Endosomes, glycosomes, and glycosylphosphatidylinositol catabolism in Leishmania major.

Glycosylphosphatidylinositols (GPIs) serve as membrane anchors of polysaccharides and proteins in the protozoan parasite Leishmania major. Free GPIs that are not attached to macromolecules are present in L. major as intermediates of protein-GPI and polysaccharide-GPI synthesis or as terminal glycolipids. The importance of the intracellular location of GPIs in vivo for functions of the glycolipids is not appreciated. To examine the roles of intracellular free GPI pools for attachment to polypeptide, a GPI-specific phospholipase C (GPI-PLCp) from Trypanosoma brucei was used to probe trafficking of GPI pools inside L. major. The locations of GPIs were determined, and their catabolism by GPI-PLCp was analyzed with respect to the intracellular location of the enzyme. GPIs accumulated on the endo-lysosomal system, where GPI-PLCp was also detected. A peptide motif [CS][CS]-x(0,2)-G-x(1)-C-x(2,3)-S-x(3)-L formed part of an endosome targeting signal for GPI-PLCp. Mutations of the endosome targeting motif caused GPI-PLCp to associate with glycosomes (peroxisomes). Endosomal GPI-PLCp caused a deficiency of protein-GPI in L. major, whereas glycosomal GPI-PLCp failed to produce the GPI deficiency. We surmise that (i) endo-lysosomal GPIs are important for biogenesis of GPI-anchored proteins in L. major; (ii) sequestration of GPI-PLCp to glycosomes protects free protein-GPIs from cleavage by the phospholipase. In T. brucei, protein-GPIs are concentrated at the endoplasmic reticulum, separated from GPI-PLCp. These observations support a model in which glycosome sequestration of a catabolic GPI-PLCp preserves free protein-GPIs in vivo.

Free protein-GPIs are synthesized in the secretory system, beginning on the cytoplasmic side of the endoplasmic reticulum (ER) (5,6) and are detectable on the plasma membrane of vertebrates and trypanosomatids (7)(8)(9)(10)(11). The exocytic system is continuous yet highly compartmentalized; it includes the ER, Golgi, secretory vesicles, and in highly polarized cells (e.g. trypanosomatids) parts of the endosomal system where the secretory and endocytic pathways converge (12,13).
In L. major, the importance of intracellular pools of GPIs has not been studied in the context of biological functions. For example, it is held that GPIs are synthesized within the ER, yet there is no direct evidence that GPIs are added to proteins in ER. Only in the yeast Saccharomyces cerevisiae has it been demonstrated conclusively that when anterograde vesicle traffic is blocked GPIs are added to protein in the ER (14).
Because GPIs are present on the plasma membrane of Leishmania, it stands to reason that many areas of the secretory pathway might contain the glycolipids (8). Unanswered questions in the field include the following: Do GPIs accumulate to different extents in different organelles of the endo/exocytic pathway of Leishmania? Are all GPIs in the secretory pathway equally competent for addition to proteins? Are GPIs added to protein in the ER? How is the biosynthesis/storage of GPIs coordinated with catabolic enzymes that might cleave the glycolipids? These questions deserve to be addressed for Leishmania and other eukaryotes because there are clear differences between vertebrate cells, for example, and S. cerevisiae in the molecular machinery and pathways for endo/exocytosis (15)(16)(17).
Trypanosoma brucei expresses a GPI-specific phospholipase C (GPI-PLCp) (18 -20). Heterologous (stable) expression of GPI-PLCp in L. major confers a GPI-negative phenotype on the cells by depleting intermediates of protein-GPI biosynthesis (21). The shortage of prefabricated GPIs for anchoring polypeptides to membranes leads to constitutive secretion of major cell surface polypeptides (e.g. gp63) that should have been GPIanchored. These proteins enter the ER because of their Nterminal signal peptide but lack signals for retention in the cell (21,22). Polysaccharide-GPIs are not cleaved in vivo by GPI-PLCp in L. major (21).
