Stage-Specific De Novo Synthesis of Very-Long-Chain Dihydroceramides Confers Dormancy to Entamoeba Parasites

ABSTRACT Amoebiasis is a parasitic disease caused by Entamoeba histolytica infection and is a serious public health problem worldwide due to ill-prepared preventive measures as well as its high morbidity and mortality rates. Amoebiasis transmission is solely mediated by cysts. Cysts are produced by the differentiation of proliferative trophozoites in a process termed “encystation.” Entamoeba encystation is a fundamental cell differentiation process and proceeds with substantial changes in cell metabolites, components, and morphology, which occur sequentially in an orchestrated manner. Lipids are plausibly among these metabolites that function as key factors for encystation. However, a comprehensive lipid analysis has not been reported, and the involved lipid metabolic pathways remain largely unknown. Here, we exploited the state-of-the-art untargeted lipidomics and characterized 339 molecules of 17 lipid subclasses. Of these, dihydroceramide (Cer-NDS) was found to be among the most induced lipid species during encystation. Notably, in encysting cells, amounts of Cer-NDS containing very long N-acyl chains (≥26 carbon) were more than 30-fold induced as the terminal product of a de novo metabolic pathway. We also identified three ceramide synthase genes responsible for producing the very-long-chain Cer-NDS molecules. These genes were upregulated during encystation. Furthermore, these ceramide species were shown to be indispensable for generating membrane impermeability, a prerequisite for becoming dormant cyst that shows resistance to environmental assault inside and outside the host for transmission. Hence, the lipid subclass of Cer-NDS plays a crucial role for Entamoeba cell differentiation and morphogenesis by alternating the membrane properties. IMPORTANCE Entamoeba is a protozoan parasite that thrives in its niche by alternating its two forms between a proliferative trophozoite and dormant cyst. Cysts are the only form able to transmit to a new host and are differentiated from trophozoites in a process termed “encystation.” During Entamoeba encystation, cell metabolites, components, and morphology drastically change, which occur sequentially in an orchestrated manner. Lipids are plausibly among these metabolites. However, the involved lipid species and their metabolic pathways remain largely unknown. Here, we identified dihydroceramides (Cer-NDSs) containing very long N-acyl chains (C26 to C30) as a key metabolite for Entamoeba encystation by our state-of-the-art untargeted lipidomics. We also showed that these Cer-NDSs are critical to generate the membrane impermeability, a prerequisite for this parasite to show dormancy as a cyst that repels substances and prevents water loss. Hence, ceramide metabolism is essential for Entamoeba to maintain the parasitic lifestyle.


RESULTS
Identification of lipid species and their fluctuating levels (increase or decrease) during encystation. To comprehensively investigate the lipid species that fluctuate during encystation, we used in vitro culture of Entamoeba invadens. Generally, studies of Entamoeba encystation have adopted the in vitro culture of E. invadens, a reptilian parasite, and not that of E. histolytica as a model system (see Fig. 1A). This is because the strains of E. histolytica available in the laboratory do not encyst after adaptation to culture conditions. The E. invadens life cycle is the same as that of E. histolytica, and the symptoms caused by E. invadens infection are similar to those of E. histolytica (4,5). Lipids were extracted from encysting E. invadens cells at designated time points after the induction of encystation and then were analyzed by untargeted lipidomics. During encystation, a series of lipid species, including ceramide, ceramide phosphatidylinositol (PI-Cer) (see Fig. 1D for the structure), lysophosphatidylserine (LPS), and lysophosphatidylinositol (LPI) was significantly increased with time ( Fig. 2A    detected throughout the Entamoeba life cycle (trophozoite and cyst stages) and mainly comprised Cer-NDSs (Fig. 2B). In mammals, the ceramide lipid class is essential to prevent water loss from the skin (31). Entamoeba cysts are also resistant to desiccation (5,32). We, therefore, focused on Cer-NDS species for further analyses.
