Metabolic perturbation of indolmycin-treated parasites during delayed death

We sought to investigate the metabolomic consequences of delayed death using specific inhibition of apicoplast translation. The inhibitor with the clearest direct evidence for an unambiguous apicoplast target is indolmycin, a specific inhibitor of the apicoplast tryptophanyl-tRNA synthetase (TrpRS^api) in P. falciparum [26]. Indolmycin inhibits TrpRS^api activity in vitro, stops protein translation in the apicoplast, and causes delayed death in P. falciparum that can be rescued by IPP supplementation [26]. To investigate the metabolic consequence of delayed death on P. falciparum, infected RBCs (iRBCs) were treated with indolmycin at a concentration that causes no death in the first IDC but complete death in the second IDC (equivalent to 50× the 96-hour EC₅₀ = 50 μM), and samples were collected at regular time intervals during the first and second IDC following treatment for analysis by liquid chromatography mass spectrometry (LC-MS) (Fig 1A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Untargeted metabolomic analysis of delayed death-parasites indicates isoprenoid defect and decreased haemoglobin turnover.

(A) Schematic of delayed death in P. falciparum. During the first IDC, parasites develop over approximately 48 hours from the ring-stage into a metabolically active trophozoite, before dividing into multiple daughter merozoites by schizogony. Parasites treated with an apicoplast inhibitor during this first IDC successfully complete schizogony, but the daughter merozoites inherit a defective apicoplast. The lethal effect of the drug manifests only after invasion of a new host cell, in which parasites fail to complete a second IDC. Delayed-death parasites in this study were sampled for metabolomic analysis by LC-MS at the indicated time intervals post drug administration. (B) Heat map summarising the metabolic effect of the delayed-death inhibitor indolmycin (50 μM) on P. falciparum-infected RBC cultures. Total metabolite pools (measured as ion counts) across two IDCs were determined relative to untreated controls. Samples were collected for analysis at 30, 58, 68, 78, and 88 hours post drug administration, equivalent to 30 hpi in the first IDC and then 14, 24, 34, and 44 hpi in the second IDC. Changes in total metabolite pools are expressed as log2 ratios of indolmycin-treated cultures compared to untreated controls. Data are presented as the means of three independent experiments. See S1 Data for raw metabolomics data underlying heat map. hpi, hours post invasion; IDC, intraerythrocytic developmental cycle; LC-MS, liquid chromatography mass spectrometry; RBC, red blood cell.

https://doi.org/10.1371/journal.pbio.3000376.g001

The relative abundance (indolmycin treated/untreated) of the 159 metabolites detected in the time series is summarised in Fig 1B. Indolmycin treatment did not cause metabolic perturbation during the first IDC following treatment (30-hour time point). During the second IDC following drug treatment, the relative abundance of the isoprenoid biosynthetic intermediates 2-C-methyl-erythritol-2,4-cyclodiphosphate and DOXP, and the downstream isoprenoid species farnesyl pyrophosphate (FPP) and geranylgeranyl pyrophosphate (GGPP) progressively decreased in the indolmycin-treated parasites. The perturbation to isoprenoid biosynthesis is consistent with parasites possessing a defective apicoplast that cannot make new isoprenoids.

Few other metabolic changes were observed between indolmycin-treated and untreated parasites during the second IDC following drug treatment, with the exception of several haemoglobin-like peptides decreasing in abundance at 58 hours (equivalent to 14 hours post invasion [hpi] in the second IDC) and beyond (Fig 1B). This perturbation suggested that catabolism of RBC-derived haemoglobin may be compromised during indolmycin-induced delayed death. Diminished haemoglobin metabolism might be caused either by aberrant proteolysis in the parasite DV or by disrupted uptake/trafficking of haemoglobin to the DV. Given the essential role of prenylated proteins in vesicular trafficking and membrane fusion [27–29], it follows that decreased abundance of the prenyl precursors FPP and GGPP (Fig 1B) could disrupt the function of proteins involved in haemoglobin metabolism; below, we show that this is indeed the case.

Protein prenylation is disrupted during delayed death

To determine if the decreased abundance of the higher isoprenoid species and prenylation substrates FPP and GGPP detected in the metabolomics study corresponds to disruption of cellular protein prenylation, we used a commercial anti-farnesyl antibody, which is cross-reactive to proteins with a farnesyl or geranylgeranyl modification. iRBCs were treated with 50 μM indolmycin for 30 hours and, cell lysates were collected at 72–78 hours during their second IDC following treatment (equivalent to 28–34 hpi in the second IDC). In untreated (DMSO only) parasite lysates, two prominent bands were detected at approximately 50 kD and 25 kD (Fig 2A), consistent with previously published prenylated species recognised by anti-farnesyl antibodies [30] and (³H)prenyl/polyprenol labelling in malaria parasites [31,32]. The parasite prenylome includes prenylated proteins at the approximate sizes detected: Rab guanosine-5'-triphosphate (GTP)ases (23–27 kD) and a HSP40 analogue (PF3D7_1437900) (40 kD) [27,28]. Indolmycin-treated parasites, collected 72–78 hrs following drug treatment (equivalent to 28–34 hpi in the second IDC), had a significant reduction in Rab proteins detected by anti-farnesyl (P < 0.001, two-tailed Student t test) (Fig 2B), suggesting a reduced capacity for delayed-death parasites to prenylate target proteins in the IDC after treatment.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Apicoplast inhibitors decrease global parasite prenylation in the second IDC after treatment.

