Characterization of human factor VIII and interaction with von Willebrand factor. An electron microscopic study. Harry F. G. HEIJNEN', Joost. A. KOEDAM' ...
Eur. J. Biochem. 194,491 -498 (1990) FEBS 1990
Characterization of human factor VIII and interaction with von Willebrand factor An electron microscopic study Harry F. G. HEIJNEN’, Joost. A. KOEDAM’, Helena SANDBERG3, Nel H. BEESER-VISSER’, Jan. W. SLOT’ and Jan J. SIXMA’ Department of Cell Biology, Medical Faculty, University of Utrecht, The Netherlands
’ Department of Haematology, University Hospital of Utrecht, The Netherlands Kabi Plasma Products, Stockholm, Sweden (Received March 12/June 25, 1990) - EJB YO 0269
Blood coagulation factor VIII is a large glycoprotein that circulates in plasma at relative low concentration (0.1 pg/ml). It consists of a heterogeneous mixture of a series heavy-chain peptides (90 - 200 kDa), each associated with a light chain of 80 kDa. To gain insight into the physical properties of the protein, we have characterized purified human factor VIII by electron microscopy and rotary shadowing. Electron microscopy of rotary shadowed factor VIII molecules showed predominantly a single globular domain structure, with a somewhat asymmetric shape, while two-domain structures were also encountered. The overall dimensions of the globular domains ranged from 4 x 6 nm to 8 x 12 nm. EDTA treatment of factor VIII reduced the overall dimensions (2.5 x 5 nm to 6 x 10 nm) while treatment with thrombin reduced the dimensions to a small extent. In complexes with von Willebrand factor, factor VIII appeared localized at the globular domains of von Willebrand factor multimers. In addition, incubation of factor VIII with Staphylococcus aureus V8 protease fragments SpII and SpIII revealed only binding to the globular domains of SpIII. In this study, the first morphological characterization of human factor VIII is presented, together with its direct localization on von Willebrand factor multimers. Native vWF circulates in plasma as a series of high-molecFactor VIII and von Willebrand factor (vWF) are two distinct proteins that circulate as a noncovalently bound com- ular-mass multimers. Each multimer consists of a variable plex in plasma [l]. Both proteins play an important role in number of 270-kDa subunits, complexed by disulfide bridges normal haemostasis. Proteolytic activation of factor VIII is a [14, 151. Substructure visualization of native vWF multimers prerequisite to its participation as a cofactor in the activation and Staphylococcus aureus V8 protease fragments revealed of factor X, together with activated factor IX, phospholipids that each subunit contains a rod domain (SpII) and a globular and calcium [2-51. It can be activated proteolytically by domain (SpIII) [16 - 191. The two subunits are disulfide-linked near the carboxy-terminal nodule to form a 500-kDa dimer. thrombin or factor Xa [6, 71. vWF mediates the primary adhesion of blood platelets to The disulfide bonds at the amino-terminal globular end form the damaged vessel wall [8]. Characterization of factor VIII the vWF multimer. Analytical velocity sedimentation revealed and its interaction with vWF has been the subject of several that each subunit of vWF can bind one factor VIII molecule investigations. In plasma, factor VIII is very susceptible to [20]. Several binding studies revealed that a major factor VIII degradation and its survival in the circulation is increased binding domain is located at the globular amino-terminus of when it is associated with vWF [9,10]. Analysis of the primary vWF [21-231. The present study was designated to characterize human structure of human factor VIII, resulting from the cloning of the factor VIII gene, has led to the designation of five regions factor VIII, using rotary shadowing and electron microscopy. with the sequence: A1-A2-B-A3-C1-C2 [ll]. SDSjPAGE Additionally the association of factor VIII with S. aureus characterization of human factor VIII, isolated from a com- V8 protease fragments SpII and SpIII and native vWF was mercial concentrate, showed that it contained one doublet investigated. First morphological data, achieved by electron chain of 80 kDa (factor VIII light chain) and a series ofpeptide microscope rotary shadowing observations are presented conchains with molecular masses of 200, 180, 160, 150, 135, 130, cerning factor VIII and its localization on vWF. 115, 105 and 90 kDa (factor VIII heavy chains) [12]. Furthermore, previous studies also showed that treatment with EDTA EXPERIMENTAL PROCEDURES dissociates the heavy chain from the light chain [12, 131. Correspondence to J. J. Sixma, Department of Haematology, G03.647, University Hospital Utrecht, P.O. Box 85500, NL-3508 GA Utrecht, The Netherlands Abbreviation. vWF, von Willebrand factor. Enzyme. Staphylococcus aureus V8 protease (EC 3.4.21.19). Note. During the preparation of the manuscript, Mosesson and coworkers have reported the ultra-structural studies of porcine factor VIII using scanning-transmission electron microscopy [Mosesson et al. (1990) J . Clin. Invest. 85, 1903-19901.
