The Role of Phosphomannose Isomerase in Leishmania mexicana Glycoconjugate Synthesis and Virulence*

Phosphomannose isomerase (PMI) catalyzes the reversible interconversion of fructose 6-phosphate and mannose 6-phosphate, which is the first step in the biosynthesis of activated mannose donors required for the biosynthesis of various glycoconjugates. Leishmania species synthesize copious amounts of mannose-con-taining glycolipids and glycoproteins, which are involved in virulence of these parasitic protozoa. To inves-tigate the role of PMI for parasite glycoconjugate synthesis, we have cloned the PMI gene ( lmexpmi ) from Leishmania mexicana , generated gene deletion mutants ( D lmexpmi ), and analyzed their phenotype. D lmexpmi mutants lack completely the high PMI activity found in wild type parasites, but are, in contrast to fungi, able to grow in media deficient for free mannose. The mutants are unable to synthesize phosphoglycan repeats [-6-Gal b 1–4Man a 1-PO 4 -] and mannose-containing glycoinositolphospholipids, and the surface expression of the glycosylphosphatidylinositol-anchored dominant surface glycoprotein leishmanolysin is strongly decreased, unless the parasite growth medium is supplemented with mannose. The D lmexpmi mutant is attenuated in infections of macrophages in vitro and of mice, suggest-ing that PMI may be a target for anti- Leishmania drug development. L. mexicana D Immunofluorescence Microscopy and FACS of Leishmania Promastigotes and Infected Macrophages— Immunofluorescence microscopy and fluorescence-activated cell sorting (FACS) studies on Leishmania promastigotes

Leishmania are protozoan parasites and the causative agents of a spectrum of animal and human diseases. Their life cycle includes flagellated promastigote stages that reside in the midgut lumen of the sandfly vector and a nonmotile amastigote stage that lives within the mammalian macrophage, where it colonizes the phagolysosomal compartment (1). Leishmania species synthesize large amounts of glycoconjugates that include the unusual glycoinositolphospholipids (GIPLs), 1 the con-served protein-linked glycosylphosphatidylinositol (GPI) membrane anchors, glycoproteins with uncommon N-linked glycans, and a unique family of phosphoglycan-modified molecules that encompasses lipid-linked (lipophosphoglycan, LPG), protein-linked (proteophosphoglycans, PPGs), and unlinked forms (PG). It is believed that these are key molecules for the remarkable resistance of Leishmania parasites against the hostile habitats within their host organisms. The structure of these parasite glycoconjugates has been analyzed in extensive detail (2)(3)(4), and some information on their biosynthesis has been generated (4 -6). The glycan backbones of these glycosylated Leishmania molecules consist predominantly of the monosaccharides Man and Gal, with smaller amounts of GlcNAc, GlcNH 2 , Glc, D-Ara, and myo-inositol present. It has been demonstrated that GDP-Man, Dol-P-Man, UDP-Gal, UDP-Glc, and GDP-D-Ara are used as monosaccharide donors for parasite glycoconjugate assembly (6 -10). Given the high biosynthesis rates, a large and/or rapidly replenishable pool of these activated sugar precursors must exist. In particular the acquisition or biosynthesis as well as the activation of Man is predicted to be of prime importance to the parasites, as all classes of Leishmania glycoconjugates contain this hexose. In yeast and in mammalian cells the activated Man donors GDP-Man and Dol-P-Man are synthesized from Man-6-PO 4 by the enzymes phosphomannomutase, GDP-Man pyrophosphorylase, and Dol-P-Man synthase (Fig. 1). Man-6-PO 4 is a central metabolite in this pathway and may be synthesized in two different ways: exogenous Man, if available, is taken up by the cells into the cytosol by glucose transporters (11) or a more specific mannose transporter and then phosphorylated by hexokinase ( Fig. 1) (12). Alternatively, Frc-6-PO 4 is converted to Man-6-PO 4 in a reaction catalyzed by phosphomannose isomerase (PMI) ( Fig.  1) (13). Lack of the latter enzyme leads to a conditional lethal phenotype in yeast (14) and a severe metabolic disease in humans (15)(16)(17), which demonstrates the essential role of PMI for these organisms. By contrast, the mannose pathway in Leishmania is largely unexplored despite the fact that it is likely to be of central importance for the synthesis of most Leishmania glycoconjugates and that it may, therefore, be a promising target for the design of new anti-parasitic drugs.
