A Heparin-binding Domain of Human von Willebrand Factor

Document Sample
A Heparin-binding Domain of Human von Willebrand Factor Powered By Docstoc
        OF         CHEMISTRY                                                                       Vol. 262, No. 4, Issue of February 5, pp. 1734-1739, 1987
0 1987 by The American Society of Biological Chemista, Inc.                                                                                Printed in U.S.A.

A Heparin-binding Domain of Human von Willebrand Factor

                                                                                                     (Received for publication, June 10, 1986)

                Yoshihiro FujimuraS, Koiti Titanif, Linda Z. Holland, James R. Roberts, Paul Kostel,
                Zaverio M.Ruggeri, and Theodore S Zimmermanll
                From the Department of Basic and Clinical Research,Scripps Clinic and Research Foundation, La Jolla, California 92037 and
                the $Department of Biochemistry, University of Washington, Seattk, Washington 98195

   We have recently shown that the domain of von                 von Willebrand factor (vWF)’ is an adhesive plasma gly-
Willebrand factor (vWF) which interactswiththe                coprotein composed of a series of multimers (0.5 x lo6 to 20
platelet glycoprotein Ib (GPIb) is located in a 52/48- X lo6daltons) which plays a critical role in the earliest phases
kDa tryptic fragment of the molecule which begins of hemostasis (1). vWF interacts with several proteins of
with amino acid residue Val-449. We have now estab- importance in the hemostatic process including platelet gly-
lished that the fragment extends to residue Lys-728 coprotein (GP) Ib, platelet GPIIb-IIIa complex, and collagen
and demonstrate here that a high affinity heparin-            (2-4). The vWF domains critical to these interactions are
binding domain of vWF also lies within this  region and currently being defined with increasing precision.
in close proximity to that for GPIb. We have used an             In 1978, Madalas et al. (5) demonstrated for the first time
assay employing heparin coupled to SepharoseCL-GB that Factor VIII-vWF complex binds to immobilized heparin.
to show that “‘I-vWF binds toheparin in a time- Fowler et al. (6) have recently used heparin-agarose affinity
dependent, saturable, and reversible  manner. Binding chromatography as a step purification of vWF from normal
could be completely inhibited by the 52/48-kDa frag- human plasma. More recently, Bockenstedt et al. (7) have
ment, but was not affected by other tryptic fragments shown that a 285-kDa fragment derived from a limited pro-
of 55, 41, 13, and 22 kDa. NHz-terminal sequencing teolytic digestion of vWF by Staphylococcus aureus V8 pro-
of these fragments showed that they were derived from         tease binds to heparin-Sepharose gel. Thislatter finding
different parts of the molecule, as follows: 13 kDa, locates a heparin-binding domain in the NH2-terminalportion
GlnzSo-Thr-Met-Val-Asp-Ser-Ser;    55 kDa, A ~ n ~ ’ ~ - S e rof the molecule (7,8).
Met-Val-Leu-Asp-Val-Ala-Phe-Val-Leu-Glu; kDa,   41               In this paper we will show that a major heparin-binding
Thr’3~2-Val-Gln-Arg-Pro-Gly-Gln-Thr-Cys-Gln-Pro-              domain of vWF is located within the same 52/48-kDa tryptic
Ile-Leu-Glu-Glu-Gln-Cys-Leu-Val; kDa, VallSz7- fragment of vWF which contains the platelet GPIb-binding
Thr-Gly-Cys-Pro-Pro. Direct binding of the purified domain. This fragment has been shown to begin with Val-449
52/48-kDa fragment to heparin-Sepharose was also (9) of the entire subunit sequence of 2050 residues recently
shown. Furthermore, crossed immunoelectrophoresis determined by Titani et al. (8).             Now we have determined that
revealed complex formation between the purified 52/ it extends to Lys-728. Studies with monoclonal antibodies
48-kDa fragment and free heparin. Twelve monoclo- reacting with this fragment place the two domains in close
nal antibodies to the 52/48-kDa fragment were evalu-          proximity to one another.
ated for their ability to block binding of ‘“1-vWF to
heparin. With the exception of one weak inhibitor of                            MATERIALS AND METHODS
heparin binding, their relative efficacy in blocking             Reagents-All chemicals used were of the best grade commercially
heparin binding was similar to that for blocking ris- available. Carrier-free Na[1261]was obtained from Amersham Corp.
