Glycosylphosphatidylinositol (GPI) structures are attached to many cell surface glycoproteins in lower and higher eukaryotes. GPI structures are particularly abundant in trypanosomatid parasites where they can be found attached to complex phosphosaccharides, as well as to glycoproteins, and as mature surface glycolipids. The high density of GPI structures at all life‐cycle stages of African trypanosomes and Leishmania suggests that the GPI biosynthetic pathway might be a reasonable target for the development of anti‐parasite drugs. In this paper we show that synthetic analogues of early GPI intermediates having the 2‐hydroxyl group of the D‐myo‐inositol residue methylated are recognized and mannosylated by the GPI biosynthetic pathways of Trypanosoma brucei and Leishmania major but not by that of human (HeLa) cells. These findings suggest that the discovery and development of specific inhibitors of parasite GPI biosynthesis are attainable goals. Moreover, they demonstrate that inositol acylation is required for mannosylation in the HeLa cell GPI biosynthetic pathway, whereas it is required for ethanolamine phosphate addition in the T.brucei GPI biosynthetic pathway.
The structure, function and biosynthesis of the GPI family of molecules has been extensively reviewed (McConville and Ferguson, 1993; Englund, 1993; Stevens, 1995; Takeda and Kinoshita, 1995; Udenfriend and Kokudula, 1995; Medof et al., 1996). The smallest GPI structure found attached to glycoproteins is NH2CH2CH2PO4H‐6Manα1‐2Manα1‐6Manα1‐4GlcNα1‐6myo‐inositol‐1‐PO4H‐lipid (EtNP‐Man3GlcN‐PI), where the lipid may be diacylglycerol, alkylacylglycerol or ceramide (McConville and Ferguson, 1993). This minimal GPI‐anchor structure may be embellished with additional ethanolamine phosphate groups and/or carbohydrate side chains in a species‐ and tissue‐specific manner (McConville and Ferguson, 1993).
The tsetse fly transmitted African parasite Trypanosoma brucei expresses a dense cell‐surface coat consisting of ∼5×106 GPI‐anchored variant surface glycoprotein (VSG) dimers (Cross, 1996). In addition to GPI‐anchored glycoproteins, other trypanosomatid parasites, such as Leishmania, Trypanosoma cruzi, Herpetomonas, Leptomonas and Phytomonas, express a wide variety of GPI structures known as glycoinositol phospholipids (GIPLs) (McConville and Ferguson, 1993; Routier et al., 1995; Redman et al., 1995 and references therein). GIPLs are metabolic end‐products expressed at the cell surface and are classified as type‐1 if they contain the motif Manα1‐6Manα1‐4GlcNα1‐6myo‐inositol‐1‐PO4H‐lipid or as type‐2 if they contain the motif Manα1‐3Manα1‐4GlcNα1‐6myo‐inositol‐1‐PO4H‐lipid or as hybrid if they contain both motifs (i.e. Manα1‐6(Manα1‐3)Manα1‐4GlcNα1‐6myo‐inositol‐1‐PO4H‐lipid) (McConville and Ferguson, 1993). The largest characterized type‐2 GIPL structures are the lipophosphoglycans (LPGs) of the Leishmania (McConville and Ferguson, 1993; McConville et al., 1995), which contain phosphosaccharide‐repeat domains. LPGs are known to be major virulence factors for these parasites (Turco and Descoteaux, 1992). GPI‐anchored glycoproteins and/or GIPLs are also abundant on non‐trypanosomatid protozoan parasites such as Plasmodium falciparum (Gerold et al., 1996), Toxoplasma gondii (Tomavo et al., 1989), Trichomonas (Singh et al., 1994) and Entamoeba (Bhattacharya et al., 1992). In contrast, higher eukaryotes express lower densities of GPI‐anchored glycoproteins and do not express GIPL structures.
