Sphingolipid degradation in Leishmania (Leishmania) amazonensis.

BACKGROUND: Human leishmaniasis is caused by more than 20 Leishmania species and has a wide range of symptoms. Our recent studies have demonstrated the essential role of sphingolipid degradation in the virulence of Leishmania (Leishmania) major, a species responsible for localized cutaneous leishmaniasis in the Old World. In this study, we investigated the function of sphingolipid degradation in Leishmania (Leishmania) amazonensis, an etiological agent of localized and diffuse cutaneous leishmaniasis in South America. METHODOLOGY/PRINCIPAL
FINDINGS: First, we identified the enzyme LaISCL which is responsible for sphingolipid degradation in L. amazonensis. Primarily localized in the mitochondrion, LaISCL shows increased expression as promastigotes progress from replicative log phase to non-replicative stationary phase. To study its function, null mutants of LaISCL (Laiscl(-)) were generated by targeted gene deletion and complemented through episomal gene add-back. In culture, loss of LaISCL leads to hypersensitivity to acidic pH and poor survival in murine macrophages. In animals, Laiscl(-) mutants exhibit severely attenuated virulence towards C57BL6 mice but are fully infective towards BALB/c mice. This is drastically different from wild type L. amazonensis which cause severe pathology in both BALB/c and C57BL 6 mice.
CONCLUSIONS/SIGNIFICANCE: A single enzyme LaISCL is responsible for the turnover of sphingolipids in L. amazonensis. LaISCL exhibits similar expression profile and biochemical property as its ortholog in L. major. Deletion of LaISCL reduces the virulence of L. amazonensis and the outcome of Laiscl(-)-infection is highly dependent on the host's genetic background. Therefore, compared to L. major, the role of sphingolipid degradation in virulence is substantially different in L. amazonensis. Future studies may reveal whether sphingolipid degradation is required for L. amazonensis to cause diffuse cutaneous infections in humans.

Introduction

Protozoan parasites of the genus Leishmania are vector-borne pathogens which infect macrophages, neutrophils, and dendritic cells of mammals [1], [2]. Human infection is caused by more than 20 Leishmania species categorized in 2 subgenera (Leishmania Leishmania and Leishmania Viannia) and 5 complexes (L. donovani, L. mexicana, L. tropica, L. hertigi, and L. braziliensis) [3]. Depending on parasite species and host immune status, Leishmania infection can cause a wide range of symptoms including localized cutaneous lesions, diffuse cutaneous lesions, destruction of mucocutaneous membranes, and visceral diseases of the hematopoietic organs. Current drugs are plagued with low efficacy, high cost, and significant toxicity [4], [5]. A better understanding of Leishmania-host interaction may facilitate the development of new cost-effective treatments.

During their life cycle, Leishmania parasites alternate between flagellated promastigotes living in the midgut of sandflies and non-flagellated amastigotes residing in mammalian phagocytes [6]. Our recent studies of an iscl mutant have demonstrated that sphingolipid (SL) degradation plays pivotal roles in both promastigote and amastigote stages of Leishmania (L.) major (commonly referred to as Leishmania major or L. major), a member of the L. tropica complex and one of the etiological agents for localized cutaneous leishmaniasis (LCL) in the Old World [7]. Briefly, L. major parasites possess a single ISCL (Inositol phosphophingolipid phospholipase -ike) protein which is responsible for the degradation of both inositol phosphorylceramide (IPC, a SL synthesized by Leishmania) and sphingomyelin (a SL synthesized by the mammalian host) [8]. ISCL-null promastigotes (iscl) survive poorly in culture during the stationary phase when cells are not replicative, and this defect is exacerbated by acidic pH [8], [9]. Importantly, iscl mutants fail to cause pathology in either immunocompetent or immunodeficient mice [9]. Virulence of iscl can be fully restored when a functional neutral sphingomyelinase (SMase) is introduced into these mutants [8], [10]. Consistent with its role as a virulence determinant, ISCL is preferentially expressed in the infective stages of L. major, i.e. stationary phase promastigotes and amastigotes [10]. The mechanism by which ISCL contributes to virulence is not well understood. One possibility is that the SMase activity is required for the generation of essential nutrients such as ceramide and phosphocholine (products of sphingomyelin degradation). Alternatively, ISCL may be used to remove excess sphingomyelin in the phagolysosome, which could be toxic for L. major.

The essentiality of ISCL in L. major virulence, in combination with its modest degree of homology to human neutral SMases, suggests that this enzyme possesses the potential of being a drug target. Because more than 20 species of Leishmania species can infect human, it is important to investigate whether the role of ISCL is conserved in these parasites. Here we extend the study of SL degradation to Leishmania (L.) amazonensis (commonly referred to as Leishmania amazonensis or L. amazonensis), which belongs to the L. mexicana complex and mostly found in South America [11], [12]. In addition to differences in geographic distribution and insect vector preference, L. amazonensis infection is distinct from L. major infection in clinical manifestation and host immune response. While L. major mainly causes LCL, L. amazonensis is associated with a range of symptoms in humans from LCL to diffuse cutaneous leishmaniasis (DCL) [13]. DCL is a rare, chronic form of leishmaniasis in which the initial cutaneous lesion is followed by the formation of secondary or satellite lesions all over the body [14]. Compared to LCL, DCL is more resistant to conventional therapy [15]. In human infections, L. amazonensis-induced DCL is characterized by high lesional macrophage-to-T cell ratio, uncontrolled parasite proliferation, and a lack of delayed hypersensitivity reaction, indicating that the cell-mediated immune mechanism is incapable of limiting the leishmanial infection (a “hyposensitivity” phenotype) [16] [12].

