Rhoptries are secretory organelles involved in the virulence of the human pathogen Toxoplasma gondii. In the present study we have used HPLC and capillary GLC to isolate and quantify lipids from whole Toxoplasma cells and their purified rhoptries. This comparative lipidomic analysis revealed an enrichment of cholesterol, sphingomyelin and, most of all, saturated fatty acids in the rhoptries. These lipids are known, when present in membranes, to contribute to their rigidity and, interestingly, fluorescence anisotropy measurements confirmed that rhoptry-derived membranes have a lower fluidity than membranes from whole T. gondii cells. Moreover, although rhoptries were initially thought to be highly enriched in cholesterol, we demonstrated that cholesterol is present in lower proportions, and we have provided additional evidence towards a lack of involvement of rhoptry cholesterol in the process of host-cell invasion by the parasite. Indeed, depleting the cholesterol content of the parasites did not prevent the secretion of protein-containing rhoptry-derived vesicles and the parasites could still establish a structure called the moving junction, which is necessary for invasion. Instead, the crucial role of host cholesterol for invasion, which has already been demonstrated [Coppens and Joiner (2003) Mol. Biol. Cell 14, 3804–3820], might be explained by the need of a cholesterol-rich region of the host cell we could visualize at the point of contact with the attached parasite, in conditions where parasite motility was blocked.
Toxoplasma gondii is a widespread opportunistic pathogen causing significant diseases in immuno-compromised individuals and neonates. It is an obligate intracellular parasite, able to actively invade virtually all nucleated mammalian cells. As for the other members of the phylum Apicomplexa (which also include other important human pathogenic species such as Plasmodium, the causative agent of malaria, or Cryptosporidium, responsible for intestinal cryptosporidiosis), the invasion process is mediated by parasite proteins secreted by specialized apical organelles in a co-ordinated manner . These include micronemal proteins, which are first secreted from small (∼50 nm×250 nm) rod-like microneme organelles, to allow the parasite to adhere to its host cell. A second wave of secretion originates from larger (∼100 nm×2500 nm) club-shaped compartments called the rhoptries, to form structures such as the MJ (moving junction), at the interface between the host cell and parasite plasma membranes, or the PV (parasitophorous vacuole), in which the parasite will grow intracellularly.
Proteomic analysis of a rhoptry-enriched fraction from T. gondii has previously identified several proteases, phosphatases and kinases that could, once secreted, be involved in the invasion process and the development of the parasites . For instance, T. gondii-specific kinases such as ROP16 and ROP18, which are secreted to the host cell nucleus and the PVM (PV membrane) respectively, have recently been shown to be crucial for proliferation and virulence [3–6]. Rhoptries are thus key players in the invasion process. Their secreted content not only contains proteins, but also includes lipids. Indeed, secreted membrane-bound vesicles originating from the rhoptries can be seen in parasites blocked for invasion by CytD (cytochalasin D; ). Membranous structures can also be visualized by electron microscopy inside the organelle . However, the exact role played by rhoptry lipids in invasion or in the pathogenicity in general is unknown. Only a few studies so far have attempted to determine the lipid composition of the rhoptries, by using TLC or enzymatic assays to perform quantitative and qualitative analyses on rhoptry-enriched fractions [9,10]. The main conclusion from these studies, was that the rhoptries seemed to be particularly enriched in cholesterol, with a cholesterol to PL (phospholipid) ratio >1, although cholesterol appeared to be non-essential for parasite virulence and host-cell invasion .
In contrast, host cell cholesterol is essential for the entry of Toxoplasma and cholesterol is found at the PVM shortly after invasion . Several studies have also shown that proteins and lipids contained cholesterol-enriched ‘raft’ microdomains on the plasma membrane of the host cell which are selectively incorporated in the PVM during parasite entry [11,12].
In the present study, we sought to determine the lipid composition of the rhoptries using more sensitive techniques than those previously used. Comparative lipid composition and quantitative analyses of whole cells and rhoptry-enriched fractions were performed by HPLC and capillary GLC. We found an enrichment of SAFAs (saturated fatty acids) in the rhoptries, suggesting an elevated rigidity for membranes derived from these organelles. We also found a lower enrichment of cholesterol than claimed previously for these organelles. The role of rhoptry cholesterol was further verified by investigations on rhoptry-secreted vesicles during invasion.
Tachyzoites of the RHΔHX strain, deleted for hypoxanthine guanine phosphoribosyl transferase , were propagated in vitro under standard procedures, by serial passage in HFF (human foreskin fibroblast) monolayers in DMEM (Dulbecco's modified Eagle's medium; Invitrogen) supplemented with 10% FBS (fetal bovine serum). Mass production of tachyzoites was achieved by infecting monolayers of Vero cells grown in DMEM supplemented with 3% (v/v) FBS at 37 °C and under 5% CO2.
