subunit interactions in human plasma fibronectin
TRANSCRIPT
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 con- stituent 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
Vol. 124, No. 3, 1984 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
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
Vol. 124, No. 3, 1984 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
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
Vol. 124, No. 3, 1904 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
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 fibro- nectin 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 coeffi- cient, 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 phos- phate, 0.15 M NaCl, respectively).
721
Vol. 124, No. 3, 1984 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
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 phos- phate, 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
Vol. 124, No. 3, 1984 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
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
Vol. 124, No. 3, 1984 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
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
Vol. 124, No. 3, 1984 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
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