Transbilayer Dynamics of Phospholipids in the Plasma Membrane of the Leishmania Genus

Protozoans of the Leishmania genus are the etiological agents of a wide spectrum of diseases commonly known as leishmaniases. Lipid organization of the plasma membrane of the parasite may mimic the lipid organization of mammalian apoptotic cells and play a role in phagocytosis and parasite survival in the mammal host. Here, we analyzed the phospholipid dynamics in the plasma membrane of both the L. (Leishmania) and the L. (Viannia) subgenera. We found that the activity and substrate specificity of the inward translocation machinery varied between Leishmania species. The differences in activity of inward phospholipid transport correlated with the different sensitivities of the various species towards the alkyl-phospholipid analogue miltefosine. Furthermore, all species exhibited a phospholipid scramblase activity in their plasma membranes upon stimulation with calcium ionophores. However, binding of annexin V to the parasite surface was only detected for a subpopulation of parasites during the stationary growth phase and only marginally enhanced by scramblase activation.


Introduction
Protozoans of the genus Leishmania cause a complex disease called leishmaniases, whose clinical manifestations have been divided into three principal types that exhibit different degrees of severity and mortality: cutaneous, mucocutaneous, and visceral [1]. Species of this genus alternate between an intracellular amastigote stage in the mammalian host and a free, flagellated promastigote stage in the gut of the sand fly vector. Lainson and Shaw [2] have divided the genus Leishmania into two subgenera, L. A critical point in the host-parasite interaction involves the attachment of the parasite to the intestinal epithelium and a posterior detachment and migration to the proboscis of the insect vector. In mammals, this host-parasite interaction involves the attachment to and invasion of host macrophages, initially by metacyclic promastigotes and later by amastigotes. Both metacyclic and amastigotes use receptor-mediated phagocytosis for mammalian cell invasion. The phospholipid phosphatidylserine (PS) has been suggested as a crucial ligand of the parasite in this process [3][4][5]. Although recent results suggest that Leishmania parasites lack PS [6][7][8][9], the detection of cell surface lipids by staining with labeled annexin V, a protein that interacts with PS but also with phosphatidic acid and phosphatidylinositol-4,5bisphosphate [10], is consistent with the hypothesis that the protozoan parasite might take advantage of a regulated surfacedisplay of specific lipids to invade and survive in host macrophages.
Typically, the eukaryotic plasma membrane displays an asymmetrical distribution of phospholipids across the bilayer, with PS and phosphatidylethanolamine (PE) restricted to the cytoplasmic leaflet and sphingolipids enriched in the exoplasmic leaflet. This distribution is maintained by energy-dependent inward translocation enzymes called flippases that use ATP hydrolysis to translocate specific lipids against a concentration gradient across the bilayer [11,12]. In addition to these energy-dependent flippases, certain eukaryotic cells contain a phospholipid scramblase: a putative membrane protein that upon activation facilitates a rapid bidirectional movement of phospholipids across the two plasma membrane leaflets, disrupting the lipid asymmetry created by the ATP-dependent flippases [13][14][15].
Not much is known about transbilayer dynamics of phospholipids and the involvement of proteins in the plasma membrane of Leishmania. Previous studies performed with L. (L.) infantum, L. (L.) donovani and L. (L.) tropica have shown an ATP-dependent, inwarddirected flippase activity [16][17][18]. This flippase activity requires at least two plasma membrane proteins: LdMT, a member of the P4 subfamily of P-type ATPases involved in phospholipid transloca-tion, and LdRos3 (the b subunit of LdMT), a member related to the Cdc50 family [19,20]. Studies on L. (L.) tropica and L. (L.) infantum have linked ATP-binding cassette transporters to active trafficking and outward transport of phospholipids across the parasite plasma membrane [21][22][23][24]. Furthermore, stimulation of L. (L.) donovani with calcium ionophores has recently been shown to trigger a phospholipid scramblase activity that facilitates bidirectional movement of phospholipids independent of cellular ATP [20].
Despite the clinical relevance of species of the L. (Viannia) subgenus, there have been few reports on their lipid organization. In this work, we compared phospholipid dynamics in the plasma membrane of the L. (Leishmania) and the L. (Viannia) subgenera. First, we studied the redistribution of fluorescent phospholipid analogues by flow cytometry. We found that the activity and substrate specificity of the inward translocation machinery varies between Leishmania species and that all species exhibit a phospholipid scramblase activity upon stimulation with calcium ionophore. Furthermore, all parasite species could be stained by FITC-labeled annexin V (annexin V-FITC) upon digitoninpermeabilization but binding of annexin V-FITC to the parasite surface was only detected for a subpopulation of parasites during the stationary growth phase and only marginally enhanced by scramblase activation.

