Proc. Nati. Acad. Sci. USA Vol. 73, No. 12, pp. 4575-4579, December 1976 Cell Biology
Concanavalin-A-induced transmembrane linkage of concanavalin A surface receptors to intracellular myosin-containing filaments (transmembrane control/membrane receptor mobility/nonmuscle cell myosin)
J. F. ASH AND S. J. SINGER Department of Biology, University of California at San Diego, La Jolla, Calif. 92093
Contributed by S. J. Singer, September 28, 1976
With normal rat kidney cells in monolayer ABSTRACT culture, we have studied the distribution on the cell surface of receptors for concanavalin A, and the distribution of the smooth muscle myosin-like protein inside the same cell, using specific fluorescence microscopic methods. The concanavalin A receptors were initially uniformly dispersed over the cell surface, but 20 min after the addition of concanavalin A at 370, the receptors showed a variety of nonuniform surface distributions, including extended parallel linear arrays. These arrays of receptors were found to be superimposed on the linear arrays of the intracellular myosin-containing filaments, indicating that a transmembrane linkage of the receptors and the filaments had occurred. This linkage required a lateral redistribution of concanavalin A receptors, since it did not occur with succinylated concanavalin A, but was subsequently induced if the cells that had been reacted with succinylated concanavalin A were then treated with antibodies to concanavalin A. The redistributions of concanavalin A receptors on the surfaces of these normal rat kidney cells, however, were much less extensive than the patching that was induced on the surfaces of the same cells infected with, and transformed by, Rous sarcoma virus.
The concept of transmembrane control-that cell surface molecules can interact with structural elements in the cell interior-has been the subject of much experimentation and discussion (for reviews, see refs. 173). Almost all of the experimental evidence supporting this concept, however, is indirect, much of it involving inferences drawn from the effects produced by drugs such as colchicine and cytochalasin B on the lateral mobility of cell surface receptors. In this paper, direct evidence is provided for such transmembrane interactions in a line of normal rat kidney (NRK) cells in monolayer culture. We have studied the distribution on the cell surface of receptors for concanavalin A (Con A) and of the smooth muscle myosin-like protein inside the cell (4-6). These distributions were determined simultaneously on the same cell by fluorescence microscopy, using fluorescein-conjugated Con A(FI-Con A) to detect the Con A receptors, and an indirect rhodamine immunofluorescence method for the myosin. Initially the Fl-Con A was found to be uniformly dispersed over the cell surface, while the intracellular myosin was organized into extended filamentous structures. After 20 min at 370, however, the FlCon A showed a variety of nonuniform distributions, ranging from a mottled pattern to extended parallel linear arrays. In the latter case, the linear arrays of Con A receptors were superimposable on the linear arrays of the intracellular myosin-containing filaments, strongly suggesting that the receptors and the filaments had become physically linked. These observations are relevant to our recently published comparative study (7) of the distributions of Con A receptors and myosin-containing structures in NRK cells infected with Rous sarcoma virus. Abbreviations: Con A, concanavalin A; Fl-Con A, fluorescein-conjugated Con A; Suc-Con A, succinylated Con A; NRK, normal rat kidney.
4575
MATERIALS AND METHODS Cell Culture. NRK cells (8) were cultured in Ham's F12 based medium [100 ml of F12, 10 ml of Tryptose Phosphate Broth (Difco Lab, Detroit, Mich.), 5 ml of fetal calf serum, and antibiotics] under an atmosphere of 10% CO2 and 90% air at 370. Cells were dissociated with trypsin for plating onto cover slips (15,000-30,000 cells per 35-mm dish) and cultured for at least 2 days before an experimental treatment. NRK cells infected with the B77 strain of Rous sarcoma virus were used in our earlier studies (7). Protein Modification. Fl-Con A was prepared as described (7). Succinylated Con A (Suc-Con A) was a single-stage modified product (9) prepared by Dr. Randy Schekman and characterized elsewhere (10). Rhodamine B conjugation of goat antibody against rabbit IgG was done by the method of Fothergill (11), who kindly supplied the rhodamine sample and assisted in making the conjugate. Cell Staining. For Con A and Suc-Con A staining, coverslip cultures in 35-mm dishes were first rinsed three times with serum-free Dulbecco's modified Eagle's medium, then incubated with 1 ml of the medium containing 50 ug of either FlCon A or Suc-Con A for 20 min at 37°. After being rinsed three times, they were then fixed with 2% formaldehyde in phosphate-buffered saline at room temperature for 20 min (7). In order to determine the effect of antibodies against Con A on the distribution of surface-bound Suc-Con A, some cultures that had been treated with Suc-Con A were not fixed, but were reacted for an additional 20 min at 370 with 1 ml of 100 gtg/ml of a rabbit IgG fraction containing antibodies against Con A in modified Eagle's medium, and were then rinsed and fixed. After fixation, the cells were rinsed with several changes of phosphate-buffered saline for 20 min, one change containing 50 mM NH4Cl to quench any unreacted aldehyde groups. To visualize the distribution of Suc-Con A, cells that had been reacted with Suc-Con A and were then fixed were treated successively with rabbit antibodies against Con A and fluorescein-conjugated goat antibodies against rabbit IgG; whereas cells that had been reacted with Suc-Con A and rabbit antibodies against Con A, and were then fixed, were treated only with the fluorescein-conjugated goat antibodies. If myosin or microtubules were to be stained, the cells were then frozen and thawed to render them permeable to antibodies. Indirect immunofluorescent staining of myosin with rabbit antibody against human uterine myosin or of microtubular protein with rabbit antibody against chick brain tubulin (a generous gift of Dr. Melvin Simon) was done as described (7), except that rhodamine-conjugated goat antibody against rabbit IgG was used instead of the fluorescein conjugate. In all of these experiments the modified Eagle's medium used in ligand binding and for rinsing was first equilibrated with 10% CO2 at 37°. Colchicine treatment was effected by adding 10-5
4576
Cell Biology: Ash and Singer
Proc. Nati. Acad. Sci. USA 73 (1976)
FIGS. 1-4. (Legend appears at bottom of the following page.)
