Specificities of Enzymes of Glycosylphosphatidylinositol Biosynthesis ...

THE JOURNAL OF BIOLOGICAL CHEMISTRY © 2002 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 277, No. 40, Issue of October 4, pp. 37147–37153, 2002 Printed in U.S.A.

Specificities of Enzymes of Glycosylphosphatidylinositol Biosynthesis in Trypanosoma brucei and HeLa Cells* Received for publication, April 8, 2002 Published, JBC Papers in Press, July 29, 2002, DOI 10.1074/jbc.M203371200

Terry K. Smith, Arthur Crossman, Michael J. Paterson‡, Charles N. Borissow‡, John S. Brimacombe, and Michael A. J. Ferguson§ From the Division of Biological Chemistry & Molecular Microbiology, The School of Life Sciences, University of Dundee, Dundee DD1 5EH, Scotland, United Kingdom

A series of synthetic analogues of D-GlcN␣1– 6-D-myoinositol-1-HPO4-sn-1,2-dipalmitoylglycerol, consisting of 22 variants of the D-GlcN or lipid components, were tested in trypanosomal and human (HeLa) cell-free systems. The assays measured the abilities of the analogues to act as substrates or inhibitors of the enzymes of glycosylphosphatidylinositol biosynthesis downstream of GlcNAc-phosphatidylinositol (GlcNAc-PI) de-N-acetylase. One compound, 4-deoxy-D-GlcN␣1– 6-D-myo-inositol1-HPO4-sn-1,2-dipalmitoylglycerol, proved to be an inhibitor of both the trypanosomal and HeLa pathways, whereas 4-O-methyl-D-GlcN␣1– 6-D-myo-inositol-1-HPO4sn-1,2-dipalmitoylglycerol and the 4ⴕ-epimer, D-GalN-␣1– 6-D-myo-inositol-1-HPO4-sn-1,2-dipalmitoylglycerol, were neither substrates nor inhibitors. The results with other analogues showed that the 6-OH of the ␣-D-GlcN residue is not required for substrate recognition in the trypanosomal and human pathways, whereas the 3-OH group is essential for both. Parasite-specific recognition of the ␤-linked analogue D-GlcN␤1– 6-D-myo-inositol-1-HPO4-sn1,2-dipalmitoylglycerol is striking. This suggests that, like the GlcNAc-PI de-N-acetylase, the trypanosomal glycosylphosphatidylinositol ␣-mannosyltransferases, inositol acyltransferse and ethanolamine phosphate transferase, do not recognize the 2-, 3-, 4-, and 5-OH groups of the D-myo-inositol residue, whereas the human inositol acyltransferase and/or first ␣-mannosyltransferase recognizes one or more of these groups. All of the various lipid analogues tested served as substrates in both the trypanosomal and HeLa cell-free systems, suggesting that a precise lipid structure and stereochemistry are not essential for substrate recognition. However, an analogue containing a single C18:0 alkyl chain in place of sn-1,2-dipalmitoylglycerol proved to be a better substrate in the trypanosomal than in the HeLa cell-free system. These findings should have a bearing on the design of future generations of specific inhibitors of the trypanosomal glycosylphosphatidylinositol biosynthetic pathway.

A significant proportion of eukaryotic cell-surface glycoproteins is attached to the plasma membrane by covalent linkage to a glycosylphosphatidylinositol (GPI)1 anchor. The structure * This work was supported by Wellcome Trust Program Grant 054491. 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. ‡ Both authors received Ph.D. studentships from the Biotechnology and Biological Sciences Research Council. § To whom correspondence should be addressed. Tel.: 44-1382344219; Fax: 44-1382-345764; E-mail: [email protected]. 1 The abbreviations used are: GPI, glycosylphosphatidylinositol; EtNThis paper is available on line at http://www.jbc.org

and biosynthesis of GPI membrane anchors and related molecules have been reviewed recently (1–5). The basic GPI core structure attached to protein is composed of NH2CH2CH2PO4H-6Man␣1–2Man␣1– 6Man␣1– 4GlcN␣1– 6D-myo-inositol-1-HPO4-lipid (EtN-P-Man3GlcN-PI), where the lipid can be diacylglycerol, alkylacylglycerol, or ceramide. This minimal GPI structure may be embellished with additional ethanolamine phosphate groups and/or carbohydrate side chains in a species- and tissue-specific manner (2). Protozoa tend to express significantly higher densities of cell-surface GPI-anchored proteins than do higher eukaryotes. For example, Trypanosoma brucei, the causative agent of African sleeping sickness, expresses a dense cell-surface coat consisting of ⬃5 ⫻ 106 dimers of a GPI-anchored variant surface glycoprotein, whose principal purpose is to protect the parasite from the alternative complement pathway of the host and, through antigenic variation, from specific immune responses (6). A variety of GPI-related structures, such as lipophosphoglycans, glycoinositolphospholipids, and mucin-like structures, are expressed by other trypanosomatid parasites (Refs. 1, 7, and 8 and references therein). Plasmodium (9, 10), Toxoplasma (11, 12), Trichomonas (13), and Entamoeba (14) also have abundant GPI-anchored glycoproteins and/or glycoinositolphospholipids. Inhibitors able to arrest the formation of GPIanchored proteins and/or GPI-related molecules on the plasma membrane of parasitic protozoa should prove useful in the development of antiparasitic agents. This notion has been validated, at least for T. brucei, where disruption of the TbGPI10 gene encoding the third mannosyltransferase of GPI anchor biosynthesis has been shown to be lethal for the bloodstream form of the parasite (15, 16).