(A) Immunoblot analysis of parasites treated with indolmycin (50 μM) with or without polyprenol rescue (5 μM GGOH), as indicated. Parasite lysates were collected after saponin isolation during their second IDC, and immunoblots were probed using anti-farnesyl (1:2,000) to analyse global parasite prenylation. Prenylated species are annotated as putative Rabs (23–27 kD) or HSP40 (40 kD) (assignment based on published P. falciparum prenylomes) [27,28]. Anti-Bip (1:1,000) were used as a loading control with an expected size of 65 kD. Immunoblot is representative of four independent experiments. (B) Quantification of the 27 kD Prenylated Rabs, normalised to average density of untreated. Data are presented as the means of four independent experiments ± SD. ***P < 0.001, two-tailed Student t test. See S2 Data for numerical data underlying figure. GGOH, geranylgeraniol; IDC, intraerythrocytic developmental cycle; ns, not significant.

https://doi.org/10.1371/journal.pbio.3000376.g002

To determine if the apparent failure to prenylate protein in delayed-death parasites is due to lack of isoprenoids used for protein prenylation (i.e., FPP/GGPP), we substituted the polyprenol analogue geranylgeraniol (GGOH) into the medium of indolmycin-treated parasites. GGOH has previously been reported to protect parasites from fosmidomycin treatment [33]. We therefore hypothesised that GGOH substitution following indolmycin treatment would facilitate further protein prenylation by complementing the depletion of de novo prenyl groups. Consistent with this hypothesis, the indolmycin-treated parasites supplemented with 5 μM GGOH show bands at the same size as detected in the untreated sample (Fig 2A).

Polyprenol substitution rescues parasites from delayed death

Supplementing the medium of indolmycin-treated parasites with the isoprenoid precursor IPP (>200 μM) is sufficient to rescue parasites from delayed death indefinitely (Fig 3A) [18]. Growth inhibition of iRBCs by indolmycin has a 48-hour EC₅₀ > 50 μM (n = 3), a 96-hour EC₅₀ = 1.1 μM (n = 3, SD ± 1.1 μM), and a 144-hour EC₅₀ = 0.1 μM (n = 3, SD ± 0.12 μM) (Fig 3B). With 200 μM IPP supplementation, iRBC are protected from the lethal effect of indolmycin up to the maximum concentration assayed (50 μM) at both 96 hours and 144 hours (Fig 3A and 3B). Given that GGOH supplementation protects against the depletion of prenylated proteins, we tested whether polyprenol precursors have a dose-dependent protective effect against indolmycin treatment. We first performed growth inhibition assays using GGOH and farnesol (FOH; the alcohol analogue of the prenyl group FPP) and determined that concentrations of GGOH exceeding 20 μM or FOH exceeding 30 μM were antagonistic to parasite growth (S1 Fig). To be used as productive substrates for protein prenylation, FOH and GGOH must first be phosphorylated (by unknown kinases) to FPP and GGPP; however, it is possible that FOH and GGOH compete directly as unproductive FPP and GGPP substrates and that this has a dose-dependent toxic effect on the parasite. Of these two polyprenol compounds, GGOH but not FOH showed protective effects when supplemented into the medium of indolmycin-treated parasites, with 5 μM GGOH being the lowest optimal protective concentration without antagonising parasite survival (S2A Fig). Likewise, supplementation of indolmycin-treated parasites with a combination of GGOH or FOH showed no synergistic protective effects (S2B Fig), suggesting that GGOH but not FOH is readily used as a prenylation substrate, which is consistent with ex vivo studies where [³H]GGOH was used preferentially to [³H]FOH as a substrate to modify protein in P. falciparum [32].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Polyprenol supplementation with GGOH protects parasites from the effect of apicoplast inhibitors in the second IDC after treatment.

(A) Dose-response curve from SYBR-Green susceptibility assay determined 48, 96, and 144 hours post indolmycin treatment, with polyprenol (5 μM GGOH) or isoprenoid (200 μM IPP) supplementation as indicated. Indolmycin causes a delayed-death effect (inhibition at 96 but not 48 hours) that is rescued by GGOH or IPP supplementation. Inhibition at 144 hours is rescued by IPP but not GGOH supplementation. Data are presented as the means of three independent experiments ± SEM. See S2 Data for numerical data underlying figure. (B) Summary table of EC_50s for delayed-death antimalarials indolmycin, clindamycin, and chloramphenicol, determined at 48, 96, and 144 hours post treatment, with polyprenol (5 μM GGOH) or isoprenoid (200 μM IPP) rescue as indicated. Data are presented as the means of three independent experiments ± SD. *P < 0.05, two-tailed Student t test. (C) Thin-blood smears were collected for Giemsa-microscopy every 6 hours during the second IDC following indolmycin treatment. Scale bar = 5 μm. GGOH, geranylgeraniol; IDC, intraerythrocytic developmental cycle; IPP, isopentenyl pyrophosphate; ns, not significant.

