Phytohormones, Isoprenoids, and Role of the Apicoplast in Recovery from Dihydroartemisinin-Induced Dormancy of Plasmodium falciparum

ABSTRACT

Many organisms undergo dormancy as a stress response to survive under unfavorable conditions that might impede development. This is observed in seeds and buds of plants and has been proposed as a mechanism of drug evasion and resistance formation in Plasmodium falciparum. We explored the effects of the phytohormones abscisic acid (ABA) and gibberellic acid (GA) on dihydroartemisinin (DHA)-induced dormant erythrocytic stages of P. falciparum parasites. Dormant ring stages exposed to ABA and GA recovered from dormancy up to 48 h earlier than parasites exposed to DHA alone. Conversely, fluridone, an herbicide inhibitor of ABA synthesis, blocked emergence from dormancy. Additionally, the role of the apicoplast was assessed in dormant parasite recovery. Apicoplast-deficient P. falciparum remained viable for up to 8 days without the organelle and recrudesced only when supplemented with isopentyl pyrophosphate (IPP). IPP was not required for survival in the dormant state. Fosmidomycin inhibition of isoprenoid biosynthesis did not prevent dormancy release from occurring in parasites with an intact apicoplast, but IPP or geranylgeranyl pyrophosphate was needed for complete recrudescence. In addition, the apicoplast and specifically the isoprenoids it produces are essential for recovery of dormant parasites. In summary, ABA and GA have significant effects on dormant parasites, and the phenotypes produced by these phytohormones and the herbicide fluridone also provide a means to explore the mechanism(s) underlying dormancy and the regulatory network that promotes cell cycle arrest in P. falciparum.

INTRODUCTION

About 216 million cases of malaria and 445,000 deaths were reported in 2016 (1). Plasmodium falciparum is responsible for the most deadly form of the disease, contributing to most morbidity and mortality. Artemisinin-based combination therapy (ACT) remains the World Health Organization's recommended antimalarial treatment due to the potent and fast activity of artemisinin (ART) and its derivatives to clear parasites. Pairing these drugs with slower-acting antimalarials increases the effectiveness of treatment while reducing the risk of recrudescence and resistance that often follows monotherapy. Despite the global use of ACTs to prevent drug resistance, ART resistance is developing as an emerging threat and has been a target of interest especially in the Southeast Asia region of Thailand and Cambodia, where resistance is spreading (2).

Dormancy, characterized by a stress-induced growth arrest in the asexual ring stages of P. falciparum, has been proposed as a mechanism for drug evasion and subsequent recrudescence following removal of drug pressure (3). Upon exposure to an artemisinin drug, early development stages known as rings are arrested as dormant forms and take on morphology characterized by condensed nuclei and reduced cytoplasm. Artemisinin drugs have been demonstrated to induce this phenotype in vitro (4, 5); removal of the drug leads to recrudescence as parasitemia increases to detectable levels (6, 7). Dormant parasites are not completely quiescent, as low levels of metabolic activity are maintained, especially in the case of fatty acid synthesis and pyruvate metabolism (6, 7). In vivo evidence of dormant parasites is supported by observations of artesunate treatment of Plasmodium vinckei petteri and Plasmodium vinckei vinckei in a rodent model (8). Growth was arrested and microscopy revealed forms morphologically similar to artemisinin-induced dormant P. falciparum ring-stage parasites. More recently, the antifolate pyrimethamine has been shown to induce a second cycle of dormancy in daughter rings following exposure in the previous life cycle (9).

Growth arrest is not exclusive to Plasmodium, as related phenomena are observed in plants where dormancy has been extensively studied (10). Dormancy can be triggered by abiotic stress (drought, salinity, or temperature) and results in an early developmental arrest that confers protection from unfavorable conditions that might otherwise disrupt seed or bud development and kill the young plant. These processes are regulated by signaling phytohormones, including abscisic acid (ABA) and gibberellic acid (GA), that promote induction or release from dormancy (11, 12). In addition to evidence showing that phytohormones have an essential role in plants, data suggest broader effects in other systems, including mammalian cells and apicomplexan parasites (13, 14). For example, ABA has been shown to block egress of Toxoplasma gondii from host cells and to affect malaria transmission by manipulating the immune system of the mosquito host (13, 15).

