Adenylosuccinate Synthetase and Adenylosuccinate Lyase Deficiencies Trigger Growth and Infectivity Deficits in Leishmania donovani*

Background: Purine salvage in Leishmania is an essential nutritional function. Results: Null mutants deficient in either adenylosuccinate synthetase or adenylosuccinate lyase impact growth and infectivity phenotypes of Leishmania donovani. Conclusion: Adenylosuccinate synthetase and adenylosuccinate lyase are central enzymes in purine salvage by L. donovani. Significance: Adenylosuccinate lyase has been validated as a potential drug target in L. donovani. Leishmania are auxotrophic for purines, and consequently purine acquisition from the host is a requisite nutritional function for the parasite. Both adenylosuccinate synthetase (ADSS) and adenylosuccinate lyase (ASL) have been identified as vital components of purine salvage in Leishmania donovani, and therefore Δadss and Δasl null mutants were constructed to test this hypothesis. Unlike wild type L. donovani, Δadss and Δasl parasites in culture exhibited a profoundly restricted growth phenotype in which the only permissive growth conditions were a 6-aminopurine source in the presence of 2′-deoxycoformycin, an inhibitor of adenine aminohydrolase activity. Although both knock-outs showed a diminished capacity to infect murine peritoneal macrophages, only the Δasl null mutant was profoundly incapacitated in its ability to infect mice. The enormous discrepancy in parasite loads observed in livers and spleens from mice infected with either Δadss or Δasl parasites can be explained by selective accumulation of adenylosuccinate in the Δasl knock-out and consequent starvation for guanylate nucleotides. Genetic complementation of a Δasl lesion in Escherichia coli implied that the L. donovani ASL could also recognize 5-aminoimidazole-(N-succinylocarboxamide) ribotide as a substrate, and purified recombinant ASL displayed an apparent Km of ∼24 μm for adenylosuccinate. Unlike many components of the purine salvage pathway of L. donovani, both ASL and ADSS are cytosolic enzymes. Overall, these data underscore the paramount importance of ASL to purine salvage by both life cycle stages of L. donovani and authenticate ASL as a potential drug target in Leishmania.

Leishmania are auxotrophic for purines, and consequently purine acquisition from the host is a requisite nutritional function for the parasite. Both adenylosuccinate synthetase (ADSS) and adenylosuccinate lyase (ASL) have been identified as vital components of purine salvage in Leishmania donovani, and therefore ⌬adss and ⌬asl null mutants were constructed to test this hypothesis. Unlike wild type L. donovani, ⌬adss and ⌬asl parasites in culture exhibited a profoundly restricted growth phenotype in which the only permissive growth conditions were a 6-aminopurine source in the presence of 2-deoxycoformycin, an inhibitor of adenine aminohydrolase activity. Although both knock-outs showed a diminished capacity to infect murine peritoneal macrophages, only the ⌬asl null mutant was profoundly incapacitated in its ability to infect mice. The enormous discrepancy in parasite loads observed in livers and spleens from mice infected with either ⌬adss or ⌬asl parasites can be explained by selective accumulation of adenylosuccinate in the ⌬asl knock-out and consequent starvation for guanylate nucleotides. Genetic complementation of a ⌬asl lesion in Escherichia coli implied that the L. donovani ASL could also recognize 5-aminoimidazole-(N-succinylocarboxamide) ribotide as a substrate, and purified recombinant ASL displayed an apparent K m of ϳ24 M for adenylosuccinate. Unlike many components of the purine salvage pathway of L. donovani, both ASL and ADSS are cytosolic enzymes. Overall, these data underscore the paramount importance of ASL to purine salvage by both life cycle stages of L. donovani and authenticate ASL as a potential drug target in Leishmania.
Protozoan parasites that infect humans constitute a phylogenetically diverse assortment of organisms that cause a variety of devastating and often fatal diseases in humans and their domestic animals. Indeed, afflictions of parasitic etiology represent some of the most consequential diseases worldwide in terms of human suffering and economic spoliation. Leishmania don-ovani is the causal agent of visceral leishmaniasis, a disease that is invariably lethal if untreated. L. donovani, like all Leishmania species, is a digenetic protozoan parasite that is present as the flagellated, extracellular promastigote in the phlebotomine sandfly vector and as the immotile, intracellular amastigote within phagolysosomes of macrophages of the infected mammalian host. There is no effective vaccine against leishmaniasis, and the standard antileishmanial drugs are far from ideal. These drugs are toxic, expensive, and necessitate prolonged and invasive administrations, and furthermore drug resistance has rendered the chemotherapies too often ineffective (1)(2)(3)(4). Thus, the need for new drugs, as well as new drug targets, is acute.
The implementation of rational, selective, and effective antiparasitic drug therapies relies upon the exploitation of underlying biochemical and/or metabolic disparities between the parasite and mammals. Among the most conspicuous of the metabolic discrepancies between Leishmania and its mammalian hosts is the pathway by which purine nucleotides are synthesized. Whereas mammals are prototrophic for purines and synthesize purine nucleotides from amino acids and one-carbon fragments, Leishmania, like all protozoan parasites, cannot generate the purine ring de novo (5)(6)(7). Accordingly, each parasite genus expresses a distinctive complement of nutritionally indispensable purine salvage and interconversion enzymes that allow the parasite to scavenge host purines.
The purine pathway of Leishmania is complex, intertwined, and capable of assimilating effectively every purine nucleobase or nucleoside from the host or culture medium into the parasite nucleotide pool (5)(6)(7)(8). Leishmania express four enzymes that are capable of converting host or extracellular purines into nucleotides: 1) hypoxanthine-guanine phosphoribosyltransferase (HGPRT) 2 ; 2) xanthine phosphoribosyltransferase (XPRT); 3) adenine phosphoribosyltransferase (APRT); 4) and adenosine kinase (2,5,9,10). HGPRT and XPRT are confined within the glycosome (11,12), a membrane-bound microbody organelle that is found exclusively among trypanosomatid parasites (13)(14)(15), whereas APRT has been definitively localized to the cytosol (12). Genetic studies in L. donovani reveal that none of the four enzymes by itself is essential for parasite survival because promastigotes deficient in the activity of any one of the four purine salvage enzymes are viable and do not exhibit a fitness deficit (16 -20). However, the construction and phenotypic characterization of a conditionally lethal ⌬hgprt/⌬xprt double mutant offers robust genetic verification for the conjecture that essentially all purine salvage by L. donovani is mediated through HGPRT or XPRT and that APRT and adenosine kinase are functionally superfluous. Whereas wild type L. donovani can grow in virtually any purine nucleobase or nucleoside as the sole purine source, the ⌬hgprt/⌬xprt double knock-out can grow only in adenine or adenosine and only when adenine aminohydrolase (AAH) is pharmacologically obstructed with 2Ј-deoxycoformycin (dCF) (21). Furthermore, the ⌬hgprt/ ⌬xprt null mutant is highly compromised in its ability to establish a visceral infection in mice (21), implying a central role for HGPRT and XPRT in purine salvage by amastigotes as well. These findings demonstrate, therefore, that virtually all host or extracellular purines are funneled into substrates of HGPRT or XPRT and strongly intimate the functional importance of the downstream nucleotide interconversion enzymes that distribute the enzymatic products of HGPRT and XPRT into adenylate and guanylate nucleotides. These nucleotide interconversion enzymes include adenylosuccinate synthetase (ADSS) and adenylosuccinate lyase (ASL), which convert IMP to AMP, IMP dehydrogenase (IMPDH) and GMP synthase, which generate GMP from IMP, and AMP deaminase and GMP reductase, which back-convert AMP and GMP into IMP, respectively (5,7,8,22) (see Fig. 1). A battery of nucleotide kinases then converts the nucleoside monophosphates to diphosphates and triphosphates (6).