Revealing a de novo ceramide synthesis pathway in Entamoeba. Very-long-chain Cer-NDSs were not detected in bovine serum, which is the major lipid source in Entamoeba encystation-inducing culture medium (33); therefore, it was unlikely that very-long-chain Cer-NDSs were derived from the external milieu. Of interest, all necessary genes for the de novo ceramide synthesis are harbored by both the E. histolytica and E. invadens genomes except for one gene encoding dihydroceramide desaturase ( Fig. 1B) (AmoebaDB, http://amoebadb.org/amoeba/); there are two types of genes encoding serine palmitoyl transferase (SPT), one gene for 3-dehydrosphinganine reductase (KDHR), and five (E. histolytica) or six (E. invadens) genes for ceramide synthase (CerS) (27).
To show the capability of Entamoeba to synthesize ceramides de novo, proliferating trophozoites and encysting cells were metabolically labeled with L-[U-14 C]serine, a substrate for the first enzyme (SPT) in the de novo pathway (see Fig. 1B). 14 C-labeled bands corresponding to ceramides were detected in both trophozoites and encysting cells (Fig. 3A). During encystation, an accumulation of radiolabeled ceramide with time was observed. A dramatic increase of radiolabeled ceramide was observed between 16 and 32 h (Fig. 3B). Alkaline treatment did not change the intensity of the detected bands, ruling out the lipids being glycerolipids (see Fig. S2). These results clearly indicated that Entamoeba synthesized ceramides by de novo biosynthesis. Notably, the time course for the accumulation of 14 C-labeled ceramide correlated well with the increased amount of very-long-chain Cer-NDSs between 16 and 24 h after encystation induction and reached a plateau after 24 h ( Fig. 2C and Fig. S1A). Consistently, during the initiation phase of encystation, expression of a series of ceramide biosynthetic enzymes was coordinately induced in Entamoeba (Fig. 3C). These results indicated that the induction of very-long-chain Cer-NDSs during Entamoeba encystation appeared to be mediated by de novo biosynthesis.
Identification of the ceramide synthase gene responsible for producing Cer-NDSs in Entamoeba. Variation in the acyl chain length of Cer-NDSs observed during Entamoeba encystation is likely to be generated by different CerS isozymes, as observed in other organisms (21,22). To identify the CerS responsible for very-longchain Cer-NDS biosynthesis in Entamoeba, we exploited an approach combining genetics and lipidomics. The genetic approach included gene knockdown mediated by transcriptional gene silencing via antisense small RNA (34,35) and gene overexpression  (Table S1). Red letters indicate ceramide species whose levels were .10-fold higher than those in trophozoites (0 h after induction). Representative data (Sample 1 in Table S1) are shown from three independent experiments. (36). We used E. histolytica instead of E. invadens as the host because the genetic systems for E. invadens have not been widely adopted. In E. histolytica trophozoites, Cer-NDS species were similarly detected as in E. invadens trophozoites (see Fig. S3A). A gene knockdown experiment was performed using five E. histolytica gene silencing (gs) transformants, EhCerS2gs to EhCerS6gs, in each of which a single gene among the five EhCerSs was knocked down. Note that E. histolytica does not have a counterpart of E. invadens CerS1 (EiCerS1) (see Fig. 1B). After verifying the level of gene knockdown in each transformant by quantitative reverse transcription-PCR (qRT-PCR) (Fig. S3B), the lipidomic profiles of Cer-NDS species in all EhCerSgs (except for EhCerS3gs) and mock transformants were individually determined ( Fig. 4A and Fig. S3C to E). One transformant, EhCerS3gs, showed a severe growth defect, which hampered long-term subculture. Among the transformants tested, only EhCerS4gs showed a significant reduction in Cer-NDS levels; the most significant reduction was observed in Cer 18:0;2O/24:1, and the amounts of Cer 17:0;2O/24:1 and Cer 19:0;2O/24:1 were also reduced (Fig. 4A). In EhCerS4gs, both EhCerS4 and EhCerS5 transcripts were significantly downregulated (0.25% 6 0.03% and 4.2% 6 0.3%, respectively, relative to the mock transformant 100% control) (Fig. S3B). However, a contribution of EhCerS5 was ruled out because the amounts of Cer-NDS species were not changed in EhCerS5gs, in which only the EhCerS5 transcript was reduced (4.4% 6 0.6%) ( Fig. S3B and D). These results indicated that EhCerS4 was responsible for synthesizing Cer-NDSs containing a C 24:1 acyl chain. None of the remaining three transformants (EhCer2gs, -5gs, and -6gs) showed obvious changes in their Cer-NDS species profile, probably because of genetic redundancy ( Fig. S3C to E).