Proteins
Factor VIII procoagulant protein was purified from commercial factor VIII concentrate (Octonativ, Kabi) as described before [12]. It contained the 200 - 90-kDa/80-kDa forms, and was stored at -20°C in 20mM Tris/HCl, 100mM NaC1, pH 7.8 (Tris/NaCl) 5 mM CaC12until use. Factor VIII procoagulant activity was measured by a one-stage clotting assay [24] or with the Coatest factor VIII chromogenic assay [25].
+
492 The specific activity of the preparation was 6000 - 6500 IU/ mg. Native vWF was purified from human cryo-precipitate by gel filtration on a Sepharose CL-4B column [12]. VWF, devoid of factor VII, was prepared from a Hyland factor VIII concentrate, from material that eluted from a dextran-sulfate Sepharose column [26] at high CaCl, concentration. The resultant protein appeared as a homogeneous preparation on SDSjPAGE under reducing conditions with a molecular mass of 250 kDa. On non-reducing agarose gels, the typical multimeric pattern was observed (data not shown). Biological activity of vWF is routinely measured using the ristocetin cofactor assay [27]. Both vWF preparations were precipitated by dialysis against 1.6 M ammonium sulfate and stored as a suspension at 4 C until use. Before use, the preparations were dissolved by dialysis against Michaelis buffer (28.5 mM sodium barbital, 25.5 mM sodium acetate, 116 mM NaC1, pH 7.35) or TrislNaCl. Fragments SpII and SpIII were obtained by limited proteolysis of the multimeric factor purified from human cryo-precipitate, with S. aureus V8 protease (Boehringer, Mannheim). SpII and SpIII fragments were separated by FPLC on a Mono-Q column, eluted with an NaCl gradient in Tris buffer (20 mM Tris/HCl pH 8.0) (unpublished results). On non-reduced SDSjpolyacrylamide gels, SpII appeared as a 220-kDa band and SpIII as three bands ranging over 270-330 kDa. These fragments have been extensively characterized by others [19,23] (unpublished results). SpII comprises the carboxy-terminal halves of two vWF subunits linked together to form a dimer. SpIII is the corresponding amino-terminal fragment. Association of factor VIII with native v WF andfragments SpII and SplII Purified vWF (340 pg/ml) was gel-filtered on a Sepharose CL-4B column just before association with factor VIII, in order to remove contaminating salt. The vWF-containing Vo fraction was used for incubation with factor VIII. Elution buffers were: TrisiNaCl + 5 mM CaC12 (pH 7.4) or 100 mM ammonium formate 5 mM CaC1, (pH 7.4). Conditions for the association were as follows: 25-p1 and 50-pl portions of factor VIII (350 U/ml in Tris/NaCl 5 mM CaC1,) were mixed with 25 pl of the vWF-containing Vo fraction and incubated for 1 h at room temperature. Immediately after incubation, the complex was prepared for rotary shadowing. s. ciureus V8 protease digestion products SpII and SpIII were diluted to 40 pglml with 100 mM ammonium formate 5 mM CaCl, (pH 7.4). Incubation with factor VIII (350 U/ ml) was carried out in a ratio of 1 : 1 (by vol.) for 1 h at room temperature. After incubation, the mixtures were prepared for rotary shadowing.