In murine models of cutaneous leishmaniasis, L. major infection induces polarized T cell response which dictates disease outcome, e.g. BALB/c mice are susceptible to L. major due to a Th2-dominated response leading to uncontrolled parasite growth and severe pathology, whereas C57BL6 mice are resistant due to a protective Th1-dominated response [17]. In contrast, L. amazonensis causes non-healing lesions in almost all inbred lab strains of mice in the absence of a Th2-dominance [18] [19]. A low level of mixed Th1/Th2 response has been observed in L. amazonensis-infected hosts [18], [20], [21]. In macrophages, L. amazonensis and other members of the L. mexicana complex are capable of forming large parasitophorous vacuoles (PV) with heavy parasite loads [22]. Such communal PVs (not formed by L. major) continuously undergo fusion with lysosomes and may protect L. amazonensis amastigotes by diluting the leishmaniacidal effects of nitric oxide (NO) and reactive oxygen species (ROS) from the host [23], [24].

To investigate the function of SL degradation in L. amazonensis, we generated a LaISCL-null mutant equivalent to the iscl mutant in L. major. Our results show that while the biochemistry of SL degradation is largely conserved in L. amazonensis, its role in parasite proliferation and disease development depends on the genetic background of the mammalian host. This study expands our understanding of SL metabolism and provides new information into the complex nature of Leishmania pathogenesis.

Methods

Materials

N-[6-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]hexanoyl]-sphingosine-1-phosphocholine (NBD C6-sphingomyelin) and MitoTracker Red 580 was purchased from Life Technologies (Grand Island, NY). N-[12-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]dodecanoyl]-sphingosine-1-phosphoinositol (NBD C12-IPC) was custom-synthesized by Avanti Polar lipids (Alabaster, AL). The rabbit anti-L. major ISCL peptide antibody was custom-produced by the Open Biosystems, Inc (Huntsville, AL). ELISA kits to measure IFN-γ, IL-4, and IL-10 production were purchased from eBioscience Inc (San Diego, CA). All other reagents were purchased from Sigma-Aldrich (St. Louis, MO) or Fisher Scientific (Pittsburgh, PA) unless specified otherwise.

Molecular constructs

The L. amazonensis ISCL (LaISCL) open reading frame was amplified from L. amazonensis genomic DNA by PCR using primers #141/#58 which were synthesized according to the sequence of L. mexicana ISCL (TryTrypDB LmxM.08.0200). The resulting 1.9-Kb fragment was cloned in pXG (a high copy expression vector in Leishmania) [25] to generate pXG-LaISCL (strain B163). To generate knockout constructs, the predicted 5′-and 3′-flanking regions of LaISCL were PCR amplified using primers #161/#143 and #144/#145, respectively. The resulting DNA fragments were cloned in tandem in the pUC18 vector. Genes conferring resistance to puromycin (PAC) and blasticidin (BSD) were then inserted between the 5′- and 3′- flanking regions to generate pUC-KO-LaISCL::PAC (strain B178) and pUC-KO-LaISCL::BSD (strain B177). Primers used in this study were summarized in Table S1. All constructs were confirmed by restriction enzyme digestion and DNA sequencing.

Leishmania culture and genetic manipulation

L. amazonensis (MHOM/BR/77/LTB0016) promastigotes were cultured at 26°C in M199 medium (pH 7.4) with 10% fetal bovine serum and additional supplements [26]. Metacyclics were isolated from day 3 stationary phase promastigotes using the density centrifugation method [27]. L. amazonensis axenic amastigotes were cultured at 33°C in Grace's insect cell culture medium (the pH was adjusted to 5.3) with L-glutamine and 20% fetal bovine serum [28], [29]. Purification of lesion amastigotes were performed as previously described [10].

To generate Laiscl mutants (▵LaISCL::PAC/▵LaISCL::BSD), the LaISCL alleles from wild type L. amazonensis parasites (La WT) were sequentially replaced by PAC and BSD resistance genes as previously described for the generation of L. major iscl mutants [8]. To confirm the deletion of LaISCL, genomic DNA was digested with SpeI, resolved on a 0.7% agarose gel, transferred to a nitrocellulose membrane, and hybridized with a [32P]-labeled DNA probe corresponding to a 550-bp downstream flanking region of LaISCL. These null mutants were maintained in 10 µg/ml of puromycin and 10 µg/ml of blasticidin. To complement these mutants, pXG-LaISCL was transfected into Laiscl to generate the episomal add-back, referred to as Laiscl. These add-back parasites were grown in 20 µg/ml of G418. To prevent virulence loss during in vitro culture and genetic manipulation, stationary phase promastigotes of La WT, Laiscl and Laiscl were passed through BALB/c mice at ∼2×107 cells/mouse and recovered one month later as previously described [30]. These parasites were then converted back to promastigotes and used in mouse footpad infection and macrophage infection.

