Phosphatidylcholine synthesis through cholinephosphate cytidylyltransferase is dispensable in Leishmania major

Phosphatidylcholine (PC) is a major cell membrane constituent and precursor of important second messengers. In Leishmania parasites, PC synthesis can occur via the choline branch of the Kennedy pathway, the N-methylation of phosphatidylethanolamine (PE), or the remodeling of exogenous phospholipids. To investigate the role of de novo PC synthesis in Leishmania major, we focused on the cholinephosphate cytidylyltransferase (CPCT) which catalyzes the formation of CDP-choline, a key intermediate in the choline branch of the Kennedy pathway. Without CPCT, L. major parasites cannot incorporate choline into PC, yet the CPCT-null mutants contain similar levels of PC and PE as wild type parasites. Loss of CPCT does not affect the growth of parasites in complete medium or their virulence in mice. These results suggest that other mechanisms of PC synthesis can compensate the loss of CPCT. Importantly, CPCT-null parasites exhibited severe growth defects when ethanolamine and exogenous lipids became limited or when they were co-cultured with certain bacteria that are known to be members of sandfly midgut microbiota. These findings suggest that Leishmania employ multiple PC synthesis pathways to utilize a diverse pool of nutrients, which may be crucial for their survival and development in the sandfly.

Cpct − mutants cannot synthesize pC from choline but can incorporate etN into pe and pC. To determine whether CPCT is required for the synthesis of CDP-choline, E. coli lysates containing recombinant L. major CPCT (LmCPCT) or S. cerevisiae CPCT (ScCPCT) were incubated in the presence of CTP and radiolabeled choline-P and the products were examined by thin layer chromatography (TLC). As shown in Fig. 4A and Fig. S3A, lysate containing ScCPCT could efficiently catalyze the formation of CDP-choline whereas lysate from empty pET vector had no activity. By comparison, LmCPCT conferred a low but clearly detectable level of CPCT activity (~26% of ScCPCT, Fig. 4A and Fig. S3A). It is not clear whether this result reflects the intrinsic difference between ScCPCT and LmCPCT in specific activity, or their ability to be functionally expressed in E. coli.
In agreement with this finding, L. major WT parasites but not cpct − mutants could incorporate [ 3 H]-choline into PC including 1,2-diacyl-PC (PtC) and lyso-PC (LPC, a hydrolytic product of PtC 49 ) ( Fig. 4B and Fig. S3B). Complementation of cpct − with CPCT, GFP-CPCT or CPCT-GFP led to robust assimilation of choline into PtC and LPC ( Fig. 4B and Fig. S3B). These findings indicate that CPCT is solely responsible for generating PC from choline in L. major. As we reported previously 23 , WT parasites were able to incorporate [ 3 H]-EtN into PE (PME + 1,2-diacyl-PE + lyso-PE or LPE 49 ) and PC ( Fig. 4C and Fig. S3C); while similar results were observed with the cpct − mutants, we detected a 1.5-1.9-fold increase in the incorporation of [ 3 H]-EtN into PC ( Fig. 4C and Fig. S3C), which could be a compensatory response to the loss of PC synthesis from choline. www.nature.com/scientificreports www.nature.com/scientificreports/ Cpct − mutants contain normal levels of pC and pe when grown in the complete medium. To examine whether the choline branch of Kennedy pathway plays a major role in the overall PC production in L. major, we extracted total lipids from promastigotes cultured in the complete M199 medium (M199 medium with 10% fetal bovine serum or FBS and other supplements) 50 . The composition of PC was assessed using electrospray ionization mass spectrometry (ESI/MS) in the positive-ion mode. As shown in Fig. 5A-F, in log phase and stationary phase, the LPC and diacyl-PC species in cpct − mutants closely resembled those found in WT and cpct − /+CPCT parasites. Through comparison with a PC standard, we estimated the overall abundance of PC in log phase cpct − mutants to be 6.5-7.1 × 10 8 molecules/cell, which was close to the average values in log phase WT and cpct − /+CPCT parasites (Fig. 5G). During the stationary phase, cpct − mutants contained 2.6-3.1 × 10 8 PC molecules/cell, whereas WT and cpct − /+CPCT parasites had 3-4 × 10 8 molecules/cell ( Fig. 5G; the difference between WT and cpct − is not statistically significant). The decrease of PC abundance in stationary phase is consistent with the global lipid remodeling when promastigotes transition from replicative procyclics to infectious metacyclics 10,51 .
