Molecular and Biochemical Parasitology, 38 (1990) 121-134 Elsevier

12i

MOLBIO 01254

Association of Plasmodium berghei proteins with the host erythrocyte membrane: binding to inside-out vesicles Mark F. Wiser, Alan C. Sartorelli and Curtis L. Patton MacArthur Centerfor Molecular Parasitology. Yale University School of Medicine, New Haven, CT, U,S.A.

Two acidic phosphoproteins of Plasmodium berghei origin, of 65 and 46 kDa, are associated with the plasma membrane of the host mouse erythrocyte. The 65-kDa protein partitions between a soluble and particulate phase upon host cell lysis, whereas the 46-kDa protein is localized exclusively in the particulate fraction. Both proteins bind to inside-out vesicles derived from erythrocyte ghosts and the conditions of the reassociation reaction indicate that the binding is specific and that the proteins interact only with the cytoplasmic face of the erythrocyte membrane. The 65-kDa protein appears to exist in two membrane-associated states; one loosely bound, which readily dissociates from the membrane, and a more tightly associated state, which does not dissociate under non-denaturing conditions. The 46-kDa protein is tightly bound to the host erythrocyte membrane and does not dissociate. Cross-linking studies suggest that both of these parasite proteins interact with the submembrane cytoskeleton of the erythrocyte, and that the 65-kDa protein also appears to interact simultaneouslywith the lipid bilayer and erythrocyte membrane proteins. However, direct interaction between the malarial proteins and distinct erythrocyte membrane proteins could not be demonstrated. In summary, these findings indicate that the acidic phosphoproteins of the malarial parasite interact with the cytoplasmic face of the erythrocyte membrane both in vivo and in vitro. Key words: Plasmodium berghei; Erythrocyte membr~:ne; Malarial parasite

Introduction One distinguishing feature of Plasmodium is that it resides within an erythrocyte during one phase of its life cycle. Obviously, during the invasion process all intracellular parasites must interact intimately with the plasma membrane of the host and, indeed, some progress has been made in defining such interactions at the cellular and molecular levels as the malarial parasite invades the host erythrocyte [1]. In addition, during the intracellular trophic period, the malarial parasite also interacts with and alters the host erythrocyte Correspondence (presenO address: M.F.Wiser, Department of Tropical Medicine and Parasitology, 1501 Canal Street, Tulane University Medical Center, New Orleans, LA 70112, U.S.A. Abbreviations: GESB, gel electrophoresis sample buffer; IOV, inside-out vesicles; Mab, monoclonal antibody; PMSF, phenylmethylsulfonylfluoride; NEM, N-ethylmaleimide;IVies, 4-morpholineethanesulfonic acid; EGTA, ethylene glycol tetraacetic acid.

membrane, inducing both structural and permeability changes [2,3]. Although it is not yet known precisely how the parasite produces these changes in the erythrocyte membrane at the molecular level, it is known that the parasite synthesizes proteins directed to the host erythrocyte membrane and that these proteins possibly actuate the alterations observed. The appearance of acidic phosphoproteins of parasite origin on the host erythrocyte m e m b r a n e has been previously reported. In particular, three proteins, a 65-kDa Plasmodium berghei protein, a 46-kDa P. berghei protein and a 93-kDa Plasmodium chabaudi protein, referred to as Pb(em)65, Pb(em)46 and Pc(era)93, respectively, have been identified and characterized [4-6]. Differential solubility studies indicate that these are not integral m e m b r a n e proteins, but are peripheral membrane proteins [4,5]. This characteristic, in combination with the inaccessibility of these proteins from the external surface of the erythrocyte, infers that these neoproteins interact with the cytoplasmic face of the erythrocyte mem-

0166-6851/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)

122

brane [6]. Therefore, we speculate that the malarial parasite induces the changes observed on the host erythrocyte membrane by interacting with and affecting its submembrane cytoskeleton. Likewise, others have speculated that Plasrnodium [alciparum synthesizes proteins which induce 'knob' formation by interacting with the host's submembrane cytoskeleton [7,8]. That the parasite can affect membrane structure and function via the host's cytoskeleton is feasible, in that the erythrocyte membrane skeleton is believed to regulate many aspects of membrane topography [9,10]. However, whether the acidic phosphoproteins, or any other malarial proteins, do indeed affect changes in the host erythrocyte membrane by interacting with the cytoskeleton has not been conclusively demonstrated. To learn about the possible role(s) of these acidic phosphoproteins, we carried out experiments to more precisely define the nature of the interaction between the P. berghei proteins, Pb(em)65 and Pb(em)46, and the erythrocyte membrane. Pb(em)65 partitions between the host erythrocyte cytosol and the host erythrocyte membrane upon host cell lysis; whereas, Pb(em)46 is found exclusively as a particulate protein. Both proteins bind in a specific manner to inside-out vesicles (IOV) derived from erythrocyte ghosts. Reversible cross-link~.ng experiments suggest that both Pb(em)65 and Pb(em)46 interact with the submembrane skeleton, but Pb(em)65 may also interact simultaneously with host cell lipid. Materials and Methods

General. Maintenance of P. berghei and the general procedures for preparing infected erythrocytes were carried out as previously reported [11]. Monoclonal antibodies (Mabs) em 16.3 and em 17.2, which recognize Pb(em)65 and Pb(em)46, respectively, have been previously characterized [6]. Polyacrylamide gel electrophoresis and immunoblotting were performed as described earlier [6]. Inside-out vesicle preparation. Erythrocyte ghosts were prepared by hypotonic lysis of erythrocytes in hemolysis buffer (5 mM phosphate, pH 8.0)

and centrifugation at 22 000 × g for 20 min. The pellet was repeatedly washed in the hemolysis buffer until ghosts were hemoglobin-free. Ghosts from infected erythrocytes were prepared in a similar fashion, except that the malarial pigmentcontaining 'button' was removed and discarded between washes. Ghosts were washed with a 1:1 mixture of hemolysis buffer and extraction buffer (0.1 mM E D T A , pH 9.2), centrifuged at 27000 x g for 20 min, and then resuspended in 7-10 vols. of extraction buffer. They were incubated at 37°C and monitored by phase contrast microscopy until vesiculation occurred (usually 10-20 rain). IOV were recovered by centrifugation at 27 000 x g for 30 rain and washed once with extraction buffer. Vesicles were also prepared from ghosts by extracting all of the peripheral membrane proteins with 0.1 M NaOH as previously described [4] and washed once with hemolysis buffer. Alkali-extracted IOV were prepared by incubation of IOV with 0.1 mM EDTA, pH 11.0, for 20 min at room temperature; centrifuged at 27000 x g for 30 min and washed once in hemolysis buffer.