Preparation of cell extracts
Tachyzoites for lipid analysis were collected by scraping the cell monolayer, released by passage through a 22 gauge needle and extracellular parasites were filtered on a 3 μm polycarbonate membrane (Whatman) to remove host-cell debris. The parasites were pelleted at 750 g for 10 min and were washed three times in PBS before being immediately frozen at −20 °C under nitrogen. The absence of contamination by cell debris was assessed by treating uninfected Vero host cells similarly and performing a subsequent lipidomic analysis. The amounts of lipids measured from lysed and 3 μm-filtered Vero cells represented at most approx. 4% that of the purified tachyzoites and hence was considered negligible.
Purification of the rhoptry-enriched fraction
Rhoptry-enriched fractions were prepared using a protocol adapted from previously published methods [2,14]. Briefly, ∼1010 purified tachyzoites were washed once in PBS and once in homogenization buffer [250 mM sucrose, 1 mM EDTA, 10 mM Mops (pH 7.2) and 2 mM DTT (dithiothreitol)] at 1300 g for 10 min. Parasites were then resuspended at 5×108 cells/ml in homogenization buffer and disrupted by passage through a French press at 35 kg/cm2. Intact cells and large debris were collected by centrifugation at 1300 g for 5 min, resuspended in homogenization buffer and submitted to a second disruption step with the French press. After a last centrifugation at 1300 g for 5 min, supernatants containing lysed cell contents were pooled. From these supernatants, an organellar fraction was obtained by centrifugation at 15000 g for 30 min. The pellet was resuspended in homogenization buffer and Percoll was added to a final concentration of 30% (v/v). The mix was centrifuged at 27000 rev./min for 30 min in a Beckman Ti-50 rotor. The rhoptry-enriched band was collected near the bottom of the gradient (see Figure 1A). Percoll was washed out by two successive centrifugations in homogenization buffer at 33000 rev./min for 30 min (in a Beckman Ti-50 rotor). All steps were carried out at 4 °C. Pellets were immediately frozen at −20 °C under nitrogen.
Isolation of a rhoptry-enriched fraction from
T. gondii tachyzoites
The purity of the fractions was determined by Western blot and immunofluorescence analysis with antibodies specific for several organellar compartments. Enrichment was calculated by band densitometry measurements using the ImageJ software (National Institutes of Health, U.S.A.) after normalization on the total protein concentration.
Determination of protein concentration
Protein concentrations were determined by the BCA (bicinchoninic acid) assay (Interchim) or the Bradford protein assay (Bio-Rad) according to the respective protocols supplied by the manufacturer.
Western blots were performed as described previously . The primary antibodies used and their respective dilutions were: mouse monoclonal anti-ROP5 diluted 1:1000 (T5 3E2; ), rabbit anti-Hsp28 diluted 1:1000 , rabbit anti-GRA3 diluted 1:1000  and mouse monoclonal anti-Sag1 diluted 1:500 (T4 1E5; ).
NL (neutral lipid) molecular species analysis by GLC
Parasite or rhoptry fractions were homogenized in 2 ml of methanol/5 mM EGTA [2:1 (v/v)] with a FastPrep homogenizer (MP Biochemicals). A 100 μl aliquot of each was evaporated, dry pellets were dissolved in 0.25 ml (or 0.05 ml) of NaOH (0.1 M) for cells (or rhoptries) overnight and used to determine the protein concentration. Lipids corresponding to 0.3 ml of the homogenate were extracted according to the method of Bligh and Dyer  in chloroform/methanol/water [2.5:2.5:2.1 by vol.], in the presence of internal standards: 2 μg of stigmasterol, 2 μg of 1,3-dimyristine, 2 μg of cholesteryl heptadecanoate and 2 μg of glyceryl triheptadecanoate. The chloroform phase was evaporated to dryness. NLs were separated over an SPE column [Strata SI-1 Silica 55 μm, 70 Å (1 Å=0.1 nm), 100 mg]. After washing the cartridge with 2 ml of chloroform, the extract was applied to the cartridge in 20 μl of chloroform, and NLs were eluted with chloroform (2 ml). The organic phase was evaporated to dryness, and then dissolved in 20 μl of ethyl acetate. A 1 μl aliquot of the lipid extract was analysed by GLC on a Focus Thermo Electron system using a Zebron-1 fused silica capillary column [5 m×0.32 mm i.d. (internal diameter), 0.50 μm film thickness; Phenomenex] . The oven temperature was programmed from 200 °C to 350 °C at a rate of 5 °C per min and the carrier gas was hydrogen (0.5 bar). The injector and the detector were at 315 °C and 345 °C respectively.
Lipids corresponding to 0.4 ml of the homogenate were extracted as described above, in the presence of the internal standard ceramide NC15 (2 μg, prepared according to Vieu et al. ), the dried lipid extract was submitted to a mild-alkaline treatment in methanolic NaOH 0.6 M (1 ml), a second organic extraction and then to silylation in 50 μl of BSTFA [N, O-bis (trimethylsilyl)trifluoroacetamide; 1% TMSCl (trimethylsilyl chloride)]-acetonitrile [1:1 (v/v)] overnight at room temperature (25 °C). The extract (5 μl) was directly analysed by GLC  on a 4890 Hewlett Packard system using a RTX-50 fused silica capillary column (30 m×0.32 mm i.d, 0.1 μm film thickness; Restek). The oven temperature was programmed from 195 °C to 310 °C at a rate of 3.5 °C per min and the carrier gas was hydrogen (0.5 bar). The injector and the detector were at 310 °C and 340 °C respectively.