ADSS catalyzes the GTP-dependent formation of adenylosuccinate from IMP and aspartic acid, whereas ASL cleaves a fumarate molecule from the adenylosuccinate product of the ADSS reaction. The native L. donovani ADSS and ASL enzymes have been partially purified from intact parasites and characterized kinetically with respect to their unusual capacities to metabolize nucleotide products of the noted antileishmanial pyrazolopyrimidine nucleobase and nucleoside analogs, e.g. allopurinol, 4-thiopurinol, and formycin B (23)(24)(25). The functional importance of either enzyme, however, to either the promastigote or the amastigote was heretofore unknown. To investigate the function of ADSS and ASL in L. donovani, ⌬adss and ⌬asl knock-outs were created by double targeted gene replacement protocols, and the growth and infectivity phenotypes of the knock-out lines were assessed. Both ⌬adss and ⌬asl parasites exhibited the same constrained growth phenotype as the previously characterized ⌬hgprt/⌬xprt double mutant (21), i.e. the only permissive growth conditions required the presence of dCF and either adenine or adenosine as the purine source. The infectivity phenotypes between the ⌬adss and ⌬asl null lines in mice were diametric opposites, however, with the former causing a robust visceral infection, whereas the parasite burdens in mice infected with the ⌬asl knock-out were greatly reduced. These infectivity data validate ASL as a promising therapeutic target for the treatment of leishmaniasis. Recombinant ASL was purified and shown to be catalytically active, and the gene complemented asl-deficient Escherichia coli in a manner that suggested that it was a bifunctional enzyme, like its human counterpart, which also could catalyze the cleavage of 5-aminoimidazole-(N-succinylocarboxamide) ribotide (SAICAR) to 5-aminoimidazole-4-carboxamide ribotide and fumarate (26). Both ADSS and ASL were localized to the parasite cytosol using immunocytochemistry and/or cell fractionation, implying a potentially intriguing portioning of the adenylate and guanylate branches of the L. donovani purine salvage pathway. This investigation is the first to validate a single component of the purine salvage pathway of Leishmania as a prospective antileishmanial drug target.

EXPERIMENTAL PROCEDURES
Materials, Chemicals, and Reagents- [8-14 C]Hypoxanthine (51 mCi/mmol) was bought from Moravek Biochemicals (Brea, CA), and [␣-32 P]dCTP was acquired from MP Biomedicals (Irvine, CA). Unlabeled purine bases, nucleosides, and nucleotides were procured from Sigma-Aldrich or Fisher Scientific, adenylosuccinate was purchased from Sigma-Aldrich, and dCF was obtained from Tocris Bioscience (Ellisville, MO). Restriction enzymes were purchased from New England Biolabs (Beverley, MA). The TOPO TA Cloning kit, pCR2.1-TOPO vector, Champion TM pET200/D-TOPO expression vector, and BL21 Star TM (DE3) One Shot TM competent cells, as well as the goat anti-rabbit Oregon Green-, goat anti-mouse Oregon Green-, goat anti-mouse Alexa Fluor 568-, and goat anti-guinea pig rhodamine red-conjugated secondary fluorescent antibodies, were bought from Invitrogen. IMPDH antiserum was produced in guinea pigs as reported (27), and mouse monoclonal anti-␣-tubulin (DM1A) antibody was obtained from EMD Millipore (Billerica, MA). Mouse monoclonal antihuman influenza hemagglutinin (HA) and mouse monoclonal (JL-8) anti-GFP primary antibodies were bought from Sigma-Aldrich and BD Biosciences, respectively. Goat anti-rabbit HRP-conjugated and goat anti-mouse HRP-conjugated sec-  ondary antibodies were purchased from Thermo Fisher Scientific. Complete Mini EDTA-free protease inhibitor tablets were procured from Roche Diagnostics, the nickel-nitriloacetic acid (Ni 2ϩ -NTA)-agarose beads were from Qiagen (Valencia, CA), and the Biosafe TM Coomassie and Bio-Rad protein assay kits were from Bio-Rad. The pX63-NEO, pX63-HYG, and pXG-BSD leishmanial expression vectors harboring the neomycin resistance (NEO), hygromycin phosphotransferase (HYG), and blasticidin deaminase (BSD) markers (28,29) were provided by Dr. Stephen Beverley (Washington University, St. Louis, MO). The pLCNEO-EGFPCO vector, 3 which contains a Leishmaniabiased codon-optimized version of the enhanced GFP (EGFP) sequence (30) and produces a gene product with an NH 2 -terminal EGFP tag, was supplied by Drs. Nicola Carter and Phillip Yates (Oregon Health and Science University, Portland, OR). All other chemicals and reagents were of the highest quality commercially available.
Axenic Parasite Culture-The 1S2D (31, 32) strain of wild type L. donovani, which originated with Dr. Dennis Dwyer (National Institutes of Health), was adapted for growth as axenic amastigotes as reported (33). The LdBob cell line cloned from the adapted 1S2D parasites was provided by Dr. Stephen Beverley (34). LdBob promastigotes were cultured at ambient temperature in modified Dulbecco's modified Eagle medium-Leishmania (35) as described (16) and supplemented with 5% FBS, whereas axenic amastigotes were propagated at 37°C as detailed (33,34). Plating methods and single-cell cloning protocols for L. donovani promastigotes are outlined elsewhere (35,36). Wild type and ⌬aah parasites were propagated in 100 M xanthine as the purine nutrient, the ⌬aah/⌬adss, ⌬adss[pADSS], and ⌬asl[pASL] cells were grown in 100 M adenine, and the ⌬adss and ⌬asl knock-outs were grown in modified Dulbecco's modified Eagle medium-Leishmania supplemented with 100 M adenine and 20 M dCF. All cell lines were maintained routinely as both promastigotes and axenic amastigotes (34) with their respective purine additions.