Ceramide metabolism in Entamoeba. To understand ceramide metabolism in Entamoeba, we investigated the effect of myriocin, a known inhibitor for the first enzyme (SPT) in the de novo pathway for ceramide biosynthesis (see Fig. 1B). Myriocin dose-dependently inhibited cyst formation in in vitro cultures of E. invadens, which was consistent with the previous report (27,28). The 50% inhibitory concentration [IC 50 ] was calculated as 68.6 6 12.5 nM (n = 3) (Fig. 5A). Also, the physiological changes during the course of encystation were monitored by flow cytometry (37). Evans blue (EB) was used as an indicator of membrane permeability, and calcofluor (CF) was used as an indicator of the level of chitin, a major component of the cyst wall (38). As shown in the control in Fig. 5A, it appeared that the CF 2 EB 1 population (proliferating trophozoites) was gradually changed to a CF 1 EB 2 population (mature cysts) through CF low EB 1 and CF 1 EB 1 populations. At 12 h after induction, myriocin treatment did not affect the phenotype, but at 16 h, cell differentiation was paused, resulting in the accumulation of an irregular CF 1 EB strong population (abnormal cells) at 20 h. These results indicated that myriocin impaired the encystation process at 16 to 20 h postinduction. Importantly, this time frame correlated well with the lipidomic changes of very-longchain Cer-NDSs dramatically increased between 16 and 24 h after induction of encystation (Fig. 2C). These results indicated that inhibition of very-long-chain Cer-NDS biosynthesis by myriocin halted cyst formation.
Next, we determined the consequence of Entamoeba encysting cells treated with myriocin. After 24 h, when the effect of myriocin was first apparent on cyst formation, both control and myriocin-treated live cells were stained by CF and EB (Fig. 5B). Flow cytometry analysis showed that the level of CF fluorescence in myriocin-treated cells was comparable to that of untreated cells (Fig. 5A). A change in the CF signal reflects the synthesis and degradation of chitin polymers. Therefore, these results indicate that chitins are synthesized and placed in the cyst wall at similar levels in both myriocintreated and untreated cells. However, a distinct physiological change was observed in myriocin-treated cells. The fluorescence signal of EB, an indicator of membrane permeability, abnormally accumulated inside the cells, indicating that myriocin treatment increased the membrane permeability of encysting cells (Fig. 5B).
To further determine the structural changes induced by myriocin treatment, we performed transmission electron microscopy analysis of encysting cells in either the presence or absence of myriocin (Fig. 5C). Cells were prepared by rapid freezing and freeze-substitution to preserve the membrane structure (39). The myriocin-treated cells were withered, and accumulations of abnormal vacuoles were observed throughout the cytoplasm. Notably, the cell membranes of myriocin-treated cells were more compressed and disconnected. Furthermore, the cyst wall regions of treated cells were swollen, and cell components randomly filled the spaces between the regions of disrupted membranes and the cyst walls. It is worth mentioning that obvious changes in lipidome of encysting cells (24 h postinduction) by myriocin treatment was observed   in Cer-NDSs containing very long N-acyl chains ($26 carbon) and their metabolites, PI-Cers. The level of LPS was also affected by myriocin treatment, but to a small extent compared to that for Cer-NDSs (see Fig. S5; Table S2). These results indicate that the off-target effect of myriocin on lipid metabolism in Entamoeba encysting cells was quite limited under this experimental condition. These results indicated that the stagespecific induction of Cer-NDSs with very long N-acyl chains (C 26 to C 30 ) were indispensable to develop membrane impermeability.