Cell Biology: Ash and Singer
Proc. Nati. Acad. Sci. USA 73 (1976)
4577
M colchicine in growth medium to the cells, and continuing culture at 370 for 1-2 hr before any other additions were carried out. Colchicine (10-5 M) was included in subsequent rinse and treatment solutions in these cases. Photomicroscopy and Analysis. Cells were observed with a Zeiss Photomicroscope using a 40 X oil immersion lens and photographed onto Kodak 35 mm Plus X film developed with Kodak Microdol X developer diluted 1:3. Fluorescence was excited with selected emissions from an Osram 200-W HBO lamp filtered through either a Zeiss FITC interference filter or a BG-12 excitor filter. When cells were stained with both Fl-Con A and rhodamine-IgG, the fluorescein fluorescence was photographed first using one or two Wratten no. 58 filters above the objective to absorb the red rhodamine emission. Then, without changing the focus, the rhodamine fluorescence was photographed using one Wratten no. 23A filter to absorb the fluorescein emission. The fluorescein conjugate was significantly bleached during the first exposure so that only one barrier filter was needed to prevent leakage into the rhodamine exposure.
In appropriate instances, a pair of fluorescein- and rhodamine-stained negatives of the same cell were reproduced as enlarged positive transparencies, and these were superimposed visually. For a more objective examination of the possible superposition of the two images, the positive transparencies were analyzed with an 12S Mini-Addcol Viewer, model 6000 (International Imaging Systems, Mountain View, Calif.), available at the Visibility Laboratory of the Scripps Institute of Oceanography. This instrument superimposes up to four individual images in register to display a composite image. The assistance of Mr. James Harris in performing this analysis is gratefully
acknowledged. RESULTS Myosin Patterns. The intracellular myosin staining patterns revealed by indirect immunofluorescence in normal NRK cells 2 days after plating were predominantly filamentous (7). Along with the filaments there were present sheet-like or striped patterns (Fig. la), which we described earlier (7). Treatment of cells with Con A or with colchicine had no obvious effects on these patterns. Con A Patterns. Shortly after the addition of Fl-Con A to the cells, the fluorescence was uniformly dispersed over the cell surface with no discernible structure (Fig. lb). After 20 min of incubation at 370, however, careful observation revealed that the surface distribution of Fl-Con A was no longer completely uniform on most cells. A mottled distribution of fluorescence was commonly observed (Fig. 2b). Furthermore, on about 10% of the cell surfaces, the fluorescence was seen in extended linear arrays (Figs. 2b and 3b). The mottled and linear patterns were often seen on different areas of the same cell surface (Fig. 2b). Photographing such linear arrays was difficult if the cell was rounded because only limited areas could be in focus at one time. These patterns were the same whether or not staining for intracellular myosin was carried out on the cells.
FIG. 5. The distribution of Con A receptors on NRK cells treated with Suc-Con A for 20 min at 370 and subsequently with rabbit antibodies to Con A for 20 min at 37°. The distribution of the receptors after the treatment with Suc-Con A alone is not shown, but was uniform as in Fig. lb. The incubation with the antibody redistributed the receptors into mottled patterns and also into extended linear arrays (arrows) much like those of Figs. 2b and 3b. These treatments caused the cells to round up, allowing only limited regions of the cell surface to be simultaneously in focus. X925.
For comparison, NRK cells infected with Rous sarcoma virus (B77-NRK cells), when treated with Fl-Con A for 20 min at 370, showed a very marked patching of fluorescence (Fig. 4b), as we described earlier for transformed cells (7). The myosin staining patterns were largely disorganized compared to normal NRK cells (7). Simultaneous Observation of Fl-Con A and Myosin. When cells treated with Fl-Con A for 20 min were then stained for myosin, in those cases where extended linear arrays of the Fl-Con A were observed, there was a remarkable similarity between these arrays and the filamentous myosin patterns. Two examples of this are presented in Figs. 2 and 3. Extensive portions of the Fl-Con A and of the myosin arrays were superimposable by visual inspection, as was confirmed by analysis in the 12S Mini-Addcol Viewer. Effects of Sue-Con A. Treatment of cells with Suc-Con A for 20 min at 370 resulted in a uniform distribution of this ligand on the cell surface. No mottling or linear arrangements were seen. However, if the cells treated with Suc-Con A for 20 min were then further reacted with rabbit antibodies against Con A for 20 min, there was significant rounding of the cells which was accompanied by a redistribution of the Suc-Con A as detected by immunofluorescence. The severe rounding of the cells made observation difficult, but mottled and linear patterns could now be seen on these cell surfaces (Fig. 5, arrows). Effects of Colchicine and Cytochalasin B. Colchicine treatment (10-5 M) of NRK cells for 1-2 hr eliminated the
FIGS. 1-4 (on preceding page). Figs. 1-3. Fluorescence patterns of the distribution of (a) the intracellular smooth muscle myosin-like protein (as visualized by rhodamine immunofluorescence) and (b) the surface receptors for Con A (as visualized by Fl-Con A) on the same NRK cell in each a, b pair. (Fig. 1) The patterns are those observed initially. (Figs. 2 and 3) Two examples of patterns observed after 20-min incubation with the Fl-Con A at 37°. The intracellular myosin is mostly arranged in extended filamentous arrays both initially (la) and after the 20-min incubation (2a and 3a). The Con A receptors are initially (lb) dispersed uniformly over the cell surface, but 20 min after the addition of Fl-Con A (2b and 3b), the receptors are no longer uniformly dispersed, but show a mottled distribution (M in 2b) or extended linear arrays (arrows). The linear arrays of myosin (3a) and of Con A receptors (3b) are mostly superimposable. In each a, b pair, the photographs were made by changing filters without changing focus. X600. Fig. 4. The distribution of surface receptors for Con A on NRK cells transformed with Rous sarcoma virus (B77-NRK): (a) initial distribution; (b) 20 min after incubation with Fl-Con A at 37°. In (b), extensive patching of the receptors on the rounded-up cells is evident. X600.
