A Soluble Pyrophosphatase, a Key Enzyme for Polyphosphate Metabolism in Leishmania*

We report the functional characterization in Leishmania amazonensis of a soluble pyrophosphatase (LaVSP1) that localizes in acidocalcisomes, a vesicular acidic compartment. LaVSP1 is preferentially expressed in metacyclic forms. Experiments with dominant negative mutants show the requirement of LaVSP1 functional expression for metacyclogenesis and virulence in mice. Depending on the pH and the cofactors Mg2+ or Zn2+, both present in acidocalcisomes, LaVSP1 hydrolyzes either inorganic pyrophosphate (Km = 92 μm, kcat = 125 s–1), tripolyphosphate (Km = 1153 μm, kcat = 131 s–1), or polyphosphate of 28 residues (Km = 123 μm, kcat = 8 s–1). Predicted structural analysis suggests that the structural orientation of the residue Lys78 in LaVSP1 accounts for the observed increase in Km compared with the yeast pyrophosphatase and for the ability of trypanosomatid VSP1 enzymes to hydrolyze polyphosphate. These results make the VSP1 enzyme an attractive drug target against trypanosomatid parasites.

Leishmania is a parasitic protozoan causing an endemic disease affecting 88 countries on four continents (www.who.int/tdr/). Treatment with pentavalent antimony-based drugs has been used for decades. However, toxicity and loss of efficacy with the emergence of resistant strains render necessary the development of alternative therapies.
Leishmania exists alternatively as a flagellated extracellular promastigote living inside the alimentary tract of the female sandfly (subfamily phlebotominae) insect vector or as an aflagellated intracellular amastigote within the acidic parasitophorous vacuoles of the mammalian host mononuclear phagocyte (1). During this digenetic life cycle Leishmania encounters major environmental changes that induce parasite differentiation and gene stage-specific expression essential for survival. Some of these dramatic changes are phosphate deprivation, pH acidification, and reduction in extracellular osmolarity. Acidocalcisomes and their polyphosphate (polyP) 2 metabolism have been shown to play an essential role in the physiological adaptations of the parasite (2)(3)(4).
Inorganic polyP is a ubiquitous molecule formed by phosphate (P i ) residues linked by high energy phosphoanhydride bond. PolyP synthesis is catalyzed by a polyphosphate kinase in prokaryotes (5,6). In eukaryotes with the exception of Dictyostelium discoideum (7), Candida humicola (8), and photosynthetic eukaryotes (9), polyphosphate kinase is absent. A system coupling phosphate acquisition and polyP synthesis has been suggested to be operative in yeast (10). In D. discoideum an actin-related protein complex called polyphosphate kinase 2 is polymerized into an actin-like filament, concurrent with its reversible synthesis of a polyP chain from ATP (11). Polyphosphate hydrolysis is catalyzed by exo-and endo-polyphosphatases (12). Exopolyphosphatases are considered as the central regulatory enzymes in polyP metabolism (13). However, the only known gene encoding for an exopolyphosphatase (14) cannot account for all the different molecular masses, substrate specificities and requirements of divalent cations described for exopolyphosphatase activities (13). We recently functionally characterized a soluble pyrophosphatase (TbVSP1) in Trypanosoma brucei that localizes in acidocalcisomes and is able to catalyze the hydrolysis of both PP i and polyP (3). This enzyme was also shown to be involved in polyP metabolism and is important for parasite virulence (3).
In the present work, we report the identification in Leishmania amazonensis of a pyrophosphatase (LaVSP1; accession number AAP74700) orthologous to TbVSP1 (3). This enzyme localized to acidocalcisomes and was preferentially expressed in metacyclic promastigotes. We used dominant negative mutants expressing nonfunctional versions of the enzyme to show that LaVSP1 is required for polyP metabolism, metacyclogenesis, and virulence in mice.