In this study, we report the cloning of the L. mexicana PMI gene (lmexpmi) and the generation of gene deletion mutants (⌬lmexpmi) that possess no PMI enzymatic activity. Lack of PMI leads to slowed growth in standard media, but in contrast to yeast, L. mexicana ⌬lmexpmi mutants do not require the addition of Man to the growth medium for viability. L. mexicana ⌬lmexpmi mutants are sensitive to high mannose concentrations, they show profound down-regulation of LPG, GIPL, and leishmanolysin (gp63) biosynthesis, and secrete undergly-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AJ300464.

Parasite Culture and Experimental Infections of Mice and Peritoneal
Macrophages-Promastigotes of the L. mexicana wild type (WT) strain MNYC/BZ/62/M379 and derived mutants were grown at 27°C in semidefined medium 79 (SDM) supplemented with 4% heat-inactivated fetal calf serum (iFCS) as described previously (18). Infection of mice with 10 7 stationary phase promastigotes and infection of mouse peritoneal macrophages were performed as outlined earlier (19). Growth curves of L. mexicana WT and mutants were obtained by seeding SDM/4% iFCS supplemented with or without various concentrations of Man (2 M to 10 mM) with 2 ϫ 10 6 promastigotes and counting the parasite numbers at 8-to 24-h intervals. In some experiments, growth curves were determined in the presence or absence of 10 M swainsonine in the standard growth medium.
Cloning of the L. mexicana lmexpmi Gene, Generation of Gene Knockout and Gene Addback Mutants, Heterologous Expression of PMI, and Generation of Antibodies-DNA techniques were performed as described previously (20). A 300-bp fragment of the L. major pmi gene was obtained from L. major LRC-L137/V121 genomic DNA by polymerase chain reaction (PCR) using the degenerate primers TT(A/G)TGGATG-GG(A/G/C/T)AC(A/G/C/T)CA(C/T)CC and GCCAT(C/T)TC(A/G/C/T)GG-(C/T)TT(A/G)TT that were derived from the conserved Candida albicans and Homo sapiens PMI peptide sequences LWMGTHP and NHKPEMA. The PCR product was subcloned into pCR2.1 (Invitrogen) and sequenced. The digoxygenin-labeled PCR product was used to screen a -DashII library (21) derived from genomic L. mexicana DNA. Positive clones were subcloned into pBSK ϩ (Stratagene) or pGEM-5z (Promega) and sequenced on both strands by the dideoxy chain termination method using an ALFexpress automated sequencer (Amersham Pharmacia Biotech) as described earlier (20). The open reading frame corresponding to lmexpmi was identified by homology to known PMI genes in the data base and by determination of the spliced leader site (20). Double-targeted gene replacement was performed by PCR amplification of the 5Ј-untranslated region (5Ј-UTR) of lmexpmi using the primers KO1 (AATGCGGCCGCATGCATCTTCTGTGCGTG-TC) and KO2 (AGTACTAGTACAGACGTAGCGGAGTTGCTTG) and by amplification of the 3Ј-UTR of lmexpmi using the primers KO3 (AGT-ACTAGTGGATCCCCGAGTTTCCTTCAACATTG) and KO4 (CTTAAG-CTTATGCATGCTCTCATCACGAGTG). The underlined sequences indicate introduced restriction sites. The NotI/SpeI-cut lmexpmi 5Ј-UTR PCR DNA fragment, the BamHI/HindIII-cut lmexpmi 3Ј-UTR PCR DNA fragment, and a SpeI/BamHI DNA fragment containing a hygromycin phosphotransferase gene (hyg) (21) were ligated consecutively into pBSK ϩ . For the second lmexpmi gene replacement cassette, a SpeI/BamHI fragment encoding the phleomycin binding protein gene (phleo) was used (19). The hyg-and phleo-containing gene replacement cassettes were excised from the plasmids by NsiI digestion and transfected into L. mexicana promastigotes as previously described (20). Selection in 96-well microtiter plates and analysis of positive clones were performed as outlined earlier (19). lmexpmi 5Ј-UTR and open reading frame DNA probes were generated by PCR using a PCR-DIG labeling kit (Roche Molecular Biochemicals) using the primer pairs GAGGGGAAGATGGTGGTGAG/GCTCCACCTTCTCCCTGCTA and TTCACCTTGGCAGACCCCTC/GAAGTTTGCCGAGGAGCTGC, respectively. For gene addback and heterologous expression studies, the open reading frame of lmexpmi was amplified from a lmexpmi genecontaining plasmid using the primers CCATGGATCCATGTCTGAGC-TCGTCAAGC and