Pb(em)65 isolation. Following passage over CF-11 and washing in Hanks' balanced salt solution, infected erythrocytes were washed once in K-1 buffer [12] wit;:out Ca 2+ and Mn 2÷ and then resuspended at approximately 5% hematocrit in K1 buffer containing 0.01% saponin, 5 mM EGTA and 1 mM phenylmethylsulfonyl fluoride (PMSF). Following incubation at 37°C for 5 min, the lysate was centrifuged at 1000 × g for 5 min, followed by decantation and centrifugation of the supernatant at 27000 x g for 20 min. Pb(em)65 was precipitated from the erythrocyte cytosol (second supernatant) by the addition of an equal volume of saturated (NH4)2SO 4 and collection by centrifugation. The protein pellet was dissolved in dialysis buffer (20 mM phosphate, pH 7.0, 0.15 M KCI, 1 mM E G T A and 0.1 mM PMSF) and insoluble material was removed by centrifugation. The supernatant was reprecipitated with 50% saturated (NH4)2SO4 as described above. The resulting protein pellet was dissolved in a volume approximately equal to the original packed cell volume and either dialyzed overnight against two changes of dialysis buffer or boiled as described below.

123

Alternatively, after the second (NH4)2SO 4 precipitation, the sample was boiled for 5 min and the denatured protein was removed by centrifugation. The supernatant was dialyzed as described above. Pb(em)65 was also isolated by two rounds of (NH4)2SO4 precipitation and boiling as described above from infected erythrocytes radiolabeled with [14C]proline as previously described [6].

Reassociation reaction. Reassociation of Pb(em)65 with IOV was carried out in a final volume of 100 Ixi containing 50 mM Mes, pH 62, 0.5 to 1.0 mg ml -I of IOV and 50 to 75 ~1 of the Pb(em)65 preparation. Following a 45-rain incubation on ice, the samp?ca were centrifuged for 15 min in a microfuge ano tl',e supernatants and pellets were separated. The supernatant and pellet were then restored to their original volumes (i.e., 100 ~1) and solubilized in gel electrophoresis sample buffer (GESB). Alternatively, samples were layered onto an equal volume of 20% sucrose (w/v) and centrifuged as previously described [13]. The supernatant was discarded and the pellet was used in other analyses. The amount of Pb(em)65 associated with the IOV was determined by SDSgel electrophoresis and immunoblotting using Mab em 16.3. Spectrin-actin isolation. Erythrocyte ghosts were extracted with 0.1 mM EDTA, pH 9.2, as described for the preparation of IOV. An equal volume of saturated (NH4)2SO4 was added to the su,';ernatant and the mixture was incubated overnight at 4°C. Precipitated protein was collected by centrifugation and the pellet was dissolved in 20 mM Tris-HCl, pH 8.0, and dialyzed for 24 h against two changes of the same buffer. Insoluble material was removed by centrifugation at 27 000 x g for 30 min and the supernatant was aliquoted and stored at -70°C. IOV were incubated with the crude spectrinactin preparation for 1 h at room temperature in 50 mM Mes, pH 6.2, layered onto 20% sucrose (w/v) and centrifuged. Soluble material was discarded and the spectrin-actin replenished IOV were reassociated with Pb(em)65 as described above.

Pb(ern)46 isolation. Erythrocyte ghosts from infected cells were prepared as described above and extracted with 20 mM Tris-HCl, pH 8.8, containing 6 M urea, 1 mM EGTA and 1 mM PMSF at 37°C for 30 min. Following centrifugation at 27 000 x g for 30 min, the supernatant was retained and dialyzed overnight against 100 vols. of 20 mM Tris-HCl, pH 8.8, containing 1 mM EGTA and 0.1 mM PMSF, and recentrifuged at 27000 × g for 30 min. The supernatant was precipitated with 30% saturated (NH4)2SO4 and then with 70% saturated (NH4)2SO 4. The 30% pellet contained the majority of Pb(em)65 and no Pb(em)46, whereas the 70% pellet contained some Pb(em)65 and all of the Pb(em)46. Both pellets were dissolved in dialysis buffer and dialyzed overnight as described for the preparation of Pb(em)65. The major erythrocyte membrane proteins distributed between both of the (NH4)2SO4 fractions. Protease pretreatment of inside-out vesicles. IOV (5 mg m1-1) were prepared as described above and incubated with 0, 0.1, 0.3, or 2.0 p,g m1-1 of papain for 30 min at 37°C. The IOV were washed 3 times in hemolysis buffer containing 5 mM Nethylmaleimide (NEM) and once in hemolysis buffer. These IOV were then reassociated with Pb(em)65 as described above. Reversible cross-linking. Ghosts were prepared from infected erythrocytes and diluted approximately 10-fold. Cross-linking was carried out with 1 mM 4,4'-dithiobis(phenylazide) (Pierce Chemical Co.) at room temperature with illumination for 2 min from a hand-held UV lamp (254 nm) at a distance of 3 cm. lodoacetamide was then added to a final concentration of 50 mM, followed by a 30-min incubation on ice. Membranes were recovered by centrifugation, solubilized in GESB without 13-mercaptoethanol and subjected to electrophoresis on 6% acrylamide tube gels. Following electrophoresis in the first dimension, the gels were incubated with GESB containing 5% 13mercaptoethanol (10 ml per gel) for 30 min at room temperature with rocking and subjected to electrophoresis in a second dimension on 9% acrylamide slab gels. The gels were stained with Coomassie Blue and the polypeptides transferred to nitrocellulose. Pb(em)65 and Pb(em)46 were detected by immunobiotting.

124

Results

m

Two acidic phosphoproteins are associated with the host erythrocyte membrane during P. berghei infection [4,6]. In these studies, the host erythrocyte cytoplasm was not examined for the presence of these proteins. The availability of specific Mabs allows the unambiguous determination of the distribution of these proteins. Pb(em)46 is found exclusively as a particulate protein, whereas Pb(em)65 distributes both in the soluble fraction and the particulate fraction when infected erythrocytes are lysed with 0.01% saponin and subjected to centrifugation (Fig. 1). The use of saponin, which only lyses the host erythrocyte and not the parasite [14], indicates that Pb(em)65 is a soluble protein present in the host erythrocyte

t

s

p

t

s

p

205 1169766-

29-

em 17.2

em 16.5

Fig. 1. Distribution of Pb(em)65 and Pb(em)46 within infected erythrocytes. Infected erythrocytes were suspended at a 10% hematocrit in K-1 buffel containing 0.01% saponin and incubated for 5 min at 37°C. Half of the sample was subjected to centrifugation [or 15 min in a microfuge, whereas the other half was solubiliz~d in GESB. Following centrifugation, the supernatant and pellet were carefully separated, restored to their original volumes and solubilized in GESB. All three fractions, total (t), supernatant (s) and pellet (p), were subjected to SDS-gei electrophoresis, transferred to nitrocellulose and analyzed by immunoblotting with Mabs em 17.2 and em 16.3 as indicated. Molecular mass standards (in kDa) are as indicated.