Cloning of L. amazonensis VSP1 Gene-A PCR amplification on genomic DNA was performed using the degenerated set of oligonucleotides LmCPPO2 (5Ј-ATGTTYTCNGCNTTY-3Ј) and LmCPP03 (5Ј-TGYTGYTGRTGRTGH-3Ј) designed from a partial sequence of the Leishmania major VSP1 (AL354096). The amplified DNA fragment of 300 bp was sequenced and used as a probe to screen a cosmid library 3 using the pCosTL vector (gift from John Kelly, London School of Hygiene and Tropical Medicine). The complete sequence of LaVSP1 gene was obtained by direct cosmid sequencing.
Overexpression of mLaVSP1 or nLaVSP1 Proteins in L. amazonensis-Expressions of the mLaVSP1 and nLaVSP1 proteins in L. amazonensis were accomplished by cloning their respective genes in the pFHTLG (gift from Ken Stuart, Seattle Biomedical Research Institute) (18), between the BglII and AvrII restriction sites and transfection of L. amazonensis (BA125). The genes were obtained by PCR amplification using the pET23a constructions as templates and the following set of primers: LDNcapase1 (5Ј-GCC-GGAGATCTATGTCCGATGACACGCGCTACAAG-3Ј) and LDNcap-ase2 (5Ј-CCTTTCGGGCTTTGTTAGCCCTAGGATCTCAGTG-3Ј). The mLaVSP1 and the nLaVSP1 proteins were expressed with a His 6 tag.
For constitutive expression of either nLaVSP1 or mLaVSP1, 10 8 exponentially growing L. amazonensis BA125 were transfected with the recombinant pFHLTG vectors. The resulting cell lines were called BA125/nLaVSP1 and BA125/mLaVSP1, respectively. To obtain an inducible expression of the same proteins, L. amazonensis was previously transfected with the pTupactet construct (18). The resulting cell lines were called BA125In/nLaVSP1 and BA125In/mLaVSP1, respectively. Protein expression was induced by culturing cells in culture medium with 50 g/ml hygromycin and 1 g/ml tetracycline for 1 week.
To integrate the Tet-R gene into the tubulin locus, pTupactet vector was digested with KpnI and NotI, and the cells were electroporated with 5 g of this DNA preparation. The cells with integrated pTupactet DNA were selected by the addition of 45 g/ml puromycin. These cells were then transfected by the recombinants pFHTLG vectors (as described above) and selected by the addition of 10 g/ml hygromycin.
Purification of the Recombinant nLaVSP1 and mLaVSP1 Proteins by Immobilized Metal Affinity Chromatography-The recombinant proteins expressed in E. coli were purified from the His.Bind column according to the manufacturer's instructions (Novagen). Before performing the PPase assay, the enzyme was desalted in MES buffer (50 mM MES/NaOH, pH 6.5, 500 mM NaCl, and 5 mM MgCl 2 ) on a PD10 column (Amersham Biosciences) and stored in that buffer.
Different affinities of PP i to Mg 2ϩ as a function of pH render the Mg 2ϩ dependences for PPases complex (19). Some buffers such as Tris modulate E. coli PPase activity (20) We therefore stored the enzyme in MES/NaOH, pH 6.5, 5 mM MgCl 2 , and 250 mM NaCl. The reactions were then performed in reaction buffers containing either Mg 2ϩ or Zn 2ϩ and at different pH levels.
PPase and Exopolyphosphatase Activity-0.020 -2 g of desalted nLaVSP1 and mLaVSP1 proteins were incubated at 25°C for 1 min to hydrolyze PP i and 2 min for either polyP 3 or polyP 28 . The reaction was stopped, and released P i was quantified as described by Shatton et al. (21). MES/NaOH buffer was used from pH 5.5 to 6.5 and Tris-HCl from pH 7.0 to 8.8. Before use, polyP 28 was purified on G25 Sephadex.
For each substrate concentration tested a blank value, corresponding to the same reaction conditions but without the enzyme, was subtracted from the values obtained in the presence of the enzyme. Because polyP 3 hydrolysis produces P i and PP i , which could be hydrolyzed to P i , we used the equation developed for an enzyme with two substrates, as published by Zyryanov et al. (22), to calculate the Michaelis constants.
Determination of Protein Concentration-Protein concentration was estimated by the method of Bradford (23), by Coomassie staining after SDS-PAGE, and/or by Western blotting.
