The N-Terminal Domain and Glycosomal Localization of Leishmania Initial Acyltransferase LmDAT Are Important for Lipophosphoglycan Synthesis

Ether glycerolipids of Leishmania major are important membrane components as well as building blocks of various virulence factors. In L. major, the first enzyme of the ether glycerolipid biosynthetic pathway, LmDAT, is an unusual, glycosomal dihydroxyacetonephosphate acyltransferase important for parasite's growth and survival during the stationary phase, synthesis of ether lipids, and virulence. The present work extends our knowledge of this important biosynthetic enzyme in parasite biology. Site-directed mutagenesis of LmDAT demonstrated that an active enzyme was critical for normal growth and survival during the stationary phase. Deletion analyses showed that the large N-terminal extension of this initial acyltransferase may be important for its stability or activity. Further, abrogation of the C-terminal glycosomal targeting signal sequence of LmDAT led to extraglycosomal localization, did not impair its enzymatic activity but affected synthesis of the ether glycerolipid-based virulence factor lipophosphoglycan. In addition, expression of this recombinant form of LmDAT in a null mutant of LmDAT did not restore normal growth and survival during the stationary phase. These results emphasize the importance of this enzyme's compartmentalization in the glycosome for the generation of lipophosphoglycan and parasite's biology.


Introduction
Worldwide Leishmania parasites cause important human and animal diseases collectively called leishmaniasis. Disease transmission occurs upon biting by an infected female sand fly. The parasite develops extracellularly as flagellated promastigotes in the midgut of the insect vector, and intracellularly as non motile amastigotes within the phagolysosomal compartment of the vertebrate host's macrophages. L. major is responsible for the cutaneous form of leishmaniasis which manifests in a local selfhealing skin lesion and affects approximately 1-1.5 million patients every year [1]. Ninety percent of cases of cutaneous leishmaniasis are found in Afghanistan, Pakistan, Syria, Saudi Arabia, Algeria, Iran, Brazil, and Peru [1].
Ether glycerolipids are major components of Leishmania membranes, representing approximately 20% of total cellular lipids [2,3]. In Leishmania major, they are found primarily in the phosphatidylethanolamine and phosphatidylinositol glycerolipids [2,3,4,5]. They are of particular importance for this parasite because ether glycerolipid based virulence factors such as lipophosphoglycan (LPG) and glycosylphosphatidylinositol-anchored proteins play critical roles throughout its life cycle (reviewed in [6,7,8,9,10]). Structurally, LPG is a complex glycolipid that is anchored to the plasma membrane via an ether lysophosphatidylinositol anchor [11]. The salient feature of LPG is the conserved domain consisting of the Galb1,4Mana1-PO 4 backbone of repeat units that in L. major are branched with galactose and arabinose residues [6,9,12,13,14].
LmDAT is a unique DHAPAT that bears a very large Nterminal extension of approximately 650 amino acids that is absent in mammalian orthologs [16]. Our previous studies demonstrated that LmDAT is important for growth, survival during stationary phase, the synthesis of ether lipids that includes the ether lipid based LPG, and for virulence, but is dispensable for raft formation [2,4,15]. All together, these data support the notion that LmDAT may represent a potential target for anti-leishmanial chemotherapy. In the present work, a rational deletion approach was applied i) to address the role of the N-terminal extension of LmDAT in enzyme stability and activity, and ii) to investigate the significance of LmDAT glycosomal localization in the synthesis of the ether lipid based LPG. Last, point mutation analysis was carried out to assess whether a catalytically active LmDAT enzyme is required to support normal growth and survival during the stationary phase of the parasite.

