- Email: [email protected]
Co-expression of spectrin and fodrin in friend erythroleukemic cells treated with DMSO
Experimental
Cell Research 152 (1984) 15-21
Co-expression of Spectrin and Fodrin Erythroleukemic Ceils Treated with JOHN Molecular
GLENNEY
and PHYLLIS
in Friend DMSO
GLENNEY
Biology and Virology Laboratory, The Salk Institute for Biological San Diego, CA 92138, USA
Studies,
Friend erythroleukemic cells can be used as a model of erythroid cell differentiation with the synthesis of the erythrocyte-specific products hemoglobin and spectrin stimulated by agents such as DMSO. In the present study we investigated the expressiou of both erythroid spectrin and non-erythroid fodrin in uninduced and DMSO-treated Friend cells. We report that both spectrin and fodrin co-exist at low levels in uninduced Friend cells and both are induced by treatment with DMSO. After longer times both spectrin and fodrin appear to undergo rearrangements into submembranous ‘patches’ and ‘caps’. Although both molecules co-localize in most of these cells, they can be independently immunoprecipitated, suggesting that significant amounts of hybrid molecules are not formed.
The molecular basis of plasma membrane-microfilament connections has received attention recently with the discovery of an analog of spectrin in nonerythroid cells [l-5]. Spectrin provides the membrane scaffold which is linked to plasma membrane via ankyrin in the mammalian erythrocyte [6]. The nonerythroid counterpart, termed fodrin, shares many properties with spectrin including the ability to interact with actin and calmodulin [7,8], as well as having an elongated double-stranded morphology with represents the head-to-head association of heterodimers [8, 91. Although clearly related, spectrin and fodrin differ in many respects. In the mammalian system, e.g., peptide maps reveal significant differences between fodrin and spectrin for both subunits and antibodies cross-react only weakly, if at all [2, 3, 9, IO]. In addition, whereas spectrin has a restricted cell-type distribution, fodrin is more general, detectable in cells as diverse as neurons, fibroblasts, hepatocytes, lymphocytes and intestinal epithelial cells [ll]. The developmental regulation of this class of proteins may be important to our understanding of their function. In intestinal epithelial cells fodrin is present before a second spectrin-like molecule TW260/240 is expressed 1111. When synthesized, TW260/240 is inserted only in the apical (brush border) membrane [ll], probably cross-linking filament bundles [12], whereas fodrin underlies the entire plasma membrane of these same cells 1111. To further explore the expression of spectrin-related proteins we have analyzed the synthesis and distribution of spectrin and fodrin in erythroleukemic cells transformed by Friend virus. This cell line provides a system in which the 2-848335
Copyright 0 1984 by Academic Press, Inc. All rights of reproduction in any form reserved 0014s4827t34 $03.00
16
Glenney and Glenney
differentiated products of the red blood cell can be experimentally induced by agents such as DMSO [13]. Antibodies which can distinguish between mammalian fodrin and spectrin were used in immunofluorescence microscopy and immunoprecipitation experiments. We report that both spectrin and fodrin are simultaneously present in Friend cells induced with DMSO.
MATERIALS
AND
METHODS
Cells and Antibodies DBA/2-inducible 745.6 Friend cells (a gift from R. Hyman, Salk Institute, see ref. [14] were grown in DME (Dulbecco’s modified Eagle medium) medium containing 10% fetal calf serum (FCS) and 2x 10m5M 2-mercaptoethanol without dimethyl sulfoxide (DMSO) or with 1.5 % DMSO for the length of time listed on the figure. 3T3 cells were a gift from Dr D. Miller (Salk Institute) and human erythrocytes were freshly drawn and washed three times with PBS. Anti-human spectrin [15] was used as an IgG fraction and was a gift from Dr Gerhardt Hiller (Max-Planck Institute, Gottingen, West Germany). Fodrin was purified by a modification of a published procedure [7] and was further purified by preparative gel electrophoresis. Bands corresponding to fodrin were visualized by precipitation in ice-cold 0.3 M KCI, excised and electrophoretically eluted, dialysed against 0.1% SDS and lyophilized. Three guinea pigs were injected with 200 ug antigen each, with 50% complete Freund’s adjuvant followed by boosters at 3 and 6 weeks. After an additional 3 weeks, 500 ug of soluble homogeneous fodrin (not electrophoretically purified) was injected per day for 3 consecutive days and the animals were exsanguinated by cardiac puncture while anesthetized 5 days after the last injection. Serum from all three animals was pooled and the antibodies purified by affinity chromatography using electrophoretically pure fodrin coupled to Sepharose 4B (Pharmacia) and elution with neutralized 4 M MgC12. After extensive dialysis against PBS (containing 0.02% NaNs), BSA was added to 1 mg/ml and the antibody was frozen at -70°C in aliquots. Rabbit antibodies to pig brain fodrin were made by a similar procedure.
