Sensing Host Arginine Is Essential for Leishmania Parasites’ Intracellular Development

In this study, we report that the ability of the human pathogen Leishmania to sense and monitor the lack of arginine in the phagolysosome of the host macrophage is essential for disease development. Phagolysosomes of macrophages are the niche where Leishmania resides and causes human leishmaniasis. During infection, the arginine concentration in the phagolysosome decreases as part of the host innate immune response. An arginine sensor on the Leishmania cell surface activates an arginine deprivation response pathway that upregulates the expression of a parasite arginine transporter (AAP3). Here, we use CRISPR/Cas9-mediated disruption of the AAP3 locus to show that this response enables Leishmania parasites to successfully compete with the host macrophage in the “hunger games” for arginine.

P rotozoan parasites of the genus Leishmania are the causative agents of a wide spectrum of human and veterinary diseases. Leishmania species cause morbidity and mortality throughout the Old and New Worlds, with clinical manifestations ranging from lesions of the skin (cutaneous leishmaniasis [CL]) and mucous membranes (mucocutaneous leishmaniasis [MCL]) to lethal infection of the spleen and liver (visceral leishmaniasis [VL]) (1). Approximately 350 million people in 88 countries are at risk of VL, with as many as 500,000 new cases diagnosed every year, 10% of which are fatal (2).
Leishmania donovani, the causative agent of kala-azar (VL), exhibits a digenetic life cycle that includes both insect and mammalian forms. Extracellular promastigotes develop in the alimentary tract of sand flies. Following infection of the mammalian host, promastigotes differentiate into intracellular amastigotes within the phagolysosome of macrophages (3,4). This differentiation process in the host can be mimicked in axenic cultures by shifting promastigotes from an insect-like (26°C and pH 7) to an intralysosomal (37°C, pH 5.5, and 5% CO 2 ) environment (5)(6)(7).
During infection, Leishmania parasites encounter macrophage defense mechanisms designed to interdict parasite invasion and block their intracellular survival, including the release of reactive oxygen species (ROS) and the synthesis of cytotoxic nitric oxide (NO) by inducible nitric oxide synthase (iNOS). Yet for macrophages to produce effective amounts of NO, they must import arginine from the extracellular environment. Countering the host NO attack, parasitic invasion activates macrophage arginase 1, which converts arginine to ornithine, the first substrate of the polyamine pathway, thereby suppressing NO synthesis and promoting parasite survival (8). However, the reduction of the host arginine pool becomes a doubleedged sword for Leishmania parasites in infected macrophages since they cannot synthesize arginine de novo. Instead, they must import exogenous arginine via a monospecific amino acid transporter (AAP3) (9), using it primarily in the polyamine pathway to provide precursors for trypanothione biosynthesis (10). Thus, they encounter an intriguing metabolic dilemma: on the one hand, emptying the host arginine pool provides an advantage (reduction of NO), but at the same time, it causes a potentially deadly disadvantage by blocking the supply of an essential amino acid (8). Therefore, to survive, Leishmania parasites need to sense and respond to changes in arginine availability.
A few years ago, we discovered that upon arginine starvation, Leishmania parasites promptly activate a mitogen-activated protein kinase 2 (MAPK2)-mediated arginine deprivation response (ADR) pathway, resulting in the upregulation of the expression and activity of the Leishmania arginine transporter (AAP3) (11). Significantly, the ADR is also activated during macrophage infection due to intracellular parasites actively depleting arginine within the host phagolysosome (12).
The L. donovani genome contains two (haploid) AAP3 gene copies that are tandemly arrayed on chromosome 31 (chr31), which is tetrasomic in L. donovani (and most other Leishmania species). While the coding sequences (CDSs) of the AAP3.1 and AAP3.2 genes are (almost) identical, their 3= untranslated regions (UTRs) are quite different, and only (4.5-kb and 3.5-kb) mRNAs from the latter (AAP3.2) are upregulated under conditions of arginine deprivation (10). Genome-scale transcriptomic analysis (RNAseq) revealed only five other changes in gene expression (mostly other transporters) associated with arginine deprivation, indicating the presence of a coordinated ADR (11).