AATAGATCTAGATTACTTGTCGCTCAAGTC. Episomal gene addback was achieved by cloning the BamHI/XbaI-cut PCR fragment into pX (22) and transfection of L. mexicana ⌬lmexpmi promastigotes with this construct as described earlier (20,23). Transfectants were selected by growth in SDM/4% iFCS containing 10 -50 g/ml G418 (Roche Molecular Biochemicals). Alternatively, the lmexpmi gene was expressed under the control of the rRNA promoter, which is known to lead to high level expression not only in promastigotes but also in amastigotes (24). For the construction of an integration vector, pFW31 (25) was first linearized by digestion with AgeI. The purified DNA fragment (9347 bp) was then subjected to partial digestion with AflII generating digestion products with lengths of 7769, 6720, 2627, 1578, and 1049 bp. The largest fragment lacking the lmxmbap open reading frame, but preserving the following spliced leader addition and polyadenylation sites, was then ligated with the annealed primers CCGGTCTAGATCTGCGGCCGCGGCGCGCC and TTAAGGCGCGCC-GCGGCCGCAGATCTAGA yielding pRIB. The gene-containing PCR product (see above) was digested with BglII and BamHI and ligated into BglII/BamHI-cut pRIB yielding pRIBlmexpmi. Correct orientation was checked by restriction enzyme digests. For chromosomal integration into the ribosomal locus of L. mexicana, the integration cassette was excised by digestion with PacI and PmeI (24,25), gel-purified, and transfected into L. mexicana. Recombinant clones were isolated by limiting dilution on 96-well plates in SDM medium containing 20 g/ml hygromycin, 2.5 g/ml phleomycin, and 20 M puromycin. The sequence data for the lmexpmi-containing DNA fragment has been submitted to the EMBL data base under accession number AJ300464.
High level expression of L. mexicana PMI in Escherichia coli M15 as inclusion bodies was achieved by cloning the BamHI/BglII-cut lmexpmi PCR fragment into pQE30 followed by transformation of the bacteria. The inclusion bodies were solubilized in 8 M urea, and denatured PMI was then purified by nickel-nitrilotriacetic acid-agarose chromatography as described by the manufacturer (Qiagen). Rabbits were immunized with 200 g of purified recombinant protein, which was dissolved in 8 M urea, 50 mM NaH 2 PO 4 , pH 4.8, and emulsified with 50% (v/v) complete Freund's adjuvant for primary immunizations and with 50% incomplete Freund's adjuvant (v/v) for all subsequent booster immunizations. Serum was obtained 10 -14 days after each booster immunization. The antiserum was affinity-purified on recombinant protein that had been electrotransferred to polyvinylidene difluoride membranes after SDS-polyacrylamide electrophoresis (PAGE) as described earlier (20).
To generate L. mexicana total cell lysates, 5 ϫ 10 9 promastigotes were resuspended in cold homogenization buffer (50 mM triethylamine/ HCl, pH 7.0, 0.1 mM EDTA, 2.5 mM MgCl 2 containing 20 g/ml leupeptin and 0.5 mM phenylmethylsulfonyl fluoride), sonicated on ice, and centrifuged at 13,000 ϫ g for 30 min. The pellet was resuspended in the same volume homogenization buffer. Enzyme assays were performed in 50 mM triethylamine/HCl, pH 7.0, 0.1 mM EDTA, 2.5 mM MgCl 2 at 25°C with 0.1% bovine serum albumin. For PMI assays, this buffer was supplemented with 0.5 mM NADP ϩ (Roche Molecular Biochemicals), 2.5 mM Man-6-PO 4 (Merck, Darmstadt, Germany), 2 units/ml phosphoglucose isomerase (Roche Molecular Biochemicals) and 2 units/ml glucose-6-phosphate dehydrogenase (Roche Molecular Biochemicals). Hexokinase was measured in assay buffer with addition of 5 mM glucose, 1 mM ATP, and 2 units/ml glucose-6-phosphate dehydrogenase, whereas the phosphoglucomutase assay required the addition of 2 mM glucose 1-phosphate and 2 units/ml glucose-6-phosphate dehydrogenase. Enzyme assays were started by the addition of 5-10 l of cell lysate, ultracentrifugation supernatants, or resuspended pellets, and the absorbance at 340 nm was recorded over 2.5-10 min. One unit of enzyme activity is defined as the amount of enzyme converting 1 mol of substrate/min into the respective product. Total protein of cell lysates was estimated according to Peterson (29).

