subunit interactions in human plasma fibronectin

Vol. 124, No. 3, 1984 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

November 14, 1984 Pages 718-725

SUBUNIT INTERACTIONS IN HUMAN PLASMA FIBRONECTIN

Randall M. Robinson and Jan Hermans

Department of Biochemistry, School of Medicine University of North Carolina, Chapel Hill, NC 27514

Received September 21, 1984

SUMMARY. The fibronectin molecule was split chemically into its two constituent chains (mol. wt. 220,000) by mild reduction with dithiothreitol. However, physical properties (molecular weight and diffusion coefficient from light scattering, and elution in gel exclusion chromatography) remained those of intact fibronectin, except (reversibly) in the presence of denaturants which also change the conformation of non-reduced fibronectin to a more open form. Similarly, during digestion of fibronectin by plasmin to fragments of molecular weight less than 200,000, the light scattering intensity drops to roughly half in 30% glycerol but not in the absence of glycerol. These results suggest that the compact conformation of native fibronectin is stabilized by specific noncovalent contacts between constituent chains. 0 1984 Academic Press. Inc.

INTRODUCTION. Human plasma fibronectin, a glycoprotein with a molecular

weight of 440,000, is composed of two polypeptide chains of approximately

equal size, connected by two disulfide bridges near their carboxyl ends (1).

The molecule is distinguished by its affinity for many biological macromole-

cules, such as collagen (2,3), heparin (4,5), and fibrin (4,6). This

affinity is an essential part of fibronectin's purported biological function,

which is to link together two or more surfaces by attaching itself-to macro-

molecules on each (7,8).

Recent experimental studies point to the existence of interactions within

or between the two subunit chains, of fibronectin (9-11). These interactions

are responsible for the compact shape of the circulating form of the protein,

which is converted to an open, extended form at high pH, at high salt concen-

tration, and in the presence of urea or glycerol (10-12). Also, binding of

fibronectin to certain collagen peptides, for which it has high affinity, is

accompanied by a transition to a more open form (10).

Abbreviations: SDS = sodium dodecyl sulfate, DTNH = 5,5'-dithio bis-(2-nitrobenzoic acid)

0006-291X/84 $1.50 Copyright 0 1984 by Academic Press. Inc. All rights of reproduction in any form reserved. 718


We here present evidence that the two chains of fibronectin interact to

form compact, more-or-less native structures even when their carboxyl ends

have been detached from one another by reduction or limited proteolytic

cleavage.

MATERIALS AND METHODS

Light scattering measurements and purification of human plasma fibronectin were performed as described previously (11). The formation of half molecules during reduction of fibronectin (at 0.5 - 1.2 mg/ml, with 16 mM dithiothreitol under nitrogen at pH 8.2) was followed by measurement of the light scattering intensity at 90°. Samples for chromatography and dynamic light scattering were prepared by reiduction of fibronectin at 5-8 mg/ml with 10 or 16 mM dithiothreitol for 15 minutes in 4 M urea at room temperature under nitrogen either at pH 7.7 or at pH 8.2. Free sulfhydryl groups were subsequently blocked by reaction in the dark with excess iodoacetamide. If required, urea was removed by gel chromatography. Alternatively, in order to measure the number of free sulfhydryl groups, the reduced mixture was reacted with DTNB, gel chromatographed and reacted with dithiothreitol to liberate the bound reagent; the DTNB content of the protein was then determined from the absorbance at 412 and 280 run (13). Electrophoresis in 4% polyacrylamide gels in solutions containing 1% SDS, followed by staining with Coomassie blue, was used to follow the progress of reduction and digestion.

RESULTS

Changes during Reduction. During reduction of fibronectin with dithiothreitol,

the scattered intensity at first remained nearly constant, then slowly began

to increase, and never decreased (Figure 1). This confirmed previous results

obtained by Williams et al. (14). SDS-gel electrophoresis of samples prepared

Figure 1. Time dependence of relative scattered intensity (at 90°) during reduction of fibronectin with 16 mM dithiothreitol. Solutions contained 0.1 M sodium phosphate, 0.15 M NaCl, 1 mM EDTA, pH 8.0, without (e) and with (0) 4 M urea.