Nutrient sensing is an essential means for parasites to adapt to and successfully settle inside the vector and host (13). Studies by several laboratories have recently identified intriguing nutrient sense response mechanisms in Leishmania. Martin et al. (14) reported that intracellular sensing of purine starvation directs promastigotes into long-term cell cycle arrest at G 1 -G 0 until a new supply of purine is provided. Starvation of Leishmania mexicana promastigotes for glucose induced a 50-fold increase in the abundance of LmGT1, a glucose transporter that localizes to the flagellum (15). In the case of the facultative intracellular parasite Leishmania, arginine availability is of particular interest since it is important for both host defense and parasite proliferation. Sensing of arginine levels in the lysosome lumen is a key mechanism that regulates mTORC1 activity in mammalian cells (16). In macrophages, mTORC1 activation induces a Th1 response (17). This pathway is the major means by which macrophages kill invading pathogenic microorganisms. Interestingly, Leishmania parasites are able to counteract this outcome by activating a Th2 response, directing arginine toward polyamine biosynthesis instead of NO production, thereby enabling the parasites to persist and cause long-term nonhealing infections (18,19). However, both Th1 and Th2 responses result in arginine depletion in the macrophage phagolysosome, presenting an existential threat to parasite survival since they are unable to synthesize this essential amino acid.
In this study, we explore the mechanism that enables intracellular Leishmania to win this "hunger game." Using CRISPR/Cas9, we created mutants that lack AAP3.2 and thereby are unable to upregulate AAP3 expression after arginine starvation. These mutants were unable to grow in either THP-1 macrophages or BALB/c mouse livers. This study shows that sensing host nutrients is essential for intracellular parasite development.

RESULTS AND DISCUSSION
The Leishmania genome contains two genes (AAP3.1 and AAP3.2) in a tandem array on chromosome 31 that encode arginine transporters. They are highly conserved (only 3 amino acid differences) within their CDSs and 5= UTRs but have very different 3= UTRs. However, only AAP3.2 is responsive to arginine deprivation (10). To assess whether the AAP3.2 response to arginine deprivation is necessary for parasite intracellular development, we aimed to delete it from the L. donovani genome. Unfortunately, chromosome 31 is tetraploid in L. donovani (20), making classical homologous gene replacement approaches cumbersome and tedious, especially since the plasticity of the Leishmania genome enables parasites to retain an additional wild-type (WT) chromosome as well as those containing the selectable marker(s) (21). Therefore, we used the Leishmaniaadapted CRISPR/Cas9 system (22) to expedite the disruption of the AAP3 genes.
WT L. donovani was transfected (separately) with CRISPR plasmids containing 21nucleotide (nt) guide RNAs (gRNAs) (G1, G2, and G3) targeting the 5= ends (positions ϩ19, 85, and 173) of both the AAP3.1 and AAP3.2 CDSs (Fig. 1A, top), and the cultures were examined 6 weeks later for AAP3 protein levels in the presence or absence of arginine. While the G1-transfected cultures showed enhanced ADR-mediated increases in AAP3 levels compared to the WT, the G2 and G3 cultures showed a diminished capacity to upregulate AAP3 levels after arginine starvation (Fig. 1B, top left). The G3 culture was seeded onto agar plates, and individual colonies were examined for an ADR, with different clones showing normal (Δap3G3-2), enhanced (Δap3G3-17), or reduced (Δap3G3-1) increases in AAP3 protein abundance compared to the WT (Fig. 1B, top right). In order to increase the efficiency of CRISPR/Cas9-mediated gene disruption, the Δap3G3-1 mutant was transfected (four times at 3-day intervals) with a 61-nt "donor" oligonucleotide consisting of an 11-nt insertion with a stop codon in all frames and 25-nt sequences flanking the Cas9 cleavage site (Fig. 1A, bottom). Single colonies were isolated 3 weeks after transfection and examined for AAP3 expression levels after arginine deprivation. Two clones (Δap3D6 and Δap3D10, D6 and D10, respectively) showed no increase in AAP3 protein abundance after arginine deprivation, while two others (Δap3D4 and -D7, D4 and D7, respectively) showed a small (less than WT) increase in AAP3 protein levels (Fig. 1B, bottom). Analyses of the initial rates of arginine transport confirmed that these mutants had lost all or some of their response to arginine deprivation ( Fig. 1C; see also Fig. S1 in the supplemental material). While WT cells showed a 2.7-fold increase in the initial rate of arginine transport 2 h after arginine starvation, most of the mutants showed little or no increase in the transport rate after arginine starvation (1.5-, 1.2-, 1.5-, and 1.1-fold for Δap3D4, Δap3D6, Δap3D7, and Δap3D10, respectively). Interestingly, the Δap3G3-17 mutant showed a 3.2-fold (larger than WT) increase in transport after arginine deprivation, consistent with additional copies of the AAP3.2 gene (see below).