4578
Cell Biology: Ash and Singer
majority of the microtubular system, as judged by indirect immunofluorescent staining with rabbit antibodies against chick brain tubulin (not shown). Colchicine treatment produced cells with a uniform antitubulin staining pattern, with occasional cells showing a few fluorescent curvilinear structures originating near the nucleus, as has been described by other workers (12, 13). This treatment did not eliminate the linear arrangements of Fl-Con A on NRK cells present after 20 min of incubation shown in Figs. 2b and 3b. If cytochalasin B (10 Atg/ml) was added with Fl-Con A before fixation many of the NRK cells were found to have a reduced amount of fluorescence on their surface. If cells were first fixed, however, and then cytochalasin B and Fl-Con A were used together to stain the fixed cells, the level of Fl-Con A binding was indistinguishable from that seen with no cytochalasin B added. Thus, the treatment of unfixed NRK cells with both cytochalasin B and Fl-Con A produced an apparent loss of Con A binding sites on the surfaces of these cells. It is therefore not possible to assess the effect of cytochalasin B on the Fl-Con A induced aggregations without further study. Myosin staining patterns also appeared to be altered in some cells after a 20-min treatment with cytochalasin B, some cells showing masses of fluorescence near their edges. DISCUSSION The binding of Con A to the surface of an NRK cell induced the Con A receptors, originally dispersed uniformly over the surface, to redistribute to a limited extent. The distribution that resulted ranged in orderliness from a mottled pattern (Fig. 2b) to a highly structured one consisting of extended linear arrays (Figs. 2b and 3b). These two patterns, often seen on adjacent portions of the same cell surface, varied in relative proportion from cell to cell in a culture. This variability is not due to cellular heterogeneity in the NRK line. It may reflect metabolic differences in the cell population, such as cell cycle differences; or, as we are more inclined to think, it may reflect different states of perfection of the same ordering process. In cases where the extended linear arrays of Con A receptors were especially clearly defined, they lined up precisely over the myosin-containing filaments inside the cell. This strongly suggests that under these circumstances, repeating physical linkages were formed which extended from the Fl-Con A molecules, at the cell surface, through its receptors in the membrane, to components in the myosin-containing filaments inside the cell. These linkages may include other intervening components as well. The extended linear arrangements of Con A receptors arose only after tetravalent Con A was specifically bound to the cell surface. Divalent Suc-Con A did not serve this function, but if antibodies against Con-A were added to the cells treated with Suc-Con A, the linear arrangement of receptors was induced (Fig. 5). This suggests that some surface redistribution of the Con A receptors, and not just ligand attachment to them, is critical for the transmembrane linkage to be formed. Microtubules are not involved in this phenomenon. The addition of lo-5 M colchicine, which disrupted the microtubule assemblies, had no discernible effect on the myosin-containing filaments, nor on the Con A-induced linkage of Con A receptors to the filaments. A number of mechanisms can be entertained to explain these effects. The Con A-induced clustering of the Con A receptors might induce some enzymatic or permeability changes in the membrane; altered concentrations of some enzyme product or permeant species might then affect the characteristics of the myosin-containing filaments so that they become linked to the
Proc. Natl. Acad. Sci. USA 73 (1976)
receptors. Alternatively, the clustering of the receptors might change the properties of the receptor molecules, allowing them to become linked to the filaments. There is insufficient information available at present to discriminate between these or other possible mechanisms. These limited redistributions of Con A receptors that are induced on the surfaces of NRK cells are to be contrasted with the much more profound Con A-induced redistributions of receptors that occur on transformed cells (7, 14-18). Such redistributions result in large patches of Fl-Con A on the cell surface. An example of such patching is shown in Fig. 4b, 20 min after the addition of Con A to an NRK cell infected with the B77 strain of Rous sarcoma virus. In a companion study (7) with NRK cells infected with a temperature-sensitive mutant (LA 23) of Rous sarcoma virus, it was found that there was a correlation between the integrity of the myosin-containing filaments and the mobility of Con A receptors on the cell surface. At permissive temperatures, the infected cells exhibited the transformed phenotype: the intracellular myosin filaments were disrupted and the Con A receptors could be clustered into large patches upon the addition of Con A. At nonpermissive temperatures, the myosin-containing filaments were intact, and Con A could not induce a patching of its receptors. The results of the present study strongly support the conclusion reached earlier (7), namely, that the mobility of Con A receptors in the
plane of the membrane is physically determined by the integrity of the myosin-containing filaments in these infected NRK cells. Do these transmembrane linkage effects occur with Con A receptors on cells other than the NRK line, or with other ligand-cell interactions? It is possible that a very similar phenomenon arises upon binding Con A to cultured rat myoblasts (19). In these studies, the lateral mobility of Fl-Con A on the surface of the myoblasts was measured by photobleaching experiments, in which the rate of recovery of fluorescence was observed in a small region of the cell surface that was initially photobleached. Among the observations made were that: (i) immediately after the attachment of the fluorescent Con A to the cell surface, the lateral mobility of the Con A at 230 was appreciable, but decreased markedly by 20 min later; (ii) Suc-Con A had a much larger lateral mobility than Con A after 20 min of attachment; and (iii) colchicine had no effect on these phenomena. These results can be explained, if, by analogy to our observations with NRK cells, the Con A receptors were originally