tocetin-induced binding to GPIb. However heparin Ristocetin, heparin sodium salt (porcine intestinal mucosa, grade 11,
failed to block ristocetin-independent binding of the 162 USP units/mg) and trypsin (bovine pancreatic Type I, 15,000
52/48-kDa fragment GPIb. It is therefore likely that units/mg) were purchased from Sigma. Crystallized bovine serum
the two binding domains are adjacent to one another, albumin (BSA) was obtained from Behring Diagnostics. IODO-GEN@
but are not precisely congruent.                              and cyanogen bromide were from Pierce. Protein A-Sepharose CL-
                                                              6B and heparin-Sepharose CL-GB gels were products of Pharmacia.
                                                                            Highly Purified v WF-vWF containing the largest vWF multimers
                                                                         was purified from cryoprecipitate (10).
                                                                            Purification of Reduced and S-Carboxymethylated Tryptic Frag-
                                                                         ments of u WF-vWF was purified by immunoadsorption and digested
   * This work wassupported in part by Grants HL31950, HL 15491,         with trypsin as previously described (9). The digest was subjected to
and HL 29595 from the National Institutes of Health. This is Publi-      size exclusion chromatography, and the resulting five fractions were
cation 4369 BCR from the Research Institute of Scripps Clinic. The       reduced and S-carboxymethylated. Chromatofocusing of the second
costs of publication of this article were defrayed in part by the        fraction, fraction “B,” yielded polypeptides of 52/48, 41, 22, and 13
payment of page charges. This articlemust therefore be hereby            kDa (9). A 55-kDa polypeptide was also purified in a similar manner
marked “advertisement” in accordance with 18 U.S.C. Section 1734         from fraction “C.”
solely to indicate this fact.                                               Antibodies-Monoclonal antibodies against the 52/48-kDa frag-
  $ Performed this work during the tenure of a Fogarty International     ment were produced in mice by an established method as previously
Fellowship of the National Institutes of Health (F 05 TW 3475-01).
   7 To whom reprint requests should be addressed Dept. Basic and            The abbreviations used are: vWF, von Willebrand factor; GP,
Clinical Research, Scripps Clinic and Research Foundation, 10666         glycoprotein; Mes, 2-(N-morpholino)ethanesulfonic acid; HPLC,
North Torrey Pines Rd., La Jolla, CA 92037.                              high pressure liquid chromatography.

                                        uon Wilbbrand Factor Heparin-binding Domain                                                   1135
described (11). Monoclonal IgGs werepurified from ascites fluid using CL-GB was measured a t 4, 23, and 37"C. There were no
a protein A-Sepharose CL-GB column according to Ey et al. (12). significant differences in the quantity of '251-vWFbound at
Polyclonal antibody against the 52/48-kDa fragment was raised in the different temperatures, and binding did not increase after
rabbits, and the specificity was determined by an immunoblotting
technique (13).
                                                                        30 min (Fig. 1). Therefore, all the succeeding experiments
   Iodination-Labeling of vWF and the 52/48-kDa fragment with were performed at room temperature with 30-min incubation
lZ5Iwas performed by the method of Fraker and Speck (14). The and occasional agitation.
specific activity of the preparations used in thesestudies varied         To determine whether binding of 1251-~WF to               heparin-
between 2.88 X lo" and 9.09 X lo" Ci/mg.                                Sepharose was saturable, thegel was incubated with amounts
   Crossed Immunoelectrophores~-This was carried out using a mod- of labeled ligand ranging from 3 to 300 pg/ml. The datawere
ification of the Laurel1 system (15). Chamber buffer consisted of 0.075
M sodium barbital and 0.015 M barbituric acid, pH 8 6 One percent
                                                        ..              evaluated by Scatchard-type analysis using the computerized
agarose gel was prepared using a 1:3 dilution of chamber buffer. For program Ligand (21). The molecular mass of the vWF              subunit
the first-dimensional electrophoresis, various concentrations of hep- (275 kDa) (8) was substituted for that of the native multi-
arin were included in the gels. Electrophoresis was carried out a t merized molecule, whichhas not been accurately determined.