Whole-genome sequencing was carried out to map the precise location of the  A small number of reads, 102, from this region are present in the Δap3D10 mutant, suggesting that some cells in this population may retain an intact copy of chr31 (or the sample was contaminated with another cell line). In contrast, the Δap3G3-17 mutant showed higher coverage in this region, consistent with recombination between the two copies of the AAP3 gene resulting in additional (2 to 3 for each copy of chr31) copies of AAP3.2 and the intergenic sequence (including the ncRNA).
Interestingly, the Δap3D6 mutant shows slightly higher than WT levels of read coverage for the AAP3.1 gene, suggesting that it retained one copy of chr31 with the AAP3.1/AAP3.2 fusion along with three copies containing both the AAP3.1 and AAP3.2 genes (one of which may contain an additional AAP3.2 gene). Examination of the read coverage in the gRNA G3 region indicated that 78% (418/536) contain the 11-bp insertion with a stop codon in all frames, consistent with the hypothesis that the Δap3D6 mutant contained seven copies of the AAP3.1 and AAP3.2 genes with stop codons that render them nonfunctional at the protein level and only one (or two) functional version(s). However, the inability of this mutant to present with even a partial ADR on the protein level (such as that of Δap3D4 or Δap3D7) suggests that the only functional copies without stop codons are AAP3.1/AAP3.2 fusions that lack ADR capacity. Northern analyses (Fig. 1D) confirmed the reduction in the AAP3.2 gene copy number in the D4, D7, and D10 cell lines by showing that its mRNA levels after arginine deprivation were significantly lower than those of the WT. These changes paralleled the changes in protein abundance observed in Fig. 1B, confirming that AAP3.2 accounts for most (if not all) of the increase in arginine transport after arginine starvation. Importantly, even though the AAP3.2 mRNA levels in the Δap3D6 mutant were similar to WT levels in both the presence and absence of arginine, AAP3 protein levels were not upregulated in response to arginine deprivation because of the stop codons in the AAP3.2 gene. Two other ADR-responsive genes (LinJ.10.1450 and LinJ.36.2900, which encode pteridine and MFS family transporters, respectively) (11) responded normally to arginine deprivation in all mutants (Fig. S2), indicating that the CRISPR/Cas9 mutations affected only the response of AAP3.2 to arginine deprivation, not the entire ADR pathway.
The analyses described above indicated that the Δap3D6 and Δap3D10 mutants are the most informative AAP3 mutants, retaining near-WT basal levels of arginine transport (when grown in normal medium), but are unable to upregulate transporter activity after arginine deprivation. Therefore, we decided to conduct further studies with these two mutants to determine whether the upregulation of AAP3 expression is necessary for intracellular growth. First, we assessed whether amastigotes retain the phenotypes observed for promastigotes. Axenic promastigotes, WT and mutants, differentiated into amastigotes in culture (7). The rates of differentiation were identical in both the WT and the Δap3D6 and Δap3D10 mutants; i.e., they reached maturation within 5 days (6). Mature amastigotes were then subjected to 2 h of arginine starvation. As shown in Fig. 3A, WT axenic amastigotes responded to arginine deprivation by increasing the AAP3 protein abundance 2.3-fold. In contrast, Δap3D6 and Δap3D10 amastigotes were insensitive to arginine deprivation, as were mutant promastigotes.
