Cytosolic NADPH Homeostasis in Glucose-starved Procyclic Trypanosoma brucei Relies on Malic Enzyme and the Pentose Phosphate Pathway Fed by Gluconeogenic Flux*

Background: NADPH production is critical for growth and oxidative stress management. Results: Redundancy of the pentose phosphate pathway and the cytosolic malic enzyme for NADPH synthesis is carbon source-independent in procyclic trypanosomes. Conclusion: The parasite has gluconeogenic capacity from proline. Significance: This work illustrates the flexible carbon source-dependent flux changes for essential NADPH supply. All living organisms depend on NADPH production to feed essential biosyntheses and for oxidative stress defense. Protozoan parasites such as the sleeping sickness pathogen Trypanosoma brucei adapt to different host environments, carbon sources, and oxidative stresses during their infectious life cycle. The procyclic stage develops in the midgut of the tsetse insect vector, where they rely on proline as carbon source, although they prefer glucose when grown in rich media. Here, we investigate the flexible and carbon source-dependent use of NADPH synthesis pathways in the cytosol of the procyclic stage. The T. brucei genome encodes two cytosolic NADPH-producing pathways, the pentose phosphate pathway (PPP) and the NADP-dependent malic enzyme (MEc). Reverse genetic blocking of those pathways and a specific inhibitor (dehydroepiandrosterone) of glucose-6-phosphate dehydrogenase together established redundancy with respect to H2O2 stress management and parasite growth. Blocking both pathways resulted in ∼10-fold increase of susceptibility to H2O2 stress and cell death. Unexpectedly, the same pathway redundancy was observed in glucose-rich and glucose-depleted conditions, suggesting that gluconeogenesis can feed the PPP to provide NADPH. This was confirmed by (i) a lethal phenotype of RNAi-mediated depletion of glucose-6-phosphate isomerase (PGI) in the glucose-depleted Δmec/Δmec null background, (ii) an ∼10-fold increase of susceptibility to H2O2 stress observed for the Δmec/Δmec/RNAiPGI double mutant when compared with the single mutants, and (iii) the 13C enrichment of glycolytic and PPP intermediates from cells incubated with [U-13C]proline, in the absence of glucose. Gluconeogenesis-supported NADPH supply may also be important for nucleotide and glycoconjugate syntheses in the insect host.

For ROS detoxification in T. brucei, electrons from NADPH are transferred through a cascade of electron carriers involving trypanothione and several small dithiol redox proteins (9). Trypanothione is a condensation product of two glutathione molecules and one spermidine molecule (10). Maintenance of the trypanothione-based redox homeostasis is also critical for production of deoxyribonucleotides required for replication and repair of DNA and for metabolism and biogenesis of ironsulfur clusters. Consequently, interfering with trypanothione biosynthesis and utilization is detrimental for trypanosomes (11)(12)(13)(14).
NADPH is the key metabolite in these processes because it is the only source of electrons for trypanothione reduction. NADPH is the product of two enzymatic reactions of the pentose phosphate pathway (PPP), the glucose-6-phosphate dehydrogenase (G6PDH)-and 6-phosphogluconate dehydrogenase-catalyzed steps. Both enzymes show a dual cytosolic and glycosomal localization in trypanosomes (15)(16)(17)(18). The oxidative branch of the PPP produces ribose 5-phosphate required for nucleic acid biosynthesis, and NADPH is a major source of reducing equivalents for biosynthetic processes, including de novo synthesis of fatty acids (19). Therefore, PPP activity should be essential for the bloodstream form of T. brucei, independently of trypanothione metabolism. In fact, in bloodstream, T. brucei cell death results from RNAi down-regulation of G6PDH or 6-phosphogluconate dehydrogenase, and incubation with dehydroepiandrosterone (DHEA), a potent uncompetitive inhibitor of G6PDH (20,21).
