The intracellular parasite Toxoplasma gondii depends on the synthesis of long chain and very long-chain unsaturated fatty acids not supplied by the host cell

SUMMARY

Apicomplexa are parasitic protozoa that cause important human diseases including malaria, cryptosporidiosis and toxoplasmosis. The replication of these parasites within their target host cell is dependent on both salvage as well as de novo synthesis of fatty acids. In T. gondii, fatty acid synthesis via the apicoplast-localized FASII is essential for pathogenesis, while the role of two other fatty acid biosynthetic complexes remains unclear. Here we demonstrate that the ER-localized fatty acid elongation (ELO) is essential for parasite growth. Conditional knock-down of the non-redundant hydroxyacyl-CoA dehydratase and enoyl-CoA reductase enzymes in the ELO pathway severely repressed intracellular parasite growth. ¹³C-glucose and ¹³C-acetate labeling and comprehensive lipidomic analyses of these mutants showed a selective defect in synthesis of unsaturated long and very long chain fatty acids (LCFAs and VLCFAs) and depletion of phosphatidylinositol and phosphatidylethanolamine species containing unsaturated LCFAs and VLCFAs. This requirement for ELO pathway was by-passed by supplementing the media with specific fatty acids, indicating active, but inefficient import of host fatty acids. Our experiments highlight a gap between the fatty acid needs of the parasite and availability of specific fatty acids in the host cell that the parasite has to close using a dedicated synthesis and modification pathway.

INTRODUCTION

Protozoan parasites belonging to the phylum Apicomplexa are the cause of several important diseases including malaria, cryptosporidiosis and toxoplasmosis. All of these parasites have obligate intracellular life cycle stages that are capable of proliferating within a variety of different host cell niches including lymphocytes and macrophages, liver cells, red blood cells, intestinal epithelial cells, muscle fibers, neurons and other cell types of the central nervous system. These cell types vary dramatically in their overall growth rate and metabolic activity, oxygen tension, and in their access to nutrients from surrounding tissues. There is increasing evidence that differences in host metabolism may impact the intracellular growth of apicomplexan parasites (Coppens, 2013). In particular, the extent to which these parasites are dependent on salvage versus de novo pathways of lipid biosynthesis are known to vary depending on both the parasite life cycle stages and the host cell involved (Tarun et al., 2009, Ramakrishnan et al., 2013). While apicomplexan parasites maintain three independent pathways for the biosynthesis of fatty acids, the expression and contribution of these pathways vary in different species and developmental stages (Ramakrishnan et al., 2013). In a number of apicomplexans, long chain fatty acids are synthesized by a type II fatty acid synthesis (FASII) pathway (Fig. 1) that is localized to the apicoplast, a unique parasite organelle derived from an algal endosymbiont (Mazumdar et al., 2006, Waller et al., 1998). As found for the cyanobacterial and chloroplast pathways from which it evolved, the separate enzymatic activities required for apicoplast FASII are encoded as individual polypeptides. Metabolomic and isotope tracer studies in T. gondii and P. falciparum demonstrate FASII is required for the de novo synthesis of long chain fatty acids, such as myristic and palmitic acid that are important components of bulk membrane lipids (Ramakrishnan et al., 2012, Botte et al., 2013). While genetic studies suggest that this pathway is essential for Toxoplasma gondii (Mazumdar et al., 2006), insect stages of P. falciparum (van Schaijk et al., 2014), and liver stages of P. yoelii (Vaughan et al., 2009) and P. berghei (Yu et al., 2008) it is dispensable for the red-blood cell stage of all these Plasmodium species suggesting that the latter can salvage all of their fatty acid requirements from the host cell (Vaughan et al., 2009, Tarun et al., 2009, Pei et al., 2010, Yu et al., 2008).

Figure 1. The apicoplast localized fatty acid synthase type II pathway and the ER associated fatty acid elongation pathway interact in the synthesis of fatty acids in T. gondii.

The enzymes of the type II fatty acid synthesis (FASII) pathway localize to apicoplast lumen (green). A starter molecule is synthesized by condensation from CoA activated precursors on acyl carrier protein (ACP). The molecule is sequentially reduced, dehydrated and reduced to yield a fully reduced fatty acyl chain. This process is repeated in multiple rounds to generate myristate or palmitate as the final products. The fatty acid elongation (FAE) pathway unfolds on the cytoplasmic face of the endoplasmic reticulum and its membrane bound enzymes catalyze a similar set of reactions. An elongase (ELO) condenses a malonyl-CoA molecule with a fatty acyl-CoA, for ELO-A the fatty acid is palmitate derived from the FASII pathway. T. gondii harbors three elongase enzymes, with catalytic specificity determined by length and saturation of the fatty acid. The products of all three elongases are then reduced, dehydrated and reduced by a ketoacyl-CoA reductase (KCR), hydroxyacyl-CoA dehydratase (DEH) and enoyl-CoA reductase enzymes (ECR). The enzymes that are the focus of this study are highlighted in red. Note that the carbon source for the two pathways is distinct, and that only the elongation pathway is efficiently labeled using acetate isotopes.

Unlike the FASII pathway, components of the type I fatty acid synthesis (FASI) pathway are encoded as a single polypeptide chain and can be found in some apicomplexans, including members of the coccidia (T. gondii, Eimeria spp) and Cryptosporidium spp. The apicomplexan FASI appears to be involved in the elongation of long chain fatty acids, rather than de novo synthesis (Zhu et al., 2004, Zhu et al., 2010). Recent studies suggest that FASI and related polyketide megasynthases may have a specialized role in the synthesis of the unique extracellular walls of infectious oocysts that can survive long term in the environment (Ramakrishnan et al., 2013, Bushkin et al., 2013).