Strains and growth conditions
Promastigotes of L. major Friedlin V1 strain (MHOM/IL/80/ Friedlin) were propagated in liquid and semi-solid M199-derived medium [2]. The null mutant Dlmdat/Dlmdat and complemented line Dlmdat/Dlmdat [LmDAT NEO] were described in [4]. Transfection was performed according to Ngo and colleagues [20] and selection was applied as appropriate in the presence of 20-40 mg/ml G418 or 25-50 mg/ml of hygromycin. To follow parasite proliferation, mid log phase parasites were diluted to 5610 5 /ml and enumerated with a hemacytometer as a function of time.
The K852L mutant form of LmDAT was created by introducing the point mutation by PCR using oligonucleotides O133 ('5-AGGAAAGCTTCATAAAGAGGGCGCCGCTG-3') and O134 ('5-TATGAAGCTTTCCTTCCTTCCGCGACGACCCG-3'), and pL-BSD.LmDAT as a template. In addition, a HindIII site was introduced to ''tag'' the mutation. The mutated fragment was swapped with the corresponding wild-type DNA using the SacI and MluI sites of pBEVY-L.LmDAT described in [16] to give pBEVY-L.LmDAT K852L (Ec276). This plasmid was subsequently digested with BamHI and the excised 4.3 kb fragment was ligated in sense orientation into the respective sites of pXG.HV-LmDAT [16] to yield pXG.HV-LmDAT K852L (Ec456). All amplified DNAs were verified by sequencing.

Enzymatic assays
Leishmania protein extracts were prepared as described previously [16,18]. Protein concentration was determined by the bicinchoninic acid assay using bovine serum albumin as a standard. DHAPAT activity was assessed by measuring the acylation rate of DHAP produced by catabolism of fructose-1,6biphosphate by the action of an aldolase and a triose phosphate isomerase, based on a protocol established by Bates and Saggerson as described in [16]. The specificity of [U-14 C]D-fructose-1,6biphosphate (MP Biomedicals) was 295 mCi/mmol.

Digitonin fractionation and electrophoresis
For digitonin treatment fresh end-log cells were harvested, washed once in phosphate buffered saline (PBS), and resuspended in 20 mM TrisHCl (pH 8.0), 1 mM EDTA, 1 mM DTT containing a protease inhibitor cocktail (Roche) at a cell density of 2610 8 /ml. Aliquots of 100 ml were made and supplemented with increasing (0 to 0.6 mg/ml) concentrations of digitonin (stock solution of 15 mg/ml in PBS) and incubated at 26uC for 10 min. Cells were then centrifuged at 20,800 g for 2 min. Supernatants were immediately removed and resolved by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE).

Immunofluorescence assay
Immunofluorescence assay was performed with wild-type parasites expressing pXG.HV-LmDAT-DC 3 as described previously [16]. The recombinant His 6 -V5 (HV) tagged HV-LmDAT-DC 3 was revealed with V5 monoclonal antibodies (Invitrogen) and hypoxanthine guanine phosphoribosyltransferase was visualized with specific rabbit polyclonal immunoglobulins [22]. Both antibodies were used at a 1:500 dilution. Images were taken with a Leica fluorescence microscope.