P-Man 3 GlcN-PI, NH 2 CH 2 CH 2 PO 4 H-6Man ␣ 1–2Man ␣ 1– 6Man ␣ 1– 4GlcN␣1– 6-D-myo-inositol-1-HPO4-lipid; GlcN-PI, D-GlcN␣1– 6-D-myoinositol-1-HPO4-sn-1,2-dipalmitoylglycerol; GlcN-␤-PI, D-GlcN␤1– 6-Dmyo-inositol-1-HPO 4 -sn-1,2-dipalmitoylglycerol; GlcN-[L]-PI, D-GlcN␣1– 6-L-myo-inositol-I-HPO4-sn-1,2-dipalmitoylglycerol; 3-dGlcN-PI, 3-deoxy-D-GlcN␣1– 6-D-myo-inositol-1-HPO4-sn-1,2-dipalmitoylglycerol; 4-dGlcN-PI, 4-deoxy-D-GlcN␣1– 6-D-myo-inositol-1-HPO4sn-1,2-dipalmitoyl-glycerol; 6-dGlcN-PI, 6-deoxy-D-GlcN␣1– 6-D-myoinositol-1-HPO4-sn-1,2-dipalmitoylglycerol; GlcN4Me-PI, 4-O-methylD-GlcN␣1– 6-D-myo-inositol-1-HPO4-sn-1,2-dipalmitoylglycerol; GalNPI, D-GalN␣1– 6-D-myo-inositol-1-HPO4-sn-1,2-dipalmitoylglycerol; ManN-PI, D-ManN␣1– 6-D-myo-inositol-1-HPO4-sn-1,2-dipalmitoylglycerol; GlcN-PI(diC8), D -GlcN ␣ 1– 6- D -myo-inositol-1-HPO 4 -sn-1,2dioctylglycerol; GlcN-PI(diC18), D-GlcN␣1– 6-D-myo-inositol-1-HPO4sn-1,2-dioctadecylglycerol; GlcN-PI(Me,C18), D -GlcN ␣ 1– 6- D -myoinositol-1-HPO 4 -sn-1-octadecyl-2-methylglycerol; GlcN-I-P-C18, D-GlcN␣1– 6-D-myo-inositol-1-HPO4-octadecyl; MT-1, the first ␣-mannosyltransferase of the GPI biosynthetic pathway; HPTLC, high performance thin layer chromatography; JB␣M, jack bean ␣-mannosidase; PI, phosphatidylinositol; GPI-PLD, glycosylphosphatidylinositolspecific phospholipase D; PI-PLC, phosphatidylinositol-specific phospholipase C; NEM, N-ethylmaleimide; AHM, 2,5-anhydromannitol.

37147

37148

Substrate Specificity in GPI Biosynthesis

The sequence of events underlying GPI biosynthesis has been studied in T. brucei (17–22), Trypanosoma cruzi (23), Toxoplasma gondii (24), Plasmodium falciparum (25), Leishmania (26, 27), Saccharomyces cerevisiae (28, 29), Cryptococcus neoformans (30), and mammalian cells (5, 31–33). In all cases, GPI biosynthesis involves the addition of GlcNAc to phosphatidylinositol (PI) to give GlcNAc-PI, which is then de-N-acetylated to form GlcN-PI. Notable differences between the T. brucei and mammalian GPI biosynthetic pathways occur from GlcN-PI onwards, including the timing of inositol acylation and deacylation (21), the addition of extra ethanolamine phosphate groups to mammalian GPI anchors (31, 32), and fatty acid remodeling of T. brucei GPI anchors (18, 34). Inositol acylation (the transfer of fatty acid to the 2-OH group of the D-myo-inositol residue) of GlcN-PI either precedes or follows the action of the first mannosyltransferase (MT-1), as in mammalian cells and T. brucei, respectively (21, 35, 36). This difference was exploited in the discovery of the first generation of specific substrates (36) and inhibitors (37) of the T. brucei GPI biosynthetic pathway in vitro. Thus, GlcN-(2-Ohexadecyl)-PI was shown to inhibit MT-1, whereas GlcN-(2-Ooctyl)-PI and its N-acetylated version inhibited inositol acylation of Man1–3GlcN-PI and prevented the subsequent addition of an ethanolamine phosphate bridge (37). Another series of parasite-specific GPI pathway inhibitors containing L-myo-inositol inhibited T. brucei MT-1 (38), and very recently, two potent suicide substrate inhibitors of the parasite GlcNAc-PI de-N-acetylase have been reported (39). Another example of a species-specific inhibitor is a terpenoid natural product that inhibits yeast and human, but not parasite, GPI biosynthesis (40). In this paper, we investigate the substrate specificities of the T. brucei and HeLa GPI biosynthesis enzymes downstream of the de-N-acetylase, particularly the recently cloned mammalian and parasite MT-1s (41), with respect to the stereochemistry and anomeric configuration of the amino sugar residue and the nature of the lipid moiety of the PI component. EXPERIMENTAL PROCEDURES

Materials—GDP-[2-3H]mannose (15.6 –17.8 Ci/mmol) was purchased from PerkinElmer Life Sciences; jack bean ␣-mannosidase and glycosylphosphatidylinositol specific phospholipase D from were from Roche Molecular Biochemicals; and Bacillus thuringiensis phosphatidylinositol-specific phospholipase C was from Glyko. All other reagents were purchased from Merck-BDH or Sigma. Substrates and Substrate Analogues—D-GlcN␣1– 6-D-myo-inositol-1HPO4-sn-1,2-dipalmitoylglycerol (GlcN-PI) was synthesized as described previously (42), as were 3-deoxy-D-GlcN␣1– 6-D-myo-inositol-1HPO4-sn-1,2-dipalmitoylglycerol (3-dGlcN-PI), 4-deoxy-D-GlcN␣1– 6-Dmyo-inositol-1-HPO4-sn-1,2-dipalmitoylglycerol (4-dGlcN-PI), 6-deoxyD -GlcN ␣ 1– 6- D -myo-inositol-1-HPO 4 -sn-1,2-dipalmitoylglycerol (6dGlcN-PI), and 4-O-methyl-D-GlcN␣1– 6-D-myo-inositol-1-HPO4-sn-1,2dipalmitoylglycerol (GlcN4Me-PI) (43, 44). The synthesis of D-GalN␣1– 6- D -myo-inositol-1-HPO 4 -sn-1,2-dipalmitoylglycerol (GalN-PI), D -ManN ␣ 1– 6- D -myo-inositol-1-HPO 4 -sn-1,2-dipalmitoylglycerol (ManN-PI), D-GlcN␤1– 6-D-myo-inositol-1-HPO4-sn-1,2-dipalmitoylglycerol (GlcN-␤-PI), and D-GlcN␣1– 6-D-myo-inositol-1-HPO4-octadecyl (GlcN-I-P-C18) will be published elsewhere.2 The lipid-modified substrate analogues D-GlcN␣1– 6-D-myo-inositol-1-HPO4-sn-1,2-dioctylglycerol (GlcN-PI(diC8)), D-GlcN␣1– 6-D-myo-inositol-1-HPO4-sn-1,2dioctadecylglycerol (GlcN-PI(diC18)), and D-GlcN␣1– 6-D-myo-inositol1-HPO4-sn-1-octadecyl-2-methylglycerol (GlcN-PI(Me,C18)) were prepared as described (45). N-Acetyl derivatives of all of the forgoing compounds were prepared by standard procedures (46). The identity and purity of each of the synthetic analogues were assessed by negativeion electrospray mass spectrometry, and the concentration of each stock solution was determined by measuring the inositol content by selected ion-monitoring gas chromatography-mass spectrometry (46, 47). 2 A. Crossman, M. J. Paterson, and J. S. Brimacombe, unpublished data.