Apicomplexan organisms share an ancestral origin with plants in the form of the apicoplast, a nonphotosynthetic organelle evolutionarily obtained following a secondary endosymbiotic relationship between a eukaryote and red algae. The apicoplast has four membranes and shared metabolic pathways (16, 17). The sole essential function of the apicoplast in the erythrocytic stages of malaria is to synthesize isoprenoids (18) via the nonmevalonate pathway using pyruvate and glyceraldehyde-3-phosphate as starting materials and has been conserved across plastids (19).

Dormancy, or cellular quiescence, is a common stress and survival mechanism and could have a regulatory mechanism(s) conserved and shared between Plasmodium and plants. In this study, we explored dormancy-regulating phytohormones and their effects on drug-induced dormant P. falciparum in vitro and observed an early recrudescence phenotype that could be blocked by an herbicide inhibitor. We also show the importance of isoprenoids in the recovery from dormancy with and without an apicoplast. Together, these findings help shed light on novel, previously unknown effects of phytohormones on parasites and a possible mechanism of cell cycle and growth regulation in P. falciparum.

RESULTS

Abscisic acid and gibberellic acid stimulate early recrudescence in dormant parasites.

Plant dormancy is a response to external stress during an early developmental stage and is regulated by endogenous phytohormones. Given the linkage of the apicoplast in malaria parasites and known effects of abscisic acid (ABA) on apicomplexan parasites (13, 15), we hypothesized that the phytohormones could have similar effects on P. falciparum cell cycle arrest and dormancy. In previous studies, dihydroartemisinin (DHA) exposure to rings induced cell cycle arrest, with subsequent recovery and normal growth (4, 6). Therefore, first we assessed the effect of exogenous supplementation of the phytohormone ABA on DHA-induced dormant ring stages of P. falciparum (3D7). Initially, ABA (100 μM) was used as a supplement for DHA-induced dormant parasites and maintained in culture media to observe the time for parasites to recover and grow. Daily thick and thin blood smears were collected and examined microscopically for parasites exhibiting normal morphology (rings, trophozoites, and schizonts) compared to dead and dormant forms. The ratio of normal to total (normal plus dead/dormant forms) was used to monitor recovery. Following DHA exposure, all ring-stage P. falciparum (3D7) organisms had become dormant or dead within 24 h and maintained this phenotype until parasites recovered and resumed normal growth on day 4 post-DHA exposure. Interestingly, normal parasite morphology was first observed by day 3 in cultures exposed to ABA, which was ∼24 h prior to the DHA-treated controls (Fig. 1A). The initial growth rates after recovery were similar for the two groups (with or without ABA). The same phenotype of early recovery from DHA-induced dormancy was observed in the chloroquine-resistant clone W2 (Fig. 1B).

FIG 1.

Effect of ABA on DHA-induced dormant P. falciparum. (A) Synchronized 3D7 parasites were exposed to DHA for 6 h to induce dormancy. Following 3 washings with RPMI medium, 100 μM ABA was added and maintained in culture. Recovery was monitored in Giemsa-stained blood smears, and the ratio of normal to total (dormant/dead plus normal) parasites was recorded. ABA-supplemented parasites recovered earlier than parasites that were exposed only to DHA. (B) A similar pattern can be observed with W2. When 50 μM FLD is used as a supplement and removed on day 2, parasites begin to emerge from dormancy 2 days later. (C) Micrograph images reveal normal parasites in smears of ABA-supplemented parasites 24 h earlier than DHA-only-treated parasites and 48 h later in FLD-supplemented parasites. Error bars represent SEMs; experiments were conducted more than 2 times with multiple parasite clones.

We next explored the effects of gibberellic acid (GA), a phytohormone that promotes germination of the seed in plants and acts to release them from dormancy. In plants GA acts antagonistically with ABA, yet when GA was used as a supplement for DHA-induced dormant rings, we observed a similar response of early recovery from dormancy (Fig. 2A). GA-treated parasites showed initial signs of normal morphology 24 h before ABA and 48 h earlier than DHA alone. In addition, GA (10 μM) was much more potent than ABA (100 μM) at stimulating early recovery from dormancy (Fig. 2A).

FIG 2.

Effects of gibberellic acid (GA) and fluridone (FLD) on DHA-induced dormant P. falciparum (3D7). (A) Dormant parasites supplemented with 10 μM GA exhibited signs of recovery earlier than when supplemented with ABA. (B) FLD was added to dormant parasites and removed at different intervals. Dormancy was maintained in the presence of FLD, but this was reversed upon its removal during the first 8 days. Parasites exposed to FLD longer than 8 days failed to recover. Error bars represent SEMs. D2, day 2.

FLD causes a delay in recrudescence of dormant parasites.