constant current of 2 mA/cm for about 2 h until a tracking dye The datapoints fit a straight line, indicating the existence of
migrated 5 cm from the well. In the second dimension, polyclonal a single class of noninteracting binding sites. It was calculated
rabbit antibody was included in the agarose gel. Electrophoresis was
performed for 6 h a t a constant current of0.5mA/cm. Gels were          that a total of 2.09 mg of vWF could bind to 1 ml of heparin-
stained with Coomassie Brilliant Blue.                                  Sepharose CL-GB with an apparent dissociation constant ( K d )
   Binding Assays Utilizing Immobilized Heparin-A heparin-Sepha- of 3.78 X                M (Fig. 2).
rose CL-GB which binds -2mg (5 units) of antithrombin 111 per             Dissociation and Displacement of Bound Labeled uWF by
milliliter of gel was washed extensively with distilled water, and then NaC1, Free Heparin, and Unlabeled vWF-As shown in Fig.
the gel was suspended at a concentration of 25% (v/v) in 0.02 M Mes- 3A, '251-vWFbinding to heparin-Sepharose was maximal at
Tris buffer containing 0.02% NaN3, pH 6.0, and stored at 4 "C until
used.                                                                   nearly physiological NaCl concentrations. In the presence of
   Assays weke performed in an incubation volume of 125 pl in plastic   1 M NaC1, binding was no greater than to the heparin-free
Eppendorf tubes. Heparin-Sepharose CL-GB gel, lZ5I-vWF,        and BSA Sepharose CL-GB control. Similarly, at a concentration of 10
were added to tubes a t final concentrations of 1.25-5% (v/v), 1-5 pg/
ml, and 0.1% (w/v), respectively. For competitive inhibition studies,
various ligands were dissolved in 0.05 M Tris. HCI, 0.15 M NaCl
buffer, pH 7.3 (TBS), and added to the incubation mixture a t con-
centrations described under "Risults." After 30-min incubation at                              3   0 V         T        l
room temperature with occasional agitation, two 50-pl aliquots of the
mixture were layered over 20% sucrose cushions (300 p l ) in 500-pl                            20
Sarstedt microfuge tubes, and the bound ligand was separated from
                                                                                                                 0 23°C
free ligand by 4-min centrifugation a t 12,000 X g in a Beckman                                 10               A 31%
Microfuge B. The tube tips containingbound ligand were amputated
and counted in a y-counter.
    Nonspecific binding was determined in the presence of 10 mg/ml                               0
heparin sodium salt (porcine intestinal mucosa, grade 11, 162 USP                                  0            30          60
units/mg; Sigma) added together with the labeled ligand and heparin-                                  IncubationTime lminl
Sepharose CL-GB gel. Nonspecific binding was always less than 5%
of the total binding. The IC& was that concentration of competing         FIG. 1. Effects of temperature and incubation time on the
ligand which inhibited the specific binding of lZSI-vWF 50%.
                                                          by            binding of v W F to heparin-Sepharose CL-GB. '251-vWF (5 pg/
   Inhibition of v WF Binding to Platelet GlycoproteinIb by Monoclonal ml, 2.93 X 1 " Ci/mg) was incubated at 4,23,and 37 "C with 5 % (v/
Antibodies-Platelets were prepared as previously described (9) and v) heparin-Sepharose. The bound radioactivity was determined at
used at a concentration of10B/ml. Purified monoclonal IgGs were         various intervals as described under "Materials and Methods." The
used at concentrations described under "Results." After 30-min in- amount of lZ5I-vWF        bound was calculated on the basis of its specific
cubation a t 37"C, ristocetin was added a t a concentration of 1.0 mg/  activity.
ml. Nonspecific and specific binding were determined as described
   Amino-terminal Analysis-The purified polypeptides were ana-
lyzed with a gas-phase sequenator (Model 470A; AppliedBiosystems),
followed by identification of the phenylthiohydantoin derivatives of
amino acids using HPLC (Perkin-Elmer)(16).