Cloning of ADSS and ASL from L. donovani-The ADSS and ASL genes were isolated from an L. donovani cosmid library that had been generated in the Supercos 1 cosmid vector (Stratagene, La Jolla, CA) (37) using the following strategy. Two putative coding sequences, one for ADSS (LmjF13.1190) and one for ASL (LmjF04.0460), were retrieved from a search of the Leishmania major genome (38). Oligonucleotide primers were designed against the L. major putative ADSS and ASL ORFs and utilized to amplify gene fragments from L. major genomic DNA via standard PCR methodology (39,40). The ADSS and ASL PCR fragments were cloned into TOPO TA and sequenced to confirm their identities. The PCR fragments were then used to probe independent platings of the L. donovani cosmid library under stringent conditions. Several colonies containing either the L. donovani ADSS or ASL gene were isolated, and the ADSS and ASL ORFs and flanking regions were sequenced in both directions from purified cosmid DNA. Multisequence and pairwise sequence alignments of primary structures were accom-plished using algorithms within the Vector NTI suite 7.1 software (Invitrogen).
Identification of the ADSS mRNA Spliced Leader Addition Site Using Rapid Amplification of cDNA Ends PCR-First strand ADSS cDNA was generated from 2 g of L. donovani total RNA via reverse transcription with Superscript III (Invitrogen) polymerase at 50°C according to the manufacturer's protocol using the ADSS gene-specific primer listed in supplemental Table S1. A portion of the cDNA reaction was then subjected to PCR using a sense primer corresponding to the spliced leader sequence, which is trans-spliced onto the 5Ј-termini of all Leishmania mRNAs (41), and the nested antisense primer listed in supplemental Table S1. The resulting PCR products were gel-purified and sequenced to identify the position of the spliced leader addition site on the ADSS mRNA relative to the predicted translation start site.
Generation of Targeting Constructs-To generate the pX63-NEO-⌬adss and pX63-HYG-⌬adss gene targeting vectors, ϳ850 and 950 bases of ADSS 5Ј-and 3Ј-UTRs were amplified from a purified ADSS cosmid by PCR, subcloned into the pCR2.1-TOPO vector (Invitrogen), sequenced to ensure fidelity, and inserted into the appropriate restriction sites of the pX63-NEO and pX63-HYG vectors (28). The primers used to amplify the ADSS UTRs are shown in supplemental Table S1. Similarly, the pX63-NEO-⌬asl and pX63-HYG-⌬asl targeting vectors were generated after ϳ900 and 1200 bases of the ASL 5Ј-and 3Ј-UTRs were PCR-amplified from a purified ASL cosmid, subcloned into the pCR2.1-TOPO vector, sequenced, and inserted into the appropriate restriction sites of the pX63-NEO or pX63-HYG vectors (29). The primers used to amplify the ASL UTRs are listed in supplemental Table S2.
Creation of Null Mutants-The ⌬adss and ⌬asl knock-outs were generated from wild type parasites by double targeted gene replacement using the transfection parameters and plating techniques reported previously (34,35). The linearized drug resistance cassettes enclosing the ADSS flanks were excised from pX63-NEO-⌬adss and pX63-HYG-⌬adss with HindIII-BglII, and the X63-NEO-⌬adss and X63-HYG-⌬adss targeting constructs were used to sequentially create ADSS/adss heterozygotes followed by ⌬adss null mutants using electroporation and plating methods as detailed (34,35). The ADSS/adss heterozygotes were isolated on semi-solid growth plates containing 20 g/ml Geneticin (G418) or 50 g/ml hygromycin supplemented with 100 M adenine, whereas the ⌬adss knockouts were selected on plates containing 20 g/ml G418, 50 g/ml hygromycin, 100 M adenine, and 20 M dCF. The ⌬aah/⌬adss double knock-out was generated within the ⌬aah parental strain (42) using linearized X63-NEO-⌬adss and X63-HYG-⌬adss targeting constructs. The ⌬aah/ADSS/adss heterozygotes and ⌬aah/⌬adss knock-outs were isolated on semi-solid medium containing selective concentrations of G418 and/or hygromycin, as appropriate for the selective marker, and supplemented with 100 M adenine in the absence of dCF.
Similarly, the X63-NEO-⌬asl and X63-HYG-⌬asl gene replacement constructs were employed to successively generate the ASL/asl heterozygotes followed by the construction of the ⌬asl knock-outs from their corresponding heterozygotes.
The heterozygotes were selected on plates containing suitable concentrations of G418 and/or hygromycin supplemented with 100 M adenine and 20 M dCF, and the ⌬asl knock-outs were selected on plates containing 20 g/ml G418, 50 g/ml hygromycin, 100 M adenine, 100 M guanine, and 20 M dCF. The ASL/asl heterozygote was created within the ⌬aah background to generate ⌬aah/ASL/asl parasites, which were then transfected with the second replacement construct in an attempt to generate a ⌬aah/⌬asl double knock-out. The putative ⌬aah/ ⌬asl parasites were selected in 20 g/ml G418, 50 g/ml hygromycin and supplemented with either 100 M adenine or a purine mixture consisting of 200 M adenine, 200 M guanine, and 200 M xanthine. The genotypes of all heterozygotes and knock-outs were verified using standard genomic DNA isolation and Southern blotting protocols (43). The hybridization probes harboring either the ADSS or ASL ORF or the 5Ј-or 3Ј-UTR were PCR-amplified from the ADSS and ASL cosmids, respectively, and gel-purified using a Wizard SV gel and PCR clean-up kit (Promega, Madison, WI).
Generation of Episomally Complemented Lines-The pXG-BSD vector (29) was used for episomal complementation of ⌬adss and ⌬asl parasites. The primers used to PCR-amplify the ADSS and ASL ORFs are shown in supplemental Tables S1 and S2, respectively. The coding sequence of each gene was subcloned independently into the pCR2.1-TOPO vector (Invitrogen), sequenced to ensure fidelity, excised from pCR2.1-TOPO with SmaI-BamHI, and inserted into the SmaI-BamHI sites of the pXG-BSD expression plasmid (29). The pXG-BSD-ADSS and pXG-BSD-ASL episomes were then used to generate ⌬adss[pADSS] and ⌬asl[pASL] "add-back" cell lines from the ⌬adss and ⌬asl knock-outs. The complemented cell lines were selected in 20 M blasticidin and supplemented with 100 M adenine. Southern blotting was performed to confirm the genotypes of the "add-backs" as described above.
Growth Phenotypes-To evaluate the capacity of wild type, ⌬adss, ⌬adss[pADSS], ⌬aah, ⌬aah/⌬adss, ⌬asl, and ⌬asl[pASL] promastigotes to grow in different purine sources, exponentially growing parasites were washed several times with PBS, resuspended at a density of 5 ϫ 10 4 cells/ml in 1.0-ml aliquots of modified Dulbecco's modified Eagle medium-Leishmania (16) containing 100 M purine and 5% dialyzed FBS, and dispensed into individual wells of 24-well tissue culture plates (Sarstedt Inc., Newton, NC). The ability of these lines to grow in either adenine or adenosine was also determined in the presence of dCF. In addition, the adenosine growth profile of the ⌬aah/ ⌬adss double knock-outs was also assessed in the presence of 20 M erythro-9-(2-hydroxy-3-nonyl)adenine (EHNA) (44) to inhibit the adenosine deaminase activity found in FBS (45), thereby preserving the nucleoside. After 7-10 days, parasites were enumerated visually by hemocytometer.