The central energy metabolism of the procyclic insect stage is more flexible and can adapt to changing carbon sources. In the midgut, this is mainly proline. For example, NADPH can theoretically be produced in the cytosol and the mitochondrion of the trypanosomatids by two isoforms of the malic enzyme (ME (Fig. 1)) (22). The relative roles of the PPP and alternative reactions to provide NADPH have not been investigated so far.
Here, we show that the oxidative branch of the PPP and the cytosolic MEc isoform can both contribute to and individually sustain the essential NADPH supply in the cytosol of procyclic trypanosomes. Surprisingly, the PPP can also operate in glucose-depleted conditions. We provide direct genetic and metabolomic evidence that this is due to gluconeogenic flux by producing glucose 6-phosphate from proline, illustrating the flexible carbon source-dependent flux changes in the procyclic stage of the parasite.

EXPERIMENTAL PROCEDURES
Trypanosomes and Cell Cultures-The procyclic form of T. brucei EATRO1125 was cultured at 27°C in SDM79 medium containing 10% (v/v) heat-inactivated fetal calf serum and 35 Metabolic network related to NADPH production. Black arrows indicate enzymatic steps of glucose and proline metabolism of procyclic trypanosomes grown in glucose-rich conditions, whereas white arrows correspond to proline metabolism in the absence of glucose (66). Dashed arrows symbolize steps for which no evidence of flux is available. The proposed NADH/NADPH converting cycle or transhydrogenase-like shunt is indicated by green arrows. For simplification, only NADP ϩ /NADPH and NAD ϩ /NADH involved in the putative transhydrogenase-like shunt are shown. Excreted end products of glucose and proline metabolism are indicated by a black background. g/ml hemin (23). The SDM79 used for glucose-depleted growth was modified by omitting glucose and the addition of 50 mM N-acetylglucosamine (GlcNAc), which is a nonmetabolized glucose analog inhibiting glucose import (24).
Inhibition of Gene Expression by RNAi-The inhibition by RNAi of gene expression in procyclic forms (25) was performed by expression of stem-loop "sense/antisense" RNA molecules of the targeted sequences (26) introduced in the pLew100 expression vector, which contains the phleomycin resistance gene (kindly provided by E. Wirtz and G. Cross) (27). Construction of the pLew-MEc/m-SAS plasmid used to simultaneously target by RNAi the mRNAs of both malic enzyme genes, which encode the cytosolic and mitochondrial isoforms (MEc: Tb11.02.3120 and MEm: Tb11.02.3130, respectively) ( RNAi MEc/m-C3 cell line), was described before (28). The pLew-MEc-SAS and pLewMEm-SAS plasmids, designed to inhibit by RNAi the expression of the MEc gene or the MEm gene, respectively, were created in the pLew100 vector with the same strategy described above, employing the same restriction sites. The targeting cassettes correspond to the end of the MEc (from position 1465 bp to 1695 bp) or MEm (from position 1476 bp to 1716 bp) coding sequence followed by the first 224 or 268 bp, respectively, of the 3Ј-UTR. The resulting plasmids (pLew-MEc-SAS and pLew-MEm-SAS) contain a sense and antisense version of the targeted gene fragment, separated by a 60-or 40-bp fragment, respectively. The pLew-PGI-SAS plasmid designed to inhibit the expression of the glucose-6-phosphate isomerase (PGI) gene (Tb927.1.3830) was made using the pLew100 vector with the same strategy described above, employing the same restriction sites. The targeting area corresponds to the midsection of the coding region (from position 889 bp to position 1366 bp) of the PGI gene. The constructed plasmid (pLew-PGI-SAS) contains a sense and antisense version of the targeted gene area, separated by a 40-bp fragment. The sense-antisense cassette designed to target expression of the G6PDH gene (21) was introduced into the HindIII and BamHI sites of the pLew100 vector to produce the pLewG6PDH-SAS plasmid.
Gene Knock-out of the MEc Gene-Replacement of both alleles of the MEc gene by the blasticidin and puromycin resistance markers via homologous recombination was performed with DNA fragments containing a resistance marker gene flanked by the MEc UTR sequences. Briefly, the pGEMt plasmid was used to clone an HpaI DNA fragment containing the blasticidin (BSD) and puromycin (PAC) resistance marker gene preceded by the MEc 5Ј-UTR fragment (581 bp) and followed by the MEc 3Ј-UTR fragment (579 bp).