In correlation with ADR-driven AAP3 expression in amastigotes, the initial rate of arginine transport in the WT increased by 3-fold following 2 h of arginine deprivation ( Fig. 1B and C). Both Δap3D6 and Δap3D10 amastigotes transported arginine at a rate similar to that of unstarved WT amastigotes but remained unchanged after arginine deprivation. Hence, the results indicate that wild-type and mutant amastigote responses to arginine deprivation were identical to those of promastigotes. This further indicates that the rate of arginine transport in L. donovani is regulated by the abundance of the AAP3 protein.
To determine whether the ADR is necessary for intracellular growth, we infected THP-1 macrophages with late-log-phase L. donovani promastigotes of the WT and Δap3D6, Δap3D10, and Δap3G3-17 mutants in medium containing a physiologically relevant (0.1 mM) concentration of arginine (12,23). As shown in Fig. 4A, the initial level of infection (i.e., the percentage of macrophages infected after 4 h of coincubation) (Fig. 4A, top) and parasite burden (i.e., the average number of parasites per infected macrophage) (bottom) with all three mutants were similar to those of the WT. However, by 48 h postinfection (Fig. 4B), the infectivity and parasite burden for both the Δap3D6 and Δap3D10 mutants were reduced by 2-fold or more compared to the WT, while the levels for Δap3G3-17 were comparable to those of the WT. Hence, our results indicate that while the mutants can infect macrophages as well as the WT, they fail to proliferate normally thereafter. Indeed, it appears that a significant number of macrophages completely cleared their internalized parasites (since the percentage of infected macrophages was lower after 48 h than that after 4 h), while the parasites in the remaining infected macrophages underwent only 1 or 2 rounds of replication (since the parasites per macrophage increased only ϳ2-fold), compared to the normal 3 to 4 rounds of replication and reinfection of new macrophages seen with WT parasites.
To assess whether the lower AAP3 expression levels (and consequently less arginine import) were responsible for the reduction in the infectivity and parasite burden of the Δap3D6 and Δap3D10 mutants, we infected THP-1 macrophages grown in medium containing 1.5 mM arginine, a concentration that prevented the ADR in intracellular WT Arginine Sensing and Leishmania Intracellular Growth ® amastigotes (12). Under these conditions, both the Δap3D6 and Δap3D10 mutants developed normally into amastigotes, with infectivity and parasitemia at 48 h similar to those of the WT (Fig. 4C). The results indicate that the inability to develop inside macrophages is due to arginine deprivation in macrophage phagolysosomes during infection and that the ability to respond to this deprivation by upregulating AAP3.2 protein levels is essential for successful intracellular Leishmania development.
To further assess the role of arginine transport in vivo, we infected BALB/c mice (8 per group) with Δap3D6, Δap3D10, Δap3G3-17, and WT parasites. On day 21 postinfection, the mice were sacrificed, and tissue parasite burdens were determined by quantitative PCR (qPCR) on the DNA extracted from the liver of each mouse. As shown in Fig. 5, the parasite burdens of the Δap3D6 and Δap3D10 mutants averaged only 20 and 24%, respectively, of that of the WT. One-way analysis of variance (ANOVA) and Tukey post hoc honestly significant difference (HSD) testing showed a significant difference (P Ͻ 0.001) from the WT for both the Δap3D6 and Δap3D10 mutants, while the Δap3G3-17 mutant was not significantly different from the WT (P ϭ 0.387). To further evaluate the level of infection, dissected mouse liver samples (n ϭ 6 for each group) were crushed in the presence of Karnovsky fixative to prepare crude liver homogenates. These were fixed and subsequently subjected to scanning electron microscopy (Fig. 6). As shown, livers from mice infected with WT parasites were highly infected with amastigotes ( Fig. 6a to d), while livers infected with the Δap3D6 mutant were almost clear of parasites (Fig. 6e and f), supporting the qPCR data that mutant parasites that lack AAP3.2 are unable to develop colonies in their host.