Proteins destined to be GPI anchored are attached to a preassembled GPI precursor in the endoplasmic reticulum in exchange for a hydrophobic COOH‐terminal peptide (Udenfriend and Kokudula, 1995). The basic sequence of events in GPI precursor biosynthesis has been studied in T.brucei (Masterson et al., 1989, 1990; Menon et al., 1990; Güther and Ferguson, 1995), T.cruzi (Heise et al., 1996), T.gondii (Tomavo et al., 1992), P.falciparum (Gerold et al., 1994), Saccharomyces cerevisiae (Sipos et al., 1994) and mammalian cells (Hirose et al., 1992; Puoti and Conzelmann, 1993; Mohney et al., 1994 and references therein). Some features of the biosynthesis of GPI‐like GIPLs and the GPI‐like anchor of the LPG of Leishmania major have also been described (Proudfoot et al., 1995; Smith et al., 1997). In all cases, GPI biosynthesis involves the addition of GlcNAc to phosphatidylinositol (PI), to give GlcNAc‐PI, which is de‐N‐acetylated to form GlcN‐PI (Doering et al., 1989; Hirose et al., 1991; Stevens, 1993; Milne et al., 1994). In T.brucei, L.major and human (HeLa) cells, de‐N‐acetylation has been shown to be a prerequisite for the mannosylation of GlcN‐PI to form later GPI intermediates (Smith et al., 1996, 1997; Sharma et al., 1997). The GlcNAc‐PI de‐N‐acetylases from these organisms show similar substrate specificities (Sharma et al., 1997; Smith et al., 1997).
From GlcN‐PI onwards there are several significant differences among the T.brucei, L.major and HeLa GPI biosynthetic pathways. For example: (i) the L.major pathway leads predominantly to the formation of GIPL/LPG intermediates containing the type‐2 GIPL sequence Manα1‐3Manα1‐4GlcNα1‐6PI, whereas the T.brucei and HeLa pathways lead to the formation of GPI anchor intermediates containing the sequence Manα1‐6Manα1‐4GlcNα1‐6PI. (ii) Acylation of the 2‐hydroxyl group of d‐myo‐inositol [to form (acyl)PI‐containing intermediates] occurs with T.brucei and HeLa GPI intermediates, but not with L.major GPI intermediates. (iii) Acylation of d‐myo‐inositol occurs only after the first mannosylation of early T.brucei GPI intermediates (Güther and Ferguson, 1995) but appears to occur before mannosylation in HeLa cells (Hirose et al., 1992; Puoti and Conzelmann, 1993; Doerrler et al., 1996). (iv) Acylation of d‐myo‐inositol can be inhibited by PMSF in T.brucei but not in HeLa cells (Güther et al., 1994). (v) Additional ethanolamine phosphate groups are added to mammalian GPI intermediates during biosynthesis (Puoti and Conzelmann, 1993; Kamitani et al., 1992), whereas no such modifications are found in T.brucei or L.major. (vi) Only T.brucei performs fatty acid remodelling, a process whereby the sn‐2 and sn‐1 fatty acids of EtNP‐Man3GlcN‐PI are removed sequentially and replaced with myristic acid (Masterson et al., 1990).
The apparent difference between the acceptor substrate specificities of the parasite Dol‐P‐Man:GlcN‐PI α1‐4 mannosytransferase and the mammalian equivalent (Dol‐P‐Man:GlcN‐(acyl)PI α1‐4 mannosytransferase) noted in (iii) above prompted us to synthesize two analogues of early GPI intermediates, namely d‐GlcNAcα1‐6(2O‐methyl)d‐myo‐inositol‐1‐PO4H–3‐sn‐1,2‐dipalmitoylglycerol [GlcNAc‐(2‐O‐Me)PI] and d‐GlcNα1‐6(2O‐methyl)d‐myo‐inositol‐1‐PO4H–3‐sn‐1,2‐dipalmitoylglycerol [GlcN‐(2‐O‐Me)PI], in which the 2‐hydroxyl group of the inositol ring is blocked by methylation. Mass spectrometric studies of human CD52 (Treumann et al., 1995), T.brucei glycolipid C (Güther et al., 1996) and procyclic acidic repetitive protein (Treumann et al., 1997) have revealed that inositol acylation occurs exclusively on the 2‐hydroxyl group. Thus, the presence of the 2‐O‐methyl group in the substrate analogues should prevent inositol acylation and so allow the involvement of the inositol 2‐hydroxyl group and the role of inositol‐acylation in GPI biosynthesis to be assessed.
GlcN‐(2‐O‐Me)PI and GlcNAc‐(2‐O‐Me)PI act as mannose acceptors in the trypanosome cell‐free system
The trypanosome cell‐free system (Masterson et al., 1989) has been modified to probe the substrate specificity of Dol‐P‐Man:GlcN‐PI α1‐4 mannosyltransferase (Smith et al., 1996). The modified assay included N‐ethylmaleimide (NEM) to inhibit the UDP‐GlcNAc:PI α1‐6 GlcNAc‐transferase (Milne et al., 1992), thereby suppressing the production of endogenous mannose acceptors and allowing the [3H]mannosylation of exogenous substrates to be examined.
