A novel pathway for phosphorylated oligosaccharide biosynthesis. Identification of an oligosaccharide-specific phosphate methyltransferase in dictyostelium discoideum.

The N-linked oligosaccharides on three lysosomal enzymes in Dictyostelium discoideum were found to contain mannose 6-phosphomethyl residues. We have identified and partially characterized a novel S-adenosylmethionine-dependent methyltransferase that is probably responsible for the synthesis of this unusual diester from Man-6-P. The enzyme selectively methylates the phosphate group of Man-6-P (Km 4.3 mM). Glucose-6-P and fructose-1-P are relatively poor acceptors; however, the enzyme is inactive against a broad array of other phosphorylated compounds. Using model di-, tri-, and pentasaccharide acceptors that include portions of the three different branches of high mannose-type oligosaccharides, we found that the enzyme prefers terminal alpha 1----2-linked Man-6-P residues (Km 0.15-1.25 mM) found on the known phosphorylated branches. The enzyme is membrane bound, has a neutral pH optimum and cofractionates on sucrose gradients with GlcNAc-1-P transferase, which resembles its mammalian counterpart, and is, presumably, the first enzyme in the phosphorylation pathway. Based on the substrate specificity and colocalization with GlcNAc-1-P transferase, the phosphate methyltransferase is likely to be responsible for the generation of mannose-6-phosphomethyldiester on Dictyostelium oligosaccharides.

phosphomonoester (10). Proteins that present an available Man-6-P can then bind to either the cation-independent or cation-dependent phosphomannosyl receptors (1).
The vegetative cells of Dictyostelium discoideum also synthesize abundant amounts of phosphorylated N-linked oligosaccharides where Man-6-P occurs as an acid-stable phosphomethyl diester (Man-6-P-OCH3)' (11)(12)(13). Lysosomal enzymes that contain an acid-stable phosphodiester bind selectively to the mammalian cation-independent phosphomannosyl receptor (14, 15) but not to the cation-dependent receptor (16,17). No comparable phosphomannosyl receptor has been found in Dictyostelium, and the role of Man-6-P in lysosomal enzyme targeting is unproven. The Dictyostelium and mammalian GlcNAc phosphotransferases closely resemble each other in that both require oligosaccharides with terminal a1-2 Man residues and both add the first GlcNAc-1-P to the a 1 4 branch of the acceptor oligosaccharide. They differ from each other in that the Dictyostelium enzyme does not preferentially recognize mammalian lysosomal enzymes (9,18). We wanted to identify the remaining enzymes involved in the biosynthetic pathway as a prelude to functional deletion analysis. Our previous studies showed that the methyl group is derived from S-adenosylmethionine (AdoMet) (19), but there was no kinetic evidence for the presence of an acidlabile GlcNAc-P-Man intermediate, or for the mammalian "uncovering enzyme" that generates the phosphomonoester. Moreover, there is no biochemical precedent for a phosphate methyltransferase. Here we report the identification of a membrane-bound, phosphorylated oligosaccharide-specific, AdoMet-dependent phosphate methyltransferase that cofractionates with GlcNAc-1-P transferase.

FIG. 1. N-Glycanase digestion and analysis of [rnethyZ-3H]methionine-labeled lysosomal enzymes.
Cells were labeled with [methyL3H]methionine in HL5 medium and a-mannosidase (panel A , 2800 cpm), @-glucosidase (panel B , 1850 cpm), and acid phosphatase (panel C, 3200 cpm) were precipitated from the medium or the cells (not shown). The immunoprecipitates were solubilized and incubated with or without 2 milliunits of N-glycanase. Following the digestion, both the control (open squares) and digests (closed circles) were analyzed by gel filtration chromatography on Sephadex G-50. The [3H]methyllabeled oligosaccharides released from amannosidase immunoprecipitated from the cells (panel D ) or the medium (panel E ) were pooled, desalted, and analyzed by QAE-Sephadex chromatography.