ADSS and ASL Activity Measurements in Intact Parasites-The ability of wild type, ⌬adss, and ⌬asl promastigotes to convert [8-14 C]hypoxanthine into various metabolites was assessed using TLC. 5.0 ϫ 10 7 parasites were washed and resuspended in 20 l of PBS; [8-14 C]hypoxanthine (51 mCi/mmol) was added to a final concentration of 100 M, and the cells were allowed to incubate at 26°C. At 30 min and 2 h, 10 l of the parasite mixture was centrifuged at 13,000 rpm in an Eppendorf Centri-fuge 5415D, the supernatant was removed, and the parasite pellet was resuspended in 5 l of glacial acetic acid to lyse the cells and terminate the reaction (46). The lysate was spotted onto a PE-SIL-G TLC plate with fluorescent indicator (GE Healthcare), so that nonradioactive standards could be detected, and developed in dioxane/ammonium hydroxide/water at a ratio of 6:1: (47). The TLC plate was exposed to X-ray film at Ϫ80°C and developed in a standard X-ray film developer.
Expression and Purification-The ASL and ADSS ORFs were amplified by PCR from the ASL and ADSS cosmids, respectively, and inserted separately into the pET200/D-TOPO E. coli expression plasmid, which automatically attaches a His 6 tag to the NH 2 terminus of the inserted gene product, and the constructs were sequenced to confirm the gene sequences. The pET200/D-TOPO-ASL and the pET200/D-TOPO-ADSS constructs were then independently transformed into BL21 Star TM (DE3) One Shot E. coli (Invitrogen). After induction in 0.5 mM IPTG, the recombinant L. donovani ASL protein was purified to virtual homogeneity over a Ni 2ϩ -NTA-agarose column from E. coli extracts that had been prepared using a French press as described (54). Recombinant ASL was eluted from the Ni 2ϩ -NTA-agarose with 250 mM imidazole as detailed (54). Separation of the purified recombinant ASL fractions on a 10% SDS-polyacrylamide gel and subsequent staining with Biosafe TM Coomassie (Bio-Rad) confirmed the purity of the recombinant protein. A Thermo Labsystems Multiskan Ascent plate reader (Thermo Scientific) was employed at 600 nm to determine the protein concentration and yield of the purified ASL after the addition of Bio-Rad protein assay reagent (54). Yields of purified recombinant ASL protein were ϳ25-30 mg/liter of bacterial culture. Efforts to purify recombinant ADSS under both native and denaturing conditions were also made.
Antibodies and Immunoblotting-Polyclonal ASL antiserum was generated in rabbits by Open Biosystems (Huntsville, AL) using purified recombinant ASL as an immunogen and standard injection protocols. Western blotting protocols were performed as detailed (43) using rabbit anti-ASL and mouse anti-␣-tubulin primary antibodies followed by goat anti-rabbit HRP-conjugated and goat anti-mouse HRP-conjugated secondary antibodies, respectively.
ASL Assay-The ability of purified, recombinant ASL to utilize adenylosuccinate as a substrate was assessed by a spectrophotometric assay in which the cleavage of adenylosuccinate to AMP and fumarate (24) was measured by a decrease in absorbance at 282 nm (24, 55) using a Beckman DU 640 spectrophotometer (Beckman Coulter). All kinetic studies with ASL were carried out under experimental conditions that were linear with time and enzyme concentration. Briefly, Ϸ1.25 g of recombinant ASL was added to a quartz cuvette containing 1.0 ml of 20 mM HEPES-KOH, pH 7.0, and adenylosuccinate vary-ing in concentration from 1 to 125 M. The reaction was carried out at 25°C and pH 7.0 over a 90-s time course, and absorbance at 282 nm was measured. The amount of adenylosuccinate consumed was calculated from the A 282 reading at each time point. The molar extinction coefficient (⑀) of adenylosuccinate was calculated experimentally to be 16,430 M Ϫ1 ⅐cm Ϫ1 under these reaction conditions. Michaelis-Menten analysis (56 ADSS Localization-PCR was employed to attach an HA tag onto either the 5Ј-or 3Ј-end of the ADSS ORF. SmaI and BamHI restriction sites (supplemental Table S1) were included on the HA-ADSS constructs, and the DNA fragments were ligated into the pXG-BSD leishmanial expression vector (29) to generate pXG-BSD-5ЈHA-ADSS and pXG-BSD-3ЈHA-ADSS plasmids, allowing the parasites to produce ADSS protein that was HA-tagged at either the NH 2 or COOH terminus. The plasmids were transfected separately into wild type L. donovani, and immunofluorescence assays were performed using polyclonal mouse anti-HA primary antibody followed by goat-anti mouse Oregon Green-(1:10,000) or goat anti-mouse Alexa Fluor 568-conjugated (1:10,000) secondary antibody to visualize the HA-tagged ADSS. Parasites were co-stained with guinea pig anti-IMPDH antibodies (1:500) and goat anti-guinea pig rhodamine red-conjugated (1:10,000) secondary antibody to differentiate glycosomal IMPDH. Cells were visualized on a Zeiss Axiovert 200 inverted microscope (Carl Zeiss Microimaging, Thornwood, NY) with a ϫ63 oil immersion lens. Photos were taken with a Zeiss AxioCam MR camera using Axiovision 4.2 software and compiled using Adobe Photoshop Creative Suite 4.
ASL Localization by Subcellular Fractionation-The separation of glycosomal and cytosolic fractions from 3-4 ϫ 10 9 L. donovani promastigotes was performed as detailed (42). A linear 20 -70% sucrose gradient was prepared as described (43) and allowed to equilibrate overnight at 4°C. Post-nuclear whole cell lysates were overlaid on the sucrose gradient and centrifuged for 6 h at 36,000 rpm in a Beckman SW41 rotor. The gradient was fractionated, and the proteins were precipitated with trichloroacetic acid and analyzed by standard Western blotting procedures (43) using polyclonal ASL and PEX14 (57) antisera.
ASL Localization by Microscopy-Purified ASL protein was bound to AminoLink coupling resin (Thermo Scientific), and crude ASL antiserum was purified against ASL protein using an AminoLink Immobilization kit (Thermo Scientific) according to the manufacturer's instructions. The immunofluorescence assay was performed on L. donovani promastigotes as described (11,12,27,48) using a 1:1000 dilution of recombinant anti-ASL antibody and a 1:10,000 dilution of secondary goat anti-rabbit Oregon Green-conjugated antibody (Invitrogen).
The ASL ORF was amplified by PCR from the ASL cosmid using primers that attached appropriate 5Ј-and 3Ј-restriction sites (supplemental Table S2), subcloned into the pCR2.1-TOPO vector, excised via an NdeI-EcoRV restriction digest, and ligated into the pLCNEO-EGFPCO vector to create pLCNEO-EFGPCO-ASL. pLCNEO-EFGPCO-ASL was then transfected into wild type L. donovani promastigotes, the transfectants were selected in 20 M G418, and EFGP-tagged ASL protein was visualized as described above.