Fluridone (FLD) is an herbicide that inhibits phytoene desaturase, carotenoid, and ABA biosynthesis (20). When DHA-induced dormant P. falciparum rings were exposed to FLD, we observed a delay in recovery from dormancy (Fig. 1B and C). FLD-exposed dormant rings did not recover and grow until 2 days later than cultures exposed to DHA alone. We next assessed the effect of FLD on dormancy by prolonged exposures of dormant rings to the herbicide. First dormancy was induced with DHA, and then FLD (50 μM) was added for 2, 8, 10, and 12 days. FLD prevented recovery under all conditions, as no dormant parasites recrudesced when FLD (50 μM) was in the media. Once FLD was removed from 2 to 8 days post-dormancy induction, parasites began to recrudesce 2 to 4 days after FLD removal; however, dormant rings did not recover following ≥8 days of exposure to FLD (Fig. 2B). These results suggest that the phytohormones ABA and GA stimulate early recovery from dormancy, whereas the herbicide inhibitor FLD blocks recovery from DHA-induced dormancy.

ABA and GA induce the early recrudescence phenotype in artemisinin-resistant P. falciparum.

We next determined if ABA, GA, and FLD have similar effects on DHA-induced dormancy of artemisinin-resistant P. falciparum. Multiple point mutations in the propeller domains of the Kelch gene on chromosome 13 (K13; PF3D7_1343700) are associated with a delayed parasite clearance phenotype and serve as a molecular marker for the spread of artemisinin resistance (21). ARC08-88 (11G) is a clone with the C580Y K13 mutation, which expresses multiple artemisinin resistance phenotypes (22). Since the 11G clone is more resistant than artemisinin-susceptible clones (Table 1), 11G parasites were exposed to two rounds of 700 nM DHA (at time zero and 24 h) for a total of 48 h of exposure to induce dormancy. ABA, GA, and FLD were then added, and FLD was washed out after 48 h. As expected, DHA-induced dormancy arrested growth of 11G by day 3 under all conditions. The first signs of parasite recovery from dormancy (when normal morphology was detectable in culture) were on day 8 in the GA-treated group and on day 9 in the ABA-treated group (Fig. 3). Parasites exposed only to DHA showed signs of recovery on day 10. FLD-treated parasites once again exhibited a delay, with recovery beginning on day 14. The findings of this study suggest that the effects of the phytohormones and FLD on DHA-induced dormancy are independent of artemisinin resistance.

TABLE 1.

IC₅₀s of artemisinin derivatives in this study

Strain	IC50 (mean ± SD), ng/ml
	Artelinic acid	Artesunate	Dihydroartemisinin	Artemisinin
Arc 08-88 (11G)	86.5 ± 48.5	7.0 ± 3.4	2.0 ± 0.38	17.6 ± 4.5
W2	20.4 ± 13.2	3.9 ± 3.5	1.3 ± 0.53	6.9 ± 1.16

FIG 3.

ABA, GA, and FLD induce similar recovery phenotypes with artemisinin-resistant P. falciparum. An artemisinin-resistant clone from Cambodia, ARC08-88 (11G), was used to assess the effects of ABA, GA, and FLD on recovery from dormancy. DHA (700 nM) was added at 0 and 24 h, and parasites were under drug pressure for a total of 48 h. After washing, ABA, GA, and FLD were added, with FLD being washed out after 48 additional hours. Dormant parasites began to exit on day 10, but when exposed to GA and ABA, parasites recovered earlier on days 8 and 9, respectively. A delay in recovery with FLD exposure was observed as normal parasites emerged by day 14. Error bars represent SEMs; the experiment was conducted 2 times.

The apicoplast is required for recrudescence from dormancy.

Next we explored the apicoplast's role in the recrudescence of dormant parasites to determine if the apicoplast was required for recovery. Isoprenoid biosynthesis in the asexual blood stages of Plasmodium is reported as the apicoplast's only essential function (18). Exposure to antibiotics such as doxycycline (DOX) results in parasites that are unable to replicate an apicoplast in the subsequent asexual life cycle and experience a delayed death effect (23). All apicoplast-related metabolism is disrupted, including production of essential isoprenoids. Fosmidomycin (FOS) inhibits isoprenoid biosynthesis and results in death in the first life cycle. In either case, parasites can be rescued with 200 μM isopentyl pyrophosphate trisammonium salt (IPP) supplementation (18). Previous work has shown that 20 μM geranylgeraniol, the alcohol analog of geranylgeranyl pyrophosphate trisammonium salt (GGPP), rescues at 10 μM FOS (18). In this study, we found that exposing P. falciparum to GGPP (0.2 to 2.0 μM) also rescued parasites exposed up to 10 μM FOS (Fig. 4A). We observed toxicity to parasites exposed to higher concentrations of GGPP (≥20 μM), and concentrations of <0.2 μM were ineffective in the FOS rescue studies. Growth was sustained in the presence of GGPP (2.0 μM), whereas parasites died when it was removed (Fig. 4B and C).

