Mitochondrial metabolism of sexual and asexual blood stages of the malaria parasite Plasmodium falciparum

The human parasite Plasmodium falciparum is the major cause of disease and death from malaria [1]. Disease is associated with the development of asexual parasite stages that undergo repeated cycles of invasion and replication in red blood cells (RBCs). Following establishment of infection, a small proportion of parasites (<1%) differentiate to gametocytes [2, 3]. Gametocytogenesis is essential for subsequent transmission because the mature gametocyte is the only stage that can undergo sexual development in the mosquito vector, which is a prerequisite for the spread of disease. Both the asexual RBC stages and gametocytes are thought to be primarily dependent on glucose uptake and glycolysis for ATP synthesis and survival. Glucose uptake in infected RBCs increases more than 75-fold compared with uninfected RBCs [4], and the resultant increased lactate production contributes to lactic acidosis, a major cause of morbidity and death during severe malaria [5]. Despite their reliance on glycolysis, the asexual stages of P. falciparum retain a single mitochondrion that is essential for parasite growth. Maintenance of the mitochondrial respiratory chain appears to be required for both the transport of proteins and metabolites into the mitochondrion and for the reoxidation of inner-membrane dehydrogenases, such as the dihydroorotate dehydrogenase involved in de novo pyrimidine biosynthesis [6]. As a result, the asexual stages and gametocytes are sensitive to electron transport chain inhibitors, including the antimalarial atovaquone [6–9].

Mitochondrial dehydrogenases require a source of reducing equivalents which could, in principal, be generated in the mitochondrion or in the cytoplasm. P. falciparum encodes all of the enzymes needed for a complete TCA cycle, but the genes encoding a mitochondrial pyruvate dehydrogenase (PDH) complex are missing [10], and it is generally assumed that a TCA cycle utilizing glycolytic pyruvate does not operate in the blood stages [11–16]. The possibility that an unusual branched TCA cycle may operate in the asexual stages of P. falciparum, fuelled by the catabolism of glutamine via both the oxidative and reductive arms of the TCA cycle, was recently proposed [13], but subsequently retracted [17]. More recent genetic studies have also suggested that operation of a complete mitochondrial TCA cycle is not required for the development of the asexual stages in the related murine parasite, Plasmodium berghei[18, 19]. The extent to which a conventional or unconventional TCA cycle operates in P. falciparum RBC stages therefore remains unresolved.

Despite lacking a recognizable mitochondrial PDH in P. falciparum, there is increasing evidence that a conventional TCA cycle can operate in the insect stages of these parasites [18, 20] and of other apicomplexan parasites. In particular, we have recently shown that the TCA cycle is essential for the growth of intracellular stages of Toxoplasma gondii[21].

In this study, we therefore reinvestigated the potential role of mitochondrial metabolism in P. falciparum asexual stages, and the possibility that the TCA cycle is important for the development of gametocytes.

Uninfected RBCs and synchronized, ring stage-infected RBCs were metabolically labeled with ¹³C-U-glucose or ¹³C-U-glutamine for 38 hours, and then rapidly chilled. This was followed by extraction of intracellular metabolites and quantification of ¹³C-enrichment by gas chromatography–mass spectrometry (GC-MS) (see Additional file 1). Incubation of uninfected and infected RBCs in medium containing ¹³C-glucose led to a high level of enrichment (>75%) in glycolytic intermediates, including phosphoenolpyruvate (PEP) and lactate (Figure 1A). Despite the parasite lacking a recognizable mitochondrial isoform of pyruvate dehydrogenase, labeling of citrate and a range of other TCA-cycle intermediates also occurred in P. falciparum-infected RBCs under these conditions. Although TCA-cycle intermediates were detected in uninfected RBCs, they were present at concentrations more than five-fold lower than those of infected RBCs, and labeling of these intermediates was negligible (Figure 1A). The predominant isotopomers of citrate in ¹³C-glucose-fed infected RBC contained +2, +4 or +6 labeled carbons, indicating the operation of a canonical TCA cycle in which pyruvate feeds into the cycle via acetyl-CoA (Figure 1B,C). Citrate isotopomers containing +3 and +5 labeled carbons were also detected, reflecting the activity or activities of the Plasmodium PEP carboxylase (PEPC) and/or PEP carboxykinase (PEPCK) that catalyze the carboxylation of ¹³C₃-phosphoenolpyruvate (PEP) to ¹³C₃-oxaloacetate. Isotopomer analysis of other intermediates in the TCA cycle provided further support for this model (Figure 1B). However, the cellular pools of these intermediates were labeled to a much lower extent than occurred with citrate, indicating entry of other unlabeled carbon sources into the TCA cycle (see below) (Figure 1B). Significant levels of labeled γ-aminobutyric acid (GABA) were also detected in infected, but not in uninfected RBCs (Figure 1A). Labeling of GABA provides evidence for the presence of a partial GABA shunt in which intermediates from the TCA cycle are used to synthesize glutamate, which is subsequently decarboxylated to GABA, as has recently been shown to occur in T. gondii[21].