Western Blot Analysis of the mLaVSP1 and nLaVSP1 as Extracted from Leishmania-3ϫ10 6 cells were lysed with SDS and a pool of protease inhibitors, boiled for 5 min, and then submitted to 12.5% SDS-PAGE with standard markers (Promega).
The proteins were transferred onto Immobilon P membranes (Amersham Biosciences). Rabbit immune serum (1:5,000 dilution) and goat anti-rabbit Ig/peroxidase (1:10,000 dilution) were used. Staining was done with ECL TM Western blotting detection reagents as described by the manufacturer (Amersham Biosciences). The membranes were scanned to produce initial images, and the protein bands were quantified using NIH Image software.
Metacyclic Promastigote Purification-Leishmania cells were harvested when in stationary phase and then washed twice in PBS; the cell suspension was adjusted to 5ϫ10 8 cells/ml and incubated 30 min at room temperature with a 1:500 dilution of monoclonal antibody 3A1, kindly provided by Dr. Elvira Maria. Saraiva (24). This antibody is directed against the L. amazonensis promastigote forms. The cell suspension was then centrifuged at 250 ϫ g for 5 min at 4°C. The supernatant was washed twice in PBS, and unagglutinated, i.e. metacyclic, promastigotes were counted using a Malassez hemacytometer.
Macrophage Infection by the Metacyclic Form-Macrophages were obtained from BALB/c mice as described (25) by in vitro differentiation of bone marrow precursor cells in 24-well plates containing 12-mmdiameter round glass coverslips for light microscopy studies. Precursors were deposited in 60-mm cultures dishes. The cells were cultured in RPMI 1640 medium (Eurobio) supplemented with 10% (v/v) heat-inactivated fetal calf serum (Eurobio), 40 g/ml gentamicin sulfate, 20% L-929 fibroblast-conditioned medium. After 5 days at 37°C in a 5% CO 2 nonadherent cells were removed, and adherent macrophages were further harvested and incubated in culture medium containing only 2% of conditioned medium at the concentration of 10 6 cell/ml in 24-well plates containing 12-mm-diameter round glass coverslips (18 -24 h before the addition of the parasites). Macrophages were infected by stationary phase promastigotes, metacyclic promastigotes, at a ratio of four parasites/macrophage. The parasites and macrophages were brought into contact by incubation at 4°C for 4 h before being washed with Dulbecco's PBS (Seromed) to remove free parasites. Macrophages were fixed either immediately or 1-72 h after infection. The cells were fixed with methanol and stained with Giemsa for 15 min at room temperature. The coverslips were finally mounted on microscope slides with Mowiol (Calbiochem). The percentage of infected macrophages (of 600 -700 cells) and the mean number of amastigotes/infected macrophage (of 150 -200 cells) were determined using an inverted microscope.
Infectivity Tests-Leishmania promastigotes were cultured for up to 8 weeks after recovery from footpad lesions and harvested in early stationary phase of growth (5-6 days after last passage). 10 6 promastigotes were injected into BALB/c mouse footpads, and lesion diameters were determined weekly with a caliper.
Immunofluorescence-12-mm-diameter round glass coverslips were treated for 30 min in 10 g/ml poly-L-lysine (Ͼ300 kDa; Sigma) in 24-well tissue culture plates (Costar) (15). Leishmania promastigotes (10 6 cells/well) were added, and the plates were centrifuged at 130 ϫ g for 10 min, fixed with 2% paraformaldehyde for 1 h, and sequentially incubated for 10 min with 50 mM NH 4 Cl for 30 min with 10% mouse serum (Jackson laboratories) plus 10% goat serum (Eurobio) in PBS, 1 h with 1:100 dilution of mouse anti-poly-His 5 antibody (Qiagen) and 1:1,000 dilution of rabbit anti-VP1 immune serum or 1:1,000 dilution of rabbit anti-exopolyphosphatase immune serum in 0.25% (w/v) gelatin (Sigma) and 1 h with 10 g/ml fluorescein-conjugated goat anti-rabbit Ig antibody (Jackson Laboratories) and 10 g/ml rhodamine-conjugated goat anti-mouse Ig antibody (Jackson Laboratories) in 0.25% (w/v) gelatin. The coverslips were mounted on microscope slides with Mowiol (Calbiochem). Cell preparations were examined under a conventional Zeiss Axioplan 2 fluorescence microscope. The images were acquired with a Princeton Instruments camera and analyzed with Metaview TM (Universal Imaging).