https://doi.org/10.1371/journal.pbio.3000376.g003

Supplementing the medium with 5 μM GGOH increased the 96-hour EC₅₀ of indolmycin to >50 μM (n = 3) (Fig 3B). Importantly, GGOH was supplemented only after replacing the medium in the second IDC. However, unlike IPP, the protective effect of GGOH is only temporary, with no difference between the 144-hour EC₅₀ values for indolmycin-treated and GGOH rescue (Fig 3B). This shows that provision of the polyprenol precursor GGOH rescues parasites during the second IDC following indolmycin treatment but that the rescued parasites still eventually succumb to the lethal effect of the drug precluding a third IDC (Fig 3C). We confirmed that temporary rescue with GGOH was common to delayed death–causing drugs by repeating the analysis with clindamycin and chloramphenicol, antibiotics that target separate components of the apicoplast protein-translation apparatus (Fig 3B), and which have previously been shown to be completely rescuable using IPP [34]. 5 μM GGOH supplementation protected parasites from delayed death up the maximum concentration of clindamycin assayed (>10 μM, n = 3, approximately 1,000× 96-hour EC₅₀), and chloramphenicol showed a significant fold reduction in its 96-hour EC₅₀ when treated parasites were supplemented with 5 μM GGOH from 27.3 μM, (n = 3, SD ± 6.6 μM) to 65.64 μM (n = 3, SD ± 10.8 μM) (P < 0.05, two-tailed Student t test) (Fig 3B and S3 Fig). However, consistent with our findings using indolmycin, GGOH did not protect parasites from treatment with clindamycin or chloramphenicol in the third IDC after treatment (Fig 3B and S3 Fig). Together, these data show that the failure to prenylate proteins during the parasite’s second IDC following indolmycin-treatment is the proximate cause of delayed death and that supplementing the polyprenol precursor GGOH is sufficient to alleviate the effects of isoprenoid fatigue in these parasites. However, restoring the capacity of delayed-death parasites to prenylate proteins alone is insufficient to indefinitely rescue parasites. As such, we hypothesise that disrupting protein prenylation is the immediate cause of parasite death in the second IDC but that even if this pathway is supplemented, another lethal insult ultimately arises, resulting in growth arrest that precludes a third IDC.

Aberrant uptake/trafficking of RBC cytoplasm to the DV during delayed death

Considering the observed disruption in haemoglobin metabolism (Fig 1B), we sought to determine whether host haemoglobin uptake or trafficking is disrupted during delayed death. We investigated parasite-mediated RBC cytoplasm internalisation using a fluorescein-dextran (F-dextran) uptake assay modified from Frankland and colleagues [35] and Baker and colleagues [36]. Uninfected RBCs were loaded with F-dextran such that their cytosolic contents are fluorescently labelled. Newly invaded, preloaded iRBCs were treated with 50 μM indolmycin for 30 hours and imaged by live-cell fluorescence microscopy 72–78 hours following drug treatment (equivalent to 28–34 hpi in the second IDC) (Fig 4A). In the untreated iRBCs (DMSO only), the F-dextran was internalised by the parasite and concentrated into one or sometimes two compartments (Fig 4B), consistent with Baker and colleagues [36]. In parasites with two F-dextran-labelled compartments, the larger compartment was typically adjacent to a smaller compartment (Fig 4A, untreated). This is consistent with a single parasite DV and transitory vesicles containing F-dextran internalised from the RBC cytoplasm. In the indolmycin-treated group, the F-dextran was more frequently concentrated into multiple compartments (P < 0.01, one-tailed Student t test) (Fig 4B). This suggests that although the initial internalisation of RBC cytoplasm by treated parasites is unaffected, the trafficking or fusion of the internalised vesicles to the DV is disrupted. This led to an accumulation of aberrant compartments that is consistent with a defective RBC cytoplasm uptake phenotype [37]. Parasites treated with indolmycin but then supplemented with GGOH after treatment internalised the F-dextran into one or two compartments, equivalent to untreated samples (Fig 4B). This demonstrates that delayed death is linked to the aberrant uptake of RBC cytoplasm, but this is ameliorated by polyprenol substitution, which suggests that the vesicle trafficking defect is prenylation dependent.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Indolmycin treated parasites aberrantly internalise F-dextran from preloaded RBCs in the second IDC after treatment.

Uninfected RBCs were preloaded with F-dextran by gentle hypotonic lysis and resealing. Following this, enriched and synchronised schizont-stage parasites were added to the loaded RBCs and merozoites were allowed to reinvade. Newly invaded ring-stage parasites were treated with indolmycin (50 μM), with and without polyprenol (5 μM GGOH) rescue as indicated. (A) Representative live-cell images taken 72–78 hours after drug administration are shown (equivalent to 28–32 hpi in the second IDC after treatment). F-dextran, green signal; Hoechst: parasite nuclei, blue signal; merge: BF and blue signal. Scale bar = 5 μm. (B) The number of F-dextran compartments per iRBC were scored in three independent experiments. **P < 0.01, one-tailed Student t test. Scatter dot plot shows error bars with mean ± SD. See S2 Data for numerical data underlying figure. BF, bright field; F-dextran; fluorescein dextran; GGOH, geranylgeraniol; hpi, hours post invasion; IDC, intraerythrocytic developmental cycle; ns, not significant; RBC, red blood cell.

https://doi.org/10.1371/journal.pbio.3000376.g004

Ultrastructural analysis of the parasite DV reveals fragmented biogenesis

Internalised RBC cytoplasm accumulates within multiple aberrant compartments of the delayed-death parasites. To investigate the nature of these aberrant compartments, we performed an analysis of parasite ultrastructure by electron microscopy. Initial analysis of ultrathin sections of indolmycin-treated parasites by transmission electron microscopy revealed aberrant DVs as well as multiple membrane-bound enclosures of RBC cytoplasm (S4 Fig). The 3D arrangement of such aberrant compartments is difficult to decipher from individual sections, so we examined the 3D ultrastructure of untreated and delayed-death parasites using reconstruction of hundreds of sections after serial block-face scanning electron microscopy (SEM) (Fig 5 and S1 Movie). 3D reconstructions of multiple trophozoite-infected RBCs supported our initial observations from 2D electron microscopy sections (Fig 5A)—that indolmycin treatment disrupts formation of the DV and that treated parasites generate multiple DV fragments, each containing haemozoin crystals (Fig 5B and S2 Movie). Remarkably, GGOH rescue completely ameliorated the fragmented DV phenotype, restoring a single DV (Fig 5B). Quantification of the number of DV compartments per cell in different conditions is shown in Fig 5C. This suggests a hereto unknown role for prenylated proteins in the formation of a mature DV.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Aberrant morphology of intracellular structures is prenylation dependent in delayed-death parasites.

