather moderate effect oxyanion of the transition state, whereas metal ion 2 presumably as- on the catalytic efficiency as long as the nucleophile is still formed sists the removal of the thiolate leaving group [377,417,427]. The [417]. In summary, Glo2 exerts a nucleophilic substitution via an addi- equilibrium of the hydrolysis is clearly on the product side, and the tion–elimination mechanism with metal/base-catalyzed nucleophile loosely associated acid is quickly released as the first product. In formation and presumably acid-catalyzed GSH formation. M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3247

7.3.3. Outlook on Glo2 catalysis and mechanistic questions and a protective function for cytosolic Glo2 was detected upon Whether a flexible quaternary structure is a common property of MG-induced cell death [435]. different Glo2-isoforms and has an influence on catalysis remains to The cytotoxic effects of exogenous 2-OA on tumors were demon- be studied. Depending on the subunit interface, a transient dimerization strated several times, culminating in the suggestion by Szent-Györgyi of Glo2-isozymes could either directly block the enzyme or regulate the and colleagues to exploit these compounds as cancerostatic agents in activity via helix α8 (Fig. 18A–D) [399,400]. Furthermore, the predicted the 1960s [30]. In 1969, Vince and Ward then proposed to directly in- base accepting the proton from the metal–water complex is so far hibit Glo1 to cause a build-up of endogenous MG in order to kill cancer unknown (Fig. 19). Residue AspM2 (maybe with the help of HisM2B) cells [436]. This strategy was later adopted for pathogens with high might be a good candidate (Fig. 18E). Regarding the leaving group it is glycolytic fluxes including malaria parasites [31,399,400,417,423] and unclear whether protonation occurs simultaneously with the break- kinetoplastid parasites [390,391,393,394]. Alternatively, inhibition of down of the transition state (Fig. 19), or whether GS− is rather stabi- Glo1-dependent osteoclastogenesis could have implications for patho- lized by metal ion 2. Further questions will presumably arise once the physiological situations such as osteoporosis or rheumatoid arthritis physiological substrates of insular Glo2 are identified. [434]. In fact, cell permeable and tight binding inhibitors of mammalian and parasite glyoxalases were demonstrated to be active in vitro [30,31,377,380,390,391,399,400,407,434,437]. Some of the anti-tumor 7.4. Physiological and medical relevance of Glo catalysis inhibitors also gave first promising results in vivo [30,407]. In addition, arecentstudyonLeishmania donovani insect stages (promastigotes) As reviewed before, traditional and current theories suggest that confirmed GLO1 to be essential for parasite survival [438], providing a glyoxalase catalysis is required for 2-OA detoxification and regulatory genetic proof of principle in contrast to the discouraging results from processes with (potential) implications for cancer, diabetes, aging the related parasite T. brucei [421] (Section 7.3.1). and infectious diseases [23,29–32]. In yeast, none of the three In summary, depending on the type of organism, glyoxalases act as a glyoxalase-encoding genes is essential, and growth phenotypes only detoxification system and/or were suggested to be involved in 2-OA- or became obvious upon challenge with exogenous MG [382,429]. How- thioester-dependent regulatory processes and signal transduction. ever, MG was suggested to modify the regulatory cysteine residues of Moreover, some Glo2-isoforms presumably exert alternative unknown Yap1, resulting in a nuclear localization of this transcription factor physiological functions. Although glyoxalases are non-essential in E. coli and an altered profile. Accordingly, in the absence and yeast, several studies suggest that Glo1 of human and selected of yeast Glo1, MG levels were elevated and Yap1 was constitutively parasites can be of medical relevance and might be exploited as drug activated in the nucleus [430]. The genes encoding Glo1 and Glo2 in target. The exact knowledge of glyoxalase properties and catalysis E. coli are also non-essential, and growth phenotypes were inconspic- might therefore be helpful for rational drug development. uous in the absence of exogenous MG [33,431]. Up-regulation of Glo2, however, surprisingly decreased the growth rate. This phenotype was 8. Glutathione transferases and MAPEG suggested to be caused by the depletion of S-D-lactoylglutathione, resulting in a decreased KefGB-mediated potassium ion efflux [33]. 8.1. Pioneers of GST and MAPEG catalysis To the best of my knowledge, glyoxalase knock-out studies in ver- tebrates have not been reported to date. Nevertheless, overexpression Studies on soluble and microsomal enzymes from thousands of of GLO1 in diabetic rats reduced the levels of hyperglycemia-induced rodent livers laid the groundwork of modern GST and MAPEG ROS markers and AGEs (Fig. 2B,D) [432]. In agreement with these research. The first GST activities were reported in 1961 for soluble results, overexpression and RNAi studies in C. elegans pointed to a enzyme preparations from rat liver with bromosulfophthalein by pro-survival function of Glo1 owing to 2-OA detoxification [433].A Combes and Stakelum and for chloronitrobenzenes by Booth et al. role of Glo1 activity for osteoclastogenesis was revealed for cultured [439]. During the following years it became obvious that there are bone marrow-derived macrophages from mice [434]. Further cell several GSH-dependent liver enzymes with overlapping conjugation culture experiments with human MCF7 and RKO cell lines indicated activities for a variety of substrates. As a consequence, the scientific an up-regulation of GLO2 by the transcription factors p63 and p73, community failed to unambiguously correlate a specific enzymatic

Fig. 19. Model of Glo2 catalysis. The position of the nucleophile-generating metal ion center corresponds to Fig. 18E. Acid–base catalysis is highlighted in red. B, unidentified base; GTE, glutathione thioester. See Section 7.3.2 for details. 3248 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 activity with a distinct soluble GST-isoform. Habig, Pabst and Jakoby 1997 [468], and MGST1 was also the first protein of the MAPEG super- therefore chose in 1974 to assign letters to the rat liver transferases family for which the detailed structure was published in 2006 [469]. based upon their order of elution from carboxymethylcellulose Structures of LTC4 followed one year later [470,471], and, in 2008, [440]. In 1976, Prohaska and Ganther reported a GSH-dependent the Hebert lab also reported the structure of human microsomal peroxidase activity with organic hydroperoxides for a soluble GST- PTGES1 [472]. isoform [441], and distinct physiological GST-dependent isomerase or oxidoreductase activities were afterwards identified or assigned, 8.2. Structure and function of GST and MAPEG i.e. by Benson et al. in 1977 [442] and by Christ-Hazelhof and Nugteren in 1979 [443]. In the next sections, I will discriminate between three protein A GST activity in mitochondria was reported in 1979 [444], and groups: (i) canonical soluble GST, (ii) distantly related soluble Kraus purified the first mitochondrial GST from rat liver in 1980 kappa class GST and (iii) hydrophobic MAPEG. Bacterial [445]. It took one more decade until Harris et al. unambiguously resistance proteins with GST activities are members of the superfam- showed that a soluble matrix GST exists [446] and five more years ily of vicinal oxygen chelate enzymes (Section 7.2.1) [77,395,448] and until the corresponding gene was cloned, and the novel kappa class will not be discussed. Depending on the organism, canonical soluble was introduced [447]. According to Allocati et al., the first evidence GST are cytosolic proteins, eukaryotic soluble kappa class GST are mi- for a GST activity in bacteria was reported for E. coli by Shishido in tochondrial and/or peroxisomal proteins, and hydrophobic MAPEG 1981 [448], but it was not until the late 1980s that soluble bacterial are found in microsomal fractions. Notably, there are exceptions to isoforms were purified and characterized [449,450]. During the these rules, for example, several canonical soluble GST-isoforms same time, studies on Schistosoma japonicum GST by Smith and of plants have mitochondrial, chloroplast or peroxisome targeting Johnson lead to a highly versatile tool in modern cell biology: GST- sequences and/or were found in these organelles or in the nucleus tagging and affinity purification [451]. Around the early 1980s, [473,474]. Mannervik and colleagues not only established the purification of human pi class GST from different tissues [69], but also demonstrated 8.2.1. Structure of GST and MAPEG that GST-isoforms within the same class exist as homo- and hetero- What are the structural differences between GST and MAPEG, and dimers, resulting in modular enzyme activities [69,452]. Based how conserved are their primary structures throughout evolution? on these findings, GST-isoforms are nowadays named according to Based on sequence similarities, Mannervik and colleagues introduced their class (e.g. GST A for alpha class) and their subunit composition a nomenclature for canonical soluble GST-isoforms in 1992, resulting (e.g. GST A1-2 for a heterodimer between two subunits encoded by in seven GST classes in mammals (alpha, mu, pi, theta, zeta, omega GSTA1 and GSTA2) [439]. In 1991, the Huber lab solved the first crystal and sigma) [439]. In addition, mammals possess a dual-targeted solu- structure of a mammalian pi class GST [453], directly followed by the ble GST-isoform of the kappa class [475] and six proteins of the structures of a mu class GST by Ji et al. [454] and a recombinant alpha non-related hydrophobic MAPEG family [75–77,466]. class GST by Sinning et al. [455]. The first structure of a kappa class GST was reported by Armstrong and colleagues in 2004, confirming (i) The combination of alpha, mu and pi class GST is restricted previous reports that these enzymes differ significantly from the to vertebrates (single class exceptions are found in parasites, other soluble GST-isoforms owing to an alternative protein fold presumably due to ). In contrast, the [456]. Today, several hundred GST structures have been deposited zeta and theta classes are conserved among eukaryotes and are in the , indicating that the crystallization of soluble also found in bacteria [77,448,474,476]. Nevertheless, owing to GST-isoforms is often feasible. the similarity-based nomenclature and the variable C-terminal In 1977, Ogino et al. partially purified the, to my knowledge, first domain of canonical GST [476], it is not surprising that the classi- GSH-dependent microsomal protein of the MAPEG superfamily fication has sometimes a rather limited value with respect (Section 2.2.5): labile microsomal prostaglandin-E synthase from to the mechanism and function. Since the introduction of bovine vesicular glands (PTGES1, also termed MPGES1) [457].A the classification, numerous additional GST classes have been microsomal GST activity from rat liver was first reported two years found—and are constantly identified—in non-related organisms earlier for the xenobiotic substrate benzo[a]pyrene-4,5-epoxide including plants (phi, tau, lambda and DHAR classes) [474], by Nemoto and Gelboin [458]. In 1979 and the following years, protists [477,478] and prokaryotes (beta and other classes) Morgenstern et al. as well as Oesch and colleagues demonstrated [77,448,479,480]. Of note, slight alterations at the active site that this enzymatic activity was not caused by soluble contaminants with significant functional consequences were even detected and purified the GST from the endoplasmic reticulum [459,460]. for highly similar members of the same GST class as exemplified The primary sequence of this integral membrane protein, termed mi- in the following sections. crosomal GST1 (MGST1), was reported in 1985 and revealed no sim- The N-terminal domain of the canonical soluble GST-isoforms ilarity with soluble GST-isoforms [461]. In the same year, based on is rather conserved, whereas the C-terminal α-helical domain previous reports by Samuelsson as well as Murphy and colleagues, is extremely variable (Fig. 20A) [476]. The glutathione-binding Yoshimoto et al. partially purified and analyzed a GSH-dependent mi- site, the so-called G-site, is accordingly formed by the conserved crosomal leukotriene-C4 synthase (LTC4) [462]. The corresponding N-terminal thioredoxin fold (Fig. 8C,D and Fig. 20A,B). In human gene was cloned in 1994 [463,464]. Its sequence revealed a contrast, the modular (hydrophobic) binding site for the electro- homology to the integral membrane protein “5-lipoxygenase activat- philic substrate, the so-called H-site, highly depends on the ing protein” (FLAP) which was discovered four years earlier by Dixon α-helical domain and is located at the domain interface et al. [465]. Notably, the activator FLAP on its own seems to neither (Fig. 20A) [71,453–455,481,482]. The GST active site residue is have a GSH-dependent nor another enzymatic activity [76]. Based either a tyrosine residue (Tyra, found in alpha, mu, pi and on the homology of LTC4 and FLAP, Jakobsson et al. discovered two sigma classes), a serine (Sera, found in theta, zeta and phi classes) additional microsomal proteins with GST and peroxidase activities, or a cysteine (Cysa, in beta, omega, lambda and plant DHAR MGST2 and MGST3, and introduced in 1999 the term MAPEG for a classes) [448,474,476]. In most cases Tyra adopts the position of novel superfamily consisting of six different human membrane pro- the basic glutathione-binding residue r1 of Grx (Fig. 8A–D), teins [466]. In the same year, twenty-two years after Ogino's discov- whereas Sera or Cysa in GST-isoforms corresponds to Cysa of ery, Jakobsson et al. also cloned human microsomal PTGES1 [467]. Grx (see also Section 4.4.2). With respect to the quaternary struc- The projection structure of MGST1 was solved by Hebert et al. in ture, members of the soluble GST classes are usually dimeric. M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3249

Exceptions are monomeric isoforms of the lambda and plant the glutathionyl-moiety (not a sulfur atom) is transferred [439]. DHAR classes [473] as well as the canonical P. falciparum GST Furthermore, not all GST and MAPEG act as transferases (EC 2), but which forms tetramers that dissociate upon GSH binding (also) add glutathione to epoxides (EC 4) or catalyze disulfide and [71,477]. peroxide reductions (EC 1) and isomerizations (EC 5). Several theta (ii) In contrast to canonical soluble GST-isoforms, the α-helical and zeta class GST were, for example, reported to have very low domain in kappa class GST is not fused to the C-terminus activity with CDNB and to weakly bind to glutathione affinity of the thioredoxin fold (Fig. 8C), but is inserted between matrices [476,492]. A conserved physiological function of zeta class

helix α2 and strand β3 [456,483]. The architecture of kappa GST is in fact the cis–trans isomerization of 4-maleylacetoacetate to class GST therefore shares similarities with DsbA in the bacte- 4-fumarylacetoacetate (EC 5.2.1.2) in the course of phenylalanine/ rial periplasm [19]. Moreover, significant similarities to bacte- tyrosine degradation [493]. Some alpha class GST isomerize the rial 2-hydroxychromene-2-carboxylate isomerases were found position of a carbon–carbon double bond in selected 3-oxo-Δ5- [77,475,483]. Armstrong and colleagues therefore suggested a steroids yielding 3-oxo-Δ4-steroids (EC 5.3.3.1) [442,494]. The isom-

DsbA-like progenitor for this enzyme class [456]. The subcellular erases prostaglandin-D2 synthase (EC 5.3.99.2) and prostaglandin-E2 localization in the mitochondrial matrix also supports an endo- synthases (EC 5.3.99.3) catalyze another intramolecular redox reac-

symbiotic/bacterial origin of the GST kappa class in eukaryotes. tion: the cleavage of the endoperoxide in prostaglandin H2 yielding Considering the active site, the CPYC-motif of Grx is replaced a hydroxyl and keto group in prostaglandin D2 or E2. Please note by a SPYS-motif in kappa class GST and the first residue (Sera) that prostaglandin-D2 synthase is a soluble GSH-dependent sigma − was shown to be crucial for GS formation [456]. class GST [443,495], whereas prostaglandin-E2 synthases include a (iii) The sequence similarities between the six mammalian MAPEG soluble GST, the MAPEG member PTGES1 and the non-related micro- classes are quite low [77]. Nevertheless, all MAPEG share a sim- somal peripheral membrane protein PTGES2 [457,472,496]. The

ilar (predicted) trimeric structure with four transmembrane he- MAPEG enzyme leukotriene-C4 synthase catalyzes a specialized reverse lices for each subunit (Fig. 20C) [76,468,470–472]. (Please note lyase reaction: the addition of GSH to the epoxide leukotriene A4 (EC that the term “membrane-associated” in MAPEG is somehow 4.4.1.20) [462–464]. In summary, among all families of glutathione- misleading because all members are in fact integral membrane dependent enzymes, GST and MAPEG are the most versatile catalysts proteins.) Mammalian MAPEG can be roughly subdivided into converting a plethora of sulfur-, oxygen- or carbon-containing electro- the MGST1/PTGES1 group and the MGST2/MGST3/LTC4/FLAP philic substances. group [75,77]. Further MAPEG families are found in insects and bacteria, but not in archaea [75]. Except for FLAP, all mammalian 8.3. The enzymatic mechanism of GST and MAPEG MAPEG possess conserved glutathione-binding residues: a first

arginine residue in helix α1 (r3′), an R-x-Q/N-x-N-x2-E-motif in Owing to the versatility of GST and MAPEG, conjugation, reduction helix α2, and a tyrosine residue (r2)inhelixα3 (Fig. 20D) [76]. and isomerization mechanisms will be discussed separately in the Residues R, Q/N and E in the motif correspond to r1,r5′ and r4,re- following two sections. Moreover, the mechanisms will be further spectively (Fig. 20D). A conserved third arginine residue (Arga) categorized based on the aromaticity and the hybridization of the at the N-terminus of helix α4 seems to generate or stabilize electrophilic center of the substrate. GS− in LTC4 [470,471,484–486] and in microsomal PTGES1 [472]. Based on the structure of MGST1, this function was also 8.3.1. The mechanism of GST-catalyzed conjugations and reductions

suggested for a fourth conserved arginine residue in helix α3 The traditional model substrate for the spectrophotometric analysis [469].Arga and the fourth arginine indeed sandwich the glutathi- of a GST conjugation activity is CDNB because of its high reactivity with one sulfur atom in crystallized PTGES1 [472], and the residue in many GST-isoforms and high extinction coefficient of 9.8 mM−1 cm−1

helix α3 was demonstrated to be highly important for MAPEG ca- at 340 nm [440]. However, numerous other (artificial) aromatic sub- talysis [472,487]. Nevertheless, a structural role has been favored strates, as well as non-aromatic substrates such as haloalkanes are and assigned to it (see also below) [76,487]. Interestingly, as also used [3,440]. As outlined below, the mechanism for aromatic and nicely illustrated by Prage et al. [488], the positions of the non-aromatic substrate moieties has to differ significantly. Further- glutathione-binding site in MAPEG as well as the conformation more, a comparative interpretation of GST steady-state kinetics leading of the bound substrate differ significantly in the crystallized pro- to a coherent mechanistic model seems almost impossible owing to teins: In MGST1 the GSH-binding site is located in a solvent ex- controversial data interpretations and the usage of alternative sub- posed part of the protein and GSH adopts a C-shaped strates and/or GST classes. Even identical enzymes were analyzed in conformation [469], whereas in MPEGS1 and LTC4 the active numerous different ways, i.e. considering or disregarding (putative) site is more buried in the membrane area and GSH adopts a additional ligand/substrate-binding sites and alternative enzyme U-shaped conformation [470–472]. The (hydrophobic) sub- conformations. strates of MAPEG were suggested to bind along a cleft that is For example, a study on rat mu class GST M1-1 in 1974 revealed bi-

formed between helices α1 and α4′ of adjacent subunits facing phasic kinetics, and Pabst et al. suggested a complex hybrid ping-pong/ the membrane (Fig. 20D) [470–472]. This theory is supported sequential mechanism with alternative substrate concentration- by MPEGS1 inhibitor studies [488–490],althoughalternative dependent reaction pathways [497]. One part of the mechanism was binding modes were also suggested for MPEGS1 [491] and in accordance with a previous study suggesting an ordered bi-bi mech- MGST1 [487]. In summary, soluble GST and hydrophobic anism with GSH and the electrophile as the first and second substrate, MAPEG are non-related proteins. Members of both groups have respectively, and the leaving group and the glutathione conjugate as highly variable primary structures, whereas variations at the ac- the first and second product, respectively [497,498]. In contrast, later tive site seem to be rather limited. experiments on rat mu class GST M1-1 [499,500] and GST M2-2 [501] as well as alpha class GST A1-1 [502] challenged these models and 8.2.2. Function of GST and MAPEG were interpreted according to a sequential random mechanism. In a Functionally, GST were classified as RX:glutathione R- recent study on mouse GST P1-1, mutations of cysteine residues (EC 2.5.1.18). R stands for the electrophilic group, including aliphatic, unmasked a GSH-dependent positive cooperativity. One of the muta- aromatic or heterocyclic compounds, and X stands for the leaving tions also switched the patterns with CDNB from sequential to apparent group such as sulfate, nitrile or halide ions (Fig. 3F). Please note that ping-pong kinetics [503]. (A similar mechanistic switch was observed the former name “glutathione S-transferase” is misleading because for different mutants of the GST-like yeast prion protein Ure2 having a 3250 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266

Fig. 20. Structures of GST and MAPEG. (A) Top view of homodimeric GST along the 2-fold axis. The bound inhibitor S-hexylglutathione (GS-C6) is highlighted at both subunits. The N-terminal thioredoxin fold and the C-terminal α-helical domain of the subunit on the left side are boxed. The orientation of the subunit on the right side is similar to Fig. 8D. The active site residue Tyra at the end of strand β1 is labeled. (B) Zoom in at the active site of GST. Residues that were demonstrated or suggested to bind the glycine moiety (1) or the

γ-glutamyl moiety (2) are highlighted. Residue Tyra is involved in deprotonation of the thiol group of GSH and is essential for catalysis. Positively charged residues at position r1 15 49 71 (Lys ) and r4 (Lys ) are conserved in several but not all GST classes, whereas residues interacting with the positively charged amino group of GSH at position r2 (Gln ) and po- 105 sition r3′ (Asp ′) are highly conserved. The binding site for the electrophilic substrate differs significantly between the GST-subclasses as far as the composition, size and acces- sibility are concerned. In the shown example, the so-called H-site is solvent accessible and hydrophobic owing to residues h1–h3. (C) Top view of homotrimeric . The four transmembrane helices are highlighted for one subunit. The three active sites with bound GSH are circled and are partially hidden by a lid from the neighboring subunit. (D) Zoom in at the active site of leukotriene C4 synthase. Left panel: Residues that were demonstrated or suggested to bind the glycine moiety (1) or the γ-glutamyl moiety 51 97 (2) are highlighted. Residue Arga was suggested to deprotonate the thiol group of GSH. The glycine moiety of GSH is bound by residues r1 (Arg ) and r2 (Tyr ). The carboxylate 30 group of the γ-glutamyl moiety is bound to r3′ (Arg ′) from a neighboring subunit, whereas both subunits interact with the positively charged amino group of GSH via residues r4 58 53 (Glu ) and r5′ (Gln '). The front part of the lid, formed by the second subunit in blue, is omitted for clarity. Right panel: The thiol(ate) group of bound glutathione points through a narrow window along a hydrophobic cleft that is formed between two subunits. The leukotriene substrate is thought to bind along this cleft. The images were generated using Swiss-Pdb viewer and the structures of P. falciparum GST and human leukotriene C4 synthase (PDB IDs: 2AAW and 2PNO, respectively [71,470]). See Sections 8.2.1 and 8.3 for details.

restored GST activity [504]. Ure2 is therefore another intriguing exam- substitution (SN2Ar) with an addition–elimination mechanism ple for the molecular evolution of glutathione-dependent enzymes is very likely (Fig. 21B). After binding of GSH to the G-site, a (po- outlined in Sections 2.3 and 4.2.1.). Since none of the cysteine residues tential) network of hydrogen bonds [505], including Tyra/Sera/ of GST P1-1 can be found in the proximity of the substrates or in direct Cysa or other residues (Section 8.3.2), lowers the pKa of the contact with the neighboring subunit in the crystallized enzyme [503], GSH thiol group to a value around 5.2-6.8 [269,501,506–511].Al- the mechanisms behind the mutational effects remain to be unraveled. ternatively, deprotonation might be facilitated upon binding of However, McManus et al. correctly emphasized that not two but three the aromatic substrate and formation of the ternary complex products are formed in the CDNB reaction—a proton, a chloride anion, [512]. The order of substrate binding, the deprotonation of GSH, and 1-(S-glutathionyl)-2,4-dinitrobenzene (GSDNB)—and therefore and a subunit cooperativity might furthermore depend on suggested that the stability of the enzyme–GS− complex and an (ir)re- whether (the C-terminus of) the protein undergoes significant versible proton release might be responsible for the kinetic patterns structural rearrangements [71,137,502,507,512–517].Assoon

[503]. In summary, several aspects of the mechanisms in the next para- as atom C1 of CDNB adopts an orientation that allows the nucle- graph are rather preliminary and could highly depend on the inves- ophilic attack of the thiolate, the conjugation occurs. The tigated system. sp2-hybridization at the electrophilic center and the aromaticity are lost, and a negatively charged σ-complex intermediate is