Irnmunojluorescence
Microscopy
Cells were washed with phosphate-buffered saline (PBS), attached to poly-L-lysine-coated coverslips, fixed and permeablized in -20°C acetone/methanol (1 : 1) and air-dried. Cells were then treated with PBS containing 1 mg/ml bovine serum albumen (BSA) (Sigma Chemical Co.), 20 @ml rabbit anti-human spectrin plus 10 ug/ml affinity-purified guinea pig antibodies to fodrin. After 30 min at 37”C, coverslips were washed three times with PBS with BSA and then incubated a further 30 min in PBS with BSA containing aIBnity-purified fluorescein-conjugated goat anti-rabbit antibodies (50 &ml, Sigma) plus rhodamine-conjugated goat anti-guinea pig antibodies (50 ug/ml, Cappel). Coverslips were then washed with PBS and mounted in glycerol containing n-propyl gallate as described [16], and examined with a Nikon fluorescence microscope. In preliminary experiments rabbit antifodrin and anti-spectrin were used separately and gave a similar distribution as in the double-labeling experiments.
Immunoprecipitation Either uninduced cells or cells treated with 1.5 % DMSO for the number of days indicated were labeled with [35S]methionine overnight in DME medium minus methionine and containing 10% FCS, 2x 10m5 M 2-mercaptoethanol, with or without 1.5% DMSO and 50 pCi [35S]methionine (New England Nuclear) per ml. Cells were harvested and washed four times in PBS and extracted in 20 mM Ttis, 1 M NaCI, 1% Tiiton X-100 0.02% NaNs, pH 8.0 as described previously [l I]. The extract was clarified by centrifugation (10000 g, 10 min) and the amount of radioactivity insoluble in 10% TCA (using 1 mg/ml BSA as carrier) was determined. The extracts were adjusted to equal volume and equal amounts TCA-insoluble radioactivity and adjusted to 10 ug/ml rabbit antisera against spectrin, fodrin or villin as a control. After 3 h at WC, 35 u1 of a 10% Sruph. a. suspension, prepared as described [17] were added and the tubes were rotated end over end for 30 min at 4°C. The Staph. a. was Exp Cell
Res 152 (1984)
Spectrin
and fodrin
in DMSO-treated
Friend
cells
17
pelleted, washed four times with extraction buffer and boiled 5 min in SDS-sample buffer. The Staph. a. was removed by centrifugation and the supernatant applied to SDS-polyacrylamide gels (4% acrylamide). Gels were stained in Coomassie Blue, destained, treated with Enhance (New England Nuclear), dried and subjected to fluorography using pre-flashed film. Pure fodrin and spectrin standards were run in adjacent lanes. In some cases the autoradiogram was carefully aligned over the dried gel and individual bands were excisted, mixed with scintillation cocktail and counted in a liquid scintillation counter. Western immunoblots were as in [1 11 using porcine or avian fodrin as antigens and rabbit anti-fodrin or anti-human spectrin as antibody.
RESULTS
AND
DISCUSSION
The antibodies used in the present study display a high degree of specificity for the protein used as immunogen. Control experiments for double label immunofluorescence microscopy demonstrate that only spectrin is detected in mature erythrocytes and only fodrin can be found in 3T3 cells (fig. 1). Similarly, immunoprecipitation demonstrates that anti-spectrin and anti-fodrin display little crossreactivity (fig. 2, see below also). In addition, ‘western blots (fig. 3) show that anti-spectrin reacts only with mammalian spectrin, whereas anti-fodrin reacts with mammalian and avian fodrin as well as avian spectrin (but not mammalian spectrin). Erythroleukemic cells transformed by Friend virus (Friend cells) have been used as a model system for erythroid cell differentiation [13] since proteins not found on mature red blood cells, such as histocompatability antigens [ 181 or T-200 [14], are lost upon DMSO treatment concomitant with the aquisition of the erythrocyte-specific products hemoglobin [19] and spectrin [20-221. Non-induced Friend cells display a low but detectable level of both spectrin and fodrin (figs. 1, 2). The anti-spectrin fluorescent staining pattern was weak and fairly diffuse and the anti-fodrin was also diffuse but with a larger amount concentrated under the membrane. Upon treatment with DMSO, not only did the level of fodrin and spectrin increase as shown by both fluorescence microscopy (fig. 1) and immunoprecipitation data (fig. 2), but the distribution appeared to change dramatically (fig. 1). After l-2 days of DMSO treatment both the spectrin and fodrin fluorescence was more punctate and concentrated under the membrane, in many cells giving the appearance of ‘patches’. Longer times (3-5 days) of exposure to DMSO in most cells led to a global rearrangement of spectrin and fodrin antigens into what appeared to be a submembrane ‘cap’. Most of these caps contained both members of this class of proteins, spectrin and fodrin. In some cases fodrin is present in surface ‘patches’ in which spectrin is lacking (see fig. 1, 3-day induced for an example); however, by a large majority of these ‘patches’ and ‘caps’ contain both proteins. No examples were found in which spectrin patches or caps were present without fodrin (fig. 1). This is the first demonstration in a mammalian cell that spectrin and fodrin can not only co-exist in the same cell but have an identical, albeit abnormal (capped) distribution. In addition, although the Friend cell line has been used as a model of erythrocyte differentiation it is Exp CellRes
