the journal of biollxical c~istry no. 4, 1994 the … · glycosylphosphatidylinositols synthesized...

THE JOURNAL OF BIOLLXICAL C ~ I S T R Y 0 1994 by The American Soeiety for Biochemistry and Molecular Biology, Inc

Vol. 269, No. 4, Issue of January 28, pp. 2597-2606, 1994 Printed in U.S.A.

Glycosylphosphatidylinositols Synthesized by Asexual Erythrocytic Stages of the Malarial Parasite, Plasmodium falciparum CANDIDATES FOR PLASMODIAL GLYCOSYLPHOSPHATIDYLINOSITOL MEMBRANE ANCHOR PRECURSORS AND PATHOGENICITY FACTORS*

(Received for publication, March 31, 1993, and in revised form, September 6, 1993)

Peter GeroldS, Angela Dieckmann-Schuppert, and Ralph T. Schwarz From Zentrum fur Hygiene und Medizinische Mikrobiologie, Philipps- Universitat Marburg, Robert-Koch-Strasse 17, 35037 Marburg, Germany

Plasmodium falciparum is the causative agent of malaria tropica in man. Biochemical studies were focused on the asexual, intraerythrocytic stages of E? falciparum, because of their role in the clinical phase of the disease and the possibility of propagation in a cell culture system. In this report, we describe the in-culture labeling of malarial glycolipids and the analysis of their hydrophilic moieties. They were identified as glycosylphosphatidylinositols (GPIs) by: l) labeling with [SH]mannose, [SH]glucosamine, and [SHlethanolamine and 2) sensitivity toward glycosylphosphatidylinositol- specific phospholipase D, phospholipase A2, and nitrous acid. Malarial GPIs are shown to be unaffected by treatment with phosphatidylinositol-specific phospholipase C, regardless of prior treatment with mild base com- monly used for inositol deacylation. Two candidates for putative GPI-anchor precursors to malarial membrane proteins with the structures ethanolamine-phosphate- 6(Manal-2)Manal-2Manal-6Mana14GlcN-P1 (Pfgl a) and ethanolamine-phosphate-6Manal-2Manal-6Man- al-4-GlcN-PI (Pfgl /3) were identified.

Malaria is the disease caused by parasitic protozoa of the genus Plasmodium. Plasmodium falciparum, the most danger- ous of the four species that are naturally infectious to man, still remains the cause of one of the most important diseases in tropical and subtropical areas. In 1985, the WHO reported 4.8 million documented cases of malaria, and more than 50% of the world’s population live in areas where malaria is endemic (Lockyer and Holder, 1989).

Glycosylphosphatidylinositols (GPIs)’ are a recently discov- ered class of glycolipids that anchor proteins and sugars into

*This study was supported by Deutsche Forschungsgemeinschaft Grants Schw 296/4-1, Schw 296/4-2, and Schw 296/4-3, Fonds der Chemischen Industrie, ARC from British CounciVDAAD, Hessisches Ministerium fiir Wissenschaft und Kunst, and P. E. Kempkes Founda- tion Marburg, Germany. This communication is presented as a partial fulfillment of the requirements of the doctoral thesis of P. G. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertise- ment” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

This work is dedicated to the memory of Bernard Fournet (1940- 1993) to whom the field of glycoconjugates owes so much.

$ Recipient of a scholarship of the Hessische Graduiertedorderung. To whom correspondence and reprint requests should be addressed: Tel.: +49-6421-285149 or +49-6421-285493; Fax: +49-6421-285482.

MSP, major surface protein; PI-PLC, phosphatidylinositol-specific phos- The abbreviations used are: GPI, glycosylphosphatidylinositol;

pholipase C; PIA,, phospholipase Az; GPI-PLD, glycosylphosphatidylinositol-specific phospholipase D; HPAEC, high pH anion exchange chromatography; Dol(?)-P-Man, dolichol(?)-phosphate-mannose (see Footnote 2); Pfg,, Z? fakiparum glycolipid; AHM, 2,5-anhydromannitol.

the plasma membrane, in a wide range of organisms. A comparison of the glycan structures of GPI membrane anchors indicates that the glycan parts of GPI-anchors contain a conserved structure consisting of ethanolamine-P04-6Manal- 2Manal-GManal-4GlcN (for review see Cross (1990)).

A huge variety of modifications on the periphery of the core glycan structure have been observed. However, the core glycan of protein-linked GPIs is highly evolutionarily conserved throughout mammalian and protozoan cells. Therefore, it is likely that identical or similar biosynthetic pathways are used in all eukaryotes.

The biosynthesis of the trypanosoma1 GPI has been elucidated using a cell-free system from lkypanosoma brucei brucei (Masterson et al., 1989; Doering et al., 1990; Menon et al., 1990a) and was also elucidated for some other biological systems (Puoti et al., 1991; Stevens and Raetz, 1991; Tomavo et al., 1992). GPI synthesis starts with the transfer of GlcNAc from UDP-GlcNAc to PI, followed by de-N-acetylation. Three mannose residues, each derived from dolichol-phosphate mannose, are then added, yielding MansGlcN-PI. Phosphoethanolamine, derived from phosphatidylethanolamine (Menon et al., 1993), is added to the terminal mannose to complete the GPI-anchor precursor which is thought to be added to the protein in exchange for a C-terminal hydrophobic domain in a transamidase reaction (Cross, 1990).

Like many parasitic protozoa, I! falciparum has a complex life cycle involving an insect vector (Anopheles spp. ). During the development in man, the intraerythrocytic, asexual stages produce the merozoite surface protein 1 (MSPl), the merozoite surface protein 2 (MSPO), and other proteins (Holder et al., 1985; Haldar et al., 1985; Haldar et al., 1986; Smythe et al., 1988; Braun-Breton et al., 1988), which were shown to be membrane-anchored involving diacylglycerol. These data lead to the proposition that some I! falciparum surface proteins are anchored by a glycosylphosphatidylinositol (GPI) membrane anchor to the parasite membrane (Haldar et al., 1985; Haldar et al., 1986; Schwarz et al., 1986; Schwarz et al., 1987).

I t has recently been demonstrated that the GPI moiety may also play an important role in the pathology of severe malaria, in addition to its role in membrane anchoring of surface proteins. Purified C-terminal GPI derived from MSPl and MSP2 by pronase hydrolysis induces tumor necrosis factor and inter- leukin 1 production by macrophages and regulates metabolism in adipocytes. When administered to mice in uiuo, it induces cytokine release, transient pyrexia, and hypoglycemia and under some conditions elicits cachexia (Schofield and Hackett, 1993).

In this report we provide evidence that the biosynthesis of GPIs in the parasitic protozoan I! falciparum involves a hydro- phobically modified inositol ring and therefore shows a simi- larity to the pathways described for mammalian cells (for a

2597

2598 Glycosylphosphatidylinositols in Plasmodium falciparum

review see Englund (1993)). The malarial pathway leads to two putative GPI-anchor precursors (ethanolamine-phosphate- 6Manal-2Mana1-6Mana14GlcN-PI and ethanolamine-phosphate-6(al-2Man)Manal-2Mana1-6Mana14GlcN-PI) whose glycan structures differ only in the number (3 versus 4) of mannose residues.

