Membrane-bound antiviral antibodies as receptors for Sendai virions in receptor-depleted erythrocytes

VIROLOGY 138, 185-197 (1984)

Membrane-Bound Antiviral Antibodies as Receptors for Sendai Virions in Receptor-Depleted Erythrocytes

OFER NUSSBAUM, NEHAMA ZAKAI, AND ABRAHAM LOYTER’

Department of Biological Chemistry, Institute of Life Sciences, The Hebrew University of Jerusalem, 91904 Jerusalem, Israel

Received February 2, 1984; accepted June 20, 19%;

Anti-Sendai virus antibodies were covalently coupled to neuraminidase-treated human erythrocytes by the use of the bifunctional crosslinking reagents, N-succinimidyl-3-(Z- pyridyldithio)propionate or succinimidyl-4-(pmaleimidophenyl)butyrate. Neuraminidase- treated erythrocytes bearing antibodies were able to bind Sendai virus particles, while treated erythrocytes lacking the antibodies failed to do so. Virus particles attached to erythrocyte membranes via the antibodies were able to cause hemolysis (virus-cell fusion) and promoted cell-cell fusion. Similar results were obtained when the antibodies were coupled to cat erythrocytes which lack receptors for Sendai virus particles. Reconstituted Sendai virus envelopes, similar to intact virus particles, were able to hemolyze and to induce fusion of neuraminidase-treated antibody-bearing erythrocytes. However, reconstituted envelopes containing inactive HN (hemagglutinin-neuraminidase) but active F (fusion) glycoproteins, despite attachment to antibody-bearing erythrocytes, failed to hemolyze or to induce cell-to-cell fusion. Fusion could be restored by insertion of an active HN glycoprotein into the membranes of the reconstituted envelopes. These results suggest that the HN glycoprotein, besides being the viral attachment protein, also participates in the membrane fusion process. :Cs 1984 Academic Press Inc

INTRODUCTION

Binding of viruses to specific cell surface molecules is the first step in the infection of mammalian cells by animal viruses (Choppin and Scheid, 1980; Dimmock, 1982). Proteins, glycolipids (Wu et al, 1980; Dimmock, 1982), and recently also membrane phospholipids (Schlegel et ah, 1983), have been suggested to serve as receptors for animal viruses. However, it is not yet clear whether the membrane receptors mediate only the binding of the virus particles to the appropriate cells or also play a role in the penetration and infection processes (Ozawa et al., 1979). Enveloped viruses belonging to the myxo- virus or paramyxovirus groups attach to sialic acid residues of membrane glycoproteins and glycolipids (Rott and Klenk, 1977; Choppin and Scheid, 1980; Dimmock,

1 Author to whom requests for reprints should be addressed.

1982). A virus-cell fusion step then intro- duces the viral nucleocapsids directly into the cell cytoplasm. Cells from which sialic acid residues are removed by treatment with neuraminidase, neither bind virus particles nor are they infected by them (Dimmock, 1982; Markwell et ah, 1981; Rott and Klenk, 1977). Recently, it has been shown that neuraminidase-treated cells can be infected by Sendai virus particles, if specific sialoglycolipids are in- serted into their membranes before the addition of the virus particles (Markwell and Paulson, 1980; Markwell et aZ., 1981). These experiments suggest that receptors for enveloped viruses can be implanted into membranes of living cells. However, in addition to sialic acid-containing molecules, other molecules that are capable of bringing Sendai virus into close contact with cell plasma membranes may serve as functional receptors. Previous experiments have shown that plant lectins, such as concanavalin A (Yamamoto and Inoue,

185 0042-6822/84 $3.00 Copyright B 1984 by Academic Press, Inc. All rights of reproduction in any form reserved.

186 NUSSBAUM, ZAKAI, AND LOYTER

1978; Yamamoto et al., 1979), can mediate binding and fusion of Sendai virus particles to horse and cat erythrocytes which lack the appropriate sialoglycoproteins or sialoglycolipids. Similarly, infection of Chinese hamster ovary cells by Newcastle disease virus (NDV) can be stimulated by the addition of phytohemagglutinin (Polos and Gallager, 1982). These observations clearly indicate that penetration of Sendai virus particles into animal cells does not necessarily require specific membrane components.

The results of the present work support the above view. It is shown here that antibodies to Sendai virions (SV-Ab) can be used as highly efficient receptors for those virions, if they are first chemically coupled to virus receptor-depleted cells. Our results show that neuraminidase- treated human erythrocytes bearing SV- Ab are not only able to bind to and be agglutinated by virions but are also hemolyzed and fused by them. Furthermore, we were able to study the involvement of the virus-binding protein, the hemagglutinin/neuraminidase (HN) glycoprotein (Choppin and Scheid, 1980; Markwell et ah, 1981; Rott and Klenk, 1977) in the fusion and penetration processes, because the binding of Sendai virus particles was mediated by SV-Ab and not by the viral HN glycoprotein.

MATERIALS AND METHODS

Cells. Human blood was obtained from volunteers by venipuncture. The cells were washed with Solution A (160 mM NaCl, 20 mM Tricine, pH 7.4) as previously described (Peretz et al., 1974), and suspended in Solution A at a concentration of 40% (v/v). Unless otherwise stated, the cells were used for up to 5 days after collection.

Cat blood was collected from slaugh- tered cats’ necks, and washed and suspended in Solution A as described above for human erythrocytes.

Virus. Sendai virus was isolated from the allantoic fluid of fertilized chicken eggs, and its hemagglutination titer was determined (Peretz et aL, 1974).

Antivirus antibodies. Rabbit SV-Ab were

obtained after injection of reconstituted viral envelopes (Volsky et ab, 1979).

Desialixation of human erythrocytes by Vibrio cholera neuraminidase. Sialic acid residues were removed from erythrocytes essentially as described before (Lalazar et al, 1977).

