Role of Cytosolic Glyceraldehyde-3-Phosphate Dehydrogenase in Visceral Organ Infection by Leishmania donovani

Abstract

The initial 7 steps of the glycolytic pathway from glucose to 3-phosphoglycerate are localized in the glycosomes in Leishmania, including step 6, catalyzed by the enzyme glyceraldehyde-3-phosphate dehydrogenase (GAPDH). In L. donovani and L. mexicana, there exists a second GAPDH enzyme present in the cytosol that is absent in L. braziliensis and that has become a pseudogene in L. major. To investigate the role of the cytosolic GAPDH (cGAPDH), an L. donovani cGAPDH-null mutant was generated, and conversely, the functional L. donovani cGAPDH was introduced into L. major and the resulting engineered parasites were characterized. The L. donovani cGAPDH-null mutant was able to proliferate at the same rate as the wild-type parasite in glucose-deficient medium. However, in the presence of glucose, the L. donovani cGAPDH-null mutant consumed less glucose and proliferated more slowly than the wild-type parasite and displayed reduced infectivity in visceral organs of experimentally infected mice. This demonstrates that cGAPDH is functional in L. donovani and is required for survival in visceral organs. Restoration of cGAPDH activity in L. major, in contrast, had an adverse effect on L. major proliferation in glucose-containing medium, providing a possible explanation of why it has evolved into a pseudogene in L. major. This study indicates that there is a difference in glucose metabolism between L. donovani and L. major, and this may represent an important factor in the ability of L. donovani to cause visceral disease.

INTRODUCTION

Unlike in higher eukaryotic cells, where the glycolytic metabolic process takes place in the cytosol, in protozoan Kinetoplastida, such as Trypanosoma and Leishmania, the glycolytic pathway is localized mainly in specialized organelles called glycosomes (1–8). The enzymes of the glycolytic pathway in the Kinetoplastida glycosomes are shown in Fig. 1 (4, 5). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (EC 1.2.1.12) is the enzyme that catalyzes the sixth step of glycolysis by converting glyceraldehyde 3-phosphate (GAP) into d-glycerate 1,3-bisphosphate (1,3BPGA). Computer modeling of glycolytic flux in bloodstream form Trypanosoma brucei suggests GAPDH and phosphoglycerate kinase (PGK) are the enzymes that could partly control the glycolytic flux (9–11). The crystal structures of T. brucei, Trypanosoma cruzi, and Leishmania mexicana glycosomal GAPDHs have been determined (12–14). Inhibitors that selectively block trypanosome glycosomal GAPDH activity, but not that of human GAPDH, have been shown to kill bloodstream T. brucei and intracellular T. cruzi (10, 15–17).

Fig 1.

Glycosomal and cytosolic glycolytic pathway in L. donovani. The glycolytic pathway between glucose and 3-phosphoglycerate (3-PGA) takes place mainly in the glycosomes. In some Leishmania species, such as L. donovani and L. mexicana, the sixth and seventh steps of the pathway between GAP and 3-PGA can also occur in the cytosol, whereas the final steps that lead to formation of pyruvate are only cytosolic. G6P, glucose-6-phosphate; F6P, fructose-6-phosphate; DHAP, dihydroxyacetone phosphate; 2-PGA, 2-phosphoglycerate. Enzymes: 1, hexokinase; 2, glucose-6-phosphate isomerase; 3, 6-phospho-1-fructokinase; 4, fructose-1,6-bis-phosphate aldolase; 5, triosephosphate isomerase; 6, glyceraldehyde 3-phosphate dehydrogenase; 7, phosphoglycerate kinase; 8, phosphoglycerate mutase; 9, enolase; 10, pyruvate kinase; 11, pentose phosphate pathway enzymes, including glucose 6-phosphate dehydrogenase, gluconolactonase, 6-phosphogluconate dehydrogenase, and ribulose 5-phosphate isomerase; 12, transketolase.