EXPERIMENTAL PROCEDURES Materials-~-[6-~H]Glucosamine hydrochloride (26.0 Ci/mmol), ~ 4 2 -

3Hlmannose (13.6 Ci/mmol), [9,10-3H]myristic acid (50.0 Ci/mmol), [9,10-3Hlpalmitic acid (53.4 Ci/mmol), my~-[~H]inositol (108.4 Ci/ mmol), and [l-3Hlethan-l-ol-2-amine hydrochloride (29.5 Ci/mmol) were purchased from Amersham. Tunicamycin was purchased from Calbiochem. Phosphatidylinositol-specific phospholipase C (PI-PLC) from Bacillus cereus and phospholipase A, (PLA,) from bee venom were from Boehringer Mannheim. Jack bean a-mannosidase was from Sigma, and Aspergillus saitoi a-mannosidase was obtained from Oxford Glycosystems. Ion exchange resins were analytical grade and purchased from Bio-Rad. All sugar and lipid standards were obtained from Sigma. All solvents were analytical or high performance liquid chromatography grade.

Parasite-The I! falciparum strain FCBR was obtained from Dr. B. Enders, Behring Co., Marburg, Germany. It was maintained in RPMI 1640 medium supplemented with 25 m Hepes, 21 m sodium bicarbonate, 0.37 m hypoxanthine, 0.1 mg/ml neomycin, and 10% defibri- nated (v/v) human A+ fresh frozen plasma (modified from Trager and Jensen (1976)). The cultures contained human A+ erythrocytes (leuko- cyte-depleted erythrocyte concentrates, Blood Donation Center, Univer- sity of Marburg, Germany) at 5% hematocrit and were incubated at 37 "C in gas-tight containers (modular incubation chamber, Flow Labo- ratories) being gassed with a mixture of 90% nitrogen/5% oxygen/5% carbon dioxide. Development and multiplication of the cultures were followed by microscopic evaluation of Giemsa-stained smears. Synchro- nous development of the parasite was achieved by the sorbitol method (Lambros and Vanderberg, 1979).

Metabolic Labeling of l? falciparum Dophozoites-About 5 x lo9 parasite-infected erythrocytes (10% parasitemia) were incubated with 250 pCi/5 ml of radiolabeled precursors. Labeling was performed in glucose-free RPMI medium (Amimed, Muttenz, Switzerland) supplemented with 10 m fructose (Schwarz et al. 1986), 25 m Hepes, 21 m sodium bicarbonate, 0.37 m hypoxanthine, 0.1 mdml neomycin, 0.1 lll~ methionine, 2 m glutamine, and 11 m glutathione. Radioactively labeled fatty acids were coupled to defatted bovine serum albumin at a molar ratio of 1:l. my~-[~H]Inositol labeling was performed in RPMI (Life Technologies Inc. Selectamine kit) without vitamins. All labeling experiments were carried out at 37 "C for 1-2 h. Controlling the viabil- ity of the parasites by microscopy of Giemsa-stained blood smears, we observed that the parasites are viable during the labeling period but die if the labeling time is prolonged for more than 4 h. For some experiments, labeling was camed out in the presence of 1 pdml tunicamycin.

Extraction and Purification-Washed parasites were harvested by saponin lysis and centrifugation at 4 "C (Goman et al., 1982). Less polar glycolipids were extracted twice with 2 ml of chlorofodmethanol (CM 2:1, v:v) per lo9 parasites to remove neutral lipids and phospholipids. This extraction was camed out for 30 min at room temperature under sonication in a water bath sonicator (Branson 3200, 47 MHz). The residual pellet was then extracted twice with 1 ml of chlorofod methanovwater (CMW 10:10:3, v:v:v) to obtain the more polar glycolipids. The CM extracts were pooled and subjected to repeated Folch wash- ing (Sharma et al., 1974). The CMW-extracted glycolipids were dried in a Speed-Vac (Savant Inc.), and the dried residue was partitioned between water and water-saturated l-butanol. The organic phase is termed "butanol component."

Preparation of GPI Standards from II: brucei-Ttypanosomes (Z brucei MIT variant clone 118) were purified from infected rat blood and labeled with [3Hlglucosamine, [3H]mannose, or PHlmyristic acid (Me- non et al., 1988). Alternatively, washed preparations of lysed trypano- somes were labeled with GDP-~-[2-~H]ma~ose (2 pCi) or UDP-D-N-[~- 3H]acetylglucosamine (2 pCi) (Masterson et al. 1989). All labeling experiments were camed out in the presence of tunicamycin. After the labeling, GPI-anchor precursors P2 and P3 (named glycolipid A and C by Krakow et al. (1986)) and their biosynthetic intermediates were obtained by sequential extraction and purified by TLC analysis as described by Menon et al. (1988).

Solvent Systems for TLC-The solvents for TLC were as follows: A, chlorofodmethanolcetic acidwater (25:15:4:2, by volume); B, chlorofodmethanovwater (10:10:2.5, by volume); C, chlorofod

methanoUO.l% KC1 (10:10:3, by volume). Analysis and Purification of Glycolipids-The Folch-washed CM ex-

tracts and the butanol components were analyzed on Silica Gel 60 TLC plates (Merck) using solvent systems A and B, respectively. After chromatography, the plates were dried and scanned for radioactivity using a Berthold LB 2842 automatic "LC scanner. The areas corresponding to the peaks of radioactivity were scraped, and the glycolipids were extracted twice with chlorofodmethanol (2:1, v:v), twice with chlorofodmethanollwater (10:10:3, v:v:v), and once with methanov pyridinelwater (2:1:1, v:v:v) for 30 min each under sonification. The yield of purified glycolipids was estimated to be approximately 40%.

Acid Hydrolysis for Component Analysis-TLC-purified glycolipids were dried to the bottom of a Reacti-vial (Pierce) and hydrolyzed for 4 h in 4 N HCl at 100 "C. HCl was removed from the samples by repeated flash evaporation with 100 pl of methanol. The released monosaccharides were analyzed by Dionex-HPAEC (Dionex, Sunnyvale, CA) using elution program 2 (see below).

Alkaline Saponification of Ester-like Bound Fatty Acids-Malarial glycolipids were dried and saponified in 200 pl of 0.05 N KOWMeOH (1:l; v:v) for 4 h a t 100 "C. The reaction was terminated by drylng the reaction mixture in a Speed-Vac evaporator. After phase partition between water and water-saturated l-butanol, the cleavage rate was estimated by liquid scintillation counting of aliquots of the organic and aqueous phases.

Enzymatic Cleavage of Glycolipids-The radiolabeled glycolipid mixtures or TLC-purified glycolipids were dried and redissolved in deter- gent-containing buffer for the corresponding enzymatic reactions. Phos- phatidylinositol-specific phospholipase C (PI-PLC) digestions were camed out in 100 pl of 0.1 M Tris-HC1 (pH 7.4), containing 0.1% sodium deoxycholate and 1 unit of B. cereus PI-PLC (Boehringer Mannheim). In some cases, TLC-purified glycolipids were treated with ammonia (32% NHJMeOH (l:l, v/v); 1 h, 37 "C) prior to PI-PLC digestion (Mayoret al., 1990). This treatment is supposed to selectively cleave palmitic acid at the inositol ring. Glycosylphosphatidyl-specific phospholipase D (GPI- PLD) treatment was performed under identical conditions, but in this case the incubation buffer contained 2 m CaCl,; 10 pl of normal rabbit serum was used as a source of GPI-PLD (Davitz et al., 1988). Phospho- lipase A, (PLA,) reactions were performed in the same incubation buffer but with 1 m CaClz and 50 units of the bee venom enzyme. After incubation for 14 h (PI-PLC, PI-PLA,) or 3 h (GPI-PLD) at 37 "C, the reactions were terminated by heating at 100 "C for 1 min, and the reaction products were extracted twice with 100 pl of water-saturated 1-butanol. The radioactivity in aliquots of the aqueous and organic phases were counted. The organic phases were analyzed by TLC using solvent system B.

