MOONLIGHTING IN APICOMPLEXANS

The phylum Apicomplexa comprises a diverse group of over 5000 obligate parasites whose near evolutionary neighbours include ciliates and dinoflagellates. Malarial parasites (e.g. Plasmodium falciparum, which is the major species responsible for human malaria mortality) and the opportunistic pathogen Toxoplasma gondii are among the most well known of the Apicomplexa and, among parasites generally, it is in these organisms that the most clear-cut examples of moonlighting are to be found.

In mammalian hosts, Toxoplasma and Plasmodium replicate intracellularly, and egress from host cells is followed by active invasion of new host cells (Soldati et al. Reference Soldati, Foth and Cowman2004; Baum et al. Reference Baum, Gilberger, Frischknecht and Meissner2008). Glycolysis provides the major source of energy production in T. gondii tachyzoites and asexual bloodstage malarial parasites, with mitochondrial oxidative phosphorylation providing little, if any, contribution to cellular processes in cultured parasites (Painter et al. Reference Painter, Morrisey, Mather and Vaidya2007; Pomel et al. Reference Pomel, Luk and Beckers2008). In both parasites, the localisation of enolase, a glycolytic enzyme, to the nucleus, as well as the cytosol, or in the case of P. falciparum the additional localisation of enolase to the parasite's food vacuole, plasma membrane, and cytoskeleton, too is suggestive of functions that extend beyond this protein's classic enzymatic role in glycolysis (Ferguson et al. Reference Ferguson, Parmley and Tomavo2002; Pal-Bhowmick et al. Reference Pal-Bhowmick, Vora and Jarori2007; Bhowmick et al. Reference Bhowmick, Kumar, Sharma, Coppens and Jarori2009). The nuclear localisation of enolase in the apicomplexans examined thus far is particularly intriguing since an alternatively-translated, truncated version of enolase has previously been implicated as a transcriptional repressor of the oncogene c-myc in a human cell line (Feo et al. Reference Feo, Arcuri, Piddini, Passantino and Giallongo2000). Another glycolytic enzyme glyceraldehyde-3-phosphate dehydrogenase (GAPDH) moonlights within the nuclei of mammalian cells in roles as diverse as transcription, DNA repair, RNA binding and transport, and telomere binding (Sirover, Reference Sirover2005; Demarse et al. Reference Demarse, Ponnusamy, Spicer, Apohan, Baatz, Ogretmen and Davies2009). These examples represent only a few of the alternative functions proposed for different glycolytic enzymes in a wide variety of organisms. However, the localisations of enolase in Plasmodium and T. gondii are only suggestive of moonlighting since one cannot rule out a possibility that differential localisation of some glycolytic enzymes helps provide energy-generating capacity at cellular sites where the demand for ATP is particularly acute. Indeed, epitope-tagging and immunofluorescence indicate that several, maybe even all, of the glycolytic enzymes are at least partially re-located from the cytosol to the cell periphery in extracellular T. gondii tachyzoites (Pomel et al. Reference Pomel, Luk and Beckers2008). Intriguingly, the cell periphery in T. gondii and P. falciparum provide the sites of the most clear-cut and best delineated examples of enzyme moonlighting in parasites.