['HIMannose-labeled oligosaccharides from a-mannosidase (cells and medium) were similarly prepared and analyzed for comparison (panel F ) . The arrows on panel D refer to the beginning of step elutions with the indicated concentration of NaCl and represent oligosaccharides with 1, 2, 3, 4, 5, and >5 charges, respectively. Uncharged chains are eluted in the first three fractions. Crude subcellular fractionation of Man-6-P-dependent methyltransferase activity and QAE-Sephadex fractionation of 3Hmethylation products. Crude sonicates of cells were prepared as described under "Experimental Procedures" and fractionated by centrifugation into 100,000 X g pellet and supernatant. Transferase assays of each fraction were conducted with various amounts of material from unfractionated lysates (panel A ) or for the 100,000 X g (1 h) supernatant (panel B ) or for the 100,OOO X g (1 h) pellet (panel C) in the presence (0) or absence (0) of Man-6-P. The products synthesized in the crude sonicate were fractionated on QAE-Sephadex and eluted with 3 X 1.5-ml washes of 0, 35, or 125 mM NaCl in 2 mM Tris base. Authentic [3H]Man-6-P-OCH, which carries one negative charge elutes exclusively in the first two 35 mM NaCl fractions. Results are shown as picomoles of Man-6-POCH3 formed during the assay. Man-6-P-OCH3 was run in an adjacent lane. 0.5-cm fractions of both sample lanes were cut and counted. The region of the plate shown contained all of the radioactivity and non-labeled standards. Standards: I, Glc-6-P; 2, Man-6-P, 3, Gal-6-P; 4, Glc-6-SO,; 5, Man-6-S04; 6, Glc; and 7, Man. Lower panel, aliquots of the "-methylated product (A) or chemically synthesized [3H]Man-6-P-OCH3 (0) were heated at 95 "C for various periods in the presence of 1.0 N HCl and then analyzed as described under "Experimental Procedures." The percent of unhydrolyzed product was measured as a function of time.

Phosphate
Axenic strain A X 4 was maintained in synthetic growth medium (HL5) and used for all experiments. Cells were grown in 3 ml of HL5 medium for 2 days in the presence of 500 pCi/ml of [2-3H]mannose or 100 pCi/ml of [methyl-3H]methionine. The cells and the medium were separated by centrifugation, and a-mannosidase, @-glucosidase, and acid phosphatase were precipitated from both using an excess of monoclonal antibody as previously described (20,21). The labeled oligosaccharides were released by digestion with N-glycanase and separated on Sephadex G-50 and analyzed as described previously (12).
Preparation of Membranes-250 ml of cells were grown to mid-log phase (5 X lo6 cells/ml), harvested by centrifugation, washed once in 50 mM Tris-HC1, pH 7.5, and resuspended at 1-3 X 10' cells/ml in the buffer. Cells were broken by sonication for 6 X 5-s bursts, and unbroken cells (1-2%) were removed by centrifugation at 800 X g for 5 min. The membranes were collected by centrifugation at 375,000 X g for 20 min on a Beckman TL-100 ultracentrifuge, resuspended at 25-40 mg/ml protein in 50 mM Tris-HC1, and immediately stored at Phosphate Methyltransferme Assay-Assays using Man-6-P were conducted in 100-pl volume in 1.5-ml microfuge tubes and contained 100,000 cpm of [meth~l-~HIAdoMet at a final concentration of 10 p~, 5 mM Man-6-P, 50 mM Tris-HC1, pH 7.5, 0.1% Nonidet P-40, and 100-400 pg of membrane protein. Samples were incubated for 1 h at room temperature (22 "C) and then stopped by heating at 100 "C for 2 min. The sample was diluted to 1 ml and spun at 10,000 X g on a Microfuge, and the supernatant was loaded onto a 0.5 X 1.5-cm column of Dowex-50 in a Pasteur pipette that was mounted on top of a similar column containing QAE-Sephadex equilibrated in 2 mM Tris base. The tandem column set was washed with 5 ml of water, and the Dowex-50 column was discarded at this point. The QAE-Sephadex column alone was washed with 4 ml of 35 mM NaCl in 2 mM Tris base, and the effluent was collected into a single 22-ml scintillation vial and counted with 15 ml of aqueous-compatible scintillation fluid. Background values of samples run in the absence of Man-6-P were subtracted and usually amounted to 10-15%.
