Role of Carbohydrate in Biological Function of the Adhesive Glycoprotein Fibronectin

Proc. Natl. Acad. Sci. USA Vol. 76, No. 7, pp. 3343-3347, July 1979 Cell Biology

Role of carbohydrate in biological function of the adhesive glycoprotein fibronectin

KENNETH OLDEN*, ROBERT M. PRATTt, AND KENNETH M. YAMADA* *Laboratory of Molecular Biology, National Cancer Institute, National Institutes of Health; and tLaboratory of Developmental Biology and Anomalies, National Institute of Dental Research, Bethesda, Maryland 20205 Communicated by DeWitt Stetten, Jr., March 28, 1979

ABSTRACT We have investigated the role of the carbohy- MATERIALS AND METHODS drate moiety in the biological activity of fibronectin in vitro by Cell Culture. Secondary chicken embryo fibroblasts (CEF) using tunicamycin to inhibit the glycosylation of this glycoprotein. Tunicamycin is a glucosamine-containing antibiotic were cultured in plastic tissue culture dishes in Ham's FlO that specifically inhibits glycosylation of protein asparaginyl medium supplemented with 10% (vol/vol) tryptose phosphate residues mediated by dolichol pyrophosphate. Fibronectin broth, 5% (vol/vol) heat-inactivated calf serum, 0.056% sodium synthesized in the presence of 0.5 ptg of tunicamycin per ml was bicarbonate, 50 units of penicillin per ml, 50 Ag of streptomycin not glycosylated, as determined by amino sugar analysis, lack per ml, and 2 mM glutamine as described (1, 12). of incorporation of [14Ciglucosamine and [3H]mannose, and Isolation of Fibronectin. Confluent primary cultures were concanavalin A binding studies. Nonglycosylated fibronectin that was isolated from chicken embryo fibroblasts and added preincubated for 3.5 hr in culture medium with or without 0.50 to transformed cells in vitro was as effective as the glycosylated ,ug of tunicamycin per ml. Cells were passaged with 0.25% protein in promoting a more normal fibroblastic phenotype, trypsin, replated at the same cell density in the preincubation including cell flattening, elongation of cell processes, and par- medium, and incubated for 24 hr. Control and tunicamycin- allel alignment of cells. The nonglycosylated protein was also treated cultures were homogenized separately in 50 mM sodium as effective as the glycosylated species in mediating cell at- phosphate, pH 11.0/1% Triton X-100, and the pH was re- tachment to collagen and spreading on plastic, as well as in to 11.0 with NaOH. The were centri- agglutination of formalin-fixed sheep erythrocytes. The non- adjusted homogenates glycosylated protein was twice as sensitive as the glycosylated fuged at 100,000 X g at 4°C for 1 hr to sediment insoluble protein to proteolytic hydrolysis in vitro as had been suggested material. The supernatant was diluted 1:10 with 100 mM Na by previous studies with intact cells [Olden, K., Pratt, R. M. & phosphate, pH 7.0/1% Triton X-100/50 mM NaCl. Fibronectin Yamada, K. M. (1978) Cell 13,461-4731. We conclude that the was isolated from the homogenates by incubation with antifi- carbohydrate moiety of fibronectin is not required for the bronectin covalently coupled to Sepharose 4B. The column was mediation of a number of biological activities characteristic of washed with 5 vol of 0.1 M borate buffer, pH 8.5/0.5 M NaCl. this glycoprotein. The supernatant containing fibronectin was slowly applied to the column (10 ml/hr), and unbound material was eluted with The role of the carbohydrate moiety covalently attached to the 0.1 M borate buffer. Nonspecifically adsorbed protein was proteins is not known. In previous studies, inhibition of glyco- eluted from the resin with 0.1 M borate buffer, pH 8.5/0.5 M sylation of fibronectin, a major cell surface glycoprotein (also NaCl. Fibronectin was then eluted from the resin with 0.2 M known as CSP or LETS), did not affect its synthesis or secretion acetic acid, pH 2.9/0.5 M NaCl. The eluates were immediately (1-4). Similar results have been obtained for other glycoproteins adjusted to pH 11.0 with NaOH, dialyzed overnight against (5-7). These results left the possibility that the biological ac- buffer A (0.15 M NaCl/1.0 mM CaCl2/10 mM cyclohexyl- tivities (e.g., the cell surface adhesive interactions affected by propane