Saturable attachment sites for type 3 mammalian reovirus on murine L cells and human HeLa cells

VirtueResearch, 1 (1984) 401-414 Elsevier

401

VRR 00132

Saturable attachment sites for type 3 mammalian reovirus on murine L cells and human HeLa cells * Jon R. Gentsch University of Pennsylvania,

and John

W. Hatfield

Department of Microbiology, School of Medicine/G2, U.S.A.

Philadeipha, PA 19104,

(Accepted 16 April 1984)

Summary Attachment of [35S]methionine-labelled mammalian type 3 reovirus to murine L cells and human HeLa cells was studied under equilibrium conditions. Cellular attachment sites could be completely saturated with 35S-labelled reovirus, indicating that specific attachment sites for reovirus are present on the surface of these cells. We calculated that L cells possess about 8600%105000 attachment sites per cell while HeLa cells possess about 126000-147 000 sites per cell for type 3 reovirus. Unlabelled reovirus was highly efficient in competing for attachment by 35S-labelled reovirus to the saturable attachment sites of both L and HeLa cells, further indicating the specificity of the interaction. We also found that unlabelled reovirus competed equally well for both binding and internalization of 35S-labelled reovirus into murine L cells, suggesting that the L cell attachment site may serve as a virus entry site. Phospholipase digestion of L cells had no effect on subsequent reovirus attachment, while treatment of L cells with moderate concentrations of bromelain (but not trypsin, proteinase K or pronase) and Vibrio cholerae neuraminidase reproducibly decreased subsequent reovirus attachment. These results and those of others (Epstein et al., 1984, Virology 133, 46-55) suggest that mammalian reoviruses attach to specific cell surface receptors on at least two species of mammalian cells to initiate the infectious cycle. mammalian ment

reovirus,

saturable

attachment

site, competitive

binding,

enzyme

treat-

* A preliminary report of a portion of this work has been presented previously (J.R. Gentsch, Abstract, Annual Meeting of the Amerian Society for Virology, 1983, W16). 0168-1702/84/$03.00

0 1984 Elsevier Science Publishers B.V.

402

Introduction Studies of animal virus adsorption to host cells in culture have led to the finding that many viruses attach to specific, different receptors present on the cell surface (Philipson et al., 1968; McGrath et al., 1978; Schlegel et al., 1982b; Wunner et al., 1984). Although these receptors are presumed to be important in viral tissue tropsims and in penetration into the cytoplasm of the host cell (Tardieu et al., 1982; Crowell, 1983), the biochemical nature of cellular receptors for viruses and their normal biological functions are for the most part not understood. The mammalian reoviruses are a group of enteric, RNA containing viruses with a very broad host range both in vivo and in vitro (Joklik, 1974; Silverstein and Dales, 1976). Early studies on reovirus type 3 attachment to murine L cell fibroblasts indicated that 60-805X of the inoculum adsorbed to the cells within 2 h (Gomatos et al., 1968; Joklik, 1972; Silverstein et al., 1976). Dales and his collaborators demonstrated, using electron microscopic and biochemical studies, that reovirus particles are phagocytized by invagination of the cell membrane around surface bound particles (Dales, 1965; Silverstein and Dales, 1968; Dales, 1973). After cultures were warmed to 37 OC virus particles were observed first in phagocytic vesicles and then in lysosomes where it is thought that reovirus becomes uncoated by lysosomal proteases (Silverstein and Dales, 1968). However, it is not possible to distinguish, based on these studies, whether reovirus enters cells by receptor-mediated endocytosis as reported for a number of other enveloped (Helenius et al., 1980; Miller and Lenard, 1982; Schlegel et al., 1982a) and unenveloped (Fitzgerald et al., 1983) animal viruses. Canning and Fields presented evidence that ammonium chloride, a lysosomotropic agent, inhibits reovirus replication at a stage subsequent to penetration of the plasma membrane (Canning and Fields, 1983). Since these agents have been reported to inhibit receptor-mediated endocytosis (Schlegel et al., 1982a) and release of viruses from lysosomes or prelysosomal vacuoles (Helenius et al., 1980; Marsh et al., 1983), it is conceivable that reoviruses may gain access to the cytoplasm by a similar mechanism, although there is no direct evidence for this. Although the kinetics of reovirus attachment to L cells have been examined (Gomatos et al., 1968), it was not known until recently whether a specific cell surface receptor for reoviruses was present on mammalian cells. Tardieu and Epstein and their coworkers (Tardieu et al., 1982; Epstein et al., 1984) have presented evidence that the attachment site on murine cells for type 1 and type 3 reoviruses is specific and saturable, suggesting that only a finite number of discrete sites are available for virus binding. These studies are consistent with the possibility that reoviruses enter mammalian cells by a receptor mediated process. In this report we have characterized the binding site for type 3 reovirus on murine L cell fibroblasts and human HeLa cells. In addition, we have analyzed the effect of enzymatic modification of the cell surface on subsequent reovirus attachment. We have found that saturable attachment sites exist for reovirus on both L cells and HeLa cells. These sites are highly specific in that unlabelled reovirus competes with 35S-labelled reovirus for a limited number of receptor sites. We have shown that the saturable binding site on L cells may serve as a virus entry site, since competition for

403

binding or internalization of 35S-labelled reovirus are both completely inhibited in an almost identical manner by unlabelled reovirus. The binding site is completely resistant to phospholipase digestion of L cells and is resistant to moderate concentrations of trypsin, proteinase K and pronase, but is partially sensitive to predigestion of L cells with bromelain and Vibrio cholerae neuraminidase.