Parasites were treated with indolmycin (50 μM), with and without polyprenol rescue (5 μM GGOH) as indicated. Enriched trophozoite-stage parasites were collected for reduced osmium fixation 72–78 hours post drug administration (equivalent to 28–32 hpi in the second IDC after treatment). (A) Representative images (top-down and cross-sectional), from SEM of each condition: untreated, indolmycin treated (+ indolmycin), and indolmycin treated with polyprenol rescue (+ indolmycin + GGOH). Structures indicated are N, DV, and CI. Scale bar = 3 μm. (B) 3D-rendered iRBCs using serial block-face scanning electron microscopy. Indicated compartments are RBC (white), parasite (blue), CIs (yellow), and DV (red). Scale bar = 3 μm. (C) Parasites were scored for number of DV compartments relative to number of nuclei in each treatment condition. Untreated: 1 nucleus n = 18, 2 nuclei n = 34, ≥ 3 nuclei n = 49; + indolmycin: 1 nucleus n = 16, 2 nuclei n = 29, ≥ 3 nuclei n = 14; + indolmycin + GGOH 1 nucleus n = 17, 2 nuclei n = 19, ≥ 3 nuclei n = 19. Scatter dot blots show error bars with mean ± SD. (D) The total volume (μm³) of the DVs and CIs were measured in each treatment condition. Bar graphs show error bars with mean ± SD. See S2 Data for numerical data underlying figure. CI, cytostomal invagination; DV, digestive vacuole; GGOH, geranylgeraniol; hpi, hours post invasion; IDC, intraerythrocytic developmental cycle; N, nucleus; RBC, red blood cell.

https://doi.org/10.1371/journal.pbio.3000376.g005

As well as confirming the fragmentation of the DV in delayed-death parasites, analysis of the 3D reconstructions dramatically altered our understanding of the membrane-bound enclosures of iRBC cytoplasm in the delayed death parasites. Rather than representing numerous discrete compartments, these were revealed as dramatic extensions of the cytostomal invagination that twisted and wound throughout the cytoplasm of the parasite (Fig 5B, and S3–S5 Movies). As with the DV fragmentation phenotype, this extended invagination phenomenon was rescued by addition of GGOH. We found that the total volume of DVs decreased while the volume of the cytostomal invaginations increased within the indolmycin-treated parasites compared to untreated and GGOH-rescued parasites (Fig 5D). These data support a model whereby the trafficking of RBC cytoplasm (and thus haemoglobin) from cytostomes is perturbed such that the cytostomal invagination itself becomes distended, and the delivery of packages of RBC cytoplasm is severely impaired resulting in a fragmented DV.

Rab GTPase-mediated vesicular trafficking is disrupted during the second IDC following indolmycin treatment

As demonstrated above, delayed death diminishes protein prenylation during the second IDC following treatment (Fig 2). We hypothesised that without prenyl modification of nascent protein, the protein substrates of prenyl-transferases that are dependent on their prenyl modification for localisation or function, such as Rab GTPases, will mislocalise during delayed death. We examined the prenylated Rab GTPase, Rab5a, which has previously been shown to be associated with haemoglobin-containing vesicles and has been hypothesised to play a role in at least one of the three or four semiredundant, but unique haemoglobin uptake pathways present in P. falciparum [29]. More recent experiments in Toxoplasma indicate a role for Rab5a in the vesicular trafficking to the secretory organelles [38]. It remains unclear what role Rab5a has in Plasmodium vesicular trafficking; nevertheless, it is a useful representative vesicular trafficking determinant whose function relies on prenylation [39].

We detected green fluorescent protein (GFP)-fusions of Rab5a in untreated parasites as discrete foci dispersed throughout trophozoite-stage parasites (Fig 6), as previously shown [39]. During the second IDC following clindamycin treatment, GFP-Rab5a mislocalised and, instead of being concentrated in discrete foci, was dispersed in a diffuse cytoplasmic pattern. Prenylation of Rab5a mediates membrane association (possibly alongside cargo-binding), and the apparent cytosolic dispersal of Rab5a in delayed-death parasites is consistent with loss of prenylation-mediated membrane attachment. This effect was less pronounced in late-stage parasites (Fig 6), suggesting that Rab5a’s localisation may be influenced by additional factors in that stage.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Rab5a shows an aberrant localisation in the second IDC after clindamycin treatment.

(A) Representative live-cell images of untreated and clindamycin (5 μM) treated parasites in the second IDC following treatment are shown. Cell populations were divided into parasites with 6 or fewer nuclei (upper panel) and parasite with more than 6 nuclei (lower panel), as the disruption is clearly most pronounced in trophozoite stage parasites. GFP-Rab5a, green signal; DAPI: parasite nuclei, blue signal; merge of green and blue signal. Scale bar = 5 μm. (B) Graph showing the ratio of the brightest GFP focus fluorescence intensity to mean fluorescence intensity distributed in the parasite cytoplasm in untreated and clindamycin treated parasites. N = 38 for untreated, 1–6 nuclei; n = 52 for + clindamycin, 1–6 nuclei; n = 20 for control, > 6 nuclei and n = 16 for + clindamycin, > 6 nuclei. *P < 0.05; ***P < 0.001, Welch t test. Scatter dot plots show error bars with mean ± SD. See S2 Data for numerical data underlying figure. DAPI, 4′,6-diamidino-2-phenylindole; DIC, differential interference contrast; GFP, green fluorescent protein; IDC, intraerythrocytic developmental cycle; iRBC, infected red blood cell.