sulfonic acid, at pH 11.0) and stored in 200-yl aliquots fibronectin) might require the carbohydrate moiety. at -60°C. Because tunicamycin inhibits the glycosylation of fibronectin Protein Synthesis and Glycosylation. Protein synthesis was (1), the major cell surface glycoprotein of chicken embryo fi- measured by the incorporation of L-[U-'4C]leucine (2 ,Ci/ml, broblasts, we used this antibiotic to evaluate the influence of 1 Ci = 3.7 X 1010 becquerels). Glycosylation of fibronectin and the carbohydrate moiety on many of the known biological ac- total macromolecular trichloroacetic acid-insoluble material tivities of fibronectin. This protein is an adhesive glycoprotein were measured by the incorporation of D-[U-'4C]glucosamine that helps maintain normal cell morphology, cell surface ar- (2 tiCi/ml) and D-[2-3H]mannose (0.5 ,Ci/ml) and also by chitecture, and cell interactions in cultured cells (2, 3, 8-11). amino sugar analysis (13). Radioactively labeled fibronectin We present evidence that the carbohydrate moiety is not was immunoprecipitated from cell homogenates and homog- required for these biological activities mediated by fibronectin enized in 2% sodium dodecyl sulfate as described (1, 4, 12). in cell culture and in assays for cell adhesiveness. In addition, Sodium dodecyl sulfate gel electrophoresis was performed we directly confirm a previous suggestion, based on in vivo with 1-mm-thick polyacrylamide slab gels according to the studies, that nonglycosylated fibronectin is more susceptible method of Laemmli (14) with stacking gels of 4% acrylamide to proteolytic degradation than the glycosylated protein (1). and resolving gels of 7.5% acrylamide as described (1, 4, 12, 15). Gels were stained with Coomassie blue and destained in 7% The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "ad- acetic acid. The polypeptide band corresponding to fibronectin vertisement" in accordance with 18 U. S. C. §1734 solely to indicate was cut from the gel and dissolved in 1.5 ml of 30% hydrogen this fact. peroxide. Radioactivity was measured by liquid scintillation 3343 Downloaded by guest on October 2, 2021 3344 Cell Biology: Olden et al. Proc. Natl. Acad. Sci. USA 76 (1979) Table 1. Incorporation of leucine and sugars into cellular fibronectin and trichloroacetic acid-insoluble fraction of cells Incorporation into fibronectin/ltg Incorporation, cpm/,ug total cell protein total cell protein Precursor Control Tunicamycin % inhibition Control Tunicamycin % inhibition L-Leucine 1432 ± 77 1045 ± 91 27 i 3 43 ± 6 23 ± 3 47 D-[U-14C]Glucosamine 1125 ± 93 259 ± 37 77 ± 2 32 ± 7 0.63 ± 1 98 D-[2-3H]Mannose 4519 + 203 181 ± 21 96 ± 3 124 ± 12 1.24 ± 1 99 Cells were incubated for 24 hr with the radioactive precursors in medium with and without tunicamycin. Then fibronectin was immunoprecipitated and electrophoresed, and radioactivity was quantitated. The data presented represent the average of three experiments (plus and minus represent the range). In a typical experiment with 0.5,ug tunicamycin per ml, the amount of [14C]leucine incorporated into total acid-precipitable material was 1365 cpm/,gg of protein for untreated cultures and 954 cpm/,ug of protein for tunicamycin-treated cultures. To correct for this difference in inhibition (27%) of protein synthesis, we applied 52 ,l of tunicamycin-treated cell homogenate to the gel compared to 40 Al of the untreated cell homogenate. Similar aliquots were used for determination of the percentage inhibition of glucosamine or mannose incorporation into total acid-precipitable material. The percentage inhibition of leucine incorporation into fibronectin was determined by dividing the cpm in fibronectin from tunicamycin-treated cultures by that found in fibronectin from untreated cultures. The percentage inhibition of glucosamine or mannose incorporation into fibronectin was determined by dividing the glucosamine/leucine ratio of tunicamycin-treated culture by the glucosamine/leucine ratio of untreated cultures. We emphasize that the inhibition of glycosylation is therefore expressed as the inhibition for the fibronectin actually synthesized and not merely as inhibition of overall carbohydrate incorporation into fibronectin.