FIG 4.

GGPP rescue in FOS-treated P. falciparum (3D7). (A) Rescue was explored in parasites exposed to a range of FOS and GGPP concentrations and compared to growth of an untreated control (luminescence in relative light units [RLU]). Ring-stage forms matching those of the control (CTRL) were observed at 96 h when 10 μM FOS-treated parasites were exposed to 0.2 and 2.0 μM GGPP. (B) Parasites exposed to FOS failed to produce merozoites and did not complete the life cycle. Conversely, GGPP (2 μM) rescued parasites exposed to 5 μM FOS. (C) When GGPP was removed on day 5, FOS inhibition resulted in death of the parasites. Error bars represent SEMs; the experiment was conducted 2 times.

We next explored the ability of GGPP to rescue DOX-treated parasites. D10-ACP_LEADER-GFP was used in the experiments because it has a green fluorescent protein (GFP)-tagged acyl carrier protein that localizes to the apicoplast (24). Thus, the apicoplast could be visualized in dormant and recovering parasites for confirmation of its presence or absence. Pretreatment with DOX for 48 h was used to generate ring-stage parasites without the organelle, and these parasites were then exposed to DHA to induce dormancy. ABA, GA, IPP, or GGPP was then used to supplement the cultures. An additional experimental group with IPP added 4 days after inducing dormancy was used to assess the effect that delaying exposure to the essential isoprenoid precursor would have on growth. DOX-treated P. falciparum without an apicoplast recovered only when exposed to IPP, while ABA- and GA-treated parasites remained dormant (Fig. 5A). For parasites exposed to IPP on day 4, normal morphology became visible by day 7 (Fig. 5B). Fluorescence imaging confirmed that the DOX-treated parasites lacked an intact apicoplast yet were still able to recover in the presence of IPP (Fig. 6). MitoTracker staining of DOX-treated P. falciparum revealed weak signals in dormant parasites that lacked distinct morphology observed in actively growing parasites. GGPP also failed to rescue recovery in DOX-treated dormant parasites (see Fig. S2 in the supplemental material).

FIG 5.

Apicoplast-deficient dormant P. falciparum (3D7) organisms recover only when supplemented with IPP. (A) Parasites were pretreated with DOX 48 h prior to DHA exposure to generate daughter parasites that lacked an apicoplast. DOX exposure was maintained in culture, and IPP, GA, and ABA were added for the duration of the experiment. IPP was also added 4 days later for dormant parasites exposed only to DOX. IPP rescued parasite recovery, whereas ABA and GA did not. Supplementation with IPP 4 days later with subsequent recovery demonstrated that the DOX-exposed dormant parasites were still viable. (B) Micrograph images reveal dormant forms in groups treated with DHA plus DOX, with signs of recovery only in groups supplemented with IPP. Bars in panel A represent SEMs; the experiment was conducted ≥2 times.

FIG 6.

Recrudescence of apicoplast-deficient dormant P. falciparum supplemented with IPP on day 4 (D4). Morphology showing recrudescence was observed in fluorescence images of D10-ACP_LEADER-GFP on day 4 (A) and day 8 (B). The lack of localization of the GFP signal is due to the absence of an apicoplast. The apicoplast-deficient parasites were still able to exit dormancy and proliferate due to the presence of IPP; GA and ABA did not rescue growth in apicoplast-deficient P. falciparum.

These results prompted us to explore if the addition of IPP is time dependent. DOX-treated parasites were exposed to IPP on days 0, 4, 8, and 16 after induction of dormancy with DHA. P. falciparum organisms supplemented with IPP from days 0 to 8 all recovered from dormancy (Fig. 7A). The length of time for normal parasite morphology to be visible was extended up 10 days when IPP was added on day 8 compared to when IPP was added on day 4. GFP signal was lost in some parasites later in the experiment; this could be due to not culturing with the selectable marker pyrimethamine for the duration of the experiment (Fig. 7B). When IPP was added on day 16, recovery was not observed. By day 30, all dormant/dead forms had been cleared from culture. These results suggest that viable dormant parasites survive at least 8 days without any supplementation with IPP.