As described previously (Smith et al., 1996), the addition of GlcNAc‐PI produced [3H]Man‐labelled Man1‐3 GlcN‐PI, Man3GlcN‐(acyl)PI and EtNP‐Man3GlcN‐PI (glycolipid A′) (Figure 1, lane 2). By contrast, the addition of the substrate analogues GlcN‐(2‐O‐Me)PI (Figure 1, lanes 3–6) and GlcNAc‐(2‐O‐Me)PI (Figure 1, lanes 7–10) produced the novel [3H]mannosylated glycolipid species T1–T5. The yield of T5 is higher with GlcN‐(2‐O‐Me)PI and, in some experiments (e.g. see Figure 3), the yield of T4 is low and an additional glycolipid (T6) is observed when using this compound. The N‐acetylated analogue GlcNAc‐(2‐O‐Me)PI appeared to be more efficiently [3H]mannosylated than its non‐N‐acetylated counterpart GlcN‐(2‐O‐Me)PI, particularly at low (5–20 μM) substrate concentrations. Similar behaviour is observed with the natural substrates, where GlcNAc‐PI is typically mannosylated six times more efficiently than GlcN‐PI (Smith et al., 1996). As with the natural substrates, it would appear that the substrate analogues are best presented to the α‐mannosyltransferases via the GlcNAc‐PI de‐N‐acetylase enzyme. Evidence for the de‐N‐acetylation of GlcNAc‐(2‐O‐Me)PI prior to mannosylation is provided below.
Characterization of glycolipids T1–T4
The structures of glycolipids T1–T4 [generated from GlcNAc‐(2‐O‐Me)PI] were investigated using chemical and enzymatic treatments (Figure 2). A conventional series of GPI intermediates (Figure 1, lane 2) were treated in parallel in order to provide positive controls for each treatment (data not shown). Glycolipids T1–T4 showed a uniform increase in Rf‐values upon N‐acetylation (Figure 2, lane 1) and were sensitive to nitrous acid deamination (Figure 2, lane 6); this is consistent with each glycolipid containing a single amino group in the form of glucosamine. The glycolipids were resistant to PI‐PLC (Figure 2, lane 3), as would be expected from the presence of the 2‐O‐methyl group on the inositol ring, and were sensitive to the action of serum GPI‐PLD (Figure 2, lane 4), confirming their identities as GPI structures. The sensitivity of the bands to JBAM indicated that ethanolamine phosphate had not been added to any of the glycolipids.
In order to define the structures of the glycan headgroups of T1–T4, each glycolipid was purified by preparative HPTLC and after deacylation by mild alkaline hydrolysis was split into two aliquots. One aliquot was subjected in turn to deamination, reduction, dephosphorylation and re‐N‐acetylation, whereas the other was subjected to dephosphorylation and N‐acetylation. The former treatment converts GlcN into 2,5‐anhydromannitol (AHM) with simultaneous cleavage of the glycosidic linkage to the 2‐O‐methyl‐myo‐inositol residue (Ferguson et al., 1985), whereas the latter retains the 2‐O‐methyl‐myo‐inositol residue and converts GlcN into GlcNAc (Ralton et al., 1993). The sizes of the glycans isolated from glycolipids T1–T4 by the treatments, as measured by Bio‐Gel P4 gel‐filtration, are shown in Table I. The sizes of the AHM‐containing glycans were identical to those of authentic Manα1‐4AHM (for T1 and T4), Manα1‐6Manα1‐4AHM (for T2) and Manα1‐2Manα1‐6Manα1‐4AHM (for T3). The hydrodynamic volumes of the 2‐O‐methyl‐inositol‐containing glycans were also consistent with these assignments (Table I).
Since glycolipids T1 and T4 have the same glycan structure, differences between their Rf‐values (Figure 1) must be attributed to differences between their lipid structures. The most likely explanation is that T4 is a lyso‐form of T1. This was confirmed by phospholipase A2 digestion, which converted glycolipids T1, T2 and T3 into glycolipids T4, T5 and T6 (Figure 3, compare lanes 2 and 3). Mild alkaline hydrolysis of glycolipids T1–T6, followed by butan‐1‐ol/water partitioning, confirmed that the lipid components of all of the glycolipids were alkali labile (Figure 3, lane 1), consistent with the presence of either diacylglycerol (T1–T3) or monoacylglycerol moieties (T4–T6).