In addition to PC, we examined the cellular levels of PE in log phase promastigotes. As summarized in Fig. 5G, no significant difference was detected between WT, cpct − and cpct − /+CPCT parasites (1.7-2.1 × 10 8 PE molecules/cell). Thus, the choline branch of Kennedy pathway is not required for bulk phospholipid synthesis when L. major parasites were cultured in the complete medium. These results suggest that Leishmania promastigotes can compensate the loss of CPCT through PE N-methylation and/or lipid uptake followed by remodeling to meet the demand of PC synthesis.

Role of CpCt in Leishmania differentiation and growth under EtN-limiting conditions.
When cultivated in the complete medium, cpct − mutants grew and replicated at a similar rate as WT and cpct − /+CPCT parasites (Fig. 6A). These mutants produced less metacyclics (the infective form to mammals) 52 than WT and add-back parasites during the stationary phase (Fig. 6B). In L. major, formation of metacyclics (metacyclogenesis) is associated with the modification of lipophosphoglycan (LPG) 53 . Western blot analysis revealed that LPG in stationary phase cpct − was of normal abundance, but migrated slightly slower than that from WT and cpct − /+CPCT parasites (Fig. S4), suggesting minor structural alteration of LPG.
To further examine the contribution of CPCT in vitro, we monitored the growth of cpct − mutants under various EtN-limiting conditions. When promastigotes were inoculated in complete M199 medium in the presence of myriocin, which inhibits the conversion of serine into EtN-P ( Fig. 1), no major growth defects were observed in the cpct − mutants (Fig. 6C). This suggests that uptake of exogenous PE and/or PC is sufficient to compensate the loss of CPCT (Fig. 1). Similarly, in a lipid-free M199 medium (M199 without FBS but with 0.4% fatty acid free bovine serum albumin), cpct − mutants proliferated nearly as well as WT and cpct − /+CPCT parasites showing a slight but not statistically significant delay (Fig. 6D). This suggests that in the absence of lipid uptake, the serine-to-EtN conversion is largely sufficient for PC synthesis (Fig. 1). However, when inoculated in lipid-free M199 (no exogenous lipids available to uptake) containing myriocin (inhibiting serine-to-EtN conversion), cpct − mutants grew significantly slower than WT and cpct − /+CPCT parasites (Fig. 6E). Finally, supplementation of EtN rescued the growth of cpct − to WT levels under this condition (Fig. 6E). Together, these results suggest that in comparison to WT and cpct − /+CPCT parasites, cpct − mutants are more dependent on PE N-methylation and the uptake of exogenous lipids for PC synthesis.

Cpct − mutants do not have virulence defects in mice.
To study the role of de novo PC synthesis in Leishmania virulence, mice were infected subcutaneously in the footpad with late stationary phase WT, cpct − , and cpct − /+CPCT promastigotes. In both BALB/c mice (susceptible to L. major; Fig. 7A) and C57BL/6 mice (resistant to L. major; Fig. 7C) 54 , cpct − mutants induced lesions of similar sizes over time as WT and cpct − /+CPCT  (Table S1). ( www.nature.com/scientificreports www.nature.com/scientificreports/ parasites. Limiting dilution assay was performed at 5, 12, or 13 weeks post infection to determine parasite burden ( Fig. 7C-D). No significant difference was observed between WT and cpct − parasites. Therefore, although CPCT affects metacyclogenesis in vitro (Fig. 6B), it is not required for the virulence of L. major, suggesting that amastigotes can fulfill their need for PC synthesis through lipid salvage or PE N-methylation ( Fig. 1).
