Alteration of Fatty Acid and Sterol Metabolism in Miltefosine-Resistant Leishmania donovani Promastigotes and Consequences for Drug-Membrane Interactions

Abstract

Miltefosine (hexadecylphosphocholine [HePC]) is the first orally active drug approved for the treatment of visceral leishmaniasis. In order to investigate the biochemical modifications occurring in HePC-resistant (HePC-R) Leishmania donovani promastigotes, taking into account the lipid nature of HePC, we investigated their fatty acid and sterol metabolisms. We found that the content of unsaturated phospholipid alkyl chains was lower in HePC-R parasite plasma membranes than in those of the wild type, suggesting a lower fluidity of HePC-R parasite membranes. We also demonstrated that HePC insertion within an external monolayer was more difficult when the proportion of unsaturated phospholipids decreased, rendering the HePC interaction with the external monolayer of HePC-R parasites more difficult. Furthermore, HePC-R parasite membranes displayed a higher content of short alkyl chain fatty acids, suggesting a partial inactivation of the fatty acid elongation enzyme system in HePC-R parasites. Sterol biosynthesis was found to be modified in HePC-R parasites, since the 24-alkylated sterol content was halved in HePC-R parasites; however, this modification was not related to HePC sensitivity. In conclusion, HePC resistance affects three lipid biochemical pathways: fatty acid elongation, the desaturase system responsible for fatty acid alkyl chain unsaturation, and the C-24-alkylation of sterols.

Leishmania spp. are the protozoan parasites responsible for leishmaniases, a family of diseases including a variety of clinical manifestations classically labeled as visceral, cutaneous, and mucocutaneous leishmaniases (9). Chemotherapy is the main tool for the control of leishmaniasis; however, the standard treatments, including pentavalent antimonials, amphotericin B (AmB), and paromomycin, are toxic and expensive. Furthermore, the development of resistance to antimonials and the lack of efficacy of these treatments against coinfections with human immunodeficiency virus are additional problems which necessitate the development of new drugs (5). Miltefosine (hexadecylphosphocholine [HePC]), an alkylphosphocholine first used for its anticancer properties, is in development for the treatment of visceral leishmaniasis (30). HePC proved to be the first drug orally active against antimony-resistant cases (11) and also against cutaneous leishmaniasis (29). Because HePC was registered in India in 2003 for the treatment of visceral leishmaniasis, it is of interest to investigate its mechanism of action and the possibility of resistance in Leishmania. However, not much information is available about the leishmanicidal mechanisms of HePC and other alkyl-lysophospholipids. These compounds have been described to cause alterations in alkyl-lipid metabolism and in the biosynthesis of alkyl-anchored glycoproteins (12). We recently described the ability of HePC to induce an apoptosis-like cell death in Leishmania donovani promastigotes (16), an effect which was confirmed to occur in amastigote forms of L. donovani (32). In addition, an HePC-resistant (HePC-R) line has been selected by drug pressure and cloned, supplying biological material to study the events responsible for HePC resistance (28). Defective inward translocation of the drug has been demonstrated to occur in HePC-resistant L. donovani (19), and the HePC transporter has been cloned and characterized as a P-type ATPase (18).

In the present paper, we describe modifications in the lipid compositions of membranes from HePC-R Leishmania donovani promastigotes from those of wild-type (WT) promastigotes which will help to identify biochemical targets potentially affected in HePC resistance. Furthermore, we have studied HePC-membrane interactions using a biomimetic phospholipid monolayer model in order to demonstrate the possible consequences of these modifications.

MATERIALS AND METHODS

Chemical compounds.

HePC (miltefosine), was kindly provided by Zentaris (Frankfurt, Germany). A mother solution of HePC was prepared daily in distilled water at 10⁻³ M and was immediately used for measurements after dilution as necessary. AmB, bis(trimethylsilyl)trifluoroacetamide, boron trifluoride etherate, 1-palmitoyl-2 oleoyl-sn-glycero-3-phosphatidylcholine (POPC), 1-palmitoyl-2 oleoyl-sn-glycero-3 phosphatidylethanolamine (POPE), 1,2 dipalmitoyl-sn-glycero-3 phosphatidylcholine (DPPC), 1,2 dipalmitoyl-sn-glycero-3 phosphatidylglycerol (DPPG), and 1,2-dipalmitoyl-sn-glycero-3-phosphatidylethanolamine (DPPE) were purchased from Sigma (Saint-Quentin Fallavier, France). For the surface pressure measurements, POPC, POPE, DPPC, and DPPG were dissolved in a chloroform-ethanol (4/1, vol/vol) mixture at a concentration of 10⁻³ M. DPPE was dissolved in a chloroform-methanol (4/1, vol/vol) mixture. Chloroform, ethanol, and methanol were purchased from Sigma (St. Louis, MO), were 99% pure, and were used without further purification.

Parasite strains and culture.

Promastigote forms of a wild-type Leishmania donovani (MHOM/ET/67/HU3) clone and an HePC-R clone, previously named R40 (28), were grown in M-199 medium supplemented with 40 mM HEPES, 100 μM adenosine, 0.5 mg/liter hemin, 10% heat-inactivated fetal bovine serum, and 50 μg/ml gentamicin at 26°C in a dark environment. HePC-R promastigotes were continuously maintained in the presence of 40 μM HePC. All experiments were performed with parasites in their logarithmic phase.

Yeast strains and culture.

A Saccharomyces cerevisiae wild-type strain and S. cerevisiae S-adenosyl-l-methionine-C-24-delta-sterol-methyl-transferase (SCMT1) (ERG6) deletion strain BKY485-C (Δleu2-3 ura3-52 erg6Δ:LEU2) were kindly given by M. Bard, Biology Department, Indianapolis University. The latter strain was used as the recipient strain for transformation with SCMT1 cDNA (20). WT S. cerevisiae was grown on liquid yeast extract-peptone-dextrose (YPD) complete medium containing 1% yeast extract (Difco), 1% Bacto peptone (Difco), and 2% glucose at 30°C.

Assessment of the role of SCMT in HePC sensitivity.