719


TABLE I MOLECULAR WEIGHT AND DIFFUSION COEFFICIENT OF FIBRONECTIN AND REDUCED FIBRONECTIN

Molecular Weight Diffusion SDS-polyacryl- Coefficient

Light Scattering amide gel (corrected) electrophoresis x 10'

Unreduced, low ionic strengtha Unreduced, ionic strength 0.35 Mb

470,000 440,000 2.5

Reduced, low ionic strengtha 470,000 440,000 2.1

Reduced, ionic strength 0.35 Mb 450,000 220,000 2.4 450,000 220,000 2.05

it 0.2 mH sodium phosphate, pH 7.2. 0.1 H sodium phosphate, 0.15 M sodium chloride, 1 mU EDTA, pH 7.2

under identical conditions resulted within about 1 hour of reduction in a

single band at molecular weight circa 220,000. Thus we found that in the

absence of denaturant, the complete reduction of intersubunit disulfide

bonds did not result in a decrease in particle size as measured by light

scattering intensity.

The reduction as assessed by increasing intensity of the 220,000 molecu-

lar weight band in SDS-gels proceeded much more rapidly in the presence of

urea. When the reduction was performed in 4 M urea, the scattered intensity

quickly dropped to about one half of the initial value, i.e., the half

molecules did not stay together under these conditions.

Analysis of Reduced Fibronectin. Both (apparent) molecular weight and dif-

fusion coefficient as measured by light scattering of reduced and blocked

fibronectin were virtually unchanged from those of native fibronectin (Table I).

The association of half molecules also was evident from results of

column chromatography experiments. For these experiments an approximately

equal mixture of whole and half fibronectin molecules (as judged from gel

electrophoresis in SDS) was prepared by reduction in 4 M urea, under nitrogen

at pH 7.7 for 15 minutes. When this mixture was chromatographed over Sepharose

4-B, separation of whole and half molecules was not achieved at low ionic

strength even in the presence of 1 M urea (Figure 2, panel B), but required a

higher concentration of either urea or salt (Figure 2, panels A and C. The

composition of the material of each peak was confirmed via gel electrophoresis

720


LA!! 20 40 60 20 40 60

-!I 20 40 60

FRACTION NUMBER

Figure 2. Chromatography of a mixture of whole and half molecules of fibronectin on a Sepharose 4-B column. (A) in 0.1 M phosphate, 0.15 H NaCl, 1 M urea; (B) in 2 mM phosphate, 1 M urea; (C) in 2 mM phosphate, 3 M urea.

in SDS). Thus, at least 4 M urea was required to effect separation of whole

and half molecules at low ionic strength, while 1 M urea was sufficient at

moderate ionic strength (0.35 M).

The effectiveness of urea and salt in dissociating half molecules of

fibronectin paralleled their effects on the conformation of the whole molecule

as determined from the diffusion coefficient of the intact fibronectin

molecule (Figure 3; for unknown reasons, the diffusion coefficient of intact

fibronectin at low ionic strength was less than measured in other experiments;

cf. Table I and reference 11).

The extent of reduction of the half molecules was determined from their

content of 14 free sulfhydryls, as measured in an experiment in which DTNB

instead of iodoacetamide was used as a blocking reagent, followed by gel

I 1 I 1 I

UREA2CONCEiTRATI06N ( M 1 8

Figure 3. Dependence of the diffusion coefficient, D (and frictional coefficient, f = kT/D) of fibronectin on urea concentration at low (0) and at moderate (0) ionic strength (2 mM phosphate, no added salt and 0.1 M phosphate, 0.15 M NaCl, respectively).

721


Y F 4 iii

% I

30 60

TIME (mid

Figure 4. Time dependence of relative scattered intensity (at 90') during digestion of fibronectin by plasmin. (Conditions were: 25Y, 0.1 M phosphate, 0.15 M NaCl, 1 mM EDTA, pH 7.2, 2.5 casein units of plasmin per mg of fibronectin, in absence (0) and in presence (0) of 30% glycerol.

exclusion chromatography and quantitation of the bound DTNB (see Methods).

