Our current knowledge of the isomerase glyoxalase I and the thioesterase glyoxalase II is based on a variety of prokaryotic and eukaryotic (model) systems with an emphasis on human glyoxalases. During the last decade, important insights on glyoxalase catalysis and structure–function relationships have also been obtained from parasitic protists. These organisms, including kinetoplastid and apicomplexan parasites, are particularly interesting, both because of their relevance as pathogens and because of their phylogenetic diversity and host–parasite co-evolution which has led to specialized organellar and metabolic adaptations. Accordingly, the glyoxalase repertoire and properties vary significantly among parasitic protists of different major eukaryotic lineages (and even between closely related organisms). For example, several protists have an insular or non-canonical glyoxalase. Furthermore, the structures and the substrate specificities of glyoxalases display drastic variations. The aim of the present review is to highlight such differences as well as similarities between the glyoxalases of parasitic protists and to emphasize the power of comparative studies for gaining insights into fundamental principles and alternative glyoxalase functions.
Introduction: the glyoxalase system
Last year was the centennial of the discovery by Neuberg, Dakin and Dudley of an often neglected enzymatic activity: the conversion of methylglyoxal (CH3COCHO) into lactic acid. The activity was initially attributed to a single enzyme termed glyoxalase, but turned out later to depend on the isomerase Glo1 (glyoxalase I), the thioesterase Glo2 (glyoxalase II) and reduced glutathione (GSH) as a coenzyme [1–5] (Figure 1). Glo1, Glo2 and GSH together compose the glyoxalase system, which forms one branch of the highly versatile glutathione system in almost every eukaryote . The major physiological role of the glyoxalase system in mammals is to decrease or circumvent the formation of glycolysis-derived AGEs (advanced glycation end-products), with implications for many pathophysiological conditions such as diabetes and cancer [1,2].
Overview of the glyoxalase system comprising the isomerase Glo1, the thioesterase Glo2 and GSH as a coenzyme
However, depending on the organism investigated, there are also sources for 2-oxoaldehydes other than glycolysis [4–6], and glyoxalases and the metabolites of the glyoxalase system can have functions beyond simple methylglyoxal detoxification. Such functions include, e.g., Glo1-dependent osteoclastogenesis of mouse macrophages , methylglyoxal-dependent osmoregulatory and transcriptional regulatory processes in yeast [8,9], and the S-lactoylglutathione-dependent efflux of potassium in Escherichia coli [10,11]. Moreover, from an evolutionary perspective, glutathione metabolism is probably much younger than glycolysis and the formation of methylglyoxal . Hence it is not surprising that alternative 2-oxoaldehyde-metabolizing enzymes, such as methylglyoxal reductase or GSH-independent glyoxalase III, are found in prokaryotes and eukaryotes [6,12–15]. In summary, we still have a quite incomplete understanding of the physiological functions of 2-oxoaldehydes and glyoxalases in many organisms.
Why parasitic protists are interesting
The diversity of eukaryotes and their biology is often misjudged because of our visual macroscopic perception of life. All animals, together with yeast and other fungi, actually belong to just one phylogenetic lineage of eukaryotes: the opisthokonts. The ‘remaining’ eukaryotes are currently grouped into five or more additional major lineages or supergroups. These lineages comprise predominantly unicellular organisms, so-called protists, which therefore account for the vast diversity of eukaryotes [16–18].
Many protists of each major eukaryotic lineage are medically and socioeconomically important parasites . Parasitism is in fact an extremely prevalent and successful survival strategy among eukaryotes and has optimized and shaped the genomes of both the parasites and their hosts. Hence parasitic protists are important study objects, not only because of their relevance as pathogens, but also because of their phylogenetic diversity and host–parasite co-evolution resulting in unique protein structure–function relationships. In the following sections, I want to demonstrate the power of comparative studies on parasitic protists for gaining insights into the diversity of glyoxalases.