Oligosaccharide Phosphate Methyltransferme Assays-The synthesis of the various phosphorylated substrates was previously described and their structures were documented by 'H NMR (22)(23)(24)(25)(26). The incubation conditions using these substrates were similar to those described above except that Man-6-P was omitted, and the final reaction volume was 20 pl using 30-50 pg of membrane protein. The reaction mixture was passed over the Dowex-50 column and then over a C-18 Sepak cartridge which was washed with 11 ml of water. The last 1-ml water wash was collected and counted to measure washing efficiency. The hydrophobic-labeled product was eluted 5 ml of MeOH and counted in a single scintillation vial and corrected for quenching by MeOH. The amount of radioactivity in the final water wash was subtracted from the MeOH wash. The C-18 cartridges were reused 10-15 times or until the water wash background values increased significantly. In all of the kinetic experiments, the same column was used for each individual oligosaccharide species tested.
Characterization of Phosphate Methyltransferme Product-The normal assay procedure was scaled up 4-fold in protein and run for 3 h to generate a large amount of product for characterization studies. The 35 mM NaCl fraction was passed over Dowex-50 (H-form) to remove sodium, and an aliquot was spotted on a 20-cm flexible cellulose thin layer plate next to chemically synthesized [3H]Man-6-P-OCH, and developed in 5:5:1:3 butanol/pyridine/water/acetic acid. Both lanes were cut into 0.5" sections and counted. Other sugar standards (5 pg each) on the plate were located by silver staining. A similar procedure was performed for a plate developed in ethyl acetate/acetic acid/water, 63:2. The product was also characterized by following the kinetics of acid hydrolysis in 1 N HC1 at 100 "C. Samples containing 1000 cpm of product were mixed with 100 cpm of 35S04, hydrolyzed, and then the [,H]CH30H was removed in a shaker evaporator and the remaining radioactivity was normalized to 35S04 recovery. The conversion of [3H]Man-6-P-OCH3 to [3H]Man-6-P was monitored by the increase in the proportion of [,HI that was converted from 1 to 2 negative charges. Under these conditions, 4 0 % Man-6-P is dephosphorylated. Base hydrolysis was performed in 20 pl of 0.1 N NaOH at 37 "C for 1-4 h. Subcellular Fractionation-Cells were grown to mid-log phase, collected by centrifugation, washed in cold 50 mM Tris-HC1 pH 7.5 and then resuspended in the same buffer at 1.5-2 X 10' cells/ml and lysed by nitrogen cavitation at 1,000 psi. The unbroken cells were removed by 800 X g spin for 5 min, and lysate was spun at 10,000 X g for 20 min. This supernatant was removed and spun at 415,000 X g for 20 min to sediment a microsomal fraction (27). This pellet was resuspended in 2.0 ml of the same buffer using 5-10 strokes of a Dounce homogenizer and fractionated by two methods. In the first method, the resuspended pellet was layered on top of a step gradient consisting of 2.0 ml of sucrose at 0.88, 1.02. 1.17, 1.32, 1.45 M in 50 mM Tris-HC1, pH 7.5, and spun at 38,000 rpm for 2.5 h in an SW-50 rotor. This was similar to a previously described method (27). In the second method, the resuspended microsomal pellet was placed on top of a linear sucrose gradient from 25 to 45% in sucrose and spun at Man-6-P Gal-6-P GIc-&P GlcNAc.6-P Fru4-P Man-1-P Gal-1-P Glc-1-P Fru-1-P bos-1-P GIyc-P Rlb-5-P

FIG. 5. Methylation of various phosphorylated compounds.