Coupling of SV-Ab (or human IgG) to desialized erythrocytes. IgG molecules (either SV-Ab or human IgG) were covalently coupled to desialized erythrocytes by a modification of the method of Godfrey et al. (1981). IgG molecules were first coupled to N-succinimidyl-3-(2-pyridyldithio)proprionate (SPDP) or to succinimidyl-4-( pmaleimidophenyl)butyrate (SMPB) at a ratio of 1gG:SPDP or 1gG:SMPB of 1:5 mol/mol, as previously described (Godfrey et al., 1981; Martin et al., 1982). Free SPDP or SMPB were removed by introducing 0.5 ml of the reaction mixture into a column (5-ml syringes) containing 5 ml of swollen Sephadex G-25 (fine) beads. After loading, the column was centrifuged (500 g for 5 min), yielding a mixture of uncoupled and coupled IgG molecules (IgG-PDP or IgG-MPB) at a concentration of 9 mg/ml.

For coupling with human or cat erythrocytes the above preparations of IgG- PDP or IgG-MPB were incubated with DTT-reduced human or cat erythrocytes as follows: erythrocytes (40%, v/v) in So- lution A (adjusted to pH 8.2) were incubated with 5 rnM of dithiothreitol (DTT) for 30 min at 37”. At the end of the incubation period, the cells were washed twice with Solution A containing 1 mg/ ml of BSA, followed by three washes with only Solution A. The final pellet was suspended in Solution A to give 40% (v/v). A volume of 1 ml of the DTT-reduced erythrocytes was then incubated with 2 mg of either IgG-PDP or IgG-MPB for 60 min at 37”, after which the erythrocytes were washed as described above. Quantitative determination of the amount of IgG which was chemically coupled to the erythrocytes was done by the use of lz51-IgG.

Preparation of reconstituted Sendai virus envelopes and membrane vesicles containing the viral F and HN glycoproteins. Reconstituted virus envelopes were prepared essentially as previously described

ANTIVIRAL ANTIBODIES AS RECEPTORS FOR SENDAI VIRIONS 18’7

(Volsky and Loyter, 1978; Vainstein et al., 1984). After solubilization of intact Sendai virus particles with Triton X-100 (10 mg virus protein in 1 ml of 2% Triton X-100, w/w), the detergent was removed by direct addition of SM-2 Bio-beads (Vainstein et al, 1984), resulting in the formation of fusogenic vesicles which contained the two viral envelope glycoproteins (F and HN). The reconstituted virus envelopes formed were collected by centrifugation (100,000 g, 60 min) and then suspended in Solution A to give 1 mg protein/ml.

Membrane vesicles containing either the HN or F glycoproteins were prepared by a modification of a previously described method (Nakanishi et al, 1982; Fukami et al., 1980). For preparation of F-containing vesicles, the HN glycoprotein was removed from a volume of 1 ml of the above detergent solution of the viral glycoproteins as follows: The pH of the solution was adjusted to 5.3 by dialyzing it against a buffer (0.5 liter) containing sodium acetate (10 m&f), pH 5.3, and 0.25% Triton X-100. After 120 min at 4”, a volume of up to 300 ~1 of the turbid dialyzate was loaded on 5 ml of dry CM-Sepharose Cl- 6B beads. The Sepharose beads were washed and equilibrated with a buffer containing 10 mM sodium acetate, pH 6.0, and 2% Triton X-100, and then packed by centrifugation in a 5-ml syringe. After loading, the column was centrifuged (500 g, 3 min), and the eluent obtained contained only the F glycoproteins, as revealed by gel electrophoresis analysis. De- tergent was removed from the eluate either by dialysis in Spectraphor tubing (Volsky and Loyter, 1978) or by direct addition of SM-2 Bio-beads (Vainstein et al, 1984). The membrane vesicles obtained were centrifuged (100,000 g, 60 min), and suspended in Solution A.

Membrane vesicles containing only the HN glycoproteins were prepared as follows: 1 ml of the detergent solution of the viral glycoproteins was dialyzed against a buffer containing 10 mM sodium phosphate, pH 6.0, and 0.25% Triton X-100. After 120 min at 4’, a volume of up to 300 ~1 of the dialyzate was loaded on 5 ml of dry DEAE-52-cellulose (5-ml syringe) (Fukami et ah, 1980). The cellulose beads

were washed and equilibrated with a buffer containing 10 mM sodium phosphate, pH 6.0, and 2% Triton X-100. All other steps of centrifugation of the columns, removal of detergent from the eluents, and formation of reconstituted vesicles were as described above for preparation of F vesicles. Gel electrophoresis analysis revealed that the eluent contained only the HN glycoprotein. Both the CM- Sepharose and DEAE-cellulose columns can be reused up to five times on the same day.

Hybrid vesicles containing both the F and HN glycoproteins were prepared by mixing detergent solutions of the F and HN glycoproteins in the appropriate weight ratios. Detergent was removed from the mixture as described above.

Human IgG (Fraction II) was obtained from Sigma. Iodination of the SV-Ab and human IgG were performed by chloramine T and Nalz51 (Lefkovitz et ah, 1970). Pro- tein was determined by the method of Lowry et al. (1951), using bovine serum albumin as standard. N-Succinimidyl-3- (2-pyridyldithio)propionate (SPDP) and succinimidyl-4-(p-maleimido-phenyl)butyrate (SMPB) were purchased from Pierce Chemical Company. Vibrio cholera neuraminidase was from BDH Biochemi- cals. Human IgG was obtained from Sigma, and anti-human IgG antibody was from Miles-Yedda, Rehovot, Israel.