Kinetoplastids: same same, but different
Kinetoplastid parasites belong to the supergroup of excavates and comprise the causative agents of trypanosomiases and leishmaniases . The parasites cycle between an insect and a vertebrate host and share several metabolic peculiarities. For example, glycolysis-derived methylglyoxal is formed in peroxisome-like organelles, termed glycosomes, to which all but the last two glycolytic enzymes are targeted [20,21]. Furthermore, trypanothione [T(SH)2≡N1,N8-bis(glutathionyl)spermidine] replaces GSH as the most abundant thiol [22–24]. This variation has drastic consequences. On the one hand, the formation of an intramolecular disulfide bond in trypanothione is entropically favoured in comparison with the formation of glutathione disulfide. In addition, trypanothione is more reactive because of a significantly lower thiol pKa value of 7.4, has an overall positive instead of negative charge and is much bulkier than GSH [4,25]. Glo1 and Glo2 isoforms of kinetoplastid parasites consequently have highly altered substrate-binding sites. First, because of the spermidine moiety, the substrate-binding sites are partially neutral or negatively (instead of positively) charged. Secondly, the binding sites of kinetoplastid glyoxalases are much wider to accommodate the additional spermidine and second glutathione moiety [23,26–28]. These structural changes result in high specificities of the kinetoplastid Glo1 and Glo2 isoforms for their trypanothione (or glutathionylspermidine) substrates as compared with glutathione [26–35]. Whether it makes sense to exploit these structural alterations for inhibitor development depends on the species investigated.
(i) African trypanosomes, such as Trypanosoma brucei, are transmitted by tsetse flies and cause nagana cattle disease or deadly sleeping sickness in humans. The vertebrate stages of T. brucei live extracellularly in the bloodstream and rely on glycolytic ATP production [21,23]. The discovery that T. brucei neither has a Glo1 isoform nor produces significant amounts of D-lactate was therefore quite surprising [35,36]. How does T. brucei maintain normal methylglyoxal levels despite extreme glycolytic fluxes ? As reviewed previously , this question remains open, since experimental data on alternative enzymatic activities are partially contradictory [35,36] or lacking . Furthermore, the effects of the organellar compartmentalization of glycolysis on methylglyoxal metabolism and the formation of AGEs have not been addressed to date. Another interesting aspect is that the genome of T. brucei encodes two Glo2 isoforms: a functional isoform, which was shown to be non-essential for parasite survival, and an uncharacterized isoform that might be inactive with standard substrates because of an unusual insertion [29,35]. The existence of such insular and/or non-canonical Glo2 isoforms obviously raises general questions regarding the physiological role of this enzyme class. It is well known that Glo2 isoforms, including the enzyme from T. brucei, hydrolyse a variety of thioester substrates. Hence these enzymes probably have alternative functions that are unrelated to the Glo1-dependent metabolism of 2-oxoaldehydes [3–5,35].
(ii) Trypanosoma cruzi causes Chagas’ disease in Latin America and is transferred to humans by the faeces of blood-feeding triatomine bugs (or by blood transfusion and transplants). Similar to Leishmania spp., the vertebrate stages of T. cruzi multiply intracellularly after the infection of macrophages and other cell types. T. cruzi furthermore travels extracellularly in the bloodstream and parasitizes the heart and colon, which can lead to lethal complications upon chronic infection. In contrast with T. brucei, T. cruzi produces D-lactate and has a functional trypanothione-dependent Glo1 and Glo2 isoform [33,36]. The Glo1 isoform was suggested to be dual-targeted to the cytosol and the parasite mitochondrion . In order to explain the glyoxalase-dependent formation of D-lactate in T. cruzi, one has to postulate an efficient diffusion or transport of glycolysis-derived methylglyoxal across the glycosome membrane. This also holds true for all Leishmania glyoxalases investigated, none of which appears to localize to glycosomes or to have a peroxisome-targeting signal. Whether one or both of the T. cruzi glyoxalases are essential for parasite survival remains to be determined.