spectrometry after dilution to 20 ml with Aquasol (New En- The radiochemicals L-[U-14C]leucine (specific activity 325 gland Nuclear). Ci/mole), D-[U-14C]glucosamine (specific activity 238 Ci/ Other Procedures. Affinity-purified goat antibodies against mole), and D-[2-3H]mannose (specific activity 2 Ci/mmol) were fibronectin were covalently coupled to CNBr-activated Seph- obtained from New England Nuclear. Phenylmethylsulfonyl arose (Pharmacia) at a ratio of 10 mg of antibody protein per fluoride, Triton X-100, and Pronase (Streptomyces griseus g of hydrated gel following the procedures recommended by protease, grade B, no. 53702) were purchased from Calbiochem, the company (Pharmacia). Hemagglutination of formalinized electrophoresis reagents from Bio-Rad, and CNBr-activated sheep erythrocytes and effects of fibronectin on cell spreading Sepharose 4B from Pharmacia. Plastic tissue culture dishes were and morphology were assayed as described (15, 16). Protein was from Costar, and nonwettable bacteriological petri dishes for determined by the method of Lowry et al. (16), with bovine the collagen attachment assay were obtained from Falcon. serum albumin as protein standard. Materials. Tunicamycin was a gift from Gakuza Tamura via RESULTS the Drug Evaluation Branch of the National Cancer Institute. Isolation of glycosylated and nonglycosylated fibronectin Table 2. Effect of tunicamycin on amino sugar composition of cellular fibronectin To quantitate the inhibition of fibronectin glycosylation by Amino sugar/ tunicamycin, we isolated the protein by immunoprecipitation cellular fibronectin from treated and untreated cultures labeled for 24 hr with ei- monomer, ther [3H]mannose or [14C]glucosamine. The results are shown Addition mol/mol % inhibition in Table 1. The average inhibition of glucosamine and mannose incorporation into fibronectin isolated by immunoprecipitation None 27 was greater than 98%. 29 A similar result (98% inhibition) was obtained in comparisons 21 of amino sugar analyses of nonglycosylated and glycosylated fibronectin that had been purified by antibody-affinity chro- Tunicamycin 0.44 98 matography (Table 2). As demonstrated previously with un- 0.53 98 purified material (1), the nonglycosylated fibronectin was no 0.39 98 longer capable of binding to concanavalin A (results not shown; Cellular fibronectin was isolated from cell homogenates by anti- see ref. 1). Additional support for this conclusion is illustrated body-affinity chromatography. Protease (Pronase) digestion of fi- in Fig. 1, in which it is evident that fibronectin isolated from bronectin was carried out as described (17). According to this proce- an molecular dure, fractions containing 50 yg of protein were lyophilized and tunicamycin treated cells migrates with apparent resuspended in a solution containing 250 ,ul of 1 mM Tris.HCl (pH weight (Mr approximately 209,000) less than that of the gly- 8.0), 100 Al of a 1% solution of protease, and 10 mM CaCl2. The di- cosylated protein (220,000) (Fig. 1A). Furthermore, the lower gestion was carried out for 48 hr at 37°C. Preparations were protected Mr species of fibronectin did not incorporate mannose (Fig. 1B) from bacterial contamination by periodic addition ofa drop of toluene. or glucosamine (Fig. 1C). The liberated amino acids were separated from the glycopeptides by These results confirm and extend our other studies (refs. 14, Bio-Gel P-6 (200-400 mesh, Bio-Rad Laboratories) gel filtration that chromatography on a 1 X 60 cm column as described (17). The gly- 18, and 19; unpublished data) which indicated tunicamycin copeptide fractions were lyophilized, and the monosaccharides were inhibited the glycosylation of fibronectin, producing a non- released by hydrolysis in 4 M HCl by heating in a boiling water bath glycosylated protein that has now been isolated by antibody- for 12 hr. The hydrolysate was neutralized with NaOH and assayed affinity chromatography. for amino sugar content (13). The N-acetylglucosamine standards were also treated with 4 M HCl and neutralized before amino sugar Biological activity of nonglycosylated fibronectin analysis. We plotted the entire spectrum of the reaction mixture and obtained a single peak at 530 nm; this indicates that all the acetylated To examine the biological activity of fibronectin, we used three hexosamines had been completely hydrolyzed and that glucose and approaches. First, we directly measured the effects of fibro- fructose were not present in the reaction mixture (13). nectin on cell-cell adhesiveness in a model system that measures Downloaded by guest on October 2, 2021 Cell Biology: Olden et al. Proc. Nati. Acad. Sci. USA 76 (1979) 3345