Although the Kinetoplastida glycolytic pathway takes place predominantly within glycosomes, GAPDH activity has also been detected in the cytosol of T. brucei and L. mexicana (18–20). Two copies of glycosomal GAPDH (gGAPDH) genes in tandem array and one copy of the cytosolic GAPDH (cGAPDH)-encoding gene have been identified in T. brucei and L. mexicana (18, 21). gGAPDH- and cGAPDH-encoding genes have also been identified in Leishmania infantum and Leishmania donovani. Interestingly, the cGAPDH gene has evolved into a pseudogene in Leishmania major and is completely absent in Leishmania braziliensis (22, 23). The gGAPDH enzymes contain sequences with glycosomal targeting signals. The gGAPDH and cGAPDH isoenzymes share only about 55% identity in amino acid sequence, and based on a phylogenetic analysis, it was proposed that the genes encoding the gGAPDH and cGAPDH isoenzymes were independently acquired by a trypanosomal ancestor (24). One gene belongs to the trypanosome lineage of origin, whereas the other may have entered the cell approximately 10⁸ years ago by horizontal gene transfer that occurred before the divergence of Trypanosoma and Leishmania (24–26).

Considering that the GAPDH and PGK activities were detected in the cytosol of T. brucei and L. mexicana and that the glycolytic intermediates can equilibrate across the glycosome membranes into the cytosol, possibly through the pore-forming channels (27, 28), and glyceraldehyde 3-phosphate can also be derived from the active pentose phosphate pathway present in the cytosol (29, 30), the parallel glycolytic pathway beginning with the sixth step of glycolysis could take place in the cytosol of trypanosomes and those Leishmania species expressing cGAPDH (31, 32) (Fig. 1).

In the present study, the cGAPDH in L. donovani was investigated, including whether it is necessary for visceral infection. We further investigated the effect of adding a functional cGAPDH back into L. major. This study demonstrates that the cGAPDH enzyme represents an evolutionary difference between L. donovani and L. major and that this difference plays a role in the ability of L. donovani to survive in visceral organs.

MATERIALS AND METHODS

Leishmania strains and culture conditions.

L. donovani 1S/Cl2D (MHOM/SD/62/1SCl2D) and L. major Friedlin V9 strains were used in this study. Leishmania promastigote culture medium and L. donovani axenic amastigote culture medium with glucose were as described previously (33–36). The RPMI glucose-deficient medium contains 1× RPMI 1640 medium without glucose (Sigma R1383), 10% heat-inactivated glucose-deficient fetal bovine serum (FBS) (dialyzed against 0.15 M NaCl with a 10,000-molecular-weight-cutoff cartridge; Sigma F0392), 100-U/ml penicillin, 100-μg/ml streptomycin, 0.1 mM adenosine, 10 μg/ml folic acid, and 25 mM HEPES, pH 7.4. The RPMI glucose medium was prepared by adding glucose (1 g/liter) to the above-mentioned RPMI glucose-deficient medium.

Leishmania promastigotes are routinely cultured at 27°C, pH 7.4, in glucose-rich medium. To determine Leishmania proliferation under different culture conditions, Leishmania promastigotes were shifted to glucose-deficient medium or to 37°C, pH 5.5, culture medium (axenic amastigote medium) to mimic the macrophage phagolysosome environment associated with the amastigote stage (37).

Construction of Leishmania expression vectors.

The L. donovani cGAPDH gene (an ortholog of the L. infantum cGAPDH gene, LinJ.36.2480) was PCR amplified from the L. donovani genomic DNA with the following primers: forward primer, 5′-cccaagcttaccATGGTCAAAGTGGGCATCAAC, and reverse primer 1, 5′-gacgagatcTCAGCGCGCGGACGTGTAGAGAA, or reverse primer 2 (for fusion with an A2 tag), 5′-cgagatctgtGCGCGCGGACGTGTAGAGAA (lowercase letters represent the restriction sites added to the 5′ end of each primer). The PCR products were double digested with HindIII and BglII and cloned into pLPneo, pLGFPC, and pLGFPN vectors (33, 34, 38, 39). pLGFPC was used to express green fluorescent protein (GFP) fusion proteins in Leishmania with GFP at the N terminus; pLGFPN has GFP at the C terminus.