Isolation of the L. mexicana lmexpmi Gene and Generation of Gene Deletion Mutants by Targeted Gene
Replacement-For the cloning of the Leishmania PMI gene, a degenerate PCR primer pair was constructed from the peptide sequences LWMGTHP and NHKPEMA, which are conserved in C. albicans and H. sapiens PMI (Fig. 2). PCR was performed using L. mexicana, L. donovani, and L. major genomic DNA as templates. Only L. major DNA yielded a PCR product of the expected size (270 -300 bp), and sequencing of this DNA fragment revealed an open reading frame with high homology to known PMIs (data not shown). The DIG-labeled PCR fragment was used to screen a -DashII library of genomic L. mexicana DNA. Sequencing of a lmexpmi gene-containing DNA fragment revealed an open reading frame of 1266 bp encoding a protein of 46.5 kDa, which showed 40 -42% identity to the Saccharomyces cerevisiae, C. albicans, and H. sapiens PMI polypeptide sequences. Complete conservation was observed for the amino acids that have been shown to serve as a ligand for a Zn 2ϩ cation in the active site and for amino acids lining the substrate binding pocket (Fig. 2) (30). The predicted involvement of Zn 2ϩ in the catalytic activity of L. mexicana PMI was corroborated by its complete inhibition by 5 mM o-phenanthroline (data not shown). Southern blots of L. mexicana genomic DNA indicated the presence of a single gene copy ( Fig. 3B and data not shown). Gene replacement cassettes containing the resistance markers phleo and hyg were constructed, and two rounds of targeted gene replacement were performed (Fig. 3A). Plating of transfected cells on SDM medium with or without additional Man (0.5 mM) yielded about the same number of clones that lacked the lmexpmi open reading frame (Fig. 3B). This result was remarkable, because deletion of the PMI gene in yeast led to a Man-dependent phenotype (14).
L. mexicana ⌬lmexpmi Mutants Are Deficient in PMI Activity and Exhibit a Growth Defect That Can Be Both Reversed and Exacerbated by Man Complementation-Immunoblots of L. mexicana promastigote and amastigote total cell lysates probed with affinity-purified antibodies raised against Escherichia coli-expressed L. mexicana PMI revealed that PMI was expressed at very similar levels in both parasite life stages (Fig. 4A). Ultracentrifugation experiments on sonicated cell lysates demonstrated that Ͼ95% of L. mexicana PMI activity is soluble (data not shown). The PMI protein band was completely absent in overexposed immunoblots of L. mexicana ⌬lmexpmi mutants (Fig. 4B), and PMI enzyme assays revealed that these mutants were completely deficient in PMI activity, whereas the activities of other metabolic enzymes like phosphoglucomutase or hexokinase showed little change or were even elevated (Fig.  4C). L. mexicana ⌬lmexpmi mutant promastigotes exhibited slowed growth in standard SDM medium (Fig. 5B) compared with wild type parasites (Fig. 5A). This growth defect could be overcome partially by the addition of 20 M (not shown) or fully by the addition of 200 M Man to the medium (Fig. 5B), but 2 mM Man led again to slow growth and a bloated shape of the cells, whereas 10 mM Man inhibited growth and ultimately killed most of the parasites (Fig. 5B). In contrast, neither growth nor cell shape of L. mexicana wild type cells and of ⌬lmexpmi mutant promastigotes carrying episomal copies of the lmexpmi gene were affected by high Man concentrations (Fig. 5C).
