Proc. Nati. Acad. Sci. USAVol. 87, pp. 801-805, January 1990Cell Biology

Trypanosome variant surface glycoprotein transfer to targetmembranes: A model for the pathogenesis of trypanosomiasis

(anemia/glycolipid membrane anchor)

MARY R. RIFKIN* AND FRANK R. LANDSBERGERThe Rockefeller University, 1230 York Avenue, New York, NY 10021

Communicated by C. de Duve, October 31, 1989 (receivedfor review August 11, 1989)

ABSTRACT The variant surface glycoprotein (VSG) oftrypanosomes is attached to the cell surface by means of aphosphatidylinositol-containing glycolipid membrane anchor.The studies presented in this paper support the hypothesis thatthe transfer of VSG from trypanosomes to erythrocytes couldlead to one of the pathological features associated with trypan-osome infection-i.e., anemia. Migration of trypanosome VSGfrom live trypanosomes to target cells (sheep erythrocytes)could be shown by preincubating erythrocytes with trypano-somes and subsequently testing the washed erythrocytes forinsertion of VSG by their susceptibility to lysis by complementin the presence of an anti-VSG antibody. Complement-mediated lysis was found to depend on the strain-specificanti-VSG antibody used. Extent of erythrocyte lysis increasedwith time of cell exposure to trypanosomes and with trypano-some concentration. No erythrocyte lysis was observed whentrypanosomes were preincubated with anti-VSG antibody be-fore adding erythrocytes. Purified membrane-form VSG(which retains the glycolipid anchor), but not soluble VSG(which no longer has the terminal diacylglycerol moiety), couldsensitize erythrocytes to anti-VSG antibody-mediated comple-ment lysis. The intermembrane transfer of VSG from trypa-nosomes to cells of the infected host could provide a molecularmechanism for the pathogenesis of trypanosomiasis.

African trypanosomes are primarily bloodstream parasites,but some of the organisms escape the vascular environmentand proliferate in the interstitial tissue. Wherever theseparasites localize, inflammatory reactions occur that lead toa multitude of pathological sequelae. While the most commonfeature of clinical and experimental African trypanosomiasisis anemia, other consequences of extravascular localizationof the parasite include meningoencephalitis, splenomegaly,and cachexia.The molecular events associated with the pathogenesis of

African trypanosomiasis are poorly understood. Intracellularparasites are not usually found, nor is there evidence that theparasites produce a toxic factor. The hypothesis examined inthis report is that the membrane form of the variant surfaceglycoprotein (mfVSG) of trypanosomes can, under the ap-propriate circumstances, transfer from the parasite plasmamembrane to that of a target cell. Such intermembranetransfer of variant surface glycoprotein (VSG) can provide amodel for the pathogenesis of trypanosomiasis.Trypanosome VSG belongs to a group of proteins that are

anchored in the plasma membrane by means of a phosphati-dylinositol-containing glycolipid covalently attached by eth-anolamine to the C-terminal end of the protein (1). Similarglycolipid moieties have been described for erythrocyteacetylcholinesterase (2) and decay accelerating factor (DAF)(3, 4). The intermembrane transfer of acetylcholinesterase

from erythrocytes to liposomes (5) and the selective incor-poration of DAF from a crude preparation into an acceptorcell (6) suggest that membrane proteins having such gly-colipid anchors are able to partition between the differentlipid environments of the donor membrane (e.g., the eryth-rocyte plasma membrane for acetylcholinesterase) and thetarget membrane (e.g., liposomes). The rate, direction, andextent of such intermembrane transfers depend on the rela-tive lipid composition and fluidity of the donor and acceptor(or target) membranes (7).

In this paper, we report that after in vitro incubation withtrypanosomes, erythrocytes become sensitive to anti-VSGantibody-mediated complement lysis. The data provide clearevidence for the transfer of mfVSG from trypanosomes to atarget membrane. This transfer of mfVSG may sensitize thecells to immune destruction and contribute to the anemiaassociated with African trypanosomiasis.