FIG 7.

Dormant P. falciparum (D10-ACP_LEADER-GFP) survive and recrudesce up to 8 days when IPP supplementation is delayed. (A) DOX was maintained in dormant cultures and IPP was added either at the start or 4, 8, or 16 days after dormancy induction. Parasites supplemented with IPP from days 0 to 8 recovered from dormancy, whereas adding IPP on day 16 failed to rescue parasite growth. (B) Fluorescence images of parasites with IPP added on day 8 reveal growing parasites. Error bars in panel A represent SEMs.

We also investigated dormancy in parasites that had isoprenoid biosynthesis inhibited but maintained an intact apicoplast. Initially 5 μM FOS, which inhibits 1-deoxy-d-xylulose 5-phosphate (DOXP) reductase of the isoprenoid biosynthetic pathway, was added at the start and maintained in media for the duration of the experiment. DHA was used to induce dormancy as previously described. These parasites also were supplemented with either 200 μM IPP or 2 μM GGPP. Interestingly, parasites exited dormancy under all conditions but fully recrudesced only in the presence of IPP or GGPP (Fig. 8A). Parasites exposed to FOS alone failed to progress past the initial asexual life cycle and were killed (Fig. 8B). These results suggest that FOS does not inhibit the viability of DHA-induced dormant rings but acts only once normal growth resumes.

FIG 8.

Inhibiting isoprenoid biosynthesis does not prevent parasites from exiting dormancy. Ring-stage parasites were exposed to DHA (to induce dormancy) and FOS to inhibit all isoprenoid biosynthesis while maintaining an apicoplast. Then 200 μM IPP or 2 μM GGPP was added, and parasites exited dormancy under all conditions; however, complete recrudescence and resumption of normal growth were observed only in FOS-treated parasites supplemented with IPP or GGPP. Without either isoprenoid precursor, FOS inhibition killed the parasites after they began to recover. Error bars in panel A represent SEMs; the experiment was conducted twice.

DISCUSSION

Dormancy is a common survival mechanism whereby cells can cease to proliferate and survive stress and other environmental conditions before recovering to grow following specific environmental queues. Dormant phenotypes are shown by some cancer cells, bacterial persisters, hynozoites in P. vivax, and seeds in plants (25–28). How cells regulate dormancy is less well understood, with the exception of plants in which phytohormones are known to regulate the processes of dormancy and seed germination. Given the link between the apicoplast of P. falciparum and plants, we explored the potential that phytohormones ABA and GA may elicit phenotypes related to DHA-induced dormancy in malaria parasites. In this study, we demonstrated that parasites are released from DHA-induced dormancy earlier when exposed to ABA or GA and that FLD, an inhibitor of ABA, blocks recovery of dormant parasites. Furthermore, the apicoplast and specifically the isoprenoids it produces are essential for recovery of dormant parasites.

The first phytohormone that we assessed was ABA since it is the primary dormancy regulator in plants. As a signaling hormone, ABA senses stress (e.g., onset of cold weather during winter and water drought) and accumulates, inducing a dormant response to prevent germination, an outgrowth of an embryo from the seed (29). Interestingly enough, ABA did not produce the expected effect with P. falciparum. Supplementing dormant parasites with ABA promoted recovery, as opposed to maintaining dormancy. This observed effect was consistent among multiple P. falciparum clones (W2, 3D7, 11G, and D10-ACP_LEADER) as microscopic signs of normal morphology were seen 24 h before they were seen in control groups (DHA treated only). An important question is why ABA seems to function differently in malaria parasites. Maternal ABA produced by the mother plant and transferred to the seed is unable to induce dormancy but rather promotes embryo growth. Only ABA produced by the seedling induced dormancy in certain species (30). For parasites to utilize exogenous ABA, there must be a mechanism to reach the parasite, and this would involve ABA crossing through the host erythrocyte and into the parasite compartment, where it acts. The human anion transporter Band 3 is expressed on red blood cells (RBCs), and ABA uses it for entry into these cells (31). G protein α subunit GPA1 has been identified as a plasma membrane receptor for ABA in Arabidopsis and yeasts (32). These transporters have not been reported for Plasmodium, and a search in PlasmoDB did not reveal potential candidates. Similarly, no analogs of ABA-specific receptors were identified.

