Summary
The deadly malaria parasite, Plasmodium falciparum, contains a non-photosynthetic plastid known as the apicoplast, that functions to produce essential metabolites. Little is known about its biology or regulation, but drugs that target the apicoplast are clinically effective. Using phylogenetic analysis, we identified a putative complex of clp (caseinolytic protease) genes. We genetically targeted members of this apicoplastregulatory complex and generated conditional mutants of the PfClpC chaperone and PfClpP protease and found that they co-localize in the apicoplast. Conditional inhibition of the PfClpC chaperone resulted in growth arrest and apicoplast loss, and was rescued by addition of the essential apicoplast-derived metabolite, IPP. Using a double conditional-mutant parasite line, we discovered that the chaperone activity is required to stabilize the active protease, revealing functional interactions. These data demonstrate the essential function of PfClpC in maintaining apicoplast integrity and its role in regulating the proteolytic activity of the Clp complex.
Introduction
Malaria is a devastating human disease caused by obligate intracellular parasites of the genus Plasmodium. This disease results in nearly 450,000 deaths each year, which are mostly caused by one species, Plasmodium falciparum (World Health Organization, 2015). The parasite has gained resistance to all clinically available antimalarial drugs, generating an urgent need to identify new drugs and potential novel drug targets (Hovlid and Winzeler, 2016; Wells et al., 2015). The parasite cell is remarkably complex with two organelles that carry their own genetic material, the mitochondrion and a unique algal endosymbiont known as the apicoplast (McFadden, 1996). The apicoplast harbors essential metabolic pathways that are required for parasite growth and survival (van Dooren and Striepen, 2013). Importantly, drugs that target cellular processes in the apicoplast are clinically effective (Erica L Dahl, 2007; Fichera and Roos, 1997; Goodman et al., 2007). Therefore, understanding the function, structure and biogenesis of the apicoplast provides a rich vein of antimalarial drug targets. One potential class of such targets are the caseinolytic protease (Clp) family of proteins that act as key regulators of the biology of bacterial cells, the evolutionary ancestors of the apicoplast. In bacteria and plant chloroplasts, Clp proteins play vital roles in cell/ organelle division, segregation, protein homeostasis and protein transport (Frees et al., 2014; Nishimura and van Wijk, 2015). Typically, they form a regulated proteolytic complex in which a Clp protease is paired with a Clp chaperone that has a AAA+ ATPase domain (also known as the Hsp100 family of chaperones) such as ClpC or ClpA (Olivares et al., 2016). There are several putative clp genes encoded in the P. falciparum genome which have been localized to the apicoplast (Bakkouri et al., 2010), and previous studies using recombinant proteins described structural features of the Clp protease and its inactive subunit (Bakkouri et al., 2013; Rathore et al., 2010). Currently, nothing is known about the apicoplast-localized Clp chaperones, and their roles in vivo as well as the interactions between the apicoplast Clp proteins remain poorly understood due to the challenging genetics of P. falciparum.
In this study, we generated conditional mutants of the P. falciparum apicoplast-targeted pfclpc and pfclpp genes and found that they localize to the apicoplast. Conditional inhibition of the PfClpC chaperone resulted in growth arrest, morphological defects, and apicoplast loss. In a series of cellular assays, we showed that PfClpC is required for apicoplast sorting into daughter cells. Addition of IPP, an essential apicoplast-derived metabolite, rescued growth, indicating that the only essential function of PfClpC is linked to the apicoplast. Using a double conditional-mutant pfclpc; pfclpp parasite line, we discovered that PfClpC activity is required to stabilize PfClpP in its active-protease form, revealing a functional interaction between the two. Our work demonstrates the essential function of PfClpC in maintaining apicoplast integrity and segregation and its role in regulating the proteolytic activity of the Clp complex.