152 (1984)
Glenney and Glenney
18
Uninduced
lday
2day
3day
4day
3T3
RBC Fig. 1. Double label immunofluorescence localization of fodrin and spectrin in Friend erythroleukemia cells. Friend cells were attached to coverslips, fixed and made permeable in methanol/acetone and treated with rabbit anti-human spectrin and affinity-purified guinea pig antibodies to fodrin. After 30 min at 37”C, coverslips were washed with PBS and then incubated for a further 30 min in PBS and affinity-purified fluorescein-conjugated goat anti-rabbit antibodies and rhodamine-conjugated goat anti-guinea pig antibodies. Coverslips were then washed with PBS and mounted and examined with a Nikon fluorescence microscope. Left, Image in the fluorescein channel; right, the same field viewed in the rhodamine channel. The double arrows (3-day DMSO-treated) shows an example in which the cell stains brightly with anti-fodrin, but weakly with anti-spectrin. The single arrows (4-day DMSOtreated) show a few examples in which brightly fluorescent ‘caps’ of anti-fodrin and anti-spectrin are apparent. Note that all of the brightly fluorescent anti-spectrin staining coincides with the anti-fodrin staining. X3, Controls in which 3T3 cells or RBC, human red blood cells were treated with antispectrin and anti-fodrin and both second antibodies, photographed and printed as above. Exp
Cell
Res 152 (1984)
Spectrin
and fodrin
in DMSO-treated
days
Friend
cells
19
& DMSO
Fig. 2. Immunoprecipitation of spectrin and fodrin from Friend cells. Either uninduced cells (v) or cells treated with 1.5 % DMSO for the number of days indicated were labeled with [35S]methionine. Cells were harvested and extracted as described in Materials and Methods. The extracts were adjusted to equal volume and amounts of TCA-insoluble radioactivity and adjusted to 10 t&ml rabbit antisera against S, spectrin; F, fodrin or V, villin as a control. After 3 h, 35 pl of a 10% Sruph. a. suspension were added and the tubes were rotated end over end for 30 min at 4°C. The Staph. a. was pelleted, washed with extraction buffer and boiled 5 min in SDS-sample buffer. The Staph. a. was removed by centrifugation and the supematant applied to SDS-polyacrylamide gels (4 % acrylamide). Gels were stained in Coomassie Blue, destained, treated with Enhance (New England Nuclear), dried and subjected to fluorography using pre-flashed film. (A) Arrows, right, indicate the position of fodrin 240, 235 K and spectrin 220 K subunits run in adjacent lanes. (A) Fluorograph of immunoprecipitates from uninduced and 3-day induced Friend cells; (B) cpm in the gel bands corresponding to O-O, fodrin 240 K; A-A, fodrin 235 K; O-O, spectrin 220 K or m-m, spectrin 240 K. Note that both spectrin and fodrin increase in amounts with DMSO treatment compared with uninduced cells, no trace of fodrin 235 K is found in anti-spectrin immunoprecipitates, and the two subunits of fodrin are found in equimolar amounts, whereas the 220 K of spectrin is synthesized in large excess over the 240 K.
interesting that a protein is induced by DMSO which is not normally present in the mature erythrocyte. Although this distribution is slightly different than the spectrin distribution originally reported by Eisen et al. [20], this may be due to the difference in clonal variants used for these studies as noted elsewhere [22]. Immunoprecipitation of 35S-labeled extracts of uninduced or DMSO-treated Friend cells reveals that both spectrin and fodrin polypeptides are synthesized. Purified spectrin or fodrin standards run in adjacent lanes and stained with Coomassie Blue (not shown) confirmed that the two polypeptides precipitated with anti-spectrin antibodies migrated in the position of the two polypeptide chains of isolated spectrin (240 and 220 K), and the two bands precipitated with anti-fodrin corresponded to the 240 and 235 K subunits of brain fodrin. Control experiments revealed that in all cases antibodies were in large excess over fodrin or spectrin (not shown). It would appear, then, that uninduced cells synthesize a low level of spectrin which then gradually increases upon treatment with DMSO. One consistent feature is that the spectrin beta chain (220 K) is produced in larger amounts over the alpha-chain, whereas the two subunits of fodrin are always Exp CellRes
152 (1984)
20
Glenney and Glenney
Fig. 3. Western blotting of spectrin and fodrin subunits using affinity purified antibodies to mammalian spectrin or fodrin. A, Porcine fodrin; B, porcine spectrin; and C avian fodrin, or D, avian spectrin were run on SDS-gels (5 % acrylamide) and stained with (a) Coomassie Blue dye; or (b) transferred to nitrocelhrlose and probed with antibodies to porcine fodrin; or (c) human spectrin followed by I*?protein A and autoradiography. Only the relevant portion of the gel and autoradiograph are shown. Note that antibodies to fodrin show no cross-reactivity with mammalian spectrin (although reacting with avian spectrin) and antibodies to human spectrin display on cross-reactivity with fodrin.