The operation of a conventional TCA cycle in the P. falciparum asexual stages was confirmed by complementary ¹³C-glutamine labeling experiments. Incubation of infected RBC with ¹³C-glutamine resulted in labeling of all detectable TCA-cycle intermediates, with the highest ¹³C-enrichment seen in α-ketoglutarate and C4 dicarboxylic acids (Figure 1A). No labeling of TCA-cycle intermediates was detected in uninfected RBCs (Figure 1A). The predominant isotopomers of succinate, malate, and fumarate in ¹³C-glutamine-labeled parasites were fully labeled, indicating that most of the carbon skeletons that enter the TCA cycle via glutamate are not continuously cycled through the TCA reactions, and are presumably exported from the mitochondrion. This was supported by the low level of labeling of citrate compared with the C4 dicarboxylic acids and the predominant presence of the +4 citrate isotopomer. However, all intermediates contained readily detectable levels of +2 isotopomers, consistent with cycling of a sub-pool of C4 dicarboxylic acids around a canonical oxidative TCA cycle. Importantly, and in contrast to a previous report [13], citrate isotopomers containing +5 labeled carbons were not detected, indicating minimal catabolism of α-ketoglutarate via the reductive arm of the TCA cycle (Figure 1B). Collectively, these analyzes show that the P. falciparum asexual stages catabolize both pyruvate and glutamate in a conventional TCA cycle, and argue against the operation of a bifurcated TCA cycle, as previously proposed [13, 17]. However, these results do support a degree of compartmentalization within this cycle, with glutamate sustaining a major flux from α-ketoglutarate to malate/oxaloacetate, and with glucose-derived pyruvate and oxaloacetate contributing to a minor flux towards citrate synthesis.

A small proportion of asexual parasite stages differentiate to gametocytes in vivo, providing a pool of transmission-competent parasites. Development of P. falciparum gametocytes involves distinct morphological transitions (stages I to V) and takes 7 to 10 days to complete [22]. During early-stage gametocytogenesis, newly invaded parasites expand in size, with concomitant depletion of host-cell cytoplasm and hemoglobin [23, 24]. These parasite stages are non-replicating, and are thought to enter a metabolically quiescent state by Stage III as they become less insensitive to current first-line anti-malarial drugs [25]. To measure the metabolic state of developing gametocytes more precisely, ¹³C-glucose uptake by Stage III gametocyte-infected RBCs was monitored by ¹³C-nuclear magnetic resonance spectroscopy (NMR). Unexpectedly, gametocyte-infected RBCs exhibited a significantly higher rate of ¹³C-glucose utilization than did RBCs infected with mature asexual stages, which was matched by increased rates of ¹³C-lactic acid secretion (Figure 2A; see Additional file 2). Gametocytes also produced significant amounts of ¹³C-acetate, indicating increased conversion of glucose into acetyl-CoA synthesis via either mitochondrial or apicoplast pathways.

To further define the metabolic state of gametocytes, Stage III gametocytes were cultured in the presence of ¹³C-glucose or ¹³C-glutamine, and enrichment in intracellular intermediates was determined by GC-MS. Marked differences were seen in the relative level of labeling of the TCA-cycle intermediates in gametocyte stages compared with the asexual stage parasites (Figure 1A). Specifically, TCA-cycle intermediates in gametocytes were strongly labeled with ¹³C-glucose, whereas the levels of enrichment in ¹³C-glutamine-fed gametocytes were low. The rapid labeling of citrate isotopomers with +2, +4, and eventually +6 labeled carbons after 2–4 hours demonstrates the operation of an active canonical TCA cycle (Figure 2B). Similarly, the rapid labeling of +3 isotopomers of oxaloacetate (indicated by aspartate labeling) and malate and the +5 isotopomer of citrate indicated that phosphoenolpyruvate may feed directly into the TCA cycle via PEP carboxylase or PEP carboxykinase (Figure 2B). Although labeling of gametocyte TCA-cycle intermediates with ¹³C-glutamine was relatively low compared with that in ¹³C-glucose-fed gametocytes, isotopomer analysis again confirmed conventional TCA-cycle operation. Specifically, a major +4 isotopomer was detected in all TCA intermediates, and this reached a maximum after 4 hours (Figure 2B). In contrast, +2 isotopomers were generated with slower kinetics, indicative of loss of labeled carbon with repeated cycles around the TCA cycle. As for the asexual stages, the absence of a predominant +5 isotopomer in citrate confirmed that the TCA cycle operates primarily in the oxidative direction in P. falciparum gametocytes (Figure 2B). GABA was labeled with both ¹³C-U-glucose and ¹³C-U-glutamine to the same maximal labeling level as other metabolites, although with slower kinetics (Figure 2B). Together, these findings suggest that there is substantial remodeling of mitochondrial metabolism in gametocytes, with glucose providing most of the carbon skeletons for operation of a complete TCA cycle (Figure 1A). Quantification of the rates of glucose consumption and lactate production provided further support for increased flux of glucose-derived pyruvate into the TCA cycle of gametocytes. Specifically, whereas more than 93% of the glucose internalized by the asexual stages was secreted as lactate, this was reduced to 80% in gametocytes, reflecting increased mitochondrial catabolism (Figure 2A).