t s

p t s p t

A

B

s p

C

t

s

p

D

Fig. 2. Low-ionic-strength extraction of membrane proteins from ghosts. Ghosts were isolated and washed free of hemoglobin from normal (A) and P. berghei infected (B-D) erythrocytes. They were then suspended in 10 vols. of 0.1 mM EDTA, pH 9.2, and incubated at 37°C for 20 rain. Half of the sample was subjected to centrifugation in a microfuge for 30 min, and the other half was solubilized in GESB. Following centrifugation, the supernatant and pellet were carefully separated, restored to their original volumes and solubilized in GESB. The three samples, total (t), supernatant (s) and pellet (p), were subjected to gel electrophoresis and either stained with Coomassie Blue (panels A and B) or transferred to nitrocellulose and analyzed by immunoblotting with Mabs em 17.2 (C) and em 16.3 (D). Molecular ,liass standards (kDa) are as indicated (lane m).

cytoplasm or possibly in the parasitophorous vacuolar space. The distribution of both proteins is the same regardless of whether the infected cells are lysed with saponin as shown, or are lysed osmotically or hypotonically (data not shown), indicating that the cytoplasmic form of Pb(em)65 is not the result of any particular lysis method, such as detergent solubilization. Differential solubility studies indicated that both Pb(em)46 and Pb(em)65 are peripheral membrane proteins [4]. Low-ionic-strength extraction can further distinguish peripheral membrane proteins of the erythrocyte. Extraction of erythrocyte ghosts with 0.1 mM EDTA, pH 9.2, results in the solubilization of spectrin and actin, whereas ankyrin and band 4.1/4.2 remain associated with the vesicle fraction along with the integral membrane proteins (Fig. 2; ref. 13). This is observed in ghosts from both normal and infected erythrocytes. Both Pb(em)46 and Pb(em)65, however, distribute between both the soluble and particulate fractions under these con-

125

ditions (Fig. 2C and D), indicating that neither protein behaves in a manner identical to that of the major peripheral membrane proteins of the erythrocyte. However, a larger percentage of Pb(em)65 is solubilized under these conditions than that of Pb(em)46, indicating a stronger association of Pb(em)46 with the erythrocyte membrane, a finding in agreement with previous resuits [4]. The soluble form of Pb(em)65 associated with the host erythrocyte cytoplasm was also characterized in terms of its ability to interact with erythrocyte membranes in vitro. Pb(em)65 binds to IOV derived from erythrocyte ghosts but not to the ghosts themselves or to vesicles prepared from NaOH extracted ghosts (Fig. 3). The binding of Pb(em)65 to IOV and not to ve~icles derived after NaOH extraction or to ghosts indicates that the results are not due to non-specific adsorption or to trapping of Pb(em)65 within the vesicles. The latter was further substantiated by treating Pb(em)65-reconstituted IOV with proteases. All the Pb(em)65 is digested under these conditions (data not shown), indicating that Pb(em)65 is accessible to the protease and therefore not trapped within the IOV. In addition, the lack of Pb(em)65 binding to ghosts implies that binding is specific for the cytoplasmic face of the erythrocyte membrane. Unfortunately, the sidedness of the vesicles derived by extraction of ghosts with NaOH is not known, but in contrast to a previous report [15], NaOH extraction did result in the vesiculation of the ghosts as observed by phase-contrast microscopy. However, IOV extracted with 0.1 mM EDTA, pH 11.0, which are equivalent in terms of Coomassie Blue-stained proteins to the NaOHextracted ghosts, still bind Pb(em)65 (see below). This suggests that the lack of Pb(em)65 binding to vesicles derived by NaOH extraction is due to an outside-out orientation, further substantiating the specificity of Pb(em)65 for the cytoplasmic face of the erythrocyte membrane. Interestingly, Pb(em)65 remains soluble after boiling and this heat-stable Pb(em)65 behaves identically to unboiled Pb(em)65 in that it binds to IOV (Fig. 3, lanes 5) and not to ghosts or NaOH-extracted ghosts (data not shown). Pb(em)65 in a crude erythrocyte cytosol prepa-

ration also binds to IOV indicating that the degree of purity of Pb(em)65 has no effect on the specific activity of IOV binding and also suggest-

I s

ms

2 p

s

I

3 p

2 p sps

s

4 p

:5

s

psp

5 p

4

s

p

5 s p

205' 116' 97' 66'

45

¸

Fig. 3. Reassociation of Pb(em)65 with host erythrocyte membranes. Partially purified Pb(em)65, prepared by two rounds of (NH4)~SO4 precipitation (lanes 1--4), or the same sample boiled for 5 rain (lanes 5), were incubated with various erythrocyte membrane preparations as described in the Materials and Methods. Following centrifugation, the supernatant (s) and the pellet (p) were separated and analyzed by SDS-gel electrophoresis. The top panels are immunoblots probed with Mab em 16.3 and the bottom panels are the corresponding Coomassie Blue-stained gels. No added membranes (lanes 1), erythrocyte ghosts (lane 2), 0.1 N NaOH-extracted ghosts (lanes 3) and IOV (lanes 4 and 5) were tested for their abilities to bind Pb(em)65. The positions of the molecular mass standards (in kDa) are as indicated (lane m) and the arrow denotes Pb(em)65.

126

ing that no macromolecular activators or inhibitots of binding are present. IOV in which spectrin and actin have been replenished also are capable of binding Pb(err,)65 (Fig. 4), inferring that the binding of Pb(em.~65 to IOV is not due to some nonphysiological binding site exposed as a result of the extraction of spectrin and actin. These results imply that Pb(em)65 can recognize the erythrocyte membrane as it appears in vivo. The coincubation of the crude spectrin-actin preparation and the

I

2

3

4

5

~~-205 fq~66

-36

q-29 q-24 Fig. 4. Binding of Pb(em)65 to spectrin-actin replenished inside-out vesicles. IOV were incubated with a crude spectrinactin preparation (lanes 3 and 5) or 20 mM Tris-HCl, pH 8.0 (lanes 2 and 4) and centrifuged through 20% sucrose as described in the Materials and Methods. These IOV were then incubated with the boiled Pb(em)65 preparation as described in Fig. 3. Following centrifugation, the pellets were analyzed by SDS-gel electrophoresis and immunoblotting with Mab em 16.3. The panels show the amount of Pb(em)65 bound in the absence of added IOV (I), control IOV (2) and spectrin-actin-replenished IOV (3), as well as the Coomassie Blue-stained control IOV (4) and the spectrin-actin-replenished IOV (5). Molecular mass standards (kDa) are indicated on the right; the arrow denotes Pb(em)65.