Results
Phospholipid transport activity in the plasma membrane of Leishmania parasites To compare the activity and characteristics of lipid transporters in the plasma membrane of various Leishmania species promastigotes, we first examined the amount of inwardly transported 7nitrobenz-2-oxa-1,3-diazole (NBD)-labeled phospholipids, which is protected against albumin extraction after the translocation, by flow cytometry. Experiments were performed at 2uC to suppress endocytosis [16]. Under these conditions, all studied species displayed pronounced differences in the internalization of NBDphosphatidylcholine (NBD-PC), NBD-PS and NBD-PE, but did not internalize NBD-sphingomyelin (NBD-SM) (Fig. 1A).
In combination with fluorescence microscopy, flow cytometry and chemical analysis we previously demonstrated that the increase of the cell-associated NBD fluorescence over time results from the transport of intact NBD-lipids to the cell interior [18,25]. For L. (L.) amazonensis, L. (L.) chagasi, and L. (L.) donovani, rapid and selective internalization of NBD-PC and NBD-PS were observed with transport rates at least 24-and 12-fold higher, respectively, than NBD-SM (Fig. 1A). This rapid and selective internalization was severely inhibited in ATP-depleted parasites (Table 1) (Table 2). Collectively, these data are consistent with the presence of an active inward transport activity in the plasma membrane of both Leishmania subgenera; however, the activity and specificity varies between species.  Phospholipid scramblase activity in the plasma membrane of ionomycin-treated parasites We recently provided evidence for a Ca 2+ -induced phospholipid scrambling activity in the plasma membrane of L. (L.) donovani during ionomycin or thapsigargin stimulation [20]. To elucidate whether this activity is a general feature of Leishmania parasites, we next analyzed the internalization of NBD-labeled phospholipids in the plasma membrane of the five studied species upon ionomycin stimulation ( Fig. 2). NBD-PS was selected for this experiment because of potential role of PS in infection [3,5,26]. NBD-SM was selected as an indicator of nonspecific transport.
To make it easier to visualize the effect, we considered the transport rate of the non-treated control cells as 1, for both NBDlipids. For all five species, stimulation with ionomycin (in a Ca 2+containing medium) enhanced the internalization rate of both NBD-lipids, resulting in transport rates at least 1.6-fold higher than the transport rates of non-treated cells (Fig. 2). For the two L. (Viannia) species, the observed increase in the NBD-lipid internalization rate was not as high as that observed for the L. (Leishmania) species (1.6 to 2.4-fold and 5.7 to 12.3-fold for L. (Viannia) and L. (Leishmania), respectively), but the rate was significantly greater than the rate observed with the untreated controls (P,0.05). In the presence of EGTA (a chelator of extracellular calcium) and BAPTA-AM (a chelator of intracellular calcium), ionomycin treatment had no effect on lipid internalization, indicating that the stimulation is Ca 2+ -dependent and not a non-specific effect of ionomycin addition (Fig. 2). Extraction and chromatographic analysis of the fluorescent lipids after 45 min of incubation at 2uC did not reveal altered lipid metabolism in the presence of ionomycin. Less than 20% of the probes were converted to other NBD-lipid derivatives, irrespective of the parasite species (data not shown). Collectively, these findings are consistent with Ca 2+induced phospholipid scrambling in the plasma membrane of all five Leishmania species.
A Leishmania subpopulation binds annexin V upon entering stationary-phase Next, we investigated the staining of log-and stationary-phase promastigotes with annexin V-FITC, a lipid-binding protein that preferentially interacts with membranes containing anionic phospholipids and that is typically used to detect PS displayed on the cell surface. For all studied species, digitonin permeabilization resulted in strong FITC labeling of the parasites, indicating the presence of at least one class of lipids with annexin V-binding capability (Fig. 3A). Quantitative analysis of propidium iodide (PI)negative promastigotes by flow cytometry (Fig. S1) showed that log-phase parasites hardly exhibited annexin V-binding at their cell surface (Fig. 3B), indicating that under these conditions lipids with annexin V-binding capability are confined to the inner plasma membrane leaflet or intracellular membranes. Upon entering the stationary growth phase, a subpopulation of PInegative promastigotes bound annexin V on its surface (Fig. S1), indicating a shift in the phospholipid distribution in the plasma membrane. This feature was observed for L. (L.) amazonensis, consistent with data previously reported [5], and also for all other tested species (Fig. 3B, compare white and gray bars). Notably, stimulation of log-phase promastigotes by ionomycin resulted only in a marginal increase in the annexin V-positive subpopulation as compared to non-treated, stationary phase cells, except for L. (L.) chagasi (Fig. 3B, compare white and striped bars). This result is in sharp contrast with the lipid analog translocation upon ionomycin stimulation, where .90% of the cell population displayed an increase in lipid analog translocation rates (data not shown).
Finally, we assessed the annexin V-binding to L. (L.) amazonensis parasites harvested in log-phase and in stationary phase as well as to a metacyclics enriched population (Fig. 4). In line with a previous report [5], the metacyclics enriched population displayed a higher frequency of annexin V-positive cells with 8.262.5%    (Fig 4B). Taken together, these data show that scramblase activation by Ca 2+ / ionophore in parasites at the different growth phases is not sufficient to trigger binding of annexin V-FITC to the cell surface of the total parasite population.