Cpct − mutants show growth delay when co-cultured with certain bacteria. In the midgut of sandfly, Leishmania promastigotes need to proliferate and differentiate in the presence of a community of microorganisms [55][56][57][58][59] . The interaction between parasites and sandfly microbiota has significant impact on Leishmania differentiation and transmission 55,60,61 . While PC is the most abundant phospholipid in eukaryotes, most bacteria synthesize PE as a major membrane lipid and have a high demand for EtN [62][63][64][65] . Here we examined whether CPCT was required when L. major promastigotes were co-cultured with Serratia marcescens or Enterobacter cloacae, two Gram-negative bacteria that have been identified in the midgut of several sandfly species 59,66,67 . In this experiment, promastigotes (1.0 × 10 6 cells/ml) and bacteria (50 cells/ml) were co-cultured in lipid-free M199 media and the density of Leishmania parasite was determined after 24 hours. As shown in Fig. 8, all parasites grew slower in the presence of S. marcescens presumably due to nutrient competition and waste/toxin production from the bacteria. For WT parasites, a 37% reduction was observed during S. marcescens co-culture, while cpct − and www.nature.com/scientificreports www.nature.com/scientificreports/ cpct − /+CPCT mutants displayed 61% and 45% reduction respectively (Fig. 8, lipid free M199 vs. S. marcescens). Thus, losing CPCT seems to reduce parasites' fitness under a competitive condition. Importantly, addition of EtN largely restored the replication of cpct − to levels close to WT and cpct − /+CPCT parasites (Fig. 8, S. marcescens vs. S. marcescens + EtN). By comparison, a less significant effect on Leishmania growth was observed during co-culture with Enterobacter cloacae, although cpct − mutants still replicated slower than WT and add-back www.nature.com/scientificreports www.nature.com/scientificreports/ parasites, and EtN supplementation improved their growth (Fig. 8, E. cloacae vs. E. cloacae + EtN). Together, these findings suggest that the choline branch of the Kennedy pathway allows Leishmania to be less dependent on EtN, which may be a limiting nutrient in the sandfly (Fig. 1).

Discussion
In this study, we characterized an ER-localized CPCT that is responsible for the incorporation of choline into PC in L. major. Deletion of CPCT had no obvious impact on the cellular levels of PC or PE when promastigotes were cultivated in complete M199 media (Fig. 6). Cpct − mutants replicated normally and did not show any defects in morphology when they were able to generate PC through PE N-methylation and/or lipid salvage. We observed a 2-4 folds reduction in metacyclogenesis during late stationary phase (Fig. 6B), which may be related to the minor alteration of LPG in cpct − (Fig. S4) 53,68 . Importantly, the fact that of cpct − mutants are fully virulent and replicative mice (Fig. 7) suggests that the de novo synthesis of PC is not required during the mammalian stage of L. major. www.nature.com/scientificreports www.nature.com/scientificreports/ Meanwhile, the proliferation of cpct − mutants was severely reduced when they were cultivated in a lipid-free medium (which eliminated the uptake, degradation and remodeling of exogenous lipids) containing myriocin (which blocked the conversion of serine into EtN-P via sphingoid base metabolism) (Fig. 6). These findings  . Cpct − mutants show significant growth delay when co-cultured with bacteria and the growth delay can be rescued by EtN supplementation. Effects of Serratia marcescens and Enterobacter cloacae on the growth of cpct − mutants were determined in a co-culture experiment. L. major promastigotes (1 × 10 6 cells/ml) were incubated in lipid-free M199 medium alone or with Serratia marcescens (50 bacteria/ml) or Enterobacter cloacae (50 bacteria/ml). EtN (250 μM) was included as indicated. The number of promastigotes/ml was recorded using a hemocytometer after 24 h. Results were averaged from three independent experiments with triplicates. Error bars represent standard deviations from three biological repeats (*p < 0.05; ***p < 0.001) based on one way Anova relative to WT. www.nature.com/scientificreports www.nature.com/scientificreports/ indicate that the choline branch of the Kennedy pathway is dispensable when parasites can synthesize PC from EtN, serine, or exogenous lipids (Fig. 1).