A striking example of apicomplexan-specific biology is the way in which parasites move across surfaces and penetrate tissues (Soldati et al. Reference Soldati, Foth and Cowman2004; Baum et al. Reference Baum, Gilberger, Frischknecht and Meissner2008). This motility, known as ‘gliding’ is dependent upon the co-ordinated interaction between adhesins exposed on the outer-face of the parasite plasma membrane and an actin-based cytoskeleton beneath the inner face of the parasite plasma membrane (Fig. 1). Related proteins that span across the plasma membrane of T. gondii and P. falciparum, respectively and provide the extra-cellular adhesin activities that engage host cells or other surfaces are MIC2 (standing for ‘micronemal protein 2’) and TRAP (standing for ‘thrombospondin-related anonymous protein’). Both proteins are cross-linked to relatively unstable F-actin filaments by the glycolytic enzyme aldolase (Buscaglia et al. Reference Buscaglia, Coppens, Hol and Nussenzweig2003; Jewett and Sibley, Reference Jewett and Sibley2003). The ability of aldolase to bind actin is seen in other organisms, too. An X-ray structure of P. falciparum aldolase bound to the C-terminal tail of the TRAP homologue from P. berghei (Bosch et al. Reference Bosch, Buscaglia, Krumm, Ingason, Lucas, Roach, Cardozo, Nussenzweig and Hol2007) and homology modelling and mutational analysis of T. gondii aldolase (Starnes et al. Reference Starnes, Coincon, Sygusch and Sibley2009) indicate that actin-binding and TRAP/MIC2-interacting regions overlap, suggesting that the tetrameric structure of catalytic aldolase is likely to be necessary for aldolase to moonlight as a structural bridge between either TRAP or MIC2 and the cytoskeleton that mediates gliding motility (Fig. 1). The substrate-binding active site of aldolase also lies in the vicinity of the actin/TRAP/MIC2-interacting region, yet site-directed mutagenesis of the Toxoplasma enzyme by the Sibley group nonetheless resulted in the identification of residues required specifically for catalysis, but not for coupling the transmembrane-spanning adhesion to actin filaments and vice versa (Starnes et al. Reference Starnes, Coincon, Sygusch and Sibley2009). The expression of various site-directed aldolase mutants (D33A, K41A, R42A, K41E:R42G, and R148A) in T. gondii tachyzoites subsequently not only underscored the importance of glycolysis as an energy-generating pathway during this life cycle stage, but moreover revealed the essentiality of the aldolase-dependent linkage between MIC2 and F-actin for efficient cell invasion and potentially in other situations where motility is challenged by resistance (e.g. during penetration between host cells), too (Starnes et al. Reference Starnes, Coincon, Sygusch and Sibley2009). Comparative genomics indicates that the protein components required for gliding by Toxoplasma tachyzoites or malarial sporozoites are conserved across all Apicomplexa examined to date (Baum et al. Reference Baum, Richard, Healer, Rug, Krnajski, Gilberger, Green, Holder and Cowman2006), suggesting that this particular example of aldolase moonlighting appeared early during apicomplexan evolution and was subsequently retained.

Fig. 1. Organisation of the ‘glidosome’ in apicomplexan parasites. The transmembrane adhesins TRAP (in malarial sporozoites) and MIC2 (in T. gondii) possess extracellular substrate-binding domains, and are cross-linked to F-actin via an aldolase bridge. During extracellular motility and cell invasion, the adhesin is released from apacial micronemes and translocated across relatively unstable actin filaments to the posterior-end of cells by a non-processive myosin motor that is attached to the microtubule-bound ‘inner membrane complex’. Posterior-directed movement of the adhesin results in the forward-directed motility or ‘gliding’ of parasites across the substrate (e.g. a host cell surface).N.B. Based on the studies published to date, aldolase is shown as a tetrameric bridge between adhesin and actin, but there is no indication of how many sub-units are likely to interface with actin and either TRAP or MIC2, respectively.

THE PROVOCATIVE IMPLICATION OF MOONLIGHTING IN TRICHOMONAS

The aldolases from T. gondii and P. falciparum provide well characterized examples of enzyme moonlighting in parasites. By contrast, a series of data pertaining to the possible expression of classic cytosolic and hydrogenosomal enzymes on the surface of Trichomonas vaginalis suggest intriguing, yet far from unequivocal examples of moonlighting in parasites.