Measuring cell growth and pH tolerance

Promastigotes or axenic amastigotes were inoculated in appropriate media (starting density for promastigotes: 1.0×105 cells/ml; for axenic amastigotes: 1.0×106 cells/ml). Growth rates were determined by counting culture density at designated times with a hemacytometer. Cell viability was determined by flow cytometry after staining with propidium iodide as previously described [31]. Percentages of round cells (defined as those promastigotes with the long axis shorter than twice the length of the short axis) were determined by microscopy as previously described [8]. To determine the sensitivity of L. amazonensis parasites to acidic pH, promastigotes were cultured in an acidic medium (same as the regular M199 medium except the pH was adjusted to 5.0 with hydrochloric acid) and growth rate and cell viability were determined as described [9].

Western blot and immunofluorescence microscopy

L. amazonensis promastigotes or lesion-derived amastigotes were suspended in phosphate buffered saline (PBS) at 5×107 cells/ml and boiled in SDS sample buffer for 5 minutes. Western blot was performed as previously described using the rabbit anti-L. major ISCL peptide antibody [10]. Results were quantified using a FluoroChem E imager (Protein Simple).

Immunofluorescence microscopy of L. amazonensis promastigotes or lesion-derived amastigotes was performed as previously described [10]. Briefly, formaldehyde fixed parasites were attached to poly-lysine coated cover slips and permeabilized with ice-cold ethanol. Cells were labeled with the rabbit anti- L. major ISCL antibody (1∶500 in 2% bovine serum albumin prepared in PBS) for 30 minutes, and then incubated with a goat anti-rabbit IgG-FITC (1∶1000 dilution) for 30 minutes. 350 nM of Mitotracker Red 580 (Molecular Probes/Life Technologies) was then applied for 30 minutes, followed by staining with 2.5 µg/ml of Hoechst 33342 for 10 minutes. Images were acquired using an Olympus BX51 Upright Fluorescence Microscope equipped with a digital camera.

Assays for SL degradation

Log phase L. amazonensis promastigotes (1–8×106 cells/ml) were suspended in a lysis buffer (25 mM Tris pH 7.5, 0.1% Triton X100, 1× protease inhibitor) at 2.0×108 cells/ml and incubated for 5 min on ice. Protein concentration was determined using a micro-BCA kit (Pierce). The neutral SMase assay and IPCase assay were performed as previously described [8]. Each reaction contained 40 µg of L. amazonensis protein and 0.8 nmol of NBD C6-sphingomyelin or 0.8 nmol of NBD C12-IPC. 0.1 unit of Bacillus cereus SMase or phosphatidylinositol phospholipase C was used as a positive control and boiled WT lysate was used as a negative control. Activities were quantified using a Storm 860 phosphoimager and converted to pmol/(µg×hour) after subtracting the value of negative control.

Phospholipid analysis by mass spectrometry

Total lipids from stationary phase promastigotes were extracted using the Bligh-Dyer method and analyzed by electrospray ionization mass spectrometry (the negative ion mode) as previously described [31].

Ethics statement for mouse use

The use of mice in this study was approved by the Animal Care and Use Committee at Texas Tech University (PHS Approved Animal Welfare Assurance NO. A3629-01). C57BL6 mice (female, 7–8 weeks old) and BALB/c mice (female, 7–8 weeks old) were purchased from Charles River Laboratories International (Wilmington, MA). Mice were housed and cared for in the facility operated by the Animal Care and Resources Center at Texas Tech University adhering to the institution's guidelines for animal husbandry. The facility was inspected monthly and animals were monitored daily by staff members. A complete range of clinical veterinary services was available on a 24-hour basis and includes consultation, diagnostic work-up and clinical care. Lab personnel are trained to use proper restraining and injection techniques to reduce pain and distress of animals.

Mice were under anesthesia (through the peritoneal injection of ketamine hydrochloride/xylazine) during recurring procedures including the injection of Leishmania parasites into footpads, the recovery of parasites from infected mice, and the measurement of lesion size using a caliper. Usually, no more than one procedure was performed on one mouse within a week. To prevent any potential secondary infections and to reduce any potential pain/distress, mice were monitored carefully (twice a week for appearance, size, movement, and general health condition) and euthanized when the lesions became too large (>2.5 mm for footpad infection). For the isolation of femur cells, draining lymph nodes (dLNs), and the determination of parasite numbers in the infected footpads, mice were euthanized by CO2 asphyxiation prior to operations.

Macrophage infection and mouse footpad infection

Bone marrow-derived macrophages were generated from the femur of BALB/c mice [10]. Macrophage infection was performed using day 3 stationary phase L. amazonensis promastigotes at a ratio of five parasites per macrophage (multiplicity of infection = 5∶1) as previously described [32].