Constructs for gene targeting.

To make the gene-targeting constructs for the L. donovani cGAPDH gene, primers 5′-gacccaagcttCTCACCTCTTCCTCCCTTCC and 5′gacgagatctGCGCAGACAAGAAGATTGGT were used to PCR amplify the 605-bp 5′-flanking DNA fragment of the L. donovani cGAPDH gene and primers 5′gacgggatccAGTCGCTATCGCTTCTCAGC and 5′gacgagatctACATCGATATGCAACGACGA for the 584-bp 3′-flanking DNA fragment. The 605-bp 5′-flanking DNA fragment digested with HindIII and BglII was inserted into the HindIII and BamH I sites of the pSPY-ble (35) or pSPY-hyg (40) vector (upstream of the bleomycin phosphotransferase gene or the hygromycin phosphotransferase gene); the 584-bp 3′-flanking DNA fragment digested with BamHI and BglII was ligated into the BglII site of the same pSPY-ble or pSPY-hyg vector (downstream of the bleomycin phosphotransferase gene or the hygromycin phosphotransferase gene). The gene-targeting fragment containing the phleomycin or hygromycin resistance gene flanked by the 5′ and 3′ L. donovani cGAPDH gene-flanking sequences was released from the targeting vectors by HindIII and BglII double digestion. The targeting fragment DNA (5 to 8 μg) purified from an agarose gel was used to transfect L. donovani promastigotes for primary or secondary gene targeting. For the first-round gene targeting, the L. donovani transfectants were selected in culture medium containing 100 μg/ml hygromycin. For the second-round gene targeting, the single L. donovani cGAPDH gene knockouts were transfected with the phleomycin targeting construct (20 μg/ml phleomycin [Cayla]).

Measurement of Leishmania proliferation in vitro culture.

The growth curves of Leishmania transfectants were measured as previously described (33, 38, 39). Briefly, promastigotes in stationary phase were seeded at a concentration of 2 × 10⁶ cells/ml in 96-well plates containing 200 μl of promastigote culture medium at 27°C, pH 7.4, or in vitro axenic amastigote culture medium at 37°C, pH 5.5. Each sample was plated in triplicate. The optical density at 600 nm (OD₆₀₀) values were measured directly from the plate daily for 5 to 7 days.

Glucose consumption assay.

To compare the glucose consumption of L. donovani wild-type and cGAPDH-null mutant cells, log-phase promastigotes were suspended in fresh promastigote or axenic amastigote culture medium containing 8 mM glucose at a cell density of 1.5 × 10⁷/ml; 200-μl cultures were taken out after culture for 6 and 22 h, respectively; and Leishmania cells were removed by centrifugation. The glucose concentration in the culture supernatants was determined with a glucose assay kit according to the manufacturer's instructions (Eton Bioscience Inc.).

Infection of BALB/c mice.

Female BALB/c mice were infected by tail vein injection with 1 × 10⁸ stationary-phase promastigotes in 100 μl PBS per mouse (33, 35, 36, 38, 39). Four weeks postinfection, the amastigotes were isolated from the livers and spleens of infected mice. The recovered amastigotes were cultured in promastigote culture medium, and the Leishmania parasite burdens were determined by limiting dilution. The liver parasite burdens were also measured by counting the amastigotes in the Giemsa-stained liver imprints, expressed as Leishman-Donovan units (LDU), as follows: number of amastigotes per 1,000 cell nuclei × liver weight (g) (35, 41).

RESULTS

GAPDH gene distribution in all sequenced Trypanosoma and Leishmania species.

