Suicidal Leishmania

Leishmania are obligate intracellular parasites known to have developed successful ways of efficient immunity evasion. Because of this, leishmaniasis, a disease caused by these flagellated protists, is ranked as one of the most serious tropical infections worldwide. Neither prophylactic medication, nor vaccination has been developed thus far, even though the infection has usually led to strong and long-lasting immunity. In this paper, we describe a “suicidal” system established in Leishmania mexicana, a human pathogen causing cutaneous leishmaniasis. This system is based on the expression and (de)stabilization of a basic phospholipase A2 toxin from the Bothrops pauloensis snake venom, which leads to the inducible cell death of the parasites in vitro. Furthermore, the suicidal strain was highly attenuated during macrophage infection, regardless of the toxin stabilization. Such a deliberately weakened parasite could be used to vaccinate the host, as its viability is regulated by the toxin stabilization, causing a profoundly reduced pathogenesis.


Introduction
Leishmaniasis is a neglected tropical disease, caused by Leishmania spp., flagellated protozoan parasites belonging to the family Trypanosomatidae [1,2]. The transmission occurs through the bite of a female sand fly (Diptera: Phlebotominae). In the vector, flagellated promastigotes differentiate into the virulent metacyclic forms, which upon injection into the mammalian host develop into the aflagellated, non-motile amastigotes that reside in the phagolysosomes of host macrophages [3,4]. The extent of clinical manifestations of the disease depends both on the parasite species and the immune response developed by the host, but typically ranges from the self-healing cutaneous lesions to fatal visceral infections [1]. Nowadays, leishmaniasis represents the second most serious tropical parasitic disease in terms of mortality and, according to the World Health Organization, over 350 million people live in at-risk endemic areas [5]. Current treatment of the disease is expensive and often associated with severe adverse side effects [6]. Moreover, in many instances a long-lasting parasite resistance develops, underlying the need for an efficient vaccination approach that would allow proper control and prevention of this debilitating disease [7].
To date, several anti-Leishmania vaccination strategies have been proposed. Leishmanization, which is based on the inoculation of low-titer or attenuated live virulent parasites, represents the most efficient approach [8,9]. However, its applicability is limited due to apparent safety concerns [10]. Vaccination The growth of the BnSP-7-ecDHFR-HA-expressing cells was significantly inhibited upon the addition of either TMP or TMP-lac, as compared to the wild type L. mexicana or the untreated promastigotes ( Figure 1A,B; Table S1). In the absence of induction, the BnSP-7-ecDHFR-HAexpressing cells grow better than the parental ones. When performing an extended growth kinetics analysis, we noticed that the starting concentration of parasites (5 × 10 5 cells/mL) was decreasing even after 24 hours of TMP treatment ( Figure 1C), with no viable cells observed on day 8, whereas the untreated group had reached the stationary log phase on day 6 ( Figure 1C). We conclude that BnSP-7 toxin is killing L. mexicana promastigotes in vitro. To confirm this, we checked the early as well as the late apoptotic events by fluorescence microscopy and flow cytometry using the FITC-Annexin V/PI method [24] (Figure 2). As a positive control, the wild type L. mexicana treated with Hygromycin B was used. Cells undergoing apoptosis became abundant in both TMP and TMP-lac-treated conditions. We also noticed that TMP-lac showed slightly lower efficacy, compared to TMP ( Figure 2). However, we continued working with TMP-lac due to its solubility in water and, thus, direct applicability for downstream in vivo experiments. To confirm this, we checked the early as well as the late apoptotic events by fluorescence microscopy and flow cytometry using the FITC-Annexin V/PI method [24] (Figure 2). As a positive control, the wild type L. mexicana treated with Hygromycin B was used. Cells undergoing apoptosis became abundant in both TMP and TMP-lac-treated conditions. We also noticed that TMP-lac showed slightly lower efficacy, compared to TMP ( Figure 2). However, we continued working with TMP-lac due to its solubility in water and, thus, direct applicability for downstream in vivo experiments. mexicana cells treated with TMP or TMP-lac for 5 days. Hygromycin B treated wild type L. mexicana cells were used as an apoptotic/necrotic control; (B) apoptosis and necrosis were quantified after FITC-Annexin V and PI labeling; bar graphs represent mean ± SD in three independent experiments; asterisks indicate p values; * ≤0.05, ** ≤0.01, *** ≤0.001; (C) fluorescence microscopy of the Annexin V-FITC, PI, and DAPI (left panels), Annexin V-FITC and DAPI (middle panels) stained, and light microscopy controls (right panels) of the L. mexicana BnSP-7-ecDHFR-HA cells prior and after induction by the stabilization ligand for 48 hours. Scale bars are 10 μm.