https://doi.org/10.1371/journal.pbio.3000376.g006

Another prenylated Rab, Rab11a, is localised to the inner membrane complex (IMC) of P. falciparum and is proposed to facilitate vesicle recycling and transport of IMC proteins such as glideosome-associated protein 45 (GAP45) [40]. Consistent with previous observations [41], we observed GAP45 at the IMC in schizont stages of untreated parasites. However, following indolmycin treatment during the second IDC, GAP45 did not localise to structures that resemble IMC (Fig 7), suggesting that there is a defect in the assembly or trafficking of the IMC in delayed death parasites. The normal localisation of GAP45 was restored with GGOH recue treatment (Fig 7). Importantly, secretory trafficking pathways remain intact in the second IDC following indolmycin or clindamycin treatment, as demonstrated by normal export of ring-exported protein 1 (REX1) to the iRBC (S5 Fig). The prenylation-dependent mislocalisation of Rab5a to the cytosol and putative disruption of Rab11a cargo delivery suggests that these trafficking proteins lose their normal association with target biological membranes during delayed death, consistent with the inactive or unbound form of these Rab GTPase proteins.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. IMC formation is disrupted in the second IDC after indolmycin treatment.

Shown are representative images of immunofluorescence assays using anti-GAP45 (1:1,000). Untreated, indolmycin (50 μM) treated and indolmycin treated with polyprenol rescue (5 μM GGOH) as indicated, collected in the second IDC after treatment. The IMC marker GAP45 localises atypically in indolmycin treated parasites. GGOH supplementation restores GAP45 localisation to the IMC equivalent to untreated. GAP45, green signal; DAPI, parasite nuclei, blue signal; merge of green and blue signal. Scale bar = 5 μm. BF, bright field; DAPI, 4′,6-diamidino-2-phenylindole; GGOH, geranylgeraniol; IDC, intraerythrocytic developmental cycle; IMC, inner membrane complex.

https://doi.org/10.1371/journal.pbio.3000376.g007

Indolmycin-induced delayed death increases osmotic fragility of iRBC

Because indolmycin treatment produces severe disruption of the morphology of the DV and haemoglobin trafficking apparatuses, we examined potential physiological effects from disrupted feeding that may contribute to delayed death. Intraerythrocytic parasites progressively internalise and digest host-derived haemoglobin to 1) acquire amino acids [42], 2) increase the available space within the iRBC into which the parasite can grow into [43], and 3) catabolise and expel excess host protein to mitigate the increased colloid-osmotic pressure caused by parasite growth within the iRBC [44,45]. Aberrant haemoglobin uptake induced by indolmycin treatment might kill parasites by disrupting these homeostatic mechanisms.

To determine whether the cellular homeostasis that maintains osmotic pressure is disrupted in indolmycin-treated parasites, we measured osmotic fragility. The propensity for iRBCs to haemolyse in varying hypotonic solutions was measured using a modified method described in Mauritz and colleagues [45] and Dennis and colleagues [46]. Indolmycin-treated parasites showed an increase in osmotic fragility relative to untreated iRBC at all tonicities assayed in the second IDC after treatment (Fig 8A). The half-maximal lytic concentration for indolmycin treated parasites increased significantly by 22.3% (n = 4, SD ± 6.07) (P < 0.05, two-tailed Student t test) (Fig 8B). Like other experiments described above, restoration of prenylation by rescue with GGOH restored measurements of osmotic fragility to those of untreated iRBCs (Fig 8A and 8B). This supports the suggestion that disruption of protein prenylation by indolmycin treatment reduces the capacity of treated parasites to internalise and degrade host-derived haemoglobin and that this has a profound effect on the physiology of the parasite and its host cell, likely contributing to growth arrest and delayed death.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Indolmycin-treated parasites have increased osmotic fragility in the second IDC after treatment.

(A) Parasites were treated with indolmycin (50 μM) and divided into two conditions, with and without polyprenol rescue (5 μM GGOH). Enriched trophozoite-stage parasites were collected 72–78 hours post drug administration (equivalent to 28–32 hpi in the second IDC after treatment) and incubated in solutions with tonicity varied by increasing the concentration of NaCl. Percentage iRBC lysis was calculated by measuring the absorbance at 415 nm (A₄₁₅) of released haem normalised to a relative tonicity of one (isotonic). Representative of four independent experiments. Scatter plots show error bars with mean ± SEM. See S2 Data for numerical data underlying figure. (B) The concentration required for Lysis C_50% for each of the three conditions: untreated, indolmycin treated (+ indolmycin), and polyprenol rescue (+ indolmycin + GGOH). The Lysis C_50% for indolmycin treated parasites increased by 22.3% (n = 4, SD ± 6.07) whereas the Lysis C_50% for the polyprenol rescued parasites did not significantly change. *P < 0.05; **P < 0.01; two-tailed Student t test. Box and whisker plots show the interquartile range. See S2 Data for numerical data underlying figure. GGOH, geranylgeraniol; hpi, hours post invasion; IDC, intraerythrocytic developmental cycle; Lysis C_50%, half-maximal lysis; ns, not significant; RBC, red blood cell.

https://doi.org/10.1371/journal.pbio.3000376.g008

Uncoupling ubiquinone and dolichol contributions during third IDC arrest

Restoring protein prenylation in delayed death using GGOH rescue treatment demonstrates that the proximate cause of delayed death is a defect in protein prenylation. However, even when rescued by GGOH, these polyprenol-rescued parasites are still unable to survive a third IDC following treatment (Fig 2 and S3 Fig) suggesting that additional causes of parasite death arise from disruption of another isoprenoid pathway. Ubiquinone (or Co-enzyme Q) is the principle and essential electron acceptor for electron transport in mitochondrion. Ubiquinone inserts into the inner membrane of the mitochondrion by its isoprene tail. In blood-stage P. falciparum, the mitochondrion’s primary function is to provide electrons for use by dihydroorotate dehydrogenase, an essential metabolic enzyme required for pyrimidine biosynthesis [47]. To determine whether disruption of ubiquinone contributes to third IDC arrest in polyprenol-rescued parasites, we used a transgenic parasite expressing Saccharomyces cerevisiae (yeast) dihydroorotate dehydrogenase (yDHODH). The yDHODH enzyme uses fumarate as an electron acceptor and is independent of ubiquinone for pyrimidine biosynthesis and hence uncouples blood-stage P. falciparum’s dependence on mitochondrial electron transport [47]. Although these transgenic yDHODH parasites are slow growing, in asexual blood stages, they are resistant to inhibition that results in mitochondrial perturbations.