Peroxisomes belong to the microbody group of organelles (for review, see Ref. 23). Important for both catabolic (e.g. ␤-oxida-tion of fatty acids, decomposition of hydrogen peroxide) and biosynthetic reactions (e.g. ether-lipid and cholesterol), peroxisomes in different eukaryotes may contain specialized pathways (e.g. glyoxylate cycle in plants and purine synthesis in mammals (for review, see Ref. 23)). The trypanosomatids (e.g. Leishmania spp. and T. brucei) are deeply diverged eukaryotes (24,25). In these organisms, peroxisomes, in addition to being the organelle in which ether-lipid synthesis and ␤-oxidation of fatty acids take place, harbor the first seven enzymes of the glycolysis pathway (for review, see Ref. 26). Consequently, peroxisomes are termed "glycosomes." Mistargeting of glycolytic enzymes to the cytosol is lethal in the bloodstream form of T. brucei (27)(28)(29), indicating that glycosomes are an essential organelle in the parasite.
We used the GPI-binding protein ␣-toxin from Clostridium septicum (30,31) to identify sites of free protein-GPI accumulation in L. major. 2 Concurrently, GPI-PLCp was used as a "catabolic probe" of free protein-GPIs in Leishmania. Both GPI-PLCp and protein-GPIs were detected on the endo-lysosomal system of L. major. However, several cysteine mutations in GPI-PLCp caused the enzyme to localize to glycosomes. Glycosomal GPI-PLCp did not cause a protein-GPI deficiency in L. major. Thus, endo-lysosomal GPIs are important for GPI addition to protein. Further, targeting of GPI-PLCp to glycosomes sequesters the enzyme from free protein-GPIs in vivo. Data obtained in T. brucei where GPI-PLCp is endogenous are consistent with the "GPI sequestration" hypothesis.
GPI-PLC Enzyme Assay-Leishmania (1 ϫ 10 7 cells/ml) growing in medium containing 200 g/ml G418 were harvested, washed in PBS and lysed in 1 ml of hypotonic buffer (10 mM Tris-HCl, pH 8.0, 2 mM EDTA) containing a protease inhibitor mixture (2.1 M leupeptin, 0.1 mM N-tosyl-L-lysine chloromethyl ketone, 0.4 unit aprotinin) (38). The cells were incubated on ice for 20 min and then centrifuged for 20 min (14,000 ϫ g, 4°C). The pellet was resuspended in 500 l of GPI-PLC assay buffer AB (1.0% Nonidet P-40, 5 mM EDTA, 50 mM Tris-HCl, pH 8.0) and incubated on ice for 20 min. After centrifugation (14,000 ϫ g, 20 min, 4°C), portions of the supernatant were added to 15 l of GPI-PLC assay buffer containing 2 g of [ 3 H]myristate-labeled membrane form variant surface glycoprotein ([ 3 H]mfVSG) (38). The reaction mixture was incubated at 37°C for 20 min and terminated by the addition of 500 l of water-saturated butanol. The amount of [ 3 H]dimyristoylglycerol released from [ 3 H]mfVSG was quantified by scintillation counting of 400 l of the organic phase. Lysates were titrated to ensure that the amount of enzyme measured was within the linear range of the assay (i.e. 0.1-1.0 unit) (38). Total protein concentration of cell lysates was determined with a bicinchoninic acid assay (Pierce).
To monitor glycoproteins to which concanavalin A (ConA) bound, the Immobilon P membrane (see preceding paragraph) was blocked in 5% nonfat milk (Carnation) in PBS and treated with 5 g/ml ConA for 1 h at 27°C. The membrane was washed with PBS followed by incubation with anti-ConA antibody (Sigma) (1:6,000 dilution) for 1 h at room temperature. After washing with PBS, the membrane was incubated with alkaline phosphatase-conjugated goat anti-rabbit IgG (Roche Applied Science) (1:1,000 dilution) for 1 h and then processed for antigen detection with 5-bromo-4-chloro-3-indolyl phosphate p-toluidine/nitro blue tetrazolium chloride (41).
Organelle Detection in Live Cells-Endo-lysosomal compartments were identified after uptake of three fluorescent markers: dextran-Texas Red, FM4-64, and LysoTracker Red. For endocytosis of dextran-Texas Red, Leishmania at a density of 1 ϫ 10 7 cells/ml were gently pelleted (2,000 ϫ g, 2 min) and resuspended in 500 l of M199 medium containing 500 g/ml dextran-Texas Red (Molecular Probes, Inc.) at 27°C for 10 or 30 min. Cells were also stained (in 500 l of M199 medium) with 10 M FM4-64 (Molecular Probes, Inc.) from a 4 mM stock in dimethyl sulfoxide). Acidic organelles (endosomes and lysosomes) were stained with 100 nM LysoTracker Red (Molecular Probes, Inc.) at 27°C for 5 min-1 h.