BnSP-7-ecDHFR-HA System Is Functional in Axenically Differentiated L. mexicana Amastigotes
To check whether stabilization of the toxin protein was not restricted to the promastigote stage, the BnSP-7-ecDHFR-HA-expressing L. mexicana cells were differentiated into amastigotes in vitro and, subsequently, treated with TMP/TMP-lac for 72 hours. Differentiation from the promastigotes into amastigotes in the presence of TMP/TMP-lac was not possible, since the cells died within several days of treatment (Figures 1 and 2). Functionality of the stabilization system was checked by Western blotting, detecting properly processed BnSP-7-HA or BnSp-7-HA3 in both promastigotes ( Figure 3A) and cells that were in vitro differentiated into the amastigotes upon the treatment with either TMP or TMP-lac ( Figure 3B). The triple HA-tag was introduced for more efficient pull-down. In all cases, protein bands of the expected size, namely ~32 kD for BnSP-7-HA and ~34 kD for BnSp-7-HA3, were documented after the TMP/TMP-lac induction. Moreover, the anti-HA antibody also detected a higher (~40 kD) and a lower molecular weight band (~20 kD) in the TMP/TMP-lac-treated parasites. Mass spectrometry analysis showed that these bands represent ubiquitylated and partially degraded forms of BnSP-7-HA or BnSP-7-HA3.  To check whether stabilization of the toxin protein was not restricted to the promastigote stage, the BnSP-7-ecDHFR-HA-expressing L. mexicana cells were differentiated into amastigotes in vitro and, subsequently, treated with TMP/TMP-lac for 72 hours. Differentiation from the promastigotes into amastigotes in the presence of TMP/TMP-lac was not possible, since the cells died within several days of treatment (Figures 1 and 2). Functionality of the stabilization system was checked by Western blotting, detecting properly processed BnSP-7-HA or BnSp-7-HA 3 in both promastigotes ( Figure 3A) and cells that were in vitro differentiated into the amastigotes upon the treatment with either TMP or TMP-lac ( Figure 3B). The triple HA-tag was introduced for more efficient pull-down. In all cases, protein bands of the expected size, namely~32 kD for BnSP-7-HA and~34 kD for BnSp-7-HA 3 , were documented after the TMP/TMP-lac induction. Moreover, the anti-HA antibody also detected a higher (~40 kD) and a lower molecular weight band (~20 kD) in the TMP/TMP-lac-treated parasites. Mass spectrometry analysis showed that these bands represent ubiquitylated and partially degraded forms of BnSP-7-HA or BnSP-7-HA 3 .

The BnSP-7-ecDHFR-HA-expressing L. mexicana Are Attenuated in Macrophage Infections
Next, J774 macrophages were infected with either the wild type or the BnSP-7-ecDHFR-HAexpressing L. mexicana, followed by the treatment 24 hours post infection (p.i.) with either 20 μM TMP-lac or saline buffer, as a control. Four hours p.i., the cells were checked for the proper phagocytic activity and no parasites were observed outside of the macrophages. Importantly, the total number of counted amastigotes (a proxy of Leishmania infectivity) 72 hours p.i. was substantially lower in the case of L. mexicana BnSP-7-ecDHFR-HA as compared to the wild type parasites even without the TMP or TMP-lac treatment ( Figure 4A). This implies higher susceptibility of the mutant cells towards macrophage phagocytosis machinery and may reflect the incomplete toxin destabilization inside the macrophages, as the percentage of infected macrophages within the examined groups did not differ ( Figure S2). The two-day cultivation of the infected macrophages with 20 μM TMP-lac resulted in a ~60% reduction of the recovered L. mexicana amastigotes, when the number of the BnSP-7-ecDHFR-HA-expressing parasites was compared to their wild type kin ( Figure 4B).  Next, J774 macrophages were infected with either the wild type or the BnSP-7-ecDHFR-HAexpressing L. mexicana, followed by the treatment 24 hours post infection (p.i.) with either 20 µM TMP-lac or saline buffer, as a control. Four hours p.i., the cells were checked for the proper phagocytic activity and no parasites were observed outside of the macrophages. Importantly, the total number of counted amastigotes (a proxy of Leishmania infectivity) 72 hours p.i. was substantially lower in the case of L. mexicana BnSP-7-ecDHFR-HA as compared to the wild type parasites even without the TMP or TMP-lac treatment ( Figure 4A). This implies higher susceptibility of the mutant cells towards macrophage phagocytosis machinery and may reflect the incomplete toxin destabilization inside the macrophages, as the percentage of infected macrophages within the examined groups did not differ ( Figure S2). The two-day cultivation of the infected macrophages with 20 µM TMP-lac resulted in a~60% reduction of the recovered L. mexicana amastigotes, when the number of the BnSP-7-ecDHFR-HA-expressing parasites was compared to their wild type kin ( Figure 4B).

