Concanavalin A

Proc. Nat. Acad. Sci. USA Vol. 71, No. 6, 2456-2460, June 1974

Receptor Mobility and the Mechanism of Cell-Cell Binding Induced by Concanavalin A (cells on nylon fiber/receptor alignment/agglutination/lymphoma cells/lymphocytes) URS RUTISHAUSER AND LEO SACHS Department of Genetics, Weizmann Institute of Science, Rehovot, Israel Communicated by Gerald M. Edelman, April 1, 1974

ABSTRACT The cell-cell binding induced by con- this interaction in terms of membrane structure and receptor canavalin A between single cells has been analyzed by use of cells attached to nylon fibers. Binding of a concanavalin dynamics. A-coated cell to an untreated cell was found to a high degree between two lymphoma tumor cells, less frequently MATERIALS AND METHODS between a lymphoma cell and a normal lymphocyte, and only rarely between two normal lymphocytes. The binding Cells. Normal lymphocytes were obtained from the lymph was inhibited by the presence of a saccharide inhibitor of nodes of 6- to 8-week-old CR/RAR rats or A strain mice. concanavalin A, but could not be reversed by addition of Cells from the two sources contained 70-85% thymus-derived the inhibitor after the cells had bound to each other. cells and gave the same results in cell-cell binding assays. Although nonbinding was obtained when both cells were Lymphocytes were obtained by teasing the cells into phos- coated with lectin or fixed with glutaraldehyde, fixation of a cell before coating with concana'valin A enhanced its phate-buffered saline, pH 7.4 (PBS), and then washing the ability to bind an untreated cell. The results indicate that cell three times with PBS. Lymphoma tumor cells were ob- cell-cell binding induced by concanavalin A requires short- tained from a Moloney virus-induced lymphoma grown in A range lateral movement of cell receptors for the lectin, strain mice (8). This is a tumor of thymus-derived cells (9). that only one cell has to have mobile receptors, and that were inoculated into adult some receptors must be unoccupied by lectin molecules Cells (105) intraperitoneally mice, before cell-cell contact. Clustering of the receptors is not and the cells were harvested 10-14 days later. The cells were necessary and seems to hinder cell-cell binding. It is sug- washed three times in PBS before they were used. gested that the short-range movement is required for alignment of individual receptors so as to form multi-point Cell-Cell Binding Assay. Nylon fibers were strung in poly- bridges between two cells by lectin molecules. The bridging ethylene frames, derivatized with concanavalin A (Con A, is then followed by the formation of irreversible bonds Miles Yeda) at 0X5 mg/ml, and incubated with cells as between the cells. The receptors on tumor cells appear to described (3). The binding of cells to fibers by shaking at 250 have a greater -ability than receptors on normal cells to align themselves for cell-cell binding. for 30 min was carried out with 2.5 X 106 cells per ml in PBS. The cell-cell binding assay (Fig. 1) involved three steps: (a) The agglutination of cells by lectins has been used as a probe binding of cells to Con A-fibers and removal of unbound cells to study changes in the cell-surface membrane (1, 2). Ag- by washing (3); (b) binding of Con A (200 jug/ml, except where glutination is generally determined by adding the lectin to a noted) for 30 min at 250 to either the fiber-bound (FB) cells cell suspension, shaking the suspension, and recording the or to free cells, followed by three washes with PBS; and (c) degree of cell aggregation. Although this agglutination assay incubation of FB cells with a cell suspension (106 cells per ml is rapid and convenient, it simultaneously measures the bind- in PBS) at 250 for 30 min with gentle shaking, as in the binding of lectin to the cells and of several cells to one another. ing of cells to fibers. In each step, a new petri dish was used The specific attachment of cells to chemically derivatized to prevent carry-over of cells or Con A. After unbound fibers (3, 4) has been used to fractionate and manipulate cells were washed away, the