Results and Discussion
The N-terminal extension is important for LmDAT activity Amino acid sequence alignment of DHAPAT enzymes from human, rat, Caenorhabditis elegans, Trypanosoma brucei and Trypanosoma cruzi shows that orthologs of parasites of the trypanosomatidae family bear a very large N-terminal extension of approximately 650 amino acids that is absent in higher eukaryotic orthologs ( [16]; Fig. 2A; data not shown). Curiously, this domain fails to exhibit any similarity to known proteins and thus, is parasite specific. The function of this N-terminal extension was first investigated by creating truncated LmDAT versions that were N-terminally tagged with a hexahistidine fused to a V5 epitope (HV) for visualization with the V5-specific monoclonal antibody. Plasmids coding for truncated proteins lacking the N-terminal 546 and 686 amino acids (HV-DN 546 -LmDAT and HV-DN 686 -LmDAT), respectively, were constructed, as well as a plasmid coding for a recombinant protein missing the 733 C-terminal amino acids (HV-LmDAT-DC 733 ). In the latter case, the LmDAT C-terminal tripeptide SKM was fused to the C-terminal portion of the truncated protein to ensure glycosomal targeting [25,26,27]. These recombinant proteins, as well as a wild-type HV tagged version (HV-LmDAT), were expressed in the null mutant Dlmdat/Dlmdat background [16].
We verified first that the HV tag did not affect the function of LmDAT; the recombinant protein HV-LmDAT was expressed as an approximately 150 kDa band as shown by Western blot analysis using the monoclonal antibody specific to the V5 epitope (Fig. 2C). Bands present at lower molecular weights very likely represent degradation products of the full length tagged protein. In addition, HV-LmDAT was enzymatically active, and supported normal growth and survival during the stationary phase (Fig. 2B, 3A; [16]). The presence of ether lipids was assessed by investigating the migration behavior of the ether lipid anchored virulence factor LPG; it has been established that Leishmania mutants lacking ether lipids synthesize slow migrating forms of LPG as a result of hyperglycosylation of its disaccharide domain [2,4]. Western blot analysis showed that HV-LmDAT restored expression of normal migrating forms of LPG while the null mutant expressed slower migrating LPG species (Fig. 2D; [4]).
Deletions of the N-terminal (first 546 or 686 amino acids) but not of the C-terminal 733 amino acids led to low intensity signals as shown by Western blot analysis using the monoclonal antibody specific to the V5 epitope, as the corresponding bands at apparent weights of approximately 100 and 83 kDa respectively, were hardly detectable (Fig. 2C). Accordingly, none of the N-terminal (HV-DN 546 -LmDAT and HV-DN 686 -LmDAT) and C-terminal (HV-LmDAT-DC 733 ) truncated proteins rescued the slow growth or the rapid death during the stationary phase (Fig. 3A). Consistent with these results, no DHAPAT activity could be measured with cells expressing any of these truncated versions of LmDAT (Fig. 2B). Accordingly, Western blot analyses in the presence of WIC79.3 for detection of LPG revealed that null mutant strains expressing any of the three truncated forms of LmDAT made glycolipids migrating slower than that of the wild type but similar to that of the null mutant ( Fig. 2D; [4]).
Our results show that N-terminal truncated versions of LmDAT gave lower intensity signals in Western blot analysis in the presence of anti-V5 antibody. A possible explanation is that the V5 antibody has lower affinity for these recombinant proteins or less access to the epitope due to conformation issues. In this case the N-terminal domain may indirectly be important for catalysis; because this domain is absent in higher eukaryotic counterparts ( Fig. 2A; data not shown; [16]), it is doubtful that it is directly involved in substrate recognition or catalysis. These events are expected to occur in the C-terminal domain of LmDAT that is shared by all DHAPAT orthologs ( Fig. 2A; [16]) and this assumption is corroborated by the fact that HV-LmDAT-DC 733 is inactive. Another possible interpretation is that the N-terminal domain somehow helps this very large protein to fold properly in order to yield a stable enzyme, and thus, affects LmDAT activity indirectly. All together these data suggest that designing a compound specific to the N-terminal extension may represent a reasonable strategy to inactivate LmDAT.

A catalytically active LmDAT enzyme is necessary for growth and survival during the stationary phase
In contrast to Dlmdat/Dlmdat, the Dads1/Dads1 mutant, carrying a genetic deletion of the alkyl DHAP synthase gene and lacking ether lipids, does not exhibit any growth phenotype, suggesting that ether lipids are dispensable for growth and survival during the stationary phase [2]. Thus, LmDAT may be a bipartite protein with an N-terminal domain involved in growth and survival during the stationary phase, and a C-terminal domain functioning in catalysis. To assess whether a catalytically active LmDAT DHAPAT is important for normal growth and survival during the stationary phase, a mutant enzyme was created that bears a leucine instead of a lysine at position 852 (K852L), based on the information that replacement of the corresponding conserved amino acid by a histidine in the human DHAPAT abrogated its enzymatic activity [28]. As described above, a HV-tagged version (HV-LmDAT K852L ) was expressed in the null mutant background. Western blot analysis was performed to verify its expression level that was similar to HV-tagged wild-type LmDAT (Fig. 2C). HV-LmDAT K852L enzyme displayed no DHAPAT activity and led to the production of slow migrating LPG glycolipids similar to that of the null mutant (Fig. 2B, 2D). In addition, HV-LmDAT K852L failed to support normal growth and survival during the stationary phase (Fig. 3B).
Altogether, these results demonstrate that the conserved lysine K852 is important for substrate recognition or catalysis of LmDAT similar to the human ortholog. Together with the fact that the Nterminal domain alone is unable to support normal growth and survival during the stationary phase, our results suggest that the acyltransferase activity of LmDAT itself is critical for normal growth and survival during the stationary phase. Our data also exclude the idea that LmDAT is a bipartite enzyme with an Nterminal domain responsible for growth and/or survival during the stationary phase, and a C-terminal part implicated in acyltransferase activity.