These results show that the inability of the mutants to express higher levels of AAP3 in order to compensate for the reduced level of arginine in the phagolysosomes of infected macrophages severely compromised their ability to develop in the liver. Since spleens were not included in this analysis, it is possible that differential expression of arginine metabolism-related enzymes in the different organs may influence the role and significance of the AAP3-mediated ADR in parasite survival. Nevertheless, we show here that the liver as a major target of this pathogen presents with significantly lower parasite burdens for mutant parasites than for both WT and AAP3.2-overexpressing parasites (G3-17), which serve as an alternative to an add-back control. Therefore, it appears that the ADR is a crucial mechanism for enabling intracellular Leishmania parasites to overcome the arginine bottleneck.
In summary, this study demonstrates that the ability to monitor metabolic depriva- . On day 21 postinfection, mice were sacrificed, and liver DNA was extracted using the proteinase K method (26). Analyses were carried out as described in Materials and Methods. Arginine Sensing and Leishmania Intracellular Growth ® tion and subsequently induce a specific response at the level of gene expression is essential for the pathogenesis of a protozoan parasite. Furthermore, we have shown that for pathogenic microorganisms to respond to host-inflicted environmental changes and survive, they must employ external sensing and response mechanisms to serve as their monitoring device.
Arginine deprivation was carried out as described previously by Pawar et al. (12). Briefly, mid-logphase promastigotes (1 ϫ 10 7 cells/ml) were washed with Earle's balanced salt solution twice and resuspended in arginine-deficient medium M-199 (Biological Industries Ltd.) at 26°C for the specified period before being transferred to ice. Arginine-deprived cells were washed twice with ice-cold Earle's balanced salt solution before being used for transport assays and Northern and Western blot analyses.
CRISPR/Cas9 guide and donor transfections. gRNA sequences were designed using the Eukaryotic Pathogen CRISPR Guide RNA/DNA Design Tool (EuPaGDT) (http://grna.ctegd.uga.edu/) and cloned into Leishmania-adapted vector pLdCN using the single-step digestion-ligation cloning protocol previously described (22), and the constructs were transfected into mid-log-phase promastigotes. Following gRNA-pLdCN transfections, cells were grown for 4 weeks with G418 at 50 g/ml and subsequently screened for their ability to increase LdAAP3 protein abundance after arginine deprivation. Next, the G3 donor sequence was introduced into a gRNA G3-originated clone exhibiting the desired phenotype. The G3 donor is a single-strand oligonucleotide donor (sense) containing 25-nt sequences flanking the Cas9 cleavage site (shown in boldface type below) and an 11-nt sequence with a stop codon in all three frames (underlined and italicized below). Three transfections with 10 l (100 m) of this oligonucleotide were performed on promastigotes at 3-day intervals as previously described (22). gRNA sequences are as follows: GTCTATTCCAGCACAGGCGG for gRNA G1, GCCGTCGATAAACACCCGAG for gRNA G2, GTGCCGA CGCCGCCAAGCCG for gRNA G3, and ATGAAAACGGTGCCGACGCCGCCAAGTGAGTAGGTAGCCGGGGCGC AACATCATCTTCCG for the G3-based donor.
Arginine deprivation. Arginine deprivation of axenic promastigotes and amastigotes was carried out as described previously by Goldman-Pinkovich et al. (11). Briefly, logarithmic-phase promastigotes or amastigotes (1 ϫ 10 7 cells/ml) were washed twice in ice-cold Earle's salt solution at pH 7 (promastigotes) or pH 5.5 (amastigotes) and subsequently suspended to a final density of 1 ϫ 10 7 cells/ml in arginine-free Earle's salt-based medium 199 (Biological Industries Ltd., Beit Haemek, Israel). This deprivation medium was supplemented with heat-inactivated dialyzed (10,000-kDa cutoff) fetal calf serum at 1% for promastigotes and 2.5% for amastigotes. Deprivation was carried out at either 26°C or 37°C for 2 h and terminated by washing twice in ice-cold Earle's salt solution at the respective pHs. Final suspension was done according to the analyses required.