6-Diamino-2-phenylindol (DAPI) Accumulation Observed by Fluorescence Microscopy-12-mm-diameter round glass poly-L-lysinetreated coverslips (Ͼ300 kDa; Sigma) were used in 24-well tissue culture plates (Costar). Leishmania promastigotes (10 6 cells/well) were washed twice Dulbecco's PBS and incubated for 10 min at 30°C in PBS with 10 g/ml DAPI. Then cells were added in the 24-well tissue culture plates containing the 12-mm-diameter round glass coverslips treated and centrifuged at 130 ϫ g for 10 min. The coverslips were mounted on microscope slides without Mowiol nor resin. PolyP and DNA were detected using the DAPI filter from Zeiss (set 02) and by exposing for 200 ms for polyP detection. DAPI has a fluorescence emission maximum at 456 nm. PolyP shift DAPI fluorescence to a higher wavelength, with a maximum of 525 nm (26).

RESULTS AND DISCUSSION
LaVSP1 Gene Cloning and Phylogenetic Analysis-Cloning and sequencing of LaVSP1 were performed as described under "Experimental Procedures." A search in the L. major genome sequence data base (www.genedb.org) with BLAST algorithms, using the LaVSP1 protein sequence, indicated the existence of only one copy of VSP1, which localized on chromosome 11 (LmjF11.0210). The LaVSP1 and LmVSP1 protein sequences only differ in 15 amino acids (ALIGN_000878), and none of them correspond to amino acids involved in the catalytic center (27).
LaVSP1 gene codes for a protein of 443 amino acids (28). Sequence comparison analysis (www.ncbi.nlm.nih.gov/BLAST) revealed that LaVSP1 exhibits 64% sequence identity with the TbVSP1 (ALIGN_000878). In addition to the pyrophosphatase domain, LaVSP1 has a large N-terminal extension region of 160 amino acids, as previously observed for TbVSP1 (28). This region contains a calcium-binding type II EF-hand domain (COG5126, FRQ1) between positions 111 and 173.
Phylogenetic protein analysis of LaVSP1 and the cytosolic soluble pyrophosphatase of Leishmania was performed with other family I sPPases of various bacteria and eukaryotes (Fig. 1). Both Leishmania acidocalcisomal (LaVSP1) and cytosolic proteins cluster with the animal/fungus PPase family (27), which also contains the plastid enzymes from plants. The mitochondrial and cytosolic PPases of the animal/ fungus family probably duplicated and diverged after the speciation (Fig.  1). In contrast, the cytosolic enzymes from plants do not cluster with the other eukaryotic enzymes, including the plastid PPases, suggesting that each plant isoform has a different origin. The origin of the two trypanosomatid PPases is not clear. Indeed, the trypanosomatid isoforms show a paraphyletic organization, and the acidocalcisomal PPases are closer to the animal/fungus PPases than to the cytosolic PPases. This observation suggests that the trypanosomatid acidocalcisomal and cytosolic PPases diverged before the trypanosomatid and animal/fungus speciation, giving rise to two different deep branches of PPases. Alternatively, they may have a different origin as proposed for the plant isoforms. One may speculate that trypanosomatid cytosolic and/or acidocalcisomal PPases have a plant origin, as previously hypothesized for many other metabolic enzymes from trypanosomatids (29). However, this view is not strongly supported by the topology of the PPase tree because the chloroplast and VSP1 PPases do not constitute a monophyletic group. In any case, this paraphyletic organization, well supported by the bootstrap values, suggests that trypanosomatid VSP1 and cytosolic PPases may carry different functions.