   Generation and Characterization of Cyanogen Bromide Frag-
ments-0.5 mg of the 52/48-kDa fragment was cleaved with 0.5 mg
of CNBr in 72% formic acid at room temperature for 15 h (17). The
digest was lyophilized after 6-fold dilution with distilled water. The
resultant fragmentswere separated ontwo TSK G 3000 SW columns
(7.5 X 600 mm) connected in series and equilibrated in 6 M guanidine                                      1"111 vWF Imglmll
HCI and 10 m sodium phosphate, pH 6.0. Further purification by
                M                                                          FIG. 2. Saturation binding of vWF to heparin-Sepharose
reverse-phase HPLC on a SynChropak RP-P(C-8) column (Syn- CL-BB. Increasing concentrations of 'T-vWF (2.88 X lo" Ci/mg)
Chrbm Western Analytical) with a gradient elution of acetonitrile were mixed with 1.25% (v/v) heparin-Sepharose and incubated for
dissolved in 0.1% trifluoroacetic acid (18) yielded fivemajor peptides. 30 min at room temperature. The amount of ' 9 - v W F bound to the
    Amino acid compositions were determined by the Waters Picotag heparin-Sepharose gel was then determined. The values shown in
method of Bidlingmeyer et al. (19) after 24-h acid hydrolysis of this graph represent total binding. The inset shows a Scatchard-type
peptides.                                                               plot of the experimental data obtained using the computer-assisted
    Protein Concentration-The concentration of purified monoclonal program Ligand (211, and the points represent specific binding ob-
IgGs was calculated from absorbance at 280 nm, assuming a value of tained by subtracting the nonspecific binding (calculated as a fitted
E% = 14.3. The concentration of other proteins was measured by the parameter) from the total binding. The best fit for the experimental
method of Bradford (20) using BSA as a standard.                        points was represented by a straight line, suggesting the existence of
                                                                        a single class of non-interacting binding sites. The relevant binding
                               RESULTS                                  parameters were, for this experiment: 'T-vWF bound at saturation
                                                                        = 2.09 m g / d of heparin-Sepharose gel; Kd= 3.78 X lo" M (calculated
    Time Course, Temperature Dependence, and Saturability of per vWF subunit assuming a molecular mass of 275 m a ) ; nonsatu-
 the Binding-The binding of 'T-vWF to heparin-Sepharose rable (nonspecific) binding = 0. B/F, bound/free.
1736                                  von Willebrand Factor Heparin-binding Domain
mg/mlfree heparin, the binding of lZ5I-vWFto heparin-
Sepharose CL-GB wasno greater than to the heparin-free
Sepharose (Fig. 3B).
   At physiological NaCl concentrations, '251-mouseIgG or
'251-humanfibrinogen bound to heparin-Sepharose CL-GB 5-
7 times more than that to Sepharose CL-GB alone. However,
the bound radioactivity was almost unchanged at concentra-
tions of 0.1-10 mg/ml free heparin (data not shown).
   The displacement of bound lZ5I-vWF pg/ml) from hepa-
rin-Sepharose was determined in 0.1 M NaC1. More than 90%
of bound radioactivity was displaced in the presence of an
                                                                                           0       200     400       600    BOO
800-fold excessof unlabeled ligand (Fig. 4A).
   Iz5I-vWF binding to heparin-Sepharose occurred in the                           L           Fold Excess of unlabeled vWF
presence of 1 m~ EDTA to thesame extent as in the absence
of EDTA, but it was enhanced approximately 40% in the
presence of 1-10 m CaClZ(data not shown).
   Location of a Heparin-binding Domain of Human uWF on
the 52/48-kDa Fragment Extending from Val-449 to Lys-728-
We have previously reported that the 52/48-kDa fragment
begins with residue Val-449. We now have determined the
NHz-terminal sequences of the reduced and S-carboxymeth-
ylated 41, 22, 13, and 55-kDa tryptic fragments of vWF and
demonstrated that they were derived from unrelated parts of
the subunit molecule (8). The NHz-terminal sequence is
GlnZW-Thr-Met-Val-Asp-Ser-Ser the 13-kDa fragment,
                                            for                                              0.1        1          10       100
Asn730-Ser-Met-Va1-Leu-A~p-Val-Ala-Phe-Val-Leu-Glu                for
                                                                                           Concentration of Competing Ligands
the 55-kDa fragment, Thr'352-Val-Gln-Arg-Pro-Gly-Gln-Thr-                                                 Id"
Cys-Gln-Pro-Ile-Leu-Glu-Glu-Gln-Cys-Leu-Val 41-          for the
                                                                        FIG. 4. Inhibition of labeled v W F binding by excess unla-
kDa fragment, and Val'927-Thr-Gly-Cys-Pro-Pro the 22- beled v W F or the 52/48-kDa fragment. Panel A , '=I-vWF (1 pg/
kDa fragment. Based on a 275-kDa molecular mass for the ml) was incubated for 30 min at room temperature with 5% (v/v)
vWF subunit, these fragments represent approximately 65% heparin-Sepharose CL-GB in the presence of excess unlabeled vWF.