(i) Mechanism for aromatic substrate moieties: The SN2 mecha- formed (Fig. 21B). This step was suggested to be rate-limiting

nism delineated in Fig. 5 is not possible for aromatic substrates for rat GST M2-2 [501]. Some of the SN2Ar reaction intermediates due to steric hindrance. The negative inductive and mesomeric (so-called Meisenheimer complexes) are rather stable and can effects of the substituents make CDNB an electron deficient ar- be characterized [518]. There are even GST crystal structures re- omatic compound (Fig. 21A). Thus, a nucleophilic aromatic sembling this state [481,512]. The chloride anion in CDNB is an M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3251

excellent leaving group, yielding GSDNB with a restored aromat- was indeed reported for human omega class GST O1-1 and bacte- ic system (Fig. 21B). The equilibrium of the elimination is on the rial beta class GST B1-1 [269,519], and ping-pong kinetics in the product side. Please note that owing to the different hybridiza- HEDS assay were demonstrated for GST B1-1 [269]. In contrast, tion states of the σ complex and GSDNB, either the glutathione peroxidase assays with GST-like Ure2—which lacks a cysteine res- moiety or the aromatic moiety of the product has to alter the po- idue at the same position [520]—revealed sequential peroxidase sition at the active site upon elimination (Fig. 21B). Crystal struc- kinetics [521] as expected. tures support the latter scenario [481]. In the last two steps of the (iii) Mechanism for non-aromatic substrate moieties with sp2- reaction cycle, the chloride anion and GSDNB leave the active hybridized electrophilic reaction centers: Two commonly used site. The last step was suggested to be rate-limiting for GST I substrates are trans-4-hydroxy-2-nonenal, a physiological product from maize [507]. from lipid peroxidation, and the diuretic drug ethacrynic acid [3]. (ii) Mechanism for non-aromatic substrate moieties with sp3- Non-aromatic 4-hydroxy-2-nonenal and aromatic ethacrynic hybridized electrophilic reaction centers: The conjugation, acid both have an α,β-unsaturated carbonyl moiety in common. thiolysis or reduction of non-aromatic substrate moieties depends Thus, they are excellent Michael acceptors, and the nucleo- on the GST-isoform and on whether a sp3-orsp2-hybridized philic addition of glutathione to these substrates is therefore electrophilic reaction center is attacked. Substrates with sp3- a special case of a Michael addition (Fig. 21F). Since the reac- hybridized reaction centers include, for example, haloalkanes, hy- tion might theoretically follow a variety of pathways, we will droperoxides (such as hydrogen peroxide) and disulfides (such as omit the substrate binding steps and directly start with the

HEDS). The reactions can, in principle, occur via a SN2 mechanism ternary complex between GST and the substrates. (One with a carbon, oxygen or sulfur atom as the electrophilic center as study suggested a rapid equilibrium random bi-bi mechanism outlined in Fig. 5 and Section 2.3.3. Please note that—owing to the with ethacrynic acid and GSH [509], but neither specified the geometry of the transition states—the overall orientation of nature of the second product nor discussed the possibility of GST-bound GS−, the electrophilic substrate and the leaving an ordered bi-uni mechanism.)

group have to be altered as compared to SN2Ar reactions (please Depending on the investigated enzyme, either the active site resi- compare panels B and C–EinFig. 21). Site-directed mutagenesis due Tyr9 in human GST P1-1 or Tyr9 and proximal Arg15 in human studies on GST from P. falciparum—having a tyrosine residue at GST A4-4 were suggested to activate the Michael donor GSH theactivesite—suggested that the GSH-dependent cleavage of [509,522,523]. After nucleophilic attack of the thiolate, the first peroxides requires the same residues at the G-site like CDNB turn- transition state and the enolate intermediate are probably stabi- over [71].Thus,GS− is probably bound in the same orientation for lized by Tyr108 (Fig. 21F). This residue is located at the end of the

SN2Ar and SN2 reactions, but the positions of the electrophilic cen- first helix in the α-helical domain and was reported to interact ter and of the leaving group are significantly altered (Fig. 21B–E). with the keto group of ethacrynic acid in crystallized GST P1-1 The leaving group of haloalkanes is usually unproblematic [524]. Accordingly, mutation to phenylalanine significantly re- (Fig. 21C), whereas the anionic leaving group of hydroperoxides duced the turnover of ethacrynic acid but not of CDNB [509] (see requires protonation by an unknown source (Fig. 21D). Moreover, alsotheroleofahistidineresidueforGSTZ1-1,Section 8.3.2). Res- the turnover of hydrogen peroxide and disulfidesrequiresasec- idue Tyr115 in rat GST M1-1 [482] and residue Tyr212 close to the

ond SN2 reaction/reduction in contrast to haloalkanes (Fig. 21D,E). C-terminus of GST A4-4 [522] seem to play a similar role for the Thefateoftheglutathionesulfenicacid(GSOH)orofGSSRseems turnover of epoxides and 4-hydroxy-2-nonenal. Subsequent pro- 15 to be rather ill-defined, and similar questions as for the Grx- and tonation of atom Cα of the substrate could depend on an Arg - GPx-catalyzed reactions outlined in Sections 4.4.2 and 5.3.2 arise activated water molecule [522] or on an unidentified acid (Fig. 21D,E). As an instable sulfenic acid, GSOH might either (Fig. 21F). The latter step is stereoselective as revealed by the anal- a) react while it is still bound to the enzyme, or b) just after it ysis of the reverse reaction of another Michael acceptor [525]. has left the active site. In the former scenario, a second GSH Finally, kinetic studies on human GST P1-1 suggested that the re- molecule would have to enter the (hydrophobic) H-site but lease of the conjugated product is not rate-limiting [509]. might regenerate the unknown proton donor. In the latter scenar- In summary, the mechanisms of GST-catalyzed conjugations and io, a soluble base catalyst is required. c) Alternatively, GSOH could reductions are extremely variable and highly depend on the GST be replaced at the G-site by an incoming GSH molecule ensuring class and on the substrate. I grouped the mechanisms based on the the efficient formation of the second nucleophile (Fig. 21D). This aromaticity and the hybridization of the electrophilic center of the mechanism somehow resembles the GST P1-1-dependent reduc- substrate. This classification might be helpful to clearly discriminate tion of PrxVI-SOH (Fig. 15C, Section 6.3.2). The situation gets even between GST-catalyzed S 2 , single and double S 2 reactions as more complicated for alkylhydroperoxide substrates such as N Ar N well as nucleophilic (Michael) additions. tBOOH and cumene hydroperoxide. Which oxygen atom is initial- ly attacked by GS−? Is GSOH formed for all alkylhydroperoxides 8.3.2. The mechanism of GST-catalyzed isomerizations or can sterical constraints and leaving group properties result The following GST-catalyzed isomerase reactions can be grouped in the formation of GSOR (see also Section 8.3.2)? Regarding the into (i) carbon–carbon double bond shifts, (ii) other intramolecular fate of GSSR (Fig. 21E), scenarios a) and c) are much more likely redox reactions and (iii) cis–trans isomerizations. than scenario b) owing to the inefficient formation of GS- at physiological pH and the lower reactivity of disulfides as com- (i) Some mammalian alpha class GST isoforms—such as human GST pared to sulfenic acids. A3-3, but not human GST A2-2—possess a significant Δ5-Δ4 Notably, for cysteine-containing omega, beta, lambda and DHAR isomerase activity with selected ketosteroid substrates including class GST-isoforms the mechanism can be different. While the the testosterone and progesterone precursors Δ5-androstene- 5 SN2Ar reaction with aromatic substrates could follow the pathway 3,17-dione and Δ -pregnene-3,17-dione, respectively (EC in Fig. 21B, the mechanisms with disulfide and hydroperoxide 5.3.3.1) [442,494,506].EventhoughtheΔ5-Δ4 isomerase substrates could be similar to Grx and GPx (Figs. 11 and 13), activity can be monitored spectrophotometrically at 248 nm including covalently modified active site cysteine residues that in vitro [442,494,506], there are, to my knowledge, no detailed are subsequently regenerated by GSH. Some GST are therefore reports on the kinetic patterns or on product inhibition studies. considered to be hybrid forms or evolutionary intermediates Thus, the presented mechanism is predominantly based on [269,519]. Glutathionylation of the active site cysteine residue crystal structures [526,527], activity measurements with wild 3252 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266

Fig. 21. Models of GST-catalyzed conjugations and reductions. (A) Structure and properties of CDNB, the aromatic model substrate for GST catalysis. (B) GST-catalyzed SN2Ar mech- anism with CDNB. (C) GST-catalyzed single SN2 mechanism with iodomethane. (D) and (E) GST-catalyzed (double) SN2 mechanism with hydrogen peroxide and disulfide sub- strates. (F) GST-catalyzed Michael additions. See Section 8.3.1 for details.

type and mutant enzymes [494,506] and a recent computational and Δ5-androstene-3,17-dione and omit the unknown steps study [528]. concerning substrate binding and product release (Fig. 22A). Mo- Since the reaction might theoretically follow a variety of lecular modeling and co-crystallization experiments of GST A3-3 pathways—including a random or an ordered mechanism—we with GSH and the product Δ4-androstene-3,17-dione showed will directly start with the ternary complex between GST, GSH that the conserved active site residue Tyr9 and the thiolate M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3253

group of glutathione face the substrate atoms C4 and C6 at the β mammalian enzyme were shown to be essential for catalysis 14 side of the steroid (the same side where the methyl groups at [534]. Residue r1 (Arg ) interacts in mammalian GST S1-1 with 2+ C10 and C13 are located) [526,527]. GSH is bound like in other Mg which is bound at the dimer interface [535]. This alpha class GST-isoforms [526,527] with a Km value of approx. ion-binding site is neither found in other GST classes nor in 0.1 mM [494]. The steroid is bound via the conserved residue helminth sigma class GST. Ion binding was demonstrated to app h1—adopting a different conformation than in Fig. 20B and in decrease the Km value for GSH, presumably owing to a GST A2-2—and additional hydrophobic (aromatic) residues reorientation of r1 from the glutathione carboxylate group to- including Phe222 at the C-terminus [526,527]. The alterations at wards the thiol group [535]. According to a postulated model of − theH-siteseemtoberesponsiblefortheactivity,becausecorre- catalysis [495],GS attacks the peroxide at atom O11 yielding a sponding modifications of the H-site of GST A2-2 turned the en- GSOR intermediate (Fig. 22B). In this model, the leaving group

zyme into an isomerase [510], therefore revealing another (atom O9) of the intermediate was not protonated which seems example for the principles outlined in Sections 2.3 and 4.2.1. ratherunlikely.Furthermore,Kanaokaetal.suggested“aGS− In contrast to all reactions described so far, GS− in GST A3-3 was in solution” as a base for subsequent proton abstraction from

suggested not to form a covalent bond with the substrate, but just atom C11 [495]. However, this scenario is also questionable at to play the role of a Brønsted base removing Hβ from the steroid physiological pH, unless there is an activation site for the second (Fig. 22A) [506,527]. (This interpretation is in agreement with GSH molecule. I therefore suggest a modified mechanism with a a GST tautomerase activity for 2-hydroxymenthofuran [529].) putative proton donor (Tyr8 or Arg14) and a putative base 14 8 The catalytic abstraction of Hβ from atom C4 by the thiolate (deprotonated Arg or Tyr ): The proton donor could first stabi- 9 and the protonation of C6 by Tyr might either occur concerted lize the GSOR intermediate and subsequently generate the base or sequential. A sequential reaction would result in a resonance- (Fig. 22B). The base abstracts the proton from atom C11,resulting stabilized dienolate intermediate. However, human GST A3-3 in the cleavage of the GSOR intermediate (see also the mecha- seems to lack residues for stabilizing such an intermediate after nism for PTGES1 described in Section 8.3.3). In the absence of deprotonation [527], in contrast to the non-related bacterial biochemical evidence for GSOR formation and reliable kinetic ketosteroid isomerases [530]. Thus, the postulated dienolate data, completely different mechanisms are of course possible: [506,526] is presumably not formed during acid–base catalysis, For example, using the isomerization in Fig. 22Aasatemplate − and the concerted mechanism is favored for GST A3-3 (Fig. 22A) mechanism, GS could first abstract the proton at atom C11 [527,528]. (As a consequence, the roles of the tyrosine residue at followed by the cleavage of the endoperoxide without GSOR for- the active sites of bacterial and human ketosteroid isomerases mation. Obviously, more wet lab data is required.

are completely different.) The pKa value of the thiol group of (iii) Evolutionary conserved zeta class GST (e.g. mammalian GST Z1-1) GST-bound GSH was estimated to be 6, supporting its role as a function as maleylacetoacetate isomerase [493] (Section 8.2.2)

base [506]. Surprisingly, the pKa value of GSH bound to GST and, in some bacteria, as maleylpyruvate isomerase. The en- A3-3 and GST A1-1 was predominantly lowered by Arg15,not zymes also possess a GSH-dependent peroxidase activity Tyr9 [506,531], exemplifying another mechanistic alteration of and a conjugase activity with α-haloacid substrates, but not some alpha class GST. Spectrophotometric titrations furthermore with CDNB [448,492,536]. The activity with haloalkanes was 9 revealed a pKa of 7.9 for the side chain of Tyr .Thevaluewas presumably altered and optimized resulting in a bacterial reported to increase to approx. 9 in the presence of GSH [506] in GST sub-class termed TCHQ-DH (tetrachlorohydroquinone accordance with a proton transfer from GSH to Tyr9, regenerating dehalogenases) [448,537].Thecis–trans isomerase function the acid at the active site of the isomerase (Fig. 22A) [528]. of GST Z1-1 is illustrated in Fig. 22C. Mutation of Tyr9 in GST A3-3 to phenylalanine did not com- Which amino acids of zeta class GST are responsible for the pecu- pletely inactivate the enzyme, and a residual activity was even liar isomerase activity? The G-site contains an YxRSSC-motif with

observed in the apparent absence of GSH. Johansson and the tyrosine at the C-terminus of strand β1,averyshortloop,and Mannervik therefore suggested that a hydroxide ion and a the two residues SC at the N-terminus of helix α1(Fig. 8C). The 9 water molecule could replace the thiolate and residue Tyr ,re- first serine residue (Sera) was reported to be absolutely essential spectively [506].Notably,theinvitroisomerizationof13-cis- for the isomerization activities of human GST Z1-1 [536] and a bac- retinoic acid to all-trans-retinoic acids by human pi class GST terial zeta class GST [538], whereas a plant mutant had a low resid- P1-1 was suggested to be even GSH-independent. Inhibition ual activity [539]. The tyrosine, the arginine and the SC-motif are studies and additional controls clearly revealed that the isomeri- not strictly conserved in zeta class GST-isoforms. Nevertheless, zation occurs at the classical active site [532].Thus,eithercataly- mutagenesis studies revealed that the latter residues as well as tic trace amounts of GSH were “contaminants” in the assay, or an arginine residue in the α-helical domain (Arg175)arealsoim- the enzyme works via an alternative isomerization mechanism portant for catalysis (i.e. by mediating substrate binding) using an unidentified base instead of GS−.Tothebestofmy [536,538,539]. In a study on bacterial TCHQ-DH, the cysteine resi- knowledge, this question is still unresolved and might have im- due was even reported to be essential for the isomerase activity plications not only for GST P1-1 but also for GST A3-3 and (and the following model might therefore not apply to other isomerases: Since GSH is not consumed during the isomer- TCHQ-DH) [537]. Moreover, a histidine at the end of the first ization (and therefore acts as a coenzyme and not as a true sub- helix of the α-helical domain was suggested to facilitate the de- strate), a periodic binding and release of GSH appears to be protonation of GSH in a subgroup of zeta class GST lacking the cys- unnecessary, and an uni-uni isomerase mechanism cannot be ex- teine residue [540] (this residue appears to be similar to Tyr108 of cluded for selected GST-isoforms. GST P1-1, Section 8.3.1). In summary, the catalytic mechanism of

(ii) The soluble hematopoietic prostaglandin-D2 synthase is a sigma zetaclassGST-isoformsispresumablynotstrictlyconserved.The class GST (GST S1-1) with a versatile repertoire of species- active site composition of these enzymes may reflect a kind of mo- specific conjugase and reductase activities [533]. The conserved lecular Swiss army knife customized for the catalytic repertoire.

GSH-dependent isomerase function for prostaglandin H2 is illus- The following mechanism is based on a pioneer study in 1979 by trated in Fig. 22B. The H-site of GST S1-1 is extended, resulting Seltzer and Lin (Fig. 22C) [541]: Once GSH is deprotonated and in a unique deep cleft where prostaglandin was suggested to the maleyl moiety is correctly bound, a nucleophilic addition of - bind with its peroxide group pointing at the thiolate of glutathi- GS at atom C2 of the substrate takes place. Since the double one [495].ResiduesTyra and r1 (Fig. 20B) at the G-site of the bond at C2 is conjugated with the at atom C6, 3254 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266

Fig. 22. Models of GST-catalyzed isomerizations. (A) Δ5-Δ4 steroid isomerization catalyzed by GST A3-3 and selected other GST-isoforms. (B) Intramolecular redox reaction of pros- taglandin H2 catalyzed by GST S1-1 and other sigma class GST. (C) Evolutionary conserved cis–trans isomerization catalyzed by zeta class GST. See Section 8.3.2 for details.

this part of the reaction is similar to the Michael addition in 8.3.3. The enzymatic mechanism of MAPEG Fig. 21F. The resulting dienediol(ate) intermediate can now ro- The activities and mechanisms of MAPEG can be classified analo-

tate around the C2\C3 single bond. Assuming rather strong asso- gously to the GST activities described above: (i) conjugations and re- ciations with glutathione and the consensus maleyl carboxylate ductions on the one hand, and (ii) isomerizations on the other hand. group, and a weaker association with the acetoacetate or pyru- Although the protein architecture and catalytic residues of MAPEG vate moiety, a loss or alteration of bonds with the latter group are completely different, the principles and limitations described for

has to be postulated. Accordingly, if the intermediate is protonat- GST are also valid for MAPEG. For example, a SN2Ar reaction requires ed and deprotonated at atom O7, two different residues would alternative positions for the substrate, the σ-complex intermediate presumably become necessary for acid–base catalysis. Alterna- and the products during catalysis (Fig. 21B), regardless whether tively, a dienediolate intermediate could be stabilized at different the electrophile is bound at an H-site in GST or between two of the

positions by alternative positively charged residues; a plausible three subunits in MAPEG (Fig. 20). Except for (artificial) SN2Ar reac- scenario considering several structurally conserved basic resi- tions, kinetic data on MAPEG catalysis are rather limited which is dues at the active site (e.g. Arg13 and Arg175 in human GST not surprising considering the availability, stability and other proper- Z1-1). Another possibility is that the keto instead of the enol tau- ties of the physiological eicosanoid substrates. The corresponding tomer is the true substrate of the isomerase. This substrate has a mechanisms are therefore predominantly based on crystal structures 3 sp -hybridized atom C5, resulting in a higher flexibility that and activity measurements with wild type and mutant enzymes. might allow a permanent association of both electrophile car- boxylate groups. In fact, such a mechanistic model has been pro- (i) MGST1-3 and presumably other MAPEG catalyze conjugations posed by Marsh et al. based on previous studies on TCHQ-DH and (hydro)peroxide reductions [458–460,466,542,543] that

[540]. However, the keto tautomer theory is in contrast to the probably follow SN2Ar or SN2 mechanisms analogous to few available kinetic data [541]. In the last part of the reaction Fig. 21B-D (with Arga instead of a tyrosine residue as illustrat- glutathione is eliminated again, yielding the trans-isomer as a ed in Fig. 20D). The groups by Morgenstern and Armstrong product. revealed that GSH binds very rapidly to free MGST1, whereas thiolate formation occurs quite slowly. A slow conformational In summary, alternative mechanisms for three different types of change resulting in tight glutathione binding and deproton- GST-catalyzed isomerizations have been suggested: one depends on ation was therefore suggested to be rate-limiting for the turn- acid–base catalysis and two depend on a conjugation-elimination over of reactive substrates such as CDNB [543,544]. Theories on mechanism. Accordingly, GS− could either play a role as a Brønstead structural dynamics of MAPEG including alternative “closed” base or form a putative GSOR or a Michael adduct intermediate. The and “open” conformations are indeed supported by a variety latter two mechanisms are preliminary and require much more of (indirect) results on MGST1 [487,488,490,545] and PTGES1 experimental evidence. [472,491,546], and only one of the three MGST1 subunits at a M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3255

110 time has a high affinity for GSH and efficiently forms the thiolate of the conserved arginine residue at helix α3 (Arg )were [543,545]. In addition to this mechanistic model of extreme neg- both detrimental for catalysis (Fig. 23B) [472,548]. According to ative subunit cooperativity (one-third-of-the-sites-reactivity), a preliminary model of PTGES1 catalysis, the thiolate attacks

the enzyme was shown to be activated upon covalent modifica- the endoperoxide at atom O9 once the ternary complex is formed tion of the only cysteine residue [542]. This residue is located in a and the substrates adopt the correct orientation (Fig. 23B). The

solvent exposed loop after helix α1 [469] covering the potential reaction is therefore different from the nucleophilic attack at GSH entry site. A preliminary model of MGST1 catalysis can be atom O11 in the mechanistic model for soluble hematopoietic summarized as follows: GSH binds with a low affinity to all prostaglandin-D2 synthase (Fig. 22B) [495],eventhough three subunits, but a conformational change at only one of the GSOR intermediates are suggested in both models (see also coupled active sites results in tight GSH binding and deproton- Section 8.3.2). Please note that the cyclopentane rings in the per- 129 ation that is presumably mediated by Arga (Arg )and—based oxide substrate, in the tension-free intermediate and in the on site-directed mutagenesis studies [487]—maybe Arg113 (the cyclopentanone product have completely different stable confor-

fourth conserved arginine at helix α3, Section 8.2.1). A SN2Ar or mations which have to be accommodated at the active site. Arga SN2 reaction takes place as soon as the nucleophile is formed in PTGES1 was suggested to stabilize and protonate the leaving 31 and the electrophile is bound in a correct orientation. Product re- group of the SN2 reaction (analogous to Arg ′ in LTC4), and to lease—for example of GSDNB for CDNB, or of GSOH and ROH for a abstract the proton from atom C9 resulting in the cleavage of hydroperoxide substrate—is presumably coupled to another con- GSOR (Fig. 23B) [472].Hence,Arga would act as a base for GSH, formational change [545]. as an acid for O11 and as a base for C9. These are obviously a lot of functions for a single residue, in particular, considering the

geometric limitations and the range of putative pKa values of The conversions of lipophilic epoxides—such as benzo[a] the groups involved. Regarding the role of tyrosines in GST catal- pyrene-4,5-epoxide, catalyzed by MGST1, or leukotriene A , catalyzed 4 ysis (Figs. 21 and 22) and the results from site-directed mutagen- by LTC4—presumably also occur via a (distorted) S 2 mechanism. N esis for PTGES1 [472,546], the functions of Arg110,Tyr117 and However, the leaving group remains attached and becomes protonat- maybe Tyr130 might have been underestimated so far. ed (Fig. 23A). Please note that these conjugation reactions are not sensu stricto nucleophilic additions, because the electrophilic center 8.3.4. Outlook on GST and MAPEG catalysis and mechanistic questions is sp3- and not sp2-hybridized (like, for example, in Fig. 21F). In con- Regarding GST- and MAPEG-catalyzed conjugations and reduc- trast to the slow activation of GSH in MGST1, the LTC4-catalyzed tions, the following aspects remain to be studied in further detail. thiolate formation is rapid and not rate-limiting [484]. The thiol pK a (i) Is there a general order of substrate-binding and product-release value of LTC4-bound GSH is roughly 6, and the thiol proton was steps depending on the GST or MAPEG class? How are the kinetic suggested to be directly released into the solvent (contrary to patterns of GST and MAPEG exactly affected by structural para- MGST1) [484]. In agreement with predictions from LTC4 crystal meters and how do they determine the rate-limiting step? (ii) A structures [470,471],residueArg of LTC4 (Arg104)wasshownto a comprehensive evaluation of the numerous mutational analyses play a central role for thiolate stabilization (Fig. 23A), although mu- [71,282,504–507,509–511,517,522,531] could furthermore reveal tants still had a residual activity at a higher pH [485,486].Notably, how the (potential) hydrogen-bonded network exactly works in the the mutation of the conserved fourth arginine residue in helix α 3 different GST classes. Moreover, when does the deprotonation of also significantly reduced the activity [486].Regardingtheelectro- GSH in GST exactly occur, is the proton transferred to a specific base philic substrate, LTC4 was reported to have a high specificity for leu- − (Fig. 21), and when and how is the proton released? Regarding GS kotriene A4 [547]. The responsible structure–function relationships formation in MAPEG, I got the impression that Arga and the fourth are unknown. Once the ternary complex is formed and both sub- conserved arginine residue in helix α3 might both stabilize the strates adopt the correct orientation, the thiolate attacks atom C 6 thiolate. The latter residue seems to approach the thiol group only of leukotriene A , and residue Arg31′ (next to r ′)stabilizesandpre- 4 3 in the closed enzyme conformation found for crystallized PTGES1 sumably protonates the leaving group at C (Fig. 23A) [470].The 5 [472]. The additional interaction might explain the more efficient de- stereospecificity of the reaction—with an inversion of the chirality protonation of GSH after a (slow) conformational change in the resulting in the (R)-configuration of leukotriene C at atom C —is 4 6 MGST1–GSH complex [543,545]. (iii) Another unsolved question is in accordance with this theory. Moreover, point mutations of the fate of GSOH after the GST- or MAPEG-dependent reduction of Arg31′ significantly reduced the enzymatic activity but had only peroxides (Fig. 21D). The same holds true for GST disulfide substrates minor effects on the Kapp value for GSH [485,486].Incontrastto m and the fate of GSSR. Whether MAPEG can efficiently reduce MGST1, each of the three active sites of LTC4 is functional, and a disulfides at all has, to my knowledge, not been thoroughly studied. cooperativity seems to be absent, at least with respect to GS− for- (iv) Furthermore, MGST1 was reported to have no significant activity mation [484]. with the Michael acceptor ethacrynic acid [542], and it might be interesting to test whether this is a general feature of MAPEG consid- (ii) The only MAPEG-catalyzed isomerization that has, to my knowl- ering that the Michael acceptor 15-deoxy-Δ12,14-prostaglandin J2 edge, been studied in detail, is the PTGES1-dependent intramolec- can act as a suicide inhibitor of PTGES1 and not as a substrate [490].