This study focuses on a third set of enzyme complexes that are expressed in apicomplexan parasites and are involved in fatty acid synthesis and elongation (Fig. 1). This ER-associated fatty acid elongation (ELO) pathway shares a number of common steps with FASI and II, but also differs from those pathways in utilizing coenzyme A (CoA) as the fatty acid carrier rather than an acyl carrier protein. Elongation is initiated by condensation of fatty acyl-CoA and malonyl-CoA followed by a reduction, dehydration and second reduction reaction to generate a fatty acid elongated by two carbon units (Jakobsson et al., 2006, Denic & Weissman, 2007). In T. gondii, the condensation steps are catalyzed by three different elongase enzymes; ELO-A, ELO-B and ELO-C (Ramakrishnan et al., 2012). These enzymes differ in their substrate specificities. While ELO-A and ELO-B are engaged in the elongation of de novo synthesized unsaturated fatty acids, ELO-C appears to primarily act on host-derived saturated fatty acids. Genetic deletion of individual ELO complexes had little effect on the intracellular growth of T. gondii tachyzoites in host cells, suggesting functional redundancy between these complexes and/or that other fatty acid biosynthetic or salvage mechanisms compensate for the loss of individual ELO complexes.

Here we report the isolation of two conditional T. gondii mutants lacking non-redundant enzymes of the ELO pathway, hydroxyacyl-CoA dehydratase and enoyl-CoA reductase. These mutants were found to have significant defects in fatty acid elongation and exhibited a marked reduction in intracellular growth. Loss of parasite viability and growth could be restored by supplementation of infected host cells with unsaturated long chain (LCFA) and very long chain fatty acids (VLCFA) suggesting that the essentiality of ELO complexes reflects the inability of these stages to scavenge sufficient amounts of these fatty acids from infected host cells under normal growth conditions. These studies highlight the extent to which the complex fatty acid demands of intracellular parasite stages are balanced by de novo and salvage pathways.

RESULTS

Isolation of conditional mutants for non-redundant components of fatty acid elongation

We have previously generated T. gondii mutants with defects in individual ELO complexes by targeted deletion of the ELO gene locus by homologous recombination (Ramakrishnan et al., 2012). However, this approach was unsuccessful when more than one ELO gene, or the genes for non-redundant components of the pathway were targeted, suggesting that parasites may be dependent on at least two of the ELO complexes. We next explored a strategy that places the native chromosomal locus directly under the control of a tetracycline inducible promoter (Sheiner et al., 2011). Using this approach we were able to isolate conditional mutants for two non-redundant components of the T. gondii ELO pathway, hydroxyacyl-CoA dehydratase (DEH) and enoyl-CoA reductase (ECR). A tetracycline promoter along with a selectable marker was targeted just upstream of the initiation codon of the gene by homologous recombination using suitable 5′ and 3′ flanks. ΔKu80/TATi parasites (Fox et al., 2009, Sheiner et al., 2011) were used as background to facilitate gene targeting and subsequent expression of the locus by transactivation. Pyrimethamine-resistant clones were isolated and tested for site-specific insertion by PCR of the wild type dehydratase gene locus resulted in the production of a diagnostic 750 bp amplicon in RH wild type and the parental ΔKu80/TATi strain. This band was no longer observed in clones where the gene was targeted (highlighted by asterisk in Fig. 2B, Fig. 2A for locus maps and primer positions). Similarly, conditional mutants for the enoyl-CoA reductase were identified based on the presence of a 2 kb amplicon (Fig. 2D) which was absent in the wild type and ΔKu80/TATi parasites. The presence of parasite genomic DNA was confirmed in all samples using a primer set for a control gene that produces a 500 bp amplicon (Fig. 2B, 2D). Single conditional mutant clones for the dehydratase (subsequently referred to as iΔDEH) and enoyl-CoA reductase (iΔECR) were selected for further analysis and characterization.

Figure 2. Parasite mutants in the dehydratase and enoyl reductase enzymes of the fatty acid elongation pathway exhibit severe growth defects.

T. gondii conditional mutants for DEH and ECR were generated by insertion of a tetracycline regulatable promoter into the chromosomal locus of the respective gene by homologous recombination of a targeting construct (A). Mutants were identified by PCR. Clones that lack a diagnostic 750 bp product were identified as dehydratase mutants (iΔDEH) (B), The abundance of DEH mRNA transcript in iΔDEH parasites was determined in ATc treated (+) and untreated parasites (−) by quantitative PCR. Transcript levels for acetyl-CoA reductases (ACCase) another enzyme involved in fatty acid synthesis were measured to control knock down specificity (C). Clones that produced a new 2 kb product were identified as enoyl-CoA reductase (iΔECR) mutants (D). All clones were tested for the presence of genomic DNA samples using a control primer set which produces a 500 bp band (B, D). The growth of conditional mutants iΔECR and iΔDEH was evaluated by plaque assay in the absence or presence of ATc in the medium as indicated. The growth of ΔKu80/TATi, the parental line used to generate these mutants is unaltered by ATc treatment (E). Growth of iΔECR and iΔDEH (F) parasites stably expressing a transgene resulting in the expression of double Tomato red fluorescent protein was evaluated by measuring parasite fluorescence in 96 well cultures daily for nine days. The average fluorescence intensity (arbitrary units) for three independent replicates is shown and error bars represent the standard deviation for each data point. Parasites were grown in the absence (−) or presence (+) of ATc or under ATc after three days of pre-incubation in the previous passage (pre).