Footpad infections of BALB/c mice and C57BL6 mice were performed as previously described [33], [34] using day 3stationary phase promastigotes (1×106 cells/mouse) or lesion-derived amastigotes (1×104 cells/mouse). Six mice were used in each group. Parasite numbers in the infected footpad were determined by limiting dilution assay [34].

Quantitation of cytokine production from lymphocytes

To prepare lymphocyte suspension, L. amazonensis-infected mice (two from each group) were sacrificed and dLNs were collected. To measure cytokine production, lymphocytes from dLNs were cultured in 24-well plates (4×106 cells/ml) and stimulated with soluble L. amazonensis antigen (SLA) (equivalent to 8×106 parasites/ml; generated by repeated freeze-thaw cycles) for 72 hours. Supernatants were assayed for IFN-γ, IL-4, or IL-10 using appropriate ELISA kits [35]. To offset potential variations in dLN cell numbers among wells, the ratio of SLA-stimulated over un-stimulated for each sample was recorded.

Statistical analysis

The difference between two experimental groups was determined by the Student's t test using Sigmaplot11.0 (Systat Software Inc, San Jose, CA). P values indicating statistical significance were grouped into values of <0.05 and <0.01.

L. amazonensis ISCL (LaISCL): GenBank JX131379

L. major ISCL (LmISCL): TriTryDB LmjF.08.0200

L. mexicana ISCL (LmexiISCL): TriTryDB LmxM.08.0200

Results

Identification and targeted deletion of LaISCL

Because the genome of L. amazonensis is not sequenced, we first synthesized oligonucleotides based on the sequence of L. mexicana ISCL (TriTryDB: LmxM.08.0200). These oligonucleotides (summarized in Table S1) were then used to amplify the open reading frame and 5′-/3′-flanking regions of LaISCL from L. amazonensis genomic DNA. The open reading frame of LaISCL (GenBank JX131379) encodes a protein of 645 amino acids with 86% identity to L. major ISCL and 98% identity to L. mexicana ISCL (Fig. S1). Similar to L. major ISCL, LaISCL possesses a P-loop motif (found in phosphatases and nucleotide-binding proteins and may be essential for catalytic efficiency) [36] and two putative transmembrane helices near the C-terminus (Fig. S1). To understand the function of this protein in L. amazonensis, null mutants of LaISCL (referred to as Laiscl) were generated through two sequential rounds of targeted gene deletion. Southern blot analysis confirmed the loss of LaISCL in Laiscl (Fig. 1; LaISCL+/ represents the heterologous parasite in which one of the two LaISCL alleles is deleted). To complement the mutant, a high copy number episome carrying LaISCL (pXG-LaISCL) was introduced into Laiscl and the add-back strain is referred to as Laiscl.

Figure 1

Replacement of LaISCL alleles by antibiotic resistance marker genes.

Genomic DNAs from La WT, LaISCL+/− (▵LaISCL::PAC/LaISCL), and Laiscl (▵LaISCL::PAC/▵LaISCL::BSD; clone #1 and #2) were digested and probed with a radioactive probe that recognized a 550-bp downstream region of LaISCL.

Growth of Laiscl mutants as promastigotes and axenic amastigotes

In culture, Laiscl promastigotes could proliferate from early log phase (<1×106 cells/ml) to stationary phase (2.8–3.2×107 cells/ml) with a doubling time of ∼7 hours (Fig. 2A and Fig. S2A). Their growth rate and maximal culture density are similar to what were observed with L. amazonensis wild type (La WT) and Laiscl parasites (Fig. 2A and Fig. S2A). After entering stationary phase (3 days in culture in Fig. 2A–B), Laiscl promastigotes became more round in shape. Microscopic observation revealed that 46–55% of Laiscl were round in late stationary phase whereas only 17–22% of La WT promastigotes showed similar morphology (Fig. 2B). Meanwhile, more dead cells were detected in Laiscl than in La WT during late stationary phase (Fig. S2B), suggesting that these round cells were not healthy. We also examined whether Laiscl promastigotes were sensitive to acidic pH in stationary phase by culturing them in an acidic medium (pH 5.0). As shown in Fig. S2C–D, Laiscl mutants died almost completely by day 5 in stationary phase and this defect was more severe than what was observed under neutral pH (Fig. S2B), indicating that these parasites are vulnerable to acidity. It is worth mentioning that when promastigotes were cultured in neutral condition (Fig. 2A–B and Fig. S2A–B), the medium was slightly acidified in late stationary phase (∼0.5 lower than in log phase). This indicates that the death of Laiscl may be partially attributed to their hypersensitivity to acidic pH, although other factors such as nutrient depletion and toxic waste build-up are likely involved as well. As an important control, restoration of LaISCL expression (Laiscl in Fig. 2 and Fig. S2) largely reversed the morphological and viability defects of Laiscl promastigotes. Finally, despite their abnormality in culture, Laiscl mutants still formed metacyclics in stationary phase at a similar rate as La WT (Fig. S3).

Figure 2

LaISCL is required for the maintenance of cell shape and viability in stationary phase.