Phosphate methyltransferase assays were conducted using indicated acceptor at 10 mM. All values were corrected for background with no added acceptor. 38,000 rpm for either 2.5 (24) or for 5.5 h. The fractions were analyzed for GlcNAc-phosphotransferase, Man-6-P phosphate methyltransferase both in the presence and the absence of Man-6-P. Assays for typical lysosomal enzymes including a-mannosidase and &glucosidase (14) showed that >98% of the activity failed to cofractionate with the above enzymes. No neutral pH a 1 4 2 Man processing mannosidase could be detected in these fractions.

RESULTS
Lysosomal Enzymes Contain Phosphomethylated Oligosaccharides-Previous studies showed that many anionic oligosaccharides in Dictyostelium contain Man-6-P-OCH, residues (13) and that individual lysosomal enzymes have an acidstable diester (11). To determine whether the acid-stable diester in these enzymes is also a phosphomethyl group, vegetative cells were labeled with [methyl-3H]methionine and they were immunoprecipitated from both the cells and the growth medium using monoclonal antibodies against a-mannosidase, @-glucosidase, or acid phosphatase. These three proteins contain most of the [methyl-3H]methionine label in the polypeptide chain and a portion in the phosphorylated oligosaccharides. The proteins were digested with proteasefree N-glycanase to release the N-linked oligosaccharides, and these were separated from the deglycosylated ,H-labeled peptide by gel filtration. Digestion specifically releases [,HI methyl-labeled material from each enzyme from the medium (shown in Fig. 1, panels A-C) and from the cells (data not shown). The [3H]methyl-labeled N-linked oligosaccharides from a-mannosidase which was precipitated from either the cells (Fig. 1,panel D) or the medium (panel E) were compared to similarly released [3H]mannose-labeled chains (panel F ) by QAE-Sephadex anion-exchange chromatography. The patterns are very similar except that no [,H]methyl label is found in neutral chains, as expected, since they do not contain Man-6-P (Fig. 1). The radioactivity was stable to mild-acid hydrolysis (0.01 N HC1, 100 "C, 30 min) and to base hydrolysis (0.1 N NaOH, 4 h at 37 "C), as previously shown for phosphomethylated oligosaccharides. We conclude that the mild acidstable phosphodiester previously described in three lysosomal enzymes is probably also a methylphosphodiester.
Preliminary Experiments-To examine the biosynthesis of the phosphodiester, we assumed that GlcNAc-1-P phosphotransferase catalyzes the first step of the pathway to form GlcNAc-1-P-6-Man. GlcNAc is probably then cleaved by another enzyme to generate the phosphomonoester. AdoMet may activate or even be required for this cleavage, but finally, "CH3 is transferred from AdoMet to Man-6-P to form Man-6-POCH3. To search for an enzyme capable of converting the diester to the monoester, we synthesized [6-3H]GlcNAc-1-P-6-ManaMe from UDP-[3H]GlcNAc and a-methylmannoside using the Dictyostelium GlcNAc phosphotransferase. Cellular extracts were then assayed for the release of [,H]GlcNAc in the presence or absence of AdoMet. A small amount of substrate was converted to [,H]GlcNAc by a membrane-bound activity at neutral pH (data not shown), but AdoMet did not stimulate the activity. The enzyme is metal-ion independent and is completely inhibited by GlcNAc-1-P, but not by other hexose 1or 6-phosphates. Further characterization of this enzyme is in progress, but its existence suggested that Man-6-P is directly phosphomethylated. To test this, [methyL3H] AdoMet was added to total cell lysates in the presence or absence of Man-6-P. Label was transferred to material that contained only a single negative charge based on its elution from QAE-Sephadex with 35 mM NaCl (Fig. 2, panel D). To determine if the transferase activity was membrane-associated, assays were done on total cell lysate, supernatant, and resuspended pellet from a 100,000 x g centrifugation. The results clearly showed that the activity was highly enriched in the pellet fraction (Fig. 2). All subsequent characterizations were done with similar membrane preparations. The activity in the soluble fraction was not further characterized.