RESULTS

Use of SV-Ab to Mediate Binding of Virus Particles to Desialixed Human Eryth- rocytes

Table 1 shows that under the conditions used in the present work, up to about 4 X lo4 molecules of IgG can be covalently attached per human erythrocyte. This was achieved by the use of SPDP or SMPB as bifunctional crosslinking reagents. SMBP (or SPDP) are first covalently bound to NHe-containing molecules such as IgG (Godfrey et ak, 1981; Martin et ah, 1982), and the resulting IgG-MPB (or IgG-PDP) complex can then be coupled to membrane exposed sulfhydryl groups which are formed after reduction of human eryth-

188 NIJSSBAUM, ZAKAI, AND LOYTER

TABLE 1

COVALENT BINDING OF IgG MOLECULES TO INTACT HUMAN ERYTHROCYTES

System

Erythroeyte-bound IgG

% of total Molecules

added Erythrocyte

Agglutination by anti-human IgG

antibodies

Erythrocyte + ‘=I-IgG 4.2

Erythrocyte + ‘251-IgG-PDP <0.2

Erytbrocyte-SH + “‘1-1gG <0.2

Erythrocyte-SH + ‘%I-IgG-PDP 1.5”

<4 x lo3

<4 x lo3

<4 x 10”

3 x 1O’O

-

-

+ (>0.3 pg anti-IgG antibody/sample)

Erythrocyte-SH + “‘1-IgG-MPB 2.5 4 x 10’ + (>O.l Kg anti-IgG antibody/sample)

Erythrocyte-PDP-‘%I-IgG + DTT 0.4 8 X lo3 -

Erythrocyte-MPB-iz51-IgG + DTT 2.5 4 x lo4 +

‘During incubation of the human erythrocytes with either IgG-PDP or IgG-MPB a certain degree of hemolysis (5-10s of total) occurs. Since the binding of IgG-PDP (but not of IgG-MPB) to the erythrocytes is reversible upon reduction, it is conceivable that some of the bound IgG were removed by reduction with glutathione which is released from the erythrocytes during hemolysis. In some experiments the amount of ‘““I-IgG-PDP which was covalently attached to DTT-reduced human erythrocytes, was low and did not exceed 8 X lo3 molecules/cell (not shown).

““I-IgG-PDP or ‘“‘I-IgG-MPB were obtained and incubated with untreated or DTT-reduced human erythrocytes (erythrocyte-SH) as described under Materials and Methods. Similar conditions were used for incubation of ““I-IgG with human erythrocytes (2 mg of “‘1-IgG/ml of 40R, v/v, human erythrocytes). Erythrocytes bearing coupled ‘“‘I-IgG were reduced with DTT as follows: After washing with Solution A (adjusted to pH 8.2), the pellet obtained was dissolved in the above buffer which contained 20 mM DTT, to give 40% (v/v) of human erythrocytes. Following 60 min of incubation at 37’, the erythrocytes were washed five times with Solution A. For induction of agglutination, erythrocytes bearing IgG (100 ~1, 2.5%) v/v, in Solution A) were incubated with several dilutions of anti-human IgG-Ab. After 60 min at 37”, agglutination was followed through observations in a phase microscope.

rocytes with DTT (Table 1). The results in Table 1 show that less than 4 X lo3 molecules of lz51-IgG were attached per nonreduced human erythrocytes and that binding of IgG to the reduced erythrocytes was dependent on the formation of IgG- MPB or IgG-PDP complexes. SPDP was less effective in mediating binding of IgG to human erythrocytes than SMPB (Table 1). This is probably due to the fact that the dithio bond formed between IgG-PDP and the erythrocyte-SH groups is reversible (Table 1, see also Fig. 2). The results in Table 1 also show that anti-human IgG agglutinated erythrocytes bearing human IgG molecules, indicating that the anti- genie properties of IgG molecules were

preserved during the chemical manipula- tions required for the coupling process and are exposed on the outer surface of the erythrocyte membrane. The latter can also be inferred from the results showing that reduction with DTT of IgG-bearing erythrocytes can remove about 75-80% of the IgG when SPDP but not SMPB was used as a crosslinking reagent (Table 1).

The number of SV-Ab bound to each erythrocyte was highly dependent upon the extent to which the erythrocytes were reduced with DTT (Fig. 1A). Our experiments show that the covalently bound SV-Ab can replace the virus’ natural receptors (sialic acid residue of membrane glycoproteins and glycolipids) in mediat-

ANTIVIRAL ANTIBODIES AS RECEPTORS FOR SENDAI VIRIONS 189

3, DTT (mM)

I , I 1

0 I 2 Sv -Ab. coupled/cell tJlo4 )

4

FIG. 1. Coupling of increasing amounts of SV-Ab to desialized human erythrocytes and their ability to mediate binding to Sendai virus particles. (A) Effect of DTT on coupling of SV-Ab: Desialized human erythrocytes (40%, v/v) were incubated with increasing concentrations of DTT for 30 min at 37’. At the end of the incubation period, the erythrocytes were washed and incubated with ‘251-SV-Ab-MPB (sp act 3 X lo5 cpm/mg). The number of IgG molecules bound per cell was calculated according to the following equation: (pg of IgG molecules coupled X Avogadro No. (6.23 X 10”())/(106 X 150,000 (MW of IgG) X No. of cells). (B) Binding of Sendai virions to desialized human erythrocytes as a function of SV-Ab concentration (number of Ah/cell): Sendai virus particles were radioiodinated by using chloramine T and Na?, as previously described (Wolf et al, 1980), resulting in 3 X lo6 cpm/mg of viral protein. For trypsinization, Sendai virus particles (2 mg/ml of viral protein) in 150 mM of NaCl, buffered with 20 m&f phosphate buffer, pH 7.4, were incubated with 60 pg of trypsin for 60 min at 37”. Trypsinization was terminated by washing the virus suspension in Eppendorf tubes, and the pellet obtained was dissolved in Solution A to give 1 mg/ml of viral proteins. ‘=I-trypsinized Sendai virus particles (3 pg) were incubated with desialized human erythrocytes (2.5%, v/v) bearing increasing amounts of SV- Ab, in a final volume of 100 ~1 of Solution A. After 30 min of incubation at 4” and 30 min at 37”, unbound particles were separated from erythrocyte- bound virus particles by centrifugation through 1 ml of 0.3 M sucrose cushion (Volsky et al, 1979). Arrows show the amount of virus particles bound to control, untreated cells.