(iii) Leishmania parasites are transmitted by sandflies and are often grouped according to the global distribution or phenotype upon infection. In humans, these phenotypes range from cutaneous lesions (e.g. Leishmania major) to defacing mucocutaneous lesions (e.g. Leishmania braziliensis) and disseminating potentially lethal visceral alterations and anaemia (Leishmania donovani and Leishmania infantum). As reviewed previously, a variety of Leishmania spp. was shown to produce D-lactate from glucose and methylglyoxal, and several trypanothione-dependent Glo1 and Glo2 isoforms from Leishmania were identified and analysed subsequently [23,37]. Studies of L. major Glo1 revealed that not only the substrate-binding site but also the catalytic metal centre differs from human Glo1 . Recombinant L. major Glo1 was most active with Ni2+ instead of Zn2+  and therefore shares several properties with bacterial Glo1 isoforms . Similar results for Glo1 from T. cruzi  suggest that the unusual Ni2+ preference could be a common feature in kinetoplastida and support the theory that their Glo1 isoforms are of prokaryotic ancestry  (in contrast with their Glo2 isoforms ). Are the metal centres of kinetoplastid Glo1 isoforms functional in vivo? This question appears to be valid considering a high local parasitaemia and a presumably limited bioavailability of Ni2+ in vertebrates and sandflies. The physiological relevance of kinetoplastid glyoxalases for methylglyoxal detoxification and their suitability as potential drug targets was furthermore challenged by a study on L. infantum insect stage cultures . However, genetic experiments on the Glo1 isoform from the closely related parasite L. donovani suggest that the enzyme is essential and contributes to the detoxification of (exogenous) methylglyoxal . Thus L. donovani Glo1 could indeed be a novel drug target with a highly altered substrate-binding site compared with the human enzyme [31,38]. Previous analyses on recombinant Glo1 from T. cruzi and L. major revealed Ki values for the competitive inhibitor S-4-bromobenzylglutathionylspermidine of approximately 5 μM  and 0.5 μM  respectively and might serve as a starting point for the rational development of specific inhibitors. Whether Glo2 isoforms from Leishmania are essential remains to be studied.
In summary, kinetoplastid parasites demonstrate a highly variable glyoxalase repertoire. Their glyoxalases have structures, metal ion requirements and substrate specificities different from those of the human isoforms. These drastic alterations result from a peculiar molecular evolution because of an unusual trypanothione metabolism and an early prokaryotic gene transfer.
Apicomplexan parasites belong to the major eukaryotic lineage of chromalveolates . They comprise numerous important human pathogens, such as the causative agents of malaria and toxoplasmosis Plasmodium and Toxoplasma respectively, as well as economically relevant animal parasites, such as Babesia, Theileria, Eimeria, Neospora and Cryptosporidium. The life cycles of apicomplexan parasites are highly evolved and complex. Two major transmission routes exist between vertebrate hosts: coccidia are directly taken up after ingestion of infected tissue or faecal contaminations (e.g. Toxoplasma or Eimeria), whereas blood parasites such as Plasmodium or Babesia are transferred by arthropod vectors, usually mosquitos or ticks. The name ‘apicomplexa’ refers to an apical organellar complex. This apparatus is required in all these intracellular parasites for host cell invasion. Another interesting feature of apicomplexan parasites is the apicoplast, a chloroplast-like organelle that was acquired by secondary endosymbiosis . As far as glyoxalases of apicomplexan parasites are concerned, experimental research has so far focused on the most important human malaria parasite Plasmodium falciparum.