Discussion
One characteristic feature of the plasma membrane of eukaryotic cells is the presence of both energy-dependent and energy-independent lipid transporters that regulate the transbilayer distribution of lipids in the two monolayers. Here, we investigated the transbilayer distribution of fluorescent phospholipid analogues in the plasma membrane of Leishmania parasites. We found pronounced differences in the phospholipid internalization rates of different species. Species of the L.  [19,27]. One possible explanation for this discrepancy is that different strains were used in the two laboratories.
The differences in activity and specificity of the inward phospholipid transport correlate with the different sensitivities of the various species towards miltefosine, an alkyl phospholipid analogue of PC used in the treatment of leishmaniasis. The mechanism by which this drug is taken up is closely related to the flippase mechanism by which lipids are translocated across membranes, and these processes seem to involve the same proteins [19,27]. How the differences in activity and specificity of the inward phospholipid transport activities relate to membrane lipid asymmetry remains to be established. Studies in the budding yeast Saccharomyces cerevisiae indicate that lipid asymmetry is created in the Golgi membrane en route to the cell membranes and that the generation of lipid asymmetry involves multiple lipid transporters, including members of the P4 subfamily of P-type ATPases [28,29]. Similar to S. cerevisiae, the sequenced genomes of L. (L.) infantum, L. (V.) braziliensis and L. (L.) major contain multiple P-type ATPases of the P4 subfamily. A future challenge is the mapping of subcellular locations of these transporters to evaluate their transport efficiency and to establish how their activities are regulated during the life cycle of the parasite.
In addition to the ATP-dependent transbilayer transport of specific phospholipids, a previous study on L. (L.) donovani revealed the presence of a Ca 2+ -dependent lipid scramblase activity that can be activated by calcium ionophores [20]. In the present study, a similar scramblase activity was found in all studied species upon stimulation with the calcium ionophore ionomycin. These data suggest that the scramblase activity is a common feature in the genus Leishmania. Although all species had been labeled by annexin V-FITC after permeabilization with digitonin and scramblase activity could be observed in most log phase cells upon activation by the Ca 2+ ionophore ionomycin, the binding of annexin V-FITC to the cell surface of the log-phase parasites upon scramblase activation by Ca 2+ /ionophores occurred in a lower level than that observed for the non-treated stationary phase cells.
On the other hand, stationary-phase cultures of all five Leishmania species contained a subpopulation of annexin V-positive parasites not present in log-phase cells. Enrichment of metacyclic forms of L. (L.) amazonensis displayed an even higher frequency of annexin V-positive cells, suggesting that this population contributes significantly to the increased numbers of annexin V-positive cells in stationary cultures. Notably, stimulation of metacyclics by ionomycin did only marginally increase the annexin V-positive subpopulation. Together, these results suggest that annexin Vbinding lipids are confined to the cell interior of the parasites during the logarithmic growth phase of the parasite. Conceivably, when parasites enter stationary phase, these lipids are trafficked to the plasma membrane in a subpopulation of the parasites, then are translocated to and displayed on the outer leaflet of the plasma membrane. Another possibility is that the change of cell surface exposed lipids in a subpopulation of the parasites during the stationary growth phase involves head group remodeling of membrane lipids. Both possibilities would explain the lack of annexin V-binding to the majority of parasites upon ionomycin treatment despite robust scramblase activation.
Although detected in different levels, the regulated display of annexin V-binding lipids on the outer leaflet of the plasma membrane appears to be a common feature of the Leishmania genus and has a strikingly similarity in phylogenetically distant species of the two subgenera. Our results indicate that the observed lipid trafficking/regulation is an important feature for the parasites life cycle. It has been previously shown that pretreatment of L. (L.) amazonensis and L. (V.) braziliensis with annexin V before exposing the parasites to macrophages decreases their virulence. This suggests that annexin V is blocking a molecule with significant importance for the infection [3,30]. Although a prime candidate, PS seems to be absent from Leishmania parasites [6][7][8].
Further analysis is warranted to define more precisely the phospholipid types exposed on the parasite cell surfaces and to fully uncover the proteins involved in controlling the lipid dynamics of the plasma membrane during the life cycle of Leishmania parasites. The characterization of proteins regulating the transbilayer lipid distribution will allow the dissection of their interactions during parasite host-cell invasion. The observed differences in the lipid translocation activities may reflect different parasite behaviors and could help define the infection course and the selection of drugs for chemotherapeutic treatment, reinforcing the importance of species identification.