Preparation of Trypanosomal and HeLa Membranes—T. brucei and HeLa cell membranes (cell-free systems) were prepared as described previously (17, 36, 46). Trypanosome membranes were washed twice and suspended at 5 ⫻ 108 cell equivalents/ml in 2 ⫻ concentrated incorporation buffer supplemented with freshly prepared N-ethylmaleimide and n-octyl-␤-Dglucopyranoside (36, 46), unless stated otherwise. The lysate was sonicated briefly and added to dry GDP-[3H]Man (0.3 ␮Ci/2.5 ⫻ 107 cell equivalents). After brief sonication, aliquots of 20 ␮l (2.5 ⫻ 107 cell equivalents) were added to reaction tubes containing an equal volume of the various GlcN-PI analogues in 10 mM n-octyl-␤-D-glucopyranoside and incubated at 30 °C for 1 h. The glycolipid products were recovered and analyzed by HPTLC before and after enzymatic and chemical treatments. HeLa cell lysate was thawed, supplemented with 2.5 mM MnCl2, 2 mg/ml leupeptin, 0.1 mM 1-chloro-3-tosylamido-7-amino-2-heptanone, 1 ␮g/ml tunicamycin, 1 mM dithiothreitol, 100 ␮M CoA, and ATP-regenerating system (100 ␮M ATP, 5 mM phosphocreatine, and 5 units of creatine phosphokinase) and then added to dry GDP-[3H]Man (1 ␮Ci per 1 ⫻ 108 cell equivalents) (36). After brief sonication, aliquots of 100 ␮l were added to reaction tubes containing the GlcN-PI analogues and incubated at 35 °C for 1.5 h. The glycolipid products were recovered and analyzed by HPTLC before and after enzymatic and chemical treatments. Inhibition studies were performed in exactly the same manner, except that the membranes were preincubated with candidate inhibitors for 5 min prior to being added to GlcN-PI or GlcNAc-PI. Glycan Headgroup Analysis—The HPTLC-purified radiolabeled glycolipids were delipidated, deaminated, reduced, dephosphorylated with aqueous hydrofluoric acid, and desalted by passage through AG50X12(H⫹) and AG3X4 (OH⫺) ion-exchange resins. The resulting neutral glycan headgroups were analyzed by Bio-Gel P4 gel filtration (21, 47). HPTLC—Samples and glycolipid standards were applied to 10-cm aluminum-backed silica gel 60 HPTLC plates, which were developed using either solvent system A: chloroform/methanol/1 M NH4Ac/13 M NH4OH/water (180/140/9/9/23, v/v) or system B: chloroform/methanol/ water (10/10/3, v/v). Radiolabeled components were detected by fluorography at ⫺70 °C, after spraying with En3HanceTM, using Kodak XAR-5 film with an intensifying screen. Enzymatic and Chemical Treatments of Radiolabeled Glycolipids— Digestions with jack bean ␣-mannosidase (JB␣M), glycosylphosphatidylinositol-specific phospholipase D (GPI-PLD) and phosphatidylinositol-specific phospholipase C (PI-PLC) and base hydrolysis, deamination, and N-acetylation were performed as described previously (36, 46, 48). RESULTS

Substrate Specificities of T. brucei MT-1 and Downstream Enzymes—The trypanosomal cell-free system was incubated with N-ethylmaleimide (NEM) to prevent the formation of endogenous GPI intermediates by inhibiting UDP-GlcNAc:PI ␣1– 6 GlcNAc-transferase (49). Thus, labeling with GDP[3H]Man (Fig. 1A, lane 1) leads only to the formation of endogenous dolichol-phosphate-[3H]mannose, allowing easy assessment of the effects of adding synthetic GlcN-PI analogues. The addition of synthetic GlcN-PI or GlcNAc-PI primed the production of GPI intermediates up to and including EtN-PMan3GlcN-PI (glycolipid A⬘) (Fig. 1A, lanes 2 and 3), as described previously (46). Various novel GlcN-PI and GlcNAc-PI analogues were tested as substrates (Fig. 1, A and B). The following analogues failed to stimulate the formation of 3Hlabeled mannosylated glycolipids and are therefore not substrates for MT-1 and/or the GlcNAc-PI de-N-acetylase: Glc-PI, 2-dGlc-PI, 3-dGlcN(Ac)-PI, 4-dGlcN(Ac)-PI, GlcN(Ac)4Me-PI, (Fig. 1A, lanes 4 –11), GalN(Ac)-PI and ManN(Ac)-PI (Fig. 1B, lanes 3, 4, 9, and 10). The only analogues that delivered 3Hlabeled mannosylated products were 6-dGlcN-PI and 6-dGlcNAc-PI (Fig. 1B, lanes 5– 8). The slightly higher Rf values of the 3 H-labeled mannosylated glycolipids generated with 6-dGlcN(Ac)-PI (Fig. 1B, lanes 5– 8), compared with those of GlcN(Ac)PI (Fig. 1B, lanes 1 and 2), are consistent with the lower polarity of the 6-deoxy compounds.


FIG. 1. Substrate specificity of T. brucei MT-1 and downstream enzymes. A, the trypanosomal cell-free system was incubated with GDP-[3H]Man and NEM alone (lane 1) or with 10 ␮M GlcN-PI or GlcNAc-PI (lanes 2 and 3) or 30 ␮M Glc-PI, 2-dGlc-PI, 3-dGlcN-PI, 4-dGlcN-PI, GlcN4Me-PI, 3-dGlcNAc-PI, 4-dGlcNAc-PI, or GlcNAc4Me-PI (lanes 4 –11, respectively). B, the trypanosomal cell-free system was incubated with GDP-[3H]Man and NEM alone (lane 11) or with GlcN-PI, GlcNAc-PI, GalN-PI, GalNAc-PI, 6-dGlcN-PI, 6-dGlcNAc-PI, ManN-PI, or ManNAc-PI at the concentrations indicated (lanes 1–10, respectively). In all cases, the radiolabeled glycolipid products were analyzed by HPTLC and fluorography. The products descriptors on the left of the chromatograms are: DPM, dolichol-phosphatemannose; M1–3, Man1–3GlcN-PI; aM2–3, Man2–3GlcN-(acyl)PI; A⬘, EtNP-Man3GlcN-PI; C⬘, EtN-P-Man3GlcN-(acyl)PI. The product descriptors on the right of B followed by (6-d) refer to products based on 6-dGlcN-PI.