ular redox reaction of prostaglandin H2 yielding prostaglandin E2 (v) In the same line of thought, is there a correlation between the

(Fig. 23B). Similar to MGST1, only one of three PTGES1 reaction loop connecting helices α1 and α2, the cysteine content in MGST1 and centers was reported to adopt an active conformation in accor- PTGES1, and the apparently absent negative subunit cooperativity in dance with an extreme negative subunit cooperativity (one- LTC4 (having neither the loop nor the cysteine residue)? According to third-of-the-sites-reactivity) [546]. Furthermore, PTGES1 has a this hypothesis, MGST2 and MGST3 should also have no one-third- cysteine residue in the same loop like MGST1, and covalent mod- of-the-sites-reactivity. ification of this residue by 15-deoxy-Δ12,14-prostaglandin J2 was There are also several aspects that remain to be addressed with shown to inhibit the enzyme in vitro [490]. respect to the GST- and MAPEG-catalyzed isomerizations. (i) The In the presence of S-methylglutathione, PTGES1 was shown to be overall mechanism often seems to be rather unclear. Mutagenesis inactive, supporting the relevance of the GSH thiol group for ca- and product inhibition experiments would be helpful to address the talysis [548]. In agreement with the crystal structure of PTGES1 substrate-binding and the product-release steps. Most important, 126 in its closed conformation [472], mutations of Arga (Arg )and which step is rate-limiting, and does GSH stay as a coenzyme at the 3256 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266

Fig. 23. Preliminary models of MAPEG catalysis. (A) LTC4-catalyzed conjugation of GSH and leukotriene A4. (B) PTGES1-catalyzed intramolecular redox reaction of prostaglandin H2. See Section 8.3.3 for details.

active sites of PTGES1 and some GST isomerases during consecutive 9. Concluding remarks catalytic cycles? (ii) What is the nucleophile in GST isomerases that apparently work without GSH [532]? Which structure–function rela- There are at least five non-related protein folds that have been tionships and evolutionary scenarios are responsible for this unusual optimized for glutathione binding and catalysis in the course of evo- property? (iii) Are the fates of the protons in the predicted mecha- lution: (i) the fold of a superfamily of flavin-dependent oxidoreduc- nism in Fig. 22A correct [527,528]? Furthermore, the role of Arg15, tases in GR, (ii) the thioredoxin fold (sometimes in combination which is not conserved among different GST classes, is in my opinion with other domains), for example, in Grx, PDI, GST, GPx and Prx, not fully resolved. Its side chain is located in the proximity of the (iii) the βαβββ-motif of vicinal oxygen chelate enzymes in Glo1 thiolate group above ring A of the substrate [526,527] facing the and fosfomycin resistance proteins A and B, (iv) the combination of exit of the substrate tunnel. It might therefore play a role beyond an α-helical domain with a β-lactamase fold in Glo2, and (v) the nucleophile formation/stabilization, for example, by supporting con- four-helix bundle in MAPEG. I tried to outline and, if possible, to com- certed acid–base catalysis via hydrogen bonding and/or by lowering pare the mechanisms of most of these enzyme classes and subclasses. the energy of the transition state. All other presented isomerization Furthermore, I emphasized several principles and open questions in mechanisms are even less clear. Many more kinetic studies are neces- glutathione catalysis regarding nucleophile activation, electrophile sary to complement the numerous structures. In summary, GST and properties, geometric constraints and leaving group properties. MAPEG have enough tricks up their sleeves to keep a whole new Homologous glutathione-dependent enzymes from different organ- generation of enzymologists busy. isms or cellular compartments can obviously exert alternative functions and/or employ significantly different mechanisms. Such functions and mechanisms often cannot be predicted from in silico analyses. Thus, po- 8.4. Physiological and medical relevance of GST and MAPEG catalysis tential substrates as well as rate constants have to be analyzed in detail in vitro. This aspect is often neglected or difficult to address, particularly The physiological relevance of GST and MAPEG catalysis ranges for complex multi-component assay systems. In addition, much more in from ubiquitous catabolism (e.g. zeta class GST), eicosanoid metabo- vivo data is required to figure out whether a predicted function or en- lism (MAPEG and sigma class GST) and xenobiotic detoxification zyme species is of any physiological relevance. For example, once the ki- (e.g. liver GST and MAPEG) to the susceptibility towards herbicides, netic parameters are known, it is highly beneficial to also determine the antibiotics or host/pathogen factors (e.g. GST-isoforms from plants, physiological concentrations of the enzyme and the substrate(s). In bacteria and parasites) [3,448,474,477,478]. Obviously, these roles summary, as long as we do not have both—sufficient in vitro and affect essential aspects of modern life, including agriculture, biotech- in vivo data—the physiological function and relevance of many nology and water quality as well as fundamental medical aspects such GSH-dependent enzymes remain nebulous. as pain, inflammation, tumor resistance mechanisms and general Several of the presented mechanistic models and of my comments, pharmacokinetics. It is nowadays sexy to hype holistic concepts. For suggestions and criticism will probably turn out to be wrong, however, example, “personalized medicine” promises to be the next gold rush my aim is not to divulge imperfect points of view, but to revitalize the for pharmaceutical industry, and “systems biology” even goes one more and more neglected research field of enzymology. Although I got step further trying to quantitatively model the whole metabolic the impression that many colleagues are still deeply interested in the network and its regulation and effects on an organismic level. How- mechanistic and evolutionary puzzles of glutathione catalysis, funding ever, considering the complexity of GST-isoforms, allelic variants agencies and advisory boards nowadays seem to focus on holistic ap- and overlapping activities on the one hand, and the knowledge (or proaches such as systems biology. I hope this review serves as a moti- complete lack thereof) regarding the enzymatic mechanisms, kinetic vation not only to connect the numerous fields of glutathione catalysis, constants and quantitative effects of GST point mutations on the but also to not forget about the importance and beauty of mechanistic other hand, I am pretty skeptical about whether such concepts details. Finally, I want to apologize to those colleagues whose contribu- make sense at all. This brings me to the final remarks. tions I might have misquoted or did not cite. M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3257

Acknowledgements [33] E. Ozyamak, S.S. Black, C.A. Walker, M.J. Maclean, W. Bartlett, S. Miller, I.R. Booth, The critical role of S-lactoylglutathione formation during methylglyoxal detoxi- fication in Escherichia coli, Mol. Microbiol. 78 (2010) 1577–1590. I thank Tobias Dick, Heiner Schirmer, Luise Krauth-Siegel and the nu- [34] M. Deponte, S. Urig, L.D. Arscott, K. Fritz-Wolf, R. Reau, C. Herold-Mende, S. merous other external and internal participants of the Heidelberg “Redox Koncarevic, M. Meyer, E. Davioud-Charvet, D.P. Ballou, C.H. Williams Jr., K. ” Becker, Mechanistic studies on a novel, highly potent gold-phosphole inhibitor Biology Seminar for their instructive presentations and discussions. of human glutathione reductase, J. Biol. Chem. 280 (2005) 20628–20637. [35] P.A. Karplus, G.E. Schulz, Substrate binding and catalysis by glutathione reduc- tase as derived from refined enzyme: substrate crystal structures at 2 A resolu- tion, J. Mol. Biol. 210 (1989) 163–180. References [36] G.E. Schulz, R.H. Schirmer, W. Sachsenheimer, E.F. Pai, The structure of the flavoenzyme glutathione reductase, Nature 273 (1978) 120–124. [1] R.D. Simoni, R.L. Hill, M. Vaughan, The discovery of glutathione by F. Gowland [37] S.M. Kanzok, A. Fechner, H. Bauer, J.K. Ulschmid, H.M. Muller, J. Botella-Munoz, S. Hopkins and the beginning of biochemistry at Cambridge University, J. Biol. Schneuwly, R. Schirmer, K. Becker, Substitution of the thioredoxin system for Chem. 277 (2002) e13. glutathione reductase in Drosophila melanogaster, Science 291 (2001) 643–646. [2] B.N. Ames, M. Profet, L.S. Gold, Nature's chemicals and synthetic chemicals: [38] S.M. Kanzok, R.H. Schirmer, I. Turbachova, R. Iozef, K. Becker, The thioredoxin comparative toxicology, Proc. Natl. Acad. Sci. U. S. A. 87 (1990) 7782–7786. system of the malaria parasite Plasmodium falciparum. Glutathione reduction [3] J.D. Hayes, J.U. Flanagan, I.R. Jowsey, Glutathione transferases, Annu. Rev. revisited, J. Biol. Chem. 275 (2000) 40180–40186. Pharmacol. Toxicol. 45 (2005) 51–88. [39] R. Pastrana-Mena, R.R. Dinglasan, B. Franke-Fayard, J. Vega-Rodriguez, M. [4] A. Bekker, H.D. Holland, P.L. Wang, D. Rumble III, H.J. Stein, J.L. Hannah, L.L. Fuentes-Caraballo, A. Baerga-Ortiz, I. Coppens, M. Jacobs-Lorena, C.J. Janse, A.E. Coetzee, N.J. Beukes, Dating the rise of atmospheric oxygen, Nature 427 Serrano, Glutathione reductase-null malaria parasites have normal blood stage (2004) 117–120. growth but arrest during development in the mosquito, J. Biol. Chem. 285 [5] R. Frei, C. Gaucher, S.W. Poulton, D.E. Canfield, Fluctuations in Precambrian (2010) 27045–27056. atmospheric oxygenation recorded by chromium isotopes, Nature 461 (2009) [40] S.X. Tan, D. Greetham, S. Raeth, C.M. Grant, I.W. Dawes, G.G. Perrone, The 250–253. thioredoxin-thioredoxin reductase system can function in vivo as an alternative [6] V. Massey, Activation of molecular oxygen by flavins and flavoproteins, J. Biol. system to reduce oxidized glutathione in Saccharomyces cerevisiae, J. Biol. Chem. Chem. 269 (1994) 22459–22462. 285 (2010) 6118–6126. [7] H.J. Forman, M. Maiorino, F. Ursini, Signaling functions of reactive oxygen [41] E.G. Muller, A glutathione reductase mutant of yeast accumulates high levels of species, Biochemistry 49 (2010) 835–842. oxidized glutathione and requires thioredoxin for growth, Mol. Biol. Cell 7 [8] J.A. Imlay, Pathways of oxidative damage, Annu. Rev. Microbiol. 57 (2003) 395–418. (1996) 1805–1813. [9] J.A. Imlay, Cellular defenses against superoxide and hydrogen peroxide, Annu. [42] K. Buchholz, E.D. Putrianti, S. Rahlfs, R.H. Schirmer, K. Becker, K. Matuschewski, Mo- Rev. Biochem. 77 (2008) 755–776. lecular genetics evidence for the in vivo roles of the two major NADPH-dependent [10] H. Sies, Strategies of antioxidant defense, Eur. J. Biochem. 215 (1993) 213–219. disulfide reductases in the malaria parasite, J. Biol. Chem. 285 (2010) 37388–37395. [11] R. Brigelius-Flohe, L. Flohe, Basic principles and emerging concepts in the redox [43] L. Marty, W. Siala, M. Schwarzlander, M.D. Fricker, M. Wirtz, L.J. Sweetlove, Y. control of transcription factors, Antioxid. Redox Signal. 15 (2011) 2335–2381. Meyer, A.J. Meyer, J.P. Reichheld, R. Hell, The NADPH-dependent thioredoxin [12] G. Groeger, C. Quiney, T.G. Cotter, Hydrogen peroxide as a cell-survival signaling system constitutes a functional backup for cytosolic glutathione reductase in molecule, Antioxid. Redox Signal. 11 (2009) 2655–2671. Arabidopsis, Proc. Natl. Acad. Sci. U. S. A. 106 (2009) 9109–9114. [13] K. Bedard, K.H. Krause, The NOX family of ROS-generating NADPH oxidases: [44] R.L. Krauth-Siegel, H. Bauer, R.H. Schirmer, Dithiol proteins as guardians of the physiology and pathophysiology, Physiol. Rev. 87 (2007) 245–313. intracellular redox milieu in parasites: old and new drug targets in trypanosomes [14] J.J. Mieyal, M.M. Gallogly, S. Qanungo, E.A. Sabens, M.D. Shelton, Molecular and malaria-causing plasmodia, Angew. Chem. Int. Ed Engl. 44 (2005) 690–715. mechanisms and clinical implications of reversible protein S-glutathionylation, [45] F.Q. Schafer, G.R. Buettner, Redox environment of the cell as viewed through the Antioxid. Redox Signal. 10 (2008) 1941–1988. redox state of the glutathione disulfide/glutathione couple, Free Radic. Biol. [15] B. Schmidt, L. Ho, P.J. Hogg, Allosteric disulfide bonds, Biochemistry 45 (2006) Med. 30 (2001) 1191–1212. 7429–7433. [46] A.J. Meyer, T.P. Dick, Fluorescent protein-based redox probes, Antioxid. Redox [16] D. Lundin, S. Gribaldo, E. Torrents, B.M. Sjoberg, A.M. Poole, Ribonucleotide Signal. 13 (2010) 621–650. reduction—horizontal transfer of a required function spans all three domains, [47] M. Gutscher, A.L. Pauleau, L. Marty, T. Brach, G.H. Wabnitz, Y. Samstag, A.J. BMC Evol. Biol. 10 (2010) 383. Meyer, T.P. Dick, Real-time imaging of the intracellular glutathione redox poten- [17] P. Nordlund, P. Reichard, Ribonucleotide reductases, Annu. Rev. Biochem. 75 tial, Nat. Methods 5 (2008) 553–559. (2006) 681–706. [48] R. Munday, Toxicity of thiols and disulphides: involvement of free-radical spe- [18] M. Deponte, K. Hell, Disulphide bond formation in the intermembrane space of cies, Free Radic. Biol. Med. 7 (1989) 659–673. mitochondria, J. Biochem. 146 (2009) 599–608. [49] A. Meister, Glutathione metabolism and its selective modification, J. Biol. Chem. [19] B.S. Mamathambika, J.C. Bardwell, Disulfide-linked protein folding pathways, 263 (1988) 17205–17208. Annu. Rev. Cell Dev. Biol. 24 (2008) 211–235. [50] L. Flohé, Fairytale of the GSH/GSSG redox equilibrium, Biochim. Biophys. Acta [20] J. Riemer, N. Bulleid, J.M. Herrmann, Disulfide formation in the ER and mito- (this issue). chondria: two solutions to a common process, Science 324 (2009) 1284–1287. [51] J.S. Hofmeyr, A. Cornish-Bowden, Regulating the cellular economy of supply and [21] M.P. Kalapos, Methylglyoxal in living organisms: chemistry, biochemistry, toxi- demand, FEBS Lett. 476 (2000) 47–51. cology and biological implications, Toxicol. Lett. 110 (1999) 145–175. [52] J.M. Rohwer, J.H. Hofmeyr, Kinetic and thermodynamic aspects of enzyme con- [22] J.P. Richard, Reaction of triosephosphate isomerase with L-glyceraldehyde trol and regulation, J. Phys. Chem. B 114 (2010) 16280–16289. 3-phosphate and triose 1,2-enediol 3-phosphate, Biochemistry 24 (1985) [53] C.H. Lillig, C. Berndt, A. Holmgren, Glutaredoxin systems, Biochim. Biophys. Acta 949–953. 1780 (2008) 1304–1317. [23] P.J. Thornalley, The glyoxalase system in health and disease, Mol. Aspects Med. [54] A. Holmgren, F. Aslund, Glutaredoxin, Methods Enzymol. 252 (1995) 283–292. 14 (1993) 287–371. [55] E. Eckers, M. Bien, V. Stroobant, J.M. Herrmann, M. Deponte, Biochemical charac- [24] D.L. Pompliano, A. Peyman, J.R. Knowles, Stabilization of a reaction intermediate terization of dithiol glutaredoxin 8 from Saccharomyces cerevisiae: the catalytic as a catalytic device: definition of the functional role of the flexible loop in redox mechanism redux, Biochemistry 48 (2009) 1410–1423. triosephosphate isomerase, Biochemistry 29 (1990) 3186–3194. [56] N.J. Bulleid, L. Ellgaard, Multiple ways to make disulfides, Trends Biochem. Sci. [25] P.J. Thornalley, Modification of the glyoxalase system in human red blood cells 36 (2011) 485–492. by glucose in vitro, Biochem. J. 254 (1988) 751–755. [57] F. Hatahet, L.W. Ruddock, Protein disulfide isomerase: a critical evaluation of its [26] R.S. Ronimus, H.W. Morgan, Distribution and phylogenies of enzymes of the function in disulfide bond formation, Antioxid. Redox Signal. 11 (2009) Embden–Meyerhof–Parnas pathway from archaea and hyperthermophilic 2807–2850. bacteria support a gluconeogenic origin of metabolism, Archaea 1 (2003) [58] A.P. Fernandes, M. Fladvad, C. Berndt, C. Andresen, C.H. Lillig, P. Neubauer, M. 199–221. Sunnerhagen, A. Holmgren, A. Vlamis-Gardikas, A novel monothiol glutaredoxin [27] R.H. White, H. Xu, Methylglyoxal is an intermediate in the biosynthesis of (Grx4) from Escherichia coli can serve as a substrate for thioredoxin reductase, J. 6-deoxy-5-ketofructose-1-phosphate: a precursor for aromatic amino acid biosyn- Biol. Chem. 280 (2005) 24544–24552. thesis in Methanocaldococcus jannaschii, Biochemistry 45 (2006) 12366–12379. [59] C. Johansson, C.H. Lillig, A. Holmgren, Human mitochondrial glutaredoxin re- [28] R.A. Cooper, Metabolism of methylglyoxal in microorganisms, Annu. Rev. duces S-glutathionylated proteins with high affinity accepting electrons Microbiol. 38 (1984) 49–68. from either glutathione or thioredoxin reductase, J. Biol. Chem. 279 (2004) [29] P.J. Thornalley, Protein and nucleotide damage by glyoxal and methylglyoxal in 7537–7543. physiological systems—role in ageing and disease, Drug Metabol. Drug Interact. [60] S. Toppo, L. Flohe, F. Ursini, S. Vanin, M. Maiorino, Catalytic mechanisms and 23 (2008) 125–150. specificities of glutathione peroxidases: variations of a basic scheme, Biochim. [30] P.J. Thornalley, N. Rabbani, Glyoxalase in tumourigenesis and multidrug resis- Biophys. Acta 1790 (2009) 1486–1500. tance, Semin. Cell Dev. Biol. 22 (2011) 318–325. [61] L. Flohe, S. Toppo, G. Cozza, F. Ursini, A comparison of thiol peroxidase mecha- [31] M. Urscher, R. Alisch, M. Deponte, The glyoxalase system of malaria parasites— nisms, Antioxid. Redox Signal. 15 (2011) 763–780. implications for cell biology and general glyoxalase research, Semin. Cell Dev. [62] R. Brigelius-Flohé, GSH peroxidases, Biochim. Biophys. Acta (this issue). Biol. 22 (2011) 262–270. [63] C. , S. Rahlfs, M. Deponte, S. Koncarevic, K. Becker, Thioredoxin networks [32] M. Xue, N. Rabbani, P.J. Thornalley, Glyoxalase in ageing, Semin. Cell Dev. Biol. 22 in the malarial parasite Plasmodium falciparum, Antioxid. Redox Signal. (2011) 293–301. 8 (2006) 1227–1239. 3258 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266