Lipids corresponding to 1 ml of the homogenate were extracted as described above, but without an internal standard. The lipid extract was dissolved in 50 μl of hexane/isopropanol, 82:18 (v/v), 10–45 μl were analysed by HPLC on a Summit system (Dionex) using an Uptisphere 6OH (5 μm) 250 mm×2 mm column (Interchim). The flow rate was 0.25 ml/min and the temperature was kept at 25 °C during all runs. A light-scattering detector was used for the detection (PL-ELS2100, Polymer Laboratories). The evaporator and nebulizator temperatures were kept at 50 and 80 °C respectively, and the nitrogen pressure was 1.8 ml/min. A binary solvent system was used with (A) hexane/isopropanol [82:18 (v/v)] and (B) isopropanol/water [85:15 (v/v)] in the presence of triethylamine [0.014% (v/v)] and acetic acid [0.5% (v/v)]. The gradient profile was started at 5% B for 5 min, was moved to 35% B for 25 min, and then to 85% B for 8 min. Finally, the gradient was returned to the starting conditions for 10 min and the column was equilibrated for 10 min before the next run.
FAME (fatty acid methyl ester) analysis
Lipids corresponding to 0.2 ml of the homogenate were extracted as described above, in the presence of the internal standards glyceryl triheptadecanoate (2 μg), and were transmethylated for 1 h in 14% boron trifluoride methanol solution (1 ml; Sigma) at 55 °C. After the addition of water (1 ml), FAMEs were extracted with hexane (3 ml), evaporated to dryness and dissolved in 20 μl of ethyl acetate. FAMEs (1 μl) were analysed by GLC  on a 5890 Hewlett Packard system using a Famewax fused silica capillary column (30 m×0.32 mm i.d., 0.25 μm film thickness; Restek). The oven temperature was programmed from 110 °C to 220 °C at a rate of 2 °C per min and the carrier gas was hydrogen (0.5 bar). The injector and the detector were at 225 °C and 245 °C respectively.
Membrane fluidity was evaluated with the fluorescent probe DPH (1,6-diphenyl-1,3,5-hexatriene) as described by Shinitzky and Barenholz . According to this method, the greater the extent of probe rotation during its excited state lifetime, the smaller will be the observed fluorescence anisotropy (r), to the extent that r=0 for complete DPH reorientation.
The DPH fluorescence anisotropy was determined by the formula:
where IVV and IVH are the intensities of the emitted light whose plane of polarization is oriented parallel and perpendicular respectively, to the plane of polarization of the excitation beam. G is the correction grating factor, IHV/IHH. The latter is instrument-dependent and accounts for the different sensitivity of photomultipliers to polarized light, measured as the intensities of emitted beams in the vertical (IHV) and horizontal (IHH) directions when the excitation beam is oriented vertically.
Whole tachyzoites, rhoptries or erythrocyte ‘ghosts’ (hypotonic ‘ghosts’ prepared according to the protocol described by Blisnick et al. ) samples corresponding to ∼100 μg of protein/ml were incubated with 2 μM of DPH in PBS at 37 °C for 30 min in the dark, with an optional methyl-β-cyclodextrin treatment (5 mM for 30 min at 37 °C) to deplete cholesterol prior to incubation. Membranes were washed once in PBS before measuring the anisotropy. The DPH anisotropy was measured at 25 °C, using a PerkinElmer LS-55 spectrophotometer, with the excitation and emission monochromators fixed at 355 and 430 nm respectively. Excitation and emission slits (band passes) were set at 2.5 and 5.0 nm respectively. The intensity values for unlabelled samples were subtracted and corrections were made for light-scattering artefacts likely to be found in the organellar or cellular suspensions.
Invasion assays, immunofluorescence and filipin labelling
Invasion assays were carried out by allowing tachyzoites to sediment on confluent HFF for 20 min at 4 °C and subsequently warming them for 2–5 min at 37 °C to trigger invasion. An excess volume of 4% (w/v) paraformaldehyde in PBS was added both to stop the invasion and fix the cells. When needed, CytD treatment was applied, by pre-incubating the parasites with 1 μM of the drug for 20 min before invasion and then allowing them to sediment on HFF monolayers for 20 min at 37 °C, still in the presence of drug. Extracellular parasites were washed and cells were fixed with 4% (w/v) paraformaldehyde in PBS.