Analysis of the 3H-labeled mannosylated glycolipids formed from 6-dGlcNAc-PI (Fig. 1B, lanes 7 and 8) showed they were all sensitive to deamination, GPI-PLD, and base treatment, indicating they are all GPI intermediates containing an amino sugar and a diacylglycerol lipid anchor (Table I). Glycolipids aM2(6-d), aM3(6-d), and C⬘(6-d) were resistant to PI-PLC, indicating that they are inositol acylated, and all glycolipids, except A⬘(6-d) and C⬘(6-d), were sensitive to JB␣M, indicating the presence of at least one terminal ␣Man residue. The 3Hlabeled mannosylated glycolipids obtained from 6-dGlcN-PI were also subjected to headgroup analysis. Neutral glycan headgroups were obtained from HPTLC-purified glycolipids by deacylation, deamination, reduction, and dephosphorylation. The desalted products were analyzed by Bio-Gel P4 Gel filtration (Table I). The headgroups are ⬃0.2 glucose units smaller than those of authentic Man1–3(2,5-anydromannitol) (47), presumably due to the presence of the 6-deoxy group, which would give rise to 2–5-anhydro-6-deoxymannnitol on deamination of the 6d-GlcN residue. Taken together, these findings suggest that the novel 6dGlcNAc-PI analogue is a substrate for the de-N-acetylase, all three ␣-mannosyltransferases, the inositol acyltransferase, ethanolamine phosphate transferase, and ultimately, the inositol deacylase. The slight build up of the M2 species suggests that the third ␣-mannosyltransferase (MT-3) has a somewhat decreased affinity for Man26-dGlcN-PI compared with Man2GlcN-PI. The Trypanosomal GPI Pathway Has No Preference for the Anomeric Configuration of GlcN-PI—We have noted previously that trypanosomal GlcNAc-PI de-N-acetylase will act on the

37149

␤-(1– 6)-linked GlcN[3H]Ac-PI as well as on the natural ␣-anomer (39), but the abilities of downstream enzymes to recognize the ␤-linked analogue are unknown. Here, we show that the addition of either GlcNAc-␤-PI (Fig. 2A, lanes 4 and 5) or GlcN-␤-PI (Fig. 2A, lanes 6 –9) to the cell-free system in the presence of GDP[3H]Man leads to the formation of 3H-labeled mannosylated glycolipids in a dose-dependent manner. These novel 3H-labeled mannosylated glycolipids had Rf values identical to those formed by the addition of the natural substrates GlcN-PI and GlcNAc-PI (Fig. 2A, lanes 1–3). The sensitivity of the products to deamination, GPI-PLD, and base treatment (Fig. 2B, lanes 1, 3, and 6, respectively) showed that they are GPI species that contain a GlcN residue and diacylglycerol-PI. Only C⬘(␤) and A⬘(␤) were resistant to JB␣M treatment (Fig. 2B, lane 5), implying the presence of ethanolamine phosphate, while aM2(␤), aM3(␤), and C⬘(␤) were resistant to PI-PLC (Fig. 2B, lane 2), implying that they are inositol-acylated. Taken together, these results show that de-N-acetylation, mannosylation, inositol acylation, and ethanolamine phosphate addition can take place on GlcNAc-PI containing either an ␣- or a ␤-glycosidic linkage. Furthermore, both the ␣ and ␤ forms of GlcNAc-PI and GlcN-PI produce closely similar patterns of 3 H-labeled mannosylated products, suggesting that there is no preference for the natural ␣-glycosidic, over the unnatural ␤-glycosidic, linkage in any of the biosynthetic steps. Substrate channeling previously observed between the de-N-acetylase and MT-1 in the trypanosomal cell-free system (46) is such that the addition of GlcNAc-PI leads to greater 3H-labeled mannosylation than that of GlcN-PI (Fig. 2A, lanes 1 and 2); this effect was also observed with GlcNAc-␤-PI (Fig. 2A, compare lanes 4 and 8). Inhibition of Trypanosomal MT-1 by GlcN-PI Substrate Analogues—To assess whether GlcN-PI analogues that were not substrates for the T. brucei GPI pathway would act as inhibitors, the cell-free system was preincubated with these analogues before the addition of GlcN-PI and GDP-[3H]Man. The 3 H-labeled mannosylation of GlcN-PI was inhibited by preincubation with 12.5 ␮M 4-dGlcN-PI and 4-dGlcNAc-PI (Fig. 3, lanes 7 and 12) but not with 20 ␮M 3-dGlcN-PI, 3-dGlcNAc-PI, GalN-PI, and ManN-PI or with 20 or 50 ␮M GlcN4Me-PI and GlcNAc4Me-PI (data not shown). The lack of inhibition with the latter two compounds suggests that the 4⬘-O-methyl group interferes with binding to the active site of MT-1. A result similar to those obtained with the 4⬘-deoxy compounds was reported previously for the L-myo-inositol-containing analogue GcN-[L]-PI (38), i.e. inhibition was not seen until the concentration of inhibitor was greater than or equal to the concentration of the GlcN-PI substrate. The reason for this type of concentration dependence is unclear. Substrate Specificities of the HeLa Inositol Acyltransferase, MT-1, and Downstream Enzymes—The HeLa cell-free system was incubated with GDP-[3H]Man and all of the GlcN-PI analogues described previously. Inositol acylaton is a prerequisite for mannosylation in mammalian GPI biosynthesis; thus the formation of 3H-labeled mannosylated glycolipids in this assay depends on substrate recognition by the insoitol acyltrasferase as well as by MT-1. In the absence of added GlcN-PI, DPM is labeled along with the GPI intermediate H5 (EtN-P-Man1GlcN-(acyl)PI) formed from endogenous GlcN-(acyl)PI (Fig. 4A, lane 1). The addition of synthetic GlcN-PI increases the incorporation of [3H]mannose into glycolipids by priming the formation of H2 (Man1GlcN-(acyl)PI) and another form of H5.3 The only GlcN-PI analogues that produced 3H-labeled mannosylated 3 The Rf value of GPI intermediates primed with synthetic (dipalmitoyl)GlcN-PI are slightly lower than their endogenous counterparts due to differences in lipid structure (36).