P. falciparum produces significant amounts of D-lactate from glucose . The parasite genome encodes four glyoxalases, a cytosolic Glo1 and Glo2 isoform, and, on the other hand, a Glo1-like protein and a second Glo2 that both localize to the apicoplast [3,41]. The cytosolic Glo1 isoform and both Glo2 isoforms are functional enzymes [3,41–45]. In contrast, the sequence of the Glo1-like protein is quite unusual, and the recombinant protein was inactive in a standard enzyme assay [3,43]. Although it is reasonable to regard the cytosolic enzymes as components of a functional glyoxalase system, there are two possibilities for the apicoplast glyoxalases. Either they form an enzyme couple with an unorthodox substrate that is accepted by the non-canonical Glo1-like protein, or they have independent unidentified functions [3,41]. In the latter scenario, the apicoplast Glo2 would be another example for an orphan glyoxalase, similar to the insular, predominantly mitochondrial, Glo2 isoforms from T. brucei, humans, yeast and many other organisms [3,4,35]. In addition to this completely unresolved topic, there are several interesting aspects regarding the cytosolic P. falciparum glyoxalases.
In contrast with human homodimeric Glo1 with its two structurally identical active sites, the enzyme from P. falciparum is monomeric and has two different active sites with similar kcat values, but distinct Km values. Moreover, both active sites of P. falciparum Glo1 adopt two discrete conformations and are allosterically coupled in a substrate concentration-dependent manner [3,44]. What might be the reasons for these peculiar properties? Malaria parasites encounter highly variable glucose concentrations during their life cycle. The allosteric regulation of the high-activity and the high-affinity conformation of the enzyme might therefore be an adaptation to altered methylglyoxal fluxes [3,44]. However, this hypothesis needs to be tested experimentally. Regarding the quaternary structure, the monomeric enzyme probably resulted from a second gene-duplication event in the course of evolution [3,42]. The advantage of such an event, which presumably occurred independently in non-related organisms, is that it allows structural adaptations at one active site, maybe for an optimized turnover of an alternative substrate, while maintaining the original enzyme activity at the other active site [3,44]. However, which alternative substrate or regulator might preferentially bind to the second active site of P. falciparum Glo1 remains to be analysed.
As far as the cytosolic Glo2 isoform of P. falciparum is concerned, mutational analyses demonstrated not only the importance of conserved residues at the glutathione-binding site, but also revealed, for the first time, the acid–base catalysis of this enzyme class . P. falciparum Glo2 works with a so-called Theorell–Chance Bi Bi mechanism, which means that the release of glutathione or the binding of the thioester substrate is a rate-limiting step for catalysis. This is why the pH-dependent nucleophile formation and thioester hydrolysis at the metal centre are usually masked in wild-type Glo2. Such a ‘hit-and-run’ mechanism could actually also explain why many Glo2 isoforms have similar activities with various metal ion compositions at the catalytic centre (as long as the hydroxide ion is still formed at a sufficient rate) [4,45]. Recombinant P. falciparum Glo2 was furthermore shown to form dimers in E. coli and in solution. Whether this is a physiologically relevant feature, e.g. for signal transduction or enzyme regulation, has to be analysed [3,41].
Is it possible to exploit the glyoxalase detoxification system as a drug target? The erythrocyte–parasite unit was shown to consume much more glucose and to produce up to 30-fold more D-lactate than uninfected erythrocytes [3,40]. Hence malaria parasites are expected to require a functional glyoxalase system to prevent the potentially toxic accumulation of methylglyoxal and AGEs. Glyoxalase inhibitors were indeed shown to kill P. falciparum blood-stage cultures [3,41,43,46], and novel tight-binding inhibitors are very potent against recombinant Glo1 . Nevertheless, the exact relevance of the host and parasite glyoxalases for parasite survival has to be studied in further detail, and several aspects remain to be considered . Preliminary analyses in my laboratory indicate that the genes encoding the cytosolic glyoxalases cannot be deleted, whereas the apicoplast Glo2 isoform is not essential for P. falciparum blood-stage cultures (R. Alisch and M. Deponte, unpublished work). However, additional controls are necessary to exclude methodological causes and to confirm the potential essentiality of these two cytosolic glyoxalases.