Assays of the susceptibilities to HePC and AmB of WT S. cerevisiae, the SCMT-null mutant, and a Leishmania SCMT1-transformed mutant were performed with YPD medium in flat-bottomed 96-well microtiter plates at 30°C in a 200-μl final volume. Briefly, logarithmic-phase cultures of yeast were diluted in YPD medium, and 195 μl was added to each well to give a final density of 10⁴ cells/ml. HePC was solubilized in water, whereas AmB was solubilized in dimethyl sulfoxide, and dilutions were performed in YPD medium before addition to the well in a volume of 5 μl; the control series received YPD medium plus 0.5% dimethyl sulfoxide. The plates were incubated at 30°C for 48 h, and the absorbance at 620 nm was read at 24 and 48 h with a microplate reader. Drug-free wells were prepared and served as references for readings of the MICs, defined as corresponding to an 80% reduction in turbidimetry.

Cell fractionation and identification of plasma and mitochondrial membranes.

Cell fractionation by differential centrifugation was performed according to the method of Hasne and Lawrence (8). Briefly, promastigotes were grown in 1 liter of medium as described above, and for HePC-R parasites, drug pressure was stopped in the subculture preceding the experiment to avoid drug contamination. Parasites were harvested by centrifugation at 5,000 × g for 5 min at 4°C and then washed twice with cold phosphate-buffered saline and resuspended in 100 mM Tris-HCl, pH 7.5, containing 1 mM EDTA and 250 mM sucrose. Parasites were lysed by three cycles of freezing (−170°C) and thawing (37°C). Parasite breakage was assessed by phase-contrast microscopy. The lysate was fractionated by differential centrifugation at 2,100 × g for 10 min, 15,800 × g for 10 min, and 100,000 × g for 1 h. The pellets were resuspended in Tris-HCl supplemented with 250 mM sucrose to give pellet fractions P1, P2, and P3, respectively, and the final supernatant, S3. Acid phosphatase was used as both a plasma membrane (tartrate-resistant) and a cytosolic (tartrate-sensitive) marker. Membrane-bound and soluble acid phosphatases were measured at 37°C with 5 mM p-nitrophenyl-phosphate in 50 mM acetate buffer, pH 5, in the presence and absence of 200 mM tartrate, respectively. Tartrate-resistant acid phosphatase was recovered mainly in the P1 fraction, whereas tartrate-sensitive acid phosphatase was found mainly in the S3 fraction. Cytochrome c oxidase activity was used as a marker for mitochondria. Cytochrome c oxidase was assayed spectrophotometrically by measuring the decrease in absorbance at 550 nm of a solution containing 35 μM reduced cytochrome c in 30 mM sodium phosphate, pH 7, and various amounts of the enzyme fraction. Cytochrome c oxidase was found in the P2 fraction.

Lipid determination.

Leishmania donovani promastigotes were cultured in Erlenmeyer flasks at an initial density of 10⁶ promastigotes/ml in 1 liter of medium as defined above. Flasks were placed in an orbital incubator under continuous shaking (150 rpm) at 27°C. When parasite cultures reached a density of 15× 10⁶ to 20 × 10⁶/ml (logarithmic phase), promastigotes were harvested by centrifugation and washed three times with large volumes of cold phosphate-buffered saline (pH 7.5). The resulting pellet was resuspended in 20 ml of dichloromethane-methanol (2:1, vol/vol) for about 24 h at 4°C. After centrifugation (11,000 × g, 1 h, 4°C), the extract was evaporated under vacuum. This is referred to as the total membranes. Pellets P1 and P2 from the cell fractionation were treated with dichloromethane-methanol in the same way to yield plasma membrane and mitochondrial membrane fractions, respectively.

Sterol determination.

The residues obtained as described above were saponified with 30% KOH in methanol at 80°C for 2 h. Sterols were extracted with hexane, and after evaporation, the residue was dissolved in dichloromethane. An aliquot of clear yellow sterol solution was added to 2 volumes of bis(trimethylsilyl)trifluoroacetamide, and the sealed tubes were heated at 80°C for 1 h. The trimethylsilyl ethers of sterols were subjected to gas chromatography/mass spectrometry (MS) analysis.

Fatty acid determination.

The aqueous phases of the methanolic KOH extracts were acidified with 6 N sulfuric acid to pH 3. Fatty acids were extracted with hexane and transesterified by boron trifluoride methanol at room temperature for 2 h. Methyl esters were extracted with hexane and redissolved in methanol-diethyl ether (1:1, vol/vol) prior to gas-liquid chromatography/MS analysis. Gas-liquid chromatography was performed with a Varian model 3400 chromatograph equipped with DB5 columns (methyl/phenylsiloxan ratio, 95/5; dimensions, 30 m by 0.25 mm). The gas carrier was helium (1 ml/min). The analysis conditions were as follows: the column was kept at 270°C, the injector was kept at 300°C, and the detector was kept at 300°C. The linear gradient for methyl esters was from 150 to 180°C at 10°C/min. MS conditions were 280°C, 70 eV, and 2.2 kV.

Monolayer study.

Monolayers of phospholipids were used as models of the external-face biological membranes. Phospholipids are amphiphilic molecules which spread at the air-water interface to form a homogenous stable monolayer. After compression of the monolayer, we obtained isotherms (pressure/area per molecule). Thereafter, exogenous molecules were injected into the subphase below the monolayer to study the interactions of these molecules with the phospholipid monolayer.

Monolayers were prepared as previously described (1) in a Teflon trough supplied by Riegler and Kirstein GmbH (Wiesbaden, Germany). The trough (6.2 by 26.3 by 0.5 cm) was filled with Millipore-purified water (pH 5.6). Twenty microliters of phospholipid solution was spread at the air/water interface. The surface pressure was measured by the Wilhelmy method, by means of a thin plate of filter paper. An electronic device enabled us to keep the surface pressure constant by controlling the displacement of the barriers. The speed of movement of the barriers (3 × 10⁻² cm · s⁻¹) was kept constant during the experiments. All experiments were performed at 21 ± 1°C.

In order to study the interaction of HePC with lipid monolayers, the following method was used.

The monolayer of phospholipid was compressed up to 25 mN/m. Generally, the surface pressure of a biological membrane is estimated at 30 mN/m, but reliable measurements must be done at 25 mN/m, because 30 mN/m is too close to the collapse pressure of HePC (22).