RESULTS
Primary Structures of L. donovani ADSS and ASL-The L. donovani ADSS (ABS11225) and ASL (ABS11226) genes and their flanking sequences were isolated from purified cosmid clones that hybridized to the putative ADSS (LmjF.13.1190) and ASL (LmjF.04.0460) ORFs, respectively, from L. major (38). The L. donovani ADSS ORF predicted a 710-amino acid protein with a molecular mass of ϳ78.2 kDa that was 93-99% identical to ADSSs from L. major, Leishmania mexicana, and Leishmania infantum and 58, 61, 16, 17, 16, and 17% identical to the ADSS sequences from Trypanosoma brucei, Trypanosoma cruzi, Plasmodium falciparum, Homo sapiens, Saccharomyces cerevisiae, and E. coli, respectively. A multiple sequence alignment among phylogenetically diverse ADSS proteins revealed that the Leishmania ADSS ORFs encompass a sizeable NH 2terminal extension (38) as well as two internal insertions of 20 and 55 residues, which render the leishmanial ADSSs much longer than ADSSs from other organisms (supplemental Fig.  S1A). A portion of the NH 2 -terminal extension, as well as the entirety of both internal insertions, is also found in the putative ADSSs from T. brucei and T. cruzi, two protozoan pathogens that are evolutionarily related to Leishmania, but not in ADSSs from other organisms (supplemental Fig. S1A). Because the length of the predicted Leishmania ssp. ADSS ORF created some uncertainty about the translation initiation site, a modified rapid amplification of cDNA ends technique was employed to determine whether the mRNA start site of ADSS was 5Ј to the predicted ATG start site. Two spliced leader addition sites were identified (data not shown) upstream of the predicted ATG start (203 and 365 bp), with no downstream sites detected. This is consistent with the predominant spliced leader addition sites subsequently reported on TriTrypDB and indicates that the predicted initiating ATG is, in fact, present on the L. donovani ADSS mRNA. These data corroborate the observation by Spector et al. (24) that the partially purified L. donovani ADSS protein appeared to have a larger particle weight than other sources of ADSS. Additional analysis of the L. donovani ADSS amino acid sequence revealed that this protein possesses many residues that are found within the active site of both the P. falciparum and E. coli ADSS crystal structures (59,60).
The L. donovani ASL encodes a protein of 479 amino acids with a molecular mass of ϳ53.7 kDa. Pairwise alignments of the L. donovani ASL with ASLs from other organisms (supplemental Fig. S1B) demonstrated 93-100% identity to ASLs from other Leishmania species (38) and 67, 71, 39, 16, 18, and 49% amino acid identity to the predicted ASL sequences from T. brucei, T. cruzi, P. falciparum, H. sapiens, S. cerevisiae, and E. coli, respectively. A multiple sequence alignment of L. donovani ASL with ASLs from other organisms revealed several residues that are conserved among all of the aligned ASL proteins (supplemental Fig. S1B).
Confirmation of ⌬adss, ⌬aah/⌬adss, and ⌬asl Genotypes-To test the functionality of ADSS and ASL in intact parasites, ⌬adss and ⌬asl null mutants were created by targeted gene replacement from wild type L. donovani as described under "Experimental Procedures." It was necessary to include dCF in the second round of transfection for both knock-outs to prevent deamination of adenine and compel its incorporation into the parasite nucleotide pool through APRT. A ⌬aah/⌬adss double knock-out was also created within the ⌬aah background in order to eliminate the pharmacological requirement for dCF and to demonstrate that AAH activity is critical to the conditionally lethal ⌬adss mutation (42). Because dCF is a specific inhibitor of AAH (42, 61), dCF was not included in the selection of ⌬aah/⌬adss parasites from the ⌬aah/ADSS/adss progenitor. Efforts to create a ⌬aah/⌬asl double knock-out from a ⌬aah/ASL/asl heterozygote were unsuccessful, even after supplementation of the culture medium with a purine mixture consisting of adenine, guanine, and xanthine that should theoretically have supplied precursors for both adenylate and guanylate nucleotides.
Southern blots of genomic DNA from wild type, ADSS/adss, ⌬adss, ⌬adss[pADSS], ⌬aah/ADSS/adss, and ⌬aah/⌬adss parasites using either the ADSS ORF or 5Ј-UTR as a hybridization probe verified the allelic replacements in the heterozygotes and null mutants (Fig. 2). The hybridization signals from the genomic DNA samples prepared from wild type and genetically manipulated parasites corresponded to the expected restriction fragments for each cell line and confirmed the specific gene rearrangements at the ADSS locus in ADSS/adss and ⌬adss parasites (Fig. 2). Western blot analysis was not carried out for the transgenic strains harboring rearrangements at the ADSS locus because of the lack of specific anti-ADSS antisera.
The ⌬asl genotype was also confirmed by Southern analysis. Southern blots of genomic DNA from wild type, ASL/asl, ⌬asl, and ⌬asl[pASL] parasites using either the ASL ORF or 3Ј-UTR as a hybridization probe confirmed the allelic substitutions (Fig.  3, A and B). The hybridization signals from the genomic DNA samples were the expected size of restriction fragments predicted from mapping of the wild type ASL locus. Additionally, the specific allelic rearrangements observed in the genetically manipulated transgenic parasites confirmed the null genotype of the ⌬asl knock-out clone (Fig. 3, A and B). Western blotting with polyclonal anti-ASL antibodies verified the specificity of the ASL antisera, the absence of ASL protein in the ⌬asl mutant, and the presence of ASL in wild type and ⌬asl[pASL] parasites (Fig. 3C).
Growth Phenotypes-The impact of ⌬adss and ⌬asl lesions on the nutritional phenotype of L. donovani promastigotes was then evaluated by ascertaining the capabilities of wild type, null mutant, and add-back parasites to grow in a variety of different purines (Fig. 4A). Whereas wild type and ⌬adss[pADSS] pro-mastigotes could proliferate in medium supplemented with any of the purine nucleobases or nucleosides tested, the ⌬adss null mutant could not grow on guanine, guanosine, hypoxanthine, inosine, or xanthine as the sole purine source (Fig. 4A). ⌬adss parasites could propagate only in medium supplemented with adenine or adenosine and only in the presence of 20 M dCF (Fig. 4A). Because the ⌬aah/⌬adss double knock-out could not deaminate adenine to hypoxanthine, this strain was able, as expected, to grow in adenine in the absence of 20 M dCF, although all of the 6-oxypurines were nonpermissive for growth. Interestingly, and somewhat unexpectedly, the ⌬aah/ ⌬adss could not salvage adenosine alone, although the parental ⌬aah null line from which the ⌬aah/⌬adss double mutant was constructed could propagate in all purines tested (Fig. 4A). This finding implied that extracellular adenosine in the culture medium was not converted to adenine by parasite nucleoside hydrolase activities but, rather, that adenosine was being deaminated to inosine by mammalian adenosine deaminase, a component of the serum-based medium (45) in which promastigotes are propagated, prior to being taken up by the parasite. To test this conjecture, the ability of all strains to salvage adenosine was also evaluated in the presence of EHNA, a specific inhibitor of mammalian adenosine deaminase (44). Wild type, ⌬adss[pADSS], ⌬aah, and ⌬aah/⌬adss promastigotes, but not ⌬adss cells, grew in adenosine plus EHNA (Fig. 4A), demonstrating both that extracellular adenosine is deaminated to inosine by serum adenosine deaminase and that EHNA is not an inhibitor of the leishmanial AAH.