produced in equal amounts (fig. 2) (both alpha and beta chains of spectrin and fodrin have approximately equal number of methionines [7]. This result is different from that seen in avian erythropoiesis [25], in which the beta-chain is synthesized in larger amounts. This could be due to the differences in labeling (0.25-3 h vs 12 h) or reflect a basic difference in normal avian distribution compared to induction of differentiated characteristics of Friend cells. Although spectrin and fodrin exist as alphaZ-beta* tetramers and spectrin and fodrin are both present in Friend cells, it is possible that heterodimers might form. Apparently this does not occur, since spectrin and fodrin are independently immunoprecipitable (fig. 2). This is surprising since (i) spectrin is always found in fodrin-positive areas, especially at later stages of induction (fig. 1); (ii) recent results have shown that hybrid molecules can be formed between subunits of spectrin and fodrin in vitro (although with a low yield) [23]; and (iii) under conditions of immunoprecipitation used in these experiments both subunits of spectrin [24] or fodrin [ 1I] remain firmly bound together as a complex, even in the presence of excess antibody [ 111. The reason that spectrin-fodrin hybrids are not found may be due to a much higher afftnity of the spectrin or fodrin subunits for each other than for the other protein. Although a small amount of a peptide comigrating with the beta-chain of spectrin is found in immunoprecipitates using anti-fodrin (fig. 2), this may reflect a low level of such hybrids. Western blots (tig. 3) and immunofluorescence controls (fig. 1) would suggest that it is not due to cross-reactivity of the antibodies. The reason for the capping of fodrin and spectrin is also not known, but may reflect a limiting amount of the membrane receptor in induced cells. Another important question is whether any integral membrane proteins co-cap with fodrin and spectrin in these cells. Future studies will have to approach this question and ihe Friend cell system may provide an excellent model to analyze the regulation of the membrane-cytoskeleton system in a cell which contains two such important structural elements. i5.x.~ Cell Res 152 (1984)
Spectrin
and fodrin
in DMSO-treated
Friend
cells
21
REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25.
Glenney, J R, Jr, Glenney, P & Weber, K, Proc natl acad sci US 79 (1982) 4002. Bennett, V, Davis, J & Fowler, W E, Nature 299 (1982) 126. But-ridge, K, Kelly, T L Mangeat, P, J cell bio195 (1982) 478. Repasky, E A, &anger, B L & Lazarides, E, Cell 29 (1982) 821. Goodman, S R, Zagon, I S C Kulikowski, R R, Proc natl acad sci US 78 (1981) 7570. Goodman, S R & Shiffer, K, J Am phys sot (1983) C-121. Glenney, J R Jr, Glenney, P & Weber, K, J biol them 257 (1982) 9781. Tsukita, S, Tsukita, S, Ishikawa, H, Kurokawa, M, Morimoto, K, Sobue, K & Kakiuchi, S, J cell bio197 (1983) 574. Glenney, J R Jr, Glenney, P & Weber, K, J mol biol 167 (1983) 275. Glenney, J R Jr 8~ Glenney, P, Cell and muscle motility. 3 (1983) 671. - Cell 34 (1983) 503. Glenney, J R Jr, Glenney, P & Weber, K, J cell biol 96 (1983) 1491. Marks, P A & Rifkind, R A, Ann rev biochem 47 (1978) 419. Hyman, R, Bowbridge, I & Cunningham, K, J cell physiol 105 (1980) 469. Hiller, G & Weber, K, Nature 266 (1977) 181. Giloh, H & Sedat, J W, Science 217 (1982) 1252. Blase, S H, Matsumura, F & Lin, J J-C, Cold Spring Harbor symp quant bio146 (1982) 455. Amdt-Jovin, D J, Ostertag, W, Eisen, H, Klimed, F & Jovin, T, J histochem cytochem 24 (1976) 332. Friend, C, Scher, W, Holland, J & Sato, T, Proc natl acad sci US 68 (1971) 378. Eisen, H, Bach, R & Emery, R, Proc natl acad sci US 74 (1977) 3898. Pfeffer, S R & Redman, C M, Biochim biophys acta 641 (1981) 254. Marshall, L M & Hunt, R C, J cell sci 54 (1982) 97. Davis, J & Bennett, V, J biol them 258 (1983) 7757. Nelson, W J, Colaco, C A & Lazarides, E, Proc natl acad sci US 80 (1983) 1626. Blikstad, I, Nelson, W J, Moon, R T & Lazarides, E, Cell 32 (1983) 1081.