+
+
+
Gel analysis and inzmunoblotting
SDS/polyacrylamide gel electrophoresis of factor VIII was carried out according to Laemmli [28] with a 4 % concentration in the stacking gel and 7.5% in the separating gel. The samples were reduced with 5% (by vol.) 2-mercaptoethanol in 2% (massivol.) SDS. The gels were stained with silver nitrate [29]. Immunoblot analysis of the factor VIII preparation was carried out as described by Towbin [30] with some modifications [ 121. Briefly, a polyclonal antibody against human factor VIII was raised in rabbits by immunization with the purified factor VIII preparation. After electrophoresis and electrotransfer to nitrocellulose, the membrane was incubated
a 1
2
b 3
200 -
116-
97 66 -
light chain
- 68 - 43
43 -
Fig. 1, ( A ) SDSIPAGE of ,factor VIII purified from commerical concentrate. ( B ) Immunoblotting of factor VIII with the use of apolyclonal factor VIIZ antibody. Lane 1, molecular mass reference molecules: myosin, Escherichia coli B-galactosidase, rabbit muscle phosphorylase, bovine serum albumin and ovalbumin (Bio-Rad). Lanes 2 and 3, two factor VIIl preparations. The resulting bands correspond to the 200-90-kDa range of factor VIII heavy chain and the 80-kDa light chain. The minor contaminating low-molecular-mass bands represent fibrinogen (triple band) and some other non-characterized proteins. Molecular mass values are expressed in kDa. (B) Heavy-chain peptides and light-chain peptides react with the polyclonal anti-(factor VIII) antibody
with the polyclonal anti-(factor VIII) antibody, followed by a second swine anti-(rabbit IgG) (Dakopatts Z 196) and rabbitPAP (Dakopatts Z 113). The antibody reactivity was visualized with 4-chloro-1 -naphthol substrate (Bio-Rad).
EDTA and thrombin treatment of factor VIII Factor VIII (350 Ujml) in TrisjNaCl + 5 mM CaCl, was either mixed with 20mM EDTA (Merck) in 100mM ammonium formate pH 7.4 or with 300 mM ammonium formate containing 5 mM CaC1,. Samples were incubated overnight at 4°C and prepared for rotary shadowing. Thrombin ( 5 U/ ml, Sigma) was added to factor VIII (350 Ujml) in Tris/NaCl and directly prepared for rotary shadowing, using factor VIII and thrombin as controls. Electron microscopy Low-angle rotary shadowing of proteins and protein complexes was performed as adapted from Shotton et al. [31] and Tyler and Branton [32]. Two methods of preparation were used: (a) glycerol spraying as described by Fowler and Erickson [16] and (b) the mica-sandwich method as described by Mould et al. [33]. Samples of proteins or protein complexes were mixed with 100% anhydrous glycerol (Fluka, Switzerland) in a ratio of 1 : 1 (by vol.); 20 - 2 5 4 portions were sprayed onto freshly cleaved mica or, alternatively, 5-pl portions were 'sandwiched' between two pieces of freshly cleaved mica (2 cm x 2 cm), adsorbed for 10 min, separated, and cut into smaller pieces. Samples were placed in an Edwards 306 vacuum evaporator, dried in vacuo (1 - 0.1 mPa), and rotary shadowed with platinum/tungsten (Pt/W) by means of resistant evaporation at an angle of 7". Carbon deposition was carried out at an angle of 90". Replicas were
493
Fig. 2. Electron micrographs of rotary shadowed factor VIII molecules. (a) General field of view, showing predominantly single-domain structures; arrows indicate occasional two-domain structures; note the heterogeneity in size. (b) Examples of selected factor VIlI molecules at higher magnification; single domains, elongated forms and occasional two-domain structures contribute to a high variation in molecular dimension. (c) Purified IgG molecules at the same magnification. (d) 3-nm and 6-nm colloidal gold particles pronounce a distinct metal shell which can be measured accurately; note that the gold particles remain attached to the replica (e, f). General field of view of factor VIII molecules, before and after treatment with 10 mM EDTA, respectively; a decrease in overall dimensions is clearly visible. Bar = 100 nm (a, e, f), 50 nm (b, c, 4)
examined in a Jeol 1200CX electron microscope and photographed at nominal magnifications of 50000 and 80 000. Final magnifications are indicated in the figure legends. The magnification was calibrated using negatively stained catalase crystals. Metal film thickness was controlled, using colloidal gold particles of 3 nm and 6 nm diameter, prepared according to the method of Slot et al. [34]. Gold particles prepared in this manner were homogeneous in size (coefficient of variation smaller than 10%) and were used as controls in every shadowing cycle, thus permitting actual measurement of the metal shell. Molecular dimensions and measurements of the colloidal gold particles were determined with a Calcomp graphic tablet (resolution 0.05 mm). Measurements of the colloidal gold particles were carried out before and after rotary shadowing. Subtracting actual gold particle dimensions from
the replicated dimensions then reveals the thickness of the replicated platinum shell. After replication both 3-nm and 6nm gold particles appeared to have a metal shell of z 2 nm. Dimensions of the molecules were corrected for this platinum shell by subtracting 4 nm from the total diameter.
