Dephospho‐CoA kinase, a nuclear‐encoded apicoplast protein, remains active and essential after Plasmodium falciparum apicoplast disruption

Abstract

Malaria parasites contain an essential organelle called the apicoplast that houses metabolic pathways for fatty acid, heme, isoprenoid, and iron–sulfur cluster synthesis. Surprisingly, malaria parasites can survive without the apicoplast as long as the isoprenoid precursor isopentenyl pyrophosphate (IPP) is supplemented in the growth medium, making it appear that isoprenoid synthesis is the only essential function of the organelle in blood‐stage parasites. In the work described here, we localized an enzyme responsible for coenzyme A synthesis, DPCK, to the apicoplast, but we were unable to delete DPCK, even in the presence of IPP. However, once the endogenous DPCK was complemented with the E. coli DPCK (EcDPCK), we were successful in deleting it. We were then able to show that DPCK activity is required for parasite survival through knockdown of the complemented EcDPCK. Additionally, we showed that DPCK enzyme activity remains functional and essential within the vesicles present after apicoplast disruption. These results demonstrate that while the apicoplast of blood‐stage P. falciparum parasites can be disrupted, the resulting vesicles remain biochemically active and are capable of fulfilling essential functions.

Production of coenzyme A is an indispensable function of the apicoplast organelle in blood‐stage malaria parasites.

Introduction

Plasmodium falciparum parasites rely on exogenous pantothenate (vitamin B5) (Geary et al, 1985; Saliba et al, 2005), which they use for the de novo synthesis of coenzyme A (CoA) (Spry & Saliba, 2009). Within the parasite, CoA functions as an essential cofactor that is used for a variety of biosynthetic pathways (Spry et al, 2008), including the generation of acetyl‐CoA, fatty acid biosynthesis and modification, and cellular oxidation and metabolism (Spry et al, 2010). The CoA biosynthetic pathway has been considered as an attractive source of antimalarial drug targets, with multiple studies showing that inhibitors of this pathway are effective against the asexual red blood cell (RBC) stage of the parasite (Spry et al, 2005; Spry et al, 2013; Macuamule et al, 2015; Schalkwijk et al, 2019).

Pantothenate is imported into malaria parasites by a proton‐coupled pantothenate transporter (Saliba & Kirk, 2001). In P. falciparum, the transporter PAT (PF3D7_0206200) has been proposed to fulfill this role (Augagneur et al, 2013). However, there is some uncertainty about the exact function of PAT since it is dispensable in blood‐stage parasites and appears to be essential for processes such as exocytosis and vesicle fusion during other stages of parasite development (Hart et al, 2014; Kehrer et al, 2016; Kenthirapalan et al, 2016). Once inside the cell, pantothenate is converted into CoA through a series of five enzyme‐mediated steps (Saliba & Kirk, 2001; Hart et al, 2014). The first four enzymes, including pantothenate kinase 1 or 2 (PanK1/2), phosphopantothenoylcysteine synthase 1 or 2 (PPCS1/2), phosphopantothenoylcysteine decarboxylase (PPCDC), and phosphopantetheine adenylyltransferase (PPAT), are believed to localize to the cytosol of the parasite (Spry et al, 2008). However, the last step of this pathway, in which dephospho‐CoA kinase (DPCK) converts dephospho‐CoA to CoA, is predicted to occur within the apicoplast (Fig 1) (Ralph et al, 2004).

Figure 1. Generation of CoA and its utilization in the apicoplast.

Pantothenate is metabolized into dephospho‐CoA through a series of enzymes within the cytosol (PanK1/2, PPCS1/2, PPCDC, and PPAT). The final step involves the generation of CoA by DPCK and is believed to occur within the apicoplast. Within the apicoplast, CoA is used to generate acetyl‐CoA and malonyl‐CoA and is used by ACPS to carry out the 4‐phosphopantetheine modification of ACP to generate holo‐ACP. The CoA produced within the apicoplast is also believed to be exported into the cytosol for utilization throughout the parasite cell. PanK1/2: pantothenate kinase 1 or 2; PPCS1/2: phosphopantothenoylcysteine synthase 1 or 2; PPCDC: phosphopantothenoylcysteine decarboxylase; PPAT: phosphopantetheine adenylyltransferase; DPCK: dephospho‐CoA kinase; ACP: acyl‐carrier protein; ACPS: acyl‐carrier protein synthase; ACC: acetyl‐CoA carboxylase; PDH: pyruvate dehydrogenase.

While not experimentally confirmed, DPCK is predicted to localize to the apicoplast due to the presence of an N‐terminal bipartite apicoplast trafficking peptide (Ralph et al, 2004). The CoA generated by DPCK is thought to be used within the apicoplast to modify the acyl‐carrier protein (ACP) with a 4‐phosphopantetheine cofactor (Spry et al, 2008). This modification is catalyzed by ACP synthase (ACPS) and is required to form the mature and biologically active form of ACP, holo‐ACP (Waters et al, 2002; Spry et al, 2008; Gallagher & Prigge, 2010). Holo‐ACP plays a central role in the type II fatty acid synthesis (FASII) pathway within the apicoplast (Ralph et al, 2004; Shears et al, 2015). Additionally, since DPCK provides the only predicted source of CoA for the parasite, the CoA generated within the apicoplast is presumably exported out of the organelle for use throughout the parasite cell by a transporter that has yet to be identified (Spry et al, 2008).

It has previously been demonstrated that the apicoplast organelle can be disrupted in blood‐stage P. falciparum parasites if supplemented with sufficient levels of the isoprenoid precursor isopentenyl pyrophosphate (IPP) (Yeh & DeRisi, 2011). These results have been interpreted as IPP being the only essential product of the apicoplast. However, with DPCK also predicted to be an essential apicoplast‐localized protein, parasites containing a disrupted apicoplast should then require supplementation with both CoA and IPP to survive (Ralph et al, 2004; Hart et al, 2017). The ability of parasites to survive on IPP supplementation alone may be explained by the finding that parasites with a disrupted apicoplast have been shown to accumulate vesicles throughout the cell (Yeh & DeRisi, 2011; Gisselberg et al, 2013). These vesicles contain nuclear‐encoded apicoplast‐specific proteins that would have presumably been trafficked to the organelle under typical conditions (Bowman et al, 2014; Pasaje et al, 2016; Florentin et al, 2017; Uddin et al, 2018). Thus, it is possible that vesicle‐localized proteins remain biochemically active and are capable of fulfilling their essential functions within the cell.

Through the work described here, we confirmed the apicoplast localization of DPCK. Additionally, we attempted to generate a genetic knockout of DPCK and were unsuccessful, even with an apicoplast metabolic bypass, suggesting an essential role beyond its function within the organelle. However, upon complementation of the endogenous DPCK with E. coli DPCK (EcDPCK), we were successful in deleting it. This not only helps to confirm the activity of the endogenous DPCK, which has not been previously enzymatically characterized, but also provides additional evidence that the protein is likely to be essential. To demonstrate that DPCK activity is indeed essential, we knocked down the complemented EcDPCK in a parasite line in which the endogenous DPCK had been deleted, which resulted in the inhibition of parasite growth. We also showed that DPCK is present and active in the parasite vesicles that are generated after the disruption of the apicoplast, with that activity being required for parasite growth. Overall, these results demonstrate that the vesicles generated after the disruption of the apicoplast in blood‐stage P. falciparum parasites remain biochemically active and capable of fulfilling essential metabolic functions.