number of cell-bound (CB) cells normal lymphoid cells for morphological and functional were counted along one edge of a 1-cm segment (Fig. 2). studies (5-7). In the present paper, the fiber-rosette method Because normal lymphocytes are smaller than the lymphoma used in antigen-binding studies (5) has been modified for the cells, it was not always possible to distinguish between tumor analysis of cell-cell binding induced by the lectin concanavalin cells bound to normal FB cells and tumor cells that may have A (Con A). This system allows the quantitative study of bound directly, to the fiber. This difficulty did not arise binding between single cells, in which the binding of Con A when tumor cells were fixed by glutaraldehyde so as to pre- to a cell and the binding of that cell to another cell can be vent their binding to the fiber. carried out as successive steps. This procedure has been used with Meta- and Cell Fixation by Glutaraldehyde and Treatment to study the cell-cell binding of normal lymphocytes bolic and Neuraminidase. cells induced Con and to Inhibitors, Trypsin, Glutaraldehyde lymphoma tumor by A, interpret fixation of fiber-bound cells and free cells (5 X 106/ml) Abbreviations: Con A, concanavalin A; F-Con A, fluorescein- was carried out with 2.5% glutaraldehyde in PBS at 40 for 2 labeled concanavalin A; CB cell, cell-bound cell; FB cell, fiber- or 4 hr (10) or with 3% glutaraldehyde in PBS at 22° for 2 bound cell; PBS, phosphate-buffered saline, pH 7.4; aNIM, hr (11) with identical results. The cells were then incubated a-methyl-D-mannopyranoside. in 0.2 M glycine at 40 for 10 min, followed by three washes 2456 Downloaded by guest on September 27, 2021 Proc. Nat. Acad. Sci. USA 71 (1974) Cell-Cell Binding Induced by Concanavalin A 2457

L- + 9 Con A + uor

FIG. 1. Procedure used in the cell-cell binding assay for the preparation of fiber-bound (FB) cells attached to Con A derivatized fibers and cell-bound (CB) cells. The shape and orientation of the Con A molecules and cell receptors for Con A is schematic.in, multivalent Con A molecule.

in PBS (10). Fiber-bound or free cells were incubated with Behringwerke AG; 25-100 ug/ml, 30-60 min, 370) was carried metabolic inhibitors (10 mM NaN3 or 1 mM 2.4-dinitro- out on fiber-bound and free cells before the addition of soluble phenol) for 30 min at 250 before and during cell-cell binding. Con A. Treatment of the cells with trypsin (3 X crystallized, Calbiochem; 5,gg/ml, 30 min, 250) or neuraminidase (purified, Fluorescence Microscopy. The distribution of Con A receptors on fiber-bound and cell-bound cells was observed in situ by fluorescence microscopy (7, 9), with 200 g.g/ml of a fluorescein-labeled Con A (F-Con A, Miles Yeda). F-Con A was first used to stain the FB cell and to induce cell-cell binding. F-Con A was then again added to stain the CB cell. ,- N RESULTS Assay of Cell-Cell Binding Induced by Con A. Fibers derivatized with Con A were used to study cell-cell binding of normal lymphocytes and lymphoma cells induced by Con A. The gen- eral procedure is summarized in Fig. 1 and illustrated in Fig. b 2. When 2.5 X 10 cells per ml were incubated with Con A fibers, the same number of normal and tumor cells were bound per unit length of fiber. With 2.5 X 106 cells per ml, both cell types formed a confluent monolayer along the fiber edge (Fig. 2a). The FB cells could not be removed by subsequent incubation with the competitive inhibitor, a-methyl-D-mannopyranoside (aMM) (6). After binding of cells to the fiber, soluble Con A was added at various concentrations to allow c d binding of the lectin to receptors not involved in the cell-fiber linkage. The cells were washed to remove unbound Con A and the addition of free cells with receptors for Con A then re- I i sulted, in certain cell combinations, in the binding of these I cells to the FB cells (Figs. 1 and 2b-d). The presence of 10 mM I aMM during the incubation of free and bound cells prevented cell-cell binding and the number of cell-bound cells was re- producible to + 10%. Increasing the concentration of Con A produced more cell-cell binding, with a plateau at about 200 gg of Con A per ml (Fig. 3). A fiber saturated with cell-