Oxidative Stress Assays-The susceptibility of trypanosomes toward oxidative stress was measured with an adapted protocol of the Alamar Blue assay (36,37). Cells were grown to densities between 6 ϫ 10 6 and 1.5 ϫ 10 7 cells/ml and diluted to 4 ϫ 10 6 cells/ml (glucose oxidase (GOX) assays) or 1.5 ϫ 10 6 cells/ml (xanthine oxidase (XOX) assays). A volume of 180 l of cell suspension was distributed to the corresponding wells of a 96-well plate. Then 20 l of GOX or XOX solutions with increasing concentrations were added to the respective wells (GOX and XOX from Sigma). After a 22-h (GOX) or 44-h (XOX) incubation, 20 l of a 0.49 mM resazurin (Sigma) solution in PBS were added to each well, and 2 h later, the fluorescence was measured in a Tecan Safire plate reader with an excitation wavelength of 530 nm and an emission wavelength of 590 nm. The background caused by the medium was measured and subtracted. All Alamar Blue assay results are mean of triplicate assays of a given biological experiment. Wells without the addition of GOX/XOX were used as reference and set to 100% cell viability. When using XOX as inducing agent, 5 mM hypoxanthine was added to the culture medium as substrate. The G6PDH inhibitor DHEA was added to the respective cultures at 15 M 24 h before inducing oxidative stress.
Mass Spectrometry Analysis-Wild-type cells grown in glucose-rich or glucose-depleted medium were washed twice with PBS and resuspended in PBS containing 2 mM [U-13 C]proline with (10 mM) or without glucose, respectively. The cells were incubated for 2 h at 27°C before fast filtration preparation of the samples for mass spectrometry analysis, as described before (38). Metabolites were analyzed by ionic-exchange chromatography coupled with tandem mass spectrometry (IC-MS/MS) using the method described by Bolten et al. (39). Retention time on the column and multiple reaction monitoring (MRM) transition of each analyzed metabolite are shown in Table 1. The 13 C mass isotopomer distribution of intracellular metabolites was determined from relevant isotopic clusters in the IC-MS/MS analysis, according to Kiefer et al. (40). 13 C mass isotopomer distribution measurements were performed using a triple quadrupole mass spectrometer (4000Qtrap, Applied Biosystems). To obtain 13 C-labeling patterns ( 13 C isotopologues), isotopic clusters were corrected for the natural abundance of isotopes other than 13 C, using the in-house software IsoCor (available at MetaSys) (41).

Subcellular Localization of the ME Activity in Procyclic T.
brucei-In procyclic trypanosomes grown in glucose as carbon source, the PPP is a major source of NADPH (15). However, other NADP-dependent enzymatic activities, such as that of ME, have been reported in insect stage trypanosomatids (42,43). Two isoforms of the NADP-dependent ME are encoded in the T. brucei genome, MEc (Tb11.02.3120) and MEm (Tb11.02.3130), and they might contribute to NADPH regeneration. One isoform (MEm) has a potential N-terminal mitochondrial targeting sequence (0.90 probability MITOPROT prediction). To address the subcellular localization of the ME activity, procyclic T. brucei cells were permeabilized with increasing concentrations of digitonin. Soluble and insoluble fractions at each concentration were separated and analyzed for ME activity ( Fig. 2A). A mitochondrial marker (TRYP2) and a cytosolic marker (cytosolic fumarase) were quantified by Western analysis of an aliquot of each fraction (Fig. 2B). ME activity is released to the supernatant at low digitonin concentrations, which correlates with the cytosolic marker. A second release of activity to the supernatant at higher digitonin concentration correlates with the appearance of the mitochondrial marker. In summary, ME activity is dually localized and equally distributed among the cytosolic and mitochondrial compartments.