0 00 X 30- x 30 -x30E E IE E C1. ~ I0. C 20 -20 20- 0

10 10. 10

o 10 20 30 0 10 20 30 0 10 20 30 Slice FIG. 1. Effect of tunicamycin on the incorporation of [14C]leucine (A), [3H]mannose (B), and D-[14C]glucosamine (C) into cellular fibronectin. Confluent monolayer cultures were incubated with tunicamycin for 3 hr, then trypsinized and replated at the same cell density in medium with or without 0.5 pg oftunicamycin per ml. After incubation for 24 hr in double-labeling medium containing [14C]leucine and [3H]mannose, fibronectin was isolated from control and tunicamycin-treated cultures by affinity chromatography. Fibronectin from the D-f14C]glucosamine-labeled cultures was isolated by immunoprecipitation. Fractions containing 2.5 ,ug of fibronectin from control and tunicamycin-treated cultures were mixed together and coelectrophoresed in sodium dodecyl sulfate/polyacrylamide gels. (A) [14C]Leucine-labeled protein profile in the high molecular weight range of double-labeled, tunicamycin-treated cultures; (B) [3H]mannose-labeled glycoprotein profile of the double-labeled, tunicamycin-treated cultures; (C) [14C]glucosamine-labeled glycoproteins isolated from a parallel culture and electrophoresed separately. The gels were subsequently stained with Coomassie blue, sectioned into 1-mm slices, and dissolved in 30% hydrogen peroxide. Radioactivity was measured in Aquasol (New England Nuclear) by liquid scintillation spectrometry. Gels were calibrated for molecular weight with the following standards: chicken gizzard filamin, Mr 250,000 (a gift from P. Davies); rabbit skeletal muscle myosin, Mr 200,000; RNA polymerase, Mr 150,000 and 160,000; phosphorylase a, Mr 94,000; bovine serum albumin, Mr 68,000; and ovalbumin, Mr 43,000.

