Differences in Charge Distribution in Leishmania tarentolae Leishmanolysin Result in a Reduced Enzymatic Activity

Leishmania tarentolae is a non-pathogenic trypanosomatid isolated from lizards widely used for heterologous protein expression and extensively studied to understand the pathogenic mechanisms of leishmaniasis. The repertoire of leishmanolysin genes was reported to be expanded in L. tarentolae genome, but no proteolytic activity was detected. Here, we analyzed L. tarentolae leishmanolysin proteins from the genome to the structural levels and evaluated the enzymatic activity of the wild-type and overexpressing mutants of leishmanolysin. A total of 61 leishmanolysin sequences were retrieved from the L. tarentolae genome. Five of them were selected for phylogenetic analysis, and for three of them, we built 3D models based on the crystallographic structure of L. major ortholog. Molecular dynamics simulations of these models disclosed a less negative electrostatic potential compared to the template. Subsequently, L. major LmjF.10.0460 and L. tarentolae LtaP10.0650 leishmanolysins were cloned in a pLEXSY expression system into L. tarentolae. Proteins from the wild-type and the overexpressing parasites were submitted to enzymatic analysis. Our results revealed that L. tarentolae leishmanolysins harbor a weak enzymatic activity about three times less abundant than L. major leishmanolysin. Our findings strongly suggest that the less negative electrostatic potential of L. tarentolae leishmanolysin can be the reason for the reduced proteolytic activity detected in this parasite.


Introduction
Neglected diseases are a major public health problem in developing countries, including Brazil [1]. In the case of leishmaniasis, there are currently 12 million infected several conserved residues of around 60% identity compared to leishmanolysins from T. brucei, L. major, and Crithidia sp., as well as the zinc-coordination motif HEXXH [21]. The insertion of 62 amino acids in the defining motif HExxHxxGxxH differs, in L. major, from that commonly found in other metallopeptidases belonging to the metzincin class. Recently, to better understand the molecular basis of the expanded variety of chromosome 10 leishmanolysins of L. braziliensis, Sutter and co-workers (2017) [22] identified two levels of structural heterogeneity that affect the electrostatic properties of leishmanolysins and the geometry of their active sites. The authors assumed that the structural plasticity of L. braziliensis leishmanolysins might constitute a crucial evolutionary adaptation related to the ability of promastigotes to interact with a broad range of substrates [22]. In this work, to understand the lack of detectable proteolytic activity of the highly expanded leishmanolysin family in L. tarentolae, we analyzed the annotated leishmanolysin sequences and then created comparative 3D-models of L. tarentolae leishmanolysins to perform molecular dynamics simulations. Then, one L. tarentolae leishmanolysin and the one from L. major were overexpressed in an L. tarentolae expression system to improve our enzymatic knowledge of these molecules.

Unveiling the L. tarenrolae Leishmanolysin Annotated Sequences
Since the first isolation of the lizard Sauroleishmania parasite in 1921 [8], a great improvement in genome sequencing has taken place [10]. In this sense, Raymond and co-workers [15] used next-generation DNA sequencing technologies (NGS) to obtain a highquality genome of L. tarentolae, revealing that more than 90% of the L. tarentolae genes are shared with all so far sequenced Leishmania species. However, some well-known limitations of this system were observed in the genome assembly, which results in a collapse of some repeat-rich regions and a tendency of repetitive expanded genes to be fragmented in the assembly, such as the leishmanolysin sequences [15]. Therefore, here we analyzed all the leishmanolysin genes annotated in the L. tarentolae Parrot-TarII genome available in the TriTrypDB.