Immunofluorescence assays were performed as follows. After fixation, cells were blocked in PBS containing 0.1% (w/v) BSA for 15 min, before incubation with the primary antibody for 1 h in PBS/BSA. Primary antibodies used were mouse monoclonal anti-RON4 antibody (T5 4H1; ) diluted at 1:1000 and rabbit anti-ROP1 antibody (J.-F. Dubremetz and O. Mercereau-Puijalon, unpublished work) at 1:1000. For organelle detection in fraction III, mouse monoclonal anti-ROP2,3,4 (T3 4A7; ) and rabbit anti-GRA3  were used at 1:1000. After three PBS washes and another blocking step in PBS/BSA, the cells were then incubated with the respective Alexa Fluor® 488- or Alexa Fluor® 594-coupled anti-rabbit or anti-mouse secondary antibodies (Invitrogen) at the dilution recommended by the manufacturer.
For the fluorescent labelling of membrane cholesterol, filipin (10 mg/ml in DMSO; Sigma) was used at 50 μg/ml on fixed cells, in addition to the primary and secondary antibodies.
Slides were mounted with Immumount (Calbiochem) and observed with a Leica DMRA2 microscope. Images were acquired with a Micromax YHS 1300 camera (Princeton Instruments) using the Metamorph software (Molecular Devices). Image acquisition was performed on workstations of the Montpellier RIO Imaging facility.
T. gondii tachyzoites
Cholesterol-depleted tachyzoites were obtained as described previously . Tachyzoites were put to invade HFF monolayers, which were subsequently treated with 4 μM lovastatin (Sigma, activated as the dihydroxy acid form) and supplemented with 250 μM mevalonate (Sigma) in complete DMEM for 2 days. Before cell lysis, culture medium was replaced by serum-free DMEM and methyl-β-cyclodextrin (Sigma) was added to a final concentration of 10 mM for 20 min. Cells were scraped and lysed by passage through a 22 gauge needle to free the parasites, which were left in the presence of methyl-β-cyclodextrin for an extra 25 min. Cholesterol depletion was routinely assessed by UV-microscopic observation of filipin-treated parasites. Quantification of the cholesterol concentration in the purified tachyzoites obtained after a typical treatment was assessed by GLC and showed a reduction of ∼70% of parasite cholesterol.
Values were expressed as means±S.E.M. Data were analysed for comparison using unpaired Student's t test with equal variance (homoscedastic) for different samples or paired Student's t test for similar samples before and after treatment. A P value of <0.005 or <0.05 was used as the level of significance.
RESULTS AND DISCUSSION
Preparation of the rhoptry-enriched fraction
Rhoptry-enriched fractions were isolated as previously described [2,10,14]. This cell fractionation on a Percoll gradient allowed us to separate the rhoptries from other subcellular components, as rhoptries were highly enriched in a distinct band near the bottom of the gradient (Figure 1A, fraction III). However, this fraction could occasionally be found to be contaminated with other organelles such as mitochondria or dense granules [2,14]. The protocol used by Bradley et al.  for the proteomic analysis of the rhoptries, providing an improved purity of the rhoptry fraction by the use of a second fractionation on a sucrose gradient, lead to extensive loss of material and could not be implemented for our lipidomic analysis. Instead, we systematically analysed the fractions by Western blot analysis using markers for rhoptries, dense granules, mitochondria and the plasma membrane (Figure 1B). In a typical fraction retained for lipidomic analysis, fraction III was indeed found to be enriched in rhoptry markers, whereas it contained, at most, only traces of mitochondria and plasma membrane components and few contaminating dense granules (Figure 1B). Quantification of the enrichment of each marker was performed by measuring the band density after Western blot analysis (Figure 1C). All markers were found to be present in fraction I, where they would be trapped in incompletely lysed parasites. In addition, the rhoptry marker ROP5 showed a highly specific, yet moderate, enrichment in fraction III. This 6.6-fold enrichment could be explained by the fact that rhoptries occupy a large part of the total cell volume (10–30%, ) and are relatively rich in proteins ( and the present study). Consequently, rhoptry proteins would already be highly represented in total cell extracts. Most importantly, the rhoptry marker showed a more than 20-fold enrichment over the marker from the most commonly contaminating organelle, the dense granules (Figure 1C). This demonstrates the high purity of the rhoptry-enriched fraction. It was further verified by immunodetection and microscopic analysis of fraction III using dense granules (anti-GRA3) or rhoptry (anti-ROP2,3,4)-specific antibodies; quantitative analysis of the labelled microscopic objects showed that dense granules generally represented less than 10% of the organelles present in fraction III (results not shown). These rhoptry-enriched fractions, as well as purified extracellular tachyzoites from T. gondii, were used for qualitative and quantitative lipidomic analyses by HPLC and capillary GLC.