ing binding of virus particles to neuraminidase-treated human erythrocytes (Fig. 1B). A quantitative estimation revealed (Fig. 1B) that binding of Sendai virus particles to desialized human erythrocytes is highly dependent upon the concentration of the covalently bound antibodies (number of SV-Ah/cell). Maximum binding of virus particles was achieved when about 1 X lo4 molecules of antibodies were attached per human erythrocyte (Fig. 1B). All the binding experiments were performed using trypsinized or heat-inactivated Sendai virus particles whose hemolytic and fusogenic activities were inhibited (Shimizu and Ishida, 1975; see also Fig. 2). This permitted the perfor- mance of binding studies at 37” without the interference of processes such as cell- cell fusion and hemolysis, which are in-

0 I2 3- lz51 - \hs added (/us)

FIG. 2. Antibody-mediated binding of Sendai virus particles to neuraminidase-treated human erythrocytes: Specificity and effect of DTT. (A) Binding of trypsinized Sendai virus particles. (B) Binding of heat-inactivated Sendai virus particles. Sendai virus particles were trypsinized as described in Fig. 1. For thermal inactivation, the particles (1 mg/ml) were incubated at 56’ for 60 min, and Senadi virus was radioiodinated as previously described (Wolf et al, 1980, and in Fig. 1). Increasing amounts of I?- Sendai virus particles were incubated with human erythrocytes (2.5%, v/v), in a final volume of 100 ~1 Solution A, for 30 min at 4” and then for 30 min at 37”. A, Human erythrocytes-SV-Ab-MPB; 0, human erythrocytes-SV-Ab-PDF; 0, human erythrocytes- SV-Ab-PDP treated with 20 mM of DTT (final concentration) for 60 min at 37”; n , human erythrocytes-human-IgG-MPB. For all other experimental conditions, see Materials and Methods, Table 1, and Fig. 1.


duced at this temperature by active virions and 0.57 pg-l for antibody and sialic acid- (see Table 2). mediated binding, respectively (Fig. 3C).

The results in Fig. 2 (A and B) strengthen the view that the binding of Sendai virus particles to desialized human erythrocytes was specifically mediated by SV-Ab. Erythrocytes bearing human IgG (instead of SV-Ab) bound very few virions, the extent of which did not exceed the amount that was found to be associated either with desialized human erythrocytes (Fig. 1) or with cells from which the bound antibodies were removed by treatment with DTT (Figs. 2A, B).

Titration studies revealed that membrane-bound SV-Ab and membrane-associated sialic acid residues (the virus’ natural receptors) can maximally mediate the binding of 5.3 and 8.1 pg of Sendai virus particles, respectively per 1 X lo7 human erythrocytes (Fig. 3). The corre- sponding K, values were found to be 0.0655

Antibody-mediated binding of Sendai virus particles to desialized human erythrocytes is also evident from the micrographs in Fig. 4. Analysis of fluorescence microscopy revealed that when desialized erythrocytes bearing antibody were incubated with fluorescein-isothiocyanate (FITC)-labeled Sendai virus particles, all the cells in the population became labeled (Fig. 4a). Very little or no fluorescence could be detected in the microscopic field when desialized cells, lacking bound SV- Ab, were incubated with fluorescently labeled Sendai virus particles (not shown).

Induction of Hemolysis and Cell-Cell Fu- siwn in Desialixed Human Erythro- cytes Beam’ng SV-Ab

Table 2 and Fig. 5 show that SV-Ab can mediate functional binding of Sendai

TABLE 2

SV-Ab-MEDIATED FUSION AND HEMOLYSIS OF HUMAN ERYTHROCYTES BY INTACT SENDAI VIRUS PARTICLES

System

Agglutination

4” 37” Hemolysis

(% of total) Cell-cell

fusion

Sendai virus + erythrocyte Sendai virus + erythrocyte-SH PMSF-Sendai virus + erythrocyte-SH Sendai virus + Ne-erythrocyte-SH Sendai virus + Ne-erythrocyte-SV-Ab Sendai virus + Ne-erythrocyte-human IgG Trypsinized-Sendai virus + Ne-erythrocyte-

SV-Ab PMSF-Sendai virus + Ne-erythrocyte-SV-Ab DTT-treated Sendai virus + Ne-erythrocyte-

+ + 92 + + + 97 + + + 5 -. - - 11 -. - + 76 + - - 9

- + 10 - + 7

SV-Ab - + 11 -

Note. Sendai virus particles (10 fig viral protein) were incubated with human erythrocytes (2.5%, v/v) in a final volume of 200 ~1 of Solution A, first, for 30 min at 4”, and then for 30 min at 37”, with gentle shaking. Agglutination, hemolysis, and cell-cell fusion were estimated as previously described (Peretz et al, 1974). Sendai virus particles were treated with 7 mM (final concentration) of PMSF (PMSF-Sendai virus) as described before (Israel et al, 1983). Trypsinization of Sendai virus was performed as described in Fig. 1 (trypsinized Sendai virus). Sendai virus particles (1 mg/ml) were incubated with DTT (10 mM final concentration) for 60 min at 37” (DTT-treated Sendai virus). After washing (Eppendorf centrifuge) with Solution A, the final pellet was suspended in Solution A to give 1 mg protein/ml. Sendai virus- antibody was coupled to human erythrocytes using SMPB as crosslinking reagent. Erythrocyte-SH, human erythrocytes treated with DTT; Ne-erythrocyte, human erythrocytes treated with neuraminidase; Ne- erythrocyte-SV-Ab, desialized human erythrocytes bearing SV-Ab, Ne-erythrocyte-human IgG, desialized human erythrocytes bearing human IgG.


