Until now, the mystery of apicoplast function has been a critical gap in our understanding of malaria pathogenesis. Our findings demonstrate that the production of isoprenoid precursors is the only essential function of the apicoplast during asexual blood-stage P. falciparum
growth (). This surprising revelation has several important implications and invites a host of new questions. Because isoprenoid precursors are building blocks to synthesize cellular isoprenoid products with diverse functions, their key role now gives added urgency to the elucidation of these products and their downstream functions. At least one essential prenylated product is ubiquinone, a component of the mitochondrial electron transport chain. There are certainly other essential, as-yet-unidentified isoprenoid products since transgenic parasites which express yeast dihydroorotic acid dehydrogenase and no longer require their electron transport chain are still susceptible to fosmidomycin and antibiotics and could be rescued with IPP supplementation (unpublished data) 
. Possible isoprenoid products include dolichols involved in protein N
-glycosylation which have been detected in Plasmodium
and prenylated proteins, such as Rab homologs required for vesicular trafficking and a recently identified tyrosine phosphatase 
The current findings also imply that several annotated apicoplast pathways are in fact non-essential. Amongst both identified pathways and the 70% of apicoplast gene products with unknown function, only isoprenoid precursor biosynthesis and any pathways supporting this function in blood-stage parasites (including those required for organelle maintenance and replication) are essential and therefore viable apicoplast drug targets 
. Assertions that type II fatty acid and, by implication, acetyl-CoA biosynthesis were essential apicoplast functions during blood-stage growth have already been disproven 
. A parasite-derived pathway for heme biosynthesis contains steps that occur in the apicoplast, mitochondria, and cytosol. Our results strongly imply that blood-stage parasites do not depend on de novo heme biosynthesis using this pathway but instead rely on an extrinsic de novo pathway utilizing imported host enzymes or salvage of heme from the host by an unidentified mechanism 
. Still other pathways such as Fe-S cluster biosynthesis supply cofactors for enzymes within the organelle but are not exported outside the organelle. These pathways become “non-essential” when the need for organelle maintenance is removed.
The complexity of the organelle and the simplicity of its blood-stage function pose an obvious contradiction. Approximately 5%–10% of the Plasmodium
genome is predicted to encode apicoplast-targeted gene products (although the localization and/or function of the majority of these gene products have not been validated) 
. In order to import these proteins into the apicoplast, parasites utilize a dedicated protein trafficking pathway 
. In addition, the organelle undergoes complex morphological development during blood stage growth, requiring cellular machinery to faithfully replicate and segregate the organelle at every cell division 
. Why are such huge resources consumed to maintain a single essential function? First, while the function of the apicoplast is limited during the blood stage, the need for more extensive organelle function during other developmental stages may dictate its maintenance in intraerythrocytic parasites as the organelle cannot be generated de novo. Fatty acid biosynthesis, for example, is an essential apicoplast function in liver stage parasites 
. Second, Plasmodium
may have been evolutionarily trapped in its bondage to the apicoplast. Having acquired the plastid early in its evolution, it may have been unable to dispense of it even after adopting an increasingly parasitic lifestyle due to the transfer of even a few essential functions to the organelle. In any case, this imbalance emphasizes the value of targeting housekeeping pathways involved in organelle maintenance and replication to interfere with its function.
An important consideration is whether our findings accurately reflect in vivo growth requirements of parasites during infection. Specifically, are there essential metabolites supplemented in culture which could not be acquired during in vivo growth and instead must be biosynthesized by the apicoplast? While parasitized RBCs during infection use human plasma as a source of extracellular nutrients, our cultures were grown in RPMI medium 1640 supplemented with purified serum substitute, Albumax. We found that Albumax could be replaced with 10% human serum with no effect on the survival of apicoplast-minus parasites in the presence of IPP (Figure S5
). RPMI medium contains salts, 20 amino acids, 11 vitamins, 4 other organic molecules, and glucose. The acquisition and biosynthesis of these nutrients by blood-stage Plasmodium
and their essentiality for intraerythrocytic growth based on available evidence is shown in Table S1
. In general, blood-stage Plasmodium
biosynthesizes very limited amounts of just 3 amino acids and is dependent on amino acids from either (1) hemoglobin degradation or (2) acquisition from patient plasma through newly established permeation pathways in the infected red cell 
. Similarly, current knowledge of Plasmodium
metabolism also suggests that the remaining organic metabolites found in RPMI medium are biosynthesized by non-apicoplast pathways or can be acquired from the host red cell or patient plasma 
. Consequently, we believe that our findings can be extrapolated to in vivo requirements for the apicoplast to support parasite growth and development. At the very least, our results define a very minimal set of potential metabolites (IPP and components found in RPMI 1640 medium) that could be biosynthesized in the apicoplast. We cannot, however, rule out additional apicoplast functions (other than those required for growth) that would not be revealed in our blood culture system, such as functions required for immune evasion.