The T. gondii fatty acid elongation pathway is required for parasite growth

Ablation of either ELO-specific hydroxyacyl-CoA dehydratase or enoyl-CoA reductase should result in inactivation of all ER-associated ELO complexes (Ramakrishnan et al., 2012). To investigate whether these complexes are essential for intracellular growth, the capacity of conditional mutant parasite lines to produce plaques in host cell monolayers was evaluated in the presence or absence of 0.5 μM anhydrous tetracycline (ATc). ATc inhibits the transactivator and blocks transcription from loci that carry the promoter modified with tet-operator sequences (Meissner et al., 2002). While the growth of the ΔKu80/TATi parental parasite line was not affected by ATc (Fig. 2E), conditional knockdown of ECR in the iΔECR mutant completely ablated plaque formation in human foreskin fibroblasts (HFF) monolayers (Fig. 2D). The growth of the iΔDEH mutant was similarly ablated in the presence of ATc, although very small plaques were occasionally observed (Fig. 2E). To further evaluate the growth phenotype of these strains we engineered mutant strains that expressed the dTomato fluorescent protein cassette (van Dooren et al., 2008). iΔDEH and iΔECR parasites expressing red fluorescent protein were cultured in either the absence or presence of ATc in 96-well plates and parasite growth was monitored by measuring fluorescence over nine days. Consistent with the plaque assay, intracellular growth of the iΔECR mutant was completely blocked following addition of ATc (Fig. 2F). In contrast, growth of iΔDEH parasites continued to grow slowly after ATc treatment until day 5 at which point parasite growth ceased. Pretreatment of iΔDEH parasites with ATc (one passage) followed by re-inoculation into fresh HFF monolayers resulted in a strong defect in parasite growth from the beginning of the experiment (Fig. 2F). These assays show that fatty acid elongation is required for growth of T. gondii tachyzoites. The basis for the differential sensitivity of iΔDEH and iΔECR mutants to ATc remains unknown, but does not appear to be due to differential loss of corresponding mRNAs as loss of DEH transcripts in the presence of ATc was confirmed by quantitative PCR (Fig. 2C).

Parasite synthesis of long chain and very long chain fatty acids is dependent on hydroxyacyl-CoA dehydratase and enoyl-CoA reductase

In order to confirm that ECR and DEH are essential for fatty acid elongation, extracellular wild type and mutant parasites grown in absence or presence of ATc were metabolically labeled with ¹⁴C-acetate in fatty acid-free media for 4 hours at 37°C under 5% CO₂. Total cellular fatty acids were prepared by saponification and their corresponding methyl esters analyzed by reverse phase thin layer chromatography. Similar patterns of ¹⁴C-labelled fatty acids were generated in the parental ΔKu80/TATi and iΔDEH and iΔECR parasites in the absence of ATc. Incorporation of ¹⁴C-acetate into fatty acids was dramatically decreased in iΔDEH parasites after pretreatment with ATc for 48 h (Fig. 3A). As pretreatment of iΔECR parasites with ATC for 48 hr results in parasite death, labeling studies in this mutant were initiated after 24 hours of ATc pretreatment. Under these labeling conditions, synthesis of long chain fatty acids was reduced (Fig. 3B), while faster migrating (shorter chain or more polar) fatty acid species were labeled more strongly. The latter fatty acid species may represent intermediates in the elongation process, which involves an initial condensation step that is not inhibited in these mutants (Fig. 1). Specifically, they may be keto-and enoyl-derivatives that are more hydrophilic than fatty acids and are expected to have faster mobility on reverse phase TLC (Lea-Smith et al., 2007). Alternatively, these species may represent shorter fatty acids that accumulate in the absence of elongation activity. However, such species were not evident when the condensation enzymes were ablated (Ramakrishnan et al., 2012).

Figure 3. Analysis of fatty acid synthesis in mutant parasites by metabolic labeling using radioactive or stable isotope precursors.

T. gondii conditional mutants for the hydroxyacyl-CoA dehydratase or enoyl-CoA reductase of the elongase pathway were labeled with [¹⁴C]-acetate and total cellular fatty acids analyzed by reverse phase TLC as their methylesters (the migration position of [¹⁴C]-palmitate methylester which was used as a standard in this experiment is indicated on the right hand side). Representative autoradiographs for the parental strain ΔKu80/TATi and the iΔDEH, mutant (A) and for the iΔECR mutant (B) are shown. The data shown is representative of 3 biological replicates for iΔECR, and more than three for iΔDEH. Parasites were grown in the absence (−) or presence (+) of ATc prior to the addition of ¹⁴C-acetate (48h for iΔDEH and 24h for iΔECR). Note that the TLC mobility of fatty acid methyl esters is inversely related to chain length. iΔDEH parasites were also subjected to comprehensive fatty acid analysis following stable isotope labeling with [¹³C]-glucose (C) or [¹³C]-acetate (D, E). Lipids were extracted and ¹³C-labeling of individual fatty acid methylesters determined by GC/MS. The level of incorporation of heavy isotope due to labeling is shown for parasites grown in the absence (white bars) or presence (black bars) of ATc in (C) and (D) respectively. Bars show the mean of three technical replicates and error bars represent the standard deviation of those measurements. (E) Changes in the overall abundance of fatty acids for [¹³C]-acetate labeled iΔDEH parasites depending on growth in the presence or absence of ATc. Please note that the absolute abundance of C24:1 is very low making the measurement of label incorporation in panels C,D and abundance in panel E susceptible to noise. Results are representative of 3 individual biological replicates. Average values for the representative replicate are listed in Table 1 and Table 2. Fatty acids for which significant changes in labeling were observed in the presence and absence of ATc using the Wilcoxon Rank Sum Test (p values less than 0.05) are indicated with an asterisk.