Pb(em)65 preparation does not affect the binding of either component to IOV (data not shown). Various parameters of the reassociation reaction were also investigated (data not shown). The pH optimum is 6.5 to 7.5 and KCI concentrations from 0 to 0.2 M have no effect, indicating that the reassociation is not simply due to electrostatic interactions. Maximal binding occurs after approximately 45 min of incubation on ice with halfmaximal binding taking place in about 5 min. Temperatures up to 37°C have no effect, except that the reaction appears to proceed a bit more rapidly at higher temperatures. Conditions which increase the ratio of bound to unbound Pb(em)65 have not been found. In addition, IOV prepared from human erythrocytes also bind Pb(em)65. Early in this study, problems with proteolysis resulted in Pb(em)65 preparations contaminated with proteolytic fragments of 60 and 55 kDa which are recognized by Mab em 16.3 (note minor bands in Figs. 1 and 3). These proteolytic fragments do not bind to IOV, indicating that intact Pb(em)65 is necessary for IOV binding activity or that the binding domain of Pb(em)65 is located near one of the termini of the protein. These results support the premise that Pb(em)65 binding is specific, since a minor degree of proteolysis completely abolishes binding. If the binding were nonspecific, a decrease in size from 65 kDa to 60 kDa (less than 10%) would not be expected to completely abolish binding. Quantitation of the binding process is not practical using immunobiotting as the detection system. Therefore, Pb(em)65 was metabolically labeled with [taC]proline and used in the reassociation reaction (Fig. 5). The results further demonstrate the specificity of Pb(em)65 binding, in that other malarial proteins present in the Pb(em)65 preparation do not bind to IOV except for a 52-kDa protein (Fig. 5A). Quantitatively, the binding of the 52-kDa protein is dearly less than that of Pb(em)65, which correlates with the rapid dissociation the 52-kDa protein (Fig. 6), suggesting that the interaction of this protein with IOV may be non-specific. Peptide mapping and the lack of recognition by Mab ern 16.3 indicate that the 52-kDa protein is not a proteolytic fragment of Pb(em)65 (data not shown). The radiolabeled Pb(em)65 preparation was also serially di-

127

S

A

P

5

P

S

116-

P

S

nonetheless, these results suggest that a high-affinity interaction occurs between Pb(em)65 and the erythrocyte membrane. To examine the dissociation reaction, IOV were reconstituted with radiolabeled Pb(em)65 and isolated by centrifugation through 20% sucrose. This procedure resulted in IOV which were easily resuspended, and problems with possible damage to the IOV were avoided. The Pb(em)65-reconstituted IOV were then incubated in the presence or absence of non-radioactive Pb(em)65. Some of the Pb(em)65 rapidly dissociates from the IOV (Fig. 6A), and the amount of Pb(em)65 which dissociates is the same for incubation periods from 10 min to 16 h. Since the ratio of bound to unbound radioactive Pb(em)65 is the same in the presence or absence of unlabeled Pb(em)65, the protein does not rapidly dissociate and then reassociate with the IOV. Instead, the results suggest that Pb(em)65 exists in two states, a loosely as-

P

--~

9 7 -

- .

.

.

.

.

66-

4529-

B

0.3

[]

~D ii

[i

~= 0,2

-s

0.1 0

,

,

,

1

2

3

4

pmoles Pb(em)65 B o u n d

Fig. 5. Quantitation of the binding of Pb(em)65 to inside-out vesicles. Pb(em)65 was prepared from [1 4 C]proline-labelcd erythrocytes and reassociated with IOV as described in Materials and Methods. Following centrifugation, the supernatant (s) and pellet (p) were analyzed by SDS-gel electrophoresis and fluorography (A). The pair of lanes on the far right represents the undiluted preparation of the [l~C]proline-labeled Pb(em)65, and moving to the left are 1:2 serial dilutions of this preparation. Molecular mass standards (kDa) are indicated on the left, and Pb(em)65 is denoted by the arrow. The amount of Pb(em)65 bound (p) and free (s) were quantitated by excising the bands from the gel and subjecting them to scintillation spectrometry and plotting according to Scatchard (B). The linear correlation coefficient is -0.93 and the intrinsic binding constant (K) calculated from the negative of the slope is 3 × 10I° M -1.

luted and the amount bound at various concentrations determined. Analysis of the resuits by a Scatchard plot [16] reveals an intrinsic binding constant (K) of 3 x 10 t° M -t (Fig. 5B). However, this value should be interpreted cautiously, since saturation was never attained;

-I-

p

s

--

p

s

"I-

p

s

p

11697 . 66 - ~

.

.

.

"

~== '==~

m

A

[3

.......

,,4

45-

29-

Fig. 6. Dissociation of Pb(em)65 from inside-out vesicles. Radiolabeled Pb(em)65 was reassociated with IOV as described in Fig. 5, except tilat the IOV were recovered by centrifugation through 20% sucrose. The Pb(em)65-reconstituted IOV were then resuspended in the absence ( - ) or presence (+) of unlabeled Pb(em)65. Following cen~.rifugation, the supernatant (s) and pellet (p) were separated and analyzed by SDS-gel electrophoresis and fluorography (A). After fluorography, the gel was rehydrated, transferred to nitrocellulose and analyzed by immunoblotting with Mab em 16.3 (B). Molecular mass standards (in kDa) are indicated on the right and the arrow denotes Pb(em)65.

128

sociated state, which readily dissociates, and a tightly associated state, which does not readily dissociate. Additional Pb(em)65 can be added to IOV previously reconstituted with Pb(em)65, indicating a lack of saturation (Fig. 6B). Pb(em)46, extracted from ghosts and partially purified, also binds to IOV (Fig. 7). In contrast to Pb(em)65, all of the Pb(em)46 present in the preparation binds to IOV, suggesting a stronger interaction of Pb(em)46 with the erythrocyte membrane. Pb(em)46 also remains soluble after boiling and the heat-treated Pb(em)46 retains IOV binding activity. In addition, Pb(em)46 does not bind to vesicles prepared by NaOH extraction or to ghosts (data not shown), indicating that the interaction with IOV is not the result of nonspecific interactions with vesicles and that

2

I s

A

p

s

3 p

s

p

A

29-

Fig. 7. Binding of Pb(em)46 to inside-out vesicles. Pb(em)46 was extracted from ghosts, prepared as described in Materials and Methods and incubated with IOV. Following centrifugation, the supernatant (s) and pellet (p) were separated and analyzed by SDS-gel electrophoresis and immunoblotting with Mab em 17.2. The panels show the pelletable Pb(em)46 in the absence (lanes 1) or the presence (lanes 2 and 3) of added IOV. Boiled Pb(em)46 also reassociates with IOV (lanes 3). Molecular mass standards (kDa) are indicated and the arrow denotes Pb(em)46.

1 2 5 4 5 205

.....

liB--..97--,~

6

I

2

5 4 5 6

~ = = ~

:

66-~

Fig. 8. The effect of papam pre-treatment of inside-out vesicles on the binding of Pb(em)65. IOV were isolated and digested with papain as described in Materials and Methods. Following incubation with the Pb(em)65 preparation, the IOV were recovered by centrifugation and analyzed by SDS-gel electrophoresis and immunoblotting. The Coomassie Bluestained proteins (left) and the amount of Pb(em)65 bound (right) are shown for no added IOV (lanes 1), IOV not subjected to any additional treatment (lanes 2), IOV treated with 0 (lanes 3), 0.1 (lanes 4), 0.5 (lanes 5) and 2.0 (lanes 6) v,g ml -t of papain and then washed in the presence of NEM. Molecular mass standards (kDa) are indicated on the left; the arrow denotes Pb(em)65.