Nitrous Acid Deamination-Glycolipid mixtures or TLC-purified glycolipids were dried and resuspended in 100 pl of freshly prepared 0.1 M sodium acetate (pH 4.01, containing 0.05% SDS, 0.25 M NaN02, and incubated for 14 h at ambient temperature. The reaction was terminated by the addition of 5 pl of 6 N HCI and extracted as described above. The organic phases were analyzed on TLC. The radioactivity released into the aqueous phases was analyzed on Bio-Gel P4 as described below.

Bio-Gel P4 Gel Filtration Analysis-Water-soluble products generated by enzymatic or nitrous acid treatment of radiolabeled glycolipids were sized on Bio-Gel P4 columns (1 x 130 cm, -400 mesh) in 0.2 M ammonium acetate, 0.02% sodium azide (Masterson et al., 1989). Frac- tions (about 850 pl) were collected at a rate of 1 fraction per 24 min. Glucose oligomers from partially hydrolyzed dextran were included as internal standards and detected by oxidation with 2 mg/ml orcinol in concentrated sulfuric acid.

Generation and Analysis of Core Glycans from Glycolipids-TLC- purified glycolipids were dephosphorylated, deaminated, and reduced according to a modification of the method described by Mayor and Menon (1990). Briefly, the glycolipids were dephosphorylated by treatment with ice-cold 48% aqueous HF for 60 h at 0 "C. After neutralization with frozen saturated LiOH, the reaction mixture was desalted, deaminated by HNO, treatment for 3 h, and reduced by sodium boro- hydride for 5 h. The reduced material was desalted on a tandem ion exchange column and filtered through a 0.4-pm and 0.25-w filter. The desalted and filtered core glycans were analyzed by Dionex HPAEC using gradient elution program 1 (Mayor and Menon, 1990).

Anion Exchange HPAEC Analysis of Labeled Glycans-Desalted glycans were analyzed by anion exchange chromatography on a Dionex Basic Chromatography System. The separation was accomplished by gradient elution (program 1) using a Carbopak"?" PA1 (4 x 250 m m ) column. Monosaccharides were analyzed using isocratic conditions (pro-

Glycosylphosphatidylinositols gram 2). Mono- and oligosaccharide standards were detected by pulsed amperometric detection. The elution programs for Dionex HPAEC were: 1) 100% bufTerA(0.1 M NaOH), 0% buffer B (0.1 M NaOH, 0.25 M NaOAc) up to 6 min after injection, then an increase of buffer B to 30% at 36 min, at a flow rate of 1 d m i n and 2) isocratic 15 m~ NaOH at a flow rate of 1 dmin.

Exoglycosidase Digestion of Glycolipids-For mannosidase digestion, TLC-purified glycolipids were deaminated and desalted as described above. The resulting water-soluble material was resuspended in 100 pl of 50 m~ sodium acetate (pH 4.5) containing 0.2 m~ ZnClz, 0.02% sodium azide, and 1.7 units of jack bean a-mannosidase and incubated at 37 "C for 24 h. Additional enzyme (0.85 unit) was added 14 h &er the beginning of the incubation. For A. saitoi a-mannosidase digestion of dephosphorylated, deaminated, and reduced glycolipids, the glycan samples were resuspended in 30 pl of 0.1 M sodium acetate (pH 5) containing 10 microunits of A. saitoi a-mannosidase and incubated for 14 h at 37 "C.

All digestions were stopped by heating at 100 "C for 1 min. The reaction mixtures were desalted, filtered, and analyzed by Bio-Gel P4 or by Dionex HPAEC chromatography.

Acetolysis ofNeutra1 Glycans-Radiolabeled glycans obtained via dephosphorylation, deamination, and reduction of labeled glycolipids were cleaved by a modification of the method described by Mayor and Menon (1990). This procedure preferentially cleaves Manal4Man linkages (Rosenfeld and Ballou, 1974). Briefly, the glycans were dissolved in 30 pl of acetic anhydriddpyridine (l:l, v/v) and incubated for 30 min at 100 "C, to peracetylate. After drying, the samples were resuspended in 50 pl of acetic acidlacetic anhydride/HZSO4 (lO:lO:l, v/v) and treated for 6 h a t 37 "C. After neutralization, the samples were partitioned between 250 pl of chloroform and 1 ml of water. The chloroform phases were dried and resuspended in 200 pl of methanol/32% ammonia (1:l. v/v) for 16 h at 37 "C to de-0-acetylate the glycans. The de-0-acetylated material was desalted and analyzed by Dionex HPAEC chromatography.

Neutral Glycan Standards-Glucose oligomer standards were prepared according to Yamashita et al. (1982) by partial acid hydrolysis of Dextran (Sigma). For oligosaccharide analysis on Bio-Gel P4 and Dionex HPAEC, these glucose oligomers were incorporated as internal standards. Neutral glycan analysis on Dionex HPAEC was correlated by internal dextran standards to the elution position of the neutral glycan portion of radiolabeled intermediates in GPI-anchor biosynthesis of I: brucei and the Thy-1 GPI-anchor, prepared by dephosphorylation, deamination, and reduction as described above.

RESULTS

Extraction and Analysis of Glycolipids-We metabolically labeled trophozoites ( 3 3 4 3 h after infection), an intraerythrocytic, asexual stage of I? falciparum, with different radioactive precursors, such as L3H1GlcN and [3HlMan. The presence of tunicamycin had no effect on the glycolipid pattern irrespective of the radioactive precursor used (Gerold et al., 1992b; Dieck- mann-Schuppert et al., 1992a). Parasites were purified and extracted as described under "Experimental Procedures." Un- infected erythrocytes showed no (<<1%) incorporation of radioactivity into glycolipids. TLC analysis of the "Fo1ch"-washed CM extracts and the butanol component of the CMW extracts (butanol phases) showed qualitatively identical glycolipid pat- terns in TLC systems A and B. Therefore, the Folch-washed CM extracts and the butanol components were pooled and analyzed together using solvent system B (Fig. 1). TLC analysis of these pooled glycolipids (termed glycolipid extract) revealed eight and seven glycolipids labeled with [3H]GlcN (Fig. lA) and L3H1Man (Fig. 1B), respectively. The relative migration (R,) values of these glycolipids were 0.25 (Pfgl a), 0.33 (Pfg, p), 0.50

(Pfgl 01, and 0.9 (dolichol(?)-phosphate-mannose)2 (Table I). We have used infected erythrocytes rather than isolated parasites for the following experiments. The glycolipid pattern is conserved throughout all experiments. However, we have observed differences in the relative amount of labeling of single malarial glycolipids in various labeling experiments (up to 30 times). The obtained glycolipid pattern is different from that described

(Pfg~ Y), 0.56 (Pfg~ 6), 0.62 (Pfgl e), 0.66 (Pfgl J), 0.71 ( P f g l 7),0.85

(?) indicates that chain length and saturation are unknown

in Plasmodium falciparum 2599

migration (cm) FTG. 1. TLC analysis of the PHIMan- and [SHIGlcN-labeled gly-

colipids extracted from I? falciparum. Asexual, intraerythrocytic stages of the malaria parasite P. falciparum were metabolically labeled for 2 h with l3H1Man and L3H1GlcN. Glycolipids were extracted and chromatographed on Silica Gel 60 plates using solvent system B. Ap- proximately 10,000 cpm were analyzed in each lane. Radioactivity was scanned using a TLC-scanner (Berthold LB 2842). Six L3H1Man and L3H1GlcN glycolipids were found and named Pfgl e, Pfgl p, Pfgl y, Pfg? 6, Pfgl E , and Pfgl I). One very hydrophobic glycolipid labeled only mth

fore designated Dol(?)-P-Man. Two with only 13H]GlcN-labeled glycolip- 13Hl-Man was identified as dolichol(?)-phosphate-mannose and there-

ids were named Pfgl ( and Pfgl 8. The glycolipids marked with * were not further characterized. 0, origin; F, front.

by Dieckmann-Schuppert et al. (1992b1, using labeling medium supplemented with 10% human serum.