L. mexicana ⌬lmexpmi Mutant Promastigotes Show Impaired Synthesis of Glycoconjugates That Can Be Reversed by Man Complementation of the Growth Medium-Immunoblots of L. mexicana WT total cell lysates probed with the anti-PO 4 -6Gal␤1-4Man␣1-repeat mAb LT6 showed a strong signal in the low molecular weight range that corresponds to LPG, and a weaker signal in the stacking gel and the top of the separating gel due to the presence of membrane-bound proteophosphoglycan (mPPG (20, 31)) (Fig. 6A, lane 1). The L. mexicana ⌬lmex-lpg1, which is specifically defective in LPG synthesis (19), showed only the signal for mPPG (Fig. 6A, lane 6). In contrast, none of the four L. mexicana ⌬lmexpmi null mutants investigated showed any reaction with mAb LT6 on immunoblots. This result was confirmed by FACS analysis of live promastigotes, where the ⌬lmexpmi mutant clones examined showed only negligible surface fluorescence, in contrast to ⌬lmexlpg1 promastigotes, which displayed at least some LT6 epitopes on their surface, most likely due to mPPG expression (Fig. 7A). Likewise, surface binding sites for LT17, an mAb most likely recognizing glucosylated disaccharide phosphate repeats (20), were down-regulated by a factor of 10 -20 (data not shown). The binding of the anti-[Man␣1-2] 0 -2 Man␣1-PO 4 cap mAb L7.25 on blots to proteins of L. mexicana ⌬lmexpmi promastigote total cell lysates was only slightly affected compared with wild type parasite lysates, but a shift of all antibody-recognized proteins to lower apparent molecular mass was clearly visible, which may indicate decreased glycosylation (Fig. 6C). A lower level of glycosylation on surface molecules in L. mexicana ⌬lmexpmi was also detected in FACS analyses with the lectin concanavalin A, where a strongly decreased signal intensity was observed compared with wild type parasites (Fig. 7C). This decrease in concanavalin A labeling was not due to the loss of LPG and mPPG in ⌬lmexpmi mutants, as L. mexicana ⌬lmex-lpg2 promastigotes, which lack both molecules on the surface display more binding sites for this lectin than wild type parasites (32). The surface binding of mAb L3.8 (33,34) directed against the GPI-anchored metalloproteinase leishmanolysin/ gp63 was decreased on ⌬lmexpmi mutants by a factor of 10 ( Fig. 7D), despite the fact that the absence of LPG in general increases accessibility of this surface glycoprotein, as shown by analysis of ⌬lmexlpg1 and ⌬lmexlpg2 mutants (19, 32) (Fig.  7D). The absence or decrease of LT6, LT17, L3.8, and concanavalin A binding sites on L. mexicana ⌬pmi mutants was corroborated by immunofluorescence labels on fixed promastigotes (Fig. 8, B, F, N, and R), and close inspection of the mutant promastigotes revealed a shorter, rounded cell shape in comparison to wild type controls (Fig. 8, I and J). Finally, the two subunits of SAP, which are N-glycosylated and O-phosphoglycosylated glycoproteins, showed strong mobility shifts in immunoblot studies for the L. mexicana ⌬lmexpmi mutant com-pared with the wild type, which is indicative of underglycosylation (Fig. 6D). The biosynthesis and surface display of LPG and mPPG (Figs. 6B, 7B, 8D, and 8H), the surface expression of leishmanolysin/gp63 and concanavalin A binding sites (Fig. 8, P and T), the glycosylation of SAP (Fig. 6D), and normal cell shape (Fig. 8, I-L) could be reconstituted in the mutants by episomal expression of lmexpmi from pX or by chromosomal integration of lmexpmi into the ribosomal locus, which demonstrated that the observed defects were due to the loss of the lmexpmi gene. Furthermore, LPG and mPPG synthesis in the lmexpmi mutants could also be reconstituted by addition of Man into the medium. At 20 M mannose, synthesis of LPG and mPPG could be detected in mutant promastigotes and at 200 M Man, almost wild type levels of these molecules were found in immunoblots (Fig. 6B), FACS analysis (Fig. 7B), and immunofluorescence microscopy (not shown). Similarly, the electrophoretic mobility of SAP decreased upon medium complementation with 20 M Man and was indistinguishable from wild type SAP when 200 M Man was added. Higher concentrations of this sugar did not lead to a further increase in PG synthesis, most likely due to its toxic effects (see above).