We have previously shown that it is possible to measure de novo fatty acid synthesis in intracellular tachyzoite stages by labeling infected host cells with ¹³C-glucose (Ramakrishnan et al., 2012). ¹³C-glucose is taken up by the intracellular tachyzoites and converted to ¹³C-triosephosphates which are transported into the apicoplast and used to synthesize long chain saturated fatty acids (myristate and palmitate) via the FASII complex (Fig. 1) (Mazumdar et al., 2006, Ramakrishnan et al., 2012). Glycolytic intermediates are also converted to ¹³C-acetyl-CoA and used by enzymes in the ELO pathway (Fig. 1) (Mazumdar et al., 2006, Ramakrishnan et al., 2012, Oppenheim et al., 2014). HFF were infected with the iΔDEH mutant (which remains viable for at least 72 h after induction) and incubated in the presence of ATc for 48 h prior to supplementation of the medium with [U-¹³C] glucose. After 24 h parasites were extracted and total fatty acids analyzed by GC/MS. As expected, the saturated long chain fatty acid, myristate (C14:0) and palmitate (C16:0) were strongly labeled in both untreated and ATc-repressed parasites (Fig. 3C), reflecting de novo synthesis via the apicoplast FASII complex, while stearate (C18:0) was minimally labeled (<5%), reflecting active salvage of this abundant host fatty acid from the medium (Mazumdar et al., 2006, Ramakrishnan et al., 2012, Charron & Sibley, 2002). It has previously been suggested that parasite synthesized C16:0 is subsequently desaturated to form C16:1 and then elongated by ELO complexes (Mazumdar et al., 2006, Ramakrishnan et al., 2012). Consistent with this pathway, C16:1 was labeled in the presence or absence of ATc, while labeling of long and very long chain, mono-unsaturated fatty acids, C18:1 to C26:1, was markedly reduced following repression of DEH expression (Fig. 3C). Reduced labeling of these monounsaturated, long and very long chain fatty acids (LCFA and VLCFA) was linked to a concomitant decrease in their absolute abundance (Fig. 3E). Decreased cellular levels were particularly apparent for those fatty acids that were strongly labeled in ¹³C-glucose-fed parasites (i.e. C20:1, C22:1 and C26:1) (Fig. 3C, E). These data provide direct evidence that the ELO enzymes are required for elongation of C16:1 to form LCFA and VLCFA

In contrast to ¹³C-glucose, ¹³C-acetate is converted directly to ¹³C-acetyl-CoA by a cytoplasmic acetyl-CoA synthetase and primarily used by enzymes in the ELO pathway (Fig. 1). Consistent with this notion, addition of ¹³C-acetate to the medium of iΔDEH-infected HFF in the absence of ATc resulted in highly selective labeling of parasite unsaturated VLCFA (C22:1, C26:1) compared to FASII synthesized C14:0 and C16:0 fatty acids (Fig. 3D). Following addition of ATc and repression of DEH expression, the labeling of C22:1 and C26:1 fatty acids was markedly reduced (Fig. 3D), coincident with reduced cellular levels of these fatty acids. These results are consistent with the ¹⁴C-acetate labeling experiments where a marked reduction in the labeling of the slower migrating very long chain fatty acid species was observed. Together, these data show that DEH is essential for all three of the previously described ELO systems and that loss of DEH activity is associated with marked decreases in the steady-state levels of selected fatty acids.

Repression of DEH results in selective loss of phosphatidylinositol and phosphatidyl-ethanolamine species containing long and very long chain mono and polyunsaturated fatty acids

Fatty acids synthesized by the parasite or scavenged from the host are primarily incorporated into membrane phospholipids. To determine whether loss of the ELO complexes results in global or selective loss of specific phospholipid molecular species, total phospholipids were profiled by LC/MS. The molecular species composition of individual phospholipid classes was determined by precursor ion and neutral loss scanning, allowing selective quantitative detection of lipids with different head groups. Repression of DEH expression had little impact on the relative abundance of the major phosphatidylcholine (PC) molecular species, but led to a marked decrease in a subset of molecular species of phosphatidylinositol (PI) and phosphatidylethanolamine (PE) containing LCFA and VLCFA (Fig. 4A–C). In particular, PI and PE molecular species containing a collective chain length/degree of unsaturation of 38:4 were decreased by >80% in the presence of ATc (Fig. 4A and B). MS/MS analysis of these species showed that both comprised predominantly C18:0–C20:4 acyl chains (data not shown). The levels of other PE species with longer acyl chains was also decreased. These data suggest that decreased synthesis of specific PI/PE molecular species may, in part, underlie the severe loss of growth phenotype following repression of DEH expression.

Figure 4. Repression of DEH leads to selective changes in tachyzoite phospholipid composition.

iDEH parasites were cultivated in HFF in the presence or absence of ATc (48hr). Parasites were purified from host debris and major phospholipid classes analyzed by LC/MS. The relative abundance of selected PI (panel A), PE (panel B) and PC (panel C) species are shown. The precise acyl composition of selected molecular species was determined by MS/MS. Molecular species of PI (38:4) and PE (38:4) contained predominantly C18:1 and C20:4. Data are represented as mean +/- SEM for three replicate analyses derived from two independent experiments.