In a first set of experiments, the surface pressure was kept constant at 25 mN/m and an aqueous solution of HePC was injected with a microsyringe under the monolayer at a final concentration between 0.2 μM and 4 μM, according to the method previously described (10). If an interaction occurred between the molecules of the subphase and the monolayer, the barriers were moved back to keep the pressure at 25 mN/m and the variation in the mean molecular area (ΔA) of lipid as a function of time was recorded over 60 min (adsorption kinetics).

In a second set of experiments, the surface area of the trough (corresponding to an initial surface pressure of 25 mN/m) was kept constant. An aqueous solution of HePC was injected under the monolayer of lipids at a final concentration between 0.2 μM and 4 μM. The variation of surface pressure (ΔP) of the lipid in the presence of HePC was recorded over 60 min.

RESULTS

Analysis of alkyl chains from fatty acids.

Fatty acids were recovered from the saponification of phospholipids, and membrane analysis of the alkyl chains in the fatty acid methyl esters was performed. In the total membrane, we identified and quantified 33 fatty acids (Table 1). Three of them represented 60% of the total fatty acid content in WT total membranes; these were hexadecanoic acid (C_16:0), octadecanoic acid (C_18:0), and 9-octadecenoic acid (C_18:1). When comparing the percentages of fatty acid alkyl chains in total, plasma, and mitochondrial membranes with those in the WT, we found that the unsaturated fatty acid content in mitochondrial membranes was only half that in total membranes (27.90% versus 54.25%). Plasma membranes also showed a significant reduction in their content of these fatty acids (36.53% versus 54.25%). It is commonly assumed that an increase of unsaturated alkyl chains increases membrane fluidity, so it was expected that mitochondrial membranes would be less fluid than the plasma membranes in WT Leishmania. Since the total membranes had a higher content of unsaturated fatty acids, we could deduce that the other internal membranes (i.e., the endoplasmic reticulum) would be more fluid than the plasma membranes. In HePC-R parasites, this effect was lessened, since plasma and mitochondrial membranes contained similar proportions of unsaturated alkyl chains. Total and plasma HePC-R parasite membranes exhibited lower unsaturated fatty acid contents than those of WT parasites (36.73% versus 54.25% and 28.88% versus 36.53%, respectively), suggesting that total and plasma membranes from HePC-R parasites were less fluid than those of WT parasites. Moreover, some significant differences were observed in the lengths of the alkyl chains. C₁₈ fatty acids, which were the most abundant fatty acids in both clones, were present in total HePC-R parasite membranes in amounts only half those in WT membranes (Table 2). For saturated C₁₈ fatty acids, we found a reduction of 36% in HePC-R parasite plasma membranes, compared with the level in WT parasites. In mitochondrial membranes, the contents of unsaturated alkyl chains were similar in HePC-R and WT parasites, suggesting similar extents of fluidity of mitochondrial membranes from the two clones (Table 1). In general, in both WT and HePC-R parasites, plasma and mitochondrial membranes had lower contents of C_18:1δ9 than did the total membranes. In contrast, the percentage of C₁₆ fatty acids was twice as high in HePC-R parasite total membranes than in WT parasites. Overall, the amounts of fatty acids from C₁₃ to C₁₆ were significantly higher in HePC-R parasite membranes than in WT, whereas those from C₁₈ to C₂₀ were lower in HePC-R parasites than in WT parasites. These results could be ascribed to a partial inhibition of fatty acid alkyl chain elongation. Moreover, we observed other significant differences between WT and HePC-R parasites. Thus, the amounts of 9-methyl-tetradecanoic acid, 14-methyl-pentadecanoic acid, and 14-methyl-hexadecanoic acid were higher in HePC-R parasites than in WT parasites (10, 7, and 3 times, respectively) (Table 1). Thus, the 9- and 14-methylation of alkyl chains occurred more easily in HePC-R parasites than in WT parasites, whereas the 17-methylation detected in WT parasites (17-methyl-octadecanoic acid) was not observed in HePC-R parasites. These methylations could perturb membrane lipid interactions by preventing hydrophobic interactions between alkyl chains. However, the relative amounts of these compounds were less than 5% of the total fatty acid content, so their importance in membrane fluidity can be considered negligible.

TABLE 1.

Identification and quantification of fatty acid methyl esters obtained from phospholipid saponification in total, plasma, and mitochondrial membranes from L. donovani promastigotes