T. vaginalis is a sexually transmitted pathogenic protozoan responsible for upwards of 250 million new cases of vaginitis in women annually, with the wider health concerns stemming from infection including increased pre-disposition to cervical cancer and increased transmissibility of HIV virus. The adaptation of Trichomonas to the microaerophilic environment of the urogenital mucosa is evident from the secondary loss of the capacity for cytochrome-dependent respiration (Carlton et al. Reference Carlton, Hirt, Silva, Delcher, Schatz, Zhao, Wortman, Bidwell, Alsmark, Besteiro, Sicheritz-Ponten, Noel, Dacks, Foster, Simillion, Van de Peer, Miranda-Saavedra, Barton, Westrop, Müller, Dessi, Fiori, Ren, Paulsen, Zhang, Bastida-Corcuera, Simoes-Barbosa, Brown, Hayes, Mukherjee, Okumura, Schneider, Smith, Vanacova, Villalvazo, Haas, Pertea, Feldblyum, Utterback, Shu, Osoegawa, de Jong, Hrdy, Horvathova, Zubacova, Dolezal, Malik, Logsdon, Henze, Gupta, Wang, Dunne, Upcroft, Upcroft, White, Salzberg, Tang, Chiu, Lee, Embley, Coombs, Mottram, Tachezy, Fraser-Liggett and Johnson2007). In these parasites, the oxidation of pyruvate occurs within degenerate forms of mitochondria known as hydrogenosomes (for further discussion see Hjort et al. Reference Hjort, Goldberg, Tsaousis, Hirt and Embley2010) resulting in the formation of hydrogen gas and, following the metabolism of pyruvate-derived acetyl-CoA to acetate, the production of ATP (from substrate level phosphorylation). Carbohydrate and to a lesser extent amino acids provide the carbon sources for energy metabolism in trichomonads.

Biochemical analysis of fractionated cells and localisation experiments using either polyclonal antisera or monoclonal antibodies raised against recombinant proteins indicate the hydrogenosomal localisation of the following key enzymes required for organellar pyruvate metabolism: pyruvate:ferredoxin oxidoreductase (or PFO), malic enzyme, α– and β–sub-units of succinyl-CoA synthetase (Williams et al. Reference Williams, Lowe and Leadlay1987; Brugerolle et al. Reference Brugerolle, Bricheux and Coffe2000; Hrdy et al. Reference Hrdy, Hirt, Dolezal, Bardonová, Foster, Tachezy and Embley2004). However, several reports document that upon contact of parasites with vaginal epithelial cells or upon exposure to high iron concentrations (250 μM, as opposed to a normal concentration in culture media of 20 μM) these hydrogenosomal enzymes, as well as the glycolytic enzymes enolase and GAPDH, can be found on the outer face of the parasite plasma membrane where they are proposed to moonlight as non-enzymatic adhesins during the adherence of parasites to epithelial cells lining the vaginal tract (Alderete et al. Reference Alderete, O'Brien, Arroyo, Engbring, Musatovova, Lopez, Lauriano and Nguyen1995; Engbring and Alderete, Reference Engbring and Alderete1998; Garcia et al. Reference Garcia, Chang, Benchimol, Klumpp, Lehker and Alderete2003; Moreno-Brito et al. Reference Moreno-Brito, Yáñez-Gómez, Meza-Cervantez, Avila-González, Rodriguez, Ortega-López, González-Robles and Arroyo2005; Mundodi et al. Reference Mundodi, Kucknoor and Alderete2008; Lama et al. Reference Lama, Kucknoor, Mundodi and Alderete2009).