RESULTS SDS-gel electrophoresis of the factor VIII protein, prepared from commercial factor VIII concentrate, revealed the expected range of molecular mass (Fig. I A , lanes 2 and 3). The 200 - 90 kDa bands represent heavy chains and the 80kDa bands light chains. The immunoblot analysis clearly shows that the different peptide chains seen on the SDS gel
494 0 20
FVIII 010
zz
020
W
3
2a
010
rL
w
2
FVIII+Thrombin
0
20
LO
60
80
100 120
6 nm, long axis from 5 nm to 10 nm) together with a less heterogenic pattern of dimensions (Fig. 3 b). The electrophoresis pattern on the SDS gel remained identical after EDTA treatment (not shown). Analysis of the number of domains was carried out before and after EDTA treatment ( n > 500). After treatment with EDTA, a decrease of twodomain structures was not observed. These results suggest that the single-domain structures represent individual factor VIII molecules. Assuming a spherical or ellipsoidal conformation of single factor VIII molecules, the total area occupied by a molecule can be estimated by: R,,, . Rmin. 71. Fig. 3 gives the relative frequency distributions of R,,, . Rmin. before and after treatment with EDTA, and after incubation with thrombin. A decrease of total area after treatment with EDTA was statistically significant ( t test P < 0.001). After incubation with thrombin, the electron microscopic images showed both globular thrombin molecules and factor VIII molecules. The globular domain structures of thrombin present in the images made a distinct identification of both molecules troublesome. The frequency distribution of total area, however, revealed two distinct peaks representing thrombin and factor VIII molecules (Fig. 3 c). Subtracting thrombin control from factor VIII after incubation with thrombin then reveals the true dimensions of factor VIII. The slight decrease in molecular dimensions observed was statistically significant ( t test P < 0.001). For comparison, purified IgG molecules (mass 160 kDa) are shown (Fig. 2c). The uniformity in molecular dimension is striking, as compared with factor VIII.
1LO
2
AREA ( n m )
Fig. 3. Frequency histogram of total area occupied by,factor VIII molecules, as measured ,from electron micrographs. Total area = R,,, . Rmin. n. (a) Factor V l I l control (FVIII);note the heterogeneity in molecular dimensions (mean = 49 nm’). (b) Factor VIII after treatment with 5 mM EDTA; the range in molecular dimensions was less pronounced and a decrease in total area was significant (mean = 25 nm’). (c) Factor VIII after incubation with 0.1 U/ml thrombin; the two distinct peaks represent thrombin and factor VIII respectively. An overall decrease in total area was significant (mean = 39 nm‘); the extra small peak falls within the range oftotal surface area of factor VIII control, and may represent non-degraded forms. (d) Thrombin control
react with the polyclonal anti-(factor VIII) antibody (Fig. 1 B). Electron micrographs of human factor VIII molecules are shown in Fig. 2. A majority of factor VIII molecules appeared as globular single-domain structures (Fig. 2 a), while sometimes a distinct two-domain structure was encountered (Fig. 2 a, arrows). Occasionally the typical trinodular structure of fibrinogen was observed in our rotary shadowed images. Both preparation methods (i. e. glycerol spraying and mica sandwich) revealed the same morphological results, except that the occurrency of two-domain structures was somewhat more predominant using the sandwich method. A clear contact between the two domains was not resolved. Since factor VIII molecules sometimes showed a somewhat elongated appearance, both long and short axes were measured. The short axis of the single domains ranged over 48 nm, while the long axis ranged over 6- 12 nm (corrected Pt shell: 4 nm, calculated from colloidal gold particles). The heterogenic pattern of dimensions was not unexpected considering the results of SDS-gel electrophoresis described above. After treatment with 10 mM EDTA, a decrease in overall dimensions was noticed (short axis from 2.5 nm to
