Characterization and storage of malaria antigens: Proteins and glycoproteins from human erythrocytes infected with Plasmodium falciparum*

Bultien of the World Health Organizarion, 57 (Suppl 1): 101-107 (1979) Proteins and glycoproteins from human erythrocytes infected with Plasmodium falciparum * A. KJLEJIAN' & J. OLSON 2 Thepurpose of this study was to determine whether metabolites ofparasite origin are incorporated into the "'knob-ike "'protrusions ofmembranes oferythrocytes infected with Plasmodium falciparum. Intraerythrocytic parasites were grown in culture and labelled metabolically with (35S)-methionine. Fractionsenrichedin membranes with knobsandfrac- tionspoor in knobs wereprepared and analysed by sodium dodecylsulfate polyacrylamide gel electrophoresis. Labelled components were visualized by autoradiography. Fractionsen- riched in membranes with knobssho weda consistent relative increase in a labelled band with an apparent relative molecular mass around 70-80 thousand. Membrane-enrichedfractions were also tested for glycoprotein components using lectin-binding. No differences were observed between wheat germ agglutinin-binding components in membranesfrom normal and parasitized erythrocytes. On the other hand, membranesfrom parasitized erythrocytes showed distinct concanavalin A-binding components around 70-80 thousand relative molecular mass range. Further experiments are required to conclude whether the metabolic- ally tabelledprotein component observed within the same relative molecular mass range is a membrane-associated glycoprotein identical to the concanavalin A-binding component. In a previous study we demonstrated that "knob- like" protrusions that develop on membranes of erythrocytes infected with Plasmnodirrn fakciparumn are antigenically different from adjacent areas devoid of them (1). The present investigation was initiated as a first step towards understanding the origin and the chemical nature of the knobs. To determine whether the knobs are a secretory product (or products) of parasite origin, parasitized erythrocytes were labelled metabolically with 3 S-methionine and the radio- active proteins of fractions enriched in membrane fragments with knobs were compared with other frac- tions devoid of them. Since secretory cellular pro- ducts that become associated with surface membranes are generally glycoprotein in nature, we also com- pared the binding of two lectins, wheat germ agglu- tinin (WGA) and concanavalin A (con A), to the iso- lated components. " This work was supported by NIH grant AJ-13728 and a grant from the UNDP/World Bank/WHO Special Programme for Re- search and Training in Tropical Diseases. Associate Professor, Laboratory of Parasitology, The Rocke- feller University, New York, NY 10021, USA. 2 Graduate Student, The Rockefeller UnLversity, New York, NY 10021, USA. MATERIALS AND METHODS Parasites P.falciparun (FMG strain) was cultured in Petri dishes in a candle jar as described by Trager & Jensen (2). Cultures were initiated from infected blood, at 8-13 No parasitaemia, which was generously provided by Dr Trager from cultures in flow-vials (2). The blood was centrifuged at low speed (10 min, 230 g) and the pellet mixed with an equal volume of fresh erythrocytes. The mixture was diluted with medium to give a 6% cell suspension and plated in 100-mm Petri dishes (10 ml of suspension per dish). After 30-40 h, the infected blood was collected and pro- cessed in an identical manner to the starting material. Through repeated selection of parasites at younger developmental stages by low-speed centrifugation of infected blood, considerable synchrony of parasite development was obtained. After the second dilution with normal erythrocyces, parasites were allowed to segment and reinvade. The infected blood was again collected and released parasite debris removed by centrifugation (10 min, 500 g). The clean pellet was replated and parasites collected when a majority had de-veloped to mature schizonts. In general, the final parasitaeniia was 20-25%o. To obtain radioactive parasite proteins, (35S)-methionine (4 uCi/ml) was added to the culture 3J76 -101 - A. KILEJ(AN & J. OLSON media 6-24 h prior to collection of parasite samples. Preparation offractions from parasitized erythrocytes Samples of young parasites were prepared either by osmotic lysis of infected blood with 10 mmol/litre phosphate buffer (pH 8) or by lysis with 0.15% (1.5 g/litre) saponin in