Transport assays. The uptake of 25 M L-[ 3 H]arginine (600 mCi/mmol) into axenic L. donovani promastigotes and amastigotes was determined using the rapid filtration technique as described previously (11). Briefly, the transport reaction mixture contained 1 ϫ 10 8 cells/ml in ice-cold Earle's balanced salt solution at pH 7 (promastigotes) or pH 5.5 (amastigotes) supplemented with 5 mM glucose. Cells were prewarmed at either 30°C for 10 min (promastigotes) or 37°C for 5 min (amastigotes) prior to the addition of radiolabeled arginine. At 0, 0.5, 1, 1.5, 2, and 3 min, the cell suspensions were filtered through glass fiber GF/C filters that were then washed twice with ice-cold Earle's solution.
The amount of radiolabel associated with the cells was linear with time over the 2-min time course of the assay, so the initial rate of transport was calculated from the slope of the line fitted by linear regression (see Fig. S1 in the supplemental material for promastigotes and Fig. 3B for amastigotes).
RNA isolation and real-time quantitative reverse transcription-PCR. RNA was isolated using Tri reagent (Sigma-Aldrich Ltd.) and a Direct-zol RNA MiniPrep kit (Zymo Research), according to the manufacturers' instructions. Eluted RNA samples were quantified using a NanoDrop One spectrophotometer (Thermo Scientific). Two micrograms of the extracted RNA was subjected to DNase treatment using RQ1 (Promega). Successful DNase treatment was verified by PCR to make sure that no residual DNA could be responsible for amplification. cDNA was synthesized from 2 g of DNase-treated RNA using a qScript cDNA synthesis kit (Quanta Biosciences) in a 40-l total volume. Real-time quantitative reverse transcription-PCR (qRT-PCR) was carried out with the reagents of SsoAdvanced universal SYBR green supermix (Bio-Rad Ltd.) in a 10-l reaction volume (5 l SYBR green, 0.5 nM forward primers, 0.5 nM reverse primers, and 2.5 l cDNA template) on a CFX96 Touch real-time PCR system (Bio-Rad). The AAP3 primers matched both AAP3.1 and AAP3.2. Primers specific for a regulatory subunit of protein kinase A (PKAR=) were used as a control. All the samples were run in triplicates, including a no-template (negative) control for all primers used. Also, RNAseq data of the arginine deprivation response indicated that PKAR= is not affected by the ADR, thus serving as a good control for testing LdAAP3 behavior under ADR and related conditions (Table 1).
PCR was performed at 95°C for 30 s, followed by 39 cycles of 95°C for 10 s and 60°C for 30 s; melt curve analysis was carried out at 65°C to 95°C with 0.5°C increments for 5 s/step; and data analysis was carried out as described previously (25). Briefly, C P values were obtained from all samples in triplicate (except negative controls). Primers were calibrated on pooled samples, and primer efficiencies (E target and E ref ) were calculated and incorporated into the equation below, where LdAAP3, pteridine, or MFS transporters were the target genes and PKAR= served as the reference gene. Ratios were calculated with the C P (crosspoint) of 0-h or 48-h infected THP-1 macrophages. WT L. donovani 48-h infected macrophages served as the control.