Promastigotes (A–B) or axenic amastigotes (C–D) were inoculated in appropriate media and culture densities were determined every 8–12 hours in A and C (•: La WT, ○: Laiscl, ▾: Laiscl). Percentages of round cells in promastigotes (B) and percentages of dead cells in axenic amastigotes (D) were analyzed daily. In B and D, black bars: La WT, white bars: Laiscl −, grey bars: Laiscl −/+LaISCL. Error bars represent standard deviations (*: p<0.05, **: p<0.01).

Next, we examined the growth and viability of Laiscl as axenic amastigotes. Unlike L. major, promastigotes of L. amazonensis can be converted into axenic amastigotes in culture [37]. When Laiscl promastigotes were exposed to an acidic amastigote-inducing medium at a higher temperature, within 24 hours, they lost flagella and became round (data not shown). This transition indicates that LaISCL is not required for the transformation of promastigotes into axenic amastigotes. However, as shown in Fig. 2C, the growth rate and maximal culture density of Laiscl axenic amastigotes were lower comparing to La WT. In addition, starting from the third day of amastigote culture, more dead cells were detected in the mutant (29–58%) than in La WT (8–35%) parasites (Fig. 2D). This defect may be related to the hypersensitivity of Laiscl to acidic pH (since the amastigote medium was maintained at pH 5.3). Finally, the add-back parasites (Laiscl) survived and grew better than La WT (Fig. 2C–D), suggesting that increased LaISCL expression can be beneficial for L. amazonensis under certain conditions such as high temperature and low pH.

Temporal and spatial expression of LaISCL

The abundance of LaISCL protein in La WT promastigotes and amastigotes was examined by Western blot, using a peptide antibody which recognizes a 16 amino acids (amino acid 241–256) epitope in L. major ISCL (Fig. S1). The high degree of similarity between LaISCL and L. major ISCL (11 out 16 amino acids within the epitope are conserved; Fig. S1) suggests that the anti-L. major ISCL antibody may cross react with LaISCL. Indeed, a 76 KD band was detected in the promastigote lysates of La WT but not Laiscl (Fig. 3A), which was consistent with the predicted molecular weight of LaISCL. Compared to La WT, Laiscl parasites synthesized 8–10 times more LaISCL protein due to the high-copy number of pXG-LaISCL episome (Fig. 3A) [25]. The cellular level of LaISCL protein increased significantly (5–6 fold) when La WT promastigotes went from replicative log phase to non-replicative but infective stationary phase and metacyclic phase (Fig. 3A). In addition, La WT amastigotes purified from infected BALB/c mice contained much more LaISCL protein than log phase promastigotes (Fig. 3B). In summary, the stage–dependent expression of LaISCL suggests that it plays a vital role in the infective forms (i.e. late stationary phase promastigotes, metacyclics, and amastigotes) of L. amazonensis.

Figure 3

Increased expression of LaISCL in the infective stages of L. amazonensis.

(A) Promastigote lysates from La WT (Log: log phase; S1–S4: day 1–4 in stationary phase; Meta: metacyclics), Laiscl (log phase), and Laiscl (log phase) were analyzed by western blot using the anti-LmISCL antibody (top) or anti-α-tubulin antibody (bottom). (B) Immunoblot of cell lysates from log phase La WT promastigotes (Pro) and lesion-derived amastigotes (Ama). Relative intensities of ISCL and tubulin bands were determined using a FluoroChem E imager and shown below the blots. Each lane contained material from 5×105 cells.

To determine the localization of LaISCL, La WT promastigotes were labeled with the anti-L. major ISCL antibody for immunofluorescence microscopy. As shown in Fig. 4, a substantial overlap between LaISCL and Mitotracker Red 580 (a mitochondrial marker) was detected in La WT and Laiscl parasites, while LaISCL was invisible in Laiscl as expected. In La WT amastigotes (isolated from infected mice), the distribution of LaISCL also resembled the pattern of mitochondrion (Fig. S4). Together, these results indicate that LaISCL is mainly localized in the mitochondria or mitochondria-associated ER membranes, which is similar to what we previously described for ISCL in L. major [8], [10].

Figure 4

Localization of LaISCL protein in promastigotes.

Day 1 stationary phase promastigotes of La WT (left column), Laiscl − (middle column), and Laiscl −/+LaISCL (right column) were analyzed by immunofluorescence microscopy. (A): differential interference contrast images; (B): DNA staining using Hoechst 33242; (C): immuno-staining with rabbit anti-LmISCL antibody, followed by goat-anti-rabbit IgG-FITC; (D): labeling with Mitotracker Red 580; (E): merge of C and D.