Associution with v W F and S . aureus V8 protease fragments SpZZ and SpZZZ
Electron microscopic visualization of human vWF has been the subject of several investigations [16-181. In our rotary shadowed preparations, vWF appeared as a thin flexible strand with alternating rod and globular domains, sometimes extended, but frequently in a ‘ball of yarn’ conformation (Fig. 4a). This multimeric appearance of human vWF was essentially the same as described previously. Electron microscopy and rotary shadowing of S. aureus V8 protease fragments SpII and SpIII was also the subject of a previous study [19] and revealed the same morphological results as the present study. Fragment SpII consisted of two flexible rods, joined at a small central nodule, designated as fragment RR (Fig. 5d). Fragment SpIII had an elongated globular structure, very similar in appearance to the alternating globular domains in native vWF, designated as fragment GG (Fig. 5a). Our electron microscopic observations revealed no morphological difference between native vWF and vWF devoid of factor VIII. Upon incubation with factor VIII, both vWF and SpIII revealed clear morphological changes, as compared with their controls (Figs 4 and 5). Electron micrographs of selected vWF multimers after incubation with factor VIII are given in Fig. 4c. A majority of the vWF multimers appeared ‘decorated’ with additional globular factor VIII molecules, while a minority remain ‘undecorated’ upon incubation with factor VIII. As many of the vWF multimers were folded into a compact ‘ball of yarn’ conformation, analyses of vWF in terms offactor VIII binding was not always feasible. Extended vWF multimers, however, permitted clear localization of factor VIII molecules at the globular domains (Fig. 4d). In order to investigate possible conformational changes of vWF multimers upon factor VIII binding, the degree of folding was characterized in conformational terms of ‘ball of yarn’, slightly extended and fully extended multimers. No significant change
495
Fig. 4. Electron micrographs of rotary shadowed human vWF before and after incubation with factor VIII. (a) ‘Ball of yarn’ conformation of vWF multimers, with alternating globular (arrows) and rod (arrowheads) domains. (b) vWF in fully extended form. (c) vWF after incubation with factor VIII. Selected multimers still show the typical ‘ball of yarn’ conformation. A clear decoration of apparent factor VIII molecules is visible. Arrows indicate factor VIII molecules bound to the globular domains of vWF multimers. (d) Example of a fully extended vWF multimer with distinct factor VIII molecules bound at the globular domains. Bar = 100 nm
in conformation was observed after factor VIII incubation (n > 200). Complexes of factor VIII and SpIII often showed an irregular appearance, suggesting that factor VIII had bound to this globular fragment (Fig. 5 b). Many complexes, however, were encountered with both proteins still distinct (Fig. 5 c). Incubation of factor VIII with SpII revealed no binding to the rod fragments (Fig. 5e). DISCUSSION In the present study, human factor VIII, purified from a commercial concentrate, was characterized using rotary shadowing and electron microscopy. Visualization of factor VIII molecules was important in order to determine the localization on vWF multimers. The heterogeneity in molecular