HPTLC analysis of a total lipid fraction of the L. mexicana lmexpmi mutant showed the absence of orcinol/H 2 SO 4 -stainable GIPLs, whereas the main L. mexicana GIPLs iM2, iM3, and iM4 (28) could be easily detected by this method in wild type parasites (Fig. 9A). Likewise, these GIPLs were largely absent from the total lipid fraction of [ 3 H]GlcNH 2 -labeled lmex- pmi mutant parasites (Fig. 9B). Instead, three new major labeled glycolipid species not present in the wild type parasites were observed, which migrated faster on the TLC indicative of a less hydrophilic nature (Fig. 9B). Like LPG and mPPG synthesis, GIPL synthesis could also be partially reconstituted by Man complementation of the growth medium, although, even at 200 M, synthesis of the largest GIPL species iM4 was not observed (Fig. 9C).

H]Man Labeling and Swainsonine Treatment of L. mexicana Promastigotes-In [ 3 H]
Man labels of L. mexicana wild type promastigotes, only about 2-3% of the offered radiolabel was incorporated in overnight labeling of parasites and could be detected in cellular and secreted macromolecules like PPGs, leishmanolysin/gp63 and SAP (Fig. 10, A-C) and in GIPLs (Fig.  9D). This low incorporation rate was independent of the presence or absence of 1 mM glucose in the labeling medium (Fig.  10A). By contrast, L. mexicana lmexpmi mutants increased Man incorporation 20-to 30-fold compared with wild type parasites, and more than 60% of the offered radiolabel was found in cellular and secreted macromolecules (Fig. 10, A-C), GIPLs, and uncharacterized compounds in the lipid fraction, most likely intermediates of GIPL, LPG, and GPI synthesis (Fig.  9D). This dramatic increase in metabolic labeling efficiency is specific for [ 3 H]Man, because radioactivity incorporation rates of [ 3 H]GlcNH 2 and myo-[ 3 H]inositol labels are very similar in the wild type and the ⌬lmexpmi mutant.
Swainsonine is a potent and specific inhibitor of lysosomal ␣-mannosidase of various eukaryotes (35). This enzyme is essential for the degradation of glycoprotein N-glycans. The addition of 10 M swainsonine had little effect on the growth of wild type promastigotes, whereas the growth of L. mexicana ⌬lmexpmi is retarded by about 50% (Fig. 11). This result is an indication that, under limiting Man supply, Leishmania may rely partially on the degradation of glycoproteins to acquire this hexose.
L. mexicana ⌬lmexpmi Mutants Are Attenuated but Remain Infectious to Macrophages and Mice-L. mexicana ⌬lmexpmi was less efficient in establishing and maintaining an infection in mouse peritoneal macrophages than the parental wild type strain (Fig. 12A). Addback of the lmexpmi gene to the mutant, either on an episome or by insertion into the chromosome at the ribosomal locus, led to increased infectivity that did not, however, reach wild type levels. Phosphoglycan synthesis of intracellular parasites in infected macrophages was down-regulated in L. mexicana ⌬lmexpmi mutants compared with wild type parasites, but both LT6 and LT17 epitopes could be clearly detected (Fig. 13, compare A-D with E-H). This result is remarkable, because cultured promastigotes exhibit no LT6 and very low levels of LT17 epitopes (Fig. 8, B and F). Wild type levels of parasite PG repeat expression in infected macrophages was observed when ⌬lmexpmi mutants with episomal copies of lmexpmi were used for infection experiments (Fig. 13, I-L). In mouse infection experiments, L. mexicana ⌬lmexpmi led to much smaller lesions compared with wild type parasites, but, surprisingly, this mutant was infectious to these mammals (Fig. 12, B and C). Complementation of ⌬lmexpmi by episomal gene copies or by integration of lmexpmi into the ribosomal locus led to markedly increased virulence, as indicated by accelerated lesion growth in infected BALB/c mice (Fig. 12, B and C).