Pb(em)46 binding is specific for the cytoplasmic face of the erythrocyte membrane. Because of problems with abundance and recovery of Pb(em)46 after extraction from the membrane, the binding of Pb(em)46 has not been as extensively characterized as that of Pb(em)65; in general, however, the conditions required for the binding of the two proteins are similar. IOV were pretreated with papain to determine if the protein components were necessary for Pb(em)65 binding. Limited digestion with papain has little effect on the binding of Pb(em)65 and more extensive digestion only results in a partial inhibition of the binding (Fig. 8). The increase in Pb(em)65 binding seen after digestion with 0.5 ixg ml -~ papain (Fig. 8, lane 5) was not reproducible, and was attributed to variability. In general, the amount of Pb(em)65 bound to such IOV was the same or slightly less than that of control IOV. The supernatants from the binding assays were also examined and no digestion of Pb(em)65 had occurred. Treatment of IOV with NEM to inhibit the papain results in a slight decrease in Pb(em)65 binding (Fig. 8, lanes 2 and 3). Pb(em)65 binding

129

is also relatively insensitive to the pretreatment of IOV with similar amounts of trypsin (data not shown). Likewise, extraction of IOV with 0.1 mM EDTA, pH 11.0, only partially inhibits the binding of Pb(em)65 (Fig. 9). However, the binding of Pb(em)46 is completely inhibited under these conditions. Since the remaining peripheral proteins (i.e., ankyrin, bands 4.1, 4.2 and 4.9) are extracted from IOV under these conditions the results imply that these peripheral proteins possibly play a partial role in the binding of Pb(em)65 and a more absolute role in the binding of Pb(em)46 to the erythrocyte membrane. It is unlikely that this treatment of IOV results in a conversion to outside-out vesicles, since similarly prepared vesicles bind band 4.2 in a saturable manner [17]. Components which interact can be chemically cross-linked and then identified after electrophoresis in a first dimension under conditions that I

2

3

123

2

3

preserve the cross-links followed by a second-dimensional separation using conditions which break the cross-links [18]. Such reversible crosslinking experiments, or nearest neighbor analyses, were employed to ascertain whether the P. berghei proteins interact with particular compo-

49-

5-

,,t

B

m

,

...... 2 0 5

~

,,,=,,'- I I 6 ~-~ ~-66 i

-4s

w-29 Pb (era) 65

Pb(em) 4 6

CB

Fig. 9. Binding of Pb(em)65 and Pb(em)46 to alkali-extracted inside-out vesicles. IOV were extracted with 0.1 mM EDTA, pH 11.0, as described in Materials and Methods, and incubated with either the Pb(em)65 or the Pb(em)46 preparation. Following centrifugation, the pellets were analyzed by SDS.gel electrophoresis and immunoblotting. Coomassie Bluestained proteins (CB) and the amounts of bound Pb(em)65 and Pb(em)46 are shown (as indicated) for no added IOV (lanes 1), control IOV (lanes 2) and alkali-extracted IOV (lanes 3). Molecular mass standards (kDa) are indicated on the right (lame m).

Fig. 10 Reversible cross-linking of Pb(em)46 and Pb(em)65 to erythrocyte membrane proteins. Hemoglobin-free ghosts were isolated from P. berghei infected erythrocytes and incubated with 1 mM 4,4'-dithiobis(phenylazide) as described in Materials and Methods. The samples were then solubilized in GESB without IB-mercaptoethanol and subjected to SDSgel electrophoresis on 6% aerylamide gels (horizontal dimension). These gels were then incubated with GESB containing 5% [3-mereaptoethanol and subjected to a second dimensional separation on 9% acrylamide gels. The gels were then stained with Coomassie Blue, transferred to nitrocellulose and analyzed by immunoblotting with Mab em 17.2 (A) or Mab em 16.3 (B). The arrows indicate the proteins recognized by the antibodies+ which stained pink in the original, whereas all of the other proteins stained blue. Pb(em)65 streaked all the way across the first dimension (B, between the two arrows), whereas Pb(em)46 did not and is localized to two distinct spots (A). The circles indicate the positions of Pb(em)46 and Pb(em)65 when the samples were not treated with cross-linking agents. The positions of the major crythrocyte membrane proteins are indicated on the left in panel A.

130

nents of the erythrocyte membrane. Isolated ghosts from P. berghei infected erythrocytes were treated with 4,4'-dithiobis(phenylazide) and analyzed by two-dimensional gel electrophoresis under non-reducing conditions in the first dimension, and then under reducing conditions in the second dimension. Major erythrocyte membrane proteins were detected by Coomassie Blue staining and Pb(em)46 and Pb(em)65 were detected by immunoblotting (Fig. 10). Pb(em)46 and Pb(em)65 are not cross-linked to any particular erythrocyte membrane proteins. Both proteins are cross-linked to very high-molecular-mass complexes which do not enter the gel, but remain at its top (left side of gels). These co~nplexes also contain all of the major erythrocyte membrane proteins (i.e., spectrin, ankyrin, band 3, band 4.1/4.2 and actin). These results suggest that Pb(em)65 and Pb(em)46 may not recognize any particular component of the erythrocyte membrane, but rather recognize the erythrocyte membrane as an entity. In addition, Pb(em)65 streaks across the first dimension when subjected to cross-linking (Fig. 10B). This streaking is not likely to be due to disulfide exchange, since treatment of the samples with iodoacetamide before electrophoresis has no effect on the streaking. One possible explanation is that Pb(em)65 interacts with lipids and multiO t t i l l - " ~tlK;'lfltlt%.lll~ represents increments in molecular mass resulting from the l - ~ . v ][I . . . . . . . . . . . . . . .

binding of increasing amounts of lipid and protein. The conditions of the cross-linking reaction (i.e., concentration, time, temperature, pH and ionic strength) were varied in an attempt to reveal specific binding. Conditions which promote cross-linking, however, always yield results identical to those presented (data not shown). Crosslinking with dimethyl-3,3'-dithiobispropionimidate and CuZ+/phenanthroline also yields results similar to those presented, except that no crosslinking of Pb(em)65 is observed with dimethyl3,3'-dithiobispropionimidate. Cross-linking with dithiobis(succinimidylpropionate) results in complexes so large that they do not enter the stacking gel of the first dimension. The ability to reassociate Pb(em)65 also allows an assessment of the specific component(s) Pb(em)65 may recognize through the use of corn-

petition experiments analogous to those employed to demonstrate the interaction between spectrin, ankyrin and band 3 [19,20]. Thus far, purified spectrin, glycophorin A and band 4.1 and antibodies against spectrin, ankyrin, band 4.2, and glycophorin A (generously provided by various members of Dr. V.T. Marchesis laboratory) have been employed in competition type assays, and none of these reagents affect the binding of Pb(em)65 (data not shown). These preliminary results are also consistent with the premise that Pb(em)65 does not recognize any particular membrane component, but the erythrocyte membrane as a whole. Discussion