of the entire subunit molecule.                                       Panel E , 'T-vWF (5 pg/ml) was incubated with 5% (v/v) heparin-
   In competitive binding assays, the 52/48-kDa fragment Sepharose gel in the presence of various concentrations of reduced
completely inhibited lZ5I-vWF                                                                         and
                                       binding to heparin-Sepharose and alkylated 52/48,41,22,13, 55-kDa tryptic fragments vW!?.   of
                                                                      Bound radioactivity is expressed as the percent of '=I-vWF binding
at a final concentration of 10 pM, with an IC50 of 2.8 pM. The in thepresence of buffer only.
other fragments of vWF showed no significant inhibition at
concentrations up to 50 p~ (Fig. 4B). Furthermore, the lZ5I-
                                                                      6B. This binding did not occur at concentrations of NaCl
52/48-kDa fragment bound directly to heparin-Sepharose CL- greater than 0.6
                                                                                        M, and was blocked by free heparin (Fig. 5,
                                                                      A and B). It should be noted that the maximum binding of
                                                                      Iz5I-52/48-kDafragment occurred from 0.2 to 0.3 M NaCI,
                                                                      whereas inhibition of binding required a higher concentration
                                                                      of NaCl (>0.6 M) than required for intact lZ5I-vWF.
                                                                        Additional evidence that the52/48-kDa fragment interacts
                                                                      with free heparin was shown by crossed immunoelectropho-

                                                                      resis experiments. In thepresence of heparin, the 52/48-kDa
                                                                      arc shiftedto a more anodal position and became asymmetric
                0     O      0.01           0.1        1
                                                                      (Fig. 6).
                                                                        COOH-terminal Sequence of the 52/48-kDa Fragment-
                                                                      Amino acid analysis of the five major cyanogen bromide
                                                                      fragments of the 52/48-kDa fragment showed that one peptide
                                                                      lacked homoserine, indicating that it originated from the
                                                                      COOH-terminal end of the fragment. The composition of that
                                                                      peptide was 0.8 Glu, 1.0 Ser, 4.2 Gly, 1.7 Thr, 1.0 Ala,2.2 Pro,
                                                                      3.0 Val, 3.1Leu, and 0.9 Lys,corresponding to theamino acid
                                                                      sequence of residues 711-728 (Ala-Gln-Val-Thr-Val-Gly-Pro-
                                                                      Gly-Leu-Leu-Gly-Val-Ser-Thr-Leu-Gly-Pro-Lys) in-          of the
                                                                      tact vWF subunit. The other four fragments corresponded in
                        1- '
                      op u.001 0.01 0.1 1 10
                                                                      the composition to amino acid residues 449-540,542-622,
                                                                      623-630, and 631-710, respectively (8).These results unam-
                                                                      biguously indicated that the52/48-kDa fragment begins with
                                  Heparin (rnglrnli                   Val-449 and ends at Lys-728.
   FIG. 3. Effects of NaCl and heparin on the binding of v W F          Effect of Monocbnul Antibodies on u WF Binding to Hepa-
to heparin-SepharoseCL-BB. IBI-vWF (5 pg/ml, 3.41 X lo" Ci/ rin-Sephurose-Evidence that the                   heparin- and GPIb-binding
mg) was incubated with 5% (v/v) heparin-Sepharose (black circles)
or Sepharose CL-GB alone (open circles) a t various concentrations of domains of vWF arelocated close to one another on the52/
NaCl ( A )or heparin ( B )for 30 min a t room temperature. The amount 48-kDa fragment was provided by a panel of monoclonal
of T - v W F bound to the gel wasthen determined.                     antibodies to the fragment (Fig. 7). Antibodies that were most
                                            von Willebrand Factor Heparin-binding                                                    1737

                          0.001    0.01   0.1       1   10
                                  Heparin (mglml)
  FIG. 5. Demonstration of direct binding of the 52/48-kDa/
tryptic fragment of v W F to heparin-SepharoseCL-GB. Panel
A , lz5I-52/48-kDa fragment (5 pg/ml; 9.09 X      Ci/mg) was incu-
bated with 5% (v/v) of heparin-Sepharose CL-GB (black circles) or