[64] D. Greetham, C.M. Grant, Antioxidant activity of the yeast mitochondrial [96] D.M. LeMaster, D.M. Kushlan, Dynamical mapping of E. coli thioredoxin via 13C one-Cys is dependent on thioredoxin reductase and glutathione NMR relaxation analysis, J. Am. Chem. Soc. 118 (1996) 9255–9264. in vivo, Mol. Cell. Biol. 29 (2009) 3229–3240. [97] A.P. Wiita, R. Perez-Jimenez, K.A. Walther, F. Grater, B.J. Berne, A. Holmgren, J.M. [65] Y. Manevich, S.I. Feinstein, A.B. Fisher, Activation of the antioxidant Sanchez-Ruiz, J.M. Fernandez, Probing the chemistry of thioredoxin catalysis enzyme 1-CYS peroxiredoxin requires glutathionylation mediated by with force, Nature 450 (2007) 124–127. heterodimerization with pi GST, Proc. Natl. Acad. Sci. U. S. A. 101 (2004) [98] J.A. Pappas, Theoretical studies of the reactions of the sulfur-sulfur bond. 1. 3780–3785. General heterolytic mechanisms, J. Am. Chem. Soc. 99 (1977) 2926–2930. [66] J.R. Pedrajas, C.A. Padilla, B. McDonagh, J.A. Barcena, Glutaredoxin participates in [99] R.E. Rosenfield, Directional preferences of nonbonded atomic contacts with the reduction of peroxides by the mitochondrial 1-CYS peroxiredoxin in divalent sulfur. 1. Electrophiles and nucleophiles, J. Am. Chem. Soc. 99 (1977) Saccharomyces cerevisiae, Antioxid. Redox Signal. 13 (2010) 249–258. 4860–4862. [67] K.J. Dietz, Peroxiredoxins in plants and cyanobacteria, Antioxid. Redox Signal. 15 [100] M. Aida, C. Nagata, An ab initio MO study on the thiol–disulfide exchange reac- (2011) 1129–1159. tion, Chem. Phys. Lett. 112 (1984) 129–132. [68] E.J. Collinson, G.L. Wheeler, E.O. Garrido, A.M. Avery, S.V. Avery, C.M. Grant, The [101] J.L. Kice, C.G. Venier, L. Heasley, Mechanisms of reactions of thiolsulfinates yeast glutaredoxins are active as glutathione peroxidases, J. Biol. Chem. 277 ( sulfenic anhydrides). I. The thiolsulfinate-sulfinic acid reaction, J. Am. Chem. (2002) 16712–16717. Soc. 89 (1967) 3557–3565. [69] B. Mannervik, U.H. Danielson, Glutathione transferases—structure and catalytic [102] S.M. Bachrach, D.C. Mulhearn, Nucleophilic substitution at sulfur: SN2 or addi- activity, CRC Crit. Rev. Biochem. 23 (1988) 283–337. tion–elimination? J. Phys. Chem. 100 (1996) 3535–3540. [70] A.E. Salinas, M.G. Wong, Glutathione S-transferases—a review, Curr. Med. Chem. [103] E. Racker, Glutathione reductase from bakers' yeast and beef liver, J. Biol. Chem. 6 (1999) 279–309. 217 (1955) 855–865. [71] N. Hiller, K. Fritz-Wolf, M. Deponte, W. Wende, H. Zimmermann, K. Becker, [104] L.W. Mapson, F.A. Isherwood, Glutathione reductase from germinated peas, Plasmodium falciparum glutathione S-transferase—structural and mechanistic Biochem. J. 86 (1963) 173–191. studies on ligand binding and enzyme inhibition, Protein Sci. 15 (2006) [105] V. Massey, C.H. Williams Jr., On the reaction mechanism of yeast glutathione 281–289. reductase, J. Biol. Chem. 240 (1965) 4470–4480. [72] M.R. Fernando, J.M. Lechner, S. Lofgren, V.N. Gladyshev, M.F. Lou, Mitochondrial [106] R. Thieme, E.F. Pai, R.H. Schirmer, G.E. Schulz, Three-dimensional structure of thioltransferase (glutaredoxin 2) has GSH-dependent and thioredoxin reductase- glutathione reductase at 2 A resolution, J. Mol. Biol. 152 (1981) 763–782. dependent peroxidase activities in vitro and in lens epithelial cells, FASEB J. 20 [107] C.H. Williams, L.D. Arscott, S. Muller, B.W. Lennon, M.L. Ludwig, P.F. Wang, D.M. (2006) 2645–2647. Veine, K. Becker, R.H. Schirmer, Thioredoxin reductase two modes of catalysis [73] D.L. Vander Jagt, The glyoxalase system, Coenzmyes Cofactors VIII (1989) have evolved, Eur. J. Biochem. 267 (2000) 6110–6117. 597–641. [108] E.F. Pai, G.E. Schulz, The catalytic mechanism of glutathione reductase as derived [74] P.J. Thornalley, The glyoxalase system: new developments towards functional from X-ray diffraction analyses of reaction intermediates, J. Biol. Chem. 258 characterization of a metabolic pathway fundamental to biological life, Biochem. (1983) 1752–1757. J. 269 (1990) 1–11. [109] P.R. Mittl, G.E. Schulz, Structure of glutathione reductase from Escherichia coli at [75] A. Bresell, R. Weinander, G. Lundqvist, H. Raza, M. Shimoji, T.H. Sun, L. Balk, R. 1.86 A resolution: comparison with the enzyme from human erythrocytes, Pro- Wiklund, J. Eriksson, C. Jansson, B. Persson, P.J. Jakobsson, R. Morgenstern, Bioin- tein Sci. 3 (1994) 799–809. formatic and enzymatic characterization of the MAPEG superfamily, FEBS J. 272 [110] G.N. Sarma, S.N. Savvides, K. Becker, M. Schirmer, R.H. Schirmer, P.A. Karplus, (2005) 1688–1703. Glutathione reductase of the malarial parasite Plasmodium falciparum: crystal [76] D. Martinez Molina, S. Eshaghi, P. Nordlund, Catalysis within the lipid structure and inhibitor development, J. Mol. Biol. 328 (2003) 893–907. bilayer-structure and mechanism of the MAPEG family of integral membrane [111] J. Yu, C.Z. Zhou, Crystal structure of glutathione reductase Glr1 from the yeast proteins, Curr. Opin. Struct. Biol. 18 (2008) 442–449. Saccharomyces cerevisiae, Proteins 68 (2007) 972–979. [77] W.R. Pearson, Phylogenies of glutathione transferase families, Methods [112] P.A. Karplus, G.E. Schulz, Refined structure of glutathione reductase at 1.54 A Enzymol. 401 (2005) 186–204. resolution, J. Mol. Biol. 195 (1987) 701–729. [78] R.G. Pearson, Chemical hardness and density functional theory, J. Chem. Sci. 117 [113] G.E. Schulz, Gene duplication in glutathione reductase, J. Mol. Biol. 138 (1980) (2005) 369–377. 335–347. [79] G.I. Giles, F.H. Fry, K.M. Tasker, A.L. Holme, C. Peers, K.N. Green, L.O. Klotz, H. Sies, [114] S.N. Savvides, P.A. Karplus, Kinetics and crystallographic analysis of human glu- C. Jacob, Evaluation of sulfur, selenium and tellurium catalysts with antioxidant tathione reductase in complex with a xanthene inhibitor, J. Biol. Chem. 271 potential, Org. Biomol. Chem. 1 (2003) 4317–4322. (1996) 8101–8107. [80] E.S. Arner, —what unique properties can arise with selenocysteine [115] C.E. Outten, V.C. Culotta, Alternative start sites in the Saccharomyces cerevisiae in place of cysteine? Exp. Cell Res. 316 (2010) 1296–1303. GLR1 gene are responsible for mitochondrial and cytosolic isoforms of glutathi- [81] S. Gromer, L. Johansson, H. Bauer, L.D. Arscott, S. Rauch, D.P. Ballou, C.H. one reductase, J. Biol. Chem. 279 (2004) 7785–7791. Williams Jr., R.H. Schirmer, E.S. Arner, Active sites of thioredoxin reductases: [116] F. Mbemba, A. Houbion, M. Raes, J. Remacle, Subcellular localization and modifi- why selenoproteins? Proc. Natl. Acad. Sci. U. S. A. 100 (2003) 12618–12623. cation with ageing of glutathione, glutathione peroxidase and glutathione re- [82] B.N. Khare, C. Sagan, Synthesis of cystine in simulated primitive conditions, ductase activities in human fibroblasts, Biochim. Biophys. Acta 838 (1985) Nature 232 (1971) 577–579. 211–220. [83] M.H. Stipanuk, Sulfur amino acid metabolism: pathways for production and [117] M. Taniguchi, T. Hara, H. Honda, Similarities between rat liver mitochondrial and removal of homocysteine and cysteine, Annu. Rev. Nutr. 24 (2004) 539–577. cytosolic glutathione reductases and their apoenzyme accumulation in ribofla- [84] G.L. Newton, N. Buchmeier, R.C. Fahey, Biosynthesis and functions of mycothiol, vin deficiency, Biochem. Int. 13 (1986) 447–454. the unique protective thiol of Actinobacteria, Microbiol. Mol. Biol. Rev. 72 [118] G. Creissen, H. Reynolds, Y. Xue, P. Mullineaux, Simultaneous targeting of pea (2008) 471–494. glutathione reductase and of a bacterial fusion protein to chloroplasts and mito- [85] R.C. Fahey, Novel thiols of prokaryotes, Annu. Rev. Microbiol. 55 (2001) chondria in transgenic tobacco, Plant J. 8 (1995) 167–175. 333–356. [119] S. Kehr, N. Sturm, S. Rahlfs, J.M. Przyborski, K. Becker, Compartmentation of [86] A.R. Sundquist, R.C. Fahey, The function of gamma-glutamylcysteine and redox metabolism in malaria parasites, PLoS Pathog. 6 (2010) e1001242. bis-gamma-glutamylcystine reductase in Halobacterium halobium, J. Biol. [120] M.J. Kelner, M.A. Montoya, Structural organization of the human glutathione re- Chem. 264 (1989) 719–725. ductase gene: determination of correct cDNA sequence and identification of a [87] C.C. Tsen, A.L. Tappel, Catalytic oxidation of glutathione and other sulfhydryl mitochondrial leader sequence, Biochem. Biophys. Res. Commun. 269 (2000) compounds by selenite, J. Biol. Chem. 233 (1958) 1230–1232. 366–368. [88] M.R. Ariyanayagam, A.H. Fairlamb, Ovothiol and trypanothione as antioxidants [121] T. Tamura, H.W. McMicken, C.V. Smith, T.N. Hansen, Mitochondrial targeting in trypanosomatids, Mol. Biochem. Parasitol. 115 (2001) 189–198. of glutathione reductase requires a leader sequence, Biochem. Biophys. Res. [89] A.H. Fairlamb, P. Blackburn, P. Ulrich, B.T. Chait, A. Cerami, Trypanothione: a Commun. 222 (1996) 659–663. novel bis(glutathionyl)spermidine for glutathione reductase in [122] R.L. Krauth-Siegel, L.D. Arscott, A. Schonleben-Janas, R.H. Schirmer, C.H. trypanosomatids, Science 227 (1985) 1485–1487. Williams Jr., Role of active site tyrosine residues in catalysis by human glutathi- [90] M. Comini, R. Radi, Trypanothiol, Biochim. Biophys. Acta (this issue). one reductase, Biochemistry 37 (1998) 13968–13977. [91] H. Tabor, C.W. Tabor, Isolation, characterization, and turnover of [123] A. Berry, N.S. Scrutton, R.N. Perham, Switching kinetic mechanism and putative glutathionylspermidine from Escherichia coli, J. Biol. Chem. 250 (1975) 2648–2654. proton donor by directed mutagenesis of glutathione reductase, Biochemistry [92] B.Y. Chiang, T.C. Chen, C.H. Pai, C.C. Chou, H.H. Chen, T.P. Ko, W.H. Hsu, C.Y. Chang, 28 (1989) 1264–1269. W.F. Wu, A.H. Wang, C.H. Lin, Protein S-thiolation by Glutathionylspermidine [124] P.W. Huber, K.G. Brandt, Kinetic studies of the mechanism of pyridine nucleotide (Gsp): the role of Escherichia coli Gsp synthetASE/amidase in redox regulation, dependent reduction of yeast glutathione reductase, Biochemistry 19 (1980) J. Biol. Chem. 285 (2010) 25345–25353. 4569–4575. [93] A. Gaballa, G.L. Newton, H. Antelmann, D. Parsonage, H. Upton, M. Rawat, A. [125] L.D. Arscott, C. Thorpe, C.H. Williams Jr., Glutathione reductase from yeast. Dif- Claiborne, R.C. Fahey, J.D. Helmann, Biosynthesis and functions of bacillithiol, a ferential reactivity of the nascent thiols in two-electron reduced enzyme and major low-molecular-weight thiol in Bacilli, Proc. Natl. Acad. Sci. U. S. A. 107 properties of a monoalkylated derivative, Biochemistry 20 (1981) 1513–1520. (2010) 6482–6486. [126] P. Rietveld, L.D. Arscott, A. Berry, N.S. Scrutton, M.P. Deonarain, R.N. Perham, C.H. [94] G.L. Newton, M. Rawat, J.J. La Clair, V.K. Jothivasan, T. Budiarto, C.J. Hamilton, A. Williams Jr., Reductive and oxidative half-reactions of glutathione reductase Claiborne, J.D. Helmann, R.C. Fahey, Bacillithiol is an antioxidant thiol produced from Escherichia coli, Biochemistry 33 (1994) 13888–13895. in Bacilli, Nat. Chem. Biol. 5 (2009) 625–627. [127] D.M. Veine, L.D. Arscott, C.H. Williams Jr., Redox potentials for yeast, Escherichia [95] C.E. Hand, J.F. Honek, Biological chemistry of naturally occurring thiols of micro- coli and human glutathione reductase relative to the NAD+/NADH redox cou- bial and marine origin, J. Nat. Prod. 68 (2005) 293–308. ple: enzyme forms active in catalysis, Biochemistry 37 (1998) 15575–15582. M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3259

[128] J.E. Bulger, K.G. Brandt, Yeast glutathione reductase. II. Interaction of oxidized [158] K. Nakashima, K. Yamauchi, S. Miwa, K. Fujimura, A. Mizutani, A. Kuramoto, Glu- and 2-electron reduced enzyme with reduced and oxidized nicotinamide ade- tathione reductase deficiency in a kindred with hereditary spherocytosis, Am. J. nine dinucleotide phosphate, J. Biol. Chem. 246 (1971) 5578–5587. Hematol. 4 (1978) 141–150. [129] K.K. Wong, J.S. Blanchard, Human erythrocyte glutathione reductase: pH depen- [159] N.M. Kamerbeek, R. van Zwieten, M. de Boer, G. Morren, H. Vuil, N. Bannink, C. dence of kinetic parameters, Biochemistry 28 (1989) 3586–3590. Lincke, K.M. Dolman, K. Becker, R.H. Schirmer, S. Gromer, D. Roos, Molecular [130] K.K. Wong, M.A. Vanoni, J.S. Blanchard, Glutathione reductase: solvent equilibri- basis of glutathione reductase deficiency in human blood cells, Blood 109 um and kinetic isotope effects, Biochemistry 27 (1988) 7091–7096. (2007) 3560–3566. [131] M.P. Deonarain, A. Berry, N.S. Scrutton, R.N. Perham, Alternative proton [160] H. Bauer, K. Fritz-Wolf, A. Winzer, S. Kuhner, S. Little, V. Yardley, H. Vezin, B. donors/acceptors in the catalytic mechanism of the glutathione reductase of Palfey, R.H. Schirmer, E. Davioud-Charvet, A fluoro analogue of the menadione Escherichia coli: the role of histidine-439 and tyrosine-99, Biochemistry 28 derivative 6-[2′-(3′-methyl)-1′,4′-naphthoquinolyl]hexanoic acid is a suicide (1989) 9602–9607. substrate of glutathione reductase. Crystal structure of the alkylated human en- [132] L.D. Arscott, D.M. Veine, C.H. Williams Jr., Mixed disulfide with glutathione as an zyme, J. Am. Chem. Soc. 128 (2006) 10784–10794. intermediate in the reaction catalyzed by glutathione reductase from yeast and [161] T. Muller, L. Johann, B. Jannack, M. Bruckner, D.A. Lanfranchi, H. Bauer, C. as a major form of the enzyme in the cell, Biochemistry 39 (2000) 4711–4721. Sanchez, V. Yardley, C. Deregnaucourt, J. Schrevel, M. Lanzer, R.H. Schirmer, E. [133] L. Canepa, A.M. Ferraris, M. Miglino, G.F. Gaetani, Bound and unbound pyridine Davioud-Charvet, Glutathione reductase-catalyzed cascade of redox reactions dinucleotides in normal and glucose-6-phosphate dehydrogenase-deficient to bioactivate potent antimalarial 1,4-naphthoquinones—a new strategy to com- erythrocytes, Biochim. Biophys. Acta 1074 (1991) 101–104. bat malarial parasites, J. Am. Chem. Soc. 133 (2011) 11557–11571. [134] C.C. Bohme, L.D. Arscott, K. Becker, R.H. Schirmer, C.H. Williams Jr., Kinetic char- [162] K. Buchholz, R.H. Schirmer, J.K. Eubel, M.B. Akoachere, T. Dandekar, K. Becker, S. acterization of glutathione reductase from the malarial parasite Plasmodium Gromer, Interactions of methylene blue with human disulfide reductases and falciparum. Comparison with the human enzyme, J. Biol. Chem. 275 (2000) their orthologues from Plasmodium falciparum, Antimicrob. Agents Chemother. 37317–37323. 52 (2008) 183–191. [135] M.A. Vanoni, K.K. Wong, D.P. Ballou, J.S. Blanchard, Glutathione reductase: com- [163] A. Zoungrana, B. Coulibaly, A. Sie, I. Walter-Sack, F.P. Mockenhaupt, B. Kouyate, parison of steady-state and rapid reaction primary kinetic isotope effects R.H. Schirmer, C. Klose, U. Mansmann, P. Meissner, O. Muller, Safety and efficacy exhibited by the yeast, spinach, and Escherichia coli enzymes, Biochemistry 29 of methylene blue combined with artesunate or amodiaquine for uncomplicated (1990) 5790–5796. falciparum malaria: a randomized controlled trial from Burkina Faso, PLoS One 3 [136] A. Serrano, J. Rivas, M. Losada, Purification and properties of glutathione reduc- (2008) e1630. tase from the cyanobacterium Anabaena sp. strain 7119, J. Bacteriol. 158 (1984) [164] M. Bountogo, A. Zoungrana, B. Coulibaly, C. Klose, U. Mansmann, F.P. 317–324. Mockenhaupt, J. Burhenne, G. Mikus, I. Walter-Sack, R.H. Schirmer, A. Sie, P. [137] V. Boggaram, K. Larson, B. Mannervik, Characterization of glutathione reductase Meissner, O. Muller, Efficacy of methylene blue monotherapy in semi-immune from porcine erythrocytes, Biochim. Biophys. Acta 527 (1978) 337–347. adults with uncomplicated falciparum malaria: a controlled trial in Burkina [138] P.M. Farber, L.D. Arscott, C.H. Williams Jr., K. Becker, R.H. Schirmer, Recombinant Faso, Trop. Med. Int. Health 15 (2010) 713–717. Plasmodium falciparum glutathione reductase is inhibited by the antimalarial [165] M. Bonilla, A. Denicola, S.M. Marino, V.N. Gladyshev, G. Salinas, Linked thioredoxin– dye methylene blue, FEBS Lett. 422 (1998) 311–314. glutathione systems in platyhelminth parasites: alternative pathways for glutathi- [139] D.S. Berkholz, H.R. Faber, S.N. Savvides, P.A. Karplus, Catalytic cycle of human one reduction and deglutathionylation, J. Biol. Chem. 286 (2011) 4959–4967. glutathione reductase near 1 A resolution, J. Mol. Biol. 382 (2008) 371–384. [166] E. Racker, Glutathione–homocystine transhydrogenase, J. Biol. Chem. 217 [140] L.I. Leichert, F. Gehrke, H.V. Gudiseva, T. Blackwell, M. Ilbert, A.K. Walker, J.R. (1955) 867–874. Strahler, P.C. Andrews, U. Jakob, Quantifying changes in the thiol redox prote- [167] H.T. Narahara, R.H. Williams, Reduction of insulin by extracts of rat liver, J. Biol. ome upon oxidative stress in vivo, Proc. Natl. Acad. Sci. U. S. A. 105 (2008) Chem. 234 (1959) 71–77. 8197–8202. [168] H.H. Tomizawa, P.T. Varandani, Glutathione-insulin transhydrogenase of human [141] C. Kumsta, M. Thamsen, U. Jakob, Effects of oxidative stress on behavior, physi- liver, J. Biol. Chem. 240 (1965) 3191–3194. ology, and the redox thiol proteome of Caenorhabditis elegans, Antioxid. Redox [169] H.M. Katzen, F. Tietze, Studies on the specificity and mechanism of action of hepatic Signal. 14 (2011) 1023–1037. glutathione-insulin transhydrogenase, J. Biol. Chem. 241 (1966) 3561–3570. [142] N. Brandes, D. Reichmann, H. Tienson, L.I. Leichert, U. Jakob, Using quantitative [170] R.B. Freedman, How many distinct enzymes are responsible for the several cel- redox proteomics to dissect the yeast redoxome, J. Biol. Chem. 286 (2011) lular processes involving thiol:protein-disulphide interchange? FEBS Lett. 97 41893–41903. (1979) 201–210. [143] M.P. Deonarain, N.S. Scrutton, R.N. Perham, Engineering surface charge. 2. A [171] A. Holmgren, Thioredoxin catalyzes the reduction of insulin disulfides by method for purifying heterodimers of Escherichia coli glutathione reductase, Bio- dithiothreitol and dihydrolipoamide, J. Biol. Chem. 254 (1979) 9627–9632. chemistry 31 (1992) 1498–1504. [172] S. Nagai, S. Black, A thiol–disulfide transhydrogenase from yeast, J. Biol. Chem. [144] B. Mannervik, A branching reaction mechanism of glutathione reductase, 243 (1968) 1942–1947. Biochem. Biophys. Res. Commun. 53 (1973) 1151–1158. [173] N. Mesecke, S. Mittler, E. Eckers, J.M. Herrmann, M. Deponte, Two novel [145] G.E. Staal, C. Veeger, The reaction mechanism of glutathione reductase from monothiol glutaredoxins from Saccharomyces cerevisiae provide further insight human erythrocytes, Biochim. Biophys. Acta 185 (1969) 49–62. into iron–sulfur cluster binding, oligomerization, and enzymatic activity of [146] N.S. Scrutton, M.P. Deonarain, A. Berry, R.N. Perham, Cooperativity induced by a glutaredoxins, Biochemistry 47 (2008) 1452–1463. single mutation at the subunit interface of a dimeric enzyme: glutathione reduc- [174] P. Askelof, K. Axelsson, S. Eriksson, B. Mannervik, Mechanism of action of en- tase, Science 258 (1992) 1140–1143. zymes catalyzing thiol–disulfide interchange. Thioltransferases rather than [147] A. Bashir, R.N. Perham, N.S. Scrutton, A. Berry, Altering kinetic mechanism and transhydrogenases, FEBS Lett. 38 (1974) 263–267. enzyme stability by mutagenesis of the dimer interface of glutathione reductase, [175] K. Axelsson, S. Eriksson, B. Mannervik, Purification and characterization of cyto- Biochem. J. 312 (Pt 2) (1995) 527–533. plasmic thioltransferase (glutathione:disulfide oxidoreductase) from rat liver, [148] C.M. Grant, L.P. Collinson, J.H. Roe, I.W. Dawes, Yeast glutathione reductase is re- Biochemistry 17 (1978) 2978–2984. quired for protection against oxidative stress and is a target gene for yAP-1 tran- [176] B. Mannervik, K. Axelsson, K. Larson, Thioltransferase, Methods Enzymol. 77 scriptional regulation, Mol. Microbiol. 21 (1996) 171–179. (1981) 281–285. [149] C.E. Outten, R.L. Falk, V.C. Culotta, Cellular factors required for protection from [177] M. Luthman, S. Eriksson, A. Holmgren, L. Thelander, Glutathione-dependent hy- hyperoxia toxicity in Saccharomyces cerevisiae, Biochem. J. 388 (2005) 93–101. drogen donor system for calf thymus ribonucleoside-diphosphate reductase, [150] C.K. Tuggle, J.A. Fuchs, Glutathione reductase is not required for maintenance Proc. Natl. Acad. Sci. U. S. A. 76 (1979) 2158–2162. of reduced glutathione in Escherichia coli K-12, J. Bacteriol. 162 (1985) [178] M. Luthman, A. Holmgren, Glutaredoxin from calf thymus. Purification to homo- 448–450. geneity, J. Biol. Chem. 257 (1982) 6686–6690. [151] E.M. Patzewitz, E.H. Wong, S. Muller, Dissecting the role of glutathione biosyn- [179] Z.R. Gan, W.W. Wells, Purification and properties of thioltransferase, J. Biol. thesis in Plasmodium falciparum, Mol. Microbiol. 83 (2012) 304–318. Chem. 261 (1986) 996–1001. [152] I. Tzafrir, R. Pena-Muralla, A. Dickerman, M. Berg, R. Rogers, S. Hutchens, T.C. [180] Z.R. Gan, W.W. Wells, Preparation of homogeneous pig liver thioltransferase by Sweeney, J. McElver, G. Aux, D. Patton, D. Meinke, Identification of genes re- a thiol:disulfide mediated pI shift, Anal. Biochem. 162 (1987) 265–273. quired for embryo development in Arabidopsis, Plant Physiol. 135 (2004) [181] Z.R. Gan, W.W. Wells, The primary structure of pig liver thioltransferase, J. Biol. 1206–1220. Chem. 262 (1987) 6699–6703. [153] N. Satoh, N. Watanabe, A. Kanda, M. Sugaya-Fukasawa, H. Hisatomi, Expression [182] Y.F. Yang, Z.R. Gan, W.W. Wells, Cloning and sequencing the cDNA encoding pig of glutathione reductase splice variants in human tissues, Biochem. Genet. 48 liver thioltransferase, Gene 83 (1989) 339–346. (2010) 816–821. [183] S.K. Katti, A.H. Robbins, Y. Yang, W.W. Wells, Crystal structure of thioltransferase [154] D.J. O'Donovan, J.P. Katkin, T. Tamura, C.V. Smith, S.E. Welty, Attenuation of at 2.2 A resolution, Protein Sci. 4 (1995) 1998–2005. hyperoxia-induced growth inhibition in H441 cells by gene transfer of [184] Y. Yang, S. Jao, S. Nanduri, D.W. Starke, J.J. Mieyal, J. Qin, Reactivity of the human mitochondrially targeted glutathione reductase, Am. J. Respir. Cell Mol. Biol. 22 thioltransferase (glutaredoxin) C7S, C25S, C78S, C82S mutant and NMR solution (2000) 732–738. structure of its glutathionyl mixed disulfide intermediate reflect catalytic speci- [155] K.M. Heyob, L.K. Rogers, S.E. Welty, Glutathione reductase targeted to type II ficity, Biochemistry 37 (1998) 17145–17156. cells does not protect mice from hyperoxic lung injury, Am. J. Respir. Cell Mol. [185] A. Holmgren, Hydrogen donor system for Escherichia coli ribonucleoside- Biol. 39 (2008) 683–688. diphosphate reductase dependent upon glutathione, Proc. Natl. Acad. Sci. U. S. A. [156] H.C. Benohr, H.D. Waller, Glutathione (author's transl), Klin. Wochenschr. 53 73 (1976) 2275–2279. (1975) 789–802. [186] A. Holmgren, Glutathione-dependent synthesis of deoxyribonucleotides. Purifi- [157] H. Loos, D. Roos, R. Weening, J. Houwerzijl, Familial deficiency of glutathione re- cation and characterization of glutaredoxin from Escherichia coli, J. Biol. Chem. ductase in human blood cells, Blood 48 (1976) 53–62. 254 (1979) 3664–3671. 3260 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266