Product Characterization-The ,H-methylated product with one negative charge (from Fig. 2, panel D) was analyzed by thin layer chromatography as shown in Fig. 3. The single major peak coincided with the synthetic standard [,H]Man-6-P-OCH3, and was well separated from other neutral and anionic sugars. Analysis in a second solvent system gave the same result (data not shown). The methylated product was stable to strong base hydrolysis (0.1 N NaOH, 4 h, 37 "C), but the methyl group was hydrolyzed with a tl/z of 65 min in 1 N HCl at 100 "C, the same as authentic [3H]Man-6-P-OCH3 (Fig. 3). Based on the Man-6-P dependence, charge, chromatographic behavior, base stability, and acid hydrolysis kinetics we conclude that the product formed from [methyl-,H] AdoMet in the presence of Man-6-P is Man-6-P-O[3H]CH3.
Optimization of Assay Conditions and General Churacteristics of the Enzyme-As shown in Fig. 4, the assay is linear for at least 5 h (panel A ) and between 0-600 pg of membrane protein (panel B ) . The pH optimum is 7.5 using either HEPES, Tris-HC1, or MES buffers (panel C), and the activity is highest at 22 "C (panel D), the normal growth temperature of Dictyostelium. The general characteristics of the methyltransferase activity are shown in Table I. It is not dependent on metal ions, but it is slightly activated by 1.0 mM Ca2+, Mn2+, or M P . It is equally active in Triton X-100 or Nonidet P-40 but is totally inhibited by octyl-a-glucoside, Zn2+ ions, and 0.2 mM S-adenosylhomocysteine, the product of AdoMetmediated methyltransferase reactions (26). Activity is partially inhibited by low concentrations of phosphate, sulfate, and acetate ions. The enzyme is totally stable for several months when stored at -70 "C in Tris-HC1 buffer at 20-40 mg/ml, but loses >50% of the activity when stored at 4 "C overnight. Repeated freezing and thawing appeared to increase activity by 50%. K , for Man-6-P is 4.3 mM and 5 p~ for AdoMet. The V,,, under optimal conditions is approximately 400 pmol/mg membrane protein (Table 11).
Alternate Acceptors-To determine whether the methyltransferase is specific for phosphomethylation of Man-6-P, a number of other compounds containing phosphomonoesters were examined. The results presented in Fig. 5 show that most of them are not methylated to a significant extent when present at 10 mM, but that a few other sugar phosphates including fructose-1-P and glucose-6-P could serve as weak acceptors; however, Man-6-P was clearly the best one (Fig.  6). Neither of these phosphorylated sugars (at 25 mM) competitively inhibits the methylation of the phosphorylated disaccharide described below (data not shown). We conclude that the methyltransferase has high specificity for Man-6-P.
Phosphomethyltramferase Activity on Synthetic Oligosaccharides-Phosphorylation occurs only on selected Man residues in both Dictyostelium and mammalian cells. To determine whether methyltransferase preferentially recognizes specific phosphorylated residues in the context of a larger oligosaccharide, we tested model di-, tri-, and pentasaccharide acceptors that include portions of known phosphorylated chains. These substrates are linked to a hydrophobic arm [(CH2)&OOCH3] for simple purification of the product on C-18 cartridges. To validate this method for our assay, we compared methyltransferase activity for a phosphorylated and non-phosphorylated mannobiose acceptor. The phosphorylated substrate was a much better acceptor for methyltransferase, and the product consisted almost entirely of material with one negative charge (85%), showing that the phosphate group was methylated (data not shown).