Results
Phylogenetic analysis of Plasmodium Clp Proteins
In order to identify Clp proteins that potentially form a proteolytic complex and compare them to other organisms, we performed phylogenetic analyses of putative Clp proteases and chaperones (Figure 1A,B). We confirmed that PfClpP is the only ClpP protease in the P. falciparum plastid, unlike other apicomplexans, which encode several copies (Figure 1A). PfClpP is closely related to ClpP of plants and cyanobacteria, in agreement with the origin of the apicoplast in the secondary endosymbiosis of a red alga. Importantly, ClpP homologs are present throughout eubacterial kingdom, including many pathogenic species, which has led to the development of potent specific inhibitors (Böttcher and Sieber, 2008; Brötz-Oesterhelt and Sass, 2014; Conlon et al., 2013; Gersch et al., 2015; Hackl et al., 2015). The genome of P. falciparum also encodes PfClpR, a ClpR homolog, which is related to ClpP but lacks catalytic residues. PfClpR is grouped with ClpR orthologues from cyanobacteria and plants, indicating that it did not result from a recent gene duplication event (Figure 1A). Phylogenetic analysis of Plasmodium Clp chaperones assigns each one of them to a different subfamily, such as ClpB1, ClpB2, ClpC and ClpM (Figure 1B). Among those, there is a single ClpC orthologue, PfClpC, which is the only Clp chaperone with a predicted ClpP binding motif. It falls into a clade together with plants, cyanobacteria, and Haemosporidia, a sub-class of apicomplexans. Other apicomplexans such as the closely related parasite Toxoplasma gondii, do not encode ClpC homologs, but contain ClpM and multiple duplications of ClpB1 and ClpB2. This suggests variations in the Clp complexes within apicomplexans, and that PfClpC may have a Plasmodium specific function.
Generating conditional mutants for PfClpC and PfClpP
Based on this analysis, we took a genetic approach to dissect the biological roles of the Plasmodium putative Clp complex, consisting of PfClpP, PfClpR, and PfClpC. To study the function of the chaperone PfClpC, we inserted a DHFR-based destabilizing domain (DDD) into the pfclpc genomic locus, a technique used for conditional auto-inhibition of chaperone function (Beck et al., 2014; Muralidharan et al., 2012). In the chaperone-DDD fusion protein, the unfolded DDD binds the chaperone intra-molecularly, thereby excluding client proteins and inhibiting normal chaperone function (Figure 1C). A small molecule ligand, trimethoprim (TMP) is used to stabilize and refold the DDD, releasing the chaperone to resume its normal function (Figure 1C). Using a single-crossover homologous recombination approach, we tagged the C-terminus of the pfclpc gene with a triple-HA tag and the DDD to produce the PfClpC-DDD parasite line (Figure 1D). Two clones from independent transfections were isolated (1G8 and 2E10) and integration was confirmed by Southern blot analysis (Figure 1D). Using an anti-HA antibody, we detected expression of the PfClpC-DDD fusion protein at the expected molecular size in tagged clones but not in the parental line (Figure 1E). As predicted by the chaperone auto-inhibition model, TMP removal did not lead to PfClpC degradation (Figure 1F).
To study PfClpP and PfClpR, we used a translational inhibition technique, involving the glmS ribozyme. In this method, a glmS sequence is inserted into the genomic locus before the 3’UTR of the gene, after the stop codon (Prommana et al., 2013). Addition of the small molecule glucosamine (GlcN) results in its conversion inside the cell to glucosamine 6-phosphate which activates the glmS ribozyme. The now-active ribozyme cleaves itself from the mRNA, leading to transcript degradation. Using CRISPR/Cas9, we appended a triple-V5 tag to the C-terminus of PfClpP and PfClpR followed by the glmS sequence (Figure 2A). We successfully modified the pfclpp genomic locus and obtained two independent parasites lines, PfClpP-glmS1 and PfClpP-glmS2, in which correct integration was verified by genomic PCR (Figure 2B). Using an anti-V5 antibody, we confirmed expression, and observed that PfClpP is proteolytically processed and appears as a full-length inactive-zymogen (I), apicoplast-localized zymogen without the transit peptide (II) and an active protease after cleavage of the pro-domain (III) (Figure 2C)(Bakkouri et al., 2010; Rathore et al., 2010). Upon addition of GlcN, PfClpP protein levels were reduced by 60% (Figure 2D). The same strategy was used for tagging pfclpr locus, however integration was inefficient, and multiple attempts to clone mutants from mixed populations via limiting dilution failed, suggesting that PfClpR is essential for asexual growth (Figure S1).
To detect the sub-cellular localization of PfClpC and PfClpP we used Immunofluorescence microscopy involving staining for known apicoplast resident proteins. We were able to determine the apicoplast localization of PfClpC by staining with anti-HA and anti-acyl carrier protein (ACP, an apicoplast marker (Waller et al., 1998)) and PfClpP by staining with anti-V5 and anti-Cpn60 (homolog of the T. gondii apicoplast chaperone Cpn60 (Agrawal et al., 2009)) (Figure 3A).