MATERIALS AND METHODSTrypanosomes. The following strains of Trypanosoma bru-

cei were used: strain 427, variant 117 (also known as MiTat1.4) (8), strain TREU 667 (9), and strain EATRO 110 (10).Three days after intraperitoneal inoculation, trypanosomeswere harvested from infected rat or mouse blood by differ-ential centrifugation and DEAE-cellulose chromatography(11). Trypanosomes, resuspended in minimum essential me-dium with Earle's salts (MEM) (GIBCO), were kept on iceand used immediately after purification.

Preincubation of Erythrocytes with Trypanosomes, SolubleVSG, and mfVSG. Fresh sheep blood was mixed with anequal volume of Alsever's bovine serum albumin and storedat 40C for at least 5 days, but <28 days, before use. Sheepblood >4 weeks old gave high antibody-independent lysisvalues and, hence, was not used in these studies. Before eachexperiment, sheep erythrocytes were washed three to fivetimes with barbital buffer (3 mM barbital/1.8 mM sodiumbarbital/145 mM NaCl/0.15 mM CaCl2/0.5 mM MgC12, pH7.4), until the supernatants were free of hemoglobin. Washedsheep erythrocytes were resuspended in barbital buffer at aconcentration of 109 cells per ml. The soluble form of VSG(sVSG), which lacks the C-terminal glycolipid moiety, wasisolated by the method ofCross (12), and mfVSG was purifiedby the method of Hereld et al. (13). Purity of sVSG andmfVSG was verified by SDS/urea acrylamide gel electro-phoresis (14).For each experiment, an aliquot of washed sheep eryth-

rocytes was centrifuged to pellet the cells. The cells wereresuspended in MEM containing either trypanosomes,sVSG, or mfVSG. In most experiments the final concentra-tions were as follows: 5 x 10' sheep erythrocytes per ml and108 trypanosomes per ml or 0.02 mg of sVSG or mfVSG perml. After incubation at 37°C for up to 30 min, cold buffer was

Abbreviations: VSG, variant surface glycoprotein; sVSG, solubleVSG; mfVSG, membrane-form VSG.

801

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

802 Cell Biology: Rifkin and Landsberger

added to cool and dilute the cell suspension. The tubes werecentrifuged at 200 x g for 10 min to pellet most of theerythrocytes while leaving most of the trypanosomes in thesupernatant. The erythrocytes were washed three times andthen resuspended in barbital buffer to a concentration of 109cells per ml.Complement Lysis Assay. Guinea pig serum, obtained from

Miles Scientific, was absorbed three times with washed sheeperythrocytes (0.3 ml packed cells per 10 ml of serum). Smallaliquots were stored at -20'C and used as a source ofcomplement. Anti-sVSG rabbit serum was prepared by im-munizing New Zealand White rabbits with purified T. bruceivariant 117 sVSG (15). The agglutination titer of this antise-rum with live trypanosomes was 1:400. Rabbit serum wasinactivated by incubation at 560C for 30 min and stored in1.0-ml aliquots at -70'C.To duplicate tubes at 0C, the following were added in

order: 0.2 ml of sheep erythrocytes (previously incubatedwith trypanosomes and washed, as described above), 0.2 mlof anti-sVSG antiserum (diluted 1:100 in barbital buffer), and2.6 ml of guinea pig complement (diluted 1:200 in barbitalbuffer). The tubes were incubated at 370C for 90 min withoccasional shaking and then centrifuged at 800 x g for 15 min.Optical density of the supernatant was determined at 540 nmto measure the amount of hemoglobin released. For eachexperiment, 100% lysis values were obtained by mixing 0.2ml of washed sheep erythrocytes with 2.8 ml of 0.1%Na2CO3. Control tubes contained sheep erythrocytes thathad been preincubated in MEM only, instead of with try-panosomes. Lysis in control tubes was usually very low(about 1-3%) but increased with the age of erythrocytes (upto 18%). The data are presented as % specific lysis, definedas [(% lysis with anti-sVSG antibody) - (% lysis withoutantibody)].