TbVSP1 expression, in both the insect and mammalian hosts, its acidocalcisome localization, and its ability to hydrolyze both PP i and polyP (30) are the representative cellular and biochemical features of TbVSP1. We therefore analyzed LaVSP1 for its expression throughout the L. amazonensis life cycle, its cellular localization, and biochemical specificities.
Differential Expression and Localization of LaVSP1-Relative expression level quantification between LACK (Leishmania homologue of receptors for activated C kinase) (31) and LaVSP1 proteins was estimated at different life cycle stages of L. amazonensis. Fig. 2 shows that LaVSP1 expression levels were regulated throughout the parasite life cycle. Namely, expression in metacyclic forms was three times greater than in the exponentially growing promastigotes and six times greater than in the amastigote stage. We determined by indirect immunofluorescence LaVSP1 localization in genetically modified L. amazonensis promastigotes (Ba125/nLaVSP1) expressing a recombinant His 6tagged nLaVSP1 using an anti-His 5 monoclonal antibody. The results shown in Fig. 3D indicate that nLaVSP1 localized to vesicle-like structures of various sizes mainly distributed around the nucleus. Such cellular organization is typical of acidocalcisome distribution (32). Next, to refine this localization, a rabbit antisera directed against the T. brucei vacuolar proton translocating pyrophosphatase (TbVP1) was used in a colocalization study (Fig. 3) with nLaVSP1 as an acidocalcisome marker on the Ba125/nLaVSP1 strain (30) (Fig. 3). The results shown in Fig. 3 indicated that LaVP1 and nLaVSP1 colocalized. In addition, LaVSP1 was also found in some vesicles where LaVP1 was not detected (Fig. 3B).
In a second step to confirm that the structures containing LaVSP1 indeed corresponded to acidocalcisomes, we performed a second experiment where we studied the colocalization of nLaVSP1 and the L. major exopolyphosphatase (LmPPX1). The latter has been reported to localize in acidocalcisomes (33). Fig. 4 shows that LaVSP1 localized in acidocalcisomes where the levels of LaPPX1 were low. However, this finding is in agreement with Fig. 3B and previously reported data showing that Leishmania acidocalcisomes contain heterogenous levels of VP1 and VSP1 (33,34).
To explain these data we may argue that different acidocalcisome populations with different structures and protein content (4, 34 -37) account for the partial colocalizations observed. Alternatively, LaVSP1 may also localize in another structure as well. It has been shown in D. discoideum (38) and in Trypanosoma cruzi (4) that acidocalcisomes and the contractile vacuole complex are functionally linked and possess common markers such as a vacuolar H ϩ -ATPase, an H ϩ -pyrophosphatase, and a Ca 2ϩ -ATPase (38). In any case additional experiments will be required to investigate the acidocalcisome heterogeneity.
Substrate Specificities of LaVSP1-If PPases display nearly absolute specificity for PP i in the presence of Mg 2ϩ , when a transition metal such as Zn 2ϩ , Co 2ϩ , or Mn 2ϩ is used, this specificity is lost. We previously observed for TbVSP (3) that depending on the cofactors used in the buffer (Mg 2ϩ or Zn 2ϩ ) and the pH buffer (7.5 or acidic), the enzyme had  We observed that LaVSP1 specifically hydrolyzed PP i in a reaction buffer containing Mg 2ϩ with an optimum pH of 7 (Fig. 5). The replacement of Mg 2ϩ for Zn 2ϩ in the reaction buffer resulted in the hydrolysis of polyP 3 and polyP 28 , in addition to PP i with optimum pH levels of 5.5, 6.5, and 5.5, respectively. The k cat and K m values for PP i hydrolysis, at pH 7.2 in the presence of Mg 2ϩ , were estimated to be 125 s Ϫ1 and 92 M, respectively (Fig. 5). The calculated catalytic constants for polyP 3 and polyP 28 were 131 and 8 s Ϫ1 for the k cat values and 1153 and 123 M for the K m values, respectively (Fig. 5). The k cat value for PP i hydrolysis using Zn 2ϩ as a cofactor was estimated to be 49 s Ϫ1 . Interestingly, the high K m value of 1153 M for polyP 3 hydrolysis could be correlated to high level of polyP 3 found in parasitic protozoa (39), which can reach 54,3 mM in T. cruzi epimastigotes (40). polyP 3 is one of the final products of the short and long chain polyP hydrolysis (41). We previously proposed that polyP 3 could be further hydrolyzed, leading to reduced polyP 3 accumulation in acidocalcisomes and to P i production (28). A conventional exopolyphosphatase, LmPPX1, was characterized in Leishmania acidocalcisomes. In the presence of Mg 2ϩ , LmPPX1 reached a maximum activity for pH values ranging from 7.0 to 8 (30). These data indicate that at pH 5-7.5, in the presence of Zn 2ϩ , short chain polyP and PP i hydrolysis could be carried out by the LaVSP1 protein. This biochemical analysis suggests that LaVSP1 could play a role in the acidocalcisome polyphosphate metabolism.