Materials and Methods Chemicals L-[ 35S]Methionine (600 Ci/mmol) was purchased from Amersham/Searle Corp., Arlington Heights, Ill. Proteinase-K and Vibrio cholerae neuraminidase were purchased from Calbiochem-Behring (San Diego, Calif.). All other enzymes (proteases and phospholipases) were purchased from Sigma Chemical Co. (St. Louis, MO.). Cells and viruses The propagation of L cells and the growth and purification of [35S]methionine labelled or unlabelled reovirus in equilibrium CsCl density gradients have been described in detail in previous publications (Ramig et al., 1977; Gentsch and Fields, 1981). HeLa cells were kindly provided by Dr. Roberto Weinmann of the Wistar Institute and propagated in the same way as mouse L cells. Binding assays Exponentially growing L or HeLa cells were pelleted at (i) Equilibrium binding. low speed, resuspended and washed once with binding buffer (Joklik modified minimal essential media (MEM) containing 3.5 mM glutamine, 2% fetal bovine serum (FCS) and 50 mM N, N-bis(hydroxymethyl)-2-aminoethane sulfonic acid buffer (BES), pH 7.4, and then resuspended in the same buffer at 2.0 X lo6 cells/ml and placed in an ice bath. The cells were kept in suspension by placing a small magnetic stirring bar in the tube and placing the ice bath and tube on a magnetic stirring apparatus. Reovirus ([ 35S]methionine labelled) was then added and samples (2.0 x lo5 cells each, in duplicate) were taken at appropriate intervals, pelleted 7 s in an Eppendorf centrifuge, washed once with 0.5 ml of binding buffer (BB) and the acid-insoluble, cell-associated 35S-labelled reovirus precipitated overnight with 0.5 ml of cold 5% trichloroacetic acid (TCA). TCA pellets were resuspended in 0.24 ml of 0.15 N NaOH and counted in 3 ml of Scintiverse scintillation cocktail (Fisher Scientific Co., Pittsburgh, Pa.). L or HeLa cells were pelleted, washed as described (ii) Saturation binding assays. above and resuspended in 0.05 ml (1 X lo5 cells/tube) or 0.10 ml (2 x lo5 cells/tube) of binding buffer containing the appropriate amount of 35S-labelled or 35S-labelled plus unlabelled reovirus type 3. Unlabelled virus was pre-dialyzed against binding buffer (without FCS) and subsequently adjusted to 2% in FCS. Radioactive virus in

404

0.15 M NaCl/l.S mM MgCl,/lO mM Tris-HCl, pH 7.4, was adjusted to 2% in FCS/SO mM BES/3.5 mM glutamine, and then diluted at least 10 fold into binding buffer (with or without unlabelled virus) to give the appropriate amount of 35S label per sample. Samples in Prosil-28 (Thomas Scientific Co., Philadelphia, Pa.) treated microfuge tubes were then incubated at 3OC until equilibrium had been achieved (about 2 h for reovirus binding to HeLa cells and 5 h for binding to L cells). Samples were pelleted, washed once with 0.5 ml binding buffer, transferred to new microfuge tubes and then processed as described above under equilibrium binding assays. Specific activities for [ 35S]methionine-labelled reovirus preparations varied from approximately 9000 to 20000 cpm per pg of viral protein. (iii) Competition for binding and internalization of reovirus. Washed L or HeLa cells (2 X lo5 cells/tube) were suspended in 0.1 ml of binding buffer containing a constant amount of 35S-labelled reovirus (0.5 pg) and increasing amounts of unlabelled reovirus between 0 and 128 pg (L cells) or between 0 and 256 pg (HeLa cells) and incubated at 3°C until equilibrium had been reached. Cell associated radioactivity was then measured as described in section (i) above. For competition for binding and internalization experiments (on L cells only), a second series of identical samples (see Fig. 3 and text) were incubated at 37 OC for 1 h so that internalization of virus could take place. These samples were washed once and resuspended in a solution of PBS containing 0.125% trypsin/0.02% EDTA (Schlegel et al., 1982b) and incubated for 30 min at 3’C and 5 min at 37 ‘C, then washed twice in binding buffer containing 20 pg/ml phenylmethyl sulfonyl fluoride (PMSF) and cell associated radioactivity measured as described before. This trypsinization procedure typically removed 84-90% of surface-bound virus. Protease treatment of mouse L cells Mouse L cells were pelleted at low speed and washed twice with PBS, pH 7.2, containing 2 mM CaCl, and 0.5 mM MgCl, (PBS-Ca-Mg) for subsequent digestion with trypsin, proteinase K or pronase. Alternatively, cells were washed with CaCl,-, MgCl,-free PBS, pH 7.6 and once with PBS containing 0.3 mM EDTA for subsequent digestion with bromelain. Cells were then suspended at 1 X 10’ cells/ml and digested with the appropriate concentration of trypsin, proteinase K or pronase in PBS-Ca-Mg, or the appropriate concentration of bromelain in PBS pH 7.6 and digested for 60 min at 24’C. Samples were washed twice with binding buffer containing 2% FCS and 20 pg/ml (PMSF) (trypsin, proteinase K and pronase samples) or twice with PBS 0.3 mM EDTA containing 20 pg/ml N-alpha-tosyl-Llysylchloromethane (TLCK) and once with binding buffer - 2% FCS (bromelain digested samples). Samples were then suspended in binding buffer, 2% FCS and an aliquot mixed with trypan blue solution and counted using a hemocytometer and an inverted microscope. 2 x lo5 total cells per tube (for both protease-treated and mock-treated cells) were then pelleted and resuspended in 0.1 ml binding buffer containing 2% FCS and 20000-30000 cpm of 35S-labelled reovirus type 3 and incubated at room temperature with frequent mixing until binding equilibrium had been reached (90 min). Samples were then pelleted, washed once with binding buffer and cell associated radioactivity measured.