Fatty acid	RTa(h)	Mean % of total fatty acids ± SD (n = 3)b
		L. donovani WT			L. donovani HePC-R
		TM	PM	MM	TM	PM	MM
Undecanoic acid	7.01	ND	ND	ND	0.28 ± 0.02	0.24 ± 0.03	0.32 ± 0.03
Dodecanoic acid	7.78	0.90 ± 0.06	4.81 ± 0.32	7.81 ± 0.82	0.33 ± 0.02	1.16 ± 0.12	1.71 ± 0.19
Tridecanoic acid	8.12	0.18 ± 0.01	0.22 ± 0.02	0.34 ± 0.05	0.24 ± 0.03	0.27 ± 0.02	0.29 ± 0.03
3-Hydroxy-decanoic acid	8.57	ND	ND	ND	0.71 ± 0.04	0.61 ± 0.05	0.62 ± 0.07
12-Methyl-tridecanoic acid	8.63	ND	ND	ND	1.65 ± 0.13	1.53 ± 0.12	1.58 ± 0.17
Tetradecanoic acid	8.82	2.83 ± 0.21	6.99 ± 0.58	7.05 ± 0.59	3.00 ± 0.41	3.21 ± 0.34	3.64 ± 0.41
3-Hydroxy-hexadecanoic acid	8.89	ND	ND	ND	1.12 ± 0.15	1.16 ± 0.10	1.10 ± 0.12
9-Methyl-tetradecanoic acid	9.12	1.35 ± 0.10	1.27 ± 0.12	1.37 ± 0.16	14.49 ± 1.13	13.37 ± 1.31	14.17 ± 1.58
Pentadecanoic acid	9.28	0.54 ± 0.06	0.78 ± 0.08	0.67 ± 0.07	1.31 ± 0.12	1.45 ± 0.15	1.40 ± 0.13
14-Methyl-pentadecanoic acid	9.57	0.36 ± 0.04	0.45 ± 0.05	0.39 ± 0.04	2.48 ± 0.21	2.56 ± 0.28	2.45 ± 0.27
9-Hexadecenoic acid	9.66	1.82 ± 0.12	1.76 ± 0.14	1.89 ± 0.21	6.90 ± 0.45	6.78 ± 0.52	6.90 ± 0.67
Hexadecanoic acid	9.74	7.36 ± 0.65	10.75 ± 1.24	17.88 ± 1.41	9.67 ± 0.83	9.12 ± 0.76	13.30 ± 1.18
2-Hexyl-cyclopropaneoctanoic acid	9.92	ND	ND	ND	2.67 ± 0.18	2.22 ± 0.24	2.16 ± 0.29
14-Methyl-hexadecanoic acid	10.02	0.99 ± 0.08	1.26 ± 0.13	1.11 ± 0.12	3.10 ± 0.41	2.91 ± 0.22	3.04 ± 0.35
Heptadecanoic acid	10.18	1.37 ± 0.21	1.30 ± 0.15	1.45 ± 0.12	1.39 ± 0.13	1.28 ± 0.14	1.22 ± 0.18
17-Octadecenoic acid	10.40	ND	ND	ND	0.42 ± 0.05	0.64 ± 0.07	0.56 ± 0.06
9-Octadecenoic acid	10.56	32.93 ± 2.96	16.91 ± 1.23	7.02 ± 0.62	18.19 ± 2.23	10.90 ± 1.08	8.86 ± 0.91
Octadecanoic acid	10.65	20.55 ± 2.17	28.31 ± 2.09	24.45 ± 2.71	11.40 ± 0.96	18.07 ± 1.96	15.06 ± 1.23
8,11-Octadecadienoic acid	10.74	ND	ND	ND	1.70 ± 0.25	1.56 ± 0.19	1.45 ± 0.12
Nonadecanoic acid	10.91	1.47 ± 0.16	1.22 ± 0.11	1.29 ± 0.14	0.52 ± 0.06	1.04 ± 0.21	0.87 ± 0.08
9,12,15-Octadecatrienoic acid	10.97	2.21 ± 0.21	2.01 ± 0.19	2.17 ± 0.21	0.65 ± 0.07	0.98 ± 0.08	0.84 ± 0.07
2-Octyl-cyclopropaneoctanoic acid	11.02	1.46 ± 0.20	1.22 ± 0.18	1.34 ± 0.16	0.88 ± 0.08	1.05 ± 0.12	0.98 ± 0.09
17-Methyl-octadecanoic acid	11.10	1.00 ± 0.08	1.25 ± 0.09	1.12 ± 0.24	ND	ND	ND
5,8,11,14-Eicosatetranoic acid	11.34	4.00 ± 0.34	3.34 ± 0.21	3.67 ± 0.38	0.19 ± 0.02	0.26 ± 0.02	0.19 ± 0.02
7,10,13-Eicosatrienoic acid	11.42	1.22 ± 0.13	1.15 ± 0.10	1.18 ± 0.16	0.94 ± 0.07	1.11 ± 0.01	1.01 ± 0.01
10,13-Eicosadienoic acid	11.50	3.43 ± 0.35	2.92 ± 0.17	3.04 ± 0.34	1.24 ± 0.15	1.56 ± 0.17	1.76 ± 0.03
9,12,15-Octadecatrienoic acid	11.54	2.21 ± 0.16	2.12 ± 0.19	2.45 ± 0.23	0.66 ± 0.07	0.77 ± 0.08	0.65 ± 0.05
Eicosanoic acid	11.63	0.66 ± 0.04	0.88 ± 0.07	0.94 ± 0.07	3.15 ± 0.02	2.92 ± 0.31	3.16 ± 0.34
Docosanoic acid	12.93	0.59 ± 0.04	0.76 ± 0.09	0.78 ± 0.09	ND	ND	ND
5,8,11,14-Eicosatetranoic acid	12.53	5.47 ± 0.61	4.98 ± 0.52	5.07 ± 0.52	1.92 ± 0.22	1.45 ± 0.13	1.67 ± 0.18
13-Docosenoic acid	12.78	0.96 ± 0.07	1.34 ± 0.12	1.41 ± 0.15	1.22 ± 0.11	1.13 ± 0.11	1.09 ± 0.13
7,10,13-Docotrienoic acid	12.89	ND	ND	ND	2.70 ± 0.22	1.74 ± 0.18	1.94 ± 0.21
Tetracosanoic acid	14.84	ND	ND	ND	0.29 ± 0.02	0.12 ± 0.01	0.17 ± 0.03
% Unsaturated fatty acids		54.25	36.53	27.90	36.73	28.88	26.92
% Not determined fatty acids		3.28	3.25	4.11	4.64	6.83	5.84

^aRT, retention time in gas liquid chromatography.

^bTM, total membrane; PM, plasma membrane; MM, mitochondrial membrane; ND, not detected.

TABLE 2.

Relative percentages of alkyl chain lengths of fatty acids from total L. donovani WT and HePC-R promastigote membranes

Length of the alkyl chain	% of alkyl chain lengtha in:
	WT	HePC-R
C-10	0	1.12
C-11	0	0.28
C-12	0.90	0.33
C-13	0.18	1.89
C-14	4.18	17.49
C-15	0.90	3.79
C-16	10.17	22.37
C-17	1.37	1.39
C-18	60.36	33.90
C-19	1.47	0.52
C-20	14.78	7.44
C-21	0	0
C-22	1.55	1.22
C-23	0	0
C-24	0	0.29

^aValues are means from three experiments.

Phospholipid isotherms.

The plasma membrane is the site at which drug is taken up by the cell, and any significant modification in its composition may have an impact on drug-membrane interactions. Since we found that plasma membranes from HePC-R parasites have a lower unsaturated fatty acid content than WT parasites, we decided to study the behavior of HePC towards phospholipid monolayers containing saturated or unsaturated alkyl chains. It is now admitted that biological membranes are composed of dynamic, more condensed (ordered) domains (rafts) included in a fluid phase (26). The difference is due to the saturation of the hydrocarbon chains in the raft lipids (3). Therefore, it was interesting to compare the behaviors of HePC parasites with fluid and condensed monolayers.