FIG. 2. Photographs of fiber-bound cells and of cell-cell binding induced by Con A. (a) Lymphoma tumor cells bound to fiber. (b) Cell-cell binding between tumor cells. (c) Two tumor cells bound to the same FB tumor cell. (d) Normal lymphocyte Soluble Con A(,Ug/ml) bound to a FB tumor cell. (e) Fluorescence microscopy of a FIG. 3. Effect of Con A concentration on the number of cell- tumor cell bound to a FB-tumor cell, both stained with F-Con A. bound cells per cm of fiber. The concentration of soluble Con A (f) Fluorescence microscopy of a normal lymphocyte bound to a is the amount used to coat the FB cells before binding of the FB tumor cell, both stained with F-Con A. a and b X332; c-f CB cells (Fig. 1). TT, TN, and NN = tumor-tumor, tumor- X996. normal, and normal-normal cell binding, respectively. Downloaded by guest on September 27, 2021 2458 Cell Biology: Rutishauser and Sachs Proc. Nat. Acad. Sci. USA 71 (1974) TABLE 1. Cell-cell binding induced by Con A between two TABLE 2. Cell-cell binding induced by Con A between two lymphoma tumor cells normal lymphocytes

Cell-cell Cell-cell Fiber-bound cell* Cejl-bound cell* binding (cm-1)t Fiber-bound cell* Cell-bound cell* binding (cm-)t No fixation: No fixation: T T 3 N N 12 Tc T 359 No N 52 Fixation of one cell: N Nc 40 Tgc T 5814 Nc Nc 15 T Tgc 598t Fixation of one cell: Tc Tg 60 Ngc N 180 Tc Tgc 18 N Ngc 189 Fixation of both cells: Nc Ng 31 Tgc Tg 4 Ng Nc 25 Tg Tgc 10 Nc Ngc 15 Fixation of both cells: * T, tumor lymphoma cell; Tc, tumor cell coated with Con A; Ngc Ng 10 Tg, tumor cell treated with glutaraldehyde; Tgc, tumor cell Ng Ngc 6 treated first with glutaraldehyde and then with Con A. t Expressed as the number of cell-bound cells observed along * N, normal lymphocyte; Nc, normal cell coated with Con A; one edge of a 1-cm fiber segment. The standard deviation of 4 Ng, normal cell treated with glutaraldehyde; Ngc, normal cell determinations was + 1Q%. treated first with glutaraldehyde and then with Con A. t Values are beyond the linear range of the assay. Extrapolation t Expressed as the number of cell-bound cells observed along indicates a value of about 650 cells per cm. one edge of a 1-cm fiber segment. The standard deviation of 4 determinations was 4-10%. bound tumor cells had 650-700 CB cells along one edge of a 1- cm fiber segment. The number of cell-bound cells was pro- portional to the number of free cells added, up to densities of cell gave intermediate values (Fig. 3). In experiments on 450 CB cells per cm. Normal lymphocytes in suspension were binding to a common cell type, unfixed tumor cells gave 2.1- hardly agglutinated by Con A, so that these cells could be 4.8 times more binding than unfixed normal cells, whereas coated with Con A in suspension, washed, and then bound to fixed tumor cells gave 1.3-1.7 times more binding than fixed FB cells. Similar results were obtained if either the FB or normal cells (Table 3). A Treatment with metabolic inhibitors can increase the the free cells were coated with Con (Tables 1-3). Capping of A of receptors was not observed at the fiber-cell or cell-cell in- agglutinability normal and tumor cells by Con (13, 14). terfaces (Fig. 2), No cell-cell binding was observed if both the Cell-cell binding between two tumor cells and between a free and the FB cells were not coated with Con A or if both tumor cell and a normal cell was also increased about 2-fold were with A. The cells did not form by the presence of NaN3 or 2,4-dinitrophenol. Cell-cell cells coated Con aggre- normal not the with bound to beads derivatized binding between lymphocytes, however, was gates of type found cells changed by metabolic inhibitors, and this may reflect mem- with D-galactose (12). brane changes induced by binding of the lymphocytes to Effect of Membrane Fixation on Cell-Cell Binding. Fixation Con A fibers (7). Treatment of cells with trypsin had no of cells inhibits Con A-induced agglutination and slightly re- effect on cell-cell binding of normal lymphocytes and only duces the number of Con A molecules