Sepharose CL-GB alone (open circles) at various concentrations of
NaCl for 30 min at room temperature, and then the amount of lZ5I-
52/48-kDa fragment bound to the gels was determined. Panel B, the
experimental conditions were similar to those described for Panel A
except that the NaCl concentration was kept constant at 0.1 M and
various concentrations of heparin were added to the incubation

effective in blocking binding of vWF to heparin were also
those which were most effective in blocking binding to GPIb,
whereas those which had noeffect on heparin bindinghad no
effect on GPIb binding. Thus, antibodies 52K-2 and 52K-8
totally inhibited ristocetin-induced platelet aggregation and
ristocetin-induced vWF binding to platelets    (Fig. 7 A )and also
completely abolished '251-vWFbinding to heparin-Sepharose
(Fig. 7B). Two other antibodies, K-21 and K-22, which par-
tially inhibited ristocetin-induced '251-vWFbinding to plate-
lets, incompletely inhibited the binding to heparin-Sepharose
even at the highest IgG concentrations tested. An additional
antibody, K-19, partially inhibited heparin binding, but did
not inhibit ristocetin-induced  binding. The corresponding ICso
values of these latter three antibodies were greater than 4             FIG. 6. Crossed immunoelectrophoresis of the purified 52/
                                                                      48-kDa fragmentusing a specific rabbit antiserum. Five pg of
mg/ml (Fig. 7B). The remaining seven antibodies failed to             the purified fragmentwas electrophoresed in each case. The first
inhibit the ristocetin-induced '251-vWFbinding to platelets           dimensions contained either no heparin ( I ) or heparin at concentra-
and did not inhibit 129-vWF    binding to heparin-Sepharose at        tions of 0.2 mg/ml(2), 0.5 mg/ml(3), or 5 mg/m ( 4 ) .Rabbit antiserum
any IgG concentration tested. The datafor one of these latter         specific for the 52/48-kDa fragment was used in the second dimen-
antibodies (K-29) areshown in Fig. 7, A and B.                        sion. Electrophoresis was performed as described under "Materials
                                                                      and Methods."
   In addition, heparin inhibited ristocetin-induced      platelet
aggregation and ristocetin-induced '251-vWFbinding to plate-
                                                               binding properties of heparin were not the result of coupling
lets. However, direct binding of the '251-52/48-kDafragment
                                                               to Sepharose was provided by the observation that free hep-
to platelet GPIb in the  absence of ristocetin was not inhibited
                                                               arin totally inhibited the binding.
by heparin alone (data notshown).                                Competitive binding assays demonstrated that a heparin-
                          DISCUSSION                           binding domain of vWF exists within the same reduced and
                                                               alkylated 52/48-kDa tryptic fragment as does the platelet
   We have developed a simple assay for measuring vWF GPIb-binding domain (9). This finding is compatible with
binding to heparin-Sepharose based ona commonly used that of Bockenstedt et al. (7) who showed that both the
platelet-ligand binding assay system (3, 9, 22). By the inclu- platelet GPIb and the heparin-binding domains are located
sion of 0.1% (w/v) BSA in the  incubation mixture,nonspecific on the same NHz-terminal 285-kDa fragment derived from S.
binding was kept to less than 5%of total binding. Binding of aureus protease V8 digestion of vWF. Interaction of free
l2'1-vWF to heparin-Sepharose occurred in a saturable man- heparin with the 52/48-kDa fragment was shown by crossed
ner at physiological ionic strength and neutral pH. The  bind- immunoelectrophoresis experiments in which the 52/48-kDa
ing occurred independently of incubation temperature and in arc shifted anodally and became asymmetric in the presence
the presence or absence of calcium ions. Evidence that the of heparin.
1738                                                                  Domain
                                          uon Willebrand Factor Heparin-binding
       120                                                                  It has been reported by others that heparin impairs risto-
              A                                                          cetin-induced platelet aggregation (24). We have now shown
       100                                                               that heparin inhibits ristocetin-induced '251-vWF binding to
                                                                         platelets. Since heparin does not inhibit the binding of the
                                                                         52/48-kDa fragment to GPIb in the absence of ristocetin, it
        80                                                               is possible that the interference of heparin with ristocetin
                                                                         induced vWF interactions is the result negatively charged
                                                                         heparin interactingwith positively charged ristocetin.