LaISCL is responsible for the SMase and IPCase activity in L. amazonensis

To determine whether LaISCL is required for the neutral SMase activity, whole cell extracts from La WT, Laiscl and Laiscl promastigotes were incubated with a NBD-labeled C6 sphingomyelin and lipid products were analyzed afterwards by TLC. As illustrated in Fig. 5A–B, fluorescent ceramide was generated by La WT parasites but not by Laiscl parasites. In addition, a higher level of SMase activity was detected in Laiscl (4–7 times more than La WT) due to the overexpression of LaISCL in these add-back cells (Figs. 3 and 5A–B). Similar results were obtained when a NBD-labeled C12 IPC was used as substrate: while cell lysates from La WT and Laiscl exhibited IPCase activity, Laiscl parasites failed to hydrolyze IPC into ceramide (Fig. 5C–D). Therefore, LaISCL is responsible for both the SMase and IPCase activity in L. amazonensis. In La WT and Laiscl cell lysates, the apparent SMase activity was 6–10 fold higher than IPCase (Fig. 5B and D), suggesting that sphingomyelin is the preferred substrate for LaISCL.

Figure 5

LaISCL is required for the hydrolysis of sphingomyelin and IPC in L. amazonensis.

Promastigote lysates were incubated with TX100-based micelles containing either NBD-SM (A–B) or NBD-IPC (C–D) as described in Methods. Lipids were then extracted and separated on TLC plates (A and C). The activity of SMase (B) or IPCase (D) was calculated based on the amount of ceramide produced and the amount of protein in each sample. 0.1 unit of B. cereus SMase (A) and B. cereus PI-PLC (C) were used as positive controls. Boiled La WT cell lysate was used as negative controls (-).

We also examined the phospholipid composition of Laiscl promastigotes by mass spectrometry in the negative ion mode. Compared to La WT, Laiscl parasites contained a higher level of IPC as the 778.60 peak representing a deprotonated IPC [31] was the strongest peak in Laiscl but not La WT (Fig. S5A–B). Notably, the level of IPC was not fully reversed in the Laiscl parasites (Fig. S5C), which could be due to the separate localization of LaISCL (mitochondrion) and IPC (plasma membrane). Besides IPC, other phospholipid species including phosphatidylethanolamine and phosphatidylinositol did not show much difference between La WT and Laiscl (Fig. S5). Therefore, deletion of LaISCL leads to accumulation of IPC but does not affect the overall lipid composition in L. amazonensis.

LaISCL is required for L. amazonensis infection in C57BL6 mice but not in BALB/c mice

To determine the role of LaISCL in virulence, L. amazonensis promastigotes were injected into the footpads of BALB/c or C57BL6 mice and the development of pathology was monitored (Fig. 6A–D). In BALB/c mice, Laiscl mutants showed a slight delay (∼2 weeks) in lesion formation but their overall disease-inducing ability was comparable to that of La WT and Laiscl promastigotes (Fig. 6A). Numbers of Laiscl parasites in the infected footpads were also similar to those of La WT and Laiscl parasites at 6, 8, and 10 weeks post infection (Fig. 6B), suggesting that LaISCL is not required for L. amazonensis infection in BALB/c mice. However, when the same experiment was performed in C57BL6 mice, Laiscl mutants did not cause any detectable disease for the first 7 weeks (Fig. 6C). After the delay, mice infected by Laiscl only developed very small lesions (<0.8 mm) (Fig. 6C). This is in steep contrast to the C57BL6 mice infected by La WT or Laiscl promastigotes, which developed noticeable pathology after 2 weeks and those lesions grew to 1.5–2.3 mm in 10 weeks (Fig. 6C). Consistent with the reduced pathology, Laiscl-infected C57BL6 mice contained significantly less parasites than those infected with La WT or Laiscl (Fig. 6D). Together, these results suggest that LaISCL plays an important role in the proliferation of L. amazonensis and the development of pathology in C57BL6 mice.

Figure 6

Laiscl − mutants are virulent in BALB/c mice but attenuated in C57BL6 mice.

BALB/c mice or C57BL6 mice were infected in the footpads with stationary phase promastigotes (1×106 parasites/mouse) (A–D) or lesion-derived amastigotes (1×104 parasites/mouse) (E–H). Footpad lesions were recorded weekly in A, C, E and G (•: WT, ○: Laiscl −, ▾: Laiscl −/+LaISCL). Parasite numbers in the infected footpads were determined at the indicated times by limiting dilution assay and summarized in B, D, F and H (black bars: La WT, white bars: Laiscl −, grey bars: Laiscl −/+LaISCL). Error bars represent standard deviations (*: p<0.05, **: p<0.01).

We also examined whether LaISCL is required for the virulence of L. amazonensis amastigotes. To do so, amastigotes were purified from the footpads of infected BALB/c mice and then immediately injected into naïve BALB/c or C57BL6 mice. As shown in Fig. 6E–F, amastigotes of Laiscl could proliferate and induce pathology in BALB/c mice at a similar rate as the amastigotes of La WT and Laiscl. In contrast, Laiscl amastigotes were severely attenuated in C57BL6 mice (Fig. 6G–H). Overall, these amastigote infection data are highly similar to the promastigote infection results.