Notwithstanding the fact that enzymes have been found to act as adhesins in other pathogens (e.g. reviewed in Alderete et al. Reference Alderete, Millsap, Lehker and Benchimol2001), the suggestion that organellar enzymes and cytosolic glycolytic enzymes in Trichomonas could act as adhesins has been met with some uncertainty (e.g. Hirt et al. Reference Hirt, Noel, Sicheritz-Ponten, Tachezy and Fiori2007). Immunofluoresence analyses using non-transgenic Trichomonas have been used to provide evidence of surface localisation of PFO, enolase, malic enzyme and succinyl-CoA synthetase (Engbring and Alderete, Reference Engbring and Alderete1998; Garcia et al. Reference Garcia, Chang, Benchimol, Klumpp, Lehker and Alderete2003; Moreno-Brito et al. Reference Moreno-Brito, Yáñez-Gómez, Meza-Cervantez, Avila-González, Rodriguez, Ortega-López, González-Robles and Arroyo2005; Mundodi et al. Reference Mundodi, Kucknoor and Alderete2008), but some concern has been voiced that the original discovery of these enzymes as putative adhesins came as a result of indirect studies, including in vitro binding assays of parasite whole cell extracts to human cells and the screening of parasite cDNA libraries with sera pooled from patients with trichomoniasis. Although the ‘stickiness’, or non-specific interactions, mediated by some glycolytic and hydrogenosomal enzymes towards surface exposed receptors on vaginal epithelial cells is conceivably sufficient to select for a moonlighting role that mediates parasite-host cell adherence, the non-specificity of these interactions coupled to an initial difficulty in conceptualising how organellar, or indeed cytosolic, enzymes lacking obvious targeting motifs arrive at, and then remain tethered to the outer-face of the parasite plasma membrane provide obvious counter-arguments to the suggestion that metabolic enzymes play a role in mediating cell adhesion in vivo. PFO is a membrane-associated enzyme (Williams et al. Reference Williams, Lowe and Leadlay1987), but again a major difficulty lies in identifying how this hydrogenosomal protein is also able to putatively locate to the cell surface Moreover, some researchers have only reported the expected hydrogenosomal location for malic enzyme, and succinyl-CoA synthetase in immunofluoresence experiments (e.g. Brugerolle et al. Reference Brugerolle, Bricheux and Coffe2000), and the reported prediction of possible trans-membrane helices in surface-exposed malic enzyme is supported by neither structural data for recombinant malic enzyme nor a report that malic enzyme excreted from Trichomonas retains a soluble, active conformation (Addis et al. Reference Addis, Rappelli, Cappuccinelli and Fiori1997; Hirt et al. Reference Hirt, Noel, Sicheritz-Ponten, Tachezy and Fiori2007). The postulation of a host cell receptor for one of the proposed Trichomonas adhesins AP120 (or PFO) could also be treated with some caution since the ‘host’ cells used in the experiments to identify AP120 (Moreno-Brito et al. Reference Moreno-Brito, Yáñez-Gómez, Meza-Cervantez, Avila-González, Rodriguez, Ortega-López, González-Robles and Arroyo2005) were HeLa cells, rather than the vaginal and ureter epithelial cells used in other studies of Trichomonas adhesins.

Yet, despite the uncertainties, the case for Trichomonas enzymes moonlighting as adhesins remains curious and somewhat compelling. For instance, it is interesting that not merely a single abundant metabolic enzyme is apparently re-located to the cell surface, but several major hydrogenosomal enzymes and cytosolic glycolytic enzymes have been described on the outer-face of the plasma membrane. This suggests large-scale re-modelling of intracellular architecture could underpin the trafficking of enzymes to the cell surface. In that regard it is relevant to note that within minutes of cytoadherence parasite morphology changes significantly from a motile multi-flagellate form to a pseudo-amoeboid form (Arroyo et al. Reference Arroyo, Gonzalez-Robles, Martinez-Palomo and Alderete1993). Rapid intracellular re-modelling of organelle and cytoskeletal architecture is not uncommon among unicellular eukaryotes, and in some instances selective autophagy pathways are used to initiate rapid organelle turnover within minutes of cells experiencing appropriate environmental cues (e.g. Herman et al. Reference Herman, Pérez-Morga, Schtickzelle and Michels2008). In their review on surface proteins in Trichomonas, Hirt and co-authors remark that an up-regulation in autophagy could result in the targeting of hydrogenosomes to the lysosome for degradation, and thus, the routing of hydrogenosomal proteins to cell surface via a lysosome-linked trafficking pathway (Hirt et al. Reference Hirt, Noel, Sicheritz-Ponten, Tachezy and Fiori2007). The signature footprint for autophagy that is evident in the Trichomonas genome (Rigden et al. Reference Rigden, Michels and Ginger2009) and suggestive experimental evidence for autophagy in Trichomonas (Benchimol, Reference Benchimol1999) lend support to this possibility.