DISCUSSION
In all eukaryotes investigated so far, the reversible isomerization of Frc-6-PO 4 and Man-6-PO 4 catalyzed by PMI is the first step in the biosynthesis of the activated Man donors GDP-Man and Dol-P-Man (Fig. 1), which are required for the biosynthesis of many glycoproteins, glycolipids, and, in the case of fungi, cell wall components.
In this study, we have cloned, sequenced, and functionally characterized the PMI of the parasitic protozoon L. mexicana as a first step in the investigation of the Leishmania Man pathway. L. mexicana promastigotes express high levels of PMI

Phosphomannose Isomerase of Leishmania mexicana
(ϳ40 milliunits/mg of protein, assay at 25°C), which exceed the average enzyme activity in mammalian tissue more than 15fold (2.6 milliunits/mg of protein, assay at 37°C) (36). Similarly, high PMI activity as in L. mexicana is found in the pathogenic yeast C. albicans (75 milliunits/mg of protein, assay at 37°C) (37). L. mexicana PMI is a soluble enzyme that is equally expressed at the protein level in insect stage promastigotes and mammalian stage amastigotes. Immunofluorescence studies on permeabilized promastigotes suggest a cytoplasmic localization, as indicated by a diffuse fluorescence signal throughout the cell body (data not shown). The gene encoding L. mexicana PMI, lmexpmi, was isolated by a PCR approach using degenerate oligonucleotide primers, and its primary structure is ϳ40% identical to that of S. cerevisiae, C. albicans, or human PMI protein sequences. Two rounds of targeted gene replacement led to lmexpmi null mutants (⌬lmexpmi) lacking detectable PMI activity. These mutants were able to multiply in standard growth medium, even when iFCS had been dialyzed extensively to remove possible traces of Man (data not shown). This lack of dependence on exogenous free Man is surprising and in contrast to S. cerevisiae, Aspergillus nidulans, and C. albicans, where loss of PMI activity is lethal, even in complex media, unless Man is provided (14, 38,39). In humans, the hereditary disease congenital disorder of glycosylation (CDG) type Ib has its basis in PMI deficiency, but although the tissues of patients contain less enzyme than that of healthy controls, complete absence of PMI has never been observed and may be incompatible with life (40). Furthermore, there are no reports about vertebrate cell lines that lack PMI activity. A possible explanation for the unexpected viability of L. mexicana ⌬lmexpmi in the absence of Man could be the degradation of serum glycoproteins by the parasites as a source of Man. Our observation that swainsonine, a potent inhibitor of lysosomal ␣-mannosidase (35), inhibits the growth of ⌬lmexpmi markedly, although affecting the growth of wild type cells only marginally, supports this view. Alternatively, L. mexicana may not need Man for viability in culture.
However, although PMI is not essential, the absence of this enzyme has several severe consequences for L. mexicana. First, growth of L. mexicana ⌬lmexpmi promastigotes in standard medium is slower compared with wild type parasites and the cells show a distinctive rounded shape. Both the growth defect and the change in morphology can be corrected by the addition of 200 M Man to the growth medium or by genetic complementation. Remarkably, addition of 2 mM Man leads again to a growth defect, and 10 mM Man completely inhibits growth and ultimately kills the cells. By contrast, wild type parasites or ⌬lmexpmi gene addback mutants are unaffected by additions of high Man concentrations to the medium. This observation is reminiscent of the toxicity of nutritional Man to apidae, the so-called honey bee syndrome. It has been suggested that an imbalance between high hexokinase and low PMI activities in honey bees, which leads to an accumulation of potentially toxic Man-6-PO 4 and, possibly, to ATP depletion, is the cause of this syndrome (41).