PL and ceramide composition
Analysis of the PL class revealed that rhoptries contain approx. 3.6 times less total PLs per mg of protein than tachyzoites (Figure 2A) and this was the case for all of the PL subclasses. The ratio of the respective PL subclasses to total PL showed some differences between the two extracts (Figure 2B). One common feature is that PC (phosphatidylcholine), a major structural component of membrane bilayers which can be both scavenged by Toxoplasma from its host cell or synthesized by the parasite , is the main PL found both in rhoptries and whole Toxoplasma extracts (Figure 2B). On the other hand, noticeable differences included the absence of PS (phosphatidylserine), reduced amounts of PE (phosphatidylethanolamine) as well as higher amounts of SMs (sphingomyelins) and, less consistently, LPLs (lysoPLs) and PI (phosphatidylinositol) in the rhoptries (Figures 2A and 2B). These results were globally in accordance with previous studies [10,27]. For both whole cells and rhoptry fractions, ceramides were also separated and quantified together with SM. The main species found were short-chain (C16, C18) sphingolipids, but no qualitative difference between the extracts could be reproducibly observed (results not shown). Among the sphingolipids, SMs were consistently found in higher amounts than ceramides (34.9 compared with 5.40 nmoles·mg of protein−1 respectively, for whole cells); it was especially the case for the rhoptries (17.88 compared with 1.81 nmoles·mg of protein−1 respectively), where the proportion of SM among the other PLs was particularly high compared with whole cells (Figure 2B).
Analysis of the PLs present in whole tachyzoite extracts and rhoptry-enriched fractions
Toxoplasma is known to be able to synthesize fatty acids de novo thanks to the presence of the FAS (fatty acid synthase) I and FASII systems, but it is also able to scavenge fatty acids from its host . For the estimation of total fatty acid composition, lipids were extracted from tachyzoites and rhoptry-enriched fractions, to be subsequently analysed as FAMEs. Several of the main FAME species were found in both types of extracts (Table 1): palmitic (C16:0) and stearic (C18:0) acids are the most abundant SAFAs found in tachyzoites or rhoptries; similarly, for both extracts, the most represented UFA (unsaturated fatty acid) is oleic acid (C18:1,n−9; Table 1). However, important qualitative differences were found. First, the rhoptry extracts were found to lack several long chain fatty acids (Table 1). Also, one striking difference was that rhoptry fractions contained high proportions of SAFA compared with the MUFA (mono-UFA) and PUFA (poly-UFA) species (Figure 3). Indeed, our results for tachyzoite extracts show a SAFA/UFA ratio of ∼0.9 (Table 3), in accordance with previously published results  describing a rich content of UFA in Toxoplasma cells, whereas in rhoptries we found a SAFA/UFA ratio of ∼2.24 (Table 3). This is of importance as fatty acid composition is known to be involved, along with cholesterol, in the regulation of the fluidity of the membranes  and this will be discussed below.
|Fatty acid .||Whole tachyzoites (n=3) .||Rhoptries (n=4) .|
|Fatty acid .||Whole tachyzoites (n=3) .||Rhoptries (n=4) .|
Proportions of SAFA, MUFAs and PUFAs compared with the total amount of fatty acids present in the extracts
Qualitative and quantitative analysis of NLs were also performed on tachyzoites and rhoptry-enriched fractions. Here, no qualitative difference was observed, both extracts were found to contain the same types of CEs (cholesterol esters), DAGs (diacylglycerols) and TAGs (triacylglycerols) (Table 2). CEs can be synthesized from cholesterol by Toxoplasma and although it was previously suggested that palmitic acid could be preferentially used to esterify cholesterol in the parasite , the main CE species we found in both extracts was a C18 (Table 2).
|Neutral lipid .||Whole tachyzoites (n=5) .||Rhoptries (n=4) .|
|Neutral lipid .||Whole tachyzoites (n=5) .||Rhoptries (n=4) .|
However, there were quantitative differences. Indeed, whole cells bore elevated amounts of TAGs compared with DAGs, which was not the case in the rhoptry fractions, where both CEs and TAGs were found in lower proportions (Figure 4). CEs and TAGs cannot integrate into the PL bilayer  so they cluster to form the hydrophobic core of the so-called lipid bodies. Lipid bodies are stores of NLs, which can be used for energy production or membrane biogenesis , and those structures are present in Apicomplexa (see  for a review). It is thus likely that the elevated levels of TAGs found in whole cell extracts reflect the presence of these lipids in the storage organelles where they could be the main constituent.
Proportions of cholesterol, CEs, DAGs and TAGs, compared with the total amount of NLs measured in the extracts
Previous studies by us  and others , have reported that rhoptries are cholesterol-rich organelles with an unusually high cholesterol/PL ratio (>1). Although we found significant amounts of cholesterol in the rhoptry-enriched fraction, and a cholesterol/PL ratio higher than for the whole cells (0.34 compared with 0.24 respectively; Table 3) the results of the present study differ from those previous findings. Because of the different methodologies used in past studies to measure the cholesterol concentrations (enzymatic assay and quantitative TLC), we also sought to use one of these alternative approaches (a colorimetric assay based on the cholesterol oxidase) to assess the cholesterol levels in our tachyzoite and rhoptry samples (see Supplementary Table S1 at http://www.BiochemJ.org/bj/415/bj4150087add.htm). Cholesterol concentrations in both whole cells and the cholesterol-enriched fraction (fraction III) were found to be in the same range as the values that were determined by chromatography (Table 2), confirming that our results were consistent. We have no obvious explanation for the differences observed with previously published results, but the reasons could be multiple; for instance, these could possibly come from experimental changes in the culture conditions, or the slightly modified protocol used for the purification of the organelles.