An experiment that should unequivocally address the intriguing possibility that several key soluble metabolic enzymes in Trichomonas also moonlight as cell surface-exposed adhesins would be to use transgenic parasites expressing GFP-tagged enolase, PFO, malic enzyme or succinyl-CoA synthetase and follow the (re-)localisation of these tagged enzymes, preferably in real-time, in response to high [Fe] exposure and contact with vaginal epithelial cells. In the absence of data from this or other relevant experiments, however, it is perhaps interesting to note that Giardia lamblia, an intestinal parasite widely accepted to be distantly related to the trichomonads, was recently observed to secrete the enzymes enolase, arginine deaminase and orthinine carbamoyl transferase in significant amounts following interaction with intestinal epithelial cells (Ringqvist et al. Reference Ringqvist, Palm, Skarin, Hehl, Weiland, Davids, Reiner, Griffiths, Eckmann, Gillin and Svärd2008). The physiological significance of this enzyme secretion is uncertain, but the characterization of enzyme secretion by cultured parasites is corroborated by data from in vivo experiments, too (Davids et al. Reference Davids, Palm, Housley, Smith, Anderson, Martin, Hendrickson, Johansen, Svärd, Gillin and Eckmann2006; Ringqvist et al. Reference Ringqvist, Palm, Skarin, Hehl, Weiland, Davids, Reiner, Griffiths, Eckmann, Gillin and Svärd2008).

TRYPANOSOMATIDS: LIMITED OPPORTUNITIES FOR GLYCOLYTIC ENZYMES TO MOONLIGHT?

The parasitic trypanosomatid family include numerous pathogens of medical, veterinary or agricultural significance. Many of these pathogens are digenetic parasites, transmitted between hosts by blood- or sap-feeding arthropod vectors. The best known are the tropical disease-causing parasites of the genera Trypanosoma and Leishmania, which are collectively responsible for African sleeping sickness (T. brucei), Chagas' disease in South and Central America (T. cruzi), and leishmaniasis. Among the many unusual, sometimes unique, biochemical and cell biological characteristics that have been described in trypanosomatids are the complex and elaborate architecture of the mitochondrial genome (or kinetoplast) (Lukeš et al. Reference Lukeš, Guilbride, Votýpka, Ziková, Benne and Englund2002), extensive mitochondrial RNA editing (Stuart et al. Reference Stuart, Schnaufer, Ernst and Panigrahi2005), and the compartmentalisation of several glycolytic enzymes, as well as other enzymes of carbohydrate metabolism, within peroxisomes (Michels et al. Reference Michels, Bringaud, Herman and Hannaert2006). Regarding examples of moonlighting, the suggestion that glutamate dehydrogenase contributes to RNA editing (Bringaud et al. Reference Bringaud, Stripecke, Frech, Freedland, Turck, Byrne and Simpson1997) has now been dismissed (Simpson et al. Reference Simpson, Sbicego and Aphasizhev2003), and claims that the Rieske iron-sulphur protein, a sub-unit of cytochrome c reductase, is a component of the mitochondrial tRNA import machinery in L. tropica have not been replicated in studies with T. brucei (Paris et al. Reference Paris, Rubio, Lukeš and Alfonzo2009). However, a recurrent theme in this short review is the ability of glycolytic enzymes to moonlight in different capacities within diverse taxa. Similar to the enolase from T. vaginalis, the enolase of Leishmania mexicana was found to be present in the cytosol and on the external face of the plasma membrane, acting in the latter location as a plasminogen-binding protein (Quiñones et al. Reference Quiñones, Peña, Domingo-Sananes, Cáceres, Michels, Avilan and Concepción2007; Vanegas et al. Reference Vanegas, Quiñones, Carrasco-López, Concepción, Albericio and Avilán2007). Again, the question of how a cytosolic enolase can also be targeted to the external face of the plasma membrane is unanswered, but in the case of Leishmania or other trypanosomatids it is also reasonable to ask whether the unique compartmentalisation of glycolysis influences the opportunities for several glycolytic enzymes to moonlight.