hemagglutinating activity. Second, we performed reconstitution Another approach to the evaluation of the role of the car- experiments in which fibronectin was added to transformed bohydrate moiety of fibronectin in mediating biological effects cells deficient in this glycoprotein and measured its capacity is to reconstitute it on the surfaces of transformed cells on which to restore the fibroblastic morphology characteristic of non- it is greatly diminished (20). For these studies, fibronectin was transformed cells. Third, we evaluated fibronectin in assays for added to the culture medium to various concentrations as de- cell attachment to collagen and cell spreading. scribed (9). Nonglycosylated fibronectin markedly altered the Purified nonglycosylated fibronectin agglutinated formali- morphology of simian virus 40-transformed 3T3 cells in culture nized sheep erythrocytes (Fig. 2). It was as effective as the (Fig. 3). Control cultures receiving control buffer alone showed glycosylated protein, with half-maximal agglutination produced no morphological changes (Fig. 3A). The morphological al- by 2-4 jig of protein per ml. This finding indicates that the terations induced by the addition of fibronectin include cell carbohydrate moiety of fibronectin is not required for its flattening, elongation of processes, and parallel alignment of hemagglutinating activity. cells, resulting in a more fibroblastic appearance (Fig. 3 B and 1 2 3 4 5 6 7 8 9 10 11 12 C). No differences were found between the glycosylated and nonglycosylated protein either qualitatively or quantitatively A Q with respect to restoration of the normal fibroblastic morphology. Both species of fibronectin were effective at concen- B~~~~~i% %t '~i# '%#XX trations as low as 1 pg/ml. C'*>8. A't,. AJW*s )e Glycosylated and nonglycosylated fibronectin were also compared with respect to their capacity to promote the spreading of baby hamster kidney (BHK) cells on the surface of plastic tissue culture dishes (21). The extent of cell spreading in the presence of various concentrations of purified glycosyl- H~~~~0 JX ated and nonglycosylated fibronectin is shown in Fig. 4. It is apparent that nonglycosylated fibronectin is as effective as the glycosylated protein in promoting cell spreading, with half- maximal effects at a protein concentration of 1 ,g/ml. FIG. 2. Hemagglutination of formalinized sheep erythrocytes Finally, glycosylated and nonglycosylated fibronectin were by fibronectin. Lanes A and B show the amount of hemagglutination compared with respect to collagen binding and cell attachment. with the following concentrations of nonglycosylated and glycosylated Fibronectin mediates the attachment of certain fibroplastic cells fibronectin, respectively: 1, 500 ,gg/ml; 2, 250 jg/ml; 3, 125 ,g/ml; 4, to collagen-coated substrata in tissue culture (22, 23). The 62 ,ug/ml; 5, 31 jg/ml; 6, 16 ,ug/ml; 7, 8 iig/ml; 8, 4 ug/ml; 9, 2 ,g/ml; nonglycosylated protein was equal in activity to the glycosyl- 10, 1 Ag/ml; 11, 0.5 ,gg/ml; and 12, no additions except phosphate- buffered saline. Well 12 shows a button ofunagglutinated cells on the ated protein in mediating attachment of Chinese hamster ovary bottom of the well; the other wells show a carpet of cells agglutinated (CHO) cells to dishes coated with type I collagen (Fig. 5). to varying degrees, with half-maximal agglutination at 2-4 l.g of fi- Previous studies on the effects of tunicamycin on fibronectin bronectin per ml. Lanes C and D are repeats of lanes A and B, re- suggested that the carbohydrate moiety was required to stabilize spectively. this protein against proteolytic degradation (1). To determine Downloaded by guest on October 2, 2021 %of%0j46 Cell Biology: Olden et al. Proc. Natl. Acad. Sci. USA 76 (1979)