|Organelle .||Cholesterol/PL .||SM/PC .||SAFA/UFA .|
|Organelle .||Cholesterol/PL .||SM/PC .||SAFA/UFA .|
Rhoptry-derived membranes have a high rigidity compared with whole cell membranes
The fluidity of biological membranes is known to be modulated by their lipid composition [30,35,36]. The main modulators are: (i) the cholesterol concentration, when cholesterol is present in large amounts it introduces conformational ordering of the lipid chains and it increases the mechanical stiffness while keeping the membrane fluid; (ii) the SM concentration, SM has the capacity to form intermolecular hydrogen bonds involving its ceramide constituent with other lipids, which is likely to stabilize and rigidify membranes; and (iii) the degree of unsaturation of the hydrocarbon tail, lipids with UFAs, which have kinks in their acyl chains are more likely to be disordered. These three parameters can be reflected by the following three ratios (the higher they are, the more rigid is the membrane) cholesterol/PL, SM/PC and SAFA/UFA respectively. When comparing with membranes from a well-studied cell type such as erythrocytes, whole Toxoplasma cells have lower cholesterol/PL, SM/PC and SAFA/UFA ratios, which could be indicative of a higher membrane fluidity (Table 3) as already described for tachyzoites [29,37]. However, as mentioned previously, lipids from rhoptry-enriched fractions display higher cholesterol/PL, SM/PC and SAFA/UFA ratios than whole tachyzoites (Table 3). When compared with erythrocytes, rhoptry lipids showed lower cholesterol/PL and SM/PC ratios, but on the other hand displayed a much higher SAFA/UFA ratio (Table 3), making it difficult to predict their fluidity.
We thus used the fluorescent lipophilic molecule DPH, which has been established as a probe to measure the fluidity of the lipid bilayer of liposomes, biological membranes and whole cells, by a fluorescence anisotropy (r) assay . DPH partitions equally between the different lipid domains in a membrane, without apparent concentration differences between ordered and disordered lipid domains, and it is thought to be evenly distributed among all of the lipidic regions of a whole living cell. The anisotropy of DPH reflects the capacity of this probe to move in its lipid surroundings and hence, the fluidity of the membranes: the higher the anisotropy value, the higher the rigidity of the membranes. DPH was used on membranes from whole Toxoplasma cells, as well as tachyzoite ‘ghosts’ and rhoptry-enriched fractions from the Percoll gradient (fractions I and III respectively; Figure 1); we also used erythrocyte ‘ghosts’ as a control, as they have been extensively studied. Our results show that although membranes from whole Toxoplasma tachyzoites and parasite ‘ghosts’ seemed to have a low DPH anisotropy value (∼0.15 and 0.17 respectively; Figure 5), indicative of a rather high-membrane fluidity, as previously observed , membranes from the rhoptry-enriched fractions displayed a significantly higher value (∼0.22), similar to the one obtained for erythrocyte ‘ghosts’ (∼0.20). Depletion of membrane cholesterol by methyl-β-cyclodextrin treatment lowered the DPH anisotropy value of membranes from parasite and erythrocyte ‘ghosts’ and whole Toxoplasma (Figure 5). This increase in membrane fluidity triggered by the depletion of cholesterol has been documented previously for the erythrocytes  and can be explained by the order-inducing effect of this molecule for the disordered (i.e. unsaturated) acyl chains of the PL . In contrast, a less consistent effect on DPH anisotropy was observed when membranes from the rhoptry-enriched fractions were treated with methyl-β-cyclodextrin (Figure 5). This could not be explained by a low cholesterol concentration, as the cholesterol/PL ratio is higher for rhoptries than for whole Toxoplasma cells for example. Instead, the presence in rhoptry membranes of high concentrations of PL with rigid acyl chains (i.e. saturated), could create a highly ordered domain where the presence of cholesterol would have a lesser effect on rigidity, or would even rather favour membrane fluidity by weakening van der Waals' interactions between hydrocarbons chains of fatty acids and preventing crystallization . Overall, our results suggest that, in contrast with the global fluidity of the membranes from the whole Toxoplasma cell, total membranes from rhoptries (possibly including internal membranous structures ) appear to have a high membrane rigidity, which could be explained by the presence of SAFAs in ordered lipid domains.