Sodium fluoroacetate (NaFAc) is a selective and potent inhibitor of the TCA-cycle enzyme aconitase [26]. Metabolite profiling of asexual and gametocyte cultures treated with 1 mmol NaFAc revealed a 7-fold and 17-fold accumulation of citrate, respectively, with a concomitant decrease in abundance of downstream TCA metabolites (Figure 3A; see Additional file 3), supporting specific inhibition of the aconitase reaction in an oxidative cycle. Interestingly, levels of glutamate and GABA were also reduced (Figure 3A; see Additional file 3), possibly reflecting increased glutaminolysis and utilization of intracellular pools of these amino acids. To examine whether perturbation of the TCA cycle affects growth of the asexual stages or gametocyte development, asexual stage parasites or Stage II/III gametocytes were cultured in the presence of 1 mmol or 10 mmol NaFAc or sodium acetate (NaAc). No significant effect on the growth of asexual stages was seen after 7 days of continuous culture in the presence of NaFAc (see Additional file 4). The resistance of this stage to NaFAc is consistent with the predominant flux in the TCA cycle being from α-ketoglutarate to malate, downstream of the aconitase reaction. By contrast, gametocyte maturation was markedly reduced in the presence of NaFAc (Figure 3B). Under controlled conditions, most Stage II/III gametocytes progressed to Stage IV by day 2, and to Stage V by day 4 of the assay, with a gradual decrease in parasitemia (see Additional files 5 and 6) [27]. Treatment with NaFAc resulted in a dramatic, dose-dependent decrease in the development of viable Stage V gametocytes (Figure 3B; see Additional files 5and 6). Treatment of gametocytes with 10 mmol NaFAc was associated with the loss of the mitochondrial membrane potential, as shown by the absence of reticular rhodamine-123 staining (Figure 3C) [28]. This was also associated with reduced labeling of intracellular membranes with a red fluorescent dye (BODIPY-TR-ceramide; Invitrogen) (Figure 3D), reflecting decreased uptake and/or integrity of intracellular organelles [29, 30]. Thus, disruption of the TCA cycle in gametocytes may have pleiotropic effects on multiple processes, leading to loss of viability.

Our data allow a major revision of current models of central carbon metabolism of P. falciparum RBC stages by showing that both asexual and sexual stages utilize a canonical oxidative mitochondrial TCA cycle to catabolize host glucose and glutamine (Figure 4). Carbon skeletons derived from either glucose or glutamine enter the TCA cycle via acetyl-CoA or anaplerotic reactions, or α-ketoglutarate, respectively. Significantly, we found no evidence for operation of a bifurcated or branched TCA cycle in which glutamine is converted to malate via both oxidative and reductive ‘arms’ of the TCA cycle, as previously proposed [13, 17]. P. falciparum asexual RBC stages seem to have a compartmentalized TCA cycle, in which carbon backbones derived from glucose sustain a minor flux from oxaloacetic acid to citrate, whereas carbon backbones derived from glutamine are used to sustain a higher flux from α-ketoglutarate to malate. Based on the rate of glucose consumption and glycolysis (lactate production), less than 7% of the internalized glucose is catabolized in the mitochondrion. A low flux of glucose and glutamate into the TCA cycle may be required for the generation of reducing equivalents for the essential respiratory chain and synthesis of succinyl-CoA for heme biosynthesis [12]. However, as shown here, chemical inhibition of the entry of glucose-derived intermediates into the TCA cycle had no detectable effect on the development of asexual stages, possibly reflecting continued operation of glutaminolysis and catabolism of the carbon backbones of glutamate in this cycle.