Received September 27, 1983 Revised received November 21, 1983
Printed
in Sweden
Exp Cell
Res I52 11984)
Cell Research 152 (1984) 15-21
Co-expression of Spectrin and Fodrin Erythroleukemic Ceils Treated with JOHN Molecular
GLENNEY
and PHYLLIS
in Friend DMSO
GLENNEY
Biology and Virology Laboratory, The Salk Institute for Biological San Diego, CA 92138, USA
Studies,
Friend erythroleukemic cells can be used as a model of erythroid cell differentiation with the synthesis of the erythrocyte-specific products hemoglobin and spectrin stimulated by agents such as DMSO. In the present study we investigated the expressiou of both erythroid spectrin and non-erythroid fodrin in uninduced and DMSO-treated Friend cells. We report that both spectrin and fodrin co-exist at low levels in uninduced Friend cells and both are induced by treatment with DMSO. After longer times both spectrin and fodrin appear to undergo rearrangements into submembranous ‘patches’ and ‘caps’. Although both molecules co-localize in most of these cells, they can be independently immunoprecipitated, suggesting that significant amounts of hybrid molecules are not formed.
The molecular basis of plasma membrane-microfilament connections has received attention recently with the discovery of an analog of spectrin in nonerythroid cells [l-5]. Spectrin provides the membrane scaffold which is linked to plasma membrane via ankyrin in the mammalian erythrocyte [6]. The nonerythroid counterpart, termed fodrin, shares many properties with spectrin including the ability to interact with actin and calmodulin [7,8], as well as having an elongated double-stranded morphology with represents the head-to-head association of heterodimers [8, 91. Although clearly related, spectrin and fodrin differ in many respects. In the mammalian system, e.g., peptide maps reveal significant differences between fodrin and spectrin for both subunits and antibodies cross-react only weakly, if at all [2, 3, 9, IO]. In addition, whereas spectrin has a restricted cell-type distribution, fodrin is more general, detectable in cells as diverse as neurons, fibroblasts, hepatocytes, lymphocytes and intestinal epithelial cells [ll]. The developmental regulation of this class of proteins may be important to our understanding of their function. In intestinal epithelial cells fodrin is present before a second spectrin-like molecule TW260/240 is expressed 1111. When synthesized, TW260/240 is inserted only in the apical (brush border) membrane [ll], probably cross-linking filament bundles [12], whereas fodrin underlies the entire plasma membrane of these same cells 1111. To further explore the expression of spectrin-related proteins we have analyzed the synthesis and distribution of spectrin and fodrin in erythroleukemic cells transformed by Friend virus. This cell line provides a system in which the 2-848335
Copyright 0 1984 by Academic Press, Inc. All rights of reproduction in any form reserved 0014s4827t34 $03.00
16
Glenney and Glenney
differentiated products of the red blood cell can be experimentally induced by agents such as DMSO [13]. Antibodies which can distinguish between mammalian fodrin and spectrin were used in immunofluorescence microscopy and immunoprecipitation experiments. We report that both spectrin and fodrin are simultaneously present in Friend cells induced with DMSO.
MATERIALS
AND
METHODS
Cells and Antibodies DBA/2-inducible 745.6 Friend cells (a gift from R. Hyman, Salk Institute, see ref. [14] were grown in DME (Dulbecco’s modified Eagle medium) medium containing 10% fetal calf serum (FCS) and 2x 10m5M 2-mercaptoethanol without dimethyl sulfoxide (DMSO) or with 1.5 % DMSO for the length of time listed on the figure. 3T3 cells were a gift from Dr D. Miller (Salk Institute) and human erythrocytes were freshly drawn and washed three times with PBS. Anti-human spectrin [15] was used as an IgG fraction and was a gift from Dr Gerhardt Hiller (Max-Planck Institute, Gottingen, West Germany). Fodrin was purified by a modification of a published procedure [7] and was further purified by preparative gel electrophoresis. Bands corresponding to fodrin were visualized by precipitation in ice-cold 0.3 M KCI, excised and electrophoretically eluted, dialysed against 0.1% SDS and lyophilized. Three guinea pigs were injected with 200 ug antigen each, with 50% complete Freund’s adjuvant followed by boosters at 3 and 6 weeks. After an additional 3 weeks, 500 ug of soluble homogeneous fodrin (not electrophoretically purified) was injected per day for 3 consecutive days and the animals were exsanguinated by cardiac puncture while anesthetized 5 days after the last injection. Serum from all three animals was pooled and the antibodies purified by affinity chromatography using electrophoretically pure fodrin coupled to Sepharose 4B (Pharmacia) and elution with neutralized 4 M MgC12. After extensive dialysis against PBS (containing 0.02% NaNs), BSA was added to 1 mg/ml and the antibody was frozen at -70°C in aliquots. Rabbit antibodies to pig brain fodrin were made by a similar procedure.