PfClpC is essential for asexual growth of parasites
To test the requirement of PfClpC for asexual replication, we removed the stabilizing ligand (TMP) and monitored the growth of PfClpC-DDD parasites. We did not see a strong effect initially, and since P. falciparum asexual life cycle takes 48 hours, we concluded that that these parasites develop normally for the first two or three growth cycles. However, a significant growth defect was observed seven days after TMP withdrawal and eventually lead to a severe growth arrest in these parasites (Figure 3B). We found that inhibition of PfClpC activity relies on TMP in a dose-dependent manner with an EC50 of 80nM (Figure 3C). Addition of GlcN to PfClpP-glmS parasites did not affect asexual growth (Figure 3D), most likely due to insufficient reduction in PfClpP levels (Figure 2D). While we could not determine the effect of PfClpP inhibition, we concluded that PfClpC activity is essential for parasite survival and asexual growth within human red blood cells.
PfClpC inhibition leads to a reduced replication rate within the red blood cells
Further characterization of PfClpC mutants revealed that they developed normally during the early stages of rings and trophozoites (Figure S2A). Late schizont stages (≥ 6 nuclei), however, developed aberrant morphology, including irregular cellular shape, empty vacuoles and fewer nuclei, suggesting that they are nonviable (Figure 4A). These morphologically abnormal parasites appeared on the 3rd replication cycle and their fraction increased over time (Figure 4B). Analysis of synchronized late-stage parasites over several replication cycles using flow cytometry revealed that instead of a typical single peak, they had a wider distribution, suggesting variation in DNA content (Figure 4C). To test the replication efficiency of these mixed populations of parasites, we monitored the rate of schizonts to rings conversion. We found that TMP removal resulted in a significant decrease in the numbers of ring-stage parasites that were formed in each successive generation (Figure 4D). This reduced replication rate accounts for the observed growth inhibition as well as the increase in the numbers of morphologically abnormal parasites with each replication cycle.
PfClpC activity is required for apicoplast integrity
To test a possible effect of PfClpC inhibition on the apicoplast, we removed TMP and performed immunofluorescence microscopy assays and observed a loss of canonical apicoplast morphology. Instead, we detected PfClpC, as well as the apicoplast resident protein ACP, in a punctate, vesicle-like pattern, suggesting that apicoplast integrity was compromised (Figure 4E). Apicoplast targeting of nuclear-encoded proteins is mediated through an N-terminal transit peptide that is cleaved upon localization to the apicoplast (Foth et al., 2003; Waller, 2000). In line with the loss of apicoplast localization, a second higher band for PfClpC appeared upon TMP removal, indicative of a cytoplasmic fraction of this protein in which the transit peptide was not cleaved (Figure S2B).
Chemical rescue of PfClpC-DDD parasites using IPP
The only essential function of the apicoplast during the blood stages is the biosynthesis of isopentenyl-pyrophosphate (IPP), the precursor for all isoprenoids, through the non-mevalonate (or MEP) pathway (Yeh and DeRisi, 2011). To test the effect of IPP on PfClpC inhibition, we removed TMP and added IPP to PfClpC-DDD parasites. We were able to observe complete restoration of normal growth as well as typical cellular morphology (Figure 5A,B). Immunofluorescence microscopy revealed, however, that IPP-treated PfClpC-DDD parasites survived in the absence of a functional apicoplast and retained the vesicle-like structures containing PfClpC and ACP (Figure 5C). To further investigate the fate of the apicoplast in PfClpC-DDD parasites, we performed quantitative Real Time PCR (qRT-PCR) experiments, comparing the nuclear, plastid and mitochondrial genomes over subsequent replication cycles without TMP. This qRT-PCR analysis revealed a significant reduction in the plastid genome, whereas the mitochondrial genome was unaffected (Figure 5D). This indicates that in addition to a functional damage to the apicoplast, there was an actual loss of the plastid genome. Overall, we concluded that the only essential activity of PfClpC is linked to apicoplast function.