Results

DPCK is refractory to deletion, even when provided with a metabolic bypass

Based on the assumption that DPCK is an essential apicoplast protein, we attempted to delete it in a metabolic bypass parasite line called PfMev. The PfMev line contains a genetically encoded apicoplast bypass system that can convert exogenously supplied mevalonate into the isoprenoid precursors isopentenyl pyrophosphate (IPP) and dimethylallyl pyrophosphate (DMAPP) (Swift et al, 2020b). In the presence of mevalonate, PfMev parasites can survive the deletion of essential apicoplast proteins and even the disruption of the apicoplast organelle (Swift et al, 2020a; Swift et al, 2020b). The PfMev parasite line also expresses the green fluorescent protein, super‐folder green (SFG), with the N‐terminal trafficking peptide of the P. falciparum acyl‐carrier protein (api‐SFG), which directs its trafficking to the apicoplast, thereby fluorescently labeling the organelle (Swift et al, 2020b).

Using the PfMev line, we attempted to delete DPCK using Cas9‐mediated genome editing (Ghorbal et al, 2014), under continuous supplementation with 50 μM mevalonate. This should allow DPCK to be deleted even if it is required for apicoplast function. However, we failed to delete DPCK despite six independent attempts (Appendix Table S1). Two of these attempts were done with direct supplementation of 5 mM CoA, which has been shown to rescue the growth of parasites treated with inhibitors that target the CoA biosynthetic pathway (Spry et al, 2010; Fletcher & Avery, 2014; Fletcher et al, 2016). As discussed further below, this result is consistent with the hypothesis that CoA supplementation may only bypass early steps of the CoA biosynthesis pathway. The inability to delete DPCK in the presence of mevalonate suggests that DPCK has an essential function beyond its potential requirement within the apicoplast or that it is not actually an apicoplast protein to begin with.

DPCK localizes to the apicoplast organelle

In P. falciparum, DPCK is predicted to localize to the apicoplast (Ralph et al, 2004), but this has yet to be experimentally confirmed. In order to determine the localization of DPCK, we generated a genetic construct for the expression of DPCK with a C‐terminal mCherry tag under control of the calmodulin promoter and integrated it into the genome of PfMev parasites. To accomplish this, we first inserted an attB site into a locus known to be dispensable in blood‐stage parasites (P230p; PF3D7_0208900) (Marin‐Mogollon et al, 2018). Plasmids pCasG (Rajaram et al, 2020) and pRS (Swift et al, 2020b) were used to insert the 40 base pair attB sequence using markerless selection (Figs 2A and EV1) followed by sequencing to verify the correct insertion. The resulting PfMev p230p attB line was used for this, and all subsequent experiments, unless otherwise noted. The DPCK‐mCherry expression construct was inserted into this line using attP/attB recombination (Nkrumah et al, 2006; Spalding et al, 2010) and verified with genotyping PCRs (Figs 2B and EV2A). Analysis via live epifluorescence microscopy revealed colocalization of the api‐SFG and mCherry signals, indicating that DPCK is indeed located in the apicoplast (Fig 2C, Table 1).

Figure 2. Colocalization of DPCK‐mCherry with the apicoplast marker api‐SFG.

1. Genotyping PCR confirming the insertion of a 40 bp attB element in the p230p gene, with amplification demonstrating integration at the Δ5’ and Δ3’ loci and lack of wild‐type parasites (as indicated by the failure to amplify at the wild‐type 5’ and 3’ loci) as compared to the parental control. A schematic describing these genotyping PCRs is shown in Fig EV1.
2. Genotyping PCR confirming the insertion of the PfDPCK‐mCherry construct using attP/attB integration. PCRs demonstrate integration at the Δ5’ and Δ3’ integration sites and lack of parental parasites (P) as compared to the parental control (P230p‐attB). A schematic describing these genotyping PCRs is shown in Fig EV2A.
3. Live epifluorescence microscopy of PfMev parasites that are expressing api‐SFG (green) and the P. falciparum DPCK‐mCherry protein (red) with DAPI marking the nuclei (blue). Colocalization statistics are shown in Table 1. The borders of the red blood cell and parasite are marked with dotted blue and white lines, respectively. Microscopy images represent fields that are 10 µm long by 10 µm wide.

Source data are available online for this figure.

Figure EV1. Schematic and details describing attB insertion into the P230p locus.

Schematic representation showing how 210 bp of P230p was replaced with a 40 bp attB element. The positions of genotyping PCR products designed to identify the wild‐type locus (blue) or the recombinant locus (red) are shown. The table describes how diagnostic primers were used to validate the insertion of the attB element.

Figure EV2. Schematic and genotyping PCRs showing insertion of plasmid p15 into the attB site of PfMev P230p‐attB parasites.

1. Schematic representation showing how the p15 plasmid containing constructs of either E. coli or P. falciparum DPCK was inserted into the attB site of PfMev P230p‐attB parasites. The positions of genotyping PCR products designed to identify the wild‐type locus (blue) or the recombinant locus (red) are shown. The table describes how diagnostic primers were used to validate integration of the p15 plasmid into the P230p‐attB locus.
2. PCRs demonstrate integration at the Δ5’ and Δ3’ integration sites and lack of parental parasites (P) as compared to the parental control (P230p‐attB).

Source data are available online for this figure.

Table 1.

Colocalization Manders M2 values for parasite lines with different treatments.

Parasite Line	Treatment	M2	SEM	N	Figure
PfDPCK‐mCherry	none	0.62	0.04	31	2C
api‐EcDPCK‐mCherry	none	0.78	0.03	6	3C
api‐EcDPCK‐mCherry	∆dpck	0.72	0.03	62	3D
api‐EcDPCK‐mCherry	∆dpck, Azith	0.73	0.02	50	3E
EcDPCK‐mCherry	none	0.17	0.03	18	4A
EcDPCK‐mCherry	∆dpck	0.15	0.01	40	4B
CLD‐EcDPCK‐mCherry‐apt	aTc	0.71	0.05	14
CLD‐EcDPCK‐mCherry‐apt	∆dpck, aTc	0.74	0.02	45	5C
CLD‐EcDPCK‐mCherry‐apt	∆dpck	*0.04	0.00	18	5D
CLD‐EcDPCK‐mCherry‐apt	∆dpck, aTc, Shield1	0.18	0.02	27	5E
CLD‐EcDPCK‐mCherry‐apt	∆dpck, Shield1	*0.02	0.01	11	5F

The Manders M2 value is shown for the colocalization of different DPCK constructs with apicoplast‐localized SFG. The standard error of the mean (SEM), total number of observations (N), and corresponding figure numbers are shown. M2 values indicating apicoplast localization are in bold. Asterisks mark values calculated with very low fluorescence signal due to knockdown of the mCherry fusion protein.

DPCK deletion is enabled by apicoplast complementation with E. coli DPCK, which is present and active within disrupted apicoplast vesicles

Since we failed to knockout DPCK directly, we attempted to delete DPCK in parasites complemented with the E. coli DPCK (EcDPCK) in the apicoplast. We generated an expression construct containing the EcDPCK with the signal and transit peptide of the P. falciparum ACP appended to the N‐terminus to direct its trafficking to the apicoplast (Waller et al, 2000), in addition to a C‐terminal mCherry tag for localization (Fig 3A). The expression of this gene was controlled by the calmodulin promoter, and the construct was integrated into the genome of the PfMev P230p attB parasite line using attP/attB integration (Nkrumah et al, 2006; Spalding et al, 2010) and verified by PCR (Fig EV2A and B). Analysis of the PfMev api‐EcDPCK‐mCherry line by live epifluorescence microscopy showed colocalization of the mCherry and api‐SFG signals, demonstrating that the E. coli DPCK was being expressed and trafficked to the apicoplast (Fig 3B, Table 1). In this parasite line, we successfully deleted the endogenous DPCK and confirmed the deletion via PCR (Figs 3C and EV3). After deleting the endogenous DPCK, we analyzed the resulting PfMev api‐EcDPCK‐mCherry Δdpck line via live epifluorescence microscopy and confirmed that the api‐SFG and mCherry signals continued to be colocalized in the apicoplast (Fig 3D, Table 1). These results indicate that apicoplast expression of EcDPCK can complement the loss of the parasite DPCK and that DPCK is likely essential because it can only be deleted when complemented. Additionally, since the E. coli DPCK (encoded by the coaE gene) has already been biochemically characterized (Mishra et al, 2001), successful complementation indicates that the putative parasite DPCK (which has not previously been biochemically characterized) has dephospho‐CoA kinase enzymatic activity.