We have retrieved a total of 61 sequences annotated as "Leishmanolysin" in the L. tarentolae reference genome (Table S1), 12 more than previously reported due to the annotated contig sequences [15]. Interpro Scan predicted that these annotated proteins contain the metallopeptidase M8 domain in their sequences. However, 34 of these sequences carry less than two hundred amino acids, a sign of fragmented assembly. Furthermore, only eight of the remaining twenty-seven protein sequences have all amino acids recognized without any unknown residues (X), which appear when the annotated gene lacks an ascribed nucleotide. Therefore, considering the missing information in some sequences and focusing on the most conserved leishmanolysin sequences in our subsequent experimental methodologies, we first concentrated our analysis on these eight L. tarentolae leishmanolysin sequences, as reported in Table 1. Among them, one belonged to chromosome 31 (LtaP31.2430), one from chromosome 33 (LtaP33.0300), three from chromosome 10 (LtaP10.0450, LtaP10.0480 and LtaP10.0650), and other three were contigs from unassigned chromosomes (LtaPcontig00585-1, LtaPcontig00616-1, and LtaPcontig03442-1) ( Table 1). Chromosomes 31 and 33 share fewer identities with L. major LmjF.10.0460, 35.6, and 58.4%, respectively. Such diversity is reasonably expected since the division of leishmanolysin sequences in different chromosomes probably happens in early steps during evolution [23]. The sequence identity for the other six sequences ranged from 63.2 to 71.2%, suggesting that the contigs could also be from L. tarentolae chromosome 10. Leishmanolysins on chromosome 10 constitute a set of gene arrays in all Leishmania species except L. donovani from the Leishmania subgenus, where it is absent. Interestingly, there are 42 leishmanolysins from chromosome 10 in L. tarentolae (Table S1), suggesting that, as it occurs in species from the Viannia subgenus, it is highly possible that this leishmanolysin expansion took place after the separation of three subgenera and could be related to an adaptation mechanism of these species to interact with a broad range of reservoirs and vectors [22,23]. Considering the protein features from L. tarentolae leishmanolysin sequences predicted by bioinformatic tools, the molecular mass ranged from 24.0 to 66.1 kDa, and the disparity can be ascribed to the shorter or even lacking N-terminal domain in most of the sequences. Two leishmanolysins in chromosome 10 (LtaP10.0450 and LtaP10.0480) presented one predicted transmembrane helix, while LtaP31.2430 and LtaP33.0300 displayed two helices. Moreover, four sequences (LtaP10.0450, LtaP10.0650, LtaP31.2430, and LtaP33.0300) had GPI anchor motifs predicted in their C-terminal domain. The presence of leishmanolysin molecules in the plasma membrane is critical to parasite adhesion and host cell interaction, as previously reported in several Leishmania species [18,19]. Only five sequences reported the presence of zinc-biding motifs (LtaP10.0480, LtaP10.0650, LtaP31.2430, LtaPcontig00585-1, and LtaPcontig03442-1), but the other sequences kept the conserved histidine and methionine residues, essential to the zinc-binding aligned with LmjF.10.0460 at positions 334 and 345, respectively [21]. Finally, among the eight L. tarentolae leishmanolysins under analysis, the five most conserved sequences reported the presence of two associated domains by Interpro Scan, the Peptidase_M8 (PF01457) and LSHMANOLYSIN (PR00782).
To better understand the phylogenetic relationship of the leishmanolysins from L. tarentolae with the pathogenic Leishmania spp., we performed a phylogenetic reconstruction based on the Maximum Likelihood, Jones-Taylor-Thornton (JTT) model ( Figure 1). Our results revealed that the five most conserved L. tarentolae sequences are closely related, supporting our hypothesis that the contig leishmanolysin sequences may have originated from chromosome 10. Moreover, L. tarentolae leishmanolysin protein sequences are phylogenetically more related to L. major than L. braziliensis and L. martiniquensis sequences. This was expected since molecular phylogenetic analysis suggests that Sauroleishmania is a sister group of the Leishmania subgenus [24]. In order to deepen the analyses, we selected the protein sequence most closely related to the L. major LmjF. 10.0460, which has its three-dimensional structure revealed by crystallization and X-ray diffraction, to perform further studies. phylogenetically more related to L. major than L. braziliensis and L. martiniquensis sequences. This was expected since molecular phylogenetic analysis suggests that Sauroleishmania is a sister group of the Leishmania subgenus [24]. In order to deepen the analyses, we selected the protein sequence most closely related to the L. major LmjF. 10.0460, which has its three-dimensional structure revealed by crystallization and Xray diffraction, to perform further studies.