405

Phospholipase treatment of mouse L cells L cells (6 x 105-1 x 106/reaction) were washed once in serum-free binding buffer, pH 7.4 (Schlegel et al., 1982b) or in 0.14 M NaCl/S mM CaCl,/S mM N-tris(hydroxymethyl)-methyl-2-aminoethanesulfonic acid (TES buffer), pH 8.5, and digested for 60 min at 37 ‘C in serum-free binding buffer (phospholipase C and D) or in TES buffer (phospholipase A,) at concentrations ranging from 0 to 250 units phospholipase per ml. At the end of digestion tubes were washed three times in binding buffer, resuspended and counted in trypan blue solution as described for protease treated cells. 2-6 x lo5 cells per tube were then pelleted and resuspended in binding buffer containing 35S-labelled reovirus and a binding assay carried out as described above for protease treatment of mouse L cells. The effect of phospholipase digestion of L cells on nonsaturable binding was measured by predigesting L cells with phospholipases as described above and then determining total and nonsaturable binding essentially as described above in the section on saturable binding and in the legend to Fig. 2. The amount of cold virus used to determine the non-saturable (linear) binding component varied from 150 pg to 1 mg per sample. Neuraminidase treatment of L cells 1 x lo6 L cells were pelleted and washed once with PBS containing 2 mM CaCl, and 5 mg/ml bovine serum albumin and resuspended in the same buffer containing O-100 mU Vibrio cholerae neuraminidase (VCN), digested for 30 min at 37 ‘C and washed twice in the same buffer without VCN. Cell samples (2 X lO’/tube) were then mixed with 35S-labelled reovirus (10 000-20 000 CPM) and cell-associated radioactivity determined at equilibrium as described above for protease-treated L cells.

Results Equilibrium binding conditions for type 3 reovirus attachment to mouse L cells In order to study the interaction of reovirus type 3 with L cells, equilibrium binding conditions were first established. Fig. 1 shows that for binding of 35S-labelled reovirus to mouse L cells at 3 “C it takes between 4 and 5 h for maximal binding (equilibrium) to be achieved. Under these conditions 30-35% of the initial input virus becomes attached to the cell surface. Little, if any, virus was internalized at 3OC as indicated by the fact that 84-90% or more of the surface-bound virus could be removed by treating the cells with trypsin after maximal attachment had occurred (not shown). This is in agreement with other studies demonstrating that little internalization of surface bound viruses occurs at 3°C (Helenius et al., 1980; Schlegel et al., 1982b). It should also be noted that about 90% of the infectivity of the 35S-labelled input virus was removed by the time equilibrium had been reached (not shown). This rules out the unlikely possibility that this type of binding assay measures only attachment

I I

3

5 TIME(h)

Fig. 1. Equilibrium binding conditions for type 3 reovirus attachment to mouse L cells. 35S-labelled reovirus type 3 (0.5 gg/ml, 41000 cpm/ml, 100 pfu/cell) was added to suspension cultures of mouse L cells (2 X lo6 cells/ml) at 3O C and cell associated acid-insoluble radioactivity was measured at intervals. For this experiment and all others to follow each time point or sample was analyzed in duplicate and each figure represents the mean of 3-6 independent experiments.

of noninfectious typical purified

virus particles (which can make up anywhere from 1 to 99% of a virus preparation) (Spendlove et al., 1970; Silverstein et al., 1976).