Figure 3. E. coli DPCK complementation in the apicoplast and the deletion of DPCK.

1. Construct used to express the E. coli DPCK in the apicoplast. Appended are the N‐terminal signal sequence and transit peptide from ACP (api) and a C‐terminal mCherry tag.
2. Live epifluorescence microscopy of PfMev parasites that are expressing api‐SFG (green) and api‐EcDPCK‐mCherry (red) with DAPI marking the nuclei (blue).
3. Genotyping PCR confirming the deletion of the endogenous DPCK, with amplification demonstrating integration at the Δ5’ and Δ3’ loci and lack of wild‐type parasites (as indicated by the failure to amplify at the wild‐type 5’ and 3’ loci) as compared to the parental control. A schematic describing these genotyping PCRs is shown in Fig EV3.
4. Live epifluorescence microscopy of the PfMev api‐EcDPCK‐mCherry Δdpck parasite line that is expressing api‐SFG (green) and api‐EcDPCK‐mCherry (red) with DAPI marking the nuclei (blue).
5. Live epifluorescence microscopy of the PfMev api‐EcDPCK‐mCherry Δdpck line after treatment with 1x azithromycin for 7 days under supplementation with 50μM mevalonate. This parasite line expresses api‐SFG (green) and api‐EcDPCK‐mCherry (red) with DAPI marking the nuclei (blue).
6. Attempted PCR detection of the ldh, sufB, and cox1 genes from the nuclear (N), apicoplast (A), and mitochondrial (M) genomes, respectively. We failed to amplify sufB from the azithromycin‐treated (+ Azith) PfMev api‐EcDPCK‐mCherry Δdpck parasites, indicating the loss of the apicoplast organellar genome.

Data information: All microscopy images represent fields that are 10 µm long by 10 µm wide. Colocalization statistics are shown in Table 1.

Source data are available online for this figure.

Figure EV3. Schematic and details describing gene deletion experiments.

Schematic representation showing how DPCK was replaced with a drug resistance cassette (hDHFR or BSD). The positions of genotyping PCR products designed to identify the wild‐type locus (blue) or the recombinant locus (red) are shown. The table describes how diagnostic primers were used to validate the ∆dpck genotype in knockout parasite lines.

We next wanted to determine whether apicoplast EcDPCK expression could complement parasites containing a disrupted apicoplast. Continued parasite survival in this state would indicate that the EcDPCK is present and active within the disrupted apicoplast vesicles. Thus, we treated the PfMev api‐EcDPCK‐mCherry Δdpck line with 100 nM azithromycin in the presence of 50 μM mevalonate for 7 days to induce apicoplast disruption (Gisselberg et al, 2013). The azithromycin‐treated parasites continued to survive in media with added mevalonate, and analysis by live epifluorescence microscopy showed the presence of multiple vesicles instead of the typical intact organellar structure, consistent with the previously reported disruption phenotype (Yeh & DeRisi, 2011; Gisselberg et al, 2013). Furthermore, in this parasite line both the api‐SFG and mCherry signals remained colocalized, indicating that EcDPCK was present within the apicoplast vesicles of the parasite (Fig 3E, Table 1). We then confirmed apicoplast disruption using a PCR assay to demonstrate loss of the sufB gene (Gisselberg et al, 2013), encoded by the apicoplast genome (Fig 3F). Apicoplast disruption was also corroborated by Western blot analysis of api‐EcDPCK‐mCherry and api‐SFG, which revealed the presence of precursor and mature proteins in untreated parasites but only the precursor proteins in parasites exposed to azithromycin (Fig EV4). Taken together, in the presence of a disrupted apicoplast, the continued parasite survival and colocalization of the api‐SFG and mCherry signals suggest that EcDPCK is present and active within the vesicles present after apicoplast disruption.

Figure EV4. Western blots of the PfMev api‐EcDPCK‐mCherry Δdpck line before and after treatment with azithromycin.

(Top panel) Untreated parasites expressing api‐EcDPCK‐mCherry contained two protein bands that were detected with anti‐mCherry antibodies, consistent with the presence of a trafficking intermediate (upper band) and an apicoplast‐localized mature protein (lower band) after cleavage of the transit peptide (lane 1). The lower band corresponding to apicoplast‐localized protein was not present after treatment with azithromycin, consistent with disruption of the apicoplast organelle (lane 2). A single band was detected in PfMev EcDPCK‐mCherry Δdpck parasites, as expected for cytosolic localization (lane 3). (Bottom panel) Anti‐HA antibodies were used to detect api‐SFG‐HA expressed by all PfMev parasite lines. In the absence of azithromycin (lanes 1 and 3), two protein bands were detected, consistent with apicoplast trafficking. Only the trafficking intermediate was present in parasites treated with azithromycin, consistent with disruption of the apicoplast organelle (lane 2).

Source data are available online for this figure.

DPCK can be knocked out if complemented with the E. coli DPCK in the cytosol

It is not clear why the last step of CoA synthesis occurs in the apicoplast. Within the apicoplast, CoA is used to form the 4'‐phosphopantetheine prosthetic group found on ACP (Waters et al, 2002), which plays a central role in the FASII fatty acid biosynthetic pathway (Prigge et al, 2003). However, this pathway is dispensable in blood‐stage parasites (Foth et al, 2005; van Schaijk et al, 2014). One would therefore expect the cytosolic expression of DPCK to be sufficient to complement blood‐stage parasites, unless DPCK is required within the organelle for a separate and essential purpose. To investigate this, we generated a second complementation line expressing the EcDPCK‐mCherry protein lacking an N‐terminal trafficking peptide, for cytosolic expression (Fig 4A). This construct was integrated into the PfMev p230p attB parasite line using attP/attB recombination (Nkrumah et al, 2006; Spalding et al, 2010) and verified by PCR (Fig EV2A and B). As expected, the cytosolic EcDPCK‐mCherry protein did not localize to the apicoplast (Fig 4B, Table 1) and Western blot analysis revealed a single band corresponding to the full‐length protein (Fig EV4). We then used this parasite line to delete the endogenous DPCK under supplementation with 50 μM mevalonate (Fig 4C). Parasites were supplemented with mevalonate to bypass any potential essential DPCK activity within the apicoplast. Analysis of the resulting apicoplast phenotype revealed that the organelle remained intact (Fig 4D and E), and the parasites were not reliant on mevalonate supplementation for survival. Overall, these results indicate that cytosolic DPCK can provide CoA for all essential processes in blood‐stage parasites.

Figure 4. E. coli DPCK complementation in the cytosol and deletion of DPCK.