that will bind to the slightly increased the binding between the tumor cells; treat- cell surface (10). Fixed cells did not bind to Con A.derivatized ment with neuraminidase had no effect on the binding of fibers. Fixation of FB cells was, therefore, carried out after either cell type. the cells were bound to the fiber. The effects of fixation on cell- Cell-Cell and Cell- cell binding are shown in Tables 1 and 2. When both the free Reversibility of Binding Agglutination. cells cell binding or agglutination induced by Con A after a 30- and the fiber-bound were treated with glutaraldehyde, min incubation was not reversed by incubation with 0.01- cell-cell binding was not observed. If, however, only one of the of irreversible its to 0.3M aMM for 30 min at 250C. The formation cells was fixed and then coated with Con A, ability bind was not the presence of in- was even than of an unfixed cell bonds prevented by metabolic untreated cells greater that to cell-cell Con A can were in solution hibitors during and subsequent binding. coated with Con A. Similar results obtained bind residues in the of two without addition of solu- to saccharide dextran polymer Sephadex by mixing the cell types further were by Con A at A. The with Con A of the fixed but not the beads (15). These beads agglutinated high ble Con coating concentrations of lectin and in contrast to unfixed cell was critical, in that incubation of fixed cells with (1-20 mg/ml), A-coated cells less cells, the agglutination of the beads was rapidly reversed by unfixed Con gave considerably binding. aMM. Experiments with mouse 3T3 fibroblasts transformed by simian virus 40 gave similar results to those obtained with the DISCUSSION cells. lymphoma Cells attached to nylon fibers (3, 4) have been used to analyze Differences in the Cell-Cell Binding of Normal and Tumor the mechanism of Con A-induced binding between single cells. Cells. At all concentrations of Con A used there was 7-10 Cell-cell binding induced by a lectin has been divided into times' higher binding between two tumor cells than between two steps, the binding of the lectin to a cell and the binding twoteofnori.al cells. Binding between a tumor cell and a normal one cell to another cell. One or both cells can be coated with Downloaded by guest on September 27, 2021 Proc. Nat. Acad. Sci. USA 71 (1974) Cell-Cell Binding Induced by Concanavalin A 2459

TABLE 3. Comparison of normal-and tumor cells in celU-cel binding induced by Con A Binding with Cell-cell tumor cell Fiber-bound Cell-bound binding Binding with cell cell (cm-l)* normal cell Comparison between unfixed ceUst Tc N 173 4.3 Nc N 40 T Nc 134 4.8 N Nc 28 T Tgc 598t 2.6 FIG. 4. Model for the mechanism of cell-cell binding induced N Tgc 251 by Con A. Two cell surfaces with receptors for Con A come into contact. One cell has been coated with Con A, so that it has T Ngc 435 2.4 binding sites for the Con A receptors on the other cell. The N Ngc 185 receptors on one or both cells are mobile and can move into a Tc T 356 2.1 position favorable for the formation of bridges between indi- Tc N 173 vidual pairs of receptors. After multiple bridge formation has Comparison between fixed cellst taken place, the cells become irreversibly bound by the formation of other bonds between the cells. E , receptor for Con A; Tgc N 299 1.7 I, multivalent Con A molecule; irreversible bonds. Ngc N 178 MA, N TOc 249 1.3 N Ngc 186 be higher on normal lymphocytes than on lymphoma cells, T Tgc 595t 1.5 and treatment of cells with trypsin increased this mobility in T Ngc 434 both types of cells to a similar value (18). Treatment with trypsin did not change the differences in cell-cell binding be- See Tables 1 and 2 for abbreviations. tween the two cell types. Although some rotation of the re- * Nonspecific binding (0-12 cells per cm) has been subtracted. ceptors may be required, changes in rotational mobility over t Cells being compared are in italics. the ranges described (18), therefore, do not have a large effect t Values are beyond the linear range of the assay. Extrap- on cell-cell binding. olation indicates a value of 650 cells