The fact that cpct − mutants can still grow (albeit at a reduced rate) and synthesize PC in lipid-free M199 containing myriocin (Fig. 6D) suggests that residue amount of EtN/EtN-P (which may be converted from serine after myriocin loses efficacy) is sufficient to sustain a low level of PC synthesis and cell proliferation. Alternatively, parasites may generate PE/EtN from serine via the activity of phosphatidylserine synthase 2 and phosphatidylserine decarboxylase 69 (Fig. 1). These two enzymes are essential for the optimal growth and mitochondrial function of procyclic Trypanosoma brucei, and their exact roles in Leishmania have yet to be determined 69 .
Since cpct − mutants are fully replicative and virulent in mice (Fig. 7), we postulate that the ability to synthesize PC from choline is dispensable during the mammalian stage when amastigotes reside within the phagolysosome of macrophages and have access to lipids, amino acids, sugars and amino sugars [70][71][72] . Meanwhile, it is of interest to determine if CPCT contributes to the proliferation of Leishmania promastigotes in the midgut of sandfly, where they must compete for nutrients with the resident microbiota 55,73 (Fig. 8). While the exact composition of carbon source in the sandfly gut is not well defined, the diet of female sandfly consists of blood meals and nectar, which is digested by the hydrolytic enzymes from the sandfly and microbiota. Because EtN or PE is an important source of carbon and nitrogen for many bacterial species 64,65 , these nutrients may be limited in the sandfly gut. In contrast, choline is not an essential nutrient for many bacteria which mainly synthesize PE, phosphatidylglycerol (PG), and cardiolipin as their membrane phospholipids 74 . Retaining the choline branch of the Kennedy pathway may allow Leishmania to synthesize PC from choline when EtN/exogenous phospholipid is limited. In support of this hypothesis, cpct − mutants showed a significant growth delay when co-cultured in a lipid-free medium with Serratia marcescens or Enterobacter cloacae, two bacteria found in both field-captured and lab-reared sandflies, and the supplementation of EtN largely restored the proliferation of cpct − mutants (Fig. 8). The reason why E. cloacae has less inhibitory effect on Leishmania than S. marcescens may involve a combination of factors including their rates on EtN consumption, nutrient depletion and toxin production.
In summary, L. major promastigotes may retain the choline branch of the Kennedy pathway to survive nutrient-limiting conditions. Such metabolic flexibility may improve their competitive fitness in the sandfly midgut.

Molecular cloning.
With the exception of cloning Saccharomyces cerevisiae CPCT into pET vector (described below), CPCT refers to L. major CPCT in this study. To facilitate CPCT assay, the ORF of Saccharomyces cerevisiae CPCT (YGR202C) was amplified from yeast genomic DNA by PCR using primers 5′-GTACTGGGATCC ATGGCAAACCCAACAACAGGGAAGTCC-3′/5′-GATCATGGATCCCAGTTCGCTGATTGTTTCTTCTTC-3′. The resulting 1280 bp DNA fragment was digested with BamHI and ligated into E. coli expression vector pET23a + to generate pET-ScCPCT. The ORF of L. major CPCT was cloned into pET23a + to generate pET23a-LmCPCT. These pET constructs were transformed into E. coli BL21 (DE3) competent cells.
All the molecular constructs were verified by restriction enzyme digestion and DNA sequencing.
Leishmania promastigote culture and genetic manipulation. Unless specified otherwise, L.
major LV39 clone 5 (Rho/Su/59/P) promastigotes were cultured at 26 °C in M199 media with or without 10% heat-inactivated FBS and other supplements 50 . Cell density was determined using a hemocytometer. Percentages of metacyclics in stationary phase culture were determined as previously described 68 . To investigate the functions of CPCT, the endogenous CPCT alleles were deleted from L. major LV39 wild type (WT) parasites by two consecutive rounds of targeted replacement as described 50  www.nature.com/scientificreports www.nature.com/scientificreports/ were confirmed by Southern blot, where genomic DNA from promastigotes was digested with SacII, separated on a 0.7% agarose gel, transferred to a nitrocellulose membrane, and probed with [ 32 P]-labeled DNA fragments corresponding to the ORF or an upstream region of CPCT. Results of Southern blot were visualized by autoradiography. To restore CPCT expression, pXG1-CPCT was transfected into cpct − mutants and referred to as cpct − /+CPCT (∆CPCT::PAC/∆CPCT::HYG/+pXG1-CPCT). For localization studies, pXG1-GFP-CPCT or pXG1-CPCT-GFP was introduced into cpct − to generate cpct − /+GFP-CPCT or cpct − /+CPCT-GFP, respectively.