Discussion
Toxins and AMPs are natural substances of both prokaryotic and eukaryotic origin [25,26]. Their role is generally related to offensive/defensive responses, resulting in damage or death of the targeted organism [27,28]. Some toxins or AMPs are able to kill trypanosomatid parasites, and their trypanocidal and/or leishmanicidal properties have been extensively studied [29]. However, these studies were mostly restricted to the in vitro killing assays with recombinant or synthetic proteins, limiting them primarily to the therapeutic applications.
The aim of this study was to express different toxins/AMPs in L. mexicana using a recently established inducible protein stabilization system [22]. Such genetically-modified suicidal strain of Leishmania would be an eligible and promising candidate for vaccine development. Our search for proteins, whose killing activity was reported either for Trypanosoma or Leishmania spp., resulted in a list of six potential candidates: α-toxin, Attacin-A, BnSP-7, Cecropin-A, Cathelicidin-5, and Exotoxin-A. Of note, three of these toxins (α-toxin, Cecropin-A, and Exotoxin-A) have been previously tested in T. cruzi [21].
Insect AMPs Attacin-A, isolated from Glossina morsitans morsitans, and Cecropin-A, obtained from Hyalophora cecropia, fused to the E. coli DHFR degron, did not change the viability of Leishmania mexicana promastigotes upon protein stabilization ( Figure S1). They have been previously demonstrated to inhibit growth of T. brucei and T. cruzi [30,31], as well as L. panamensis and L. major amastigotes [32] in vitro. A Bos taurus-derived antimicrobial peptide Cathelicidin-5, whose lethal activity was previously demonstrated for L. major promastigotes [33], was similarly not effective in L. mexicana ( Figure S1). AMPs are mostly cationic agents disrupting the membrane integrity by the electrostatic interactions with their anionic phospholipid counterparts [34,35]. Different membrane composition in the Leishmania and Trypanosoma cells (or in different Leishmania spp.) may be responsible for the discrepancy observed for these parasites. This implies that there must be other mechanisms (for example, post-translational modifications, protein stability and folding, etc.), determining whether a toxin will be active against a given parasite species. To illustrate this further, the α-toxin from Staphylococcus aureus was the best-choice candidate in T. cruzi studies [21], whereas it had no effect on L. mexicana ( Figure S1). This toxin is known to form large Ca 2+ -permissive pores in the cellular membrane, resulting in a massive homeostasis disruption and cell death [36]. Because the