Two acidic phosphoproteins of P. berghei origin are associated with the host membrane of infected erythrocytes. The nature of the association between the erythrocyte membrane and these two proteins was investigated in this study. Pb(em)46 is always found as a membrane-associated protein and never in a soluble form under non-denaturing conditions. In contrast, Pb(em)65 is found both as a soluble protein associated with the host erythrocyte cytosol and as an insoluble protein associated with the host erythrocyte membrane. The origin of the host cytosolic form of Pb(em)65 is not absolutely clear. One possibility is that the soluble form represents Pb(em)65 being extracted from the erythrocyte membrane during lysis. Another possibility is that Pb(em)65 exists as both a membrane and a host cytoplasmic protein and that the two forms are in equilibrium. The data presented here indicate that Pb(em)65 both associates with and dissociates from the erythrocyte membrane in vitro. Furthermore, a 93-kDa P. chabaudi protein, which may be equivalent to Pb(em)65 (see below), is associated with the host erythrocyte cytoplasm in a free form and not associated with any structures, as determined by immunoelectron microscopy [21]. Together, these results suggest that Pb(em)65 exists both as a soluble and a membrane associated protein in vivo and that the soluble form of Pb(em)65 is not an artefact of lysis. The ability of the P. berghei proteins to reassociate with IOV derived from erythrocyte mem-

131

branes appears to be specific and reflec6ve of the in vivo situation. The binding is not simply due to electrostatic interactions, since binding is not affected by the concentration of KC1 and, likewise, the binding is most probably not due to hydrophobic interactions, since the proteins do not bind to vesicles prepared by extraction with NaOH or to etythrocyte ghosts. In addition, the binding is specific for the cytoplasmic face of the erythrocyte membrane. Also, interestingly, Pb(em)65 exists in both a soluble and a membrane-associated form and exhibits a similar partitioning in the reassociation reaction, whereas the related protein Pb(em)46 is found exclusively associated with the membranes, both in the reassociation reaction and during membrane isolation. Thus, the in vitro binding properties of these proteins appear to reflect a biological activity. An interaction of Pb(em)65 with the lipid moieties of the erythrocyte membrane is implied by the reversible cross-linking experiments and from the binding of Pb(em)65 to IOV pretreated with proteases. Clearly, experiments analyzing a direct interaction between lipids or liposomes and Pb(em)65 need to be carried out before concluding that Pb(em)65 binds to lipids. On the other hand, some dependence on erythrocyte membrane proteins is indicated by the decreased binding of Pb(em)65 after extensive proteolysis of the IOV and the partial inhibition of Pb(em)65 binding resulting from extraction of the IOV at pH 11.0. This may also explain the two states of membrane association, in that interactions with erythrocyte lipids represent predominantly one state and interactions with erythrocytic membrane proteins predominantly the other. It was previously speculated that the acidic phosphoproteins interact with and affect the submembrane cytoskeleton of the erythrocyte [6]. Interestingly, other erythrocyte membrane skeletal proteins interact with both Kpids and proteins [22]. Our results support the concept that these malarial parasite proteins interact with the erythrocyte cytoskeleton. Indeed, both Pb(em)46 and Pb(em)65 can be cross-linked to high-molecular-mass complexes consisting primarily of the erythrocyte membrane skeletal elements. However, it is not clear whether the P. berghei proteins interact with distinct erythrocyte mcmbrane

components. Thus far, the evidence suggests that these malarial proteins, especially Pb(em)65, recognize the erythrocyte membrane as a unit rather than recognizing specific components of the erythrocyte membrane. The binding of Pb(em)46, however, does appear to require ankyrin, bands 4.1, 4.2 and/or 4.9. Similar acidic phosphoproteins are also associated with the host erythrocyte membrane of P. chabaudi-infected erythrocytes [5,6]. The 93-kDa protein Pc(em)93 has been analyzed by methods similar to those employed in this study. Like Pb(em)65, Pc(em)93 exists in both a soluble and particulate form [21], is stable to boiling and binds to IOV (M. Wiser, unpublished data). One difference between Pc(em)93 and Pb(em)65 is that a rauch larger percentage (80-90%) of the ghostassociated Pc(em)93 is extracted at low ionic strength [21]. However, since there is 10-fold more Pc(em)93 than Pb(em)65 per infected erythrocyte, approximately the same amounts of both proteins are associated with the IOV prepared from infected erythrocytes. In summary, although Pc(em)93 and Pb(em)65 are not identical in terms of their characteristics, they are more similar to each other than either is to Pb(em)46, suggesting that the two proteins may be functionally equivalent. Since Pb(em)65 is phosphorylated by spectrin kinase [23], it is conceivable that phosphorylation regulates the association/dissociation of Pb(em)65 with IOV. However, inclusion of Mg2+ and ATP under phosphorylating conditions had no effect on the binding of Pb(em)65 to IOV. Furthermore, addition of spectrin kinase or pretreatment of Pb(em)65 with phosphatases had no effect on the binding of Pb(em)65 to IOV (M. Wiser, unpubl.ished data). These negative results, however, do not necessarily mean that phosphorylation plays no role in the association of Pb(em)65 with the host erythrocyte membrane. It is possible that the conditions necessary to demonstrate such a role for phosphorylation have not yet been realized. It should also be pointed out that roles for the phosphorylation of spectrin or any other erythrocyte membrane protein have yet to be clearly defined [9]. It was suggested previously that these acidic phosphoproteins may be Ca2+-binding proteins

132

[6]. However, neither Ca 2÷ nor EGTA had any effect on the reassociation of Pb(em)65 with IOV (M.F. Wiser, unpublished data). In a manner analogous to that discussed above for the phosphorylation, these negative results are not definitive as to possible roles for Ca 2÷ in the association of Pb(em)65 with the host erythrocyte membrane. In light of the possible simultaneous interaction of Pb(em)65 with lipid and membrane skeletal proteins, attention should be drawn to the extensive family of Ca2+-regulated phospholipid binding proteins which also interact with F-actin [24]. The data presented here do not shed any light on the mechanism(s) by which these proteins are transported from the parasite to the host erythrocyte membrane. However, the in vitro binding results are consistent with the concept that the parasite secretes soluble proteins into the host erythrocyte cytosol, which then recognize and bind to the cytoplasmic face of the erythrocyte membrane. The transport of Pc(era)93 via a soluble intermediate has recently been suggested [21]. Other malarial proteins have been speculated to be transported via ill-defined carrier mechanisms [25-29]. The malarial parasite may utilize multiple mechanisms of 'extraparasite' transport. Furthermore, the early synthesis and transport of these acidic phosphoproteins to the ~r,,,~, . . . . . . . . m ~ , . u ~k._ ,.~juax,,,.,yt~ a n ~ ^ r,~, t u b compared to the later appearance of the putative carriers [25-29], suggests a stage specificity in the extraparasite transport mechanism. Thus far, the role that these acidic phosphoproteins play in malarial physiology is unknown. However, known Plasmodium-induced alterations of the erythrocyte membrane provide a basis for speculation about possible roles and in particular the involvement of the membrane skeleton. For example, the erythrocyte membrane exhibits an increase in the phospholipid flip-rate after P. knowlesi infection [30]. Interestingly, this alteration is first manifested during the ring stage and persists throughout schizogony. The acidic phosphoproteins described herein are likewise synthesized during the ring stage and are rapidly transported to the erythrocyte membrane, where they persist throughout schizogony [6]. Since phospholipid asymmetry is controlled in part by