Comparison of trypanosomatid genomes in TriTrypDB (http://tritrypdb.org/tritrypdb) revealed that all sequenced Trypanosoma and Leishmania species contain one or two copies of gGAPDH genes in tandem array. One copy of a cGAPDH-encoding gene is also present in L. mexicana, the L. donovani complex species, Leishmania tarentolae, and all sequenced Trypanosoma species. Surprisingly, the cGAPDH gene, together with the cytosolic phosphoglycerate kinase gene, is absent in L. braziliensis, and the cGAPDH gene has become a pseudogene in L. major (22, 23, 42). Interestingly, although L. tarentolae possesses the cGAPDH gene, it does not have the cytosolic phosphoglycerate kinase gene. The L. infantum cGAPDH has 80 to 85% identity with its counterparts in L. mexicana and Trypanosoma species. As the L. major genome sequence shows, our PCR and sequencing analysis confirmed that a stop codon is present at amino acid position 58 in both alleles of the cGAPDH gene in L. major. Given that the next available ATG translation initiation site is further down at amino acid position 118, the gene product will miss the first 117-amino-acid sequence, which contains the NAD⁺ binding domain critical for GAPDH function. Interestingly, this L. major pseudogene still has 82% amino acid identity to L. infantum cGAPDH (Fig. 2). Since there is no selective pressure to retain a nonfunctional sequence, this suggests that the cGAPDH gene turned into a pseudogene is a recent evolution event in L. major. L. infantum cGAPDH shares only 55% identity with its gGAPDH (Fig. 2). It is important to note that recent resequencing of L. major and L. braziliensis genomes further confirmed that the cGAPDH gene is a pseudogene in L. major and is missing in L. braziliensis (23).

Fig 2.

Amino acid sequence alignments of L. infantum cGAPDH (Lic) (LinJ.36.2480) with L. mexicana cGAPDH (Lmxc) (LmxM.36.2350), the L. major cGAPDH pseudogene product (Lmc) (LmjF.36.2350), and L. infantum gGAPDH (Lig) (LinJ.30.3000). The dashes represent identical amino acids; the dots represent amino acid gaps introduced for better alignment. The stop codon at amino acid 58 in the L. major cGAPDH pseudogene coding sequence is marked with an asterisk. The tripeptide AKM (boldface) is the glycosomal targeting signal present in the C terminus of L. infantum gGAPDH. Note that the following data are not shown: L. donovani BPK282A1 cGAPDH has only two amino acid differences from L. infantum cGAPDH (Ld:A287G:Li and Ld:D302E:Li).

Initially, it was necessary to confirm whether L. donovani cGAPDH was indeed present in the cytosol. Fusion proteins with GFP were therefore generated and expressed in transfected L. donovani. As shown in Fig. 3, expression of cGAPDH-GFP fusion proteins from L. donovani (genetically closely related to L. infantum) confirmed that cGAPDH is evenly distributed throughout the L. donovani promastigotes and absent from the nucleus and kinetoplast, demonstrating that the enzyme is not confined to glycosomes.

Fig 3.

Cytosolic localization of L. donovani cGAPDH. (A) Western blot with anti-GFP antibodies showing expression of cGAPDH-GFP or GFP-cGAPDH fusion protein in transfected L. donovani promastigotes. (B) Cytosolic localization of cGAPDH-GFP and GFP-cGAPDH fusion proteins in L. donovani promastigotes. Note that L. donovani cGAPDH is evenly distributed in the cytosol and not present in the nucleus or kinetoplast. The nuclear (N) and kinetoplast (K) DNAs are stained white with 4′,6-diamidino-2-phenylindole (DAPI) fluorescent dye in these cells. The bars represent 2 μM.

Generation of an L. donovani cGAPDH-null mutant.