Given the presence of 2 free sulfhydryls (15,16) per native chain, and the

formation of 2 additional ones when the interchain disulfides are reduced

(1,17), one concludes that an average of 5 reduced (of approximately 30

total, 18) intrachain disulfides are present in the half molecules studied

here.

Digestion with Plasmin. Qualitatively similar effects were noticed when

fibronectin was digested with plasmin in the absence and in the presence of

30% glycerol (Figure 4). Gel electrophoresis in SDS of the digests showed

the presence of the expected fragments (19), and the absence of fragments

with a molecular weight above 200,000 after approximately 30 minutes in both

solutions. In the absence of glycerol, the decrease in apparent molecular

weight is less even than the small amount expected due to removal of a

27,000 dalton domain from the amino terminus of each chain (19). Only in

the presence of glycerol does the apparent particle size approach anywhere

near 50% of the starting value as the digestion proceeds.

DISCUSSION

Evidence about the molecular shape of fibronectin includes the extended,

somewhat variable shapes seen by electron microscopy (20-24), the fragmentation

722


into globular, partly functional, domains by specific proteolytic cleavage (25),

and physical measurements of properties related to molecular size and shape in

solution (9-11). From these studies a description of the fibronectin molecule

has emerged as a chain of independent globular domains connected by short

flexible segments of polypeptide; the length of the molecule measured along this

chain is circa 140 nm, and its width varies around 2 nm. The overall shape of

the molecule varies reversibly from a compact form to an almost extended

conformation, depending on solvent conditions.

The conformation is compact under physiological conditions. From an

absence of changes of the circular dichroism when 1.5 M guanidinium chloride,

a stronger denaturant than urea, is added to the solution (12), it appears

that the change of the molecular shape does not depend on changes of the

conformation of folded domains, but on changes of the interaction between

domains.

A compact model of fibronectin was first proposed by Hb'rmann (28); in

this model each of the two polypeptide chains is folded back on itself so

that domains of opposite net charge are juxtaposed. In addition , presumably

more specific, complementary binding sites can maintain intra- or inter-

molecular connections, the former in the compact form, and the latter in

fibrils formed by fibronectin in the presence of heparin (28) or polyamines

(29) - An alternate model has been proposed by Rocco et al. (11,30); in this

model the two halves of the molecule interact with one another to produce a

compact particle. Both models are adequate to explain the measured solution

properties, while the model in which the chains cross is the more compatible

with the electron micrographs of the compact form (11,22-24).

The lack of change of apparent molecular weight upon limited reduction or

proteolysis indeed suggests that the association is, at least in part, quite

specific, and not simply a matter of attraction between domains with opposite

net charge. It is also more than likely that inter-chain contacts are important

in forming the compact conformation, as in the model proposed by Rocco et al.

(11,30).

723


Our observations also indicate that it is difficult, if not im-

possible, to establish a requirement of complete versus one of partial

structure for expression of fibronectin's various biological functions, by

doing experiments with half molecules. Under the conditions of such experi-

ments the two chains will most probably associate to form a particle very

much like native fibronectin. Similar specific associations may affect the

results of experiments done with proteolytic fragments (31).

ACKNOWLEDGEMENTS

This is a contribution from the Specialized Center of Research in Thrombosis, supported by a grant from the National Institutes of Health, HL-20319. Randall Robinson was a recipient of a National Research Service Award from the National Institutes of Health (HL-06791). We thank Gillian Payne for assistance.

1.

2.

3.

4. 5.

6.

7. 8. 9.

10.

11.

12.

13.

14.

15. 16.

17.

18. 19.

20.