Laiscl parasites survived poorly in murine macrophages

To better understand the role of LaISCL in parasites-host interaction, we conducted murine macrophage infection experiments in vitro. Macrophages were induced from the bone marrow cells of BALB/c mice and infected with L. amazonensis promastigotes (Fig. S6A–B). Fractions of infected macrophages and the number of parasites in 100 macrophages were monitored at 2–72 hours post infection (Fig. S6A–B). Compared to La WT and Laiscl − /+LaISCL parasites, Laiscl − mutants survived much poorly in macrophages, especially during the first 24 hours of infection (Fig. S6A–B). Similar results were obtained using bone marrow macrophages from C57BL6 mice (Fig. S6C–D). Therefore, although Laiscl − mutants are fully virulent against BALB/c mice, they are compromised in macrophage infection.

Cytokine response in Laiscl infected mice

The reduced virulence of Laiscl − parasites in C57BL6 mice but not in BALB/c mice prompted us to examine whether these mutants elicited a different cytokine response from La WT parasites. At 10 weeks post infection, the dLN cells of infected mice were stimulated with SLA and the production of IFN-γ, IL-4 and IL-10 were measured as previously described [9]. In C57BL6 mice, L. amazonensis infection induced high levels of IFN-γ (Fig. 7A). Comparing to La WT and Laiscl − /+LaISCL, Laiscl − parasites triggered more of IL-4 and IL-10 production (the left side of Fig. 7B–C), although only the difference in IL-4 level is statistically significant. In BALB/c mice, L. amazonensis infection led to over production of IL-4 and IL-10 but very low levels of IFN-γ, while no statistically significant difference was detected between Laiscl − and La WT or Laiscl −/+LaISCL parasites (the right side of Fig. 7A–C). As indicated by the ratios of IL-4/IFN-γ and IL-10/IFN-γ, L. amazonensis parasites induced a Th1-biased response in C57BL6 mice and a Th2-biased response in BALB/c mice (Fig. 7D–E). It is of note that these biases are not as extreme as those with L. major infection, which is strictly Th1-dominated in C57BL6 mice and Th2-dominated in BALB/c mice [9], [17]. In summary, loss of SL degradation in L. amazonensis did not significantly alter the cytokine production in BALB/c mice, but it did result in a slight increase in IL-4 and IL-10 expression in C57BL6 mice.

Figure 7

Cytokine production in Laiscl −-infected mice.

C57BL6 mice or BALB/c mice were infected in the footpads and sacrificed after 10 weeks. Lymphocytes (dLNs) were isolated and plated on 24-well dishes. After SLA stimulation for 3 days, culture supernatants were collected to measure the level of IFN-γ (A), IL-4 (B), and IL-10 (C). Ratios of SLA-stimulated/un-stimulated were calculated for each cytokine (A–C). Ratios of IL-4/IFN-γ and IL-10/IFN-γ (both from SLA-treated samples) were also calculated (D–E). Error bars represent standard deviations from 3 replicates (*: p<0.05).

Discussion

SL degradation plays multiple roles in L. major: while the IPCase activity is important for promastigote survival and acid tolerance, the SMase activity is required for amastigote proliferation in mice and the manifestation of disease [9]. In this study, we investigated whether the function of SL degradation is conserved in another Leishmania species, L. amazonensis. Unlike L. major, L. amazonensis parasites cause non-healing lesions in almost all inbred strains of mice in the absence of a Th2 dominance [18] [19].

Biochemically, LaISCL resembles L. major ISCL in that it is responsible for the hydrolysis of both IPC and sphingomyelin (Fig. 5). LaISCL protein is highly expressed in the infective stages and is strongly associated with the mitochondria (Figs. 3, 4). For promastigotes, losing LaISCL leads to hypersensitivity to acidic pH and poor viability in late stationary phase (Figs. 2 and S2). As axenic amastigotes, Laiscl mutants exhibit a slower growth rate and reduced survival than La WT (Fig. 2). Importantly, both promastigotes and lesion-derived amastigotes of Laiscl were fully infective towards BALB/c mice yet showed severely attenuated virulence towards C57BL6 mice (Fig. 6). This phenotype makes the Laiscl mutants somewhat similar to the L. major wild type parasites but drastically different from the L. major iscl mutants which are completely avirulent in both BALB/c and C57BL6 mice [9].

While the degradation of sphingomyelin or IPC is dispensable for L. amazonensis to establish infection in BALB/c mice, Laiscl mutants do not survive well in BALB/c macrophages in vitro and this defect is clearly due to the loss of LaISCL (Fig. S6). Potentially, in BALB/c mice, other host factors such as cytokines, neutrophils, dendritic cells, or T cells may interact with macrophages in a way that benefits the survival and proliferation of Laiscl. It is not clear why the outcome of Laiscl infection is dependent on mouse genetic background. Among common inbred mouse strains, there is widespread variation in the number and function of natural killer T cells (NKT-cells) [38]. These cells can recognize and be activated by the CD1d-presented glycosphingophospholipid antigen from Leishmania to promote parasite killing [39], [40]. The SL degradation defect in Laiscl could alter the production and/or presentation of glycosphingophospholipid antigen, and thus affects its interaction with NKT cells. Compared to BALB/c mice, C57BL6 mice contain a higher number of NKT cells [38] which may lead to a more effective control of Laiscl parasites.