                                                                            Several well-characterized plasma proteins contain more
                                                                         than one heparin-binding domain which remain intact after
        40                                                               limited enzymaticcleavage (25-28). However, fibronectin and
                                                                         thrombospondin lose their heparin-binding properties after
                                                                         reduction and alkylation (25, 27). After reduction and alkyl-
                                                                         ation in 7 M guanidine HCl and purification in6 M urea (9),
                                                                         the 52/48-kDa fragment retains the ability bind directly to
                                                                         heparin-Sepharose,to complex withfree heparin,andto
                                                                         competitively inhibit binding of intact lZ5I-vWFto heparin.
                                                                         It is thereforepossible that an amino  acid sequence contained
                                                                         withinthe 52/48-kDa fragment definesa heparin-binding
                                                                            Although the physiological role of the vWF heparin-binding
                                                                         domain is uncertainat present, it  may prove to be instrumen-
                                                                         tal in the bindingvWF toexposed subendothelium. Wagner
                                                                         et al. (29) have shown that vWF binds to endothelial cell-
                                                                         produced extracellular matrices even after collagen has been
                                                                         enzymatically digested away. In addition, Fauvel et al. (30)
                                                                         demonstrated that vWF binds ato      microfibrillar extract from
                                                                         adult bovine aorta. It ispossible that heparin-like glycosami-
                                                                         noglycans are responsible for both of these interactions and
                                                                         that they contribute to binding of vWF to the denuded sub-

                                                                           Acknowledgments-We thank Armour Pharmaceutical         (Rorer
                             I                   I         I   K-21      Group, Fort Washington, PA) for their generous gift of commercial
          0   0.01          0.1                  1        10             factor VIII/vWF concentrate, Santosh Kumar for her excellenttech-
                                                                         nical assistance, and DonnaDiorio and Claire Jackson for expert
                                  IgG (mglrnl)                           secretarial assistance.
   FIG. 7. Panel A , inhibition of vWF binding to heparin-Sepharose
CL-GBbymonoclonal antibodies against the 52/48-kDa fragment.                                     REFERENCES
Panel B, inhibition of ristocetin-induced vWF binding platelets by 1. Ruggeri, Z. M., and Zimmerman, T. S. (1985) Semin. Hemutol.
the same antibodies. '251-~WF pg/ml) was preincubated with serial
                                (5                                      22,203-218
dilutions of IgGs of six monoclonal  antibodies againstthe 52/48-kDa
                                                                   2. Coller. B. S.. Peerschke, E. I.. Scudder, L. E., and Sullivan,C. A.
fragment (or buffer ina control mixture)for 30 min at 37 "C. Either     (1983) Blood 61,99-110
heparin-Sepharose CL-GB (5% v/v) or platelets (lOs/ml) plus risto- 3. Ruszszeri.Z. M.. Bader. R.. and DeMarco. , L. (1982) Proc. Natl.
                                                                                                                        .   .
cetin (1 mg/ml) was then added. After an additional 30-min incuba-     ,&-ad.'Sci. U.'S. A . 7 9 , 6038-6041
tion at room temperature, the radioactivity bound to the gel or to the
                                                                   4. Nyman, D. (1977) Thromb. Res. 11,433-438
platelets wasmeasured and expressed as the percent of lZ5I-vWF     5. Madalas, F., Bell, W. R., and Castaldi, P. A. (1978) Huemostasis
binding in the control mixture.                                         7,321-331
                                                                   6. Fowler,W.E., Fretto, L. J., Hamilton, K. K., Erickson, H. P.,
   Roberts et al. (23) have recently reported that vWF    specif-       and McKee, P. A. (1985) J. Clin. Znuest. 7 6 , 1491-1500