In initial experiments mouse, rat, and human erythrocyteswere used as target cells with human, rat, or guinea pig serumas a source of complement (in heterologous combinations). Inthe best of these experiments, specific anti-VSG antibody-mediated lysis was noted at levels only twice the lysis in theabsence of antibody. In contrast, the classical assay combi-nation for complement-mediated immune lysis consisting ofsheep erythrocytes, rabbit antibody, and guinea pig comple-ment (16) gave very little lysis (<2%) in the absence ofanti-VSG antibody and up to 90% lysis in the presence ofanti-VSG antibody. Therefore, all experiments reported hereused this combination of sheep erythrocytes, rabbit antibody,and guinea pig complement; experiments with cow erythro-cytes as target cells gave similar results but the percentspecific lysis was about half that observed with sheep cells.These results are consistent with observations by others thaterythrocytes from different species vary widely in theirsensitivity to immune hemolysis, even when purified anti-bodies and complement components are used; human eryth-rocytes, in particular, appear to be relatively resistant toimmune hemolysis, whereas sheep and bovine cells lyse morereadily (17, 18).Neuraminidase Treatment. Sheep erythrocytes were

washed twice with 20 mM Tris maleate buffer, pH 5.5/145mM NaCl (TM buffer) and resuspended in TM buffer to 5 x108 cells per ml. After addition of neuraminidase (Vibriocholerae, Calbiochem) (0.1 unit per 4 x 109 cells), theerythrocytes were incubated at 37°C. After either 30 or 60min, the incubation tubes were placed on ice, and an equalvolume of cold TM buffer was added. Control cells wereincubated without enzyme at 37°C for 60 min. The cells werepelleted by centrifugation (800 x g, 10 min) at 4°C, washedthree times in barbital buffer, and resuspended to 109 cells perml. A portion of the neuraminidase-treated cells was used forcomplement lysis assays as described above. The remainderof the treated erythrocytes was pelleted, resuspended in 40

vol of distilled water to lyse the erythrocytes, and centrifuged(25,000 x g, 20 min) to pellet the ghosts. The ghosts wereresuspended in 0.4 M H2SO4 and digested at 80'C for 60 minto release sialic acid. The ghosts were pelleted by centrifu-gation, and the supernatant was saved for sialic acid deter-mination by the thiobarbituric acid assay (19).

RESULTSFig. 1 outlines the experimental protocol used to demonstratethe intermembrane transfer of VSG from trypanosomes to atarget, or acceptor, cell. In the experiments reported here,sheep erythrocytes were used as the target cell; similarresults were obtained with erythrocytes from cattle, a naturalhost for African trypanosomes. The proposed model predictsthat during the first incubation (step 1) VSG transfers fromlive trypanosomes to erythrocytes. The presence of trans-ferred VSG on the surface oferythrocytes is then tested in thesecond incubation (step 2) by the ability of added anti-VSGantibody and complement to lyse the sensitized erythrocytes.The results in Table 1 show that erythrocytes, after incu-

bation with trypanosomes, become sensitive to lysis bycomplement in the presence of anti-VSG antibody, as de-picted in Fig. 1, and illustrate the immunological specificityof the phenomenon. When preimmune serum was used orwhen anti-VSG antibody was omitted, no lysis was observed.Moreover, negligible erythrocyte lysis was observed whentwo different strains of T. brucei, not recognized by theanti-117 VSG antibody, were used at 10-fold and 30-foldhigher concentrations. Thus, lysis is specific with respect tothe antigenic properties of the trypanosomes used.To assess whether the transfer ofVSG from trypanosomes

to erythrocytes depends on the incubation medium, thegrowth medium of Baltz (20) was compared with MEM alone.Phase-contrast microscopy of the trypanosomes after incu-bation showed no difference in the percentage of dead cells(1-2%). However, lysis of target cells was found to bedecreased by a factor of 3-5 when Baltz medium was used.In a similar experiment performed with sheep whole blood,lysis of erythrocytes also occurred at a reduced level (20%specific lysis after 30 min). These results suggest that eitherless VSG is released when trypanosomes are incubated inmore physiological media or that the albumin in these mediabinds some of the released VSG.The kinetics of transfer of VSG from trypanosomes to