Additional radiolabeling experiments were performed in or- der to define the composition of these glycolipids. In addition to labeling with mannose and glucosamine, the two most hydrophilic glycolipids, Pfg l a and Pfgl /3, could also be labeled with L3Hlmyristic acid, L3Hlpalmitic acid, [3Hlethanolamine, and my~-[~H]inositol, as expected for glycosylinositol lipids (Table I). Labeling of the parasites with L3H1ethanolamine or myo- L3H1inositol resulted in more than 98% incorporation of radioactivity into phosphatidylethanolamine or phosphatidylinositol, respectively (Gerold et al., 1992b; Dieckmann-Schuppert et al., 1992b). No incorporation of galactose or fucose into malarial glycolipids was observed.

Analysis of the Radioactivity Incorporated into the Glycolip- ids-Acid hydrolysis products of the neutral glycans of malarial glycolipids labeled with L3H1GlcN or L3HlMan were analyzed on Dionex-HPAEC using elution program 2. The analysis showed that the incorporated radioactivity was not metabolically converted (more than 98% of the radioactivity is detected as glucosamine or mannose, respectively), in contrast to the results described for Toxoplasma gondii GPIs (Schwarz and Tomavo, 1993).

Identification of most Glycolipids as GPZs-To investigate whether the [3HIGlcN- or L3H1Man-labeled glycolipids are GPIs, we tested the sensitivity of the glycolipid extracts and also of the single TLC-purified glycolipids toward PI-specific phospholipase C (PI-PLC) before and &r mild alkaline treatment, the sensitivity toward GPI- specific phospholipase D (GPI-PLD) and nitrous acid (HN02) treatment (Table I). &r


TABLE I Characterization of malarial glycolipids

Parasites were labeled with C3H1GlcN, L3H1Man, or [3Hlethanolamine, and glycolipids were sequentially extracted with CM and CMW. TLC-

The results summarized in this table correspond to the average of two ([3Hlethanolamine) and four (C3H1Man, I3H1GlcN) radiolabeling experiments. purified glycolipids were subjected to different treatments. After butanol-water phase separation, cleavage was measured by scintillation counting.

N L , not labeled; +, sensitive; -, insensitive.

Glycolipid species

Labeled via Sensitivity to Susceptibility to

[3HlGlcN 13H1Man L3H1EtN PI-PLC +p3Lc PLD PLAz HNOz NaOH Proposed structure R, values

pf; e (Dol(?)P-Man)

EtN-P-Man,-GlcN-Ino(x)-PA EtN-P-Man,-GlcN-Ino(x) - P A

Man,-GlcN-Ino(x) -PA Man,-GlcN-Ino(x) -PA Man,-GlcN-Ino (x) -PA

GlcN-Ino-PA Man,-GlcN-Ino (x) -PA

GlcN-Ino (x) -PA

0.25 L 0.33 L 0.50 L 0.56 L 0.62 L 0.66 L 0.71 L 0.85 L 0.90 NL

L L L L L NL L NL L

L L NL NL NL NL NL NL NL

+ f + + + f + + + f + + + f + + + f + + + + + + + * + + + * + + " - -

the treatments, reaction mixtures were partitioned between water and water-saturated 1-butanol. The organic phases contained uncleaved glycolipids or lipid products. They were analyzed by TLC (Fig. 2, left panels), The corresponding aqueous phases contained glycan fragments and were analyzed on Bio- Gel P4 (Fig. 2, right panels). Control reactions were performed in parallel. As shown in Fig. 2 (left panel ), only glycolipid 6 is cleaved by PI-PLC. All other glycolipids are not susceptible to PI-PLC treatment. To ascertain whether the insensitivity of the other malarial GPIs toward PI-PLC was due to the presence of palmitoylated inositol, as it is described for some GPIs of ?: brucei (e.g. P3) and several mammalian GPIs (for review, see Ferguson (1991) and Englund (1993)), the glycolipids Pfgl a, p, y, 6, E , q, and 8 were subjected to mild alkaline hydrolysis by ammonia prior to PI-PLC treatment. Only glycolipid 8 became 45% susceptible, whereas the glycolipids Pfgl a, p, y, 6, E , and q remained insensitive (Table I).

The glucosamine-inositol fragment of Pfgl a, generated by dephosphorylation and treatment with jack bean a-mannosidase, was completely (>go%) recovered in the organic phase after phase partition. Therefore, Pfgl a (and presumably all other PI-PLC-resistant malarial glycolipids) carry a hydrophobic modification at the inositol ring.

The treatment of the glycolipids with GPI-PLD and HNOz leads to recovery of >90% (only ~10% in the case of Pfgl 6 ) and 60% of the radioactivity, respectively, in the organic phases. The comparison of the GPI-PLD-generated hydrophobic fragments of Pfgl p and the putative GPI-anchor precursor P3 of 2'. brucei in the TLC system C showed significant differences in the chromatographic behavior (Rf = 0.07 and Rf = 0.75, respectively) although both GPIs share the same hydrophilic moieties (see below).

However, after HN02 treatment, we found one major and several minor products in the organic phases. The generation of alternative deamination products appears to be a common fea- ture of nitrous acid deamination (Krakow et al., 1986; Mayor et al., 1990). Based on the results that the efficiency of deamination in our experiments is similar to that of the deamination of the GPI-anchor precursors P2 and P3 of ?: brucei, we conclude that the malarial glycolipids are sensitive to deamination, indicating that they have a nonacetylated hexosamine, as is described for GPIs.

Treatment of the malarial glycolipids with phospholipase A2 (PLA2) leads to the formation of lyso-forms of each malarial glycolipid with the exception of dolichol(?)-phosphatee-mannose (Dol(?)-P-Man; Table I). However, only about 3040% of each glycolipid, except P f g l 6, was converted to its lyso-form, consistent with the results described for I: brucei GPI-anchor precursor P3, which was described to have a palmitoylated inositol

ring (Menon et al., 1988). Pfgl 6 is totally susceptible to PIA2 treatment, consistent with the assumption that this glycolipid has no modification at the inositol.

One [3Hlmannose-labeled lipid was shown to be insensitive to PI-PLC, GPI-PLD, and HNOZ (Table I). This glycolipid was identified as Dol(?)-P-Man by the following observations: 1) mild acid hydrolysis (2 N HCV1-propanol, 1:1, v/v, 50 "C, 15 min) quantitatively released mannose and 2) co-migration on TLC (solvent systems A and B) with Dol-P-Man from 2: brucei and l? falciparum Dol(?)-P-Man synthesized in a cell-free system prepared from asexual intraerythrocytic stages of l? falciparum (Gerold et al., 1992a).

After PI-PLC, GPI-PLD, or HNOz treatment of malarial glycolipids, the released oligosaccharide moieties were analyzed on Bio-Gel P4 (Fig. 2, right panel). In the cases of GPI-PLD and HN02, five peaks with different apparent sizes (depending on the cleavage method) were obtained. In contrast, the radioactivity released by PI-PLC elutes as one peak on the Bio-Gel P4 columns. This is consistent with the observation that only Pfgl 6 disappeared from the organic phases after PI-PLC treatment.