Trager's buffer (3). Contami- nation by large stages of parasites was removed by centrifugation (10 min, 200 g). To prepare samples of large trophozoites and schizonts, parasitized blood was centrifuged (10 min, 230 g) and the pellet was saved for further processing. The supernatant was centrifuged (10 min, 900 g) and the pellet suspended in 10 volumes of Trager's buffer. Erythrocytes were disrupted eitber by rapid forcing of the suspension 3 times through a 26-gauge hypo- dermic needle or by sonicating for 2 minutes at 10 pulses per minute using a "Sonicator" cell disrupter (Model W-220F), fitted with a microtip and set at out- put control 2. The homogenates were layered over a discontinuous sucrose gradient as described below. The pellet from the initial low-speed centrifugation, which showed a considerable number of parasitized erythrocytes in addition to uninfected cells, was lysed with saponin and the mature stages of parasites col- lected by centrifugation (10 nin, 500 g). The pellets were suspended in Trager's buffer and the erythrocyte membranes disrupted by shearing through a 26-gauge needle. The suspensions were layered on a discontinu- ous sucrose gradient consisting of 50/o, 40%7, 300o and 20%1f sucrose made in 10 mmolllitre phosphate buffer, pH 8. Gradients were centrifuged (40 min, 20 000 g, Beckman SW 27 rotor). Membrane frag- ments with knobs were collected from the boundary of 30%7 and 40% sucrose. Merozoites and small para- sites trapped within membranes banded at the inter- phase of 40% and 5O0. sucrose, while intact multi- nucleate parasites pelleted. Fractions collected from each interphase were diluted with 10 mmol/litre phosphate buffer, pH 8, and pelleted by centrifu- gation (60 min, 1l0 000 g). All procedures were carried out at ice-bath tem- perature. Analytical procedures Sodium dodecyl sulfate polyacrylamide gel electro- phoresis (SDS-PAGE) was carried out as described by Laemmli (4). Lectin bindiing to components in poly- acrylamide gels was carried out according to the pro- cedure of Burridge (5). Samples for electron micro- scopy were processed as previously described (6). To obtain a representative picEure of the subcellular con- stituents in a sample, minute amounts were fixed and dehydrated in cellulose nitrate centrifuge tubes. The plastic was solubilized with propylene oxide and the total sample was embedded in flat boats. Perpen- dicular sections were cut through several depths ofthe embedded sample and examined. ..tf .. .4. - ,~~W *,1i-- jllwRP4. Fig. 1. Giemsa-stained smears of P. falciparum from culture. la) Infection showing mainly young parasites. (b) Infection showing mainly mature schizonts and some free merozoites, probably released by mechanical pressure while making the smear. 102 A. KILEJ1AN & J OLSON a b c di p e fr~~~~~ o14 A __4 _' ; 3 - t t4i;t 4 B~~~~~ae- - = n4e_ o - Fig. 2. Stained an-d autoradiographed proteins of young andi mature stages of P falciparum analysed by SDS-PAGE. (a) Stained proteins of erythrocytes parasitized mainly with uninucleate P. faciparum and analysed after osmotic lysis. (b) Autoradiogram of gel band (a) (c) Autoradiogram of proteins from erythrocytes parasitized with multinucleate P. falci- parumand analysedafterosmotic lysis. Id) Stained proteinsof gel band Ic?, 'le) Autoradiogram oferythrocytesinfectedwith multinucleate parasites, analysed after lysis with saponin. if) Autoradiogaram of erythrocytes infected with uninucleate parasites, analysed after lysis with saponin. PROTEINS FROM ERYTHROCYTES INFECTED WITH P. FALCIPARUM a b C d e 4'-5000 Fig 3- Autor-adiograrm of drfferent fractionis fr-om P. flc,parum prepared by centrifugation of homogenates through a discontinuous sucrose gradient and analysed by SDS-PAGE. (a) Erytthrocytes infected with uninucleate parasites prepared by osmotic lysis. (b) Intraerythrocytic multinucleate parasites obtained from a pellet of discontinuous sucrose gradient which was layered with a homogenate prepared by shearing infected blood through a hypodermic needle. (c-Fraction enriched in host membranes with knobs, collected at interphase between 30 and 40% sucrose from the same gradient as that used in (b) (d) Fraction enriched in erythrocyte membranes with knobs and merozoite ghosts, collected at interphase between 40 and 50% sucrose from the gradient used for fractionation of a sonicated homogenate of infected blood.