An apicoplast sorting defect in PfClpC-DDD parasites
The reduced replication rate upon loss of PfClpC function could be indicative of a problem in apicoplast segregation into daughter cells. To test this, we aimed to visualize apicoplast presence or absence in early stage parasites (≤ 5 hours post-invasion). We used imaging flow cytometry to analyze these parasites and found that PfClpC was present in all parasites (Figure S3A and Table S1). Higher resolution microscopy confirmed that these early stage parasites contained PfClpC and it was present in vesicle-like punctate pattern rather than in the canonical apicoplast structure (Figure S3B). This suggests that either young parasites inherit these vesicles from their mother cell, or that these vesicles represent very early de novo synthesis of these proteins. Either way, nuclear-encoded apicoplast proteins are expressed and present in the cell, even in the earliest stages, regardless of apicoplast presence or absence.
We therefore employed a functional assay to differentiate between parasites containing an apicoplast and parasites that have lost it, relying on the fact that the plastid must be inherited and cannot be synthesized de novo. We removed TMP and allowed PfClpC-DDD parasites to grow in the presence or absence of IPP for two weeks. As expected, PfClpC-DDD parasites grown without TMP and supplemented with IPP grew normally, whereas parasites incubated without TMP and without IPP were unable to grow and were undetectable for several days (Figure 5E, left). On day 14 we removed IPP and relieved PfClpC inhibition by adding back TMP, and monitored the growth of the parasites. Upon adding back TMP to the media, the parasites that were grown without IPP recovered and resumed normal growth, indicating that a small fraction of inhibited parasites still possessed a functional apicoplast (Figure 5E, right). Conversely, parasites that were grown with IPP started dying 48 hours after removing IPP and adding back TMP (Figure 5E, right). These parasites lost their apicoplast but used IPP to survive, and therefore died despite restoration of PfClpC activity. Overall, these data indicate that PfClpC-DDD mutants lose the apicoplast from most parasites, with the exception of a small population that retains the plastid, suggesting a sorting defect.
PfClpC is required to stabilize PfClpP active protease form
Since PfClpC is the only Clp chaperone in the apicoplast with a ClpP binding motif (Figure 1B) (Bakkouri et al., 2010; Kim et al., 2001), we were interested in possible interactions between the chaperone and the other Clp complex subunits. Multiple attempts to immunoprecipitate PfClpC or PfClpP in their native-state from parasite-cell lysates have failed, probably due to the compact nature of the Clp complex that renders the epitope tags inaccessible to the antibodies. We also explored the effect of a previously described bacterial ClpP inhibitor, U1-lactone, that was shown to inhibit PfClpP in vitro (Rathore et al., 2010). We observed, however, that inhibition of parasite growth by U1-lactone could not be rescued by IPP, indicating that in vivo it is not specific for PfClpP and has at least one target outside of the apicoplast (Figure S4). Therefore, to explore functional interactions between PfClpC and PfClpP, we generated pfclpc; pfclpp double-conditional mutants. In these parasites, PfClpC is tagged with HA and is controlled by TMP removal and PfClpP is tagged with V5 and is controlled by addition of GlcN (Figure 6A,B). These parasites responded to PfClpC inhibition but not to partial reduction of PfClpP, similar to the single mutants (Figure 6C). Interestingly, PfClpC inhibition affected PfClpP localization, and it appeared in vesicle-like structures together with PfClpC (Figure 6D). This was accompanied with accumulation of the of the full-length PfClpP-zymogen, as seen by the appearance of the band with the highest apparent molecular weight on an SDS-PAGE (I) (Figure 6E). These observations were made seven days or more after TMP removal, and were consistent with the growth arrest and apicoplast loss in these parasites. Unexpectedly, the second processing event of PfClpP, i.e. removal of the pro-domain to produce the active protease, was also inhibited upon chaperone inactivation, but with much faster kinetics (Figure 6E). One day after TMP removal we observed significant reduction in the active protease form (III), and accumulation of the inactive zymogen (II) (Figure 6E). Analysis at shorter time points revealed a rapid decrease in the active form of PfClpP within 4 hours after TMP removal (Figure 6F). This rapid effect suggests that PfClpC activity is required to stabilize PfClpP in its active form.
Discussion
The deadly malaria parasite, Plasmodium falciparum, is a eukaryotic pathogen and as such, it shares conserved basic biology with its human host. It is therefore both challenging and essential, in the search for potential drug targets, to identify key components that are absent or significantly different from the human host.