1. Construct used to express the E. coli DPCK protein with a C‐terminal mCherry tag in the cytosol of the PfMev parasite line.
2. Live epifluorescence microscopy of the PfMev EcDPCK‐mCherry line. This parasite line expresses api‐SFG (green) and EcDPCK‐mCherry (red) with DAPI marking the nuclei (blue).
3. Genotyping PCR confirming the deletion of the endogenous DPCK using primer pairs described in Fig EV3.
4. Live epifluorescence microscopy of the PfMev EcDPCK‐mCherry Δdpck line. This parasite line expresses api‐SFG (green) and EcDPCK‐mCherry (red) with DAPI marking the nuclei (blue).
5. Attempted PCR detection of the ldh, sufB, and cox1 genes from the nuclear (N), apicoplast (A), and mitochondrial (M) genomes, respectively. We were successful in amplifying sufB from the PfMev EcDPCK‐mCherry Δdpck parasite line grown under continuous supplementation with 50μM mevalonate, indicating retention of the apicoplast organellar genome.

Data information: All microscopy images represent fields that are 10 µm long by 10 µm wide. Colocalization statistics are shown in Table 1.

Source data are available online for this figure.

DPCK activity is required for parasite growth

While the previous results strongly suggest that DPCK activity is required for parasite survival, in order to more conclusively demonstrate this, we attempted to knockdown the complemented EcDPCK in parasites in which the endogenous DPCK had been deleted. The expression of EcDPCK‐mCherry was made controllable through the previously established tetR‐DOZI system, which relies on a C‐terminal aptamer array to inhibit translation upon the removal of anhydrotetracycline (aTc) from the parasite culture (Goldfless et al, 2014). Additionally, for a second form of inducible control of EcDPCK, we appended an N‐terminal conditional localization domain (CLD). The CLD directs protein trafficking to the apicoplast, but is rerouted and secreted into the parasitophorous vacuole upon the addition of the ligand Shield1 (Fig 5A) (Gallagher et al, 2011; Roberts et al, 2019). Both systems were simultaneously employed to ensure that we could achieve sufficient downregulation of EcDPCK, which is expressed by the strong bidirectional calmodulin promoter (Crabb & Cowman, 1996; Epp et al, 2008). This construct was integrated into the PfMev p230p attB parasite line (Fig EV2A and B), followed by deletion of the endogenous DPCK, confirmed via PCR (Fig 5B), generating the resulting PfMev CLD‐EcDPCK‐mCherry‐apt (aptamer) Δdpck parasite line. Expression and apicoplast trafficking of the CLD‐EcDPCK‐mCherry protein were confirmed via live epifluorescence microscopy (Fig 5C, Table 1). We then verified the functionality of the individual knockdown systems via live epifluorescence microscopy through the removal of aTc (Fig 5D), the addition of Shield1 (Fig 5E), or the simultaneous removal of aTc and the addition of Shield1 (Fig 5F). Removal of aTc resulted in a diminished mCherry signal, consistent with knockdown of the protein, and the addition of Shield1 resulted in mislocalization of the protein away from the apicoplast (Fig 5, Table 1).

Figure 5. Characterization of the PfMev CLD‐EcDPCK‐mCherry‐apt Δdpck line under permissive and non‐permissive conditions.

1. Construct used to generate the PfMev CLD‐EcDPCK‐mCherry‐apt parasite line.
2. Genotyping PCR confirming the deletion of the endogenous DPCK (Δdpck) using primer pairs described in Fig EV3.
3. Live epifluorescence microscopy of the PfMev CLD‐EcDPCK‐mCherry‐apt Δdpck line. This parasite line expresses api‐SFG (green) and CLD‐EcDPCK‐mCherry (red) with DAPI marking the nuclei (blue). This parasite line was grown in the presence of 0.5 μM aTc.
4. The PfMev CLD‐EcDPCK‐mCherry‐apt Δdpck parasite line grown in the absence of aTc for 48 h. The brightness of the red fluorescence image was increased to demonstrate the lack of mCherry signal associated with the parasite.
5. The PfMev CLD‐EcDPCK‐mCherry‐apt Δdpck parasite line grown in the presence of 0.5 μM aTc and in the presence of 0.5 μM Shield1 (Shld) for 48 h.
6. The PfMev CLD‐EcDPCK‐mCherry‐apt Δdpck parasite line grown in the absence of aTc and in the presence of 0.5 μM Shield1 (Shld) for 48 h. The brightness of the red fluorescence image was increased to demonstrate the lack of mCherry signal associated with the parasite.
7. Growth curve of the PfMev CLD‐EcDPCK‐mCherry‐apt Δdpck parasite line grown either in the presence of 0.5 μM aTc and the absence 0.5 μM Shield1 (permissive condition, blue), or the absence of 0.5 μM aTc and the presence of 0.5 μM Shield1 (non‐permissive condition, red). Pantothenate was supplied at 50 nM in both conditions, and the cultures were diluted 1:5 on day 4. Treatment under non‐permissive conditions resulted in reduced parasite growth beginning at day 1 (two‐way ANOVA, followed by Bonferroni’s correction; *P < 0.05). Error bars represent the standard error of the mean from two independent experiments, each conducted in quadruplicate.

Data information: All microscopy images represent fields that are 10 µm long by 10 µm wide. Colocalization statistics are shown in Table 1.

Source data are available online for this figure.

In the non‐permissive state for both systems (aTc removed, Shield1 added), we conducted an assay, tracking parasite growth over 6 days, and only observed a ˜50% reduction in growth on day 6 (Appendix Fig S1). We reasoned that this mild growth inhibition may be due to incomplete knockdown of EcDPCK, with high expression driven by the calmodulin promoter. Thus, to potentially exacerbate the observed growth phenotype, we limited the amount of exogenously provided pantothenate from the culture medium in order to sensitize the parasites to reduced levels of DPCK. While typical parasite media contains ˜1 μM pantothenate, we found that as little as 50 nM was sufficient to allow parasites to grow at near wild‐type levels (Fig EV5). This is consistent with previous work which found the concentration of pantothenate required to support parasite growth at half the level of control parasites (SC₅₀) to be ˜9.4 nM (Tjhin et al, 2018). Under limiting pantothenate concentrations, we observed a severe growth defect for the PfMev CLD‐EcDPCK‐mCherry‐apt Δdpck parasite line under non‐permissive conditions, indicating that DPCK activity is required for parasite growth (Fig 5G).

Figure EV5. The effect of pantothenate concentration on PfMev parasite growth.

PfMev parasites were seeded at ˜0.5% parasitemia and grown in media containing different concentrations of pantothenate (Pant). Parasites grown in complete medium with AlbuMAX II (CMA) were included as a control. Parasite media was refreshed every 24 h, at which time samples were collected for analysis via flow cytometry. All parasite cultures were cut 1:5 on day 4. Cultures supplemented with 50 nM pantothenate or higher concentrations grew significantly better than unsupplemented parasites by day 4 of the experiment (two‐way ANOVA, followed by Bonferroni’s correction; *P < 0.05; ***P < 0.001). Cultures supplemented with 10 nM pantothenate were not significantly different from those without supplement over the course of the 6‐day experiment. Error bars represent the standard error of the mean from two independent experiments, each conducted in quadruplicate.

DPCK activity in disrupted apicoplast vesicles is essential for parasite growth

While the previous results demonstrated that DPCK activity is essential, it is not clear whether disrupted apicoplast vesicles remain biochemically active and if DPCK activity remains functional within these vesicles. To help elucidate this, we induced apicoplast disruption in the PfMev CLD‐EcDPCK‐mCherry‐apt Δdpck line by treating the parasites with 100 nM azithromycin in the presence of 50 μM mevalonate for 7 days, with organelle disruption confirmed via PCR and epifluorescence microscopy (Fig 6A and B). We observed by microscopy that the api‐SFG and mCherry signals colocalized, indicating that EcDPCK‐mCherry was present within the disrupted apicoplast vesicles. We then repeated the aforementioned growth experiment with this line in the presence of 50 μM mevalonate and 50 nM pantothenate, and observed a severe inhibition of parasite growth under the non‐permissive condition (Fig 6C). Taken together, these results show that EcDPCK is present and active within the disrupted apicoplast vesicles and that DPCK activity is required for parasite survival. Additionally, these results demonstrate that these vesicles remain biochemically active, as enzymes within them continue to be capable of fulfilling their essential activity.