The results presented in Fig. 7 and in Table I1 show that phosphate methyltransferase has a marked preference for terminal Man-6-P residues that are linked a 1 4 2 to an underlying Man residue. The VmaX is about &fold higher, and the K,,, is >25 times lower than for Man-6-P. The linkage of the next sugar is also important. Those with a142 > a 1 4 >> a1+3 Man whether this is part of a trisaccharide or pentasaccharide acceptor. The enzyme methylates a diphosphorylated pentasaccharide, but its activity is about the same as an acceptor containing P on the a 1 4 3 branch, suggesting that phosphorylation on the branch inhibits methylation. As shown in Fig. 8, the substrate preference corresponds well, but not precisely (13,34), with known locations of methylphosphate residues in Dictyostelium. Since Man processing is rare and uncharacterized in vegetative Dictyostelium cells (32), it is not known when the terminal Man on branch I is removed. This may be a significant issue because phosphates located in subterminal positions are poor acceptors (Fig. 7). Moreover, the unusual kinetic curves suggest that their interaction with the transferase is different than the others with Man-6-P in terminal positions, which show typical saturation ) and highlighted residues on this oligosaccharide indicate that these are favored locations for phosphorylation and methylation as determined by previous studies (13,34) and the results presented here using model acceptors. Other residues may also be phosphorylated, but this has not been demonstrated.

I II
kinetics. These results provide further evidence that this methyltransferase is involved in phosphorylated oligosaccharide synthesis. Subcellular Localization of the Phosphate Methyltransferme Activity-We fractionated cells lysed by nitrogen cavitation into 10,000 X g and 100,000 X g pellets, and 100,000 X g supernatant as described previously for the isolation of membranes that sulfate N-linked oligosaccharides on endogenous acceptors (27). The distribution and specific activities of phosphate methyltransferase and GlcNAc-1-P transferase were compared to alkaline phosphatase and cAMP phosphodiesterase as markers for the plasma membranes. Under these lysis conditions, the lysosomal enzymes are solubilized. As shown in Table 111, >80% of both phosphotransferase and methyltransferase activities are found in the 100,000 X g pellet, compared to the plasma membrane markers which are about equally divided between the 100,000 X g and the 10,000 x g pellet. The two transferases cofractionate in continuous (not shown) or discontinuous sucrose gradients (Fig. 9). The 100,000 x g pellet was resuspended and applied to a gradient consisting of 0.88, 1.02, 1.17, 1.32, and 1.45 M sucrose, and individual fractions were assayed for GlcNAc phosphotransferase and phosphate methyltransferase. The activities cofractionate in the 1.17 and 1.02 M sucrose layers which is very similar to the distribution previously reported for N-linked oligosaccharide sulfation-competent membranes (27). Although Dictyostelium is thought to have a Golgi apparatus, no specific enzyme markers have been found.

DISCUSSION
Previous studies showed that Dictyostelium N-linked oligosaccharides released from a collection of secreted glycoproteins contain a novel Man-6-P-OCH3 phosphomethyldiester (13). Cells metabolically labeled with [methyl-3H]methionine produced labeled oligosaccharides that were essentially identical to those labeled with [3H]Man (19). Here, we confirmed that the acid-stable phosphodiesters previously described on the oligosaccharides of three frequently studied lysosomal enzymes is also Man-6-P-OCH3 (11,29). Very little of the Man-6-P occurs as either an acid-labile diester or a monoester in these molecules.
Protein, nucleic acid, and carbohydrate methylation are well known (30), but phosphomethylation has only been previously described in these oligosaccharides and in small nuclear RNA where it occurs as a 5'-end cap on GTP (33). Identification of the Dictyostelium phosphate methyltransferase activity is the first enzyme of its kind and appears to be selective for Man-6-P. Fructose-1-P, which is a close structural analogue of Man-6-P (31), and glucose-6-P appear to be much weaker acceptors, but neither of these inhibits the methylation of the preferred oligosaccharide substrates even at 25 mM.
The strongest evidence supporting the role of the methyltransferase in phosphorylated oligosaccharide biosynthesis is its marked preference for methyl group transfer to specific  di-, tri-, and pentasaccharides. The best acceptors have terminal Man-6-P a1-2 linked to the underlying Man. A similar preference is seen for in vitro assays of mammalian GlcNAc phosphotransferase (9). Trisaccharide acceptors in which the third sugar is linked either al-2 or a 1 4 are much better than when it is linked a1-3. The same preference holds true when the terminal Man-6-P is part of a branched pentasaccharide. These results clearly show a preference for either branches I or I11 of the oligosaccharide shown in Fig. 9.