Discussion
Toxins and AMPs are natural substances of both prokaryotic and eukaryotic origin [25,26]. Their role is generally related to offensive/defensive responses, resulting in damage or death of the targeted organism [27,28]. Some toxins or AMPs are able to kill trypanosomatid parasites, and their trypanocidal and/or leishmanicidal properties have been extensively studied [29]. However, these studies were mostly restricted to the in vitro killing assays with recombinant or synthetic proteins, limiting them primarily to the therapeutic applications.
The aim of this study was to express different toxins/AMPs in L. mexicana using a recently established inducible protein stabilization system [22]. Such genetically-modified suicidal strain of Leishmania would be an eligible and promising candidate for vaccine development. Our search for proteins, whose killing activity was reported either for Trypanosoma or Leishmania spp., resulted in a list of six potential candidates: α-toxin, Attacin-A, BnSP-7, Cecropin-A, Cathelicidin-5, and Exotoxin-A. Of note, three of these toxins (α-toxin, Cecropin-A, and Exotoxin-A) have been previously tested in T. cruzi [21].
Insect AMPs Attacin-A, isolated from Glossina morsitans morsitans, and Cecropin-A, obtained from Hyalophora cecropia, fused to the E. coli DHFR degron, did not change the viability of Leishmania mexicana promastigotes upon protein stabilization ( Figure S1). They have been previously demonstrated to inhibit growth of T. brucei and T. cruzi [30,31], as well as L. panamensis and L. major amastigotes [32] in vitro. A Bos taurus-derived antimicrobial peptide Cathelicidin-5, whose lethal activity was previously demonstrated for L. major promastigotes [33], was similarly not effective in L. mexicana ( Figure S1). AMPs are mostly cationic agents disrupting the membrane integrity by the electrostatic interactions with their anionic phospholipid counterparts [34,35]. Different membrane composition in the Leishmania and Trypanosoma cells (or in different Leishmania spp.) may be responsible for the discrepancy observed for these parasites. This implies that there must be other mechanisms (for example, post-translational modifications, protein stability and folding, etc.), determining whether a toxin will be active against a given parasite species. To illustrate this further, the α-toxin from Staphylococcus aureus was the best-choice candidate in T. cruzi studies [21], whereas it had no effect on L. mexicana ( Figure S1). This toxin is known to form large Ca 2+ -permissive pores in the cellular membrane, resulting in a massive homeostasis disruption and cell death [36]. Because the α-toxin forms a heptameric transmembrane pore [37], we presumed that its fusion with a destabilizing domain has interfered with its proper folding. In contrast, all our attempts to generate a cell line, expressing Pseudomonas aeruginosa Pathogens 2020, 9, 79 7 of 12 Exotoxin-A, have invariably failed. We explained this by insufficient toxin destabilization, followed by the inhibition of protein synthesis and cell death. Overall, the data presented here indicate that the functionality of a suicidal system relies both on the selected toxin and the host used for its expression.
The BnSP-7, a catalytically inactive Phospholipase A2 (PLA 2 ) homolog from the Bothrops pauloensis snake venom, was the only, albeit very effective, toxin, acting against L. mexicana promastigotes and amastigotes, identified in this study. Although lacking the catalytic activity due to a single amino acid substitution (Asp49Lys), myotoxic PLA2 homologs have the ability to disrupt biological membranes in a Ca 2+ -independent manner not involving the hydrolysis of phospholipids [38][39][40]. In cytotoxicity assays, its activity has also been demonstrated against several Leishmania spp. [41][42][43]. The pilot TMP/TMP-lac treatment experiments with the BnSP-7-ecDHFR-HA-expressing parasites have confirmed the anti-leishmanial potential of this toxin (Figure 1). Cells did not multiply and their morphology dramatically changed from long promastigotes to round-shaped cells, indicating apoptosis. Moreover, the number of parasites was decreasing after 72 hours of continuous TMP treatment, reaching an undetectable level by day 8 ( Figure 1C). Flow cytometry and fluorescence microscopy evaluation [44,45] confirmed the apoptotic nature of the toxin-induced Leishmania cell death ( Figure 2).
Amastigotes could not be differentiated in the presence of BnSP-7. To overcome this, we first produced amastigotes in vitro and then treated them with either TMP or TMP-lac. The readout of this experiment was a stabilized toxin, detected by immunoblotting. Both the promastigotes and the amastigotes showed the same expression pattern with three bands upon addition of the stabilizing agent (Figure 3). Immunoprecipitation followed by mass spectrometry analysis confirmed the identity of BnSP-7-ecDHFR-HA, its degradation product, and the post-translationally modified mono-ubiquitylated form. In both promastigotes and amastigotes, the functional toxin was predominant after stabilization. Recent studies have shown that mono-ubiquitylation of smaller proteins can target them for proteasomal recognition and processing [46]. We assume that mono-ubiquitylation of the fused destabilized protein drives its proteolysis, and the domain stabilization by TMP/TMP-lac interferes with this process.
When performing the macrophage infections, we noticed a surprisingly reduced infectivity of the BnSP-7-ecDHFR-HA-expressing L. mexicana even in the absence of the TMP-facilitated stabilization ( Figure 4A). Since the phagocytic activity of macrophages did not differ between the wild type L. mexicana and the mutant, we presume that the BnSP-7-ecDHFR-HA-expressing parasites possess all antigens required for macrophage recognition, but their viability and/or division rate was lowered after internalization into the phagolysosomes. The most parsimonious explanation of such an attenuated phenotype is that the toxin was partly stabilized in the phagolysosomal environment, thus resulting in a reduced infectivity. As expected, the TMP-lac treatment further decreased the number of recovered amastigotes in the case of the mutant cell line compared to the wild type L. mexicana ( Figure 4B). The BnSP-7 toxin expression causes elevated susceptibility upon macrophage engulfment and this can be enhanced via an inducible (de)stabilization system. We realize that such a highly attenuated mutant strain of L. mexicana needs to be reexamined in vivo in animal models of leishmaniasis, as the immune response is too complex to be reliably predicted from the in vitro studies only.
In conclusion, this is the first report of the basic PLA 2 homologue, derived from the Bothrops pauloensis snake venom, with lethal activity against L. mexicana. We believe that these results may ultimately lead to the development of novel vaccination strategies based on the controllably suicidal Leishmania. (Sigma-Aldrich), 2 µg/mL Hemin (Jena Bioscience GmbH, Jena, Germany), 25 mM HEPES, 50 units/mL of penicillin/streptomycin (Life Technologies, Carlsbad, CA, USA), and 10% Fetal Bovine Serum (FBS, BioSera Europe, Nuaillé, France). The amastigotes of L. mexicana were differentiated in vitro and cultivated at 32 • C in the complete M199 medium (pH 5.5), supplemented with 20% FBS [47,48]. The following ligands were tested: Trimethoprim (Sigma-Aldrich, 10 mM stock in DMSO) and Trimethoprim-lactate (TMP-lac, Sigma-Aldrich, 10 mM stock in M199, or 1 mM stock in a saline buffer).