Immunofluorescence Assays-Log phase cells (1 ϫ 10 7 /ml) were fixed in 2% paraformaldehyde (8 min, on ice), washed in PBS, and adhered to poly-L-lysine-coated coverslips. The cells were permeabilized with 0.25% (w/v) Triton X-100 (200 l/coverslip) for 5 min at 4°C. Nonspecific sites were blocked with 1% bovine serum albumin and PBS for 1 h at room temperature. For studies with T. brucei bloodstream, cells were fixed as described but permeabilized with methanol (5 min, Ϫ20°C) before processing as described below.
To detect GPI-PLCp, anti-GPIPLC (anti-peptide) antibody RC300 (1:1,200 dilution in blocking solution; 200 l total volume) or monoclonal 2A6-6 (1:1,500 dilution in blocking solution; 200 l total volume) (32) was added for 1 h at room temperature. Cells were washed once with PBS, twice with high salt buffer (PBS plus 500 mM NaCl), and twice with PBS. The cells were then immersed in 200 l of a 1:1,000 dilution of Alexa Fluor 488-conjugated goat anti-rabbit IgG or a 1:2,000 dilution of Alexa Fluor 594-conjugated goat anti-mouse IgG for 1 h at 4°C.
Labeled cells were viewed with a fluorescence microscope (Leica DMIRBE). Images were captured using an interline chip-cooled CCD camera (Orca 9545: Hamamatsu) and processed with OpenLab 3.1.2 software (Improvision, Inc.).
In both organelle detection and immunofluoresence assays, more than 100 cells were visualized microscopically. Data presented in each figure are representative of those found in at least 90% of the population, unless otherwise noted.

RESULTS
Exploration of GPI Utilization in L. major-To investigate functions of intracellular pools of free GPIs in L. major, a "probe" for GPIs was needed in living cells. GPI-PLCp from T. brucei was highly suitable as the in vivo probe of GPIs because heterologous expression of the enzyme in L. major causes a deficiency of (free) protein-GPI without affecting polysaccharide-GPI metabolism (21,46). L. major expressing GPI-PLCp or cysteine mutants of the enzyme were analyzed for two reasons. First, most Cys mutants of GPI-PLCp (expressed in Escherichia coli) retain enzyme activity (47,48). Second, in some biological systems Cys mutations can alter intracellular location of polypeptides (49 -52). Therefore, we reasoned that Cys mutants of GPI-PLCp, compared with unmutated enzyme, could have different intracellular locations that might lead to divergent effects of the enzyme on GPIs in L. major.
For our work, mutations of Cys to alanine (Ala) or serine (Ser) in GPI-PLCp were constructed. It was important to test whether the altered proteins retained enzyme activity in L. major. For this purpose, in vitro experiments were performed; lysates of L. major transfectants expressing GPI-PLCp or Cys mutants of the enzyme (e.g. C80A) were used to cleave exogenous [ 3 H]myristate-labeled VSG (38) (Fig. 1A). Cys mutants of GPI-PLCp exhibited enzyme activity in the range detected for the unmutated protein (Fig. 1A). No GPI-PLCp activity was detected in L. major transfectants expressing pUTK (34), the expression vector without a gene for GPI-PLCp. A Q81L mutant of GPI-PLCp was devoid of enzyme activity (Fig. 1A). These observations are consistent with data from earlier studies of enzyme activity in E. coli (34,47,48).
Heterologous expression of GPI-PLCp in L. major produces a GPI deficiency that is marked by accelerated secretion of gp63, the major GPI-anchored protein (21,53). Loss of gp63 from the cells was used as a "read-out" of increased catabolism of (free) protein-GPIs in vivo. From Western blot analysis (Fig. 1B), gp63 was reduced in cells expressing GPI-PLCp (Fig. 1B, com-FIG. 1. A, activity of GPI-PLCp cysteine mutants in lysates of L. major. Detergent extract from L. major expressing GPI-PLC was assayed for enzyme activity and corrected for total protein. 100% activity corresponded to 652 units/mg activity. B, effect of GPI-PLCp on GPI-anchoring of Gp63p in L. major. Parasites were lysed hypotonically in the presence of p-chloromercuriphenyl sulfonate, and the membranebound proteins were separated by SDS-PAGE. Proteins were transferred to Immobilon P membrane and detected with anti-gp63 polyclonal antibody or by ConA immunoblotting.