Saturation of L cell receptor sites for type 3 reovirus To determine if the attachment site on L cells for reovirus was specific, a saturation binding analysis was carried out. In these experiments increasing amounts

,LQ REOVIRUS

ADDED

Fig. 2. Saturation of cell receptor sites for type 3 reovirus. (2500-40000 cpm) was added to replicate cultures of 2X lo5 (0) of excess (250 pg/tube) unlabelled reovirus type 3 and was measured for each concentration after 5 h at 3O C. Total binding, (0); inset (specific binding).

Increasing amounts of 35S-labelled reovirus L cells either in the absence (A) or presence TCA-insoluble, cell-associated radioactivity 35S-labelled virus bound, (A); non-saturable

407

of 35S-labelled reovirus (prepared by mixing different amounts of radiolabelled and unlabelled virus to give the same specific activity) were added to replicate samples of 2 x lo5 L cells and binding to equilibrium then carried out. A second series of tubes contained identical amounts of 35S-labelled virus and a large excess (250 pg/tube) of unlabelled virus. Fig. 2 shows that in the presence of increasing 35S-labelled virus alone the attachment curve is initially hyperbolic, then begins to level off. In the presence of the same amounts of 35S-labelled virus and a large excess of unlabelled virus, binding increased linearly (non-saturable binding component) with increasing 35S-labelled virus concentration. When the non-saturable binding component was subtracted from total binding a characteristic saturation binding curve (Schlegel et al., 1982b) was obtained (inset). Specific cellular attachment sites were fully saturated at 45 pg input 35S-labelled virus per 2 X lo5 cells. Using the method of Lonberg-Holm and Korant (1972) and the knowledge that there are approximately 1.1 x 10” reovirus particles per pg of viral protein (Smith et al., 1969), it was calculated that there are approximately 86 000 cellular receptor (attachment) sites (CRS) per murine L cell for type 3 reovirus (1.5 pg reovirus bound per 2 X lo5 L cells). Multiple experiments done over a period of months with different batches of virus showed that binding site number varied from 8.6 X lo4 to 1.1 X lo5 per L cell. We have also measured the number of CRS using increasing amounts of 35Slabelled virus only and calculated that there are about 94000 CRS per cell (not shown), in good agreement with the result obtained with mixtures of 35S-labelled and unlabelled virus (Fig. 2). This results suggested that the binding properties of 35S-labelled reovirus and unlabelled reovirus are very similar. Therefore, to limit the amount of 35S-labelled virus needed, we usually used mixtures of “S-1abelled and unlabelled virus for saturation experiments. Competition for binding and internalization of 35S-labelled reovirus by unlabelled reovirus To determine if the saturable binding site on L cells was specific and if it could serve as a possible virus entry site, we carried out an experiment designed to measure competition for the binding and internalization of 35S-labelled reovirus by unlabelled reovirus. Constant amounts (0.5 pg) of 35S-labelled virus and increasing amounts of unlabelled reovirus were added to replicate cultures of L cells. One set of cultures was then incubated at 3°C and the other set at 37 “C. Competition for binding at 3°C was determined by measuring cell associated radioactivity at equilibrium. For the 37 “C cultures, competition for binding and internalization was measured by trypsinizing any remaining surface bound virus after equilibrium had been reached. As described earlier trypsinization removes 84-90% of surface bound virus. As shown in Fig. 3 competition for binding at 3’C and for binding and internalization at 37 OC were inhibited in a similar manner by increasing amounts of unlabelled reovirus, suggesting that the CRS for reovirus may also serve as a virus entry site (Schlegel et al., 1982b). Fig. 3 also shows the high efficiency of competition in this system, suggesting that the intraction between reovirus and its CRS is highly specific. At a ratio of 64 yg unlabelled to 0.5 pg labelled virus, attachment of 35S-labelled reovirus is reduced about 95%. With different batches of labelled and

unlabelled reovirus we have observed that competition above varies between 80-95% reduction of 35S-labelled

under the conditions virus binding.

shown

Characterization of HeLa cell receptor for type 3 reovirus To determine if other mammalian species possess specific receptors for reovirus type 3 we characterized the attachment of reovirus to human HeLa cells. HeLa cells are a human cell strain derived from a patient with cervical carcinoma (Gey et al., 1952) and like L cells, support reovirus replication efficiently. To determine if HeLa cells also possess a specific receptor for reoviruses, we examined the saturability and specificity of reovirus attachment under equilibrium conditions (shown in Fig. 4A). Fig. 4B shows that the HeLa cell receptor can be completely saturated with x lo5 cells or about 126000 35S-labelled typ e 3 reovirus (2.2 pg virus bound/2 CRS/cell). This estimate is in the same range (1.5 fold higher) as the number of CRS for reovirus present on murine L cells. Using 35S-labelled virus alone, instead of mixtures of 35S-labelled and unlabelled virus we obtained a figure of 147000 CRS per cell (not shown). Binding of 35S-labelled reovirus to HeLa cells was also highly specific since increasing amounts of unlabelled reovirus (Fig. 4C), competed for binding only slightly less efficiently than the observed competition in the reovirus-L cell system (see Fig. 3). Using different batches of 35S-labelled and unlabelled reovirus we found that maximum competition varied from 60 to 95% inhibition of binding of 35S-labelled reovirus. We have not yet examined whether the saturable receptor site on HeLa cells could serve as a virus entry site.