per cm, which has been The inhibition of cell-cell binding by fixation of both the used to calculate the tumor cell per normal cell binding ratio. fiber-bound and cell-bound cells indicates that cell-cell binding does require receptor mobility. Our experiments also show lectin and binding can be studied between different types of that fixed cells coated with Con A can bind and agglutinate cells. The cells can be recovered by plucking the fibers (3, 6). with unfixed cells that have not been coated with Con A. Pre- This method should be generally applicable to the analysis of fixed cells have been shown by electron microscopy to have a cell-cell interactions in immunology, developmental biology, diffuse distribution of Con A-binding sites (11, 19). It there- and neurophysiology. fore appears that clustering of receptors is not a requirement Cell-cell binding between a Con A-coated cell and an un- for cell-cell binding or agglutination. It is also unlikely that treated cell was rarely observed between normal lymphocytes cluster formation is required after cell-cell contact, because the but occurred with a high frequency between lymphoma tumor Con A molecules are held in a diffuse distribution by the fixed cells. This observation is in agreement with the higher degree cell. The observation that prefixation of Con A-coated cells of Con A agglutinability of the tumor cells than the normal increases their ability to bind unfixed cells suggests that clus- cells in suspension (9). Normal cells also bound better to tering of receptors by Con A can actually hinder cell-cell tumor cells than to normal cells. There was, however, no binding. This conclusion is also supported by the finding that difference between the normal and tumor cells in the number binding was inhibited if the Con A was on the unfixed instead of Con A-binding sites per unit cell protein (9) or in their of the fixed cell. The enhancement of cell-cell binding by ability to bind to Con A-derivatized fibers. metabolic inhibitors is consistent with the suggestion that It has been suggested that mobility of membrane receptors cellular ATP levels affect the stability of the plasma membrane for Con A is required for Con A-induced cell agglutination (10, (13, 20) and the mobility of Con A receptors (13). 14, 16) . Three types of movements of Con A receptors on the The fiber experiments suggest a model for the mechanism of cell surface have been described: large lateral movements that Con A-induced cell-cell binding (Fig. 4). This model requires result in cap formation (9, 17), molecular rotation (18), and short-range lateral movement of Con A receptors, over the small lateral movements than can result in clustering of average distance between two receptors on the same cell, so receptors (10, 11, 14, 16, 17, 19). No cap formation was ob- that Con A molecules bound to one cell can find unoccupied served on cells bound to fibers and capping was not induced by receptors on the other cell. Mobility of receptors is necessary, the binding of one cell to another. This result indicates that therefore in only one of the two cells, and at least some of the cell-cell binding does not require cap formation by Con A receptors must be unoccupied by lectin molecules before cell- receptors and that the low cell-cell binding between normal cell contact. The speed of receptor movement is important be- lymphocytes was not due to cap formation induced by Con A. cause alignment must occur before the colliding cells separate. The rotational mobility of Con A receptors has been found to In contrast to previous suggestions (10, 14, 16, 21), the cluster- Downloaded by guest on September 27, 2021 2460 Cell Biology: Rutishauser and Sachs Proc. Nat. Acad. Sci. USA 71 (1974)