FIG. 2. GPI-PLCp is not detected at the Golgi or ER. A-D,
Leishmania transfected with unmutated GPI-PLCp was fixed, permeabilized, and incubated with anti-GPIPLCp monoclonal antibody, followed by Alexa Fluor 594 goat anti-mouse-conjugated IgG (red). E-H, L. major expressing GPI-PLCp was fixed, permeabilized, and incubated with anti-GPIPLCp and anti-BiP antibodies. Anti-GPIPLC was detected using Alexa Fluor 594 goat anti-mouse IgG conjugate (G, red). Anti-BiP was detected by an Alexa Fluor 488 goat-conjugated antirabbit IgG (F, green). I-L, cells expressing GPIPLC-GFP were incubated with the Golgi marker, BODIPY-TR ceramide. GPIPLC-GFP was detected by standard protocols (J, green); BODIPY-TR ceramide (K, red) and a merge of images from J and K are shown on L. Blue, DAPI staining of DNA; n, nucleus; k, kinetoplast (mitochondrial DNA). More than 100 cells were visualized microscopically. Data presented in each panel are representative of at least 90% of the population.
Altered Intracellular Location of Class I Cys Mutants of GPI-PLCp-We were interested in learning why some but not all Cys mutants of GPI-PLCp failed to produce GPI-deficient L. major (Fig. 1B). Because all of the proteins cleaved GPIs in vitro (Fig. 1A), an explanation that some Cys mutants were enzymatically inactivated was ruled out, in favor of an alternative theory that Cys mutations in GPI-PLCp altered intracellular targeting of the enzyme. To set the stage for testing this second hypothesis, we first determined the intracellular location of unmutated GPI-PLCp in L. major.
In immunofluorescence analysis, GPI-PLCp was detected at the anterior end of Leishmania, frequently positioned between the nucleus and the kinetoplast (mitochondrial DNA). A smaller amount of the protein was detected posterior to the cell nucleus (Fig. 2, C and D). In control experiments, cells containing the expression vector (pUTK) alone had no fluorescence (data not shown). These results demonstrate that GPI-PLCp associates with an organelle (or structure) in the anterior region of L. major.
Protein-GPI biosynthesis begins on the cytoplasmic side of membranes of the secretory system, which includes the ER (5, 6). One hypothesis to explain the GPI deficiency in Leishmania expressing GPI-PLCp is that the enzyme binds to the ER where it digests cytoplasmically oriented (free) protein-GPIs (21). (Lacking an N-terminal signal sequence (19,32), GPI-PLCp cannot enter the secretory system.) To test this hypothesis, Leishmania expressing GPI-PLCp were double labeled with antibodies against BiP, an ER marker (44) and GPI-PLCp, and observed by fluorescence microscopy. Some BiP was found at the periphery of the cell nucleus, but the protein was concen- Leishmania transfected with cysteine mutants of GPI-PLC were fixed and permeabilized. Anti-GPIPLC monoclonal antibody was added followed by detection with Alexa Fluor 594 goat anti-mouseconjugated IgG (red). DNA was stained with DAPI (blue). In each pair of images, the phase-contrast image is on the left, and a merged fluorescence image is on the right; n, nucleus; k, kinetoplast. More than 100 cells were visualized microscopically. Data presented in each panel are representative of at least 90% of the population.
Endosomes, Glycosomes, and GPI Catabolism in L. major trated in a region anterior to the kinetoplast (Fig. 2F). In contrast, GPI-PLCp was located largely between the nucleus and the kinetoplast (Fig. 2G). Therefore, the location of GPI-PLCp is distinct from the ER. A small fraction of GPI-PLCp appeared to overlap with a minor proportion of BiP in the region between the nucleus and the kinetoplast. Nonetheless, the general staining patterns of GPI-PLCp and BiP are different, justifying a preliminary conclusion that GPI-PLCp is not an ER protein. These data alone, however, cannot exclude the possibility that GPI-PLCp is targeted to a subdomain of the ER.