Since the glycolytic pathway largely occurs in glycosomes and cGAPDH is not present in all Leishmania species, it was of considerable interest to determine the role of the cGAPDH enzyme in L. donovani. To investigate this, we undertook a genetic approach by generating an L. donovani cGAPDH-null mutant using gene targeting. As shown in Fig. 4, two rounds of gene targeting were successfully carried out to delete both alleles of the cGAPDH gene. After the first round of gene targeting with the hygromycin resistance gene-targeting construct, PCR analysis with one primer specific for the hygromycin resistance gene (Hyg) and another primer (E) for the flanking sequence downstream of 3′cGAPDH obtained a predicted 725-bp PCR fragment (Hyg + E), demonstrating that the Hyg resistance gene was correctly targeted to one of the cGAPDH gene alleles. Similarly, a bleomycin resistance gene-targeting construct was used for replacing the second cGAPDH gene allele. As shown in Fig. 4B, the predicted 834-bp PCR fragment (Ble + E) was obtained with a primer specific for the bleomycin resistance gene and flanking primer E, demonstrating that the bleomycin resistance gene was correctly targeted to the remaining cGAPDH gene allele. Thus, as expected, no cGAPDH gene PCR fragment was present after PCR analysis with primer pairs (F and R) specific for the cGAPDH gene in the resulting L. donovani mutant.

Fig 4.

Generation of the L. donovani cGAPDH-null mutant. (A) Homologous replacement strategy with hygromycin and bleomycin resistance gene-targeting constructs. The thick black lines represent the 5′ and 3′ sequences flanking the cGAPDH gene. The primers used to verify the correct gene replacement events and the sizes of PCR products are indicated: F, forward primer; R, reverse primer; Hyg, hygromycin resistance gene-specific primer; Ble, bleomycin resistance gene-specific primer; E, primer from downstream of the 3′-flanking targeting sequence. (B) PCR analysis of wild-type L. donovani cells (+/+), single cGAPDH-targeting (hygromycin) cell line (+/−), and double cGAPDH replacement cell line (null mutant; −/−), with primer pairs as indicated.

Proliferation of the L. donovani cGAPDH-null mutant in the presence and absence of glucose.

We initially compared the proliferation of the wild type and the cGAPDH-null mutant as axenic promastigotes and amastigotes in normal glucose-containing media. As shown in Fig. 5, the null mutant displayed a significant reduction in proliferation compared to the control. Addition of the cGAPDH gene in an episomal vector back to the null mutant was able to restore the growth rate to near the level of the wild-type cells. The reduced proliferation for the cGAPDH-null mutant could not be corrected by supplementing the medium with pyruvate, suggesting the growth defect was not a result of reduced capacity to synthesize pyruvate in the cell (see Discussion).

Fig 5.

In vitro proliferation curves of the L. donovani cGAPDH-null mutant in glucose-containing medium with or without added pyruvate. (A) Promastigote growth curves. (B) Axenic amastigote growth curves. LdWT, wild-type L. donovani; cGAPDH−/−, L. donovani cGAPDH-null mutant; cGAPDH−/−+, cGAPDH-null mutant plus the L. donovani cGAPDH episomal expression plasmid; cGAPDH−/− (Pyru), cGAPDH−/− mutant in medium supplemented with 1 mM pyruvate. The data are the means of triplicate OD₆₀₀ values ± the standard errors of the mean (SEM). The cell density differences between the wild type and the cGAPDH-null mutant are statistically significant starting from day 2 in both promastigote and axenic amastigote cultures (*, P < 0.05; **, P < 0.01). The data are representative of four independent experiments.

To investigate whether reduced proliferation of the cGAPDH-null mutant resulted from factors unrelated to glycolysis, the cGAPDH-null mutant and wild-type cells were cultured in glucose-deficient medium to direct the cells to use other energy sources, such as fatty acid, proline, or threonine, as previously described in procyclic T. brucei (43). As expected, both the wild type and the cGAPDH-null mutants proliferated much more slowly in glucose-deficient medium than in glucose-containing medium (Fig. 6). However, despite the lower proliferation rate, the cGAPDH-null mutant was able to proliferate at a rate similar to that of the wild-type L. donovani cells in the glucose-deficient medium. This indicates that cGAPDH is required for normal proliferation in glucose-containing medium, suggesting that cGAPDH participates in the cytosolic glycolysis pathway to provide energy for optimal proliferation.