37150


TABLE I Characterisation of the [3H]mannosylated glycolipids formed in the trypanosomal cell-free system with various analogues of GlcN-PI Abbreviationsa

JB␣M

PI-PLC

Baseb

Glycan sizec

Assignment

M1(6-d) M2(6-d) M3(6-d) aM2(6-d) aM3(6-d) C⬘(6-d) A⬘(6-d) M1(diC18) M2(diC18) M3(diC18) aM3(diC18) C⬘(diC18) A⬘(diC18) M3(diC8) aM3(diC8) C⬘(diC8) M3(Me,C18) aM3(Me,C18) C⬘(Me,C18) A⬘(Me,C18) M3(IPC18) aM2(IPC18) aM3(IPC18) C⬘(IPC18) A⬘(IPC18)

⫹ ⫹ ⫹ ⫹ ⫹ ⫺ ⫺ ⫹ ⫹ ⫹ ⫹ ⫺ ⫺ ⫹ ⫹ ⫺ ⫹ ⫹ ⫺ ⫺ ⫹ ⫹ ⫹ ⫺ ⫺

⫹ ⫹ ⫹ ⫺ ⫺ ⫺ ⫹ ⫹ ⫹ ⫹ ⫺ ⫺ ⫹ ⫹ ⫺ ⫺ ⫹ ⫺ ⫺ ⫹ ⫹ ⫺ ⫺ ⫺ ⫹

⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫺ ⫺ ⫺ ⫹(M3(diC18)) ⫹(A⬘(diC18)) ⫺ ⫺ ⫹(M3(diC8)) ⫹(A⬘(diC8)) ⫺ ⫹(M3(Me,C18)) ⫹(A⬘(Me,C18)) ⫺ ⫺ ⫹(M2(IPC18)) ⫹(M3(IPC18)) ⫹(A⬘(IPC18)) ⫺

2.1 3.0 3.9 ND 3.9 4.1 4.1 2.3 3.2 4.3 4.2 4.2 4.1 4.2 4.3 4.2 ND ND ND ND 4.3 3.1 4.1 4.2 4.1

Man16-dGlcN-PI Man26-dGlcN-PI Man36-dGlcN-PI Man26-dGlcN-(acyl)PI Man36-dGlcN-(acyl)PI EtNP-Man36-dGlcN-(acyl)PI EtNP-Man36-dGlcN-PI Man1GlcN-PI(diC18) Man2GlcN-PI(diC18) Man3GlcN-PI(diC18) Man3GlcN-(acyl)PI(diC18) EtNP-Man3GlcN-(acyl)PI(diC18) EtNP-Man3GlcN-PI(diC18) Man3GlcN-PI(diC8) Man3GlcN-(acyl)PI(diC8) EtNP-Man3GlcN-(acyl)PI(diC8) Man3GlcN-PI(Me,C18) Man3GlcN-(acyl)PI(Me,C18) EtNP-Man3GlcN-(acyl)PI(Me,C18) EtNP-Man3GlcN-PI(Me,C18) Man3GlcN-I-P-C18 Man2GlcN-(acyl)I-P-C18 Man3GlcN-(acyl)I-P-C18 EtNP-Man3GlcN-(acyl)I-P-C18 EtNP-Man3GlcN-I-P-C18

d

a

See Fig. 5 for corresponding bands. All of the glycolipids listed here were sensitive to GPI-PLD and deamination. The products of base treatment if sensitive are given in parentheses. c The size of the neutral glycans terminating in 2,5-anhydromannitol (AHM) were determined as described under “Experimental Procedures” and are expressed in glucose units. The published values for authentic Man␣1– 4AH, Man␣1– 6Man␣1– 4AHM, and Man␣1–2Man␣1– 6Man␣1– 4AHM are 2.4, 3.2, and 4.2 glucose units, respectively (52). ND, not determined d ⫹ indicates a positive digestion, i.e. loss (or shift) of the radiolabeled band: ⫺ indicates a negative digestion, i.e. no loss or shift of band. b

FIG. 2. Both GlcN-␣-PI and GlcN-␤-PI are substrates for the trypanosomal GPI pathway. A, the trypanosomal cell-free system was incubated with GDP-[3H]Man and NEM alone (lane 10) or with GlcN-PI, GlcNAc-PI, GlcNAc-␤-PI, or GlcN-␤-PI (lanes 1–9) at the concentrations indicated. B, the [3H]Man-labeled glycolipids formed from GlcNAc-␤-PI (lane 4) were subjected to deamination, PI-PLC, GPI-PLD, and JB␣M digestions (lanes 1–3 and 5) and base hydrolysis (lane 6). The products that partitioned into butan-1-ol were analyzed by HPTLC and fluorography. The product descriptors are the same as those in Fig. 1 except that the descriptor (␤) denotes products based on GlcN-␤-PI.

glycolipids were 6-dGlcN-PI (Fig. 4A, lane 6) and 6-dGlcNAc-PI (data not shown). Enzymatic and chemical digests of the mannosylated products, together with their Rf values, suggest that they are H2(6-d) and H5(6-d). This, in turn, implies that 6-

dGlcNAc-PI and 6-dGlcN-PI are substrates for the de-N-acetylase, inositol acyltransferase, MT-1, and the first ethanolamine phosphate transferase of the HeLa cell GPI pathway. None of the other GlcN-PI analogues, including GlcN-␤-PI, afforded 3 H-labeled mannosylated glycolipids in the HeLa cell-free system (Fig. 4A, lanes 3–5 and 7–9). Thus the substrate specificities of the HeLa and trypanosomal cell-free systems appear to be very similar except that GlcN-␤-PI is a specific primer of the trypanosomal GPI biosynthetic pathway. Inhibition of the HeLa Cell-free System by GlcN-PI Substrate Analogues—To assess whether GlcN-PI analogues that were not substrates for the HeLa GPI pathway would act as inhibitors, the cell-free system was preincubated with these analogues before the addition of 100 ␮M GlcN-PI and GDP[3H]Man. The formation of exogenous H2 and H5 was inhibited by preincubation with 100 ␮M 4-dGlcN-PI (Fig. 4B, lane 5) but not with 100 ␮M Glc-PI, 3-dGlcN-PI, GlcN4Me-PI, GalN-PI, ManN-PI, or GlcN-␤-PI (Fig. 4B, lanes 3, 4, and 6 –9). The inhibition by 4-dGlcN-PI cannot be ascribed to a nonspecific effect resulting from an increased concentration of synthetic lipid in the system, since 200 ␮M GlcN-PI is also effective in priming the pathway (Fig. 4B, lane 10). However, as with the trypanosomal system, 4-dGlcN-PI failed to inhibit the HeLa cell-free system at concentrations below that of the GlcN-PI substrate (data not shown). Interestingly, 4-dGlcN-PI does not appear to inhibit the labeling of endogenous H5 (compare Fig. 4B, lanes 1 and 5), suggesting that it inhibits the HeLa inositol acyltransferase rather than MT-1. The Effects of Lipid Structure on Substrate Recognition by the Trypanosomal and HeLa GPI Pathways—The trypanosomal cell-free system was assayed with a variety of GlcN-PI and GlcNAc-PI analogues with variations in the lipid structure. All of these analogues produced 3H-labeled mannosylated glycolipids (Fig. 5, A–D), which were characterized by their Rf values and JB␣M, PI-PLC, and base treatments and, in most cases, by headgroup analysis (Table I). The data show that all of the analogues produced GPIs up to and including the correspond-


37151

FIG. 3. Inhibition of trypanosomal MT-1 by GlcN-PI analogues. The trypanosomal cell-free system was incubated with GDP-[3H]Man alone (lane 1) or with GlcN-PI after preincubation with 4-dGlcN-PI (lanes 3–7) or 4-dGlcNAc-PI (lanes 8 –12) at the concentrations indicated. [3H]Man-labeled glycolipids were extracted and analyzed by HPTLC and fluorography. The product descriptors are the same as those in Fig. 1.