Whole-genome sequencing. Genomic DNA was prepared from Leishmania promastigotes using proteinase K digestion followed by phenol-chloroform extraction, ethanol precipitation, and fragmentation to 200 bp using a Covaris S2 sonicator according to the manufacturer's protocol. Next-generation sequencing (NGS) libraries were prepared using a New England BioLabs (NEB) NextUltra II kit, quantified on Agilent bioanalyzer and Qubit instruments, and sequenced on an Illumina HiSeq 4000 instrument to generate 46 million to 96 million paired-end 75-bp reads, depending on the sample. Reads were aligned to the L. donovani 1S genome sequence assembled from a combination of PacBio and Illumina reads (Sur et al., unpublished) using Bowtie2 within Geneious, and the gene copy number was estimated by normalizing the read counts per gene to account for library size and assuming 4 copies of chr31.
Infectivity assays. THP-1 cells grown in RPMI 1640 medium with 10% FBS (catalog no. R8758) and a 1% penicillin-streptomycin solution (catalog no. 03-031-1B; Biological Industries ) in a humidified 37°C, 5% CO 2 air atmosphere were seeded onto glass coverslips (1 ϫ 10 6 cells/well) in a 6-well plate and treated with 50 ng/ml of phorbol myristate acetate (PMA) (catalog no. P8139; Sigma-Aldrich, USA) for 48 h. The cells were infected as described previously (12), and the intracellular parasite load was visualized by Giemsa staining. Briefly, differentiated THP-1 macrophages were infected with log-phase L. donovani promastigotes at a multiplicity of infection (MOI) of 10 for 4 h. Parasites were washed three times in ice-cold Earle's solution (X1) prior to infection to wash out arginine in the growth medium. Infection and subsequent incubation were performed with either 0.1 or 1.5 mM arginine in RPMI 1640 medium. Totals of 0.1 and 1.5 mM arginine in RPMI 1640 medium were prepared from no-arginine, no-leucine, and no-lysine RPMI 1640 medium (catalog no. R1780; Sigma) supplemented with 0.1 or 1.5 mM arginine, leucine (120 g/liter), lysine (70 g/liter), 10% FBS, and a 1% penicillin-streptomycin solution (catalog no. 03-031-1B; Biological Industries). Following 4 h of coincubation, the medium was aspirated from the wells. Cells were washed a total of five times with warm phosphate-buffered saline (PBS) and either collected (0-h time point) or incubated for 48 h (48-h time point) in RPMI 1640 medium containing either 0.1 mM or 1.5 mM arginine until coverslips were collected for Giemsa staining (catalog no. 32884; Sigma), and RNA was collected by direct resuspension of well contents in 600 l Tri reagent (catalog no. T9424; Sigma).
In vivo BALB/c mouse infections and analysis. L. donovani infections were done by tail intravenous (i.v.) injection of 10 8 stationary-phase promastigotes per mouse. On day 21 of infection, the mice were sacrificed, and mouse liver DNA was isolated using the proteinase K method (26). Quantitative PCR was performed on mouse liver DNA using mouse BDNF as the reference gene and parasite gp63 as the target gene with previously reported primer sequences (27,28).
qPCR was carried out with the reagents of SYBR green (catalog no. A25776; Thermo Fisher Scientific) in a 10-l reaction volume (5 l SYBR green, final primer concentrations of 300 nM forward primer and Reverse 5=-GAGGCAGTGCAATGAGAAGC-3= 300 nM reverse primer, and 125 ng DNA template) on a CFX96 real-time PCR system (Bio-Rad). All samples were run in triplicate separately for the 2 primer sets, and the PCR cycle included a 30-s incubation step at 95°C and then 40 cycles of 5 s at 95°C and 30 s at 60°C. The output of normalized expression was determined using the Bio-Rad software of the instrument. Scanning electron microscopy images from crude liver homogenates. Crude liver homogenates were prepared from dissected mouse livers (n ϭ 6) of different cohorts of infection. The dissected mouse liver samples were crushed in the presence of Karnowsky fixative to prepare a crude liver homogenate. The homogenate was processed further with tannic acid fixation followed by 4% osmium tetroxide. The samples were serially dehydrated in ethanol and freon dried. The grids were coated with gold and visualized using an FEI Quanta 200 field emission gun (FEG) electron microscope. The samples were screened for the presence of amastigotes in 200 fields per grid.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.