Long and very long chain monounsaturated fatty acids complement the loss of the elongation system

Loss of DEH or ECR appears to result in selective loss of specific unsaturated fatty acids that may be present at limiting amounts in the parasitophorous vacuole (PV), a specialized compartment that forms during parasite invasion and is subsequently remodeled. We therefore investigated whether it was possible to by-pass the requirement for ELO activity by supplementation of the medium with exogenous fatty acids. iΔDEH parasites expressing the red fluorescent protein were cultured in the presence of ATc and individual fatty acids or fatty acid mixtures were delivered in complex with bovine serum albumin carrier and parasite growth was monitored over 6 days. Changes in the fluorescence of ATc-treated parasites were normalized to non-ATc-treated cultures to control for plate-to-plate variation. Strikingly, growth of fluorescent iΔDEH parasites in the presence of ATc was rescued by supplementation of the medium with a mixture of saturated- and mono-unsaturated LCFA and VLCFA at a concentration of 250 μM (Fig. 5A). These findings suggest that exogenous fatty acids can be transported to the parasitophorous vacuoles and utilized by intracellular parasite stages. They also provide direct evidence that growth arrest of DEH-depleted parasites was due to the lack of one or more of these fatty acids.

Figure 5. Chemical complementation of fatty acid elongation mutants depends on the enzyme lost and the specific fatty acid species provided.

HFF were infected with T. gondii iΔDEH (A–F) and iΔECR (G) parasites stably expressing a dTomato-RFP transgene and cultivated in the presence of BSA-fatty acid conjugates. Intracellular parasite growth following repression of DEH or ECR in the presence of different fatty acids was monitored by fluorescence in 96 well cultures. Bars show percent fluorescence (mean of three measurements) compared to untreated controls included in the same plate after six days of culture. The concentration of fatty acids (conjugated to BSA) in the medium was 250 μM, unless stated otherwise. The BSA conjugates included individual fatty species or mixtures of saturated fatty acids (Sat FA; C18:0, C20:0, C22:0 and C24:0), unsaturated fatty acids (Unsat FA; C18:1, C20:1, C22:1 and C24:1) or combined fatty acids (FA; C18:0, C20:0, C22:0, C24:0, C18:1, C20:1, C22:1 and C24:1). Intracellular growth of i DEH parasites in HFF cultured in the presence of ATc and (A) the combined fatty acid mixture, (B) saturated fatty acids (Sat) or unsaturated fatty acids (Unsat) alone, or (C) individual saturated and (D) individual monounsaturated fatty acids. The effect of adding higher concentrations of FA on i DEH growth was also tested (E). Addition of unsat FA to iΔDEH infected HFF also restored the capacity of these parasites to form plaques (F). In contrast, growth of the iΔECR mutant was not restored when infected HFF were cultured in the presence of BSA-FA mix (G). Similar results were obtained with other FA mixtures (not shown). The results were analysed using an unpaired t-test. Differences with a P value < 0.05 were deemed statistically significant and are indicated by an asterisk.

To further investigate the fatty acid requirements of the ELO mutants, the growth media of HFF infected with fluorescent iΔDEH parasites was supplemented with either a mixture of saturated or unsaturated LCFAs and VLCFAs at a final fatty acid concentration of 250 μM. Supplementation of the medium with saturated fatty acids (C20:0, C22:0 and C24:0) did not rescue parasite growth, while addition of the four unsaturated fatty acids, C18:1, C20:1, C22:1 and C24:1 restored growth completely (Fig. 5B). These results were further supported by supplementation experiments with individual fatty acids. None of the saturated fatty acids complemented intracellular iΔDEH growth (Fig. 5C) while individual unsaturated fatty acids provided partial rescue, with C20:1 and C22:1 being most effective (Fig. 5D). Complementation with individual unsaturated fatty acid did not rescue growth as efficiently as when a mixture of unsaturated fatty acids was provided (even when concentrations were increased Fig. 5E), suggesting that exogenous fatty acids may not be processed by the ELO machinery as efficiently as de novo synthesized saturated/non-unsaturated fatty acid precursors. Lastly, media supplementation with a mixture of unsaturated fatty acids also fully restores parasite plaque formation (Fig. 5F).

We also attempted to chemically complement the growth of the iΔECR mutant. In stark contrast to iΔDEH, supplementing media with fatty acids did not rescue growth. Fig. 5G shows one representative example using a mixture of unsaturated fatty acids that was highly effective for the iΔDEH mutant. The rapid death of parasites upon ECR suppression combined with the clear lack of chemical complementation suggests that ECR mutants may accumulate toxic intermediates and/or deplete important cofactors like CoA resulting in a dominant lethal phenotype.

DISCUSSION

T. gondii tachyzoites are highly dependent on fatty acid biosynthesis and/or salvage pathways in order to sustain the synthesis of new membrane and high proliferation rates. However, the extent to which these parasite stages are dependent on host or de novo synthesized lipids still remains poorly defined and may vary depending on the host cell involved. Genomic, biochemical and genetic studies have shown that T. gondii and other apicomplexan parasites express three distinct fatty acid biosynthetic complexes that have overlapping, as well as non-redundant, roles in generating the complete spectrum of fatty acids needed for growth (Tarun et al., 2009, Zhu et al., 2004, Zhu et al., 2010, Zhu, 2004, Ramakrishnan et al., 2012, Coppens, 2013). The apicoplast-located FASII complex is involved in the synthesis of saturated long chain fatty acids (up to C16:0), while the three ER-located ELO complexes act sequentially on both saturated and unsaturated fatty acids to form species with chain length up to C26:0. The function of the cytoplasmic FASI complex remains to be fully defined. We have recently shown that individual ELO complexes are not essential for tachyzoite growth in human fibroblasts indicating substantial redundancy in the substrate specificity of these complexes and/or the capacity of intracellular parasites stages to salvage specific fatty acid species from the host cell. In contrast, we now show that knockdown of DEH or ECR diminishes the activity of all three ELO complexes and results in a profound defect in parasite growth in host cells demonstrating lack of redundancy between the parallel pathways of de novo synthesis and salvage of long and very long chain fatty acids. Our results suggest that salvage pathways may allow sufficient uptake to by-pass the loss of single ELO complexes, but not all three complexes. The constitutive expression of all three ELO complexes is thus likely to be essential in allowing these parasites to grow in different PV lipid environments in the wide variety of different host cells that sustain T. gondii infection.