Determination of the Size and Presence of Terminal Manno- sidase-blocking Residues-The presence of the terminal ethanolamine in GPI-anchor precursors renders the structure resistant to jack bean a-mannosidase which requires an unsubstituted terminal mannose residue (Li and Li, 1972). Gly- colipids Pfgl a and p appear to contain ethanolamine since they can be radiolabeled with [3H]ethanolamine. To determine whether this ethanolamine is blocking the terminal mannose in glycolipids Pfgl a and p and whether the other malarial glycolipids carry blocking residues, the lipids were tested for their susceptibility toward jack bean a-mannosidase. TLC-purified [3H]Man-labeled Pfgl a, p, y, 6, E, and q and the r3H1GlcN- labeled Pfgl 6 and 8 were treated with HN02. The released oligosaccharides were sized on Bio-Gel P4 before and after treatment with jack bean a-mannosidase (Fig. 3, A and B) . HN02-generated fragments of Pfgl a and p (corresponding to 7.7 and 6.7 glucose units, respectively; Fig. 3 A , panels A and B 1 are only partly sensitive (Pfgl a; Fig. 3A, panel A1 or insensitive (Pfgl 0; Fig. 3A, panel B1 ), respectively. The HN02-generated and a-mannosidase-treated fragments of Pfgl a elute corresponding to 6.7 and 0.9 glucose units. This indicated that Pfgl

a possesses one additional a-mannose side chain which is not blocked like the other mannosyl residues of Pfgl a and are. The larger fragments of Pfgl a and p produced by jack bean a-mannosidase treatment co-elute on Bio-Gel P4 with the HN02-generated fragment of the structure ethanolamine-phosphate-mannose3-anhydromannose of the T brucei precursors P2 and P3 (6.7 glucose units). The peak intensities of the a-mannosidase-generated fragments of Pfgl a (Fig. 3A, panel

Glycosylphosphatidylinositols in Plasmodium fakiparum 2601

PI-PLC

a 0

3

migration (cm) Fraction number FIG. 2. Characterization of the [SH]GlcN-labeled glycolipids. [3HlGlcN-labeled glycolipids were treated with PI-PLC, GPI-PLD, and HNOp

The organic phases were analyzed on silica TLC plates using solvent system B (left panel ). The aqueous phases were analyzed on Bio-Gel P4 as described under “Experimental Procedures.”ARer incubation, the reaction mixture was partitioned between water and water-saturated butanol.

glycolipids marked with * were not further characterized. columns (right panel) with internal dextran standards (see “Experimental Procedures”). Positions of the standards are shown at the top. The

A1 ) show that the side chain mannose is preferentially labeled. The HN02-generated fragments of the [3Hlmannose-labeled

glycolipids Pfgl y , 6 , E , and q (which elute at 4.7,3.7,2.7, and 1.8 glucose units, respectively; Fig. 3B) are sensitive to a-mannosidase treatment (i.e. elution at 0.9 glucose unit afbr treatment), confirming that their terminal mannosyl residues are not blocked.

The glycolipids Pfgl ( and Pfgl 8, labeled only with L3H]glucosamine, are not affected by digestion with jack bean a-mannosidase (Fig. 3B, panels F and HI. Together with the failure to label these two glycolipids with [3Hlmannose and their elution size (1.2 glucose units), these results indicate that Pfgl 5 and 0 do not contain mannose residues.

Based on the labeling experiments with L3Hlethanolamine, the sensitivity toward dephosphorylation with aqueous HF (see below), the a-mannosidase digestions, and the conserved structure of other GPI-anchor precursors indicate that the terminal mannose residues of Pfgl a and p are blocked by ethanolamine- phosphate.

Analysis of the Glycans by Dionex-HPAEC-For further char-

acterization, the L3H]GlcN-labeled glycolipids Pfgl a, p, y, 6, E , (, q, and 0 were dephosphorylated by treatment with aqueous HF. This procedure selectively cleaves phosphodiester linkages. The HF-generated glycan fragments were deaminated and reduced (as described under “Experimental Procedures”), then analyzed by Dionex-HPAEC using elution program 1 (Fig. 4, A and B).

The neutral glycans derived from glycolipids Pfgl a and y (Fig. 4A, panels A and C ) co-eluted with a Man4-2,5-anhydromannitol (AHM) standard generated from the Thy-1 GPI-anchor (a generous gift from Dr. M. Ferguson). ARer treatment of the neutral glycans of Pf,, a and y with A. saitoi a-mannosidase (an exo-mannosidase specific for Manal-2Man linkages), the resulting fragments were analyzed on the Dionex-HPAEC system and found to co-elute with a Man2-AHM standard generated from I: brucei (Fig. 4A, panels A1 and C1 ). This indicates the removal of two al-2-linked mannoses from the glycans of Pfgl a and y.

Recent experiments show that the malarial surface proteins MSPl and MSP2 possess GPI-anchors with the same neutral


A P.I. gl (L

200

IO0

E

200

IO0

0

B P.f. gl t

"0 lP.B.!. ?. ? .

.. . . . , . . . .

P.I. gl p P.I. gl y

'p I?.?.!. t . ?.

.... .. .. . .

" 1 Fraction number

P.f. 01 c

P.I. gl 6

....... ., .

45 55 65 75 e5 95 105

P.I. gl 11

"0 I ? . ! . ? . ? . ?

R

...... .. * .

45 55 65 7 5 hlUdJ" U5 95 105 I P.f. gl 6

n

o_

3 . . . . . . . . . D D 3

5 Fraction number

FIG. 3. A, Bio-Gel P4 chromatography of HN02-generated hydrophilic moieties before and after jack bean a-mannosidase treatment. TLC- purified, [SH]Man-labeled glycolipids Pf,, a, Pf,, p, FYgl y, and Pf,, 8 were deaminated, and an aliquot (1000-1500 cpm for each control; A-D) of the HNOz-generated hydrophilic fragment was loaded on a Bio-Gel P4 column. A second aliquot (1500 cpm) was incubated with jack bean a-manno- Ridase and analyzed on the same column (panels A,-&). Glucose oligomers were added to the samples as internal standards, and their elution positions are indicated on the top of each profile. Radioactivity (y-axis) in the fractions (850 111) was plotted against the fraction number. B, Bio-Gel P4 chromatography of €€NO,-generated hydrophilic moieties before and after jack bean a-mannosidase treatment. TLC-purified, ISH1Man-labeled glycolipids Pf,, < and Pf,, and [SH]GlcN-labeled FY,, and Pfzl 8 were deaminated, and an aliquot (1000-1500 cpm for each control; E-H) of the HN0,-generated hydrophdic fragment was loaded on a Bio-Gel P4 column. A second aliquot (1500 cpm) was incubated with jack bean a-mannosidase and analyzed on the same column (panels E1-HI). Glucose oligomers were added to the samples as internal standards, and their elution positions are indicated on the top of each profile. Radioactivity (y-axis) in the fractions (850 pl) was measured and plotted against the fraction number.

glycan as Pfgl a.3 dard (Fig. 4A, panels BI and Dl ), confirming the removal of one The neutral glycans of the glycolipids Pfgl p and 6 both co- terminal al-2-linked mannose residue, as described for the

eluted with a Man,-AHM standard prepared from I: brucei neutral glycan of the T brucei GPI-anchor precursors P2 and (Fig. 4.4, panels B and D ) . When the malarial neutral glycans P3 (Mayor et al., 1990). were treated with A. saitoi a-mannosidase, the resulting frag- The HF-generated, deaminated, and reduced glycan fragments co-eluted on Dionex-HPAEC with a Man2-AHM stan- ments from Pfgl e and q co-eluted with Man2-" and Manl-