1

,.,g COMPETING

REOVIRUS

reovirus. Fig. 3. Competition for binding and internalization of 35S-labelled reovirus by unlabelled amounts of Constant amounts (0.5 pg) (5000-10004 cpm) of 35S-labelled reovirus and increasing unlabelled reovirus (O-128 pg) were added to replicate cultures of L cells (2 X lo5 cells/tube) and one set of samples incubated at 3 o C for 5 h and the other set at 37 o C for 1 h. The 3O C samples (A) were then processed to measure TCA-insoluble, cell-associated radioactivity. The 37 o C samples (0) were treated with trypsin to remove surface-bound virus and then TCA-insoluble radioactivity determined as a measure of internalized virus.

OoI TIME

/.~g REOVIRUS

(hi

40

80

120

@ COMPETING

160

200

240

REOVIRUS

ADDED

Fig. 4. Reovints type 3 attachment to HeLa cells. (A) Binding equilibrium conditions. 35S-labelled reovirus (1.6 pg/ml, 40000 cpm/ml, 320 pfu/cell) was added to HeLa cell suspension cultures and TCA-insoluble radioactivity determined at intervals as described in Fig. 1 and Materials and Methods. Three-fold more 35S-labelled virus was added in these experiments (compared to mouse L cell equilibrium used in this experiment. (B) experiments) because of a lower specific activity of the 35S virus preparations Saturation of HeLa cell receptor by type 3 reovirus. Increasing amounts of 35S-labelled reovirus or 35S-labelled reovirus (2400-47000 cpm) plus unlabelled reovirus were added to HeLa L cell cultures and total (A), non-saturable (0), and specific (inset) (0). reovirus attachment measured as described in Fig. 2 and Materials and Methods. (C) Competition for 35S-labelled reovirus binding by unlabelled reovirus. Constant amounts of 35S-labelled reovirus (0.5 ng) (10000-13000 cpm) and increasing amounts of unlabelled reovirus (O-256 pg) were added to HeLa cell suspension cultures and after equilibrium was reached TCA-insoluble, cell-associated radioactivity was measured as described in Fig. 3 and Materials and Methods.

Effect of enzyme treatment of mouse L cells on the attachment of type 3 reovirus In an effort to determine the chemical nature of the reovirus attachment site(s) on mammalian cells we have analyzed the effect of digesting L cells with several enzymes on subsequent reovirus attachment. These results are summarized in Table 1. It can be seen that digestion of living L cells with moderate concentrations of trypsin, proteinase K, or pronase, under conditions where cell viability (as measured by trypan blue exclusion) was minimally affected, had little effect on subsequent attachment of reovirus type 3. High concentrations (1 mg/ml) of pronase, proteinase K and bromelain inhibited subsequent attachment 40-8056 but in these cases cell

410 TABLE

1

EFFECT OF ENZYME 3 REOVIRUS Enzyme

Concentration

Trypsin

Proteinase

TREATMENTS

0

Pronase

Bromelain

Neuraminidase

Phosphohpase

C

a,b

L CELLS

% Maximal

ON THE ATTACHMENT

attachment

’

Cell viability

10 0 10 100 0 25 250

100 143 119 100 124 101 100 109 96 100 12 67

83 85 96 90 90 81 85 87 92 69 86 85

0 100’

100 48

90 89

0 20 50 120

100 99 106 83

92 90 83 66

10 0

K

OF MOUSE

OF TYPE

d

Protease concentrations are in pg/ml. Phospholipase and neuraminidase concentrations are in units/ml and milliunits/ml, respectively. Notes: One international unit of Neuraminidase (Calbiochem-Behring) is defined as the amount of enzyme that will release 1 pmol of N-acetylneuraminic acid from human acid cx,-glycoprotein at 37 o C per minute under conditions specified by the manufacturer. One unit of phopsholipase C (Sigma) is defined as the amount of enzyme that will release 1 pmol of organic phosphorous from L-a-phosphatidyl choline per minute at pH 7.3 and 37 OC. Attachment was measured under equilibrium binding conditions (90 min at room temperature), each point represents the median value of 4 or more independent experiments. Attached 35S-labelled reovirus cpm varied from about 2500-13000 for different experiments. The effect of each enzyme was analyzed as a separate experiment. Measured by trypan blue exclusion. Other neuraminidase concentrations between 1 and 35 mU also removed between 50 and 70% of the attachment site (not shown).