Function of long and very long chain fatty acids

Our studies suggest that LCFA and VLCFA are essential for intracellular tachyzoite growth. These fatty acids account for less than 20% of the total parasite FA pool and are incorporated into all phospholipid classes. Interestingly, comprehensive lipidomic analysis of parasite membrane lipids after conditional knock-down of DEH suggested that depletion of LCFA and VLCFA leads to a selective decrease in some PE and PI molecular species. Specifically, abundant PE and PI molecular species containing C20:4 fatty acyl chains were markedly decreased following conditional knock-down of DEH. Less abundant PE species containing longer acyl chain combinations were also decreased. In contrast, the levels of expression of major molecular species of phosphatidylcholine, including PC (38:4) were largely unchanged. It is notable that each of the major phospholipid classes detected in these analyses contained distinctive acyl chain compositions which likely reflects differences in their subcellular sites of synthesis as well as the extent to which their acyl chain composition is remodelled through acyl exchange reactions. For example, PE is synthesized via multiple pathways in the ER, mitochondrion, dense granules and the parasitophorous vacuole (Hartmann et al., 2014), while PC is primarily synthesized by the Kennedy pathway in the ER or late secretory pathway (Gupta et al., 2005). PC and other phospholipids may also be actively scavenged from the host cell. These class-specific differences in phospholipid synthesis and salvage may account for the selective depletion of PE and PI molecular species following DEH repression. The finding that PI species containing C20:4 species is selectively depleted is of particular interest given the well-defined role of lipase-released diacylglycerol containing this fatty acid in signalling (Thardin et al., 1993). Recent studies have also highlighted the role of phosphoinositides in membrane trafficking and apicoplast biogenesis in Toxoplasma (Tawk et al., 2011). Therefore changes in the levels of specific PE or PI molecular species may underlie the lethal phenotype observed following repression of DEH.

Salvage mechanisms

The intracellular growth phenotype induced by repression of DEH activity could be rescued by supplementation of the growth medium of infected host cells with unsaturated LCFA and VLCFA, but not with saturated species. These findings demonstrate that intracellular tachyzoites can salvage unsaturated LCFA and VLCFA and suggest that these fatty acids are normally present at levels that are limiting for parasite growth in the absence of at least some de novo biosynthesis. They also suggest that saturated VLCFA are not efficiently converted to unsaturated VLCFA, either because they are taken up less efficiently than unsaturated VLCFA or because internalized saturated VLCFA are not substrates or accessible to the tachyzoite desaturases. We have previously proposed that FASII-synthesized palmitate is converted to C18:1 via the activities of ELO1 and a stearate desaturase, although the precise sequence of these reactions remains uncertain and there could potentially be additional routes of synthesis. Regardless of the precise sequence, de novo synthesized C18:1 appears to be preferentially utilized by other ELO complexes to form C20:1 to C26:1. While C18:1 can be salvaged from the host (as shown by reduced labelling compared to C16:1 in ¹³C-glucose-fed parasites) it is poorly extended to unsaturated VLCFA. Taken together, these studies suggest that the regulation of fatty acid biosynthesis in intracellular tachyzoites is complex and may involve the selective channelling of FASII products from the apicoplast to the ER-located ELO enzymes. Substrate channelling may in turn regulate the extent to which host-derived saturated and unsaturated VLCFAs are modified and/or used directly. How apicoplast-derived fatty acids could be channelled to ER-localized ELO complexes remains unclear, although there is evidence that parts of the apicoplast outer membrane may be closely apposed to the ER (Tomova et al., 2009). These junctions are generally proposed to be sites of transport of nuclear-encoded protein from the ER to the apicoplast, but could equally be involved in the transport of apicoplast synthesized fatty acids to the ER.

Intracellular T. gondii tachyzoites have been shown to scavenge a number of essential lipids from the host cell (Coppens, 2014), including sterols (Robibaro et al., 2002), ceramides (Romano et al., 2013) and isoprenoids (Li et al., 2013). These salvage pathways are needed to acquire lipids for which the parasite lacks the enzymatic machinery, or to supplement endogenous pathways of de novo synthesis. Recent studies have highlighted multiple pathways by which host lipids can be delivered to the parasitophorous vacuole (Coppens, 2013). These include the diversion of host endocytic vesicles containing LDL particles to the PV, the internalization of host endo-lysosomes into tubular invaginations into the PV, and the transfer of lipids from host ER and mitochondria that surround the PV (Sinai et al., 1997, Coppens et al., 2000). There is accumulating evidence that the transfer of lipids from host membrane vesicles and organelles to the PV is complex and requires multiple lipid-binding proteins and transporters. Whether tachyzoites acquire LCFAs and VLCFAs via direct uptake of free fatty acids, as demonstrated here, or via the uptake and breakdown of phospholipids during the maintenance and turnover of the PV membrane remains to be determined. The ELO mutants generated in this study provide a useful experimental system for investigating these processes.