Analysis of membrane fluidity by analysis of DPH anisotropy in
T. gondii tachyzoites, T. gondii ghosts from fraction I, rhoptries isolated from fraction III; values for erythrocyte ghosts are shown as a reference
Cholesterol and cell invasion
Several microbial pathogens use cholesterol-rich domains (for instance ‘raft’ domains or caveolae) as a gateway for entry into their host cells [41,42]. During the process of Toxoplasma invasion, the limiting membrane from the parasite-containing PV, which is largely derived from the plasma membrane of the host cell but also contains rhoptry-derived material, is known to be rich in cholesterol . Hence, the presence of cholesterol in the rhoptries was initially though to be of importance for the virulence of the parasite through putative secretion at the plasma membrane or for insertion into the nascent PV . These hypotheses have been tested and cholesterol-depletion assays have shown that parasite cholesterol (in particular from the rhoptries) was not necessary for infection, whereas host-cell cholesterol was . Our lipidomic data, indicating lower amounts of rhoptry cholesterol, could also point towards a low impact of rhoptry cholesterol from the organelle on the virulence of the parasite. Thus we wanted to investigate more precisely whether cholesterol could be found in secreted rhoptry material. To assess this, we took advantage of specific experimental conditions where the parasite mobility is blocked by CytD (a potent inhibitor of actin polymerization). In these conditions, tachyzoites cannot invade, although they can bind to host cells, form an MJ and secrete protein-rich rhoptry-derived vesicles that can be visualized in the cytosol of the host cell and are termed ‘evacuoles’ . We could label the evacuoles through the rhoptry protein marker ROP1, and perform co-localization experiments with filipin, a fluorescent dye specific for membrane-bound cholesterol (Figure 6A). As observed previously, CytD-treated parasites displayed an apical punctate signal for the rhoptry neck protein RON4, showing that the parasites had successfully attached and formed the MJ . Additionally, ROP1 was found to label clustered evacuoles forming a network extending into the host-cell cytoplasm . Labelling of the membrane-bound cholesterol with filipin highlighted the plasma membrane of the parasite and of the host cell (Figure 6A). An intense filipin labelling could often be seen at the interface between the parasite and the host cell, which could either correspond to the apical part of the parasite or to the plasma membrane of the host cell. If present at the apex of the parasites, the filipin labelling could possibly be internal to the rhoptries; however, it is of note that the organelles were not found to be labelled in extracellular parasites, suggesting they would possess a low amount of membrane-bound cholesterol, in accordance with our lipidomic data (Table 2).
Cholesterol localization during invasion in CytD-treated tachyzoites
When the evacuole clusters were close to the plasma membrane (suggesting an early time point after contact, as the evacuole network is known to extend further into the host cell with time ), both evacuoles and the MJ were in the close vicinity of the cholesterol-rich region, even showing some partial co-localization at the point of contact of the parasite (Figure 6A). However, when the network of ROP1-positive evacuoles was more developed inside the host cell cytoplasm, no obvious co-localization could be seen between the farthest evacuoles and filipin (Figure 6A, bottom row), suggesting that these evacuoles are indeed not enriched in cholesterol.
To further analyse this process, we prepared cholesterol-depleted parasites by growing them into fibroblasts that had been treated with statin [an inhibitor of HMG-CoA (3-hydroxy-3-methylglutaryl-CoA) reductase in the cholesterol biosynthesis pathway] and methyl-β-cyclodextrin (a cholesterol-segregating drug) to lower considerably (by ∼70%) their cholesterol content . These parasites were used to assess whether the cholesterol seen at the Toxoplasma–host-cell interface is of parasite origin, by performing the same type of CytD-blocked invasion assay as described above. In these conditions, cholesterol could still be seen at the interface between the parasite and the host-cell plasma membrane, whereas the parasite plasma membrane appeared devoid of peripheral or internal filipin labelling, suggesting efficient and considerable reduction of cholesterol (Figure 6B). Both the MJ and the evacuoles were still found to be forming with cholesterol-depleted parasites and apart from the point of contact, no consistent co-localization was found with the filipin signal (Figure 6B). This would mean that the cholesterol-rich membranes visualized at the interface between the host and parasite during the initial stages of invasion are of host origin. It could either be that the parasite is preferentially selecting cholesterol-rich domains of the plasma membrane of the host to initiate invasion, or that Toxoplasma could trigger the recruitment of host cholesterol at the site of invasion. This could be mediated simply by lateral diffusion through the membrane, by vesicle-mediated transport or a soluble protein carrier (see  for a review).
Finally, as filipin-labelled cholesterol was found close to the MJ at the point of contact between Toxoplasma and the host cell, we wanted to assess whether the junction remained bound to cholesterol-rich membranes during parasite invasion. Invasions were then performed without CytD (Figure 7), to allow the active penetration of the tachyzoites. Filipin was generally not found to be co-localizing with the MJ throughout the invasive process and with the residual junction found on the recently closed PVM , although it was found together with ROP1 at the PVM as described previously .