Parasites expressing yDHODH were not innately protected from second IDC arrest after indolmycin treatment (Fig 9). This suggests that isoprenoid disruption of parasite DHODH is not the cause of delayed death and indeed causes no further insult in the second IDC after indolmycin treatment even after the initial insult of disruption of protein prenylation. Furthermore, indolmycin-treated yDHODH parasites supplemented with GGOH were not protected in the third IDC after treatment. This result indicates that arrest in the third IDC after treatment cannot be attributed either to disruption of parasite DHODH or protein prenylation. The remaining known fate for isoprenoids in Plasmodium is in the synthesis of dolichols, so P. falciparum survival during the third IDC following indolmycin treatment is likely co-contingent with dolichol biosynthesis from isoprenoid precursors.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 9. Supplementing ubiquinone-independent yDHODH-expressing transgenic parasites with GGOH does not protect from delayed death in the third IDC after treatment.

Dose-response curve from SYBR-Green susceptibility assay determined 48, 96, and 144 hours post indolmycin treatment, with and without polyprenol (5 μM GGOH) rescue. Assays were performed in parallel with 3D7 and transgenic parasites expressing ubiquinone-independent yDHODH. yDHODH parasites are susceptible to delayed death equivalent to 3D7 (inhibition at 96 hours). GGOH supplementation does not further protect yDHODH parasites in the third IDC (inhibition at 144 hours). Representative of at least three independent experiments. Scatter plots show error bars with mean ± SEM. See S2 Data for numerical data underlying figure. GGOH, geranylgeraniol; IDC, intraerythrocytic developmental cycle; yDHODH, yeast dihydroorotate dehydrogenase.

https://doi.org/10.1371/journal.pbio.3000376.g009

P. falciparum culture

P. falciparum 3D7 parasites were maintained in continuous culture with minor modifications to the method described by [56]. Briefly, parasites were cultured in human O+ RBCs (leukocyte-depleted by filtration, provided by the Australian Red Cross Blood Service) at 2% haematocrit in RMPI-1640 supplemented with 25 mM sodium bicarbonate, 25 mM HEPES, 150 μM hypoxanthine, 20 μg/mL gentamicin, and 0.5% (w/v) Albumax II (Invitrogen) (complete medium). Cultures were maintained in flasks and petri dishes inside a sealed container filled with a low oxygen malaria-mix gas (1% O₂, 5% CO₂, and 94% N₂) at 37°C with or without mechanical shaking.

Synchronised ring-stage cultures were obtained by double (14-hr interval) treatment with 5% (w/v) D-sorbitol during the parasite’s IDC prior to experimentation, and a single sorbitol treatment of ring-stage parasites immediately prior to experimentation. Alternatively, 0–4 hpi ring-stage parasites were generated by first enriching for schizonts using a custom magnetic separation apparatus described previously [57], incubating the enriched parasites with fresh RBCs for 4 hrs, and then treating the newly invaded RBCs with 5% (w/v) D-sorbitol.

LC-MS metabolic profiling

Metabolite profiling across the development of delayed death was initiated by treating 0–4-hr ring-stage cultures (prepared as described above) with 50 μM indolmycin (BioAustralis Fine Chemicals Product 21200-24-8) or dimethyl sulfoxide vehicle control for 30 hrs and then the culture medium was replaced. Samples were collected at 30, 58, 68, 78, and 88 hrs post drug administration (equivalent to 30 hpi in the first IDC and then 14, 24, 34, and 44 hpi in the second IDC) for metabolite extraction and analysis by LC-MS. Samples with equivalent parasitemia were collected and 1×10⁸ RBCs per conditions were centrifuged at 14,000 g for 30 seconds (4°C), washed with 1 mL of ice-cold phosphate buffered saline (pH 7.4), centrifuged again and the cell pellet extracted with 200 μL of 80% (v/v) acetonitrile (in water containing 5 μM ¹³C-aspartate as the internal standard). The cell extracts were rapidly resuspended, vortexed and then centrifuged at 14,000 g for 10 min at 4°C. The metabolite extract was transferred to an MS vial and LC-MS analysis was preformed using methods previously described [58], with an Agilent Q-TOF mass spectrometer 6550 operating in negative ESI mode. In an effort to detect dolichol we also purchased a dolichol standard (Avanti) but were unable to detect this by LC-MS. Metabolomics data have been deposited at Metabolomics workbench [59], study ST001188 (http://dx.doi.org/10.21228/M88M3Q).