(eJ Fraction enriched in erythrocyte membranes with knobs obtained from interphase between 30 and 40% sucrosefrom the gradient used for fractionation of homogenate prepared by shearing saponin-lysed infected erythrocytes through a hypodermic needle PROTEINS FROM ERYTHROCYTES INFECTED WITH P FALCIPARUM RESULTS AND DISCUSSION The lack of synchronous development of P.fatci- parum in culture along with the fragility and stickiness of erythrocytes parasitized with schizonts posed sev- eral technical problems in our studies. As shown in Fig. 1, the selective removal of large parasites from initial cultures by differential centrifugation intro- duced some degree of synchrony. Once parasitaemia increased above 12-15%, centrifugation became in- effective as a means of separating parasite stages; erythrocytes infected with scbizonts tend to agglutin- ate and pellet with uninfected cells. Accumulation of residual bodies and other debris released during emergence of merozoites from in- fected erythrocytes appears to have a distinct lytic ef- fect on erythrocytes in culture. This was evident from the presence of large numbers of normal erythrocyte ghosts and masses of fragments of denatured haemo- globulin. Repeated removal of debris from our cut- tui es improved parasite growth and eliminated most of thebreakdowniproducts from normal erythrocytes. Several species of malaria parasites induce morpho- logical alterations of the host erythrocyte membrane (7). A basic question that arises is whether these changes are the result of chemical alterations of eiyth- rocyte membrane components due to interaction wsith metabolites produced by the parasites or represent in- sertion of a biochemical entity of parasite origin into the membrane. Even though immunocytochemical studies suggest that these host surface membrane structures contain malarial antigens (1, 8, 9) it could be argued that the hosts from which antisera were obtained could also have produced antibodies against chemically altered erythrocyte membrane constitu- ents to which they were exposed. The most direct approach would be the demonstration of a metabolic- ally-labelled component in punfied host cell mem- branes; since erythrocytes have lost their synthetic ability, all labelled constituents have to originate from the parasite. We were not able to obtain highly puri- fied fractions from erythrocytes infected with P.falci- parum; therefore, we resorted to a comparative approach. Labelled proteins from osmotically-lysed erythro- cytes infected either mainly with uninucleate or with multinucleate parasites were analysed by SDS-PAGE. The gels were stained, dried, and autoradiographed. In general, the younger parasites showed many more minor labelled components than the multinucleate ones (Fig. 2; compare (b) and (c), (e) and (I)). This could ieflect less efficient removal of soluble proteins from the former. For clarity, the labelled bands that will be discussed have been numbered from 1 to 4. Comparison of relative intensities of labelled com- ponents showed only one conspicuously augmented band in mature stages; this difference was consistent, regardless of the methods used for preparation of the samples (Fig. 2, band 3). Fractions enriched in erythrocyte membranes with knobs showed a pattern of labelling almost identical to that observed in lysed erythrocytes infected with multinucleate parasites (Fig. 3, (b)-(e)). The intensity of band 3 did not seem to be correlated with a reduc- tion in labelled components of higher molecular weight. The most variable component in analysed fractions was band 2; sample d (Fig. 3), a mixture of merozoites and membranes, showed a relative in- crease in this band. which seemed to be correlated with a parallel reduction in band 1. On the other hand, sample e (Fig. 3), which showed a sharp band I, did not show a band 2. Band 2 could reflect either dif- ferent forms of band I or origilate as a degradation product from it. Since a glycoprotein is the most likely candidate for a secretory cell product associated uwth surface msnem- branes, the presence of carbohydrate in membrane- enriched fractions from infected erythrocytes was compared with normal red cell ghosts. The samples were analysed by SDS-PAGE and incubated witb tritium-labelled con A and WGA. These two lectins were selected because their affianities cover mnost of the common carbohydrate constituents in glycoproteins; con A binds to glucose, mannose, fructose and N-acetylglucosamine, while WGA binds to galactose and