Irnmunojluorescence
Microscopy
Cells were washed with phosphate-buffered saline (PBS), attached to poly-L-lysine-coated coverslips, fixed and permeablized in -20°C acetone/methanol (1 : 1) and air-dried. Cells were then treated with PBS containing 1 mg/ml bovine serum albumen (BSA) (Sigma Chemical Co.), 20 @ml rabbit anti-human spectrin plus 10 ug/ml affinity-purified guinea pig antibodies to fodrin. After 30 min at 37”C, coverslips were washed three times with PBS with BSA and then incubated a further 30 min in PBS with BSA containing aIBnity-purified fluorescein-conjugated goat anti-rabbit antibodies (50 &ml, Sigma) plus rhodamine-conjugated goat anti-guinea pig antibodies (50 ug/ml, Cappel). Coverslips were then washed with PBS and mounted in glycerol containing n-propyl gallate as described [16], and examined with a Nikon fluorescence microscope. In preliminary experiments rabbit antifodrin and anti-spectrin were used separately and gave a similar distribution as in the double-labeling experiments.
Immunoprecipitation Either uninduced cells or cells treated with 1.5 % DMSO for the number of days indicated were labeled with [35S]methionine overnight in DME medium minus methionine and containing 10% FCS, 2x 10m5 M 2-mercaptoethanol, with or without 1.5% DMSO and 50 pCi [35S]methionine (New England Nuclear) per ml. Cells were harvested and washed four times in PBS and extracted in 20 mM Ttis, 1 M NaCI, 1% Tiiton X-100 0.02% NaNs, pH 8.0 as described previously [l I]. The extract was clarified by centrifugation (10000 g, 10 min) and the amount of radioactivity insoluble in 10% TCA (using 1 mg/ml BSA as carrier) was determined. The extracts were adjusted to equal volume and equal amounts TCA-insoluble radioactivity and adjusted to 10 ug/ml rabbit antisera against spectrin, fodrin or villin as a control. After 3 h at WC, 35 u1 of a 10% Sruph. a. suspension, prepared as described [17] were added and the tubes were rotated end over end for 30 min at 4°C. The Staph. a. was Exp Cell
Res 152 (1984)
Spectrin
and fodrin
in DMSO-treated
Friend
cells
17
pelleted, washed four times with extraction buffer and boiled 5 min in SDS-sample buffer. The Staph. a. was removed by centrifugation and the supernatant applied to SDS-polyacrylamide gels (4% acrylamide). Gels were stained in Coomassie Blue, destained, treated with Enhance (New England Nuclear), dried and subjected to fluorography using pre-flashed film. Pure fodrin and spectrin standards were run in adjacent lanes. In some cases the autoradiogram was carefully aligned over the dried gel and individual bands were excisted, mixed with scintillation cocktail and counted in a liquid scintillation counter. Western immunoblots were as in [1 11 using porcine or avian fodrin as antigens and rabbit anti-fodrin or anti-human spectrin as antibody.
RESULTS
AND
DISCUSSION
The antibodies used in the present study display a high degree of specificity for the protein used as immunogen. Control experiments for double label immunofluorescence microscopy demonstrate that only spectrin is detected in mature erythrocytes and only fodrin can be found in 3T3 cells (fig. 1). Similarly, immunoprecipitation demonstrates that anti-spectrin and anti-fodrin display little crossreactivity (fig. 2, see below also). In addition, ‘western blots (fig. 3) show that anti-spectrin reacts only with mammalian spectrin, whereas anti-fodrin reacts with mammalian and avian fodrin as well as avian spectrin (but not mammalian spectrin). Erythroleukemic cells transformed by Friend virus (Friend cells) have been used as a model system for erythroid cell differentiation [13] since proteins not found on mature red blood cells, such as histocompatability antigens [ 181 or T-200 [14], are lost upon DMSO treatment concomitant with the aquisition of the erythrocyte-specific products hemoglobin [19] and spectrin [20-221. Non-induced Friend cells display a low but detectable level of both spectrin and fodrin (figs. 1, 2). The anti-spectrin fluorescent staining pattern was weak and fairly diffuse and the anti-fodrin was also diffuse but with a larger amount concentrated under the membrane. Upon treatment with DMSO, not only did the level of fodrin and spectrin increase as shown by both fluorescence microscopy (fig. 1) and immunoprecipitation data (fig. 2), but the distribution appeared to change dramatically (fig. 1). After l-2 days of DMSO treatment both the spectrin and fodrin fluorescence was more punctate and concentrated under the membrane, in many cells giving the appearance of ‘patches’. Longer times (3-5 days) of exposure to DMSO in most cells led to a global rearrangement of spectrin and fodrin antigens into what appeared to be a submembrane ‘cap’. Most of these caps contained both members of this class of proteins, spectrin and fodrin. In some cases fodrin is present in surface ‘patches’ in which spectrin is lacking (see fig. 1, 3-day induced for an example); however, by a large majority of these ‘patches’ and ‘caps’ contain both proteins. No examples were found in which spectrin patches or caps were present without fodrin (fig. 1). This is the first demonstration in a mammalian cell that spectrin and fodrin can not only co-exist in the same cell but have an identical, albeit abnormal (capped) distribution. In addition, although the Friend cell line has been used as a model of erythrocyte differentiation it is Exp CellRes