One such class of potential candidates is the apicoplast-associated prokaryotic Clp family of chaperones and proteases. In the bacterial ancestors, as well as in other organellar descendants such as the mitochondria and chloroplast, these proteins serve a variety of basic molecular functions ranging from protein degradation, transport across membranes, protein folding, cell division, stress response and pathogenicity (Frees et al., 2014). Little is known, however, about the physiological roles of the apicoplastresident Clp proteins in the biology of Plasmodium falciparum.
In this study, we performed a thorough phylogenetic analysis, and identified the Clp proteins that potentially form a proteolytic complex in the apicoplast of P. falciparum. These include the protease PfClpP, the inactive subunit PfClpR, and PfClpC, a AAA+ chaperone with a unique ClpP binding motif (Figure 1). This analysis revealed homology to the Clp proteins of several pathogenic bacteria as well as to those that are in plants chloroplasts, but not necessarily to other apicomplexan parasites such as Toxoplasma gondii. This suggests that parasites have adopted and use this machinery in different ways, pointing, in particular in the case of the PfClpC chaperone, to a Plasmodium specific function.
Taking a genetic approach, we were able to demonstrate the localization of PfClpP and PfClpC to the apicoplast, and showed that the chaperone activity is essential to parasite viability and asexual replication (Figure 3). In a series of cellular assays, we were able to show that it is required for apicoplast integrity. Inhibition of PfClpC resulted in the loss of distinct apicoplast morphology and in the presence of vesicular-like structures (Figure 4). Several studies reported the appearance of such structures when the apicoplast integrity is compromised, for example with the use of certain antibiotics (Gisselberg et al., 2013; Yeh and DeRisi, 2011). This was interpreted as stalled vesicular transport from the ER that cannot dock to the apicoplast membrane due to loss of the organelle (van Dooren and Striepen, 2013). Moreover, chemical rescue using IPP restored PfClpC-DDD growth and cellular morphology despite the absence of the apicoplast as could be seen by microscopy and qRT-PCR (Figure 5). It has been shown that isoprenoid biosynthesis is the only essential metabolic function of the apicoplast, and supplementing IPP can replace a non-functional apicoplast in living parasites (Yeh and DeRisi, 2011). We therefore concluded that the only essential role of PfClpC is linked to apicoplast function. Inhibition of essential apicoplast metabolic pathways with drugs like Fosmidomycin, kills parasites immediately and does not lead to the loss of the organelle (Bowman et al., 2014). Conversely, inhibition of apicoplast translation or replication with drugs like Doxycycline, allows the parasites to complete one replication cycle, proceed through the next, and die only during the second schizogony (Fichera and Roos, 1997; Yeh and DeRisi, 2011). Similar to the effect of drugs that inhibit apicoplast replication, PfClpC mutants developed normally during the early stages of rings and trophozoites but late schizont stages developed aberrant morphology (Figure 4). These non-viable parasites did not manifest uniformly at the end of the second cycle but appeared on the 3rd replication cycle, and their fraction increased over time (Figure 4B). A possible explanation for the delayed growth arrest, as well as the gradual increase in abnormal parasites, is that PfClpC inhibition interferes with the segregation or division of functional apicoplast into daughter merozoites. As a consequence, a mixed population of viable and non-viable daughter cells is forming after each cycle, diluting overtime the viable parasites in the total culture. Indeed, we observed a significant decrease in the rate of schizont to ring conversion in each successive generation, clarifying the delayed growth inhibition, and suggesting a defect in apicoplast sorting (Figure 4).
To support that, we performed functional analysis, by inhibiting and then reactivating the chaperone function. This was achieved by removing TMP from the cultures for two weeks, reduce parasite numbers below detection, and then adding back TMP to reactivate the chaperone in any possible remaining parasites (Figure 5E). In the event of a uniform functional damage to all parasites in the culture, PfClpC re-activation would not lead to viable parasites, as the apicoplast must be inherited and its de novo synthesis is impossible. Nonetheless, we observed that re-addition of TMP could restore parasites growth, indicating the presence of a small yet undetectable sub-population of parasites that contains a functional apicoplast. This observation further supports a sorting defect rather than a general apicoplast dysfunction in the entire parasite population. Interestingly, TMP addition had the opposite effect on parasites that were rescued with IPP (Figure 5E). These parasites started dying 48 hours after removing IPP despite addition of TMP, indicating that re-activation of PfClpC was not enough to sustain viability in a population of parasites that permanently lost the apicoplast.