0 5 IO 15 20 25 0 2 4 6 8 VIW added (,ug ) Virus bound ( ,A& )

FIG. 3. Binding of ‘%I-Sendai virus particles to neuraminidase-treated human erythrocytes bearing SV-Ab: Determination of association constant. SV- Ab were coupled to desialized human erythrocytes. Each desialized human erythrocyte had 3.6 X 10’ molecules of SV-Ab. Sendai virus particles were radiolabeled and trypsinized as described in Fig. 1. Trypsinized ‘“I-Sendai virus particles at different concentrations were incubated with human erythrocytes (2.5%, v/v), in a final volume of 200 ~1, for 30 min at 4”, followed by 30 min at 37”. Unbound virus was removed and erythrocyte-associated ra- dioactivity was determined in a gamma counter, as described in Fig. 1. (A) A, Control, nontreated human erythrocytes; 0, neuraminidase-treated human erythrocytes bearing SV-Ab; n , neuraminidase- treated human erythrocytes (without SV-Ab). In order to estimate specific binding of Sendai virus particles to human erythrocytes, mediated either by the viral natural receptors (sialic acid residues) or by SV-Ab, the following was performed: The data plotted in A giving the amount of ‘=I-Sendai virus particles bound to control (A) or to neurminidase- treated human erythrocytes bearing antibody (O), were subtracted from the amount of Sendai virus particles bound to neuraminidase-treated human erythrocytes @) (unspecific adsorption). The results obtained are plotted in (B) as specific binding to untreated human erythrocytes (A) or to neuraminidase-treated human erythrocytes bearing SV-Ab (0). (C) Scatchard plot analysis of binding of ‘%I-Sendai virus particles to untreated human erythrocytes (A) or to neuraminidase-treated human erythrocytes bearing SV-Ab (0).

virus particles to desialized human erythrocytes. This can be inferred from the results showing that Sendai virus particles induce hemolysis and promote cell-cell fusion of desialized erythrocytes bearing SV-Ab. It was previously shown that virus-induced hemolysis reflects a process

of virus-cell fusion (Shimizu et al, 1976). Indeed, the electron micrographs in Fig. 4b show a virus particle in the process of fusion with a membrane of desialized human erythrocytes.

Figure 5 shows that the extent of hemolysis (similar to the extent of binding, see Fig. 1B) was dependent upon the concentration of the SV-Ab (number of SV- Ah/cell). Maximum hemolysis was obtained with cells bearing 3 X lo4 Ah/cell and above. Clearly, virus-induced agglutination and hemolysis were absolutely dependent upon the presence of covalently coupled SV-Ab. Desialized human erythrocytes bearing human IgG were neither agglutinated nor hemolyzed when incubated with Sendai -virus particles (Table 2).

Agglutination of sialic acid-bearing cells occurs at 4”, whereas agglutination of desialized human erythrocytes bearing SV-Ab was induced only at 37”, probably after the establishment of an antibody- antigen complex (Table 2). Moreover, ag- glutinates formed by desialized human erythrocytes did not undergo a process of disagglutination (not shown) as do agglu- tinates of control erythrocytes, due to the removal of the sialic acid residues by the virus-associated neuraminidase (Rott and Klenk, 1977).

Hemolysis was not induced by virus particles which were either treated with phenylmethyl-sulfonyl-fluoride (PMSF) or with trypsin (Table 2). These treatments are known to inactivate the virus fusion factor (Shimizu and Ishida, 19’75; Israel et ah, 1983). Treatment with DTT, which reduced the viral envelope glycoproteins (Ozawa et al., 1979), also substantially inhibited the viral hemolytic activity without affecting the ability to bind to the membrane bearing SV-Ab (Table 2).

The results in Table 3 show that reconstituted Sendai virus envelopes which contain only the two viral envelope glycoproteins, F and HN (Volsky and Loyter, 1978), behave similarly to intact virus particles when incubated with desialized human erythrocytes. Only erythrocytes bearing SV-Ab were agglutinated and hemolyzed by the reconstituted Sendai virus envelopes. Reconstituted envelopes which

NUSSBAUM, ZAKAI, AND LOYTER

b

FIG. 4. Binding of Sendai virus to neuraminidase-treated human erythrocytes bearing SV-Ab: Microscopic studies. (a) Fluorescent microscopy observations: Sendai virus were labeled by fluorescein-isothiocyanate (FITC), as previously described (Chejanovsky et aL, 1983). Incubation of FITC-labeled Sendai virus with human erythrocytes bearing SV-Ab was as described under Materials and Methods and in Table 2. Magnification X880. (b) Electron microscopy observations: All experimental conditions for preparation of electron microscopy were as described before (Peretz et al, 1974). Note a virus particle (V) fusing with a neuraminidase-treated human erythrocyte bearing SV-Ab (HE-M). Magnification X72,000.

were either heat inactivated, trypsinized, or reduced by DTT, did not induce hemolysis of desialyzed human erythrocytes bearing SV-Ab (Table 3).

The results in Fig. 6 further support the view that SV-Ab can serve as functional receptors of Sendai virus particles. Cat erythrocytes whose membranes contain low amounts of sialic acid residues (Yamamoto and Inoue, 1978), are hardly agglutinated or hemolyzed at 4 and 37”, respectively, even when incubated with relatively high concentrations of Sendai virus particles (Yamamoto and Inoue,

19’78; see also Fig. 6). However, coupling of SV-Ab to cat erythrocytes confers upon them the ability to react with Sendai virus particles (Fig. 6). A high degree of agglutination and hemolysis was promoted when Sendai virus particles were incubated with cat erythrocytes bearing SV- Ab (Fig. 6).