Analysis of ME-deficient Cell Lines-To investigate the respective functional roles of the isoforms, we created three different inducible RNAi knockdown cell lines. One RNAi hair-pin construct targets both ME isoforms ( RNAi MEc/m), and the other constructs each target specifically the individual isoforms ( RNAi MEc or RNAi MEm). RNAi against both isoforms led to a severe growth phenotype 4 days after induction with tetracycline (Fig. 3A). Samples were taken in parallel for ME activity assays. The ME activity is decreased to a very low level after 3-4 days, correlating with the growth phenotype (Fig. 3A). In the RNAi cell lines targeting only the cytosolic ME, we did not see a growth phenotype (Fig. 3C), but targeting the mitochondrial isoform resulted in the severe phenotype (Fig. 3B), similar to that observed when targeting both isoforms (Fig. 3A). All noninduced clones showed normal growth (Fig. 3, A-C). The ME activities were reduced to ϳ40% after tetracycline induction in the cell lines with singly targeted MEc or MEm (Fig. 3, B and C). We concluded that under standard culture conditions, the

Carbon Source-dependent Flux Changes for NADPH Production
mitochondrial ME isoform is essential, but not the cytosolic ME. This was fully confirmed by the successful creation of a ⌬mec/⌬mec knock-out cell line, named ⌬mec (Fig. 3D). As expected, this line proliferated as the parental control, and the remaining total ME activity within the ⌬mec line was similar to the activity in the induced RNAi MEc line (Fig. 3D).
Cytosolic ME-deficient (⌬mec) Cells Cannot Tolerate Depletion or Inhibition of G6PDH-To explore an alternative source of NADPH of possible physiological relevance in procyclic T. brucei, we depleted the major NADPH source, the PPP in the wild-type background and in the ⌬mec mutant background. There are two NADPH-producing steps in the PPP, G6PDH and 6-phosphogluconate dehydrogenase. By depleting or inhibiting G6PDH, the downstream 6-phosphogluconate dehydrogenase activity is also affected due to substrate limitation. Knockdown of G6PDH by RNAi in the wild-type background had no effect on the growth of the cells (Fig. 4A), in contrast to the same treatment in the bloodstream stage (20,21). However, the induction of the G6PDH RNAi in the ⌬mec background resulted in one of the most dramatic growth phenotypes we observed so far in procyclic trypanosomes (Fig. 4B). As MEc can rescue the G6PDH deficiency, the main function of the MEc is obviously NADPH production. We also conclude that the degree of RNAi-mediated repression of G6PDH is limiting for NADPH production but not for ribose synthesis. The genetic interaction was confirmed by chemical inhibition of T. brucei G6PDH by DHEA, a specific inhibitor of G6PDH in T. brucei (20,21,44). We determined the LD 50 of this compound for the procyclic ⌬mec line to be 18 M after 48 h of incubation (Fig.  4D). The LD 50 for wild-type cells was at least 10-fold higher (data not shown). Using 15 M DHEA, we observed a growth defect and cell death after 5-6 days in ⌬mec cells but not in wild-type cells, exactly as upon RNAi induction (Fig. 4C). The IC 50 value for G6PDH inhibition by DHEA in vitro is 2.8 Ϯ 0.6 M (44). Assuming efficient uptake of the drug, 15 M DHEA would result in 86% inhibition of G6PDH. As wild-type cells can perfectly tolerate much higher concentrations of DHEA (not shown) as well as efficient RNAi depletion of G6PDH (Fig. 4A), we attempted to delete the G6PDH gene. Although we obtained double drug-resistant lines with G6PDH locus-specific marker integration, a G6PDH allele or G6PDH protein was always retained in these clones, suggestive of triploidization (supplemental Table S1). In trypanosomes, this type of locus triploidization upon targeting is indicative of gene essentiality (45), which can be also observed in Leishmania (46). We concluded that G6PDH and ME are redundant with respect to NADPH production, yet the PPP must have an additional and essential function. The cells might well tolerate PPP deficiency over the short time of RNAi-dependent knockdown or DHEA inhibition experiments, but for clonal outgrowth of cells, the PPP seems to be essential, probably because of its function in nucleotide synthesis (47,48).