Structural Heterogeneity on Charge Distributions in L. tarentolae Leishmanolysins
Comparative modeling resolved the L. tarentolae leishmanolysin three-dimensional shape to verify whether their differences affected the structural properties. The threedimensional models were built based on the crystallographic structure of L. major LmjF.10.0460 leishmanolysin, displaying 72, 68 and 65% identity and 93, 90 and 96% coverage with LtaP10.0650, LtaP10.0480, and LtaPcontig00616-1, respectively ( Figure S1). All models were refined, and their structural quality was assessed before and after the optimization procedure (Table S2). Moreover, we obtained MolProbity scores ranging from 0.71 to 1.8 among the models, indicating no stereochemical clashes. In addition, all the obtained QMEAN values were higher than −4.23, which is the low-quality model threshold, making it appropriate for further study.
Like the L. major leishmanolysin three-dimensional structure, the L. tarentolae models are composed of three domains: N-terminal, central region, and C-terminal. These structures contain a zinc peptidase motif HExxHxxGxxH, such as Metzincin class zinc peptidases. However, an unexpected 62 amino acid insertion between the glycine and last histidine residue (Gly271-His334) of the metzincin motif represents most of the central domain [21]. The comparison between the template and the models revealed that the obtained structures displayed a root-mean-square deviation (RMSD) below 0.4 Å, meaning that the structures share a characteristic folding and are highly conserved ( Figure 2). Leishmanolysins are rich in disulfide bonds. All structures presented nine conserved pairs, except for LtaPcontig00616-1, which presented eight pairs. The absence of the first disulfide bond pair is explained by the truncated N-terminal region, with 33 residues less than LmjF.10.0460 ( Figure 2).
The L. major template and L. tarentolae modeled 3D structures were submitted to molecular dynamics simulations to disclose the influence of pH on the net charge and stability of leishmanolysin. The pKa values of ionizable groups of Leishmania proteins were assessed for pH values 5.5 and 7.4. The RMSD values were computed to evaluate the influence of pH on leishmanolysin stability, resulting in an irrelevant difference below 0.3 Å ( Figure S2). The template showed the highest variability when compared to the systems, oscillating at about 1.5 Å throughout the run. The models had an increasing mean of RMSD according to the lowest sequential identity with the template, displaying 2.3 ± 0.39 Å, 2.5 ± 0.47 Å, and 3.7 ± 0.75 Å for LtaP10.0650, LtaP10.0480, and LtaPcontig00616-1, respectively.
The pH influenced the net charge of L. tarentolae proteins, with more positive values for pH 5.5 and negative for pH 7.4 (Table S3). The three active-site histidine residues that coordinate the zinc ion were assigned Nd neutral tautomers for all pH values. Non-catalytic histidine residues were positively charged at pH 5.5 and neutral at pH 7.4. (Figure S3). We also evaluated the electrostatic potential distribution on the surface of leishmanolysins. The L. major leishmanolysin showed an extensive region of negative electrostatic potential surrounding the active site ( Figure 3). At less intensity, we find a similar pattern in LtaP10.0650, while LtaP10.0480 and LtaPcontig00616-1 showed a predominantly positive pattern of charge density. This difference between the electrostatic potential of L. major and L. tarentolae leshmanolysins could affect the protein orientation, as demonstrated for other proteins using experimental and theoretical approaches [25,26], as well as the impact on their affinities for substrates with distinct charge distributions [27]. Therefore, the proteolytic activity of L. tarentolae leishmanolysins, or at least the measurement protocols, might be affected by these less negative charge distributions. To confirm the charge distribution influence, L. braziliensis and L. martiniquensis leishmanolysin had their electrostatic potential analyzed. Similar to L. major, both leishmanolysins displayed a predominantly negative pattern of charge density ( Figure S4).