Involvement of the Sendai Virus HN Gly- coprotein in the Process of Virus-In- duced Hemolysis

The results in Fig. ‘7 show that SV-Ab mediated the binding of membrane vesi-


‘OOv

I-1 --.2-l - -~ j 0 / 2 3 4

SWAb coupled/cell (~10~)

FIG. 5. Virus-induced hemolysis of desialized human erythrocytes: Dependence on antibody concentration (number of SV-Ah/cell). Sendai virus particles (3 pg) were incubated with desialized human erythrocytes (2.5%, v/v) bearing increasing amounts of SV-Ab, in a final volume of 100 ~1, for 30 min at 4’ and then for 30 min at 37“. At the end of the incubation periods, hemolysis was determined at 540 nm (Peretz et al, 1974). Coupling of SV-Ab to desialized human erythrocytes was as described in Fig. 1. The arrow indicates the degree of hemolysis obtained by 3 pg Sendai virus under the same conditions as with control, untreated human erythrocytes.

cles-containing only the viral HN glycoprotein- to desialized human erythrocytes. On the other hand, SV-Ab did not support the binding of F vesicles (vesicles including only the F glycoprotein) to desialized human erythrocytes. This clearly indicates that the SV-Ab preparation con- tains mainly anti-HN antibodies and very little anti-F antibodies. Neither F nor HN vesicles induced hemolysis of either control or desialized cells bearing SV-Ab (Fig. 7). Hemolysis of desialized cells was induced, however, only by hybrid vesicles formed by coreconstitution of F and HN glycoproteins (Fig. 7), and was absolutely dependent upon the presence of SV-Ab (compare Fig. 7A with B). Inactivation of the viral HN glycoprotein, either by treatment with DTT or by heating at 56” prior to its coreconstitution with the F glycoprotein, led to the formation of nonhemo- lytic hybrid vesicles, although their binding properties were not affected (Fig. 7). Evidently, inactivation of the HN glycoprotein did not modify its antigenic char- acteristics. Interestingly, coreconstitution

of native, untreated HN together with inactive HN and active F glycoproteins (F:HN:HNsGo or F:HN:HNnTT in Fig. 7), although it did not change the binding of the hybrid vesicles, restored their hemolytic activity. Figure 8 shows that similar results were obtained when F, HN, or F-HN hybrid vesicles were incubated with cat erythrocytes bearing SV-Ab.

DISCUSSION

In the present work, the interaction between fusogenic Sendai virus particles and desialized human erythrocytes has been used as a model system to study questions related to the molecular mechanism of virus-cell fusion and infection. Incubation of Sendai virus particles with human erythrocytes causes their agglutination and promotes their hemolysis. Ag- glutination results from a process of virus-cell binding, while hemolysis reflects a process of virus-cell fusion (Rott and Klenk, 1977; Shimizu et ak, 1976). Human erythrocytes are a convenient system for investigating the molecular mechanism of virus-cell fusion, since under the experimental conditions used, most of the erythrocyte-associated virus particles fuse with the cell membrane.

The results of the present work may suggest a new way to convert cells which are resistant to virus attachment and infection, to become highly susceptible to it. This is by chemically coupling SV-Ab to the appropriate cell plasma membranes.

At first glance it may seem paradoxical that antibodies which usually are used to inhibit infection of cells by viruses, support it once they are coupled to the cell membranes. It should be emphasized, however, that soluble antibodies exert their effect forming an immune complex, resulting in precipitation of the virus particles and thus avoiding the association between the virions and the recipient cells. It also appears that the antibody preparations used in this work contain mainly antibodies against the HN glycoprotein and very few against the F protein. Therefore, after attachment to erythrocytes, these antibody preparations should


TABLE 3

THE ABILITY OF RECONSTITUTED SENDAI VIRUS ENVELOPES TO INDUCE HEMOLYSIS IN NEURAMINIDASE-

TREATED HUMAN ERYTHROCYTES BEARING SENDAI VIRUS ANTIBODIES

System

Neuraminidase-treated human erythrocytes Nontreated human

erythrocytes Lacking SV-Ab Bearing SV-Ab

Aggluti- Hemolysis Aggluti- Hemolysis Aggluti- Hemolysis nation (% of total) nation (% of total) nation (% of total)

Intact Sendai virus + 72 - 12 t 64 Reconstituted Sendai

virus envelopes + 86 - 14 + 79 Heated reconstituted

Sendai virus envelopes - 9 - 3 + 3 Trypsinized

reconstituted virus envelopes + 2 - 2 + 3

DTT-treated reconstituted virus envelopes - 3 - 3 + 5

Note. Experimental conditions were as described in Table 2 for intact Sendai virus and under Materials and Methods. A suspension of reconstituted Sendai virus envelopes (1 mg/ml) was heat inactivated or treated with trypsin, as described in Figs. 1 and 2 for intact Sendai virus particles. SV-Ab were coupled to DTT-reduced, neuraminidase-treated human erythrocytes, using SMPB as the bifunctional crosslinking reagent.

support binding of virus particles without affecting the fusion process. Similar observations were reported before when

-‘““P--- - --- -1 I

Virus added (fig 1

FIG. 6. The effect of SV-Ab on virus-induced hemolysis of cat erythrocytes. SV-Ab were coupled to cat erythrocytes as described for human erythrocytes in Table 1. A quantitative estimation revealed that 4.1 X lo4 IgG molecules were coupled to each cat erythrocyte (not shown). Cat erythrocytes, bearing SV-Ab, were hemolyzed by increasing concentrations of Sendai virus particles, as described for human erythrocytes in Fig. 5. Hemolysis (0) and agglutination (A) of cat erythrocytes. Arrows indicate the amount of hemolysis (0) and agglutination (A) obtained by incubation of untreated (lacking antibodies) cat erythrocytes with 10 ng of Sendai virus particles.

monoclonal antiviral antibodies were an- alyzed (Orvall and Grandien, 1982).