Taken together, the data suggest that glycolipids T1–T6 have the structures indicated in Table I, namely, Man1‐3GlcN‐(2‐O‐Me)PI (T1–T3) and Man1‐3GlcN‐lyso(2‐O‐Me)PI (T4–T6).
The formation of these glycolipids from GlcNAc‐(2‐O‐Me)PI shows that methylation of the 2‐hydroxyl group of the d‐myo‐inositol residue of GlcNAc‐PI has no effect on substrate recognition and turnover by the trypanosomal GlcNAc‐PI de‐N‐acetylase and α‐mannosyltransferases. On the other hand, methylation of this hydroxyl group appears to prevent the addition of ethanolamine phosphate to Man3GlcN‐(2‐O‐Me)PI. Finally, the Man1‐3GlcN‐(2‐O‐Me)PI products can undergo significant deacylation at the sn‐2 position of the diacylglycerol moiety, indicating that the presence of the 2‐O‐methyl group does not affect substrate recognition and turnover by a trypanosomal PLA2‐like activity.
GlcN‐(2‐O‐Me)PI and GlcNAc‐(2‐O‐Me)PI act as mannose acceptors in the Leishmania cell‐free system
The Leishmania cell‐free system (Brown et al., 1996) has been modified to probe the substrate specificity of Dol‐P‐Man:GlcN‐PI α1‐4 mannosyltransferase (Smith et al., 1997). The modified assay included dithiothreitol to stimulate the mannosylation of exogenous acceptor substrates.
As previously described (Smith et al., 1997), both Dol‐P‐Man and endogenous Manα1‐4GlcN‐PI (glycolipid E) were labelled with [3H]Man in the absence of an exogenous acceptor (Figure 4, lane 3), whereas the presence of exogenous GlcN‐PI or GlcNAc‐PI produced [3H]Man‐labelled Manα1‐4GlcN‐PI (glycolipid Y) and Manα1‐4GlcN‐lysoPI (glycolipid Z) (Figure 4, lanes 1 and 2). The addition of GlcN‐(2‐O‐Me)PI (Figure 4, lane 4) and GlcNAc‐(2‐O‐Me)PI (Figure 4, lane 5) produced two novel glycolipids, Y(2‐O‐Me) and Z(2‐O‐Me) having Rf‐values slightly higher than those of glycolipids Y and Z, repectively. Glycolipids Y(2‐O‐Me) and Z(2‐O‐Me) were characterized by chemical and enzymatic treatments (summarized in Table II) and were assigned the structures Manα1‐4GlcN‐(2‐O‐Me)PI and Manα1‐4GlcN‐lyso(2‐O‐Me)PI, respectively.
The formation of these glycolipids from GlcNAc‐(2‐O‐Me)PI shows that methylation of the 2‐hydroxyl group of the d‐myo‐inositol residue of GlcNAc‐PI does not prevent substrate recognition and turnover by the Leishmania GlcNAc‐PI de‐N‐acetylase and Dol‐P‐Man:GlcN‐PI α1‐4 mannosyltransferase. Furthermore, some deacylation at the sn‐2 position of the diacylglycerol moiety of Man1GlcN‐(2‐O‐Me)PI is evident, suggesting that the presence of the 2‐O‐methyl group does not prevent substrate recognition and turnover by a Leishmania PLA2‐like activity. As previously observed with the natural GlcN‐PI and GlcNAc‐PI substrates (Smith et al., 1997 and Figure 4, lanes 1 and 2), the presence of the N‐acetyl group results in increased mannosylation of GlcNAc‐(2‐0‐Me)PI compared with GlcN‐(2‐O‐Me)PI (Figure 4, compare lanes 4 and 5), suggesting a degree of substrate channelling between the de‐N‐acetylase and the α1‐4 mannosyltransferase.
GlcN‐(2‐O‐Me)PI and GlcNAc‐(2‐O‐Me)PI do not act as mannose acceptors in the HeLa cell‐free system
The HeLa cell‐free system (Hirose et al., 1992; Güther et al., 1994) has been modified to probe the substrate specificity of human Dol‐P‐Man:GlcN‐(acyl)PI α1‐4 mannosyltransferase (Sharma et al., 1997). The modified assay included CoA to stimulate inositol acylation of GPI intermediates (Stevens and Zhang, 1994).