⌬mec Mutant Cells Control Oxidative Stress in the Absence of Glucose-As a sensitive assay for the contributions of G6PDH and MEc to NADPH production, we quantified the oxidative stress sensitivity of the respective mutants. Detoxification of ROS depends on NADPH supply. Oxidative stress was applied by continuous H 2 O 2 production with GOX or XOX in the growth medium (49,50). This results in a more physiological stress type when compared with the bolus application of H 2 O 2 , which acts only transiently with a half-life time of about 10 min under standard culture conditions (51). DHEA-mediated inhibition of G6PDH or deletion of MEc alone resulted in a very moderate increase in susceptibility (Fig. 5A). However, incubation of the ⌬mec mutant cell line with 15 M DHEA for 2 days caused a severe increase in H 2 O 2 susceptibility (Fig. 5A) and growth inhibition (Fig. 4D). These data confirm the redundancy of the PPP and the MEc activities for NADPH production and for maintenance of the cytosolic redox state when glucose is available as carbon source.
Procyclic trypanosomes grow in essentially glucose-free conditions in the insect gut. Therefore, we tested the ⌬mec mutant in glucose-depleted medium for susceptibility to oxidative stress. We expected increased susceptibility due to low PPP activity under glucose-depleted conditions. Fig. 5B shows that the ⌬mec mutant was not significantly more susceptible to oxidative stress than wild-type cells. By inhibition of G6PDH (and hence the PPP) with DHEA, we obtained the same phenotype as in glucose-fed conditions. One possible explanation would be that uptake of residual glucose from the serum is sufficient to feed the PPP. This is, however, extremely unlikely as we added a large excess of GlcNAc (50 mM) to the medium to inhibit the uptake of the residual glucose present in FCS (0.5 mM). GlcNAc has been shown to inhibit glucose-uptake (24). The alternative would be glucose-independent supply of glucose 6-phosphate (G6P) via gluconeogenic flux. Production of glucose 6-phosphate from a non-glycolytic source through gluconeogenesis has been suspected in trypanosomatids but not experimentally verified to date (52).
Inhibition of Glucose-independent G6P Production Phenocopies G6PDH Deficiency-To address the possibility of glucose-independent supply of G6P, expression of the PGI gene was down-regulated by RNAi in the wild-type and the ⌬mec mutant background. PGI and hexokinase are the only enzymes in T. brucei that can give rise to G6P as product (Fig. 1). Therefore, in glucose-depleted medium, the tetracycline-induced ⌬mec/ RNAi PGI double mutant but not the single RNAi PGI mutant should phenocopy the combined MEc and G6PDH deficiency. The growth behavior of the respective cell lines in glucose-rich and glucose-depleted conditions is shown in Fig. 6. Under glucose-depleted conditions (Fig. 6B), only the induced ⌬mec/ RNAi PGI but not the induced RNAi PGI mutant died. Thus, cell death is likely to result from NADPH depletion in analogy to the ⌬mec/ RNAi G6PDH double mutant (Fig. 4B). Tetracycline induction in the presence of glucose (Fig. 6A) resulted in cell death in both the ⌬mec/ RNAi PGI and the RNAi PGI mutants. This is expected due to glycosomal ATP depletion as glycolysis is blocked. The ATP consumed to produce G6P cannot be regained in the pathway. The same mutant cell lines were then subjected to oxidative stress assays in glucose-depleted conditions (Fig. 6C). The induced RNAi PGI mutant showed the same susceptibility as the uninduced control, but the induced ⌬mec/ RNAi PGI had significantly increased oxidative stress susceptibility. It should be noted that the result in Fig. 6C is virtually indistinguishable from the analogous experiment with MEc and G6PDH double deficiency (Fig. 4B). This further supports the redundant roles of MEc and the PPP to provide the essential cytosolic NADPH. Most interestingly, under glucose-depleted culture conditions, the PPP seems clearly active. This reveals gluconeogenic flux up to the level of G6P. In the physiological context, this G6P should derive from proline. Metabolic Evidence of Gluconeogenesis-To investigate the gluconeogenic flux from proline as carbon source, procyclic cells were