[187] A. Holmgren, Glutathione-dependent synthesis of deoxyribonucleotides. Char- [215] K.F. Discola, M.A. de Oliveira, J.R. Rosa Cussiol, G. Monteiro, J.A. Barcena, P. acterization of the enzymatic mechanism of Escherichia coli glutaredoxin, J. Porras, C.A. Padilla, B.G. Guimaraes, L.E. Netto, Structural aspects of the distinct Biol. Chem. 254 (1979) 3672–3678. biochemical properties of glutaredoxin 1 and glutaredoxin 2 from Saccharomyces [188] J.O. Hoog, H. Jornvall, A. Holmgren, M. Carlquist, M. Persson, The primary struc- cerevisiae, J. Mol. Biol. 385 (2009) 889–901. ture of Escherichia coli glutaredoxin. Distant homology with thioredoxins in a [216] M. Luo, Y.L. Jiang, X.X. Ma, Y.J. Tang, Y.X. He, J. Yu, R.G. Zhang, Y. Chen, C.Z. Zhou, superfamily of small proteins with a redox-active cystine disulfide/cysteine Structural and biochemical characterization of yeast monothiol glutaredoxin dithiol, Eur. J. Biochem. 136 (1983) 223–232. Grx6, J. Mol. Biol. 398 (2010) 614–622. [189] J.O. Hoog, H. von Bahr-Lindstrom, H. Jornvall, A. Holmgren, Cloning and expres- [217] M.A. Comini, J. Rettig, N. Dirdjaja, E.M. Hanschmann, C. Berndt, R.L. Krauth-Siegel, sion of the glutaredoxin (grx) gene of Escherichia coli, Gene 43 (1986) 13–21. Monothiol glutaredoxin-1 is an essential iron–sulfur protein in the [190] V.A. Sandberg, B. Kren, J.A. Fuchs, C. Woodward, Escherichia coli glutaredoxin: of African trypanosomes, J. Biol. Chem. 283 (2008) 27785–27798. cloning and overexpression, thermodynamic stability of the oxidized and re- [218] V. Noguera, O. Walker, N. Rouhier, J.P. Jacquot, I. Krimm, J.M. Lancelin, NMR re- duced forms, and report of an N-terminal extended species, Biochemistry 30 veals a novel glutaredoxin–glutaredoxin interaction interface, J. Mol. Biol. 353 (1991) 5475–5484. (2005) 629–641. [191] T.H. Xia, J.H. Bushweller, P. Sodano, M. Billeter, O. Bjornberg, A. Holmgren, K. Wuthrich, [219] J.J. Kelley III, T.M. Caputo, S.F. Eaton, T.M. Laue, J.H. Bushweller, Comparison of NMR structure of oxidized Escherichia coli glutaredoxin: comparison with reduced backbone dynamics of reduced and oxidized Escherichia coli glutaredoxin-1 E. coli glutaredoxin and functionally related proteins, Protein Sci. 1 (1992) 310–321. using 15N NMR relaxation measurements, Biochemistry 36 (1997) 5029–5044. [192] J.H. Bushweller, M. Billeter, A. Holmgren, K. Wuthrich, The nuclear magnetic res- [220] C.M. Grant, Role of the glutathione/glutaredoxin and thioredoxin systems in yeast onance solution structure of the mixed disulfide between Escherichia coli growth and response to stress conditions, Mol. Microbiol. 39 (2001) 533–541. glutaredoxin(C14S) and glutathione, J. Mol. Biol. 235 (1994) 1585–1597. [221] J.J. Song, Y.J. Lee, Differential role of glutaredoxin and thioredoxin in metabolic [193] U. Muhlenhoff, S. Molik, J.R. Godoy, M.A. Uzarska, N. Richter, A. Seubert, Y. oxidative stress-induced activation of apoptosis signal-regulating kinase 1, Zhang, J. Stubbe, F. Pierrel, E. Herrero, C.H. Lillig, R. Lill, Cytosolic monothiol Biochem. J. 373 (2003) 845–853. glutaredoxins function in intracellular iron sensing and trafficking via their [222] S. Gon, M.J. Faulkner, J. Beckwith, In vivo requirement for glutaredoxins and bound iron–sulfur cluster, Cell Metab. 12 (2010) 373–385. thioredoxins in the reduction of the ribonucleotide reductases of Escherichia [194] N. Rouhier, S.D. Lemaire, J.P. Jacquot, The role of glutathione in photosynthetic coli, Antioxid. Redox Signal. 8 (2006) 735–742. organisms: emerging functions for glutaredoxins and glutathionylation, Annu. [223] J. Shi, A. Vlamis-Gardikas, F. Aslund, A. Holmgren, B.P. Rosen, Reactivity of Rev. Plant Biol. 59 (2008) 143–166. glutaredoxins 1, 2, and 3 from Escherichia coli shows that glutaredoxin 2 is the [195] J.L. Martin, Thioredoxin—a fold for all reasons, Structure 3 (1995) 245–250. primary hydrogen donor to ArsC-catalyzed arsenate reduction, J. Biol. Chem. [196] E. Herrero, M.A. de la Torre-Ruiz, Monothiol glutaredoxins: a common domain 274 (1999) 36039–36042. for multiple functions, Cell. Mol. Life Sci. 64 (2007) 1518–1530. [224] F. Zahedi Avval, A. Holmgren, Molecular mechanisms of thioredoxin and [197] N. Mesecke, A. Spang, M. Deponte, J.M. Herrmann, A novel group of glutaredoxin as hydrogen donors for mammalian S phase ribonucleotide reduc- glutaredoxins in the cis-Golgi critical for oxidative stress resistance, Mol. Biol. tase, J. Biol. Chem. 284 (2009) 8233–8240. Cell 19 (2008) 2673–2680. [225] L. Brautigam, L.D. Schutte, J.R. Godoy, T. Prozorovski, M. Gellert, G. Hauptmann, [198] A. Izquierdo, C. Casas, U. Muhlenhoff, C.H. Lillig, E. Herrero, Saccharomyces A. Holmgren, C.H. Lillig, C. Berndt, Vertebrate-specific glutaredoxin is essential cerevisiae Grx6 and Grx7 are monothiol glutaredoxins associated with the for brain development, Proc. Natl. Acad. Sci. U. S. A. 108 (2011) 20532–20537. early secretory pathway, Eukaryot. Cell 7 (2008) 1415–1426. [226] S. Luikenhuis, G. Perrone, I.W. Dawes, C.M. Grant, The yeast Saccharomyces [199] B. Hoffmann, M.A. Uzarska, C. Berndt, J.R. Godoy, P. Haunhorst, C.H. Lillig, R. Lill, cerevisiae contains two glutaredoxin genes that are required for protection U. Muhlenhoff, The multidomain thioredoxin-monothiol glutaredoxins repre- against reactive oxygen species, Mol. Biol. Cell 9 (1998) 1081–1091. sent a distinct functional group, Antioxid. Redox Signal. 15 (2011) 19–30. [227] K.H. Wolfe, D.C. Shields, Molecular evidence for an ancient duplication of the en- [200] J.H. Bushweller, F. Aslund, K. Wuthrich, A. Holmgren, Structural and functional tire yeast genome, Nature 387 (1997) 708–713. characterization of the mutant Escherichia coli glutaredoxin (C14–S) and its [228] J.R. Pedrajas, P. Porras, E. Martinez-Galisteo, C.A. Padilla, A. Miranda-Vizuete, J.A. mixed disulfide with glutathione, Biochemistry 31 (1992) 9288–9293. Barcena, Two isoforms of Saccharomyces cerevisiae glutaredoxin 2 are expressed [201] C. Johansson, K.L. Kavanagh, O. Gileadi, U. Oppermann, Reversible sequestration in vivo and localize to different subcellular compartments, Biochem. J. 364 of active site cysteines in a 2Fe–2S-bridged dimer provides a mechanism for (2002) 617–623. glutaredoxin 2 regulation in human mitochondria, J. Biol. Chem. 282 (2007) [229] W.K. Huh, J.V. Falvo, L.C. Gerke, A.S. Carroll, R.W. Howson, J.S. Weissman, E.K. 3077–3082. O'Shea, Global analysis of protein localization in budding yeast, Nature 425 [202] M. Fladvad, M. Bellanda, A.P. Fernandes, S. Mammi, A. Vlamis-Gardikas, A. (2003) 686–691. Holmgren, M. Sunnerhagen, Molecular mapping of functionalities in the solu- [230] P. Porras, B. McDonagh, J.R. Pedrajas, J.A. Barcena, C.A. Padilla, Structure and tion structure of reduced Grx4, a monothiol glutaredoxin from Escherichia coli, function of yeast glutaredoxin 2 depend on postranslational processing and J. Biol. Chem. 280 (2005) 24553–24561. are related to subcellular distribution, Biochim. Biophys. Acta 1804 (2010) [203] R. Alves, E. Vilaprinyo, A. Sorribas, E. Herrero, Evolution based on domain com- 839–845. binations: the case of glutaredoxins, BMC Evol. Biol. 9 (2009) 66. [231] P. Porras, C.A. Padilla, M. Krayl, W. Voos, J.A. Barcena, One single in-frame AUG [204] M. Maiorino, F. Ursini, V. Bosello, S. Toppo, S.C. Tosatto, P. Mauri, K. Becker, A. codon is responsible for a diversity of subcellular localizations of glutaredoxin Roveri, C. Bulato, L. Benazzi, A. De Palma, L. Flohe, The thioredoxin specificity 2inSaccharomyces cerevisiae, J. Biol. Chem. 281 (2006) 16551–16562. of Drosophila GPx: a paradigm for a peroxiredoxin-like mechanism of many [232] S. Ghaemmaghami, W.K. Huh, K. Bower, R.W. Howson, A. Belle, N. Dephoure, glutathione peroxidases, J. Mol. Biol. 365 (2007) 1033–1046. E.K. O'Shea, J.S. Weissman, Global analysis of protein expression in yeast, Nature [205] J. Yu, N.N. Zhang, P.D. Yin, P.X. Cui, C.Z. Zhou, Glutathionylation-triggered con- 425 (2003) 737–741. formational changes of glutaredoxin Grx1 from the yeast Saccharomyces [233] S. Camier, E. Ma, C. Leroy, A. Pruvost, M. Toledano, M.C. Marsolier-Kergoat, Visual- cerevisiae, Proteins 72 (2008) 1077–1083. ization of ribonucleotide reductase catalytic oxidation establishes thioredoxins as [206] W.F. Li, J. Yu, X.X. Ma, Y.B. Teng, M. Luo, Y.J. Tang, C.Z. Zhou, Structural basis for its major reductants in yeast, Free Radic. Biol. Med. 42 (2007) 1008–1016. the different activities of yeast Grx1 and Grx2, Biochim. Biophys. Acta 1804 [234] E.J. Collinson, C.M. Grant, Role of yeast glutaredoxins as glutathione (2010) 1542–1547. S-transferases, J. Biol. Chem. 278 (2003) 22492–22497. [207] J. Couturier, C.S. Koh, M. Zaffagnini, A.M. Winger, J.M. Gualberto, C. Corbier, P. [235] M.M. Molina, G. Belli, M.A. de la Torre, M.T. Rodriguez-Manzaneque, E. Herrero, Decottignies, J.P. Jacquot, S.D. Lemaire, C. Didierjean, N. Rouhier, Structure–func- Nuclear monothiol glutaredoxins of Saccharomyces cerevisiae can function as tion relationship of the chloroplastic glutaredoxin S12 with an atypical WCSYS mitochondrial glutaredoxins, J. Biol. Chem. 279 (2004) 51923–51930. active site, J. Biol. Chem. 284 (2009) 9299–9310. [236] R. Lopreiato, S. Facchin, G. Sartori, G. Arrigoni, S. Casonato, M. Ruzzene, L.A. [208] M. Deponte, K. Becker, S. Rahlfs, Plasmodium falciparum glutaredoxin-like pro- Pinna, G. Carignani, Analysis of the interaction between piD261/Bud32, an evo- teins, Biol. Chem. 386 (2005) 33–40. lutionarily conserved protein kinase of Saccharomyces cerevisiae, and the Grx4 [209] T. Iwema, A. Picciocchi, D.A. Traore, J.L. Ferrer, F. Chauvat, L. Jacquamet, Structur- glutaredoxin, Biochem. J. 377 (2004) 395–405. al basis for delivery of the intact [Fe2S2] cluster by monothiol glutaredoxin, Bio- [237] M.T. Rodriguez-Manzaneque, J. Ros, E. Cabiscol, A. Sorribas, E. Herrero, Grx5 chemistry 48 (2009) 6041–6043. glutaredoxin plays a central role in protection against protein oxidative damage [210] C. Johansson, A.K. Roos, S.J. Montano, R. Sengupta, P. Filippakopoulos, K. Guo, F. in Saccharomyces cerevisiae, Mol. Cell. Biol. 19 (1999) 8180–8190. von Delft, A. Holmgren, U. Oppermann, K.L. Kavanagh, The crystal structure of [238] J. Tamarit, G. Belli, E. Cabiscol, E. Herrero, J. Ros, Biochemical characterization of human GLRX5: iron–sulfur cluster co-ordination, tetrameric assembly and yeast mitochondrial Grx5 monothiol glutaredoxin, J. Biol. Chem. 278 (2003) monomer activity, Biochem. J. 433 (2011) 303–311. 25745–25751. [211] M.T. Rodriguez-Manzaneque, J. Tamarit, G. Belli, J. Ros, E. Herrero, Grx5 is a mi- [239] R. Lill, Function and biogenesis of iron–sulphur proteins, Nature 460 (2009) tochondrial glutaredoxin required for the activity of iron/sulfur enzymes, Mol. 831–838. Biol. Cell 13 (2002) 1109–1121. [240] D. Shenton, G. Perrone, K.A. Quinn, I.W. Dawes, C.M. Grant, Regulation of protein [212] M.J. Berardi, J.H. Bushweller, Binding specificity and mechanistic insight into S-thiolation by glutaredoxin 5 in the yeast Saccharomyces cerevisiae, J. Biol. glutaredoxin-catalyzed protein disulfide reduction, J. Mol. Biol. 292 (1999) Chem. 277 (2002) 16853–16859. 151–161. [241] R.A. Wingert, J.L. Galloway, B. Barut, H. Foott, P. Fraenkel, J.L. Axe, G.J. Weber, K. [213] K. Nordstrand, A. Sandstrom, F. Aslund, A. Holmgren, G. Otting, K.D. Berndt, NMR Dooley, A.J. Davidson, B. Schmid, B.H. Paw, G.C. Shaw, P. Kingsley, J. Palis, H. structure of oxidized glutaredoxin 3 from Escherichia coli, J. Mol. Biol. 303 (2000) Schubert, O. Chen, J. Kaplan, L.I. Zon, Deficiency of glutaredoxin 5 reveals Fe–S 423–432. clusters are required for vertebrate haem synthesis, Nature 436 (2005) [214] J.P. Bacik, B. Hazes, Crystal structures of a poxviral glutaredoxin in the oxidized 1035–1039. and reduced states show redox-correlated structural changes, J. Mol. Biol. 365 [242] S. Bandyopadhyay, F. Gama, M.M. Molina-Navarro, J.M. Gualberto, R. Claxton, (2007) 1545–1558. S.G. Naik, B.H. Huynh, E. Herrero, J.P. Jacquot, M.K. Johnson, N. Rouhier, M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3261

Chloroplast monothiol glutaredoxins as scaffold proteins for the assembly and [268] F. Aslund, K.D. Berndt, A. Holmgren, Redox potentials of glutaredoxins and other delivery of [2Fe–2S] clusters, EMBO J. 27 (2008) 1122–1133. thiol–disulfide oxidoreductases of the thioredoxin superfamily determined by [243] K.D. Kim, W.H. Chung, H.J. Kim, K.C. Lee, J.H. Roe, Monothiol glutaredoxin Grx5 direct protein–protein redox equilibria, J. Biol. Chem. 272 (1997) 30780–30786. interacts with Fe–S scaffold proteins Isa1 and Isa2 and supports Fe–S assembly [269] A.M. Caccuri, G. Antonini, N. Allocati, C. Di Ilio, F. De Maria, F. Innocenti, M.W. and DNA integrity in mitochondria of fission yeast, Biochem. Biophys. Res. Parker, M. Masulli, M. Lo Bello, P. Turella, G. Federici, G. Ricci, GSTB1-1 from Commun. 392 (2010) 467–472. Proteus mirabilis: a snapshot of an enzyme in the evolutionary pathway from a [244] N. Pujol-Carrion, G. Belli, E. Herrero, A. Nogues, M.A. de la Torre-Ruiz, redox enzyme to a conjugating enzyme, J. Biol. Chem. 277 (2002) 18777–18784. Glutaredoxins Grx3 and Grx4 regulate nuclear localisation of Aft1 and the oxida- [270] Y.S. Ho, Y. Xiong, D.S. Ho, J. Gao, B.H. Chua, H. Pai, J.J. Mieyal, Targeted disruption tive stress response in Saccharomyces cerevisiae, J. Cell Sci. 119 (2006) of the glutaredoxin 1 gene does not sensitize adult mice to tissue injury induced 4554–4564. by ischemia/reperfusion and hyperoxia, Free Radic. Biol. Med. 43 (2007) [245] L. Ojeda, G. Keller, U. Muhlenhoff, J.C. Rutherford, R. Lill, D.R. Winge, Role of 1299–1312. glutaredoxin-3 and glutaredoxin-4 in the iron regulation of the Aft1 transcrip- [271] J.R. Godoy, M. Funke, W. Ackermann, P. Haunhorst, S. Oesteritz, F. Capani, H.P. tional activator in Saccharomyces cerevisiae, J. Biol. Chem. 281 (2006) Elsasser, C.H. Lillig, Redox atlas of the mouse. Immunohistochemical detection 17661–17669. of glutaredoxin-, peroxiredoxin-, and thioredoxin-family proteins in various tis- [246] A. Mercier, S. Labbe, Both Php4 function and subcellular localization are regulat- sues of the , Biochim. Biophys. Acta 1810 (2011) 2–92. ed by iron via a multistep mechanism involving the glutaredoxin Grx4 and the [272] S. Yoshitake, H. Nanri, M.R. Fernando, S. Minakami, Possible differences in the exportin Crm1, J. Biol. Chem. 284 (2009) 20249–20262. regenerative roles played by thioltransferase and thioredoxin for oxidatively [247] C. Peggion, R. Lopreiato, E. Casanova, M. Ruzzene, S. Facchin, L.A. Pinna, G. damaged proteins, J. Biochem. 116 (1994) 42–46. Carignani, G. Sartori, Phosphorylation of the Saccharomyces cerevisiae Grx4p [273] S. Mohr, H. Hallak, A. de Boitte, E.G. Lapetina, B. Brune, Nitric oxide-induced glutaredoxin by the Bud32p kinase unveils a novel signaling pathway involving S-glutathionylation and inactivation of glyceraldehyde-3-phosphate dehydro- Sch9p, a yeast member of the Akt / PKB subfamily, FEBS J. 275 (2008) genase, J. Biol. Chem. 274 (1999) 9427–9430. 5919–5933. [274] S. Reddy, A.D. Jones, C.E. Cross, P.S. Wong, A. Van Der Vliet, Inactivation of crea- [248] S. Witte, M. Villalba, K. Bi, Y. Liu, N. Isakov, A. Altman, Inhibition of the c-Jun tine kinase by S-glutathionylation of the active-site cysteine residue, Biochem. J. N-terminal kinase/AP-1 and NF-kappaB pathways by PICOT, a novel protein ki- 347 (Pt 3) (2000) 821–827. nase C-interacting protein with a thioredoxin homology domain, J. Biol. Chem. [275] E. Cabiscol, R.L. Levine, The phosphatase activity of carbonic anhydrase III is reversi- 275 (2000) 1902–1909. bly regulated by glutathiolation, Proc. Natl. Acad. Sci. U. S. A. 93 (1996) 4170–4174. [249] A. Vlamis-Gardikas, F. Aslund, G. Spyrou, T. Bergman, A. Holmgren, Cloning, [276] Y.R. Chen, C.L. Chen, D.R. Pfeiffer, J.L. Zweier, Mitochondrial complex II in the overexpression, and characterization of glutaredoxin 2, an atypical glutaredoxin post-ischemic heart: oxidative injury and the role of protein S-glutathionylation, from Escherichia coli, J. Biol. Chem. 272 (1997) 11236–11243. J. Biol. Chem. 282 (2007) 32640–32654. [250] F. Aslund, B. Ehn, A. Miranda-Vizuete, C. Pueyo, A. Holmgren, Two additional [277] E.R. Taylor, F. Hurrell, R.J. Shannon, T.K. Lin, J. Hirst, M.P. Murphy, Reversible glutaredoxins exist in Escherichia coli: glutaredoxin 3 is a hydrogen donor for ri- glutathionylation of complex I increases mitochondrial superoxide formation, bonucleotide reductase in a thioredoxin/glutaredoxin 1 double mutant, Proc. J. Biol. Chem. 278 (2003) 19603–19610. Natl. Acad. Sci. U. S. A. 91 (1994) 9813–9817. [278] J. Wang, E.S. Boja, W. Tan, E. Tekle, H.M. Fales, S. English, J.J. Mieyal, P.B. Chock, [251] C. Berndt, C. Hudemann, E.M. Hanschmann, R. Axelsson, A. Holmgren, C.H. Lillig, Reversible glutathionylation regulates actin polymerization in A431 cells, How does iron–sulfur cluster coordination regulate the activity of human J. Biol. Chem. 276 (2001) 47763–47766. glutaredoxin 2? Antioxid. Redox Signal. 9 (2007) 151–157. [279] J. Wang, E. Tekle, H. Oubrahim, J.J. Mieyal, E.R. Stadtman, P.B. Chock, Stable and [252] M.M. Gallogly, D.W. Starke, A.K. Leonberg, S.M. Ospina, J.J. Mieyal, Kinetic and controllable RNA interference: investigating the physiological function of mechanistic characterization and versatile catalytic properties of mammalian glutathionylated actin, Proc. Natl. Acad. Sci. U. S. A. 100 (2003) 5103–5106. glutaredoxin 2: implications for intracellular roles, Biochemistry 47 (2008) [280] T. Adachi, R.M. Weisbrod, D.R. Pimentel, J. Ying, V.S. Sharov, C. Schoneich, R.A. 11144–11157. Cohen, S-Glutathiolation by peroxynitrite activates SERCA during arterial relax- [253] S. Hopper, R.S. Johnson, J.E. Vath, K. Biemann, Glutaredoxin from rabbit bone ation by nitric oxide, Nat. Med. 10 (2004) 1200–1207. marrow. Purification, characterization, and amino acid sequence determined [281] V. Anathy, S.W. Aesif, A.S. Guala, M. Havermans, N.L. Reynaert, Y.S. Ho, R.C. Budd, Y.M. by tandem mass spectrometry, J. Biol. Chem. 264 (1989) 20438–20447. Janssen-Heininger, Redox amplification of apoptosis by caspase-dependent cleavage [254] M. Lundberg, C. Johansson, J. Chandra, M. Enoksson, G. Jacobsson, J. Ljung, M. of glutaredoxin 1 and S-glutathionylation of Fas, J. Cell Biol. 184 (2009) 241–252. Johansson, A. Holmgren, Cloning and expression of a novel human glutaredoxin [282] P. Aracena-Parks, S.A. Goonasekera, C.P. Gilman, R.T. Dirksen, C. Hidalgo, S.L. (Grx2) with mitochondrial and nuclear isoforms, J. Biol. Chem. 276 (2001) Hamilton, Identification of cysteines involved in S-nitrosylation, S-glutathionylation, 26269–26275. and oxidation to disulfides in ryanodine receptor type 1, J. Biol. Chem. 281 (2006) [255] J.J. Mieyal, D.W. Starke, S.A. Gravina, B.A. Hocevar, Thioltransferase in human red 40354–40368. blood cells: kinetics and equilibrium, Biochemistry 30 (1991) 8883–8891. [283] C.S. Queiroga, A.S. Almeida, C. Martel, C. Brenner, P.M. Alves, H.L. Vieira, [256] S. Rahlfs, M. Fischer, K. Becker, Plasmodium falciparum possesses a classical Glutathionylation of adenine nucleotide induced by carbon monox- glutaredoxin and a second, glutaredoxin-like protein with a PICOT homology ide prevents mitochondrial membrane permeabilization and apoptosis, J. Biol. domain, J. Biol. Chem. 276 (2001) 37133–37140. Chem. 285 (2010) 17077–17088. [257] N. Rouhier, E. Gelhaye, J.P. Jacquot, Exploring the active site of plant [284] Y. Yang, W. Shi, N. Cui, Z. Wu, C. Jiang, Oxidative stress inhibits vascular K(ATP) glutaredoxin by site-directed mutagenesis, FEBS Lett. 511 (2002) 145–149. channels by S-glutathionylation, J. Biol. Chem. 285 (2010) 38641–38648. [258] S. Sha, K. Minakuchi, N. Higaki, K. Sato, K. Ohtsuki, A. Kurata, H. Yoshikawa, M. [285] W.C. Barrett, J.P. DeGnore, S. Konig, H.M. Fales, Y.F. Keng, Z.Y. Zhang, M.B. Yim, Kotaru, T. Masumura, K. Ichihara, K. Tanaka, Purification and characterization P.B. Chock, Regulation of PTP1B via glutathionylation of the active site cysteine of glutaredoxin (thioltransferase) from rice (Oryza sativa L.), J. Biochem. 121 215, Biochemistry 38 (1999) 6699–6705. (1997) 842–848. [286] K.M. Humphries, C. Juliano, S.S. Taylor, Regulation of cAMP-dependent protein [259] M. Zaffagnini, L. Michelet, V. Massot, P. Trost, S.D. Lemaire, Biochemical charac- kinase activity by glutathionylation, J. Biol. Chem. 277 (2002) 43505–43511. terization of glutaredoxins from Chlamydomonas reinhardtii reveals the unique [287] T. Kambe, T. Song, T. Takata, N. Hatano, Y. Miyamoto, N. Nozaki, Y. Naito, H. properties of a chloroplastic CGFS-type glutaredoxin, J. Biol. Chem. 283 (2008) Tokumitsu, Y. Watanabe, Inactivation of Ca2+/calmodulin-dependent protein 8868–8876. kinase I by S-glutathionylation of the active-site cysteine residue, FEBS Lett. [260] M. Filser, M.A. Comini, M.M. Molina-Navarro, N. Dirdjaja, E. Herrero, R.L. 584 (2010) 2478–2484. Krauth-Siegel, Cloning, functional analysis, and mitochondrial localization of [288] R.K. Rao, L.W. Clayton, Regulation of protein phosphatase 2A by hydrogen peroxide Trypanosoma brucei monothiol glutaredoxin-1, Biol. Chem. 389 (2008) 21–32. and glutathionylation, Biochem. Biophys. Res. Commun. 293 (2002) 610–616. [261] X.H. Gao, M. Zaffagnini, M. Bedhomme, L. Michelet, C. Cassier-Chauvat, P. [289] N.L. Reynaert, A. van der Vliet, A.S. Guala, T. McGovern, M. Hristova, C. Decottignies, S.D. Lemaire, Biochemical characterization of glutaredoxins from Pantano, N.H. Heintz, J. Heim, Y.S. Ho, D.E. Matthews, E.F. Wouters, Y.M. Chlamydomonas reinhardtii: kinetics and specificity in deglutathionylation reac- Janssen-Heininger, Dynamic redox control of NF-kappaB through glutaredoxin- tions, FEBS Lett. 584 (2010) 2242–2248. regulated S-glutathionylation of inhibitory kappaB kinase beta, Proc. Natl. Acad. [262] S.A. Gravina, J.J. Mieyal, Thioltransferase is a specific glutathionyl mixed disul- Sci. U. S. A. 103 (2006) 13086–13091. fide oxidoreductase, Biochemistry 32 (1993) 3368–3376. [290] N.E. Ward, J.R. Stewart, C.G. Ioannides, C.A. O'Brian, Oxidant-induced [263] Y.F. Yang, W.W. Wells, Identification and characterization of the functional S-glutathiolation inactivates protein kinase C-alpha (PKC-alpha): a potential amino acids at the active center of pig liver thioltransferase by site-directed mu- mechanism of PKC isozyme regulation, Biochemistry 39 (2000) 10319–10329. tagenesis, J. Biol. Chem. 266 (1991) 12759–12765. [291] S. Bandyopadhyay, D.W. Starke, J.J. Mieyal, R.M. Gronostajski, Thioltransferase [264] Y.F. Yang, W.W. Wells, Catalytic mechanism of thioltransferase, J. Biol. Chem. (glutaredoxin) reactivates the DNA-binding activity of oxidation-inactivated 266 (1991) 12766–12771. nuclear factor I, J. Biol. Chem. 273 (1998) 392–397. [265] U. Srinivasan, P.A. Mieyal, J.J. Mieyal, pH profiles indicative of rate-limiting nu- [292] T. Adachi, D.R. Pimentel, T. Heibeck, X. Hou, Y.J. Lee, B. Jiang, Y. Ido, R.A. Cohen, cleophilic displacement in thioltransferase catalysis, Biochemistry 36 (1997) S-glutathiolation of Ras mediates redox-sensitive signaling by angiotensin II in 3199–3206. vascular smooth muscle cells, J. Biol. Chem. 279 (2004) 29857–29862. [266] M. Zaffagnini, M. Bedhomme, C.H. Marchand, J.R. Couturier, X.H. Gao, N. Rouhier, [293] E. Pineda-Molina, P. Klatt, J. Vazquez, A. Marina, M. Garcia de Lacoba, D. P. Trost, S.P. Lemaire, Glutaredoxin s12: unique properties for redox signaling, Perez-Sala, S. Lamas, Glutathionylation of the p50 subunit of NF-kappaB: a Antioxid. Redox Signal. 16 (2012) 17–32. mechanism for redox-induced inhibition of DNA binding, Biochemistry 40 [267] M.J. Peltoniemi, A.R. Karala, J.K. Jurvansuu, V.L. Kinnula, L.W. Ruddock, Insights (2001) 14134–14142. into deglutathionylation reactions. Different intermediates in the glutaredoxin [294] S. Qanungo, D.W. Starke, H.V. Pai, J.J. Mieyal, A.L. Nieminen, Glutathione supplemen- and protein disulfide isomerase catalyzed reactions are defined by the tation potentiates hypoxic apoptosis by S-glutathionylation of p65-NFkappaB, gamma-linkage present in glutathione, J. Biol. Chem. 281 (2006) 33107–33114. J. Biol. Chem. 282 (2007) 18427–18436. 3262 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266