As reported previously (Sharma et al., 1997), Dol‐P‐Man and endogenous EtNP‐Manα1‐4GlcN‐(acyl)PI (glycolipid H5) were labelled with [3H]Man in the absence of an exogenous acceptor (Figure 5, lane 1) whereas the addition of exogenous GlcN‐PI resulted in the formation of exogenous H2 (Manα1‐4GlcN‐(acyl)PI and exogenous H5 that has a slightly lower Rf‐value than that of endogenous H5 (Figure 5, lanes 2 and 7). In contrast, the addition of GlcN‐(2‐O‐Me)PI or GlcNAc‐(2‐O‐Me)PI did not result in the formation of any additional glycolipids (Figure 5, lanes 3 and 4). Furthermore, the presence of these substrate analogues (at an equimolar concentration to GlcN‐PI) did not affect the processing of GlcN‐PI to yield exogenous H2 and H5 (Figure 5, lanes 5 and 6), even when the cell‐free system was preincubated with the substrate analogues prior to the addition of GlcN‐PI (Figure 5, lanes 7–9).
These data show that methylation of the 2‐hydroxyl group of the d‐myo‐inositol residue of GlcN‐PI prevents substrate recognition and turnover by the HeLa cell Dol‐P‐Man:GlcN‐(acyl)PI α1‐4 mannosyltransferase. Furthermore they show that the 2‐O‐methyl group cannot substitute for the 2‐O‐acyl group in GlcN‐(acyl)PI with respect to substrate recognition.
The results presented in this paper are summarized in Figure 6. They support the following conclusions about GPI biosynthesis: (i) both the trypanosomal and Leishmania GlcNAc‐PI de‐N‐acetylases can tolerate methylation of the 2‐hydroxyl group of the d‐myo‐inositol residue. However, the failure of the HeLa cell‐free system to mannosylate GlcN‐(2‐O‐Me)PI (see below) prevented assessment of the ability of the HeLa cell de‐N‐acetylase to recognise GlcNAc‐(2‐O‐Me)PI. (ii) Both the trypanosomal and Leishmania Dol‐P‐Man:GlcN‐PI α1‐4 mannosyltransferases can tolerate methylation of the 2‐hydroxyl group of the d‐myo‐inositol residue, whereas the comparable HeLa cell enzyme cannot recognise this 2‐O‐methylated substrate. (iii) The trypanosomal Dol‐P‐Man: Man1GlcN‐PI α1‐6 mannosyltransferase and Dol‐P‐Man: Man2GlcN‐PI α1‐2 mannosyltransferase can tolerate methylation of the 2‐hydroxyl group of the d‐myo‐inositol residue, whereas the trypanosomal phosphatidylethanolamine:Man3GlcN‐PI ethanolamine phosphotransferase cannot recognise this 2‐O‐methylated substrate. (iv) Both the trypanosomal and Leishmania PLA2‐like enzymes can tolerate methylation of the 2‐hydroxyl group of the d‐myo‐inositol residue.
The observations made in point (ii) are perhaps the most significant since they show that GlcN‐(2‐O‐Me)PI and GlcNAc‐(2‐O‐Me)PI are selective substrates for the parasite GPI pathways. This suggests that the discovery and development of inhibitors that are selective for the parasite GPI pathways are attainable goals.
The observations made in points (i), (ii) and (iii) also clarify several fundamental features of GPI biosynthesis. Firstly, the suggestion that the trypanosomal Dol‐P‐Man:GlcN‐PI α1‐4 mannosyltransferase requires a free hydroxyl group at the 2‐position of the d‐myo‐inositol residue for substrate recognition (Güther and Ferguson, 1995; Smith et al., 1996) must be revised. This notion arose from the observation that inositol‐acylation of GlcN‐PI does not occur until after the first αMan residue is added (Güther and Ferguson, 1995). However, in the light of the successful mannosylation of GlcN‐(2‐O‐Me)PI, this could be reinterpreted in terms of the accessibility of substrates to the inositol‐acyltransferase. This view would be consistent with the suggestion of substrate channelling between the de‐N‐acetylase and the first α‐mannosyltransferase (Smith et al., 1996) such that GPI intermediates might only be available to the inositol acyltransferase after emerging from a de‐N‐acetylase/α‐mannosyltransferase complex.
Secondly, the prediction that inositol acylation is required in trypanosomes for efficient ethanolamine‐phosphate addition (Güther and Ferguson, 1995) appears to be correct. The presence of the methyl group in GlcN‐(2‐O‐Me)PI precludes 2‐O‐acylation of the inositol ring and appears to block completely ethanolamine‐phosphate addition. This observation suggests that the trypanosomal inositol acyltransferase might be a potential target for the development of potent GPI‐pathway inhibitors, particularly as the mammalian inositol acyltransferase appears to differ in its donor‐substrate and inhibition characteristics (Güther et al., 1994, 1996). However, since there is no evidence that Leishmania perform inositol acylation (Proudfoot et al., 1995 and references therein), an inositol acyltransferase inhibitor would not function as a general anti‐trypanosomatid agent.