Glycosomal expression of LmDAT is critical for LPG synthesis
LmDAT resides in glycosomes consistent with the presence of a typical type 1 C-terminal glycosomal targeting signal sequence (SKM), that is sufficient for targeting proteins to this organelle [16,19]. Hence, the importance of the glycosomal subcellular localization of LmDAT for ether lipid biosynthesis was assessed by expressing a HV-tagged LmDAT recombinant enzyme lacking the C-terminal glycosomal targeting tripeptide SKM (HV-LmDAT-DC 3 ) in the null mutant background [25,27,29]. Western blot analysis was performed in the presence of anti-V5 monoclonal antibodies demonstrated that the levels of HV-LmDAT-DC 3 protein were similar to that of wild-type tagged HV-LmDAT expressed in the null mutant background (Fig. 2C).
Subcellular localization of HV-LmDAT-DC 3 was assessed by immunofluorescence assay. HV-LmDAT-DC 3 was revealed in the presence of V5 monoclonal antibodies and gave a signal that was partially overlapping with that obtained with antibodies specific to the glycosomal resident hypoxanthine guanine phosphoribosyltransferase suggesting a glycosomal association ( [22] ; Fig. 4). However, digitonin fractionation provided evidence for HV-LmDAT-DC 3 localizing outside the glycosomes. Digitonin specif- ically permeabilizes the cytoplasmic membrane at low concentrations while high doses of this detergent are needed to solubilize the glycosomal membrane [30]. HV-LmDAT required a minimal concentration of digitonin of 0.45 mg/ml in order to be properly released in the cell supernatant and behaved similarly as the glycosomal resident arginase ( [31]; Fig. 5), In contrast, HV-LmDAT-DC 3 readily fractionated in the cell supernant at a low digitonin concentration of 0.075 mg/ml, as the cytosolic enzyme phosphomannomutase [23]. Our data suggest that HV-LmDAT-DC 3 is partially associated with glycosomes and possibly other organelles but resides on the cytosolic side of the organellar membrane. This is consistent with previous studies that demonstrated that abrogation of the type 1 C-terminal glycosomal targeting signal sequence resulted in a cytoplasmic localization [25,27,29].
In vitro DHAPAT assays showed that HV-LmDAT-DC 3 was enzymatically active as the wild-type HV-tagged version of LmDAT (Fig. 2B), demonstrating that the C-terminal glycosomal targeting sequence is dispensable for enzymatic activity. Surprisingly, Western blot analysis performed in the presence of WIC79.3 demonstrated that, in contrast to HV-LmDAT, HV-LmDAT-DC 3 expression failed to restore the synthesis of normal migrating LPG (Fig. 2D). Consistent with this result, expression of HV-LmDAT-DC 3 did not ameliorate the slow growth and survival during the stationary phase of the null mutant (Fig. 3B). These data suggest that the proper acyl donor for LmDAT, palmitoyl-CoA, may not be available in the cytosol [16]. This is unlikely because fatty acyl-CoAs are made either in the endoplasmic reticulum by elongases or in the mitochondria by type II fatty acyl-CoA synthases, and have to be transported via the cytosol to the endoplasmic reticulum and glycosomes where lipid biosynthesis occurs [16,32,33,34]. Alternatively, the role of the glycosomal compartmentalization of   LmDAT is to sequester its product, 1-acyl-DHAP, in this organelle for conversion into 1-alkyl-DHAP by the glycosomal alkyl DHAP synthase ADS1 rather than being metabolized to 1-acyl-G3P by the cytosolic alkyl/acyl-DHAP reductase for the synthesis of ester glycerolipids [2,17]. Last, DHAP produced in the glycosome may not be available in the cytosol where HV-LmDAT-DC 3 accumulates [19]. These results are also in accordance with the idea that the glycosomal membrane does not allow transport of 1-acyl-DHAP from the cytosol into the glycosome for conversion to 1alkyl-DHAP by the alkyl-DHAP synthase ADS1.