EXPERIMENTAL PROCEDURES

Parasite culture and construction of mutants

T. gondii tachyzoites derived from strain RH were cultured and genetically manipulated as described previously (Striepen & Soldati, 2007). Conditional mutants for hydroxyaxyl-CoA dehydratase (Genbank accession number: NP_012438) and enoyl-CoA reductase (Genbank accession number: NP_010269) were generated using a promoter insertion strategy in a ΔKu80/TATi mutant parasite line (Sheiner et al., 2011). Stable transgenesis in this parasite line occurs primarily through homologous recombination and a tetracycline repressible transactivator systems permit conditional gene expression. For the dehydratase locus, cosmid PSBMG05 was modified by recombineering (Brooks et al., 2010) so that a tetracycline-regulated promoter was inserted prior to the start codon of the coding sequence. The modified cosmid was then transfected into the ΔKu80/TATi parasite line and transgenic parasites selected in the presence of 1 μM pyrimethamine. Mutants were identified by PCR (detailed in Fig. 2A, 2B and 2D; See supplementary table S1 for primer details). To generate a conditional mutant for the enoyl-CoA reductase locus, a targeting plasmid was developed. This plasmid contained the tetracycline-regulated promoter and a pyrimethamine resistance cassette flanked by suitable sequences derived from the ECR locus (See table S1 for primer details). Following transfection of ΔKu80/TATi parasites with this plasmid, transgenic parasites were selected in presence of pyrimethamine and mutants were identified by PCR (Fig. 2D). Parasite growth was evaluated by plaque assay in absence or presence of 0.5 μM ATc (Fig. 2E) (Striepen & Soldati, 2007). Both mutants were transfected with a dTomato-RFP expression plasmid (van Dooren et al., 2008) and stable transgenics were isolated by cell sorting. Growth was measured in 96 well plates using a fluorescence assay (Gubbels et al., 2003).

[¹⁴C]-Acetate radiolabelling and thin layer chromatography

Parasites were grown in the absence or presence of 0.5 μM ATc for 48 hours and free tachyzoites (10⁸) were metabolically labeled with 10 μCi of sodium [¹⁴C]-acetate (specific activity 57mCimmol⁻¹, Perkin Elmer) in 1 ml of Dulbecco’s modified eagle medium for 4 hr at 37°C and 5% CO₂. Total lipids were extracted with chloroform/methanol (2:1 v/v), dried under nitrogen, and fatty acids released as their methyl esters after acidic methanolysis (0.5 mM HCl in anhydrous methanol, 80°C, 3hours). The labeled fatty acid methyl esters were extracted with hexane and analyzed on RP-18 HPTLC plates developed in methanol/chloroform/water (75:25:5 v/v), sprayed with Enhance scintillator and exposed to film for 1 week at −80°C.

Stable isotope labeling and metabolomics analyses

T. gondii infected fibroblasts were cultivated in DMEM medium in a T175 flask in the absence or presence of ATc. For labeling studies, the medium was supplemented with 8 mM [U-¹³C]-glucose (final concentration of 16 mM [¹³C/¹²C] glucose) 48 hr after infection and parasites were grown for another 24 hours. Following host cell lysis and parasite egress, free parasites were separated from host cells by filtration through a 3 μm polycarbonate membrane and rapidly chilled by immersion of the cell suspension in a dry-ice/ethanol bath (Ramakrishnan et al., 2012, MacRae et al., 2012). Parasites were recovered by centrifugation (4,000 X g, 25 min, 0 °C) and cell pellets washed three times with ice-cold PBS. Cell aliquots (2 X 10⁸ cells) were transferred to microcentrifuge tubes, harvested by centrifugation (10,000 X g, 30 s, 0 °C) and pellets extracted in chloroform/methanol (300μl, 2:1 v/v). Lipid extraction was facilitated by vigorous vortex-mixing, followed by sonication (5 min, RT) and incubation at 60 °C for 20 min. Extracts were centrifuged at 10,000 X g for 5 min and the supernatant was dried and subsequently washed twice with methanol. Dried residues were dissolved in 25 μl of methprep II (alltech) and analyzed directly on an Agilent 7890A-5975C GC-MS system. Samples were injected (split ratio 10:1, injection temperature 250 °C) onto a 30 m + 10 m X 0.25 mm DB-5MS + DG column (J&W, Agilent Technologies) using helium as the carrier gas. The initial oven temperature was 70 °C (1 min), followed by temperature gradients to 230 °C at 17 °C2mins⁻¹, and from there to 325 °C at 25 °Cmin⁻¹. The final temperature was held for 10 min. Data analysis was performed using Chemstation software (MSD Chemstation D.01.02.16, Agilent Technologies). Abundance and label incorporation was calculated as described previously (Saunders et al., 2011). Data shown are the average of three technical replicates and their standard deviation. Statistical significance was evaluated by Wilcoxon Rank Sum Testing with continuity corrections, using the R data analysis package (version 2.14.0), where p values of <0.05 were considered significant.

For LC-MS analysis, lipids were extracted in chloroform: methanol: water (1:2:0.6 v/v). After removal of insoluble material by centrifugation (15,000 x g, 15 min), extracts were dried under nitrogen and dried cellular extracts were resuspended in 100 μl of butanol/methanol (1:1, v/v) containing 5 μM ammonium formate. Cellular lipids were separated by injecting 5-μl aliquots onto a 50 mm × 2.1 mm × 2.7 μm Ascentis Express RP Amide column (Supelco, Sigma) at 35 °C using an Agilent LC 1200. Lipids were eluted at 0.18 ml/min over a 5 min gradient of water/methanol/tetrahydrofuran (50:20:30, v/v) to water/methanol/tetrahydrofuran (5:20:75, v/v), and the final buffer held for 3 min. Lipids were analyzed by electrospray ionization-mass spectrometry (ESI-MS) using an Agilent Triple Quad 6460. The molecular species of each lipid classes were identified using precursor and neutral loss ion scanning from 100 to 1000 m/z in positive ion mode and negative mode. Identified lipid species were quantified using multiple reaction monitoring (MRM) with a 10 ms dwell time for the simultaneous measurements of ~20 to 50 compounds and the chromatographic peak width of 30 sec to 45 sec, the minimum of data points collected across the peak was 12 to 16. Optimized parameters for capillary, fragmentor, and collision voltages were 4000 V, 140–380, and 15–60 V, respectively. In all cases, the collision gas was nitrogen at 7 lmin⁻¹. ESI-MS data was processed using Agilent Mass Hunter Quantitative software (Mulgrave, Australia). Detected lipid species were annotated as follows; lipid class (sum of carbon atoms in the two fatty acid chains: sum of double bonds in the fatty acid chains). The peak area of each lipid species was normalized to abundance of total peak area of lipid species within the lipid class.