152 (1984)
Glenney and Glenney
18
Uninduced
lday
2day
3day
4day
3T3
RBC Fig. 1. Double label immunofluorescence localization of fodrin and spectrin in Friend erythroleukemia cells. Friend cells were attached to coverslips, fixed and made permeable in methanol/acetone and treated with rabbit anti-human spectrin and affinity-purified guinea pig antibodies to fodrin. After 30 min at 37”C, coverslips were washed with PBS and then incubated for a further 30 min in PBS and affinity-purified fluorescein-conjugated goat anti-rabbit antibodies and rhodamine-conjugated goat anti-guinea pig antibodies. Coverslips were then washed with PBS and mounted and examined with a Nikon fluorescence microscope. Left, Image in the fluorescein channel; right, the same field viewed in the rhodamine channel. The double arrows (3-day DMSO-treated) shows an example in which the cell stains brightly with anti-fodrin, but weakly with anti-spectrin. The single arrows (4-day DMSOtreated) show a few examples in which brightly fluorescent ‘caps’ of anti-fodrin and anti-spectrin are apparent. Note that all of the brightly fluorescent anti-spectrin staining coincides with the anti-fodrin staining. X3, Controls in which 3T3 cells or RBC, human red blood cells were treated with antispectrin and anti-fodrin and both second antibodies, photographed and printed as above. Exp
Cell
Res 152 (1984)
Spectrin
and fodrin
in DMSO-treated
days
Friend
cells
19
& DMSO
Fig. 2. Immunoprecipitation of spectrin and fodrin from Friend cells. Either uninduced cells (v) or cells treated with 1.5 % DMSO for the number of days indicated were labeled with [35S]methionine. Cells were harvested and extracted as described in Materials and Methods. The extracts were adjusted to equal volume and amounts of TCA-insoluble radioactivity and adjusted to 10 t&ml rabbit antisera against S, spectrin; F, fodrin or V, villin as a control. After 3 h, 35 pl of a 10% Sruph. a. suspension were added and the tubes were rotated end over end for 30 min at 4°C. The Staph. a. was pelleted, washed with extraction buffer and boiled 5 min in SDS-sample buffer. The Staph. a. was removed by centrifugation and the supematant applied to SDS-polyacrylamide gels (4 % acrylamide). Gels were stained in Coomassie Blue, destained, treated with Enhance (New England Nuclear), dried and subjected to fluorography using pre-flashed film. (A) Arrows, right, indicate the position of fodrin 240, 235 K and spectrin 220 K subunits run in adjacent lanes. (A) Fluorograph of immunoprecipitates from uninduced and 3-day induced Friend cells; (B) cpm in the gel bands corresponding to O-O, fodrin 240 K; A-A, fodrin 235 K; O-O, spectrin 220 K or m-m, spectrin 240 K. Note that both spectrin and fodrin increase in amounts with DMSO treatment compared with uninduced cells, no trace of fodrin 235 K is found in anti-spectrin immunoprecipitates, and the two subunits of fodrin are found in equimolar amounts, whereas the 220 K of spectrin is synthesized in large excess over the 240 K.
interesting that a protein is induced by DMSO which is not normally present in the mature erythrocyte. Although this distribution is slightly different than the spectrin distribution originally reported by Eisen et al. [20], this may be due to the difference in clonal variants used for these studies as noted elsewhere [22]. Immunoprecipitation of 35S-labeled extracts of uninduced or DMSO-treated Friend cells reveals that both spectrin and fodrin polypeptides are synthesized. Purified spectrin or fodrin standards run in adjacent lanes and stained with Coomassie Blue (not shown) confirmed that the two polypeptides precipitated with anti-spectrin antibodies migrated in the position of the two polypeptide chains of isolated spectrin (240 and 220 K), and the two bands precipitated with anti-fodrin corresponded to the 240 and 235 K subunits of brain fodrin. Control experiments revealed that in all cases antibodies were in large excess over fodrin or spectrin (not shown). It would appear, then, that uninduced cells synthesize a low level of spectrin which then gradually increases upon treatment with DMSO. One consistent feature is that the spectrin beta chain (220 K) is produced in larger amounts over the alpha-chain, whereas the two subunits of fodrin are always Exp CellRes
152 (1984)
20
Glenney and Glenney
Fig. 3. Western blotting of spectrin and fodrin subunits using affinity purified antibodies to mammalian spectrin or fodrin. A, Porcine fodrin; B, porcine spectrin; and C avian fodrin, or D, avian spectrin were run on SDS-gels (5 % acrylamide) and stained with (a) Coomassie Blue dye; or (b) transferred to nitrocelhrlose and probed with antibodies to porcine fodrin; or (c) human spectrin followed by I*?protein A and autoradiography. Only the relevant portion of the gel and autoradiograph are shown. Note that antibodies to fodrin show no cross-reactivity with mammalian spectrin (although reacting with avian spectrin) and antibodies to human spectrin display on cross-reactivity with fodrin.