In summary, functional monomeric Glo1 from P. falciparum differs significantly from its human homologue, and there are numerous apicomplexan parasites with uncharacterized and/or unorthodox glyoxalases. This also includes the enigmatic plastid enzymes from P. falciparum with their peculiar molecular evolution resulting from the endosymbiotic acquisition of the apicoplast.
Other parasitic protists
Many fascinating parasitic protists have an unspecified glyoxalase repertoire or 2-oxoaldehyde metabolism. Such pathogens include (i) additional excavates, e.g. sexually transmitted Trichomonas vaginalis and Giardia lamblia (a common cause of gastroenteritis), (ii) chromalveolates, e.g. the apicomplexan parasites denoted above, (ii) rhizaria, e.g. economically relevant haplosporidian parasites of molluscs, (iv) amoebozoa, e.g. Entamoeba histolytica (the causative agent of amoebiasis), and (v) numerous microsporidian parasites (which are now classified as fungal opisthokonts) . The glyoxalase repertoire of several of these parasites can be predicted using the eukaryotic pathogen database (http://eupathdb.org) comprising many finished and unfinished genome projects . A recent in silico overview, for example, suggested that the genomes of E. histolytica and G. lamblia do not encode a Glo1 isoform , similar to T. brucei [35,36]. Nevertheless, new insights will in the end require the analysis of the corresponding organisms and enzymes.
The biochemical characterization of glyoxalases from parasitic protists also revealed some valuable practical insights. First of all, it might be beneficial to use Mops buffers instead of phosphate buffers to characterize the enzymatic activity of Glo1 isoforms, especially of active site mutants . It is furthermore important to check the metal requirements of the studied enzyme and to analyse whether the active sites are saturated [13,30,33]. For example, one of the two active sites in homodimeric kinetoplastid or bacterial Glo1 isoforms can lack a functional metal centre [26,49]. Because mechanistic peculiarities of glyoxalases are often overlooked, it is also important to publish kinetic datasets (not just tables) at different substrate and/or inhibitor concentrations. This is particularly relevant for monomeric glyoxalases (e.g. from P. falciparum or yeast) because the kinetic parameters reflect an average of two different active sites [41,44,47].
Parasitic protists teach us the following lessons. First, some eukaryotes are absolutely fine without a functional glyoxalase system despite high glycolytic fluxes. This lifestyle might be supported by an organellar compartmentalization of glycolysis and requires alternative enzymes for methylglyoxal removal. Secondly, the glyoxalase repertoire and its physiological relevance are highly variable, even among closely related organisms. Some eukaryotes have insular and/or non-canonical glyoxalases. Their exact functions still represent one of the most fascinating unresolved issues in glyoxalase research. Furthermore, it is difficult to predict whether a glyoxalase is essential or not, and the potential of glyoxalases as drug targets depends on the investigated organism. Thirdly, the structures, metal ion requirements and substrate specificities of eukaryotic glyoxalases are all but strictly conserved. Metabolic adaptations, such as an altered thiol pool, or genetic events, such as gene duplications or the acquisition of glyoxalases owing to endosymbiosis, can result in drastically altered enzyme properties. Comparative studies are therefore extremely helpful to identify and define anomalies and common principles in order to understand the functions and mechanisms of glyoxalases.
Glyoxalase Centennial: 100 Years of Glyoxalase Research and Emergence of Dicarbonyl Stress: A Biochemical Society Focused Meeting held at the University of Warwick, U.K., 27–29 November 2013. Organized and Edited by Naila Rabbani and Paul Thornalley (University of Warwick, U.K.).
I am grateful to Naila Rabbani and Paul Thornalley for organizing the Glyoxalase Centennial meeting and for bringing us all together. Furthermore, I thank all participants of the meeting for their instructive presentations, posters and lively discussions.