[295] R. Dringen, J. Hirrlinger, Glutathione pathways in the brain, Biol. Chem. 384 tryparedoxin peroxidases in Trypanosoma brucei, J. Biol. Chem. 280 (2005) (2003) 505–516. 14385–14394. [296] K.J. Dietz, Redox signal integration: from stimulus to networks and genes, Phys- [324] R.S. Esworthy, R. Aranda, M.G. Martin, J.H. Doroshow, S.W. Binder, F.F. Chu, Mice iol. Plant. 133 (2008) 459–468. with combined disruption of Gpx1 and Gpx2 genes have colitis, Am. J. Physiol. [297] N. Sturm, E. Jortzik, B.M. Mailu, S. Koncarevic, M. Deponte, K. Forchhammer, S. Gastrointest. Liver Physiol. 281 (2001) G848–G855. Rahlfs, K. Becker, Identification of proteins targeted by the thioredoxin super- [325] X.G. Lei, Glutathione peroxidase-1 gene knockout on body antioxidant defense family in Plasmodium falciparum, PLoS Pathog. 5 (2009) e1000383. in mice, Biofactors 14 (2001) 93–99. [298] G.C. Mills, Hemoglobin catabolism. I. Glutathione peroxidase, an erythrocyte en- [326] R.S. Esworthy, K.M. Swiderek, Y.S. Ho, F.F. Chu, Selenium-dependent glutathione zyme which protects hemoglobin from oxidative breakdown, J. Biol. Chem. 229 peroxidase-GI is a major glutathione peroxidase activity in the mucosal epithe- (1957) 189–197. lium of rodent intestine, Biochim. Biophys. Acta 1381 (1998) 213–226. [299] D. Neubert, A.B. Wojtczak, A.L. Lehninger, Purification and enzymatic identity of [327] L.J. Yant, Q. Ran, L. Rao, H. Van Remmen, T. Shibatani, J.G. Belter, L. Motta, A. mitochondrial contraction-factors I and II, Proc. Natl. Acad. Sci. U. S. A. 48 (1962) Richardson, T.A. Prolla, The selenoprotein GPX4 is essential for mouse develop- 1651–1658. ment and protects from radiation and oxidative damage insults, Free Radic. Biol. [300] L. Flohe, G. Loschen, W.A. Gunzler, E. Eichele, Glutathione peroxidase, V, The ki- Med. 34 (2003) 496–502. netic mechanism, Hoppe Seylers Z. Physiol. Chem. 353 (1972) 987–999. [328] M. Maiorino, M. Scapin, F. Ursini, M. Biasolo, V. Bosello, L. Flohe, Distinct pro- [301] W.A. Gunzler, H. Vergin, I. Muller, L. Flohe, Glutathione peroxidase VI: the reac- moters determine alternative transcription of gpx-4 into phospholipid- tion of glutahione peroxidase with various hydroperoxides, Hoppe Seylers Z. hydroperoxide glutathione peroxidase variants, J. Biol. Chem. 278 (2003) Physiol. Chem. 353 (1972) 1001–1004. 34286–34290. [302] L. Flohe, W.A. Gunzler, H.H. Schock, Glutathione peroxidase: a selenoenzyme, [329] T.R. Pushpa-Rekha, A.L. Burdsall, L.M. Oleksa, G.M. Chisolm, D.M. Driscoll, Rat FEBS Lett. 32 (1973) 132–134. phospholipid-hydroperoxide glutathione peroxidase. cDNA cloning and identifi- [303] R. Ladenstein, O. Epp, K. Bartels, A. Jones, R. Huber, A. Wendel, Structure analysis cation of multiple transcription and translation start sites, J. Biol. Chem. 270 and molecular model of the selenoenzyme glutathione peroxidase at 2.8 A res- (1995) 26993–26999. olution, J. Mol. Biol. 134 (1979) 199–218. [330] J.M. Herrmann, K. Hell, Chopped, trapped or tacked—protein translocation into [304] O. Epp, R. Ladenstein, A. Wendel, The refined structure of the selenoenzyme glu- the IMS of mitochondria, Trends Biochem. Sci. 30 (2005) 205–211. tathione peroxidase at 0.2-nm resolution, Eur. J. Biochem. 133 (1983) 51–69. [331] C. Godeas, G. Sandri, E. Panfili, Distribution of phospholipid hydroperoxide glu- [305] J.W. Forstrom, J.J. Zakowski, A.L. Tappel, Identification of the catalytic site of rat tathione peroxidase (PHGPx) in rat testis mitochondria, Biochim. Biophys. Acta liver glutathione peroxidase as selenocysteine, Biochemistry 17 (1978) 1191 (1994) 147–150. 2639–2644. [332] M. Arai, H. Imai, D. Sumi, T. Imanaka, T. Takano, N. Chiba, Y. Nakagawa, Import [306] W.A. Gunzler, G.J. Steffens, A. Grossmann, S.M. Kim, F. Otting, A. Wendel, L. into mitochondria of phospholipid hydroperoxide glutathione peroxidase re- Flohe, The amino-acid sequence of bovine glutathione peroxidase, Hoppe quires a leader sequence, Biochem. Biophys. Res. Commun. 227 (1996) 433–439. Seylers Z. Physiol. Chem. 365 (1984) 195–212. [333] M. Schneider, H. Forster, A. Boersma, A. Seiler, H. Wehnes, F. Sinowatz, C. [307] I. Chambers, J. Frampton, P. Goldfarb, N. Affara, W. McBain, P.R. Harrison, The struc- Neumuller, M.J. Deutsch, A. Walch, M. Hrabe de Angelis, W. Wurst, F. Ursini, A. ture of the mouse glutathione peroxidase gene: the selenocysteine in the active site Roveri, M. Maleszewski, M. Maiorino, M. Conrad, Mitochondrial glutathione per- is encoded by the ‘termination’ codon, TGA, EMBO J. 5 (1986) 1221–1227. oxidase 4 disruption causes male infertility, FASEB J. 23 (2009) 3233–3242. [308] F. Ursini, M. Maiorino, M. Valente, L. Ferri, C. Gregolin, Purification from pig liver [334] H. Imai, N. Hakkaku, R. Iwamoto, J. Suzuki, T. Suzuki, Y. Tajima, K. Konishi, S. of a protein which protects liposomes and biomembranes from peroxidative Minami, S. Ichinose, K. Ishizaka, S. Shioda, S. Arata, M. Nishimura, S. Naito, Y. degradation and exhibits glutathione peroxidase activity on phosphatidylcho- Nakagawa, Depletion of selenoprotein GPx4 in spermatocytes causes male infer- line hydroperoxides, Biochim. Biophys. Acta 710 (1982) 197–211. tility in mice, J. Biol. Chem. 284 (2009) 32522–32532. [309] F. Ursini, M. Maiorino, C. Gregolin, The selenoenzyme phospholipid hydroperox- [335] K. Nomura, H. Imai, T. Koumura, T. Kobayashi, Y. Nakagawa, Mitochondrial phos- ide glutathione peroxidase, Biochim. Biophys. Acta 839 (1985) 62–70. pholipid hydroperoxide glutathione peroxidase inhibits the release of cyto- [310] C. Rocher, J.L. Lalanne, J. Chaudiere, Purification and properties of a recombinant chrome c from mitochondria by suppressing the peroxidation of cardiolipin in sulfur analog of murine selenium-glutathione peroxidase, Eur. J. Biochem. 205 hypoglycaemia-induced apoptosis, Biochem. J. 351 (2000) 183–193. (1992) 955–960. [336] C. Godeas, F. Tramer, F. Micali, A. Roveri, M. Maiorino, C. Nisii, G. Sandri, E. [311] S. Toppo, S. Vanin, V. Bosello, S.C. Tosatto, Evolutionary and structural insights Panfili, Phospholipid hydroperoxide glutathione peroxidase (PHGPx) in rat tes- into the multifaceted glutathione peroxidase (Gpx) superfamily, Antioxid. tis nuclei is bound to chromatin, Biochem. Mol. Med. 59 (1996) 118–124. Redox Signal. 10 (2008) 1501–1514. [337] S.G. Moreno, G. Laux, M. Brielmeier, G.W. Bornkamm, M. Conrad, Testis-specific [312] H. Sztajer, B. Gamain, K.D. Aumann, C. Slomianny, K. Becker, R. Brigelius-Flohe, L. expression of the nuclear form of phospholipid hydroperoxide glutathione per- Flohe, The putative glutathione peroxidase gene of Plasmodium falciparum codes oxidase (PHGPx), Biol. Chem. 384 (2003) 635–643. for a thioredoxin peroxidase, J. Biol. Chem. 276 (2001) 7397–7403. [338] M. Conrad, S.G. Moreno, F. Sinowatz, F. Ursini, S. Kolle, A. Roveri, M. Brielmeier, [313] T. Tanaka, S. Izawa, Y. Inoue, GPX2, encoding a phospholipid hydroperoxide glu- W. Wurst, M. Maiorino, G.W. Bornkamm, The nuclear form of phospholipid hy- tathione peroxidase homologue, codes for an atypical 2-Cys peroxiredoxin in droperoxide glutathione peroxidase is a protein thiol peroxidase contributing to Saccharomyces cerevisiae, J. Biol. Chem. 280 (2005) 42078–42087. sperm chromatin stability, Mol. Cell. Biol. 25 (2005) 7637–7644. [314] T. Schlecker, M.A. Comini, J. Melchers, T. Ruppert, R.L. Krauth-Siegel, Catalytic [339] H. Liang, S.E. Yoo, R. Na, C.A. Walter, A. Richardson, Q. Ran, Short form glutathi- mechanism of the glutathione peroxidase-type tryparedoxin peroxidase of one peroxidase 4 is the essential isoform required for survival and somatic mi- Trypanosoma brucei, Biochem. J. 405 (2007) 445–454. tochondrial functions, J. Biol. Chem. 284 (2009) 30836–30844. [315] B.G. Jung, K.O. Lee, S.S. Lee, Y.H. Chi, H.H. Jang, S.S. Kang, K. Lee, D. Lim, S.C. Yoon, [340] A. Seiler, M. Schneider, H. Forster, S. Roth, E.K. Wirth, C. Culmsee, N. Plesnila, E. D.J. Yun, Y. Inoue, M.J. Cho, S.Y. Lee, A Chinese cabbage cDNA with high sequence Kremmer, O. Radmark, W. Wurst, G.W. Bornkamm, U. Schweizer, M. Conrad, Gluta- identity to phospholipid hydroperoxide glutathione peroxidases encodes a thione peroxidase 4 senses and translates oxidative stress into 12/15-lipoxygenase novel isoform of thioredoxin-dependent peroxidase, J. Biol. Chem. 277 (2002) dependent- and AIF-mediated cell death, Cell Metab. 8 (2008) 237–248. 12572–12578. [341] Q. Ran, H. Liang, M. Gu, W. Qi, C.A. Walter, L.J. Roberts II, B. Herman, A. [316] C.S. Koh, C. Didierjean, N. Navrot, S. Panjikar, G. Mulliert, N. Rouhier, J.P. Jacquot, Richardson, H. Van Remmen, Transgenic mice overexpressing glutathione per- A. Aubry, O. Shawkataly, C. Corbier, Crystal structures of a poplar thioredoxin oxidase 4 are protected against oxidative stress-induced apoptosis, J. Biol. peroxidase that exhibits the structure of glutathione peroxidases: insights into Chem. 279 (2004) 55137–55146. redox-driven conformational changes, J. Mol. Biol. 370 (2007) 512–529. [342] H.Z. Chae, S.J. Chung, S.G. Rhee, Thioredoxin-dependent peroxide reductase [317] P. Scheerer, A. Borchert, N. Krauss, H. Wessner, C. Gerth, W. Hohne, H. Kuhn, from yeast, J. Biol. Chem. 269 (1994) 27670–27678. Structural basis for catalytic activity and enzyme polymerization of phospholip- [343] S.G. Rhee, H.Z. Chae, K. Kim, Peroxiredoxins: a historical overview and specula- id hydroperoxide glutathione peroxidase-4 (GPx4), Biochemistry 46 (2007) tive preview of novel mechanisms and emerging concepts in cell signaling, Free 9041–9049. Radic. Biol. Med. 38 (2005) 1543–1552. [318] J. Melchers, M. Diechtierow, K. Feher, I. Sinning, I. Tews, R.L. Krauth-Siegel, C. [344] M. Deponte, S. Rahlfs, K. Becker, Peroxiredoxin systems of protozoal parasites, Muhle-Goll, Structural basis for a distinct catalytic mechanism in Trypanosoma Subcell. Biochem. 44 (2007) 219–229. brucei tryparedoxin peroxidase, J. Biol. Chem. 283 (2008) 30401–30411. [345] L.B. Poole, The catalytic mechanism of peroxiredoxins, Subcell. Biochem. 44 [319] F.F. Chu, J.H. Doroshow, R.S. Esworthy, Expression, characterization, and tissue (2007) 61–81. distribution of a new cellular selenium-dependent glutathione peroxidase, [346] P.A. Karplus, A. Hall, Structural survey of the peroxiredoxins, Subcell. Biochem. GSHPx-GI, J. Biol. Chem. 268 (1993) 2571–2576. 44 (2007) 41–60. [320] K. Takahashi, N. Avissar, J. Whitin, H. Cohen, Purification and characterization of [347] B. Knoops, E. Loumaye, V. Van Der Eecken, Evolution of the peroxiredoxins, human plasma glutathione peroxidase: a selenoglycoprotein distinct from the Subcell. Biochem. 44 (2007) 27–40. known cellular enzyme, Arch. Biochem. Biophys. 256 (1987) 677–686. [348] M. Deponte, K. Becker, Biochemical characterization of Toxoplasma gondii 1-Cys [321] F. Ursini, S. Heim, M. Kiess, M. Maiorino, A. Roveri, J. Wissing, L. Flohe, Dual func- with mechanistic similarities to typical 2-Cys Prx, Mol. tion of the selenoprotein PHGPx during sperm maturation, Science 285 (1999) Biochem. Parasitol. 140 (2005) 87–96. 1393–1396. [349] N. Rouhier, E. Gelhaye, J.P. Jacquot, Glutaredoxin-dependent peroxiredoxin from [322] N. Navrot, V. Collin, J. Gualberto, E. Gelhaye, M. Hirasawa, P. Rey, D.B. Knaff, E. poplar: protein–protein interaction and catalytic mechanism, J. Biol. Chem. 277 Issakidis, J.P. Jacquot, N. Rouhier, Plant glutathione peroxidases are functional (2002) 13609–13614. peroxiredoxins distributed in several subcellular compartments and regulated [350] F. Pauwels, B. Vergauwen, F. Vanrobaeys, B. Devreese, J.J. Van Beeumen, Purifica- during biotic and abiotic stresses, Plant Physiol. 142 (2006) 1364–1379. tion and characterization of a chimeric enzyme from Haemophilus influenzae Rd [323] T. Schlecker, A. Schmidt, N. Dirdjaja, F. Voncken, C. Clayton, R.L. Krauth-Siegel, Sub- that exhibits glutathione-dependent peroxidase activity, J. Biol. Chem. 278 strate specificity, localization, and essential role of the glutathione peroxidase-type (2003) 16658–16666. M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3263