the membrane skeleton [31,32], it is possible that malarial proteins enhance the transbilayer motility of lipids via perturbation of the erythrocyte membrane skeleton. Similarly, an increase in the fluidity of the erythrocyte membrane is associated with malaria infection [33-36]. Although this increased fluidity has been attributed to a decrease in the ratio of cholesterol to phospholipid [36], the possibility that this increase in disorder results at least in part from a perturbation of the membrane skeleton cannot be ruled out. New permeability pathways appear on the erythrocyte membrane during the late ring stage and increase throughout schizogony [37]. Biophysical data suggest that these new permeability pathways may result from the improper packing of lipids with parasite proteins inserted into the host erythrocyte membrane. However, it is feasible that alteration of the cytoskeleton could affect membrane topography in such a way to increase permeability. Alternatively, the early synthesis of the acidic phosphoproteins may indicate that they are involved in the targeting of other malarial proteins, such as integral membrane proteins, to the erythrocyte membrane and thereby influence permeability indirectly. In other words, these early proteins may in general serve as a 'scaffold' for ensuing host membrane changes. Likewise, the early synthesis of these proteins during erythrocytic schizogony may imply that these proteins are involved in the assembly of supramolecular complexes on the host erythrocyte membrane and/or within the host cytoplasm; for example, the putative extraparasite carriers [25-29] discussed above. In a like manner, the ability of Pb(em)65 to associate with and dissociate from the erythrocyte membrane is reminiscent of clathrin assembly and disassembly [38]. Interestingly, P. falciparum has been reported to synthesize its own transferrin receptor [39,40], which is likely a needed replacement of the transferrin receptor that is lost as retieulocytes mature into erythrocytes [41,42]. The parasite may also need to synthesize components which carry out or participate in receptoi'-mediated endocytosis. Although assembled clathrin is found within erythroeytes [43], the equilibrium lies toward the disassembled state, possibly due to the relatively higher concentration of the uncoating ATPase [44].

133

It is clear from the above discussion that the host-parasite interaction at the level of the erythrocyte membrane is quite complex. The ability to reconstitute malarial proteins with the erythrocyte membrane is clearly a necessary step in determining the function of these proteins and in elucidating the molecular mechanisms by which the malaria parasite alters the erythrocyte membrane. Hopefully, this approach will allow a correlation between functional alterations of the erythrocyte membrane and specific malarial proteins. Furthermore, these acidic phosphoproteins, and possibly other malarial proteins, may also assist in delineating structure-function relationships of the erythrocyte membrane. If indeed

these proteins do interact with the submembrane cytoskeleton, they may be useful as probes for investigating the role of the cytoskeleton in membrane physiology.

Acknowledgements We would like to thank Drs. B.J. Bormann and William Horne for their help and advice during the initial phases of this work. This research was supported in part by grants from the John D. and Catherine T. MacArthur Foundation and from National Institute of Allergy and Infectious Diseases Grant # AI-21862.

References 1 Hadley, T.J., Kiotz, F.W. and Miller, L.H. (1986) Invasion of erythrocytes by malaria parasites: a cellular and molecular overview. Annu. Rev. Microbiol. 40, 451-477. 2 Howard, R.J. (1982) Alterations in the surface membrane of red blood cells during malaria. Immunol. Rev. 61, 67-107. 3 Sherman, I.W. (1985) Membrane structure and function of malaria parasites and the infected erythrocyte. Parasitology 91,609-645. 4 Wiser, M.F., Wood, P.A., Eaton, J.W. and Sheppard, J.R. (1983) Membrane associated phosph~,proteins in Plasmodium berghei infected murine red cells. J. Cell Biol. 97, 196-201. 5 Wiser, M.F. (1987) Phosphoproteins associated with the host erythrocyte membrane during Plasmodium chabaudi infection. In: Host-Parasite Cellular and Molecular Interactions in Protozoal Infections (Chang, K.-P. and Snary, D., Eds.), NATO ASI Series, pp. 315-320, Springer-Verlag, Heidelberg. 6 Wiser, M.F., Leible, M.B. and Plitt, B. (1988) Acidic phosphoproteins associated with the host erythrocyte membrane of erythrocytes infected with Plasmodium berghei and P. chabaudi. Mol. Biochem. Parasitol. 27, 11-22. 7 Leech, J.H., Barnwell, J.W., Miller, L.H. ant~ Howard, R.J. (1984) Plasmodium falciparum malaria: association of knobs on the surface of infected erythrocytes with a histidine-ricb protein and the erythrocyte skeleton. J. Cell Biol. 98, 1256--1264. 8 Taylor, D.W,, Parra, M., Chapman, G.B., Stearns, M.E., Rener, J., Aikawa, M., Uni, S., Aley, S.B., Panton, L.J. and Howard, R.J. (1987) Localization of Plasmodium falciparum histidine-rich protein 1 in the erythrocyte skeleton under knobs. Mol. Biochem. Parasitol. 25, 165-174. 9 Bennett, V. (1985) The membrane skeleton of human erythrocytes and its implications for more complex cells. Annu. Rev. Biochem. 54, 273--304.

10 Marchesi, V.T. (1985) Stabilizing infrastructure of cell membranes. Annu. Rev. Cell Biol. 1,531-561. 11 Wiser, M.F. and Schweiger, H.-G. (1985) Cytosolic protein kinase activity associated with the maturation of the malaria parasite Plasmodium berghei. Mol. Biochem. Parasitol. 17, 179-189. 12 Trager, W. (1971) Malaria parasites (Plasmodium lophurae) developing extracellularly in vitro: incorporation of labeled precursors. J. Protozool. 18, 392-399. 13 Bennett, V. and Branton, D. (1977) Selective association of spectrin with the cytoplasmic surface of human erythrocyte plasma membranes. J. Biol. Chem. 252, 2~53--2763. 14 Ancelin, M.L. and Vail, H.J. (!987) Choline kinase activity in Plasmodium infected erythrocytes: characterization and utilization as a parasite specific marker in maia~t~ fractionation studies. Biochim. Biophys. Acta 875, 52-58. 15 Steck, T.L. and Yu, I. (1973) Selective solubilization of proteins from red blood cell membranes by protein perturbants. J. Supramol. Struct. 1,220-232. 16 Scatchard, G. (1949) The attraction of proteins for small molecules and ions. Ann. NY Acad. Sci. 51,661)-672. i7 Korsgren, C. and Cohen, C.M. (1986) Purification and properties of human erythrocyte band 4.2. J. Biol. Chem. 261, 5536-5543. 18 Peters, K. and Richards, F.M. (1977) Chemical crosslinking: reagents and problems in studies of membrane structure. Annu. Rev. Biochem. 46, 523--551. 19 Bew'ett, V. and Stenbuck, P.J. (1979) Identification and partial purification of ankyrin, the high-affinity membrane attachment site for human erythrocyte spectrin. J. Biol. Chem. 254, 2533--2541. 20 Bennett, V. and Stenbuck, P.J. (1980) Association between ankyrin and cytoplasmic domain of Band 3 isolated from the human erythrocyte membrane. J. Biol. Chem. 255, 6424-6432. 21 Wundeflich, F., Helwig, M., Schillinger,G., Speth, V. and Wiser, M.F. (1988) Expression of the parasite protein Pc90