dimensions observed in our images coincided very well with the range of proteolytic forms of native factor VIII as seen by SDS-gel electrophoresis (Fig. 1) [12]. However, we do not have data to support that the high number of single-domain structures represent intact factor VIII molecules under all conditions. As factor VIII is very susceptible to degradation, forces that might distort it when it contacts the mica surface may have influenced the appearance of the molecules in our images. Carrel1 et al. [35]observed an apparent dissociation of factor XI11 dimers during specimen preparation, leaving subunits remaining in pairs but too far apart to be in contact. In our factor VIII preparations we occasionally observed twodomain structures. A majority of the molecules appeared as single-domain structures, situated at relatively large intermolecular distances. The number of two-domain structures did not alter upon EDTA treatment, while the total area of factor
496
Fig. 5. Electron microgruphs of rotary shudowed S. aureus V8 protease,frugments SpIU and SpIl before and ufter incubation with fuctor VIII. (a) Field of view of SpIII before incubation with factor VIII; predominant elongated globular domain structures represent the dimeric domain. also present in vWF multimers. (b) Small aggregates of SpIII and factor VTII are visible; a distinct recognition of both proteins was not always feasible. (c) Selected examples of Sp-111- factor-VTII complexes; factor VIII molecules are still recognized. (d) Field of view of rotary shadowed SpII fragments, showing the typical rod-like structure, also present in vWF multimers. (e) After incubation with factor VIII no binding is observed. Bar = 100 nm
VIII components decreased significantly. Therefore we believe that the single-domain structures represent both heavy and light chains of factor VIII. The decrease in molecular dimension may possibly reflect the dissociation of the 80-kDa light chain from the 200-90-kDa heavy chains. Incubation of factor VIII with thrombin revealed a slight decrease in total area. Previous studies already reported a decrease in the molecular mass of factor VIII after thrombin incubation [36, 371. Anderson et al. [I21 showed a rapid degradation of factor VIII heavy and light chains after thrombin incubation, producing new peptide chains of 52 kDa, 43 kDa and 70 kDa. The decrease in total area after incubation with thrombin in turn may reflect the degradation of factor VIII after thrombin incubation. Measurements of the overall dimensions of factor
VIII molecules revealed a spherical to ellipsoidal shape (Rmax/ R,,, = 1.49). In comparison, Lollar et al. [38] calculated a frictional coefficient ratio of 1.39 for thrombin-activated factor VIII, a common value for asymmetric globular molecules [39].In an attempt to deduce the approximate molecular mass of factor VIII molecules, we compared the surface area of the molecules with that of IgG molecules of molecular mass 160 kDa (average surface area 38 nm’). The factor VIII molecules appeared to have surface areas ranging from 20 nm2 for the smaller molecules to 75 nm2 for the larger molecules, corresponding with molecular masses ranging over 80 320 kDa. Although these calculated molecular masses of factor VIII molecules represent an estimation of true values, they reflect the heterogeneous molecular range seen on SDS/
497 polyacrylamide gels. Lollar et al. [20] calculated the radii of factor VIII heavy and light chains from the corresponding masses of 166 kDa and 76 kDa as 3.6 nm and 2.8 nm, respectively. Factor VIlI shares a considerable structural similarity with factor V [40, 411; recent studies suggest a similar quaternary structure for both porcine factor VIII and bovine factor V after activation with thrombin [38]. However, morphological studies of both human and bovine factor V revealed a multidomain structure [42, 431 which is not consistent with our morphological data concerning human factor VIII. A sub-nanomolar concentration of factor VIII in