At first, the isotherms of DPPE, DPPG, DPPC, POPE, and POPC were recorded at 21 ± 1°C (Fig. 1). At this temperature, monolayers of DPPE and DPPG were in a liquid condensed state and POPC was in a liquid expanded state. POPE and DPPC were in an intermediate state, and a phase transition liquid expanded-liquid condensed state appeared at 35 and 6 mN/m, respectively. The state of the monolayer depends on the fatty acid chain saturation and on the size and the charge of the polar head group.

FIG. 1.

Compression isotherms of phospholipids. Δ, DPPE; •, DPPGl; □, DPPC; ♦, POPE; ×, POPC. The subphase was distilled water (pH 5.6) at a temperature of 21 ± 1°C. molec, molecule.

The mean molecular areas obtained at 25 mN/m for each phospholipid are given in Table 3. At this surface pressure, the mean molecular areas of POPE and of POPC were superior to 50 Ų because these molecules were in the liquid expanded (fluid) phase, whereas DPPE, DPPG, and DPPC were in liquid condensed (ordered) phase. At this temperature, the most condensed lipid was DPPE.

TABLE 3.

Mean molecular areas obtained from isotherms of phospholipids at a surface pressure of 25 mN/m^a

Phospholipid	DPPE	DPPG	DPPC	POPE	POPC
Mean molecular area (Å2/molecule ± 0.5 Å2)	31.2	35.8	43.5	52.1	69
Lipid state	Condensed	Condensed	Condensed	Fluid	Fluid

^aThe subphase was distilled water (pH 5.6) at a temperature of 21 ± 1°C.

Once the physical state of the different lipid monolayers had been determined, the interaction of HePC with condensed and fluid monolayers was investigated both at constant surface pressure and at constant surface area. The two techniques are complementary and together mimic biological conditions. When exogenous molecules interact, by adsorption and/or insertion with the membrane, two phenomena may occur; the surface area may increase without a change of the pressure, and the surface pressure may increase without area modification. It is probable that the real situation is intermediate; i.e., both pressure and area increase.

In Fig. 2 and 3, the variations of the mean molecular areas (ΔA) obtained 1 hour after injection of HePC under different phospholipid monolayers are plotted as a function of HePC concentration.

FIG. 2.

Variations of the mean ΔA obtained 1 h after injection of HePC monomers (○, •) and monomers and micelles (□, ▪) under a monolayer of POPE (○, □) and POPC (•, ▪) versus HePC concentration at a constant surface pressure (25 mN/m). The subphase was distilled water (pH 5.6) at a temperature of 21 ± 1°C. molec, molecule.

FIG. 3.

Variations in the mean ΔA obtained 1 h after injection of HePC under a monolayer of DPPE (○), DPPG (Δ), and DPPC (▪) versus HePC concentration at a constant surface pressure (25 mN/m). The subphase was distilled water (pH 5.6) at a temperature of 21 ± 1°C. molec, molecule.

The behavior of HePC in the presence of a fluid phospholipid monolayer (POPE) is shown in Fig. 2 and compared to the results obtained with POPC, another fluid monolayer published in a previous work (22). Both fluid phospholipids exhibited similar behaviors in the presence of HePC. HePC monomers progressively inserted into these monolayers only up to a concentration of 2 μM, corresponding to the critical micellar concentration (CMC) of HePC. Above the CMC, an irregular insertion of HePC molecules occurred, because not only monomers but also groups of monomers were transferred into the fluid monolayer.

With the different condensed phospholipid monolayers (DPPE, DPPG, and DPPC), we observed that molecular area increased smoothly over the whole range of HePC concentrations (Fig. 3). Thus, a progressive insertion of HePC monomers into these condensed monolayers occurred even beyond the CMC. This progressive insertion had not reached a plateau at a concentration of 4 μM HePC, however, at this concentration, the barriers of trough were expanded at their maximum, so the insertion of HePC was artificially limited and the experiment could not be continued beyond this limit. So, although HePC micelles were able to insert into fluid monolayers (POPE and POPC), this was not possible when the phospholipids were in the condensed state (DPPE, DPPG, and DPPC).

These results were compared to those obtained with the complementary experiment set-up in which the molecular area was kept constant. The ΔP induced by HePC were studied in the presence of monolayers of DPPE (the most condensed phospholipid) and POPC (the most fluid one), at constant molecular areas of 31.2 ± 0.5 Ų and 69.0 ± 0.5 Ų for DPPE and POPC, respectively (Table 3). In Table 4, ΔP values measured instantaneously and 1 h after injection of HePC under the phospholipid monolayer are reported. It can be seen that the pressure change due to the adsorption of HePC monomers into DPPE reached 5 mN/m immediately with a low HePC concentration of 0.5 μM and increased rapidly with an HePC concentration up to 12 mN/m. Since phospholipids were condensed, the insertion of some HePC molecules created a relatively large surface pressure variation. In contrast, with the fluid phospholipid (POPC), at the same HePC concentration of 0.5 μM, the ΔP remained at 2 mN/m even after 1 h. The instantaneous ΔP increased more slowly with the HePC concentration. The phospholipid fluidity allowed an insertion of monomers, with a smaller effect on the initial surface pressure of the monolayer, and more time was required to reach the maximum ΔP (ΔPmax). Whatever the phospholipid, this maximal ΔP was 12 mN/m. As we have previously demonstrated, if we add this ΔPmax to the initial surface pressure of 25 mN/m, we find the pressure of maximal cohesion of HePC molecules to be 37 mN/m (22).

TABLE 4.

Variations in the surface pressure of DPPE or POPC instantaneously and 1 hour after injection of HePC at a constant molecular area^a

HePC concn (μM)	ΔP (±1 mN/m)
	DPPE monolayer		POPC monolayer
	Instantaneous	After 1 h	Instantaneous	After 1 h
0.5	5	7.5	2	2
1	7.5	8	3.5	7
1.5	9	12	5	9
2	11	11	5.2	10
2.5	12	12	8	10
3	12	11	8	12
4	12	12	8	12

^aThe subphase was distilled water (pH 5.6) at a temperature of 21 ± 1°C.