It is generally accepted that the function of the viral HN glycoprotein is only to mediate binding of the virus particles to the recipient cell receptors, while the F protein is essential for the fusion process (Choppin and Scheid, 1980; Loyter and Volsky, 1982). However, membrane vesicles containing active F but inactive HN glycoproteins are able to attach to neuraminidase-treated human erythrocytes bearing SV-Ab but can neither induce hemolysis nor promote cell-cell fusion. Insertion of an active HN glycoprotein to these vesicles restores their hemolytic and fusogenic activities, without significantly affecting their binding properties. These observations suggest an active role for the viral HN glycoprotein in the virus- cell fusion process, which is unrelated to its function as the viral binding protein.

It is tempting to divide the process of virus-cell attachment into three steps: First, the virus particles are bound to exposed cell surface sialoglycoproteins,


w :w 3:l 1. 1 1:l 2.1:1 1.1 2.1.1

FIG. ‘7. Induction of hemolysis in desialized human erythrocytes bearing SV-Ab by hybrid vesicles containing F and HN viral glycoproteins. Detergent solutions of only the viral HN glycoprotein (850 pg protein/ml in 2% Triton X-100) were either reduced by 5 mMof DTT (HNorr) for 30 min at 37” or thermally inactivated at 56” for 60 min (HN5&. DTT was removed by gel chromatography of the detergent solution through Sephadex G-25 (fine, 5-ml column). The detergent solubilized viral glycoproteins were radiolabeled by chloramine T and Na?, as described for intact Sendai virus particles (Wolf et al, 1980), giving a sp act for HN of 5 X lo6 cpm/mg and for F of 7 X lo6 cpm/mg. Hybrid vesicles containing F and HN glycoproteins within the same membrane were prepared by mixing detergent solutions of the individual glycoproteins (850 pg of viral glycoproteins in 2% Triton X-100) at the required ratio (w/w). After removal of the detergent (Vainstein et al, 1984) and centrifugation, the pellet formed was suspended in Solution A to give 0.5 mg viral protein/ml. Binding of vesicles to desialized human erythrocytes as well as determination of vesicle-induced hemolysis were performed as described for intact Sendai virus particles in Figs. 1 and 5. F, vesicles containing only the F glycoprotein; HN, vesicles containing only the HN glycoprotein; F:HN, hybrid vesicles containing both the F and HN glycoproteins at a ratio (w/w) designated in the figure. (A) Desialized human erythrocytes bearing SV-Ab. (B) Desialized human erythrocytes.

resulting in virus-induced cell-cell agglu- that antivirus antibodies substitute for tination. Second, the virus particles are the membrane sialoglycoproteins. After brought into close proximity to the mem- the above steps are completed, membrane- brane phospholipids. This step may occur associated virus particles are transferred in neuraminidase-treated cells, provided to another membrane component such as

-60

50$J

OL Elm. Virus Iv

F HN F HII f HW f H$yF. HI H

GlyCOprOkln 31 11

&&k Id&l 1 10

EdaIl 0 ln&=f HflOIl F HN Illlo,,

1 1 211 11 ,l 2 1

W:W

FIG. 8. The ability of reconstituted vesicles, containing the Sendai virus F and HN glycoproteins, to bind to and hemolyze cat erythrocytes bearing SV-Ab. All experimental conditions were as described for human erythrocytes in Fig. 1 and under Materials and Methods. The first dashed column shows the amount of hemolysis induced by intact Sendai virus particles.


ganglioside, whose presence is necessary for the fusion process and cannot be re- placed by SV-Ab. For this step, an active HN glycoprotein may be required, even when desialized human erythrocytes bearing antibodies are used. It is note- worthy that Kohn et al. (1980) have recently provided evidence for two-component receptor systems of thyrotropin con- sisting of a glycoprotein and gangliosides.

It may therefore be possible-as has already been suggested before from experiments using anti-HN monoclonal antibodies (Miura et aZ., 1982) or hybrid vesicles containing the Sendai virus F glycoprotein and influenza HN glycoprotein (Ozawa et aZ., 1979)-that the Sendai virus HN glycoprotein also plays an active role in the membrane fusion step itself.

Finally, membrane-coupled antibodies may support functional binding of other active molecules such as polypeptide hor- mones or toxins. Experiments to study this possibility are currently being con- ducted in our laboratory.

ACKNOWLEDGMENTS

This work was supported by a grant from the National Council for Research and Development, Israel, and from the GSF, Munich, Germany.

REFERENCES

CHEJANOVSKY, N., BEIGEL, M., and LOYTER, A. (1984). Attachment of Sendai virus particles to cell membranes: Dissociation of adsorbed virus particles with dithiothreitol. J. Viral. 49, 1009-1013.

CHOPPIN, P. W., and SCHEID, A. (1980). The role of viral glycoproteins in adsorption, penetration, and photogenicity of viruses. In “Reviews of Infectious Diseases,” Vol. 2, pp. 40-58.

DIMMOCK, N. J. (1982). Initial stages in infection with animal viruses. J. Gen. ViroL 59, l-22.

FUKAMI, Y., HOSAKA, Y., and YAMAMOTO, K. (1980). Separation of Sendai virus glycoproteins by CM- Sepharose column chromatography. FEBS Z.&t. 114,342-347.

GODFREY, W., DOE, B., WALLACE, E. F., BREDT, B., and WOLFSY, L. (1981). Affinity targeting of membrane vesicles to cell surfaces. Exp. Cell Res. 135, 137-145.

ISRAEL, S., GINSBERG, D., LASTER, Y., ZAKAI, N., MILNER, Y., and LOYTER, A. (1983). A possible involvement of virus associated protease in the fusion of Sendai virus envelopes with human erythrocytes. Biochim. Biophgs. Acta 723,337-346.