134 in plasma membranes of erythrocytes infected with Plasmodium chabaudi. Eur. J. Cell Biol. 47, 157-164. 22 Haest, C.W.M. (1982) Interactions between membrane skeleton proteins and the intrinsic domain of the erythrocyte membrane. Biochim. Biophys. Acta 694, 331-352. 23 Wiser, M.F. (1988) Phosphorylation of Plasmodium berghei erythrocyte membrane associated proteins by the spectrin kinase. Mol. Cell. Biochem. 84, 51-57. 24 Geisow, M.J., Walker, J.H., Boustead, C. and Taylor, W. (1987) Annexins - new family of Ca2+-regulated phospholipid binding protein. Biosci. Rep. 7, 289-298. 25 Aikawa, M., Uni, Y., Andrutis, A.T. and Howard, R.J. (1984) Membrane-associated electron-dense material of the asexual stages of Plasmodium falciparum: evidence for movement from the intraeellular parasite to the erythrocyte membrane. Am. J. Trop. Med. Hyg. 35, 30-36. 26 Howard, R.J., Lyon, J.A., Uni, S., Saul, A.J., Aley, S.B., Klotz, F., Panton, L.J., Sherwood, J.A., Marsh, K., Aikawa, M. and Rock, E.P. (1987) Transport of an Mr 300000 Plasmodium falciparum protein (pfEMP 2) from the intraerythrocytie asexual parasite to the cytoplasmic face of the host cell membrane. J. Cell Biol. 104, 1269-1280. 27 Howard, R.J., Uni, S., Aikawa, M., Aley, S.B., Leech, J.H., Lew, A.M., Wellems, T.E., Rener, J. and Taylor, D.W. (1986) Secretion of a malarial histidine-ricn protein (PfHRP II) from Plasmodium falciparum-infected erythrocytes. J. Cell Biol. 103, 1269-1277. 28 Hui, G.S.N. and Siddiqui, W.A. (1988) Characterization of a Plasraodium falciparura polypeptide associated with membrane vesicles in the infected erythrocytes. Mol. Biochem. Parasito!• 29, 283-293. 29 Kara, U.A.K., Stenzel, D.J., Ingrain, L.T. and Kidson, C. (1988) The parasitophorous vacuole membrane o[ Plasmodium falciparum: demonstration of vesicle formation using an immunoprobe ~:'-- J. ,-,~1, n:_, A~ , ,-, 30 Beaumelle, B.D., Vial, H.J. and Bienvenue, A. (1988) Enhanced transbilayer mobility of phospholipids in malaria-infected monkey erythrocytes: a spin label study. J. Cell. Physiol. 135, 94--100. 31 Haest, C.W.M., Plasa, G., Kamp, D. and Deuticke, B. (1978) Spectrin as a stabilizer of the phospholipid asymmetry in the human erythrocyte membrane. Biochim. Biophys. Acta 509, 21-32. 32 WiUiamson, P., Bateman, J., Kozarsky, K., Mattacks, K., Hermanowicz, N., Choe, H.R. and Schlegel, R.A. (1982) Involvement of spectrin in the maintenance of phasestate asymmetry in the erythrocyte membrane. Cell 30, 725--733. • l t . d l J ,l .

It.~l~ll

uIUI.

"lu t

PJL

I.

33 Howard, R.J. and Sawyer, W. (1980) Changes in the membrane microviscosity of mouse red blood cells infected with Plasmodium berghei detected using n-(9-anthroyloxy) fatty acid fluorescent probes. Parasitology 80, 331-342. 34 Allred, D., Sterling, C. and Morse, P. (1983) Increased fluidity of Plasmodium berghei-infected mouse red blood cell membranes detected by electron spin resonance spectroscopy. Mol. Biochem. Parasitol. 7, 27-39. 35 Sherman, I.W. and Greenan, J.R.T. (1984) Altered red cell membrane fluidity during schizogonic development of malarial parasites (Plasmodium falciparum and P. lophurae). Trans. R. Soc. Trop. Med. Hyg. 78, 641-644. 36 Butler, K.W., Deslauries, R. and Smith, I.C.P. (1984) Plasmodium berghei: electron spin resonance and lipid analysis of infected mouse erythrocyte membranes. Exp. Parasitol. 57, 178--184. 37 Ginsburg, H. and Stein, W.D. (1987) New permeability pathways induced by the malarial parasite in the membrane of its host erythrocyte: potential routes for targeting drugs into infected cells. Biosci. Rep. 7, 455--463. 38 Pearse, B.M.F. and Crowther, R.A. (1987) Structure and assembly of coated vesicles. Annu. Rev. Biophys. Chem. 16, 49-68. 39 Haldar, K., Henderson, C.L. and Cross, G.A.M. (1986) Identification of the parasite transferrin receptor of Plasmodium falciparum-infected erythrocytes and its acylation via 1,2-diacyl-sn-glycerol. Proc. Natl. Acad. Sci. USA 83, 8565--8569. 40 Rodriguez, M.H. and Jungery, M. (1986) A protein of Plasmodium falciparum infected erythrocytes functions as a transferrin receptor. Nature 324, 388--391. 41 Nunez, M.T., Glass, J., Fischer, S., Lavidier, L.M., Lenk, E.M. and Robinson, S.H. (1977) Transferrin receptors in developing murine erythroid cells. Br. J. Haematol. 36, 519-526. 42 Davis, J.Q., Dansereau, D., Johnstone, R.M. and Bennett, V. (1986) Selective externalization of an ATP-binding protein structurally related to the clathrin-uncoating ATPase/heat shock protein in vesicles containing te~minal transferrin receptors during reticulocyte maturation• J, Biol. Chem. 261, 15368-15371. 43 Bar-Zvi, D., Levin, A.E. and Branton, D. (1987) Assembled clathrin in erythrocytes. J. Biol. Chem. 262, 17719-17723. 44 Davis, J.Q. and Bennett, V. (1985) Human erythrocyte clathrin and dathrin-uncoating protein. J. Biol. Chem. 260, 14850-14856.