                                                                   7. Bockenstedt, P., Greenberg, J. M., and Handin, R. I. (1986) J.
ically binds to sulfated glycolipids. They have also  shown that
                                                                        Clin. Inuest. 77,743-749
the bindingof lZ5I-vWF to sulfated     glycolipids was prevented   8. Titani, K., Kumar, S., Takio, K., Ericsson, L. H., Wade, R. D.,
by high molecular weight dextran sulfate, but notby heparin.            Ashida, K., Walsh, K. A., Chopek, M. W., Sadler, J. E., and
Both dextran sulfate and heparin are sulfated    glycoconjugates,       Fujikawa, K. (1986) Biochemistry 25, 3171-3184
but differ insugar composition and position of sulfation.          9. Fujimura, Y., Titani, K., Holland, L. Z., Russell, S. R., Roberts,
                                                                        J. R., Elder, J. H., Ruggeri,Z. M., and Zimmerman, S. (1986)
Thus, these two binding sites on the vWF molecule for sul-              J. Biol. Chem. 2 6 1 , 381-385
fated glycolipids and heparin seem to be different.               10. De Marco, L., and Shapiro, S. S. (1981) J. Clin. Znuest. 6 8 , 321-
   Monoclonal antibodiesagainstthe         52/48-kDa fragment           328
which inhibit ristocetin-inducedlZ5I-vWFbinding to platelets 11. Fulcher, C. A., and Zimmerman, T. S. (1982) Proc. Natl. Acad.
also abolished '251-vWFbinding to heparin-Sepharosea t sim-             Sci. U. S. A. 7 9 , 1648-1652
                                                                  12. Ey, P. L., Prowse, S. J., and Jenkin, C. R. (1978) Imrnunochern-
ilar IgG concentrations. With one    exception, antibodies which        istry 15,429-436
failed to inhibit ristocetin-inducedvWF binding alsofailed to 13. Towbin, H., Staehelin, and Gordon,J. (1979) Proc. Natl. Acad.
inhibit vWF binding to heparin. This suggests that the two              Sci. U. S. A . 76,4350-4354
domains lie in close proximity to one another on the native       14. Fraker, P. J., and Speck, J. C., Jr. (1978) Biochem. Biophys. Res.
vWF molecule. However, direct binding of the 52/48-kDa                  Commun. 80,849-857
                                                                  15. Laurell, C.-B. (1965) Anal. Biochem. 10, 358-361
fragment to platelet GPIb the absenceof ristocetin was not 16. Ericsson, L. H., Wade, R. D.,Gagnon, J., McDonald, R. M.,
inhibited by heparin. Thus, it is unlikely that the binding
            .     .                                                     Granbera, R. R.. andWalsh, K. A. (1977) in Solid Phase
domains are precisely congruent.                                        Methods-inProteinSequenceAnalysis           (Previero, A. I., and
                               Heparin-binding                                                                                             1739
        Coletti-Previero, M. A., eds) p. 137, Elsevier/North Holland,     24. Pekcelen, Y., and Inceman, S. (1976) Thromb. Haemostasis 35,
        Amsterdam                                                               485-487
17.   Titani, K., Sasagawa, T., Ericsson, L. H., Kumar, S., Smith, S.     25. Hynes, R. O., and Yamada;K. M. (1982) J. Cell Biol. 95, 369-
        B., Krebs, E. G., and Walsh, K.A. (1984) Biochemistry 23,               377
        4193-4199                                                         26. Weisgraber, K.H., Rall, S. C., Mahley, R. W., Milne, R. W.,
18.   Mahoney, W. C., and Hermodson, M. A. (1980) J. Biol. Chem.                Marcel, Y. L., and Sparrow, J. T. (1986) J. Biol. Chem. 261,
                                                                          27. Haverstick, D. M., Dixit, V. M., Grant, G. A., Frazier, W. A., and
19.   Bidlingmeyer, B.A., Cohen, S. A., and Tarvin, T. L. (1984) J.             Santoro, S. A. (1984) Biochemistry 23, 5597-5603
        Chromutogr. 336,93-104                                            28. Koide, T., Foster, D., and Odani, S. (1986) FEBS Lett. 194,242-
20.   Bradford, M. M. (1976) Anal. Biochem. 7 2 , 248-254                       244
21.   Munson, P. J. (1983) Methods Enzymology 92, 542-576                 29. Wagner, D. D., Urban-Pickering, M., and Marder, V. J. (1984)
22.   Marguerie, G. A., Edgington, T. S., and Plow, E. F. (1980)J.Biol.         Proc. Natl. Acad. Sci. U. S. A. 8 1 , 471-475
        Chem. 2 5 5 , 154-161                                             30. Fauvel, F.,Grant, M. E., Lenand, Y. J., Souchon. H.. Tobelem.
23.   Roberts, D. D., Williams, S. B., Gralnick, H. R., and Ginsburg,           G., Jackson, D..S., and Cain,J. P . (1983) Proc. fiatl.'Acad. Sci:
        V. (1986) J. BWL. Chem. 261,3306-3309                                   U. S. A. 80, 551-554