erythrocytes was studied by incubating trypanosomes with

DONOR TARGET

trypanosomes rmc-. \v!SG roc

remnovai from-circulation

ysis withcomplemernl,

)/ / "'7.1IOC antta

SG

FIG. 1. Schematic diagram of the experimental protocol to dem-onstrate the spontaneous transfer of trypanosome VSG to a targetmembrane. (Step 1) Intermembrane transfer of VSG from trypano-somes (donor) to erythrocytes (rbc) (target) at 37TC. (Step 2) Immunelysis of washed erythrocytes with anti-VSG antibody and comple-ment to demonstrate the presence of transferred VSG on the eryth-rocyte surface.

Proc. Natl. Acad. Sci. USA 87 (1990)

I

I

i+

4c

Proc. Natl. Acad. Sci. USA 87 (1990) 803

Table 1. Trypanosome strain specificity of anti-VSG andcomplement-mediated erythrocyte lysis

Trypanosomespreincubated

with erythrocytes

Strain Cells per ml Antibody Specific lysis, %117 107 live Anti-117 VSG 67.7117 107 dead* Anti-117 VSG 86.8117 107 live Preimmune serum 0.6117 107 live No antibody 0.8667 108 live Anti-117 VSG 4.4110 3 x 108 live Anti-117 VSG 3.4110 3 x 108 dead* Anti-117 VSG 5.4None Anti-117 VSG 0.4Sheep erythrocytes at 5 x 108 cells per ml were preincubated with

trypanosomes at 37TC for 30 min and then washed three times beforeincubation with antibody and complement.*Trypanosomes were killed by three cycles of freezing and thawing.

erythrocytes for varying lengths of time at both 370 and 40C.Fig. 2A shows the time-dependent, linear increase in theextent of specific lysis of the erythrocytes after incubationwith live trypanosomes at 370C. When trypanosomes anderythrocytes were kept at 40C, the increase in lysis at 20 minover the time 0 value was <3%. Control erythrocytes incu-bated at 370C for 30 min without trypanosomes showed anegligible increase (<2%) in susceptibility to lysis.The extent of lysis of erythrocytes that had been resus-

pended in MEM containing trypanosomes and immediatelyplaced on ice (time 0) was significant (-12%). This resultcould have been due either to a very rapid initial rate oftransfer that we could not measure with the current experi-mental protocol or to the presence of some released VSG inthe trypanosome-containing solution used to resuspend theerythrocytes. The rapid insertion of such molecules into thetarget erythrocyte membrane might be expected. To test thishypothesis, high-speed supernatants were prepared frommedium in which purified trypanosomes were resuspendedand either immediately placed on ice (time 0) or incubated for15 and 30 min. The results in Fig. 2B show that VSG wasreleased into the medium in a time-dependent fashion and

A B

,°4° <40;~~50

Time (min)

FIG. 2. Extent of lysis of sheep erythrocytes after incubation withtrypanosomes for different lengths of time (A) or with high-speedsupernatants of medium in which trypanosomes had been incubatedfor different lengths of time (B). (A) Washed sheep erythrocytes wereresuspended in MEM containing 107 trypanosomes per ml andincubated at 370C. At the indicated times, samples were placed onice. (B) High-speed supernatants (134,000 X gave, 60 min, 4C) ofBaltz medium in which trypanosomes (1 x 108 per ml) had beenincubated at 370C for 0, 15, and 30 min were prepared. Washed sheeperythrocytes were resuspended in the high-speed supernatant prep-arations and incubated at 370C for 30 min. Control erythrocytes,resuspended in MEM or in Baltz medium without trypanosomes,were processed simultaneously. Complement lysis was assayed on

the washed, incubated sheep erythrocytes.

that significant amounts of VSG were already present in themedium at time 0. VSG release was much greater when thecells had been killed before centrifugation, resulting in asmuch as 82% lysis at time 0. These results are consistent withthe hypothesis that transfer of VSG occurs by way of theaqueous phase rather than by transient fusion or collision oftrypanosomes or trypanosome-derived small vesicles withthe target cell.