The growth phenotypes of ⌬asl and ⌬asl[pASL] parasites were similarly assessed. Restrictive and permissive growth conditions for the ⌬asl knock-out were comparable to those of the ⌬adss parasites. Predictably, wild type and ⌬asl[pASL] promastigotes could proliferate in medium supplemented with any of the purine nucleobases or nucleosides appraised, whereas the The absence of a hybridization signal in genomic DNA from ⌬adss and ⌬aah/ ⌬adss parasites (lanes 4 and 9) confirmed the null genotype of these cell lines. B, the 1650-bp hybridization signal produced from genomic DNA that had been cut with HindIII/BglII/EcoRI and probed with the AAH ORF (lanes 1-5) encompasses 765 bp of the AAH ORF and 885 bp of the AAH 5Ј-UTR. The absence of an AAH hybridization signal in DNA from parasites containing the ⌬aah lesion (lanes 6 -9) verified the genotype of these cell lines. ⌬asl null mutant could not grow with guanine, guanosine, hypoxanthine, inosine, or xanthine as the purine in the medium (Fig. 4B). The only permissive growth conditions for the ⌬asl line were a 6-aminopurine source, either adenine or adenosine, in medium to which 20 M dCF was added. ADSS and ASL Activity Measurements in Vivo-To demonstrate that ADSS and ASL are functional in intact parasites, the capacities of wild type and null mutant L. donovani promastigotes to metabolize [ 14 C]hypoxanthine into various products was determined. Wild type parasites converted [ 14 C]hypoxanthine into numerous detectable products, including di-and triphosphorylated nucleotides, adenylosuccinate, IMP, AMP, and nucleosides (Fig. 5). In contrast, ⌬asl parasites amassed high levels of adenylosuccinate, synthesized no AMP, and produced diminished amounts of total nucleotides (Fig. 5 and Table 1). Interestingly, a band that migrates above the nucleosides but below hypoxanthine (indicated by an arrow in Fig. 5; see also Table 1) appeared after TLC separation of the ⌬asl cellular metabolites. The nature of this band is unclear, but its location in the chromatogram suggests that it is a nonphosphorylated product, perhaps dephosphorylated adenylosuccinate. As expected, ⌬adss parasites did not metabolize [ 14 C]hypoxanthine into any detectable amounts of adenylosuccinate or AMP, and the majority of the radiolabeled substrate was converted to IMP as well as other nucleotides (Fig. 5). The extra spot that accumulates in ⌬asl cells is absent in the ⌬adss lanes of the chromatogram. These data confirm the biochemical activities of ADSS and ASL in L. donovani promastigotes and the metabolic deficits in metabolism incurred by the ⌬adss and ⌬asl genetic lesions.
Macrophage Infections-To ascertain whether a genetic deficiency in either ⌬adss or ⌬asl compromised the ability of L. donovani to infect host cells, infectivity assays with stationary phase wild type, knock-out, and add-back promastigotes were performed in murine peritoneal macrophages. Whereas wild type, ⌬adss[pADSS], and ⌬asl[pASL] parasites were capable of infecting and sustaining robust infections in the macrophages, the parasite loads in macrophages infected with ⌬adss and ⌬asl parasites were 4.2 and 19.8%, respectively, relative to what was obtained for the wild type parasite infections (Fig. 6A).  . Hypoxanthine metabolism in wild type, ⌬asl, and ⌬adss parasites. Intact wild type, ⌬asl, and ⌬adss parasites were evaluated for their ability to metabolize [ 14 C]hypoxanthine. Samples were collected at 30 and 120 min, and the metabolites were separated using the TLC methodology described under "Experimental Procedures." Lysate from a total of 2.5 ϫ 10 7 parasites was spotted in each lane. The unused metabolites can be seen at the top of the TLC radiographs, and the products are separated below. Standards are shown to the left of the chromatograms. AMPS, adenylosuccinate; NS, nucleosides; and Hyp, hypoxanthine. The unusual metabolite produced in the ⌬asl parasites is indicated by a black arrow. The relative percentages of each product are quantified in Table 1.

TABLE 1 Percentage of products produced by each cell line when given [ 14 C]hypoxanthine
Values are expressed as a percentage of the total amount of products produced by each cell line at 30 and 120 min. The products are listed in the left-most column. Hyp, hypoxanthine; AMPS, adenylosuccinate.

Products
Wild type ⌬asl ⌬adss Mouse Infections-Because both ⌬adss and ⌬asl parasites exhibited markedly reduced parasite loads in macrophages, the ability of the null mutants and their respective complements to infect BALB/c mice, a well characterized rodent model for L. donovani infections (48,49,(62)(63)(64), was evaluated. Unexpectedly, despite the reduced parasite numbers observed in peritoneal macrophages infected with ⌬adss parasites (Fig. 6A), the inoculation of BALB/c mice with stationary phase ⌬adss promastigotes resulted in a robust infection in which parasite loads were equivalent to those obtained from mice injected with wild type parasites (Fig. 6B). The parasite burdens in livers and spleens from mice infected with the ⌬adss[pADSS] addback were predictably equivalent to those obtained from mice inoculated with wild type parasites (Fig. 6B). The ⌬aah/⌬adss double knock-out was also evaluated for its ability to establish an infection in mice. Parasite burdens were reduced by 1 to 2 orders of magnitude in livers but were equivalent to wild type levels in the spleens (Fig. 6B). Parasite loads in the livers and spleens of mice infected with the ⌬asl knock-out were 4 and 3.5 orders of magnitude lower, respectively, relative to the parasite loads obtained from mice inoculated with wild type L. donovani (Fig. 6C). Complementation of the genetic defect in the ⌬asl[pASL] add-back cell line essentially restored parasite burdens in livers and spleens to wild type levels (Fig. 6C).