freely mobile in the membrane, but upon binding tetravalent Con A, became redistributed and linked over a period of 20 min to myosin-containing filament networks underneath the membrane. In other studies with lymphocytes (20,21) it has been shown that relatively large amounts Con A bound to the cell surface inhibit the capping that is produced by smaller amounts of Con A. These large amounts of Con A also inhibit the capping of several other receptors which is produced by their respective antibodies. These immobilization effects are fairly specific for Con A, and do not occur with the other receptor-ligand interactions. These effects may involve a transmembrane linkage of Con A receptors to myosin-containing intracellular filaments that is induced upon binding sufficiently large amounts of Con A to the lymphocyte surface. In this case, however, these receptor immobilizations are reversed by the addition of colchine, and as a result, microtubules have been implicated in these lymphocyte surface effects. It may be that the intracellular myosin-containing filaments in lymphocytes are less extensive than in the NRK cells, and are in some manner linked together by microtubules. It is interesting to note that long linear arrays
Cell Biology: Ash and Singer of transmembrane linkages could produce extended "picketfence" structures in the cross section of the membrane, consisting of regularly spaced receptor molecules spanning the membrane (see Fig. 2b in ref. 5). Those membrane molecules that were incapable of penetrating the interstices in the "picket-fence" would be prevented from large-scale lateral movement in the membrane even if they were not bound to the Con A receptors. This mechanism might account in part for the apparent immobilization of many other receptors in the lymphocyte surface that is induced by Con A binding to the cell. Transmembrane linkage phenomena of the kind we have described may therefore be characteristic of Con A receptors on a variety of different cells. Whether any other kinds of cell surface receptors act similarly remains to be determined. We gratefully acknowledge the use of tissue culture facilities generously provided by Dr. Immo Scheffler. The NRK and B77-NRK cells were from stocks originally provided by Dr. Peter K. Vogt. J.F.A. is a USPHS Postdoctoral Fellow, 1974 to present. S.J.S. is a American Cancer Society Research Professor. Studies supported by USPHS Grants AI-06659 and GM-15972. 1. Singer, S. J. (1974) "The molecular organization of membranes," Annu. Rev. Biochem. 43, 805-833. 2. Nicolson, G. L. (1976) "Transmembrane control of the receptors on normal and tumor cells. I. Cytoplasmic influence over cell surface components," Biochim. Biophys. Acta 457, 57-108. 3. Edelman, G. M. (1976) "Surface modulation in cell recognition and cell growth," Science 192, 218-226. 4. Weber, K. & Groeschel-Stewart, U. (1974) "Antibody to myosin: The specific visualization of myosin-containing filaments in non-muscle cells," Proc. Natl. Acad. Sci. USA 71, 4561-4564. 5. Painter, R. G., Sheetz, M. & Singer, S. J. (1975) "Detection and ultrastructural localization of human smooth muscle myosin-like molecules in human non-muscle cells by specific antibodies," Proc. Natl. Acad. Sci. USA 72, 1359-1363. 6. Wang, K., Ash, J. F. & Singer, S. J. (1975) "Filamin, a new high-molecular-weight protein found in smooth muscle and non-muscle cells," Proc. Natl. Acad. Sci. USA 72, 4483-4486. 7. Ash, J. F., Vogt, P. K. & Singer, S. J. (1976) "The reversion from transformed to normal phenotype by inhibition of protein synthesis in rat kidney cells infected with a temperature-sensitive Rous sarcoma virus mutant," Proc. Natl. Acad. Sci. USA 73, 3603-3607. 8. Duc-Nguyen, H., Rosenblum, E. N. & Zeigel, R. F. (1966) "Persistent infection of a rat kidney cell line with Rauscher
Proc. Natl. Acad. Sci. USA 73 (1976)
4579
murine leukemia virus," J. Bacteriol. 92, 1133-1140. 9. Gunther, G. R., Wang, J. L., Yahara, I., Cunningham, B. A. & Edelman, G. M. (1973) "Concanavalin A derivatives with altered biological activities," Proc. Natl. Acad. Sci. USA 70, 10121016. 10. Schekman, R. & Singer, S. J. (1976) "Clustering and endocytosis of membrane receptors can be induced in mature erythrocytes of neonatal humans but not adults," Proc. Natl. Acad. Sci. USA 73,4075-4079. 11. Fothergill, J. E. (1969) "Properties of conjugated serum proteins," in Fluorescent Protein Tracing, ed. Nairn, R. C. (Livingstone, London), 3rd ed., pp. 35-0. 12. Brinkley, B. R., Fuller, G. M. & Highfield, D. P. (1975) "Cytoplasmic microtubules in normal and transformed cells in culture: Analysis by tubulin antibody immunofluorescence," Proc. Natl. Acad. Sci. USA 72,4981-4985. 13. Osborn, M. & Weber, K. (1976) "Cytoplasmic microtubules in tissue culture cells appear to grow from an organizing structure towards the plasma membrane," Proc. Nati. Acad. Sci. USA 73, 867-871. 14. Rosenblith, J. Z., Ukena, T. E., Yin, H. H., Berlin, R. D. & Karnovsky, M. J. (1973) "A comparative evaluation of the distribution of concanavalin A-binding sites on the surfaces of normal, virally-transformed, and protease-treated fibroblasts," Proc. Nati. Acad. Sci. USA 70, 1625-1629. 15. Nicolson, G. L. (1973) "Temperature-dependent mobility of concanavalin A sites on tumour cell surfaces," Nature New Biol. 243,218-220. 16. Noonan, K. D. & Burger, M. M. (1973) "The relationship of concanavalin A binding to lectin-initiated cell agglutination," J. Cell Biol. 59, 134-142. 17. Inbar, M. & Sachs, L. (1973) "Mobility of carbohydrate containing sites on the surface membrane in relation to the control of cell growth," FEBS Lett. 32, 124-128. 18. Edelman, G. M. & Yahara, I. (1976) "Temperature-sensitive changes in surface modulating assemblies of fibroblasts transformed by mutants of Rous sarcoma virus," Proc. Natl. Acad. Scd. USA 73,2047-2051. 19. Schlessinger, J., Koppel, D. E., Axelrod, D., Jacobson, K., Webb, W. W. & Elson, E. L. (1976) "Lateral transport on cell membranes: Mobility of concanavalin A receptors on myoblasts," Proc. Natl. Acad. Sci. USA 73,2409-2413. 20. Yahara, I. & Edelman, G. M. (1972) "Restriction of the mobility of lymphocyte immunoglobulin receptors by concanavalin A," Proc. Natl. Acad. Sci. USA 69,608-612. 21. Yahara, I. & Edelman, G. M. (1975) "Modulation of lymphocyte receptor mobility by locally bound concanavalin A," Proc. Natl. Acad. Sci. USA 72, 1579-1583.