Specific lysis of erythrocytes also depends on the trypano-some concentration during the first incubation, as shown inFig. 3. Similar results were obtained whether live or freeze-thawed trypanosomes were used, except that the extent oflysis was always higher when equivalent numbers of freeze-thawed trypanosomes were used (cf. Table 1). Preincubationof erythrocytes with dead strain 117 at 107 cells per ml led tosignificant lysis, whereas preincubation with dead strain 110at 3 x 108 cells per ml did not (Table 1). Thus, it is unlikelythat the enhanced lysis observed with freeze-thawed trypa-nosomes was due to release of a toxic, nonspecific lyticmolecule.Trypanosomes disrupted by sonication or osmotic lysis are

known to release 90% of their VSG, mostly as sVSG (12). Aslow release of sVSG from live trypanosomes has also beendescribed (21, 22). To find out whether this form of VSG,which lacks a hydrophobic anchor, can sensitize erythro-cytes to antibody-mediated complement lysis, purified sVSGand mfVSG were incubated with erythrocytes. As shown inTable 2, only mfVSG was effective, indicating that transfer ofVSG from trypanosomes to erythrocytes requires the pres-ence ofthe hydrophobic anchor. The results also indicate thatthe insertion of mfVSG into the acceptor (target) membraneoccurs in an orientation that allows the protein to be recog-nized by specific antibody.Table 2 also shows that anti-VSG antibodies completely

block the transfer when present in the first incubation step.This finding is consistent with the transfer of mfVSG fromtrypanosomes to erythrocytes by means of a transient par-titioning of mfVSG into the aqueous phase because thepresence of antibody would not only crosslink VSG mole-cules on the surface of trypanosomes but would also formcomplexes with any VSG molecules that might have de-sorbed from the trypanosome membrane. Thus, this mech-anism of VSG transfer is similar to that described for inter-membrane protein transfer in other systems (23, 24)-i.e., viaprotein monomers or dimers in the aqueous phase.Because the presence of antibody during the incubation of

erythrocytes with trypanosomes prevents the transfer of

:> 60F80 I

0/60~~

204

00i o6 7o i80 105 106 107 108Tryps/m

FIG. 3. Extent of lysis depends on concentration of trypano-somes (Tryps) incubated with sheep erythrocytes. Washed sheeperythrocytes were resuspended to 5 x 108 cells per ml in MEMcontaining trypanosomes at the indicated concentrations. Afterincubation at 37TC for 20 min, the erythrocytes were washed and usedin a complement lysis assay. *, Live trypanosomes; 0, freeze-thawed trypanosomes.

Cell Biology: Rifkin and Landsberger

804 Cell Biology: Rifkin and Landsberger

Table 2. Comparison of strain 117 cells, sVSG- ormfVSG-induced erythrocyte lysis

SpecificPreincubation with erythrocytes lysis, %

107 live cells per ml 61.5sVSG 3.0mfVSG 84.6107 live cells per ml + anti-117 VSG 3.7108 dead cells per ml 80.4108 dead cells per ml + anti-117 VSG 2.3

sVSG and mfVSG were added at 0.02 mg per ml, which isequivalent to the total concentration of VSG on 1.3 x 107 cells perml. Trypanosomes were killed by three sequential freeze-thawcycles.

VSG, trypanosomes remaining in the second incubation stepwould probably not contribute to the observed erythrocytelysis. This hypothesis was tested directly by adding trypano-somes at 106 per ml to the second incubation step. Nodifference in lysis was found with or without added trypano-somes. Because in most experiments at least 90o ofthe initialtrypanosomes are washed out, we can conclude that thecontribution of contaminating trypanosomes to the specificlysis is negligible.To determine whether the charge on the erythrocyte plays

a role in the ability of erythrocytes to act as acceptor cells fortransfer of mfVSG from trypanosomes, erythrocytes weretreated with neuraminidase and then tested for the presenceof VSG on their surface by antibody-mediated complementlysis after incubation with trypanosomes. No difference wasfound between control cells and neuraminidase-treated cells(Table 3).