Immunoblot analysis

Synchronised ring-stage parasites (prepared by sorbitol treatment described as per P. falciparum culture) were treated with 50 μM indolmycin (BioAustralis Fine Chemicals) or dimethyl sulfoxide vehicle control for 30 hrs. The medium was replaced 30 hrs post drug administration, supplementing with the polyprenol rescue compound GGOH (5 μM) (Sigma Aldrich, G3278) as indicated. Trophozoites (5–10% parasitaemia) were collected at approximately 72–78 hrs post drug administration by lysis with 0.05% (w/v) saponin. Cells were then pelleted by centrifugation, washed with PBS, and then resuspended in Laemmli Sample Buffer (Bio-RAD) with 2.5% (v/v) 2-Mercaptoethanol and complete EDTA-free Protease Inhibitor (Roche). Samples were then either snap-frozen in liquid nitrogen and stored at -80°C or loaded directly onto Mini-PROTEAN™ TGX Precast Gels (Bio-RAD) in Tris-glycine buffer. Following electrophoresis at 180 Volts for 35 min, proteins were transferred to nitrocellulose membranes using the iBlot 2 western transfer system (Thermo Fisher Scientific).

PVDF membranes with bound protein were then blocked in 5% (w/v) skim milk in PBS, incubated with rabbit anti-farnesyl polyclonal antibody at 1:1,000 dilution (Life Technologies, PA1-12554) or mouse anti-BiP at 1:1,000 [60]. Following multiple washes with PBS-tween, horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG (PerkinElmer) or HRP-conjugated rabbit anti-mouse IgG (Promega) at 1:10,000 dilution was applied to membranes. Enhanced chemiluminescent detection was performed with SuperSignal® West Pico/Femto Sensitivity Substrate (Thermo Fisher Scientific) and analysed using the ChemiDoc™ Imaging System (Bio-RAD).

Fluorescein dextran uptake assay

Using a modified version of the methodology described in Frankland and colleagues [35] and Baker and colleagues [36], packed RBCs were lysed with ice-cold lysis buffer (5 mM Na phosphate, 1 mM ATP; pH 7.5) containing 50 μM F-dextran (Life Technologies, D1820), and then incubated for 10 min at 4°C. The cells were then re-sealed by adding NaCl to 150 mM and incubating them for 45 min at 37°C. The re-sealed RBCs, loaded with F-dextran, were then washed in complete media (see above, P. falciparum culture) and stored at 4°C for up to one week. Internalisation assays were commenced by inoculating the preloaded RBCs with enriched schizonts (prepared as per P. falciparum culture), treating the newly invade ring-stage parasites with 50 μM indolmycin (BioAustralis Fine Chemicals) or dimethyl sulfoxide vehicle control for 30 hrs. The medium was replaced 30 hrs post drug administration, supplementing with the polyprenol rescue compound GGOH (5 μM) (Sigma Aldrich) as indicated. Trophozoites were collected at approximately 72–78 hrs post drug administration for live-cell microscopy. The iRBCs were washed for 5 min in 1 μg/mL Hoescht 33342 (Life Technologies) and mounted to directly to slides. Microscopy was performed using the DeltaVision Elite™ Widefield deconvolution (GE Healthcare) imaging platform. ImageJ (v1.51n) [61] was used to merge multiple microscopy channels. GraphPad Prism (V. 7.01) was used to plot data.

Fluorescent microscopy

Fluorescent imaging of cells was performed either by live microscopy or indirect immunofluorescence assay. For live cell microscopy: iRBCs were collected, stained with 1 μg/mL Hoechst 33342 (Life Technologies), and mounted directly to glass slides. For indirect immunofluorescence assay: glass coverslips were pre-coated with 0.1 mg/mL pHAE (erythroagglutinating phytohemagglutinin) (Sigma Aldrich), and iRBCs were applied at 5% haematocrit in PBS to form a monolayer. The cells were then fixed on the coverslip with 2% (w/v) paraformaldehyde and 0.006% (w/v) glutaraldehyde for 20 min, permeabilised with 0.1% (v/v) Triton X-100 for 10 min, and then incubated with primary antibodies: 1:300 rabbit anti-GAP45 [41], 1:1,000 rabbit anti-REX1 [62] in 3% (w/v) BSA for 2 hrs. Following washing, Alexa Fluor 488 and 594-conjugated (Thermo Fisher scientific) anti-rabbit or anti-mouse secondary antibodies in 3% (w/v) BSA were applied to cells for 1 hr, washed with 300 nM DAPI (Sigma Aldrich), and mounted to slides with ProLong® Gold Antifade (Thermo Fisher scientific). Microscopy was performed using the DeltaVision Elite™ Widefield deconvolution (GE Healthcare) imaging platform. ImageJ (v1.51n) [61] was used to merge multiple microscopy channels.

GFP-Rab5a localisation

Endogenously tagged GFP-2xFKBP-Rab5a [39] parasites were maintained in O+ RBCs as described [56] under selection with 4 nM WR99210 (Jacobus Pharmaceuticals). Parasite cultures were synchronized to 0–5 hpi and parental cultures were split into two dishes. Cells were exposed to 5 μM clindamycin (Ratiopharm) or vehicle (dH₂O) for 35 hrs. Parasites were washed three times in complete RPMI medium (see above, P. falciparum culture), incubated for a further 51 hrs until 86–91 hpi and subsequently prepared for imaging.

Imaging and analysis of GFP-Rab5a cell line

All live cell images of the GFP-Rab5a cell line haven been taken with a Carl Zeiss Axio Imager A1 and imaging was performed as described [63]. A 100×/1.4–numerical aperture lens combined with a Hamamatsu Orca C4742-95 camera was used. Parasite cultures were stained with 1 μg/mL DAPI for 15 min before imaging. Exposure time was identical for all acquired images on the GFP channel. Parasites to be analysed were selected from DIC images. The corresponding GFP image were loaded into ImageJ (v1.51n) [61]. Setting to be recorded were max grey value and mean grey value. The background fluorescence of the images as well as the maximum fluorescent intensity of the brightest focus within the parasite and the background fluorescence of each parasite (excluding GFP foci) were measured. Image background fluorescence intensity was subtracted from the latter two values and the ratio was calculated. For plotting of data GraphPad Prism (V. 7.01) was used. Representative images were processed with Corel Photo-Paint x6 by adjusting brightness and intensity (V. 16.4.1.1281).