In addition, we describe here a mechanism by which PfClpC regulates the activity of the Clp complex. We found that PfClpC activity is necessary and regulates the processing of PfClpP protease in several ways (Figure 6). Chaperone inhibition resulted in PfClpP miss-localization and reduced the cleavage of its transit peptide 7 days or more after TMP removal. These observations are correlated and consistent with the kinetics of apicoplast loss in PfClpC mutants (Figure 3, 4 and 5). But it also affected the active form of PfClpP with much faster kinetics. We detected a significant reduction in PfClpP active-protease form (in which the transit peptide and the pro-domain have been removed) 4 hours after TMP removal, suggesting degradation of the complex subunits when the chaperone is inhibited (Figure 6F). In this model, the activity of the Clp chaperone is required to stabilize the proteolytic Clp complex, representing a novel mode of regulation, which may be relevant to bacterial and chloroplasts systems.
Finally, the roles that bacterial Clp proteins play in cell division, stress response and pathogenicity (Frees et al., 2014), have placed them at the center of several drug discovery programs (Böttcher and Sieber, 2008; Brötz-Oesterhelt and Sass, 2014; Conlon et al., 2013; Gersch et al., 2015; Hackl et al., 2015). Our data demonstrate that targeting the P. falciparum plastid-localized Clp proteins is a viable strategy for antimalarial drug development and future work will allow us to repurpose highly active antibacterial compounds as effective anti-malarial agents.
Experimental Procedures
Multiple sequence alignments and phylogenetic analyses
Candidate proteins were chosen based on the presence of conserved Pfam domains (Punta et al., 2012), CLP_protease for the Clp protease family and ClpB_D2-small for the Clp chaperone family. Representative genomes were searched using hmmscan, then candidate full length proteins were aligned with clustalo (Sievers and Higgins, 2014) using the Pfam generated hmm as a guide, then the alignment was trimmed with trimal (Capella-Gutiérrez et al., 2009) with the following settings “-st 0.00001 -gt 0.01”. Exact duplicate alignments were temporarily removed prior to the creation of trees and added back as 0.00001 distance sister taxa.
Maximum likelihood trees of candidate proteins were creating using FastTree2 (Price et al., 2010) with options “-spr 4 -mlacc 2 -slownni -gamma” or RaXML (Stamatakis, 2014) using the best estimate protein model selection using the script ProteinModelSelection.pl from RaXML and RaXML options “-f a -x 12345 -p 12345 -N autoMRE“. Rogue taxa (those taxa with uncertain positions in phylogenetic trees) were identified with RagNaRok (Aberer et al., 2013) and excluded from trees. Trees were then visualized and interrogated with Archeopetryx (Han and Zmasek, 2009).
Complete trees and full sequence alignemnts can be found in supplemental information.
Plasmids construction
Genomic DNAs were isolated from P. falciparum using the QIAamp DNA blood kit (QIAGEN). Constructs utilized in this study were confirmed by sequencing. PCR products were inserted into the respective plasmids using the In-Fusion cloning system (Clonetech) or using SLIC (Sequence and Ligation Independent Cloning). Briefly, insert and cut vector were mixed with a T4 DNA polymerase and incubated for 2.5 minutes at room temperature, followed by 10 minutes incubation on ice and then transformed into bacteria. All restriction enzymes used for this study were purchased from New England Biolabs. All oligonucleotides and detailed cloning procedures can be found under Supplemental experimental procedures
Cell culture and transfections
Parasites were cultured in RPMI medium supplemented with Albumax I (Gibco) and transfected as described earlier (Drew et al., 2008; Russo et al., 2009). To generate PfClpC-DDD parasites, pPfClpC-HADB was transfected in duplicates into 3D7-derived parental strain PM1KO which contains a hDHFR expression cassette conferring resistance to TMP (Liu et al., 2005). Selection, drug cycling and cloning were performed as described (Muralidharan et al., 2012) in the presence of 10 μM of TMP (Sigma). Integration was detected after one round of drug cycling with blasticidin (Sigma). Two clones from 2 independent transfections, 1G8 and 2E10, were isolated via limiting dilutions and used for subsequent experiments.