KOHEN, L. D., CONSIGLIO, E., DOWOLL, N. J. S.,

GROLLMAN, E. F., LEDLEY, F. D., and MORAIS, N. P. (1980). Thyrotropin receptors and gangliosides. Adv. Exp. Med BioL 125, 487-503.

LALAZAR, A., MICHAELI, D., and LOYTER, A. (1977). Restoration of the fusion capacity of human erythrocyte ghosts hy SH-blocking reagents. Exp. Cell Res. 107, 79-98.

LOWRY, 0. H., ROSEBROUGH, N. J., FARR, A. L., and RANDALL, R. J. (1951). Protein measurement with the folin phenol reagent. J. BioL Chem. 193, 265- 275.

LOYTER, A., and VOLSKY, D. J. (1982). Reconstituted Sendai virus envelopes as carriers for the introduction of biological material into animal cells. In “Cell Surface Reviews” (G. Poste and G. L. Nicol- son, eds.), Vol. 8, pp. 215-266, Elsevier, Amsterdam/ New York.

MARKWELL, M., and PAULSON, J. C. (1980). Sendai virus utilizes specific sialyloligosaccharides as host cell receptor determinants. Proc N&l. Acud. Sci. USA 77, 5693-5697.

MARKWELL, M. A., SVENNERHOLM, L., and PAULSON, J. C. (1981). Specific ganglioside function as host cell receptor for Sendai virus. Proc. NutL Acad. Sci. USA 78, 5406-5410.

MARTIN, F. J., PAPAHADJOPOULOS, D., and VAIL, W. J. (1982). Irreversible coupling of immunoglob- ulin fragments to preformed vesicles. J. Biol Chmr~

257, 286-288.

MIURA, N., UCHIDA, T., and OKADA, Y. (1982). HVJ (Sendai virus)-induced envelope fusion and cell fusion are blocked by anti-HN protein antibody that does not inhibit hemagglutination activity of HVJ. Exp. Cell Res. 141, 409-420.

NAKANISHI, M., UCHIDA, T., KIM, J., and OKADA, Y. (1982). Glycoproteins of Sendai virus (HVJ) have a critical ratio for fusion between virus envelopes and cell membranes. Exp. Cell Rex 142, 95-101.

ORVALL, C., and GRANDIEN, M. (1982). The effect of monoclonal antibodies on activity of Sendai virus. J. ZmmunoL 129, 2779-2787.

OZAWA, M., ASANO, A., and OKADA, Y. (1979). Bio- logical activities of glycoproteins of HVJ (Sendai virus) studied by reconstitution of hybrid envelope and by concanavalin A-mediated binding: A new function of HANA protein and structural require- ment for F protein hemolysis. Virology 99, 197s 202.

PERETZ, H., TOISTER, Z., LASTER, Y., and LOYTER, A. (1974). Fusion of intact human erythrocytes and erythrocyte ghosts. J. Cell BioL 63, l-11.

POLOS, P. G., and GALLAGHER, W. R. (1982). Inde- pendent effect of phytohemagglutinin on cell fusion and attachment by Newcastle disease virus. Virol- ogy 119, 268-275.

ROTH, J., LEFKOWITZ, R. J., PRICE, W., and PASTAN, I. (1970). Acth receptors in the adrenal: Specific binding of Acth-‘&I and its relation to adenylate cyclase. Proc. NatL Acad Sci. USA 65, 745-749.


ROTT, R., and KLENK, H. 0. (1977). Structure and assembly by viral envelopes. In “Cell Surface Reviews,” Vol. 2: “Virus Infection and the Cell Surface” (G. Poste and G. L. Nicolson, eds.), pp. 47-87, North Holland, Amsterdam.

SCHLEGEL, R., TRALKA, T. S., WILLIGHAM, M. C., and PASTAN, I. (1983). Inhibition of VSV binding and infectivity by phosphatidylserine: Is phosphatidylserine a VSV binding site? Cell 32, 639-646.

SHIMIZU, Y. K., SHIMIZU, K., ISHIDA, N., and HOMMA, M. (1976). On the study of Sendai virus hemolysis. II. Morphological study of envelope fusion and hemolysis. Virology 27, 48-60.

SHIMIZU, K., and ISHIDA, N. (1975). The smallest protein of Sendai virus: Its candidate function of binding nucleocapsid to envelope. Virology 67,427- 437.

VAINSTEIN, A., HERSHKOVITZ, M., ISRAEL, S., RABIN, S., and LOYTER, A. (1984). A new method for reconstitution of highly fusogenic Sendai virus envelopes. Biochim. Biophys. Acta 773, 181-188.

VOLSKY, D. J., CABANTCHIK, 2. I., BEIGEL, M., and

LOYTER, A. (1979). :Implantation of isolated human erythrocyte anion channel into plasma membranes of Friend erythroleukemic cells. Proc, NutL Acad Sci. USA 77,5453-5457.

VOLSKY, D. J., and LOYTER, A. (1978). An efficient method for assembly of fusogenic Sendai virus envelopes after solubilization of intact virions with Triton X-100. FEBS L&t. 92, 190-194.

WOLF, D., KAHANA, I., NIR, S., and LOYTER, A. (1980). The interaction between Sendai virus and cell membranes. JZr,n Cell Res. 130, 361-369.

WIJ, P. S., LEDEE~, R. W., UDEM, S., and ISAACSON, Y. A. (1980). Nature of the Sendai virus receptor: Glycoproteins versus gangliosides. J. Viral. 33, 304-310.

YAhlAhlOTO, K., and INOUE, K. (1978). Interaction of paramyxoviruses with concanavalin A-modified erythrocyte membranes. Virology 84, 203-206.

YAMAMOTO, K., INOUE, K., and SUZUKI, K. (1979). Interaction of paramyxovirus with erythrocyte membranes modified by concanavalin A. Nature (Lmdon) 250, 511513.

Membrane-bound antiviral antibodies as receptors for Sendai virions in receptor-depleted erythrocytes

Documents