Concanavalin-A-induced transmembrane linkage of concanavalin A surface receptors to intracellular myosin-containing filaments (transmembrane control/membrane receptor mobility/nonmuscle cell myosin)
J. F. ASH AND S. J. SINGER Department of Biology, University of California at San Diego, La Jolla, Calif. 92093
Contributed by S. J. Singer, September 28, 1976
With normal rat kidney cells in monolayer ABSTRACT culture, we have studied the distribution on the cell surface of receptors for concanavalin A, and the distribution of the smooth muscle myosin-like protein inside the same cell, using specific fluorescence microscopic methods. The concanavalin A receptors were initially uniformly dispersed over the cell surface, but 20 min after the addition of concanavalin A at 370, the receptors showed a variety of nonuniform surface distributions, including extended parallel linear arrays. These arrays of receptors were found to be superimposed on the linear arrays of the intracellular myosin-containing filaments, indicating that a transmembrane linkage of the receptors and the filaments had occurred. This linkage required a lateral redistribution of concanavalin A receptors, since it did not occur with succinylated concanavalin A, but was subsequently induced if the cells that had been reacted with succinylated concanavalin A were then treated with antibodies to concanavalin A. The redistributions of concanavalin A receptors on the surfaces of these normal rat kidney cells, however, were much less extensive than the patching that was induced on the surfaces of the same cells infected with, and transformed by, Rous sarcoma virus.
The concept of transmembrane control-that cell surface molecules can interact with structural elements in the cell interior-has been the subject of much experimentation and discussion (for reviews, see refs. 173). Almost all of the experimental evidence supporting this concept, however, is indirect, much of it involving inferences drawn from the effects produced by drugs such as colchicine and cytochalasin B on the lateral mobility of cell surface receptors. In this paper, direct evidence is provided for such transmembrane interactions in a line of normal rat kidney (NRK) cells in monolayer culture. We have studied the distribution on the cell surface of receptors for concanavalin A (Con A) and of the smooth muscle myosin-like protein inside the cell (4-6). These distributions were determined simultaneously on the same cell by fluorescence microscopy, using fluorescein-conjugated Con A(FI-Con A) to detect the Con A receptors, and an indirect rhodamine immunofluorescence method for the myosin. Initially the Fl-Con A was found to be uniformly dispersed over the cell surface, while the intracellular myosin was organized into extended filamentous structures. After 20 min at 370, however, the FlCon A showed a variety of nonuniform distributions, ranging from a mottled pattern to extended parallel linear arrays. In the latter case, the linear arrays of Con A receptors were superimposable on the linear arrays of the intracellular myosin-containing filaments, strongly suggesting that the receptors and the filaments had become physically linked. These observations are relevant to our recently published comparative study (7) of the distributions of Con A receptors and myosin-containing structures in NRK cells infected with Rous sarcoma virus. Abbreviations: Con A, concanavalin A; Fl-Con A, fluorescein-conjugated Con A; Suc-Con A, succinylated Con A; NRK, normal rat kidney.
4575
MATERIALS AND METHODS Cell Culture. NRK cells (8) were cultured in Ham's F12 based medium [100 ml of F12, 10 ml of Tryptose Phosphate Broth (Difco Lab, Detroit, Mich.), 5 ml of fetal calf serum, and antibiotics] under an atmosphere of 10% CO2 and 90% air at 370. Cells were dissociated with trypsin for plating onto cover slips (15,000-30,000 cells per 35-mm dish) and cultured for at least 2 days before an experimental treatment. NRK cells infected with the B77 strain of Rous sarcoma virus were used in our earlier studies (7). Protein Modification. Fl-Con A was prepared as described (7). Succinylated Con A (Suc-Con A) was a single-stage modified product (9) prepared by Dr. Randy Schekman and characterized elsewhere (10). Rhodamine B conjugation of goat antibody against rabbit IgG was done by the method of Fothergill (11), who kindly supplied the rhodamine sample and assisted in making the conjugate. Cell Staining. For Con A and Suc-Con A staining, coverslip cultures in 35-mm dishes were first rinsed three times with serum-free Dulbecco's modified Eagle's medium, then incubated with 1 ml of the medium containing 50 ug of either FlCon A or Suc-Con A for 20 min at 37°. After being rinsed three times, they were then fixed with 2% formaldehyde in phosphate-buffered saline at room temperature for 20 min (7). In order to determine the effect of antibodies against Con A on the distribution of surface-bound Suc-Con A, some cultures that had been treated with Suc-Con A were not fixed, but were reacted for an additional 20 min at 370 with 1 ml of 100 gtg/ml of a rabbit IgG fraction containing antibodies against Con A in modified Eagle's medium, and were then rinsed and fixed. After fixation, the cells were rinsed with several changes of phosphate-buffered saline for 20 min, one change containing 50 mM NH4Cl to quench any unreacted aldehyde groups. To visualize the distribution of Suc-Con A, cells that had been reacted with Suc-Con A and were then fixed were treated successively with rabbit antibodies against Con A and fluorescein-conjugated goat antibodies against rabbit IgG; whereas cells that had been reacted with Suc-Con A and rabbit antibodies against Con A, and were then fixed, were treated only with the fluorescein-conjugated goat antibodies. If myosin or microtubules were to be stained, the cells were then frozen and thawed to render them permeable to antibodies. Indirect immunofluorescent staining of myosin with rabbit antibody against human uterine myosin or of microtubular protein with rabbit antibody against chick brain tubulin (a generous gift of Dr. Melvin Simon) was done as described (7), except that rhodamine-conjugated goat antibody against rabbit IgG was used instead of the fluorescein conjugate. In all of these experiments the modified Eagle's medium used in ligand binding and for rinsing was first equilibrated with 10% CO2 at 37°. Colchicine treatment was effected by adding 10-5
4576
Cell Biology: Ash and Singer
Proc. Nati. Acad. Sci. USA 73 (1976)
FIGS. 1-4. (Legend appears at bottom of the following page.)