100

75 ccm ._

50 U

251 7ffi

**_-<0oS j i i

2.0 4.0 6.0 8.0 10.0 Fibronectin, ,Ag/ml ant'X/fj¢t.; mug-'ffiW'b§>S;tjob S F 11 me FIG. 4. Effect of isolated fibronectin on the spreading of non- Liszt_5S/09"j;*j-t transformed cells in vitro. BHK cells grown in monolayer cultures were removed from the dish by treatment with trypsin; then serum was added to inhibit further proteolysis. The cells were collected by centrifugation and washed twice in adhesion medium (150 mM eAt'A NaCl/3 mM KCl/1 mM CaCl2/0.5 mM MgCl_/6 mM Na2HPO4/1 mM M-_-Am""--s'= ''-''An; =_=-tt¢ KH2PO4), and 0.5 X 106 cells in 1.0 ml of adhesion medium were In wo ;=%S > < In w As; s-t-*|r.E.S. _. In A__ transferred to 35-mm culture dishes and incubated for 45 min in _SAB0iSa;pi/ medium containing the following concentrations of glycosylated (0) 5R..'..._ .>.>s .A_..t or nonglycosylated (0) fibronectin: 0jig/ml (control), 1 ug/ml, 5 ,:,,,,.__ . and . 10 Ofas Cell 4^.-He ,ug/ml, jsg/ml. spreading at each concentration was scored by counting 250 cells and calculating the percentage ofcells that were ? *_ *,_ .Of * ..,.#Jan . > fully spread (no longer refractile by phase contrast microscopy and surrounded by a continuous zone of lamellar cytoplasm). The error bars represent the range. increases cell-cell and cell-substratum adhesiveness in a variety * m.s .. _...... I'._ of cellular adhesion assays (2, 3, 8, 9, 11, 12). Consistent with this function, the reconstitution of purified fibronectin on transformed cells results in reversion to a more normal fibroblastic morphology, adhesiveness, cell surface architecture, motility, microfilament bundle organization, and alignment at confluence (2, 3, 8, 9, 11, 12). The results presented in this paper indicate that the carbohydrate moiety of cellular fibronectin is not for the FIG. 3. Effect of isolated fibronectin on the morphology and required alignment of transformed cells in vitro. Glycosylated or nonglycosylated fibronectin was added to a final concentration of 50 pg/ml to 70 cultures of SV1-3T3 cells and incubated for 48 hr. (A) Control to which was added buffer A;,(B) cultures containing glycosylated fibronectin; (C) cultures containing nonglycosylated fibronectin. 60 0 50 whether nonglycosylated fibronectin is inherently more sen- x sitive to proteases than the glycosylated protein, we incubated both species of the isolated protein with protease (Pronase). The 40 rate of release of trichloroacetic acid-soluble ['4C]leucine from each preparation is shown in Fig. 6. It is apparent that the " 30 nonglycosylated protein fraction is degraded to soluble peptides and amino acids approximately 2-fold more rapidly than the 0 20 / glycosylated protein. zI6 The studies described here support the conclusion that the 10 carbohydrate moiety of fibronectin is not required for a variety of biological activities mediated by this glycoprotein. / 0 5 10 15 20 25 DISCUSSION Protein, jig/ml We have isolated fibronectin, the major cell surface glycopro- FIG. 5. Effect of isolated fibronectin on attachment of cells to collagen. Monolayer cultures of CHO cells were trypsinized, washed, tein present on the surface of chicken embryo fibroblasts in and plated at a concentration of 105 cells per ml of alpha modified culture, and have investigated the role of its carbohydrate Eagle's minimal essential medium (Flow Laboratories, Rockville, MD) moiety in the mediation of its biological functions. Fibronectin in 35-mm plastic bacteriological dishes coated with 10 jig of type I is reported to be biologically active as an adhesive protein that collagen as described (23). Downloaded by guest on October 2, 2021 Cell Biology: Olden et al. Proc. Natl. Acad. Sci. USA 76 (1979) 3347 have shown that the carbohydrate moiety of the vesicular sto- 40 matitis virus glycoprotein is not required for the specific in- fectivity of the virion (28). The effect of the carbohydrate moiety on solubility and sensitivity to proteases suggests that it is important to preserve 30 the structural features of the glycoprotein but may have no other functional significance.