Figure 6. Characterization of azithromycin‐treated PfMev CLD‐EcDPCK‐mCherry‐apt Δdpck parasites under permissive and non‐permissive conditions.

1. Attempted PCR detection of the ldh, sufB, and cox1 genes from the nuclear (N), apicoplast (A), and mitochondrial (M) genomes, respectively. We were unsuccessful in amplifying sufB from the PfMev CLD‐EcDPCK‐mCherry‐apt Δdpck parasite line after treatment with 100 nM azithromycin (Azith) in the presence of 50 μM mevalonate, indicating the disruption of the apicoplast organelle.
2. Live epifluorescence microscopy of the apicoplast‐negative PfMev CLD‐EcDPCK‐mCherry‐apt Δdpck line post‐treatment with 100 nM azithromycin in the presence of 50 μM mevalonate. This parasite line expresses api‐SFG (green) and the CLD‐EcDPCK‐mCherry construct (red) with DAPI marking the nuclei (blue). Images represent fields that are 10 µm long by 10 µm wide.
3. Growth curve of the PfMev CLD‐EcDPCK‐mCherry‐apt Δdpck parasite line containing a disrupted apicoplast grown either in the presence of 0.5 μM aTc and the absence 0.5 μM Shield1 (permissive condition, blue), or the absence of 0.5 μM aTc and the presence of 0.5 μM Shield1 (non‐permissive condition, red). Pantothenate (50 nM) and mevalonate (50 μM) were supplied in both conditions, and the cultures were diluted 1:5 on day 4. Treatment under non‐permissive conditions resulted in reduced parasite growth beginning at day 2 (two‐way ANOVA, followed by Bonferroni’s correction; **P < 0.01). Error bars represent the standard error of the mean from two independent experiments, each conducted in quadruplicate.
4. Upon apicoplast disruption, DPCK is located within vesicles that contain apicoplast‐specific proteins and may still be able to generate CoA.

Source data are available online for this figure.

Discussion

Plasmodiumfalciparum parasites rely on the supplementation of exogenous pantothenate, which is used by the parasite to generate CoA (Spry & Saliba, 2009). This process involves the cytosolic enzymes PanK1/2, PPCS1/2, PPCDC, and PPAT, while the apicoplast‐localized DPCK mediates the final step (Ralph et al, 2004; Spry et al, 2008). Most of these enzymes have been targeted for deletion in P. yoelli or P. berghei, and all except for the last two (PPAT and DPCK) have proven to be dispensable for blood‐stage parasite survival (Hart et al, 2014; Hart et al, 2016; Kehrer et al, 2016; Kenthirapalan et al, 2016; Srivastava et al, 2016; Hart et al, 2017). These results are largely concordant with forward genetic screens in P. berghei and P. falciparum, which suggest that DPCK is essential, and the majority of upstream enzymes in the pathway are dispensable (Schwach et al, 2015; Zhang et al, 2018) (Table 2). One notable exception is PanK1, which has been reported to be refractory to deletion in P. falciparum (Tjhin et al, 2018). The dispensability of other upstream enzymes has been attributed to the potential ability of parasites to import pantetheine, which is presumably phosphorylated by PanK1 or another kinase to generate the PPAT substrate 4′‐phosphopantetheine (Hart et al, 2017). A similar phenomenon may explain why supplementation with CoA can bypass certain pathway inhibitors (Fletcher & Avery, 2014; Fletcher et al, 2016), but did not allow us to delete DPCK. If CoA cannot be imported directly into the parasite at sufficient levels, it is possible that supplemented CoA may break down into pathway intermediates, as previously reported (Spry et al, 2010), bypassing early pathway enzymes, but not DPCK.

Table 2.

Predicted essentiality of genes based on forward genetic screens.

	P. falciparum PiggyBac (Zhang et al, 2018)			P. berghei PlasmoGem (Schwach et al, 2015)
Gene Name	Gene ID PF3D7_#	Mutagenesis Index Score	Mutagenesis Fitness Score	Gene ID PBANKA_#	Phenotype
PAT	0206200	1	−2.21	0303900	Not studieda
PanK1	1420600	0.18	−2.77	1022600	Dispensableb
PanK2	1437400	0.92	−2.93	0611400	Dispensableb
PPCS1	0412300	0.98	−2.84	0613600	Dispensable
PPCS2	1102400	1	−0.55	Absent in P. berghei
PPCDC	0816100	0.16	−2.91	0714700	Dispensable
PPAT	0704700	1	−2.7	0802400	Not studied
DPCK	1443700	0.23	−3.44	1307600	Essential
ACP	0208500	0.23	−3.47	030560	Essential
ACPS	0420200	0.23	−2.64	0521600	Not studied

^a PAT was deleted successfully in P. berghei (Kehrer et al, 2016; Kenthirapalan et al, 2016; Srivastava et al, 2016).

^b PanK1 and PanK2 were deleted successfully in P. berghei (Srivastava et al, 2016).

While the P. falciparum DPCK has been predicted to be one of the few enzymes within the pantothenate pathway that is absolutely required for parasite survival, it has not been well studied. DPCK is predicted to reside within the apicoplast, performing the essential function of CoA generation for the parasite. However, the P. falciparum DPCK had previously neither been localized nor functionally characterized, and its essentiality had not been conclusively demonstrated. Through the work outlined here, we were able to localize DPCK to the apicoplast and show that it is essential for parasite survival. Complementation experiments confirmed that the essential activity of P. falciparum DPCK is shared by the previously characterized dephospho‐CoA kinase (coaE) from E. coli (Mishra et al, 2001).

It has been shown that the apicoplast can be disrupted as long as the parasites are supplemented with IPP (Yeh & DeRisi, 2011). The disrupted apicoplast phenotype includes loss of the apicoplast genome (Gisselberg et al, 2013) and the presence of vesicles containing apicoplast‐specific proteins (Bowman et al, 2014; Pasaje et al, 2016; Florentin et al, 2017; Uddin et al, 2018). Once the apicoplast is disrupted, parasites are entirely reliant on IPP supplementation for survival (Yeh & DeRisi, 2011). From these results, it has been assumed that IPP is the only essential apicoplast product in blood‐stage parasites and that parasites in this state are “apicoplast‐negative”, without functional apicoplast proteins and pathways. However, this conclusion is hard to reconcile with the need for CoA that is generated inside the apicoplast, as a lack of CoA should not be bypassable through IPP supplementation. Our results resolve this issue by demonstrating that DPCK cannot be bypassed by IPP because the enzyme remains functional in the vesicles remaining after apicoplast disruption.

While DPCK may still be active in disrupted apicoplast vesicles, it is important to note that continued activity is not possible for all apicoplast pathways. For instance, the apicoplast‐localized MEP isoprenoid precursor pathway that is responsible for the generation of IPP (Guggisberg et al, 2014) would no longer be functional upon disruption, as it relies on factors encoded by the apicoplast organellar genome. Specifically, the MEP pathway relies on essential iron–sulfur (Fe‐S) cluster cofactors generated by the apicoplast‐localized SUF pathway, with one of the proteins within this pathway, SufB, being encoded by the organellar genome (Gisselberg et al, 2013). Thus, while some apicoplast‐localized biochemical pathways may remain functional upon organelle disruption, this would not hold true for all pathways, especially those reliant on factors encoded by the apicoplast genome.