Our knowledge of the role of ABA in Plasmodium is limited, and the earliest clues were first reported for the related apicomplexan parasite Toxoplasma gondii. Nagamune et al. detected endogenous ABA in parasite extracts and studied its effects on development (13). Bradyzoites, the dormant cyst forms, go dormant as a response to stress induced by the host immune response and are unaffected by most drugs, similar to what occurs with dormant P. falciparum. Supplementation with exogenous ABA mediated calcium (Ca²⁺) release by inducing production of second-messenger cyclic ADP ribose (cADPR) (13). This led to elevated levels of intracellular Ca²⁺ that stimulated protein secretion required for development and egress. Ca²⁺ is required for asexual Plasmodium growth and invasion of erythrocytes (33). Transient receptor potential (Trp) channel homologs and inositol 1,4,5-triphosphate (IP₃) for Ca²⁺ release have been identified in Plasmodium spp. (34, 35). Ca²⁺-mediated release by ABA could possibly explain its mechanism in P. falciparum recovery from dormancy. Further studies are warranted to identify the mechanism(s) whereby ABA stimulates parasite recovery from dormancy.

Recently, ABA was used to investigate malaria disease severity and parasite transmission in mosquito and mouse hosts using Plasmodium yoelii (15). There was a reduction in blood parasitemia and gametocytemia following ABA supplementation as well as reduced transmission to mosquitos. After exploration of the in vitro effects with P. falciparum, it was determined that this decrease was likely due to ABA stimulating the host's immune system. There was no direct effect assessed on in vitro development of the parasite as asexual growth, nor was gametocyte formation altered. This study sheds light on the effect of the hormone on host systems, which ABA affects, reducing parasite burden. Combined with its effect on dormancy, ABA could work to stimulate recrudescence and reduce parasite burden by further stimulating the host's immune response to help clear residual parasites.

GA is a phytohormone that during seed development ends dormancy and promotes germination and growth (11, 36). In our experiments, GA produced a similar phenotype in DHA-induced dormant P. falciparum. GA not only stimulated early release from dormancy but also did so more potently than ABA (≥10-fold) and acted 24 h earlier (Fig. 2A). Very little work has been reported on gibberellins and apicomplexan parasites. In one study, GA was found to increase the multiplication rate 3- to 4-fold, resulting in higher parasitemia when used at 5 μM to supplement P. falciparum blood-stage cultures (37). This role as a growth promoter might help explain why the effect of GA on dormant parasites is more pronounced than that of ABA. Several gibberellin biosynthesis inhibitors in plants (inabenfide and AMO-1618) were effective at killing P. falciparum parasites by causing cells to swell and rupture (38), yet GA supplementation did not reverse these effects, suggesting a different mechanism of action than observed with GA rescue from dormancy. As a growth promoter, GA might stimulate P. falciparum growth signals in dormant parasites to release them from the growth arrest induced during dormancy. In plants, the hormone is theorized to activate enzymes that release dormancy (39). Mechanisms associated with parasite recrudescence are unknown, but phytohormone influence on expression might reveal upregulation in genes that are key to this process.

FLD is an herbicide that inhibits ABA and carotenoid biosynthesis in plants; thus, it was not surprising that FLD would have the opposite effect from ABA by keeping parasites dormant during FLD exposure but releasing them upon removal. This result is consistent with a previous study showing that FLD induces a dormant response in T. gondii by induction of dormant bradyzoite forms that encyst and grow slowly (13). FLD exposure prevented dormant P. falciparum from recovering, but the effect was reversible if FLD was removed within 8 days. ABA or GA supplementation of FLD-treated parasites was unable to reverse this effect. We also explored the effect of constant FLD pressure on nondormant cultures. Progression through the life cycle was slow and parasitemia remained low, but dormant parasites did not develop (data not shown).

Isoprenoids are a large class of molecules found in all living organisms. The apicoplast is required for isoprenoid biosynthesis, and ABA and GA are both derived from isoprenoids in plant species. In Plasmodium, isoprenoids serve as building blocks for many metabolites, such as ubiquinone, and are vital for various cellular functions (40–42). They are essential for parasite survival in the blood stages, but this has not been examined in dormant forms for viability. In this study, we aimed to explore how essential the apicoplast and isoprenoids were to the recovery of dormant parasites. We exploited the fact that IPP can compensate for isoprenoid inhibition in apicoplast-deficient parasites and those that had the biosynthetic pathway directly disrupted with FOS (18). We observed that the downstream product GGPP, which is used for carotenoid formation and protein prenylation, can achieve the same effect at lower concentrations when exposed to up to 10 μM FOS.