Finally, to get a deeper understanding, we computed the residue solvent accessibility. Positively charged residues (lysine and arginine) presented higher SASA values (>1.5 nm/S2/N), with 6, 11, 11, and 13 residues for LmjF.10.0460, LtaP10.0650, LtaP10.0480, and LtaPcontig00616-1, respectively. We discovered that these residues led to the largest variance in charge density patterns compared to the electrostatic potential ( Figure S5). As a result, changes in the electrostatic potential of amino acids occurred, particularly at the active site of leishmanolysins, changing their affinities for molecules.   Finally, to get a deeper understanding, we computed the residue solvent accessibility. Positively charged residues (lysine and arginine) presented higher SASA values (>1.5 nm/S2/N), with 6, 11, 11, and 13 residues for LmjF. 10.0460, LtaP10.0650, LtaP10.0480, and LtaPcontig00616-1, respectively. We discovered that these residues led to the largest variance in charge density patterns compared to the electrostatic potential ( Figure S5). As a result, changes in the electrostatic potential of amino acids occurred, particularly at the active site of leishmanolysins, changing their affinities for molecules. The comparative molecular dynamic analysis was performed using the crystal structure of L. major LmjF.10.0460 (PDBid: 1LML) and the modeled L. tarentolae LtaP10.0480, LtaP10.0650, and LtaPcontig00616_1 in pH 5.5 and 7.4. The molecular surface is colored according to electrostatic potential using the Chimera software, where red, white, and blue correspond to acidic, neutral, and basic potentials, respectively.

L. tarentolae Leishmanolysin Has a Low Proteolytic Activity
Aiming to experimentally evaluate the molecular dynamics results of L. tarentolae leishmanolysins and to compare their proteolytic activity with L. major leishmanolysin (LmjF. 10.0460), we cloned the most closely related leishmanolysin (LtaP10.0650) and the L. major one in pLEXSY expression systems to overexpress these peptidases ( Figure S6). A hemagglutinin tag (HA tag) was added at the C-terminal of the cloned leishmanolysins, and two overexpressed clones were generated, as confirmed by the detection of the HA-tag ( Figure S7). This expression system has been widely used for a broad range of biotechnological and biomedical applications due to the advantages of L. tarentolae. Its relatively easy and cost-effective cultivation, the mammalian-like post-translational protein modifications capable of producing substantial recombinant protein yields, as well as the potential to be extended to an industrial production scale are the main benefits that have attracted the attention of many researchers and make the L. tarentolae expression system such a good platform of heterologous protein production [28].
The metallopeptidase activity of wild-type L. tarentolae and L. major were first evaluated by gelatin-SDS-PAGE zymography. By incubating the soluble extracts in the gelatin gels for 72 h in phosphate buffer pH 5.5, we could observe a weak halo of degradation by the wildtype L. tarentolae extracts migrating in the 63 kDa range, which is remarkably close to the wild-type L. major polypeptides (Figure 4). These 63 kDa degradations were inhibited by the metallopeptidase inhibitor 1,10-phenanthroline and partially inhibited by the ion chelator EDTA, strongly suggesting the metallopeptidase activity ( Figure 4). Although previous reports failed to detect proteolytic activity in L. tarentolae, this result is not surprising. Our research group has reported proteolytic activity in all monoxenic trypanosomatids and the phytomonads examined up to now, as well as all Leishmania spp. assessed so far [16]. Nevertheless, the activity is minimal compared to L. major (Figure 4) or to any other studied trypanosomatid. No difference was observed in gels incubated at pH 7.4 or even by the addition of DTT (data not shown). As expected, the metallopeptidase activity was higher in the two overexpressed L. tarentolae leishmanolysin strains and much higher in the two strains overexpressing L. major metallopeptidase, being incapable of complete inhibition by the metallopeptidase inhibitor at the concentration of 10 mM, probably due to the high amount of enzyme produced (Figure 4). This huge difference between the leishmanolysin activity from L. tarentolae and L. major might be attributable to the higher negative potential observed in the molecular dynamic simulations. Moreover, considering that several factors influence peptidase detection through gelatin-SDS-PAGE zymography [29], additional parameters were evaluated in our enzymatic assays, such as the buffer pH and the time of incubation, which could explain the differences between these results and those reported previously [15]. However, few details on the methodological approach were described, and it was assumed that the high sequence variability of the L. tarentolae leishmanolysin genes might have affected the peptidase activity. Accordingly, our research group previously observed no metallopeptidase activity incubating the gelatin gels with L. tarentolae extracts in an acidic buffer (pH 5.5) supplemented with 2 mM DTT for 24 h [30].