AHM standards, respectively (Fig. 4B, panels E and G ) . Con- 3 p. Gerald, L. Schofield, and R. T. Shwarz, manuscript in prepara- sistent with the linkages in the Co~esPonding GPI-anchor

tion. biosynthesis intermediates from I: brucei, the two malarial

Glycosylphosphatidylinositols in Plasmodium falciparum

A P.f.gl a P.I. gl p P.t. gl 7 P.I. gl 6

2603

150

s 100. s

50

0 . . . 0 10 20 30 40 50 EO 0 10 20 30 40 50 EO 0 10 20 30 40 50 EO 0 10 20 30 40 50 EO

Fraction number

E u a

B P.f. gl E P.f. gl 5 P.f. gI ll p.f. pl e

6 2

LOO ~ 2 1

. . . . . . . . . . . ,. . . . ,. . . . .. , , v

2oo - R 'Ell I 150. L

P) El tr

100 - 3 50 -E

0 . . 6 IO 20 30 40 50 60 0 10 20 30 4 0 50 60 0 10 20 30 40 50 60 0 10 20 30 40 50 60 , ,." ~ 1 . . .

Fraction number FIG. 4. A, Dionex-HPAEC analysis of the HF-generated, dephosphorylated, deaminated, and reduced core glycans. TLC-purified, L3H1GlcN-

labeled glycolipids Pf,, 01, Pf,, p, REI 7, and Hgl 6 were dephosphorylated, deaminated, and reduced as described under "Experimental Procedures." The neutral glycans were desalted and divided into aliquots. One aliquot (1200-1500 cpm) of each sample was analyzed by Dionex-HPAEC using

to the analysis on Dionex-HPAEC (Al-&). The standards indicated on the top of each profile are generated from Thy-1 GPI-anchor and GPI-anchor gradient elution program 1 (profiles A-D). A second aliquot (1500 cpm) was treated with A. saitoi ul-2-specific mannosidase and desalted prior

precursors of I: brucei. The flow rate was 1 d m i n , and fractions were collected every 0.4 min. B , Dionex-HPAEC analysis of the HF-generated dephosphorylated, deaminated, and reduced core glycans. TLC-purified, [3HlGlcN-labeled glycolipids Pfgl B , Pf,, 5, Pfgl q, and H,, 8 were dephosphorylated, deaminated, and reduced as described under "Experimental Procedures." The neutral glycans were desalted and divided into aliquots. One aliquot (1200-1500 cpm) of each sample was analyzed by Dionex-HPAEC using gradient elution program 1 (profiles E-H). A second aliquot (1500 cpm) was treated with A. saitoi ul-2-specific mannosidase and desalted prior to analysis on Dionex-HPAEC (El-Hl). The standards indicated on the top of each profile are generated from Thy-1 GPI-anchor and GPI-anchor precursors of I: brucei. The flow rate was 1 d m i n , and fractions were collected every 0.4 min.

glycans were found to be resistant to A. saitoi a-mannosidase labeled by [3H]mannose, Fig. L4) showed that both fragments (Fig. 4B, panels E l and G1). were identical and co-eluted with 2,5-anhydroma~itol (Fig.

Dionex-HPAEC analysis of the neutral glycans derived from 4B, panels F and H ) . Treatment with A. saitoi a-mannosidase the [3H]GlcN-labeled glycolipids PfBl &' and 0 (which were not had no effect on the elution behavior of the glycans of both


P.f.g1 (1 P.i.gl p

FIG. 5. Dionex-WAEC analysis of the dephosphorylated core glycans prior to and after jack bean a-mannosidase and acetolysis treatment. TLC- purified, [3HlGlcN-labeled glycolipids Pfgl

a, Pfgl p, and FYgl y were dephosphorylated, deaminated, and reduced. The desalted neutral glycans were divided into aliquots. The first aliquot (1200-1500 cpm) was analyzed by Dionex-HPAEC using elution program 1 (profiles A X ) . A second aliquot (1200-1500 cpm) was di- gested with a-mannosidase from jack bean. The desalted fragments were also analyzed on Dionex-HPAEC (A,C1). The third aliquot (3000 cpm) was subjected to acetolysis and analyzed on Dionex- HPAEC (A+?& using the same elution program. The standards indicated on the top of each profile are generated from Thy-1 GPI-anchor and GPI-anchor precursors of I: brucei. The flow rate was 1 mumin, and fractions were collected every 0.4 min.

250

200

I50

100

50

0

200

E I50

3 100

50

0

200

150

IO0

50

0

4 -g 3 .' :- I .. . . .

a P

.. .

glycolipids, indicating that they have no terminal al-2 mannose residue (Fig. 48, panels F l and HI ).

Jack Bean a-Mannosidase Digestion of Neutral Glycans -Digestion of HF-generated neutral glycans of the glycolipids Pfgl a, Pfgl p, and Pfgl y with jack bean a-mannosidase produced single glycans for each malarial neutral glycan (Fig. 5, panels A l , B l , and C1 ). These single glycans co-elute in each case with a 2,5-anhydromannitol standard. These data show that the neutral glycans of Pfgl a, /3, and y consist of a-mannose residues without additional non-mannose components.

Acetolysis of the Neutral Glycans of Malarial Glycolipids"T0 test for Manal-GMan linkages, we made use of the relatively selective acetolysis conditions described by Rosenfeld and Bal- lou (1974). The HF-generated deaminated and reduced neutral glycans of [3H]GlcN-labeled Pfgl a, Pfgl p, and Pfgl y were subjected to partial acetolysis and then analyzed on Dionex HPAEC (Fig. 5, panels A 2 , B2, and C2). Besides the unhydro- lyzed neutral glycans of these three malarial glycolipids, we found in all cases material co-eluting with Manl-AHM and AHM standards. The identification of a Manl-AHM fragment generated by acetolysis is consistent with the presence of a Manal-GMan bond between the first and second mannose residues of the malarial glycolipids Pfgl a, Pfgl p, and Pfgl y. The observation of under- and overdigestion products in acetolysis studies is confirmed by the results of Mayor et al. (1990).

DISCUSSION In this report we provide evidence that the asexual, intra-

erythrocytic late stages of the malaria parasite P. falciparum

Fraction number

P.1.91 -f

L .. . . .

IO 20 30 40 50 60

produce two putative glycosylphosphatidylinositol membrane anchor precursors ethanolamine-phosphate-6(al-2Man) Mancul-2Manal-GMana1-4GlcN-PI (Pfgl a) and ethanolamine-phosphate-6Manal-2Mana1~Mana1-4GlcN-PI (Pfgl p) and a series of minor glycosylphosphatidylinositols (GPIs) showing a lesser degree of glycosylation.