The targeting of between six and eight (if one counts pyruvate phosphate dikinase) glycolytic enzymes into Trypanosoma and Leishmania peroxisomes (Fig. 2) means these metabolically specialised micro-bodies are better known as glycosomes (reviewed by Michels et al. Reference Michels, Bringaud, Herman and Hannaert2006). Although it is only possible to speculate on the selective pressure(s) and mechanism(s) through which glycolytic enzymes became peroxisomal components in an ancestor of the trypanosomatids and their nearest free-living relations (Ginger et al. Reference Ginger, McFadden and Michels2010), mathematical modelling and subsequent experimental studies revealed that re-compartmentalisation was accompanied by significant changes in the regulation of trypanosome glycolysis (Bakker et al. Reference Bakker, Michels, Opperdoes and Westerhoff1999a, Reference Bakker, Walsh, ter Kuile, Mensonides, Michels, Opperdoes and Westerhoffb, Reference Bakker, Mensonides, Teusink, van Hoek, Michels and Westerhoff2000; Furuya et al. Reference Furuya, Kessler, Jardim, Schnaufer, Crudder and Parsons2002; Albert et al. Reference Albert, Haanstra, Hannaert, Van Roy, Opperdoes, Bakker and Michels2005; Haanstra et al. Reference Haanstra, van Tuijl, Kessler, Reijnders, Michels, Westerhoff, Parsons and Bakker2008). In particular, the allosteric feedback regulation of hexokinase and phosphofructokinase seen in other organisms does not occur in trypanosomes – instead the peroxisomal membrane provides a regulatory barrier that prevents uncontrolled hexokinase and phosphofructokinase activities rapidly depleting the cytosolic ATP concentration (Bakker et al. Reference Bakker, Mensonides, Teusink, van Hoek, Michels and Westerhoff2000; Furuya et al. Reference Furuya, Kessler, Jardim, Schnaufer, Crudder and Parsons2002). ATP/ADP balance within glycosomes requires either organellar phosphoglycerate kinase (PGK) activity or, in trypanosomatids other than bloodstream form T. brucei, potentially a combination of several glycosomal enzymes, including adenylate kinase, pyruvate phosphate dikinase, and phosphoenolpyruvate carboxykinase. Some, maybe most, glycolytic intermediates, however, are able to move between glycosomes and the cytosol and the consequence of ectopic cytosolic expression normally of glycolytic enzymes can be dire for trypanosome cells (Blattner et al. Reference Blattner, Helfert, Michels and Clayton1998; Helfert et al. Reference Helfert, Estévez, Bakker, Michels and Clayton2001). Since glycolytic enzymes, which are generally abundant in many organisms or cell-types and can be subjected to dynamic intracellular re-localisation are, to date, the most commonly observed moonlighters, one could reasonably ask whether the unique compartmentalisation of glycolysis and re-programming of glycolysis regulation in trypanosomatids limits the possibilities for glycosomal glycolytic enzymes to moonlight.

Fig. 2. Peroxisomal compartmentalisation of carbohydrate metabolism in trypansomatids. The scheme illustrates the compartmentalisation of the first seven glycolytic enzymes in the glycosomes of bloodstream T. brucei. In other life cycle stages, or indeed in other trypanosomatid species, PGK is not necessarily a glycosomal enzyme, but PPDK (catalysing conversion of phospho-enolpyruvate to pyruvate) is present (Michels et al. Reference Michels, Bringaud, Herman and Hannaert2006).Key: HK, hexokinase; PGI, phosphoglucoisomerase; PFK, phosphofructokinase; ALD, aldolase; TPI, triose-phosphate isomerase; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; PGK, phosphoglycerate kinase; PGM, phosphoglycerate mutase; ENO, enolase; PK, pyruvate kinase.