DISCUSSION
Entamoeba encystation is a crucial process for maintaining the life cycle of this parasitic species. Encystation is a fundamental cell differentiation and morphogenesis process that involves a variety of pathways, which function as an orchestrated network (5). Here, we performed an untargeted lipidomic analysis of encysting cells over time. This approach enabled us to reveal that the levels of Cer-NDSs were significantly induced during encystation. We also unraveled unique features of an Entamoeba metabolic pathway and its physiology. First, Entamoeba possesses an atypical de novo ceramide synthesis pathway that produces Cer-NDSs possessing a broad range of acyl chains (C 16 , C 20 , C 22 , C 23 , C 24 , C 26 , C 28 , C 30:0 , C 24:1 , C 28:1 , C 28:2 , C 28:3 , C 30:1 , and C 30:2 ) as terminal metabolites. Second, during encystation, the amounts of very-long-chain Cer-NDSs with an acyl chain length of $26 were increased, coinciding with transcriptional upregulation of the three genes encoding CerS2, -5, and -6. Those enzymes were functionally redundant and responsible for producing those ceramides. Third, inhibition of de novo ceramide synthesis caused damage to the plasma membrane and increased membrane permeability to a nonphysiological level, resulting in the formation of aberrant cysts.
Differentiation into dormant cysts from proliferative trophozoites is necessary for Entamoeba to be resistant to environmental assaults inside as well as outside the host and to be transmitted to a new host. This dormant process involves rounding and strengthening of the cells, which requires alternations to properties of the plasma membrane, such as changes to the composition and topology of the lipid bilayer. Halting Cer-NDS production by adding myriocin to in vitro encystation induction cultures resulted in proliferating trophozoites becoming rounded as normal but that differentiated into aberrant cysts. Therefore, newly synthesized Cer-NDSs in encysting cells are not involved in the cell rounding process, but they do have crucial roles after the encysting cells became rounded. The underlying molecular mechanisms need to be elucidated; however, in myriocin-treated encysting cells, a partly disconnected plasma membrane and abnormal accumulation in the cytoplasm of EB, a membrane impermeable dye, were simultaneously observed. Furthermore, we can estimate the content of very-long-chain Cer-NDSs with the acyl chain length of $26 in a single encysting cell to be significantly elevated because cell numbers should not theoretically increase, as encystation is a differentiation process, but they empirically decreased a bit. Cell volume also became around half during encystation. These findings indicate that very-long-chain Cer-NDSs have critical roles in maintaining the plasma membrane impermeability and that Entamoeba regulates intracellular ceramide metabolism to provide very-long-chain Cer-NDSs to enable transmission to a new host. Cer-NSs were also present in Entamoeba despite the absence of dihydroceramide desaturase gene in Entamoeba genome (AmoebaDB). Therefore, these ceramides containing nonhydroxy fatty acid and sphingosines (Cer-NSs) are plausibly acquired from the culture medium. Interestingly, a species difference of Entamoeba in the ratio of Cer-NS to ceramide was . This may reflect that the E. histolytica capability for de novo ceramide synthesis is less than that of E. invadens.
Furthermore, we showed evidence that Cer-NDS species may also be essential for Entamoeba to multiply as trophozoites. Both E. histolytica and E. invadens trophozoite proliferations were impaired by myriocin with IC 50 values of 46. 7 6 11.5 nM (n = 3) and 1.90 6 0.10 mM (n = 3), respectively. The growth impairment by myriocin was complemented by the gene knockdown analyses targeting the enzymes in the E. histolytica de novo ceramide biosynthesis pathway (see Fig. 1B). Both knockdown strains, EhSPT1gs and EhSPT2gs (see Fig. S6A), showed severe growth defects, which hampered longterm subculture. Another knockdown strain, EhKDHRgs, also showed a growth defect ( Fig. S6A and B). Another knockdown strain, EhCerS3gs, in which the downregulation of EhCerS3 was confirmed in an early subculture (Fig. S3B), showed severe growth defects, similar to those of EhSPT1gs and EhSPT2gs. Alternating its forms between a proliferating trophozoite and dormant cyst is a parasitic strategy for surviving in different niches. Hence, ceramide metabolism plays crucial roles in the Entamoeba life cycle.