GPI-PLCp did not localize to the Golgi complex. This fact was demonstrated by labeling L. major expressing a GPIPLC-GFP chimera with the Golgi marker BODIPY-TR ceramide (42). The fluorescence from GPIPLC-GFP (Fig. 2J) was not coincident with the signal from BODIPY-TR ceramide (Fig. 2, K and L). To test this hypothesis, endosomes in L. major expressing GPIPLC-GFP were detected with the styryl dye FM4-64 (43,54). After endocytosis of FM4-64 for 30 min, the dye and GPI-PLC-GFP were visualized in live cells. FM4-64 encompassed the region where GPI-PLCp was found (Fig. 3, B-D), indicating that GPI-PLCp associated with the endo-lysosomal system of Leishmania. To distinguish between early and late endosomal compartments, parasites were allowed to endocytose fluorescent dextran-Texas Red for 1, 3, 5, 30, and 45 min. Dextran-Texas Red was not detected between the kinetoplast and the nucleus before 10 min (data not shown). After 30 min, the dye colocalized with GPIPLC-GFP (Fig. 3, F-H). We conclude that GPI-PLCp associates with late endosomes (and possibly the lysosomal system) but not early endosomes.
Further evidence that GPI-PLCp was present on the endolysosomal system was obtained by staining L. major with Lyso-Tracker Red, a dye that accumulates in organelles with low pH (12,43). In double labeling studies, the green fluorescence from GPIPLC-GFP was coincident predominantly with zones detected with LysoTracker Red (Fig. 3L). Similar results were obtained in experiments using a nontagged form of GPI-PLCp (data not presented).
From these observations, we conclude that GPI-PLCp is targeted specifically to the endo-lysosomal system in L. major. GPI-PLCp is highly likely to bind to the cytoplasmic side of endosomes because it lacks an N-terminal ER signal peptide for entry into the secretory pathway (19,32). (Figs. 2 and 3) it was possible to test whether or not Cys mutations in the polypeptide redirected the enzyme to a different site. Following protocols used to determine the location of unmutated GPI-PLCp, the C184A and C347A mutants of GPI-PLCp (both of which are class II Cys mutants, defined as enzymes that produced a protein-GPI deficiency in L. major) were found in the anterior region of L. major, similar to unmutated GPI-PLCp (Fig. 4, A and B). The Q81L mutant was also detected in the anterior region (Fig. 4C).

Glycosomal Targeting of Class I Cysteine Mutants of GPI-PLCp-Knowing the location of GPI-PLCp in Leishmania
Class I Cys mutants of GPI-PLCp (defined as those proteins that failed to produce GPI-deficient L. major) included C80A, C269S/C270S, C269S/C273S, C270S/C273S, and C269S/ C270S/C273S. Those proteins were associated with punctate structures throughout the cell (Fig. 4, D-H). Apparent displace- Endosomes, Glycosomes, and GPI Catabolism in L. major ment of GPI-PLCp from the anterior region of Leishmania in the class I mutants correlates with reversal of the GPI deficiency normally associated with GPI-PLCp expression in vivo (Fig. 1B).
The organelle that GPI-PLCp bound in L. major was identified in double labeling experiments with anti-GPIPLCp (antibody) and one of four markers of subcellular organelles:FM4-64, dextran-Texas Red, anti-HGPRT, and anti-aldolase. (HGPRT and aldolase are marker proteins for peroxisomes (glycosomes) of Leishmania (45), whereas FM4-64 and dextran-Texas Red are taken up into endosomes and lysosomes (12,55).) C184A and C347A mutants of GPI-PLCp colocalized with endocytosed FM4-64 (Fig. 5). Similar data were obtained with dextran-Texas Red (not presented). These observations demonstrate that class II Cys mutants of GPI-PLCp, which retain the ability to cause a GPI deficiency, associate with the endolysosomal system. However, the class I mutants of GPI-PLCp (i.e. C80A and C269S/C270S/C273S) did not colocalize with the endosomal system (data not presented).