While the exact mechanism is unclear, DPCK can likely continue to fulfill its essential metabolic role within these vesicles if transporters within the membrane of the vesicles allow for the import of substrates and export of products (Fig 6D). In T. gondii, vesicles remaining after apicoplast disruption contain apicoplast membrane proteins (Tawk et al, 2011), suggesting that apicoplast transporters could be functional in these vesicles. Additionally, an energy source in the form of ATP is required for DPCK activity, which suggests that ATP is imported or generated by an enzyme such as pyruvate kinase II (PyrKII) in the vesicles (Ralph et al, 2004; Swift et al, 2020a). One could also further speculate that the environmental conditions (pH, redox potential, ion concentration) are also maintained in the vesicles at levels compatible with enzymatic activity.

Previous experiments using the PfMev bypass line interrogated the effect apicoplast disruption has on the parasite through the transcriptomic and metabolomic analyses of parasites containing a disrupted apicoplast and maintained on mevalonate (Swift et al, 2020b). These results revealed that the transcriptomic profile of parasites in this state does not appear to be meaningfully perturbed, as the nuclear transcript levels of even apicoplast‐specific proteins are comparable to parasites containing an intact apicoplast (Swift et al, 2020b). Similarly, metabolomic experiments show that apicoplast disruption has a surprisingly small effect on metabolite levels (Swift et al, 2020b). These results suggest that hundreds of apicoplast‐specific proteins are still generated, packaged, and sent out into the cell, but due to the disruption of the organelle, accumulate in vesicles throughout the cell. While parasites containing a disrupted apicoplast have previously been referred to as “apicoplast‐negative”, we have shown that the resulting vesicles remain biochemically active.

Interestingly, the expression of the cytosolic EcDPCK successfully complemented the loss of the endogenous DPCK; parasites lacking apicoplast DPCK did not require mevalonate supplementation for growth or apicoplast maintenance. This suggests that CoA is not required within the apicoplast during the blood stage of the parasite. Inside the apicoplast, CoA is used to generate acetyl‐CoA and malonyl‐CoA (Prigge et al, 2003; Foth et al, 2005; Dellibovi‐Ragheb et al, 2018) that feed into the FASII pathway, which has previously been shown to be dispensable in blood‐stage parasites (Spalding & Prigge, 2008; Yu et al, 2008; Vaughan et al, 2009; Pei et al, 2010). CoA is also predicted to be used by acyl‐carrier protein synthase (ACPS) to modify ACP with a 4′‐phosphopantetheine modification, with the modified ACP playing a key role in the FASII pathway (Waters et al, 2002; Ralph et al, 2004; Spry et al, 2008; Gallagher & Prigge, 2010; Shears et al, 2015). While the only predicted function of ACP and ACPS within the apicoplast of P. falciparum parasites involves their role in the FASII pathway, both ACP and ACPS appear to be essential according to forward genetic screens in Plasmodium (Table 2) (Schwach et al, 2015; Zhang et al, 2018). Thus, if CoA is indeed required within the apicoplast, it is possible that the CoA generated within the cytosol by EcDPCK is imported back into the apicoplast. Presumably, there is at least one transporter that brings dephospho‐CoA into the apicoplast and exports CoA during the normal operation of the apicoplast. This transporter may either be sufficiently nonspecific or possess the ability to work in reverse to bring cytosolic CoA into the organelle, thus allowing the production of CoA in the cytosol to fulfill any potential requirements within the apicoplast.

Overall, the work here demonstrates that DPCK is an essential apicoplast‐localized protein that is required for parasite survival and that its function, generating CoA, cannot be bypassed with IPP. We have also shown that the vesicles derived from the disruption of the apicoplast continue to be biochemically active and capable of fulfilling essential metabolic requirements within the parasite.

Materials and Methods

P. falciparum culture and maintenance

Unless otherwise noted, blood‐stage P. falciparum parasites were cultured in human erythrocytes at 1% hematocrit in a 10 ml total volume of CMA (Complete Medium with AlbuMAX II) containing RPMI 1640 medium with l‐glutamine (USBiological Life Sciences), supplemented with 20 mM HEPES, 0.2% sodium bicarbonate, 12.5 μg/ml hypoxanthine, 5 g/l AlbuMAX II (Life Technologies), and 25 μg/ml gentamicin. Cultures were maintained in 25‐cm² gassed flasks (94% N₂, 3% O₂, 3% CO₂) and incubated at 37°C.

Generating markerless PfMev P230p‐attB parasites

A 1.2 kB region of the p230p gene (PF3D7_0208900) was amplified from PfMev genomic DNA using primers P230p.HA.F and P230p.HA.R (Appendix Table S2). These primers were designed to contain ˜15 bp overhangs which allowed insertion of the amplicon into the NgoMIV site of pRS (Swift et al, 2020b) by ligation independent cloning (In‐Fusion, Clontech). The plasmid was digested with BglII to excise a 210bp fragment between two BglII sites in the p230p gene. The excised region was replaced with an oligo comprised of complementary primers attBr. InF.F and attBr. InF.R (Appendix Table S2) using In‐Fusion to generate an attB site flanked by p230p homology arms. The plasmid was then digested with BsaI to insert a segment of DNA encoding a guide RNA targeting the excised region of p230p. Complementary primers P230p.gRNA.F and P230p.gRNA.R (Appendix Table S2) were annealed and inserted using ligation independent cloning with In‐Fusion (Clontech). The resulting plasmid pRS‐P230p was used along with pCasG (Rajaram et al, 2020) in transfections for markerless insertion of the attB element into the P230p locus of PfMev parasites. After 48 h, transfectants were selected with 1.5 μM DSM1 for 7 days and 2.5 nM WR99210 for 10 days. Parasite clones were characterized by genotyping PCRs (Fig 2A) using primers described in Fig EV1 and Appendix Table S2.

Generation of the PfMev EcDPCK complementation lines

The E. coli complementation constructs were generated from the p15‐Mev‐aSFG plasmid (GenBank: MN822298) containing a bidirectional Cam/HOP promoter (Swift et al, 2020b). On one side of the promoter, a drug resistance marker can be inserted into the BamHI/HindIII sites; on the other side, transgenes can be inserted into the AvrII/AflII sites. The plasmid also contains an attP site for integration into PfMev P230p‐attB parasites (Nkrumah et al, 2006; Spalding et al, 2010). Unless noted, the construction of the complementation plasmids used ligation independent cloning with In‐Fusion (Clontech).

The hDHFR drug resistance cassette was amplified from the pRS‐LacZ plasmid (GenBank: MN822297) (Swift et al, 2020b) using the primers listed in Appendix Table S2 (hDHFR.F and hDHFR.R) and inserted into the BamHI/HindIII sites of p15‐Mev‐aSFG to generate p15‐hDHFR. The sequence encoding E. coli DPCK (EcDPCK) was amplified from E. coli genomic DNA using the primers in Appendix Table S2 (EcDPCK.F and EcDPCK.R) and inserted into the AvrII/BsiWI sites of the pLN‐TP‐ACP‐mCherry plasmid (Gisselberg et al, 2013) to generate pLN‐EcDPCK‐mCherry. The entire EcDPCK‐mCherry region was then amplified from this plasmid using the primers in Appendix Table S2 (CamEcDpmCry.F and CamEcDpmCry.R) and inserted into the AvrII/AflII sites of p15‐hDHFR to generate p15‐EcDPCK‐mCherry.