DISCUSSIONThe molecular events associated with the diverse patholog-ical sequelae of trypanosomiasis remain poorly understood.The hypothesis examined in this paper was whether changesin antigenic properties or functions of cells of the infectedhost might be explained by the insertion of trypanosomemfVSG into the plasma membrane of the host cell by meansof an intermembrane protein transfer mechanism.Intermembrane protein transfer is a phenomenon that has

been described previously for a variety of membrane pro-teins, not only those anchored via a phosphatidylinositol-glycan moiety (acetylcholinesterase) (7) but also those withamphipathic membrane binding properties (apolipoproteinsC) (22) and with hydrophobic peptide membrane-spanningregions (erythrocyte band 3, cytochrome b5) (25-27). Kineticstudies ofthese systems are consistent with models of proteintransfer through an aqueous phase and rule out models thatrequire membrane fusion or collision to effect transfer. Anaqueous transfer mechanism also readily explains the find-ings that protein is transferred in native orientation-i.e., theprotein on the target, or acceptor, membrane exhibits thesame susceptibility to enzyme digestion as on the donormembrane (25). These published findings suggested by anal-ogy that trypanosome VSG could transfer to a target cell andthat insertion of VSG into the target cell could result in

changes in the antigenic, metabolic, or functional propertiesof that cell.

In this paper we have demonstrated the intermembranetransfer of trypanosome VSG from living trypanosomes to atarget cell (the erythrocyte). This intermembrane transferconfers to the target cell antigenic properties that allow it tobe recognized by specific anti-VSG antibody and lysed bycomplement, indicating that VSG inserts into the targetmembrane in native orientation. The insertion into the targetmembrane requires the hydrophobic glycolipid anchor be-cause erythrocytes become sensitive to anti-VSG-mediatedcomplement lysis only after incubation with mfVSG but notwith sVSG.The role of charge interactions between the protein being

transferred and charges on the acceptor plasma membranecontributed by phospholipid headgroups or amino acid orsugar side chains of membrane proteins has not been exten-sively studied in other systems. We have shown (Table 3) thatremoval of sialic acid groups on the acceptor membrane doesnot appear to influence transfer of VSG to erythrocytes. Boththe importance of membrane charge, as well as membranefluidity, in intermembrane VSG transfer can be more readilystudied with liposomes of defined composition. Preliminaryexperiments (unpublished observations) have shown that VSGtransfers to a greater extent to lecithin liposomes containing10o phosphatidic acid than to pure lecithin liposomes.The conditions used in the experiments to demonstrate

transfer of VSG from trypanosomes to erythrocytes closelymimic those found in the infected host. In natural infectionsin cattle the parasitemia can reach 107 trypanosomes per mlof blood, whereas in experimental infections parasitemias upto 109 trypanosomes per ml are observed. Thus, the range oferythrocyte/trypanosome ratios in our experiments spansthe same range as that observed in infected animals. Fur-thermore, transfer ofVSG was observed when trypanosomeswere incubated in whole blood.The ability of trypanosome VSG to transfer to the cell

membranes of the infected host may provide a molecularexplanation for the pathological sequelae associated withtrypanosomiasis. Because anti-VSG antibody can block thetransfer ofVSG (Table 2), the in vivo transfer ofVSG to cellsof the infected host must precede anti-VSG antibody pro-duction. After the host mounts an immune response to theinfection, increased erythrocyte destruction by phagocytosisofopsonized erythrocytes could occur, resulting ultimately inanemia. With each successive wave of antigenically distinctparasites, progressively more erythrocytes would be de-stroyed. In fact, extensive erythrophagocytosis and spleno-megaly have been frequently reported to be associated withtrypanosome infections (28, 29). Moreover, the degree ofanemia in acute infections can be correlated with the numberof circulating parasites (30). Our hypothesis that intermem-brane transfer of VSG from trypanosomes to erythrocytesoccurs in vivo also explains the previously published findingsthat parasite antigens and immune complexes were found onerythrocytes during trypanosome infections (28, 31). Throm-bocytopenia, another common clinical complication of try-panosomiasis (32), can be explained by a similar mechanismof transfer of VSG to platelets, followed by anti-VSG anti-body-mediated aggregation and destruction. Cachectin/