plasma was suggested to contribute to the mechanism of nonproteolytic dissociation of activated factor VIII (factor VIIIa) [l,381. This high degree of instability of the factor VIII molecule may partially explain the morphological controversy concerning factors VIIl and V. Although the substructure of factor VIII heavy chain and light chain was not resolved using the rotary shadowing technique, visualization of the individual factor VIII molecules permitted us the unique possibility of localizing it on vWF multimers. Incubation of factor VIII with native vWF and the S. aureus V8 protease fragments SpII and SpIII revealed a distinct binding to the amino-terminal globular domains of vWF multimers and binding to the SpIII fragments. Our morphological data support previous findings that the binding domain of factor VIII is situated at the aminoterminal globular ends of vWF multimers [21,23,44]. Conformational changes of vWF multimers upon factor VIII binding were not observed in our electron microscopic images. Since vWF multimers frequently exist in a ‘ball of yarn’ conformation, quantification of the number of factor VIII molecules bound/vWF multimer was not feasible. Since both SpIII and vWF multimers can bind factor VIII molecules, multimerisation of vWF seems not to be a prerequisite for factor VIII binding. Previous studies on the substructure of human vWF revealed no morphological difference between vWF devoid of factor VIII or containing it [45]. This is in agreement with our studies where no morphological difference was observed between native human vWF purified from human cryoprecipitate (containing factor VIII) and vWF prepared from Hyland factor VIII concentrate (devoid of factor VIII). Since factor VIII circulates in plasma at sub-nanomolar concentrations in complex with vWF, this was not expected. Hamer et al. [36] have calculated from their binding studies that, under saturating conditions, factor VIII binds to vWF in a 1 :4 ratio. Lollar et al. [20] used purified porcine factor VIII heterodimer and light chain in their velocity sedimentation studies and found a stoichiometry of 1 : 1 for both heterodimer and light chainlsubunit vWF. In conclusion, we have presented the first morphological data of purified human factor VIII and demonstrated binding to vWF multimers and S. oureus V8 protease fragment SpIII. The tangled conformation of vWF has been suggested to represent the conformation in solution [19]. Although both inultimeric vWF and SpIIl can bind factor VIII, the dimeric conformation may still be a partial prerequisite for factor VIII protection. Folding of the rodlike SpII domain at the central carboxy-terminal nodule may function cooperatively with the formation of dimers in order to maintain complete protection of factor VIII against degradation. In our images we observed both ‘decorated’ and ‘undecorated’ vWF multimers upon incubation with factor VIII. Although not likely, the possibility of a sub-population of vWF multimers, with a predominant ability to bind factor VIII molecules cannot be-excluded.
REFERENCES 1. Hoyer, L. W. (1981) BloodS8, 1-13. 2. Osterud, B. & Rapaport, S. I. (1970) Biochemistry 9, 1854- 1862. 3. Hougie, C., Denson, K. W. E. & Biggs, R. (1967) Thromb. Diath. Huernorrh. 18, 211 -222. 4. Van Dieijen, G., Tans, G., Rosing, J. & Hemker, H. C. (1981) J . Biol. Chem. 256, 3433 - 3442. 5. Hultin, M. B. (1982) J . Clin. Invest. 69, 950-958. 6. Lollar, P., Knutson, G. J . & Fass, D. N. (1985) Biochemistry 24, 8056 - 8064. 7. Eaton, D., Rodriguez, H. & Vehar, G. A. (1986) Biochemistry 25, 505-512. 8. Ruggeri, Z. M. & Zimmerman, T. S. (1987) Blood 70, 895-904. 9. Tuddenham, E. G. D., Lane, R. S., Rotblat, F., Johnson, A. J., Snape, T. J., Middleton, S. & Kernoff, P. B. A. (1982) BY. J . Haematol. 52, 259 -267. 