incubated with uniformly 13 C-enriched proline ([U-13 C]proline) in the presence and absence of glucose. Incorporation of 13 C into intermediate metabolites was quantified by IC-MS/MS, and the values for selected glycolytic and PPP metabolites are shown in Fig. 7. The incorporation of 13 C atoms into glycolytic metabolites in the presence of glucose was low (Fig. 7A), as expected, as the proline consumption is repressed by the presence of glucose (53). High 13 C incorporation was observed with proline as the only carbon source (Fig.  7B). For all glycolytic intermediates analyzed (phosphoenolpyruvate (PEP); 2-or 3-phosphoglycerate (2/3-PGA); 1,3-bisphosphoglycerate (1,3-BGA); fructose 1,6-bisphosphate (FBP); fructose 6-phosphate (F6P), and G6P), the fraction of 13 C-enriched molecules is Ͼ85% in glucose-depleted conditions when compared with ϳ10% in the presence of glucose. For additional control, glutamine was added to the medium in glucose-depleted conditions. Glutamine has been shown to serve as carbon source and share with proline the same degradation pathway from glutamate (53). As expected, isotopic dilution was seen (Fig. 7B). In contrast, the addition of threonine only resulted in a minimal dilution effect (Fig. 7B). Threonine can be degraded to acetyl-CoA but can neither repress proline consumption, as does glucose, nor compete with proline degradation. The relative low amounts of [U-13 C]hexose phosphate (FBP, F6P, and G6P) and [U-13 C]triose phosphate glycolytic intermediates when compared with partially 13 C-enriched molecules (Fig. 7A) are probably due to introduction of 12 C carbons at the carboxylation/decarboxylation step catalyzed by PEP carboxykinase (PEPCK) and the complete reversibility of the PEPCK/malate dehydrogenase/fumarase branch (54). 5 The key result of this experiment is the evidence for production of [U-13 C]G6P from [U-13 C]proline. As this occurs only in glucose-depleted conditions, clear evidence for flux in the gluconeogenic direction is derived. Interestingly, the same degree of enrichment is seen for hexose phosphates (FBP, F6P, and G6P) and the PPP intermediates (6-phosphogluconolactone and sedoheptulose 7-phosphate), confirming that proline-derived G6P feeds the PPP.

DISCUSSION
Although glucose is sparse or absent in the natural environment of procyclic trypanosomes, the insect midgut (55), the common in vitro culture conditions are glucose-rich. Here, we have identified the metabolic pathways used to produce cytosolic NADPH, in glucose-rich and glucose-depleted conditions. The reduced cofactor NADPH is essential for biosynthetic pathways and detoxification of ROS generated by oxidative stress. The candidate enzymatic reactions providing NADPH in the cytosol, i.e. the first and third steps of the PPP (G6PDH and 6-phosphogluconate dehydrogenase, respectively) and MEc, were shown to have a redundant function in glucose-rich conditions. The ⌬mec/ RNAi G6PDH.i (where "i" stands for tetracycline-induced) double mutant dies rapidly after a few days of induction, whereas the ⌬mec and RNAi G6PDH.i single mutants show no significant growth alteration. Hypersusceptibility to oxidative stress of the ⌬mec cell line incubated with DHEA, a specific inhibitor of G6PDH, indicates that the common function of the two pathways is production of NADPH, which is essential for defense against ROS. This redundancy may compensate in trypanosomatids for the lack of a transhydrogenase gene (56), which in other organisms catalyzes the reversible conversion of NADH into NADPH. In contrast to procyclic cells, the bloodstream form cells of T. brucei are highly susceptible to DHEA (21). This implies that the contribution of MEc to NADPH production is minor in bloodstream trypanosomes, although MEc is expressed in this developmental stage (22). The essential function of the PPP for NADPH production in bloodstream forms is in agreement with the high glycolytic and PPP fluxes, whereas metabolic flux through MEc is expected to be relatively low (15). A low level of PPP function may also be essential upon long term culture of procyclic cells due to the role in nucleotide biosynthesis through ribose 5-phosphate. This is supported by our inability to knock out the G6PDH gene.