To examine cell growth under various nutrient-limiting conditions, we prepared M199 media that are supplemented with either 10% FBS (complete medium), 0.4% fatty acid free bovine serum albumin (lipid-free M199 medium, which contains 23-24 μM of serine but no EtN), 0.4% bovine serum albumin plus 4 μM of myriocin, or 0.4% bovine serum albumin plus 4 μM of myriocin and 250 μM of EtN. At low densities (<1.0 × 10 6 cells/ ml), cells were concentrated 10-fold before counting, and at high densities (>1 × 10 7 cells/ml), cells were passed through a 27 ½ G needle three times to break up clumps before counting. To measure the percentage of dead cells parasites were labeled with 5.6 µg/ml propidium iodide followed by flow cytometry.
Metabolic labeling, CpCt assay and thin layer chromatography (tLC). To  To examine the activity of recombinant CPCT, E. coli BL21 (DE3) transformants containing pET23-LmCPCT, pET23-ScCPCT, or empty pET23a + vector were grown in LB media until OD 600 nm reached 0.6 and induced with 1 mM of IPTG for 3 hours. Bacteria were resuspened in a lysis buffer (Tris HCl 30 mM at pH7.4, 5% glycerol, 1 mM EDTA, and 1 x protease inhibitor) and sonicated on ice. Protocol of CPCT assay was adapted from a previous report 76 .
Briefly, E. coli lysate containing 60 μg of bacterial protein was incubated in a 20 μl reaction mix containing 25 mM of MgCl 2 , 5 mM of CTP, and 0.1 mM of phosphorylcholine [methyl-14 C] for 30 minutes at 30 °C. Reaction mix (10 μl each) was then analyzed by TLC using silica-60 plates and a solvent made of ethanol:0.5% NaCl:25% ammonium hydroxide (10:10:1 by volume). Radioactive signals were detected using a Personal Molecular Imager (Bio-Rad).
Fluorescence microscopy. Promastigotes expressing GFP-CPCT or CPCT-GFP were attached to poly-L-lysine coated coverslips, fixed with 3.7% formaldehyde, and then permeabilized on ice with ethanol. Incubation with rabbit anti-T. brucei BiP antiserum (1:1000) was performed at room temperature for 40 minutes. After washing, coverslips were incubated with a goat anti-rabbit-Texas Red (1:2000) antiserum for 40 minutes. An Olympus Fluoview FV3000 Laser Scanning Confocal Microscope was used to visualize the expression and localization of GFP-CPCT. To quantify the overlap between GFP-CPCT and anti-BiP staining, 30 randomly selected cells were analyzed using Image J JACoP (Just Another Colocalization Plugin) 48 .

Mouse infection.
Use of mice in this study was approved by the Animal Care and Use Committee at Texas Tech University (US PHS Approved Animal Welfare Assurance NO. A3629-01). BALB/c and C57BL/6 mice (female, 8 weeks old) were purchased from Charles River Laboratories International. Mice were housed and cared for in the facility operated by the Animal Care and Resources Center at Texas Tech University adhering to the Guide for the Care and Use of Laboratory Animals (the 8th Edition, NRC 2011) for animal husbandry. To maintain virulence, promastigotes were injected into the footpads of BALB/c mice and recovered after 3-4 weeks to start low passage in vitro cultures. To assess virulence, day 3 stationary phase promastigotes (cultured for less than five passages after recovery from mice) were resuspended in DMEM and injected into the left hind footpads of mice (1.0 × 10 6 cells per mouse, 5 mice per group). Lesion sizes were measured weekly using a Vernier caliper and parasite loads were determined by limiting dilution assay 79 .