REFERENCES

Garcia-Pardo, A., Pearlstein E., and Frangione, B. (1984) Biochem. Biophys. Res. Conan. 120, 1015-1021. Jilek, F., and HGrmann, H. (1978) Hoppe-Seyler's Z. Physiol. Chem. 359, 247-250. Engvall, E., Ruoslathi, E. and Miller, E. J. (1978), J. Exp. Med. 147, 1584-1595. Stathakis, N. E. and Mosesson, M. W. (1977) J. Clin. Invest. 60, 855-865 Jilek, F. and Hiirmann, H. (1979) Hoppe-Seyler's Z. Physiol. Chem. 360, 597-603. Stathakis, N. E., Mosesson, M. W., Chen, A. B. and Galanakis, D. K. (1978 Blood 2, 1211-1222. Yamada, K. M. and Olden, K. (1978) Nature 275, 179-184. McDonagh, J. M. (1981) Arch. Pathol. Lab. iYed. 105, 393-396. Odermatt, J., Engel, J.,

. . . ~. Richter, H. and Hormann, H. (1982) J. Mol. Biol.

1

2, 109-123. Williams, E. C., Janmey, P. A., Ferry, J. D., and Mosher, D. F. (1982) J. Biol. Chem. 257, 14973-14878. ROCCO, M., Carson, M., Hantgan, R., McDonagh, J., and Hermans, J. (1983) J. Biol. Chem. 258, 14545-14549. Alexander, S. S., Colonna, G. and Edelhoch, H. (1979) J. Biol. Chem. 253, 1501-1505. Habeeb, A.F.S.A. (1972) (Hirs, C.H.W., and Timasheff, S. N., in Enzymology, Vol. 25, (Eds) pp. 457-464, Academic Press, New York. Williams, E. C., Janmey, P. A., Johnson, R. B. and Mosher, D. F. (1983) J. Biol. Chem. 258, 5911-5914. Hayashi, M., and Yamada, K. M. (1983) J. Biol. Chem. 258, 3332-3340. Smith, D. E., Mosher, D. F., Johnson, R. B., and Furcht, L. J. (1982) J. Biol. Chem. 257, 5831-5838. Petersen, T. E., Thbgersen, H. C., Skorstensgaard, K., Vibe-Pedersen, K., Sahl, P., Sottrup-Jensen, L., and Magnusson, S. (1983) Proc. Natl. Acad. Sci. USA, 80, 137-141. McDonald, J. A., and Kelley, D. G. (1980) J. Biol. Chem. 255, 8848-8858. Jilek, F., and Hb'rmann, H. (1977) Hoppe-Seyler's Z. Physiol. Chem. 358, 1165-1168. Engel, J., Odermatt, E., Engel, A., Madri, J. A., Furthmayr, H., Rohde, H., and Timpl, R. (1981) J. Mol. Biol. x, 97-120.

724


21.

22.

23. 24.

25. 26.

27.

28. 29.

30.

31.

Erickson, H. P., Carrell, N., and McDonagh, J. (1981) J. Cell Biol. 91, 673-678. Price, T. W., Rudee, M. L., Piersbacher, M., and Ruoslahti, E. (1982) Eur. J. Biochem. 2, 358-363. Erickson, H. P., and Carrell, N. (1983) J. Biol. Chem. 258, 14539-14544. Tooney, N. M., Mosesson, M. W., Amrani, D. L., Hainfeld, J. F., and Wall, J. S. (1984) J. Cell Mol. 97, 1686-1692. Hynes, R. O., and Yamada, K. M. (1982) J. Cell Biol. 95, 369-377. Kleinman, H. K., McGoodwin, E. B., and Klebe, R. J. (1976) Biochem. Biophys. Res. Comm. 11, 426-432. Kleinman, H. K., McGoodwin, E. B., Martin, G. R., Klebe, R. J., Fietzek, P., and Woolley, D. E. (1978) J. Biol. Chem. 253, 5642-5646. Hb'rmann, H. (1981) Klin. Wochenschr. 60, 1265-1277. Vuento, M., Vartio, T., Saraste, M., von Bonsdorff, C. H., and Vaheri, A. (1980) Eur. J. Biochem. 105, 33-42. Hermans, J. (1984) in Plasma Fibronectin: Structure and Function, McDonagh, J. (Ed.), Marcel1 Dekker, New York, Ch. 2. Molnar, J., Lai, M. Z., Siefring, G. E., and Lorand, L. (1983) Bio- chemistry, 22, 5704-5709.

725

subunit interactions in human plasma fibronectin

Documents