Class I Cys mutants of GPI-PLCp were targeted to glycosomes because they colocalized predominantly with HGPRT (Fig. 6, A-H) (and aldolase, not presented). Control (unmutated) GPI-PLCp did not colocalize with HGPRT (Fig. 6, I-L). Together these results demonstrate that specific Cys mutants of GPI-PLCp associate with glycosomes.
Endosomal GPIs: Colocalization with GPI-PLCp Causes a GPI Deficiency-Stable expression of GPI-PLCp in Leishmania causes a protein-GPI deficiency (21,46). Interestingly, enzyme activity of GPI-PLCp (Fig. 1B) and the presence of the enzyme on the endo-lysosomal system (Figs. 3 and 4) are necessary to produce GPI-deficient cells. Based on these observations, we envisioned that both GPIs and GPI-PLCp might be detected concurrently on endosomes if the enzyme did not cleave all of the glycolipids. To test this theory, we attempted to find (free) protein-GPIs and GPI-PLCp in L. major expressing either unmutated or the enzymatically inactive Q81L mutant of GPI-PLCp. Similar studies were performed with a C269S/C270S/ C273S mutant of GPI-PLCp as a control.
Intracellular GPIs detected with the GPI-binding protein ␣-toxin were found on the endo-lysosomal system. 2 Those GPIs colocalized with unmutated GPI-PLCp (Fig. 7, B-D). Unfortunately, because unmutated GPI-PLCp cleaves GPIs, the inten- sity of the ␣-toxin fluorescence was routinely less than that found in cells that did not contain the phospholipase. To preserve the ␣-toxin fluorescence signal, we performed similar studies with the Q81L mutant of GPI-PLCp, which does not cleave GPIs (48). In L. major containing the Q81L mutant, colocalization of GPI and GPI-PLCp was clear (Fig. 7, I-L) compared with cells containing unmutated enzyme (Fig. 7,  B-D). These data indicate that GPI-PLCp binds to (free) protein-GPIs on endosomes. In contrast, the vast majority of the C269S/C270S/C273S mutant (Fig. 7F) failed to colocalize with ␣-toxin (Fig. 7, G and H). Thus, the inability of C269S/C270S/ C273S GPI-PLCp (and other class I Cys mutants) to produce GPI-deficient Leishmania correlates with failure of the bulk of the proteins to bind endosomes.
T. brucei: GPIs Accrue at the ER and Are Sequestered from GPI-PLCp-The experiments in Leishmania led to the conclusion that free protein-GPIs can be protected from cleavage by GPI-PLCp if the enzyme is targeted to glycosomes in vivo (Figs. 1B and 6). In T. brucei where GPI-PLCp is endogenous, the enzyme does not cause a GPI deficiency that results in constitutive secretion of the major GPI-anchored protein VSG (18,56,57). Therefore, we hypothesized that free protein-GPIs may be separated from GPI-PLCp in T. brucei. To test this theory, we examined the location of both intracellular GPIs and GPI-PLCp (Fig. 8). Intracellular GPIs were detected with the GPI-binding protein ␣-toxin, 2 whereas endosomes were detected after endocytosis of FM4-64 (38,55,59).
GPIs were detected in a subregion (i.e. the posterior) of the ER, which was identified with the marker protein TbSec61p (Fig. 8, A-D). (Sec61p forms the protein import pore on the ER of eukaryotes (60).) GPIs were not concentrated on the endolysosomal system of T. brucei (Fig. 8, E-H). GPI-PLCp was associated with glycosomes, which were detected with the marker protein aldolase (61) (Fig. 8, I-L). Consequently, GPIs in T. brucei did not colocalize with GPI-PLCp (Fig. 8, M-P).
In conclusion, the bulk of free protein-GPIs in T. brucei were detected at the ER. These GPIs are sequestered from GPI-PLCp, which is glycosomal in T. brucei (Fig. 8L). Separation of enzyme from substrate is likely to help preserve the protein-GPIs that are needed for attachment to proteins (e.g. transferrin receptor and VSG) in T. brucei (for review, see Refs. 2 and 62).