Previous studies on Dictyostelium oligosaccharides have shown that these branches are the preferred sites of Man-6-P-OCH3 (30). The correlation is not exact, since that report shows that the terminal Man residue on branch I is missing. It is not known whether this Mannose trimming was due to biosynthetic processing or to extracellular degradation of the sugar chains.
When the Man-6-P is found in a subterminal position on the trisaccharide synthetic substrates, it is a relatively poor acceptor, even compared to Man-6-P. These substrates and those in which the third sugar is in a1+3 linkage, in fact, do not yield typical substrate saturation curves. Moreover, phosphorylation of the a1+3 arm inhibits the methylation of the preferred a 1 4 arm when the diphosphorylated pentasaccharide is used as the acceptor. These two observations might be explained by allosteric interactions or by strong, but nonproductive binding of the acceptors. Alternatively, poor methylation could simply be an artifact that results from unfavorable positioning of the hydrophobic aglycone relative to the a1+3 branch.
In vitro assays of mammalian, amoeba, and Dictyostelium GlcNAc phosphotransferase are dependent on the presence of terminal al-2-linked Man residues on the oligosaccharide acceptors (9). In mammalian cells, terminal and subterminal Man residues are phosphorylated in uiuo, but in vitro assays using Man a1-2 ManaMe disaccharide as an acceptor show that only the terminal al-2-linked Man is phosphorylated (35).
The Golgi apparatus is difficult to visualize in vegetative cells (32), but is clearly seen in later development (37). Many studies in vegetative cells have shown the existence of an intermediate compartment between the endoplasmic reticulum and the lysosomes where lysosomal enzyme oligosaccha-rides become resistant to endo-B-N-acetylglucosaminidase H digestion and acquire sulfate residues (38). These newly synthesized enzymes appear to gain Man-6-P, immediately (<2 min) prior to the addition of sulfate (39). It is possible that GlcNAc phosphotransferase and phosphate methyltransferase are located in the Golgi, and they may be good candidates for Golgi marker enzymes. The results of preliminary subcellular fractionation studies support this but further technical refinements are needed The function of phosphorylated oligosaccharides in lysosomal enzyme targeting in Dictyostelium is an unanswered question, primarily because there are no mutants strains that lack all Man-6-P. Many mutants show a partial loss of Man-6-P or a partial/complete loss of oligosaccharide sulfation, and still properly target their lysosomal enzymes, albeit more slowly. No phosphomannosyl receptor that binds phosphomethylated oligosaccharides has been identified (38), but similar studies have not yet been done with non-methylated intermediates. The similarities in substrate specificity of mammalian and Dictyostelium GlcNAc phosphotransferase are striking. Both recognize the same positions on the oligosaccharide acceptors and add the first and second GlcNAc-1-P residues to the same corresponding branches, respectively. They differ in that mammalian phosphotransferase recognizes the peptide portion of mammalian lysosomal enzymes, but the Dictyostelium enzyme does not. However, it is not known whether the Dictyostelium transferase recognizes the peptide portion of Dictyostelium lysosomal enzymes. As shown here, the methyltransferase has a strong preference for the known positions of phosphate groups. Based on these specificities and the fact that 98% of the Man-6-P occurs as Man-6-P-OCH, and not as GlcNAc-P-Man diester or Man-6-P monoester in lysosomal enzymes, it seems likely that phosphorylation/methylation are tightly coupled in vegetative cells. Typical N-linked oligosaccharide processing pathways usually produce variable amounts of the partially processed intermediates (40). It is very uncommon for 98% of the molecules to be converted into the fully mature form. The results found in Dictyostelium could imply that the reactions all occur without ever releasing the intermediates, such as a single multienzyme complex, or, alternatively, that the proper function of the phosphate group, whatever it is, requires methyl-diester covers. Now that the pathway is known, it will be possible to examine its regulation during development and to isolate mutants that are blocked at different steps in the pathway.