There have been reports of glycosomal glycolytic enzymes being found bound to microtubules in biochemically fractionated cytoskeletons and flagella, but these observations are almost certainly an experimental artefact of abundant enzymes with a high pI being able to bind the trypanosome microtubule cytoskeleton once peroxisomal matrix components are solubilised following detergent-extraction of live cells (Robinson et al. Reference Robinson, Beattie, Sherwin and Gull1991). Yet, as we discuss below, partial localisation outside of glycosomes of one isoform of hexokinase is suggestive of an additional cellular function beyond glycolysis. This scenario is even more intriguing given how the properties of the hexokinase isoform in question suggest it is catalytically inactive when expressed without an appropriate binding partner. One can therefore pose the question of whether this absence of catalytic activity guards against leaving cells vulnerable to ATP depletion due to unregulated hexokinase activity when this particular trypanosome hexokinase is on moonlighting duties.

Thus, T. brucei possess two, tandemly duplicated hexokinase genes; both genes are expressed in bloodstream and procyclic trypanosomes. The protein product from the second of these genes, TbHK2, has independently been found by two groups to be present both inside and outside of glycosomes. Most recently, HK2 protein was shown to partially localise to the trypanosome flagellum (Lyda et al. Reference Lyda, Dodson, Coley, Morris, Michels and Morris2009). Several explanations for the dual targeting of HK2 protein to glycosomes and flagella can be put forward, but perhaps more important is that the differences between the C-termini of the otherwise near identical paralogous HK1, which is exclusively glycosomal, and HK2 are sufficient to render the latter protein incapable of phosphorylating glucose when expressed in recombinant form (Morris et al. Reference Morris, DeBruin, Yang, Chambers, Smith and Morris2006). Gene knockout studies reveal that HK2 is necessary for optimal growth of procyclic T. brucei (Morris et al. Reference Morris, DeBruin, Yang, Chambers, Smith and Morris2006) although it is presently not known whether changes in growth rate and morphology reflect the loss of an important moonlighting function or the effect on HK2 depletion on glycolysis: interestingly catalytically inactive recombinant HK2, when mixed with recombinant TbHK1 produces a hexameric enzyme that exhibits kinetic parameters which are more similar to those reported for native T. brucei hexokinase activity purified from cells than to the kinetic parameters of recombinant HK1 hexamers (Chambers et al. Reference Chambers, Kearns, Morris and Morris2008). The observation that hexokinase activity is reconstituted when TbHK2 is mixed with a recombinant S160A HK1 mutant that is unable to catalyse phospho-group transfer further supports the likelihood that HK2 contributes to glycolytic flux (or its regulation) within glycosomes (Chambers et al. Reference Chambers, Kearns, Morris and Morris2008).