Thirdly, the prediction that inositol acylation of GlcN‐PI is a prerequisite for mannosylation in HeLa (and other mammalian) cells appears to be correct. This notion was originally supported by the observations that the majority of GPI intermediates in mammalian cells are inositol acylated (Hirose et al., 1992; Puoti and Conzelmann, 1993) and that mannosylation is greatly stimulated by CoA (Stevens and Zhang, 1994; Sharma et al., 1997) or acyl‐CoA (Doerrler et al., 1996). In the latter study, the use of synthetic GlcN‐PI substrates provided strong evidence for this model but could not entirely rule out the possibility that mannosylation precedes inositol acylation. For example, acyl‐CoA‐dependent mannosylation would be observed if Dol‐P‐Man:GlcN‐PI α1‐4 mannosyltransferase underwent product inhibition such that mannosylation was undetectable until such inhibition was relieved by inositol acylation. In contrast, the inability of GlcN‐(2‐O‐Me)PI to undergo mannosylation by the HeLa cell‐free system provides more direct evidence that inositol 2‐O‐acylation is a prerequisite for mannosylation. It is worth noting that, unlike the trypanosomal and Leishmania cell‐free systems, there is no evidence for substrate channelling between the de‐N‐acetylase and mannosyltransferases in the HeLa cell‐free system (Sharma et al., 1997). This is also consistent with the need for GlcN‐PI to access the inositol acyltransferase prior to reaching the mannosytransferases.
The significance of point (iv) is less clear, because it is not known whether the PLA2‐like activities that act on the [3H]mannosylated products derived from exogenous substrates are specific for GPI structures. However, it is possible that the PLA2‐like activities are those of GPI fatty acid remodelling (Masterson et al., 1990) and/or myristate exchange (Buxbaum et al., 1996; Werbovetz and Englund, 1997) for the trypanosome cell‐free system and of LPG biosynthesis for the Leishmania cell‐free system.
Materials and methods
GDP‐[2‐3H]mannose (14.9–17.8 Ci/mmol) and En3HanceTM were purchased from Dupont NEN. Jack bean α‐mannosidase (JBAM) and pig pancreas phospholipase A2 (PLA2) were purchased from Boehringer Mannheim and Bacillus thuringiensis phosphatidylinositol‐specific phospholipase C (PI‐PLC) and Aspergillus phoenicis α‐mannosidase (APAM) from Oxford GlycoSystems. Whole human serum was used as a source of glycosylphosphatidylinositol specific phospholipase D (GPI‐PLD). n‐Octyl β‐d‐glucopyranoside was obtained from Calbiochem. Ion‐exchange resins (AG–50X12 and AG–3X4) were obtained from Bio‐Rad. All the other reagents were purchased from Merck‐BDH or Sigma.
Substrates and substrate analogues
d‐GlcNα1‐6d‐myo‐inositol‐1‐HPO4–3‐sn‐1,2‐dipalmitoylglycerol (GlcN‐PI) was synthesized according to Cottaz et al. (1993). d‐GlcNα1‐6d‐(2‐O‐methyl)myo‐inositol‐1‐HPO4–3‐sn‐1,2‐dipalmitoylglycerol [GlcN‐(2‐OMe)PI] was prepared according to Crossman et al. (1997). These compounds were N‐acetylated as described below for the radiolabelled glycolipids. The purity of the synthetic substrates was checked by negative ion electrospray mass spectrometry prior to use and the concentrations of stock solutions of the synthetic substrates were measured by analysis of the myo‐inositol content by GC‐MS, as described in Smith et al. (1996).
Preparation of Trypanosomes membranes
Bloodstream forms of T.brucei (variant MITat.1.4) were isolated from infected rats and mice. Trypanosome membranes (trypanosome cell‐free system) were prepared as previously described by Masterson et al. (1989), except that the cells were not pre‐incubated with tunicamycin prior to lysis. Aliquots (5×108 cells/ml) were snap‐frozen in liquid N2 and stored at −70°C.