Our lipidomic analysis detected PE-Cers, PI-Cers, and SMs, the precursors of which are ceramides ( Fig. 2A and Fig. S1B to D), which is consistent with the previous studies (29,30). Furthermore, a drastic increase of some very-long-chain PE-Cer species, such as PE-Cer 18:0;2O/26:0 and PE-Cer 18:0;2O/28:1, was observed during E. invadens encystation (Fig. S1B), although the total amount of PE-Cers in cells did not change ( Fig. 2A). Because changes in the level of PE-Cer-NDSs and Cer-NDSs levels were well correlated during the course of cyst formation (Fig. 2C andS1A and B), PE-Cer-NDSs appeared to be synthesized de novo via Cer-NDSs. Note that previous studies determined the effects of E. histolytica and E. invadens CerS2 gene knockdown or overexpression on trophozoite proliferation, encystation, and excystation (25,26). The observed phenotypes, at least for E. histolytica trophozoite proliferation, were inconsistent with our present results from the E. histolytica genetic study (Fig. S4B). We attribute this inconsistency to the functional redundancy among EhCerS2, -5, and -6. This genetic redundancy may also affect the encystation and excystation, because E. invadens possesses all of these counterparts (AmoebaDB) (26) (Fig. 1B). However, the possibility that CerS2 specifically functions in these processes cannot be ruled out; therefore, alternative approaches, such as pharmacological blockage of specific CerS, are required for elucidating the roles of Cer-NDS species, products of CerS, during Entamoeba encystation and excystation. Taken together, Entamoeba provides the necessary diversity of sphingolipids, such as Cer, PE-Cer, PI-Cer, and SM. However, the precise physiology of these sphingolipids in Entamoeba, including identification and characterization of sphingolipid synthase(s) and the uptake mechanism of SM from the host, needs to be unraveled.
As well as ceramides, sphingolipid and glycerophospholipid diversity are generated by variations in acyl chains, i.e., the number of carbon atoms and the level of unsaturation (Fig. S1E to K). The acyl chain variations in these lipids are principally introduced by a ubiquitous enzyme, acyl-CoA synthetase, which uses various fatty acids as a substrate. Organisms typically utilize fatty acids per se, which are either scavenged from the external milieu or synthesized by a de novo pathway. After elongation and desaturation by fatty acid elongases and desaturases, respectively, these provide fatty acids. Unlike typical organisms, such as human and yeast, Entamoeba relies totally on the external milieu as the fatty acid source because genes for neither type I nor II fatty acid synthases, responsible for de novo synthesis, are present in the genome (34,40,41). Furthermore, fatty acid desaturases are not encoded. In contrast, all enzymes necessary for fatty acid elongation, which proceeds via a four-step biochemical cycle (42,43), are encoded in Entamoeba genomes (AmoebaDB) (34, 40) (see Fig. S7A). Consistently, during encystation, significant upregulation of E. invadens genes that encode enzymes involved in fatty acid elongation was observed (Fig. S7B). Notably, knockdown of the gene encoding the second enzyme of the pathway in E. histolytica produced a severe growth defect. Therefore, Entamoeba fatty acid elongation, along with other lipid metabolism, such as sphingolipid and sulfolipid metabolism, is crucial for maintenance of the life cycle (this study, 34).
In conclusion, we have shown the overall scheme of Entamoeba sphingolipid metabolism and its unique features. These findings substantiate the importance of lipid metabolism in Entamoeba encystation and indicate a new role for ceramides in organism homeostasis. This contributes not only to the advances in understanding Entamoeba physiology but also to the field of sphingolipid and membrane biology.

MATERIALS AND METHODS
Parasite cultures. E. histolytica (G3 and HM-1:IMSS cl6) were routinely maintained as previously described (44). E. invadens (IP-1) was routinely maintained in a glass tube filled with 6 ml BI-S-33 (proliferation medium). To induce encystation, 2.5 Â 10 5 E. invadens trophozoites were seeded in a Nunc cell culture flask with a solid cap (catalog number [no.] 163371; Thermo Fisher Scientific, Waltham, MA, USA) filled with 56 ml BI-S-33 medium and cultivated at 26°C for 5 days. Trophozoites were harvested from the required numbers of flasks and transferred to encystation medium (37) at a final concentration of 6 Â 10 5 /ml.