Table 3. Effect of neuraminidase pretreatment of erythrocytes on VSG transferSialic acid, Specific Iysis, %nmol per 109 Sialic acid

Enzyme treatment erythrocytes removed, % - Trypanosomes + TrypanosomesNone 5.1 1.1 65.1Neuraminidase (37°C at 30 min) 1.9 63 9.6 68.1No difference was observed between erythrocytes treated with neuraminidase for 30 min or 60 min in the amount of sialic

acid removed or their sensitivity to lysis.

Proc. Natl. Acad. Sci. USA 87 (1990)

Proc. Natl. Acad. Sci. USA 87 (1990) 805

tumor necrosis factor production is stimulated by addition offreeze-thawed trypanosomes to macrophages (33). Thus,cachexia may be the end result of mfVSG transfer to mac-rophages. Of interest is whether transfer ofmfVSG to cells ofthe central nervous system can alter the function of thesecells and be responsible for the neurological disorders asso-ciated with late stages of trypanosomiasis.Intermembrane protein transfer may also contribute to the

pathogenesis of other parasitic infections, such as malariaand schistosomiasis. In malaria, splenic clearance of nonpar-asitized erythrocytes is well documented (34) and appears tobe, in part, responsible for the anemia associated withmalaria infections (35). The degree of anemia that develops,even in subclinical infections, is greater than that expectedfrom peripheral blood parasitemia. Because merozoites areknown to have a phospholipid-linked membrane protein ontheir surface (36), this protein could also transfer to nonpar-asitized erythrocytes, with consequent binding of antibodyand clearance of these cells by the reticuloendothelial sys-tem. In schistosomiasis, the schistosomule membrane ac-quires host molecules, such as glycolipids and major histo-compatibility antigens (37), and thereby contributes to anti-gen mimicry of the parasite. The mechanism of transfer ofhost molecules to the worm surface remains poorly under-stood. A mechanism involving intermembrane transfer, sim-ilar to that which we have described for trypanosome VSG,may also be operative in the acquisition of host antigens byschistosomes.There remain many unanswered questions associated with

the proposed model that transfer of VSG from trypanosomesto target cells of the infected host is responsible for thepathogenesis of trypanosomiasis. What are the properties ofthe target cell that regulate transfer? Clearly not all host cellsacquire trypanosome antigens. Does insertion ofmfVSG intoa target cell also affect the cell's function or metabolism? Ifso, is the altered state permanent or transient? Is therecooperativity in the transfer ofVSG to target cells-i.e., doesthe first transferred VSG molecule confer properties to thetarget cell that render it a more effective acceptor of otherVSG molecules?

The technical assistance of Lacey Washington, Jules Feledy, andMohinie Bharosay is gratefully acknowledged. We thank Dr. JudyFox for help with the preparation ofmfVSG and Dr. Anthony Ceramifor encouragement and critical reading of the manuscript. This workwas supported by National Institutes of Health Grants AI-20324 andGM-37778 and National Science Foundation Grant PCM8409213.

1. Cross, G. A. M. (1987) Cell 48, 179-181.2. Haas, R., Brandt, P. T., Knight, J. & Rosenberry, T. L. (1986)

Biochemistry 25, 3098-3105.3. Medof, M. E., Walter, E. I., Roberts, W. L., Haas, R. &

Rosenberry, T. L. (1986) Biochemistry 25, 6740-6747.4. Shak, S., Davitz, M. A., Wolinsky, M. L., Nussenzweig, V.,

Turner, M. J. & Gurnett, A. (1988) J. Immunol. 140, 2046-2050.