[351] S.A. Reeves, D. Parsonage, K.J. Nelson, L.B. Poole, Kinetic and thermodynamic fea- [381] A. Bito, M. Haider, P. Briza, P. Strasser, M. Breitenbach, Heterologous expression, tures reveal that Escherichia coli BCP is an unusually versatile peroxiredoxin, purification, and kinetic comparison of the cytoplasmic and mitochondrial Biochemistry 50 (2011) 8970–8981. glyoxalase II enzymes, Glo2p and Glo4p, from Saccharomyces cerevisiae, Protein [352] D. Richard, R. Bartfai, J. Volz, S.A. Ralph, S. Muller, H.G. Stunnenberg, A.F. Expr. Purif. 17 (1999) 456–464. Cowman, A genome-wide chromatin-associated nuclear peroxiredoxin from [382] A. Bito, M. Haider, I. Hadler, M. Breitenbach, Identification and phenotypic anal- the malaria parasite Plasmodium falciparum, J. Biol. Chem. 286 (2011) ysis of two glyoxalase II encoding genes from Saccharomyces cerevisiae, GLO2 11746–11755. and GLO4, and intracellular localization of the corresponding proteins, J. Biol. [353] B. Schremmer, Y. Manevich, S.I. Feinstein, A.B. Fisher, Peroxiredoxins in the lung Chem. 272 (1997) 21509–21519. with emphasis on peroxiredoxin VI, Subcell. Biochem. 44 (2007) 317–344. [383] E.M. Frickel, P. Jemth, M. Widersten, B. Mannervik, Yeast glyoxalase I is a mono- [354] L.A. Ralat, Y. Manevich, A.B. Fisher, R.F. Colman, Direct evidence for the forma- meric enzyme with two active sites, J. Biol. Chem. 276 (2001) 1845–1849. tion of a complex between 1-cysteine peroxiredoxin and glutathione [384] M.J. Penninckx, C.J. Jaspers, M.J. Legrain, The glutathione-dependent glyoxalase S-transferase pi with activity changes in both enzymes, Biochemistry 45 pathway in the yeast Saccharomyces cerevisiae, J. Biol. Chem. 258 (1983) (2006) 360–372. 6030–6036. [355] A.B. Fisher, C. Dodia, Y. Manevich, J.W. Chen, S.I. Feinstein, Phospholipid hydro- [385] M.W. Crowder, M.K. Maiti, L. Banovic, C.A. Makaroff, Glyoxalase II from A. peroxides are substrates for non-selenium glutathione peroxidase, J. Biol. Chem. thaliana requires Zn(II) for catalytic activity, FEBS Lett. 418 (1997) 351–354. 274 (1999) 21326–21334. [386] P. Limphong, G. Nimako, P.W. Thomas, W. Fast, C.A. Makaroff, M.W. Crowder, [356] H.J. Choi, S.W. Kang, C.H. Yang, S.G. Rhee, S.E. Ryu, Crystal structure of a novel Arabidopsis thaliana mitochondrial glyoxalase 2–1 exhibits beta-lactamase ac- human peroxidase enzyme at 2.0 A resolution, Nat. Struct. Biol. 5 (1998) tivity, Biochemistry 48 (2009) 8491–8493. 400–406. [387] G.P. Marasinghe, I.M. Sander, B. Bennett, G. Periyannan, K.W. Yang, C.A. [357] S.J. Kim, J.R. Woo, Y.S. Hwang, D.G. Jeong, D.H. Shin, K. Kim, S.E. Ryu, The tetra- Makaroff, M.W. Crowder, Structural studies on a mitochondrial glyoxalase II, J. meric structure of Haemophilus influenza hybrid Prx5 reveals interactions be- Biol. Chem. 280 (2005) 40668–40675. tween electron donor and acceptor proteins, J. Biol. Chem. 278 (2003) [388] M. Ridderstrom, B. Mannervik, Molecular cloning and characterization of the 10790–10798. thiolesterase glyoxalase II from Arabidopsis thaliana, Biochem. J. 322 (Pt 2) [358] V. Noguera-Mazon, J. Lemoine, O. Walker, N. Rouhier, A. Salvador, J.P. Jacquot, J.M. (1997) 449–454. Lancelin, I. Krimm, Glutathionylation induces the dissociation of 1-Cys D- [389] A. Sommer, P. Fischer, K. Krause, K. Boettcher, P.M. Brophy, R.D. Walter, E. peroxiredoxin non-covalent homodimer, J. Biol. Chem. 281 (2006) 31736–31742. Liebau, A stress-responsive glyoxalase I from the parasitic nematode Onchocerca [359] H. Castro, F. Teixeira, S. Romao, M. Santos, T. Cruz, M. Florido, R. Appelberg, P. volvulus, Biochem. J. 353 (2001) 445–452. Oliveira, F. Ferreira-da-Silva, A.M. Tomas, mitochondrial peroxiredoxin [390] A. Ariza, T.J. Vickers, N. Greig, K.A. Armour, M.J. Dixon, I.M. Eggleston, A.H. plays a crucial peroxidase-unrelated role during infection: insight into its novel Fairlamb, C.S. Bond, Specificity of the trypanothione-dependent Leishmania chaperone activity, PLoS Pathog. 7 (2011) e1002325. major glyoxalase I: structure and biochemical comparison with the human en- [360] H.H. Jang, K.O. Lee, Y.H. Chi, B.G. Jung, S.K. Park, J.H. Park, J.R. Lee, S.S. Lee, J.C. zyme, Mol. Microbiol. 59 (2006) 1239–1248. Moon, J.W. Yun, Y.O. Choi, W.Y. Kim, J.S. Kang, G.W. Cheong, D.J. Yun, S.G. [391] N. Greig, S. Wyllie, T.J. Vickers, A.H. Fairlamb, Trypanothione-dependent Rhee, M.J. Cho, S.Y. Lee, Two enzymes in one; two yeast peroxiredoxins display glyoxalase I in , Biochem. J. 400 (2006) 217–223. oxidative stress-dependent switching from a peroxidase to a molecular chaper- [392] T. Irsch, R.L. Krauth-Siegel, Glyoxalase II of African trypanosomes is one function, Cell 117 (2004) 625–635. trypanothione-dependent, J. Biol. Chem. 279 (2004) 22209–22217. [361] X. Wang, S.A. Phelan, K. Forsman-Semb, E.F. Taylor, C. Petros, A. Brown, C.P. Lerner, [393] M.S. Silva, L. Barata, A.E. Ferreira, S. Romao, A.M. Tomas, A.P. Freire, C. Cordeiro, B. Paigen, Mice with targeted mutation of peroxiredoxin 6 develop normally but Catalysis and structural properties of Leishmania infantum glyoxalase II: are susceptible to oxidative stress, J. Biol. Chem. 278 (2003) 25179–25190. trypanothione specificity and phylogeny, Biochemistry 47 (2008) 195–204. [362] F. Pauwels, B. Vergauwen, J.J. Van Beeumen, Physiological characterization of [394] T.J. Vickers, N. Greig, A.H. Fairlamb, A trypanothione-dependent glyoxalase I Haemophilus influenzae Rd deficient in its glutathione-dependent peroxidase with a prokaryotic ancestry in , Proc. Natl. Acad. Sci. U. S. A. PGdx, J. Biol. Chem. 279 (2004) 12163–12170. 101 (2004) 13186–13191. [363] H.D. Dakin, H.W. Dudley, On glyoxalase, J. Biol. Chem. 14 (1913) 423–431. [395] R.N. Armstrong, Mechanistic diversity in a metalloenzyme superfamily, Bio- [364] C. Neuberg, The destruction of lactic and methylglyoxal by animal or- chemistry 39 (2000) 13625–13632. gans, Biochem. Z. 49 (1913) 502–506. [396] M. Bergdoll, L.D. Eltis, A.D. Cameron, P. Dumas, J.T. Bolin, All in the family: struc- [365] M. Jowett, J.H. Quastel, The glyoxalase activity of the red blood cell: the function tural and evolutionary relationships among three modular proteins with diverse of glutathione, Biochem. J. 27 (1933) 486–498. functions and variable assembly, Protein Sci. 7 (1998) 1661–1670. [366] K. Lohmann, Beitrag zur enzymatischen Umwandlung von synthetischem [397] M.M. He, S.L. Clugston, J.F. Honek, B.W. Matthews, Determination of the struc- Methylglyoxal in Milchsäure, Biochem. Z. 254 (1932) 332–354. ture of Escherichia coli glyoxalase I suggests a structural basis for differential [367] E. Racker, The mechanism of action of glyoxalase, J. Biol. Chem. 190 (1951) metal activation, Biochemistry 39 (2000) 8719–8727. 685–696. [398] R. Iozef, S. Rahlfs, T. Chang, H. Schirmer, K. Becker, Glyoxalase I of the malarial [368] D.L. Vander Jagt, E. Daub, J.A. Krohn, L.P. Han, Effects of pH and thiols on the ki- parasite Plasmodium falciparum: evidence for subunit fusion, FEBS Lett. 554 netics of yeast glyoxalase I. An evaluation of the random pathway mechanism, (2003) 284–288. Biochemistry 14 (1975) 3669–3675. [399] M. Urscher, S.S. More, R. Alisch, R. Vince, M. Deponte, Tight-binding inhibitors [369] D.L. Vander Jagt, L.P. Han, C.H. Lehman, Kinetic evaluation of substrate specificity efficiently inactivate both reaction centers of monomeric Plasmodium falciparum in the glyoxalase-I-catalyzed disproportionation of -ketoaldehydes, Biochemis- glyoxalase 1, FEBS J. 279 (2012) 2568–2578. try 11 (1972) 3735–3740. [400] M. Urscher, J.M. Przyborski, M. Imoto, M. Deponte, Distinct subcellular localiza- [370] M. Deponte, N. Sturm, S. Mittler, M. Harner, H. Mack, K. Becker, Allosteric cou- tion in the cytosol and apicoplast, unexpected dimerization and inhibition of pling of two different functional active sites in monomeric Plasmodium Plasmodium falciparum glyoxalases, Mol. Microbiol. 76 (2010) 92–103. falciparum glyoxalase I, J. Biol. Chem. 282 (2007) 28419–28430. [401] M. Ridderstrom, A.D. Cameron, T.A. Jones, B. Mannervik, Involvement of an [371] K. Ekwall, B. Mannervik, The stereochemical configuration of the lactoyl group of active-site Zn2+ ligand in the catalytic mechanism of human glyoxalase I, J. S-lactoylglutathionine formed by the action of glyoxalase I from porcine eryth- Biol. Chem. 273 (1998) 21623–21628. rocytes and yeast, Biochim. Biophys. Acta 297 (1973) 297–299. [402] S.L. Clugston, R. Yajima, J.F. Honek, Investigation of metal binding and activation [372] L. Uotila, Purification and characterization of S-2-hydroxyacylglutathione hy- of Escherichia coli glyoxalase I: kinetic, thermodynamic and mutagenesis studies, drolase (glyoxalase II) from human liver, Biochemistry 12 (1973) 3944–3951. Biochem. J. 377 (2004) 309–316. [373] C.E. Griffis, L.H. Ong, L. Buettner, D.J. Creighton, Nonstereospecific substrate [403] N. Sukdeo, S.L. Clugston, E. Daub, J.F. Honek, Distinct classes of glyoxalase I: usage by glyoxalase I, Biochemistry 22 (1983) 2945–2951. metal specificity of the Yersinia pestis, Pseudomonas aeruginosa and Neisseria [374] J.A. Landro, E.J. Brush, J.W. Kozarich, Isomerization of (R)- and (S)- meningitidis enzymes, Biochem. J. 384 (2004) 111–117. glutathiolactaldehydes by glyoxalase I: the case for dichotomous stereo- [404] R.E. Allen, T.W. Lo, P.J. Thornalley, A simplified method for the purification of chemical behavior in a single active site, Biochemistry 31 (1992) 6069–6077. human red blood cell glyoxalase. I. Characteristics, immunoblotting, and inhibi- [375] C. Rae, S.I. O'Donoghue, W.A. Bubb, P.W. Kuchel, Stereospecificity of substrate tor studies, J. Protein Chem. 12 (1993) 111–119. usage by glyoxalase 1: nuclear magnetic resonance studies of kinetics and [405] A.C. Aronsson, G. Tibbelin, B. Mannervik, Purification of glyoxalase I from human hemithioacetal substrate conformation, Biochemistry 33 (1994) 3548–3559. erythrocytes by the use of affinity chromatography and separation of the three [376] A.D. Cameron, B. Olin, M. Ridderstrom, B. Mannervik, T.A. Jones, Crystal structure isoenzymes, Anal. Biochem. 92 (1979) 390–393. of human glyoxalase I—evidence for gene duplication and 3D domain swapping, [406] C.M. Schimandle, D.L. Vander Jagt, Isolation and kinetic analysis of the multiple EMBO J. 16 (1997) 3386–3395. forms of glyoxalase-I from human erythrocytes, Arch. Biochem. Biophys. 195 [377] A.D. Cameron, M. Ridderstrom, B. Olin, B. Mannervik, Crystal structure of human (1979) 261–268. glyoxalase II and its complex with a glutathione thiolester substrate analogue, [407] D.J. Creighton, D.S. Hamilton, Brief history of glyoxalase I and what we have Structure 7 (1999) 1067–1078. learned about metal ion-dependent, enzyme-catalyzed isomerizations, Arch. [378] U. Suttisansanee, J.F. Honek, Bacterial glyoxalase enzymes, Semin. Cell Dev. Biol. Biochem. Biophys. 387 (2001) 1–10. 22 (2011) 285–292. [408] F. Himo, P.E. Siegbahn, Catalytic mechanism of glyoxalase I: a theoretical study, J. [379] A.C. Aronsson, E. Marmstal, B. Mannervik, Glyoxalase I, a zinc metalloenzyme of Am. Chem. Soc. 123 (2001) 10280–10289. mammals and yeast, Biochem. Biophys. Res. Commun. 81 (1978) 1235–1240. [409] U. Richter, M. Krauss, Active site structure and mechanism of human glyoxalase I [380] A.D. Cameron, M. Ridderstrom, B. Olin, M.J. Kavarana, D.J. Creighton, B. —an ab initio theoretical study, J. Am. Chem. Soc. 123 (2001) 6973–6982. Mannervik, Reaction mechanism of glyoxalase I explored by an X-ray crystallo- [410] Y. Lan, T. Lu, P.S. Lovett, D.J. Creighton, Evidence for a (triosephosphate graphic analysis of the human enzyme in complex with a transition state ana- isomerase-like) “catalytic loop” near the active site of glyoxalase I, J. Biol. logue, Biochemistry 38 (1999) 13480–13490. Chem. 270 (1995) 12957–12960. 3264 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266

[411] D.J. Creighton, M. Migliorini, T. Pourmotabbed, M.K. Guha, Optimization of effi- [440] W.H. Habig, M.J. Pabst, W.B. Jakoby, Glutathione S-transferases. The first enzy- ciency in the glyoxalase pathway, Biochemistry 27 (1988) 7376–7384. matic step in mercapturic acid formation, J. Biol. Chem. 249 (1974) 7130–7139. [412] Z. Su, N. Sukdeo, J.F. Honek, 15N-1H HSQC NMR evidence for distinct specificity [441] J.R. Prohaska, H.E. Ganther, Glutathione peroxidase activity of glutathione- of two active sites in Escherichia coli glyoxalase I, Biochemistry 47 (2008) s-transferases purified from rat liver, Biochem. Biophys. Res. Commun. 76 13232–13241. (1976) 437–445. [413] V.A. Campos-Bermudez, N.R. Leite, R. Krog, A.J. Costa-Filho, F.C. Soncini, G. Oliva, [442] A.M. Benson, P. Talalay, J.H. Keen, W.B. Jakoby, Relationship between the soluble A.J. Vila, Biochemical and structural characterization of Salmonella typhimurium glutathione-dependent delta 5-3-ketosteroid isomerase and the glutathione glyoxalase II: new insights into metal ion selectivity, Biochemistry 46 (2007) S-transferases of the liver, Proc. Natl. Acad. Sci. U. S. A. 74 (1977) 158–162. 11069–11079. [443] E. Christ-Hazelhof, D.H. Nugteren, Purification and characterisation of prosta- [414] H. Daiyasu, K. Osaka, Y. Ishino, H. Toh, Expansion of the zinc metallo- glandin endoperoxide D-isomerase, a cytoplasmic, glutathione-requiring en- family of the beta-lactamase fold, FEBS Lett. 503 (2001) 1–6. zyme, Biochim. Biophys. Acta 572 (1979) 43–51. [415] N. Mitic, S.J. Smith, A. Neves, L.W. Guddat, L.R. Gahan, G. Schenk, The catalytic [444] A. Wahllander, S. Soboll, H. Sies, I. Linke, M. Muller, Hepatic mitochondrial and cy- mechanisms of binuclear metallohydrolases, Chem. Rev. 106 (2006) 3338–3363. tosolic glutathione content and the subcellular distribution of GSH-S-transferases, [416] J. O'Young, N. Sukdeo, J.F. Honek, Escherichia coli glyoxalase II is a binuclear FEBS Lett. 97 (1979) 138–140. zinc-dependent metalloenzyme, Arch. Biochem. Biophys. 459 (2007) 20–26. [445] P. Kraus, Resolution, purification and some properties of three glutathione [417] M. Urscher, M. Deponte, Plasmodium falciparum glyoxalase II: Theorell-Chance transferases from rat liver mitochondria, Hoppe Seylers Z. Physiol. Chem. 361 product inhibition patterns, rate-limiting substrate binding via Arg(257)/Lys(260), (1980) 9–15. and unmasking of acid–base catalysis, Biol. Chem. 390 (2009) 1171–1183. [446] J.M. Harris, D.J. Meyer, B. Coles, B. Ketterer, A novel glutathione transferase [418] T.M. Zang, D.A. Hollman, P.A. Crawford, M.W. Crowder, C.A. Makaroff, (13–13) isolated from the matrix of rat liver mitochondria having structural Arabidopsis glyoxalase II contains a zinc/iron binuclear metal center that is es- similarity to class theta enzymes, Biochem. J. 278 (Pt 1) (1991) 137–141. sential for substrate binding and catalysis, J. Biol. Chem. 276 (2001) 4788–4795. [447] S.E. Pemble, A.F. Wardle, J.B. Taylor, Glutathione S-transferase class Kappa: char- [419] O. Schilling, N. Wenzel, M. Naylor, A. Vogel, M. Crowder, C. Makaroff, W. acterization by the cloning of rat mitochondrial GST and identification of a Meyer-Klaucke, Flexible metal binding of the metallo-beta-lactamase domain: human homologue, Biochem. J. 319 (Pt 3) (1996) 749–754. glyoxalase II incorporates iron, , and zinc in vivo, Biochemistry 42 [448] N. Allocati, L. Federici, M. Masulli, C. Di Ilio, Glutathione transferases in bacteria, (2003) 11777–11786. FEBS J. 276 (2009) 58–75. [420] R.E. Allen, T.W. Lo, P.J. Thornalley, Purification and characterisation of glyoxalase [449] C. Di Ilio, A. Aceto, R. Piccolomini, N. Allocati, A. Faraone, L. Cellini, G. Ravagnan, II from human red blood cells, Eur. J. Biochem. 213 (1993) 1261–1267. G. Federici, Purification and characterization of three forms of glutathione trans- [421] A. Wendler, T. Irsch, N. Rabbani, P.J. Thornalley, R.L. Krauth-Siegel, Glyoxalase II ferase from Proteus mirabilis, Biochem. J. 255 (1988) 971–975. does not support methylglyoxal detoxification but serves as a general [450] M. Iizuka, Y. Inoue, K. Murata, A. Kimura, Purification and some properties of trypanothione thioesterase in African trypanosomes, Mol. Biochem. Parasitol. glutathione S-transferase from Escherichia coli B, J. Bacteriol. 171 (1989) 163 (2009) 19–27. 6039–6042. [422] P.A. Cordell, T.S. Futers, P.J. Grant, R.J. Pease, The Human hydroxyacylglutathione [451] D.B. Smith, K.S. Johnson, Single-step purification of polypeptides expressed in hydrolase (HAGH) gene encodes both cytosolic and mitochondrial forms of Escherichia coli as fusions with glutathione S-transferase, Gene 67 (1988) 31–40. glyoxalase II, J. Biol. Chem. 279 (2004) 28653–28661. [452] B. Mannervik, H. Jensson, Binary combinations of four protein subunits with dif- [423] M. Akoachere, R. Iozef, S. Rahlfs, M. Deponte, B. Mannervik, D.J. Creighton, H. ferent catalytic specificities explain the relationship between six basic glutathi- Schirmer, K. Becker, Characterization of the glyoxalases of the malarial parasite one S-transferases in rat liver cytosol, J. Biol. Chem. 257 (1982) 9909–9912. Plasmodium falciparum and comparison with their human counterparts, Biol. [453] P. Reinemer, H.W. Dirr, R. Ladenstein, J. Schaffer, O. Gallay, R. Huber, The Chem. 386 (2005) 41–52. three-dimensional structure of class pi glutathione S-transferase in complex [424] V. Talesa, L. Uotila, M. Koivusalo, G. Principato, E. Giovannini, G. Rosi, Isolation of with glutathione sulfonate at 2.3 A resolution, EMBO J. 10 (1991) 1997–2005. glyoxalase II from two different compartments of rat liver mitochondria. Kinetic [454] X. Ji, P. Zhang, R.N. Armstrong, G.L. Gilliland, The three-dimensional structure of and immunochemical characterization of the enzymes, Biochim. Biophys. Acta a glutathione S-transferase from the mu gene class. Structural analysis of the 993 (1989) 7–11. binary complex of isoenzyme 3–3 and glutathione at 2.2-A resolution, Biochem- [425] F.R. Opperdoes, P. Borst, Localization of nine glycolytic enzymes in a istry 31 (1992) 10169–10184. microbody-like organelle in Trypanosoma brucei: the glycosome, FEBS Lett. 80 [455] I. Sinning, G.J. Kleywegt, S.W. Cowan, P. Reinemer, H.W. Dirr, R. Huber, G.L. (1977) 360–364. Gilliland, R.N. Armstrong, X. Ji, P.G. Board, et al., Structure determination and re- [426] M. Ridderstrom, P. Jemth, A.D. Cameron, B. Mannervik, The active-site residue finement of human alpha class glutathione transferase A1-1, and a comparison tyr-175 in human glyoxalase II contributes to binding of glutathione derivatives, with the Mu and Pi class enzymes, J. Mol. Biol. 232 (1993) 192–212. Biochim. Biophys. Acta 1481 (2000) 344–348. [456] J.E. Ladner, J.F. Parsons, C.L. Rife, G.L. Gilliland, R.N. Armstrong, Parallel evolu- [427] S.L. Chen, W.H. Fang, F. Himo, Reaction mechanism of the binuclear zinc enzyme tionary pathways for glutathione transferases: structure and mechanism of the glyoxalase II — a theoretical study, J. Inorg. Biochem. 103 (2009) 274–281. mitochondrial class kappa enzyme rGSTK1-1, Biochemistry 43 (2004) 352–361. [428] M.K. Guha, D.L. Vander Jagt, D.J. Creighton, Diffusion-dependent rates for the hy- [457] N. Ogino, T. Miyamoto, S. Yamamoto, O. Hayaishi, Prostaglandin endoperoxide E drolysis reaction catalyzed by glyoxalase II from rat erythrocytes, Biochemistry isomerase from bovine vesicular gland microsomes, a glutathione-requiring 27 (1988) 8818–8822. enzyme, J. Biol. Chem. 252 (1977) 890–895. [429] Y. Inoue, A. Kimura, Identification of the structural gene for glyoxalase I from [458] N. Nemoto, H. Gelboin, Assay and properties of glutathione-S-benzo(a) Saccharomyces cerevisiae, J. Biol. Chem. 271 (1996) 25958–25965. pyrene-4,5-oxide transferase, Arch. Biochem. Biophys. 170 (1975) 739–742. [430] Y. Inoue, K. Maeta, W. Nomura, Glyoxalase system in : structure, function, [459] T. Friedberg, P. Bentley, P. Stasiecki, H.R. Glatt, D. Raphael, F. Oesch, The identi- and physiology, Semin. Cell Dev. Biol. 22 (2011) 278–284. fication, solubilization, and characterization of microsome-associated glutathi- [431] M.J. MacLean, L.S. Ness, G.P. Ferguson, I.R. Booth, The role of glyoxalase I in the one S-transferases, J. Biol. Chem. 254 (1979) 12028–12033. detoxification of methylglyoxal and in the activation of the KefB K+ efflux sys- [460] R. Morgenstern, C. Guthenberg, J.W. Depierre, Microsomal glutathione tem in Escherichia coli, Mol. Microbiol. 27 (1998) 563–571. S-transferase. Purification, initial characterization and demonstration that it is [432] O. Brouwers, P.M. Niessen, I. Ferreira, T. Miyata, P.G. Scheffer, T. Teerlink, P. not identical to the cytosolic glutathione S-transferases A, B and C, Eur. J. Schrauwen, M. Brownlee, C.D. Stehouwer, C.G. Schalkwijk, Overexpression of Biochem. 128 (1982) 243–248. glyoxalase-I reduces hyperglycemia-induced levels of advanced glycation end [461] R. Morgenstern, J.W. DePierre, H. Jornvall, Microsomal glutathione transferase. products and oxidative stress in diabetic rats, J. Biol. Chem. 286 (2011) 1374–1380. Primary structure, J. Biol. Chem. 260 (1985) 13976–13983. [433] M. Morcos, X. Du, F. Pfisterer, H. Hutter, A.A. Sayed, P. Thornalley, N. Ahmed, J. [462] T. Yoshimoto, R.J. Soberman, R.A. Lewis, K.F. Austen, Isolation and characteriza- Baynes, S. Thorpe, G. Kukudov, A. Schlotterer, F. Bozorgmehr, R.A. El Baki, D. tion of leukotriene C4 synthetase of rat basophilic leukemia cells, Proc. Natl. Stern, F. Moehrlen, Y. Ibrahim, D. Oikonomou, A. Hamann, C. Becker, M. Zeier, Acad. Sci. U. S. A. 82 (1985) 8399–8403. V. Schwenger, N. Miftari, P. Humpert, H.P. Hammes, M. Buechler, A. Bierhaus, [463] B.K. Lam, J.F. Penrose, G.J. Freeman, K.F. Austen, Expression cloning of a cDNA for M. Brownlee, P.P. Nawroth, Glyoxalase-1 prevents mitochondrial protein modi- human leukotriene C4 synthase, an integral membrane protein conjugating reduced fication and enhances lifespan in Caenorhabditis elegans, Aging Cell 7 (2008) glutathione to leukotriene A4, Proc. Natl. Acad. Sci. U. S. A. 91 (1994) 7663–7667. 260–269. [464] D.J. Welsch, D.P. Creely, S.D. Hauser, K.J. Mathis, G.G. Krivi, P.C. Isakson, Molecu- [434] M. Kawatani, H. Okumura, K. Honda, N. Kanoh, M. Muroi, N. Dohmae, M. Takami, lar cloning and expression of human leukotriene-C4 synthase, Proc. Natl. Acad. M. Kitagawa, Y. Futamura, M. Imoto, H. Osada, The identification of an Sci. U. S. A. 91 (1994) 9745–9749. osteoclastogenesis inhibitor through the inhibition of glyoxalase I, Proc. Natl. [465] R.A. Dixon, R.E. Diehl, E. Opas, E. Rands, P.J. Vickers, J.F. Evans, J.W. Gillard, D.K. Acad. Sci. U. S. A. 105 (2008) 11691–11696. Miller, Requirement of a 5-lipoxygenase-activating protein for leukotriene syn- [435] Y. Xu, X. Chen, Glyoxalase II, a detoxifying enzyme of glycolysis byproduct thesis, Nature 343 (1990) 282–284. methylglyoxal and a target of p63 and p73, is a pro-survival factor of the p53 [466] P.J. Jakobsson, R. Morgenstern, J. Mancini, A. Ford-Hutchinson, B. Persson, Com- family, J. Biol. Chem. 281 (2006) 26702–26713. mon structural features of MAPEG — a widespread superfamily of membrane as- [436] R. Vince, W.B. Wadd, Glyoxalase inhibitors as potential anticancer agents, sociated proteins with highly divergent functions in eicosanoid and glutathione Biochem. Biophys. Res. Commun. 35 (1969) 593–598. metabolism, Protein Sci. 8 (1999) 689–692. [437] S.S. More, R. Vince, Inhibition of glyoxalase I: the first low-nanomolar [467] P.J. Jakobsson, S. Thoren, R. Morgenstern, B. Samuelsson, Identification of human tight-binding inhibitors, J. Med. Chem. 52 (2009) 4650–4656. prostaglandin E synthase: a microsomal, glutathione-dependent, inducible en- [438] S.C. Chauhan, R. Madhubala, Glyoxalase I gene deletion mutants of Leishmania zyme, constituting a potential novel drug target, Proc. Natl. Acad. Sci. U. S. A. donovani exhibit reduced methylglyoxal detoxification, PLoS One 4 (2009) e6805. 96 (1999) 7220–7225. [439] B. Mannervik, P.G. Board, J.D. Hayes, I. Listowsky, W.R. Pearson, Nomenclature for [468] H. Hebert, I. Schmidt-Krey, R. Morgenstern, K. Murata, T. Hirai, K. Mitsuoka, Y. mammalian soluble glutathione transferases, Methods Enzymol. 401 (2005) 1–8. Fujiyoshi, The 3.0 A projection structure of microsomal glutathione transferase M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266 3265