To better evaluate the differences between the proteolytic activity of L. tarenrolae and L. major leishmanolysins, we performed solution assays with the extracts of one clone of each overexpressed peptidase employing a specific matrix metallopeptidase substrate. After 40 min, the substrate was fully hydrolyzed by the soluble extract of the L. major leishmanolysin clone, and this proteolytic activity was almost completely inhibited in the presence of 1,10-phenanthroline, corroborating our previous findings in gelatin zymography ( Figure 5). However, the overexpressing L. tarentolae leishmanolysin degraded about three times less substrate than the L. major enzyme. Furthermore, the presence of the serine peptidase inhibitor PMSF and the cysteine peptidase inhibitor E-64 had a slight effect on the activity of both leishmanolysin clones, while the ion chelators EDTA and EGTA partially affected the L. major peptidase ( Table 2). Similar results were observed in the supernatant extracts from the spent culture medium ( Figure S8). These results corroborate our hypothesis of reduced proteolytic activity of L. tarentolae leishmanolysin and raise a question about an indirect correlation between gene expansion and proteolytic activity. For instance, similarly to L. tarentolae, T. cruzi leishmanolysin suffer an extensive expansion to more than 420 predicted genes and pseudogenes, but the metallopeptidase activity of T. cruzi extracts is rather weak, only detected when the parasites have their soluble extracts obtained in the presence of proteolytic inhibitors from the other peptidase classes, such as PMSF, E-64, and the aspartyl inhibitor pepstatin A [31]. In contrast, pathogenic Leishmania spp. usually presents abundant detectable leishmanolysin activity and a lower number of leishmanolysin genes. To better evaluate the differences between the proteolytic activity of L. tarenrolae and L. major leishmanolysins, we performed solution assays with the extracts of one clone of each overexpressed peptidase employing a specific matrix metallopeptidase substrate. After 40 min, the substrate was fully hydrolyzed by the soluble extract of the L. major leishmanolysin clone, and this proteolytic activity was almost completely inhibited in the presence of 1,10-phenanthroline, corroborating our previous findings in gelatin zymography ( Figure 5). However, the overexpressing L. tarentolae leishmanolysin degraded about three times less substrate than the L. major enzyme. Furthermore, the presence of the serine peptidase inhibitor PMSF and the cysteine peptidase inhibitor E-64 had a slight effect on the activity of both leishmanolysin clones, while the ion chelators EDTA and EGTA partially  In solution assays were performed using the fluorogenic substrate MCA-Pro-Cha-Gly-Nva-His-Ala-Dpa-NH2 at 37 • C for 1 h. Activity of the cloned leishmanolysin from L. tarentolae (LtaP10.0650) and from L. major (LmjF.10.0460) in 100 mM glycine-NaOH buffer pH 10.0 buffer alone was taken as 100%. The effects of proteolytic inhibitors described below were measured according to the Materials and Methods. more than 420 predicted genes and pseudogenes, but the metallopeptidase activity of T. cruzi extracts is rather weak, only detected when the parasites have their soluble extracts obtained in the presence of proteolytic inhibitors from the other peptidase classes, such as PMSF, E-64, and the aspartyl inhibitor pepstatin A [31]. In contrast, pathogenic Leishmania spp. usually presents abundant detectable leishmanolysin activity and a lower number of leishmanolysin genes.