Bacterial Complementation-To confirm that the L. donovani ASL encodes a functional ASL protein, the leishmanial ASL was used to complement the asl deficiency of S⌽200 E. coli (26). The S⌽200 bacteria are incapable of de novo purine synthesis because they lack the bifunctional E. coli enzyme that catalyzes both the adenylosuccinate and SAICAR lyase activities. SAICAR lyase is an early step in the purine biosynthetic pathway that occurs prior to the synthesis of IMP in purine prototrophic organisms (26). S⌽200 cells are, therefore, auxotrophic for purines but can be propagated in minimal medium containing both adenine and hypoxanthine, with the former necessary to bypass ASL to synthesize adenylate nucleotides and the latter to circumvent SAICAR lyase to make guanylate nucleotides. S⌽200 cells, however, cannot grow in either adenine or hypoxanthine alone (26). S⌽200 E. coli transformed with pBAce-ASL were capable of growing on minimal medium plates lacking purine or supplemented with adenine or hypoxanthine, whereas cells transformed with pBAce alone were only capable of growth on minimal medium when provided with both adenine and hypoxanthine (Fig. 7). In contrast, transformation of the ADSS chimeric construct pBAce-ADSS into a variety of adss-deficient E. coli strains harboring a purA mutation, including ES4, PC1523, JW4135-2, TX595, and H1238, did not complement the purine auxotrophy conferred by the adss lesion. Protein Expression, Purification, and Kinetic Analysis-Recombinant His 6 -tagged L. donovani ASL was purified to essential homogeneity over a Ni 2ϩ -NTA column and its catalytic activity assessed spectrophotometrically by monitoring the conversion of adenylosuccinate to AMP. ASL activity was first tested over a range of pH values and was found to be optimal at pH 7.0 (data not shown). ASL kinetics were then determined by varying the adenylosuccinate concentration at pH 7.0, the pH at which ASL displayed optimal activity. Michaelis-Menten analysis (56) (Fig. 8) revealed that the apparent K m of ASL for adenylosuccinate was ϳ24 M with a V max of 2.1 mol⅐min Ϫ1 ⅐mg protein Ϫ1 . A k cat value of 1680 min Ϫ1 was then calculated from the kinetic data (Fig. 8).
Attempts to purify sufficient, recombinant, His 6 -tagged L. donovani ADSS using a Ni 2ϩ -NTA column were unproductive. Although ADSS production from pET200/D-TOPO-ADSS was robust, virtually all of the recombinant ADSS protein was insoluble and not amenable to kinetic analysis, despite a number of induction and purification schemes that were implemented. The presence of ADSS as inclusion bodies could account for the failure of the pBAce-ADSS vector to complement multiple adss-deficient E. coli strains.
ADSS and ASL Localization-ADSS was localized to the cytosol by immunofluorescence of L. donovani parasites that expressed either NH 2 -or COOH-terminal HA-tagged ADSS using anti-HA primary and Oregon Green-or Alexa Fluor 568-conjugated secondary antibodies (Fig. 9A). The cytosolic location of ADSS was also supported by the lack of colocalization with IMPDH, a known glycosomal protein (27) (data not shown). Western blot analysis of whole cell lysates from 5Ј-HA-ADSS-and 3Ј-HA-ADSS-expressing parasites ensured that the HA tag was not cleaved from ADSS and that an HA-ADSS product of the correct molecular weight was produced in these parasites (data not shown).
ASL was localized to the cytosol using a variety of techniques including subcellular fractionation, immunofluorescence assay, and direct fluorescence. Separation of L. donovani lysates into organellar and cytosolic components by differential centrifugation and subsequent immunoblotting revealed that ASL was localized almost entirely to the cytosolic compartment (Fig.  9B). The differential centrifugation protocol that was employed has been shown previously to yield intact glycosomes (65,66). Further fractionation of the post-nuclear cell lysates on a sucrose gradient revealed that PEX14, a known glycosomal marker, cosedimented with the glycosomes (lower portion of the gradient, fractions 15-28) (57), whereas ASL could be detected exclusively at the top of the sucrose gradient (fractions [1][2][3][4][5][6][7][8][9][10][11][12][13][14] with other cytosolic enzymes ( Fig. 9C and data not shown). An immunofluorescence assay validated the cytosolic context of ASL in intact wild type and ⌬asl[pASL] promastigotes and confirmed the absence of ASL in ⌬asl parasites (Fig.  9D). The cytosolic location of ASL was also corroborated by direct fluorescence in wild type L. donovani that had been transfected with pLCNEO-EGFPCO-ASL (Fig. 9D). The cytosolic milieu for ASL was further reinforced by the lack of colocalization with IMPDH, a glycosomal enzyme (27) (data not shown).

DISCUSSION
The striking non-infectious phenotype of the ⌬hgprt/⌬xprt double knock-out implicated the downstream nucleotide interconversion enzymes ADSS, ASL, IMPDH, and GMP synthase as crucial bottlenecks for purine metabolism by L. donovani and, therefore, potential therapeutic targets. To test this hypothesis, ⌬adss and ⌬asl mutants were generated by double targeted gene replacement and their growth and infectivity phenotypes determined. As predicted, both null strains exhibited a strikingly restricted growth phenotype that was virtually identical to that of the previously characterized ⌬hgprt/⌬xprt cell line. The only permissive growth conditions for ⌬adss and ⌬asl promastigotes were a 6-aminopurine, either adenine or adenosine, as the purine source, and dCF, the AAH inhibitor (Fig. 4). These growth studies clearly establish that both ADSS and ASL are essential enzymes for the conversion of IMP to AMP. Episomal complementation of the null lines restored the wild type growth spectrum, authenticating that the nonpermissive growth conditions could be ascribed to the specific genetic lesions in the mutant lines and not to secondary genetic alteration(s) that may have been introduced through the genetic manipulations required for strain construction.
To assess the impacts of the ⌬adss and ⌬asl lesions on the viability of the amastigote stage of L. donovani, the abilities of  the null strains to establish infection were assessed in murine macrophages and in BALB/c mice. Intracellular parasite numbers were substantially reduced in peritoneal macrophages infected with either ⌬adss or ⌬asl parasites, and these infectivity deficits were eliminated in macrophages infected with the episomally complemented add-back lines (Fig. 6A). Interestingly, ⌬adss parasites retained their capacity to establish a robust infection in mouse livers and spleens with parasite loads that were equivalent to those obtained from mice infected with either wild type or ⌬adss[pADSS] parasites. The disparate capacity of ⌬adss parasites to infect macrophages in vitro and mice in vivo could be explained by discrepant purine availability in the two experimental models. The capability of ⌬adss L. donovani to establish visceral infection levels equivalent to those of wild type parasites (Fig. 6B) intimates that L. donovani amastigotes are able to scavenge purines from the mouse that fulfill both their adenylate and guanylate nucleotide requirements using ADSS-independent routes. Presumably, the guanylate nucleotide requirements of the ⌬adss parasites could be fulfilled by guanine or guanosine derived directly from host breakdown of guanylate nucleotides. Alternatively, hypoxanthine, a breakdown product of host nucleotide metabolism, can be converted to GMP by the successive actions of HGPRT or XPRT, IMPDH, and GMP synthase (see Fig. 1). To fulfill its adenylate nucleotide requirement, however, alternative metabolic routes using 6-aminopurine sources derived from the host must be utilized, because IMP to AMP conversion is blocked in the ⌬adss knock-out. These alternative routes include adenosine kinase and/or APRT (see Fig. 1). In contrast, mice inoculated with ⌬asl parasites displayed parasite loads in livers and spleens that were 4 and 3 orders of magnitude lower, respectively, than those in mice infected with wild type parasites. These reduced parasite burdens in mice infected with ⌬asl parasites were effectively restored to wild type levels by episomal complementation with ASL (Fig. 6C). Thus, the infectivity data in mice are analogous to those obtained in vitro with the peritoneal macrophages (Fig. 6, A and C) and bolster ASL as a possible target for therapeutic manipulation.