viability was markedly decreased (not shown). As seen in Table 1 bromelain digestion of L cells at 25 or 250 pg/ml reproducibly decreased subsequent reovirus attachment under conditions where the viability of bromelain treated cells was equal to or greater than the viability of control cultures incubated in the same buffer. It should be noted that the effect of these proteases on reovirus attachment are not directly comparable since the number of units of protease used for the digestions varied from one protease preparation to another. It has been reported that Vibrio cholerue neuraminidase pretreatment of ox erythrocytes inhibited subsequent hemagglutination reactions with type 3 reovirus (Gomatos and Tamm, 1968). We therefore tested the effect of neuraminidase

411

pretreatment of L cells on subsequent reovirus attachment (Table 1). We observed a highly reproducible effect which ranged from 50-60% inhibition of binding. As for other enzyme treatments this effect was observed under conditions where cell viability was comparable to control values. In addition, the observed effect could be prevented by including excess fetuin (a sialic acid containing glycoprotein which is a good neuraminidase substrate) or 2,3-dehydro-2-deoxy-N-acetylneuraminic acid ( a potent neuraminidase inhibitor) in the reaction (not shown), strongly suggesting that the effect is mediated by neuraminidase (Rogers and Paulson, 1983). It has recently been demonstrated that phosphatidyl serine and possibly other lipids may be components of the cell receptors for vesicular stomatitis virus (Schlegel et al., 1983) and rabies virus (Wunner et al., 1984). In the case of vesicular stomatitis virus, these conclusions were partially based on the complete sensitivity of the receptor to digestion of intact cells with phospholipase C. As seen in Table 1, reovirus attachment to mouse L cells is completely insensitive to pretreatment of L cells with phospholipase C. The concentrations used were up to 2.5 fold higher than those used by Schlegel and his coworkers. Phospholipase C pretreatment (at 120 U/ml) caused a slight reduction in binding (17%), but only under conditions where cell viability had substantially decreased (to 66%). We have also determined that nonsaturable binding (see Fig. 2 for conditions for measuring nonsaturable binding) of reovirus to L cells is unaffected by pretreatment of cells with phospholipase C at 50-75 units/ml (not shown). Phospholipase A, and D treatment of L cells by identical protocols also had no effect (not shown) on subsequent reovirus attachment. However, living cells have been reported to be poor substrates for phospholipases A, and D (Raelofsen et al., 1974) so these data were not included.

Discussion Although it has been known for some time that reovirus attaches efficiently to murine L cell suspension cultures (Gomatos et al., 1968; Joklik, 1972; Silverstein et al., 1976) it had not been determined whether the reovirus attachment site behaved as a specific, saturable receptor with properties similar to other animal virus receptors (Phillipson et al., 1968; McGrath et al., 1978; Schlegel et al., 1982b; Wunner et al., 1984). In this report we have characterized the properties of murine L cell and human HeLa cell attachment sites for type 3 reovirus. Our results demonstrate that the binding sites on murine L cells and human HeLa cells for type 3 reovirus possess characteristics typical of other animal virus receptors in that they are specific and can be fully saturated with 35S-labelled virus (Phillipson et al., 1968). We found that L and HeLa cells possess 8.6 X 104-1.1 X lo5 and 1.3 x 105-1.5 x 10’ specific attachment sites per cell, respectively, for type 3 reovirus. Our results confirm those of Epstein and coworkers (Epstein et al., 1984) who reported that murine L cells possess about 8.6 X lo4 specific attachment sites per cell for lz51labelled type 3 reovirus and extend to a second species of mammalian cells the finding of specific cell surface receptors for type 3 reovirus. Competitive binding studies indicated that L and HeLa cell attachment sites