Sequence Analysis and Phylogenetic Reconstruction
Protein sequences of Leishmania major, L. braziliensis, L. martiniquensis, and L. tarentolae Parrot-TarII reference strain annotated as leishmanolysin (or GP63 or PSP or MSP) were retrieved from the TriTryp database (DB) [32]. These proteins were analyzed for In-terPro families and domains databases by Interproscan v5.51-85.0 to confirm the presence of LSHMANOLYSIN (PR00782) and peptidase_M8 (PF01457) domains. The proteins containing both domains and more than 200 residues were aligned using MUS-CLE v3. 8

Comparative Modeling and Molecular Dynamics Simulations
The structure of three L. tarentolae leishmanolysin proteins (LtaP10.0650, LtaP10.0480, and LtaPcontig00616-1), L. braziliensis LbrM.10.0600, and L. martiniquensis LMARLEM2494_ 000016500 were modeled. Initially, BLASTP was used to search the PDB database and select the 1.86 Å resolution crystal structure of L. major leishmanolysin (PDB ID: 1LML) [21] as a template. Then, the leishmanolysin sequences were aligned using the PSI-Coffee mode of the T-Coffee program, and five hundred homology models were created using the "standard auto model" routine of the Modeller v. 9.23 [33]. All modeled structures were optimized via the variable target function method (VTFM) with up to 300 iterations. Molecular dynamic (MD) optimization was conducted in the slow level mode, repeating the whole cycle twice to produce optimized conformations of the model. Finally, the modeled structures were filtered according to their discrete optimized protein energy (DOPE) score and refined using the locPREFMD server [34]. Original and optimized models were evaluated by the QMean server, while the ERRAT and Ramachandran plots were calculated on the SAVES server of UCLA-DOE Lab for stereochemistry analysis. Sequence alignment and three-dimensional structure figures were generated using ALINE and UCSF ChimeraX.
The three models of L. tarentolae and the template of L. major were simulated in the pH values of 5.5 and 7.4. MD simulations were carried out using AMBER 20.0; protein interactions were represented using amber ff14SB force field [35]. Protonation states of titratable residues were assigned using the PDB2PQR server. The disulfide bonds were specified. Electrostatic interactions were treated using the particle mesh Ewald (PME) algorithm with a cutoff of 12 Å. Each system was simulated under periodic boundary conditions in a triclinic box considering 12 Å from the outermost protein atoms in all cartesian directions. The simulation box was filled with TIP3P water molecules [36]. An appropriate number of Na + and Cl − counter ions were added to neutralize the systems. Subsequently, a two-step energy minimization procedure was performed: (i) 2000 steps (1000 steepest descent + 1000 conjugate-gradient) with all heavy atoms harmonically restrained with force constant of 5 kcal mol −1 Å −2 ; (ii) 5000 steps (2500 steepest descent + 2500 conjugategradient) without position restraints. All systems were further equilibrated during nine successive 500 ps equilibration simulations and simulated with no restraints at 300 K in the Gibbs ensemble with a 1 atm pressure using isotropic coupling. All chemical bonds containing hydrogen atoms were restricted using the SHAKE algorithm [37]. Three independent MD runs of 300 ns for each system were simulated using different initial velocities, achieving 7.2 µs simulation time. Trajectories were analyzed with GROMACS package tools. Root-mean-square deviation (RMSD) values were calculated separately for each system fitting their backbone atoms. Conformational clusterization for protein was performed using the GROMOS method with a cutoff of 2.0 Å, considering the backbone from the whole trajectory of each MD simulation. The central structure of the largest cluster was taken for the electrostatic potential analysis conducted with the Adaptive Poisson-Boltzmann Solver (APBS) program. Root-mean-square fluctuation (RMSF) and solvent accessible surface areas (SASA) were calculated for each system's concatenated three independent MD simulations.
As previously described, crude extracts of L. tarentolae mutants were analyzed by Western blotting to confirm the overexpression [31]. Briefly, proteins were separated in 10% SDS-PAGE, and the polypeptides were electrophoretically transferred to nitrocellulose membranes and blocked in 10% low-fat dried milk dissolved in PBS containing 2% Tween 20 (Sigma-Aldrich). The membranes were washed and incubated with an anti-HA antibody (Thermo Scientific) (1:1000 dilution) for 1 h. The secondary peroxidase-conjugated goat anti-rabbit immunoglobulin G (1:1500 dilution) (Thermo Scientific) was used, followed by chemiluminescence immunodetection. An anti-α-tubulin polyclonal antibody (Sigma-Aldrich) produced in rabbit (1:5000 dilution) was used as a loading control.