Implication of LaVSP in the Regulation of the Polyphosphate Metabolism in Acidocalcisomes-Because of their N-terminal domain, VSP1 proteins form functional oligomers. We therefore decided to use the dominant negative mutant strategy to study the role of LaVSP1 in the acidocalcisome polyphosphate metabolism. To attain such a goal, promastigote dominant negative mutants were obtained by overexpressing a mutated recombinant LaVSP1 (mLaVSP1). The His 6 -tagged recombinant D325E-LaVSP1 (mLaVSP1) mutant was expressed in E. coli and purified by immobilized metal affinity chromatography. This mutation was selected because the corresponding D120E variant in yeast had no activity (42). First, we confirmed that the recombinant protein could not hydrolyze PP i (data not shown). Secondly, the mutated gene was cloned into the pFHLTG vector to obtain a tetracycline-dependent expression of mLaVSP1 in the L. amazonensis BA125 strain (BA-Ind/mLaVSP1), previously transfected with the pTupactet vector (18). Western blot analysis (Fig. 6, A and B) showed that LaVSP1 expression in noninduced cells was four times less than that in the induced parasites. The same samples analyzed with an anti-His 5 monoclonal antibody indicated expression of mLaVSP1 in the noninduced cells (Fig. 6B). However, this level of expression is much lower than that in  induced cells and was quantified as representing only a third of the endogenous native protein level. DAPI staining, which allowed the detection of polyP in acidocalcisomes (40), revealed a much lower accumulation of polyP in the induced than in the noninduced BA-Ind/mLaVSP1 cells (Fig.  6, C and D). These data confirm a role for LaVSP1 in the acidocalcisome polyphosphate metabolism.
LaVSP1 Is Not Essential for Leishmania Promastigote Growth and Macrophage Infection but for Metacyclogenesis-WT, noninduced, and induced mLaVSP1 promastigotes grew in vitro with an equivalent doubling time of around eight h, and at the stationary phase they reached the same cell density and presented the same viability. However, metacyclic promastigotes, a nondividing stage, which differentiate during the stationary phase (43), were shown to express the highest level of LaVSP1 during the parasite life cycle stage (Fig. 2). We therefore analyzed the importance of LaVSP1 for metacyclogenesis. As monitored by purifica-tion with the 3A1 monoclonal antibody (24), metacyclic promastigotes appeared after entry into stationary phase and peaked after 3-4 days (Fig. 7). Morphological change (43) was also obvious and distinguished them from procyclic promastigotes. Fig. 7 shows that Tet-induced promastigotes produced 10 times less metacyclics when compared with either the noninduced or WT cells. This phenotype reminded what was observed for spt2 Ϫ promastigote mutants that failed to differentiate into metacyclic promastigotes. These mutants died following an autophagic process (44), and their acidocalcisomes were devoid of long chain polyp (45). Is polyP depletion in acidocalcisomes responsible for that defect? An answer can be found in the ability of polyP to respond to cellular stresses such as pH and nutrient changes (40,46), which are known to be crucial for parasite differentiation (47,48).