Several aspects of the chemical rescue with isoprenoid precursors are notable. During chemical rescue, exogenous IPP could enter the parasite through permeation pathways established in the parasitized erythrocyte or other uncharacterized membrane transporters 
. The RBC is largely metabolically inactive and unlikely to have significant ongoing isoprenoid precursor biosynthesis via the host mevalonate pathway or stores of these metabolites 
. It is also unlikely that these high-energy pyrophosphorylated molecules would accumulate to appreciable levels in plasma (200 µM was required for rescue in our experiments). Consistent with this notion, IPP was not present in the Serum Metabolome Database (SMDB), which contains 4,229 detectable metabolites 
. Therefore, acquisition of isoprenoid precursors in vivo by salvage of IPP from infected blood is improbable. Once in the parasite, exogenous IPP may fulfill its function in the cytoplasm with or without uptake into the apicoplast 
Although both IPP and DMAPP are required to synthesize isoprenoid products, supplementation with IPP alone is sufficient to fulfill endogenonous isoprenoid precursor biosynthesis, implying the presence of an IPP isomerase in the cytoplasm that converts IPP to DMAPP. This activity may be important in establishing the optimal cellular ratio of IPP to DMAPP, as toxicity was noted with increasing DMAPP concentrations in our experiments. A putative IPP isomerase has been identified in the Plasmodium
. A recent report suggested that geranylgeraniol, the alcohol analog of a C20
prenyl chain, could rescue fosmidomycin inhibition 
. We were unable to rescue fosmidomycin inhibition with alcohol analogs of IPP and DMAPP, indicating either poor cellular penetration of the alcohols or the absence of a kinase to convert the alcohol analogs to the pyrophosphorylated and active metabolites (Figure S1
). Even with conversion of geranylgeraniol to geranylgeranyl pyrophosphate in the cell, it would seem that a C5
building block, such as IPP, would almost certainly be required to extend the supplemented C20
unit for construction of polyprenyl chains, such as that found in ubiquinone, and to construct smaller prenyl chains, such as for protein farnesylation. The reported rescue with geranylgeraniol was performed at 1.5 µM fosmidomycin, which is above the concentration required for 50% growth inhibition but may be below that required for adequate inhibition of the biosynthetic pathway (since phenotypic growth inhibition can be apparent even at low levels of inhibition of the biosynthetic pathway) 
. Therefore, the reported results may be complicated by ongoing biosynthesis of IPP and DMAPP contributing to the precursor pool. Consistent with this, neither farnesol nor geranylgeraniol was able to rescue fosmidomycin concentrations >10 µM, and both showed dose-related parasite toxicity (Figure S6
). In contrast, we were able to demonstrate IPP rescue at fosmidomycin concentrations exceeding 100 µM, well above its EC90
for growth inhibition.
The consequences of apicoplast loss following antibiotic treatment and IPP rescue are no less intriguing. In the parasites that survive antibiotic treatment by chemical rescue, the organelle is irreversibly lost when it fails to segregate to daughter cells 
. In these apicoplast-minus parasites, apicoplast gene products encoded in the nucleus may continue to be transcribed and translated. These products may properly shuttle into the secretory pathway but cannot be diverted to the organelle 
. Based on the microscopy results, we hypothesize that proteins may be packaged into transport vesicles bound for the organelle but are unable to localize to the missing structure and therefore accumulate in the cytoplasm appearing as numerous foci. While we cannot rule out the presence of structural remnants of the apicoplast, the observed foci are unlikely to support apicoplast functions. Apicoplast-targeted proteins may require both cleavage of the long basic transit peptide and chaperones in the lumen of the apicoplast for proper folding. We observed that cleavage of the transit peptide from targeted proteins, a critical apicoplast function, does not occur in rescued parasites ().
The close physical and functional relationship between the apicoplast and the mitochondria raises the possibility that loss of the apicoplast might affect the ability of the mitochondria to replicate and divide. We were able to detect the mitochondrial genome by qPCR for the cytB3 gene and observe labeling of the mitochondria with Mitotracker by fluorescence microscopy in apicoplast-minus parasites (; unpublished data). Despite the loss of the apicoplast, these parasites do appear to contain mitochondria.
While the survival of apicoplast-minus P. falciparum invokes a slew of intriguing questions, these same parasites will be a powerful and indispensable tool for further dissection of apicoplast biology. Apicoplast-minus P. falciparum strains generated in this study can be used to assess organelle requirement during gametocytogenesis and mosquito stage development. These strains also provide novel avenues to identify isoprenoid products, generate conditional mutants of essential genes involved in apicoplast maintenance and replication, conduct metabolomic or proteomic profiling, and study protein trafficking to the organelle.
With regard to drug development, our chemical rescue strategy also addresses the critical deficiency of current cell growth screening assays, namely lack of knowledge of the drug target. Candidate drug hits detected in phenotypic assays can be screened for chemical rescue of the growth inhibition. The reversal of growth inhibition by IPP supplementation specifically identifies inhibitors that target pathways involved in MEP pathway function, replication, or maintenance of the apicoplast, providing a pathway-specific drug screen to aid in discovery of new classes of anti-malarials. The ability to chemically complement the cell death phenotype will prevent false leads from off-target effects, like that seen with triclosan and its misconstrued effect on type II fatty acid biosynthesis 
Finally, the apicoplast-minus strains dependent on IPP for continued growth are a unique and ideal candidate for an attenuated blood-stage vaccine 
. Unlike irradiated or drug-treated whole parasite vaccines, apicoplast-minus parasites would continue to develop in blood at most one cycle, including a single erythrocyte rupture and reinvasion, thereby stimulating a stronger immune response. However, judging by the effects of IPP withdrawal in culture, they would be unable to develop further in the absence of exogenous IPP (). Lending support to this notion, a similar “limited survival” strategy targeting the apicoplast in liver-stage parasites has proven effective as a liver-stage vaccine candidate 
. A significant advantage of our approach is that attenuation is achieved chemically and does not require difficult or costly genetic manipulation (as is the case with genetically modified vaccine strains), thereby allowing for the possibility of incorporating circulating field strains of Plasmodium
in a vaccine formulation 
. There would also be very little risk of reversion as it would be extremely difficult to reacquire apicoplast function by mutation.
In summary, we believe that the current study ushers in a new era in the investigation of the apicoplast in Plasmodium with exciting opportunities to counteract the malarial scourge on human lives.