Growth Kinetics and Differentiation
Growth kinetics analysis of the wild type and transgenic cell lines was performed for 5 days from the starting density of 5 × 10 5 cells/mL in three biological replicates, as described previously [51]. The counting was also performed for all transfectants after the addition of 20 µM TMP or TMP-lac on day 0, using wild type L. mexicana as a control. The BnSP-7-ecDHFR-HA expressing L. mexicana cells were additionally counted in the same manner every 24 hours for 8 days with daily replenishment of TMP.
Differentiation of the BnSP-7-ecDHFR-HA expressing L. mexicana mutants into amastigotes was performed by varying pH and temperature as described previously [48]. Once the parasites reached the amastigote stage (day 17), they were treated with 20 µM TMP/TMP-lac and analyzed by Western blotting after 48 hours.

Apoptosis Assays
The analysis of L. mexicana apoptosis was carried out using the eBioscience Annexin V Apoptosis Detection Kit FITC (Thermo Fisher Scientific) according to the manufacturer's instructions. Briefly, 5 × 10 6 /mL of the BnSP-7-ecDHFR-HA expressing promastigotes, incubated with or without 20 µM TMP, TMP-lac, or DMSO, were harvested, washed, and treated with 5 µL of FITC-Annexin V for 15 min at room temperature in the dark, followed by treatment with 5 µL propidium iodide (PI, 20 µg/mL). The cells were then analyzed on a FACSCanto II flow cytometer (Becton Dickinson, Franklin Lakes, NJ, USA) or visualized directly by fluorescent microscopy using the Olympus BX51 microscope equipped with a DP70 charge-coupled device camera (Olympus, Tokyo, Japan). As a positive control, L. mexicana wild type cells, treated by Hygromycin B, were used. Data were collected from three independent biological replicates and tested for the significance using two-tailed, paired Student t-test. p values ≤0.05 were considered statistically significant.
The immunoprecipitation of triple HA-tagged BnSP-7-ecDHFR was carried out with the L. mexicana protein lysates using Pierce Anti-HA magnetic beads (Thermo Fisher Scientific), following the manufacturer's instructions. Elution of the sample was performed by heat denaturation in Laemmli sample buffer (Bio-Rad, Hercules, CA, USA) for 10 minutes. Samples were separated in 10% SDS-PAGE gels, stained with Coomassie Brilliant Blue G-250 (Thermo Fisher Scientific), cut out and sent for liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis (CEITEC, Brno, Czech Republic).

Macrophage Infection
The L. mexicana wild type promastigotes were cultured in the M199 medium (Sigma-Aldrich) containing 10% FBS supplemented with 1% Basal Medium Eagle vitamins (Sigma-Aldrich), 2% sterile urine, and 250 µg/mL Amikacin (Bristol-Myers Squibb, New York, NY, USA), and 200 µg/mL of Hygromycin B (Carl Roth GmbH) for BnSP-7-ecDHFR-HA mutant cell lines. Infection of the J774 macrophages was performed as described previously [49,52]. In addition, infection was evaluated by counting Giemsa-stained slides as described previously [51]. All experiments were performed with two independent biological replicates and samples were analyzed in triplicate.