In contrast to expectation, the PPP and MEc pathways can both provide NADPH to procyclic cells, also in glucose-depleted conditions. As a control, the ⌬mec/ RNAi G6PDH.i double mutant died in glucose-depleted medium (data not shown) as it did in glucose-rich medium (Fig. 4B). However, the susceptibility to oxidative stress of the wild-type cells incubated or not with DHEA was the same as the sensitivity of ⌬mec mutant cells in glucose-depleted conditions. This provided strong evidence for a flux through the PPP in the absence of glucose, suggesting an alternative hexokinase-independent source of G6P. The obvious hypothesis, gluconeogenesis, was confirmed by the observed oxidative stress hypersensitivity of the ⌬mec/ RNAi PGI.i mutant in a glucose-depleted environment. PGI, the enzyme reversibly converting F6P into G6P, is an essential step of gluconeogenesis. In addition to this genetic evidence, we directly demonstrated for the first time bona fide gluconeogenesis in T. brucei by ϳ9-fold increase of 13 C incorporation into glycolytic and PPP intermediates from [U-13 C]proline. Proline is the main available carbon source in the absence of glucose (53). Hannaert and colleagues (52,57) had proposed gluconeogenic capacity in the glycosomes of T. brucei because the genome encodes a glycosomal fructose-1,6-bisphosphatase. However, no experimental evidence of gluconeogenesis has been reported for either procyclic or bloodstream forms of T. brucei.
What is the benefit of redundant NADPH sources in procyclic T. brucei in both glucose-rich and glucose-depleted conditions? The primary role of the irreversible reaction catalyzed by MEc is apparently cytosolic NADPH production. Therefore, MEc can provide a high flexibility to adapt NADPH production to cellular need, whatever carbon source is available. In glucosedepleted conditions, carbon flow from proline metabolism can glucose-rich 13   be redistributed toward the MEc route without affecting production of the essential gluconeogenic precursor PEP. Indeed, PEP can be produced from three different pathways starting from cytosolic malate (MEc and the glycosomally located pyruvate phosphate dikinase (PPDK)), mitochondrial malate (MEm and PPDK), or glycosomal malate (malate dehydrogenase and PEPCK) (28). A single route is sufficient to feed gluconeogenesis (Fig. 1, white arrows). This redundancy is confirmed by the nearly wild-type growth rates of ⌬mec, ⌬pepck, and ⌬ppdk null mutants (38) (Fig. 6B and data not shown). In glucose-rich conditions, the MEc reaction is a single-step bridge between two main branches of glucose metabolism that lead to mitochondrial succinate and acetate production (Fig. 1, black arrows).

C-MID
Redistribution of the carbon flow toward acetate production from glucose-derived malate has no significant impact on the growth rate and allows an increase of NADPH production through MEc. Flux reduction in the mitochondrial succinate branch, due to this redistribution, is well tolerated because ablation of this branch by ⌬pepck, RNAi FRDm (mitochondrial NADH-dependent fumarate reductase), or RNAi FHm (mitochondrial fumarase) mutants does not affect the growth rate of procyclic cells (33,38,58). Thus, an increase of NADPH demand may lead to an increase of acetate production from glucose metabolism. The suggested role of MEc is consistent with a recent in silico analysis of flux distribution between the different metabolic branches using a multiobjective-criteria bioinformatics approach (59). The simulations predict that ME activity is primarily responsible for the high flexibility observed for the excreted succinate/acetate ratio (33,54).