Fig 6.

In vitro growth curves of L. donovani cGAPDH-null mutant promastigotes in medium with or without glucose. LdWT, wild-type L. donovani; cGAPDH−/−, cGAPDH-null mutant; +Glu, medium with 1 g/liter glucose; −Glu, medium with no glucose. The data are the means of triplicate OD₆₀₀ values ± SEM. The cell density differences between the wild type and the cGAPDH-null mutant are statistically significant starting from day 2 in glucose-containing culture and from day 3 in glucose-deficient culture (*, P < 0.05; **, P < 0.01). The data are representative of four independent experiments. Note that we also counted the cells at each time point by microscope; we found that the cell numbers were in good agreement with the OD₆₀₀ values, although the OD₆₀₀ values appeared to be low in glucose-deficient medium.

To further investigate the role of cGAPDH in glycolysis in L. donovani, we compared the glucose consumption of wild-type L. donovani, the cGAPDH-null mutant, and the cGAPDH add-back cells in culture medium containing glucose. Equal numbers of log-phase wild-type L. donovani cells and the cGAPDH mutants were suspended in fresh axenic promastigote or amastigote culture medium, and glucose consumption was determined by measuring the glucose concentrations in the culture medium after 6 and 22 h. As shown in Fig. 7, after 6 h, the glucose concentration was slightly lower in wild-type L. donovani and in cGAPDH add-back cell cultures than in cGAPDH-null mutant culture. However, a more significant difference in the glucose concentration was observed after culture for 22 h. The cGAPDH-null mutant utilizes less glucose than wild-type and cGAPDH add-back cells in both promastigote and axenic amastigote culture, strongly suggesting that cGAPDH is involved in glycolysis in L. donovani.

Fig 7.

The L. donovani cGAPDH-null mutant consumes less glucose than wild-type L. donovani in culture medium containing glucose. Log-phase promastigotes were suspended in fresh promastigote or axenic amastigote culture medium containing 8 mM glucose at a cell density of 1.5 × 10⁷/ml. Glucose concentrations in these media after culture for 6 and 22 h were determined. (A) Glucose consumption in promastigote culture. (B) Glucose consumption in axenic amastigote culture. LdWT, wild-type L. donovani; cGAPDH−/−, cGAPDH-null mutant; cGAPDH−/−+, cGAPDH-null mutant plus the L. donovani cGAPDH episomal expression plasmid. The data are the means of glucose concentrations in triplicate cultures ± SEM. The glucose concentration differences in culture media between the wild type and the cGAPDH-null mutant are statistically significant after culture for 22 h in both promastigote culture and axenic amastigote culture (*, P < 0.05; **, P < 0.01). The data are representative of three independent experiments.

The L. donovani cGAPDH-null mutant displays reduced infectivity in visceral organs.

To examine whether L. donovani requires cGAPDH for survival in visceral organs, BALB/c mice were infected with wild-type and cGAPDH-null mutant promastigotes by intravenous injection, and the infection levels were determined after 4 weeks. As shown in Fig. 8, the liver and spleen parasite burdens of mice infected with the cGAPDH-null mutant were significantly reduced compared with wild-type L. donovani-infected mice. Addition of the cGAPDH gene in an episomal vector back to the null mutant was able to partially restore infectivity, confirming the requirement for cGAPDH in visceral organ survival.

Fig 8.