LC-MS/MS-based lipidomics. E. invadens cyst formation was induced as previously described in either the absence or presence of 1 mM myriocin (37). One micromolar myriocin was freshly diluted from 5 mM stock, which was prepared by dissolving myriocin powder (Cayman, MI, USA) in dimethyl sulfoxide (DMSO) and stored at 230°C. Sample containing 0.02% DMSO was used as a control of myriocin treatment. Briefly, trophozoites suspended in encystation medium (6 Â 10 5 cells/ml) were seeded in 24-well culture plates (2 ml per well) and sealed as described (45) using Parafilm (Bemis Company, Inc., Oshkosh, WI, USA). Then, plates were incubated at 26°C for the period indicated in the text and figures. Cell pellets from two wells of a 24-well plate were collected in a single 15-ml tube using 10 ml phosphate-buffered saline (PBS) and then centrifuged at 770 Â g for 5 min at 4°C. The cell pellet was washed with 6 ml PBS and resuspended in 4 ml PBS. One milliliter of the cell suspension was then dispensed into each of four 1.5-ml tubes, and cells were repelleted by centrifugation. Cell pellets in tubes were kept at 280°C until use.
For E. histolytica transformants, stably subculturing cells (1.5 Â 10 6 ) in the presence of 20 mg/ml G418 disulfate (Nacalai Tesque, Kyoto, Japan) were collected in a 5-ml tube by centrifugation at 440 Â g for 5 min at 4°C. The cell pellet of each transformant was washed with 4 ml PBS and resuspended in 1.5 ml PBS. Five hundred microliters of the cell suspension was dispensed into each of three 1.5-ml tubes, and cells were repelleted by centrifugation at 770 Â g for 5 min at 4°C. Cell pellets were then kept at 280°C until use.
Lipids were extracted from cells using single-phase extraction as previously described (46) with minor modifications. The cell pellet prepared as described above was mixed with 0.5 ml methanol, sonicated for 2 min, and incubated for 1 h at ambient temperature. After 0.2 ml of the obtained suspension was mixed with 0.1 ml CHCl 3 in a new glass tube, the sample was incubated for 1 h at ambient temperature. Then, 20 ml water was added to the sample, and the mixture was incubated for 15 min at ambient temperature. After the extract was centrifuged at 2,000 Â g for 10 min at ambient temperature, the supernatants were collected and dried. The obtained lipids were resuspended in 50 ml methanol (MeOH)-CHCl 3 -H 2 O solution (2:1:0.2 [vol/vol/vol]), and were then kept at 4°C until use.
LC-MS/MS analysis was carried out using a quadrupole time of flight mass spectrometer, TripleTOF 6600 (SCIEX, Framingham, MA, USA) coupled with an ACQUITY ultraperformance liquid chromatography (UPLC) system (Waters, Milford, MA, USA). All analyses were performed using data-dependent MS/MS acquisition (DDA) at the high-resolution mode in MS1 and at the high sensitivity mode in MS2. The UPLC peptide ethylene-bridged hybrid (BEH) C 18 (50 by 2.1 mm; 1.7 mm) column was maintained at 45°C at a flow rate of 0.3 ml/min. The LC separation was performed with a gradient elution of mobile phase A (methanol-acetonitrile-water, 1:1:3 [vol/vol/vol] containing 5 mM ammonium acetate [Wako Chemicals, Osaka, Japan] and 10 nM EDTA [Dojindo, Kumamoto, Japan]) and mobile phase B (isopropanol containing 5 mM ammonium acetate and 10 nM EDTA). The LC gradient and mass spectrometer settings were the same as previously described (46). The data analysis was performed as previously described (20). The obtained data were, as a result, normalized by adjusting the cell numbers processed; for encysting cells, the cell numbers were those treated for encystation induction, whereas for transformants, those treated for lipid extraction were used.