5. Bouma, S. R., Drislane, F. W. & Huestis, W. H. (1977) J. Biol.Chem. 252, 6759-6763.

6. Medof, M. E., Kinoshita, T. & Nussenzweig, V. (1984) J. Exp.Med. 160, 1558-1578.

7. Cook, S. L., Bouma, S. R. & Huestis, W. H. (1980) Biochem-istry 19, 4601-4607.

8. Cross, G. A. M. (1975) Parasitology 71, 393-417.9. Clarkson, A. B., Bacchi, C. J., Mellow, G. H., Nathan, H. C.,

McCann, P. P. & Sjoerdsma, A. (1983) Proc. Natl. Acad. Sci.USA 80, 5729-5733.

10. Rifkin, M. R. (1978) Exp. Parasitol. 46, 207-212.11. Lanham, S. M. & Godfrey, D. G. (1970) Exp. Parasitol. 28,

511-534.12. Cross, G. A. M. (1984) J. Cell. Biochem. 24, 79-90.13. Hereld, D., Krakow, J. L., Bangs, J. D., Hart, G. W. &

Englund, P. T. (1986) J. Biol. Chem. 261, 13813-13819.14. Cardoso de Alameida, M. L. & Turner, M. J. (1983) Nature

(London) 302, 349-352.15. Lamont, G. S., Tucker, R. S. & Cross, G. A. M. (1986) Par-

asitology 92, 355-367.16. Kabat, E. A. & Mayer, M. M. (1961) Experimental Immuno-

chemistry (Thomas, Springfield, IL).17. Yamamoto, K. (1977) J. Immunol. 119, 1482-1485.18. Linscott, W. D. (1967) J. Immunol. 98, 991-1003.19. Warren, L. (1959) J. Biol. Chem. 234, 1971-1975.20. Baltz, T., Baltz, D., Giroud, C. & Crockett, J. (1985) EMBO J.

4, 1273-1277.21. Shapiro, S. Z. (1986) Exp. Parasitol. 61, 432-437.22. Bulow, R., Nonnengasser, C. & Overath, P. (1989) Mol.

Biochem. Parasitol. 32, 85-92.23. Huestis, W. H., Daleke, D. L., Ferrell, J. E., Jr., & Lee, K. J.

(1986) Biophys. J. 49, 137 (abstr.).24. McKeone, B. J., Massey, J. B., Knapp, R. D. & Pownall,

H. J. (1988) Biochemistry 27, 4500-4505.25. Huestis, W. H. & Newton, A. C. (1986) J. Biol. Chem. 261,

16274-16278.26. Leto, T. L., Roseman, M. A. & Holloway, P. W. (1980) Bio-

chemistry 19, 1911-1916.27. Enoch, H. G., Fleming, P. J. & Strittmatter, P. (1977) J. Biol.

Chem. 252, 5656-5660.28. Jarvinen, J. A. & Dalmasso, A. P. (1977) Exp. Parasitol. 43,

203-210.29. Sadun, E. H., Johnson, A. J., Nagle, R. B. & Duxbury, R. E.

(1973) Am. J. Trop. Med. Hyg. 22, 323-330.30. Amole, B. O., Clarkson, A. B., Jr., & Shear, H. L. (1982)

Infect. Immun. 36, 1060-1068.31. Herbert, W. J. & Inglis, M. D. (1973) Trans. R. Soc. Trop.

Med. Hyg. 67, 268.32. Davis, C. E. (1982) Acta Trop. 39, 123-133.33. Hotez, P. J., Le Trang, N., Fairlamb, A. H. & Cerami, A.

(1984) Parasite Immunol. 6, 203-209.34. Greenwood, B. M. (1987) Parasitol. Today 3, 206-214.35. Phillips, R. E. & Warrell, D. A. (1986) Parasitol. Today 2,

271-282.36. Haldar, K., Ferguson, M. A. J. & Cross, G. A. M. (1985) J.

Biol. Chem. 260, 4969-4974.37. Capron, A., Dessaint, J. P., Capron, M., Ouma, J. H. &

Butterworth, A. E. (1987) Science 238, 1065-1072.

Cell Biology: Rifkin and Landsberger