For generation of PfClpP-glmS and PfClpR-glmS parasites, a mix of two plasmids was transfected in duplicates into 3D7 parasites. The plasmids mix contained pUF1-Cas9-guide (Spillman et al., 2017) which contains the DHOD resistance gene, and pV5-glmS-PfClpP or pV5-glmS-PfClpR, which are marker-free. To generate the double-mutant parasites, PfClpC-DDD; PfClpP-glmS, the same plasmid mix was transfected into PfClpC-DDD parasites. Drug pressure was applied 48 hours post transfection, using 1 μM DSM1 (Ganesan et al., 2011), selecting only for Cas9 expression. DSM1 was removed from the culturing media once parasites were detected in the culture, usually around 3 weeks post transfection.
Growth assays
For asynchronous growth assays, parasites were washed twice and incubated without TMP (for PfClpC-DDD parasites) or with 5 mM GlcN (Sigma) (for PfClpP-glmS parasites). Throughout the course of the experiment parasites were sub-cultured to maintain the parasitemia between 1-5% and parasitemia was monitored every 24 hours via flow cytometry. Relative parasitemia at each time point was back calculated based on actual parasitemia multiplied by the relevant dilution factors. Parasitemia in the presence of TMP (PfClpC-DDD) or in the absence of GlcN (PfClpP-glmS) at the end of each experiment was set as the highest relative parasitemia and was used to normalize parasites growth. Data were fit to exponential growth equations using Prism (GraphPad Software, Inc.)
For IPP rescue, media was supplemented with 200 μM of IPP (Isoprenoids LC) in PBS. To test for the ClpP inhibitor, media was supplemented with or without 60 μM of U1-lactone and 200 μM of IPP.
To generate an EC50 curve for TMP, asynchronous PfClpC-DDD parasites were incubated for 11 days without TMP, and on day 12 were seeded in a 96 well plate with varying concentrations of TMP. Parasitemia was measured after 5 days using flow cytometry. Data were fit to a dose-response equation using Prism (GraphPad Software, Inc.).
Replication rate analysis
To determine replication rate (rings: schizonts ratio) of PfClpC-DDD parasites, TMP was removed from percoll-isolated schizonts-stage parasites and parasites were allowed to egress and reinvade fresh RBCs. Parasitemia was monitored by flow cytometry and microscopy. The ratio of rings to schizonts was calculated using number of rings arising from schizonts in the previous generation. At the beginning of each replication cycle parasites were re-synchronized using Sorbitol, and sub-cultured when required. For each replication cycle, data were normalized to rings: schizonts ratio in the presence of TMP.
The fraction of morphologically aberrant schizonts was determined using Giemsastained thin blood smears of synchronized PfClpC-DDD parasites at the final stages of each replication cycle and the fraction of defective cells was calculated based on the total late schizont stage parasite counts.
Southern blot
Southern blots were performed with genomic DNA isolated using the Qiagen Blood and Cell Culture kit. 10 μg of DNA isolated from PfClpC-DDD parasites was digested overnight with NcoI and XmnI (New England Biolabs) and integrants were screened using biotin-labeled probes against the 3’-end of the pfclpc ORF. Southern blot was performed as described earlier (Klemba et al., 2004). The probe was labeled using biotinylated Biotin-16-dUTP (Sigma). The biotinylated probe was detected on blots using IRDye 800CW Streptavidin conjugated dye (LICOR Biosciences) and was imaged, processed and analyzed using the Odyssey infrared imaging system software (LICOR Biosciences).
Western blot
Western blots were performed as described previously (Muralidharan et al., 2011). Briefly, parasites were collected and host red blood cells were permeabilized selectively by treatment with ice-cold 0.04% saponin in PBS for 10 min, followed by a wash in ice-cold PBS. Cells were lysed using RIPA buffer, sonicated, and cleared by centrifugation at 4°C. The antibodies used in this study were rat monoclonal anti-HA, 3F10 (Roche, 1:3000), rabbit anti-V5, D3H8Q (Cell Signaling, 1:1000), mouse monoclonal anti-PMV (from D. Goldberg, 1:400), and rabbit polyclonal anti-EF1α (from D. Goldberg, 1:2000). The secondary antibodies that were used are IRDye 680CW goat anti-rabbit IgG and IRDye 800CW goat anti-mouse IgG (LICOR Biosciences, 1:20,000). The Western blot images were processed and analyzed using the Odyssey infrared imaging system software (LICOR Biosciences).