The essential role of the apicoplast in parasite recovery from dormancy was evaluated, and our results show that the apicoplast is required for recrudescence from DHA-induced dormancy. In the absence of the organelle, parasites fail to recover, but this effect was overcome with IPP supplementation. ABA, GA, or GGPP alone does not appear to be sufficient to promote full recovery from dormancy without the apicoplast present. This result was surprising, because initially we expected parasites exposed to ABA and GA to exit dormancy but then die, as neither can compensate for the absence of IPP. Interestingly, apicoplast-deficient parasites can survive for up to 3 days when supplemented with GGPP before dying; however, these parasites do not exit dormancy in the presence of GGPP. This suggests that IPP plays a larger role in recovery and that isoprenoid products not derived specifically from GGPP are necessary to “awaken” parasites and promote growth. IPP is the basic precursor for all isoprenoid products, and a network of these may be needed to exit and survive. One possible explanation is mitochondrion function. IPP is a precursor of ubiquinone, an essential metabolite of the electron transport chain. Dormant parasites exhibited low mitochondrial membrane potential, and when exposed to atovaquone, an inhibitor of the Q_O site of cytochrome b, recovery of dormant parasites was delayed (43). Low levels of pyruvate metabolism are maintained in dormant parasites (7). Pyruvate serves as one of the starting materials for IPP production, so this pathway possibly may be ramped up in parasites as they prepare to exit dormancy. We also noted that dormant parasites remained viable for up to 8 days without the organelle or IPP supplementation. This result confirmed that the parasites were not dead but had their development arrested in a dormant state. IPP was, then, essential for parasite growth after the signals for parasite recovery from dormancy occurred.

Our second approach to inhibiting isoprenoid biosynthesis in dormancy involved using FOS to directly target DOXP reductase in the apicoplast, preventing the pathway from completion. These parasites retained the organelle, and with it present, recrudescence followed DHA exposure. Merely inhibiting isoprenoid production did not prevent a release from their dormant state; however, IPP or GGPP was required for full recovery. This was due to the toxic activity of FOS: even though parasites appear to be protected from FOS inhibition while dormant, they become susceptible again as rings recover to resume growth. At this stage, isoprenoids were needed to reverse FOS inhibition. High percentages of trophozoites and schizonts were observed to exit dormancy in the FOS-only groups. The few rings that were observed were likely parasites that had just emerged from dormancy and not newly invaded parasites from these schizonts. In these studies, FOS (5 μM) may not be sufficient for a complete block of isoprenoid inhibition, and therefore, low levels of IPP present in the parasites before they go dormant might be used to recover.

Susceptibility of ring stages of P. falciparum to drugs has emerged as an important pharmacodynamic property of antimalarial drugs. Artemisinin drugs are the most effective and rapidly acting drugs because unlike most other classes of drugs, they inhibit ring stages. Similarly reduced susceptibility of ring stages is associated with artemisinin resistance, and artemisinin-induced dormancy in rings is thought to play a role in recrudescence (8). Herein we show that the phytohormones ABA and GA have significant effects on dormant parasites and promote early recovery from DHA-induced dormancy. These results suggest a strategy for new drug combinations with an artemisinin drug to prevent or reduce recrudescence caused by dormant parasites that are more difficult to eliminate than growing cells. The phenotypes produced by these phytohormones and the herbicide FLD also provide a means to explore the mechanism(s) underlying dormancy and the regulatory network that promotes cell cycle arrest in P. falciparum. The findings of this study could potentially have larger implications on dormant forms seen in different Plasmodium species. The hypnozoites of P. vivax in hepatocytes remain elusive, as the precise mechanism behind their formation and subsequent relapses is not fully understood. If ABA and/or GA were able to stimulate a release from dormancy, theoretically a drug with potent activity against growing liver stages could be used to clear infection. Further work is warranted to better understand the effects that phytohormones may have in regulating dormancy and promoting growth in malaria parasites.

MATERIALS AND METHODS

In vitro parasite culture.

P. falciparum clones W2 (Indochina), 3D7 (Netherlands), D10-ACP_LEADER-GFP (MRA-568), and ARC08-88 (11G; Cambodia) (22) were maintained in culture using previously described methods (44). The 3D7 clone used in these studies expresses luciferase. Cultures were maintained in complete medium consisting of RPMI 1640, 0.5% Albumax II, 0.2% sodium bicarbonate, 50 mg/liter of hypoxanthine, 25 mM HEPES, and 2% O⁺ human red blood cells (RBCs; Interstate Blood Bank, Memphis, TN). D10 was cultured in the presence of 100 nM pyrimethamine. For dormancy experiments and drug assays, cultures were synchronized to ring stage using 5% (wt/vol) d-sorbitol (Sigma-Aldrich) and washed three times with RPMI medium (45).