Second, lack of PMI activity leads to drastic down-regulation of glycoconjugate synthesis in L. mexicana promastigotes: A combination of immunoblot, FACS, and immunofluorescence microscopy studies suggest that ⌬lmexpmi promastigotes grown in standard medium are unable to synthesize phosphoglycan repeat-modified LPG and PPGs. Furthermore, a decrease in surface binding sites for concanavalin A as well as mobility shifts in SDS-PAGE of many cellular proteins and the dominant secreted glycoprotein SAP suggest a general underglycosylation of glycoproteins in these mutants. Finally, the synthesis of the dominant promastigote GIPLs iM2, iM3, iM4, and EPiM3 (28) is down-regulated to such an extent that they could not be detected by the methods employed in this study. It has been suggested that Man-containing GIPLs are essential for L. mexicana viability (42). The results of this study demonstrate, however, that their biosynthesis can be down-regulated to undetectable levels without affecting parasite viability in culture. Metabolic [ 3 H]GlcNH 2 labeling revealed new, more hydrophobic GIPL species in the ⌬lmexpmi mutants, which are not present in wild type parasites. It is likely that they represent GIPL precursors like GlcNH 2 -phosphatidylinositol or Gl-cNAc-phosphatidylinositol species, although this suggestion requires proof by structure analysis. The defect in glycoconjugate synthesis can be partially or fully restored by lmexpmi gene addback or by the addition of 200 M Man to the growth medium. In metabolic labeling experiments with [ 3 H]Man, ⌬lmexpmi promastigotes incorporate about 30 times more radioactivity into their glycoconjugates than wild type parasites, which use this hexose very inefficiently for biosynthesis. This result suggests that, under normal culture conditions, the majority of Man-6-PO 4 in L. mexicana originates from the Glc-6-PO 4 /Frc-6-PO 4 pool (Fig. 1) of the promastigote and not from Man. Whether this hexosemonophosphate pool is fed by exogenous Glc or by gluconeogenesis, or both, remains unclear. This is in contrast to the situation in humans, where, in most tissues, the bulk of Man-6-PO 4 utilized for glycoprotein synthesis is not derived from Glc-6-PO 4 but originates from Man in serum, where its concentration is around 50 M. A specific Man transporter, which is only weakly affected by the large excess of Glc in serum (ϳ5 mM), ensures efficient uptake of this hexose  (12,42). Our finding that [ 3 H]Man labeling of L. mexicana promastigotes is also only marginally affected by a more than 10,000-fold excess of Glc over [ 3 H]Man suggests that the parasites may also possess a similarly efficient and specific Man uptake system.
A third consequence of the lack of PMI for the parasites is a marked decrease in virulence. In macrophage infection studies, ⌬lmexpmi parasites are less successful in colonizing host cells than the parental wild type line, and in mouse infections, lesion growth is much slower. Episomal and, in particular, chromosomal integration addback of the lmexpmi gene to ⌬lmexpmi mutants improves their ability to persist and multiply within macrophages and leads to increased virulence to BALB/c mice. The fact that the ⌬lmexpmi mutants, although severely attenuated, are still infectious to macrophages and mice is at first  LT17 (B, F, J). Counterstaining of the same specimen for DNA was performed with 4,6-diamidino-2-phenylindol (C, D, G, H, K, L). Leishmania-infected macrophages that can be recognized by the intracellular spot-like 4,6-diamidino-2-phenylindol signals of the parasite kinetoplasts are marked by arrows, whereas uninfected cells are marked by asterisks. The exposure times are identical for specimens stained with the same antibody. unexpected, because the general impairment of glycoconjugate synthesis (2), in particular the defect of GIPL assembly (43), should preclude virulence completely. However, those ⌬lmexpmi promastigotes that survive within macrophages transform into amastigotes and, remarkably, synthesize glycoconjugates displaying mAb LT6 and LT17 epitopes, although at lower levels than either wild type or lmexpmi gene addback mutants. Because L. mexicana ⌬lmexpmi promastigotes in culture need exogenous Man added to their medium to regain this biosynthetic ability, this result suggests that amastigotes have access to some macrophage-derived Man within the parasitophorous vacuole.
PMI of the pathogenic fungus C. albicans is currently being investigated as a target to combat fungal infections (30). Our results suggest that the inhibition of this enzyme in Leishmania amastigotes colonizing a human lesion may lead to drastically slowed parasite growth, which may enable the immune system to eradicate the infection. Therefore, L. mexicana PMI may be a valid target for the development of new anti-Leishmania drugs. The ⌬lmexpmi mutant generated in this study provides the first Man-dependent conditional system for the synthesis of Leishmania phosphoglycans, GIPLs, and possibly GPI anchors and N-glycans, which could be exploited for the investigation of biosynthesis, intracellular transport, and biological functions of these parasite glycoconjugates.