In the present study, we have shown that the behavior of HePC differs in the presence of condensed and fluid phospholipids. First, at a constant surface pressure above the CMC of HePC, only fluid phospholipids allowed the recruitment of deployed micelles as groups of monomers and abrupt deployment of micelles as groups of monomers into the monolayer, whereas continuous progressive insertion of HePC parasite monomers into condensed phospholipid monolayers was observed even beyond the CMC. Second, at a constant molecular area, a similar HePC concentration created a larger increase in ΔP, and the ΔPmax was reached more rapidly with condensed phospholipids than with fluid ones. However, in the presence of both fluid and condensed monolayers, HePC micelles constituted a reservoir of monomers both for monomer insertion into condensed phospholipids and for groups of monomers into fluid phospholipids. Since biological membranes are composed of dynamically condensed domains included in a fluid phase, we suggest that, above the CMC, HePC can insert into both kinds of phase: as monomers into the condensed phases and as a group of monomers into the fluid phase. In both cases, micelles of HePC act as a reservoir of monomers.

Sterol composition.

Eighteen sterols were identified and quantified in total, plasma, and mitochondrial membranes (Table 5). In WT parasites, the plasma membranes had a sterol composition similar to that of the total membranes, whereas mitochondrial membranes showed some characteristic differences. Thus, the cholesterol content was significantly higher than that found in total membranes (52.2% versus 34.0%). Ergosterol was not detected in mitochondrial membranes but was replaced by β-sitosterol (24-ethyl-cholesterol) at a similar percentage (about 25%).

TABLE 5.

Composition of free sterols in total, plasma, and mitochondrial membranes from L. donovani WT and HePC-R promastigotes

Sterol	RTa (h)	Mean % of total sterols ± SD (n = 3)b
		WT			HePC-R
		TM	PM	MM	TM	PM	MM
Cholesterol	1.00	34.0 ± 2.7	25.1 ± 2.2	52.2 ± 4.8	58.8 ± 4.5	34.3 ± 3.1	30.8 ± 3.2
14α-Methylcholesta-5,7,24-trien-3β-ol	1.02	0.2 ± 0.1	0.2 ± 0.1	0.3 ± 0.1	0.3 ± 0.1	0.2 ± 0.1	0.2 ± 0.1
14α-Methylcholesta-8,24-dien-3β-ol	1.04	0.1 ± 0.1	0.2 ± 0.1	0.1 ± 0.1	0.2 ± 0.1	0.1 ± 0.1	0.1 ± 0.1
Cholesta-5,7,24-trien-3β-ol	1.07	0.2 ± 0.1	4.1 ± 0.5	0.3 ± 0.1	0.2 ± 0.1	0.3 ± 0.1	0.3 ± 0.1
Ergosterol (ergosta-5,7,22-trien-3β-ol)	1.10	24.9 ± 1.8	26.7 ± 1.8	ND	16.1 ± 1.5	26.5 ± 2.2	33.0 ± 3.2
Ergosta-5,7,22,24(24′)-tetraen-3β-ol	1.13	1.8 ± 0.1	1.6 ± 0.2	0.9 ± 0.1	0.6 ± 0.1	1.3 ± 0.1	4.1 ± 0.4
Ergosta-5,7,24(24′)-trien-3β-ol	1.17	24.5 ± 2.2	22.4 ± 2.1	ND	9.2 ± 1.1	20.5 ± 0.3	4.0 ± 0.4
Stigmasta-7,24(24′)-dien-3β-ol	1.18	ND	3.9 ± 0.3	6.1 ± 0.6	ND	6.1 ± 0.5	3.1 ± 0.4
Ergosta-7,22-dien-3β-ol	1.19	ND	3.1 ± 0.4	5.8 ± 0.6	ND	1.1 ± 0.1	21.7 ± 1.8
Ergosta-7,24(24′)-dien-3β-ol	1.20	5.7 ± 0.6	7.2 ± 0.7	0.6 ± 0.1	1.9 ± 0.2	3.4 ± 0.3	0.5 ± 0.1
14α-Methylergosta-5,7,24(24′)-trien-3β-ol	1.21	0.2 ± 0.1	0.3 ± 0.1	0.1 ± 0.1	0.1 ± 0.1	0.1 ± 0.1	0.2 ± 0.1
4,4α-Dimethylcholesta-8,24-dien-3β-ol	1.22	0.3 ± 0.1	0.3 ± 0.1	0.2 ± 0.2	0.3 ± 0.1	0.2 ± 0.1	0.2 ± 0.1
β-Sitosterol (stigmast-5-en-3β-ol)	1.225	ND	2.5 ± 0.3	26.1 ± 2.2	ND	1.1 ± 0.1	ND
4α,14α-Dimethylcholesta-8,24-dien-3β-ol	1.23	0.2 ± 0.1	0.1 ± 0.1	0.2 ± 0.1	0.1 ± 0.1	0.1 ± 0.1	0.2 ± 0.1
14α-Methylergosta-8,24(24′)-dien-3β-ol	1.25	0.2 ± 0.1	0.2 ± 0.1	0.1 ± 0.1	0.2 ± 0.1	0.2 ± 0.1	0.3 ± 0.1
4,4α-Dimethylergosta-8,24(24′)-dien-3β-ol	1.27	0.8 ± 0.1	0.6 ± 0.1	0.4 ± 0.1	0.8 ± 0.1	0.1 ± 0.1	0.3 ± 0.3
4α,14α-Dimethylergosta-8,24(24′)-dien-3β-ol	1.28	0.4 ± 0.1	0.3 ± 0.1	0.3 ± 0.1	0.2 ± 0.1	0.2 ± 0.1	0.2 ± 0.1
24-Methylenedihydrolanosterol	1.45	1.7 ± 0.1	1.1 ± 0.1	0.8 ± 0.1	1.1 ± 0.1	0.7 ± 0.1	0.6 ± 0.1
Total C-24-alkylated sterols		60.2	66.8	35.4	30.2	60.2	46.3
Other sterols (not identified)		4.8	0.1	5.5	9.9	3.5	0.2

^aRT, retention time relative to that of cholesterol in gas chromatography.

^bTM, total membrane; PM, plasma membrane; MM, mitochondrial membrane; ND, not detected.