412

(CRS) for type 3 reovirus were highly specific, since binding of 35S-labelled reovirus could be reduced by as much as 95% with excess unlabelled reovirus. These results are also in good agreement with those of Epstein and coworkers who reported that unlabelled reovirus preparations competed efficiently for the attachment of ‘2sIlabelled reovirus to murine cells (Epstein et al., 1984). In addition we have presented data showing that competition for binding of 35S-labelled type 3 reovirus to L cells (measured at 4’C so that internalization could not take place) and competition for binding and internalization (measured at 37 ‘C so that internalization takes place) are inhibited in the same manner by increasing amounts of unlabelled reovirus type 3. This result suggests that the specific L cell attachment site for reovirus could serve as a major virus entry site, by analogy with the studies reported for vesicular stomatitis virus binding to Vero cells (Schlegel et al., 1982b). However, our system (as for the VSV-Vero cell systems) possesses a linear or nonsaturable binding component which represents 20-30s of the total observed binding (Epstein also observed about 25% nonsaturable binding for type 3 reovirus). We cannot at this time rule out that this could also serve as a virus entry site. In this regard, Borsa and his collaborators have demonstrated that reovirus subviral particles prepared in vitro by chymotrypsin digestion, or in vivo in infected cells are capable of direct penetration of the L-cell membrane (Borsa et al., 1979). Since all virus stocks (cell lysates) usually contain small numbers of these partially uncoated particles, it is conceivable that reovirus may have several entry pathways into L cells. We previously demonstrated that type 3 reovirus binds avidly to acidic L cell phospholipids, phosphatidylserine, and several other phospholipids (but not to neutral lipids) which have been resolved on thin layer silica plates (Gentsch and Higgins, unpublished results). To determine if lipids might play a major role in reovirus attachment as described for rhabdoviruses (Schlegel et al., 1983; Wunner et al., 1984), we measured reovirus attachment to cells treated with phospholipase C. Treatment of L cells with phospholipase C had no effect on either the specific or the nonsaturable binding sites for type 3 reovirus, even at concentrations up to 2.5 times greater than those used by Schlegel and coworkers. Although this evidence cannot be regarded as conclusive, we presently believe that cellular lipids are not major components of the reovirus saturable binding site on murine L cells. We also analyzed the protease sensitivity of the reovirus binding site on mouse L cells under conditions which had minimal effect on cell viability. The L cell attachment site for reovirus was completely resistant to moderate concentrations of one highly specific (trypsin) or two non-specific (pronase and proteinase K) proteases. The mechanism whereby trypsin (and possibly proteinase K) pretreatment of L cells at low concentrations causes a slight enhancement of virus binding is not known. In contrast to the results with most proteases, bromelain was able to reproducibly reduce subsequent reovirus attachment by a moderate amount. This is not conclusive evidence for the involvement of a protein in the L cell attachment site for reovirus. Various cell surface alterations caused by the protease-mediated removal of glycoproteins and proteoglycans could alter binding of virions to the surface. It should also be noted that bromelain preparations are not highly purified and we cannot yet rule out that other enzyme activities are present in these

413

preparations. We are currently doing more detailed experiments to optimize the observed bromelain effect, to characterize the inhibitory activity and to analyze the saturability and specificity of attachment after such treatments. Preliminary experiments indicate that reovirus attachment can be reduced up to 80% under conditions where cell viability is also decreased (not shown). Epstein and her coworkers reported susceptibility of type I reovirus attachment to L cells to pretreatment of cells with 1 mg/ml pronase. We observed a similar effect for type 3 reovirus (at 1 mg/ml pronase, proteinase K or bromelain) but cell viability was variable and low under these conditions. Since we only found decreased attachment under conditions where cell death and cell lysis were extensive, we are not sure of the meaning of this result. However we cannot rule out that a highly protease resistant protein or glycoprotein is involved in attachment. We have also found that neuraminidase digestion of L cells reproducibly decreases subsequent reovirus type 3 attachment suggesting that cell surface, sialic acid containing glycoconjugates may be involved in virus attachment. This result is also consistent with the finding that neuraminidase pretreatment of erythrocytes destroys the ability of these cells to be hemagglutinated by type 3 reovirus (Gomatos and Tamm, 1968). We are currently analyzing the possibility that cell surface glycoconjugates may be a component of the type 3 reovirus attachment site on mouse L cells.

Acknowledgements The authors thank Ms. Remonia N. Tarpley for her help in preparing this manuscript and Drs. Howard Weiner and William Wunner for providing preprints of two manuscripts before their publication. This work was supported in part by NIH grant AI 18848, by BRSG S07-RR-05415-21 awarded by the Biomedical Research Support Grant Program, Division of Research Resources, NIH, by ACS IN-135C, awarded to the University of Pennsylvania Cancer Center by the American Cancer Society. References Borsa, J., Morash, B.D., Sargent, M.D., Copes, T.P., Lievar, R.T. and Szekeley, J.G. (1979) Two modes of entry of reovirus particles into L cells. J. Gen. Virol. 45, 161-170. Canning, W.M. and Fields, B.N. (1983) Ammonium chloride prevents lytic growth of reovirus and facilitates the establishment of persistent infection in mouse L cells. Science 219, 987-988. Crowell, R.L. and Landau, B.J. (1982) Receptors in the initiation of picomavirus infections. In: Comprehensive Virology, Vol. 18 (Fraenkel-Conrat, H. and Wagner, R.R., eds.), pp. l-42. Plenum Press, New York. Dales, S. (1973) Early events in cell-animal virus interactions. Bacterial. Rev. 37, 103-135. Dales, S., Gomatos, P.J. and Hsu, T.C. (1965) The uptake and development of reovirus in strain L cells followed with labelled viral nucleic acid and ferritin antibody conjugates. Virology 25, 193-211. Epstein, R.L., Powers, M.L., Bogart, R.B. and Weiner, H.L. (1984) Binding of ‘251-labelled reovirus to cell surface receptors. Virology 133, 46-55. Fitzgerald, D.J.P., Padmanabhan, R., Pastan, I. and Willingham, M.C. (1983) Adenovirus induced release of epidermal growth factor and pseudomonas toxin into the cytosol of KB cells during receptor-mediated endocytosis. Cell 30, 607-617.