Metabolic labeling of E. invadens and lipid analysis. E. invadens trophozoites suspended in proliferation medium (1.5 Â 10 5 /ml) or encystation medium (6 Â 10 5 cells/ml) were seeded in 96-well culture plates (240 ml per well). After adding U-14 C-labeled L-serine (173.6 mCi/mmol) (Moravek, Brea, CA, USA) to each well (final radioactivity, 3 mCi/ml), the plates were sealed and incubated at 26°C for the period indicated as described above. For each time indicated, cell cultures from four wells of a 96-well plate were collected in a single 6-ml glass tube, and cells were pelleted by centrifugation at 1,500 Â g for 5 min at 4°C. The cell pellet in each tube was washed twice with PBS. Then lipids were extracted by successive addition of 3.8 ml chloroform-methanol-0.15 N HCl (5:10:4 [vol/vol/vol]), 1 ml chloroform, and 1 ml 1% KCl (wt/vol deionized water) with thorough mixing at each addition. Phases were separated by centrifugation at 770 Â g for 5 min at ambient temperature, and the organic phase was recovered and dried. The lipids extracted from 2.88 Â 10 5 cells were resolved by thin-layer chromatography (TLC) on Silica Gel 60 high-performance TLC plates (Merck, Darmstadt, Germany) with chloroform-methanol-15 N Alkaline treatment of lipids. The lipids obtained from 2.88 Â 10 5 cells, as described above, were suspended in 600 ml 0.1 M KOH in chloroform-methanol (2:1 [vol/vol]) and incubated for 2 h at 37°C. After incubation, the lipid solution was sequentially mixed with 21 ml 4 M formic acid, 200 ml chloroform, and 400 ml deionized water. Then, the phases were separated by centrifugation at 770 Â g for 5 min at ambient temperature, and the organic phase was recovered, dried, and dissolved in 50 ml chloroformmethanol (1:1 [vol/vol]) (47). The obtained lipids were resolved by TLC on Silica Gel 60 high-performance TLC plates (Merck, Darmstadt, Germany) with chloroform-methanol-15 N NH 3 (60: 35:8 [vol/vol/vol]). Each spot on the TLC plates was analyzed as described above.
Real-time qRT-PCR. Real-time qRT-PCR was performed as previously described (48) with minor modifications. Total RNA from Entamoeba cells was extracted with RNAiso Plus (TaKaRa Bio Inc., Kyoto, Japan), and then cDNA was synthesized using the ReverTra Ace qPCR RT master mix with genomic DNA (gDNA) remover (Toyobo Co. Ltd., Osaka, Japan). Real-time PCR was performed using StepOnePlus (Thermo Fisher Scientific, Waltham, MA, USA), Thunderbird SYBR qPCR mix (Toyobo), and appropriate primer sets (see Table S3 in the supplemental material). For encysting E. invadens cell analysis, the cell pellets from one aliquot prepared for lipidomics were suspended in 1 ml RNAiso Plus (TaKaRa), and the resulting samples were similarly processed and analyzed.
Gene knockdown in E. histolytica. Construction of pSAP2-g-multi-based plasmids for gene silencing using appropriate primers (Table S3) was performed essentially as described previously (34). Plasmid transfection into E. histolytica (G3) trophozoites using Lipofectamine LTX and establishment of stable transformants were performed as described (34). The knockdown levels of the targeted genes in the established transformants were evaluated by real-time qRT-PCR using suitable primer sets (Table S3) as described above.
Overexpression in E. histolytica. Overexpression of HA-tagged proteins was achieved using the pEhEx-m-HA vector, which was derived from pEhEx-HA (36). The EcoRI-BglII fragment of pEhEx-HA was replaced with a newly PCR-amplified 59 conserved sequence (CS) region appended with EcoRI and NheI-HindIII-BglII sites at the either end. PCR amplicons harboring open reading frames (ORFs) of target genes were obtained using suitable primers sets (Table S3), digested with NheI and BglII, and inserted into the corresponding sites of pEhEx-m-HA. The resulting correct plasmids were then introduced into E. histolytica (HM-1:IMSS cl6) trophozoites, and stable transformants were established as described above. The levels of overexpression of the targeted genes in the established transformants were evaluated by realtime qRT-PCR using suitable primer sets (Table S3) as described above.