For the generation of the apicoplast‐localized E. coli DPCK‐mCherry protein, the pLN‐TP‐ACP‐mCherry plasmid (Gisselberg et al, 2013) was cut with AvrII/BsiWI. The N‐terminal signal and transit peptide, corresponding to the first 55 amino acids of the P. falciparum ACP protein (Api55), was amplified from p15‐Mev‐aSFG using the primers in Appendix Table S2 (pLN. Api55.F and pLN. Api55.R) and inserted upstream of the mCherry sequence to form the pLN‐Api‐mCherry plasmid. The sequence encoding EcDPCK was then amplified from E. coli genomic DNA using the primers in Appendix Table S2 (Api. EcDPCK.F and Api. EcDPCK.R) and inserted into the BsiWI/BspEI sites of the previous plasmid. The resulting sequence, comprised of Api55, EcDPCK, and mCherry, was amplified from the pLN‐Api‐EcDPCK‐mCherry plasmid using the primers in Appendix Table S2 (CamApEcDpmC.F and CamApEcDpmC.R) and then inserted into the AvrII/AflII sites of p15‐hDHFR to generate p15‐Api‐EcDPCK‐mCherry.

The PfMev P230p‐attB parasite line was transfected with either the p15‐EcDPCK‐mCherry or p15‐Api‐EcDPCK‐mCherry plasmids along with the pINT plasmid encoding the bxb1 integrase (Nkrumah et al, 2006; Spalding et al, 2010). Integrants were selected for with 2.5 nM WR99210 for 7 days, after which point drug pressure was removed. Infected RBCs were then observed ˜20–25 days post‐transfection, at which point 2.5 nM WR99210 was added back, with the parasites maintained in the presence of this drug. Parasite lines were characterized by genotyping PCR (Fig EV2A and B) using primers listed in Appendix Table S2.

Generation of the P. falciparum DPCK localization line

The gene encoding P. falciparum DPCK (PF3D7_1443700) was amplified from cDNA using the primers (DPCK.loc.F and DPCK.loc.R) listed in Appendix Table S2 and inserted into the AvrII/BsiWI sites of p15‐Api55‐EcDPCK‐mCherry to generate p15‐PfDPCK‐mCherry. This plasmid was transfected into the PfMev P230p‐attB parasite line along with pINT (Nkrumah et al, 2006; Spalding et al, 2010). Parasites were selected and validated as described above.

Live cell epifluorescence microscopy

Approximately 100 μl of resuspended parasite cultures was incubated with 1 µg/ml 4′, 6‐diamidino‐2‐phenylindole (DAPI) for 30 min at 37°C. Cells were then washed three times with 100 μl of CMA medium and incubated for 5 min at 37°C after each wash. Cells were resuspended in 20 μl of CMA and then pipetted onto slides and sealed with wax for observation on a Zeiss AxioImager M2 microscope. A series of images spanning 5 µm in the z‐plane were acquired with 0.2 µm spacing, and images were deconvolved using the VOLOCITY software (PerkinElmer) to report a single image in the z‐plane.

Generation of P. falciparum plasmid constructs for gene deletion

DPCK was targeted for deletion through Cas9‐mediated gene editing, using the pL8 plasmid (Swift et al, 2020b), in combination with the pUF1‐Cas9 plasmid, which was generously provided by Dr. Jose‐Juan Lopez‐Rubio (Ghorbal et al, 2014). To generate pL8‐DPCK, two DPCK homology arms (HA1 and HA2) and a 20 bp guide RNA were inserted into pL8. HA1 (562bp) was amplified with primers DPCK.HA1.F and DPCK.HA1.R, and HA2 (393bp) was amplified with primers DPCK.HA2.F and DPCK.HA2.R (Appendix Table S2). HA1 was inserted into the NotI site, and HA2 was inserted into the NgoMIV site of pL8 using In‐Fusion. Primers DPCK.gRNA.F and DPCK.gRNA.R (Appendix Table S2) were annealed and inserted into the BtgZI sites of pL8 to form pL8‐DPCK. For deletion of DPCK in the EcDPCK complementation lines, the gene encoding hDHFR in pL8‐DPCK was replaced by a codon harmonized gene encoding residues 2–129 of the Aspergillus terreus blasticidin‐S deaminase (BSD) synthesized by LifeSct. The harmonized bsd gene (Appendix Fig S2) was amplified using the primers pRsBSD.F and pRsBSD.R (Appendix Table S2) and inserted into the BamHI/HindIII sites of pL8‐DPCK using ligation independent cloning (In‐Fusion, Clontech).

P. falciparum transfections for attempted gene deletion of DPCK

Transfections were conducted as previously described (Spalding et al, 2010). Briefly, 400 μl of RBCs was electroporated with 75 μg each of the Cas9 expression plasmid and the pL8‐DPCK homology repair plasmid. The transfected RBCs were then mixed with synchronized schizont‐stage PfMev parasites. After ˜48 h, drug‐selection was initiated by the addition of 1.5 μM DSM1, 2.5 nM WR99210, and 50 µM mevalonate, with or without 5 mM CoA.

Generation of the PfMev CLD‐EcDPCK‐mCherry‐apt parasite line

In order to generate the p15‐CLD‐EcDPCK‐mCherry‐apt plasmid, the p15‐Api‐EcDPCK‐mCherry plasmid was digested with AvrII and BspEI in order to remove the ACP signal and transit peptide sequence. The conditional localization domain sequence was then amplified from the pLN‐CLD2‐HCS1 plasmid (Roberts et al, 2019) using the primers listed in Appendix Table S2 (CLD.F and CLD.R). The CLD‐EcDPCK‐mCherry gene sequence was then amplified using the CLD.F and EcDPCK.CLD.R primers (Appendix Table S2), and inserted into the p15‐aFluc‐mCh plasmid (Swift et al, 2020a), previously digested with AvrII and PspOMI, using In‐Fusion cloning. See Appendix Fig S3 for plasmid sequence. This plasmid, along with the pINT plasmid encoding the bxb1 integrase for attP/attB integration, was transfected into p230p attB PfMev parasites and selected with 2.5 µg/ml blasticidin for 7 days, after which drug pressure was removed. Parasites were observed via Giemsa stain 20–25 days later, at which point the parasites were again cultured in the presence of 2.5 µg/ml blasticidin.

Deletion of DPCK in the PfMev EcDPCK‐mCherry, PfMev api‐EcDPCK‐mCherry, and PfMev CLD‐EcDPCK‐mCherry‐apt lines

The PfMev EcDPCK‐mCherry and PfMev api‐EcDPCK‐mCherry lines were transfected with the pL8‐DPCK (BSD) plasmid along with the pUF1‐Cas9 plasmid. The parasites were selected with 2.5 µg/ml blasticidin and 1.5 μM DSM1 for 7 days, also in the presence of 2.5 nM WR99210. After 7 days, blasticidin and DSM1 were removed. Infected RBCs were then typically seen after ˜20–25 days. After parasites were observed, the cultures were then maintained in the presence of 2.5 nM WR99210 and 2.5 µg/ml blasticidin.

For the deletion of DPCK in the PfMev CLD‐EcDPCK‐mCherry‐apt parasite line, the original pL8‐DPCK plasmid containing the DPCK homology arms flanking the hDHFR drug selectable marker was used. Parasites were transfected with the pL8‐DPCK (hDHFR) plasmid along with the pUF1‐Cas9 plasmid. Parasites were then selected with 2.5 nM WR99210 and 1.5 μM DSM1 for 7 days, also in the presence of 2.5 µg/ml blasticidin. After 7 days, WR99210 and DSM1 were removed. Infected RBCs were then typically seen after ˜20–25 days. After parasites were observed, the cultures were then maintained in the presence of 2.5 nM WR99210 and 2.5 µg/ml blasticidin. Parasites were continuously cultured in the presence of 0.5 µM aTc.