3D block-face scanning electron microscopy

Synchronised 0–4 hpi ring-stage parasites (prepared as per P. falciparum culture) were treated with 50 μM indolmycin (BioAustralis Fine Chemicals) or dimethyl sulfoxide vehicle control for 30 hrs. Media was replaced 30 hrs post drug administration, and where indicated, cultures were supplemented with the polyprenol rescue compound GGOH (5 μM) (Sigma Aldrich). Trophozoites (5–10% parasitaemia) were collected at approximately 72–78 hrs post drug administration (equivalent to 28–34 hpi in the second IDC) by magnetic separation (as described above, P. falciparum culture) and fixed with 2.5% (w/v) glutaraldehyde in 0.1 M sodium cacodylate for at least 2 hrs at 4°C. Cells were washed in 0.175 M sodium cacodylate and then incubated en bloc in 1% (w/v) low-melt agarose with 1.5% (w/v) potassium ferrocyanide and 2% (w/v) osmium tetroxide in 0.15 M sodium cacodylate for 1 hr at 4°C. Cells were washed in Milli-Q® water and then incubated with 1% (w/v) thiocarbohydrazide for 20 min. Following further washes, cells were incubated with 2% (w/v) osmium textroxide for 30 min, 2% (w/v) uranyl acetate overnight at 4°C, and then Walton’s lead aspartate for 30 min at 60°C. Supernatant was then removed, samples were washed thoroughly with Milli-Q water before gradually dehydrated in ethanol, then acetone, and finally infiltrated with Procure 812 resin (ProSciTech). Samples were then polymerised by curing at 60°C for at least 24 hrs. Each polymerised block-face was trimmed to 1 mm³, and then serially sectioned (50 nm) and imaged using the Teneo Volume Scope (FEI) at 3kV. Image processing was performed using the IMOD software package (V. 4.9) [64]. For plotting of data GraphPad Prism (V. 7.01) was used.

Dose-response assay

Synchronised ring-stage 3D7 parasites (prepared by sorbitol treatment described as per P. falciparum culture), or yDHODH parasites [65] maintained under selection with 5 nM WR99210 (Jacobus Pharmaceuticals), were set up in triplicate within V-bottom 96-well plates (1% haematocrit, 1% parasitaemia: 48 hrs; 0.1% parasitaemia: 96 hrs; 0.01% parasitaemia: 144 hrs) and treated with varying concentrations of indolmycin (BioAustralis Fine Chemicals), clindamycin hydrochloride (Sigma Aldrich), and chloramphenicol (Sigma Aldrich), prepared by serial dilution in complete medium (See above, P. falciparum culture). Equivalent iRBC wells were set up in complete medium only and positive-KILL controls were initiated in triplicate using these additional wells by adding 200 nM dihydroartemisinin (Sigma Aldrich) to ring-stage parasites 48 hrs prior to assaying growth inhibition. Complete medium was replaced after each IDC, and rescue compounds GGOH (Sigma Aldrich, G3278), FOH (Sigma Aldrich, F203), and IPP were added at varying concentration in complete medium for the cycles indicated. Growth inhibitions were analysed at 48, 96, and 144 hrs as indicated using SYBR Green assay as previously described [66], briefly RBC pellets were lysed and incubated with SYBR Green I (Thermo Fisher Scientific) for 1 hr and analysed using FLUOstar Omega plate reader (BMG Labtech). Percentage survival was determined by normalising the SYBR green signal of the treated conditions to the untreated controls, and then subtracting the background signal derived from the positive-KILL controls. EC50s given here are the concentration of compound required to inhibit 50% of this normalised SYBR green signal. Corrected values were plotted by XY scatter using GraphPad Prism (V. 7.01) and EC₅₀ values were attained from the dose-response curves.

Osmotic fragility assay

The osmotic fragility of indolmycin treated iRBCs were analysed using a protocol modified from Lew, Tiffert and Ginsburg [44], and Dennis and colleagues [46]. Synchronised 0–4 hpi ring-stage parasites (prepared as per P. falciparum culture) were treated with 50 μM indolmycin (BioAustralis Fine Chemicals) or dimethyl sulfoxide vehicle control for 30 hrs. Media was replaced 30 hrs post drug administration, and where indicated, cultures were supplemented with the polyprenol rescue compound GGOH (5 μM) (Sigma Aldrich). Trophozoites (2% haematocrit, 5–10% parasitaemia) were enriched at approximately 72–78 hrs post drug administration (equivalent to 28–34 hpi in the second IDC) by magnetic separation (as described above, P. falciparum culture). iRBC were resuspended in 400 μL of solution A (150 mM NaCl, 2 mM HEPES-Na, pH 7.4, relative tonicity (RT) = 1:300 mOsm). In a V-bottom 96-well plate, 200 μL of varying saline-tonic solutions were prepared in triplicate by diluting Solution A with Solution B (2 mM HEPES-Na, pH 7.4, RT = 0.04; 12 mOsm) to produce a range of saline tonic-solutions with RTs of 0.04 to 1. 10 μL of enriched iRBCs were then aliquoted into each tonic solution and incubated at room temperature for 10 min. Plates were then centrifuged at 1,200 g for 10 min and 175 μL of supernatants were transferred to a flat-bottom 96-well plate for analysis as described in Dennis and colleagues [46] using a FLUOstar Omega plate reader (BMG Labtech). GraphPad Prism (V. 7.01) was used to plot data.