Parasite Cultivation for Protein Extraction
L. tarentolae Parrot-TarII strain wild-type strain (WT) and L. major (MHOM/SU/73/5-ASKH) promastigotes were routinely maintained in Schneider's medium supplemented with 10% FBS at 26 • C. In addition, L. tarentolae mutants overexpressing leishmanolysin (Lta + LtaP10.0650 and Lta + LmjF.10.0460) were maintained with 100 µg/mL hygromycin up to 20 passages, according to manufacturer's instructions. To obtain the protein crude extracts, 10 9 promastigotes from each strain were resuspended in 2% CHAPS (Sigma-Aldrich) diluted in 50 mM Tris-HCl, pH 7.2, and broken in a vortex by alternating 30 s shaking and 30 s cooling intervals [29]. The crude extract was centrifuged at 4 • C, and the soluble and insoluble fractions were separated by centrifugation (14,000× g). For the supernatant extraction, the spent culture medium was concentrated 50-fold in a 10,000-molecular-weight cutoff Amicon micropartition system (Amicon ® , Sigma-Aldrich). The protein concentration was determined following the manufacturer's instructions of Pierce™ BCA Protein Assay Kit (Thermo Scientific), using bovine serum albumin (BSA) as standard, and the samples were stored at −80 • C.

Enzymatic Activity Assays
The leishmanolysin activity of L. tarentolae mutants was assayed simultaneously by gel zymography and fluorimetry. First, approximately 40 µg of the soluble extracts from the supernatant were submitted to 12% polyacrylamide-SDS gel electrophoresis (SDS-PAGE) containing 0.1% gelatin from porcine skin (Sigma-Aldrich) in an ice-bath. Samples were diluted (v/v) in SDS-PAGE sample buffer 4× without reducing agent and loaded onto gels. Electrophoresis was carried out under constant current (25 mA). After running, the gels were washed 4 times in 2.5% Triton X-100 (Sigma-Aldrich) under agitation for 1 h at room temperature for SDS removal and enzyme renaturation. Subsequently, the proteolytic activity was determined by incubating the gels at 37 • C for 72 h in 50 mM phosphate buffer, pH 5.5 and pH 7.4, containing or not 2 mM dithiothreitol (DTT) (Sigma-Aldrich). To determine the peptidase class, the lysates were incubated with the inhibitors: (i) 1,10-phenanthroline (Phe, 10 mM), ethylene-bis(oxyethylenenitrilo)tetraacetic acid (EGTA, 10 mM) and ethylenediaminetetraacetic acid (EDTA, 10 mM) for metallopeptidases; (ii) Nα-tosyl-Lys chloromethyl ketone (TLCK, 100 µM) and phenylmethanesulfonyl fluoride (PMSF, 1 mM) for serine peptidases, (iii) and trans-epoxysuccinyl-L-leucylamido(4-guanidino)butane (E-64, 10 µM) for cysteine peptidases. The gels were stained with 0.2% Coomassie Brilliant Blue R-250 (Sigma-Aldrich) in methanol-acetic acid-water (40:10:50) and destained in the same solution without the dye. The gels were scanned and analyzed with Image Scanner III (GE HealthCare, Chicago, IL, USA). The molecular masses of the peptidases were estimated by comparison with the mobility of low molecular mass standards [29].
Alternatively, the enzymatic activity was measured continuously using the fluorogenic peptide substrate MCA-Pro-Cha-Gly-Nva-His-Ala-Dpa-NH2 (Sigma-Aldrich) at 37 • C in 0.1 M glycine-NaOH buffer pH 10.0 containing 1 mM CaCl 2 [31]. The assays were performed in black round bottom 96-microwell plates in the Spectra Max Gemini spectrofluorometer (Molecular Devices, San Jose, CA, USA), with excitation wavelength at 328 nm and emission at 393 nm. The reaction started with the addition of 10 µM substrate to 10 µg of the parasite cellular extract, or the supernatant extract, in a total volume of 100 µL. The inhibition studies were performed by incubation with the following inhibitors: 10 mM Phe, 10 mM EDTA, 10 mM EGTA, 1 mM PMSF, or 10 µM E-64. In parallel, the addition of divalent ions ZnCl 2 , ZnSO 4 , CaCl 2, and MnCl 2 at 5 mM was also evaluated.