ing of receptors is not required for cell-cell binding and, in fact, technical assistance. U.R. is a Fellow of the Jane Coffin Childs may hinder receptor alignment by reducing mobility (17) and Memorial Fund for Medical Research. blocking receptor sites. The ability of receptors on a cell to 1. Sachs, L. (1974) Harvey Lectures (Academic Press, New cluster after Con A binding, therefore, need not correlate York), Vol. 68, in press. with that cell's capacity for cell-cell binding. 2. Sachs, L., Inbar, M. & Shinitzky, M. (1974) Control of Proliferation in Animal Cells (Cold Spring Harbor Lab., Our alignment model also involves the formation of ir- Cold Spring Harbor, New York), in press. reversible bonds between the cells after alignment of receptors 3. Edelman, G. M., Rutishauser, U. & Millette, C. F. (1971) and multi-point bridging by Con A molecules. These irrevers- Proc. Nat. Acad. Sci. USA 68, 2153-2157. ble bonds, which also occur between cells and fibers (6), can be 4. Edelman, G. M. & Rutishauser, U. (1974) Methods in formed in the presence of metabolic inhibitors. The complete Enzymology, eds. Jakoby, W. B. & Wilchek, M. (Academic Press, New York), Vol. 34, in press. dissociation of Con A-aggregated Sephadex beads by aMM 5. Rutishauser, U., Millette, C. F. & Edelman, G. M. (1972) suggests that multi-point bridging by itself is not sufficient Proc. Nat. Acad. Sci. USA 69, 1596-1600. to cause irreversibility. The possibility that the cell-cell bind- 6. Rutishauser, U., d'Eustachio, P. & Edelman, G. M. (1973) ing is initiated only by these irreversible interactions induced Proc. Nat. Acad. Sci. USA 70, 3894-3899. 7. Rutishauser, U., Yahara, I. & Edelman, G. M. (1974) by Con A is unlikely, because binding did not occur between Proc. Nat. Acad. Sci. USA, in press. two cells both of which had been coated with Con A. 8. Klein, E. & Klein, G. (1964) J. Nat. Cancer Inst. 32, 547-568. Differences in the ability for the alignment of receptors can 9. Inbar, M., Ben-Bassat, H. & Sachs, L. (1973) Int. J. explain differences in cell-cell binding of normal and tumor Cancer 12, 93-99. cells. The fluidity of lipids in the surface membrane of the 10. Inbar, M., Huet, C., Oseroff, A. R., Ben-Bassat, H. & Sachs, L. (1973) Biochim. Biophys. Acta 311, 594-599. lymphoma cells is greater than that of the normal lymphocytes 11. de Petris, S., Raff, M. C. & Mallucci, L. (1973) Nature (22) and this is associated with a higher cholesterol to phos- New Biol. 244, 275-278. pholipid ratio in the normal lymphocytes (23). Assuming 12. Chipowsky, S., Lee, Y. C. & Roseman, S. (1973) Proc. Nat. that Con A receptors include structures that penetrate into Acad. Sci. lTSA 70, 2309-2312. 13. Vlodavsky, I., Inbar, M. & Sachs, L. (1973) Proc. Nat. Acad. the lipid phase of the membrane, receptor alignment could be Sci. USA 70, 1780-1784. facilitated by this higher lipid fluidity. Cell-cell binding might 14. Inbar, M. & Sachs, L. (1973) FEBS Lett. 32, 124-128. also be influenced by surface charge and cell size. The bind- 15. Agrawal, B. B. L. & Goldstein, I. J. (1965) Biochemistry 96, ing of fixed to unfixed normal lymphocytes and the lack of 23-28. effect of neuraminidase treatment suggest that charge re- 16. Nicolson, G. L. (1973) Nature New Biol. 243, 218-220. 17. Yahara, I. & Edelman, G. M. (1973) Exp. Cell Res. 81, pulsion alone cannot explain the low binding obtained between 143-155. normal cells. The differences in binding of normal and tumor 18. Inbar, M., Shinitzky, M. & Sachs, L. (1973) J. Mol. Biol. cells was decreased by membrane fixation. This finding sug- 81, 245-253. gests that these differences involve receptor alignment and are 19. Rosenblith, J. Z., Ukena, T. E., Yin, H. H., Berlin, R. D. & Karnovsky, M. J. (1973) Proc. Nat. Acad. Sci. USA 70, not only due to the smaller size of the normal lymphocytes. It, 1625-1629. therefore, appears that receptors on tumor cells have a greater 20. Yanovsky, A. & Loyter, A. (1972) J. Biol. Chem. 247, ability than receptors on normal cells to align themselves for 4021-4028. the formation of cell-lectin-cell bridges. 21. Noonan, K. D. & Burger, M. M. (1973) J. Cell Biol. 59, 134-142. This work was supported by Contract no. NoI-CP43241 with 22. Inbar, M., Shinitzky, M. & Sachs, L. (1973) FEBS Lett. 38, the Virus Cancer Program of the National Cancer Institute, 268-270. USPHS. We are indebted to Miss Ziva Misolovin for skillful 23. Vlodavsky, I. & Sachs, L. (1974) Nature, in press. Downloaded by guest on September 27, 2021