Confirmation of knockout genotype

Primers were designed to screen for 5’ integration (Δ5’ reaction primers DPCK.5.F and pL8HA1.R) and 3’ integration (Δ3’ reaction primers pL8HA2.F and DPCK.3.R) of the gene disruption cassette, and the 5’ region (primers DPCK.5.F and DPCK.5.WT.R) and 3’ region (primers DPCK.3.WT.F and DPCK.3.R) of the WT gene (Appendix Table S2). The parental PfMev line was used as a control for these reactions.

Generation and purification of anti‐mCherry rabbit antibodies

Recombinant mCherry was produced to generate affinity‐purified antibodies. Primers mCh.pMAL. EcoRI. For and mCh.pMAL. HindIII. Rev (Appendix Table S2) were used to amplify mCherry for insertion into the pMALcHT E. coli expression vector (Muench et al, 2003). mCherry was expressed and purified using the same protocol previously described for the purification of recombinant GFP (Roberts et al, 2019). Pure recombinant mCherry was used to generate rabbit antiserum using the custom antibody service of Cocalico Biologicals Inc. Briefly, 250 µg of mCherry mixed with Complete Freund’s Adjuvant was used for the initial inoculation followed by boosts of 125 µg of antigen 2, 3, and 7 weeks later. Final exsanguination was performed on day 56. Specific antibodies were purified from antiserum with a mCherry affinity column using previously described methods (Roberts et al, 2019). A total of 18.6 mg of anti‐mCherry IgG was concentrated to 3.1 mg/ml and stored at −80°C in storage buffer (PBS, 40% glycerol, 0.02% NaN₃).

Western blotting

Parasite samples were centrifuged at 500 g for 5 min at room temperature (RT), and pellets were stored at −20°C until the time of protein extraction. To isolate parasites from RBCs, the pellets were thawed and resuspended in 0.15% saponin for 5 min at RT. Lysed RBCs were removed by washing three times with PBS. Parasite pellets were resuspended in NuPAGE LDS sample buffer (Thermo Fisher) containing 2% β‐mercaptoethanol and incubated at 95°C for 5 min. Proteins were resolved by SDS–PAGE on 4–12% gradient gels and transferred to nitrocellulose membranes. To detect EcDPCK‐mCherry constructs, membranes were blocked in 5% milk and probed overnight at 4°C with 1:10,000 rabbit anti‐mCherry (vide supra). They were then incubated for an hour at RT with 1:10,000 donkey anti‐rabbit HRP‐linked secondary antibodies (GE healthcare, NA934). Protein bands were detected on X‐ray film using SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific), according to the manufacturer’s protocol. Membranes were stripped of antibody with 200 mM glycine (pH 2.0) for 5 min and probed with 1:2,500 rat anti‐HA mAb 3F10 (Roche) in order to detect api‐SFG which contains a C‐terminal HA tag (Swift et al, 2020b). After incubation with 1:5,000 goat anti‐rat HRP‐linked secondary antibodies (GE healthcare, NA935), proteins were detected as described above.

Growth curve for the PfMev CLD‐EcDPCK‐mCherry‐apt Δdpck line

A culture of PfMev CLD‐EcDPCK‐mCherry‐apt Δdpck parasites was washed four times with 10 ml of CMA in order to remove aTc. Parasites were then used to seed two separate 25‐cm² flasks, each at 1% hematocrit in a 10 ml total volume of CMA media, and a starting parasitemia of ˜0.5%. One flask was supplemented with 0.5 µM aTc, representing the permissive condition, while the other was treated with 0.5 µM Shield1, representing the non‐permissive condition. The appropriate media were replaced approximately every 24 h, at which point a blood‐film was made, and the parasitemia was determined via Giemsa stain. On day 4 of this growth curve, the parasitemia was diluted 1:10.

Pantothenate titration growth curve

Pantothenate‐free media was generated, consisting of RPMI 1640 media with L‐glutamine, and without pantothenate (USBiological Life Sciences #R8999‐01A), supplemented with 20 mM HEPES, 0.2% sodium bicarbonate, 12.5 μg/ml hypoxanthine, 5 g/l AlbuMAX II (Life Technologies), and 25 μg/ml gentamicin. PfMev parasites were washed with 10 ml pantothenate‐free media three times in order to remove the preexisting pantothenate from the culture. Parasites were then seeded into a 96‐well plate at a starting parasitemia of 0.5, 2% hematocrit, and a total volume of 250 µl per well. Pantothenate was then added back to the wells containing pantothenate‐free medium at a concentration of 0, 10, 50, 250 nM, or 1 μM, with CMA as a control. The appropriate media were replaced approximately every 24 h, at which point samples were collected for analysis via flow cytometry using previously described methods (Swift et al, 2020b). On day 4 of this growth curve, the parasitemia was diluted 1:5. In Fig EV5, error bars represent the standard error of the mean from two independent experiments, each conducted in quadruplicate.

Growth curves for the PfMev CLD‐EcDPCK‐mCherry‐apt Δdpck line under limiting pantothenate concentrations

The PfMev CLD‐EcDPCK‐mCherry‐apt Δdpck line was washed with 10 ml pantothenate‐free media four times in order to remove aTc and pantothenate. Parasites were then seeded into a 96‐well plate (Corning) at a 0.5% starting parasitemia, 2% hematocrit, and a total volume of 250 µl per well in quadruplicate. Pantothenate was added back to the pantothenate‐free medium to yield a final concentration of 50 nM. Two conditions were tested: The parasites were either supplemented with 0.5 µM aTc, representing the permissive condition, or with 0.5 µM Shield1, representing the non‐permissive condition. The appropriate media were replaced approximately every 24 h, at which point samples were collected for analysis via flow cytometry using previously described methods (Swift et al, 2020b). On day 4 of this growth curve the parasitemia was diluted 1:5. In Figs 5G and 6C, error bars represent the standard error of the mean from two independent experiments, each conducted in quadruplicate.

Inducing apicoplast loss in the PfMev api‐EcDPCK‐mCherry Δdpck and PfMev CLD‐EcDPCK‐mCherry‐apt Δdpck parasite lines

The PfMev CLD‐EcDPCK‐mCherry‐apt Δdpck line was cultured in media containing 100 nM azithromycin in the presence of 50 μM mevalonate for 7 days. After 7 days, the parasites were cultured continuously in the presence of 50 μM mevalonate. Loss of the apicoplast was confirmed via PCR using the methods described below.

Confirmation of apicoplast loss

The presence of the apicoplast organellar genome was detected by PCR using primers specific for the sufB gene (SufB.F and SufB.R). Control PCRs amplified the lactate dehydrogenase (ldh) gene from the nucleus (LDH.F and LDH.R) and cytochrome c oxidase subunit 1 (cox1) from the mitochondrial genome (Cox1.F and Cox1.R) (Appendix Table S2). For each PCR, 1 µl of parasite culture was added to a 50 µl reaction volume. The parental PfMev line was used as a positive control for apicoplast genome detection.

Author contributions

RPS and STP arranged the figures and wrote the manuscript, with input from all listed coauthors. RPS, HBL, and KR carried out the experiments displayed in this manuscript.

Conflict of interest

The authors declare that they have no conflict of interest.

Supporting information

Review Process File

Appendix

Expanded View Figures PDF

Source Data for Expanded View

Source Data for Figure 2

Source Data for Figure 3

Source Data for Figure 4

Source Data for Figure 5

Source Data for Figure 6

The EMBO Journal (2021) 40: e107247.

Data availability

This study includes no data deposited in external repositories.