In HePC-R parasites, cholesterol was the major sterol found in total membranes (58.8%) and ergosterol was the second-most abundant, representing only 16.1% of total sterols. When WT and HePC-R parasites are compared, it can be seen that the proportion of cholesterol was enhanced in total HePC-R parasite membranes (58.8%) with respect to the level in WT parasites (34.0%) but that the content of C-24-methylated sterols was halved in HePC-R parasites (30.2%) compared with that in WT parasites (60.2%). However, since the 24-alkylation of sterols was not greatly affected in plasma membranes from HePC-R parasites, we suggest that this reflects a marked reduction of the 24-alkylated sterol content in internal membranes, such as the endoplasmic reticulum.

Relationship between the 24-alkylation of sterols and HePC sensitivity.

The enzyme systems which are responsible for the alkylation of sterols at C-24 are S-adenosyl-l-methione-C-24-δ-sterol-methyltransferases, known as SCMT. SCMT1 catalyses the first methylation at the C-24 position, and SCMT2 is able to promote a second methylation leading to ethyl-sterols (2, 27). In the light of the reduction of 24-alkylated sterol content in total HePC-R parasite membranes, we decided to investigate whether the SCMT1 system was involved in HePC sensitivity. To this end, we compared the sensitivities to HePC of three strains: a Saccharomyces cerevisiae WT strain, an SCMT-null mutant strain, and an L. donovani SCMT1-transformed S. cerevisiae strain. Whatever the strain, the sensitivities to HePC were similar, whereas that to amphotericin B seemed linked to SCMT expression (Table 6). Although our test was not performed in a homologous system, we can conclude that the presence of the Leishmania SCMT gene did not change the HePC sensitivity of S. cerevisiae.

TABLE 6.

SCMT1 activity and drug sensitivity in WT S. cerevisiae, an SCMT1 (ERG6)-null mutant, and L. donovani SCMT1-transformed S. cerevisiae cultured with or without an inductor

S. cerevisiae strain	Ergosterol content as a % of total sterols ± SDa	MIC (μM)
		AmB		HePC
		24 h	48 h	24 h	48 h
WT	27.4 ± 3.1	0.78 ± 0.05	1.56 ± 0.12	6.25 ± 0.71	12.5 ± 1.27
SCMT1 (ERG6)-null mutant	0	3.12 ± 0.18	12.5 ± 1.15	6.25 ± 0.74	12.5 ± 1.31
SCMT1-transformed strain
With inductorb	31.1 ± 2.8	1.56 ± 0.11	6.25 ± 0.79	6.25 ± 0.51	12.5 ± 0.99
Without inductor	0	3.12 ± 0.22	12.5 ± 1.22	6.25 ± 0.69	12.5 ± 1.44

^aResults are from a recent study (20). The activity of SCMT1 in yeasts was assessed by determining the ergosterol content expressed as a percentage of total sterols. Values are means of results of three independent experiments ± standard deviations.

^bTo induce expression, 2% galactose replaced 2% glucose in the medium.

DISCUSSION

A mechanism for HePC resistance in Leishmania was proposed recently as the result of the identification and characterization of the HePC transporter in L. donovani (18). This novel P-type ATPase belonging to the aminophospholipid translocase subfamily and called LdMT was transfected into HePC-R parasites, which recovered their sensitivity to HePC and the ability to take up radiolabeled HePC normally. The mechanism of HePC resistance in Leishmania clearly relies on this HePC transporter (18). However, the resistant parasites which grow in the presence of 40 μM HePC were killed at a concentration of 100 μM (16). Thus, HePC is active against resistant parasites which possess a nonfunctional LdMT. This observation suggests that another mechanism becomes predominant at high concentrations, probably involving direct drug-membrane interactions which compensate for the transporter deficiency. The present investigation highlights some metabolic alterations in HePC-resistant parasites based on lipid analysis. A monolayer membrane model allowed us to demonstrate how one of these changes could modulate drug-membrane interactions.

In this work, we have studied the modifications in lipid composition which accompany HePC resistance in order to identify some biochemical targets which might be affected in Leishmania. Lipid composition is linked to membrane fluidity, which may influence drug-membrane interactions. Since HePC is an amphiphilic and zwitterionic molecule, it may directly interact with the membrane. Therefore, we focused on the main parameters determining fluidity: the unsaturation of the fatty acid alkyl chains in membrane components (phospholipids, monoacylglycerols, diacylglycerols, triacylglycerols, and glycolipids) and the side chain alkylation of sterols (13). However, our experimental protocol did not allow analysis of ether-lipids that cannot be obtained from saponification. The plasma membrane is the first pharmacokinetic barrier to be crossed by the drug before it can be active. Therefore, the alkyl chain composition of the lipids located in the plasma membrane may modulate drug-membrane interactions and possibly be involved in drug resistance. The results obtained here suggest that plasma membranes from HePC-R parasites could be less fluid than those from WT parasites, based on their unsaturated fatty acid content. This decrease of unsaturated alkyl chain content in HePC-R parasites could be explained by a partial inactivation of desaturases. However, desaturation reactions have not yet been described to occur in Leishmania. In mammalian cells, these enzymes are present in the endoplasmic reticulum membranes and include four broad-specificity fatty acyl coenzyme A (acyl-CoA) desaturases (non-heme iron-containing enzymes) that introduce unsaturation at C-4, C-5, C-6, or C-9 (24). The fatty acyl-CoA desaturases in WT and HePC-R L. donovani parasites now have to be identified by cloning and expression to provide a molecular explanation for our observations in this study.

As far as the lengths of fatty acid alkyl chains were concerned, we observed a partial inactivation of alkyl chain elongation in HePC-R parasites. In particular, we found a C₁₆ content twice as high in HePC-R parasites as in WT parasites and a C₁₈ content in HePC-R parasites only half that in WT parasites. Two types of fatty acid synthesis have been identified. In type I fatty acid synthesis, as it occurs in mammalian cells, the enzymatic machinery is a multidomain polypeptide including acyl carrier protein (ACP), acetyl-CoA-ACP transacylase, malonyl-CoA-ACP-transacylase, β-ketoacyl-ACPsynthase, β-ketoacyl-ACP reductase, β-hydroxyacyl-ACP dehydrase, and enoyl-ACP reductase, whereas in type II fatty acid synthesis, discrete mono-functional proteins are involved (23). The recent finding that Trypanosoma brucei has a type II fatty acid synthesis pathway indicates that kinetoplastids have a fatty acid synthesis pathway completely different from that observed in mammalian cells (14, 17). The fatty acid elongation system operating in Leishmania needs to be elucidated before the significance of our observations can be understood.