The discrepant parasite loads obtained from mice infected with L. donovani harboring either a ⌬adss or ⌬asl lesion was surprising but can be explained by their respective metabolic functions. When exposed to [ 14 C]hypoxanthine as the extracellular purine, ⌬adss promastigotes primarily accumulate [ 14 C]IMP, nucleosides, and other (presumably guanylate) nucleotides (Fig. 5). Conversely, the ⌬asl promastigotes accrue, as expected, adenylosuccinate and exhibit decreased di-and FIGURE 9. Localization of ADSS and ASL. A, L. donovani promastigotes harboring ether the pXG-5ЈHA-ADSS (I-III) or pXG-ADSS-3Ј-HA (IV-VI) episome were subjected to immunofluorescence analysis using mouse anti-HA antibodies, and goat anti-mouse Oregon Green-conjugated secondary antibody was used to detect the anti-HA antibodies. Parasites were also stained with DAPI (II and V) and phase contrast images are shown in panels III and VI. B, L. donovani promastigote cell lysates were fractionated by sedimentation at 45,000 ϫ g for 1 h at 4°C, and the total cytosolic (lane 1 (C)) and crude organellar (lane 2 (O)) fractions were subjected to Western blot analysis using anti-ASL antibody. C, post-nuclear whole cell lysate pellets were resolved by centrifugation on a linear sucrose density gradient. All 28 fractions collected from the sucrose gradient were then subjected to Western blot analysis using anti-ASL and anti-PEX14 antibodies. The gradient fractions are indicated on the top of the panels. D, fixed and permeabilized exponentially growing wild type (I-III), ⌬asl (IV-VI), or ⌬asl[pASL] (VII-IX) L. donovani promastigotes were incubated with rabbit anti-ASL antisera, and the primary antibodies were visualized with goat anti-rabbit Oregon Green-conjugated secondary antibody (I, IV, and VII). EGFP-tagged ASL was also visualized by direct fluorescence (X). DAPI staining is shown in panels II, V, VIII, and XI, and phase contrast images are illustrated in panels III, VI, IX, and XII. triphosphorylated nucleotide synthesis as compared with wild type parasites (Fig. 5). One can speculate, therefore, that ⌬asl L. donovani amastigotes also build up adenylosuccinate at the expense of nucleotide production; thus the adenylosuccinate serves as a metabolic "dead end" sink for ⌬asl parasites. Conversely, because adenylosuccinate synthesis is blocked in ⌬adss parasites, IMP is converted to guanylate nucleotides in both life cycle stages, whereas the amastigote adenylate pool is replenished from adenosine and/or adenine supplied by the host through the parasite adenosine kinase and/or APRT (Fig. 1). Thus, the ⌬adss knock-out can generate adenylate and guanylate nucleotides from host purines, whereas the ⌬asl amastigotes funnel host purines into adenylosuccinate via ADSS, thereby precluding sufficient nucleotide production in the parasite. Consistent with this scenario is the fact that we were unable to generate a ⌬aah/⌬asl double knock-out in promastigotes under any experimental condition, presumably because exogenous purines were shunted to adenylosuccinate, a metabolic dead end for L. donovani harboring a ⌬asl mutation.
Because the infectivity data bolstered ASL as a promising therapeutic target, the ASL enzyme was characterized further. Recombinant ASL was purified to effective homogeneity, kinetically and functionally characterized, and localized. The recombinant enzyme catalyzed the conversion of adenylosuccinate to AMP and fumarate with high affinity (Fig. 8), and the L. donovani ASL gene was capable of complementing an asl-deficient (purB) strain of E. coli (Fig. 7). Because the bacterial ASL catalyzes two discrete steps in the purine biosynthesis pathway of E. coli, the SAICAR lyase activity that precedes IMP formation as well as adenylosuccinate lyase (26), the ability of the L. donovani ASL to complement the bacterial purB lesion demonstrates that the parasite enzyme catalyzes both activities. The fact that the L. donovani ASL enzyme can cleave SAICAR to 5-aminoimidazole-4-carboxamide ribotide and fumarate, the antepenultimate step in IMP biosynthesis in purine prototrophs, is a bit of a curiosity, because SAICAR lyase would be effectively unproductive in an organism such as Leishmania that lacks the remainder of the purine biosynthetic machinery.
ADSS was localized to the cytosol via indirect immunofluorescence (Fig. 9A). ASL was also shown to be cytosolic by subcellular fractionation, indirect immunofluorescence, and direct fluorescence (Fig. 9B-D). Many other vital components of the purine salvage pathway, however, including HGPRT, XPRT, and IMPDH (11,12,27), are sequestered within the glycosome. The targeting of HGPRT, XPRT, and IMPDH to the glycosome is mediated by an archetypical COOH-terminal triad, peroxisomal targeting signal 1 (11,12,27), which is lacking in ADSS and ASL. ADSS and ASL also lack the well characterized NH 2terminal peroxisomal targeting signal 2 (67). Thus, the purine salvage pathway is distributed between the cytosol and glycosomal compartments. The location of ADSS and ASL within the cytosolic milieu proves that the glycosomal membrane is permeable to nucleotides, as all salvaged purines are phosphoribosylated to the nucleotide level by glycosomal HGPRT and/or XPRT (2,11,12,21). To generate adenylate nucleotides, therefore, the IMP product of HGPRT and XPRT must exit the glycosome intact in order to be converted in the cytosol to AMP by the sequential actions of ADSS and ASL. This genetic dissection of the adenylate pathway advances ASL as a critical bottleneck in purine salvage by L. donovani and intimates that this last enzyme in AMP synthesis could serve as a possible drug target for visceral leishmaniasis and perhaps other forms of leishmaniasis triggered by other Leishmania species. Whether ASL is also a putative target for other genera of protozoan parasites remains to be explored, although ASL is not expressed by all protozoan parasites that infect humans (58). The purine salvage pathways of T. brucei and T. cruzi, protozoan pathogens derived from a genus closely related to Leishmania, are similar but not identical to that of Leishmania, suggesting that ASL could be essential for these pathogens, but this has not been tested. Despite the complexity of the purine pathway of Leishmania, this is, at least to date, the only purinemetabolizing enzyme that, by itself, is vital to both life cycle stages of the parasite. Previous studies have established that HGPRT and XPRT together are indispensable to the parasite, but neither is essential on its own (16,21). Further therapeutic validation of ASL is therefore warranted.