N-acetylgalactosamine. Con A-binding receptors ofhuman erythrocyte membranes have been shown to be primarily associated with the electrophoretic lead- ing end ofwhat has been called Band I11 by Fairbanks et al. (J0); in addition, some unspecific binding has been observed to bands I, II, V, and VI (11). Con A- blnding components of membrane-enriched fractions from parasitized erychrocytes showed some distinct differences from those of normal erythrocytes; a very high relative mo]ecular mass component (larger than 250 000) and three bands witbin 80 000-60 000 rela- tive molecular mass range were apparent (Fig. 4a). The binding seemed to be specific since it was nhibited by the addition of2%o r-methyl-D-glucoside to the reaction mixture with con A (Fig. 4b). No major differences were apparent in WGA-binding components of the two membrane samples tested (Fig. 4c). It is tempting to speculate on the identity between the metabolically-labelled componenlt 3 and one of the major con A-binding components observed in membrane fractions of infected erythrocytes. Such an association would certainly facilitate isolation and purification of this glycoprotein. However, until further conclusive experimental evidence is obtained, this observation can be regard only as an interesting coincidence. 10S A. KJLLEJIAN & J. OLSON 2- 250000 89000 68000 b w(_, __m _ s_w ,, 3 10 20 30 40 50 60 SLICE NUMBER Fig 4. Distribution of radioactive lectins bound to components of normal erythrocyte membranes and membrane-enriched fraction from P fa/ciparum-infected cells. (a) Concanavalin A (b) Concanavalin A in the presence of2% a-methyl-D-gluco- side. (c) Wheat germ agglutinin. Broken curves- normal erythrocyte membranes; solid curves- membrane-enriched frac- tion from infected erythrocytes. CM4 I0 x C) 106 PROTEINS FROM ERYT13ROCYTES INFECIED WITH P. F-ILCIPARUM 107 ACKNOWLEDGEMENTS We thank Dr W. Trager for his generous supply of infected blood for initiating our cultures. RtSUME LES PROTELNES ET GLYCOPROTIINES DES tRYTHROCYTES HUIMAINS INFECTES PAR PLASMODIUM FALCIPARUM L'objet de cette Etude &laLt de determiner si des m6ta- bolites d'origine parasitaire se trouvent incorpores dans les protuberances arrondies apparaissant sui la mcmbrane des 6rythrocytes infectes par Plasmodium faciparumn. Lcs forTmes inura-cellulaires du parasite, obtenues par culture, ont Wt6 marquees metaboliquement i la (3' S)-mekhionine Des fractions enrichics de membrane "a protub6ranccs" et des fractions pauvres en protuberances ont ete preparees et analys6cs par 6lectrophorese en gel de dodecvlsulfaLe de sodium-polyacryJamide (SDS/ PAGE). Les constituants niarques ont &6 obserne&s par aulo-adiograph.e. Pour les fractions cnrichies en membrane a proEubcrances. on a consEate une augmentation reJative coMsuanic d'intensite d'une bande marqu6c de poids mol&culaire apparent se situant aulour de 70-80 mille. D'autre part, on a utilisb des lectines pour mettre en evidence les constiwuants glyco- prot6iques dans ces mtmes fractions. Avec L'agglutinine de germe de bl6, on n'a pas constalc de diff6rence entre les conslituants des mcmbranes d'&rthirocytes normaux et de celles pi ovenant d'6rythrocytes infectes En ievanche, des conslituants distmncts se lianm A la concanavaline A sur une chtelle dc poids moleculaire de 70 a 80 mille ont cte observes sur les niembrancs d'&rvthroc3 tes infect6s. D'autres cxp&- riences seiont n6cessaircs pour dcterminer Sl le constituanlt proteique marquc metaboliquement obser%e dans la meme zone de poids rnoWculairc eSI urne glycoprot6ine associ6e A [a membrane identique au constitlLanL se hiant A la concanava- line A REFERENCES 1. KILEJIAN. A. El AL. Experimi7eiinal pcrasitolog.} 42: 157-164 (1977). 2. TRACER, W. & JENSF.N, J. B. Science, 193: 674-675 (1976) 3. TRAGLR. W. Experitental parasitology, 8 265-273 (1959). 4. LAEMMLI, U. K. Naruire (Lonidoni), 227. 680-685 (1970). 5. BURRIDUL, K. Proceedings of the NatonalAcademe7y of Sciences ofthe UnnredfStates ofAmzerica, 73: 4457-4461 (1976). 6. KIIEJIAN. A. Journal of biological chemnistry; 249: 4650-4655 (1974) 7. AIkAWA. WI. Btilletin of the World Health Organiz- ation, 55 139-156 (1977). 8. ToBij, J. E. & COATNEY, G. R. Expermenfal para- strology, 11: 128-132(1961). 9. AIKAWA. M. LI AL Amiie,ican journal of pathology, 79: 285-294 (1975). 10. FAlRBANKS, G LI AL Bioclietnistry, 10: 2606-2617 (1971). 11. FINDLAY, J. Journal ofbiological cheriislry, 249: 4398- 4403 (1974).