The glycolipid extracts contain Dol(?)-Man, six f3H]Man- and eight [3HlGlcN-labeled glycolipids which are shown to be GPIs by different chemical and enzymatic cleavages. All glycolipids, except Dol(?)-P-Man, are sensitive to GPI-PLD, PLA2, HNO2, and NaOH. Only the glucosamine-labeled Pfgl 5 is sensitive to PI-PLC. All other glycolipids are insensitive to PI-PLC treatment suggesting that these glycolipids might have a modified inositol structure, which is described to hinder PI-PLC cleavage (Roberts et al., 1988). Additional evidence for a modified inositol structure is provided by the only partial sensitivity of the malarial GPIs toward PLA2 which is similar to that of the I: brucei GPI-anchor precursor P3 (Menon et al., 1988,1990b). Different protein-bound GPIs (Roberts et al., 1988; Walter et al., 1990; Clayton and Mowatt, 1989) and GPI-lipids (Krakow et al., 1989; Mayor et al., 1990; Field et al., 1991) are described which have inositol structures modified by palmitic acid. In these cases, pretreatment with ammonia prior to PI-PLC treatment leads to partial susceptibility toward PI-PLC (Krakow et al., 1989; Mayor et al., 1990). Malarial glycolipids, however, could not be rendered PI-PLC-sensitive by pretreating with ammonia. Partition of the glucosamine-inositol-X fragment of Pfgl a into the organic phase and the different chromatographic behavior of the GPI-PLD-generated fragment of Pfgl p and the

Glycosylphosphatidylinositols in Plasmodium falciparum 2605

respective fragment of P3 indicate that all malarial GPIs, except Pfgl (, carry a hydrophobic modification on the inositol ring different from palmitic acid which had been described for try- panosomal P3 (Menon et al., 1988).

We describe two malarial glycolipids with a single glucosamine as their hydrophilic head group (Pfgl ( and e) . Both glycolipids differ in their chromatographic behavior and their sensitivity toward PI-PLC. Therefore, we postulate an additional hydrophobic modification at the inositol of Pfgl 8. This would be consistent with a biosynthetic pathway beginning with glucosamine-PI (Pf,, () followed by the addition of a hydrophobic modification (Pf,, 0) prior to the addition of the first mannose residue. A similar biosynthetic pathway has been described for mammalian GPIs (Urakaze et al., 1992; Hirose et al., 1992). At this stage, it is difficult to assess the exact location of this modification on the inositol ring, but it is assumed to be an ester-bound hydrophobic moiety because of its sensitivity to treatment with NaOH.

One glycolipid releases mannose upon mild acid hydrolysis specific for Dol-P-Man (McDowell and Schwarz, 1988). This glycolipid is presumably Dol(?)-P-Man as judged by its chromatographic properties. I t co-chromatographs with a manno- lipid synthesized in the cell-free system (Gerold et al . , 1992a) showing properties of Dol-P-Man (alkali stability, chromatographic behavior on DEAE-cellulose), the synthesis of which is inhibited in vitro by amphomycin. The donor of the mannose residues to the core glycan of l! brucei GPIs has been shown to be Dol-P-Man (Schwarz et al., 1989; Menon et al., 1990b). Therefore, by analogy, an isoprenoid-P-Man can be postulated to play a similar role for at least some of the mannose residues of the malarial GPIs. The lack of Dol-PP-(GlcNAc), Dol-PP- (GlcNAc)2, and higher glycosylated intermediates is consistent with the findings that the asexual, intraerythrocytic stages of €? falciparum do not produce detectable amounts of dolichol cycle intermediates as observed in this and previous (Dieck- mann-Schuppert et al., 1992a; Gerold et al., 1992a, 1992b) studies.

We show that €? falciparum produces mannose- and glucosamine-labeled GPIs, whose glycan backbones are mannosylated to a variable degree. The dephosphorylated, deaminated, and reduced neutral glycans analyzed by HPAEC (Dionex) and the deaminated hydrophilic fragments analyzed on Bio-Gel P4 of the different malarial glycolipids co-elute with the respective glycans or hydrophilic fragments of the GPI-anchor of Thy-1 (hexosaminidase-treated) and the GPI-anchor precursor P2 of I: brucei and its biosynthetic intermediates. Labeling with f3HIethanolamine and the analysis of the neutral glycans after HF dephosphorylation strongly suggest Pfgl a and Pfgl p to be ethanolamine-phosphate-carrying, putative GPI-anchor precursors. The results of the linkage type analysis of the terminal mannoses in the malarial glycolipids Pfgl a and Pfgl y using al-2-specific mannosidase from A. saitoi and acetolysis are consistent with P f g l a and y carrying an additional al-2-linked mannose side chain at the terminal mannose. The linkage of the additional mannose residue to the terminal mannose is suggested by the identification of a Manl-AHM fragment generated by acetolysis. These results also indicate that there is a Mana-1,6-Man linkage between the second and first mannose residue.

The neutral glycans of Pfgl p and Pfgl 6 generated by dephosphorylation, deamination, and reduction show identical frag- mentation products upon incubation with A. saitoi al-2 mannosidase, jack bean a-mannosidase and after acetolysis are identical to the respective fragment of the IC: brucei GPI-anchor precursor P2. Therefore, they are supposed to have the structures ethanolamine-phosphate-6Manal-2Manal-GMana1-

4GlcN-PI (Pf,, p ) and Manal-2Mana1-6Manal-tGlcN-PI

A pathway of reactions in the synthesis of GPI-anchors involving sequential addition of monosaccharides, derived from activated precursors, has been described for T brucei (Master- son et al., 1989; Doering et al., 1990; Mayor and Menon, 1990; Mayor et al., 1990; Field et al., 1992; Menon et al., 1993), Toxo- plasma gondii (Tomavo et al., 19921, and some mammalian cells (Puoti et al., 1991; Stevens and Raetz, 1991). The Bio-Gel P4 and the Dionex-HPAEC analysis of the different hydrophilic fragments of the malarial glycolipids Pfgl c, Pfgl q, Pf,, (, and Pfgl e, in comparison to neutral glycans generated from intermediates in the GPI-anchor biosynthesis of T brucei, their insensitivity to A. saitoi a1-2-mannosidase, and the remarkable conservation of the core glycan on all investigated GPIs leads us to propose the structures Manal-GManal-tGlcN (Pf,, e), GlcN (Pf,, (), Manal4GlcN (Pf,, q), and GlcN (Pfgl 0 ) for these malarial glycolipids. The identification of potential biosynthetic intermediates of GPI-anchor synthesis and two putative GPI-anchor precursors in I! falciparum labeled in culture which, with one exception, are substituted at the inositol ring is reminiscent of the pathways shown for mammalian cells.

In this paper we present preliminary data which suggest that the inositol ring of malarial GPIs (except Pfgl (1 is modified by a hydrophobic residue different from palmitic acid, the only hydrophobic modification on the inositol ring of GPIs described so far (for review, see Englund (1993)). Especially the hydrophobic elements of malarial GPIs may be responsible for some of the pathological properties of I? falciparum in severe malaria. For instance, the non-protein-bound malarial GPIs Pfgl a and Pfgl p are potential inducers of cytokine r e l e a ~ e . ~ Together with findings that purified GPIs derived from pronase hydrolysis of MSPl and MSP2 induce cytokine release, transient pyrexia, hypoglycemia, and lethal cachexia in mice (Schofield and Hackett, 1993) and with other evidence (Kwiatkowski et al., 1990), these results point to an important role of this class of molecules in the pathology of malaria.

Elucidation of the structures and biosynthetic pathways of malarial GPIs may provide a basis for the development of a glycolipid-based “antidisease” vaccine (Playfair et al., 1990; Schofield and Hackett, 1993).

Additional studies in progress using a cell-free system will provide more detailed information of the biosynthetic pathway of malarial GPIs and help to determine whether Pfgl p is an additional possible GPI-anchor precursor. Also, the design and testing of inhibitors of GPI biosynthesis should be facilitated by studying some of the reactions involved in the cell-free system.

Acknowledgments-We thank S. Koslowski, M. Eppinger, and S. Kauer for technical assistance and B. Striepen, C. Zinecker, and Dr. M. A. J. Ferguson for helpful discussions. We gratefully acknowledge Drs. A. K. Menon and S. Tomavo for critical reading of the manuscript. We would like to thank Dr. V. Kretschmer, Blood Donation Center, Univer- sity of Marburg, for providing human erythrocytes and Dr. M. A. J. Ferguson for a generous gift of Thy-1 neutral glycans. P. G. thanks the Hessische Graduiertenfdrderung for a fellowship and V. G, A. G. and A. B. for continuous support.