produced in equal amounts (fig. 2) (both alpha and beta chains of spectrin and fodrin have approximately equal number of methionines [7]. This result is different from that seen in avian erythropoiesis [25], in which the beta-chain is synthesized in larger amounts. This could be due to the differences in labeling (0.25-3 h vs 12 h) or reflect a basic difference in normal avian distribution compared to induction of differentiated characteristics of Friend cells. Although spectrin and fodrin exist as alphaZ-beta* tetramers and spectrin and fodrin are both present in Friend cells, it is possible that heterodimers might form. Apparently this does not occur, since spectrin and fodrin are independently immunoprecipitable (fig. 2). This is surprising since (i) spectrin is always found in fodrin-positive areas, especially at later stages of induction (fig. 1); (ii) recent results have shown that hybrid molecules can be formed between subunits of spectrin and fodrin in vitro (although with a low yield) [23]; and (iii) under conditions of immunoprecipitation used in these experiments both subunits of spectrin [24] or fodrin [ 1I] remain firmly bound together as a complex, even in the presence of excess antibody [ 111. The reason that spectrin-fodrin hybrids are not found may be due to a much higher afftnity of the spectrin or fodrin subunits for each other than for the other protein. Although a small amount of a peptide comigrating with the beta-chain of spectrin is found in immunoprecipitates using anti-fodrin (fig. 2), this may reflect a low level of such hybrids. Western blots (tig. 3) and immunofluorescence controls (fig. 1) would suggest that it is not due to cross-reactivity of the antibodies. The reason for the capping of fodrin and spectrin is also not known, but may reflect a limiting amount of the membrane receptor in induced cells. Another important question is whether any integral membrane proteins co-cap with fodrin and spectrin in these cells. Future studies will have to approach this question and ihe Friend cell system may provide an excellent model to analyze the regulation of the membrane-cytoskeleton system in a cell which contains two such important structural elements. i5.x.~ Cell Res 152 (1984)
Spectrin
and fodrin
in DMSO-treated
Friend
cells
21
REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25.
Glenney, J R, Jr, Glenney, P & Weber, K, Proc natl acad sci US 79 (1982) 4002. Bennett, V, Davis, J & Fowler, W E, Nature 299 (1982) 126. But-ridge, K, Kelly, T L Mangeat, P, J cell bio195 (1982) 478. Repasky, E A, &anger, B L & Lazarides, E, Cell 29 (1982) 821. Goodman, S R, Zagon, I S C Kulikowski, R R, Proc natl acad sci US 78 (1981) 7570. Goodman, S R & Shiffer, K, J Am phys sot (1983) C-121. Glenney, J R Jr, Glenney, P & Weber, K, J biol them 257 (1982) 9781. Tsukita, S, Tsukita, S, Ishikawa, H, Kurokawa, M, Morimoto, K, Sobue, K & Kakiuchi, S, J cell bio197 (1983) 574. Glenney, J R Jr, Glenney, P & Weber, K, J mol biol 167 (1983) 275. Glenney, J R Jr 8~ Glenney, P, Cell and muscle motility. 3 (1983) 671. - Cell 34 (1983) 503. Glenney, J R Jr, Glenney, P & Weber, K, J cell biol 96 (1983) 1491. Marks, P A & Rifkind, R A, Ann rev biochem 47 (1978) 419. Hyman, R, Bowbridge, I & Cunningham, K, J cell physiol 105 (1980) 469. Hiller, G & Weber, K, Nature 266 (1977) 181. Giloh, H & Sedat, J W, Science 217 (1982) 1252. Blase, S H, Matsumura, F & Lin, J J-C, Cold Spring Harbor symp quant bio146 (1982) 455. Amdt-Jovin, D J, Ostertag, W, Eisen, H, Klimed, F & Jovin, T, J histochem cytochem 24 (1976) 332. Friend, C, Scher, W, Holland, J & Sato, T, Proc natl acad sci US 68 (1971) 378. Eisen, H, Bach, R & Emery, R, Proc natl acad sci US 74 (1977) 3898. Pfeffer, S R & Redman, C M, Biochim biophys acta 641 (1981) 254. Marshall, L M & Hunt, R C, J cell sci 54 (1982) 97. Davis, J & Bennett, V, J biol them 258 (1983) 7757. Nelson, W J, Colaco, C A & Lazarides, E, Proc natl acad sci US 80 (1983) 1626. Blikstad, I, Nelson, W J, Moon, R T & Lazarides, E, Cell 32 (1983) 1081.
Received September 27, 1983 Revised received November 21, 1983
Printed
in Sweden
Exp Cell
Res I52 11984)