10. Brinkhous, K. M., Sandberg, H., Garis, J. B., Mattson, C., Palm, M., Griggs, T. & Read, M. S. (1985) Proc. Natl Acad. Sci. U S A 82,8752- 8756. 11. Vehar, G. A., Keyt, B., Eaton, D., Rodriguez, H., O’Brien, D. P., Rotblat, F., Oppermann, H., Keck, R., Wood, W. I., Harkins, R. N., Tuddenham, E. G. D., Lawn, R. M. & Capon, D. J. (1985) Nuture 312, 337-342. 12. Andersson, L. O., Forsman, N., Huang, K., Larsen, K., Lundin, A,, Pavlu, B., Sandberg, H., Sewerin, K. & Smart, J. (1986) Proc. Nut1 Acad. Sci. USA 83,2979-2983. 13. Hamer, R. J., Koedam, J. A., Beeser-Visser, N. H., Bertha, R. M., van Mourik, J. A. & Sixma, J. J. (1987) Eur. J . Biochem. 166, 37-43. 14. Chopek, M. W., Girma, J. P., Fujikawa, K., Davie, E. W. & Titani, K. (1986) Biochemistry 25, 3146-3155. 15. Van Mourik, J. A. & Mochter, I. A. (1970) Biochim. Biophys. Acts 221, 677-679. 16. Fowler, W. E. & Erickson, H. P. (1979) J. Mol. Biol. 134, 241 249. 17. Fowler, W. E., Fretto, L. J., Hamilton, K. K., Erickson, H. P. & McKee, P. A. (1985) J. Clin. Invest. 76, 1491-1500. 18. Slayter, H., Loscalzo, J., Bockenstedt, P. & Handin, R. I. (1985) J . Bid. Chem. 260,8559-8563. 19. Fretto, L. J., Fowler, W. E., McCaslin, D. R., Erickson, H. P. & McKee, P. A. (1986) J . Biol. Chem. 261, 15679-15689. 20. Lollar, P. & Parker, C. G. (1987) J . Bid. Chem. 262, 1757217 576. 21. Foster, P. A,, Fulcher, C. A,, Marti, T., Titani, K. & Zimmerman, T. S. (1987) .I. Biol. Chem. 262, 8443-8446. 22. Takahashi, Y., Kalafatis, M., Girma, J. P., Sewerin, K., Andersson, L. 0. & Meyer, D. (1987) Blood 70, 1679- 1682. 23. Girma, J. P., Chopek, M. W., Titanie, K. & Davie, E. W. (1986) Biochemistry 25,3156-3163. 24. Rosen, S., Andersson, M., Blomback, M., Hagglund, U., Larrieu, M. J., Wolf, M., Boyer, C., Rotschild, C., Nilsson, 1. M., Sjorin, E. & Vinazzer, H. (1985) Thromh. Huemost. 54,818-823. 25. Margolis, J. (1979) Pathology 11, 149-159. 26. Koedam, J. A,, Meyers, J. C. M., Sixma, J. J. & Bouma, B. N. (1988) J . Clin.Invest. 82, 1236-1243. 27. Ruggeri, U. M., Mannucci, P. M., Jeffcoate, S. C. & Ingram, G. I. C. (1976) Br. J . Huematol. 33,221 -232. 28. Laemmli, U. K. (1970) Nuture 227, 680-685. 29. Morrissey, J. H. (1 981) Anal. Biochem. 117, 307 - 310. 30. Towbin, H., Staehelin, T. & Gordon, J. (1979) Proc. Natl Acad. Sci. U S A 76, 4350-4354. 31. Shotton, D. M., Burke, B. & Branton, D. (1979) J . Mol. Biol. 131, 303 - 329. 32. Tyler, J. & Branton, D. (1980) J. Ultrustruct. Res. 71, 95-102. 33. Mould, P. A,, Holmes, D. F., Kadler, K. E. & Chapman, J. A. (1985) J . Ullrastr. Res. 91, 66-76. 34. Slot, J. W. & Geuze, H. J. (1985) Eur. J . Cell Bid. 38, 87-93. 35. Carrell, N. A,, Erickson, H. P. & McDonagh, J. (1989) J . Biol. Chern. 264, 551 -556. 36. Hamer, R. J.. Koedam, J. A., Beeser-Visser, N. H. & Sixma, J. J. (1987) Eur. J . Biochem. 167,253-259.
49 8 37. Fulchner, C. A.. Roberts, J . R. & Zimmerman, T. S. (1983) Blood 61, 807-811. 38. Lollar, P. & Parker, C. G. (1989) Biochemistry 28,666-674. 39. Creighton, T. E. (1984) Proteins. Structures andrnoleculurproperties, pp. 265 - 286, W. H. Freeman & Company, New York. 40. Church, W. R., Jernigan, R. L., Toole, J., Hewick, R. M., Knopfig, J., Knutson. G. J., Nesheim, M. E., Mann, K. G. & Fass, D. N. (1984) Proc. Natl Acad. Sci. U S A 81,6934-6937. 41. Jenny, R. J., Pittman, D. D., Toole, J. J., Kriz, R. W., Aldape, R. A., Hewick, R. M., Kaufman, R. J. & Mann, K. G. (1987) P ~ o cNut1 . A c ~Sci. . USA 84, 4846-4850.
42. Mosseson, M. W., Nesheim, M. E., DiOrio, J., Hainfeld, J. F., Wall, J. S. & Mann, K. G. (1985) Blood65, 1158-1162. 43. Dahlback, B. (1986) J . Biol. Chenz. 261,9495-9501. 44. Titani, K. & Walsh, K. A. (1988) Trends Biochem. Sci. 13, 9497. 45. Ohmori, K., Fretto. L. J., Harrison, R. L., Switzer, M. E. P., Erickson, H. P. & McKee, P. A. (1982) J . Cell Biol. 95, 632640.