Cell growth assay. The cell growth assay was performed as described previously (48,49) except that the starting cell number was 1,666 or 3,000 trophozoites/ml for overexpression or gene knockdown transformants, respectively. The IC 50 of myriocin for trophozoite proliferation was also similarly determined. In detail, trophozoites of either E. histolytica (HM-1:IMSS cl6) or E. invadens (IP-1) were collected from routine cultures and suspended in proliferation medium at 3,333 cells/ml. Three milliliters of the cell suspension was dispensed into a 7-ml glass tube, which was filled with 6 ml proliferation medium containing various concentrations of myriocin. Each myriocin stock, prepared as described below, was added to the medium at 1/100 before trophozoite inoculation. For a solvent control, in place of the myriocin stock solution, DMSO was added at final concentration of 1% (vol/vol). After incubating for 72 h, viable cells were manually counted under a phase-contrast microscope.
Flow cytometry and fluorescence and electron microscopy. E. invadens trophozoites treated for encystation were suspended in encystation medium (37) containing either various concentrations of myriocin or the solvent control, DMSO. Note that the DMSO content in all wells was 1% (vol/vol). Then, the cell suspensions were seeded in 96-well culture plates, sealed, and incubated at 26°C for the designated period, as described above. The medium containing myriocin was prepared by adding each myriocin stock solution at 1/100 (vol/vol). The series of stock solutions was prepared by serial dilution with DMSO from a 5 mM stock.
For flow cytometry using Evans blue (EB) and calcofluor (CF), cells in the above-described cultures were treated and processed, and the obtained data were analyzed as described previously (37). The IC 50 of myriocin for the cyst formation at 72 h after inducing encystation was also determined by this flow cytometry method.
For fluorescence microscopy, a portion of the flow cytometry samples (1 mM myriocin and control) was examined at 24 h under a fluorescence microscope (Zeiss Axio Imager 2; Carl Zeiss, Germany) equipped with a Zeiss AxioCam 305 mono camera (Carl Zeiss). The obtained images were processed using ZEN software (Carl Zeiss).
Transmission electron microscopy analysis, based on a rapid freezing and freeze-fixation method, was outsourced to Tokai Electron Microscopy, Inc. (Nagoya, Japan). The cells treated for encystation were cultivated either in the presence of 1 mM myriocin or DMSO for 24 h in 6 wells of a 96-well plate as described above. Then, the cells were collected in a single 1.5-ml tube using 1 ml PBS and pelleted by centrifugation at 5,200 Â g for 1 min at 4°C. Each cell pellet was sandwiched between copper disks and quickly frozen in liquid propane at 2175°C. The resulting samples were then freeze-substituted with 2% glutaraldehyde and 1% tannic acid in ethanol containing 2% distilled water (vol/vol) at 280°C for 48 h. Subsequently, they were transferred to a 220°C freezer and kept at 220°C for 3 h, followed by warming to 4°C by 4 h. The samples were then dehydrated for 30 min three times in absolute ethanol at ambient temperature and left in absolute ethanol at ambient temperature overnight. On the following day, the samples were soaked for 30 min twice in propylene oxide (PO) and once in a mixture of PO and resin (Quetol-812; Nisshin EM Co., Tokyo, Japan) (70:30 [vol/vol]) for 1 h. The tube cap was left open overnight to completely volatilize the PO. Then, the samples were freshly soaked in 100% resin and incubated at 60°C for 48 h to polymerize the resin. The samples embedded in the polymerized resins were sectioned at an ultrathin thickness of 70 nm using an ultramicrotome (Ultracut UCT; Leica, Vienna, Austria). These samples were stained with 2% uranyl acetate at ambient temperature for 15 min and washed with distilled water. Then, the samples were secondarily stained with lead stain solution (Sigma-Aldrich, St. Louis, MO, USA) at ambient temperature for 3 min. The resulting samples were then examined at 100 kV acceleration voltage under a transmission electron microscope (JEM-1400 Plus; JEOL, Tokyo, Japan). Images were acquired using a charge-coupled-device (CCD) camera (EM-14830RUBY2; JEOL).
Data availability. All raw mass spectrometry data are freely available on the RIKEN DROP Met website (http://prime.psc.riken.jp/menta.cgi/prime/drop_index), under index number DM0036.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.