Microscopy and image processing
For IFA cells were fixed using a mix of 4% Paraformaldehyde and 0.015% glutaraldehyde and permeabilized using 0.1% Triton-X100. Primary antibodies used are rat anti-HA clone 3F10 (Roche, 1:100), rabbit anti-V5, D3H8Q (Cell Signaling, 1:100), mouse anti-V5, TCM5 (eBioscience™, 1:100), rabbit anti-Cpn60 (1:1,000) and rabbit anti-ACP (from G. Mcfadden, 1:10,000). Secondary antibodies used are Alexa Fluor 488 and Alexa Fluor 546 (Life Technologies, 1:100). Cells were mounted on ProLong Diamond with DAPI (Invitrogen) and were imaged using DeltaVision II microscope system with an Olympus IX-71 inverted microscope using a 100X objective. All images were collected as Z-stack, were deconvolved using the DVII acquisition software SoftWorx and displayed as maximum intensity projection. Image processing, analysis and display were preformed using SoftWorx and Adobe Photoshop. Adjustments to brightness and contrast were made for display purposes. Thin blood smears were stained using Hema 3 stain set (PRTOCOL/ Fisher Diagnostics) and were imaged on a Nikon Eclipse E400 microscope.
Flow cytometry
Aliquots of parasite cultures (5μl) were stained with 1.5 mg/ml Acridine Orange (Molecular Probes) in PBS. The fluorescence profiles of infected erythrocytes were measured by flow cytometry on a CyAn ADP (Beckman Coulter, Hialeah, Florida) and analyzed by FlowJo software (Treestar, Inc., Ashland, Oregon). The parasitemia data were fit to standard growth curve or dose–response using Prism (GraphPad Software, Inc.).
Quantitative Real Time PCR
Synchronized ring stage parasites samples were collected at the beginning of each replication cycle and genomic DNA was isolated by saponin lysis to remove extracellular DNA. Genomic DNA was purified using QIAamp blood kits (Qiagen). Primers that amplify segments from genes encoded by nuclear or organelles genomes were designed using RealTime qPCR Assay Entry (IDT). The following primer sequences were used: cht1 (nuclear): 21 + 22. tufA (apicoplast): 23 + 24. cytb3 (mitochondria): 25 + 26. Reactions contained template DNA, 0.5 μM of gene specific primers, and IQ™ SYBR Green Supermix (BIORAD). Quantitative real-time PCR was carried out in triplicates and was performed at a 2-step reaction with 95°C denaturation and 56°C annealing and extension for 35 cycles on a CFX96 Real-Time System (BIORAD). Relative quantification of target genes was determined using Bio-Rad CFX manager 3.1 software. Standard curves for each primers set were obtained by using different dilutions of control gDNA isolated from parasites grown in the presence of TMP (20 to 0.2 ng) as template, and these standard curves were used to determine primers efficiency. For each replication cycle number, the organelle: nuclear genome ratio of the –TMP+IPP treated parasites was calculated relative to that of the +TMP control.
Imaging flow cytometry
Synchronized PfClpC-DDD parasites incubated for 10 days without TMP and then were isolated on a percoll gradient following by a sorbitol treatment 5 hours later to obtain early rings (0-5 hours post invasion). Cells were fixed and stained with anti HA antibody as described above and nuclei were stained using DAPI from Amnis Intracellular staining kit (EMD MILIPORE). Data were collected on ImageStream X Mark II (EMD MILIPORE) and an automated collection of a statistically large number of cells (10,000) was performed. Data were analyzed using IDEAS software version 6.2.
Acknowledgements
We thank Geoffrey McFadden for anti-ACP antibody and Dan Goldberg for pUF1-Cas9 plasmid, anti-PMV and anti EF1α antibodies; Stephan A. Sieber for providing the U1-lactone inhibitor. Drew Etheridge and Roberto Docampo for comments on the manuscript; Julie Nelson at the CTEGD Cytometry Shared Resource Laboratory for help with flow cytometry and analysis; and Muthugapatti Kandasamy at the Biomedical Microscopy Core at the University of Georgia for help with microscopy. We acknowledge the assistance of the Children’s Healthcare of Atlanta and Emory University Pediatric Flow Cytometry Core for imaging flow cytometry. This work was supported by grants from the March of Dimes Foundation (Basil O’Connor Starter Scholar Research Award), the US National Institutes of Health (R00AI099156 and R21AI128195), and UGA startup funds to V. M.