In Leishmania donovani, but not T. brucei (Vanhollebeke et al. Reference Vanhollebeke, de Muylder, Nielsen, Pays, Tebabi, Dieu, Raes, Moestrup and Pays2008), hexokinase has been suggested to moonlight as a haemoglobin receptor within the flagellar pocket, and thus potentially contribute to haem and/or iron acquisition (Krishnamurthy et al. Reference Krishnamurthy, Vikram, Singh, Patel, Agarwal, Mukhopadhyay, Basu and Mukhopadhyay2005). As cytosolic hexokinase activity is toxic to Leishmania promastigotes (Kumar et al. Reference Kumar, Gupta, Srivastava, Sahasrabuddhe and Gupta2009), the catalytic activity of hexokinase might need to be masked within the flagellar pocket. If, however, T. brucei HK2 provides an example of a gene where mutation is likely to have been necessary to facilitate a moonlighting function, then one can reasonably ask whether the evolution of GAPDH-like proteins that are encoded in all trypanosomatid species for which genome sequences are available (e.g. the protein encoded by Tb09.211.1370) and genes in Trypanosoma species, but not Leishmania, that encode large PGK-like proteins (e.g. Tb11.22.0003) are at least in part a consequence of constraints that the glycosomal compartmentalisation of glycolysis places upon the ability of some glycolytic enzymes to moonlight. Identifying functions for these proteins will answer this question, but the lack of amino acid conservation at many positions known to contribute to catalysis in bona fide GAPDH and PGK enzymes, respectively strongly suggests neither GAPDH-like nor PGK-like proteins in trypanosomatids are likely to be catalytically active. Interestingly, epitope-tagged variants of T. brucei GAPDH-like and PGK-like proteins localise or co-purify with cytoskeletal structures, and the PGK-like protein is expressed in both bloodstream and tsetse form trypanosomes (our unpublished data). The function(s) of these novel proteins is under investigation in our laboratory.

CLOSING PERSPECTIVES

Completely sequenced genomes and stable transformation are available for the parasites discussed in this review. Thus, our survey of bona fide and putative examples of enzyme moonlighting in parasitic protists provide a timely reminder that in the post-genomic analysis of gene function it is important to remember that for some, maybe many, metabolic enzymes their cellular functions extend beyond classic roles in intermediary metabolism. The widespread use of sensitive mass spectrometry approaches in molecular cell biology means that enzymes will continue to be found in unexpected intracellular locales or structures, although in many instances rather than being due to moonlighting, this will be an artefact of contamination during biochemical fractionation (as in the example of abundant glycosomal enzymes binding to microtubules in trypanosomatid cell-free extracts (Robinson et al. Reference Robinson, Beattie, Sherwin and Gull1991)). Work with the apicomplexan aldolases (Bosch et al. Reference Bosch, Buscaglia, Krumm, Ingason, Lucas, Roach, Cardozo, Nussenzweig and Hol2007; Starnes et al. Reference Starnes, Coincon, Sygusch and Sibley2009), however, illustrates how the individual functions of enzymes that do moonlight can be teased apart experimentally.

In addition to moonlighting enzymes, there are also many examples in diverse taxa of proteins that resemble enzymes, but which perform regulatory or structural tasks rather than catalytic functions. These examples are somewhat tangential and too numerous to cover in detail in a review such as this, but in the case of many regulatory proteins, such as pseudo-protein kinases (Boudeau et al. Reference Boudeau, Miranda-Saavedra, Barton and Alessi2006) or the S-adenosylmethionine decarboxylase (AdoMetDC)-like protein that regulates AdoMetDC activity in trypanosomes (Willert et al. Reference Willert, Fitzpatrick and Phillips2007), it is hard to rationalise how some regulatory functions could have evolved from moonlighting functions. In other instances, however, gene duplication of a multi-functional protein has been followed by paralogue speciation. A classic example is the regulation of galactose metabolism in yeast: in some species a single protein is both a galactokinase and transcriptional activator of galactose metabolism, whereas in other yeast species closely-related paralogues perform either the enzymatic task of galactose phosphorylation or transcriptional activation (Hittinger and Carroll, Reference Hittinger and Carroll2007; Campbell et al. Reference Campbell, Leverentz, Ryan and Reece2008). Time will tell whether the examples of GAPDH-like and PGK-like proteins in trypanosomes, or indeed other enzyme-like proteins, such as a GAPDH-like protein from Giardia lamblia (Yang et al. Reference Yang, Yong, Lee, Im and Park2002), are more likely to have evolved as a consequence of ancestral genes that encoded multifunctional proteins (or moonlighters) being subject to duplication or whether acquisition of novel function(s) and loss of catalytic activity are merely the consequences of sequence drift in a duplicated paralogue.