P. falciparum and other apicomplexan parasites lack a mitochondrial isoform of PDH, and the enzyme involved in converting pyruvate to acetyl-CoA remains to be characterized. A possible candidate for this function is the mitochondrion-located branched chain α-keto acid dehydrogenase (BCKDH) complex [31]. This complex is retained in all apicomplexa (except Cryptosporidium spp.), as well as more distantly related protists that lack a mitochondrial PDH [32], and has been shown to utilize pyruvate in some organisms [33, 34]. The transcriptional upregulation of key subunits in the BCKDH complex (PF3D7_1312600) in P. falciparum gametocytes [35], may underlie the increased flux of pyruvate into the TCA cycle seen in this stage. However, entry of glutamine/glutamate into the P. falciparum TCA cycle could be regulated by either the glutamate dehydrogenase or by amino acid transaminases [36, 37]. Because the P. falciparum glutamate dehydrogenase is not required for growth of asexual stages [37], it is likely that amino acid transaminases alone can fulfill this role. We have previously shown that P. falciparum encodes a putative glutamate/GABA transaminase (PF3D7_0608800), as well as the proteins needed for synthesis of GABA (PF3D7_0405700) and import of this metabolite into the mitochondrion (PF3D7_0209600) [21]. The finding that GABA is synthesized in both asexual and sexual RBC stages, and that pyruvate is converted to alanine is consistent with both metabolites contributing to transamination reactions in the mitochondrion, which convert glutamate to α-ketoglutarate (Figure 4) [38]. It is notable that P. falciparum lacks an obvious homolog for succinate semi-aldehyde dehydrogenase, which is required for the conversion of GABA to the TCA-cycle intermediate succinate [21]. This differs from the situation in the related apicomplexan parasite, Toxoplasma gondii, which is able to utilize GABA as a short-term energy reserve [21], highlighting further specializations in the carbon metabolism of the Apicomplexa.

Despite entering a non-proliferative state, gametocytes exhibit increased levels of glucose utilization, glycolytic flux, and TCA catabolism of pyruvate. This switch to a more efficient method of energy generation may be needed to sustain gametocyte development under conditions of hypoglycemia, which is common in patients with severe malaria [39, 40]. The upregulation of TCA function in gametocytes might also reflect increased energy demands in female gametes preparing for the post-fertilization stages, during which access to glucose in the mosquito hemolymph may be limited [41]. Indeed, there is accumulating evidence that a functional TCA cycle is essential for the development of the ookinete [18]. Interestingly, deletion of the TCA-cycle enzyme succinate dehydrogenase had no observable effect on the production of P. berghei gametocytes [18]. By contrast, we found that partial inhibition of the TCA cycle with NaFAc inhibits gametocyte development, and high concentrations of NaFAc are lethal to this stage. These apparent differences could reflect species-specific differences in their dependence on the TCA for energy or maintenance of the respiratory chain, or the possibility that inhibition of the aconitase reaction results in a stronger metabolic phenotype, including reduced yield of NADH and/or the accumulation of intermediates such as citrate to toxic levels. Further gene knockout studies are required to definitively validate these possibilities.

The metabolic switch to increased TCA-cycle catabolism of glucose in gametocytes could also reflect changes in carbon-source availability in the infected RBC. Glutamine uptake in trophozoite-infected RBC is mediated by a specific RBC amino acid transporter [42] and novel permeation pathways (NPPs) induced by the parasite [43]. There is evidence that the NPPs are downregulated in developing gametocytes [44], resulting in reduced glutamine uptake. The reduced availability of glutamine in gametocytes might explain the switch to glucose as the preferred carbon source for TCA metabolism.

Recent transcriptional studies have suggested the existence of at least two physiological states of blood stage parasites in the human host, including a glycolysis-dependent asexual state and a ‘weakly gametocyte/sporozoite-like’ state that seems to have upregulated non-glycolytic energy metabolism, including the TCA cycle and oxidative phosphorylation [40, 45, 46]. Our results add to these findings, and suggest that the transcriptional upregulation of TCA cycle enzymes in response to glucose limitation or other metabolic stresses possibly represents a switch to a more energy-efficient metabolism.

Our findings reveal major changes in metabolic fluxes in P. falciparum bloodstream stages, which are not reflected in transcriptional-profiling studies, highlighting the importance of post-transcriptional regulatory mechanisms in controlling parasite metabolism. Stage-specific changes in metabolic fluxes in core metabolic pathways may be exploited in drug development. In particular, our results suggest that inhibitors of the mitochondrial enzyme responsible for converting pyruvate to acetyl-CoA could lead to a defect in gametocytogenesis. Field studies and mathematical modeling show that such transmission-blocking inhibitors will be needed to achieve the long-term goal of malaria elimination [47, 48].