PI(diC8) and GlcN(Ac)-I-P-C18 are processed faster than GlcN(Ac)-PI, since early intermediates (i.e. M1 and M2) do not accumulate with these analogues (Fig. 5B, lanes 4 and 5; Fig. 5D, lanes 2, 3, 5, and 6). However, the absence of glycolipid A⬘(diC8) (Fig. 5B, lanes 4 and 5) suggests that the inositol deacylase may not act on glycolipid C⬘(diC8), for reasons that remain unclear. The poor processing of GlcN(Ac)-PI(Me,C18) is in contrast to the results obtained for the de-N-acetylation of GlcN[3H]Ac-PI(Me,C18), which showed the same kinetics as GlcN[3H]Ac-PI (39). It is by no means clear why the enzymes downstream of the de-N-acetylase should recognize GlcNPI(Me,C18) so poorly. Results similar to those described above were obtained with the HeLa system. All of the lipid analogues afforded 3H-labeled mannosylated products that were identified as the corresponding H2 and H5 products (data not shown). This indicates that they are all substrates for the HeLa cell de-N-acetylase and/or inositol acyltransferase, MT-1, and first ethanolamine phosphate transferase. The yields of H2 and H5 produced by the various lipid modified substrates are in the order GlcN-PI ⬃ GlcN-PI(diC18) ⬎ GlcN-PI(diC8) ⬃ GlcN-PI(Me,C18) ⬎ GlcNI-P-C18. It is worth noting that the simple analogue GlcN-I-PC18 is a much better substrate in the trypanosomal system than in the HeLa system. DISCUSSION

FIG. 4. Substrate specificity of HeLa inositol acyltransferase, MT-1, and downstream enzymes. A, the HeLa cell-free system was incubated with GDP-[3H]Man alone (lane 1) or with 100 ␮M GlcN-PI, 3-dGlcN-PI, 4-dGlcN-PI, GlcN4Me-PI, 6-dGlcN-PI, GalN-PI, ManN-PI, or GlcN-␤-PI (lanes 2–9). B, the HeLa cell-free system was incubated with GDP-[3H]Man either alone (lane 1) or with GlcN-PI before (lanes 2 and 10) or after preincubation with Glc-PI, 3-dGlcN-PI, 4-dGlcN-PI, GlcN4Me-PI, GalN-PI, ManN-PI, or GlcN-␤-PI (lanes 3–9, respectively). The products were extracted and analyzed by HPTLC and fluorography. The product descriptors are: DPM, dolichol-phosphate-mannose; H2, Man1GlcN-(acyl)PI; ENDO H5 and EXOG H5, endogenous and exogenous EtN-P-Man1GlcN-(acyl)PI derived from endogenous GlcN(acyl)PI or exogenous synthetic GlcN-PI, respectively.

ing glycolipid C⬘ analogues, and all, except GlcN(Ac)-PI(diC8), produced the corresponding glycolipid A⬘. It appears that the sn-1,2-diacylglycerol component of GlcN(Ac)-PI is not a stringent requirement for substrate recognition by any of the enzymes leading to glycolipid A⬘ and/or C⬘ and that it can be replaced by long chain (diC18:0) and short chain (diC8:0) sn1,2-dialkylglycerols (Fig. 5, A and B), sn-1-octadecyl-2-methylglycerol (Fig. 5C), and even by a simple C18:0 alkyl chain (Fig. 5D). In each case, the efficiency of product formation for the free amino and N-acetyl derivatives was compared with those of sn-1,2-dipalmitoylglycerol-containing GlcN-PI and GlcNAcPI. With the exception of GlcN(Ac)-PI(Me,C18), which were relatively poor substrates (Fig. 5C), the analogues were processed with similar efficiencies. It seems that GlcN(Ac)-

A number of GlcN-PI and GlcNAc-PI analogues, previously tested for recognition by T. brucei and HeLa GlcNAc-PI de-Nacetylases (39), have now been tested for recognition by the downstream enzymes of GPI biosynthesis in the same organisms. A number of salient conclusions can be drawn for these results, which together with information gained using other GlcN-PI analogues (36 –39, 46) are summarized in (Fig. 6). The mannosylation, inositol acylation, and addition of ethanolamine phosphate stemming from precursor GlcN-␤-PI by the trypanosomal (but not by the HeLa) cell-free system (Fig. 2A) may be rationalized in the same way as for the parasitespecific de-N-acetylation of GlcN[3H]Ac-␤-PI (39); by postulating that the precise disposition of the 2-, 3-, 4-, and 5-OH groups of the D-myo-inositol residue relative to the D-GlcN residue is not important for substrate recognition by any of the relevant trypanosomal enzymes. This is somewhat unexpected for the inositol acyltransferase, which attaches a fatty acyl group to the 2-OH of the D-myo-inositol residue. However, this might be explained if the parasite inositol acyltransferase was located in the lumen of the endoplasmic reticulum, which is essentially devoid of other PI phospholipids. In this location the GPI inositol acyltransferase would need only to be specific for the PI moiety of GPIs to provide the appropriate specificity. The inositol acyltransferase is yet to be isolated or cloned from any organism so that this proposal cannot be readily tested. But, it is worth pointing out that T. brucei inositol acyltrans-

37152


FIG. 5. The effect of lipid structure on substrate recognition by the trypanosomal GPI pathway. A, the trypanosomal cell-free system was incubated with GDP-[3H]Man alone (lanes 1 and 6) or with 20 ␮M GlcN-PI or GlcNAc-PI (lanes 2 and 3) or with 30 ␮M GlcN-PI(diC18) or GlcNAc-PI(diC18) (lanes 4 and 5). B, the trypanosomal cell-free system was incubated with GDP-[3H]Man alone (lane 1) or with 10 ␮M GlcN-PI or GlcNAc-PI (lanes 2 and 3) or with 10 ␮M GlcN-PI(diC8) or GlcNAc-PI(diC8) (lanes 4 and 5). C, the trypanosomal cell-free system was incubated with GDP-[3H]Man alone (lane 1) or with 20 ␮M GlcN-PI or 10 ␮M GlcNAc-PI (lanes 2 and 3, respectively) or with 30 ␮M GlcN-PI(Me,C18) or 30 ␮M GlcNAc-PI(Me,C18) (lanes 4 and 5, respectively). D, the trypanosomal cell-free system was incubated with GDP-[3H]Man alone (lane 4) or with GlcNAc-PI (lane 1) or GlcN-I-P-C18 (lanes 2 and 3) or GlcNAc-I-P-C18 (lanes 5 and 6) at the concentrations indicated. In all cases, the radiolabeled glycolipid products were analyzed by HPTLC and fluorography. The identities of the bands produced from GlcN(Ac)-PI are indicated on the left of the chromatogram (see legend to Fig. 1). The 3H-labeled mannosylated products obtained from the various GlcN-PI lipid analogues are indicated on the right of the chromatogram (see also Table I).