Cell Biology: Ash and Singer
Proc. Nati. Acad. Sci. USA 73 (1976)
4577
M colchicine in growth medium to the cells, and continuing culture at 370 for 1-2 hr before any other additions were carried out. Colchicine (10-5 M) was included in subsequent rinse and treatment solutions in these cases. Photomicroscopy and Analysis. Cells were observed with a Zeiss Photomicroscope using a 40 X oil immersion lens and photographed onto Kodak 35 mm Plus X film developed with Kodak Microdol X developer diluted 1:3. Fluorescence was excited with selected emissions from an Osram 200-W HBO lamp filtered through either a Zeiss FITC interference filter or a BG-12 excitor filter. When cells were stained with both Fl-Con A and rhodamine-IgG, the fluorescein fluorescence was photographed first using one or two Wratten no. 58 filters above the objective to absorb the red rhodamine emission. Then, without changing the focus, the rhodamine fluorescence was photographed using one Wratten no. 23A filter to absorb the fluorescein emission. The fluorescein conjugate was significantly bleached during the first exposure so that only one barrier filter was needed to prevent leakage into the rhodamine exposure.
In appropriate instances, a pair of fluorescein- and rhodamine-stained negatives of the same cell were reproduced as enlarged positive transparencies, and these were superimposed visually. For a more objective examination of the possible superposition of the two images, the positive transparencies were analyzed with an 12S Mini-Addcol Viewer, model 6000 (International Imaging Systems, Mountain View, Calif.), available at the Visibility Laboratory of the Scripps Institute of Oceanography. This instrument superimposes up to four individual images in register to display a composite image. The assistance of Mr. James Harris in performing this analysis is gratefully
acknowledged. RESULTS Myosin Patterns. The intracellular myosin staining patterns revealed by indirect immunofluorescence in normal NRK cells 2 days after plating were predominantly filamentous (7). Along with the filaments there were present sheet-like or striped patterns (Fig. la), which we described earlier (7). Treatment of cells with Con A or with colchicine had no obvious effects on these patterns. Con A Patterns. Shortly after the addition of Fl-Con A to the cells, the fluorescence was uniformly dispersed over the cell surface with no discernible structure (Fig. lb). After 20 min of incubation at 370, however, careful observation revealed that the surface distribution of Fl-Con A was no longer completely uniform on most cells. A mottled distribution of fluorescence was commonly observed (Fig. 2b). Furthermore, on about 10% of the cell surfaces, the fluorescence was seen in extended linear arrays (Figs. 2b and 3b). The mottled and linear patterns were often seen on different areas of the same cell surface (Fig. 2b). Photographing such linear arrays was difficult if the cell was rounded because only limited areas could be in focus at one time. These patterns were the same whether or not staining for intracellular myosin was carried out on the cells.
FIG. 5. The distribution of Con A receptors on NRK cells treated with Suc-Con A for 20 min at 370 and subsequently with rabbit antibodies to Con A for 20 min at 37°. The distribution of the receptors after the treatment with Suc-Con A alone is not shown, but was uniform as in Fig. lb. The incubation with the antibody redistributed the receptors into mottled patterns and also into extended linear arrays (arrows) much like those of Figs. 2b and 3b. These treatments caused the cells to round up, allowing only limited regions of the cell surface to be simultaneously in focus. X925.
For comparison, NRK cells infected with Rous sarcoma virus (B77-NRK cells), when treated with Fl-Con A for 20 min at 370, showed a very marked patching of fluorescence (Fig. 4b), as we described earlier for transformed cells (7). The myosin staining patterns were largely disorganized compared to normal NRK cells (7). Simultaneous Observation of Fl-Con A and Myosin. When cells treated with Fl-Con A for 20 min were then stained for myosin, in those cases where extended linear arrays of the Fl-Con A were observed, there was a remarkable similarity between these arrays and the filamentous myosin patterns. Two examples of this are presented in Figs. 2 and 3. Extensive portions of the Fl-Con A and of the myosin arrays were superimposable by visual inspection, as was confirmed by analysis in the 12S Mini-Addcol Viewer. Effects of Sue-Con A. Treatment of cells with Suc-Con A for 20 min at 370 resulted in a uniform distribution of this ligand on the cell surface. No mottling or linear arrangements were seen. However, if the cells treated with Suc-Con A for 20 min were then further reacted with rabbit antibodies against Con A for 20 min, there was significant rounding of the cells which was accompanied by a redistribution of the Suc-Con A as detected by immunofluorescence. The severe rounding of the cells made observation difficult, but mottled and linear patterns could now be seen on these cell surfaces (Fig. 5, arrows). Effects of Colchicine and Cytochalasin B. Colchicine treatment (10-5 M) of NRK cells for 1-2 hr eliminated the
FIGS. 1-4 (on preceding page). Figs. 1-3. Fluorescence patterns of the distribution of (a) the intracellular smooth muscle myosin-like protein (as visualized by rhodamine immunofluorescence) and (b) the surface receptors for Con A (as visualized by Fl-Con A) on the same NRK cell in each a, b pair. (Fig. 1) The patterns are those observed initially. (Figs. 2 and 3) Two examples of patterns observed after 20-min incubation with the Fl-Con A at 37°. The intracellular myosin is mostly arranged in extended filamentous arrays both initially (la) and after the 20-min incubation (2a and 3a). The Con A receptors are initially (lb) dispersed uniformly over the cell surface, but 20 min after the addition of Fl-Con A (2b and 3b), the receptors are no longer uniformly dispersed, but show a mottled distribution (M in 2b) or extended linear arrays (arrows). The linear arrays of myosin (3a) and of Con A receptors (3b) are mostly superimposable. In each a, b pair, the photographs were made by changing filters without changing focus. X600. Fig. 4. The distribution of surface receptors for Con A on NRK cells transformed with Rous sarcoma virus (B77-NRK): (a) initial distribution; (b) 20 min after incubation with Fl-Con A at 37°. In (b), extensive patching of the receptors on the rounded-up cells is evident. X600.