Preparation of Leishmania major membranes
Leishmania major (V121) promastigotes were grown to 1.25×107 cells/ml in Schneider's medium supplemented with 10% heat‐inactivated fetal calf serum. The cells were pelleted, washed with ice‐cold phosphate‐buffered saline and suspended in 0.1 mM N‐α‐p‐tosyl‐l‐lysine chloromethyl ketone (TLCK) containing 1 μg/ml leupeptin to give a final density of 1×109 cells/ml. The cells were were disrupted twice in a nitrogen cavitation bomb at 2.8 MPa and an equal volume of 0.1 M HEPES (pH 7.4), 50 mM KCl, 10 mM MgCl2, 10 mM MnCl2, 20% (v/v) glycerol, 0.1 mM TLCK, 1 μg/ml leupeptin was then added (Brown et al., 1996). Aliquots (5×108 cell equivalents/ml) were snap‐frozen in liquid N2 and stored at −70°C.
Preparation of HeLa membranes
HeLa cells were grown at 37°C in Dulbecco's modified minimal essential medium supplemented with 10% fetal calf serum in a 5% CO2 atmosphere. The HeLa cell‐free system was prepared according to Güther et al. (1994) with the following modifications. Subconfluent HeLa cells were treated with 5 μg/ml tunicamycin for 2 h at 37°C and were harvested after incubation (10 min at 37°C) with phosphate buffered saline (PBS) containing 0.5 mM EDTA instead of trypsin. The cells were washed twice with 30 ml of PBS to remove EDTA and were then hypotonically lysed in water containing 0.1 mM TLCK and 0.1 μg/ml of leupeptin. An equal volume of 100 mM HEPES–NaOH buffer (pH 7.4), 50 mM KCl, 10 mM MgCl2, 0.1 mM TLCK, 0.1 μg/ml leupeptin and 20% (w/v) glycerol was added. Aliquots (1×107 cell equivalents/ml) were snap frozen in liquid N2 and stored at −70°C.
Trypanosome cell‐free system assay
Trypanosome membranes were washed twice in 0.1 M HEPES buffer (pH 7.4), containing 25 mM KCl, 5 mM MgCl2, 0.1 mM TLCK and 2 μg/ml leupeptin, and were then suspended at 5×108 cell equivalents/ml in 2× concentrated incorporation buffer: 0.1 M HEPES (pH 7.4), 50 mM KCl, 10 mM MgCl2, 10 mM MnCl2, 20% (v/v) glycerol, 2.5 μg/ml tunicamycin, 0.2 mM TLCK and 2 μg/ml leupeptin (Masterson et al., 1989). Unless stated otherwise, the 2× concentrated incorporation buffer was supplemented with freshly prepared 0.2 M N‐ethylmalemide (NEM) and 10 mM (0.3% w/v) n‐octyl β‐d‐glucopyranoside. The resuspended lysate was vortexed, briefly sonicated and added to a tube containing dry GDP‐[3H]Man (0.3 μCi per 107 cell equivalents). After sonication for 1 min, aliquots of 20 μl (1×107 cell equivalents) were withdrawn and added to the reaction tubes containing an equal volume of 10–100 μM solutions of the various GlcN‐PI analogues in 10 mM n‐octyl β‐d‐glucopyranoside. The reaction tubes were incubated at 30°C for 1 h whereafter the reactions were terminated by the addition of 270 μl of chloroform:methanol (1:1, v/v). The glycolipid products were recovered in the chloroform/methanol/water‐soluble fraction, which was evaporated and partitioned between butan‐1‐ol and water, as previously described (Smith et al., 1996). Aliquots of the butan‐1‐ol phase containing the glycolipid products were subjected to HPTLC analysis both before and after enzymatic and chemical treatments.
Leishmania major cell‐free system assay
Leishmania major membranes were thawed and washed twice as described for the trypanosome assay. The pelleted membranes were suspended in 2× incorporation buffer, as described above, except that the the buffer was supplemented with 2 mM DTT instead of NEM and did not contain n‐octyl β‐d‐glucopyranoside. The suspension of membranes was added to a tube containing dry GDP‐[3H]Man (0.5 μCi per 2×108 cell equivalents) and sonicated for 1 min. Aliquots of 50 μl (2×108 cell equivalents) were added to the reaction tubes containing an equal volume of 60 μM solutions of the various GlcN‐PI analogues in water. The reaction tubes were incubated at 30°C for 1 h. After termination of the reactions by the addition of 666 μl of chloroform: methanol (1:1, v/v), the glycolipids were extracted and processed as described above.