Fatty acid complementation

Individual fatty acids (100 mM) were dissolved in ethanol and these stock solutions used to generate mixtures that were dried under a stream of nitrogen. Fatty acid-free BSA (Sigma-Aldrich) was dissolved in PBS to a concentration of 0.5 mM, filter-sterilized and added to dried fatty acids to obtain a FA:BSA molar ratio of 2:1. Samples were sonicated in a water bath until fatty acids were dissolved (~1 hr). Volumes were adjusted with 0.5 mM fatty acid free BSA solution to a final total fatty acid concentration of 250 μM. iΔDEH parasites were grown for 72 hours in T25 HFF cultures in the presence or absence of 0.5 μM ATC. iΔECR parasites were not preincubated with ATc as they die rapidly. Parasites were lyzed, filtered to remove host cell debris and washed twice with PBS before inoculation of 96-well optical plates (5,000 parasites per well). Wells were incubated in 200 μl DMEM supplemented with ATc and BSA/fatty acid complex as indicated. Growth was monitored by following fluorescence using a plate reader (Gubbels et al., 2003).

Table 1.

Percent heavy isotope labeling and abundance of fatty acids derived from iΔDEH parasite grown in media supplemented with [U-¹³C]glucose in the presence or absence of ATc.

Cell type	Percent labeling		Overall abundance (nM)
+/− ATc	−	+	−	+
C14:0	76.28 (± 0.22)	77.48 (± 0.42)	18.69 (± 5.2)	18.02 (± 1.83)
C16:0	44.85 (± 1.09)	39.66 (± 2.04)	86.51 (± 7.78)	101.10 (± 22.14)
C18:0	4.85 (± 0.5)	5.37 (± 2.14)	65.8 (± 3.35)	76.62 (± 6.71)
C20:0	4.43 (±0.73)	Nd	1.71 (± 0.12)	1.56 (± 0.12)
C22:0	17.56 (± 1.9)	14.26 (± 4.4)	0.54 (± 0.05)	0.6 (± 0.04)
C24:0	29.08 (± 1.46)	11.85 (± 0.5)	0.81 (± 0.12)	0.73 (± 0.12)
C16:1	46.44 (± 0.5)	39.66 (± 2.04)	10.19 (±1.23)	16.06 (±1.59)
C18:1	36.04 (± 1.02)	16.58 (± 0.29)	161.62 (± 20.25)	133.19 (± 18.59)
C20:1	49.93 (± 1.82)	19.62 (± 0.79)	9.7 (± 1.26)	4.46 (± 0.6)
C22:1	50.33 (± 2.62)	8.14 (± 1.65)	0.56 (± 0.07)	0.42 (± 0.03)
C24:1	15.86 (± 2.12)	9.30 (± 2.79)	0.45 (± 0.15)	0.56 (± 0.27)
C26:1	68.47 (± 2.3)	19.63 (± 2.3)	6.64 (± 1.8)	0.78 (± 0.04)

Table 2.

Percent labeling and abundance of fatty acids derived from iΔDEH parasite grown in media supplemented with [U-¹³C]acetate in the presence or absence of ATc.

Cell type	Percent labeling		Overall abundance (nM)
+/− ATc	−	+	−	+
C14:0	5.5 (± 1.76)	5.28 (± 2.14)	150.84 (± 7.33)	160.70 (± 2.07)
C16:0	9.6 (± 0.54)	7.5 (± 1.15)	838.53 (± 113.09)	901.4 (± 12.33)
C18:0	7.85 (± 1.45)	8.3 (± 1.01)	588.46 (± 99.32)	717.33 (± 13.59)
C20:0	31.35 (±2.28)	3.25 (±1.885)	13.5 (± 3.5)	14.44 (± 0.96)
C22:0	25.83 (± 1.39)	20.73 (± 4.06)	7.56 (± 1.66)	9.49 (± 1.62)
C24:0	30.05 (± 1.63)	7.59 (± 9.80)	15.53 (± 4.34)	19.39 (± 9.37)
C16:1	11.62 (± 0.9)	10.12 (± 1.61)	142.06 (± 11.58)	174.03 (± 4.43)
C18:1	38.59 (± 1.31)	29.36 (± 0.56)	1712.5 (± 234.32)	1659.11 (± 137.59)
C20:1	83.45 (± 0.112)	55.44 (± 2.77)	118.77 (± 15.20)	77.46 (± 21.92)
C22:1	59.85 (± 1.22)	0.69 (± 0.62)	11.37 (± 2.06)	6.21 (± 1.87)
C24:1	46.4 (± 3.8)	23.4 (± 3.7)	9.97 (± 3.26)	15.4 (±10.03)
C26:1	88.57 (± 8.14)	−3.3 (± 4.65)	55.02 (± 29.46)	4.16 (± 1.61)