In vitro infections of bone marrow macrophages were performed with purified metacyclic promastigotes. The results indicated that soon (up to 5 h) after macrophage infection with either noninduced, induced BA125Ind/mLaVSP1 or WT cells, 80 -90% of macrophages appeared with attached or internalized parasites (data not shown). We found some difficulties in maintaining the overexpression of mLaVSP1 in the BA125In/mLaVSP1 strain when the parasites were not in axenic cultures. Forty-eight h following infection, 4 and 16% of the macrophages were still infected by the induced and noninduced BA125Ind/mLaVSP1 parasites. This last piece of data suggests that under these conditions, metacyclic parasites expressing a mutated nonactive LaVSP1 were still able to infect, differentiate, and survive inside macrophages.
Virulence toward Mice-To obtain a constitutive high level of mLaVSP1 expression, WT BA125 were only transfected with the pFHLTG-D330ELaVSP1 construction (Fig. 8, inset). Overexpression under the same conditions of nLaVSP1 was used as a control (BA125/ nLaVSP1). Two independent transfections were performed, and we did not clone the transfected cells. For both transfections, only mice inoculated with BA125/nLaVSP1 late stationary phase promastigotes displayed cutaneous lesions (Fig. 8). We did not attribute this virulence defect to metacyclogenesis deficiency only because metacyclic promastigotes (up to 500) were sufficiently abundant in the inoculum to develop lesions. This information indicates that LaVSP1 is essential to maintain virulence in mice and that acidocalcisomes and their polyP metabolism as previously suggested (3,49) may represent interesting drug targets.
According to the in vitro macrophage infection study, the virulence defect of BA125/mLaVSP1 in mice could be explained by different environmental conditions found in the two systems. Several characteristics reveal the presence of a hypoxic microenvironment in lesions (50) and indicate that macrophages exposed to hypoxia showed a reduction in the percentage of infected cells and the number of intracellular parasites/cell BA125Ind/mLaVSP1 strain were loaded in these gels. LaVSP1 expressions were quantified by Western blotting using rabbit antisera directed against T. brucei VSP1 (28). The LACK antigen was used as a loading marker (28). B, immunoblot analysis of mLaVSP1. Monoclonal antibodies directed against the poly-His tag (Qiagen) were used to probe the recombinant protein mLaVSP1. C and D, overlapping images of DAPI staining in noninduced (C) and induced (D) L. amazonensis mutants. PolyP appears in green, and DNA appears in blue. (51). Moreover, observations on the kinetics of infection showed that hypoxia did not depress L. amazonensis phagocytosis but induced macrophages to reduce intracellular parasitism (51). Polyphosphates were shown to regulate the mammalian TOR (target of rapamycin) kinase (52) involved in signaling pathways implicated in cellular stress responses such as hypoxia (53). Thus BA125/mLaVSP1 might fail to develop in mice because of their polyphosphatase metabolism deficiency, which did not allow parasite to survive under the hypoxic conditions found in the lesions.
Predicted Structure Analysis of TbVSP1, LaVSP1, and T. cruzi VSP1 (TcVSP1)-The animal/fungus and acidocalcisome PPases differ by the ability of the former ones to hydrolyze polyP (Fig. 5). This substrate specificity difference may offer the opportunity to develop molecules specifically inhibiting the polyP activity of the acidocalcisomal PPases. Despite the primary sequence homology observed between trypanosomatid and the animal/fungus PPases (27), we looked for structural differences existing between the catalytic center of the two sets of PPases.