HeLa cell‐free system assay
HeLa cell lysate was thawed and supplemented with 0.5 mM DTT, 1 mM Coenzyme‐A, 10 mM ATP, 5 mM MnCl2, 2 μg/ml leupeptin, 0.1 mM TLCK and 1 μg/ml tunicamycin. Aliquots of 100 μl (1×106 cell equivalents) were added to tubes containing dry GDP‐[3H]Man (2.5 μCi) and the synthetic GlcN‐PI analogues to give a final concentration of 100 μM. The reaction tubes were incubated at 35°C for 1.5 h whereafter and the glycolipids were extracted and processed as described above. Note: the results in Figure 5 and elsewhere (Sharma et al., 1997) show that the HeLa cell‐free system used in these studies only generates the mannosylated GPI intermediates H2 and H5, even if excess UDP‐GlcNAc is added to the membranes instead of GlcN‐PI (data not shown). This is different from the results of Hirose et al. (1992) and our own results (Güther et al., 1994) where larger GPI intermediates were observed. This discrepancy indicates that the HeLa cell‐free system is culture and/or membrane‐preparation dependent with respect to the observed final products.
Samples and glycolipid standards were applied to 10 cm aluminium‐backed silica gel 60 HPTLC plates (Merck) which were developed with chloroform/methanol/1 M ammonium acetate/13 M ammonium hydroxide/water (180:140:9:9:23, v/v), except for the HPTLC plate in Figure 1 that was developed using chloroform/methanol/water (10:10:3, v/v). Radiolabelled components were detected by fluorography at −70°C using Kodak XAR–5 film and an intensifying screen after spraying the plates with En3HanceTM.
Enzyme treatments of radiolabelled glycolipids
Digestions with APAM, JBAM, PI‐PLC and GPI‐PLD and processing of the products for analysis by HPTLC were performed as described previously (Güther et al., 1994; Smith et al., 1996). Pig pancreas PLA2 digests were performed at 37°C in 40 μl of 25 mM Tris–HCl (pH 8.0), 2 mM CaCl2, and 0.1% sodium deoxycholate with the addition of 8 U of enzyme at hourly intervals over 3 h, followed by further incubation at 37°C for 12 h.
Chemical treatments of radiolabelled glycolipids
Deamination of glycolipids was carried out in 20 μl of 0.1 M sodium acetate (pH 4.0), containing 0.01% Zwittergent 3–16. Aliquots (10 μl) of freshly prepared 0.5 M NaNO2 were added at hourly intervals with incubation at 60°C for 4 h. Lipidic products were extracted into butan‐1‐ol for analysis by HPTLC.
Glycolipids were N‐acetylated at 0°C in 100 μl of saturated NaHCO3 by the addition of three aliquots (2.5 μl) of acetic anhydride over 20 min. The reaction mixture was warmed to room temperature and N‐acetylated glycolipids were extracted into butan‐1‐ol. Residual salts were removed by washing the butan‐1‐ol phase with water.
Radiolabelled glycolipids from the trypanosome assay were purified by preparative HPTLC. They were eluted from the excised silica with chloroform/methanol/water (10:10:3, v/v), dried and delipidated by incubation (5 h, 50°C) with 300 μl of concentrated aqueous ammonia/50% propan‐1‐ol (1:1, v/v). The radiolabelled soluble glycan products were recovered in the aqueous phase of a butan‐1‐ol/water partition and were treated in one of two ways: (i) deamination, reduction, aqueous HF dephosphorylation, re‐N‐acetylation and desalting by passage through AG50X12(H+) over AG3X4(OH−) ion‐exchange resins (Ferguson, 1992) to yield neutral glycans terminating in 2,5‐anhydromannitol (AHM). (ii) Aqueous HF dephosphorylation, N‐acetylation and desalting by passage through AG50X12(H+) over AG3X4 (OH−) ion‐exchange resins (Ferguson, 1992) to yield neutral glycans containing GlcNAc and 2‐O‐methyl‐myo‐inositol.
The neutral glycans resulting from these procedures were dissolved in water containing glucose oligomer internal standards and the aqueous solution was filtered through a 0.2 μm membrane and analysed by Bio‐Gel P4 gel filtration using an Oxford Glycosystems GlycoMap. Fractions (250 μl) were collected and counted for radioactivity.
The biochemical work was supported by a programme grant from the Wellcome Trust and the synthetic work by a project grant from the MRC. M.A.J.F. is a Howard Hughes International Research Scholar. Deepak Sharma thanks the MRC and Alex Dix thanks the BBSRC for their PhD studentships.
- Copyright © 1997 European Molecular Biology Organization