(Pf,, 6).

REFERENCES

Braun-Breton, C., Rosenberry, T. L., and Pereira da Silva, L. (1988) Nature 332,

Clayton, C. E., and Mowatt, M. R. (1989) J. Bwl. Chem. 264, 15088-15093 Cross, G. A. M. (1990) Annu. Reu. Cell Bwl. 6, 1-39 Davitz, M. A., Hereld, D., Shak, S., Krakow, J., Englund, P. T., and Nussenzweig,

Dieckmann-Schuppert, A., Bender, S., Odenthal-Schnittler, M., Bause, E., and

457459

V. (1988) Science 238, 81-84

Schwarz, R. T. (1992a) Eu,: J. Biochern. 206,815-825

S. Tachado, P. Gerold, R. T. Schwarz, and L. Schofield, manuscript in preparation.


(1992b) Paras&l. Res. 78,416422 Doering, T. L., Masterson, W. J., Hart, G. W., and Englund, P. T. (1990) J. Biol.

Englund, P. T. (1993)Annu. Reu. Biochem. 62, 121-138 Ferguson, M. A. J. (1991) Curr. Opinion Strut . Bwl. 1, 5 2 x 2 9 Field, M. C., Menon, A. K, and Cmss, G. A. M. (1991) J. Bwl. Chem. 266,8392-

Field, M. C., Menon, A. K, and Cross, G. A. M. (1992) J. Biol. Chem. 267,5324-

Chern. 286,611-614

8400

5329 GeGld, P., Dieckmann-Schuppert, A., and Schwarz, R. T. (1992a) Biochem. Soc.

Trans. 20.2975 Gedd, P., heekmann-Schuppert, A, and Schwan, R. T. (1992b) Biol. Chem.

Hoppe-Seyler 373,844 Goman, M., Langsley, G., Hyde, J. E., Yankovsky, N. K, Zolg, J. W., and Scaife, J.

G. (1982) Mol. Biochem. Pamsitol. 6,391400 Haldar, K, Ferguson, M. A. J., and Cross, G. A. M. (1985) J. Bwl. Chem. 260,

4969-4974 Haldar, 11, Henderson, C. L., and Cross, G. A. M. (1986) h.oc. Natl. Acad. Sci. U. S. A 89,8565-8569

Hirose, S., Prince, G. M., Sevlever, D., Ravi, L., Rosenberg, T. L., Ueda, E., and Medof, M. E. (1992) J. Biol. Chem. 267, 16968-16974

Holder, A. A, Loekyer, M. J., Odink, K G., Sandhu, J. S., Riveros-Moreno, C., Davey, L. S., Tizard, M. L. V., Schwarz, R. T., and Freeman, R. R. (1985) Nature

Kwiatkowski, D., Hill, A. V., Sambou, I., Twwmasi, B., Castraeane, J., Manogue, K 317,27&273

R., Cerami, A., Brewster, D. R., and Greenwood, B. M. (1990) Lancet SSS, 1201-1204

Krakow, J. L.. Hereld, D., Banm, J. D., Hart, G. W., and Englund, P. T. (1986) J. Bwl. Chem. 261,12147-12153

Krakow, J. L., Doering, T. L., Masterson, W. J., Hart, G. W., and Englund, P. T.

h b m s , C., and Vanderberg, J. P. (1979) J. Parasitol. 86,418420 Li, Y. T., and Li, S. C. (1972) Methods Enzymol. 28,702-713 Loekyer, M. J., and Holder, A. A. (1989) in Vaccination Strategies of %pica1

Masterson, W. J., Doering, T. L., Hart, G. W., and Englund, P. T. (1989) Cell 66,

Mayor, S., and Menon, A. K (1990) Methwls: A Companion to Methods Enzyml . 1,

(1989) Mol. Biochern. Parasitol. 36,263-270

Diswses (Liu, F. Y., ed) pp. 124-148, CRC Press, Inc., Boca Raton, FL

793-800

297-305

Mayor, S., Menon, A. K, and Cross, G. A. M. (1990) J. Biol. Chern. 286,6174-6181 McDowell, W., and S c h w a , R. T. (1988) in Posttranslational Modification of

Menon, A. K, Mayor, S., Ferguson, M. A. J., Duszenko, M., and Cmss, G. A. M. Proteins by Lipids (Brodbeck, U., and Bordier, C., eds) Springer-Verlag, Berlin

Menon, A. K, S c h w a , R. T., Mayor, S., and Cross, G. A. M. (1990a) J. Biol. Chem. (1988) J. Biol. Chem. 263,1970-1977

Menon, A. K, Mayor, S., and S c h w a , R. T. (199Ob) EMBO J. 9,42494258 286, 9033-9042

Menon, A. K, Eppinger, M., Mayor, S., and S c h w a , R. T. (1993) EMBO J. 12.

Playfair, J. H. L., Taverne, J., Bate, C. A. W., and de Souza, J. B. (1990) Zmrnunol.

Puoti, A., Desponds, C., Fankhauser, C., and Conzelmann, A. (1991) J. Biol. Chem.

Roberts, W. L., Myher, J. J., Kukais, A., Law, M. G., and Rosenberry, T. L. (1988) J.

Rosenfeld, L., and Ballou, C. E. (1974) Carbohydr. Res. 32,287-298 Schofield, L., and Hackett, F. (1993) J. Exp. Med. 177, 145-153 S c h w a , R. T., and lbmavo, S. (1993) Res. Zmmunol. 144,24-32 S c h w a , R. T., Riveros-Moreno, V., Loclryer, M. J., Nicholls, S. C., Davey, L. S.,

B i d . 6,964-968 Hillman, Y., Sandhu, J. S., Freeman, R. R., and Holder, A. A (1986) Mol. Cell

S c h w a , R. T., Lockyer, M. J., and Holder, A. A. (1987) Host-Parasitic Cellular and Molecular Zntemtions in Protozoal Infections, pp. 275-281, Springer Verlag,

S c h w a , R. T., Mayor, S., and Menon, A. K (1989) Biochern. Soc. hns. 17,746s New York

Sharma, C. B., Babczinski, P., Lehle, L., and h e r , W. (1974) Eur. J. Biochem. 46,

Smythe, J. A,, Coppel, R. L., Ramasamy, R., Kemp, D. J., andhders , R. F. (1988)

lbmavo, S., Dubremetz, J. F., and S c h w a , R. T. (1992) J. Biol. Chem. 267, Stevens, V. L., and Raetz, C. R. H. (1991) J. Biol. Chem. 266, 10039-10042

Trager, W.. and Jensen, J. B. (1976) Science 193.673-675 Urakaze, M., Kamitani, T., DeGaspari, R., Sugiyama, E., Chang, H. M., Warren, C.

Walter, E. I., Roberts, W. L., Rosenberry, T. L., Ratnoff, W. D., and Medof, M. E.

Yamashita, K, Mizuochi, T., and Kobata, A. (1982) Methods Enzyml . 83,105-126

1907-1914

!lb&y 11,2627

286,21051-21059

Biol. Chem. 263, 18766-18775

35-41

Proe. Natl. Acad. Sci. U. S. A 85,5195-5199

21446-21458

D., and Yeh, E. T. H. (1992) J. Biol. Chem. 267,6459-6462

(1990) J. Zmmunol. 244,1030-1036

the journal of biollxical c~istry no. 4, 1994 the … · glycosylphosphatidylinositols synthesized...

Documents