Based on x-ray crystallographic analyses, it has been shown that 14 or 15 charged or polar residues are conserved in structural alignment between the E. coli apo PPase and the yeast PPase (Y-PPase) (42,54). Assuming that PP i and polyP could be bound by PPases according to different modes, we focused our attention on the two phosphate-binding subsites, P1 and P2. In Y-PPase, P1 is formed by Arg 78 , Tyr 192 , Lys 193 , and metal ions M3 and M4, whereas P2 interacted with Lys 56 , Tyr 93 , and metal ions, M1-M4 (54). The SWISS-MODEL (swissmodel.expasy.org) was used as an automated protein structure homology modeling server to identify theoretical structural differences of the phosphate-binding sites of LaVSP1 and Y-PPase. The data presented in Fig. 9A indicate a major difference in the positioning of Arg 78 . Two other significant changes concern the side chains of Lys 56 and Glu 58 . Because such structural differences, depending on the P1 and P2 phosphate-binding subsites were previously reported for the yeast R78K variant (55), a predictive structure homology between LaVSP1 and R78K was investigated. According to the P1 and P2 phosphate-binding subsites, the SWISS-MODEL predicts a better structural homology between LaVSP1 and R78K (Fig. 9B) than between LaVSP1 and Y-PPase. In particular, the R78K and LaVSP1 resting enzymes favor the out position of the loop 71-78 (55). These changes are consistent with the reduced affinity for PP i of the R78K variant (K m ϭ 27 M) and LaVSP1 (K m ϭ 92 M) compared with that of the WT Y-PPase (K m ϭ 1.45 M). The need for LaVSP1 and R78K to overcome an unfavorable equilibrium constant in binding substrate could account for the observed increase in K m (55).  One set of data contained five BALB/C mice infected in both footpads with 10 6 late stationary phase promastigotes BA125 WT (black squares), Leishmania over expressing a native LaVSP1 (BA125/nLaVSP1, black diamonds), and Leishmania overexpressing a mutated LaVSP1 (BA125/mLaVSP1, black triangles). Lesion sizes were monitored weekly. The inset represents a Western blot showing the expression of recombinant His-tagged native and mutated LaVSP1 from protein lysates of 3 ϫ 10 6 cells, monoclonal antibodies directed against the poly-His tag (Qiagen) were used. IND, induced. FIGURE 9. Structural analysis. A, structures of the PP i binding site of the Y-PPase (gray) superimposed with that of R78Y-PPase (red) (55). B, structures of the PP i binding site of R78Y-PPase (red) superimposed with a predicted structure of the PP i binding sites of L. amazonensis (blue). C, structures of the PP i -binding site of R78Y-PPase (red) and Y-PPase (gray) superimposed with a predicted structure of the PP i -binding sites of L. amazonensis (blue) and T. cruzi (green) VSP1. Pi2 is represented in gray, and Pi1 is absent from the mutant structure (55). Metal ions are indicated by black dots. M3 is absent because it is the most labile metal ion in the wild type product complex (55).
Of course, to confirm that hypothesis it would have been interesting to develop variants. However nature often offers natural variants. In the case of VSP1 proteins, we analyzed the Arg 78 position for TbVSP1 and TcVSP1 (GenBank TM accession number DQ090844) proteins. The SWISS-MODEL predicted (Fig. 9C) a perfect structure homology between LaVSP1 and TbVSP1. However, TcVSP1 significantly differed from the two others at position Arg 78 , which is now completely oriented outside the catalytic center (Fig. 9C). Interestingly, TcVSP1 did not hydrolyze PP i but P 3 and polyP (data not shown), which correlates with TcVSP1 structural differences. In conclusion, to hydrolyze polyP the active site underwent significant distortion to accommodate the larger polyP sizes, suggesting that PP i and polyP are bound by VSP1 by different modes. As a matter of fact the Arg 78 residue was shown to be directly involved in binding and optimum orientation for catalysis of one phosphate group of PP i (11,55).
In addition, a mitochondrial family I pyrophosphatase was reported in L. major (LmsPPase) with a R78G mutation (56). The catalytic consequence was a Ca 2ϩ dependence of the pyrophosphatase activity, whereas Ca 2ϩ was known to be a very strong inhibitor of all family I pyrophosphatases (57), including the VSP1 subtype (3). This difference was attributed to the fact that the R78G mutation gave a different conformation change in the Gly 78 neighboring active site region after Ca 2ϩ binding, resulting in a more active enzyme (11).
Because VSP1 has been shown to be essential for virulence (3) and because VSP1 substrate specificity correlated with potential substrate binding structural characteristics, we consider this protein to be an attractive drug target. We recently published a study (58) about bisphosphonates (noncleavable pyrophosphate (P-O-P)) analogues in which the central oxygen is replaced by a carbon (P-C-P) with various side chains as inhibitors of the exopolyphosphatase activity of TbVSP1 and for their ability to protect mice against parasite infection.