NBD-lipid uptake
NBD-lipid uptake experiments were performed essentially as described [20]. Briefly, promastigotes (2610 7 mL 21 ) were preincubated in HPMI supplemented with 6 mM 3-(4-octadecyl)benzoylacrylic acid (Biomol, Hamburg, Germany) and 1 mM phenylmethanesulfonyl fluoride (PMSF) for 15 min at 25uC to block the conversion of NBD-lipids by phospholipases [16]. Unless otherwise indicated, parasites were cooled to 2uC and then labeled with 4 mM NBD-lipid. At the indicated times, 50-mL samples of the cell suspension (10 6 parasites) were transferred to a vial containing 250 mL of HPMI with 1% (w/v) fatty acid-free bovine serum albumin, to extract NBD lipids from the cell surface. The cells were then analyzed by flow cytometry. For analysis of NBDlipid metabolism, lipids were extracted from the cells and culture medium as previously described [32] and separated by thin-layer chromatography using a chloroform/methanol/water mixture (65:25:4, v/v/v). Fluorescent spots were visualized under UV light and quantified with ImageJ 1.44p (Wayne Rasband, NIH, USA).

Metacyclic promastigotes enrichment
Metacyclic promastigotes from L. (L.) amazonensis were enriched from stationary cultures using a modified Ficoll density gradient protocol [33]. Briefly, stationary promastigotes cultures were washed in Phosphate buffered saline (PBS), ressuspended at 5610 8 cells mL 21 in 2 mL serum-free DMEM. Cells were loaded on top of a Ficoll gradient prepared in a Falcon 15 mL conical centrifuge tube and consisting of the following steps: 2 ml 20%, 2 mL 15%, and 2 mL of 10% Ficoll in serum-free M199. After centrifugation at 750 x g for 30 min, metacyclic parasites were recovered from the top of the 10% Ficoll fraction (1 ml), washed in Annexin binding-buffer (without CaCl 2 ) and analyzed for annexin Vbinding. Purification efficiency was assessed based on the flow cytometry methods [5,34] and was greater than 60%.

Flow cytometry
Flow cytometry was performed with a Becton Dickinson fluorescence-activated cell sorter (FACS, San Jose, CA, USA) equipped with an argon laser (488 nm). The data were analyzed with the Cell Quest software (BD Biosciences, San Jose, CA, USA). One microliter of 1 mg mL 21 PI in H 2 O was added to 200 mL of cell suspension just before analysis. Ten thousand cells were analyzed at room temperature with gating during data acquisition. Live cells were selected based on forward and sidescatter gating as well as PI exclusion. The following fluorescence channels (in log scale) were used: FL1 (530/30 nm, FITC, NBD) and FL2 (585/42 nm, PI). Data were analyzed with the Cyflogic (BD Biosciences, San Jose, CA, USA) software and the geometric mean of the fluorescence intensity was calculated.

Microscopy
Confocal laser scanning microscopy (Zeiss LSM 510, Carl Zeiss, Oberkochen, Germany) was performed using a 100-fold magnification (a numerical aperture of 1.3) with an oil-immersion objective. FITC fluorescence was excited by an argon laser at 488 nm and then recorded between 505 and 530 nm. PI fluorescence was excited with a 559 nm HeNe1 laser and then recorded between 560 and 615 nm. Pinholes of 288 and 296 mm were used for the green and red channels, respectively.

Data analysis
All assays were performed in triplicate and data are presented as the mean 6 standard deviation (SD), unless otherwise stated. For NBD-lipid uptake experiments, data were fit to a singleexponential curve with the OriginPro 8 analysis tools (OriginLab Corporation, Northampton, MA USA) using the equation y = A * [1 -exp(-k * t)], where t is the time after NBD-lipid addition, A is the intracellular concentration at steady state, and k is the rate coefficient. The initial influx rate was derived from the product A*k and used consistently for presenting and comparing results. All data comparisons were tested for significance using two-tailed Student t-tests calculated with the GraphPad software; P values,0.05 were considered significant.