FIG. 6. Summary of the structural features of GlcN-PI required for processing in trypanosomal and HeLa cell-free systems.

ferase differs from the corresponding HeLa enzyme in a number of ways, notably with respect to its sensitivity to inhibition by phenylmethylsulfonyl fluoride, its use of a membrane-bound (non-acyl-CoA) acyl donor, and its strict requirement for prior mannosylation of the GlcN-PI substrate (21, 48). The latter

point may be particularly relevant, since the active sites of mammalian and T. brucei MT-1 have been suggested to be lumenally disposed in the endoplasmic reticulum (41). The cleavage of the 3H-labeled mannosylated products derived from GlcN-␤-PI by GPI-PLD (Fig. 2B, lane 3) is also

Substrate Specificity in GPI Biosynthesis worthy of note. Human serum GPI-PLD recognizes all natural GPI structures from GlcNAc-PI onwards, regardless of whether they are inositol acylated or not. The molecular models previously described (39) indicate that the spatial disposition of the GlcN(Ac) residue and the scissile phosphodiester bond (between the D-myo-inositol and the phosphatidyl residues) are almost identical in ␣- and ␤-linked GlcN(Ac)-PIs; these positions provide the key features recognized by GPI-PLD. By contrast, the spacial dispositions of the 2-, 3-, 4-, and 5-OH groups of D-myo-inositol are completely different in ␣- and ␤-linked GlcN(Ac)-PI; these groups appear to play no role in substrate recognition and this is consistent with the ability of GPI-PLD to cleave GPIs with or without inositol acylation (50). This study has uncovered another competitive MT-1 inhibitor in the form of 4-dGlcN-PI, although it is not parasitespecific. More importantly, the data reported here have established: (i) the 3-OH group unlike the 6-OH group of the of the GlcN residue is essential for substrate recognition by trypanosomal and HeLa MT-1 enzymes. (ii) Parasite-specific recognition of GPI intermediates containing ␤-GlcN, suggesting the that the 2-, 3-, 4-, and 5-OH groups4 of the D-myo-inositol residue are not essential for substrate recognition by trypanosomal MT-1. (iii) The efficient recognition by trypanosomal enzymes of substrates with a simple C18:0 alkyl chain linked directly to the D-myo-inositol-1-phosphate (replacing the synthetically complex chiral sn-1,2-diacylglycerol moiety). Future candidates as parasite-specific inhibitors might exploit these findings by removing non-essential hydroxyl groups, thus reducing the polarity of the compounds, and utilizing a simple alkyl chain in place of sn-1,2-diacylglycerol for ease of synthesis and for chemical and biological stability. We note that some of the results reported herein are at variance with some of the observations of Lehrman and colleagues (35, 51), who find that GlcN-␤-PI is a substrate for inositol acylation and for MT-1 in a mammalian (Chinese hamster ovary) cell-free system but not in a trypanosomal cell-free system; we find otherwise. However, their GlcN-␤-PI analogue contained sn-1,2-dioctanoylglycerol rather than sn-1,2-dipalmitoylglycerol, and this might affect substrate presentation. Also, in their experiments, a GlcN-PI containing sn-1,2-dioctanoylglycerol was more efficient than one containing sn-1,2-dipalmitoylglycerol in a trypanosomal cell-free system. In our experiments, there is little or no difference between sn-1,2dipalmitoylglycerol, sn-1,2-octadecylglycerol, and sn-1,2dioctylglycerol variants of GlcN-PI (Fig. 5, A, B, and D). The reasons for these apparent inconsistencies are not clear. Acknowledgment—We are grateful to Dr. A. P. Dix for his interest and a number of preliminary experiments. REFERENCES 1. Ferguson, M. A. J. (1999) J. Cell Sci. 112, 2799 –2808 2. Ferguson, M. A. J., Brimacombe, J. S., Brown, J. R., Crossman, A., Dix, A., Field, R. A., Gu¨ ther, M. L. S., Milne, K. G., Sharma, D. K., and Smith, T. K. (1999) Biochim. Biophys. Acta 1455, 327–340 3. McConville, M. J., and Menon, A. K. (2000) Mol. Membr. Biol. 17, 1–16 4. Morita, Y. S., Acosta-Serrano, A., and Englund PT. (2000) in Oligosaccharides in Chemistry and Biology-A Comprehensive Handbook (Ernst, P., Sinay, P., and Hart, G. W., eds) pp. 417– 433, Wiley-VCH, Weinheim, Germany 5. Kinoshita, T., and Inoue, N. (2000) Curr. Opin. Chem. Biol. 4, 632– 638 6. Cross, G. A. M. (1996) Bioessays 18, 283–291 7. McConville, M. J., and Ferguson, M. A. J. (1993) Biochem. J. 294, 305–324

4

A previous study suggested that the 3- and/or 5-OH groups of the residue might be involved in substrate recognition by trypanosomal MT-1 (38). The data reported here refines that model and suggest that this is not the case. D-myo-inositol