414

Gentsch, J.R. and Fields, B.N. (1981) Tryptic peptide analysis of outer capsid polypeptides of mammalian reovirus serotypes 1, 2, and 3. J. Virol. 38, 208-218. Gey, G.O., Coffman, W.D. and Bobicek, M.T. (1952) Tissue culture studies of the proliferative capacity of cervical carcinoma and normal epithelium. Cancer Res. 12, 264-265. Gomatos, P.J. and Tamm, I. (1968) Reactive sites of reovirus type 3 and their interaction with receptor substances. Virology 17, 455-461. Gomatos, P.J., Tamm, I., Dales, S. and Franklin, R.M. (1968) Reovirus type 3: Physical characteristics and interaction with L cells. Virology 17, 441-445. Helenius, A., Kartenbeck, J., Simons, K. and Fries, E. (1980) On the entry of Semliki-Forest virus into BHK-21 cells. J. Cell. Biol. 84, 404-420. Joklik, W. (1972) Studies on the effect of chymotrypsin on reovirions. Virology 49, 700-715. Joklik, W. (1974) Reproduction of reoviridae. In: Comprehensive Virology, Vol. 18 (Fraenkel-Conrat, H. and Wagner, R.R., eds.), pp. 231-230. Plenum Press, New York. Lee, P.W.K., Hayes, E.C. and Joklik, W.K. (1981) Protein eI is the reovirus cell attachment protein. Virology 108, 156-163. Marsh, M., Bolzau, E. and Helenius, A. (1983) Penetration of Semliki-Forest virus from acid prelysosomal vacuoles. Cell 32, 931-940. McGrath, MS., Decleve, M., Kaplan, H.S. and Weissman, I.L. (1978) Specificity of cell surface receptors on radiation Leukemia virus and radiation induced thymomas. J. Virol. 28, 819-827. Miller, D.K. and Lenard, J. (1980) Inhibition of vesicular stomatitis virus infection by spike glycoprotein. J. Cell. Biol. 84, 430-437. Miller, D.K. and Lenard, J. (1982) Uncoating of enveloped viruses. Cell 28, 5-6. Philipson, L., Lonberg-Holm, K. and Petterson, U. (1968) Virus-receptor interaction in the Adenovirus system. J. Virol. 2, 1064-1075. Ramig, R.F., Cross, R.K. and Fields, B.N. (1977) Genome RNAs and polypeptides of reovirus serotypes 1, 2, and 3. J. Virol. 22, 726-733. Raelofsen, B., Zwaal, R.F.A. and Woodward, C.B. (1974) The action of pure phospholipases on native and ghost red cell membranes. Methods Enzymol. 32, 131-139. Rogers, G.N. and Paulson, J.C. (1983) Receptor determinants of human and animal influenza virus isolates: differences in receptor specificity of the H3 hemagglutinin based on species of origin. Virology 127, 361-373. Schlegel, R., Dickson, R.B., Willingham, M.C. and Pastan, I.H. (1982a) Amantidine and dansylcadaverine inhibit vesicular stomatitis virus uptake and receptor mediated endocytosis of 2-macroglobulin. Proc. Natl. Acad. Sci. 79, 2291-2295. Schlegel, R., Willingham, M.C. and Pastan, 1.H. (1982b) Saturable binding sites for vesicular stomatitis virus on the surface of vero cells. J. Virol. 43, 871-875. Schlegel, R., Tralka, T.S., Willingham, M.C. and Pastan, I.H. (1983) Inhibition of VSV and infectivity by phosphatidylserine. Is phosphatidylserine a VSV binding site? Cell 32, 639-646. Silverstein, SC., Christman, J.K. and Acs, G. (1976) The reovirus replicative cycle. Annu. Rev. B&hem. 45, 375-408. Silverstein, S.C. and Dales, S. (1968) The penetration of reovirus RNA and initiation of its genetic function by L strain fibroblasts. J. Cell. Biol. 36, 197-230. Smith, R.E., Zweerink, H.J. and Joklik, W.K. (1969) Polypeptide components of virions, top component and cores of reovirus type 3. Virology 39, 791-810. Spendlove, R.S., McClain, M.E. and Lennette, E.H. (1970) Enhancement of reovirus infectivity by extracellular removal or alteration of the virus capsid by proteolytic enzymes. J. Gen. Virol. 8, 83-94. Tardieu, M., Epstein, R.L. and Weiner, H.W. (1982) Interaction of viruses with cell surface receptors. Int. Rev. Cytol. 80, 27-61. Wunner, W.H., Reagan, K.J. and Koprowski, H. (1984) Characterization of saturable binding sites for Rabies virus. J. Virol. 50. 691-697. (Manuscript

received

27 February

1984)