As an extension of the flux redistribution model discussed above, we propose a cycle able to increase the metabolic flux through MEc in both glucose-depleted and glucose-rich conditions. Given the reversibility of the PPDK reaction (60), a cycle made up of MEc and three glycosomal steps (PPDK working in the gluconeogenic direction, PEPCK, and malate dehydrogenase) may operate without impact on the energy metabolism and glycosomal ATP balance to produce cytosolic NADPH sustained by NADH production in the glycosomes (Fig. 1, green  arrows). The relevance of this hypothetical cycle, which may also include the mitochondrial MEm instead of MEc, is its ability to substitute for the absence of cytosolic and mitochondrial transhydrogenases. A similar cycle for converting NADH into NADPH has been engineered in Saccharomyces cerevisiae, and due to its function, it was named transhydrogenase-like shunt (61). A S. cerevisiae strain was modified to optimize ethanol production, which was limited by the redox imbalance created during fermentation. By overexpressing ME, malate dehydrogenase, and pyruvate carboxylase, it was possible to decrease NADH levels by using it for NADPH for the cost of ATP, which is used during the reaction catalyzed by pyruvate carboxylase. The cycle we propose in T. brucei is able to use glycosomal NADH for production of cytosolic/mitochondrial NADPH without a net cost of ATP as PPDK and PEPCK replace pyruvate carboxylase (Fig. 1, green arrows). The ATP used in the PPDK reaction is regained in the PEPCK reaction. Our proposed transhydrogenase-like shunt is therefore potentially more efficient than the synthetic one used for metabolic engineering of S. cerevisiae. The cycle depends on glycosomal NADH availability and may only be active temporarily to compensate peak demands of cytosolic NADPH, e.g. during oxidative stress. Irrespective of the precise model, MEc clearly serves to ensure the essential NADPH production in procyclic trypanosomes in different metabolic situations. This may be most relevant in the insect vector, where a relatively low flux through the oxidative branch of the PPP has to be supported by gluconeogenesis.
A second ME isoform (MEm) is expressed in the T. brucei mitochondrion and accounts for approximately half of the total ME activity in procyclic cells (Figs. 1 and 2). In contrast to MEc, MEm is essential, whatever carbon source is provided. MEm also builds a single-step bridge between the mitochondrial succinate and acetate pathways and is not critical for the metabolic network. This strongly supports the view that MEm is essential mainly for NADPH production. The argument implies that there is no other significant source of NADPH in the mitochondrion of procyclic cells. The absence of transhydrogenases in trypanosomatids has already been mentioned.
The metabolism of trypanosomes in the tsetse vector or under conditions mimicking the tsetse environment is a largely unexplored field of research and of key interest to understand the developmental adaptations in the life cycle. In this context, our direct demonstration of a gluconeogenic flux is an important advance. Gluconeogenesis seems to be essential for virulence of the related pathogen Leishmania major in macrophages and establishment of infection in mice. When intracellular stages that reside in the parasitophorous vacuole cannot support gluconeogenesis (62) or synthesize specific sugars (63), the amastigotes stop replicating but remain viable. This cell cycle arrest phenotype may be caused by depletion of ribulose 5-phosphate for nucleotide synthesis. Supplementation with exogenous amino acids stimulates the growth of intracellular amastigotes (64), suggesting adaptation of the amastigotes to the sugar-poor but amino acid-rich environment. In the gut and hemolymph of the insect vector of T. brucei, a gluconeogenic flux may not only contribute to maintain the redox balance of the cell, but be crucial for synthesizing ribulose 5-phosphate and certain sugars for cell surface glycoconjugates, as shown in Leishmania (62). As discussed above, NADPH can also be provided by ME as an alternative source. Recent studies on cancer cells demonstrated that ME plays a central metabolic role in controlling NADPH levels and furthermore plays a regulatory role in preventing senescence of mammalian cells (65). The surprising similarity of metabolic adaptation mechanisms to meet NADPH demand in response to growth conditions highlights the importance of investigating metabolism in defined functional states, be it developmental stages of a parasite, host environment, or the proliferative capacity and malignancy of differentiated mammalian cell populations in the tissue context.