4578
Cell Biology: Ash and Singer
majority of the microtubular system, as judged by indirect immunofluorescent staining with rabbit antibodies against chick brain tubulin (not shown). Colchicine treatment produced cells with a uniform antitubulin staining pattern, with occasional cells showing a few fluorescent curvilinear structures originating near the nucleus, as has been described by other workers (12, 13). This treatment did not eliminate the linear arrangements of Fl-Con A on NRK cells present after 20 min of incubation shown in Figs. 2b and 3b. If cytochalasin B (10 Atg/ml) was added with Fl-Con A before fixation many of the NRK cells were found to have a reduced amount of fluorescence on their surface. If cells were first fixed, however, and then cytochalasin B and Fl-Con A were used together to stain the fixed cells, the level of Fl-Con A binding was indistinguishable from that seen with no cytochalasin B added. Thus, the treatment of unfixed NRK cells with both cytochalasin B and Fl-Con A produced an apparent loss of Con A binding sites on the surfaces of these cells. It is therefore not possible to assess the effect of cytochalasin B on the Fl-Con A induced aggregations without further study. Myosin staining patterns also appeared to be altered in some cells after a 20-min treatment with cytochalasin B, some cells showing masses of fluorescence near their edges. DISCUSSION The binding of Con A to the surface of an NRK cell induced the Con A receptors, originally dispersed uniformly over the surface, to redistribute to a limited extent. The distribution that resulted ranged in orderliness from a mottled pattern (Fig. 2b) to a highly structured one consisting of extended linear arrays (Figs. 2b and 3b). These two patterns, often seen on adjacent portions of the same cell surface, varied in relative proportion from cell to cell in a culture. This variability is not due to cellular heterogeneity in the NRK line. It may reflect metabolic differences in the cell population, such as cell cycle differences; or, as we are more inclined to think, it may reflect different states of perfection of the same ordering process. In cases where the extended linear arrays of Con A receptors were especially clearly defined, they lined up precisely over the myosin-containing filaments inside the cell. This strongly suggests that under these circumstances, repeating physical linkages were formed which extended from the Fl-Con A molecules, at the cell surface, through its receptors in the membrane, to components in the myosin-containing filaments inside the cell. These linkages may include other intervening components as well. The extended linear arrangements of Con A receptors arose only after tetravalent Con A was specifically bound to the cell surface. Divalent Suc-Con A did not serve this function, but if antibodies against Con-A were added to the cells treated with Suc-Con A, the linear arrangement of receptors was induced (Fig. 5). This suggests that some surface redistribution of the Con A receptors, and not just ligand attachment to them, is critical for the transmembrane linkage to be formed. Microtubules are not involved in this phenomenon. The addition of lo-5 M colchicine, which disrupted the microtubule assemblies, had no discernible effect on the myosin-containing filaments, nor on the Con A-induced linkage of Con A receptors to the filaments. A number of mechanisms can be entertained to explain these effects. The Con A-induced clustering of the Con A receptors might induce some enzymatic or permeability changes in the membrane; altered concentrations of some enzyme product or permeant species might then affect the characteristics of the myosin-containing filaments so that they become linked to the
Proc. Natl. Acad. Sci. USA 73 (1976)
receptors. Alternatively, the clustering of the receptors might change the properties of the receptor molecules, allowing them to become linked to the filaments. There is insufficient information available at present to discriminate between these or other possible mechanisms. These limited redistributions of Con A receptors that are induced on the surfaces of NRK cells are to be contrasted with the much more profound Con A-induced redistributions of receptors that occur on transformed cells (7, 14-18). Such redistributions result in large patches of Fl-Con A on the cell surface. An example of such patching is shown in Fig. 4b, 20 min after the addition of Con A to an NRK cell infected with the B77 strain of Rous sarcoma virus. In a companion study (7) with NRK cells infected with a temperature-sensitive mutant (LA 23) of Rous sarcoma virus, it was found that there was a correlation between the integrity of the myosin-containing filaments and the mobility of Con A receptors on the cell surface. At permissive temperatures, the infected cells exhibited the transformed phenotype: the intracellular myosin filaments were disrupted and the Con A receptors could be clustered into large patches upon the addition of Con A. At nonpermissive temperatures, the myosin-containing filaments were intact, and Con A could not induce a patching of its receptors. The results of the present study strongly support the conclusion reached earlier (7), namely, that the mobility of Con A receptors in the
plane of the membrane is physically determined by the integrity of the myosin-containing filaments in these infected NRK cells. Do these transmembrane linkage effects occur with Con A receptors on cells other than the NRK line, or with other ligand-cell interactions? It is possible that a very similar phenomenon arises upon binding Con A to cultured rat myoblasts (19). In these studies, the lateral mobility of Fl-Con A on the surface of the myoblasts was measured by photobleaching experiments, in which the rate of recovery of fluorescence was observed in a small region of the cell surface that was initially photobleached. Among the observations made were that: (i) immediately after the attachment of the fluorescent Con A to the cell surface, the lateral mobility of the Con A at 230 was appreciable, but decreased markedly by 20 min later; (ii) Suc-Con A had a much larger lateral mobility than Con A after 20 min of attachment; and (iii) colchicine had no effect on these phenomena. These results can be explained, if, by analogy to our observations with NRK cells, the Con A receptors were originally