Fate of cholesterol during invasion in permissive conditions
Not only has the PVM been found to be enriched in cholesterol at the time of invasion by Toxoplasma , but it has also been shown to incorporate several other host cell membrane reporter molecules belonging to the cholesterol-rich raft [i.e. GPI (glycerophosphatidylinositol)-anchored proteins CD55, GPI-ICAM1 (GPI-intercellular adhesion molecule 1), GM1 ganglioside or DiIC16 lipid], and also to non-raft lipid domains (i.e. tail-less ICAM1 protein, FAST DiOC18 lipid) . Although the MJ could play a role in sorting the protein components of raft domains (such as GPI-anchored proteins) to be incorporated into the PVM , we have shown that the MJ itself does not seem to co-localize with cholesterol-rich domains of the host cell during invasion. In Plasmodium, several raft-associated host proteins seem to be dispensable for parasite infection, but total raft depletion blocks parasite entry . Similarly, depletion of cholesterol from the host cell totally blocks invasion by Toxoplasma , suggesting that host lipid rafts could have a role in invasion. Our finding that cholesterol-rich membranes can be visualized at the interface between Toxoplasma and its host cell seems to go along the same line. However, the fact that they can still be observed when using cholesterol-depleted parasites, that the parasite-derived evacuoles do not seem to be enriched in cholesterol and that, as observed before , cholesterol-depleted Toxoplasma can still invade its host cell suggests little or no involvement of rhoptry cholesterol in invasion.
The lipid composition of Toxoplasma tachyzoites and of their rhoptries has been the object of a few past studies, which were aimed at determining the composition and properties of the membranes from this pathogen and these organelles that are considered important for virulence [9,10,37]. With improved technological tools at hand, the results of our present study provide a more exhaustive analysis and substantial additional information. The present study revealed several features of the rhoptries lipidome: (i) globally, we found a lower lipid to protein ratio in the rhoptries than for the whole tachyzoites, suggesting a rather rich protein content of the organelle; (ii) choline-containing PLs (i.e. PC, SM) appear to be highly represented, with a lower percentage of PE and no detectable PS; (iii) low levels of TAGs probably reflect an absence of lipid bodies in the organelle; (iv) fatty acid analysis of the whole lipids revealed that there was a high proportion of short chain and highly saturated fatty acids; (v) in contrast with previous studies, we did not find a particularly high enrichment of cholesterol in the rhoptries.
Consistent with this, our study confirmed that the role played by cholesterol in the rhoptries does not appear to be essential for the virulence of the tachyzoites, in contrast with host-cell cholesterol which seems to be present at the contact point between parasite and the plasma membrane during early stages of invasion. A major challenge is now to understand the precise role of host-cell cholesterol and the molecular interactions between host and parasite lipids during invasion.
On the other hand, cholesterol may have, along with SM and the particularly abundant SAFA, a structural role for the organelle. Indeed, the presence of such lipids, assuming they are incorporated into the rhoptry membranes, is likely to affect their physical properties, which, as we have shown, are rather rigid compared with the global membranous content of the parasite, and this might contribute to the peculiar elongated morphology of rhoptries. In contrast, high membrane fluidity of the tachyzoites might be essential for the plasticity of their plasma membrane, as they undergo extensive deformability during host-cell invasion. It might also facilitate the movement of surface-adhesive proteins which are translocated along the parasite cell body during the invasion process.
Finally, our results reflect a global lipidomic analysis of the organelle and, obviously, there must be local variations within membrane domains of the rhoptries. For instance, these organelles appear to be highly polarized, as they harbour proteins that are specifically localized to their neck (RON) or bulbous (ROP) parts, which seem to have different roles for invasion or the establishment of the parasite in the PV and hence could be secreted sequentially in time. Thus rhoptry lipids might be involved in sorting and segregation of these proteins. There are several examples where the lipid composition of the membranes drives protein sorting and trafficking, such as in the photoreceptor rod, where there is a cholesterol and fatty acid unsaturation gradient conferring different physical properties and functional roles to the organelle , or with the cis-to-trans cholesterol gradient of the Golgi . The lipids we detected in the rhoptries might also be present in internal structures that could be involved in the packaging of the proteins. Indeed, solubilization treatments of the organelle show that many rhoptry proteins could be embedded in membranes . Some of these might be resident proteins of the rhoptries and structural components of the membrane, but others such as ROP2 , are known to be secreted after invasion. Hence an additional role of rhoptry lipids could be organizing and packing the insoluble rhoptry proteins as lipoproteins or within vesicular membranes, before secretion into the host cell. Rhoptry cholesterol does not appear to be essential for secretion of RON or ROP proteins, but determining the lipid subdomains within the rhoptry and the putative role of local lipids in sorting or secreting those proteins is certainly a future objective to further characterize the many virulence factors present in the organelle.
Our thanks go to Dr S.O. Angel (Laboratorio de Parasitología Molecular, IIB-INTECH, CONICET-UNSAM, Buenos Aires, Argentina) for contributing antibodies. We also gratefully acknowledge Veronique Roques for helpful assistance at the lipidomic platform. This work was supported by the CNRS, INSERM and ANR grant 06-MIME-024-01.
Dulbecco's modified Eagle's medium
fatty acid methyl ester
fetal bovine serum
human foreskin fibroblast
internal diameter, LPL, lysophospholipid
monounsaturated fatty acid
polyunsaturated fatty acid
saturated fatty acid
unsaturated fatty acid