The L. donovani cGAPDH-null mutant displayed reduced infectivity in the livers and spleens of BALB/c mice. L. donovani cGAPDH mutants were used to infect BALB/c mice by tail vein injection (5 mice per group); 4 weeks after infection, the mice were examined for liver and spleen parasite burden. (A) Liver parasite burden. (B) Spleen parasite burden. LdWT, wild-type L. donovani; cGAPDH−/−, cGAPDH-null mutant; cGAPDH−/−+, cGAPDH-null mutant plus the L. donovani cGAPDH episomal expression plasmid. The data are the means plus SEM. The LDU and spleen parasite burden differences are statistically significant (P < 0.01) between mice infected with the cGAPDH-null mutant and mice infected with wild-type parasites or the cGAPDH add-back mutant. The experiment was repeated two times with similar results.

Restoration of cGAPDH activity suppressed L. major proliferation in glucose-containing, but not glucose-deficient, medium.

Unlike in L. donovani, the cGAPDH gene has become a pseudogene in L. major (Fig. 2). It was therefore interesting to investigate whether restoring cGAPDH activity in L. major would result in increased proliferation in the presence of glucose. An episomal expression vector containing the L. donovani cGAPDH gene was transfected into L. major, and the transfectants were placed in glucose-containing or glucose-deficient medium. Interestingly, in glucose-containing medium, the presence of cGAPDH suppressed L. major proliferation and reduced cell density at stationary phase compared to the control (Fig. 9). In contrast, the proliferation of L. major containing cGAPDH was similar to that of the control in glucose-deficient medium. This indicates that, unlike in L. donovani, cGAPDH activity is not beneficial and may be harmful to L. major. cGAPDH-containing L. major consumed no more glucose than control transfected L. major in glucose-containing culture medium, and this could be due to the adverse effect of cGAPDH on L. major (data not shown).

Fig 9.

In vitro promastigote proliferation curves of L. major containing the L. donovani cGAPDH gene (Lm+cGAPDH) and control transfected L. major (Lm neo) in medium with (+Glu) or without (−Glu) glucose. The data are the means of triplicates ± SEM. The cell density differences between Lm+cGAPDH and Lm neo are statistically significant starting from day 3 in glucose-containing culture (*, P < 0.05; **, P < 0.01). The data are representative of four independent experiments.

DISCUSSION

Since the glycolytic pathway from glucose to 3-phosphoglycerate can take place exclusively in the glycosomes, it is not clear why the cytosolic counterpart of GAPDH is present in only some Leishmania species, such as L. donovani. In this study, we demonstrate that deletion of the L. donovani cGAPDH gene reduced glucose uptake, suppressed proliferation in glucose-containing medium, and, most significantly, attenuated survival in visceral organs. These observations indicate that the cGAPDH gene in L. donovani is functional and plays a role in visceral disease pathogenesis.

It has been previously observed that glycolytic intermediates can equilibrate across the glycosome membranes into the cytosol (28). Indeed, a recent study has suggested the glycosome membrane contains several types of pore-forming channels that allow movement of small solutes across the membrane, including intermediate metabolites of glycolysis with molecular masses up to 300 to 400 Da (27). Together with the observations in this study, the parallel glycolytic pathway beginning with the sixth step of glycolysis must also take place in the cytosol of these Leishmania species expressing cGAPDH (Fig. 1). Since in silico analysis of glycolytic flux in bloodstream form T. brucei shows GAPDH is one of the enzymes that could play a role in the control of the glycolytic pathway (9–11), the presence of the partial parallel glycolytic pathway in the L. donovani cytosol could increase the overall glycolytic flux and thus the generation of ATP. The partial parallel glycolytic pathway may also help to balance the cytosolic NAD⁺/NADH ratio. Throughout this study, we observed that growth of the L. donovani cGAPDH-null mutants appeared to be more profoundly affected in late log and stationary phases than during the early log phase (Fig. 5 and 6). This could be due to the fact that the fast-growing Leishmania cells in log phase require more active glycolytic flux and GAPDH activity for energy production (11, 44), and the accumulating effect makes the cell density difference more obvious in late log and stationary phases than in early log phase.