0~~~~~ We thank Dr. Gakuzo Tamura for the generous gift of tunicamycin, Miss Elizabeth Lovelace and Miss Annie Harris for aid in cell culture, 20 / Mrs. Wilma Davis for typing the manuscript, Dr. Hynda Kleinman for type I lathyritic rat skin collagen, and Dr. Peter Davies and Dr. Michael Gottesman for critical review of the manuscript. 0~~/ 0~~~~ 10 * 1. Olden, K., Pratt, R. M. & Yamada, K. M. (1978) Cell 13, 461- 0-1 473. 2. Yamada, K. M., Olden, K. & Pastan, I. (1978) Ann. N. Y. Acad. Sci. 312, 256-277. 3. Yamada, K. M. & Olden, K. (1978) Nature (London) 275, 0 5 10 15 20 25 30 179-184. Time, min 4. Olden, K. & Olden, A. T. (1978) in Recent Advances in Cancer and Molecular Biology, ed. Guillory, W. (Univ. of Utah Press, FIG. 6. Protease digestion of fibronectin. Fifty micrograms of Salt Lake City, UT), pp. 19-41. fibronectin, isolated by antibody-affinity chromatography, was in- 5. Eagon, P. K. & Heath, E. C. (1977) J. Biol. Chem. 252, 2372- cubated with 0.05 sg of the broad spectrum protease Pronase per ml 2383. in 0.1 M sodium phosphate buffer (pH 7.4) at 37°C. Aliquots (100-,ul) 6. Struck, D. K., Suita, P. B., Lane, M. D. & Lennarz, W. J. (1978) were removed at the indicated intervals and the radioactivity in both J. Biol. Chem. 253,5332-5337. the soluble and precipitable fractions was determined after precipi- 7. Trowbridge, I. S., Hyman, R. & Mazauskas, C. (1978) Cell 14, tation with trichloroacetic acid.O 0, Nonglycosylated fibronectin; 21-32. * --- 0, glycosylated fibronectin. 8. Yamada, K. M., Yamada, S. S. & Pastan, I. (1975) Proc. Natl. Acad. Sci. USA 72,3158-3162. expression of a number of the biological activities thought to 9. Yamada, K. M., Yamada, S. S. & Pastan, I. (1970) Proc. Natl. Acad. Sci. USA 73, 1217-1221. of this For no apparent be.characteristic glycoprotein. example, 10. Ali, I. U., Mautner, V., Lanza, R. & Hynes, R. 0. (1977) Cell 11, differences were detected between glycosylated and nongly- 115-126. cosylated fibronectin when tested for activity in mediating 11. Vaheri, A. & Mosher, D. F. (1978) Biochim. Biophys. Acta 516, attachment of cells to collagen. In addition, both species of fi- 1-25. bronectin were equally effective in hemagglutination of sheep 12. Olden, K. & Yamada, K. M. (1977) Cell 11, 957-969. erythrocytes and in the restoration of a more normal fibroblastic 13. Ashwell, G. (1957) Methods Enzymol. 3,95-97. phenotype to transformed cells. However, nonglycosylated 14. Laemmli, U. K. (1970) Nature (London) 227,680-685. fibronectin was shown to be more sensitive to proteases in 15. Olden, K., Willingham, M. C. & Pastan, I. (1976) Cell 8, 383- vitro. 390. In a previous study from this laboratory, we demonstrated 16. Lowry, O. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. no requirement for the carbohydrate moiety in the secretion (1951) J. Biol. Chem. 193,265-275. & T. we the 17. Robbins, P. W., Krag, S. S. Liu, (1977) J. Biol. Chem. 252, of fibronectin (1), and in the present study show that 1780-1785. carbohydrate does not appear to perform a specific biological 18. Olden, K., Pratt, R. M., Jaworski, C. & Yamada, K. M. (1979) function. However, although the carbohydrate unit is not re- Proc. Natl. Acad. Sci. USA 76,791-795. quired for the secretion or adhesive biological activity of fi- 19. Pratt, R. M., Yamada, K. M., Olden, K., Ohanian, S. H. & Hascall, bronectin, it appears to stabilize the protein against proteolytic V. C. (1979) Exp. Cell Res. 118, 245-252. degradation (1). In addition, its covalent attachment to the 20. Yamada, K. M. & Weston, J. A. (1974) Cell 5,75-81. protein may still serve some undetermined general or specific 21. Grinnell, F., Hays, D. G. & Minter, D. (1977) Exp. Cell Res. 110, function. For example, the carbohydrate unit may be required 175-190. for increased solubility (24) or for receptor or immunological 22. Klebe, R. J. (1974) Nature (London) 250,248-251. specificity (7, 25, 26). 23. Kleinman, H. K., McGoodwin, E. B., Rennard, S. I. & Martin, G. The findings reported here are consistent with earlier studies R. (1979) Anal. Biochem. 94,308-312. 24. Leavitt, R., Schlesinger, S. & Kornfeld, S. (1977) J. Biol. Chem. that suggested that several enzymes have biological activities 252,9018-9023. that appear to be independent of the bound carbohydrate (27). 25. Ashwell, G. & Morell, A. G. (1974) Adv. Enzymol. 41,99-128. However, most of these studies involved the destruction of the 26. Neufeld, E. F., Lim, T. W. & Shapiro, L. J. (1975) Annu. Rev. carbohydrate moiety with glycosidases or periodate. These Biochem. 44, 357-376. types of studies are important but are not definitive because not 27. Eylar, E. H. (1965) J. Theor. Biol. 10, 89-113. all of the carbohydrate units are removed or destroyed by these 28. Gibson, R., Leavitt, R., Kornfeld, S. & Schlesinger, S. (1978) Cell procedures. In addition, more recent studies with tunicamycin 13,671-679. Downloaded by guest on October 2, 2021