Dormancy induction.

Dormancy was induced by adding 700 nM dihydroartemisinin (DHA) to synchronized rings. Parasitemia was adjusted to 2% in 4% hematocrit, and parasites were exposed to DHA for 6 h except where otherwise noted. After DHA was washed out, 100 μM (±)-abscisic acid (Acros Chemicals), 50 μM fluridone (Sigma-Aldrich), or 10 μM gibberellic acid (Sigma-Aldrich) was added to culture. Experiments were done with different exposures to phytohormones and conditions; 1-ml cultures were seeded in 12-well plates in duplicates. A sham control with a vehicle (dimethyl sulfoxide [DMSO]) was also grown in parallel. Giemsa-stained thick and thin blood smears were obtained daily and used to monitor recrudescence and growth. Complete medium was changed every 48 h. Experiments were conducted with two replicates at least twice.

Drug susceptibility assays.

Drugs were serially diluted 1:2 in complete medium in 96-well plates, and 10 μl of drug was transferred to a new plate with 90 μl/well of synchronized ring-stage parasites at 0.5% parasitemia and 2% hematocrit. After 48 h of incubation, 1 μCi of [³H]hypoxanthine monochloride was added to each well and allowed to be incorporated for an additional 24 h. Plates were then frozen at −80°C. For assays with artemisinin drugs, hypoxanthine was added at time zero and allowed to incubate for 48 h prior to freezing (6). Plates were thawed, washed, and harvested in a FilterMate cell harvester, 50 μl of scintillation cocktail was added to each well, and plates were read in a TopCount NXT scintillation counter. Data were imported into TriFox DataAspects Plate Manager software for calculation of 50% inhibitory concentration (IC₅₀) and curve fitting analysis.

GGPP rescue in isoprenoid-inhibited parasites.

Ring-stage synchronized luciferase-expressing 3D7 was exposed to 1 to 100 μM fosmidomycin (FOS [Sigma-Aldrich]) to inhibit isoprenoid biosynthesis or pretreated with 2 μM doxycycline (DOX [Sigma-Aldrich]) for 48 h to generate parasites without an apicoplast (23). Parasites were plated in 24-well plates at a starting parasitemia of 0.5% in 2% hematocrit, and 200 μM isopentyl pyrophosphate trisammonium salt (IPP [Isoprenoids, LC]) or 0.02, 0.2, 2 or 20 μM geranylgeranyl pyrophosphate trisammonium salt (GGPP [Isoprenoids, LC]) was added to each drug treatment. Aliquots were taken daily and combined with 150 μg/ml of d-luciferin (PerkinElmer) for 10 min prior to luminescence reading on a TopCount NXT scintillation counter. In a separate experiment, 2 μM GGPP was given to parasites treated with 5 μM FOS. On day 4, GGPP was removed from one of these groups, while FOS was maintained in culture. Micrograph images were obtained daily.

Isoprenoid inhibition in dormant parasites.

Synchronous ring-stage D10-ACP_LEADER-GFP parasites were pretreated with 2 μM DOX for 48 h and then sorbitol synchronized again. Parasitemia was adjusted to 2% in 4% hematocrit, 700 nM DHA was used to induce dormancy, and then DHA was washed out (3 times) after 6 h. Next 200 μM IPP, 2 μM GGPP, and 100 μM ABA or 10 μM GA was added and cultures were maintained in the presence of DOX. In a separate experiment, 5 μM FOS was used to inhibit isoprenoid synthesis in dormant parasites. FOS was added at the same time as DHA and maintained in culture. Then 200 μM IPP or 2 μM GGPP was added following DHA removal and kept in media. Fluorescence images and slides were taken daily.

Fluorescence microscopy.

For fluorescence images, D10-ACP_LEADER-GFP parasites in 0.1% hematocrit were incubated with 1 μg/ml of Hoechst 33342 nuclear stain and 50 μM MitoTracker Red CMXRos for 20 min at 37°C, followed by 3 washes with RPMI medium. Cells were resuspended in complete medium and allowed to settle onto 8-well chambered slides (Thermo Scientific) coated with 0.05 mg/ml of poly-l-lysine. Images were obtained on a DeltaVision CORE microscope (Applied Precision) using a 100× objective.