DISCUSSION
Endosomal GPIs are Important for Biogenesis of GPI-anchored Proteins in L. major-In this study, we have found two conditions that facilitate catabolism of (free) GPIs in L. major (outlined in Fig. 9). First, enzyme activity of GPI-PLCp is necessary because a Q81L mutant of the protein which cannot cleave GPIs (48) fails to produce GPI-deficient L. major (Fig. 1). Second, targeting of GPI-PLCp to intracellular GPIs is important, as expected. In L. major, free protein-GPIs accumulate at endosomes (Fig. 3). 2 Consequently, cells become GPI-deficient only if GPI-PLCp associates with the endo-lysosomal system (Figs. 3 and 5). When Cys mutants of GPI-PLCp are targeted to glycosomes, the GPI deficiency is averted (Figs. 6 and 9). These observations highlight the importance of endosomal GPI in biogenesis of GPI-anchored proteins in L. major: deficiency of protein-GPIs is associated with loss of cell-associated gp63 (Fig. 1B) (21).
In T. brucei where GPI-PLCp is endogenous (18,20,57), each cell contains ϳ10 7 molecules of VSG, a GPI-anchored protein (63). How does T. brucei retain sufficient GPIs needed for anchoring VSG to the plasma membrane when GPI-PLCp is present in the same cell? The data obtained from the class I Cys mutants of GPI-PLCp in our studies in L. major suggest a plausible model: separation of GPI-PLCp from areas of the cell where (free) GPIs accumulate may help prevent cleavage of GPIs. This so-called GPI-PLCp sequestration hypothesis was evaluated in T. brucei. There, we discovered that GPIs accumulate at the ER, whereas GPI-PLCp is a glycosome protein (Fig. 8). These data are consistent with a proposal that sequestration of GPI-PLCp to glycosomes contributes to prevention of excessive cleavage of GPIs in T. brucei (as first observed in L. major). (Signals that target GPI-PLCp to glycosomes in T. brucei and L. major are under investigation because the enzyme lacks peroxisome targeting sequences; for review, see Refs. 23 and 26.) Our studies reveal two previously unknown differences in free protein-GPI metabolism in L. major and T. brucei. First, whereas (free) protein-GPIs accumulate at the ER of T. brucei (Fig. 8), the glycolipids concentrate on the endo-lysosomal system of L. major. Second, organelle-targeting signals on GPI-PLCp are deciphered differently by the two trypanosomatids. In T. brucei, unmutated GPI-PLCp is glycosomal (Fig. 8), but in L. major the protein is targeted to the endo-lysosomal system (Fig. 3). Thus, organelle signal peptides from T. brucei may not (always) be recognized in L. major.
A Protein-GPIs are associated mainly with the endo-lysosomal system. Colocalization of enzymatically active GPI-PLCp and protein-GPIs on endosomes leads to a GPI deficiency, arising from cleavage of protein-GPIs by the phospholipase C. Peroxisomal targeting of GPI-PLCp separates the enzyme from protein-GPIs present on the endo-lysosomal system. A GPInegative phenotype can be obviated by sequestration of GPI-PLCp from its potential intracellular substrate. Although directed to endosomes, a catalytically inactive mutant (Q81L) of GPI-PLCp does not produce a protein-GPI deficiency in Leishmania because it does not cleave GPIs efficiently.
Endosomes, Glycosomes, and GPI Catabolism in L. major of GPI-PLCp are part of a novel targeting signal that directs GPI-PLCp to the endo-lysosomal system in L. major (Fig. 3); their mutation to Ser or Ala abolishes GPI-PLCp binding to the organelle. On the contrary, mutation of Cys-184 and Cys-347 did not block targeting of GPI-PLCp to endosomes (Fig. 5).
To determine what might be unique about Cys residues whose mutations redirected GPI-PLCp away from to the endolysosomal system in L. major, an alignment of 20 amino acids surrounding Cys-80 and Cys-273 was performed (Fig. 10). Each of the two Cys residues was found to be part of a motif [CS][CS]-x(0,2)-G-x(1)-C-x(2,3)-S-x(3)-L (PROSITE nomenclature) (64). The motif was not present around those Cys residues (i.e. Cys-184 and Cys-347) whose mutations did not influence endosomal targeting of GPI-PLCp. In the genome of L. major four proteins with the endosome targeting motif were found (Table I). It will be interesting to find out whether those L. major proteins (Table I)     Endosomes, Glycosomes, and GPI Catabolism in L. major