As shown in Fig. 7A–C, BALB/c mice infected by Laiscl produced similar levels of IFN-γ, IL-4, and IL-10 as those infected by La WT or Laiscl/+ LaISCL parasites. In C57BL6 mice, however, Laiscl triggered significantly more IL-4 production than La WT or Laiscl/+ LaISCL parasites (Fig. 7B). While IL-4 is primarily a susceptibility factor for L. amazonensis infection in BALB/c mice [18], [41], its role in C57BL6 or C3H mice is less clear [18], [42]. A previous study indicates that the level of IL-4 correlates with lesion development in L. amazonensis-infected C57BL10 mice [18]. Therefore, the increased IL-4 production from Laiscl-infection may be the consequence, rather than the cause of reduced virulence (Figs. 6, 7).

Clearly, L. amazonensis infection of C57BL6 mice led to significant IFN-γ production (Fig. 7A). Effect of IFN-γ on L. amazonensis infection can be complex. On one hand, this cytokine can activate murine macrophages and inhibit parasite growth when it is applied in combination with LPS [21], [43]. On the other hand, without LPS, IFN-γ alone can improve parasite invasion and replication in macrophages [43], [44]. Mechanism of such an infection-promoting effect is not well defined, although the induction of autophagy via IFN-γ treatment could be involved [45]. Loss of LaISCL did not affect the production of IFN-γ in C57BL6 mice but did limit the replication of Laiscl (Figs. 6, 7). This result suggests that SL degradation may be involved in balancing the dual effects of IFN-γ, which is crucial for L. amazonensis proliferation in C57BL6 mice.

In summary, we demonstrate that the role of SL degradation in Leishmania virulence can vary significantly among different parasite species and is highly dependent on the mammalian host. Similar to murine infections, the genetic background of human host also has a major influence in the outcome of Leishmania infection [46]–[48]. To further evaluate the potential of SL degradation as a drug target, future studies may expand the investigation to other Leishmania species and other hosts. Another potential point of interest from our study is that the Laiscl mutants resemble wild type L. major parasites in mouse infections (fully virulent in BALB/c mice but severely attenuated in C57BL6 mice). Since L. amazonensis–infection can cause a wider range of symptoms in humans than L. major–infection, it would be interesting to examine whether LaISCL is required for the dissemination of L. amazonensis and the manifestation of DCL, a rare but difficult disease to treat.

Sequence alignment of ISCL (LaISCL) and ISCL (LmISCL). Alignment was done using the NCBI BLASTp program. Non-identical amino acids are shown in red. The underlined sequence indicates the P-loop motif. Amino acids 241–256 of LmISCL (boxed area) represent the epitope recognized by the anti-LmISCL peptide antibody. The braces represent predicted transmembrane helices. Asterisks mark amino acids that are essential for catalysis based on a recent study of LmISCL [10].

(PDF)

Click here for additional data file.

Ability of Promastigotes were cultured to stationary phase in either regular media (pH 7.4, A–B) or acidic media (pH 5.0, C–D). Cell density (A and C) and viability (B and D) were measured daily after entry into stationary phase. Black bars: La WT; white bars: Laiscl; grey bars: Laiscl. Experiments were repeated three times and error bars represent standard deviations (*: p<0.05, **: p<0.01).

(PDF)

Click here for additional data file.

Metacyclogenesis is normal in mutants. LaWT (•), Laiscl (○) and Laiscl (▾) promastigotes (in vitro passage numbers <5) were cultured to stationary phase. Metacyclics were purified using the density centrifugation method [27] and percentages of metacyclics were determined daily.

(PDF)

Click here for additional data file.

Localization of LaISCL in amastigotes. La WT amastigotes were isolated from infected BALB/c mice and analyzed by immunofluorescence microscopy. (A) phase contrast image; (B) DNA staining using Hoechst 33242; (C) labeling with Mitotracker Red 580; (D) immuno-staining with rabbit anti-LmISCL antibody, followed by FITC conjugated goat-anti-rabbit IgG; (E) merge of C and D.

(PDF)

Click here for additional data file.

Total lipids were extracted from stationary phase promastigotes of La WT (A), Laiscl − (B), or Laiscl (C) and analyzed by electrospray ionization mass spectrometry as previously described [31]. Representative spectra (negative ion mode) are shown with major phospholipids labeled (IPC: inositol phosphorylceramide, PE: phosphatidylethanolamine, PI: phosphatidylinositol).

(PDF)

Click here for additional data file.

Bone marrow MÖs from BALB/c mice (A–B) or C57BL6 mice (C–D) were infected by stationary phase promastigotes of La WT(•), Laiscl − (○), or Laiscl −/+LaISCL (▾). As a control, La WT parasites were also used to infect MΦs that were activated with 50 ng/ml of LPS and 50 ng/ml of IFN-γ (▵). Fraction of infected MΦs (A, C) and number of parasites per 100 MΦs (B, D) were recorded. Error bars represent standard deviations.

(PDF)

Click here for additional data file.

List of oligonucleotides. Sequences shown in lowercase represent restriction enzyme sites.

(PDF)

Click here for additional data file.