37153

8. Guha-Niyogi, A., Sullivan, D. R., and Turco, S. J. (2001) Glycobiology 11, 45R–59R 9. Gerold, P., Schofield, L., Blackman, M. J., Holder, A. A., and Schwarz, R. T. (1996) Mol. Biochem. Parasitol. 75, 131–143 10. Naik, R. S., Branch, O. H., Woods, A. S., Vijaykumar, M., Perkins, D. J., Nahlen, B. L., Lal, A. A., Cotter, R. J., Costello, C. E., Ockenhouse, C. F., Davidson, E. A., and Gowda, D. C. (2000) J. Exp. Med. 192, 1563–1576 11. Striepen, B., Zinecker, C. F., Damm, J. B., Melgers, P. A., Gerwig, G. J., Koolen, M., Vliegenthart, J. F. G., Dubremetz, J. F., and Schwarz, R. T. (1997) J. Mol. Biol. 266, 797– 813 12. Zinecker, C. F., Striepen, B., Geyer, H., Geyer, R., Dubremetz, J., and Schwarz, R. T. (2001) Mol. Biochem. Parasitol. 116, 127–135 13. Singh, B. N., Beach, D. H., Lindmark, D. G., and Costello, C. E. (1994) Arch. Biochem. Biophys. 309, 273–280 14. Moody-Haupt, S., Patterson, J. H., Mirelman, D., and McConville, M. J. (2000) J. Mol. Biol. 297, 409 – 420 15. Nagamune, K., Nozaki, T., Maeda, Y., Ohishi, K., Fukuma, T., Hara, T., Schwarz, R. T., Sutterlin, C., Brun, R., Riezman, H., and Kinoshita, T. (2000) Proc. Natl. Acad. Sci. U. S. A. 97, 10336 –10341 16. Ferguson, M. A. J. (2000) Proc. Natl. Acad. Sci. U. S. A. 97, 10673–10675 17. Masterson, W. J., Doering, T. L., Hart, G. W., and Englund, P. T. (1989) Cell 56, 793– 800 18. Masterson, W. J., Raper, J., Doering, T. L., Hart, G. W., and Englund, P. T. (1990) Cell 62, 73– 80 19. Menon, A. K., Schwarz, R. T., Mayor, S., and Cross, G. A. M. (1990) J. Biol. Chem. 265, 9033–9042 20. Menon, A. K., Mayor, S., and Schwarz, R. T. (1990) EMBO J. 9, 4249 – 4258 21. Gu¨ ther, M. L. S., and Ferguson, M. A. J. (1995) EMBO J. 14, 3080 –3093 22. Morita, Y. S., Acosta-Serrano, A., Buxbaum, L. U., and Englund, P. T. (2000) J. Biol. Chem., 275, 14147–14154 23. Heise, N., Raper, J., Buxbaum, L. U., Peranovich, T. M., and de Almeida, M. L. (1996) J. Biol. Chem. 271, 16877–16887 24. Striepen, B., Dubremetz, J.-F., and Schwarz, R. T. (1999) Biochemistry 38, 1478 –1487 25. Gerold, P., Jung, N., Azzouz, N., Freiberg, N., Kobe, S., and Schwarz, R. T. (1999) Biochem. J. 344, 731–738 26. Smith, T. K., Milne, F. C., Sharma, D. K., Crossman, A., Brimacombe, J. S., and Ferguson, M. A. J. (1997) Biochem. J. 326, 393– 400 27. Ralton, J. E., and McConville, M. J. (1998) J. Biol. Chem. 273, 4245– 4257 28. Flury, I., Benachour, A., and Conzelmann, A. (2000) J. Biol. Chem. 275, 24458 –24465 29. Grimme, S. J., Westfall, B. A., Wiedman, J. M., Taron, C. H., and Orlean, P. (2001) J. Biol. Chem. 276, 27731–27739 30. Franzot, S. P., and Doering, T. L. (1999) Biochem. J. 340, 25–32 31. Hirose, S., Prince, G. M., Sevlever, D., Ravi, L., Rosenberry, T. L., Ueda, E., and Medof, M. E. (1992) J. Biol. Chem. 267, 16968 –16974 32. Puoti, A., and Conzelmann, A. (1993) J. Biol. Chem. 268, 7215–7224 33. Chen, R., Walter, E. I., Parker, G., Lapurga, J. P., Millan, J. L., Ikehara, Y., Udenfriend, S., and Medof, M. E. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 9512–9517 34. Morita, Y. S., and Englund, P. T. (2001) Mol. Biochem. Parasitol. 115, 157–164 35. Doerrler, W. T., Ye, J., Falck, J. R., and Lehrman, M. A. (1996) J. Biol. Chem. 271, 27031–27038 36. Smith, T. K., Sharma, D. K., Crossman, A., Dix, A., Brimacombe, J. S., and Ferguson, M. A. J. (1997) EMBO J. 16, 6667– 6675 37. Smith, T. K., Sharma, D. K., Crossman, A., Brimacombe, J. S., and Ferguson, M. A. J. (1999) EMBO J. 18, 5922–5930 38. Smith, T. K., Paterson, M. J., Crossman, A., Brimacombe, J. S., and Ferguson, M. A. J. (2000) Biochemistry 39, 11801–11807 39. Smith, T. K., Crossman, A., Borissow, C. N., Paterson, M. J., Dix, A., Brimacombe, J. S., and Ferguson, M. A. J. (2001) EMBO J. 20, 3322–3332 40. Su¨ tterlin, C., Horvath, A., Gerold, P., Schwarz, R. T., Wang, Y., Dreyfuss, M., and Riezman, H. (1997) EMBO J. 16, 6374 – 6383 41. Maeda, Y., Watanabe, R., Harris, C. L., Hong, Y., Ohishi, K., Kinoshita, K., and Kinoshita, T. (2001) EMBO J. 20, 250 –261 42. Cottaz, S., Brimacombe, J. S., and Ferguson, M. A. J. (1993) J. Chem. Soc. Perkin Trans. 1, 2945–2951 43. Dix, A. P., Borissow, C. N., Ferguson, M. A. J., and Brimacombe, J. S. (2001) Tetrahedron Lett. 42, 117–119 44. Borissow, C. N., Smith, T. K., Ferguson, M. A. J., and Brimacombe, J. S. (2001) Tetrahedron Lett. 42, 121–123 45. Crossman, A., Brimacombe, J. S., Ferguson, M. A. J., and Smith, T. K. (1999) Carbohydr. Res. 321, 42–51 46. Smith, T. K., Cottaz, S., Brimacombe, J. S., and Ferguson, M. A. J. (1996) J. Biol. Chem. 271, 6476 – 6482 47. Ferguson, M. A. J. (1994) in Glycobiology: A Practical Approach (Fukuda, M., and Kobata, A., eds) pp. 349 –383, IRL Oxford University Press, Oxford 48. Gu¨ ther, M. L. S., Masterson, W. J., and Ferguson, M. A. J. (1994) J. Biol. Chem. 269, 18694 –18701 49. Milne, K. G., Ferguson, M. A. J., and Masterson, W. J. (1992) Eur. J. Biochem. 208, 309 –314 50. Roberts, W. L., Myher, J. J., Kuksis, A., Low, M. G., and Rosenberry, T. L. (1988) J. Biol. Chem. 263, 18766 –18775 51. Ye, J., Doerrler, W. T., Lehrman, M. A., and Falck, J. R. (1996) Bioorg. Med. Chem. Lett. 14, 1715–1718 52. Ferguson, M. A. J. (1992) in Lipid Modification of Proteins: A Practical Approach (Hooper, N. M., and Turner, A. J., eds) pp. 191–230, IRL Oxford University Press, New York