as determined by electron crystallography of p 21212 two-dimensional crystals, [493] J.M. Fernandez-Canon, M.A. Penalva, Characterization of a fungal maleylacetoacetate J. Mol. Biol. 271 (1997) 751–758. isomerase gene and identification of its human homologue, J. Biol. Chem. 273 (1998) [469] P.J. Holm, P. Bhakat, C. Jegerschold, N. Gyobu, K. Mitsuoka, Y. Fujiyoshi, R. 329–337. Morgenstern, H. Hebert, Structural basis for detoxification and oxidative stress [494] A.S. Johansson, B. Mannervik, Human glutathione transferase A3-3, a highly effi- protection in membranes, J. Mol. Biol. 360 (2006) 934–945. cient catalyst of double-bond isomerization in the biosynthetic pathway of ste- [470] H. Ago, Y. Kanaoka, D. Irikura, B.K. Lam, T. Shimamura, K.F. Austen, M. Miyano, roid hormones, J. Biol. Chem. 276 (2001) 33061–33065. Crystal structure of a human membrane protein involved in cysteinyl leukotri- [495] Y. Kanaoka, H. Ago, E. Inagaki, T. Nanayama, M. Miyano, R. Kikuno, Y. Fujii, N. ene biosynthesis, Nature 448 (2007) 609–612. Eguchi, H. Toh, Y. Urade, O. Hayaishi, Cloning and crystal structure of hemato- [471] D. Martinez Molina, A. Wetterholm, A. Kohl, A.A. McCarthy, D. Niegowski, E. poietic prostaglandin D synthase, Cell 90 (1997) 1085–1095. Ohlson, T. Hammarberg, S. Eshaghi, J.Z. Haeggstrom, P. Nordlund, Structural [496] T. Yamada, J. Komoto, K. Watanabe, Y. Ohmiya, F. Takusagawa, Crystal structure basis for synthesis of inflammatory mediators by human leukotriene C4 and possible catalytic mechanism of microsomal prostaglandin E synthase type synthase, Nature 448 (2007) 613–616. 2 (mPGES-2), J. Mol. Biol. 348 (2005) 1163–1176. [472] C. Jegerschold, S.C. Pawelzik, P. Purhonen, P. Bhakat, K.R. Gheorghe, N. Gyobu, K. [497] M.J. Pabst, W.H. Habig, W.B. Jakoby, Glutathione S-transferase A. A novel kinetic Mitsuoka, R. Morgenstern, P.J. Jakobsson, H. Hebert, Structural basis for induced mechanism in which the major reaction pathway depends on substrate concen- formation of the inflammatory mediator prostaglandin E2, Proc. Natl. Acad. Sci. tration, J. Biol. Chem. 249 (1974) 7140–7147. U. S. A. 105 (2008) 11110–11115. [498] B. Gillham, The mechanism of the reaction between glutathione and [473] D.P. Dixon, B.G. Davis, R. Edwards, Functional divergence in the glutathione 1-menaphthyl sulphate catalysed by a glutathione S-transferase from rat liver, transferase superfamily in plants. Identification of two classes with putative Biochem. J. 135 (1973) 797–804. functions in redox homeostasis in Arabidopsis thaliana, J. Biol. Chem. 277 [499] I. Jakobson, M. Warholm, B. Mannervik, Multiple inhibition of glutathione (2002) 30859–30869. S-transferase A from rat liver by glutathione derivatives: kinetic analysis [474] D.P. Dixon, R. Edwards, Glutathione transferases, Arabidopsis Book 8 (2010) supporting a steady-state random sequential mechanism, Biochem. J. 177 e0131. (1979) 861–868. [475] F. Morel, C. Rauch, E. Petit, A. Piton, N. Theret, B. Coles, A. Guillouzo, Gene [500] I. Jakobson, M. Warholm, B. Mannervik, The binding of substrates and a product and protein characterization of the human glutathione S-transferase kappa of the enzymatic reaction to glutathione S-transferase A, J. Biol. Chem. 254 and evidence for a peroxisomal localization, J. Biol. Chem. 279 (2004) (1979) 7085–7089. 16246–16253. [501] W.J. Chen, G.F. Graminski, R.N. Armstrong, Dissection of the catalytic mechanism [476] D. Sheehan, G. Meade, V.M. Foley, C.A. Dowd, Structure, function and evolution of isozyme 4–4 of glutathione S-transferase with alternative substrates, Bio- of glutathione transferases: implications for classification of non-mammalian chemistry 27 (1988) 647–654. members of an ancient enzyme superfamily, Biochem. J. 360 (2001) 1–16. [502] V.L. Schramm, R. McCluskey, F.A. Emig, G. Litwack, Kinetic studies and active [477] M. Deponte, K. Becker, Glutathione S-transferase from malarial parasites: struc- site-binding properties of glutathione S-transferase using spin-labeled glutathi- tural and functional aspects, Methods Enzymol. 401 (2005) 241–253. one, a product analogue, J. Biol. Chem. 259 (1984) 714–722. [478] P.K. Fyfe, G.D. Westrop, A.M. Silva, G.H. Coombs, W.N. Hunter, Leishmania TDR1 [503] G. McManus, M. Costa, A. Canals, M. Coll, T.J. Mantle, Site-directed mutagenesis structure, a unique trimeric glutathione transferase capable of deglutathionylation of mouse glutathione transferase P1-1 unlocks masked cooperativity, introduces and antimonial prodrug activation, Proc. Natl. Acad. Sci. U. S. A. 109 (2012) a novel mechanism for ‘ping pong’ kinetic behaviour, and provides further struc- 11693–11698. tural evidence for participation of a water molecule in proton abstraction from [479] N. Allocati, L. Federici, M. Masulli, C. Di Ilio, Distribution of glutathione transfer- glutathione, FEBS J. 278 (2011) 273–281. ases in Gram-positive bacteria and Archaea, Biochimie 94 (2012) 588–596. [504] Z.R. Zhang, M. Bai, X.Y. Wang, J.M. Zhou, S. Perrett, “Restoration” of glutathione [480] L. Xun, S.M. Belchik, R. Xun, Y. Huang, H. Zhou, E. Sanchez, C. Kang, P.G. Board, transferase activity by single-site mutation of the yeast prion protein Ure2, J. S-Glutathionyl-(chloro)hydroquinone reductases: a novel class of glutathione Mol. Biol. 384 (2008) 641–651. transferases, Biochem. J. 428 (2010) 419–427. [505] G. Xiao, S. Liu, X. Ji, W.W. Johnson, J. Chen, J.F. Parsons, W.J. Stevens, G.L. [481] X. Ji, R.N. Armstrong, G.L. Gilliland, Snapshots along the reaction coordinate of an Gilliland, R.N. Armstrong, First-sphere and second-sphere electrostatic effects SNAr reaction catalyzed by glutathione transferase, Biochemistry 32 (1993) in the active site of a class mu gluthathione transferase, Biochemistry 35 12949–12954. (1996) 4753–4765. [482] X. Ji, W.W. Johnson, M.A. Sesay, L. Dickert, S.M. Prasad, H.L. Ammon, R.N. [506] A.S. Johansson, B. Mannervik, Active-site residues governing high steroid isom- Armstrong, G.L. Gilliland, Structure and function of the xenobiotic substrate bind- erase activity in human glutathione transferase A3-3, J. Biol. Chem. 277 (2002) ing site of a glutathione S-transferase as revealed by X-ray crystallographic analysis 16648–16654. of product complexes with the diastereomers of 9-(S-glutathionyl)-10- [507] N.E. Labrou, L.V. Mello, Y.D. Clonis, Functional and structural roles of the hydroxy-9,10-dihydrophenanthrene, Biochemistry 33 (1994) 1043–1052. glutathione-binding residues in maize (Zea mays) glutathione S-transferase I, [483] A. Robinson, G.A. Huttley, H.S. Booth, P.G. Board, Modelling and bioinformatics Biochem. J. 358 (2001) 101–110. studies of the human Kappa-class glutathione transferase predict a novel third glu- [508] S. Liu, P. Zhang, X. Ji, W.W. Johnson, G.L. Gilliland, R.N. Armstrong, Contribution tathione transferase family with similarity to prokaryotic 2-hydroxychromene- of tyrosine 6 to the catalytic mechanism of isoenzyme 3–3 of glutathione 2-carboxylate isomerases, Biochem. J. 379 (2004) 541–552. S-transferase, J. Biol. Chem. 267 (1992) 4296–4299. [484] . A. Rinaldo-Matthis, S. Ahmad, A. Wetterholm, P. Lachmann, R. Morgenstern, [509] M. Lo Bello, A.J. Oakley, A. Battistoni, A.P. Mazzetti, M. Nuccetelli, G. J.Z. Haeggstrom, Pre-steady-state kinetic characterization of thiolate anion Mazzarese,J.Rossjohn,M.W.Parker,G.Ricci, Multifunctional role of Tyr 108 formation in human leukotriene C(4) synthase, Biochemistry 51 (2012) in the catalytic mechanism of human glutathione transferase P1-1. Crystallo- 848–856. graphic and kinetic studies on the Y108F mutant enzyme, Biochemistry 36 [485] A. Rinaldo-Matthis, A. Wetterholm, D. Martinez Molina, J. Holm, D. Niegowski, E. (1997) 6207–6217. Ohlson, P. Nordlund, R. Morgenstern, J.Z. Haeggstrom, Arginine 104 is a key catalyt- [510] P.L. Pettersson, A.S. Johansson, B. Mannervik, Transmutation of human glutathi- ic residue in leukotriene C4 synthase, J. Biol. Chem. 285 (2010) 40771–40776. one transferase A2-2 with peroxidase activity into an efficient steroid isomerase, [486] H. Saino, Y. Ukita, H. Ago, D. Irikura, A. Nisawa, G. Ueno, M. Yamamoto, Y. J. Biol. Chem. 277 (2002) 30019–30022. Kanaoka, B.K. Lam, K.F. Austen, M. Miyano, The catalytic architecture of leukotri- [511] A. Vararattanavech, A.J. Ketterman, A functionally conserved basic residue in ene C4 synthase with two arginine residues, J. Biol. Chem. 286 (2011) glutathione transferases interacts with the glycine moiety of glutathione and 16392–16401. is pivotal for , Biochem. J. 406 (2007) 247–256. [487] L.S. Busenlehner, J. Alander, C. Jegerscohld, P.J. Holm, P. Bhakat, H. Hebert, R. [512] L. Prade, R. Huber, T.H. Manoharan, W.E. Fahl, W. Reuter, Structures of class pi Morgenstern, R.N. Armstrong, Location of substrate binding sites within the in- glutathione S-transferase from human placenta in complex with substrate, tegral membrane protein microsomal glutathione transferase-1, Biochemistry transition-state analogue and inhibitor, Structure 5 (1997) 1287–1295. 46 (2007) 2812–2822. [513] C.S. Allardyce, P.D. McDonagh, L.Y. Lian, C.R. Wolf, G.C. Roberts, The role of [488] E.B. Prage, S.C. Pawelzik, L.S. Busenlehner, K. Kim, R. Morgenstern, P.J. Jakobsson, tyrosine-9 and the C-terminal helix in the catalytic mechanism of Alpha-class R.N. Armstrong, Location of inhibitor binding sites in the human inducible pros- glutathione S-transferases, Biochem. J. 343 (Pt 3) (1999) 525–531. taglandin E synthase, MPGES1, Biochemistry 50 (2011) 7684–7693. [514] T.K. Hitchens, B. Mannervik, G.S. Rule, Disorder-to-order transition of the active [489] S.C. Pawelzik, N.R. Uda, L. Spahiu, C. Jegerschold, P. Stenberg, H. Hebert, R. site of human class Pi glutathione transferase, GST P1-1, Biochemistry 40 (2001) Morgenstern, P.J. Jakobsson, Identification of key residues determining species 11660–11669. differences in inhibitor binding of microsomal prostaglandin E synthase-1, J. [515] A.J. Oakley, M. Lo Bello, G. Ricci, G. Federici, M.W. Parker, Evidence for an Biol. Chem. 285 (2010) 29254–29261. induced-fit mechanism operating in pi class glutathione transferases, Biochem- [490] E.B. Prage, R. Morgenstern, P.J. Jakobsson, D.F. Stec, M.W. Voehler, R.N. Armstrong, istry 37 (1998) 9912–9917. Observation of two modes of inhibition of human microsomal prostaglandin [516] L. Stella, M. Nicotra, G. Ricci, N. Rosato, E.E. Di Iorio, Molecular dynamics simula- E synthase 1 by the cyclopentenone 15-deoxy-Delta(12,14)-prostaglandin J(2), tions of human glutathione transferase P1-1: analysis of the induced-fit mecha- Biochemistry 51 (2012) 2348–2356. nism by GSH binding, Proteins 37 (1999) 1–9. [491] A. Hamza, M. Tong, M.D. AbdulHameed, J. Liu, A.C. Goren, H.H. Tai, C.G. Zhan, Un- [517] B. Wang, Y. Peng, T. Zhang, J. Ding, Crystal structures and kinetic studies of derstanding microscopic binding of human microsomal prostaglandin E human Kappa class glutathione transferase provide insights into the catalytic synthase-1 (mPGES-1) trimer with substrate PGH2 and cofactor GSH: insights mechanism, Biochem. J. 439 (2011) 215–225. from computational scanning and site-directed mutagenesis, J. Phys. [518] G.F. Graminski, P.H. Zhang, M.A. Sesay, H.L. Ammon, R.N. Armstrong, Formation Chem. B 114 (2010) 5605–5616. of the 1-(S-glutathionyl)-2,4,6-trinitrocyclohexadienate anion at the active [492] P.G. Board, R.T. Baker, G. Chelvanayagam, L.S. Jermiin, Zeta, a novel class of glu- site of glutathione S-transferase: evidence for enzymic stabilization of sigma- tathione transferases in a range of species from plants to humans, Biochem. J. complex intermediates in nucleophilic aromatic substitution reactions, Bio- 328 (Pt 3) (1997) 929–935. chemistry 28 (1989) 6252–6258. 3266 M. Deponte / Biochimica et Biophysica Acta 1830 (2013) 3217–3266

[519] P.G. Board, M. Coggan, G. Chelvanayagam, S. Easteal, L.S. Jermiin, G.K. Schulte, [538] T. Fang, D.F. Li, N.Y. Zhou, Identification and clarification of the role of key active D.E. Danley, L.R. Hoth, M.C. Griffor, A.V. Kamath, M.H. Rosner, B.A. Chrunyk, site residues in bacterial glutathione S-transferase zeta/maleylpyruvate isomer- D.E. Perregaux, C.A. Gabel, K.F. Geoghegan, J. Pandit, Identification, characteriza- ase, Biochem. Biophys. Res. Commun. 410 (2011) 452–456. tion, and crystal structure of the Omega class glutathione transferases, J. Biol. [539] R. Thom, D.P. Dixon, R. Edwards, D.J. Cole, A.J. Lapthorn, The structure of a zeta Chem. 275 (2000) 24798–24806. class glutathione S-transferase from Arabidopsis thaliana: characterisation of a [520] L. Bousset, V. Redeker, P. Decottignies, S. Dubois, P. Le Marechal, R. Melki, Struc- GST with novel active-site architecture and a putative role in tyrosine catabo- tural characterization of the fibrillar form of the yeast Saccharomyces cerevisiae lism, J. Mol. Biol. 308 (2001) 949–962. prion Ure2p, Biochemistry 43 (2004) 5022–5032. [540] M. Marsh, D.K. Shoemark, A. Jacob, C. Robinson, B. Cahill, N.Y. Zhou, P.A. [521] M. Bai, J.M. Zhou, S. Perrett, The yeast prion protein Ure2 shows glutathione per- Williams, A.T. Hadfield, Structure of bacterial glutathione-S-transferase maleyl oxidase activity in both native and fibrillar forms, J. Biol. Chem. 279 (2004) pyruvate isomerase and implications for mechanism of isomerisation, J. Mol. 50025–50030. Biol. 384 (2008) 165–177. [522] C.M. Bruns, I. Hubatsch, M. Ridderstrom, B. Mannervik, J.A. Tainer, Human gluta- [541] S. Seltzer, M. Lin, Maleylacetone cis–trans-isomerase. Mechanism of the interac- thione transferase A4-4 crystal structures and mutagenesis reveal the basis of tion of coenzyme glutathione and substrate maleylacetone in the presence and high catalytic efficiency with toxic lipid peroxidation products, J. Mol. Biol. 288 absence of enzyme, J. Am. Chem. Soc. 101 (1979) 3091–3097. (1999) 427–439. [542] R. Morgenstern, J.W. DePierre, Microsomal glutathione transferase. Purification in [523] D.S. Hamilton, X. Zhang, Z. Ding, I. Hubatsch, B. Mannervik, K.N. Houk, B. Ganem, unactivated form and further characterization of the activation process, substrate D.J. Creighton, Mechanism of the glutathione transferase-catalyzed conversion specificity and amino acid composition, Eur. J. Biochem. 134 (1983) 591–597. of antitumor 2-crotonyloxymethyl-2-cycloalkenones to GSH adducts, J. Am. [543] R. Svensson, J. Alander, R.N. Armstrong, R. Morgenstern, Kinetic characterization Chem. Soc. 125 (2003) 15049–15058. of thiolate anion formation and chemical catalysis of activated microsomal glu- [524] A.J. Oakley, J. Rossjohn, M. Lo Bello, A.M. Caccuri, G. Federici, M.W. Parker, The tathione transferase 1, Biochemistry 43 (2004) 8869–8877. three-dimensional structure of the human Pi class glutathione transferase [544] R. Morgenstern, R. Svensson, B.A. Bernat, R.N. Armstrong, Kinetic analysis of the P1-1 in complex with the inhibitor ethacrynic acid and its glutathione conju- slow ionization of glutathione by microsomal glutathione transferase MGST1, gate, Biochemistry 36 (1997) 576–585. Biochemistry 40 (2001) 3378–3384. [525] J. Chen, R.N. Armstrong, Stereoselective catalysis of a retro-Michael reaction by [545] J. Alander, J. Lengqvist, P.J. Holm, R. Svensson, P. Gerbaux, R.H. Heuvel, H. Hebert, class mu glutathione transferases. Consequences for the internal distribution W.J. Griffiths, R.N. Armstrong, R. Morgenstern, Microsomal glutathione transfer- of products in the active site, Chem. Res. Toxicol. 8 (1995) 580–585. ase 1 exhibits one-third-of-the-sites-reactivity towards glutathione, Arch. [526] Y. Gu, J. Guo, A. Pal, S.S. Pan, P. Zimniak, S.V. Singh, X. Ji, Crystal structure of Biochem. Biophys. 487 (2009) 42–48. human glutathione S-transferase A3-3 and mechanistic implications for its [546] S. He, Y. Wu, D. Yu, L. Lai, Microsomal prostaglandin E synthase-1 exhibits high steroid isomerase activity, Biochemistry 43 (2004) 15673–15679. one-third-of-the-sites reactivity, Biochem. J. 440 (2011) 13–21. [527] K. Tars, B. Olin, B. Mannervik, Structural basis for featuring of steroid isomer- [547] T. Izumi, Z. Honda, N. Ohishi, S. Kitamura, S. Tsuchida, K. Sato, T. Shimizu, Y. ase activity in alpha class glutathione transferases, J. Mol. Biol. 397 (2010) Seyama, Solubilization and partial purification of leukotriene C4 synthase from 332–340. guinea-pig lung: a microsomal enzyme with high specificity towards 5,6- [528] M. Calvaresi, M. Stenta, M. Garavelli, P. Altoé, A. Bottoni, Computational evi- epoxide leukotriene A4, Biochim. Biophys. Acta 959 (1988) 305–315. dence for the catalytic mechanism of human glutathione S-transferase A3-3: a [548] T. Hammarberg, M. Hamberg, A. Wetterholm, H. Hansson, B. Samuelsson, J.Z. QM/MM investigation, ACS Catal. 2 (2012) 280–286. Haeggstrom, Mutation of a critical arginine in microsomal prostaglandin E [529] S.C. Khojasteh-Bakht, S.D. Nelson, W.M. Atkins, Glutathione S-transferase cata- synthase-1 shifts the isomerase activity to a reductase activity that converts lyzes the isomerization of (R)-2-hydroxymenthofuran to mintlactones, Arch. prostaglandin H2 into prostaglandin F2alpha, J. Biol. Chem. 284 (2009) 301–305. Biochem. Biophys. 370 (1999) 59–65. [549] R.C. Fahey, R.M. Buschbacher, G.L. Newton, The evolution of glutathione metab- [530] Z.R. Wu, S. Ebrahimian, M.E. Zawrotny, L.D. Thornburg, G.C. Perez-Alvarado, P. olism in phototrophic microorganisms, J. Mol. Evol. 25 (1987) 81–88. Brothers, R.M. Pollack, M.F. Summers, Solution structure of 3-oxo-delta5-steroid [550] R.N. Ondarza, J.L. Rendon, M. Ondarza, Glutathione reductase in evolution, J. Mol. isomerase, Science 276 (1997) 415–418. Evol. 19 (1983) 371–375. [531] R. Bjornestedt, G. Stenberg, M. Widersten, P.G. Board, I. Sinning, T.A. Jones, B. [551] A.R. Sundquist, R.C. Fahey, Evolution of antioxidant mechanisms: thiol- Mannervik, Functional significance of arginine 15 in the active site of human dependent peroxidases and thioltransferase among procaryotes, J. Mol. Evol. class alpha glutathione transferase A1-1, J. Mol. Biol. 247 (1995) 765–773. 29 (1989) 429–435. [532] H. Chen, M.R. Juchau, Recombinant human glutathione S-transferases catalyse [552] J. Hu, L. Dong, C.E. Outten, The redox environment in the mitochondrial enzymic isomerization of 13-cis-retinoic acid to all-trans-retinoic acid in vitro, intermembrane space is maintained separately from the cytosol and matrix, J. Biochem. J. 336 (Pt 1) (1998) 223–226. Biol. Chem. 283 (2008) 29126–29134. [533] I.R. Jowsey, A.M. Thomson, J.U. Flanagan, P.R. Murdock, G.B. Moore, D.J. Meyer, G.J. [553] K. Kojer, M. Bien, H. Gangel, B. Morgan, T.P. Dick, J. Riemer, Glutathione redox Murphy, S.A. Smith, J.D. Hayes, Mammalian class Sigma glutathione S-transferases: potential in the mitochondrial intermembrane space is linked to the cytosol catalytic properties and tissue-specific expression of human and rat and impacts the Mia40 redox state, EMBO J. 31 (2012) 3169–3182. GSH-dependent prostaglandin D2 synthases, Biochem. J. 359 (2001) 507–516. [554] H. Ostergaard, C. Tachibana, J.R. Winther, Monitoring disulfide bond formation [534] E. Pinzar, M. Miyano, Y. Kanaoka, Y. Urade, O. Hayaishi, Structural basis of hema- in the eukaryotic cytosol, J. Cell Biol. 166 (2004) 337–345. topoietic prostaglandin D synthase activity elucidated by site-directed muta- [555] B.M. Dixon, S.H. Heath, R. Kim, J.H. Suh, T.M. Hagen, Assessment of endoplasmic genesis, J. Biol. Chem. 275 (2000) 31239–31244. reticulum glutathione redox status is confounded by extensive ex vivo oxida- [535] T. Inoue, D. Irikura, N. Okazaki, S. Kinugasa, H. Matsumura, N. Uodome, M. tion, Antioxid. Redox Signal. 10 (2008) 963–972. Yamamoto, T. Kumasaka, M. Miyano, Y. Kai, Y. Urade, Mechanism of metal acti- [556] C. Hwang, A.J. Sinskey, H.F. Lodish, Oxidized redox state of glutathione in the en- vation of human hematopoietic prostaglandin D synthase, Nat. Struct. Biol. 10 doplasmic reticulum, Science 257 (1992) 1496–1502. (2003) 291–296. [557] D.P. Jones, J.L. Carlson, P.S. Samiec, P. Sternberg Jr., V.C. Mody Jr., R.L. Reed, L.A. [536] P.G. Board, M.C. Taylor, M. Coggan, M.W. Parker, H.B. Lantum, M.W. Anders, Clar- Brown, Glutathione measurement in human plasma. Evaluation of sample col- ification of the role of key active site residues of glutathione transferase lection, storage and derivatization conditions for analysis of dansyl derivatives zeta/maleylacetoacetate isomerase by a new spectrophotometric technique, by HPLC, Clin. Chim. Acta 275 (1998) 175–184. Biochem. J. 374 (2003) 731–737. [558] A. Echalier, X. Trivelli, C. Corbier, N. Rouhier, O. Walker, P. Tsan, J.P. Jacquot, A. [537] K. Anandarajah, P.M. Kiefer Jr., B.S. Donohoe, S.D. Copley, Recruitment of a dou- Aubry, I. Krimm, J.M. Lancelin, Crystal structure and solution NMR dynamics of ble bond isomerase to serve as a reductive dehalogenase during biodegradation a D (type II) peroxiredoxin glutaredoxin and thioredoxin dependent: a new in- of pentachlorophenol, Biochemistry 39 (2000) 5303–5311. sight into the peroxiredoxin oligomerism, Biochemistry 44 (2005) 1755–1767.