freely mobile in the membrane, but upon binding tetravalent Con A, became redistributed and linked over a period of 20 min to myosin-containing filament networks underneath the membrane. In other studies with lymphocytes (20,21) it has been shown that relatively large amounts Con A bound to the cell surface inhibit the capping that is produced by smaller amounts of Con A. These large amounts of Con A also inhibit the capping of several other receptors which is produced by their respective antibodies. These immobilization effects are fairly specific for Con A, and do not occur with the other receptor-ligand interactions. These effects may involve a transmembrane linkage of Con A receptors to myosin-containing intracellular filaments that is induced upon binding sufficiently large amounts of Con A to the lymphocyte surface. In this case, however, these receptor immobilizations are reversed by the addition of colchine, and as a result, microtubules have been implicated in these lymphocyte surface effects. It may be that the intracellular myosin-containing filaments in lymphocytes are less extensive than in the NRK cells, and are in some manner linked together by microtubules. It is interesting to note that long linear arrays
Cell Biology: Ash and Singer of transmembrane linkages could produce extended "picketfence" structures in the cross section of the membrane, consisting of regularly spaced receptor molecules spanning the membrane (see Fig. 2b in ref. 5). Those membrane molecules that were incapable of penetrating the interstices in the "picket-fence" would be prevented from large-scale lateral movement in the membrane even if they were not bound to the Con A receptors. This mechanism might account in part for the apparent immobilization of many other receptors in the lymphocyte surface that is induced by Con A binding to the cell. Transmembrane linkage phenomena of the kind we have described may therefore be characteristic of Con A receptors on a variety of different cells. Whether any other kinds of cell surface receptors act similarly remains to be determined. We gratefully acknowledge the use of tissue culture facilities generously provided by Dr. Immo Scheffler. The NRK and B77-NRK cells were from stocks originally provided by Dr. Peter K. Vogt. J.F.A. is a USPHS Postdoctoral Fellow, 1974 to present. S.J.S. is a American Cancer Society Research Professor. Studies supported by USPHS Grants AI-06659 and GM-15972. 1. Singer, S. J. (1974) "The molecular organization of membranes," Annu. Rev. Biochem. 43, 805-833. 2. Nicolson, G. L. (1976) "Transmembrane control of the receptors on normal and tumor cells. I. Cytoplasmic influence over cell surface components," Biochim. Biophys. Acta 457, 57-108. 3. Edelman, G. M. (1976) "Surface modulation in cell recognition and cell growth," Science 192, 218-226. 4. Weber, K. & Groeschel-Stewart, U. (1974) "Antibody to myosin: The specific visualization of myosin-containing filaments in non-muscle cells," Proc. Natl. Acad. Sci. USA 71, 4561-4564. 5. Painter, R. G., Sheetz, M. & Singer, S. J. (1975) "Detection and ultrastructural localization of human smooth muscle myosin-like molecules in human non-muscle cells by specific antibodies," Proc. Natl. Acad. Sci. USA 72, 1359-1363. 6. Wang, K., Ash, J. F. & Singer, S. J. (1975) "Filamin, a new high-molecular-weight protein found in smooth muscle and non-muscle cells," Proc. Natl. Acad. Sci. USA 72, 4483-4486. 7. Ash, J. F., Vogt, P. K. & Singer, S. J. (1976) "The reversion from transformed to normal phenotype by inhibition of protein synthesis in rat kidney cells infected with a temperature-sensitive Rous sarcoma virus mutant," Proc. Natl. Acad. Sci. USA 73, 3603-3607. 8. Duc-Nguyen, H., Rosenblum, E. N. & Zeigel, R. F. (1966) "Persistent infection of a rat kidney cell line with Rauscher
Proc. Natl. Acad. Sci. USA 73 (1976)
4579
murine leukemia virus," J. Bacteriol. 92, 1133-1140. 9. Gunther, G. R., Wang, J. L., Yahara, I., Cunningham, B. A. & Edelman, G. M. (1973) "Concanavalin A derivatives with altered biological activities," Proc. Natl. Acad. Sci. USA 70, 10121016. 10. Schekman, R. & Singer, S. J. (1976) "Clustering and endocytosis of membrane receptors can be induced in mature erythrocytes of neonatal humans but not adults," Proc. Natl. Acad. Sci. USA 73,4075-4079. 11. Fothergill, J. E. (1969) "Properties of conjugated serum proteins," in Fluorescent Protein Tracing, ed. Nairn, R. C. (Livingstone, London), 3rd ed., pp. 35-0. 12. Brinkley, B. R., Fuller, G. M. & Highfield, D. P. (1975) "Cytoplasmic microtubules in normal and transformed cells in culture: Analysis by tubulin antibody immunofluorescence," Proc. Natl. Acad. Sci. USA 72,4981-4985. 13. Osborn, M. & Weber, K. (1976) "Cytoplasmic microtubules in tissue culture cells appear to grow from an organizing structure towards the plasma membrane," Proc. Nati. Acad. Sci. USA 73, 867-871. 14. Rosenblith, J. Z., Ukena, T. E., Yin, H. H., Berlin, R. D. & Karnovsky, M. J. (1973) "A comparative evaluation of the distribution of concanavalin A-binding sites on the surfaces of normal, virally-transformed, and protease-treated fibroblasts," Proc. Nati. Acad. Sci. USA 70, 1625-1629. 15. Nicolson, G. L. (1973) "Temperature-dependent mobility of concanavalin A sites on tumour cell surfaces," Nature New Biol. 243,218-220. 16. Noonan, K. D. & Burger, M. M. (1973) "The relationship of concanavalin A binding to lectin-initiated cell agglutination," J. Cell Biol. 59, 134-142. 17. Inbar, M. & Sachs, L. (1973) "Mobility of carbohydrate containing sites on the surface membrane in relation to the control of cell growth," FEBS Lett. 32, 124-128. 18. Edelman, G. M. & Yahara, I. (1976) "Temperature-sensitive changes in surface modulating assemblies of fibroblasts transformed by mutants of Rous sarcoma virus," Proc. Natl. Acad. Scd. USA 73,2047-2051. 19. Schlessinger, J., Koppel, D. E., Axelrod, D., Jacobson, K., Webb, W. W. & Elson, E. L. (1976) "Lateral transport on cell membranes: Mobility of concanavalin A receptors on myoblasts," Proc. Natl. Acad. Sci. USA 73,2409-2413. 20. Yahara, I. & Edelman, G. M. (1972) "Restriction of the mobility of lymphocyte immunoglobulin receptors by concanavalin A," Proc. Natl. Acad. Sci. USA 69,608-612. 21. Yahara, I. & Edelman, G. M. (1975) "Modulation of lymphocyte receptor mobility by locally bound concanavalin A," Proc. Natl. Acad. Sci. USA 72, 1579-1583.