A recent glucose metabolic labeling study of L. mexicana promastigotes has shown that, instead of being converted into pyruvate, the usual final product of aerobic glycolysis, the majority of phosphoenolpyruvate (PEP) generated by glycolysis enters the glycosome and is fermented into succinate to regenerate the ATP and NAD⁺ consumed by upper glycolytic pathway reactions, and most of the C₄ dicarboxylic acids generated during succinate fermentation are further catabolized in the tricarboxylic acid (TCA) cycle (32). This could explain why the cGAPDH-null mutant proliferation rate was not improved by supplementing the medium with pyruvate, as shown in Fig. 5.

It has been suggested from proteomic analysis that during differentiation from extracellular promastigotes into intracellular amastigotes, Leishmania parasites shift from glucose to amino acids and fatty acids as the main energy sources (45). However, several studies have shown that glucose and/or amino sugars are required for the amastigote stage, though they might be mainly needed for the formation of glycoconjugates (46–49). Leishmania amastigotes are able to utilize amino sugars in the phagolysosome of mammalian macrophages as a source of carbon and energy (47). Interestingly, gene deletion study of the gluconeogenic enzyme fructose-1,6-bisphosphatase (FBP) showed Leishmania also requires gluconeogenesis for its virulence (50). Since GAPDH is also able to catalyze the reverse reaction from 3-phospho-d-glyceroyl phosphate into GAP in gluconeogenesis. The observation in this study that the L. donovani cGAPDH-null mutant had reduced infectivity in visceral organs suggests that cGAPDH could be involved in glycolysis and/or gluconeogenesis pathways when L. donovani resides as an amastigote in the macrophage phagolysosome, where the sugar levels are relatively low (47, 51).

The available crystal structures of T. brucei, T. cruzi, and L. mexicana gGAPDHs have enabled the design of inhibitors that effectively block trypanosomatid gGAPDH, and they were lethal to bloodstream T. brucei and intracellular T. cruzi (10, 15–17). The bloodstream stage of T. brucei uses glycolysis as its only energy source and the cytosolic phosphoglycerate kinase expression is suppressed, which would explain why inhibitors of gGAPDH are lethal to bloodstream T. brucei. However, it is unknown whether these gGAPDH inhibitors could also inhibit cGAPDH and whether they were lethal to L. mexicana (10). Together with the observations in this study, it would be appropriate to develop inhibitors to cGAPDH in L. donovani, since the null mutant was significantly attenuated in visceral organs. However, the ideal inhibitor should be capable of blocking both the glycosomal and cytosolic GAPDH activities. This may be difficult with a single inhibitor, since the enzymes share only 55% homology.

It is not clear why the cGAPDH gene has evolved into a pseudogene in L. major and is absent in L. braziliensis but present in L. donovani and L. mexicana. Unlike other Leishmania species, L. braziliensis also does not have cytosolic phosphoglycerate kinase (cPGK). In the present study, it was not expected that introduction of L. donovani cGAPDH into L. major would result in an adverse effect on proliferation in glucose-containing medium. This could provide some rationale as to why the gene has evolved into a pseudogene in L. major.

The principal function of GAPDH is to catalyze the sixth step of glycolysis to produce energy and carbon molecules. In addition, GAPDH has recently been implicated in several nonmetabolic processes in higher eukaryotic cells, including transcription activation, initiation of apoptosis, and endoplasmic reticulum (ER)-to-Golgi apparatus vesicle shuttling (52, 53). An indication of GAPDH involvement in nonmetabolic processes in higher eukaryotic cells is the translocation of the enzyme to the nucleus in response to cellular stress (53). Our studies showed that there was no translocation into the nucleus for L. donovani GFP-cGAPDH fusion proteins (Fig. 3) when these Leishmania cells were exposed to cellular stress, such as serum withdrawal or stress induced by exposure to hydrogen peroxide (data not shown). It is nevertheless not possible to rule out the possibility that L. donovani cGAPDH plays some roles independent of glycolysis.