Although it is commonly assumed that elongation and unsaturation of fatty acids occurs in both the mitochondria and the endoplasmic reticulum, this assumption has to be verified for Leishmania. However, it has been demonstrated that acetate is the major precursor of phospholipids in L. donovani, although leucine can also be used for phospholipid biosynthesis (7). Using the phospholipid monolayer model, we demonstrated previously that HePC had an affinity for membranes and inserted freely into a phospholipid monolayer (22). In the present study, we found that HePC inserts more easily into a monolayer rich in unsaturated phospholipids, which are more fluid. Therefore, the decrease in the unsaturated alkyl chains content in HePC-R parasite plasma membranes is in favor of reduced HePC-external phospholipid monolayer interactions. Perez-Victoria et al. (19), working on living parasites, described similar levels of binding of HePC to parasites in both WT and HePC-R parasite clones. Our results are not contradictory with those of Perez-Victoria et al. (19) but rather confirm them. Their binding measurements took into account HePC molecules weakly and nonspecifically adsorbed to the external membrane of the parasite. On the other hand, the monolayer model measures HePC insertion within the external layer of the membrane, which was considered part of the uptake fraction in the Perez-Victoria study (19). The bovine serum albumin used in this study (19) to remove surface-adsorbed HePC was not able to eliminate HePC included within the external monolayer. So, their uptake fraction included HePC inserted within the external monolayer of the parasite membrane.

We found a significant increase in the cholesterol contents of total membranes of HePC-R parasites. Since cholesterol is not a product of de novo sterol biosynthesis but is derived from the culture medium, we can assume that cholesterol can enter the HePC-R parasites more easily than the WT parasites. Moreover, cholesterol, due to its ordering effect, may contribute to the decrease in total and plasma membrane fluidity. In a recent study, we demonstrated that HePC was able to condense sterols in monolayers, with similar effects for cholesterol and ergosterol, a C-24-alkylated sterol (22).

Moreover, the similar contents of 24-alkylated sterols in WT and HePC-R parasite plasma membranes explains the equivalent sensitivities of the two lines to amphotericin B (50% inhibitory concentration, 0.01 μM) found by Seifert et al. (28), since these sterols are the main target for amphotericin B at the membrane level. Nevertheless, the amount of 24-alkylated sterol found in total membranes was reduced to half in HePC-R parasites compared with that in WT parasites. This observation implies a drastic reduction of sterol methylation, leading to low levels of 24-alkylated sterols in internal membranes by partial inactivation of SCMT1, the enzyme responsible for the first methylation of sterols at the C-24 position. SCMT1, which does not exist in mammalian cells, is considered an interesting target for kinetoplastid chemotherapy (23, 31). We decided to study the potential role played by SCMT1 in HePC sensitivity. With this aim, we assessed the HePC sensitivities of an SCMT-null mutant, Leishmania SCMT-transformed, and wild-type Saccharomyces cerevisiae clones. The presence of the SCMT gene was not linked to HePC sensitivity. Therefore, we can conclude from these data that the decrease of 24-methylated sterols in HePC-R parasite membranes had no effect on the HePC-membrane interactions since the contents of these sterols remained similar in WT and HePC-R parasite plasma membranes, and no effect on HePC sensitivity was observed.

In mitochondrial membranes of WT parasites, we did not find ergosterol, but β-sitosterol (24β-ethylcholesterol) is the most abundant sterol after cholesterol. The presence of an ethyl group at the C-24 position suggested the presence and the functionality of one SCMT2, the enzyme responsible for a second methylation of C-24-methylated sterols, operating after SCMT1. SCMT2 is frequently found in plants (2). In HePC-R parasites, we did not find sitosterol, leading to the hypothesis of SCMT2 inactivation. Although the role of sitosterol in WT mitochondrial membranes is not known, its presence is a characteristic of plants (2).

Our results demonstrating a difference in lipid composition between WT and HePC-R parasite membranes and the study of HePC interaction with some relevant lipids are in favor of reduced interactions of HePC with the external layer of the plasma membrane in HePC-R parasites. At the moment, we do not know whether HePC insertion within the external monolayer is a prerequisite for the interaction with its transporter. Further studies should be performed using phospholipid vesicles to determine whether HePC transbilayer movement can occur in the absence of the transporter. The LdMT transporter is the key factor in the mechanism of resistance to HePC (18); however, our results can explain why the resistant parasites growing at 40 μM with a nonfunctional LdMT transporter were killed at 100 μM. Therefore, we propose that drug-membrane interactions are negligible at low concentrations and that the LdMT transporter plays the major role in drug uptake. At higher concentrations, drug-membrane interactions may play a more important role in the biological activity of HePC.

The situation we described in this study and previously (22) could be a model for what occurs more specifically in lipid rafts and in the parasite flagellar pocket, where HePC may interact more easily with the membrane. Lipid rafts have been described to occur in both Leishmania and Trypanosoma brucei, suggesting a broad distribution throughout the kinetoplastida (6, 15). These microdomains have been demonstrated to contain lipophosphoglycans (25) and glycosylphosphatidylinositol-anchored proteins, and their size (about 50 nm) is comparable to that of lipid rafts in mammalian cells (21).

The modifications in lipid metabolism that we observed in the HePC-R clone (i.e., the decrease in the level of fatty acid alkyl chain unsaturation and in the amounts of 24-alkylsterols in total membranes) exhibit some similarities to our previous observations of an atovaquone-resistant Leishmania infantum clone (4) and an amphotericin B-resistant L. donovani clone (13). In the latter amphotericin B-resistant clone, 24-alkylsterols were absent, and we recently found that the transcript of SCMT1 did not possess the spliced leader sequence found in the transcript highly expressed in L. donovani WT parasites (20). In conclusion, HePC resistance in Leishmania is accompanied by significant changes in C-24-sterol alkylation and fatty acid elongation and unsaturation. Some of these phenomena are also present in Leishmania clones that are resistant to other drugs (i.e., amphotericin B and atovaquone), indicating that resistance towards HePC, amphotericin B, and atovaquone affects the parasite membrane structure. In order to elucidate the biochemical modifications involved in HePC resistance, further studies will be focused mainly on Leishmania acyl-CoA synthetase and desaturase systems.